Sympy Docs PDF 0.7.6
Uploaded by
JaviOrellanaBarrosoSympy Docs PDF 0.7.6
Uploaded by
JaviOrellanaBarrosoSymPy Documentation
Release 0.7.6
SymPy Development Team
November 20, 2014
CONTENTS
1 Installation
1.1 Source . . . . .
1.2 Git . . . . . . . .
1.3 Anaconda . . .
1.4 Other Methods
1.5 Run SymPy . .
1.6 Questions . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
1
1
1
2
2
2
3
2 SymPy Tutorial
2.1 Preliminaries . . . . . . . . . . . . . .
2.2 Introduction . . . . . . . . . . . . . . .
2.3 Gotchas . . . . . . . . . . . . . . . . .
2.4 Basic Operations . . . . . . . . . . . .
2.5 Printing . . . . . . . . . . . . . . . . .
2.6 Simplication . . . . . . . . . . . . . .
2.7 Calculus . . . . . . . . . . . . . . . . .
2.8 Solvers . . . . . . . . . . . . . . . . . .
2.9 Matrices . . . . . . . . . . . . . . . . .
2.10 Advanced Expression Manipulation .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
5
5
7
10
15
18
23
37
43
46
52
3 Gotchas and Pitfalls
3.1 Introduction . . . . . . . . . . . .
3.2 Equals Signs (=) . . . . . . . . .
3.3 Variables . . . . . . . . . . . . . .
3.4 Symbolic Expressions . . . . . .
3.5 Special Symbols . . . . . . . . .
3.6 Getting help from within SymPy
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
61
61
61
62
65
70
73
4 SymPy Users Guide
4.1 Introduction . . . . . .
4.2 Learning SymPy . . .
4.3 SymPys Architecture
4.4 Contributing . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
75
75
75
77
84
5 SymPy Modules Reference
5.1 SymPy Core . . . . . . . . . .
5.2 Combinatorics Module . . .
5.3 Number Theory . . . . . . . .
5.4 Basic Cryptography Module
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
85
85
187
276
303
.
.
.
.
.
.
.
.
.
.
.
.
5.5
5.6
5.7
5.8
5.9
5.10
5.11
5.12
5.13
5.14
5.15
5.16
5.17
5.18
5.19
5.20
5.21
5.22
5.23
5.24
5.25
5.26
5.27
5.28
5.29
5.30
5.31
5.32
5.33
5.34
5.35
5.36
5.37
5.38
5.39
Concrete Mathematics . . . . . . . . . . . . . . . . .
Numerical evaluation . . . . . . . . . . . . . . . . . .
Numeric Computation . . . . . . . . . . . . . . . . . .
Functions Module . . . . . . . . . . . . . . . . . . . .
Geometric Algebra Module . . . . . . . . . . . . . . .
Geometry Module . . . . . . . . . . . . . . . . . . . .
Symbolic Integrals . . . . . . . . . . . . . . . . . . . .
Numeric Integrals . . . . . . . . . . . . . . . . . . . .
Logic Module . . . . . . . . . . . . . . . . . . . . . . .
Matrices . . . . . . . . . . . . . . . . . . . . . . . . . .
Welcome to mpmaths documentation! . . . . . . . .
Polynomials Manipulation Module . . . . . . . . . .
Printing System . . . . . . . . . . . . . . . . . . . . . .
Plotting Module . . . . . . . . . . . . . . . . . . . . . .
Pyglet Plotting Module . . . . . . . . . . . . . . . . .
Assumptions module . . . . . . . . . . . . . . . . . . .
Term rewriting . . . . . . . . . . . . . . . . . . . . . .
Series Expansions . . . . . . . . . . . . . . . . . . . .
Sets . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Simplify . . . . . . . . . . . . . . . . . . . . . . . . . .
Details on the Hypergeometric Function Expansion
Stats . . . . . . . . . . . . . . . . . . . . . . . . . . . .
ODE . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
PDE . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Solvers . . . . . . . . . . . . . . . . . . . . . . . . . . .
Diophantine . . . . . . . . . . . . . . . . . . . . . . . .
Inequality Solvers . . . . . . . . . . . . . . . . . . . .
Tensor Module . . . . . . . . . . . . . . . . . . . . . .
Utilities . . . . . . . . . . . . . . . . . . . . . . . . . . .
Parsing input . . . . . . . . . . . . . . . . . . . . . . .
Calculus . . . . . . . . . . . . . . . . . . . . . . . . . .
Physics Module . . . . . . . . . . . . . . . . . . . . . .
Category Theory Module . . . . . . . . . . . . . . . .
Dierential Geometry Module . . . . . . . . . . . . .
Contributions to docs . . . . . . . . . . . . . . . . . .
6 Development Tips: Comparisons in
6.1 Introduction . . . . . . . . . . . . .
6.2 Hashing . . . . . . . . . . . . . . .
6.3 Method Resolution . . . . . . . . .
6.4 General Notes and Caveats . . . .
6.5 Script To Verify This Guide . . . .
Python
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
Module .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
1861
. 1861
. 1861
. 1862
. 1864
. 1864
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
319
331
337
340
450
489
583
619
625
635
713
1046
1248
1270
1283
1287
1302
1304
1312
1327
1350
1361
1397
1460
1468
1483
1505
1508
1532
1588
1592
1597
1829
1846
1860
7 Wiki
1869
7.1 FAQ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1869
8 SymPy Papers
1871
9 Planet SymPy
1873
10 SymPy logos
1875
11 Projects using SymPy
1877
12 Blogs, News, Magazines
1879
ii
13 About
1881
13.1 SymPy Development Team . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1881
13.2 Financial and Infrastructure Support . . . . . . . . . . . . . . . . . . . . . . . . . . 1891
13.3 License . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1892
14 SymPy Special Topics
1893
14.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1893
14.2 Finite Dierence Approximations to Derivatives . . . . . . . . . . . . . . . . . . . . 1893
Bibliography
1899
Python Module Index
1919
Index
1923
iii
iv
CHAPTER
ONE
INSTALLATION
The SymPy CAS can be installed on virtually any computer with Python 2.6 or above. SymPy
does not require any special Python modules: let us know if you have any problems with
SymPy on a standard Python install. The current recommended method of installation is
directly from the source les. Alternatively, executables are available for Windows, and some
Linux distributions have SymPy packages available.
SymPy ocially supports Python 2.6, 2.7, 3.2, 3.3, 3.4, and PyPy.
1.1 Source
SymPy currently recommends that users install directly from the source les. You will rst
have to download the source les via the archive. Download the latest release (tar.gz) from
the downloads site and open it with your operating systems standard decompression utility.
After the download is complete, you should have a folder called sympy. From your favorite
command line terminal, change directory into that folder and execute the following:
$ python setup.py install
Alternatively, if you dont want to install the package onto your computer, you may run SymPy
with the isympy console (which automatically imports SymPy packages and denes common
symbols) by executing within the sympy folder:
$ ./bin/isympy
You may now run SymPy statements directly within the Python shell:
>>>
>>>
>>>
>>>
>>>
>>>
x
from __future__ import division
from sympy import *
x, y, z, t = symbols(x y z t)
k, m, n = symbols(k m n, integer=True)
f, g, h = symbols(f g h, cls=Function)
diff(x**2/2, x)
1.2 Git
If you are a developer or like to get the latest updates as they come, be sure to install from
git. To download the repository, execute the following from the command line:
SymPy Documentation, Release 0.7.6
$ git clone git://github.com/sympy/sympy.git
Then, execute either the setup.py or the bin/isympy scripts as demonstrated above.
To update to the latest version, go into your repository and execute:
$ git pull origin master
If you want to install SymPy, but still want to use the git version, you can run from your
repository:
$ setupegg.py develop
This will cause the installed version to always point to the version in the git directory.
1.3 Anaconda
Although SymPy does not have any hard dependencies, many nice features are only enabled
when certain libraries are installed. For example, without Matplotlib, only simple text-based
plotting is enabled. With the IPython notebook or qtconsole, you can get nicer LATEX printing
by running init printing(). An easy way to get all these libraries in addition to SymPy is to
install Anaconda, which is a free Python distribution from Continuum Analytics that includes
SymPy, Matplotlib, IPython, NumPy, and many more useful packages for scientic computing.
1.4 Other Methods
An installation executable (.exe) is available for Windows users at the downloads site. In
addition, various Linux distributions have SymPy available as a package. Others are strongly
encouraged to download from source (details above).
1.5 Run SymPy
After installation, it is best to verify that your freshly-installed SymPy works. To do this, start
up Python and import the SymPy libraries:
$ python
>>> from sympy import *
From here, execute some simple SymPy statements like the ones below:
>>> x = Symbol(x)
>>> limit(sin(x)/x, x, 0)
1
>>> integrate(1/x, x)
log(x)
For a starter guide on using SymPy eectively, refer to the SymPy Tutorial (page 5).
Chapter 1. Installation
SymPy Documentation, Release 0.7.6
1.6 Questions
If you have a question about installation or SymPy in general, feel free to visit our chat on
Gitter. In addition, our mailing list is an excellent source of community support.
If you think theres a bug or you would like to request a feature, please open an issue ticket.
1.6. Questions
SymPy Documentation, Release 0.7.6
Chapter 1. Installation
CHAPTER
TWO
SYMPY TUTORIAL
2.1 Preliminaries
This tutorial assumes that the reader already knows the basics of the Python programming
language. If you do not, the ocial Python tutorial is excellent.
This tutorial assumes a decent mathematical background. Most examples require knowledge
lower than a calculus level, and some require knowledge at a calculus level. Some of the
advanced features require more than this. If you come across a section that uses some mathematical function you are not familiar with, you can probably skip over it, or replace it with a
similar one that you are more familiar with. Or look up the function on Wikipedia and learn
something new. Some important mathematical concepts that are not common knowledge will
be introduced as necessary.
2.1.1 Installation
Quick Tip
You do not need to install SymPy to try it.
You can use the online shell at
http://live.sympy.org, or the shell at the bottom right of this documentation page.
You will need to install SymPy rst. See the installation guide (page 1).
Alternately, you can just use the SymPy Live Sphinx extension to run the code blocks in the
browser. For example, click on the green Run code block in SymPy Live button below
>>> from sympy import *
>>> x = symbols(x)
>>> a = Integral(cos(x)*exp(x), x)
>>> Eq(a, a.doit())
Integral(exp(x)*cos(x), x) == exp(x)*sin(x)/2 + exp(x)*cos(x)/2
The SymPy Live shell in the bottom corner will pop up and evaluate the code block. You can
also click any individual line to evaluate it one at a time.
The SymPy Live shell is a fully interactive Python shell. You can type any expression in the
input box to evaluate it. Feel free to use it throughout the tutorial to experiment.
To show or hide the SymPy Live shell at any time, just click the green button on the bottom
right of the screen.
SymPy Documentation, Release 0.7.6
By default, the SymPy Live shell uses LATEX for output. If you want the output to look more like
the output in the documentation, change the output format to Str or Unicode in the settings.
If you wish to modify an example before evaluating it, change the evaluation mode to copy
in the SymPy Live settings. This will cause clicking on an example to copy the example to
the SymPy Live shell, but not evaluate it, allowing you to change it before execution. You can
also use the up/down arrow keys on your keyboard in the input box to move through the shell
history.
The SymPy Live shell is also available at http://live.sympy.org, with extra features, like a
mobile phone enhanced version and saved history.
2.1.2 Exercises
This tutorial was the basis for a tutorial given at the 2013 SciPy conference in Austin, TX.
The website for that tutorial is here. It has links to videos, materials, and IPython notebook
exercises. The IPython notebook exercises in particular are recommended to anyone going
through this tutorial.
2.1.3 About This Tutorial
This tutorial aims to give an introduction to SymPy for someone who has not used the library
before. Many features of SymPy will be introduced in this tutorial, but they will not be exhaustive. In fact, virtually every functionality shown in this tutorial will have more options
or capabilities than what will be shown. The rest of the SymPy documentation serves as API
documentation, which extensively lists every feature and option of each function.
These are the goals of this tutorial:
To give a guide, suitable for someone who has never used SymPy (but who has used
Python and knows the necessary mathematics).
To be written in a narrative format, which is both easy and fun to follow. It should read
like a book.
To give insightful examples and exercises, to help the reader learn and to make it entertaining to work through.
To introduce concepts in a logical order.
To use good practices and idioms, and avoid antipatterns. Functions or methodologies
that tend to lead to antipatterns are avoided. Features that are only useful to advanced
users are not shown.
To be consistent. If there are multiple ways to do it, only the best way is shown.
To avoid unnecessary duplication, it is assumed that previous sections of the tutorial
have already been read.
Feedback on this tutorial, or on SymPy in general is always welcome. Just write to our mailing
list.
Chapter 2. SymPy Tutorial
SymPy Documentation, Release 0.7.6
2.2 Introduction
2.2.1 What is Symbolic Computation?
Symbolic computation deals with the computation of mathematical objects symbolically. This
means that the mathematical objects are represented exactly, not approximately, and mathematical expressions with unevaluated variables are left in symbolic form.
Lets take an example. Say we wanted to use the built-in Python functions to compute square
roots. We might do something like this
>>> import math
>>> math.sqrt(9)
3.0
9 is a perfect square, so we got the exact answer, 3. But suppose we computed the square
root of a number that isnt a perfect square
>>> math.sqrt(8)
2.82842712475
Here we got an approximate result. 2.82842712475 is not the exact square root of 8 (indeed,
the actual square root of 8 cannot be represented by a nite decimal, since it is an irrational
number). If all we cared about was the decimal form of the square root of 8, we would be
done.
But suppose we want to go further. Recall that 8 = 4 2 = 2 2. We would have a hard
time deducing this from the above result. This is where symbolic computation comes in. With
a symbolic computation system like SymPy, square roots of numbers that are not perfect
squares are left unevaluated by default
>>> import sympy
>>> sympy.sqrt(3)
sqrt(3)
Furthermoreand this is where we start to see the real power of symbolic computationsymbolic results can be symbolically simplied.
>>> sympy.sqrt(8)
2*sqrt(2)
2.2.2 A More Interesting Example
The above example starts to show how we can manipulate irrational numbers exactly using
SymPy. But it is much more powerful than that. Symbolic computation systems (which by
the way, are also often called computer algebra systems, or just CASs) such as SymPy are
capable of computing symbolic expressions with variables.
As we will see later, in SymPy, variables are dened using symbols. Unlike many symbolic
manipulation systems, variables in SymPy must be dened before they are used (the reason
for this will be discussed in the next section (page 10)).
Let us dene a symbolic expression, representing the mathematical expression x + 2y.
>>> from sympy import symbols
>>> x, y = symbols(x y)
>>> expr = x + 2*y
2.2. Introduction
SymPy Documentation, Release 0.7.6
>>> expr
x + 2*y
Note that we wrote x + 2*y just as we would if x and y were ordinary Python variables. But
in this case, instead of evaluating to something, the expression remains as just x + 2*y. Now
let us play around with it:
>>> expr + 1
x + 2*y + 1
>>> expr - x
2*y
Notice something in the above example. When we typed expr - x, we did not get x + 2*y x, but rather just 2*y. The x and the -x automatically canceled one another. This is similar to
how sqrt(8) automatically turned into 2*sqrt(2) above. This isnt always the case in SymPy,
however:
>>> x*expr
x*(x + 2*y)
Here, we might have expected x(x + 2y) to transform into x2 + 2xy, but instead we see that the
expression was left alone.
This
is a common theme in SymPy. Aside from obvious simplications like x x = 0 and 8 = 2 2, most simplications are not performed automatically. This
is because we might prefer the factored form x(x + 2y), or we might prefer the expanded form
x2 + 2xy. Both forms are useful in dierent circumstances. In SymPy, there are functions to
go from one form to the other
>>> from sympy import expand, factor
>>> expanded_expr = expand(x*expr)
>>> expanded_expr
x**2 + 2*x*y
>>> factor(expanded_expr)
x*(x + 2*y)
2.2.3 The Power of Symbolic Computation
The real power of a symbolic computation system such as SymPy is the ability to do all sorts
of computations symbolically. SymPy can simplify expressions, compute derivatives, integrals, and limits, solve equations, work with matrices, and much, much more, and do it all
symbolically. It includes modules for plotting, printing (like 2D pretty printed output of math
formulas, or LATEX), code generation, physics, statistics, combinatorics, number theory, geometry, logic, and more. Here is a small sampling of the sort of symbolic power SymPy is
capable of, to whet your appetite.
>>> from sympy import *
>>> x, t, z, nu = symbols(x t z nu)
This will make all further examples pretty print with unicode characters.
>>> init_printing(use_unicode=True)
Take the derivative of sin (x)ex .
>>> diff(sin(x)*exp(x), x)
x
x
e sin(x) + e cos(x)
Chapter 2. SymPy Tutorial
SymPy Documentation, Release 0.7.6
Compute (ex sin (x) + ex cos (x)) dx.
>>> integrate(exp(x)*sin(x) + exp(x)*cos(x), x)
x
e sin(x)
Compute
sin (x2 ) dx.
>>> integrate(sin(x**2), (x, -oo, oo))
___
___
\ 2 \
----------2
Find limx0
sin (x)
x .
>>> limit(sin(x)/x, x, 0)
1
Solve x2 2 = 0.
>>> solve(x**2 - 2, x)
___
___
-\ 2 , \ 2
Solve the dierential equation y 00 y = et .
>>> y = Function(y)
>>> dsolve(Eq(y(t).diff(t, t) - y(t), exp(t)), y(t))
-t
t t
y(t) = C2e
+ C1 + -e
Find the eigenvalues of [ 12 22 ].
>>> Matrix([[1, 2], [2, 2]]).eigenvals()
____
____
3
\ 17
\ 17
3
- + ------: 1, - ------ + -: 1
2
2
2
2
Rewrite the Bessel function J (z) in terms of the spherical Bessel function j (z).
>>> besselj(nu, z).rewrite(jn)
___
___
\ 2 \ z jn( - 1/2, z)
-------------------------___
\
cos2 (x) dx using LATEX.
>>> latex(Integral(cos(x)**2, (x, 0, pi)))
\int_{0}^{\pi} \cos^{2}{\left (x \right )}\, dx
2.2. Introduction
SymPy Documentation, Release 0.7.6
2.2.4 Why SymPy?
There are many computer algebra systems out there. This Wikipedia article lists many of
them. What makes SymPy a better choice than the alternatives?
First o, SymPy is completely free. It is open source, and licensed under the liberal BSD
license, so you can modify the source code and even sell it if you want to. This contrasts
with popular commercial systems like Maple or Mathematica that cost hundreds of dollars in
licenses.
Second, SymPy uses Python. Most computer algebra systems invent their own language.
Not SymPy. SymPy is written entirely in Python, and is executed entirely in Python. This
means that if you already know Python, it is much easier to get started with SymPy, because
you already know the syntax (and if you dont know Python, it is really easy to learn). We
already know that Python is a well-designed, battle-tested language. The SymPy developers
are condent in their abilities in writing mathematical software, but programming language
design is a completely dierent thing. By reusing an existing language, we are able to focus
on those things that matter: the mathematics.
Another computer algebra system, Sage also uses Python as its language. But Sage is large,
with a download of over a gigabyte. An advantage of SymPy is that it is lightweight. In
addition to being relatively small, it has no dependencies other than Python, so it can be used
almost anywhere easily. Furthermore, the goals of Sage and the goals of SymPy are dierent.
Sage aims to be a full featured system for mathematics, and aims to do so by compiling all the
major open source mathematical systems together into one. When you call some function in
Sage, such as integrate, it calls out to one of the open source packages that it includes. In
fact, SymPy is included in Sage. SymPy on the other hand aims to be an independent system,
with all the features implemented in SymPy itself.
A nal important feature of SymPy is that it can be used as a library. Many computer algebra
systems focus on being usable in interactive environments, but if you wish to automate or
extend them, it is dicult to do. With SymPy, you can just as easily use it in an interactive
Python environment or import it in your own Python application. SymPy also provides APIs
to make it easy to extend it with your own custom functions.
2.3 Gotchas
To begin, we should make something about SymPy clear. SymPy is nothing more than a Python
library, like NumPy, Django, or even modules in the Python standard library sys or re. What
this means is that SymPy does not add anything to the Python language. Limitations that are
inherent in the Python language are also inherent in SymPy. It also means that SymPy tries
to use Python idioms whenever possible, making programming with SymPy easy for those
already familiar with programming with Python. As a simple example, SymPy uses Python
syntax to build expressions. Implicit multiplication (like 3x or 3 x) is not allowed in Python,
and thus not allowed in SymPy. To multiply 3 and x, you must type 3*x with the *.
2.3.1 Symbols
One consequence of this fact is that SymPy can be used in any environment where Python is
available. We just import it, like we would any other library:
>>> from sympy import *
10
Chapter 2. SymPy Tutorial
SymPy Documentation, Release 0.7.6
This imports all the functions and classes from SymPy into our interactive Python session.
Now, suppose we start to do a computation.
>>> x + 1
Traceback (most
...
NameError: name
Traceback (most
...
NameError: name
recent call last):
x is not defined
recent call last):
x is not defined
Oops! What happened here? We tried to use the variable x, but it tells us that x is not dened.
In Python, variables have no meaning until they are dened. SymPy is no dierent. Unlike
many symbolic manipulation systems you may have used, in SymPy, variables are not dened
automatically. To dene variables, we must use symbols.
>>> x = symbols(x)
>>> x + 1
x + 1
symbols takes a string of variable names separated by spaces or commas, and creates Symbols out of them. We can then assign these to variable names. Later, we will investigate some
convenient ways we can work around this issue. For now, let us just dene the most common
variable names, x, y, and z, for use through the rest of this section
>>> x, y, z = symbols(x y z)
As a nal note, we note that the name of a Symbol and the name of the variable it is assigned
to need not have anything to do with one another.
>>> a, b = symbols(b a)
>>> a
b
>>> b
a
Here we have done the very confusing thing of assigning a Symbol with the name a to the
variable b, and a Symbol of the name b to the variable a. Now the Python variable named a
points to the SymPy Symbol named b, and visa versa. How confusing. We could have also
done something like
>>> crazy = symbols(unrelated)
>>> crazy + 1
unrelated + 1
This also shows that Symbols can have names longer than one character if we want.
Usually, the best practice is to assign Symbols to Python variables of the same name, although
there are exceptions: Symbol names can contain characters that are not allowed in Python
variable names, or may just want to avoid typing long names by assigning Symbols with long
names to single letter Python variables.
To avoid confusion, throughout this tutorial, Symbol names and Python variable names will
always coincide. Furthermore, the word Symbol will refer to a SymPy Symbol and the word
variable will refer to a Python variable.
Finally, let us be sure we understand the dierence between SymPy Symbols and Python
variables. Consider the following:
2.3. Gotchas
11
SymPy Documentation, Release 0.7.6
x = symbols(x)
expr = x + 1
x = 2
print(expr)
What do you think the output of this code will be? If you thought 3, youre wrong. Lets see
what really happens
>>>
>>>
>>>
>>>
x +
x = symbols(x)
expr = x + 1
x = 2
print(expr)
1
Changing x to 2 had no eect on expr. This is because x = 2 changes the Python variable
x to 2, but has no eect on the SymPy Symbol x, which was what we used in creating expr.
When we created expr, the Python variable x was a Symbol. After we created, it, we changed
the Python variable x to 2. But expr remains the same. This behavior is not unique to SymPy.
All Python programs work this way: if a variable is changed, expressions that were already
created with that variable do not change automatically. For example
>>> x = abc
>>> expr = x + def
>>> expr
abcdef
>>> x = ABC
>>> expr
abcdef
Quick Tip
To change the value of a Symbol in an expression, use subs
>>> x = symbols(x)
>>> expr = x + 1
>>> expr.subs(x, 2)
3
In this example, if we want to know what expr is with the new value of x, we need to reevaluate
the code that created expr, namely, expr = x + 1. This can be complicated if several lines
created expr. One advantage of using a symbolic computation system like SymPy is that we
can build a symbolic representation for expr, and then substitute x with values. The correct
way to do this in SymPy is to use subs, which will be discussed in more detail later.
>>> x = symbols(x)
>>> expr = x + 1
>>> expr.subs(x, 2)
3
2.3.2 Equals signs
Another very important consequence of the fact that SymPy does not extend Python syntax is
that = does not represent equality in SymPy. Rather it is Python variable assignment. This is
hard-coded into the Python language, and SymPy makes no attempts to change that.
12
Chapter 2. SymPy Tutorial
SymPy Documentation, Release 0.7.6
You may think, however, that ==, which is used for equality testing in Python, is used for
SymPy as equality. This is not quite correct either. Let us see what happens when we use ==.
>>> x + 1 == 4
False
Instead of treating x + 1 == 4 symbolically, we just got False. In SymPy, == represents exact
structural equality testing. This means that a == b means that we are asking if a = b. We
always get a bool as the result of ==. There is a separate object, called Eq, which can be used
to create symbolic equalities
>>> Eq(x + 1, 4)
x + 1 == 4
There is one additional caveat about == as well. Suppose we want to know if (x+1)2 = x2 +2x+1.
We might try something like this:
>>> (x + 1)**2 == x**2 + 2*x + 1
False
We got False again. However, (x + 1)2 does equal x2 + 2x + 1. What is going on here? Did we
nd a bug in SymPy, or is it just not powerful enough to recognize this basic algebraic fact?
Recall from above that == represents exact structural equality testing. Exact here means
that two expressions will compare equal with == only if they are exactly equal structurally.
Here, (x + 1)2 and x2 + 2x + 1 are not the same symbolically. One is the power of an addition of
two terms, and the other is the addition of three terms.
It turns out that when using SymPy as a library, having == test for exact symbolic equality is
far more useful than having it represent symbolic equality, or having it test for mathematical
equality. However, as a new user, you will probably care more about the latter two. We have
already seen an alternative to representing equalities symbolically, Eq. To test if two things
are equal, it is best to recall the basic fact that if a = b, then a b = 0. Thus, the best way to
check if a = b is to take a b and simplify it, and see if it goes to 0. We will learn later (page 23)
that the function to do this is called simplify. This method is not infalliblein fact, it can
be theoretically proven that it is impossible to determine if two symbolic expressions are
identically equal in generalbut for most common expressions, it works quite well.
>>>
>>>
>>>
0
>>>
>>>
4*x
a = (x + 1)**2
b = x**2 + 2*x + 1
simplify(a - b)
c = x**2 - 2*x + 1
simplify(a - c)
There is also a method called equals that tests if two expressions are equal by evaluating
them numerically at random points.
>>> a = cos(x)**2 - sin(x)**2
>>> b = cos(2*x)
>>> a.equals(b)
True
2.3.3 Two Final Notes: and /
You may have noticed that we have been using ** for exponentiation instead of the standard .
Thats because SymPy follows Pythons conventions. In Python, represents logical exclusive
2.3. Gotchas
13
SymPy Documentation, Release 0.7.6
or. SymPy follows this convention:
>>> True ^ False
True
>>> True ^ True
False
>>> x^y
Xor(x, y)
Finally, a small technical discussion on how SymPy works is in order. When you type something like x + 1, the SymPy Symbol x is added to the Python int 1. Pythons operator rules
then allow SymPy to tell Python that SymPy objects know how to be added to Python ints, and
so 1 is automatically converted to the SymPy Integer object.
This sort of operator magic happens automatically behind the scenes, and you rarely need to
even know that it is happening. However, there is one exception. Whenever you combine a
SymPy object and a SymPy object, or a SymPy object and a Python object, you get a SymPy
object, but whenever you combine two Python objects, SymPy never comes into play, and so
you get a Python object.
>>> type(Integer(1) + 1)
<class sympy.core.numbers.Integer>
>>> type(1 + 1)
<... int>
This is usually not a big deal. Python ints work much the same as SymPy Integers, but there
is one important exception: division. In SymPy, the division of two Integers gives a Rational:
>>> Integer(1)/Integer(3)
1/3
>>> type(Integer(1)/Integer(3))
<class sympy.core.numbers.Rational>
But in Python / represents either integer division or oating point division, depending on
whether you are in Python 2 or Python 3, and depending on whether or not you have run from
future import division:
>>> from __future__ import division
>>> 1/2
0.5
To avoid this, we can construct the rational object explicitly
>>> Rational(1, 2)
1/2
This problem also comes up whenever we have a larger symbolic expression with int/int in
it. For example:
>>> x + 1/2
x + 0.5
This happens because Python rst evaluates 1/2 into 0.5, and then that is cast into a SymPy
type when it is added to x. Again, we can get around this by explicitly creating a Rational:
>>> x + Rational(1, 2)
x + 1/2
There are several tips on avoiding this situation in the Gotchas and Pitfalls (page 61) document.
14
Chapter 2. SymPy Tutorial
SymPy Documentation, Release 0.7.6
2.3.4 Further Reading
For more discussion on the topics covered in this section, see Gotchas and Pitfalls (page 61).
2.4 Basic Operations
Here we discuss some of the most basic operations needed for expression manipulation in
SymPy. Some more advanced operations will be discussed later in the advanced expression
manipulation (page 52) section.
>>> from sympy import *
>>> x, y, z = symbols(x y z)
2.4.1 Substitution
One of the most common things you might want to do with a mathematical expression is
substitution. Substitution replaces all instances of something in an expression with something
else. It is done using the subs method. For example
>>> expr = cos(x) + 1
>>> expr.subs(x, y)
cos(y) + 1
Substitution is usually done for one of two reasons:
1. Evaluating an expression at a point. For example, if our expression is cos(x) + 1 and
we want to evaluate it at the point x = 0, so that we get cos(0) + 1, which is 2.
>>> expr.subs(x, 0)
2
2. Replacing a subexpression with another subexpression. There are two reasons we might
want to do this. The rst is if we are trying to build an expression that has some symmexx
try, such as xx . To build this, we might start with x**y, and replace y with x**y. We
would then get x**(x**y). If we replaced y in this new expression with x**x, we would
get x**(x**(x**x)), the desired expression.
>>> expr = x**y
>>> expr
x**y
>>> expr = expr.subs(y, x**y)
>>> expr
x**(x**y)
>>> expr = expr.subs(y, x**x)
>>> expr
x**(x**(x**x))
The second is if we want to perform a very controlled simplication, or perhaps a simplication that SymPy is otherwise unable to do. For example, say we have sin(2x)+cos(2x),
and we want to replace sin(2x) with 2 sin(x) cos(x). As we will learn later, the function
expand trig does this. However, this function will also expand cos(2x), which we may
not want. While there are ways to perform such precise simplication, and we will learn
some of them in the advanced expression manipulation (page 52) section, an easy way
is to just replace sin(2x) with 2 sin(x) cos(x).
2.4. Basic Operations
15
SymPy Documentation, Release 0.7.6
>>> expr = sin(2*x) + cos(2*x)
>>> expand_trig(expr)
2*sin(x)*cos(x) + 2*cos(x)**2 - 1
>>> expr.subs(sin(2*x), 2*sin(x)*cos(x))
2*sin(x)*cos(x) + cos(2*x)
There are two important things to note about subs. First, it returns a new expression. SymPy
objects are immutable. That means that subs does not modify it in-place. For example
>>> expr = cos(x)
>>> expr.subs(x, 0)
1
>>> expr
cos(x)
>>> x
x
Quick Tip
SymPy expressions are immutable. No function will change them in-place.
Here, we see that performing expr.subs(x, 0) leaves expr unchanged. In fact, since SymPy
expressions are immutable, no function will change them in-place. All functions will return
new expressions.
To perform multiple substitutions at once, pass a list of (old, new) pairs to subs.
>>> expr = x**3 + 4*x*y - z
>>> expr.subs([(x, 2), (y, 4), (z, 0)])
40
It is often useful to combine this with a list comprehension to do a large set of similar replacements all at once. For example, say we had x4 4x3 + 4x2 2x + 3 and we wanted to replace
all instances of x that have an even power with y, to get y 4 4x3 + 4y 2 2x + 3.
>>> expr = x**4 - 4*x**3 + 4*x**2 - 2*x + 3
>>> replacements = [(x**i, y**i) for i in range(5) if i % 2 == 0]
>>> expr.subs(replacements)
-4*x**3 - 2*x + y**4 + 4*y**2 + 3
2.4.2 Converting Strings to SymPy Expressions
The sympify function (thats sympify, not to be confused with simplify) can be used to
convert strings into SymPy expressions.
For example
>>> str_expr = x**2 + 3*x - 1/2
>>> expr = sympify(str_expr)
>>> expr
x**2 + 3*x - 1/2
>>> expr.subs(x, 2)
19/2
Warning: sympify uses eval. Dont use it on unsanitized input.
16
Chapter 2. SymPy Tutorial
SymPy Documentation, Release 0.7.6
2.4.3 evalf
To evaluate a numerical expression into a oating point number, use evalf.
>>> expr = sqrt(8)
>>> expr.evalf()
2.82842712474619
SymPy can evaluate oating point expressions to arbitrary precision. By default, 15 digits of
precision are used, but you can pass any number as the argument to evalf. Lets compute
the rst 100 digits of .
>>> pi.evalf(100)
3.141592653589793238462643383279502884197169399375105820974944592307816406286208998628034825342117068
To numerically evaluate an expression with a Symbol at a point, we might use subs followed
by evalf, but it is more ecient and numerically stable to pass the substitution to evalf
using the subs ag, which takes a dictionary of Symbol: point pairs.
>>> expr = cos(2*x)
>>> expr.evalf(subs={x: 2.4})
0.0874989834394464
Sometimes there are roundo errors smaller than the desired precision that remain after an
expression is evaluated. Such numbers can be removed at the users discretion by setting the
chop ag to True.
>>> one = cos(1)**2 + sin(1)**2
>>> (one - 1).evalf()
-0.e-124
>>> (one - 1).evalf(chop=True)
0
2.4.4 lambdify
subs and evalf are good if you want to do simple evaluation, but if you intend to evaluate
an expression at many points, there are more ecient ways. For example, if you wanted to
evaluate an expression at a thousand points, using SymPy would be far slower than it needs
to be, especially if you only care about machine precision. Instead, you should use libraries
like NumPy and SciPy.
The easiest way to convert a SymPy expression to an expression that can be numerically
evaluated is to use the lambdify function. lambdify acts like a lambda function, except it
converts the SymPy names to the names of the given numerical library, usually NumPy. For
example
>>> import numpy
>>> a = numpy.arange(10)
>>> expr = sin(x)
>>> f = lambdify(x, expr, numpy)
>>> f(a)
[ 0.
0.84147098 0.90929743
-0.2794155
0.6569866
0.98935825
0.14112001 -0.7568025
0.41211849]
-0.95892427
You can use other libraries than NumPy. For example, to use the standard library math module, use math.
2.4. Basic Operations
17
SymPy Documentation, Release 0.7.6
>>> f = lambdify(x, expr, math)
>>> f(0.1)
0.0998334166468
To use lambdify with numerical libraries that it does not know about, pass a dictionary of
sympy name:numerical function pairs. For example
>>> def mysin(x):
...
...
My sine. Not only accurate for small x.
...
...
return x
>>> f = lambdify(x, expr, {sin:mysin})
>>> f(0.1)
0.1
2.5 Printing
As we have already seen, SymPy can pretty print its output using Unicode characters. This
is a short introduction to the most common printing options available in SymPy.
2.5.1 Printers
There are several printers available in SymPy. The most common ones are
str
repr
ASCII pretty printer
Unicode pretty printer
LaTeX
MathML
Dot
In addition to these, there are also printers that can output SymPy objects to code, such as
C, Fortran, Javascript, Theano, and Python. These are not discussed in this tutorial.
2.5.2 Setting up Pretty Printing
If all you want is the best pretty printing, use the init printing() function. This will automatically enable the best printer available in your environment.
>>> from sympy import init_printing
>>> init_printing()
Quick Tip
You an also change the printer used in SymPy Live. Just change the Output Format in
the settings.
18
Chapter 2. SymPy Tutorial
SymPy Documentation, Release 0.7.6
If you plan to work in an interactive calculator-type session, the init session() function will
automatically import everything in SymPy, create some common Symbols, setup plotting, and
run init printing().
>>> from sympy import init_session
>>> init_session()
Python console for SymPy 0.7.3 (Python 2.7.5-64-bit) (ground types: gmpy)
These commands were executed:
>>> from future import division
>>> from sympy import *
>>> x, y, z, t = symbols(x y z t)
>>> k, m, n = symbols(k m n, integer=True)
>>> f, g, h = symbols(f g h, cls=Function)
>>> init printing() # doctest: +SKIP
Documentation can be found at http://www.sympy.org
>>>
In any case, this is what will happen:
In the IPython QTConsole, if LATEX is installed, it will enable a printer that uses LATEX.
If LATEX is not installed, but Matplotlib is installed, it will use the Matplotlib rendering
engine. If Matplotlib is not installed, it uses the Unicode pretty printer.
In the IPython notebook, it will use MathJax to render LATEX.
2.5. Printing
19
SymPy Documentation, Release 0.7.6
In an IPython console session, or a regular Python session, it will use the Unicode pretty
printer if the terminal supports Unicode.
In a terminal that does not support Unicode, the ASCII pretty printer is used.
20
Chapter 2. SymPy Tutorial
SymPy Documentation, Release 0.7.6
To explicitly not use LATEX, pass use latex=False to init printing() or init session(). To
explicitly not use Unicode, pass use unicode=False.
2.5.3 Printing Functions
In addition to automatic printing, you can explicitly use any one of the printers by calling the
appropriate function.
str
To get a string form of an expression, use str(expr). This is also the form that is produced
by print(expr). String forms are designed to be easy to read, but in a form that is correct
Python syntax so that it can be copied and pasted. The str() form of an expression will
usually look exactly the same as the expression as you would enter it.
>>> from sympy import *
>>> x, y, z = symbols(x y z)
>>> str(Integral(sqrt(1/x), x))
Integral(sqrt(1/x), x)
>>> print(Integral(sqrt(1/x), x))
Integral(sqrt(1/x), x)
repr
The repr form of an expression is designed to show the exact form of an expression. It will
be discussed more in the Advanced Expression Manipulation (page 52) section. To get it, use
2.5. Printing
21
SymPy Documentation, Release 0.7.6
srepr() 1 .
>>> srepr(Integral(sqrt(1/x), x))
Integral(Pow(Pow(Symbol(x), Integer(-1)), Rational(1, 2)), Tuple(Symbol(x)))
The repr form is mostly useful for understanding how an expression is built internally.
ASCII Pretty Printer
The ASCII pretty printer is accessed from pprint(). If the terminal does not support Unicode,
the ASCII printer is used by default. Otherwise, you must pass use unicode=False.
>>> pprint(Integral(sqrt(1/x), x), use_unicode=False)
/
|
|
___
|
/ 1
|
/ - dx
| \/
x
|
/
pprint() prints the output to the screen. If you want the string form, use pretty().
>>> pretty(Integral(sqrt(1/x), x), use_unicode=False)
/
\n |
\n |
___
\n |
/ 1
\n |
>>> print(pretty(Integral(sqrt(1/x), x), use_unicode=False))
/
|
|
___
|
/ 1
|
/ - dx
| \/
x
|
/
dx\n | \\/
Unicode Pretty Printer
The Unicode pretty printer is also accessed from pprint() and pretty(). It the terminal
supports Unicode, it is used automatically. If pprint() is not able to detect that the terminal
supports unicode, you can pass use unicode=True to force it to use Unicode.
>>> pprint(Integral(sqrt(1/x), x), use_unicode=True)
___
- dx
\
x
1 SymPy does not use the Python builtin repr() function for repr printing, because in Python str(list) calls
repr() on the elements of the list, and some SymPy functions return lists (such as solve()). Since srepr() is so
verbose, it is unlikely that anyone would want it called by default on the output of solve().
22
Chapter 2. SymPy Tutorial
\n |
SymPy Documentation, Release 0.7.6
LATEX
To get the LATEX form of an expression, use latex().
>>> print(latex(Integral(sqrt(1/x), x)))
\int \sqrt{\frac{1}{x}}\, dx
The latex() function has many options to change the formatting of dierent things. See its
documentation (page 1260) for more details.
MathML
There is also a printer to MathML, called print mathml().
sympy.printing.mathml.
It must be imported from
>>> from sympy.printing.mathml import print_mathml
>>> print_mathml(Integral(sqrt(1/x), x))
<apply>
<int/>
<bvar>
<ci>x</ci>
</bvar>
<apply>
<root/>
<apply>
<power/>
<ci>x</ci>
<cn>-1</cn>
</apply>
</apply>
</apply>
print mathml() prints the output. If you want the string, use the function mathml().
Dot
The dotprint() function in sympy.printing.dot prints output to dot format, which can be
rendered with Graphviz. See the Advanced Expression Manipulation (page 52) section for
some examples of the output of this printer.
2.6 Simplication
To make this document easier to read, we are going to enable pretty printing.
>>> from sympy import *
>>> x, y, z = symbols(x y z)
>>> init_printing(use_unicode=True)
2.6.1 simplify
Now lets jump in and do some interesting mathematics. One of the most useful features of
a symbolic manipulation system is the ability to simplify mathematical expressions. SymPy
2.6. Simplication
23
SymPy Documentation, Release 0.7.6
has dozens of functions to perform various kinds of simplication. There is also one general
function called simplify() that attempts to apply all of these functions in an intelligent way
to arrive at the simplest form of an expression. Here are some examples
>>> simplify(sin(x)**2 + cos(x)**2)
1
>>> simplify((x**3 + x**2 - x - 1)/(x**2 + 2*x + 1))
x - 1
>>> simplify(gamma(x)/gamma(x - 2))
(x - 2)(x - 1)
Here, gamma(x) is (x), the gamma function. We see that simplify() is capable of handling
a large class of expressions.
But simplify() has a pitfall. It just applies all the major simplication operations in SymPy,
and uses heuristics to determine the simplest result. But simplest is not a well-dened
term. For example, say we wanted to simplify x2 + 2x + 1 into (x + 1)2 :
>>> simplify(x**2 + 2*x + 1)
2
x + 2x + 1
We did not get what we want. There is a function to perform this simplication, called factor(), which will be discussed below.
Another pitfall to simplify() is that it can be unnecessarily slow, since it tries many kinds
of simplications before picking the best one. If you already know exactly what kind of simplication you are after, it is better to apply the specic simplication function(s) that apply
those simplications.
Applying specic simplication functions instead of simplify() also has the advantage that
specic functions have certain guarantees about the form of their output. These will be discussed with each function below. For example, factor(), when called on a polynomial with
rational coecients, is guaranteed to factor the polynomial into irreducible factors. simplify() has no guarantees. It is entirely heuristical, and, as we saw above, it may even miss
a possible type of simplication that SymPy is capable of doing.
simplify() is best when used interactively, when you just want to whittle down an expression to a simpler form. You may then choose to apply specic functions once you see what
simplify() returns, to get a more precise result. It is also useful when you have no idea what
form an expression will take, and you need a catchall function to simplify it.
2.6.2 Polynomial/Rational Function Simplication
expand
expand() is one of the most common simplication functions in SymPy. Although it has a lot
of scopes, for now, we will consider its function in expanding polynomial expressions. For
example:
>>> expand((x + 1)**2)
2
x + 2x + 1
>>> expand((x + 2)*(x - 3))
2
x - x - 6
24
Chapter 2. SymPy Tutorial
SymPy Documentation, Release 0.7.6
Given a polynomial, expand() will put it into a canonical form of a sum of monomials.
expand() may not sound like a simplication function. After all, by its very name, it makes
expressions bigger, not smaller. Usually this is the case, but often an expression will become
smaller upon calling expand() on it due to cancellation.
>>> expand((x + 1)*(x - 2) - (x - 1)*x)
-2
factor
factor() takes a polynomial and factors it into irreducible factors over the rational numbers.
For example:
>>> factor(x**3 - x**2 + x - 1)
2
(x - 1)x + 1
>>> factor(x**2*z + 4*x*y*z + 4*y**2*z)
2
z(x + 2y)
For polynomials, factor() is the opposite of expand(). factor() uses a complete multivariate factorization algorithm over the rational numbers, which means that each of the factors
returned by factor() is guaranteed to be irreducible.
If you are interested in the factors themselves, factor list returns a more structured output.
>>> factor_list(x**2*z + 4*x*y*z + 4*y**2*z)
(1, [(z, 1), (x + 2y, 2)])
Note that the input to factor and expand need not be polynomials in the strict sense. They
will intelligently factor or expand any kind of expression (though note that the factors may
not be irreducible if the input is no longer a polynomial over the rationals).
>>> expand((cos(x) + sin(x))**2)
2
2
sin (x) + 2sin(x)cos(x) + cos (x)
>>> factor(cos(x)**2 + 2*cos(x)*sin(x) + sin(x)**2)
2
(sin(x) + cos(x))
collect
collect() collects common powers of a term in an expression. For example
>>> expr = x*y + x
>>> expr
3
2
2
x - x z + 2x +
>>> collected_expr
>>> collected_expr
3
2
x + x (-z + 2) +
- 3 + 2*x**2 - z*x**2 + x**3
xy + x - 3
= collect(expr, x)
x(y + 1) - 3
collect() is particularly useful in conjunction with the .coeff() method. expr.coeff(x,
n) gives the coecient of x**n in expr:
2.6. Simplication
25
SymPy Documentation, Release 0.7.6
>>> collected_expr.coeff(x, 2)
-z + 2
cancel
cancel() will take any rational function and put it into the standard canonical form, pq , where
p and q are expanded polynomials with no common factors, and the leading coecients of p
and q do not have denominators (i.e., are integers).
>>> cancel((x**2 + 2*x + 1)/(x**2 + x))
x + 1
----x
>>> expr = 1/x + (3*x/2 - 2)/(x - 4)
>>> expr
3x
--- - 2
2
1
------- + x - 4
x
>>> cancel(expr)
2
3x - 2x - 8
-------------2
2x - 8x
>>> expr = (x*y**2 - 2*x*y*z + x*z**2 + y**2 - 2*y*z + z**2)/(x**2 - 1)
>>> expr
2
2
2
2
xy - 2xyz + xz + y - 2yz + z
--------------------------------------2
x - 1
>>> cancel(expr)
2
2
y - 2yz + z
--------------x - 1
Note that since factor() will completely factorize both the numerator and the denominator
of an expression, it can also be used to do the same thing:
>>> factor(expr)
2
(y - z)
-------x - 1
However, if you are only interested in making sure that the expression is in canceled form,
cancel() is more ecient than factor().
apart
apart() performs a partial fraction decomposition on a rational function.
26
Chapter 2. SymPy Tutorial
SymPy Documentation, Release 0.7.6
>>> expr = (4*x**3 + 21*x**2 + 10*x + 12)/(x**4 + 5*x**3 + 5*x**2 + 4*x)
>>> expr
3
2
4x + 21x + 10x + 12
-----------------------4
3
2
x + 5x + 5x + 4x
>>> apart(expr)
2x - 1
1
3
---------- - ----- + 2
x + 4
x
x + x + 1
2.6.3 Trigonometric Simplication
Note:
SymPy follows Pythons naming conventions for inverse trigonometric functions,
which is to append an a to the front of the functions name. For example, the inverse cosine, or arc cosine, is called acos().
>>> acos(x)
acos(x)
>>> cos(acos(x))
x
>>> asin(1)
trigsimp
To simplify expressions using trigonometric identities, use trigsimp().
>>> trigsimp(sin(x)**2 + cos(x)**2)
1
>>> trigsimp(sin(x)**4 - 2*cos(x)**2*sin(x)**2 + cos(x)**4)
cos(4x)
1
-------- + 2
2
>>> trigsimp(sin(x)*tan(x)/sec(x))
2
sin (x)
trigsimp() also works with hyperbolic trig functions.
>>> trigsimp(cosh(x)**2 + sinh(x)**2)
cosh(2x)
>>> trigsimp(sinh(x)/tanh(x))
cosh(x)
Much like simplify(), trigsimp() applies various trigonometric identities to the input expression, and then uses a heuristic to return the best one.
2.6. Simplication
27
SymPy Documentation, Release 0.7.6
expand trig
To expand trigonometric functions, that is, apply the sum or double angle identities, use
expand trig().
>>> expand_trig(sin(x + y))
sin(x)cos(y) + sin(y)cos(x)
>>> expand_trig(tan(2*x))
2tan(x)
------------2
- tan (x) + 1
Because expand trig() tends to make trigonometric expressions larger, and trigsimp()
tends to make them smaller, these identities can be applied in reverse using trigsimp()
>>> trigsimp(sin(x)*cos(y) + sin(y)*cos(x))
sin(x + y)
2.6.4 Powers
Before we introduce the power simplication functions, a mathematical discussion on the
identities held by powers is in order. There are three kinds of identities satised by exponents
1. xa xb = xa+b
2. xa y a = (xy)a
3. (xa )b = xab
Identity 1 is always true.
Identity 2 is not always true. For example, if x = y = 1 and a = 12 , then xa y a = 1 1 =
i i = 1, whereas (xy)a = 1 1 = 1 = 1. However, identity 2 is true at least if x and y
are nonnegative and a is real (it may also be true
other conditions as well). A common
under
consequence of the failure of identity 2 is that x y 6= xy.
(
)1/2
Identity 3 is not always true. For example, if x = 1, a = 2, and b = 12 , then (xa )b = (1)2
=
ab
21/2
1
1 = 1 and x = (1)
= (1) = 1. However, identity 3 is true when b is an integer (again,
it may also hold in other cases
as well). Two common consequences of the failure of identity
2
3 are that x 6= x and that x1 6= 1x .
To summarize
Identity
1. xa xb = xa+b
2. xa y a = (xy)a
3. (xa )b = xab
Sucient conditions to hold
Always true
x, y 0 and a R
bZ
Counterexample
when
conditions
are not met
None
(1)1/2 (1)1/2 6= (1
1)1/2
(
(1)2
)1/2
6= (1)21/2
Important consequences
None
x y 6=
xy in general
x2 6= x and x1 6= 1x
in general
This is important to remember, because by default, SymPy will not perform simplications if
they are not true in general.
28
Chapter 2. SymPy Tutorial
SymPy Documentation, Release 0.7.6
In order to make SymPy perform simplications involving identities that are only true under
certain assumptions, we need to put assumptions on our Symbols. We will undertake a full
discussion of the assumptions system later, but for now, all we need to know are the following.
By default, SymPy Symbols are assumed to be complex (elements of C). That is, a simplication will not be applied to an expression with a given Symbol unless it holds for all
complex numbers.
Symbols can be given dierent assumptions by passing the assumption to symbols().
For the rest of this section, we will be assuming that x and y are positive, and that a and
b are real. We will leave z, t, and c as arbitrary complex Symbols to demonstrate what
happens in that case.
>>> x, y = symbols(x y, positive=True)
>>> a, b = symbols(a b, real=True)
>>> z, t, c = symbols(z t c)
Note: In SymPy, sqrt(x) is just a shortcut to x**Rational(1, 2). They are exactly the
same object.
>>> sqrt(x) == x**Rational(1, 2)
True
powsimp
powsimp() applies identities 1 and 2 from above, from left to right.
>>> powsimp(x**a*x**b)
a + b
x
>>> powsimp(x**a*y**a)
a
(xy)
Notice that powsimp() refuses to do the simplication if it is not valid.
>>> powsimp(t**c*z**c)
c c
t z
If you know that you want to apply this simplication, but you dont want to mess with assumptions, you can pass the force=True ag. This will force the simplication to take place,
regardless of assumptions.
>>> powsimp(t**c*z**c, force=True)
c
(tz)
Note that in some instances, in particular, when the exponents are integers or rational numbers, and identity 2 holds, it will be applied automatically
>>> (z*t)**2
2 2
t z
>>> sqrt(x*y)
___
___
\ x \ y
2.6. Simplication
29
SymPy Documentation, Release 0.7.6
This means that it will be impossible to undo this identity with powsimp(), because even if
powsimp() were to put the bases together, they would be automatically split apart again.
>>> powsimp(z**2*t**2)
2 2
t z
>>> powsimp(sqrt(x)*sqrt(y))
___
___
\ x \ y
expand power exp / expand power base
expand power exp() and expand power base() apply identities 1 and 2 from right to left,
respectively.
>>> expand_power_exp(x**(a + b))
a b
x x
>>> expand_power_base((x*y)**a)
a a
x y
As with powsimp(), identity 2 is not applied if it is not valid.
>>> expand_power_base((z*t)**c)
c
(tz)
And as with powsimp(), you can force the expansion to happen without ddling with assumptions by using force=True.
>>> expand_power_base((z*t)**c, force=True)
c c
t z
As with identity 2, identity 1 is applied automatically if the power is a number, and hence
cannot be undone with expand power exp().
>>> x**2*x**3
5
x
>>> expand_power_exp(x**5)
5
x
powdenest
powdenest() applies identity 3, from left to right.
>>> powdenest((x**a)**b)
ab
x
As before, the identity is not applied if it is not true under the given assumptions.
30
Chapter 2. SymPy Tutorial
SymPy Documentation, Release 0.7.6
>>> powdenest((z**a)**b)
b
a
z
And as before, this can be manually overridden with force=True.
>>> powdenest((z**a)**b, force=True)
ab
z
2.6.5 Exponentials and logarithms
Note: In SymPy, as in Python and most programming languages, log is the natural logarithm,
also known as ln. SymPy automatically provides an alias ln = log in case you forget this.
>>> ln(x)
log(x)
Logarithms have similar issues as powers. There are two main identities
1. log (xy) = log (x) + log (y)
2. log (xn ) = n log (x)
Neither identity is true for arbitrary complex x and y, due to the branch cut in the complex
plane for the complex logarithm. However, sucient conditions for the identities to hold are
if x and y are positive and n is real.
>>> x, y = symbols(x y, positive=True)
>>> n = symbols(n, real=True)
As before, z and t will be Symbols with no additional assumptions.
( )
( )
Note that the identity log xy = log(x)log(y) is a special case of identities 1 and 2 by log xy =
(
)
(
)
log x y1 = log(x) + log y 1 = log(x) log(y), and thus it also holds if x and y are positive,
but may not hold in general.
We also see that log (ex ) = x comes from log (ex ) = x log(e) = x, and thus holds when x is real
(and
( it can
) be veried that it does not hold in general for arbitrary complex x, for example,
log ex+2i = log (ex ) = x 6= x + 2i).
expand log
To apply identities 1 and 2 from left to right, use expand log(). As always, the identities will
not be applied unless they are valid.
>>> expand_log(log(x*y))
log(x) + log(y)
>>> expand_log(log(x/y))
log(x) - log(y)
>>> expand_log(log(x**2))
2log(x)
>>> expand_log(log(x**n))
nlog(x)
2.6. Simplication
31
SymPy Documentation, Release 0.7.6
>>> expand_log(log(z*t))
log(tz)
As with powsimp() and powdenest(), expand log() has a force option that can be used to
ignore assumptions.
>>> expand_log(log(z**2))
2
logz
>>> expand_log(log(z**2), force=True)
2log(z)
logcombine
To apply identities 1 and 2 from right to left, use logcombine().
>>> logcombine(log(x) + log(y))
log(xy)
>>> logcombine(n*log(x))
n
logx
>>> logcombine(n*log(z))
nlog(z)
logcombine() also has a force option that can be used to ignore assumptions.
>>> logcombine(n*log(z), force=True)
n
logz
2.6.6 Special Functions
SymPy implements dozens of special functions, ranging from functions in combinatorics to
mathematical physics.
An extensive list of the special functions included with SymPy and their documentation is at
the Functions Module (page 342) page.
For the purposes of this tutorial, lets introduce a few special functions in SymPy.
Lets dene x, y, and z as regular, complex Symbols, removing any assumptions we put on
them in the previous section. We will also dene k, m, and n.
>>> x, y, z = symbols(x y z)
>>> k, m, n = symbols(k m n)
The factorial function is factorial. factorial(n) represents n! = 1 2 (n 1) n. n! represents the number of permutations of n distinct items.
>>> factorial(n)
n!
( )
The binomial coecient function is binomial. binomial(n, k) represents nk , the number
of ways to choose k items from a set of n distinct items. It is also often written as nCk, and is
pronounced n choose k.
32
Chapter 2. SymPy Tutorial
SymPy Documentation, Release 0.7.6
>>> binomial(n, k)
n
k
The factorial
function is closely related to the gamma function, gamma. gamma(z) represents
(z) = 0 tz1 et dt, which for positive integer z is the same as (z 1)!.
>>> gamma(z)
(z)
The generalized hypergeometric
function
is hyper. hyper([a 1, ..., a p], [b 1, ...,
)
(
a1 , . . . , ap
b q], z) represents p Fq
z . The most common case is 2 F1 , which is often referred
b1 , . . . , bq
to as the ordinary hypergeometric function.
>>> hyper([1, 2], [3], z)
- 1, 2 |
-
| z
2 1 3
|
rewrite
A common way to deal with special functions is to rewrite them in terms of one another. This
works for any function in SymPy, not just special functions. To rewrite an expression in terms
of a function, use expr.rewrite(function). For example,
>>> tan(x).rewrite(sin)
2
2sin (x)
--------sin(2x)
>>> factorial(x).rewrite(gamma)
(x + 1)
For some tips on applying more targeted rewriting, see the Advanced Expression Manipulation (page 52) section.
expand func
To expand special functions in terms of some identities, use expand func(). For example
>>> expand_func(gamma(x + 3))
x(x + 1)(x + 2)(x)
hyperexpand
To rewrite hyper in terms of more standard functions, use hyperexpand().
>>> hyperexpand(hyper([1, 1], [2], z))
-log(-z + 1)
------------z
2.6. Simplication
33
SymPy Documentation, Release 0.7.6
hyperexpand() also works on the more general Meijer G-function (see its documentation
(page 428) for more information).
>>> expr = meijerg([[1],[1]], [[1],[]], -z)
>>> expr
-1, 1 1 1 |
| -z
-2, 1 1
|
>>> hyperexpand(expr)
1
z
e
combsimp
To simplify combinatorial expressions, use combsimp().
>>> combsimp(factorial(n)/factorial(n - 3))
n(n - 2)(n - 1)
>>> combsimp(binomial(n+1, k+1)/binomial(n, k))
n + 1
----k + 1
combsimp() also simplies expressions with gamma.
>>> combsimp(gamma(x)*gamma(1 - x))
-------sin(x)
2.6.7 Example: Continued Fractions
Lets use SymPy to explore continued fractions. A continued fraction is an expression of the
form
1
a0 +
a1 +
a2 +
1
..
.+
1
an
where a0 , . . . , an are integers, and a1 , . . . , an are positive. A continued fraction can also be innite, but innite objects are more dicult to represent in computers, so we will only examine
the nite case here.
A continued fraction of the above form is often represented as a list [a0 ; a1 , . . . , an ]. Lets write
a simple function that converts such a list to its continued fraction form. The easiest way
to construct a continued fraction from a list is to work backwards. Note that despite the
apparent symmetry of the denition, the rst element, a0 , must usually be handled dierently
from the rest.
34
Chapter 2. SymPy Tutorial
SymPy Documentation, Release 0.7.6
>>> def list_to_frac(l):
...
expr = Integer(0)
...
for i in reversed(l[1:]):
...
expr += i
...
expr = 1/expr
...
return l[0] + expr
>>> list_to_frac([x, y, z])
1
x + ----1
y + z
We use Integer(0) in list to frac so that the result will always be a SymPy object, even if
we only pass in Python ints.
>>> list_to_frac([1, 2, 3, 4])
43
-30
Every nite continued fraction is a rational number, but we are interested in symbolics here,
so lets create a symbolic continued fraction. The symbols() function that we have been using
has a shortcut to create numbered symbols. symbols(a0:5) will create the symbols a0, a1,
..., a5.
>>> syms = symbols(a0:5)
>>> syms
(a, a1, a2, a, a)
>>> a0, a1, a2, a3, a4 = syms
>>> frac = list_to_frac(syms)
>>> frac
1
a + ----------------1
a1 + -----------1
a2 + ------1
a + -a
This form is useful for understanding continued fractions, but lets put it into standard rational
function form using cancel().
>>> frac = cancel(frac)
>>> frac
aa1a2aa + aa1a2 + aa1a + aaa + a + a2aa + a2 + a
------------------------------------------------------------------------a1a2aa + a1a2 + a1a + aa + 1
Now suppose we were given frac in the above canceled form. In fact, we might be given the
fraction in any form, but we can always put it into the above canonical form with cancel().
Suppose that we knew that it could be rewritten as a continued fraction. How could we do
this with SymPy? A continued fraction is recursively c + f1 , where c is an integer and f is
a (smaller) continued fraction. If we could write the expression in this form, we could pull
out each c recursively and add it to a list. We could then get a continued fraction with our
list to frac() function.
2.6. Simplication
35
SymPy Documentation, Release 0.7.6
The key observation here is that we can convert an expression to the form c + f1 by doing a
partial fraction decomposition with respect to c. This is because f does not contain c. This
means we need to use the apart() function. We use apart() to pull the term out, then
subtract it from the expression, and take the reciprocal to get the f part.
>>> l = []
>>> frac = apart(frac, a0)
>>> frac
a2aa + a2 + a
a + --------------------------------------a1a2aa + a1a2 + a1a + aa + 1
>>> l.append(a0)
>>> frac = 1/(frac - a0)
>>> frac
a1a2aa + a1a2 + a1a + aa + 1
--------------------------------------a2aa + a2 + a
Now we repeat this process
>>> frac = apart(frac, a1)
>>> frac
aa + 1
a1 + -----------------a2aa + a2 + a
>>> l.append(a1)
>>> frac = 1/(frac - a1)
>>> frac = apart(frac, a2)
>>> frac
a
a2 + --------aa + 1
>>> l.append(a2)
>>> frac = 1/(frac - a2)
>>> frac = apart(frac, a3)
>>> frac
1
a + -a
>>> l.append(a3)
>>> frac = 1/(frac - a3)
>>> frac = apart(frac, a4)
>>> frac
a
>>> l.append(a4)
>>> list_to_frac(l)
1
a + ----------------1
a1 + -----------1
a2 + ------1
a + -a
36
Chapter 2. SymPy Tutorial
SymPy Documentation, Release 0.7.6
Quick Tip
You can execute multiple lines at once in SymPy Live. Typing Shift-Enter instead of
Enter will enter a newline instead of executing.
Of course, this exercise seems pointless, because we already know that our frac is
list to frac([a0, a1, a2, a3, a4]). So try the following exercise. Take a list of symbols
and randomize them, and create the canceled continued fraction, and see if you can reproduce
the original list. For example
>>>
>>>
>>>
>>>
>>>
import random
l = list(symbols(a0:5))
random.shuffle(l)
orig_frac = frac = cancel(list_to_frac(l))
del l
Click on Run code block in SymPy Live on the denition of list to frac) above, and then
on the above example, and try to reproduce l from frac. I have deleted l at the end to
remove the temptation for peeking (you can check your answer at the end by calling cancel(list to frac(l)) on the list that you generate at the end, and comparing it to orig frac.
See if you can think of a way to gure out what symbol to pass to apart() at each stage (hint:
1
think of what happens to a0 in the formula a0 + a1 +
when it is canceled).
2.7 Calculus
This section covers how to do basic calculus tasks such as derivatives, integrals, limits, and
series expansions in SymPy. If you are not familiar with the math of any part of this section,
you may safely skip it.
>>> from sympy import *
>>> x, y, z = symbols(x y z)
>>> init_printing(use_unicode=True)
2.7.1 Derivatives
To take derivatives, use the diff function.
>>> diff(cos(x), x)
-sin(x)
>>> diff(exp(x**2), x)
2
x
2xe
diff can take multiple derivatives at once. To take multiple derivatives, pass the variable as
many times as you wish to dierentiate, or pass a number after the variable. For example,
both of the following nd the third derivative of x4 .
>>> diff(x**4, x, x, x)
24x
>>> diff(x**4, x, 3)
24x
2.7. Calculus
37
SymPy Documentation, Release 0.7.6
You can also take derivatives with respect to many variables at once. Just pass each derivative
in order, using the same syntax as for single variable derivatives. For example, each of the
7
following will compute xy 2 z4 exyz .
>>> expr = exp(x*y*z)
>>> diff(expr, x, y, y, z, z,
3 2 3 3 3
2 2 2
x y x y z + 14x y z
>>> diff(expr, x, y, 2, z, 4)
3 2 3 3 3
2 2 2
x y x y z + 14x y z
>>> diff(expr, x, y, y, z, 4)
3 2 3 3 3
2 2 2
x y x y z + 14x y z
z, z)
xyz
+ 52xyz + 48e
xyz
+ 52xyz + 48e
xyz
+ 52xyz + 48e
diff can also be called as a method. The two ways of calling diff are exactly the same, and
are provided only for convenience.
>>> expr.diff(x, y, y, z, 4)
3 2 3 3 3
2 2 2
xyz
x y x y z + 14x y z + 52xyz + 48e
To create an unevaluated derivative, use the Derivative class. It has the same syntax as
diff.
>>> deriv = Derivative(expr, x, y, y, z, 4)
>>> deriv
7
xyz
----------e
4
2
z y x
To evaluate an unevaluated derivative, use the doit method.
>>> deriv.doit()
3 2 3 3 3
2 2 2
xyz
x y x y z + 14x y z + 52xyz + 48e
These unevaluated objects are useful for delaying the evaluation of the derivative, or for printing purposes. They are also used when SymPy does not know how to compute the derivative
of an expression (for example, if it contains an undened function, which are described in the
Solving Dierential Equations (page 45) section).
2.7.2 Integrals
To compute an integral, use the integrate function. There are two kinds of integrals, denite
and indenite. To compute an indenite integral, that is, an antiderivative, or primitive, just
pass the variable after the expression.
>>> integrate(cos(x), x)
sin(x)
Note that SymPy does not include the constant of integration. If you want it, you can add
one yourself, or rephrase your problem as a dierential equation and use dsolve to solve it,
which does add the constant (see Solving Dierential Equations (page 45)).
38
Chapter 2. SymPy Tutorial
SymPy Documentation, Release 0.7.6
Quick Tip
in SymPy is oo (thats the lowercase letter oh twice). This is because oo looks like
, and is easy to type.
To compute a denite integral, pass the argument (integration variable, lower limit,
upper limit). For example, to compute
ex dx,
0
we would do
>>> integrate(exp(-x), (x, 0, oo))
1
As with indenite integrals, you can pass multiple limit tuples to perform a multiple integral.
For example, to compute
2
2
ex y dx dy,
do
>>> integrate(exp(-x**2 - y**2), (x, -oo, oo), (y, -oo, oo))
If integrate is unable to compute an integral, it returns an unevaluated Integral object.
>>> expr = integrate(x**x, x)
>>> print(expr)
Integral(x**x, x)
>>> expr
x
x dx
As with Derivative, you can create an unevaluated integral using Integral. To later evaluate
this integral, call doit.
>>> expr = Integral(log(x)**2, x)
>>> expr
2
log (x) dx
>>> expr.doit()
2
xlog (x) - 2xlog(x) + 2x
integrate uses powerful algorithms that are always improving to compute both denite and
indenite integrals, including heuristic pattern matching type algorithms, a partial implementation of the Risch algorithm, and an algorithm using Meijer G-functions that is useful
for computing integrals in terms of special functions, especially denite integrals. Here is a
sampling of some of the power of integrate.
2.7. Calculus
39
SymPy Documentation, Release 0.7.6
>>> integ = Integral((x**4 + x**2*exp(x) - x**2 - 2*x*exp(x) - 2*x ...
exp(x))*exp(x)/((x - 1)**2*(x + 1)**2*(exp(x) + 1)), x)
>>> integ
4
2 x
2
x
x x
x + x e - x - 2xe - 2x - e e
---------------------------------------- dx
2
2 x
(x - 1) (x + 1) e + 1
>>> integ.doit()
x
x
e
loge + 1 + -----2
x - 1
>>> integ = Integral(sin(x**2), x)
>>> integ
2
sinx dx
>>> integ.doit()
___
___
___
\ 2 x
3\ 2 \ fresnels-------(3/4)
___
\
-------------------------------------8(7/4)
>>> integ = Integral(x**y*exp(-x), (x, 0, oo))
>>> integ
y -x
x e
dx
0
>>> integ.doit()
(y + 1)
for -re(y) < 1
y -x
x e
dx
otherwise
This last example returned a Piecewise expression because the integral does not converge
unless <(y) > 1.
40
Chapter 2. SymPy Tutorial
SymPy Documentation, Release 0.7.6
2.7.3 Limits
SymPy can compute symbolic limits with the limit function. The syntax to compute
lim f (x)
xx0
is limit(f(x), x, x0).
>>> limit(sin(x)/x, x, 0)
1
limit should be used instead of subs whenever the point of evaluation is a singularity. Even
though SymPy has objects to represent , using them for evaluation is not reliable because
they do not keep track of things like rate of growth. Also, things like and
return nan
(not-a-number). For example
>>> expr = x**2/exp(x)
>>> expr.subs(x, oo)
nan
>>> limit(expr, x, oo)
0
Like Derivative and Integral, limit has an unevaluated counterpart, Limit. To evaluate
it, use doit.
>>> expr = Limit((cos(x) - 1)/x, x, 0)
>>> expr
cos(x) - 1
lim ---------x-0
x
>>> expr.doit()
0
To evaluate a limit at one side only, pass + or - as a third argument to limit. For example,
to compute
lim
x0+
1
,
x
do
>>> limit(1/x, x, 0, +)
As opposed to
>>> limit(1/x, x, 0, -)
-
2.7.4 Series Expansion
SymPy can compute asymptotic series expansions of functions around a point. To compute
the expansion of f (x) around the point x = x0 terms of order xn , use f(x).series(x, x0, n).
x0 and n can be omitted, in which case the defaults x0=0 and n=6 will be used.
>>> expr = exp(sin(x))
>>> expr.series(x, 0, 4)
2
2.7. Calculus
41
SymPy Documentation, Release 0.7.6
x
4
1 + x + -- + Ox
2
( )
The O x4 term at the end represents the Landau order term at x = 0 (not to be confused with
big O notation used in computer science, which generally represents the Landau order term
at x = ). It means that all x terms with power greater than or equal to x4 are omitted. Order
terms can be created and manipulated outside of series. They automatically absorb higher
order terms.
>>> x + x**3 + x**6 + O(x**4)
3
4
x + x + Ox
>>> x*O(1)
O(x)
If you do not want the order term, use the removeO method.
>>> expr.series(x, 0, 4).removeO()
2
x
-- + x + 1
2
The O notation supports arbitrary limit points (other than 0):
>>> exp(x - 6).series(x, x0=6)
2
3
4
5
(x - 6)
(x - 6)
(x - 6)
(x - 6)
-5 + -------- + -------- + -------- + -------- + x + O(x - 6) ; x 6
2
6
24
120
2.7.5 Finite dierences
So far we have looked at expressions with analytical derivatives and primitive functions respectively. But what if we want to have an expression to estimate a derivative of a curve for
which we lack a closed form representation, or for which we dont know the functional values
for yet. One approach would be to use a nite dierence approach.
You can use the as finite diff method of on any Derivative instance to generate approximations to derivatives of arbitrary order:
>>> f = Function(f)
>>> dfdx = f(x).diff(x)
>>> as_finite_diff(dfdx)
-f(x - 1/2) + f(x + 1/2)
here the rst order derivative was approximated around x using a minimum number of points
(2 for 1st order derivative) evaluated equidistantly using a step-size of 1. We can use arbitrary
steps (possibly containing symbolic expressions):
>>> f = Function(f)
>>> d2fdx2 = f(x).diff(x, 2)
>>> h = Symbol(h)
>>> as_finite_diff(d2fdx2, [-3*h,-h,2*h])
f(-3h)
f(-h)
2f(2h)
------- - ----- + --------
42
Chapter 2. SymPy Tutorial
SymPy Documentation, Release 0.7.6
2
5h
2
3h
2
15h
If you are just interested in evaluating the weights, you can do so manually:
>>> finite_diff_weights(2, [-3, -1, 2], 0)[-1][-1]
[1/5, -1/3, 2/15]
note that we only need the last element in the last sublist returned from nite di weights.
The reason for this is that nite di weights also generates weights for lower derivatives and
using fewer points (see the documentation of finite diff weights for more details).
if using finite diff weights directly looks complicated and the as finite diff function
operating on Derivative instances is not exible enough, you can use apply finite diff
which takes order, x list, y list and x0 as parameters:
>>> x_list = [-3, 1, 2]
>>> y_list = symbols(a b c)
>>> apply_finite_diff(1, x_list, y_list, 0)
3a
b
2c
- --- - - + --20
4
5
2.8 Solvers
>>> from sympy import *
>>> x, y, z = symbols(x y z)
>>> init_printing(use_unicode=True)
2.8.1 A Note about Equations
Recall from the gotchas (page 12) section of this tutorial that symbolic equations in SymPy
are not represented by = or ==, but by Eq.
>>> Eq(x, y)
x = y
However, there is an even easier way. In SymPy, any expression is not in an Eq is automatically
assumed to equal 0 by the solving functions. Since a = b if and only if a b = 0, this means
that instead of using x == y, you can just use x - y. For example
>>> solve(Eq(x**2, 1), x)
[-1, 1]
>>> solve(Eq(x**2 - 1, 0), x)
[-1, 1]
>>> solve(x**2 - 1, x)
[-1, 1]
This is particularly useful if the equation you wish to solve is already equal to 0. Instead of
typing solve(Eq(expr, 0), x), you can just use solve(expr, x).
2.8. Solvers
43
SymPy Documentation, Release 0.7.6
2.8.2 Solving Equations Algebraically
The main function for solving algebraic equations, as we saw above, is solve. The syntax is
solve(equations, variables), where, as we saw above, equations may be in the form of
Eq instances or expressions that are assumed to be equal to zero.
When solving a single equation, the output of solve is a list of the solutions.
>>> solve(x**2 - x, x)
[0, 1]
If no solutions are found, an empty list is returned, or NotImplementedError is raised.
>>> solve(exp(x), x)
[]
Note: If solve returns [] or raises NotImplementedError, it doesnt mean that the equation
has no solutions. It just means that it couldnt nd any. Often this means that the solutions
cannot be represented symbolically. For example, the equation x = cos(x) has a solution, but
it cannot be represented symbolically using standard functions.
>>> solve(x - cos(x), x)
Traceback (most recent call last):
...
NotImplementedError: multiple generators [x, exp(I*x)]
No algorithms are implemented to solve equation exp(I*x)
Traceback (most recent call last):
...
NotImplementedError: multiple generators [x, exp(I*x)]
In fact, solve makes no guarantees whatsoever about the completeness of the solutions it
nds. Much of solve is heuristics, which may nd some solutions to an equation or system
of equations, but not all of them.
solve can also solve systems of equations. Pass a list of equations and a list of variables to
solve for.
>>>
{x:
>>>
solve([x - y + 2, x + y - 3], [x, y])
1/2, y: 5/2}
solve([x*y - 7, x + y - 6], [x, y])
___
___
___
___
\ 2 + 3, \ 2 + 3, \ 2 + 3, - \ 2 + 3
Note: The type of the output of solve when solving systems of equations varies depending
on the type of the input. If you want a consistent interface, pass dict=True.
>>> solve([x - y + 2, x + y - 3], [x, y], dict=True)
[{x: 1/2, y: 5/2}]
>>> solve([x*y - 7, x + y - 6], [x, y], dict=True)
___
___
___
___
x: - \ 2 + 3, y: \ 2 + 3, x: \ 2 + 3, y: - \ 2 + 3
solve reports each solution only once. To get the solutions of a polynomial including multiplicity use roots.
44
Chapter 2. SymPy Tutorial
SymPy Documentation, Release 0.7.6
>>>
[0,
>>>
{0:
solve(x**3 - 6*x**2 + 9*x, x)
3]
roots(x**3 - 6*x**2 + 9*x, x)
1, 3: 2}
The output {0: 1, 3: 2} of roots means that 0 is a root of multiplicity 1 and 3 is a root of
multiplicity 2.
2.8.3 Solving Dierential Equations
To solve dierential equations, use dsolve. First, create an undened function by passing
cls=Function to the symbols function.
>>> f, g = symbols(f g, cls=Function)
f and g are now undened functions. We can call f(x), and it will represent an unknown
function.
>>> f(x)
f(x)
Derivatives of f(x) are unevaluated.
>>> f(x).diff(x)
d
--(f(x))
dx
(see the Derivatives (page 37) section for more on derivatives).
To represent the dierential equation f 00 (x) 2f 0 (x) + f (x) = sin(x), we would thus use
>>> diffeq = Eq(f(x).diff(x, x) - 2*f(x).diff(x) + f(x), sin(x))
>>> diffeq
2
d
d
f(x) - 2--(f(x)) + ---(f(x)) = sin(x)
dx
2
dx
To solve the ODE, pass it and the function to solve for to dsolve.
>>> dsolve(diffeq, f(x))
x
cos(x)
f(x) = (C1 + C2x)e + -----2
dsolve returns an instance of Eq. This is because in general, solutions to dierential equations cannot be solved explicitly for the function.
>>> dsolve(f(x).diff(x)*(1 - sin(f(x))), f(x))
f(x) + cos(f(x)) = C1
The arbitrary constants in the solutions from dsolve are symbols of the form C1, C2, C3, and
so on.
2.8. Solvers
45
SymPy Documentation, Release 0.7.6
2.9 Matrices
>>> from sympy import *
>>> init_printing(use_unicode=True)
To make a matrix in SymPy, use the Matrix object. A matrix is constructed by providing a list
of row vectors that make up the matrix. For example, to construct the matrix
1 1
3 4
0 2
use
>>>
1
Matrix([[1, -1], [3, 4], [0, 2]])
-1
To make it easy to make column vectors, a list of elements is considered to be a column vector.
>>> Matrix([1, 2, 3])
1
2
3
Matrices are manipulated just like any other object in SymPy or Python.
>>> M = Matrix([[1, 2, 3], [3, 2, 1]])
>>> N = Matrix([0, 1, 1])
>>> M*N
5
3
One important thing to note about SymPy matrices is that, unlike every other object in SymPy,
they are mutable. This means that they can be modied in place, as we will see below. The
downside to this is that Matrix cannot be used in places that require immutability, such as
inside other SymPy expressions or as keys to dictionaries. If you need an immutable version
of Matrix, use ImmutableMatrix.
2.9.1 Basic Operations
Shape
Here are some basic operations on Matrix. To get the shape of a matrix use shape
>>> M = Matrix([[1, 2, 3], [-2, 0, 4]])
>>> M
1
2 3
-2 0 4
46
Chapter 2. SymPy Tutorial
SymPy Documentation, Release 0.7.6
>>> M.shape
(2, 3)
Accessing Rows and Columns
To get an individual row or column of a matrix, use row or col. For example, M.row(0) will
get the rst row. M.col(-1) will get the last column.
>>> M.row(0)
[1 2 3]
>>> M.col(-1)
3
4
Deleting and Inserting Rows and Columns
To delete a row or column, use row del or col del. These operations will modify the Matrix
in place.
>>>
>>>
2
0
>>>
>>>
[2
M.col_del(0)
M
3
4
M.row_del(1)
M
3]
To insert rows or columns, use row insert or col insert. These operations do not operate
in place.
>>>
[2
>>>
>>>
2
0
>>>
>>>
1
-2
M
3]
M = M.row_insert(1, Matrix([[0, 4]]))
M
3
4
M = M.col_insert(0, Matrix([1, -2]))
M
2 3
0 4
Unless explicitly stated, the methods mentioned below do not operate in place. In general,
a method that does not operate in place will return a new Matrix and a method that does
operate in place will return None.
2.9.2 Basic Methods
As noted above, simple operations like addition and multiplication are done just by using +,
*, and **. To nd the inverse of a matrix, just raise it to the -1 power.
2.9. Matrices
47
SymPy Documentation, Release 0.7.6
>>> M = Matrix([[1, 3], [-2, 3]])
>>> N = Matrix([[0, 3], [0, 7]])
>>> M + N
1
6
-2 10
>>> M*N
0 24
0 15
>>> 3*M
3
9
-6 9
>>> M**2
-5 12
-8 3
>>> M**-1
1/3 -1/3
2/9 1/9
>>> N**-1
Traceback (most recent call last):
...
ValueError: Matrix det == 0; not invertible.
Traceback (most recent call last):
...
ValueError: Matrix det == 0; not invertible.
To take the transpose of a Matrix, use T.
>>>
>>>
1
4
>>>
1
M = Matrix([[1, 2, 3], [4, 5, 6]])
M
2 3
5 6
M.T
4
2.9.3 Matrix Constructors
Several constructors exist for creating common matrices. To create an identity matrix, use
eye. eye(n) will create an n n identity matrix.
>>>
1
0
>>>
1
48
eye(3)
0 0
1 0
0 1
eye(4)
0 0 0
Chapter 2. SymPy Tutorial
SymPy Documentation, Release 0.7.6
To create a matrix of all zeros, use zeros. zeros(n, m) creates an n m matrix of 0s.
>>> zeros(2, 3)
0 0 0
0 0 0
Similarly, ones creates a matrix of ones.
>>>
1
ones(3, 2)
1
To create diagonal matrices, use diag. The arguments to diag can be either numbers or
matrices. A number is interpreted as a 1 1 matrix. The matrices are stacked diagonally. The
remaining elements are lled with 0s.
>>>
1
0
>>>
-1
diag(1, 2, 3)
0 0
2 0
0 3
diag(-1, ones(2, 2), Matrix([5, 7, 5]))
0 0 0
1 1 0
1 1 0
0 0 5
0 0 7
0 0 5
2.9.4 Advanced Methods
Determinant
To compute the determinant of a matrix, use det.
>>> M = Matrix([[1, 0, 1], [2, -1, 3], [4, 3, 2]])
>>> M
1 0
1
2.9. Matrices
49
SymPy Documentation, Release 0.7.6
2 -1 3
4 3
2
>>> M.det()
-1
RREF
To put a matrix into reduced row echelon form, use rref. rref returns a tuple of two elements.
The rst is the reduced row echelon form, and the second is a list of indices of the pivot
columns.
>>> M = Matrix([[1, 0, 1, 3], [2, 3, 4, 7], [-1, -3, -3, -4]])
>>> M
1
0
1
3
2
3
4
7
-1 -3 -3 -4
>>> M.rref()
1 0
1
3 , [0, 1]
0 1 2/3 1/3
0 0
0
0
Note: The rst element of the tuple returned by rref is of type Matrix. The second is of
type list.
Nullspace
To nd the nullspace of a matrix, use nullspace. nullspace returns a list of column vectors
that span the nullspace of the matrix.
>>> M = Matrix([[1, 2, 3, 0, 0], [4, 10, 0, 0, 1]])
>>> M
1 2
3 0 0
4 10 0 0 1
>>> M.nullspace()
-15, 0, 1
6 0 -1/2
1 0 0
0 1 0
0 0 1
50
Chapter 2. SymPy Tutorial
SymPy Documentation, Release 0.7.6
Eigenvalues, Eigenvectors, and Diagonalization
To nd the eigenvalues of a matrix, use eigenvals. eigenvals returns a dictionary of eigenvalue:algebraic multiplicity pairs (similar to the output of roots (page 44)).
>>> M = Matrix([[3, -2,
>>> M
3 -2 4
-2
5 3
-3 -2
5 -2 2
-2
5 -2 -3 3
>>> M.eigenvals()
{-2: 1, 3: 1, 5: 2}
4, -2], [5,
3, -3, -2], [5, -2,
2, -2], [5, -2, -3,
3]])
This means that M has eigenvalues -2, 3, and 5, and that the eigenvalues -2 and 3 have algebraic multiplicity 1 and that the eigenvalue 5 has algebraic multiplicity 2.
To nd the eigenvectors of a matrix, use eigenvects. eigenvects returns a list of tuples of
the form (eigenvalue:algebraic multiplicity, [eigenvectors]).
>>> M.eigenvects()
-2, 1, 0, 3, 1, 1,
1
1
1
1
1
1
5, 2, 1,
0
-1
0
1
This shows us that, for example, the eigenvalue 5 also has geometric multiplicity 2, because
it has two eigenvectors. Because the algebraic and geometric multiplicities are the same for
all the eigenvalues, M is diagonalizable.
To diagonalize a matrix, use diagonalize. diagonalize returns a tuple (P, D), where D is
diagonal and M = P DP 1 .
>>>
>>>
0
1
>>>
-2
0
>>>
3
P, D = M.diagonalize()
P
1 1 0
1 1 -1
1 1 0
1 0 1
D
0 0 0
3 0 0
0 5 0
0 0 5
P*D*P**-1
-2 4
-2
2.9. Matrices
51
SymPy Documentation, Release 0.7.6
5 3
-3 -2
5 -2 2
-2
5 -2 -3 3
>>> P*D*P**-1 == M
True
Quick Tip
lambda is a reserved keyword in Python, so to create a Symbol called , while using the
same names for SymPy Symbols and Python variables, use lamda (without the b). It will
still pretty print as .
Note that since eigenvects also includes the eigenvalues, you should use it instead of eigenvals if you also want the eigenvectors. However, as computing the eigenvectors may often
be costly, eigenvals should be preferred if you only wish to nd the eigenvalues.
If all you want is the characteristic polynomial, use charpoly. This is more ecient than
eigenvals, because sometimes symbolic roots can be expensive to calculate.
>>> lamda = symbols(lamda)
>>> p = M.charpoly(lamda)
>>> factor(p)
2
( - 5) ( - 3)( + 2)
2.10 Advanced Expression Manipulation
In this section, we discuss some ways that we can perform advanced manipulation of expressions.
2.10.1 Understanding Expression Trees
Quick Tip
To play with the srepr form of expressions in the SymPy Live shell, change the output
format to Repr in the settings.
Before we can do this, we need to understand how expressions are represented in SymPy.
A mathematical expression is represented as a tree. Let us take the expression x2 + xy, i.e.,
x**2 + x*y. We can see what this expression looks like internally by using srepr
>>> from sympy import *
>>> x, y, z = symbols(x y z)
>>> expr = x**2 + x*y
>>> srepr(expr)
Add(Pow(Symbol(x), Integer(2)), Mul(Symbol(x), Symbol(y)))
52
Chapter 2. SymPy Tutorial
SymPy Documentation, Release 0.7.6
The easiest way to tear this apart is to look at a diagram of the expression tree:
Add
Symbol('x')
Mul
Pow
Symbol('y')
Integer(2)
Symbol('x')
Note: The above diagram was made using Graphviz and the dotprint (page 1269) function.
First, lets look at the leaves of this tree. Symbols are instances of the class Symbol. While
we have been doing
>>> x = symbols(x)
we could have also done
>>> x = Symbol(x)
Either way, we get a Symbol with the name x 2 . For the number in the expression, 2, we
got Integer(2). Integer is the SymPy class for integers. It is similar to the Python built-in
type int, except that Integer plays nicely with other SymPy types.
When we write x**2, this creates a Pow object. Pow is short for power.
>>> srepr(x**2)
Pow(Symbol(x), Integer(2))
We could have created the same object by calling Pow(x, 2)
>>> Pow(x, 2)
x**2
Note that in the srepr output, we see Integer(2), the SymPy version of integers, even though
technically, we input 2, a Python int. In general, whenever you combine a SymPy object with
a non-SymPy object via some function or operation, the non-SymPy object will be converted
into a SymPy object. The function that does this is sympify 3 .
>>> type(2)
<... int>
>>> type(sympify(2))
<class sympy.core.numbers.Integer>
2 We have been using symbols instead of Symbol because it automatically splits apart strings into multiple Symbols.
symbols(x y z) returns a tuple of three Symbols. Symbol(x y z) returns a single Symbol called x y z.
3 Technically, it is an internal function called sympify, which diers from sympify in that it does not convert
strings. x + 2 is not allowed.
2.10. Advanced Expression Manipulation
53
SymPy Documentation, Release 0.7.6
We have seen that x**2 is represented as Pow(x, 2). What about x*y? As we might expect,
this is the multiplication of x and y. The SymPy class for multiplication is Mul.
>>> srepr(x*y)
Mul(Symbol(x), Symbol(y))
Thus, we could have created the same object by writing Mul(x, y).
>>> Mul(x, y)
x*y
Now we get to our nal expression, x**2 + x*y. This is the addition of our last two objects,
Pow(x, 2), and Mul(x, y). The SymPy class for addition is Add, so, as you might expect, to
create this object, we use Add(Pow(x, 2), Mul(x, y)).
>>> Add(Pow(x, 2), Mul(x, y))
x**2 + x*y
SymPy expression trees can have many branches, and can be quite deep or quite broad. Here
is a more complicated example
>>> expr = sin(x*y)/2 - x**2 + 1/y
>>> srepr(expr)
Add(Mul(Integer(-1), Pow(Symbol(x), Integer(2))), Mul(Rational(1, 2),
sin(Mul(Symbol(x), Symbol(y)))), Pow(Symbol(y), Integer(-1)))
Here is a diagram
Add
Pow
Symbol('y')
Integer(-1)
Mul
Rational(1, 2)
Mul
sin
Integer(-1)
Mul
Symbol('y')
Symbol('x')
Pow
Integer(2)
Symbol('x')
This expression reveals some interesting things about SymPy expression trees. Lets go
through them one by one.
Lets rst look at the term x**2. As we expected, we see Pow(x, 2). One level up, we see we
have Mul(-1, Pow(x, 2)). There is no subtraction class in SymPy. x - y is represented as
x + -y, or, more completely, x + -1*y, i.e., Add(x, Mul(-1, y)).
>>> expr = x - y
>>> srepr(x - y)
Add(Symbol(x), Mul(Integer(-1), Symbol(y)))
54
Chapter 2. SymPy Tutorial
SymPy Documentation, Release 0.7.6
Add
Symbol('x')
Mul
Symbol('y')
Integer(-1)
Next, look at 1/y. We might expect to see something like Div(1, y), but similar to subtraction,
there is no class in SymPy for division. Rather, division is represented by a power of -1. Hence,
we have Pow(y, -1). What if we had divided something other than 1 by y, like x/y? Lets
see.
>>> expr = x/y
>>> srepr(expr)
Mul(Symbol(x), Pow(Symbol(y), Integer(-1)))
Mul
Symbol('x')
Pow
Symbol('y')
2.10. Advanced Expression Manipulation
Integer(-1)
55
SymPy Documentation, Release 0.7.6
We see that x/y is represented as x*y**-1, i.e., Mul(x, Pow(y, -1)).
Finally, lets look at the sin(x*y)/2 term. Following the pattern of the previous example,
we might expect to see Mul(sin(x*y), Pow(Integer(2), -1)). But instead, we have
Mul(Rational(1, 2), sin(x*y)). Rational numbers are always combined into a single
term in a multiplication, so that when we divide by 2, it is represented as multiplying by 1/2.
Finally, one last note. You may have noticed that the order we entered our expression and the
order that it came out from srepr or in the graph were dierent. You may have also noticed
this phenonemon earlier in the tutorial. For example
>>> 1 + x
x + 1
This because in SymPy, the arguments of the commutative operations Add and Mul are stored
in an arbitrary (but consistent!) order, which is independent of the order inputted (if youre
worried about noncommutative multiplication, dont be. In SymPy, you can create noncommutative Symbols using Symbol(A, commutative=False), and the order of multiplication
for noncommutative Symbols is kept the same as the input). Furthermore, as we shall see in
the next section, the printing order and the order in which things are stored internally need
not be the same either.
Quick Tip
The way an expression is represented internally and the way it is printed are often not
the same.
In general, an important thing to keep in mind when working with SymPy expression trees
is this: the internal representation of an expression and the way it is printed need not be the
same. The same is true for the input form. If some expression manipulation algorithm is not
working in the way you expected it to, chances are, the internal representation of the object
is dierent from what you thought it was.
2.10.2 Recursing through an Expression Tree
Now that you know how expression trees work in SymPy, lets look at how to dig our way
through an expression tree. Every object in SymPy has two very important attributes, func,
and args.
func
func is the head of the object. For example, (x*y).func is Mul. Usually it is the same as the
class of the object (though there are exceptions to this rule).
Two notes about func. First, the class of an object need not be the same as the one used to
create it. For example
>>> expr = Add(x, x)
>>> expr.func
<class sympy.core.mul.Mul>
We created Add(x, x), so we might expect expr.func to be Add, but instead we got Mul. Why
is that? Lets take a closer look at expr.
56
Chapter 2. SymPy Tutorial
SymPy Documentation, Release 0.7.6
>>> expr
2*x
Add(x, x), i.e., x + x, was automatically converted into Mul(2, x), i.e., 2*x, which is a Mul.
SymPy classes make heavy use of the new class constructor, which, unlike init , allows
a dierent class to be returned from the constructor.
Second, some classes are special-cased, usually for eciency reasons 4 .
>>> Integer(2).func
<class sympy.core.numbers.Integer>
>>> Integer(0).func
<class sympy.core.numbers.Zero>
>>> Integer(-1).func
<class sympy.core.numbers.NegativeOne>
For the most part, these issues will not bother us. The special classes Zero, One, NegativeOne,
and so on are subclasses of Integer, so as long as you use isinstance, it will not be an issue.
args
args are the top-level arguments of the object. (x*y).args would be (x, y). Lets look at
some examples
>>> expr = 3*y**2*x
>>> expr.func
<class sympy.core.mul.Mul>
>>> expr.args
(3, x, y**2)
From this, we can see that expr == Mul(3, y**2, x). In fact, we can see that we can
completely reconstruct expr from its func and its args.
>>> expr.func(*expr.args)
3*x*y**2
>>> expr == expr.func(*expr.args)
True
Note that although we entered 3*y**2*x, the args are (3, x, y**2). In a Mul, the Rational
coecient will come rst in the args, but other than that, the order of everything else follows
no special pattern. To be sure, though, there is an order.
>>> expr = y**2*3*x
>>> expr.args
(3, x, y**2)
Muls args are sorted, so that the same Mul will have the same args. But the sorting is based
on some criteria designed to make the sorting unique and ecient that has no mathematical
signicance.
The srepr form of our expr is Mul(3, x, Pow(y, 2)). What if we want to get at the args of
Pow(y, 2). Notice that the y**2 is in the third slot of expr.args, i.e., expr.args[2].
4 Classes like One and Zero are singletonized, meaning that only one object is ever created, no matter how many
times the class is called. This is done for space eciency, as these classes are very common. For example, Zero
might occur very often in a sparse matrix represented densely. As we have seen, NegativeOne occurs any time we
have -x or 1/x. It is also done for speed eciency because singletonized objects can be compared by is. The unique
objects for each singletonized class can be accessed from the S object.
2.10. Advanced Expression Manipulation
57
SymPy Documentation, Release 0.7.6
>>> expr.args[2]
y**2
So to get the args of this, we call expr.args[2].args.
>>> expr.args[2].args
(y, 2)
Now what if we try to go deeper. What are the args of y. Or 2. Lets see.
>>> y.args
()
>>> Integer(2).args
()
They both have empty args. In SymPy, empty args signal that we have hit a leaf of the
expression tree.
So there are two possibilities for a SymPy expression. Either it has empty args, in which case
it is a leaf node in any expression tree, or it has args, in which case, it is a branch node of any
expression tree. When it has args, it can be completely rebuilt from its func and its args.
This is expressed in the key invariant.
Key Invariant
Every well-formed SymPy expression must either have empty args or satisfy expr ==
expr.func(*expr.args).
(Recall that in Python if a is a tuple, then f(*a) means to call f with arguments from the
elements of a, e.g., f(*(1, 2, 3)) is the same as f(1, 2, 3).)
This key invariant allows us to write simple algorithms that walk expression trees, change
them, and rebuild them into new expressions.
Walking the Tree
With this knowledge, lets look at how we can recurse through an expression tree. The nested
nature of args is a perfect t for recursive functions. The base case will be empty args. Lets
write a simple function that goes through an expression and prints all the args at each level.
>>> def pre(expr):
...
print(expr)
...
for arg in expr.args:
...
pre(arg)
See how nice it is that () signals leaves in the expression tree. We dont even have to write
a base case for our recursion; it is handled automatically by the for loop.
Lets test our function.
>>> expr = x*y + 1
>>> pre(expr)
x*y + 1
1
x*y
x
y
58
Chapter 2. SymPy Tutorial
SymPy Documentation, Release 0.7.6
Can you guess why we called our function pre? We just wrote a pre-order traversal function
for our expression tree. See if you can write a post-order traversal function.
Such traversals are so common in SymPy that the generator functions preorder traversal
and postorder traversal are provided to make such traversals easy. We could have also
written our algorithm as
>>> for arg in preorder_traversal(expr):
...
print(arg)
x*y + 1
1
x*y
x
y
2.10. Advanced Expression Manipulation
59
SymPy Documentation, Release 0.7.6
60
Chapter 2. SymPy Tutorial
CHAPTER
THREE
GOTCHAS AND PITFALLS
3.1 Introduction
SymPy runs under the Python Programming Language, so there are some things that may
behave dierently than they do in other, independent computer algebra systems like Maple
or Mathematica. These are some of the gotchas and pitfalls that you may encounter when
using SymPy. See also the FAQ, the Tutorial (page 5), the remainder of the SymPy Docs, and
the ocial Python Tutorial.
If you are already familiar with C or Java, you might also want to look at this 4 minute Python
tutorial.
Ignore #doctest: +SKIP in the examples. That has to do with internal testing of the examples.
3.2 Equals Signs (=)
3.2.1 Single Equals Sign
The equals sign (=) is the assignment operator, not equality. If you want to do x = y, use
Eq(x, y) for equality. Alternatively, all expressions are assumed to equal zero, so you can
just subtract one side and use x - y.
The proper use of the equals sign is to assign expressions to variables.
For example:
>>>
>>>
>>>
x -
from sympy.abc import x, y
a = x - y
print(a)
y
3.2.2 Double Equals Signs
Double equals signs (==) are used to test equality. However, this tests expressions exactly,
not symbolically. For example:
61
SymPy Documentation, Release 0.7.6
>>> (x + 1)**2 == x**2 + 2*x + 1
False
>>> (x + 1)**2 == (x + 1)**2
True
If you want to test for symbolic equality, one way is to subtract one expression from the
other and run it through functions like expand(), simplify(), and trigsimp() and see if the
equation reduces to 0.
>>>
>>>
0
>>>
>>>
0
>>>
0
from sympy import simplify, cos, sin, expand
simplify((x + 1)**2 - (x**2 + 2*x + 1))
eq = sin(2*x) - 2*sin(x)*cos(x)
simplify(eq)
expand(eq, trig=True)
Note: See also Why does SymPy say that two equal expressions are unequal? in the FAQ.
3.3 Variables
3.3.1 Variables Assignment does not Create a Relation Between Expressions
When you use = to do assignment, remember that in Python, as in most programming languages, the variable does not change if you change the value you assigned to it. The equations you are typing use the values present at the time of creation to ll in values, just like
regular Python denitions. They are not altered by changes made afterwards. Consider the
following:
>>>
>>>
>>>
>>>
a +
>>>
>>>
4
>>>
a +
from sympy import Symbol
a = Symbol(a) # Symbol, a, stored as variable a
b = a + 1
# an expression involving a stored as variable b
print(b)
1
a = 4
# a now points to literal integer 4, not Symbol(a)
print(a)
print(b)
1
# b is still pointing at the expression involving a
Changing quantity a does not change b; you are not working with a set of simultaneous equations. It might be helpful to remember that the string that gets printed when you print a
variable referring to a SymPy object is the string that was given to it when it was created;
that string does not have to be the same as the variable that you assign it to.
>>> from sympy import var
>>> r, t, d = var(rate time short_life)
>>> d = r*t
>>> print(d)
rate*time
>>> r = 80
62
Chapter 3. Gotchas and Pitfalls
SymPy Documentation, Release 0.7.6
>>> t = 2
>>> print(d)
rate*time
>>> d = r*t
>>> print(d)
160
# We havent changed d, only r and t
# Now d is using the current values of r and t
If you need variables that have dependence on each other, you can dene functions. Use the
def operator. Indent the body of the function. See the Python docs for more information on
dening functions.
>>>
>>>
c
>>>
d
>>>
...
...
...
...
...
>>>
c*d
>>>
>>>
2
>>>
2*d
c, d = var(c d)
print(c)
print(d)
def ctimesd():
This function returns whatever c is times whatever d is.
return c*d
ctimesd()
c = 2
print(c)
ctimesd()
If you dene a circular relationship, you will get a RuntimeError.
>>> def a():
...
return b()
...
>>> def b():
...
return a()
...
>>> a()
Traceback (most recent call last):
File ..., line ..., in ...
compileflags, 1) in test.globs
File <...>, line 1, in <module>
a()
File <...>, line 2, in a
return b()
File <...>, line 2, in b
return a()
File <...>, line 2, in a
return b()
...
RuntimeError: maximum recursion depth exceeded
Traceback (most recent call last):
File ..., line ..., in ...
compileflags, 1) in test.globs
File <...>, line 1, in <module>
a()
File <...>, line 2, in a
return b()
3.3. Variables
63
SymPy Documentation, Release 0.7.6
File <...>, line 2, in b
return a()
File <...>, line 2, in a
return b()
...
RuntimeError: maximum recursion depth exceeded
Note: See also Why doesnt changing one variable change another that depends on it? in
the FAQ.
3.3.2 Symbols
Symbols are variables, and like all other variables, they need to be assigned before you can
use them. For example:
>>> import sympy
>>> z**2 # z is not defined yet
Traceback (most recent call last):
File <stdin>, line 1, in <module>
NameError: name z is not defined
>>> sympy.var(z) # This is the easiest way to define z as a standard symbol
z
>>> z**2
z**2
Traceback (most recent call last):
File <stdin>, line 1, in <module>
NameError: name z is not defined
If you use isympy, it runs the following commands for you, giving you some default Symbols
and Functions.
>>>
>>>
>>>
>>>
>>>
from __future__ import division
from sympy import *
x, y, z, t = symbols(x y z t)
k, m, n = symbols(k m n, integer=True)
f, g, h = symbols(f g h, cls=Function)
You can also import common symbol names from sympy.abc.
>>> from sympy.abc import w
>>> w
w
>>> import sympy
>>> dir(sympy.abc)
[A, B, C, D, E, F, G, H, I, J, K, L, M, N, O,
P, Q, R, S, Symbol, T, U, V, W, X, Y, Z,
__builtins__, __doc__, __file__, __name__, __package__, _greek,
_latin, a, alpha, b, beta, c, chi, d, delta, e,
epsilon, eta, f, g, gamma, h, i, iota, j, k, kappa,
l, m, mu, n, nu, o, omega, omicron, p, phi, pi,
psi, q, r, rho, s, sigma, t, tau, theta, u, upsilon,
v, w, x, xi, y, z, zeta]
If you want control over the assumptions of the variables, use Symbol() and symbols(). See
Keyword Arguments (page 72) below.
64
Chapter 3. Gotchas and Pitfalls
SymPy Documentation, Release 0.7.6
Lastly, it is recommended that you not use I, E, S, N, C, O, or Q for variable or symbol names, as
those are used for the imaginary unit (i), the base of the natural logarithm (e), the sympify()
function (see Symbolic Expressions (page 65) below), numeric evaluation (N() is equivalent
to evalf() (page 331) ), the class registry (for things like C.cos(), to prevent cyclic imports
in some code), the big O order symbol (as in O(n log n)), and the assumptions object that
holds a list of supported ask keys (such as Q.real), respectively. You can use the mnemonic
QCOSINE to remember what Symbols are dened by default in SymPy. Or better yet, always
use lowercase letters for Symbol names. Python will not prevent you from overriding default
SymPy names or functions, so be careful.
>>> cos(pi) # cos and pi are a built-in sympy names.
-1
>>> pi = 3
# Notice that there is no warning for overriding pi.
>>> cos(pi)
cos(3)
>>> def cos(x): # No warning for overriding built-in functions either.
...
return 5*x
...
>>> cos(pi)
15
>>> from sympy import cos # reimport to restore normal behavior
To get a full list of all default names in SymPy do:
>>> import sympy
>>> dir(sympy)
# A big list of all default sympy names and functions follows.
# Ignore everything that starts and ends with __.
If you have IPython installed and use isympy, you can also press the TAB key to get a list of all
built-in names and to autocomplete. Also, see this page for a trick for getting tab completion
in the regular Python console.
Note: See also What is the best way to create symbols? in the FAQ.
3.4 Symbolic Expressions
3.4.1 Python numbers vs. SymPy Numbers
SymPy uses its own classes for integers, rational numbers, and oating point numbers instead
of the default Python int and float types because it allows for more control. But you have
to be careful. If you type an expression that just has numbers in it, it will default to a Python
expression. Use the sympify() function, or just S(), to ensure that something is a SymPy
expression.
>>> 6.2 # Python float. Notice the floating point accuracy problems.
6.2000000000000002
>>> type(6.2) # <type float> in Python 2.x, <class float> in Py3k
<... float>
>>> S(6.2) # SymPy Float has no such problems because of arbitrary precision.
6.20000000000000
>>> type(S(6.2))
<class sympy.core.numbers.Float>
3.4. Symbolic Expressions
65
SymPy Documentation, Release 0.7.6
If you include numbers in a SymPy expression, they will be sympied automatically, but there
is one gotcha you should be aware of. If you do <number>/<number> inside of a SymPy
expression, Python will evaluate the two numbers before SymPy has a chance to get to them.
The solution is to sympify() one of the numbers, or use Rational.
>>> x**(1/2) # evaluates to x**0 or x**0.5
x**0.5
>>> x**(S(1)/2) # sympyify one of the ints
sqrt(x)
>>> x**Rational(1, 2) # use the Rational class
sqrt(x)
With a power of 1/2 you can also use sqrt shorthand:
>>> sqrt(x) == x**Rational(1, 2)
True
If the two integers are not directly separated by a division sign then you dont have to worry
about this problem:
>>> x**(2*x/3)
x**(2*x/3)
Note: A common mistake is copying an expression that is printed and reusing it. If the
expression has a Rational (i.e., <number>/<number>) in it, you will not get the same result,
obtaining the Python result for the division rather than a SymPy Rational.
>>> x = Symbol(x)
>>> print(solve(7*x -22, x))
[22/7]
>>> 22/7 # If we just copy and paste we get int 3 or a float
3.142857142857143
>>> # One solution is to just assign the expression to a variable
>>> # if we need to use it again.
>>> a = solve(7*x - 22, x)
>>> a
[22/7]
The other solution is to put quotes around the expression and run it through S() (i.e., sympify
it):
>>> S(22/7)
22/7
Also, if you do not use isympy, you could use from
the / sign from performing integer division.
future
import division to prevent
>>> from __future__ import division
>>> 1/2
# With division imported it evaluates to a python float
0.5
>>> 1//2 # You can still achieve integer division with //
0
But be careful: you will now receive oats where you might have desired a Rational:
>>> x**(1/2)
x**0.5
66
Chapter 3. Gotchas and Pitfalls
SymPy Documentation, Release 0.7.6
Rational only works for number/number and is only meant for rational numbers. If you
want a fraction with symbols or expressions in it, just use /. If you do number/expression or
expression/number, then the number will automatically be converted into a SymPy Number.
You only need to be careful with number/number.
>>> Rational(2, x)
Traceback (most recent call last):
...
TypeError: invalid input: x
>>> 2/x
2/x
Traceback (most recent call last):
...
TypeError: invalid input: x
3.4.2 Evaluating Expressions with Floats and Rationals
SymPy keeps track of the precision of Float objects. The default precision is 15 digits. When
an expression involving a Float is evaluated, the result will be expressed to 15 digits of
precision but those digits (depending on the numbers involved with the calculation) may not
all be signicant.
The rst issue to keep in mind is how the Float is created: it is created with a value and
a precision. The precision indicates how precise of a value to use when that Float (or an
expression it appears in) is evaluated.
The values can be given as strings, integers, oats, or rationals.
strings and integers are interpreted as exact
>>> Float(100)
100.000000000000
>>> Float(100, 5)
100.00
to have the precision match the number of digits, the null string can be used
for the precision
>>> Float(100, )
100.
>>> Float(12.34)
12.3400000000000
>>> Float(12.34, )
12.34
>>> s, r = [Float(j, 3) for j in (0.25, Rational(1, 7))]
>>> for f in [s, r]:
...
print(f)
0.250
0.143
Next, notice that each of those values looks correct to 3 digits. But if we try to evaluate them
to 20 digits, a dierence will become apparent:
The 0.25 (with precision of 3) represents a number that has a non-repeating binary
decimal; 1/7 is repeating in binary and decimal it cannot be represented accurately
too far past those rst 3 digits (the correct decimal is a repeating 142857):
3.4. Symbolic Expressions
67
SymPy Documentation, Release 0.7.6
>>> s.n(20)
0.25000000000000000000
>>> r.n(20)
0.14285278320312500000
It is important to realize that although a Float is being displayed in decimal at aritrary precision, it is actually stored in binary. Once the Float is created, its binary
information is set at the given precision. The accuracy of that value cannot be subsequently changed; so 1/7, at a precision of 3 digits, can be padded with binary
zeros, but these will not make it a more accurate value of 1/7.
If inexact, low-precision numbers are involved in a calculation with with higher precision
values, the evalf engine will increase the precision of the low precision values and inexact
results will be obtained. This is feature of calculations with limited precision:
>>> Float(0.1, 10) + Float(0.1, 3)
0.2000061035
Although the evalf engine tried to maintain 10 digits of precision (since that was the highest
precision represented) the 3-digit precision used limits the accuracy to about 4 digits not all
the digits you see are signicant. evalf doesnt try to keep track of the number of signicant
digits.
That very simple expression involving the addition of two numbers with dierent precisions
will hopefully be instructive in helping you understand why more complicated expressions
(like trig expressions that may not be simplied) will not evaluate to an exact zero even
though, with the right simplication, they should be zero. Consider this unsimplied trig
identity, multiplied by a big number:
>>> big = 12345678901234567890
>>> big_trig_identity = big*cos(x)**2 + big*sin(x)**2 - big*1
>>> abs(big_trig_identity.subs(x, .1).n(2)) > 1000
True
When the cos and sin terms were evaluated to 15 digits of precision and multiplied by the
big number, they gave a large number that was only precise to 15 digits (approximately) and
when the 20 digit big number was subtracted the result was not zero.
There are three things that will help you obtain more precise numerical values for expressions:
1) Pass the desired substitutions with the call to evaluate. By doing the subs rst,
the Float values can not be updated as necessary. By passing the desired substitutions with the call to evalf the ability to re-evaluate as necessary is gained and the
results are impressively better:
>>> big_trig_identity.n(2, {x: 0.1})
-0.e-91
2) Use Rationals, not Floats. During the evaluation process, the Rational can be
computed to an arbitrary precision while the Float, once created at a default of
15 digits cannot. Compare the value of -1.4e+3 above with the nearly zero value
obtained when replacing x with a Rational representing 1/10 before the call to
evaluate:
>>> big_trig_identity.subs(x, S(1/10)).n(2)
0.e-91
3) Try to simplify the expression. In this case, SymPy will recognize the trig identity
and simplify it to zero so you dont even have to evaluate it numerically:
68
Chapter 3. Gotchas and Pitfalls
SymPy Documentation, Release 0.7.6
>>> big_trig_identity.simplify()
0
3.4.3 Immutability of Expressions
Expressions in SymPy are immutable, and cannot be modied by an in-place operation. This
means that a function will always return an object, and the original expression will not be
modied. The following example snippet demonstrates how this works:
def main():
var(x y a b)
expr = 3*x + 4*y
print(original =, expr)
expr_modified = expr.subs({x: a, y: b})
print(modified =, expr_modified)
if __name__ == __main__:
main()
The output shows that the subs() function has replaced variable x with variable a, and variable y with variable b:
original = 3*x + 4*y
modified = 3*a + 4*b
The subs() function does not modify the original expression expr. Rather, a modied copy of
the expression is returned. This returned object is stored in the variable expr modified. Note
that unlike C/C++ and other high-level languages, Python does not require you to declare a
variable before it is used.
3.4.4 Mathematical Operators
SymPy uses the same default operators as Python. Most of these, like */+-, are standard.
Aside from integer division discussed in Python numbers vs. SymPy Numbers (page 65) above,
you should also be aware that implied multiplication is not allowed. You need to use * whenever you wish to multiply something. Also, to raise something to a power, use **, not as
many computer algebra systems use. Parentheses () change operator precedence as you
would normally expect.
In isympy, with the ipython shell:
>>> 2x
Traceback (most recent call last):
...
SyntaxError: invalid syntax
>>> 2*x
2*x
>>> (x + 1)^2 # This is not power. Use ** instead.
Traceback (most recent call last):
...
TypeError: unsupported operand type(s) for ^: Add and int
>>> (x + 1)**2
(x + 1)**2
>>> pprint(3 - x**(2*x)/(x + 1))
2*x
3.4. Symbolic Expressions
69
SymPy Documentation, Release 0.7.6
x
- ----- + 3
x + 1
Traceback (most recent call last):
...
TypeError: unsupported operand type(s) for ^: Add and int
3.4.5 Inverse Trig Functions
SymPy uses dierent names for some functions than most computer algebra systems. In particular, the inverse trig functions use the python names of asin(), acos() and so on instead
of the usual arcsin and arccos. Use the methods described in Symbols (page 64) above to
see the names of all SymPy functions.
3.5 Special Symbols
The symbols [], {}, =, and () have special meanings in Python, and thus in SymPy. See the
Python docs linked to above for additional information.
3.5.1 Lists
Square brackets [] denote a list. A list is a container that holds any number of dierent
objects. A list can contain anything, including items of dierent types. Lists are mutable,
which means that you can change the elements of a list after it has been created. You access
the items of a list also using square brackets, placing them after the list or list variable. Items
are numbered using the space before the item.
Note: List indexes begin at 0.
Example:
>>>
>>>
[x,
>>>
x
>>>
>>>
[2,
>>>
[-1
a = [x, 1] # A simple list of two items
a
1]
a[0] # This is the first item
a[0] = 2 # You can change values of lists after they have been created
print(a)
1]
print(solve(x**2 + 2*x - 1, x)) # Some functions return lists
+ sqrt(2), -sqrt(2) - 1]
Note: See the Python docs for more information on lists and the square bracket notation for
accessing elements of a list.
70
Chapter 3. Gotchas and Pitfalls
SymPy Documentation, Release 0.7.6
3.5.2 Dictionaries
Curly brackets {} denote a dictionary, or a dict for short. A dictionary is an unordered list of
non-duplicate keys and values. The syntax is {key: value}. You can access values of keys
using square bracket notation.
>>> d = {a: 1, b: 2} # A dictionary.
>>> d
{a: 1, b: 2}
>>> d[a] # How to access items in a dict
1
>>> roots((x - 1)**2*(x - 2), x) # Some functions return dicts
{1: 2, 2: 1}
>>> # Some SymPy functions return dictionaries. For example,
>>> # roots returns a dictionary of root:multiplicity items.
>>> roots((x - 5)**2*(x + 3), x)
{-3: 1, 5: 2}
>>> # This means that the root -3 occurs once and the root 5 occurs twice.
Note: See the Python docs for more information on dictionaries.
3.5.3 Tuples
Parentheses (), aside from changing operator precedence and their use in function calls, (like
cos(x)), are also used for tuples. A tuple is identical to a list (page 70), except that it is not
mutable. That means that you can not change their values after they have been created. In
general, you will not need tuples in SymPy, but sometimes it can be more convenient to type
parentheses instead of square brackets.
>>> t = (1, 2, x) # Tuples are like lists
>>> t
(1, 2, x)
>>> t[0]
1
>>> t[0] = 4 # Except you can not change them after they have been created
Traceback (most recent call last):
File <console>, line 1, in <module>
TypeError: tuple object does not support item assignment
Traceback (most recent call last):
File <console>, line 1, in <module>
TypeError: tuple object does not support item assignment
Single element tuples, unlike lists, must have a comma in them:
>>> (x,)
(x,)
Without the comma, a single expression without a comma is not a tuple:
>>> (x)
x
integrate takes a sequence as the second argument if you want to integrate with
limits (and a tuple or list will work):
3.5. Special Symbols
71
SymPy Documentation, Release 0.7.6
>>> integrate(x**2, (x, 0, 1))
1/3
>>> integrate(x**2, [x, 0, 1])
1/3
Note: See the Python docs for more information on tuples.
3.5.4 Keyword Arguments
Aside from the usage described above (page 61), equals signs (=) are also used to give named
arguments to functions. Any function that has key=value in its parameters list (see below
on how to nd this out), then key is set to value by default. You can change the value of the
key by supplying your own value using the equals sign in the function call. Also, functions
that have ** followed by a name in the parameters list (usually **kwargs or **assumptions)
allow you to add any number of key=value pairs that you want, and they will all be evaluated
according to the function.
sqrt(x**2) doesnt auto simplify to x because x is assumed to be complex by default, and, for example, sqrt((-1)**2) == sqrt(1) == 1 != -1:
>>> sqrt(x**2)
sqrt(x**2)
Giving assumptions to Symbols is an example of using the keyword argument:
>>> x = Symbol(x, positive=True)
The square root will now simplify since it knows that x >= 0:
>>> sqrt(x**2)
x
powsimp has a default argument of combine=all:
>>> pprint(powsimp(x**n*x**m*y**n*y**m))
m + n
(x*y)
Setting combine to the default value is the same as not setting it.
>>> pprint(powsimp(x**n*x**m*y**n*y**m, combine=all))
m + n
(x*y)
The non-default options are exp, which combines exponents...
>>> pprint(powsimp(x**n*x**m*y**n*y**m, combine=exp))
m + n m + n
x
*y
...and base, which combines bases.
>>> pprint(powsimp(x**n*x**m*y**n*y**m, combine=base))
m
n
(x*y) *(x*y)
72
Chapter 3. Gotchas and Pitfalls
SymPy Documentation, Release 0.7.6
Note: See the Python docs for more information on function parameters.
3.6 Getting help from within SymPy
3.6.1 help()
Although all docs are available at docs.sympy.org or on the SymPy Wiki, you can also get info
on functions from within the Python interpreter that runs SymPy. The easiest way to do this
is to do help(function), or function? if you are using ipython:
In [1]: help(powsimp)
# help() works everywhere
In [2]: # But in ipython, you can also use ?, which is better because it
In [3]: # it gives you more information
In [4]: powsimp?
These will give you the function parameters and docstring for powsimp(). The output will
look something like this:
3.6.2 source()
Another useful option is the source() function. This will print the source code of a function,
including any docstring that it may have. You can also do function?? in ipython. For
example, from SymPy 0.6.5:
>>> source(simplify) # simplify() is actually only 2 lines of code.
In file: ./sympy/simplify/simplify.py
def simplify(expr):
Naively simplifies the given expression.
...
Simplification is not a well defined term and the exact strategies
this function tries can change in the future versions of SymPy. If
your algorithm relies on simplification (whatever it is), try to
determine what you need exactly - is it powsimp()? radsimp()?
together()?, logcombine()?, or something else? And use this particular
function directly, because those are well defined and thus your algorithm
will be robust.
...
expr = Poly.cancel(powsimp(expr))
return powsimp(together(expr.expand()), combine=exp, deep=True)
3.6. Getting help from within SymPy
73
SymPy Documentation, Release 0.7.6
74
Chapter 3. Gotchas and Pitfalls
CHAPTER
FOUR
SYMPY USERS GUIDE
4.1 Introduction
If you are new to SymPy, start with the Tutorial (page 5). If you went through it, now its time
to learn how SymPy works internally, which is what this guide is about. Once you grasp the
ideas behind SymPy, you will be able to use it eectively and also know how to extend it and
x it. You may also be just interested in SymPy Modules Reference (page 85).
4.2 Learning SymPy
Everyone has dierent ways of understanding the code written by others.
4.2.1 Ondejs approach
Lets say Id like to understand how x + y + x works and how it is possible that it gets
simplied to 2*x + y.
I write a simple script, I usually call it t.py (I dont remember anymore why I call it that way):
from sympy.abc import x, y
e = x + y + x
print e
And I try if it works
$ python t.py
y + 2*x
Now I start winpdb on it (if youve never used winpdb its an excellent multiplatform debugger, works on Linux, Windows and Mac OS X):
$ winpdb t.py
y + 2*x
and a winpdb window will popup, I move to the next line using F6:
75
SymPy Documentation, Release 0.7.6
Then I step into (F7) and after a little debugging I get for example:
Tip: Make the winpdb window larger on your screen, it was just made smaller to t in this
76
Chapter 4. SymPy Users Guide
SymPy Documentation, Release 0.7.6
guide.
I see values of all local variables in the left panel, so its very easy to see whats happening. You
can see, that the y + 2*x is emerging in the obj variable. Observing that obj is constructed
from c part and nc part and seeing what c part contains (y and 2*x). So looking at the line
28 (the whole line is not visible on the screenshot, so here it is):
c_part, nc_part, lambda_args, order_symbols = cls.flatten(map(_sympify, args))
you can see that the simplication happens in cls.flatten. Now you can set the breakpoint
on the line 28, quit winpdb (it will remember the breakpoint), start it again, hit F5, this will
stop at this breakpoint, hit F7, this will go into the function Add.flatten():
@classmethod
def flatten(cls, seq):
Takes the sequence seq of nested Adds and returns a flatten list.
Returns: (commutative_part, noncommutative_part, lambda_args,
order_symbols)
Applies associativity, all terms are commutable with respect to
addition.
terms = {}
# term -> coeff
# e.g. x**2 -> 5
for ... + 5*x**2 + ...
coeff = S.Zero
# standalone term
# e.g. 3 + ...
lambda_args = None
order_factors = []
while seq:
o = seq.pop(0)
and then you can study how it works. I am going to stop here, this should be enough to get
you going with the above technique, I am able to understand almost any Python code.
4.3 SymPys Architecture
We try to make the sources easily understandable, so you can look into the sources and read
the doctests, it should be well documented and if you dont understand something, ask on the
mailinglist.
You can nd all the decisions archived in the issues, to see rationale for doing this and that.
4.3.1 Basics
All symbolic things are implemented using subclasses of the Basic class. First, you need to
create symbols using Symbol(x) or numbers using Integer(5) or Float(34.3). Then you
construct the expression using any class from SymPy. For example Add(Symbol(a), Symbol(b)) gives an instance of the Add class. You can call all methods, which the particular
class supports.
For easier use, there is a syntactic sugar for expressions like:
4.3. SymPys Architecture
77
SymPy Documentation, Release 0.7.6
cos(x) + 1 is equal to cos(x). add (1) is equal to Add(cos(x), Integer(1))
or
2/cos(x) is equal to cos(x). rdiv (2) is equal to Mul(Rational(2),
Rational(-1))).
Pow(cos(x),
So, you can write normal expressions using python arithmetics like this:
a = Symbol(a)
b = Symbol(b)
e = (a + b)**2
print e
but from the SymPy point of view, we just need the classes Add, Mul, Pow, Rational, Integer.
4.3.2 Automatic evaluation to canonical form
For computation, all expressions need to be in a canonical form, this is done during the creation of the particular instance and only inexpensive operations are performed, necessary to
put the expression in the canonical form. So the canonical form doesnt mean the simplest
possible expression. The exact list of operations performed depend on the implementation.
Obviously, the denition of the canonical form is arbitrary, the only requirement is that all
equivalent expressions must have the same canonical form. We tried the conversion to a
canonical (standard) form to be as fast as possible and also in a way so that the result is what
you would write by hand - so for example b*a + -4 + b + a*b + 4 + (a + b)**2 becomes
2*a*b + b + (a + b)**2.
Whenever you construct an expression, for example Add(x, x), the Add. new () is called
and it determines what to return. In this case:
>>>
>>>
>>>
>>>
2*x
from sympy import Add
from sympy.abc import x
e = Add(x, x)
e
>>> type(e)
<class sympy.core.mul.Mul>
e is actually an instance of Mul(2, x), because Add. new () returned Mul.
4.3.3 Comparisons
Expressions can be compared using a regular python syntax:
>>> from sympy.abc import x, y
>>> x + y == y + x
True
>>> x + y == y - x
False
We made the following decision in SymPy: a = Symbol(x) and another b = Symbol(x)
(with the same string x) is the same thing, i.e a == b is True. We chose a == b, because
it is more natural - exp(x) == exp(x) is also True for the same instance of x but dierent
instances of exp, so we chose to have exp(x) == exp(x) even for dierent instances of x.
78
Chapter 4. SymPy Users Guide
SymPy Documentation, Release 0.7.6
Sometimes, you need to have a unique symbol, for example as a temporary one in some
calculation, which is going to be substituted for something else at the end anyway. This is
achieved using Dummy(x). So, to sum it up:
>>> from sympy import Symbol, Dummy
>>> Symbol(x) == Symbol(x)
True
>>> Dummy(x) == Dummy(x)
False
4.3.4 Debugging
Starting with 0.6.4, you can turn on/o debug messages with the environment variable
SYMPY DEBUG, which is expected to have the values True or False. For example, to turn on
debugging, you would issue:
[user@localhost]: SYMPY DEBUG=True ./bin/isympy
4.3.5 Functionality
There are no given requirements on classes in the library. For example, if they dont implement the fdiff() method and you construct an expression using such a class, then trying to
use the Basic.series() method will raise an exception of not nding the fdiff() method
in your class. This duck typing has an advantage that you just implement the functionality
which you need.
You can dene the cos class like this:
class cos(Function):
pass
and use it like 1 + cos(x), but if you dont implement the fdiff() method, you will not be
able to call (1 + cos(x)).series().
The symbolic object is characterized (dened) by the things which it can do, so implementing
more methods like fdiff(), subs() etc., you are creating a shape of the symbolic object.
Useful things to implement in new classes are: hash() (to use the class in comparisons),
fdiff() (to use it in series expansion), subs() (to use it in expressions, where some parts
are being substituted) and series() (if the series cannot be computed using the general
Basic.series() method). When you create a new class, dont worry about this too much
- just try to use it in your code and you will realize immediately which methods need to be
implemented in each situation.
All objects in sympy are immutable - in the sense that any operation just returns a new instance
(it can return the same instance only if it didnt change). This is a common mistake to change
the current instance, like self.arg = self.arg + 1 (wrong!). Use arg = self.arg +
1; return arg instead. The object is immutable in the sense of the symbolic expression it
represents. It can modify itself to keep track of, for example, its hash. Or it can recalculate
anything regarding the expression it contains. But the expression cannot be changed. So you
can pass any instance to other objects, because you dont have to worry that it will change,
or that this would break anything.
4.3. SymPys Architecture
79
SymPy Documentation, Release 0.7.6
4.3.6 Conclusion
Above are the main ideas behind SymPy that we try to obey. The rest depends on the current
implementation and may possibly change in the future. The point of all of this is that the
interdependencies inside SymPy should be kept to a minimum. If one wants to add new
functionality to SymPy, all that is necessary is to create a subclass of Basic and implement
what you want.
4.3.7 Functions
How to create a new function with one variable:
class sign(Function):
nargs = 1
@classmethod
def eval(cls, arg):
if isinstance(arg, Basic.NaN):
return S.NaN
if isinstance(arg, Basic.Zero):
return S.Zero
if arg.is_positive:
return S.One
if arg.is_negative:
return S.NegativeOne
if isinstance(arg, Basic.Mul):
coeff, terms = arg.as_coeff_mul()
if not isinstance(coeff, Basic.One):
return cls(coeff) * cls(Basic.Mul(*terms))
is_finite = True
def _eval_conjugate(self):
return self
def _eval_is_zero(self):
return isinstance(self[0], Basic.Zero)
and thats it. The eval * functions are called when something is needed. The eval is
called when the class is about to be instantiated and it should return either some simplied instance of some other class or if the class should be unmodied, return None
(see core/function.py in Function. new for implementation details). See also tests in
sympy/functions/elementary/tests/test interface.py that test this interface. You can use them
to create your own new functions.
The applied function sign(x) is constructed using
sign(x)
both inside and outside of SymPy. Unapplied functions sign is just the class itself:
sign
both inside and outside of SymPy. This is the current structure of classes in SymPy:
80
Chapter 4. SymPy Users Guide
SymPy Documentation, Release 0.7.6
class BasicType(type):
pass
class MetaBasicMeths(BasicType):
...
class BasicMeths(AssumeMeths):
__metaclass__ = MetaBasicMeths
...
class Basic(BasicMeths):
...
class FunctionClass(MetaBasicMeths):
...
class Function(Basic, RelMeths, ArithMeths):
__metaclass__ = FunctionClass
...
The exact names of the classes and the names of the methods and how they work can be
changed in the future.
This is how to create a function with two variables:
class chebyshevt_root(Function):
nargs = 2
@classmethod
def eval(cls, n, k):
if not 0 <= k < n:
raise ValueError(must have 0 <= k < n)
return C.cos(S.Pi*(2*k + 1)/(2*n))
Note: the rst argument of a @classmethod should be cls (i.e. not self).
Here its how to dene a derivative of the function:
>>> from sympy import Function, sympify, cos
>>> class my_function(Function):
...
nargs = 1
...
...
def fdiff(self, argindex = 1):
...
return cos(self.args[0])
...
...
@classmethod
...
def eval(cls, arg):
...
arg = sympify(arg)
...
if arg == 0:
...
return sympify(0)
So guess what this my function is going to be? Well, its derivative is cos and the function
value at 0 is 0, but lets pretend we dont know:
>>> from sympy import pprint
>>> pprint(my_function(x).series(x, 0, 10))
3
5
7
9
x
x
x
x
/ 10\
x - -- + --- - ---- + ------ + O\x /
6
120
5040
362880
Looks familiar indeed:
4.3. SymPys Architecture
81
SymPy Documentation, Release 0.7.6
>>> from sympy import sin
>>> pprint(sin(x).series(x, 0, 10))
3
5
7
9
x
x
x
x
/ 10\
x - -- + --- - ---- + ------ + O\x /
6
120
5040
362880
Lets try a more complicated example. Lets dene the derivative in terms of the function
itself:
>>> class what_am_i(Function):
...
nargs = 1
...
...
def fdiff(self, argindex = 1):
...
return 1 - what_am_i(self.args[0])**2
...
...
@classmethod
...
def eval(cls, arg):
...
arg = sympify(arg)
...
if arg == 0:
...
return sympify(0)
So what is what am i? Lets try it:
>>> pprint(what_am_i(x).series(x, 0, 10))
3
5
7
9
x
2*x
17*x
62*x
/ 10\
x - -- + ---- - ----- + ----- + O\x /
3
15
315
2835
Well, its tanh:
>>> from sympy import tanh
>>> pprint(tanh(x).series(x, 0, 10))
3
5
7
9
x
2*x
17*x
62*x
/ 10\
x - -- + ---- - ----- + ----- + O\x /
3
15
315
2835
The new functions we just dened are regular SymPy objects, you can use them all over
SymPy, e.g.:
>>> from sympy import limit
>>> limit(what_am_i(x)/x, x, 0)
1
4.3.8 Common tasks
Please use the same way as is shown below all across SymPy.
accessing parameters:
>>> from sympy import sign, sin
>>> from sympy.abc import x, y, z
>>> e = sign(x**2)
>>> e.args
(x**2,)
82
Chapter 4. SymPy Users Guide
SymPy Documentation, Release 0.7.6
>>> e.args[0]
x**2
Number arguments (in Adds and Muls) will always be the first argument;
other arguments might be in arbitrary order:
>>> (1 + x + y*z).args[0]
1
>>> (1 + x + y*z).args[1] in (x, y*z)
True
>>> (y*z).args
(y, z)
>>> sin(y*z).args
(y*z,)
Never use internal methods or variables, prexed with (example: dont use args, use
.args instead).
testing the structure of a SymPy expression
Applied functions:
>>> from sympy import sign, exp, Function
>>> e = sign(x**2)
>>> isinstance(e, sign)
True
>>> isinstance(e, exp)
False
>>> isinstance(e, Function)
True
So e is a sign(z) function, but not exp(z) function.
Unapplied functions:
>>> from sympy import sign, exp, FunctionClass
>>> e = sign
>>> f = exp
>>> g = Add
>>> isinstance(e, FunctionClass)
True
>>> isinstance(f, FunctionClass)
True
>>> isinstance(g, FunctionClass)
False
>>> g is Add
True
So e and f are functions, g is not a function.
4.3. SymPys Architecture
83
SymPy Documentation, Release 0.7.6
4.4 Contributing
We welcome every SymPy user to participate in its development. Dont worry if youve never
contributed to any open source project, well help you learn anything necessary, just ask on
our mailinglist.
Dont be afraid to ask anything and dont worry that you are wasting our time if you are new
to SymPy and ask questions that maybe most of the people know the answer to you are
not, because thats exactly what the mailinglist is for and people answer your emails because
they want to. Also we try hard to answer every email, so youll always get some feedback and
pointers what to do next.
4.4.1 Improving the code
Go to issues that are sorted by priority and simply nd something that you would like to get
xed and x it. If you nd something odd, please report it into issues rst before xing it.
Feel free to consult with us on the mailinglist. Then send your patch either to the issues or
the mailinglist.
Please read our excellent SymPy Patches Tutorial at our wiki for a guide on how to write
patches to SymPy, how to work with Git, and how to make your life easier as you get started
with SymPy.
4.4.2 Improving the docs
Please see the documentation (page 85) how to x and improve SymPys documentation. All
contribution is very welcome.
84
Chapter 4. SymPy Users Guide
CHAPTER
FIVE
SYMPY MODULES REFERENCE
Because every feature of SymPy must have a test case, when you are not sure how to use
something, just look into the tests/ directories, nd that feature and read the tests for it,
that will tell you everything you need to know.
Most of the things are already documented though in this document, that is automatically
generated using SymPys docstrings.
Click the modules (modindex) link in the top right corner to easily access any SymPy module,
or use this contens:
5.1 SymPy Core
5.1.1 sympify
sympify
sympy.core.sympify.sympify(a, locals=None, convert xor=True, strict=False, rational=False, evaluate=None)
Converts an arbitrary expression to a type that can be used inside SymPy.
For example, it will convert Python ints into instance of sympy.Rational, oats into instances of sympy.Float, etc. It is also able to coerce symbolic expressions which inherit
from Basic. This can be useful in cooperation with SAGE.
It currently accepts as arguments:
any object dened in sympy
standard numeric python types: int, long, oat, Decimal
strings (like 0.09 or 2e-19)
booleans, including None (will leave None unchanged)
lists, sets or tuples containing any of the above
If the argument is already a type that SymPy understands, it will do nothing but return
that value. This can be used at the beginning of a function to ensure you are working
with the correct type.
>>> from sympy import sympify
85
SymPy Documentation, Release 0.7.6
>>> sympify(2).is_integer
True
>>> sympify(2).is_real
True
>>> sympify(2.0).is_real
True
>>> sympify(2.0).is_real
True
>>> sympify(2e-45).is_real
True
If the expression could not be converted, a SympifyError is raised.
>>> sympify(x***2)
Traceback (most recent call
...
SympifyError: SympifyError:
Traceback (most recent call
...
SympifyError: SympifyError:
last):
could not parse ux***2
last):
could not parse ux***2
Notes
Sometimes autosimplication during sympication results in expressions that are very
dierent in structure than what was entered. Until such autosimplication is no longer
done, the kernS function might be of some use. In the example below you can see how
an expression reduces to -1 by autosimplication, but does not do so when kernS is used.
>>> from sympy.core.sympify import kernS
>>> from sympy.abc import x
>>> -2*(-(-x + 1/x)/(x*(x - 1/x)**2) - 1/(x*(x - 1/x))) - 1
-1
>>> s = -2*(-(-x + 1/x)/(x*(x - 1/x)**2) - 1/(x*(x - 1/x))) - 1
>>> sympify(s)
-1
>>> kernS(s)
-2*(-(-x + 1/x)/(x*(x - 1/x)**2) - 1/(x*(x - 1/x))) - 1
Locals
The sympication happens with access to everything that is loaded by from sympy import *; anything used in a string that is not dened by that import will be converted
to a symbol. In the following, the bitcount function is treated as a symbol and the O
is interpreted as the Order object (used with series) and it raises an error when used
improperly:
>>> s = bitcount(42)
>>> sympify(s)
bitcount(42)
>>> sympify(O(x))
O(x)
>>> sympify(O + 1)
Traceback (most recent call last):
...
86
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
TypeError: unbound method...
Traceback (most recent call last):
...
TypeError: unbound method...
In order to have bitcount be recognized it can be imported into a namespace dictionary
and passed as locals:
>>>
>>>
>>>
>>>
6
from sympy.core.compatibility import exec_
ns = {}
exec_(from sympy.core.evalf import bitcount, ns)
sympify(s, locals=ns)
In order to have the O interpreted as a Symbol, identify it as such in the namespace dictionary. This can be done in a variety of ways; all three of the following are possibilities:
>>>
>>>
>>>
>>>
>>>
O +
from sympy import Symbol
ns[O] = Symbol(O) # method 1
exec_(from sympy.abc import O, ns) # method 2
ns.update(dict(O=Symbol(O))) # method 3
sympify(O + 1, locals=ns)
1
If you want all single-letter and Greek-letter variables to be symbols then you can use the
clashing-symbols dictionaries that have been dened there as private variables: clash1
(single-letter variables), clash2 (the multi-letter Greek names) or clash (both single and
multi-letter names that are dened in abc).
>>> from sympy.abc import _clash1
>>> _clash1
{C: C, E: E, I: I, N: N, O: O, Q: Q, S: S}
>>> sympify(C & Q, _clash1)
And(C, Q)
Strict
If the option strict is set to True, only the types for which an explicit conversion has
been dened are converted. In the other cases, a SympifyError is raised.
>>> print(sympify(None))
None
>>> sympify(None, strict=True)
Traceback (most recent call last):
...
SympifyError: SympifyError: None
Traceback (most recent call last):
...
SympifyError: SympifyError: None
Evaluation
If the option evaluate is set to False, then arithmetic and operators will be converted
into their SymPy equivalents and the evaluate=False option will be added. Nested Add
5.1. SymPy Core
87
SymPy Documentation, Release 0.7.6
or Mul will be denested rst. This is done via an AST transformation that replaces operators with their SymPy equivalents, so if an operand redenes any of those operations,
the redened operators will not be used.
>>> sympify(2**2 / 3 + 5)
19/3
>>> sympify(2**2 / 3 + 5, evaluate=False)
2**2/3 + 5
Extending
To extend sympify to convert custom objects (not derived from Basic), just dene a
sympy method to your class. You can do that even to classes that you do not own by
subclassing or adding the method at runtime.
>>> from sympy import Matrix
>>> class MyList1(object):
...
def __iter__(self):
...
yield 1
...
yield 2
...
raise StopIteration
...
def __getitem__(self, i): return list(self)[i]
...
def _sympy_(self): return Matrix(self)
>>> sympify(MyList1())
Matrix([
[1],
[2]])
If you do not have control over the class denition you could also use the converter
global dictionary. The key is the class and the value is a function that takes a single
argument and returns the desired SymPy object, e.g. converter[MyList] = lambda x:
Matrix(x).
>>> class MyList2(object):
# XXX Do not do this if you control the class!
...
def __iter__(self): #
Use _sympy_!
...
yield 1
...
yield 2
...
raise StopIteration
...
def __getitem__(self, i): return list(self)[i]
>>> from sympy.core.sympify import converter
>>> converter[MyList2] = lambda x: Matrix(x)
>>> sympify(MyList2())
Matrix([
[1],
[2]])
5.1.2 assumptions
This module contains the machinery handling assumptions.
All symbolic objects have assumption attributes that can be accessed via .is <assumption
name> attribute.
Assumptions determine certain properties of symbolic objects. Assumptions can have 3 possible values: True, False, None. None is returned when it is impossible to say something about
the property. For example, a generic Symbol is not known beforehand to be positive.
88
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
By default, all symbolic values are in the largest set in the given context without specifying
the property. For example, a symbol that has a property being integer, is also real, complex,
etc.
Here follows a list of possible assumption names:
commutative object commutes with any other object with respect to multiplication operation.
complex object can have only values from the set of complex numbers.
imaginary object value is a number that can be written as a real number multiplied by the
imaginary unit I. See [R26] (page 1899). Please note, that 0 is not considered to be an
imaginary number, see issue #7649.
real object can have only values from the set of real numbers.
integer object can have only values from the set of integers.
odd, even object can have only values from the set of odd (even) integers [R25] (page 1899).
prime object is a natural number greater than 1 that has no positive divisors other than 1
and itself. See [R29] (page 1899).
composite object is a positive integer that has at least one positive divisor other than 1 or
the number itself. See [R27] (page 1899).
zero, nonzero object is zero (not zero).
rational object can have only values from the set of rationals.
algebraic object can have only values from the set of algebraic numbers 1 .
transcendental object can have only values from the set of transcendental numbers 2 .
irrational object value cannot be represented exactly by Rational, see [R28] (page 1899).
nite, innite object absolute value is bounded (is value is arbitrarily large). See [R30]
(page 1899), [R31] (page 1899), [R32] (page 1899).
negative, nonnegative
(page 1899).
object can have only negative (only nonnegative) values [R24]
positive, nonpositive object can have only positive (only nonpositive) values.
hermitian, antihermitian object belongs to the eld of hermitian (antihermitian) operators.
Examples
>>> from sympy import Symbol
>>> x = Symbol(x, real = True); x
x
>>> x.is_real
True
>>> x.is_complex
True
1
2
http://en.wikipedia.org/wiki/Algebraic number
http://en.wikipedia.org/wiki/Transcendental number
5.1. SymPy Core
89
SymPy Documentation, Release 0.7.6
See Also
See Also:
sympy.core.numbers.ImaginaryUnit (page 137) sympy.core.numbers.Zero (page 132)
sympy.core.numbers.One (page 133)
Notes
Assumption values are stored in obj. assumptions dictionary or are returned by getter methods (with property decorators) or are attributes of objects/classes.
References
5.1.3 cache
cacheit
sympy.core.cache.cacheit(func)
5.1.4 basic
Basic
class sympy.core.basic.Basic
Base class for all objects in SymPy.
Conventions:
1.Always use .args, when accessing parameters of some instance:
>>> from sympy import cot
>>> from sympy.abc import x, y
>>> cot(x).args
(x,)
>>> cot(x).args[0]
x
>>> (x*y).args
(x, y)
>>> (x*y).args[1]
y
2.Never use internal methods or variables (the ones prexed with ):
>>> cot(x)._args
(x,)
# do not use this, use cot(x).args instead
args
Returns a tuple of arguments of self.
90
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Notes
Never use self. args, always use self.args. Only use args in new when creating
a new function. Dont override .args() from Basic (so that its easy to change the
interface in the future if needed).
Examples
>>> from sympy import cot
>>> from sympy.abc import x, y
>>> cot(x).args
(x,)
>>> cot(x).args[0]
x
>>> (x*y).args
(x, y)
>>> (x*y).args[1]
y
as content primitive(radical=False)
A stub to allow Basic args (like Tuple) to be skipped when computing the content
and primitive components of an expression.
See docstring of Expr.as content primitive
as poly(*gens, **args)
Converts self to a polynomial or returns None.
>>> from sympy import sin
>>> from sympy.abc import x, y
>>> print((x**2 + x*y).as_poly())
Poly(x**2 + x*y, x, y, domain=ZZ)
>>> print((x**2 + x*y).as_poly(x, y))
Poly(x**2 + x*y, x, y, domain=ZZ)
>>> print((x**2 + sin(y)).as_poly(x, y))
None
assumptions0
Return object type assumptions.
For example:
Symbol(x, real=True) Symbol(x, integer=True)
are dierent objects. In other words, besides Python type (Symbol in this case), the
initial assumptions are also forming their typeinfo.
5.1. SymPy Core
91
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import Symbol
>>> from sympy.abc import x
>>> x.assumptions0
{commutative: True}
>>> x = Symbol(x, positive=True)
>>> x.assumptions0
{commutative: True, complex: True, hermitian: True,
imaginary: False, negative: False, nonnegative: True,
nonpositive: False, nonzero: True, positive: True, real: True,
zero: False}
atoms(*types)
Returns the atoms that form the current object.
By default, only objects that are truly atomic and cant be divided into smaller pieces
are returned: symbols, numbers, and number symbols like I and pi. It is possible to
request atoms of any type, however, as demonstrated below.
Examples
>>> from sympy import Number, NumberSymbol, Symbol
>>> (1 + x + 2*sin(y + I*pi)).atoms(Symbol)
set([x, y])
>>> (1 + x + 2*sin(y + I*pi)).atoms(Number)
set([1, 2])
>>> (1 + x + 2*sin(y + I*pi)).atoms(Number, NumberSymbol)
set([1, 2, pi])
>>> (1 + x + 2*sin(y + I*pi)).atoms(Number, NumberSymbol, I)
set([1, 2, I, pi])
Note that I (imaginary unit) and zoo (complex innity) are special types of number
symbols and are not part of the NumberSymbol class.
The type can be given implicitly, too:
>>> (1 + x + 2*sin(y + I*pi)).atoms(x) # x is a Symbol
set([x, y])
Be careful to check your assumptions when using the implicit option since
S(1).is Integer = True but type(S(1)) is One, a special type of sympy atom,
while type(S(2)) is type Integer and will nd all integers in an expression:
>>> from sympy import S
>>> (1 + x + 2*sin(y + I*pi)).atoms(S(1))
set([1])
>>> (1 + x + 2*sin(y + I*pi)).atoms(S(2))
set([1, 2])
Finally, arguments to atoms() can select more than atomic atoms: any sympy type
(loaded in core/ init .py) can be listed as an argument and those types of atoms
as found in scanning the arguments of the expression recursively:
92
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> from sympy import Function, Mul
>>> from sympy.core.function import AppliedUndef
>>> f = Function(f)
>>> (1 + f(x) + 2*sin(y + I*pi)).atoms(Function)
set([f(x), sin(y + I*pi)])
>>> (1 + f(x) + 2*sin(y + I*pi)).atoms(AppliedUndef)
set([f(x)])
>>> (1 + x + 2*sin(y + I*pi)).atoms(Mul)
set([I*pi, 2*sin(y + I*pi)])
canonical variables
Return a dictionary mapping any variable dened in self.variables as underscoresuxed numbers corresponding to their position in self.variables. Enough underscores are added to ensure that there will be no clash with existing free symbols.
Examples
>>>
>>>
>>>
{x:
from sympy import Lambda
from sympy.abc import x
Lambda(x, 2*x).canonical_variables
0_}
classmethod class key()
Nice order of classes.
compare(other)
Return -1, 0, 1 if the object is smaller, equal, or greater than other.
Not in the mathematical sense. If the object is of a dierent type from the other
then their classes are ordered according to the sorted classes list.
Examples
>>>
>>>
-1
>>>
0
>>>
1
from sympy.abc import x, y
x.compare(y)
x.compare(x)
y.compare(x)
count(query)
Count the number of matching subexpressions.
count ops(visual=None)
wrapper for count ops that returns the operation count.
doit(**hints)
Evaluate objects that are not evaluated by default like limits, integrals, sums and
products. All objects of this kind will be evaluated recursively, unless some species
were excluded via hints or unless the deep hint was set to False.
>>> from sympy import Integral
>>> from sympy.abc import x
5.1. SymPy Core
93
SymPy Documentation, Release 0.7.6
>>> 2*Integral(x, x)
2*Integral(x, x)
>>> (2*Integral(x, x)).doit()
x**2
>>> (2*Integral(x, x)).doit(deep = False)
2*Integral(x, x)
dummy eq(other, symbol=None)
Compare two expressions and handle dummy symbols.
Examples
>>> from sympy import Dummy
>>> from sympy.abc import x, y
>>> u = Dummy(u)
>>> (u**2 + 1).dummy_eq(x**2 + 1)
True
>>> (u**2 + 1) == (x**2 + 1)
False
>>> (u**2 + y).dummy_eq(x**2 + y, x)
True
>>> (u**2 + y).dummy_eq(x**2 + y, y)
False
find(query, group=False)
Find all subexpressions matching a query.
free symbols
Return from the atoms of self those which are free symbols.
For most expressions, all symbols are free symbols. For some classes this is not true.
e.g. Integrals use Symbols for the dummy variables which are bound variables, so
Integral has a method to return all symbols except those. Derivative keeps track of
symbols with respect to which it will perform a derivative; those are bound variables,
too, so it has its own symbols method.
Any other method that uses bound variables should implement a symbols method.
classmethod fromiter(args, **assumptions)
Create a new object from an iterable.
This is a convenience function that allows one to create objects from any iterable,
without having to convert to a list or tuple rst.
Examples
>>> from sympy import Tuple
>>> Tuple.fromiter(i for i in range(5))
(0, 1, 2, 3, 4)
94
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
func
The top-level function in an expression.
The following should hold for all objects:
>> x == x.func(*x.args)
Examples
>>> from sympy.abc import x
>>> a = 2*x
>>> a.func
<class sympy.core.mul.Mul>
>>> a.args
(2, x)
>>> a.func(*a.args)
2*x
>>> a == a.func(*a.args)
True
has(*args, **kwargs)
Test whether any subexpression matches any of the patterns.
Examples
>>> from sympy import sin
>>> from sympy.abc import x, y, z
>>> (x**2 + sin(x*y)).has(z)
False
>>> (x**2 + sin(x*y)).has(x, y, z)
True
>>> x.has(x)
True
Note that expr.has(*patterns) is exactly equivalent to any(expr.has(p) for p
in patterns). In particular, False is returned when the list of patterns is empty.
>>> x.has()
False
is comparable
Return True if self can be computed to a real number with precision, else False.
Examples
>>> from sympy import exp_polar, pi, I
>>> (I*exp_polar(I*pi/2)).is_comparable
True
>>> (I*exp_polar(I*pi*2)).is_comparable
False
iter basic args(*args, **kwargs)
Iterates arguments of self.
5.1. SymPy Core
95
SymPy Documentation, Release 0.7.6
match(pattern, old=False)
Pattern matching.
Wild symbols match all.
Return None when expression (self) does not match with pattern. Otherwise return
a dictionary such that:
pattern.xreplace(self.match(pattern)) == self
Examples
>>> from sympy import Wild
>>> from sympy.abc import x, y
>>> p = Wild(p)
>>> q = Wild(q)
>>> r = Wild(r)
>>> e = (x+y)**(x+y)
>>> e.match(p**p)
{p_: x + y}
>>> e.match(p**q)
{p_: x + y, q_: x + y}
>>> e = (2*x)**2
>>> e.match(p*q**r)
{p_: 4, q_: x, r_: 2}
>>> (p*q**r).xreplace(e.match(p*q**r))
4*x**2
The old ag will give the old-style pattern matching where expressions and patterns
are essentially solved to give the match. Both of the following give None unless
old=True:
>>> (x - 2).match(p - x, old=True)
{p_: 2*x - 2}
>>> (2/x).match(p*x, old=True)
{p_: 2/x**2}
matches(expr, repl dict={}, old=False)
Helper method for match() that looks for a match between Wild symbols in self and
expressions in expr.
Examples
>>> from sympy import symbols, Wild, Basic
>>> a, b, c = symbols(a b c)
>>> x = Wild(x)
>>> Basic(a + x, x).matches(Basic(a + b, c)) is None
True
>>> Basic(a + x, x).matches(Basic(a + b + c, b + c))
{x_: b + c}
rcall(*args)
Apply on the argument recursively through the expression tree.
This method is used to simulate a common abuse of notation for operators. For
instance in SymPy the the following will not work:
96
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
(x+Lambda(y, 2*y))(z) == x+2*z,
however you can use
>>>
>>>
>>>
x +
from sympy import Lambda
from sympy.abc import x,y,z
(x + Lambda(y, 2*y)).rcall(z)
2*z
replace(query, value, map=False, simultaneous=True, exact=False)
Replace matching subexpressions of self with value.
If map = True then also return the mapping {old: new} where old was a subexpression found with query and new is the replacement value for it. If the expression
itself doesnt match the query, then the returned value will be self.xreplace(map)
otherwise it should be self.subs(ordered(map.items())).
Traverses an expression tree and performs replacement of matching subexpressions
from the bottom to the top of the tree. The default approach is to do the replacement in a simultaneous fashion so changes made are targeted only once. If this is
not desired or causes problems, simultaneous can be set to False. In addition, if
an expression containing more than one Wild symbol is being used to match subexpressions and the exact ag is True, then the match will only succeed if non-zero
values are received for each Wild that appears in the match pattern.
The list of possible combinations of queries and replacement values is listed below:
See Also:
subs (page 99) substitution of subexpressions as dened by the objects themselves.
xreplace (page 101) exact node replacement in expr tree; also capable of using
matching rules
Examples
Initial setup
>>> from sympy import log, sin, cos, tan, Wild, Mul, Add
>>> from sympy.abc import x, y
>>> f = log(sin(x)) + tan(sin(x**2))
1.1. type -> type obj.replace(type, newtype)
When object of type type is found, replace it with the result of passing its argument(s) to newtype.
>>> f.replace(sin, cos)
log(cos(x)) + tan(cos(x**2))
>>> sin(x).replace(sin, cos, map=True)
(cos(x), {sin(x): cos(x)})
>>> (x*y).replace(Mul, Add)
x + y
1.2. type -> func obj.replace(type, func)
When object of type type is found, apply func to its argument(s). func must be
written to handle the number of arguments of type.
5.1. SymPy Core
97
SymPy Documentation, Release 0.7.6
>>> f.replace(sin, lambda arg: sin(2*arg))
log(sin(2*x)) + tan(sin(2*x**2))
>>> (x*y).replace(Mul, lambda *args: sin(2*Mul(*args)))
sin(2*x*y)
2.1. pattern -> expr obj.replace(pattern(wild), expr(wild))
Replace subexpressions matching pattern with the expression written in terms
of the Wild symbols in pattern.
>>> a = Wild(a)
>>> f.replace(sin(a), tan(a))
log(tan(x)) + tan(tan(x**2))
>>> f.replace(sin(a), tan(a/2))
log(tan(x/2)) + tan(tan(x**2/2))
>>> f.replace(sin(a), a)
log(x) + tan(x**2)
>>> (x*y).replace(a*x, a)
y
When the default value of False is used with patterns that have more than one
Wild symbol, non-intuitive results may be obtained:
>>> b = Wild(b)
>>> (2*x).replace(a*x + b, b - a)
2/x
For this reason, the exact option can be used to make the replacement only
when the match gives non-zero values for all Wild symbols:
>>> (2*x + y).replace(a*x + b, b - a, exact=True)
y - 2
>>> (2*x).replace(a*x + b, b - a, exact=True)
2*x
2.2. pattern -> func obj.replace(pattern(wild), lambda wild: expr(wild))
All behavior is the same as in 2.1 but now a function in terms of pattern variables
is used rather than an expression:
>>> f.replace(sin(a), lambda a: sin(2*a))
log(sin(2*x)) + tan(sin(2*x**2))
3.1. func -> func obj.replace(lter, func)
Replace subexpression e with func(e) if filter(e) is True.
>>> g = 2*sin(x**3)
>>> g.replace(lambda expr: expr.is_Number, lambda expr: expr**2)
4*sin(x**9)
The expression itself is also targeted by the query but is done in such a fashion that
changes are not made twice.
>>> e = x*(x*y + 1)
>>> e.replace(lambda x: x.is_Mul, lambda x: 2*x)
2*x*(2*x*y + 1)
98
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
rewrite(*args, **hints)
Rewrite functions in terms of other functions.
Rewrites expression containing applications of functions of one kind in terms of
functions of dierent kind. For example you can rewrite trigonometric functions as
complex exponentials or combinatorial functions as gamma function.
As a pattern this function accepts a list of functions to to rewrite (instances of DenedFunction class). As rule you can use string or a destination function instance
(in this case rewrite() will use the str() function).
There is also the possibility to pass hints on how to rewrite the given expressions.
For now there is only one such hint dened called deep. When deep is set to False
it will forbid functions to rewrite their contents.
Examples
>>> from sympy import sin, exp
>>> from sympy.abc import x
Unspecied pattern: >>> sin(x).rewrite(exp) -I*(exp(I*x) - exp(-I*x))/2
Pattern as a single function: >>> sin(x).rewrite(sin, exp) -I*(exp(I*x) - exp(-I*x))/2
Pattern as a list of functions: >>> sin(x).rewrite([sin, ], exp) -I*(exp(I*x) - exp(-I*x))/2
sort key(*args, **kwargs)
Return a sort key.
Examples
>>> from sympy.core import S, I
>>> sorted([S(1)/2, I, -I], key=lambda x: x.sort_key())
[1/2, -I, I]
>>> S([x, 1/x, 1/x**2, x**2, x**(1/2), x**(1/4), x**(3/2)])
[x, 1/x, x**(-2), x**2, sqrt(x), x**(1/4), x**(3/2)]
>>> sorted(_, key=lambda x: x.sort_key())
[x**(-2), 1/x, x**(1/4), sqrt(x), x, x**(3/2), x**2]
subs(*args, **kwargs)
Substitutes old for new in an expression after sympifying args.
args is either:
two arguments, e.g. foo.subs(old, new)
one iterable argument, e.g. foo.subs(iterable). The iterable may be
o an iterable container with (old, new) pairs. In this case the
replacements are processed in the order given with successive patterns
possibly aecting replacements already made.
o a dict or set whose key/value items correspond to old/new pairs.
In this case the old/new pairs will be sorted by op count and in case of a
tie, by number of args and the default sort key. The resulting sorted list
is then processed as an iterable container (see previous).
5.1. SymPy Core
99
SymPy Documentation, Release 0.7.6
If the keyword simultaneous is True, the subexpressions will not be evaluated until
all the substitutions have been made.
See Also:
replace (page 97) replacement capable of doing wildcard-like matching, parsing
of match, and conditional replacements
xreplace (page 101) exact node replacement in expr tree; also capable of using
matching rules
evalf calculates the given formula to a desired level of precision
Examples
>>> from
>>> from
>>> (1 +
pi*y + 1
>>> (1 +
1 + 2*pi
>>> (1 +
1 + 2*pi
>>> reps
>>> (x +
6
>>> (x +
x**2 + 2
sympy import pi, exp, limit, oo
sympy.abc import x, y
x*y).subs(x, pi)
x*y).subs({x:pi, y:2})
x*y).subs([(x, pi), (y, 2)])
= [(y, x**2), (x, 2)]
y).subs(reps)
y).subs(reversed(reps))
>>> (x**2 + x**4).subs(x**2, y)
y**2 + y
To replace only the x**2 but not the x**4, use xreplace:
>>> (x**2 + x**4).xreplace({x**2: y})
x**4 + y
To delay evaluation until all substitutions have been made, set the keyword simultaneous to True:
>>> (x/y).subs([(x, 0), (y, 0)])
0
>>> (x/y).subs([(x, 0), (y, 0)], simultaneous=True)
nan
This has the added feature of not allowing subsequent substitutions to aect those
already made:
>>> ((x + y)/y).subs({x + y: y, y: x + y})
1
>>> ((x + y)/y).subs({x + y: y, y: x + y}, simultaneous=True)
y/(x + y)
In order to obtain a canonical result, unordered iterables are sorted by count op
length, number of arguments and by the default sort key to break any ties. All other
iterables are left unsorted.
>>> from sympy import sqrt, sin, cos
>>> from sympy.abc import a, b, c, d, e
100
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>>
>>>
>>>
>>>
>>>
A
B
C
D
E
=
=
=
=
=
(sqrt(sin(2*x)), a)
(sin(2*x), b)
(cos(2*x), c)
(x, d)
(exp(x), e)
>>> expr = sqrt(sin(2*x))*sin(exp(x)*x)*cos(2*x) + sin(2*x)
>>> expr.subs(dict([A,B,C,D,E]))
a*c*sin(d*e) + b
The resulting expression represents a literal replacement of the old arguments with
the new arguments. This may not reect the limiting behavior of the expression:
>>> (x**3 - 3*x).subs({x: oo})
nan
>>> limit(x**3 - 3*x, x, oo)
oo
If the substitution will be followed by numerical evaluation, it is better to pass the
substitution to evalf as
>>> (1/x).evalf(subs={x: 3.0}, n=21)
0.333333333333333333333
rather than
>>> (1/x).subs({x: 3.0}).evalf(21)
0.333333333333333314830
as the former will ensure that the desired level of precision is obtained.
xreplace(rule)
Replace occurrences of objects within the expression.
Parameters rule : dict-like
Expresses a replacement rule
Returns xreplace : the result of the replacement
See Also:
replace (page 97) replacement capable of doing wildcard-like matching, parsing
of match, and conditional replacements
subs (page 99) substitution of subexpressions as dened by the objects themselves.
Examples
>>> from sympy import symbols, pi, exp
>>> x, y, z = symbols(x y z)
>>> (1 + x*y).xreplace({x: pi})
pi*y + 1
>>> (1 + x*y).xreplace({x:pi, y:2})
1 + 2*pi
5.1. SymPy Core
101
SymPy Documentation, Release 0.7.6
Replacements occur only if an entire node in the expression tree is matched:
>>> (x*y + z).xreplace({x*y: pi})
z + pi
>>> (x*y*z).xreplace({x*y: pi})
x*y*z
>>> (2*x).xreplace({2*x: y, x: z})
y
>>> (2*2*x).xreplace({2*x: y, x: z})
4*z
>>> (x + y + 2).xreplace({x + y: 2})
x + y + 2
>>> (x + 2 + exp(x + 2)).xreplace({x + 2: y})
x + exp(y) + 2
xreplace doesnt dierentiate between free and bound symbols. In the following,
subs(x, y) would not change x since it is a bound symbol, but xreplace does:
>>> from sympy import Integral
>>> Integral(x, (x, 1, 2*x)).xreplace({x: y})
Integral(y, (y, 1, 2*y))
Trying to replace x with an expression raises an error:
>>> Integral(x, (x, 1, 2*x)).xreplace({x: 2*y})
ValueError: Invalid limits given: ((2*y, 1, 4*y),)
Atom
class sympy.core.basic.Atom
A parent class for atomic things. An atom is an expression with no subexpressions.
Examples
Symbol, Number, Rational, Integer, ... But not: Add, Mul, Pow, ...
C
sympy.core.basic.C
5.1.5 singleton
S
sympy.core.singleton.S
102
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
5.1.6 expr
5.1.7 Expr
class sympy.core.expr.Expr
Base class for algebraic expressions.
Everything that requires arithmetic operations to be dened should subclass this class,
instead of Basic (which should be used only for argument storage and expression manipulation, i.e. pattern matching, substitutions, etc).
See Also:
sympy.core.basic.Basic (page 90)
apart(x=None, **args)
See the apart function in sympy.polys
args cnc(cset=False, warn=True, split 1=True)
Return [commutative factors, non-commutative factors] of self.
self is treated as a Mul and the ordering of the factors is maintained. If cset is True
the commutative factors will be returned in a set. If there were repeated factors
(as may happen with an unevaluated Mul) then an error will be raised unless it is
explicitly supressed by setting warn to False.
Note: -1 is always separated from a Number unless split 1 is False.
>>> from sympy import symbols, oo
>>> A, B = symbols(A B, commutative=0)
>>> x, y = symbols(x y)
>>> (-2*x*y).args_cnc()
[[-1, 2, x, y], []]
>>> (-2.5*x).args_cnc()
[[-1, 2.5, x], []]
>>> (-2*x*A*B*y).args_cnc()
[[-1, 2, x, y], [A, B]]
>>> (-2*x*A*B*y).args_cnc(split_1=False)
[[-2, x, y], [A, B]]
>>> (-2*x*y).args_cnc(cset=True)
[set([-1, 2, x, y]), []]
The arg is always treated as a Mul:
>>> (-2 + x + A).args_cnc()
[[], [x - 2 + A]]
>>> (-oo).args_cnc() # -oo is a singleton
[[-1, oo], []]
as coeff Add()
Eciently extract the coecient of a summation.
as coeff Mul(rational=False)
Eciently extract the coecient of a product.
as coeff add(*deps)
Return the tuple (c, args) where self is written as an Add, a.
c should be a Rational added to any terms of the Add that are independent of deps.
args should be a tuple of all other terms of a; args is empty if self is a Number or if
self is independent of deps (when given).
5.1. SymPy Core
103
SymPy Documentation, Release 0.7.6
This should be used when you dont know if self is an Add or not but you want to
treat self as an Add or if you want to process the individual arguments of the tail of
self as an Add.
if you know self is an Add and want only the head, use self.args[0];
if you dont want to process the arguments of the tail but need the tail then use
self.as two terms() which gives the head and tail.
if you want to split self into an independent and dependent parts use
self.as independent(*deps)
>>> from sympy import S
>>> from sympy.abc import x, y
>>> (S(3)).as_coeff_add()
(3, ())
>>> (3 + x).as_coeff_add()
(3, (x,))
>>> (3 + x + y).as_coeff_add(x)
(y + 3, (x,))
>>> (3 + y).as_coeff_add(x)
(y + 3, ())
as coeff exponent(x)
c*x**e -> c,e where x can be any symbolic expression.
as coeff mul(*deps)
Return the tuple (c, args) where self is written as a Mul, m.
c should be a Rational multiplied by any terms of the Mul that are independent of
deps.
args should be a tuple of all other terms of m; args is empty if self is a Number or if
self is independent of deps (when given).
This should be used when you dont know if self is a Mul or not but you want to treat
self as a Mul or if you want to process the individual arguments of the tail of self as
a Mul.
if you know self is a Mul and want only the head, use self.args[0];
if you dont want to process the arguments of the tail but need the tail then use
self.as two terms() which gives the head and tail;
if you want to split self into an independent and dependent parts use
self.as independent(*deps)
>>> from sympy import S
>>> from sympy.abc import x, y
>>> (S(3)).as_coeff_mul()
(3, ())
>>> (3*x*y).as_coeff_mul()
(3, (x, y))
>>> (3*x*y).as_coeff_mul(x)
(3*y, (x,))
>>> (3*y).as_coeff_mul(x)
(3*y, ())
as coefficient(expr)
Extracts symbolic coecient at the given expression. In other words, this functions
separates self into the product of expr and expr-free coecient. If such separation is not possible it will return None.
104
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
See Also:
coeff (page 110) return sum of terms have a given factor
as coeff Add (page 103) separate the additive constant from an expression
as coeff Mul (page 103) separate the multiplicative constant from an expression
as independent (page 107) separate x-dependent terms/factors from others
sympy.polys.polytools.coeff monomial eciently nd the single coecient of a
monomial in Poly
sympy.polys.polytools.nth like coe monomial but powers of monomial terms
are used
Examples
>>> from sympy import E, pi, sin, I, Poly
>>> from sympy.abc import x
>>> E.as_coefficient(E)
1
>>> (2*E).as_coefficient(E)
2
>>> (2*sin(E)*E).as_coefficient(E)
Two terms have E in them so a sum is returned. (If one were desiring the coecient
of the term exactly matching E then the constant from the returned expression could
be selected. Or, for greater precision, a method of Poly can be used to indicate the
desired term from which the coecient is desired.)
>>> (2*E + x*E).as_coefficient(E)
x + 2
>>> _.args[0] # just want the exact match
2
>>> p = Poly(2*E + x*E); p
Poly(x*E + 2*E, x, E, domain=ZZ)
>>> p.coeff_monomial(E)
2
>>> p.nth(0,1)
2
Since the following cannot be written as a product containing E as a factor, None is
returned. (If the coecient 2*x is desired then the coeff method should be used.)
>>> (2*E*x + x).as_coefficient(E)
>>> (2*E*x + x).coeff(E)
2*x
>>> (E*(x + 1) + x).as_coefficient(E)
>>> (2*pi*I).as_coefficient(pi*I)
2
>>> (2*I).as_coefficient(pi*I)
as coefficients dict()
Return a dictionary mapping terms to their Rational coecient. Since the dictio-
5.1. SymPy Core
105
SymPy Documentation, Release 0.7.6
nary is a defaultdict, inquiries about terms which were not present will return a
coecient of 0. If an expression is not an Add it is considered to have a single term.
Examples
>>> from sympy.abc import a, x
>>> (3*x + a*x + 4).as_coefficients_dict()
{1: 4, x: 3, a*x: 1}
>>> _[a]
0
>>> (3*a*x).as_coefficients_dict()
{a*x: 3}
as content primitive(radical=False)
This method should recursively remove a Rational from all arguments and return
that (content) and the new self (primitive). The content should always be positive and Mul(*foo.as content primitive()) == foo. The primitive need no be
in canonical form and should try to preserve the underlying structure if possible
(i.e. expand mul should not be applied to self).
Examples
>>> from sympy import sqrt
>>> from sympy.abc import x, y, z
>>> eq = 2 + 2*x + 2*y*(3 + 3*y)
The as content primitive function is recursive and retains structure:
>>> eq.as_content_primitive()
(2, x + 3*y*(y + 1) + 1)
Integer powers will have Rationals extracted from the base:
>>>
(4,
>>>
(1,
((2 + 6*x)**2).as_content_primitive()
(3*x + 1)**2)
((2 + 6*x)**(2*y)).as_content_primitive()
(2*(3*x + 1))**(2*y))
Terms may end up joining once their as content primitives are added:
>>> ((5*(x*(1 + y)) + 2*x*(3 + 3*y))).as_content_primitive()
(11, x*(y + 1))
>>> ((3*(x*(1 + y)) + 2*x*(3 + 3*y))).as_content_primitive()
(9, x*(y + 1))
>>> ((3*(z*(1 + y)) + 2.0*x*(3 + 3*y))).as_content_primitive()
(1, 6.0*x*(y + 1) + 3*z*(y + 1))
>>> ((5*(x*(1 + y)) + 2*x*(3 + 3*y))**2).as_content_primitive()
(121, x**2*(y + 1)**2)
>>> ((5*(x*(1 + y)) + 2.0*x*(3 + 3*y))**2).as_content_primitive()
(1, 121.0*x**2*(y + 1)**2)
Radical content can also be factored out of the primitive:
106
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> (2*sqrt(2) + 4*sqrt(10)).as_content_primitive(radical=True)
(2, sqrt(2)*(1 + 2*sqrt(5)))
as expr(*gens)
Convert a polynomial to a SymPy expression.
Examples
>>> from sympy import sin
>>> from sympy.abc import x, y
>>> f = (x**2 + x*y).as_poly(x, y)
>>> f.as_expr()
x**2 + x*y
>>> sin(x).as_expr()
sin(x)
as independent(*deps, **hint)
A mostly naive separation of a Mul or Add into arguments that are not are dependent
on deps. To obtain as complete a separation of variables as possible, use a separation
method rst, e.g.:
separatevars() to change Mul, Add and Pow (including exp) into Mul
.expand(mul=True) to change Add or Mul into Add
.expand(log=True) to change log expr into an Add
The only non-naive thing that is done here is to respect noncommutative ordering
of variables.
The returned tuple (i, d) has the following interpretation:
i will has no variable that appears in deps
d will be 1 or else have terms that contain variables that are in deps
if self is an Add then self = i + d
if self is a Mul then self = i*d
if self is anything else, either tuple (self, S.One) or (S.One, self) is returned.
To force the expression to be treated as an Add, use the hint as Add=True
Examples
self is an Add
>>> from sympy import sin, cos, exp
>>> from sympy.abc import x, y, z
>>> (x + x*y).as_independent(x)
(0, x*y + x)
>>> (x + x*y).as_independent(y)
(x, x*y)
>>> (2*x*sin(x) + y + x + z).as_independent(x)
(y + z, 2*x*sin(x) + x)
5.1. SymPy Core
107
SymPy Documentation, Release 0.7.6
>>> (2*x*sin(x) + y + x + z).as_independent(x, y)
(z, 2*x*sin(x) + x + y)
self is a Mul
>>> (x*sin(x)*cos(y)).as_independent(x)
(cos(y), x*sin(x))
non-commutative terms cannot always be separated out when self is a Mul
>>> from sympy import symbols
>>> n1, n2, n3 = symbols(n1 n2 n3, commutative=False)
>>> (n1 + n1*n2).as_independent(n2)
(n1, n1*n2)
>>> (n2*n1 + n1*n2).as_independent(n2)
(0, n1*n2 + n2*n1)
>>> (n1*n2*n3).as_independent(n1)
(1, n1*n2*n3)
>>> (n1*n2*n3).as_independent(n2)
(n1, n2*n3)
>>> ((x-n1)*(x-y)).as_independent(x)
(1, (x - y)*(x - n1))
self is anything else:
>>> (sin(x)).as_independent(x)
(1, sin(x))
>>> (sin(x)).as_independent(y)
(sin(x), 1)
>>> exp(x+y).as_independent(x)
(1, exp(x + y))
force self to be treated as an Add:
>>> (3*x).as_independent(x, as_Add=True)
(0, 3*x)
force self to be treated as a Mul:
>>>
(1,
>>>
(1,
(3+x).as_independent(x, as_Add=False)
x + 3)
(-3+x).as_independent(x, as_Add=False)
x - 3)
Note how the below diers from the above in making the constant on the dep term
positive.
>>> (y*(-3+x)).as_independent(x)
(y, x - 3)
use .as independent() for true independence testing instead of .has(). The
former considers only symbols in the free symbols while the latter considers all
symbols
>>> from sympy import Integral
>>> I = Integral(x, (x, 1, 2))
>>> I.has(x)
True
>>> x in I.free_symbols
108
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
False
>>> I.as_independent(x) == (I, 1)
True
>>> (I + x).as_independent(x) == (I, x)
True
Note: when trying to get independent terms, a separation method might need to be
used rst. In this case, it is important to keep track of what you send to this routine
so you know how to interpret the returned values
>>> from sympy import separatevars, log
>>> separatevars(exp(x+y)).as_independent(x)
(exp(y), exp(x))
>>> (x + x*y).as_independent(y)
(x, x*y)
>>> separatevars(x + x*y).as_independent(y)
(x, y + 1)
>>> (x*(1 + y)).as_independent(y)
(x, y + 1)
>>> (x*(1 + y)).expand(mul=True).as_independent(y)
(x, x*y)
>>> a, b=symbols(a b,positive=True)
>>> (log(a*b).expand(log=True)).as_independent(b)
(log(a), log(b))
See also: .separatevars(), .expand(log=True), .as two terms(), .as coe add(),
.as coe mul()
as leading term(*args, **kwargs)
Returns the leading (nonzero) term of the series expansion of self.
The eval as leading term routines are used to do this, and they must always return
a non-zero value.
Examples
>>> from sympy.abc import x
>>> (1 + x + x**2).as_leading_term(x)
1
>>> (1/x**2 + x + x**2).as_leading_term(x)
x**(-2)
as numer denom()
expression -> a/b -> a, b
This is just a stub that should be dened by an objects class methods to get anything
else.
See Also:
normal return a/b instead of a, b
as ordered factors(order=None)
Return list of ordered factors (if Mul) else [self].
as ordered terms(order=None, data=False)
Transform an expression to an ordered list of terms.
5.1. SymPy Core
109
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import sin, cos
>>> from sympy.abc import x
>>> (sin(x)**2*cos(x) + sin(x)**2 + 1).as_ordered_terms()
[sin(x)**2*cos(x), sin(x)**2, 1]
as powers dict()
Return self as a dictionary of factors with each factor being treated as a power.
The keys are the bases of the factors and the values, the corresponding exponents.
The resulting dictionary should be used with caution if the expression is a Mul and
contains non- commutative factors since the order that they appeared will be lost in
the dictionary.
as real imag(deep=True, **hints)
Performs complex expansion on self and returns a tuple containing collected both
real and imaginary parts. This method cant be confused with re() and im() functions,
which does not perform complex expansion at evaluation.
However it is possible to expand both re() and im() functions and get exactly the
same results as with a single call to this function.
>>> from sympy import symbols, I
>>> x, y = symbols(x,y, real=True)
>>> (x + y*I).as_real_imag()
(x, y)
>>> from sympy.abc import z, w
>>> (z + w*I).as_real_imag()
(re(z) - im(w), re(w) + im(z))
as terms()
Transform an expression to a list of terms.
cancel(*gens, **args)
See the cancel function in sympy.polys
coeff(x, n=1, right=False)
Returns the coecient from the term(s) containing x**n or None. If n is zero then
all terms independent of x will be returned.
When x is noncommutative, the coe to the left (default) or right of x can be returned.
The keyword right is ignored when x is commutative.
See Also:
as coefficient (page 104) separate the expression into a coecient and factor
as coeff Add (page 103) separate the additive constant from an expression
as coeff Mul (page 103) separate the multiplicative constant from an expression
as independent (page 107) separate x-dependent terms/factors from others
sympy.polys.polytools.coeff monomial eciently nd the single coecient of a
monomial in Poly
110
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
sympy.polys.polytools.nth like coe monomial but powers of monomial terms
are used
Examples
>>> from sympy import symbols
>>> from sympy.abc import x, y, z
You can select terms that have an explicit negative in front of them:
>>> (-x + 2*y).coeff(-1)
x
>>> (x - 2*y).coeff(-1)
2*y
You can select terms with no Rational coecient:
>>> (x + 2*y).coeff(1)
x
>>> (3 + 2*x + 4*x**2).coeff(1)
0
You can select terms independent of x by making n=0; in this case
expr.as independent(x)[0] is returned (and 0 will be returned instead of None):
>>> (3 + 2*x + 4*x**2).coeff(x, 0)
3
>>> eq = ((x + 1)**3).expand() + 1
>>> eq
x**3 + 3*x**2 + 3*x + 2
>>> [eq.coeff(x, i) for i in reversed(range(4))]
[1, 3, 3, 2]
>>> eq -= 2
>>> [eq.coeff(x, i) for i in reversed(range(4))]
[1, 3, 3, 0]
You can select terms that have a numerical term in front of them:
>>> (-x - 2*y).coeff(2)
-y
>>> from sympy import sqrt
>>> (x + sqrt(2)*x).coeff(sqrt(2))
x
The matching is exact:
>>>
2
>>>
4
>>>
0
>>>
z
>>>
0
(3 + 2*x + 4*x**2).coeff(x)
(3 + 2*x + 4*x**2).coeff(x**2)
(3 + 2*x + 4*x**2).coeff(x**3)
(z*(x + y)**2).coeff((x + y)**2)
(z*(x + y)**2).coeff(x + y)
In addition, no factoring is done, so 1 + z*(1 + y) is not obtained from the following:
5.1. SymPy Core
111
SymPy Documentation, Release 0.7.6
>>> (x + z*(x + x*y)).coeff(x)
1
If such factoring is desired, factor terms can be used rst:
>>> from sympy import factor_terms
>>> factor_terms(x + z*(x + x*y)).coeff(x)
z*(y + 1) + 1
>>>
>>>
1
>>>
3
>>>
1 +
>>>
m
n, m, o = symbols(n m o, commutative=False)
n.coeff(n)
(3*n).coeff(n)
(n*m + m*n*m).coeff(n) # = (1 + m)*n*m
m
(n*m + m*n*m).coeff(n, right=True) # = (1 + m)*n*m
If there is more than one possible coecient 0 is returned:
>>> (n*m + m*n).coeff(n)
0
If there is only one possible coecient, it is returned:
>>> (n*m + x*m*n).coeff(m*n)
x
>>> (n*m + x*m*n).coeff(m*n, right=1)
1
collect(syms,
func=None,
evaluate=True,
tribute order term=True)
See the collect function in sympy.simplify
exact=False,
dis-
combsimp()
See the combsimp function in sympy.simplify
compute leading term(x, logx=None)
as leading term is only allowed for results of .series() This is a wrapper to compute
a series rst.
could extract minus sign()
Canonical way to choose an element in the set {e, -e} where e is any expression.
If the canonical element is e, we have e.could extract minus sign() == True, else
e.could extract minus sign() == False.
For
any
expression,
the
set
{e.could extract minus sign(),
e).could extract minus sign()} must be {True, False}.
(-
>>> from sympy.abc import x, y
>>> (x-y).could_extract_minus_sign() != (y-x).could_extract_minus_sign()
True
count ops(visual=None)
wrapper for count ops that returns the operation count.
equals(other, failing expression=False)
Return True if self == other, False if it doesnt, or None. If failing expression is True
then the expression which did not simplify to a 0 will be returned instead of None.
If self is a Number (or complex number) that is not zero, then the result is False.
112
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
If self is a number and has not evaluated to zero, evalf will be used to test whether
the expression evaluates to zero. If it does so and the result has signicance (i.e.
the precision is either -1, for a Rational result, or is greater than 1) then the evalf
value will be used to return True or False.
expand(*args, **kwargs)
Expand an expression using hints.
See the docstring of the expand() function in sympy.core.function for more information.
extract additively(c)
Return self - c if its possible to subtract c from self and make all matching coecients move towards zero, else return None.
See Also:
extract multiplicatively (page 114),
(page 104)
coeff (page 110),
as coefficient
Examples
>>> from sympy.abc import x, y
>>> e = 2*x + 3
>>> e.extract_additively(x + 1)
x + 2
>>> e.extract_additively(3*x)
>>> e.extract_additively(4)
>>> (y*(x + 1)).extract_additively(x + 1)
>>> ((x + 1)*(x + 2*y + 1) + 3).extract_additively(x + 1)
(x + 1)*(x + 2*y) + 3
Sometimes auto-expansion will return a less simplied result than desired;
gcd terms might be used in such cases:
>>> from sympy import gcd_terms
>>> (4*x*(y + 1) + y).extract_additively(x)
4*x*(y + 1) + x*(4*y + 3) - x*(4*y + 4) + y
>>> gcd_terms(_)
x*(4*y + 3) + y
extract branch factor(allow half=False)
Try to write self as exp polar(2*pi*I*n)*z in a nice way. Return (z, n).
>>> from sympy import exp_polar, I, pi
>>> from sympy.abc import x, y
>>> exp_polar(I*pi).extract_branch_factor()
(exp_polar(I*pi), 0)
>>> exp_polar(2*I*pi).extract_branch_factor()
(1, 1)
>>> exp_polar(-pi*I).extract_branch_factor()
(exp_polar(I*pi), -1)
>>> exp_polar(3*pi*I + x).extract_branch_factor()
(exp_polar(x + I*pi), 1)
>>> (y*exp_polar(-5*pi*I)*exp_polar(3*pi*I + 2*pi*x)).extract_branch_factor()
(y*exp_polar(2*pi*x), -1)
>>> exp_polar(-I*pi/2).extract_branch_factor()
(exp_polar(-I*pi/2), 0)
5.1. SymPy Core
113
SymPy Documentation, Release 0.7.6
If allow half is True, also extract exp polar(I*pi):
>>>
(1,
>>>
(1,
>>>
(1,
>>>
(1,
exp_polar(I*pi).extract_branch_factor(allow_half=True)
1/2)
exp_polar(2*I*pi).extract_branch_factor(allow_half=True)
1)
exp_polar(3*I*pi).extract_branch_factor(allow_half=True)
3/2)
exp_polar(-I*pi).extract_branch_factor(allow_half=True)
-1/2)
extract multiplicatively(c)
Return None if its not possible to make self in the form c * something in a nice way,
i.e. preserving the properties of arguments of self.
>>> from sympy import symbols, Rational
>>> x, y = symbols(x,y, real=True)
>>> ((x*y)**3).extract_multiplicatively(x**2 * y)
x*y**2
>>> ((x*y)**3).extract_multiplicatively(x**4 * y)
>>> (2*x).extract_multiplicatively(2)
x
>>> (2*x).extract_multiplicatively(3)
>>> (Rational(1,2)*x).extract_multiplicatively(3)
x/6
factor(*gens, **args)
See the factor() function in sympy.polys.polytools
getO()
Returns the additive O(..) symbol if there is one, else None.
getn()
Returns the order of the expression.
The order is determined either from the O(...) term. If there is no O(...) term, it
returns None.
Examples
>>>
>>>
>>>
2
>>>
from sympy import O
from sympy.abc import x
(1 + x + O(x**2)).getn()
(1 + x).getn()
integrate(*args, **kwargs)
See the integrate function in sympy.integrals
invert(g)
See the invert function in sympy.polys
114
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
is algebraic expr(*syms)
This tests whether a given expression is algebraic or not, in the given symbols, syms.
When syms is not given, all free symbols will be used. The rational function does
not have to be in expanded or in any kind of canonical form.
This function returns False for expressions that are algebraic expressions with
symbolic exponents. This is a simple extension to the is rational function, including
rational exponentiation.
See Also:
is rational function (page 117)
References
http://en.wikipedia.org/wiki/Algebraic expression
Examples
>>> from sympy import Symbol, sqrt
>>> x = Symbol(x, real=True)
>>> sqrt(1 + x).is_rational_function()
False
>>> sqrt(1 + x).is_algebraic_expr()
True
This function does not attempt any nontrivial simplications that may result in an
expression that does not appear to be an algebraic expression to become one.
>>> from sympy import exp, factor
>>> a = sqrt(exp(x)**2 + 2*exp(x) + 1)/(exp(x) + 1)
>>> a.is_algebraic_expr(x)
False
>>> factor(a).is_algebraic_expr()
True
is constant(*wrt, **ags)
Return True if self is constant, False if not, or None if the constancy could not be
determined conclusively.
If an expression has no free symbols then it is a constant. If there are free symbols
it is possible that the expression is a constant, perhaps (but not necessarily) zero.
To test such expressions, two strategies are tried:
1) numerical evaluation at two random points. If two such evaluations give two different values and the values have a precision greater than 1 then self is not constant.
If the evaluations agree or could not be obtained with any precision, no decision is
made. The numerical testing is done only if wrt is dierent than the free symbols.
2) dierentiation with respect to variables in wrt (or all free symbols if omitted) to
see if the expression is constant or not. This will not always lead to an expression
that is zero even though an expression is constant (see added test in test expr.py).
If all derivatives are zero then self is constant with respect to the given symbols.
If neither evaluation nor dierentiation can prove the expression is constant, None
is returned unless two numerical values happened to be the same and the ag failing number is True in that case the numerical value will be returned.
5.1. SymPy Core
115
SymPy Documentation, Release 0.7.6
If ag simplify=False is passed, self will not be simplied; the default is True since
self should be simplied before testing.
Examples
>>> from sympy import cos, sin, Sum, S, pi
>>> from sympy.abc import a, n, x, y
>>> x.is_constant()
False
>>> S(2).is_constant()
True
>>> Sum(x, (x, 1, 10)).is_constant()
True
>>> Sum(x, (x, 1, n)).is_constant()
False
>>> Sum(x, (x, 1, n)).is_constant(y)
True
>>> Sum(x, (x, 1, n)).is_constant(n)
False
>>> Sum(x, (x, 1, n)).is_constant(x)
True
>>> eq = a*cos(x)**2 + a*sin(x)**2 - a
>>> eq.is_constant()
True
>>> eq.subs({x:pi, a:2}) == eq.subs({x:pi, a:3}) == 0
True
>>> (0**x).is_constant()
False
>>> x.is_constant()
False
>>> (x**x).is_constant()
False
>>> one = cos(x)**2 + sin(x)**2
>>> one.is_constant()
True
>>> ((one - 1)**(x + 1)).is_constant() in (True, False) # could be 0 or 1
True
is number
Returns True if self has no free symbols. It will be faster than if notself.f rees ymbols,
however, since isn umber will fail as soon as it hits a free symbol.
Examples
>>> from sympy import log, Integral
>>> from sympy.abc import x
>>> x.is_number
False
>>> (2*x).is_number
False
>>> (2 + log(2)).is_number
True
>>> (2 + Integral(2, x)).is_number
116
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
False
>>> (2 + Integral(2, (x, 1, 2))).is_number
True
is polynomial(*syms)
Return True if self is a polynomial in syms and False otherwise.
This checks if self is an exact polynomial in syms. This function returns False for
expressions that are polynomials with symbolic exponents. Thus, you should be
able to apply polynomial algorithms to expressions for which this returns True, and
Poly(expr, *syms) should work if and only if expr.is polynomial(*syms) returns True.
The polynomial does not have to be in expanded form. If no symbols are given, all
free symbols in the expression will be used.
This is not part of the assumptions system.
mial=True).
You cannot do Symbol(z, polyno-
Examples
>>> from sympy import Symbol
>>> x = Symbol(x)
>>> ((x**2 + 1)**4).is_polynomial(x)
True
>>> ((x**2 + 1)**4).is_polynomial()
True
>>> (2**x + 1).is_polynomial(x)
False
>>> n = Symbol(n, nonnegative=True, integer=True)
>>> (x**n + 1).is_polynomial(x)
False
This function does not attempt any nontrivial simplications that may result in an
expression that does not appear to be a polynomial to become one.
>>> from sympy import sqrt, factor, cancel
>>> y = Symbol(y, positive=True)
>>> a = sqrt(y**2 + 2*y + 1)
>>> a.is_polynomial(y)
False
>>> factor(a)
y + 1
>>> factor(a).is_polynomial(y)
True
>>> b = (y**2 + 2*y + 1)/(y + 1)
>>> b.is_polynomial(y)
False
>>> cancel(b)
y + 1
>>> cancel(b).is_polynomial(y)
True
See also .is rational function()
is rational function(*syms)
Test whether function is a ratio of two polynomials in the given symbols, syms. When
5.1. SymPy Core
117
SymPy Documentation, Release 0.7.6
syms is not given, all free symbols will be used. The rational function does not have
to be in expanded or in any kind of canonical form.
This function returns False for expressions that are rational functions with symbolic exponents. Thus, you should be able to call .as numer denom() and apply polynomial algorithms to the result for expressions for which this returns True.
This is not part of the assumptions system.
nal function=True).
You cannot do Symbol(z, ratio-
Examples
>>> from sympy import Symbol, sin
>>> from sympy.abc import x, y
>>> (x/y).is_rational_function()
True
>>> (x**2).is_rational_function()
True
>>> (x/sin(y)).is_rational_function(y)
False
>>> n = Symbol(n, integer=True)
>>> (x**n + 1).is_rational_function(x)
False
This function does not attempt any nontrivial simplications that may result in an
expression that does not appear to be a rational function to become one.
>>> from sympy import sqrt, factor
>>> y = Symbol(y, positive=True)
>>> a = sqrt(y**2 + 2*y + 1)/y
>>> a.is_rational_function(y)
False
>>> factor(a)
(y + 1)/y
>>> factor(a).is_rational_function(y)
True
See also is algebraic expr().
leadterm(x)
Returns the leading term a*x**b as a tuple (a, b).
Examples
>>>
>>>
(1,
>>>
(1,
from sympy.abc import x
(1+x+x**2).leadterm(x)
0)
(1/x**2+x+x**2).leadterm(x)
-2)
limit(x, xlim, dir=+)
Compute limit x->xlim.
118
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
lseries(x=None, x0=0, dir=+, logx=None)
Wrapper for series yielding an iterator of the terms of the series.
Note: an innite series will yield an innite iterator. The following, for exaxmple,
will never terminate. It will just keep printing terms of the sin(x) series:
for term in sin(x).lseries(x):
print term
The advantage of lseries() over nseries() is that many times you are just interested
in the next term in the series (i.e. the rst term for example), but you dont know
how many you should ask for in nseries() using the n parameter.
See also nseries().
nseries(x=None, x0=0, n=6, dir=+, logx=None)
Wrapper to eval nseries if assumptions allow, else to series.
If x is given, x0 is 0, dir=+, and self has x, then eval nseries is called. This calculates n terms in the innermost expressions and then builds up the nal series just
by cross-multiplying everything out.
The optional logx parameter can be used to replace any log(x) in the returned series
with a symbolic value to avoid evaluating log(x) at 0. A symbol to use in place of
log(x) should be provided.
Advantage its fast, because we dont have to determine how many terms we need
to calculate in advance.
Disadvantage you may end up with less terms than you may have expected, but
the O(x**n) term appended will always be correct and so the result, though perhaps
shorter, will also be correct.
If any of those assumptions is not met, this is treated like a wrapper to series which
will try harder to return the correct number of terms.
See also lseries().
Examples
>>>
>>>
>>>
x >>>
x -
from sympy import sin, log, Symbol
from sympy.abc import x, y
sin(x).nseries(x, 0, 6)
x**3/6 + x**5/120 + O(x**6)
log(x+1).nseries(x, 0, 5)
x**2/2 + x**3/3 - x**4/4 + O(x**5)
Handling of the logx parameter in the following example the expansion fails since
sin does not have an asymptotic expansion at -oo (the limit of log(x) as x approaches
0):
>>> e = sin(log(x))
>>> e.nseries(x, 0, 6)
Traceback (most recent call last):
...
PoleError: ...
...
>>> logx = Symbol(logx)
>>> e.nseries(x, 0, 6, logx=logx)
sin(logx)
5.1. SymPy Core
119
SymPy Documentation, Release 0.7.6
Traceback (most recent call last):
...
PoleError: ...
In the following example, the expansion works but gives only an Order term unless
the logx parameter is used:
>>> e = x**y
>>> e.nseries(x, 0, 2)
O(log(x)**2)
>>> e.nseries(x, 0, 2, logx=logx)
exp(logx*y)
nsimplify(constants=[], tolerance=None, full=False)
See the nsimplify function in sympy.simplify
powsimp(deep=False, combine=all)
See the powsimp function in sympy.simplify
primitive()
Return the positive Rational that can be extracted non-recursively from every term
of self (i.e., self is treated like an Add). This is like the as coe Mul() method but
primitive always extracts a positive Rational (never a negative or a Float).
Examples
>>> from sympy.abc import x
>>> (3*(x + 1)**2).primitive()
(3, (x + 1)**2)
>>> a = (6*x + 2); a.primitive()
(2, 3*x + 1)
>>> b = (x/2 + 3); b.primitive()
(1/2, x + 6)
>>> (a*b).primitive() == (1, a*b)
True
radsimp()
See the radsimp function in sympy.simplify
ratsimp()
See the ratsimp function in sympy.simplify
refine(assumption=True)
See the rene function in sympy.assumptions
removeO()
Removes the additive O(..) symbol if there is one
round(p=0)
Return x rounded to the given decimal place.
If a complex number would results, apply round to the real and imaginary components of the number.
Notes
Do not confuse the Python builtin function, round, with the SymPy method of the
same name. The former always returns a oat (or raises an error if applied to a
120
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
complex value) while the latter returns either a Number or a complex number:
>>> isinstance(round(S(123), -2), Number)
False
>>> isinstance(S(123).round(-2), Number)
True
>>> isinstance((3*I).round(), Mul)
True
>>> isinstance((1 + 3*I).round(), Add)
True
Examples
>>> from sympy import pi, E, I, S, Add, Mul, Number
>>> S(10.5).round()
11.
>>> pi.round()
3.
>>> pi.round(2)
3.14
>>> (2*pi + E*I).round()
6. + 3.*I
The round method has a chopping eect:
>>> (2*pi + I/10).round()
6.
>>> (pi/10 + 2*I).round()
2.*I
>>> (pi/10 + E*I).round(2)
0.31 + 2.72*I
separate(deep=False, force=False)
See the separate function in sympy.simplify
series(x=None, x0=0, n=6, dir=+, logx=None)
Series expansion of self around x = x0 yielding either terms of the series one by
one (the lazy series given when n=None), else all the terms at once when n != None.
Returns the series expansion of self around the point x = x0 with respect to x up
to O((x - x0)**n, x, x0) (default n is 6).
If x=None and self is univariate, the univariate symbol will be supplied, otherwise
an error will be raised.
>>> from sympy import cos, exp
>>> from sympy.abc import x, y
>>> cos(x).series()
1 - x**2/2 + x**4/24 + O(x**6)
>>> cos(x).series(n=4)
1 - x**2/2 + O(x**4)
>>> cos(x).series(x, x0=1, n=2)
cos(1) - (x - 1)*sin(1) + O((x - 1)**2, (x, 1))
>>> e = cos(x + exp(y))
>>> e.series(y, n=2)
cos(x + 1) - y*sin(x + 1) + O(y**2)
>>> e.series(x, n=2)
cos(exp(y)) - x*sin(exp(y)) + O(x**2)
5.1. SymPy Core
121
SymPy Documentation, Release 0.7.6
If n=None then a generator of the series terms will be returned.
>>> term=cos(x).series(n=None)
>>> [next(term) for i in range(2)]
[1, -x**2/2]
For dir=+ (default) the series is calculated from the right and for dir=- the series
from the left. For smooth functions this ag will not alter the results.
>>> abs(x).series(dir=+)
x
>>> abs(x).series(dir=-)
-x
simplify(ratio=1.7, measure=None)
See the simplify function in sympy.simplify
taylor term(n, x, *previous terms)
General method for the taylor term.
This method is slow, because it dierentiates n-times. Subclasses can redene it to
make it faster by using the previous terms.
together(*args, **kwargs)
See the together function in sympy.polys
trigsimp(**args)
See the trigsimp function in sympy.simplify
5.1.8 AtomicExpr
class sympy.core.expr.AtomicExpr
A parent class for object which are both atoms and Exprs.
For example: Symbol, Number, Rational, Integer, ... But not: Add, Mul, Pow, ...
5.1.9 symbol
Symbol
class sympy.core.symbol.Symbol
Assumptions: commutative = True
You can override the default assumptions in the constructor:
>>> from sympy import symbols
>>> A,B = symbols(A,B, commutative = False)
>>> bool(A*B != B*A)
True
>>> bool(A*B*2 == 2*A*B) == True # multiplication by scalars is commutative
True
as dummy()
Return a Dummy having the same name and same assumptions as self.
122
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Wild
class sympy.core.symbol.Wild
A Wild symbol matches anything, or anything without whatever is explicitly excluded.
Examples
>>> from sympy import Wild, WildFunction, cos, pi
>>> from sympy.abc import x, y, z
>>> a = Wild(a)
>>> x.match(a)
{a_: x}
>>> pi.match(a)
{a_: pi}
>>> (3*x**2).match(a*x)
{a_: 3*x}
>>> cos(x).match(a)
{a_: cos(x)}
>>> b = Wild(b, exclude=[x])
>>> (3*x**2).match(b*x)
>>> b.match(a)
{a_: b_}
>>> A = WildFunction(A)
>>> A.match(a)
{a_: A_}
Tips
When using Wild, be sure to use the exclude keyword to make the pattern more precise.
Without the exclude pattern, you may get matches that are technically correct, but not
what you wanted. For example, using the above without exclude:
>>> from sympy import symbols
>>> a, b = symbols(a b, cls=Wild)
>>> (2 + 3*y).match(a*x + b*y)
{a_: 2/x, b_: 3}
This is technically correct, because (2/x)*x + 3*y == 2 + 3*y, but you probably wanted
it to not match at all. The issue is that you really didnt want a and b to include x and
y, and the exclude parameter lets you specify exactly this. With the exclude parameter,
the pattern will not match.
>>> a = Wild(a, exclude=[x, y])
>>> b = Wild(b, exclude=[x, y])
>>> (2 + 3*y).match(a*x + b*y)
Exclude also helps remove ambiguity from matches.
>>> E = 2*x**3*y*z
>>> a, b = symbols(a b, cls=Wild)
>>> E.match(a*b)
{a_: 2*y*z, b_: x**3}
>>> a = Wild(a, exclude=[x, y])
>>> E.match(a*b)
{a_: z, b_: 2*x**3*y}
5.1. SymPy Core
123
SymPy Documentation, Release 0.7.6
>>> a = Wild(a, exclude=[x, y, z])
>>> E.match(a*b)
{a_: 2, b_: x**3*y*z}
Dummy
class sympy.core.symbol.Dummy
Dummy symbols are each unique, identied by an internal count index:
>>> from sympy import Dummy
>>> bool(Dummy(x) == Dummy(x)) == True
False
If a name is not supplied then a string value of the count index will be used. This is
useful when a temporary variable is needed and the name of the variable used in the
expression is not important.
>>> Dummy()
_Dummy_10
symbols
sympy.core.symbol.symbols(names, **args)
Transform strings into instances of Symbol (page 122) class.
symbols() (page 124) function returns a sequence of symbols with names taken from
names argument, which can be a comma or whitespace delimited string, or a sequence
of strings:
>>> from sympy import symbols, Function
>>> x, y, z = symbols(x,y,z)
>>> a, b, c = symbols(a b c)
The type of output is dependent on the properties of input arguments:
>>> symbols(x)
x
>>> symbols(x,)
(x,)
>>> symbols(x,y)
(x, y)
>>> symbols((a, b, c))
(a, b, c)
>>> symbols([a, b, c])
[a, b, c]
>>> symbols(set([a, b, c]))
set([a, b, c])
If an iterable container is needed for a single symbol, set the seq argument to True or
terminate the symbol name with a comma:
>>> symbols(x, seq=True)
(x,)
124
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
To reduce typing, range syntax is supported to create indexed symbols. Ranges are
indicated by a colon and the type of range is determined by the character to the right of
the colon. If the character is a digit then all contiguous digits to the left are taken as the
nonnegative starting value (or 0 if there is no digit left of the colon) and all contiguous
digits to the right are taken as 1 greater than the ending value:
>>> symbols(x:10)
(x0, x1, x2, x3, x4, x5, x6, x7, x8, x9)
>>> symbols(x5:10)
(x5, x6, x7, x8, x9)
>>> symbols(x5(:2))
(x50, x51)
>>> symbols(x5:10,y:5)
(x5, x6, x7, x8, x9, y0, y1, y2, y3, y4)
>>> symbols((x5:10, y:5))
((x5, x6, x7, x8, x9), (y0, y1, y2, y3, y4))
If the character to the right of the colon is a letter, then the single letter to the left (or
a if there is none) is taken as the start and all characters in the lexicographic range
through the letter to the right are used as the range:
>>> symbols(x:z)
(x, y, z)
>>> symbols(x:c) # null range
()
>>> symbols(x(:c))
(xa, xb, xc)
>>> symbols(:c)
(a, b, c)
>>> symbols(a:d, x:z)
(a, b, c, d, x, y, z)
>>> symbols((a:d, x:z))
((a, b, c, d), (x, y, z))
Multiple ranges are supported; contiguous numerical ranges should be separated by
parentheses to disambiguate the ending number of one range from the starting number
of the next:
>>> symbols(x:2(1:3))
(x01, x02, x11, x12)
>>> symbols(:3:2) # parsing is from left to right
(00, 01, 10, 11, 20, 21)
Only one pair of parentheses surrounding ranges are removed, so to include parentheses
around ranges, double them. And to include spaces, commas, or colons, escape them
with a backslash:
>>> symbols(x((a:b)))
(x(a), x(b))
>>> symbols(x(:1\,:2))
(x(0,0), x(0,1))
# or x((:1)\,(:2))
All newly created symbols have assumptions set according to args:
5.1. SymPy Core
125
SymPy Documentation, Release 0.7.6
>>> a = symbols(a, integer=True)
>>> a.is_integer
True
>>> x, y, z = symbols(x,y,z, real=True)
>>> x.is_real and y.is_real and z.is_real
True
Despite its name, symbols() (page 124) can create symbol-like objects like instances of
Function or Wild classes. To achieve this, set cls keyword argument to the desired type:
>>> symbols(f,g,h, cls=Function)
(f, g, h)
>>> type(_[0])
<class sympy.core.function.UndefinedFunction>
var
sympy.core.symbol.var(names, **args)
Create symbols and inject them into the global namespace.
This calls symbols() (page 124) with the same arguments and puts the results into the
global namespace. Its recommended not to use var() (page 126) in library code, where
symbols() (page 124) has to be used:
.. rubric:: Examples
>>> from sympy import var
>>> var(x)
x
>>> x
x
>>> var(a,ab,abc)
(a, ab, abc)
>>> abc
abc
>>> var(x,y, real=True)
(x, y)
>>> x.is_real and y.is_real
True
See symbol() documentation for more details on what kinds of arguments can be passed
to var() (page 126).
5.1.10 numbers
Number
class sympy.core.numbers.Number
Represents any kind of number in sympy.
126
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Floating point numbers are represented by the Float class. Integer numbers (of any size),
together with rational numbers (again, there is no limit on their size) are represented by
the Rational class.
If you want to represent, for example, 1+sqrt(2), then you need to do:
Rational(1) + sqrt(Rational(2))
as coeff Add()
Eciently extract the coecient of a summation.
as coeff Mul(rational=False)
Eciently extract the coecient of a product.
cofactors(other)
Compute GCD and cofactors of self and other.
gcd(other)
Compute GCD of self and other.
lcm(other)
Compute LCM of self and other.
Float
class sympy.core.numbers.Float
Represents a oating point number. It is capable of representing arbitrary-precision
oating-point numbers.
Notes
Floats are inexact by their nature unless their value is a binary-exact value.
>>> approx, exact = Float(.1, 1), Float(.125, 1)
For calculation purposes, evalf needs to be able to change the precision but this will not
increase the accuracy of the inexact value. The following is the most accurate 5-digit
approximation of a value of 0.1 that had only 1 digit of precision:
>>> approx.evalf(5)
0.099609
By contrast, 0.125 is exact in binary (as it is in base 10) and so it can be passed to Float
or evalf to obtain an arbitrary precision with matching accuracy:
>>> Float(exact, 5)
0.12500
>>> exact.evalf(20)
0.12500000000000000000
Trying to make a high-precision Float from a oat is not disallowed, but one must keep
in mind that the underlying oat (not the apparent decimal value) is being obtained with
high precision. For example, 0.3 does not have a nite binary representation. The closest
rational is the fraction 5404319552844595/2**54. So if you try to obtain a Float of 0.3
to 20 digits of precision you will not see the same thing as 0.3 followed by 19 zeros:
>>> Float(0.3, 20)
0.29999999999999998890
5.1. SymPy Core
127
SymPy Documentation, Release 0.7.6
If you want a 20-digit value of the decimal 0.3 (not the oating point approximation of
0.3) you should send the 0.3 as a string. The underlying representation is still binary but
a higher precision than Pythons oat is used:
>>> Float(0.3, 20)
0.30000000000000000000
Although you can increase the precision of an existing Float using Float it will not increase the accuracy the underlying value is not changed:
>>> def show(f): # binary rep of Float
...
from sympy import Mul, Pow
...
s, m, e, b = f._mpf_
...
v = Mul(int(m), Pow(2, int(e), evaluate=False), evaluate=False)
...
print(%s at prec=%s % (v, f._prec))
...
>>> t = Float(0.3, 3)
>>> show(t)
4915/2**14 at prec=13
>>> show(Float(t, 20)) # higher prec, not higher accuracy
4915/2**14 at prec=70
>>> show(Float(t, 2)) # lower prec
307/2**10 at prec=10
The same thing happens when evalf is used on a Float:
>>> show(t.evalf(20))
4915/2**14 at prec=70
>>> show(t.evalf(2))
307/2**10 at prec=10
Finally, Floats can be instantiated with an mpf tuple (n, c, p) to produce the number
(-1)**n*c*2**p:
>>> n, c, p = 1, 5, 0
>>> (-1)**n*c*2**p
-5
>>> Float((1, 5, 0))
-5.00000000000000
An actual mpf tuple also contains the number of bits in c as the last element of the tuple:
>>> _._mpf_
(1, 5, 0, 3)
This is not needed for instantiation and is not the same thing as the precision. The mpf
tuple and the precision are two separate quantities that Float tracks.
Examples
>>> from sympy import Float
>>> Float(3.5)
3.50000000000000
>>> Float(3)
3.00000000000000
Floats can be created from a string representations of Python oats to force ints to Float
or to enter high-precision (> 15 signicant digits) values:
128
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> Float(.0010)
0.00100000000000000
>>> Float(1e-3)
0.00100000000000000
>>> Float(1e-3, 3)
0.00100
Float can automatically count signicant gures if a null string is sent for the precision;
space are also allowed in the string. (Auto- counting is only allowed for strings, ints and
longs).
>>> Float(123 456 789 . 123 456, )
123456789.123456
>>> Float(12e-3, )
0.012
>>> Float(3, )
3.
If a number is written in scientic notation, only the digits before the exponent are
considered signicant if a decimal appears, otherwise the e signies only how to move
the decimal:
>>> Float(60.e2, ) # 2 digits significant
6.0e+3
>>> Float(60e2, ) # 4 digits significant
6000.
>>> Float(600e-2, ) # 3 digits significant
6.00
Attributes
is irrational
is rational
Rational
class sympy.core.numbers.Rational
Represents integers and rational numbers (p/q) of any size.
See Also:
sympify, sympy.simplify.simplify.nsimplify (page 1340)
Examples
>>> from sympy import Rational, nsimplify, S, pi
>>> Rational(3)
3
>>> Rational(1, 2)
1/2
Rational is unprejudiced in accepting input. If a oat is passed, the underlying value of
the binary representation will be returned:
5.1. SymPy Core
129
SymPy Documentation, Release 0.7.6
>>> Rational(.5)
1/2
>>> Rational(.2)
3602879701896397/18014398509481984
If the simpler representation of the oat is desired then consider limiting the denominator to the desired value or convert the oat to a string (which is roughly equivalent to
limiting the denominator to 10**12):
>>> Rational(str(.2))
1/5
>>> Rational(.2).limit_denominator(10**12)
1/5
An arbitrarily precise Rational is obtained when a string literal is passed:
>>> Rational(1.23)
123/100
>>> Rational(1e-2)
1/100
>>> Rational(.1)
1/10
>>> Rational(1e-2/3.2)
1/320
The conversion of other types of strings can be handled by the sympify() function, and
conversion of oats to expressions or simple fractions can be handled with nsimplify:
>>> S(.[3]) # repeating digits in brackets
1/3
>>> S(3**2/10) # general expressions
9/10
>>> nsimplify(.3) # numbers that have a simple form
3/10
But if the input does not reduce to a literal Rational, an error will be raised:
>>> Rational(pi)
Traceback (most recent call last):
...
TypeError: invalid input: pi
Traceback (most recent call last):
...
TypeError: invalid input: pi
Low-level
Access numerator and denominator as .p and .q:
>>>
>>>
3/4
>>>
3
>>>
4
130
r = Rational(3, 4)
r
r.p
r.q
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Note that p and q return integers (not SymPy Integers) so some care is needed when
using them in expressions:
>>> r.p/r.q
0.75
as content primitive(radical=False)
Return the tuple (R, self/R) where R is the positive Rational extracted from self.
Examples
>>> from sympy import S
>>> (S(-3)/2).as_content_primitive()
(3/2, -1)
See docstring of Expr.as content primitive for more examples.
factors(limit=None, use trial=True, use rho=False, use pm1=False, verbose=False, visual=False)
A wrapper to factorint which return factors of self that are smaller than limit (or
cheap to compute). Special methods of factoring are disabled by default so that
only trial division is used.
limit denominator(max denominator=1000000)
Closest Rational to self with denominator at most max denominator.
>>> from sympy import Rational
>>> Rational(3.141592653589793).limit_denominator(10)
22/7
>>> Rational(3.141592653589793).limit_denominator(100)
311/99
Integer
class sympy.core.numbers.Integer
NumberSymbol
class sympy.core.numbers.NumberSymbol
approximation(number cls)
Return an interval with number cls endpoints that contains the value of NumberSymbol. If not implemented, then return None.
RealNumber
sympy.core.numbers.RealNumber
alias of Float (page 127)
5.1. SymPy Core
131
SymPy Documentation, Release 0.7.6
igcd
sympy.core.numbers.igcd(*args)
Computes positive integer greatest common divisor.
The algorithm is based on the well known Euclids algorithm. To improve speed, igcd()
has its own caching mechanism implemented.
Examples
>>> from sympy.core.numbers import igcd
>>> igcd(2, 4)
2
>>> igcd(5, 10, 15)
5
ilcm
sympy.core.numbers.ilcm(*args)
Computes integer least common multiple.
Examples
>>>
>>>
10
>>>
21
>>>
30
from sympy.core.numbers import ilcm
ilcm(5, 10)
ilcm(7, 3)
ilcm(5, 10, 15)
seterr
sympy.core.numbers.seterr(divide=False)
Should sympy raise an exception on 0/0 or return a nan?
divide == True .... raise an exception divide == False ... return nan
Zero
class sympy.core.numbers.Zero
The number zero.
Zero is a singleton, and can be accessed by S.Zero
References
[R33] (page 1899)
132
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import S, Integer, zoo
>>> Integer(0) is S.Zero
True
>>> 1/S.Zero
zoo
One
class sympy.core.numbers.One
The number one.
One is a singleton, and can be accessed by S.One.
References
[R34] (page 1899)
Examples
>>> from sympy import S, Integer
>>> Integer(1) is S.One
True
NegativeOne
class sympy.core.numbers.NegativeOne
The number negative one.
NegativeOne is a singleton, and can be accessed by S.NegativeOne.
See Also:
One (page 133)
References
[R35] (page 1899)
Examples
>>> from sympy import S, Integer
>>> Integer(-1) is S.NegativeOne
True
5.1. SymPy Core
133
SymPy Documentation, Release 0.7.6
Half
class sympy.core.numbers.Half
The rational number 1/2.
Half is a singleton, and can be accessed by S.Half.
References
[R36] (page 1899)
Examples
>>> from sympy import S, Rational
>>> Rational(1, 2) is S.Half
True
NaN
class sympy.core.numbers.NaN
Not a Number.
This represents the corresponding data type to oating point nan, which is dened in
the IEEE 754 oating point standard, and corresponds to the Python float(nan).
NaN serves as a place holder for numeric values that are indeterminate. Most operations
on NaN, produce another NaN. Most indeterminate forms, such as 0/0 or oo - oo
produce NaN. Two exceptions are 0**0 and oo**0, which all produce 1 (this is
consistent with Pythons oat).
NaN is mathematically not equal to anything else, even NaN itself. This explains the
initially counter-intuitive results with Eq and == in the examples below.
NaN is a singleton, and can be accessed by S.NaN, or can be imported as nan.
References
[R37] (page 1899)
Examples
>>> from sympy import nan, S, oo, Eq
>>> nan is S.NaN
True
>>> oo - oo
nan
>>> nan + 1
nan
>>> Eq(nan, nan)
# mathematical equality
False
>>> nan == nan
# structural equality
True
134
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Attributes
is
is
is
is
is
is
is
is
is
is
algebraic
nite
integer
negative
positive
prime
rational
real
transcendental
zero
Innity
class sympy.core.numbers.Infinity
Positive innite quantity.
In real analysis the symbol denotes an unbounded limit: x means that x grows
without bound.
Innity is often used not only to dene a limit but as a value in the anely extended
real number system. Points labeled + and can be added to the topological space of
the real numbers, producing the two-point compactication of the real numbers. Adding
algebraic properties to this gives us the extended real numbers.
Innity is a singleton, and can be accessed by S.Infinity, or can be imported as oo.
See Also:
NegativeInfinity (page 135), NaN (page 134)
References
[R38] (page 1899)
Examples
>>>
>>>
oo
>>>
0
>>>
>>>
oo
from sympy import oo, exp, limit, Symbol
1 + oo
42/oo
x = Symbol(x)
limit(exp(x), x, oo)
NegativeInnity
class sympy.core.numbers.NegativeInfinity
Negative innite quantity.
NegativeInnity is a singleton, and can be accessed by S.NegativeInfinity.
5.1. SymPy Core
135
SymPy Documentation, Release 0.7.6
See Also:
Infinity (page 135)
ComplexInnity
class sympy.core.numbers.ComplexInfinity
Complex innity.
In complex analysis the symbol ,
called complex innity, represents a quantity with
innite magnitude, but undetermined complex phase.
ComplexInnity is a singleton, and can be accessed by S.ComplexInfinity, or can be
imported as zoo.
See Also:
Infinity (page 135)
Examples
>>>
>>>
zoo
>>>
0
>>>
nan
>>>
zoo
from sympy import zoo, oo
zoo + 42
42/zoo
zoo + zoo
zoo*zoo
Exp1
class sympy.core.numbers.Exp1
The e constant.
The transcendental number e = 2.718281828 . . . is the base of the natural logarithm and
of the exponential function, e = exp(1). Sometimes called Eulers number or Napiers
constant.
Exp1 is a singleton, and can be accessed by S.Exp1, or can be imported as E.
References
[R39] (page 1899)
Examples
>>> from sympy import exp, log, E
>>> E is exp(1)
True
>>> log(E)
1
136
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
ImaginaryUnit
class sympy.core.numbers.ImaginaryUnit
The imaginary unit, i = 1.
I is a singleton, and can be accessed by S.I, or can be imported as I.
References
[R40] (page 1899)
Examples
>>>
>>>
I
>>>
-1
>>>
-I
from sympy import I, sqrt
sqrt(-1)
I*I
1/I
Pi
class sympy.core.numbers.Pi
The constant.
The transcendental number = 3.141592654 . . . represents the ratio of a circles circumference to its diameter, the area of the unit circle, the half-period of trigonometric functions,
and many other things in mathematics.
Pi is a singleton, and can be accessed by S.Pi, or can be imported as pi.
References
[R41] (page 1899)
Examples
>>> from sympy import S, pi, oo, sin, exp, integrate, Symbol
>>> S.Pi
pi
>>> pi > 3
True
>>> pi.is_irrational
True
>>> x = Symbol(x)
>>> sin(x + 2*pi)
sin(x)
>>> integrate(exp(-x**2), (x, -oo, oo))
sqrt(pi)
5.1. SymPy Core
137
SymPy Documentation, Release 0.7.6
EulerGamma
class sympy.core.numbers.EulerGamma
The Euler-Mascheroni constant.
= 0.5772157 . . . (also called Eulers constant) is a mathematical constant recurring in
analysis and number theory. It is dened as the limiting dierence between the harmonic
series and the natural logarithm:
( n
)
1
= lim
ln n
n
k
k=1
EulerGamma is a singleton, and can be accessed by S.EulerGamma.
References
[R42] (page 1899)
Examples
>>> from sympy import S
>>> S.EulerGamma.is_irrational
>>> S.EulerGamma > 0
True
>>> S.EulerGamma > 1
False
Attributes
is irrational
Catalan
class sympy.core.numbers.Catalan
Catalans constant.
K = 0.91596559 . . . is given by the innite series
K=
k=0
(1)k
(2k + 1)2
Catalan is a singleton, and can be accessed by S.Catalan.
References
[R43] (page 1899)
138
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import S
>>> S.Catalan.is_irrational
>>> S.Catalan > 0
True
>>> S.Catalan > 1
False
Attributes
is irrational
GoldenRatio
class sympy.core.numbers.GoldenRatio
The golden ratio, .
= 1+2 5 is algebraic number. Two quantities are in the golden ratio if their ratio is the
same as the ratio of their sum to the larger of the two quantities, i.e. their maximum.
GoldenRatio is a singleton, and can be accessed by S.GoldenRatio.
References
[R44] (page 1899)
Examples
>>> from sympy import S
>>> S.GoldenRatio > 1
True
>>> S.GoldenRatio.expand(func=True)
1/2 + sqrt(5)/2
>>> S.GoldenRatio.is_irrational
True
5.1.11 power
Pow
class sympy.core.power.Pow
Denes the expression x**y as x raised to a power y
Singleton denitions involving (0, 1, -1, oo, -oo):
5.1. SymPy Core
139
SymPy Documentation, Release 0.7.6
expr
z**0
z**1
(-oo)**(1)
(-1)**-1
S.Zero**1
value reason
1
Although arguments over 0**0 exist, see [2].
z
0
1**-1
oo**-1
0**oo
0**-oo
1
0
0
oo
1**oo
1**-oo
(-1)**oo
(-1)**(oo)
oo**oo
oo**-oo
(-oo)**oo
(-oo)**-oo
nan
-1
oo
nan
This is not strictly true, as 0**-1 may be undened, but is
convenient is some contexts where the base is assumed to be
positive.
Because for all complex numbers z near 0, z**oo -> 0.
This is not strictly true, as 0**oo may be oscillating between
positive and negative values or rotating in the complex plane. It is
convenient, however, when the base is positive.
Because there are various cases where lim(x(t),t)=1, lim(y(t),t)=oo
(or -oo), but lim( x(t)**y(t), t) != 1. See [3].
Because of oscillations in the limit.
oo
0
nan
Because symbolic computations are more exible that oating point calculations and we
prefer to never return an incorrect answer, we choose not to conform to all IEEE 754
conventions. This helps us avoid extra test-case code in the calculation of limits.
See Also:
Infinity, NegativeInfinity, NaN
References
[R45] (page 1899), [R46] (page 1899), [R47] (page 1899)
as base exp()
Return base and exp of self.
If base is 1/Integer, then return Integer, -exp. If this extra processing is not needed,
the base and exp properties will give the raw arguments
Examples
>>> from sympy import Pow, S
>>> p = Pow(S.Half, 2, evaluate=False)
>>> p.as_base_exp()
(2, -2)
>>> p.args
(1/2, 2)
as content primitive(radical=False)
Return the tuple (R, self/R) where R is the positive Rational extracted from self.
140
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>>
>>>
(2,
>>>
(1,
from sympy import sqrt
sqrt(4 + 4*sqrt(2)).as_content_primitive()
sqrt(1 + sqrt(2)))
sqrt(3 + 3*sqrt(2)).as_content_primitive()
sqrt(3)*sqrt(1 + sqrt(2)))
>>> from sympy import expand_power_base, powsimp, Mul
>>> from sympy.abc import x, y
>>> ((2*x + 2)**2).as_content_primitive()
(4, (x + 1)**2)
>>> (4**((1 + y)/2)).as_content_primitive()
(2, 4**(y/2))
>>> (3**((1 + y)/2)).as_content_primitive()
(1, 3**((y + 1)/2))
>>> (3**((5 + y)/2)).as_content_primitive()
(9, 3**((y + 1)/2))
>>> eq = 3**(2 + 2*x)
>>> powsimp(eq) == eq
True
>>> eq.as_content_primitive()
(9, 3**(2*x))
>>> powsimp(Mul(*_))
3**(2*x + 2)
>>> eq = (2 + 2*x)**y
>>> s = expand_power_base(eq); s.is_Mul, s
(False, (2*x + 2)**y)
>>> eq.as_content_primitive()
(1, (2*(x + 1))**y)
>>> s = expand_power_base(_[1]); s.is_Mul, s
(True, 2**y*(x + 1)**y)
See docstring of Expr.as content primitive for more examples.
integer nthroot
sympy.core.power.integer nthroot(y, n)
Return a tuple containing x = oor(y**(1/n)) and a boolean indicating whether the result
is exact (that is, whether x**n == y).
>>>
>>>
(4,
>>>
(5,
from sympy import integer_nthroot
integer_nthroot(16,2)
True)
integer_nthroot(26,2)
False)
5.1. SymPy Core
141
SymPy Documentation, Release 0.7.6
5.1.12 mul
Mul
class sympy.core.mul.Mul
as coeff Mul(rational=False)
Eciently extract the coecient of a product.
as content primitive(radical=False)
Return the tuple (R, self/R) where R is the positive Rational extracted from self.
Examples
>>> from sympy import sqrt
>>> (-3*sqrt(2)*(2 - 2*sqrt(2))).as_content_primitive()
(6, -sqrt(2)*(-sqrt(2) + 1))
See docstring of Expr.as content primitive for more examples.
as ordered factors(order=None)
Transform an expression into an ordered list of factors.
Examples
>>> from sympy import sin, cos
>>> from sympy.abc import x, y
>>> (2*x*y*sin(x)*cos(x)).as_ordered_factors()
[2, x, y, sin(x), cos(x)]
as two terms(*args, **kwargs)
Return head and tail of self.
This is the most ecient way to get the head and tail of an expression.
if you want only the head, use self.args[0];
if you want to process the arguments of the tail then use self.as coef mul() which
gives the head and a tuple containing the arguments of the tail when treated as
a Mul.
if you want the coecient when self is treated as an Add then use
self.as coe add()[0]
>>> from sympy.abc import x, y
>>> (3*x*y).as_two_terms()
(3, x*y)
classmethod flatten(seq)
Return commutative, noncommutative and order arguments by combining related
terms.
142
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Notes
In an expression like a*b*c, python process this through sympy as Mul(Mul(a,
b), c). This can have undesirable consequences.
Sometimes terms are not combined as
https://github.com/sympy/sympy/issues/4596}
one
would
like:
{c.f.
>>> from sympy import Mul, sqrt
>>> from sympy.abc import x, y, z
>>> 2*(x + 1) # this is the 2-arg Mul behavior
2*x + 2
>>> y*(x + 1)*2
2*y*(x + 1)
>>> 2*(x + 1)*y # 2-arg result will be obtained first
y*(2*x + 2)
>>> Mul(2, x + 1, y) # all 3 args simultaneously processed
2*y*(x + 1)
>>> 2*((x + 1)*y) # parentheses can control this behavior
2*y*(x + 1)
Powers with compound bases may not nd a single base to combine with
unless all arguments are processed at once. Post-processing may be necessary in such cases. {c.f. https://github.com/sympy/sympy/issues/5728}
>>> a = sqrt(x*sqrt(y))
>>> a**3
(x*sqrt(y))**(3/2)
>>> Mul(a,a,a)
(x*sqrt(y))**(3/2)
>>> a*a*a
x*sqrt(y)*sqrt(x*sqrt(y))
>>> _.subs(a.base, z).subs(z, a.base)
(x*sqrt(y))**(3/2)
If more than two terms are being multiplied then all the previous terms will
be re-processed for each new argument. So if each of a, b and c were Mul
(page 142) expression, then a*b*c (or building up the product with *=) will
process all the arguments of a and b twice: once when a*b is computed and
again when c is multiplied.
Using Mul(a, b, c) will process all arguments once.
The results of Mul are cached according to arguments, so atten will only be
called once for Mul(a, b, c). If you can structure a calculation so the arguments
are most likely to be repeats then this can save time in computing the answer. For
example, say you had a Mul, M, that you wished to divide by d[i] and multiply
by n[i] and you suspect there are many repeats in n. It would be better to
compute M*n[i]/d[i] rather than M/d[i]*n[i] since every time n[i] is a repeat,
the product, M*n[i] will be returned without attening the cached value will
be returned. If you divide by the d[i] rst (and those are more unique than the
n[i]) then that will create a new Mul, M/d[i] the args of which will be traversed
again when it is multiplied by n[i].
{c.f. https://github.com/sympy/sympy/issues/5706}
This consideration is moot if the cache is turned o.
5.1. SymPy Core
143
SymPy Documentation, Release 0.7.6
Nb
The validity of the above notes depends on the implementation details of Mul and
atten which may change at any time. Therefore, you should only consider them
when your code is highly performance sensitive.
Removal of 1 from the sequence is already handled by AssocOp. new .
prod
sympy.core.mul.prod(a, start=1)
Return product of elements of a. Start with int 1 so if only ints are included then an int
result is returned.
Examples
>>> from sympy import prod, S
>>> prod(range(3))
0
>>> type(_) is int
True
>>> prod([S(2), 3])
6
>>> _.is_Integer
True
You can start the product at something other than 1: >>> prod([1, 2], 3) 6
5.1.13 add
Add
class sympy.core.add.Add
as coeff Add()
Eciently extract the coecient of a summation.
as coeff add(*args, **kwargs)
Returns a tuple (coe, args) where self is treated as an Add and coe is the Number
term and args is a tuple of all other terms.
Examples
>>>
>>>
(7,
>>>
(0,
144
from sympy.abc import x
(7 + 3*x).as_coeff_add()
(3*x,))
(7*x).as_coeff_add()
(7*x,))
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
as coefficients dict(a)
Return a dictionary mapping terms to their Rational coecient. Since the dictionary is a defaultdict, inquiries about terms which were not present will return a
coecient of 0. If an expression is not an Add it is considered to have a single term.
Examples
>>> from sympy.abc import a, x
>>> (3*x + a*x + 4).as_coefficients_dict()
{1: 4, x: 3, a*x: 1}
>>> _[a]
0
>>> (3*a*x).as_coefficients_dict()
{a*x: 3}
as content primitive(radical=False)
Return the tuple (R, self/R) where R is the positive Rational extracted from self. If
radical is True (default is False) then common radicals will be removed and included
as a factor of the primitive expression.
Examples
>>> from sympy import sqrt
>>> (3 + 3*sqrt(2)).as_content_primitive()
(3, 1 + sqrt(2))
Radical content can also be factored out of the primitive:
>>> (2*sqrt(2) + 4*sqrt(10)).as_content_primitive(radical=True)
(2, sqrt(2)*(1 + 2*sqrt(5)))
See docstring of Expr.as content primitive for more examples.
as real imag(deep=True, **hints)
returns a tuple representing a complex number
Examples
>>> from sympy import I
>>> (7 + 9*I).as_real_imag()
(7, 9)
>>> ((1 + I)/(1 - I)).as_real_imag()
(0, 1)
>>> ((1 + 2*I)*(1 + 3*I)).as_real_imag()
(-5, 5)
as two terms(*args, **kwargs)
Return head and tail of self.
This is the most ecient way to get the head and tail of an expression.
if you want only the head, use self.args[0];
5.1. SymPy Core
145
SymPy Documentation, Release 0.7.6
if you want to process the arguments of the tail then use self.as coef add() which
gives the head and a tuple containing the arguments of the tail when treated as
an Add.
if you want the coecient when self is treated as a Mul then use
self.as coe mul()[0]
>>> from sympy.abc import x, y
>>> (3*x*y).as_two_terms()
(3, x*y)
classmethod class key()
Nice order of classes
extract leading order(*args, **kwargs)
Returns the leading term and its order.
Examples
>>> from sympy.abc import x
>>> (x + 1 + 1/x**5).extract_leading_order(x)
((x**(-5), O(x**(-5))),)
>>> (1 + x).extract_leading_order(x)
((1, O(1)),)
>>> (x + x**2).extract_leading_order(x)
((x, O(x)),)
classmethod flatten(seq)
Takes the sequence seq of nested Adds and returns a atten list.
Returns: (commutative part, noncommutative part, order symbols)
Applies associativity, all terms are commutable with respect to addition.
NB: the removal of 0 is already handled by AssocOp. new
See Also:
sympy.core.mul.Mul.flatten (page 142)
primitive()
Return (R, self/R) where R is the Rational GCD of self.
R is collected only from the leading coecient of each term.
Examples
>>> from sympy.abc import x, y
>>> (2*x + 4*y).primitive()
(2, x + 2*y)
>>> (2*x/3 + 4*y/9).primitive()
(2/9, 3*x + 2*y)
>>> (2*x/3 + 4.2*y).primitive()
(1/3, 2*x + 12.6*y)
146
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
No subprocessing of term factors is performed:
>>> ((2 + 2*x)*x + 2).primitive()
(1, x*(2*x + 2) + 2)
Recursive subprocessing can be done with the as content primitive() method:
>>> ((2 + 2*x)*x + 2).as_content_primitive()
(2, x*(x + 1) + 1)
See also: primitive() function in polytools.py
5.1.14 mod
Mod
class sympy.core.mod.Mod
Represents a modulo operation on symbolic expressions.
Receives two arguments, dividend p and divisor q.
The convention used is the same as Pythons: the remainder always has the same sign
as the divisor.
Examples
>>> from sympy.abc import x, y
>>> x**2 % y
Mod(x**2, y)
>>> _.subs({x: 5, y: 6})
1
5.1.15 relational
Rel
sympy.core.relational.Rel
alias of Relational
Eq
sympy.core.relational.Eq
alias of Equality (page 148)
Ne
sympy.core.relational.Ne
alias of Unequality (page 155)
5.1. SymPy Core
147
SymPy Documentation, Release 0.7.6
Lt
sympy.core.relational.Lt
alias of StrictLessThan (page 159)
Le
sympy.core.relational.Le
alias of LessThan (page 152)
Gt
sympy.core.relational.Gt
alias of StrictGreaterThan (page 156)
Ge
sympy.core.relational.Ge
alias of GreaterThan (page 149)
Equality
class sympy.core.relational.Equality
An equal relation between two objects.
Represents that two objects are equal. If they can be easily shown to be denitively equal
(or unequal), this will reduce to True (or False). Otherwise, the relation is maintained
as an unevaluated Equality object. Use the simplify function on this object for more
nontrivial evaluation of the equality relation.
See Also:
sympy.logic.boolalg.Equivalent (page 631) for representing equality between two
boolean expressions
Notes
This class is not the same as the == operator. The == operator tests for exact structural
equality between two expressions; this class compares expressions mathematically.
If either object denes an e valE q method, it can be used in place of the default algorithm.
If lhs.e valE q(rhs) or rhs.e valE q(lhs) returns anything other than None, that return value
will be substituted for the Equality. If None is returned by e valE q, an Equality object will
be created as usual.
Examples
148
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> from sympy import Eq
>>> from sympy.abc import x, y
>>> Eq(y, x+x**2)
y == x**2 + x
GreaterThan
class sympy.core.relational.GreaterThan
Class representations of inequalities.
The *Than classes represent inequal relationships, where the left-hand side is generally
bigger or smaller than the right-hand side. For example, the GreaterThan class represents an inequal relationship where the left-hand side is at least as big as the right side,
if not bigger. In mathematical notation:
lhs >= rhs
In total, there are four *Than classes, to represent the four inequalities:
Class Name
GreaterThan
LessThan
StrictGreaterThan
StrictLessThan
Symbol
(>=)
(<=)
(>)
(<)
All classes take two arguments, lhs and rhs.
Signature Example
GreaterThan(lhs, rhs)
LessThan(lhs, rhs)
StrictGreaterThan(lhs, rhs)
StrictLessThan(lhs, rhs)
Math equivalent
lhs >= rhs
lhs <= rhs
lhs > rhs
lhs < rhs
In addition to the normal .lhs and .rhs of Relations, *Than inequality objects also have
the .lts and .gts properties, which represent the less than side and greater than side
of the operator. Use of .lts and .gts in an algorithm rather than .lhs and .rhs as an
assumption of inequality direction will make more explicit the intent of a certain section
of code, and will make it similarly more robust to client code changes:
>>>
>>>
>>>
>>>
>>>
from
from
from
from
from
sympy import GreaterThan, StrictGreaterThan
sympy import LessThan,
StrictLessThan
sympy import And, Ge, Gt, Le, Lt, Rel, S
sympy.abc import x, y, z
sympy.core.relational import Relational
>>> e = GreaterThan(x, 1)
>>> e
x >= 1
>>> %s >= %s is the same as %s <= %s % (e.gts, e.lts, e.lts, e.gts)
x >= 1 is the same as 1 <= x
Notes
There are a couple of gotchas when using Pythons operators.
The rst enters the mix when comparing against a literal number as the lhs argument.
Due to the order that Python decides to parse a statement, it may not immediately nd
5.1. SymPy Core
149
SymPy Documentation, Release 0.7.6
two objects comparable. For example, to evaluate the statement (1 < x), Python will rst
recognize the number 1 as a native number, and then that x is not a native number. At
this point, because a native Python number does not know how to compare itself with
a SymPy object Python will try the reective operation, (x > 1). Unfortunately, there is
no way available to SymPy to recognize this has happened, so the statement (1 < x) will
turn silently into (x > 1).
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
x >
x <
x <
x >
e1 = x > 1
e2 = x >= 1
e3 = x < 1
e4 = x <= 1
e5 = 1 > x
e6 = 1 >= x
e7 = 1 < x
e8 = 1 <= x
print(%s
1
x >= 1
1
x <= 1
1
x <= 1
1
x >= 1
%s\n*4 % (e1, e2, e3, e4, e5, e6, e7, e8))
If the order of the statement is important (for visual output to the console, perhaps),
one can work around this annoyance in a couple ways: (1) sympify the literal before
comparison, (2) use one of the wrappers, or (3) use the less succinct methods described
above:
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
1 >
1 <
1 >
1 <
e1 = S(1) > x
e2 = S(1) >= x
e3 = S(1) < x
e4 = S(1) <= x
e5 = Gt(1, x)
e6 = Ge(1, x)
e7 = Lt(1, x)
e8 = Le(1, x)
print(%s
%s\n*4 % (e1, e2, e3, e4, e5, e6, e7, e8))
x
1 >= x
x
1 <= x
x
1 >= x
x
1 <= x
The other gotcha is with chained inequalities. Occasionally, one may be tempted to write
statements like:
>>> e = x < y < z
Traceback (most recent call
...
TypeError: symbolic boolean
Traceback (most recent call
...
TypeError: symbolic boolean
last):
expression has no truth value.
last):
expression has no truth value.
Due to an implementation detail or decision of Python [R48] (page 1899), there is no way
for SymPy to reliably create that as a chained inequality. To create a chained inequality,
the only method currently available is to make use of And:
>>> e = And(x < y, y < z)
>>> type( e )
And
>>> e
150
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
And(x < y, y < z)
Note that this is dierent than chaining an equality directly via use of parenthesis (this
is currently an open bug in SymPy [R49] (page 1900)):
>>> e = (x < y) < z
>>> type( e )
<class sympy.core.relational.StrictLessThan>
>>> e
(x < y) < z
Any code that explicitly relies on this latter functionality will not be robust as this behaviour is completely wrong and will be corrected at some point. For the time being
(circa Jan 2012), use And to create chained inequalities.
Examples
One generally does not instantiate these classes directly, but uses various convenience
methods:
>>> e1 = Ge( x, 2 )
>>> print(e1)
x >= 2
# Ge is a convenience wrapper
>>> rels = Ge( x, 2 ), Gt( x, 2 ), Le( x, 2 ), Lt( x, 2 )
>>> print(%s\n%s\n%s\n%s % rels)
x >= 2
x > 2
x <= 2
x < 2
Another option is to use the Python inequality operators (>=, >, <=, <) directly. Their
main advantage over the Ge, Gt, Le, and Lt counterparts, is that one can write a more
mathematical looking statement rather than littering the math with oddball function
calls. However there are certain (minor) caveats of which to be aware (search for
gotcha, below).
>>> e2 = x >= 2
>>> print(e2)
x >= 2
>>> print(e1: %s,
e2: %s % (e1, e2))
e1: x >= 2,
e2: x >= 2
>>> e1 == e2
True
However, it is also perfectly valid to instantiate a *Than class less succinctly and less
conveniently:
>>> rels = Rel(x, 1, >=), Relational(x, 1, >=), GreaterThan(x, 1)
>>> print(%s\n%s\n%s % rels)
x >= 1
x >= 1
x >= 1
>>> rels = Rel(x, 1, >), Relational(x, 1, >), StrictGreaterThan(x, 1)
>>> print(%s\n%s\n%s % rels)
x > 1
5.1. SymPy Core
151
SymPy Documentation, Release 0.7.6
x > 1
x > 1
>>> rels = Rel(x, 1, <=), Relational(x, 1, <=), LessThan(x, 1)
>>> print(%s\n%s\n%s % rels)
x <= 1
x <= 1
x <= 1
>>>
>>>
x <
x <
x <
rels = Rel(x, 1, <), Relational(x, 1, <), StrictLessThan(x, 1)
print(%s\n%s\n%s % rels)
1
1
1
LessThan
class sympy.core.relational.LessThan
Class representations of inequalities.
The *Than classes represent inequal relationships, where the left-hand side is generally
bigger or smaller than the right-hand side. For example, the GreaterThan class represents an inequal relationship where the left-hand side is at least as big as the right side,
if not bigger. In mathematical notation:
lhs >= rhs
In total, there are four *Than classes, to represent the four inequalities:
Class Name
GreaterThan
LessThan
StrictGreaterThan
StrictLessThan
Symbol
(>=)
(<=)
(>)
(<)
All classes take two arguments, lhs and rhs.
Signature Example
GreaterThan(lhs, rhs)
LessThan(lhs, rhs)
StrictGreaterThan(lhs, rhs)
StrictLessThan(lhs, rhs)
Math equivalent
lhs >= rhs
lhs <= rhs
lhs > rhs
lhs < rhs
In addition to the normal .lhs and .rhs of Relations, *Than inequality objects also have
the .lts and .gts properties, which represent the less than side and greater than side
of the operator. Use of .lts and .gts in an algorithm rather than .lhs and .rhs as an
assumption of inequality direction will make more explicit the intent of a certain section
of code, and will make it similarly more robust to client code changes:
>>>
>>>
>>>
>>>
>>>
from
from
from
from
from
sympy import GreaterThan, StrictGreaterThan
sympy import LessThan,
StrictLessThan
sympy import And, Ge, Gt, Le, Lt, Rel, S
sympy.abc import x, y, z
sympy.core.relational import Relational
>>> e = GreaterThan(x, 1)
>>> e
x >= 1
152
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> %s >= %s is the same as %s <= %s % (e.gts, e.lts, e.lts, e.gts)
x >= 1 is the same as 1 <= x
Notes
There are a couple of gotchas when using Pythons operators.
The rst enters the mix when comparing against a literal number as the lhs argument.
Due to the order that Python decides to parse a statement, it may not immediately nd
two objects comparable. For example, to evaluate the statement (1 < x), Python will rst
recognize the number 1 as a native number, and then that x is not a native number. At
this point, because a native Python number does not know how to compare itself with
a SymPy object Python will try the reective operation, (x > 1). Unfortunately, there is
no way available to SymPy to recognize this has happened, so the statement (1 < x) will
turn silently into (x > 1).
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
x >
x <
x <
x >
e1 = x > 1
e2 = x >= 1
e3 = x < 1
e4 = x <= 1
e5 = 1 > x
e6 = 1 >= x
e7 = 1 < x
e8 = 1 <= x
print(%s
1
x >= 1
1
x <= 1
1
x <= 1
1
x >= 1
%s\n*4 % (e1, e2, e3, e4, e5, e6, e7, e8))
If the order of the statement is important (for visual output to the console, perhaps),
one can work around this annoyance in a couple ways: (1) sympify the literal before
comparison, (2) use one of the wrappers, or (3) use the less succinct methods described
above:
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
1 >
1 <
1 >
1 <
e1 = S(1) > x
e2 = S(1) >= x
e3 = S(1) < x
e4 = S(1) <= x
e5 = Gt(1, x)
e6 = Ge(1, x)
e7 = Lt(1, x)
e8 = Le(1, x)
print(%s
%s\n*4 % (e1, e2, e3, e4, e5, e6, e7, e8))
x
1 >= x
x
1 <= x
x
1 >= x
x
1 <= x
The other gotcha is with chained inequalities. Occasionally, one may be tempted to write
statements like:
>>> e = x < y < z
Traceback (most recent call last):
...
TypeError: symbolic boolean expression has no truth value.
Traceback (most recent call last):
5.1. SymPy Core
153
SymPy Documentation, Release 0.7.6
...
TypeError: symbolic boolean expression has no truth value.
Due to an implementation detail or decision of Python [R50] (page 1900), there is no way
for SymPy to reliably create that as a chained inequality. To create a chained inequality,
the only method currently available is to make use of And:
>>> e = And(x < y, y < z)
>>> type( e )
And
>>> e
And(x < y, y < z)
Note that this is dierent than chaining an equality directly via use of parenthesis (this
is currently an open bug in SymPy [R51] (page 1900)):
>>> e = (x < y) < z
>>> type( e )
<class sympy.core.relational.StrictLessThan>
>>> e
(x < y) < z
Any code that explicitly relies on this latter functionality will not be robust as this behaviour is completely wrong and will be corrected at some point. For the time being
(circa Jan 2012), use And to create chained inequalities.
Examples
One generally does not instantiate these classes directly, but uses various convenience
methods:
>>> e1 = Ge( x, 2 )
>>> print(e1)
x >= 2
# Ge is a convenience wrapper
>>> rels = Ge( x, 2 ), Gt( x, 2 ), Le( x, 2 ), Lt( x, 2 )
>>> print(%s\n%s\n%s\n%s % rels)
x >= 2
x > 2
x <= 2
x < 2
Another option is to use the Python inequality operators (>=, >, <=, <) directly. Their
main advantage over the Ge, Gt, Le, and Lt counterparts, is that one can write a more
mathematical looking statement rather than littering the math with oddball function
calls. However there are certain (minor) caveats of which to be aware (search for
gotcha, below).
>>> e2 = x >= 2
>>> print(e2)
x >= 2
>>> print(e1: %s,
e2: %s % (e1, e2))
e1: x >= 2,
e2: x >= 2
>>> e1 == e2
True
However, it is also perfectly valid to instantiate a *Than class less succinctly and less
conveniently:
154
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> rels = Rel(x, 1, >=), Relational(x, 1, >=), GreaterThan(x, 1)
>>> print(%s\n%s\n%s % rels)
x >= 1
x >= 1
x >= 1
>>>
>>>
x >
x >
x >
rels = Rel(x, 1, >), Relational(x, 1, >), StrictGreaterThan(x, 1)
print(%s\n%s\n%s % rels)
1
1
1
>>> rels = Rel(x, 1, <=), Relational(x, 1, <=), LessThan(x, 1)
>>> print(%s\n%s\n%s % rels)
x <= 1
x <= 1
x <= 1
>>>
>>>
x <
x <
x <
rels = Rel(x, 1, <), Relational(x, 1, <), StrictLessThan(x, 1)
print(%s\n%s\n%s % rels)
1
1
1
Unequality
class sympy.core.relational.Unequality
An unequal relation between two objects.
Represents that two objects are not equal. If they can be shown to be denitively equal,
this will reduce to False; if denitively unequal, this will reduce to True. Otherwise, the
relation is maintained as an Unequality object.
See Also:
Equality (page 148)
Notes
This class is not the same as the != operator. The != operator tests for exact structural
equality between two expressions; this class compares expressions mathematically.
This class is eectively the inverse of Equality. As such, it uses the same algorithms,
including any available e valE q methods.
Examples
>>> from sympy import Ne
>>> from sympy.abc import x, y
>>> Ne(y, x+x**2)
y != x**2 + x
5.1. SymPy Core
155
SymPy Documentation, Release 0.7.6
StrictGreaterThan
class sympy.core.relational.StrictGreaterThan
Class representations of inequalities.
The *Than classes represent inequal relationships, where the left-hand side is generally
bigger or smaller than the right-hand side. For example, the GreaterThan class represents an inequal relationship where the left-hand side is at least as big as the right side,
if not bigger. In mathematical notation:
lhs >= rhs
In total, there are four *Than classes, to represent the four inequalities:
Class Name
GreaterThan
LessThan
StrictGreaterThan
StrictLessThan
Symbol
(>=)
(<=)
(>)
(<)
All classes take two arguments, lhs and rhs.
Signature Example
GreaterThan(lhs, rhs)
LessThan(lhs, rhs)
StrictGreaterThan(lhs, rhs)
StrictLessThan(lhs, rhs)
Math equivalent
lhs >= rhs
lhs <= rhs
lhs > rhs
lhs < rhs
In addition to the normal .lhs and .rhs of Relations, *Than inequality objects also have
the .lts and .gts properties, which represent the less than side and greater than side
of the operator. Use of .lts and .gts in an algorithm rather than .lhs and .rhs as an
assumption of inequality direction will make more explicit the intent of a certain section
of code, and will make it similarly more robust to client code changes:
>>>
>>>
>>>
>>>
>>>
from
from
from
from
from
sympy import GreaterThan, StrictGreaterThan
sympy import LessThan,
StrictLessThan
sympy import And, Ge, Gt, Le, Lt, Rel, S
sympy.abc import x, y, z
sympy.core.relational import Relational
>>> e = GreaterThan(x, 1)
>>> e
x >= 1
>>> %s >= %s is the same as %s <= %s % (e.gts, e.lts, e.lts, e.gts)
x >= 1 is the same as 1 <= x
Notes
There are a couple of gotchas when using Pythons operators.
The rst enters the mix when comparing against a literal number as the lhs argument.
Due to the order that Python decides to parse a statement, it may not immediately nd
two objects comparable. For example, to evaluate the statement (1 < x), Python will rst
recognize the number 1 as a native number, and then that x is not a native number. At
this point, because a native Python number does not know how to compare itself with
a SymPy object Python will try the reective operation, (x > 1). Unfortunately, there is
no way available to SymPy to recognize this has happened, so the statement (1 < x) will
turn silently into (x > 1).
156
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
x >
x <
x <
x >
e1 = x > 1
e2 = x >= 1
e3 = x < 1
e4 = x <= 1
e5 = 1 > x
e6 = 1 >= x
e7 = 1 < x
e8 = 1 <= x
print(%s
1
x >= 1
1
x <= 1
1
x <= 1
1
x >= 1
%s\n*4 % (e1, e2, e3, e4, e5, e6, e7, e8))
If the order of the statement is important (for visual output to the console, perhaps),
one can work around this annoyance in a couple ways: (1) sympify the literal before
comparison, (2) use one of the wrappers, or (3) use the less succinct methods described
above:
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
1 >
1 <
1 >
1 <
e1 = S(1) > x
e2 = S(1) >= x
e3 = S(1) < x
e4 = S(1) <= x
e5 = Gt(1, x)
e6 = Ge(1, x)
e7 = Lt(1, x)
e8 = Le(1, x)
print(%s
%s\n*4 % (e1, e2, e3, e4, e5, e6, e7, e8))
x
1 >= x
x
1 <= x
x
1 >= x
x
1 <= x
The other gotcha is with chained inequalities. Occasionally, one may be tempted to write
statements like:
>>> e = x < y < z
Traceback (most recent call
...
TypeError: symbolic boolean
Traceback (most recent call
...
TypeError: symbolic boolean
last):
expression has no truth value.
last):
expression has no truth value.
Due to an implementation detail or decision of Python [R52] (page 1901), there is no way
for SymPy to reliably create that as a chained inequality. To create a chained inequality,
the only method currently available is to make use of And:
>>> e = And(x < y, y < z)
>>> type( e )
And
>>> e
And(x < y, y < z)
Note that this is dierent than chaining an equality directly via use of parenthesis (this
is currently an open bug in SymPy [R53] (page 1901)):
>>> e = (x < y) < z
>>> type( e )
5.1. SymPy Core
157
SymPy Documentation, Release 0.7.6
<class sympy.core.relational.StrictLessThan>
>>> e
(x < y) < z
Any code that explicitly relies on this latter functionality will not be robust as this behaviour is completely wrong and will be corrected at some point. For the time being
(circa Jan 2012), use And to create chained inequalities.
Examples
One generally does not instantiate these classes directly, but uses various convenience
methods:
>>> e1 = Ge( x, 2 )
>>> print(e1)
x >= 2
# Ge is a convenience wrapper
>>> rels = Ge( x, 2 ), Gt( x, 2 ), Le( x, 2 ), Lt( x, 2 )
>>> print(%s\n%s\n%s\n%s % rels)
x >= 2
x > 2
x <= 2
x < 2
Another option is to use the Python inequality operators (>=, >, <=, <) directly. Their
main advantage over the Ge, Gt, Le, and Lt counterparts, is that one can write a more
mathematical looking statement rather than littering the math with oddball function
calls. However there are certain (minor) caveats of which to be aware (search for
gotcha, below).
>>> e2 = x >= 2
>>> print(e2)
x >= 2
>>> print(e1: %s,
e2: %s % (e1, e2))
e1: x >= 2,
e2: x >= 2
>>> e1 == e2
True
However, it is also perfectly valid to instantiate a *Than class less succinctly and less
conveniently:
>>> rels = Rel(x, 1, >=), Relational(x, 1, >=), GreaterThan(x, 1)
>>> print(%s\n%s\n%s % rels)
x >= 1
x >= 1
x >= 1
>>>
>>>
x >
x >
x >
rels = Rel(x, 1, >), Relational(x, 1, >), StrictGreaterThan(x, 1)
print(%s\n%s\n%s % rels)
1
1
1
>>> rels = Rel(x, 1, <=), Relational(x, 1, <=), LessThan(x, 1)
>>> print(%s\n%s\n%s % rels)
x <= 1
158
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
x <= 1
x <= 1
>>>
>>>
x <
x <
x <
rels = Rel(x, 1, <), Relational(x, 1, <), StrictLessThan(x, 1)
print(%s\n%s\n%s % rels)
1
1
1
StrictLessThan
class sympy.core.relational.StrictLessThan
Class representations of inequalities.
The *Than classes represent inequal relationships, where the left-hand side is generally
bigger or smaller than the right-hand side. For example, the GreaterThan class represents an inequal relationship where the left-hand side is at least as big as the right side,
if not bigger. In mathematical notation:
lhs >= rhs
In total, there are four *Than classes, to represent the four inequalities:
Class Name
GreaterThan
LessThan
StrictGreaterThan
StrictLessThan
Symbol
(>=)
(<=)
(>)
(<)
All classes take two arguments, lhs and rhs.
Signature Example
GreaterThan(lhs, rhs)
LessThan(lhs, rhs)
StrictGreaterThan(lhs, rhs)
StrictLessThan(lhs, rhs)
Math equivalent
lhs >= rhs
lhs <= rhs
lhs > rhs
lhs < rhs
In addition to the normal .lhs and .rhs of Relations, *Than inequality objects also have
the .lts and .gts properties, which represent the less than side and greater than side
of the operator. Use of .lts and .gts in an algorithm rather than .lhs and .rhs as an
assumption of inequality direction will make more explicit the intent of a certain section
of code, and will make it similarly more robust to client code changes:
>>>
>>>
>>>
>>>
>>>
from
from
from
from
from
sympy import GreaterThan, StrictGreaterThan
sympy import LessThan,
StrictLessThan
sympy import And, Ge, Gt, Le, Lt, Rel, S
sympy.abc import x, y, z
sympy.core.relational import Relational
>>> e = GreaterThan(x, 1)
>>> e
x >= 1
>>> %s >= %s is the same as %s <= %s % (e.gts, e.lts, e.lts, e.gts)
x >= 1 is the same as 1 <= x
5.1. SymPy Core
159
SymPy Documentation, Release 0.7.6
Notes
There are a couple of gotchas when using Pythons operators.
The rst enters the mix when comparing against a literal number as the lhs argument.
Due to the order that Python decides to parse a statement, it may not immediately nd
two objects comparable. For example, to evaluate the statement (1 < x), Python will rst
recognize the number 1 as a native number, and then that x is not a native number. At
this point, because a native Python number does not know how to compare itself with
a SymPy object Python will try the reective operation, (x > 1). Unfortunately, there is
no way available to SymPy to recognize this has happened, so the statement (1 < x) will
turn silently into (x > 1).
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
x >
x <
x <
x >
e1 = x > 1
e2 = x >= 1
e3 = x < 1
e4 = x <= 1
e5 = 1 > x
e6 = 1 >= x
e7 = 1 < x
e8 = 1 <= x
print(%s
1
x >= 1
1
x <= 1
1
x <= 1
1
x >= 1
%s\n*4 % (e1, e2, e3, e4, e5, e6, e7, e8))
If the order of the statement is important (for visual output to the console, perhaps),
one can work around this annoyance in a couple ways: (1) sympify the literal before
comparison, (2) use one of the wrappers, or (3) use the less succinct methods described
above:
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
1 >
1 <
1 >
1 <
e1 = S(1) > x
e2 = S(1) >= x
e3 = S(1) < x
e4 = S(1) <= x
e5 = Gt(1, x)
e6 = Ge(1, x)
e7 = Lt(1, x)
e8 = Le(1, x)
print(%s
%s\n*4 % (e1, e2, e3, e4, e5, e6, e7, e8))
x
1 >= x
x
1 <= x
x
1 >= x
x
1 <= x
The other gotcha is with chained inequalities. Occasionally, one may be tempted to write
statements like:
>>> e = x < y < z
Traceback (most recent call
...
TypeError: symbolic boolean
Traceback (most recent call
...
TypeError: symbolic boolean
last):
expression has no truth value.
last):
expression has no truth value.
Due to an implementation detail or decision of Python [R54] (page 1901), there is no way
for SymPy to reliably create that as a chained inequality. To create a chained inequality,
160
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
the only method currently available is to make use of And:
>>> e = And(x < y, y < z)
>>> type( e )
And
>>> e
And(x < y, y < z)
Note that this is dierent than chaining an equality directly via use of parenthesis (this
is currently an open bug in SymPy [R55] (page 1902)):
>>> e = (x < y) < z
>>> type( e )
<class sympy.core.relational.StrictLessThan>
>>> e
(x < y) < z
Any code that explicitly relies on this latter functionality will not be robust as this behaviour is completely wrong and will be corrected at some point. For the time being
(circa Jan 2012), use And to create chained inequalities.
Examples
One generally does not instantiate these classes directly, but uses various convenience
methods:
>>> e1 = Ge( x, 2 )
>>> print(e1)
x >= 2
# Ge is a convenience wrapper
>>> rels = Ge( x, 2 ), Gt( x, 2 ), Le( x, 2 ), Lt( x, 2 )
>>> print(%s\n%s\n%s\n%s % rels)
x >= 2
x > 2
x <= 2
x < 2
Another option is to use the Python inequality operators (>=, >, <=, <) directly. Their
main advantage over the Ge, Gt, Le, and Lt counterparts, is that one can write a more
mathematical looking statement rather than littering the math with oddball function
calls. However there are certain (minor) caveats of which to be aware (search for
gotcha, below).
>>> e2 = x >= 2
>>> print(e2)
x >= 2
>>> print(e1: %s,
e2: %s % (e1, e2))
e1: x >= 2,
e2: x >= 2
>>> e1 == e2
True
However, it is also perfectly valid to instantiate a *Than class less succinctly and less
conveniently:
>>> rels = Rel(x, 1, >=), Relational(x, 1, >=), GreaterThan(x, 1)
>>> print(%s\n%s\n%s % rels)
x >= 1
5.1. SymPy Core
161
SymPy Documentation, Release 0.7.6
x >= 1
x >= 1
>>>
>>>
x >
x >
x >
rels = Rel(x, 1, >), Relational(x, 1, >), StrictGreaterThan(x, 1)
print(%s\n%s\n%s % rels)
1
1
1
>>> rels = Rel(x, 1, <=), Relational(x, 1, <=), LessThan(x, 1)
>>> print(%s\n%s\n%s % rels)
x <= 1
x <= 1
x <= 1
>>>
>>>
x <
x <
x <
rels = Rel(x, 1, <), Relational(x, 1, <), StrictLessThan(x, 1)
print(%s\n%s\n%s % rels)
1
1
1
5.1.16 multidimensional
vectorize
class sympy.core.multidimensional.vectorize(*mdargs)
Generalizes a function taking scalars to accept multidimensional arguments.
For example
>>>
>>>
>>>
>>>
from sympy import diff, sin, symbols, Function
from sympy.core.multidimensional import vectorize
x, y, z = symbols(x y z)
f, g, h = list(map(Function, fgh))
>>> @vectorize(0)
... def vsin(x):
...
return sin(x)
>>> vsin([1, x, y])
[sin(1), sin(x), sin(y)]
>>> @vectorize(0, 1)
... def vdiff(f, y):
...
return diff(f, y)
>>> vdiff([f(x, y, z), g(x, y, z), h(x, y, z)], [x, y, z])
[[Derivative(f(x, y, z), x), Derivative(f(x, y, z), y), Derivative(f(x, y, z), z)], [Derivative(
162
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
5.1.17 function
Lambda
class sympy.core.function.Lambda
Lambda(x, expr) represents a lambda function similar to Pythons lambda x: expr. A
function of several variables is written as Lambda((x, y, ...), expr).
A simple example:
>>>
>>>
>>>
>>>
16
from sympy import Lambda
from sympy.abc import x
f = Lambda(x, x**2)
f(4)
For multivariate functions, use:
>>> from sympy.abc import y, z, t
>>> f2 = Lambda((x, y, z, t), x + y**z + t**z)
>>> f2(1, 2, 3, 4)
73
A handy shortcut for lots of arguments:
>>>
>>>
>>>
x +
p = x, y, z
f = Lambda(p, x + y*z)
f(*p)
y*z
expr
The return value of the function
is identity
Return True if this Lambda is an identity function.
variables
The variables used in the internal representation of the function
WildFunction
class sympy.core.function.WildFunction(name, **assumptions)
A WildFunction function matches any function (with its arguments).
Examples
>>> from sympy import WildFunction, Function, cos
>>> from sympy.abc import x, y
>>> F = WildFunction(F)
>>> f = Function(f)
>>> F.nargs
Naturals0()
>>> x.match(F)
>>> F.match(F)
{F_: F_}
>>> f(x).match(F)
{F_: f(x)}
5.1. SymPy Core
163
SymPy Documentation, Release 0.7.6
>>> cos(x).match(F)
{F_: cos(x)}
>>> f(x, y).match(F)
{F_: f(x, y)}
To match functions with a given number of arguments, set nargs to the desired value at
instantiation:
>>> F = WildFunction(F, nargs=2)
>>> F.nargs
{2}
>>> f(x).match(F)
>>> f(x, y).match(F)
{F_: f(x, y)}
To match functions with a range of arguments, set nargs to a tuple containing the desired
number of arguments, e.g. if nargs = (1, 2) then functions with 1 or 2 arguments will
be matched.
>>> F = WildFunction(F, nargs=(1, 2))
>>> F.nargs
{1, 2}
>>> f(x).match(F)
{F_: f(x)}
>>> f(x, y).match(F)
{F_: f(x, y)}
>>> f(x, y, 1).match(F)
Derivative
class sympy.core.function.Derivative
Carries out dierentiation of the given expression with respect to symbols.
expr must dene . eval derivative(symbol) method that returns the dierentiation result.
This function only needs to consider the non-trivial case where expr contains symbol and
it should call the di() method internally (not eval derivative); Derivative should be the
only one to call eval derivative.
Simplication of high-order derivatives:
Because there can be a signicant amount of simplication that can be done when multiple dierentiations are performed, results will be automatically simplied in a fairly
conservative fashion unless the keyword simplify is set to False.
>>>
>>>
>>>
>>>
136
>>>
30
from sympy import sqrt, diff
from sympy.abc import x
e = sqrt((x + 1)**2 + x)
diff(e, x, 5, simplify=False).count_ops()
diff(e, x, 5).count_ops()
Ordering of variables:
If evaluate is set to True and the expression can not be evaluated, the list of dierentiation symbols will be sorted, that is, the expression is assumed to have continuous
derivatives up to the order asked. This sorting assumes that derivatives wrt Symbols
164
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
commute, derivatives wrt non-Symbols commute, but Symbol and non-Symbol derivatives dont commute with each other.
Derivative wrt non-Symbols:
This class also allows derivatives wrt non-Symbols that have di wrt set to True, such
as Function and Derivative. When a derivative wrt a non- Symbol is attempted, the nonSymbol is temporarily converted to a Symbol while the dierentiation is performed.
Note that this may seem strange, that Derivative allows things like f(g(x)).di(g(x)), or
even f(cos(x)).di(cos(x)). The motivation for allowing this syntax is to make it easier
to work with variational calculus (i.e., the Euler-Lagrange method). The best way to
understand this is that the action of derivative with respect to a non-Symbol is dened
by the above description: the object is substituted for a Symbol and the derivative is
taken with respect to that. This action is only allowed for objects for which this can be
done unambiguously, for example Function and Derivative objects. Note that this leads
to what may appear to be mathematically inconsistent results. For example:
>>>
>>>
>>>
2
>>>
0
from sympy import cos, sin, sqrt
from sympy.abc import x
(2*cos(x)).diff(cos(x))
(2*sqrt(1 - sin(x)**2)).diff(cos(x))
This appears wrong because in fact 2*cos(x) and 2*sqrt(1 - sin(x)**2) are identically
equal. However this is the wrong way to think of this. Think of it instead as if we have
something like this:
>>> from sympy.abc import c, s
>>> def F(u):
...
return 2*u
...
>>> def G(u):
...
return 2*sqrt(1 - u**2)
...
>>> F(cos(x))
2*cos(x)
>>> G(sin(x))
2*sqrt(-sin(x)**2 + 1)
>>> F(c).diff(c)
2
>>> F(c).diff(c)
2
>>> G(s).diff(c)
0
>>> G(sin(x)).diff(cos(x))
0
Here, the Symbols c and s act just like the functions cos(x) and sin(x), respectively.
Think of 2*cos(x) as f(c).subs(c, cos(x)) (or f(c) at c = cos(x)) and 2*sqrt(1 - sin(x)**2)
as g(s).subs(s, sin(x)) (or g(s) at s = sin(x)), where f(u) == 2*u and g(u) == 2*sqrt(1
- u**2). Here, we dene the function rst and evaluate it at the function, but we can
actually unambiguously do this in reverse in SymPy, because expr.subs(Function, Symbol) is well-dened: just structurally replace the function everywhere it appears in the
expression.
This is actually the same notational convenience used in the Euler-Lagrange method
when one says F(t, f(t), f(t)).di(f(t)). What is actually meant is that the expression
5.1. SymPy Core
165
SymPy Documentation, Release 0.7.6
in question is represented by some F(t, u, v) at u = f(t) and v = f(t), and F(t, f(t),
f(t)).di(f(t)) simply means F(t, u, v).di(u) at u = f(t).
We do not allow derivatives to be taken with respect to expressions where this is not
so well dened. For example, we do not allow expr.di(x*y) because there are multiple
ways of structurally dening where x*y appears in an expression, some of which may
surprise the reader (for example, a very strict denition would have that (x*y*z).di(x*y)
== 0).
>>> from sympy.abc import x, y, z
>>> (x*y*z).diff(x*y)
Traceback (most recent call last):
...
ValueError: Cant differentiate wrt the variable: x*y, 1
Traceback (most recent call last):
...
ValueError: Cant differentiate wrt the variable: x*y, 1
Note that this denition also ts in nicely with the denition of the chain rule. Note how
the chain rule in SymPy is dened using unevaluated Subs objects:
>>> from sympy import symbols, Function
>>> f, g = symbols(f g, cls=Function)
>>> f(2*g(x)).diff(x)
2*Derivative(g(x), x)*Subs(Derivative(f(_xi_1), _xi_1),
(_xi_1,), (2*g(x),))
>>> f(g(x)).diff(x)
Derivative(g(x), x)*Subs(Derivative(f(_xi_1), _xi_1),
(_xi_1,), (g(x),))
Finally, note that, to be consistent with variational calculus, and to ensure that the denition of substituting a Function for a Symbol in an expression is well-dened, derivatives
of functions are assumed to not be related to the function. In other words, we have:
>>> from sympy import diff
>>> diff(f(x), x).diff(f(x))
0
The same is actually true for derivatives of dierent orders:
>>> diff(f(x), x, 2).diff(diff(f(x), x, 1))
0
>>> diff(f(x), x, 1).diff(diff(f(x), x, 2))
0
Note, any class can allow derivatives to be taken with respect to itself. See the docstring
of Expr. di wrt.
Examples
Some basic examples:
>>>
>>>
>>>
>>>
>>>
166
from sympy import Derivative, Symbol, Function
f = Function(f)
g = Function(g)
x = Symbol(x)
y = Symbol(y)
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> Derivative(x**2, x, evaluate=True)
2*x
>>> Derivative(Derivative(f(x,y), x), y)
Derivative(f(x, y), x, y)
>>> Derivative(f(x), x, 3)
Derivative(f(x), x, x, x)
>>> Derivative(f(x, y), y, x, evaluate=True)
Derivative(f(x, y), x, y)
Now some derivatives wrt functions:
>>> Derivative(f(x)**2, f(x), evaluate=True)
2*f(x)
>>> Derivative(f(g(x)), x, evaluate=True)
Derivative(g(x), x)*Subs(Derivative(f(_xi_1), _xi_1),
(_xi_1,), (g(x),))
doit numerically(a, b)
Evaluate the derivative at z numerically.
When we can represent derivatives at a point, this should be folded into the normal
evalf. For now, we need a special method.
di
sympy.core.function.diff(f, *symbols, **kwargs)
Dierentiate f with respect to symbols.
This is just a wrapper to unify .di() and the Derivative class; its interface is similar to that
of integrate(). You can use the same shortcuts for multiple variables as with Derivative.
For example, di(f(x), x, x, x) and di(f(x), x, 3) both return the third derivative of f(x).
You can pass evaluate=False to get an unevaluated Derivative class. Note that if there
are 0 symbols (such as di(f(x), x, 0), then the result will be the function (the zeroth
derivative), even if evaluate=False.
See Also:
Derivative (page 164)
sympy.geometry.util.idiff computes the derivative implicitly
References
http://reference.wolfram.com/legacy/v5 2/Built-inFunctions/AlgebraicComputation/Calculus/D.html
Examples
>>> from sympy import sin, cos, Function, diff
>>> from sympy.abc import x, y
>>> f = Function(f)
>>> diff(sin(x), x)
cos(x)
>>> diff(f(x), x, x, x)
Derivative(f(x), x, x, x)
5.1. SymPy Core
167
SymPy Documentation, Release 0.7.6
>>> diff(f(x), x, 3)
Derivative(f(x), x, x, x)
>>> diff(sin(x)*cos(y), x, 2, y, 2)
sin(x)*cos(y)
>>> type(diff(sin(x), x))
cos
>>> type(diff(sin(x), x, evaluate=False))
<class sympy.core.function.Derivative>
>>> type(diff(sin(x), x, 0))
sin
>>> type(diff(sin(x), x, 0, evaluate=False))
sin
>>> diff(sin(x))
cos(x)
>>> diff(sin(x*y))
Traceback (most recent call last):
...
ValueError: specify differentiation variables to differentiate sin(x*y)
Traceback (most recent call last):
...
ValueError: specify differentiation variables to differentiate sin(x*y)
Note that diff(sin(x)) syntax is meant only for convenience in interactive sessions and
should be avoided in library code.
FunctionClass
class sympy.core.function.FunctionClass(*args, **kwargs)
Base class for function classes. FunctionClass is a subclass of type.
Use Function(<function name> [ , signature ]) to create undened function classes.
nargs
Return a set of the allowed number of arguments for the function.
Examples
>>> from sympy.core.function import Function
>>> from sympy.abc import x, y
>>> f = Function(f)
If the function can take any number of arguments, the set of whole numbers is returned:
>>> Function(f).nargs
Naturals0()
If the function was initialized to accept one or more arguments, a corresponding set
will be returned:
>>> Function(f, nargs=1).nargs
{1}
>>> Function(f, nargs=(2, 1)).nargs
{1, 2}
168
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
The undened function, after application, also has the nargs attribute; the actual
number of arguments is always available by checking the args attribute:
>>> f = Function(f)
>>> f(1).nargs
Naturals0()
>>> len(f(1).args)
1
Function
class sympy.core.function.Function
Base class for applied mathematical functions.
It also serves as a constructor for undened function classes.
Examples
First example shows how to use Function as a constructor for undened function classes:
>>> from sympy import Function, Symbol
>>> x = Symbol(x)
>>> f = Function(f)
>>> g = Function(g)(x)
>>> f
f
>>> f(x)
f(x)
>>> g
g(x)
>>> f(x).diff(x)
Derivative(f(x), x)
>>> g.diff(x)
Derivative(g(x), x)
In the following example Function is used as a base class for my func that represents a
mathematical function my func. Suppose that it is well known, that my func(0) is 1 and
my func at innity goes to 0, so we want those two simplications to occur automatically.
Suppose also that my func(x) is real exactly when x is real. Here is an implementation
that honours those requirements:
>>> from sympy import Function, S, oo, I, sin
>>> class my_func(Function):
...
...
@classmethod
...
def eval(cls, x):
...
if x.is_Number:
...
if x is S.Zero:
...
return S.One
...
elif x is S.Infinity:
...
return S.Zero
...
...
def _eval_is_real(self):
...
return self.args[0].is_real
...
>>> x = S(x)
5.1. SymPy Core
169
SymPy Documentation, Release 0.7.6
>>> my_func(0) + sin(0)
1
>>> my_func(oo)
0
>>> my_func(3.54).n() # Not yet implemented for my_func.
my_func(3.54)
>>> my_func(I).is_real
False
In order for my func to become useful, several other methods would need to be implemented. See source code of some of the already implemented functions for more complete examples.
Also, if the function can take more than one argument, then nargs must be dened, e.g.
if my func can take one or two arguments then,
>>> class my_func(Function):
...
nargs = (1, 2)
...
>>>
as base exp()
Returns the method as the 2-tuple (base, exponent).
fdiff(argindex=1)
Returns the rst derivative of the function.
is commutative
Returns whether the functon is commutative.
Note: Not all functions are the same
SymPy denes many functions (like cos and factorial). It also allows the user to create
generic functions which act as argument holders. Such functions are created just like symbols:
>>> from sympy import Function, cos
>>> from sympy.abc import x
>>> f = Function(f)
>>> f(2) + f(x)
f(2) + f(x)
If you want to see which functions appear in an expression you can use the atoms method:
>>> e = (f(x) + cos(x) + 2)
>>> e.atoms(Function)
set([f(x), cos(x)])
If you just want the function you dened, not SymPy functions, the thing to search for is
AppliedUndef:
>>> from sympy.core.function import AppliedUndef
>>> e.atoms(AppliedUndef)
set([f(x)])
170
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Subs
class sympy.core.function.Subs
Represents unevaluated substitutions of an expression.
Subs(expr, x, x0) receives 3 arguments: an expression, a variable or list of distinct
variables and a point or list of evaluation points corresponding to those variables.
Subs objects are generally useful to represent unevaluated derivatives calculated at a
point.
The variables may be expressions, but they are subjected to the limitations of subs(), so
it is usually a good practice to use only symbols for variables, since in that case there
can be no ambiguity.
Theres no automatic expansion - use the method .doit() to eect all possible substitutions
of the object and also of objects inside the expression.
When evaluating derivatives at a point that is not a symbol, a Subs object is returned.
One is also able to calculate derivatives of Subs objects - in this case the expression is
always expanded (for the unevaluated form, use Derivative()).
A simple example:
>>> from sympy import Subs, Function, sin
>>> from sympy.abc import x, y, z
>>> f = Function(f)
>>> e = Subs(f(x).diff(x), x, y)
>>> e.subs(y, 0)
Subs(Derivative(f(x), x), (x,), (0,))
>>> e.subs(f, sin).doit()
cos(y)
An example with several variables:
>>> Subs(f(x)*sin(y) + z, (x, y), (0, 1))
Subs(z + f(x)*sin(y), (x, y), (0, 1))
>>> _.doit()
z + f(0)*sin(1)
expr
The expression on which the substitution operates
point
The values for which the variables are to be substituted
variables
The variables to be evaluated
expand
sympy.core.function.expand(e, deep=True, modulus=None, power base=True,
mul=True,
log=True,
multinopower exp=True,
mial=True, basic=True, **hints)
Expand an expression using methods given as hints.
Hints evaluated unless explicitly set to False are: basic, log, multinomial, mul,
power base, and power exp The following hints are supported but not applied unless
set to True: complex, func, and trig. In addition, the following meta-hints are supported by some or all of the other hints: frac, numer, denom, modulus, and force. deep
5.1. SymPy Core
171
SymPy Documentation, Release 0.7.6
is supported by all hints. Additionally, subclasses of Expr may dene their own hints or
meta-hints.
The basic hint is used for any special rewriting of an object that should be done automatically (along with the other hints like mul) when expand is called. This is a catch-all hint
to handle any sort of expansion that may not be described by the existing hint names.
To use this hint an object should override the eval expand basic method. Objects may
also dene their own expand methods, which are not run by default. See the API section
below.
If deep is set to True (the default), things like arguments of functions are recursively
expanded. Use deep=False to only expand on the top level.
If the force hint is used, assumptions about variables will be ignored in making the
expansion.
See Also:
expand log (page 178), expand mul (page 178), expand multinomial (page 179), expand complex (page 179), expand trig (page 179), expand power base (page 180), expand power exp (page 180), expand func (page 178), hyperexpand
Notes
You can shut o unwanted methods:
>>> (exp(x + y)*(x + y)).expand()
x*exp(x)*exp(y) + y*exp(x)*exp(y)
>>> (exp(x + y)*(x + y)).expand(power_exp=False)
x*exp(x + y) + y*exp(x + y)
>>> (exp(x + y)*(x + y)).expand(mul=False)
(x + y)*exp(x)*exp(y)
Use deep=False to only expand on the top level:
>>> exp(x + exp(x + y)).expand()
exp(x)*exp(exp(x)*exp(y))
>>> exp(x + exp(x + y)).expand(deep=False)
exp(x)*exp(exp(x + y))
Hints are applied in an arbitrary, but consistent order (in the current implementation, they are applied in alphabetical order, except multinomial comes before mul,
but this may change). Because of this, some hints may prevent expansion by other
hints if they are applied rst. For example, mul may distribute multiplications and
prevent log and power base from expanding them. Also, if mul is applied before
multinomial, the expression might not be fully distributed. The solution is to use the various expand hint helper functions or to use hint=False
to this function to nely control which hints are applied. Here are some examples:
>>> from sympy import expand, expand_mul, expand_power_base
>>> x, y, z = symbols(x,y,z, positive=True)
>>> expand(log(x*(y + z)))
log(x) + log(y + z)
Here, we see that log was applied before mul. To get the mul expanded form, either
of the following will work:
172
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> expand_mul(log(x*(y + z)))
log(x*y + x*z)
>>> expand(log(x*(y + z)), log=False)
log(x*y + x*z)
A similar thing can happen with the power base hint:
>>> expand((x*(y + z))**x)
(x*y + x*z)**x
To get the power base expanded form, either of the following will work:
>>> expand((x*(y + z))**x, mul=False)
x**x*(y + z)**x
>>> expand_power_base((x*(y + z))**x)
x**x*(y + z)**x
>>> expand((x + y)*y/x)
y + y**2/x
The parts of a rational expression can be targeted:
>>> expand((x + y)*y/x/(x + 1), frac=True)
(x*y + y**2)/(x**2 + x)
>>> expand((x + y)*y/x/(x + 1), numer=True)
(x*y + y**2)/(x*(x + 1))
>>> expand((x + y)*y/x/(x + 1), denom=True)
y*(x + y)/(x**2 + x)
The modulus meta-hint can be used to reduce the coecients of an expression postexpansion:
>>> expand((3*x + 1)**2)
9*x**2 + 6*x + 1
>>> expand((3*x + 1)**2, modulus=5)
4*x**2 + x + 1
Either expand() the function or .expand() the method can be used. Both are equivalent:
>>> expand((x + 1)**2)
x**2 + 2*x + 1
>>> ((x + 1)**2).expand()
x**2 + 2*x + 1
Hints
These hints are run by default
Mul
Distributes multiplication over addition:
>>>
>>>
>>>
x*y
from sympy import cos, exp, sin
from sympy.abc import x, y, z
(y*(x + z)).expand(mul=True)
+ y*z
5.1. SymPy Core
173
SymPy Documentation, Release 0.7.6
Multinomial
Expand (x + y + ...)**n where n is a positive integer.
>>> ((x + y + z)**2).expand(multinomial=True)
x**2 + 2*x*y + 2*x*z + y**2 + 2*y*z + z**2
Power exp
Expand addition in exponents into multiplied bases.
>>> exp(x + y).expand(power_exp=True)
exp(x)*exp(y)
>>> (2**(x + y)).expand(power_exp=True)
2**x*2**y
Power base
Split powers of multiplied bases.
This only happens by default if assumptions allow, or if the force meta-hint is used:
>>> ((x*y)**z).expand(power_base=True)
(x*y)**z
>>> ((x*y)**z).expand(power_base=True, force=True)
x**z*y**z
>>> ((2*y)**z).expand(power_base=True)
2**z*y**z
Note that in some cases where this expansion always holds, SymPy performs it automatically:
>>> (x*y)**2
x**2*y**2
Log
Pull out power of an argument as a coecient and split logs products into sums of logs.
Note that these only work if the arguments of the log function have the proper assumptionsthe arguments must be positive and the exponents must be realor else the force
hint must be True:
>>> from sympy import log, symbols
>>> log(x**2*y).expand(log=True)
log(x**2*y)
>>> log(x**2*y).expand(log=True, force=True)
2*log(x) + log(y)
>>> x, y = symbols(x,y, positive=True)
>>> log(x**2*y).expand(log=True)
2*log(x) + log(y)
174
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Basic
This hint is intended primarily as a way for custom subclasses to enable expansion by
default.
These hints are not run by default:
Complex
Split an expression into real and imaginary parts.
>>> x, y = symbols(x,y)
>>> (x + y).expand(complex=True)
re(x) + re(y) + I*im(x) + I*im(y)
>>> cos(x).expand(complex=True)
-I*sin(re(x))*sinh(im(x)) + cos(re(x))*cosh(im(x))
Note that this is just a wrapper around as real imag(). Most objects that wish to redene eval expand complex() should consider redening as real imag() instead.
Func
Expand other functions.
>>> from sympy import gamma
>>> gamma(x + 1).expand(func=True)
x*gamma(x)
Trig
Do trigonometric expansions.
>>> cos(x + y).expand(trig=True)
-sin(x)*sin(y) + cos(x)*cos(y)
>>> sin(2*x).expand(trig=True)
2*sin(x)*cos(x)
Note that the forms of sin(n*x) and cos(n*x) in terms of sin(x) and cos(x) are not
unique, due to the identity sin2 (x) + cos2 (x) = 1. The current implementation uses the
form obtained from Chebyshev polynomials, but this may change. See this MathWorld
article for more information.
Api
Objects can dene their own expand hints by dening eval expand hint(). The function should take the form:
def _eval_expand_hint(self, **hints):
# Only apply the method to the top-level expression
...
5.1. SymPy Core
175
SymPy Documentation, Release 0.7.6
See also the example below. Objects should dene eval expand hint() methods only
if hint applies to that specic object. The generic eval expand hint() method dened
in Expr will handle the no-op case.
Each hint should be responsible for expanding that hint only. Furthermore, the expansion
should be applied to the top-level expression only. expand() takes care of the recursion
that happens when deep=True.
You should only call eval expand hint() methods directly if you are 100% sure that
the object has the method, as otherwise you are liable to get unexpected AttributeErrors.
Note, again, that you do not need to recursively apply
the hint to args of your object:
this is handled automatically by
expand(). eval expand hint() should generally not be used at all outside of an
eval expand hint() method. If you want to apply a specic expansion from within another method, use the public expand() function, method, or expand hint() functions.
In order for expand to work, objects must be rebuildable by their args, i.e.,
obj.func(*obj.args) == obj must hold.
Expand methods are passed **hints so that expand hints may use metahintshints
that control how dierent expand methods are applied. For example, the force=True
hint described above that causes expand(log=True) to ignore assumptions is such a
metahint. The deep meta-hint is handled exclusively by expand() and is not passed to
eval expand hint() methods.
Note that expansion hints should generally be methods that perform some kind of expansion. For hints that simply rewrite an expression, use the .rewrite() API.
Example
>>> from sympy import Expr, sympify
>>> class MyClass(Expr):
...
def __new__(cls, *args):
...
args = sympify(args)
...
return Expr.__new__(cls, *args)
...
...
def _eval_expand_double(self, **hints):
...
...
Doubles the args of MyClass.
...
...
If there more than four args, doubling is not performed,
...
unless force=True is also used (False by default).
...
...
force = hints.pop(force, False)
...
if not force and len(self.args) > 4:
...
return self
...
return self.func(*(self.args + self.args))
...
>>> a = MyClass(1, 2, MyClass(3, 4))
>>> a
MyClass(1, 2, MyClass(3, 4))
>>> a.expand(double=True)
MyClass(1, 2, MyClass(3, 4, 3, 4), 1, 2, MyClass(3, 4, 3, 4))
>>> a.expand(double=True, deep=False)
MyClass(1, 2, MyClass(3, 4), 1, 2, MyClass(3, 4))
176
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> b = MyClass(1, 2, 3, 4, 5)
>>> b.expand(double=True)
MyClass(1, 2, 3, 4, 5)
>>> b.expand(double=True, force=True)
MyClass(1, 2, 3, 4, 5, 1, 2, 3, 4, 5)
PoleError
class sympy.core.function.PoleError
count ops
sympy.core.function.count ops(expr, visual=False)
Return a representation (integer or expression) of the operations in expr.
If visual is False (default) then the sum of the coecients of the visual expression will
be returned.
If visual is True then the number of each type of operation is shown with the core class
types (or their virtual equivalent) multiplied by the number of times they occur.
If expr is an iterable, the sum of the op counts of the items will be returned.
Examples
>>> from sympy.abc import a, b, x, y
>>> from sympy import sin, count_ops
Although there isnt a SUB object, minus signs are interpreted as either negations or
subtractions:
>>> (x - y).count_ops(visual=True)
SUB
>>> (-x).count_ops(visual=True)
NEG
Here, there are two Adds and a Pow:
>>> (1 + a + b**2).count_ops(visual=True)
2*ADD + POW
In the following, an Add, Mul, Pow and two functions:
>>> (sin(x)*x + sin(x)**2).count_ops(visual=True)
ADD + MUL + POW + 2*SIN
for a total of 5:
>>> (sin(x)*x + sin(x)**2).count_ops(visual=False)
5
Note that what you type is not always what you get. The expression 1/x/y is translated
by sympy into 1/(x*y) so it gives a DIV and MUL rather than two DIVs:
>>> (1/x/y).count_ops(visual=True)
DIV + MUL
5.1. SymPy Core
177
SymPy Documentation, Release 0.7.6
The visual option can be used to demonstrate the dierence in operations for expressions
in dierent forms. Here, the Horner representation is compared with the expanded form
of a polynomial:
>>> eq=x*(1 + x*(2 + x*(3 + x)))
>>> count_ops(eq.expand(), visual=True) - count_ops(eq, visual=True)
-MUL + 3*POW
The count ops function also handles iterables:
>>> count_ops([x, sin(x), None, True, x + 2], visual=False)
2
>>> count_ops([x, sin(x), None, True, x + 2], visual=True)
ADD + SIN
>>> count_ops({x: sin(x), x + 2: y + 1}, visual=True)
2*ADD + SIN
expand mul
sympy.core.function.expand mul(expr, deep=True)
Wrapper around expand that only uses the mul hint. See the expand docstring for more
information.
Examples
>>> from sympy import symbols, expand_mul, exp, log
>>> x, y = symbols(x,y, positive=True)
>>> expand_mul(exp(x+y)*(x+y)*log(x*y**2))
x*exp(x + y)*log(x*y**2) + y*exp(x + y)*log(x*y**2)
expand log
sympy.core.function.expand log(expr, deep=True, force=False)
Wrapper around expand that only uses the log hint. See the expand docstring for more
information.
Examples
>>> from sympy import symbols, expand_log, exp, log
>>> x, y = symbols(x,y, positive=True)
>>> expand_log(exp(x+y)*(x+y)*log(x*y**2))
(x + y)*(log(x) + 2*log(y))*exp(x + y)
expand func
sympy.core.function.expand func(expr, deep=True)
Wrapper around expand that only uses the func hint. See the expand docstring for more
information.
178
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import expand_func, gamma
>>> from sympy.abc import x
>>> expand_func(gamma(x + 2))
x*(x + 1)*gamma(x)
expand trig
sympy.core.function.expand trig(expr, deep=True)
Wrapper around expand that only uses the trig hint. See the expand docstring for more
information.
Examples
>>> from sympy import expand_trig, sin
>>> from sympy.abc import x, y
>>> expand_trig(sin(x+y)*(x+y))
(x + y)*(sin(x)*cos(y) + sin(y)*cos(x))
expand complex
sympy.core.function.expand complex(expr, deep=True)
Wrapper around expand that only uses the complex hint. See the expand docstring for
more information.
See Also:
Expr.as real imag
Examples
>>> from sympy import expand_complex, exp, sqrt, I
>>> from sympy.abc import z
>>> expand_complex(exp(z))
I*exp(re(z))*sin(im(z)) + exp(re(z))*cos(im(z))
>>> expand_complex(sqrt(I))
sqrt(2)/2 + sqrt(2)*I/2
expand multinomial
sympy.core.function.expand multinomial(expr, deep=True)
Wrapper around expand that only uses the multinomial hint. See the expand docstring
for more information.
5.1. SymPy Core
179
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import symbols, expand_multinomial, exp
>>> x, y = symbols(x y, positive=True)
>>> expand_multinomial((x + exp(x + 1))**2)
x**2 + 2*x*exp(x + 1) + exp(2*x + 2)
expand power exp
sympy.core.function.expand power exp(expr, deep=True)
Wrapper around expand that only uses the power exp hint.
See the expand docstring for more information.
Examples
>>> from sympy import expand_power_exp
>>> from sympy.abc import x, y
>>> expand_power_exp(x**(y + 2))
x**2*x**y
expand power base
sympy.core.function.expand power base(expr, deep=True, force=False)
Wrapper around expand that only uses the power base hint.
See the expand docstring for more information.
A wrapper to expand(power base=True) which separates a power with a base that is a
Mul into a product of powers, without performing any other expansions, provided that
assumptions about the powers base and exponent allow.
deep=False (default is True) will only apply to the top-level expression.
force=True (default is False) will cause the expansion to ignore assumptions about the
base and exponent. When False, the expansion will only happen if the base is nonnegative or the exponent is an integer.
>>> from sympy.abc import x, y, z
>>> from sympy import expand_power_base, sin, cos, exp
>>> (x*y)**2
x**2*y**2
>>> (2*x)**y
(2*x)**y
>>> expand_power_base(_)
2**y*x**y
>>> expand_power_base((x*y)**z)
(x*y)**z
>>> expand_power_base((x*y)**z, force=True)
x**z*y**z
>>> expand_power_base(sin((x*y)**z), deep=False)
180
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
sin((x*y)**z)
>>> expand_power_base(sin((x*y)**z), force=True)
sin(x**z*y**z)
>>> expand_power_base((2*sin(x))**y + (2*cos(x))**y)
2**y*sin(x)**y + 2**y*cos(x)**y
>>> expand_power_base((2*exp(y))**x)
2**x*exp(y)**x
>>> expand_power_base((2*cos(x))**y)
2**y*cos(x)**y
Notice that sums are left untouched. If this is not the desired behavior, apply full expand() to the expression:
>>> expand_power_base(((x+y)*z)**2)
z**2*(x + y)**2
>>> (((x+y)*z)**2).expand()
x**2*z**2 + 2*x*y*z**2 + y**2*z**2
>>> expand_power_base((2*y)**(1+z))
2**(z + 1)*y**(z + 1)
>>> ((2*y)**(1+z)).expand()
2*2**z*y*y**z
noat
sympy.core.function.nfloat(expr, n=15, exponent=False)
Make all Rationals in expr Floats except those in exponents (unless the exponents ag
is set to True).
Examples
>>> from sympy.core.function import nfloat
>>> from sympy.abc import x, y
>>> from sympy import cos, pi, sqrt
>>> nfloat(x**4 + x/2 + cos(pi/3) + 1 + sqrt(y))
x**4 + 0.5*x + sqrt(y) + 1.5
>>> nfloat(x**4 + sqrt(y), exponent=True)
x**4.0 + y**0.5
5.1.18 evalf
PrecisionExhausted
class sympy.core.evalf.PrecisionExhausted
5.1. SymPy Core
181
SymPy Documentation, Release 0.7.6
N
class sympy.core.evalf.N
Calls x.evalf(n, **options).
Both .n() and N() are equivalent to .evalf(); use the one that you like better. See also the
docstring of .evalf() for information on the options.
Examples
>>> from sympy import Sum, oo, N
>>> from sympy.abc import k
>>> Sum(1/k**k, (k, 1, oo))
Sum(k**(-k), (k, 1, oo))
>>> N(_, 4)
1.291
5.1.19 containers
Tuple
class sympy.core.containers.Tuple
Wrapper around the builtin tuple object
The Tuple is a subclass of Basic, so that it works well in the SymPy framework. The
wrapped tuple is available as self.args, but you can also access elements or slices with
[:] syntax.
>>>
>>>
>>>
>>>
(b,
>>>
(d,
from sympy import symbols
from sympy.core.containers import Tuple
a, b, c, d = symbols(a b c d)
Tuple(a, b, c)[1:]
c)
Tuple(a, b, c).subs(a, d)
b, c)
index(value[, start[, stop ]]) integer return rst index of value.
Raises ValueError if the value is not present.
tuple count(value)
T.count(value) -> integer return number of occurrences of value
Dict
class sympy.core.containers.Dict
Wrapper around the builtin dict object
The Dict is a subclass of Basic, so that it works well in the SymPy framework. Because it
is immutable, it may be included in sets, but its values must all be given at instantiation
and cannot be changed afterwards. Otherwise it behaves identically to the Python dict.
>>> from sympy.core.containers import Dict
182
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> D = Dict({1: one, 2: two})
>>> for key in D:
...
if key == 1:
...
print(%s %s % (key, D[key]))
1 one
The args are sympied so the 1 and 2 are Integers and the values are Symbols. Queries
automatically sympify args so the following work:
>>> 1 in D
True
>>> D.has(one) # searches keys and values
True
>>> one in D # not in the keys
False
>>> D[1]
one
get(k[, d ]) D[k] if k in D, else d. d defaults to None.
items() list of Ds (key, value) pairs, as 2-tuples
keys() list of Ds keys
values() list of Ds values
5.1.20 compatibility
iterable
sympy.core.compatibility.iterable(i,
exclude=((<type
str>,
<type
unicode>),
<type
dict>,
<class
sympy.core.compatibility.NotIterable
at
0x9906bfc>))
Return a boolean indicating whether i is SymPy iterable. True also indicates that the
iterator is nite, i.e. you e.g. call list(...) on the instance.
When SymPy is working with iterables, it is almost always assuming that the iterable is
not a string or a mapping, so those are excluded by default. If you want a pure Python
denition, make exclude=None. To exclude multiple items, pass them as a tuple.
See also: is sequence
Examples
>>> from sympy.utilities.iterables import iterable
>>> from sympy import Tuple
>>> things = [[1], (1,), set([1]), Tuple(1), (j for j in [1, 2]), {1:2}, 1, 1]
>>> for i in things:
...
print(%s %s % (iterable(i), type(i)))
True <... list>
True <... tuple>
True <... set>
True <class sympy.core.containers.Tuple>
True <... generator>
False <... dict>
5.1. SymPy Core
183
SymPy Documentation, Release 0.7.6
False <... str>
False <... int>
>>> iterable({}, exclude=None)
True
>>> iterable({}, exclude=str)
True
>>> iterable(no, exclude=str)
False
is sequence
sympy.core.compatibility.is sequence(i, include=None)
Return a boolean indicating whether i is a sequence in the SymPy sense. If anything
that fails the test below should be included as being a sequence for your application, set
include to that objects type; multiple types should be passed as a tuple of types.
Note: although generators can generate a sequence, they often need special handling to
make sure their elements are captured before the generator is exhausted, so these are
not included by default in the denition of a sequence.
See also: iterable
Examples
>>> from sympy.utilities.iterables import is_sequence
>>> from types import GeneratorType
>>> is_sequence([])
True
>>> is_sequence(set())
False
>>> is_sequence(abc)
False
>>> is_sequence(abc, include=str)
True
>>> generator = (c for c in abc)
>>> is_sequence(generator)
False
>>> is_sequence(generator, include=(str, GeneratorType))
True
as int
sympy.core.compatibility.as int(n)
Convert the argument to a builtin integer.
The return value is guaranteed to be equal to the input. ValueError is raised if the input
has a non-integral value.
184
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.core.compatibility import as_int
>>> from sympy import sqrt
>>> 3.0
3.0
>>> as_int(3.0) # convert to int and test for equality
3
>>> int(sqrt(10))
3
>>> as_int(sqrt(10))
Traceback (most recent call last):
...
ValueError: ... is not an integer
Traceback (most recent call last):
...
ValueError: ... is not an integer
5.1.21 exprtools
gcd terms
sympy.core.exprtools.gcd terms(terms,
isprimitive=False,
tion=True)
Compute the GCD of terms and put them together.
clear=True,
frac-
terms can be an expression or a non-Basic sequence of expressions which will be handled
as though they are terms from a sum.
If isprimitive is True the gcd terms will not run the primitive method on the terms.
clear controls the removal of integers from the denominator of an Add expression. When
True (default), all numerical denominator will be cleared; when False the denominators
will be cleared only if all terms had numerical denominators other than 1.
fraction, when True (default), will put the expression over a common denominator.
See Also:
factor terms (page 186), sympy.polys.polytools.terms gcd (page 1065)
Examples
>>> from sympy.core import gcd_terms
>>> from sympy.abc import x, y
>>> gcd_terms((x + 1)**2*y + (x + 1)*y**2)
y*(x + 1)*(x + y + 1)
>>> gcd_terms(x/2 + 1)
(x + 2)/2
>>> gcd_terms(x/2 + 1, clear=False)
x/2 + 1
>>> gcd_terms(x/2 + y/2, clear=False)
(x + y)/2
>>> gcd_terms(x/2 + 1/x)
(x**2 + 2)/(2*x)
5.1. SymPy Core
185
SymPy Documentation, Release 0.7.6
>>> gcd_terms(x/2 + 1/x, fraction=False)
(x + 2/x)/2
>>> gcd_terms(x/2 + 1/x, fraction=False, clear=False)
x/2 + 1/x
>>> gcd_terms(x/2/y + 1/x/y)
(x**2 + 2)/(2*x*y)
>>> gcd_terms(x/2/y + 1/x/y, fraction=False, clear=False)
(x + 2/x)/(2*y)
The clear ag was ignored in this case because the returned expression was a rational
expression, not a simple sum.
factor terms
sympy.core.exprtools.factor terms(expr,
radical=False,
clear=False,
fraction=False, sign=True)
Remove common factors from terms in all arguments without changing the underlying structure of the expr. No expansion or simplication (and no processing of noncommutatives) is performed.
If radical=True then a radical common to all terms will be factored out of any Add subexpressions of the expr.
If clear=False (default) then coecients will not be separated from a single Add if they
can be distributed to leave one or more terms with integer coecients.
If fraction=True (default is False) then a common denominator will be constructed for
the expression.
If sign=True (default) then even if the only factor in common is a -1, it will be factored
out of the expression.
See Also:
gcd terms (page 185), sympy.polys.polytools.terms gcd (page 1065)
Examples
>>> from sympy import factor_terms, Symbol
>>> from sympy.abc import x, y
>>> factor_terms(x + x*(2 + 4*y)**3)
x*(8*(2*y + 1)**3 + 1)
>>> A = Symbol(A, commutative=False)
>>> factor_terms(x*A + x*A + x*y*A)
x*(y*A + 2*A)
When clear is False, a rational will only be factored out of an Add expression if all terms
of the Add have coecients that are fractions:
>>> factor_terms(x/2 + 1, clear=False)
x/2 + 1
>>> factor_terms(x/2 + 1, clear=True)
(x + 2)/2
This only applies when there is a single Add that the coecient multiplies:
186
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> factor_terms(x*y/2 + y, clear=True)
y*(x + 2)/2
>>> factor_terms(x*y/2 + y, clear=False) == _
True
If a -1 is all that can be factored out, to not factor it out, the ag sign must be False:
>>> factor_terms(-x - y)
-(x + y)
>>> factor_terms(-x - y, sign=False)
-x - y
>>> factor_terms(-2*x - 2*y, sign=False)
-2*(x + y)
5.2 Combinatorics Module
5.2.1 Contents
Partitions
class sympy.combinatorics.partitions.Partition
This class represents an abstract partition.
A partition is a set of disjoint sets whose union equals a given set.
See Also:
sympy.utilities.iterables.partitions (page 1566), sympy.utilities.iterables.multiset parti
(page 1563)
Attributes
is
is
is
is
Complement
EmptySet
Intersection
UniversalSet
RGS
Returns the restricted growth string of the partition.
The RGS is returned as a list of indices, L, where L[i] indicates the block in which
element i appears. For example, in a partition of 3 elements (a, b, c) into 2 blocks
([c], [a, b]) the RGS is [1, 1, 0]: a is in block 1, b is in block 1 and c is in block
0.
Examples
>>>
>>>
>>>
(1,
>>>
(0,
from sympy.combinatorics.partitions import Partition
a = Partition([1, 2], [3], [4, 5])
a.members
2, 3, 4, 5)
a.RGS
0, 1, 2, 2)
5.2. Combinatorics Module
187
SymPy Documentation, Release 0.7.6
>>> a + 1
{{3}, {4}, {5}, {1, 2}}
>>> _.RGS
(0, 0, 1, 2, 3)
classmethod from rgs(rgs, elements)
Creates a set partition from a restricted growth string.
The indices given in rgs are assumed to be the index of the element as given in
elements as provided (the elements are not sorted by this routine). Block numbering
starts from 0. If any block was not referenced in rgs an error will be raised.
Examples
>>> from sympy.combinatorics.partitions import Partition
>>> Partition.from_rgs([0, 1, 2, 0, 1], list(abcde))
{{c}, {a, d}, {b, e}}
>>> Partition.from_rgs([0, 1, 2, 0, 1], list(cbead))
{{e}, {a, c}, {b, d}}
>>> a = Partition([1, 4], [2], [3, 5])
>>> Partition.from_rgs(a.RGS, a.members)
{{2}, {1, 4}, {3, 5}}
partition
Return partition as a sorted list of lists.
Examples
>>> from sympy.combinatorics.partitions import Partition
>>> Partition([1], [2, 3]).partition
[[1], [2, 3]]
rank
Gets the rank of a partition.
Examples
>>> from sympy.combinatorics.partitions import Partition
>>> a = Partition([1, 2], [3], [4, 5])
>>> a.rank
13
sort key(order=None)
Return a canonical key that can be used for sorting.
Ordering is based on the size and sorted elements of the partition and ties are broken
with the rank.
Examples
188
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> from sympy.utilities.iterables import default_sort_key
>>> from sympy.combinatorics.partitions import Partition
>>> from sympy.abc import x
>>> a = Partition([1, 2])
>>> b = Partition([3, 4])
>>> c = Partition([1, x])
>>> d = Partition(list(range(4)))
>>> l = [d, b, a + 1, a, c]
>>> l.sort(key=default_sort_key); l
[{{1, 2}}, {{1}, {2}}, {{1, x}}, {{3, 4}}, {{0, 1, 2, 3}}]
class sympy.combinatorics.partitions.IntegerPartition
This class represents an integer partition.
In number theory and combinatorics, a partition of a positive integer, n, also called an
integer partition, is a way of writing n as a list of positive integers that sum to n. Two
partitions that dier only in the order of summands are considered to be the same partition; if order matters then the partitions are referred to as compositions. For example,
4 has ve partitions: [4], [3, 1], [2, 2], [2, 1, 1], and [1, 1, 1, 1]; the compositions [1, 2,
1] and [1, 1, 2] are the same as partition [2, 1, 1].
See Also:
sympy.utilities.iterables.partitions (page 1566), sympy.utilities.iterables.multiset parti
(page 1563)
Reference http://en.wikipedia.org/wiki/Partition %28number theory%29
as dict()
Return the partition as a dictionary whose keys are the partition integers and the
values are the multiplicity of that integer.
Examples
>>> from sympy.combinatorics.partitions import IntegerPartition
>>> IntegerPartition([1]*3 + [2] + [3]*4).as_dict()
{1: 3, 2: 1, 3: 4}
as ferrers(char=#)
Prints the ferrer diagram of a partition.
Examples
>>> from sympy.combinatorics.partitions import IntegerPartition
>>> print(IntegerPartition([1, 1, 5]).as_ferrers())
#####
#
#
conjugate
Computes the conjugate partition of itself.
5.2. Combinatorics Module
189
SymPy Documentation, Release 0.7.6
Examples
>>>
>>>
>>>
[5,
from sympy.combinatorics.partitions import IntegerPartition
a = IntegerPartition([6, 3, 3, 2, 1])
a.conjugate
4, 3, 1, 1, 1]
next lex()
Return the next partition of the integer, n, in lexical order, wrapping around to [n]
if the partition is [1, ..., 1].
Examples
>>> from sympy.combinatorics.partitions import IntegerPartition
>>> p = IntegerPartition([3, 1])
>>> print(p.next_lex())
[4]
>>> p.partition < p.next_lex().partition
True
prev lex()
Return the previous partition of the integer, n, in lexical order, wrapping around to
[1, ..., 1] if the partition is [n].
Examples
>>> from sympy.combinatorics.partitions import IntegerPartition
>>> p = IntegerPartition([4])
>>> print(p.prev_lex())
[3, 1]
>>> p.partition > p.prev_lex().partition
True
sympy.combinatorics.partitions.random integer partition(n, seed=None)
Generates a random integer partition summing to n as a list of reverse-sorted integers.
Examples
>>> from sympy.combinatorics.partitions import random_integer_partition
For the following, a seed is given so a known value can be shown; in practice, the seed
would not be given.
>>> random_integer_partition(100, seed=[1, 1, 12, 1, 2, 1, 85, 1])
[85, 12, 2, 1]
>>> random_integer_partition(10, seed=[1, 2, 3, 1, 5, 1])
[5, 3, 1, 1]
>>> random_integer_partition(1)
[1]
sympy.combinatorics.partitions.RGS generalized(m)
Computes the m + 1 generalized unrestricted growth strings and returns them as rows
in matrix.
190
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.combinatorics.partitions import RGS_generalized
>>> RGS_generalized(6)
Matrix([
[ 1,
1,
1, 1, 1, 1, 1],
[ 1,
2,
3, 4, 5, 6, 0],
[ 2,
5, 10, 17, 26, 0, 0],
[ 5, 15, 37, 77, 0, 0, 0],
[ 15, 52, 151, 0, 0, 0, 0],
[ 52, 203,
0, 0, 0, 0, 0],
[203,
0,
0, 0, 0, 0, 0]])
sympy.combinatorics.partitions.RGS enum(m)
RGS enum computes the total number of restricted growth strings possible for a superset
of size m.
Examples
>>>
>>>
>>>
15
>>>
52
>>>
203
from sympy.combinatorics.partitions import RGS_enum
from sympy.combinatorics.partitions import Partition
RGS_enum(4)
RGS_enum(5)
RGS_enum(6)
We can check that the enumeration is correct by actually generating the partitions. Here,
the 15 partitions of 4 items are generated:
>>>
>>>
>>>
...
...
...
>>>
a = Partition(list(range(4)))
s = set()
for i in range(20):
s.add(a)
a += 1
assert len(s) == 15
sympy.combinatorics.partitions.RGS unrank(rank, m)
Gives the unranked restricted growth string for a given superset size.
Examples
>>>
>>>
[0,
>>>
[0,
from sympy.combinatorics.partitions import RGS_unrank
RGS_unrank(14, 4)
1, 2, 3]
RGS_unrank(0, 4)
0, 0, 0]
sympy.combinatorics.partitions.RGS rank(rgs)
Computes the rank of a restricted growth string.
5.2. Combinatorics Module
191
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.combinatorics.partitions import RGS_rank, RGS_unrank
>>> RGS_rank([0, 1, 2, 1, 3])
42
>>> RGS_rank(RGS_unrank(4, 7))
4
Permutations
class sympy.combinatorics.permutations.Permutation
A permutation, alternatively known as an arrangement number or ordering is an arrangement of the elements of an ordered list into a one-to-one mapping with itself. The
permutation of a given arrangement is given by indicating the positions of the elements
after re-arrangement [R8] (page 1902). For example, if one started with elements [x, y,
a, b] (in that order) and they were reordered as [x, y, b, a] then the permutation would
be [0, 1, 3, 2]. Notice that (in SymPy) the rst element is always referred to as 0 and the
permutation uses the indices of the elements in the original ordering, not the elements
(a, b, etc...) themselves.
>>> from sympy.combinatorics import Permutation
>>> Permutation.print_cyclic = False
See Also:
Cycle (page 213)
References
[R7] (page 1902), [R8] (page 1902), [R9] (page 1902), [R10] (page 1902), [R11]
(page 1902), [R12] (page 1902), [R13] (page 1902)
Permutations Notation
Permutations are commonly represented in disjoint cycle or array forms.
Array Notation And 2-line Form
In the 2-line form, the elements and their nal positions are shown as a matrix with 2
rows:
[0 1 2 ... n-1] [p(0) p(1) p(2) ... p(n-1)]
Since the rst line is always range(n), where n is the size of p, it is sucient to represent
the permutation by the second line, referred to as the array form of the permutation.
This is entered in brackets as the argument to the Permutation class:
>>> p = Permutation([0, 2, 1]); p
Permutation([0, 2, 1])
Given i in range(p.size), the permutation maps i to ip
192
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> [i^p for i in range(p.size)]
[0, 2, 1]
The composite of two permutations p*q means rst apply p, then q, so i(p*q) = (ip)q
which is ipq according to Python precedence rules:
>>>
>>>
[2,
>>>
[2,
q = Permutation([2, 1, 0])
[i^p^q for i in range(3)]
0, 1]
[i^(p*q) for i in range(3)]
0, 1]
One can use also the notation p(i) = ip, but then the composition rule is (p*q)(i) = q(p(i)),
not p(q(i)):
>>>
[2,
>>>
[2,
>>>
[1,
[(p*q)(i) for i in range(p.size)]
0, 1]
[q(p(i)) for i in range(p.size)]
0, 1]
[p(q(i)) for i in range(p.size)]
2, 0]
Disjoint Cycle Notation
In disjoint cycle notation, only the elements that have shifted are indicated. In the above
case, the 2 and 1 switched places. This can be entered in two ways:
>>> Permutation(1, 2) == Permutation([[1, 2]]) == p
True
Only the relative ordering of elements in a cycle matter:
>>> Permutation(1,2,3) == Permutation(2,3,1) == Permutation(3,1,2)
True
The disjoint cycle notation is convenient when representing permutations that have several cycles in them:
>>> Permutation(1, 2)(3, 5) == Permutation([[1, 2], [3, 5]])
True
It also provides some economy in entry when computing products of permutations that
are written in disjoint cycle notation:
>>> Permutation(1, 2)(1, 3)(2, 3)
Permutation([0, 3, 2, 1])
>>> _ == Permutation([[1, 2]])*Permutation([[1, 3]])*Permutation([[2, 3]])
True
Entering a singleton in a permutation is a way to indicate the size of the permutation.
The size keyword can also be used.
Array-form entry:
>>> Permutation([[1, 2], [9]])
Permutation([0, 2, 1], size=10)
>>> Permutation([[1, 2]], size=10)
Permutation([0, 2, 1], size=10)
5.2. Combinatorics Module
193
SymPy Documentation, Release 0.7.6
Cyclic-form entry:
>>> Permutation(1, 2, size=10)
Permutation([0, 2, 1], size=10)
>>> Permutation(9)(1, 2)
Permutation([0, 2, 1], size=10)
Caution: no singleton containing an element larger than the largest in any previous cycle
can be entered. This is an important dierence in how Permutation and Cycle handle the
call syntax. A singleton argument at the start of a Permutation performs instantiation
of the Permutation and is permitted:
>>> Permutation(5)
Permutation([], size=6)
A singleton entered after instantiation is a call to the permutation a function call and
if the argument is out of range it will trigger an error. For this reason, it is better to start
the cycle with the singleton:
The following fails because there is is no element 3:
>>> Permutation(1, 2)(3)
Traceback (most recent call last):
...
IndexError: list index out of range
Traceback (most recent call last):
...
IndexError: list index out of range
This is ok: only the call to an out of range singleton is prohibited; otherwise the permutation autosizes:
>>> Permutation(3)(1, 2)
Permutation([0, 2, 1, 3])
>>> Permutation(1, 2)(3, 4) == Permutation(3, 4)(1, 2)
True
Equality Testing
The array forms must be the same in order for permutations to be equal:
>>> Permutation([1, 0, 2, 3]) == Permutation([1, 0])
False
Identity Permutation
The identity permutation is a permutation in which no element is out of place. It can be
entered in a variety of ways. All the following create an identity permutation of size 4:
>>> I = Permutation([0, 1, 2, 3])
>>> all(p == I for p in [
... Permutation(3),
... Permutation(range(4)),
... Permutation([], size=4),
... Permutation(size=4)])
True
194
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Watch out for entering the range inside a set of brackets (which is cycle notation):
>>> I == Permutation([range(4)])
False
Permutation Printing
There are a few things to note about how Permutations are printed.
1) If you prefer one form (array or cycle) over another, you can set that with the
print cyclic ag.
>>> Permutation(1, 2)(4, 5)(3, 4)
Permutation([0, 2, 1, 4, 5, 3])
>>> p = _
>>> Permutation.print_cyclic = True
>>> p
Permutation(1, 2)(3, 4, 5)
>>> Permutation.print_cyclic = False
2) Regardless of the setting, a list of elements in the array for cyclic form can be obtained
and either of those can be copied and supplied as the argument to Permutation:
>>> p.array_form
[0, 2, 1, 4, 5, 3]
>>> p.cyclic_form
[[1, 2], [3, 4, 5]]
>>> Permutation(_) == p
True
3) Printing is economical in that as little as possible is printed while retaining all information about the size of the permutation:
>>> Permutation([1, 0, 2, 3])
Permutation([1, 0, 2, 3])
>>> Permutation([1, 0, 2, 3], size=20)
Permutation([1, 0], size=20)
>>> Permutation([1, 0, 2, 4, 3, 5, 6], size=20)
Permutation([1, 0, 2, 4, 3], size=20)
>>> p = Permutation([1, 0, 2, 3])
>>> Permutation.print_cyclic = True
>>> p
Permutation(3)(0, 1)
>>> Permutation.print_cyclic = False
The 2 was not printed but it is still there as can be seen with the array form and size
methods:
>>> p.array_form
[1, 0, 2, 3]
>>> p.size
4
5.2. Combinatorics Module
195
SymPy Documentation, Release 0.7.6
Short Introduction To Other Methods
The permutation can act as a bijective function, telling what element is located at a given
position
>>>
>>>
2
>>>
2
>>>
{0:
q = Permutation([5, 2, 3, 4, 1, 0])
q.array_form[1] # the hard way
q(1) # the easy way
dict([(i, q(i)) for i in range(q.size)]) # showing the bijection
5, 1: 2, 2: 3, 3: 4, 4: 1, 5: 0}
The full cyclic form (including singletons) can be obtained:
>>> p.full_cyclic_form
[[0, 1], [2], [3]]
Any permutation can be factored into transpositions of pairs of elements:
>>> Permutation([[1, 2], [3, 4, 5]]).transpositions()
[(1, 2), (3, 5), (3, 4)]
>>> Permutation.rmul(*[Permutation([ti], size=6) for ti in _]).cyclic_form
[[1, 2], [3, 4, 5]]
The number of permutations on a set of n elements is given by n! and is called the
cardinality.
>>> p.size
4
>>> p.cardinality
24
A given permutation has a rank among all the possible permutations of the same elements, but what that rank is depends on how the permutations are enumerated. (There
are a number of dierent methods of doing so.) The lexicographic rank is given by the
rank method and this rank is used to increment a permutation with addition/subtraction:
>>> p.rank()
6
>>> p + 1
Permutation([1, 0, 3, 2])
>>> p.next_lex()
Permutation([1, 0, 3, 2])
>>> _.rank()
7
>>> p.unrank_lex(p.size, rank=7)
Permutation([1, 0, 3, 2])
The product of two permutations p and q is dened as their composition as functions,
(p*q)(i) = q(p(i)) [R12] (page 1902).
>>>
>>>
>>>
[2,
>>>
[3,
>>>
[3,
196
p = Permutation([1, 0, 2, 3])
q = Permutation([2, 3, 1, 0])
list(q*p)
3, 0, 1]
list(p*q)
2, 1, 0]
[q(p(i)) for i in range(p.size)]
2, 1, 0]
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
The permutation can be applied to any list-like object, not only Permutations:
>>> p([zero, one, four, two])
[one, zero, four, two]
>>> p(zo42)
[o, z, 4, 2]
If you have a list of arbitrary elements, the corresponding permutation can be found with
the from sequence method:
>>> Permutation.from_sequence(SymPy)
Permutation([1, 3, 2, 0, 4])
array form
Return a copy of the attribute array form Examples ========
>>>
>>>
>>>
>>>
[2,
>>>
[3,
>>>
[2,
>>>
[0,
from sympy.combinatorics.permutations import Permutation
Permutation.print_cyclic = False
p = Permutation([[2,0], [3,1]])
p.array_form
3, 0, 1]
Permutation([[2,0,3,1]]).array_form
2, 0, 1]
Permutation([2,0,3,1]).array_form
0, 3, 1]
Permutation([[1, 2], [4, 5]]).array_form
2, 1, 3, 5, 4]
ascents()
Returns the positions of ascents in a permutation, ie, the location where p[i] < p[i+1]
See Also:
descents (page 199), inversions (page 203), min (page 207), max (page 207)
Examples
>>>
>>>
>>>
[1,
from sympy.combinatorics.permutations import Permutation
p = Permutation([4,0,1,3,2])
p.ascents()
2]
atoms()
Returns all the elements of a permutation
Examples
>>> from sympy.combinatorics import Permutation
>>> Permutation([0, 1, 2, 3, 4, 5]).atoms()
set([0, 1, 2, 3, 4, 5])
>>> Permutation([[0, 1], [2, 3], [4, 5]]).atoms()
set([0, 1, 2, 3, 4, 5])
cardinality
Returns the number of all possible permutations.
5.2. Combinatorics Module
197
SymPy Documentation, Release 0.7.6
See Also:
length (page 206), order (page 208), rank (page 209), size (page 211)
Examples
>>> from sympy.combinatorics.permutations import Permutation
>>> p = Permutation([0,1,2,3])
>>> p.cardinality
24
commutator(x)
Return the commutator of self and x: x*self*x*self
If f and g are part of a group, G, then the commutator of f and g is the group identity
i f and g commute, i.e. fg == gf.
References
http://en.wikipedia.org/wiki/Commutator
Examples
>>> from sympy.combinatorics.permutations import Permutation
>>> Permutation.print_cyclic = False
>>> p = Permutation([0, 2, 3, 1])
>>> x = Permutation([2, 0, 3, 1])
>>> c = p.commutator(x); c
Permutation([2, 1, 3, 0])
>>> c == ~x*~p*x*p
True
>>> I = Permutation(3)
>>> p = [I + i for i in range(6)]
>>> for i in range(len(p)):
...
for j in range(len(p)):
...
c = p[i].commutator(p[j])
...
if p[i]*p[j] == p[j]*p[i]:
...
assert c == I
...
else:
...
assert c != I
...
commutes with(other)
Checks if the elements are commuting.
Examples
>>> from sympy.combinatorics.permutations import Permutation
>>> a = Permutation([1,4,3,0,2,5])
>>> b = Permutation([0,1,2,3,4,5])
>>> a.commutes_with(b)
True
198
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> b = Permutation([2,3,5,4,1,0])
>>> a.commutes_with(b)
False
cycle structure
Return the cycle structure of the permutation as a dictionary indicating the multiplicity of each cycle length.
Examples
>>>
>>>
>>>
{1:
>>>
{2:
from sympy.combinatorics import Permutation
Permutation.print_cyclic = True
Permutation(3).cycle_structure
4}
Permutation(0, 4, 3)(1, 2)(5, 6).cycle_structure
2, 3: 1}
cycles
Returns the number of cycles contained in the permutation (including singletons).
See Also:
sympy.functions.combinatorial.numbers.stirling (page 374)
Examples
>>> from sympy.combinatorics import Permutation
>>> Permutation([0, 1, 2]).cycles
3
>>> Permutation([0, 1, 2]).full_cyclic_form
[[0], [1], [2]]
>>> Permutation(0, 1)(2, 3).cycles
2
cyclic form
This is used to convert to the cyclic notation from the canonical notation. Singletons
are omitted.
See Also:
array form (page 197), full cyclic form (page 200)
Examples
>>> from sympy.combinatorics.permutations import Permutation
>>> Permutation.print_cyclic = False
>>> p = Permutation([0, 3, 1, 2])
>>> p.cyclic_form
[[1, 3, 2]]
>>> Permutation([1, 0, 2, 4, 3, 5]).cyclic_form
[[0, 1], [3, 4]]
descents()
Returns the positions of descents in a permutation, ie, the location where p[i] >
p[i+1]
5.2. Combinatorics Module
199
SymPy Documentation, Release 0.7.6
See Also:
ascents (page 197), inversions (page 203), min (page 207), max (page 207)
Examples
>>>
>>>
>>>
[0,
from sympy.combinatorics.permutations import Permutation
p = Permutation([4,0,1,3,2])
p.descents()
3]
classmethod from inversion vector(inversion)
Calculates the permutation from the inversion vector.
Examples
>>> from sympy.combinatorics.permutations import Permutation
>>> Permutation.print_cyclic = False
>>> Permutation.from_inversion_vector([3, 2, 1, 0, 0])
Permutation([3, 2, 1, 0, 4, 5])
classmethod from sequence(i, key=None)
Return the permutation needed to obtain i from the sorted elements of i. If custom
sorting is desired, a key can be given.
Examples
>>> from sympy.combinatorics import Permutation
>>> Permutation.print_cyclic = True
>>> Permutation.from_sequence(SymPy)
Permutation(4)(0, 1, 3)
>>> _(sorted(SymPy))
[S, y, m, P, y]
>>> Permutation.from_sequence(SymPy, key=lambda x: x.lower())
Permutation(4)(0, 2)(1, 3)
full cyclic form
Return permutation in cyclic form including singletons.
Examples
>>> from sympy.combinatorics.permutations import Permutation
>>> Permutation([0, 2, 1]).full_cyclic_form
[[0], [1, 2]]
get adjacency distance(other)
Computes the adjacency distance between two permutations.
This metric counts the number of times a pair i,j of jobs is adjacent in both p and p.
If n adj is this quantity then the adjacency distance is n - n adj - 1 [1]
200
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
[1] Reeves, Colin R. Landscapes, Operators and Heuristic search, Annals of Operational Research, 86, pp 473-490. (1999)
See Also:
get precedence matrix (page 202),
get adjacency matrix (page 201)
get precedence distance
(page
202),
Examples
>>>
>>>
>>>
>>>
3
>>>
>>>
4
from sympy.combinatorics.permutations import Permutation
p = Permutation([0, 3, 1, 2, 4])
q = Permutation.josephus(4, 5, 2)
p.get_adjacency_distance(q)
r = Permutation([0, 2, 1, 4, 3])
p.get_adjacency_distance(r)
get adjacency matrix()
Computes the adjacency matrix of a permutation.
If job i is adjacent to job j in a permutation p then we set m[i, j] = 1 where m is the
adjacency matrix of p.
See Also:
get precedence matrix (page 202),
get adjacency distance (page 200)
get precedence distance
(page
202),
Examples
>>> from sympy.combinatorics.permutations import Permutation
>>> p = Permutation.josephus(3,6,1)
>>> p.get_adjacency_matrix()
Matrix([
[0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 1, 0],
[0, 0, 0, 0, 0, 1],
[0, 1, 0, 0, 0, 0],
[1, 0, 0, 0, 0, 0],
[0, 0, 0, 1, 0, 0]])
>>> q = Permutation([0, 1, 2, 3])
>>> q.get_adjacency_matrix()
Matrix([
[0, 1, 0, 0],
[0, 0, 1, 0],
[0, 0, 0, 1],
[0, 0, 0, 0]])
get positional distance(other)
Computes the positional distance between two permutations.
See Also:
get precedence distance (page 202), get adjacency distance (page 200)
5.2. Combinatorics Module
201
SymPy Documentation, Release 0.7.6
Examples
>>>
>>>
>>>
>>>
>>>
12
>>>
12
from sympy.combinatorics.permutations import Permutation
p = Permutation([0, 3, 1, 2, 4])
q = Permutation.josephus(4, 5, 2)
r = Permutation([3, 1, 4, 0, 2])
p.get_positional_distance(q)
p.get_positional_distance(r)
get precedence distance(other)
Computes the precedence distance between two permutations.
Suppose p and p represent n jobs. The precedence metric counts the number of
times a job j is preceded by job i in both p and p. This metric is commutative.
See Also:
get precedence matrix (page 202),
get adjacency distance (page 200)
get adjacency matrix
(page
201),
Examples
>>>
>>>
>>>
>>>
7
>>>
7
from sympy.combinatorics.permutations import Permutation
p = Permutation([2, 0, 4, 3, 1])
q = Permutation([3, 1, 2, 4, 0])
p.get_precedence_distance(q)
q.get_precedence_distance(p)
get precedence matrix()
Gets the precedence matrix. This is used for computing the distance between two
permutations.
See Also:
get precedence distance (page 202),
get adjacency distance (page 200)
get adjacency matrix
(page
201),
Examples
>>> from sympy.combinatorics.permutations import Permutation
>>> p = Permutation.josephus(3,6,1)
>>> p
Permutation([2, 5, 3, 1, 4, 0])
>>> p.get_precedence_matrix()
Matrix([
[0, 0, 0, 0, 0, 0],
[1, 0, 0, 0, 1, 0],
[1, 1, 0, 1, 1, 1],
[1, 1, 0, 0, 1, 0],
[1, 0, 0, 0, 0, 0],
[1, 1, 0, 1, 1, 0]])
202
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
index()
Returns the index of a permutation.
The index of a permutation is the sum of all subscripts j such that p[j] is greater than
p[j+1].
Examples
>>> from sympy.combinatorics.permutations import Permutation
>>> p = Permutation([3, 0, 2, 1, 4])
>>> p.index()
2
inversion vector()
Return the inversion vector of the permutation.
The inversion vector consists of elements whose value indicates the number of elements in the permutation that are lesser than it and lie on its right hand side.
The inversion vector is the same as the Lehmer encoding of a permutation.
See Also:
from inversion vector (page 200)
Examples
>>>
>>>
>>>
[4,
>>>
>>>
[3,
from sympy.combinatorics.permutations import Permutation
p = Permutation([4, 8, 0, 7, 1, 5, 3, 6, 2])
p.inversion_vector()
7, 0, 5, 0, 2, 1, 1]
p = Permutation([3, 2, 1, 0])
p.inversion_vector()
2, 1]
The inversion vector increases lexicographically with the rank of the permutation,
the -ith element cycling through 0..i.
>>> p = Permutation(2)
>>> while p:
...
print(%s %s %s %
...
p = p.next_lex()
...
Permutation([0, 1, 2]) [0,
Permutation([0, 2, 1]) [0,
Permutation([1, 0, 2]) [1,
Permutation([1, 2, 0]) [1,
Permutation([2, 0, 1]) [2,
Permutation([2, 1, 0]) [2,
(p, p.inversion_vector(), p.rank()))
0]
1]
0]
1]
0]
1]
0
1
2
3
4
5
inversions()
Computes the number of inversions of a permutation.
An inversion is where i > j but p[i] < p[j].
For small length of p, it iterates over all i and j values and calculates the number of
inversions. For large length of p, it uses a variation of merge sort to calculate the
number of inversions.
5.2. Combinatorics Module
203
SymPy Documentation, Release 0.7.6
See Also:
descents (page 199), ascents (page 197), min (page 207), max (page 207)
References
[1] http://www.cp.eng.chula.ac.th/piak/teaching/algo/algo2008/count-inv.htm
Examples
>>>
>>>
>>>
0
>>>
6
from sympy.combinatorics.permutations import Permutation
p = Permutation([0,1,2,3,4,5])
p.inversions()
Permutation([3,2,1,0]).inversions()
is Empty
Checks to see if the permutation is a set with zero elements
See Also:
is Singleton (page 204)
Examples
>>> from sympy.combinatorics import Permutation
>>> Permutation([]).is_Empty
True
>>> Permutation([0]).is_Empty
False
is Identity
Returns True if the Permutation is an identity permutation.
See Also:
order (page 208)
Examples
>>> from sympy.combinatorics.permutations import Permutation
>>> p = Permutation([])
>>> p.is_Identity
True
>>> p = Permutation([[0], [1], [2]])
>>> p.is_Identity
True
>>> p = Permutation([0, 1, 2])
>>> p.is_Identity
True
>>> p = Permutation([0, 2, 1])
>>> p.is_Identity
False
204
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
is Singleton
Checks to see if the permutation contains only one number and is thus the only
possible permutation of this set of numbers
See Also:
is Empty (page 204)
Examples
>>> from sympy.combinatorics import Permutation
>>> Permutation([0]).is_Singleton
True
>>> Permutation([0, 1]).is_Singleton
False
is even
Checks if a permutation is even.
See Also:
is odd (page 205)
Examples
>>> from sympy.combinatorics.permutations import Permutation
>>> p = Permutation([0,1,2,3])
>>> p.is_even
True
>>> p = Permutation([3,2,1,0])
>>> p.is_even
True
is odd
Checks if a permutation is odd.
See Also:
is even (page 205)
Examples
>>> from sympy.combinatorics.permutations import Permutation
>>> p = Permutation([0,1,2,3])
>>> p.is_odd
False
>>> p = Permutation([3,2,0,1])
>>> p.is_odd
True
classmethod josephus(m, n, s=1)
Return as a permutation the shuing of range(n) using the Josephus scheme in
which every m-th item is selected until all have been chosen. The returned permutation has elements listed by the order in which they were selected.
The parameter s stops the selection process when there are s items remaining and
these are selected by continuing the selection, counting by 1 rather than by m.
5.2. Combinatorics Module
205
SymPy Documentation, Release 0.7.6
Consider selecting every 3rd item from 6 until only 2 remain:
choices
========
012345
01 345
01 34
01 4
0
4
0
chosen
======
2
25
253
2531
25314
253140
References
1.http://en.wikipedia.org/wiki/Flavius Josephus
2.http://en.wikipedia.org/wiki/Josephus problem
3.http://www.wou.edu/burtonl/josephus.html
Examples
>>> from sympy.combinatorics import Permutation
>>> Permutation.josephus(3, 6, 2).array_form
[2, 5, 3, 1, 4, 0]
length()
Returns the number of integers moved by a permutation.
See Also:
min (page 207), max (page 207), support (page 212), cardinality (page 197), order
(page 208), rank (page 209), size (page 211)
Examples
>>> from sympy.combinatorics import Permutation
>>> Permutation([0, 3, 2, 1]).length()
2
>>> Permutation([[0, 1], [2, 3]]).length()
4
list(size=None)
Return the permutation as an explicit list, possibly trimming unmoved elements if
size is less than the maximum element in the permutation; if this is desired, setting
size=-1 will guarantee such trimming.
Examples
>>>
>>>
>>>
>>>
[0,
206
from sympy.combinatorics.permutations import Permutation
Permutation.print_cyclic = False
p = Permutation(2, 3)(4, 5)
p.list()
1, 3, 2, 5, 4]
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> p.list(10)
[0, 1, 3, 2, 5, 4, 6, 7, 8, 9]
Passing a length too small will trim trailing, unchanged elements in the permutation:
>>> Permutation(2, 4)(1, 2, 4).list(-1)
[0, 2, 1]
>>> Permutation(3).list(-1)
[]
max()
The maximum element moved by the permutation.
See Also:
min (page 207), descents (page 199), ascents (page 197), inversions (page 203)
Examples
>>> from sympy.combinatorics.permutations import Permutation
>>> p = Permutation([1,0,2,3,4])
>>> p.max()
1
min()
The minimum element moved by the permutation.
See Also:
max (page 207), descents (page 199), ascents (page 197), inversions (page 203)
Examples
>>> from sympy.combinatorics.permutations import Permutation
>>> p = Permutation([0,1,4,3,2])
>>> p.min()
2
mul inv(other)
other*self, self and other have array form
next lex()
Returns the next permutation in lexicographical order. If self is the last permutation
in lexicographical order it returns None. See [4] section 2.4.
See Also:
rank (page 209), unrank lex (page 212)
Examples
>>>
>>>
>>>
17
>>>
18
from sympy.combinatorics.permutations import Permutation
p = Permutation([2, 3, 1, 0])
p = Permutation([2, 3, 1, 0]); p.rank()
p = p.next_lex(); p.rank()
5.2. Combinatorics Module
207
SymPy Documentation, Release 0.7.6
next nonlex()
Returns the next permutation in nonlex order [3]. If self is the last permutation in
this order it returns None.
See Also:
rank nonlex (page 209), unrank nonlex (page 213)
Examples
>>> from sympy.combinatorics.permutations import Permutation
>>> Permutation.print_cyclic = False
>>> p = Permutation([2, 0, 3, 1]); p.rank_nonlex()
5
>>> p = p.next_nonlex(); p
Permutation([3, 0, 1, 2])
>>> p.rank_nonlex()
6
next trotterjohnson()
Returns the next permutation in Trotter-Johnson order. If self is the last permutation
it returns None. See [4] section 2.4. If it is desired to generate all such permutations,
they can be generated in order more quickly with the generate bell function.
See Also:
rank trotterjohnson (page 210),
unrank trotterjohnson
sympy.utilities.iterables.generate bell (page 1557)
(page
213),
Examples
>>> from sympy.combinatorics.permutations import Permutation
>>> Permutation.print_cyclic = False
>>> p = Permutation([3, 0, 2, 1])
>>> p.rank_trotterjohnson()
4
>>> p = p.next_trotterjohnson(); p
Permutation([0, 3, 2, 1])
>>> p.rank_trotterjohnson()
5
order()
Computes the order of a permutation.
When the permutation is raised to the power of its order it equals the identity permutation.
See Also:
identity, cardinality (page 197), length (page 206), rank (page 209), size
(page 211)
Examples
208
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> from sympy.combinatorics.permutations import Permutation
>>> Permutation.print_cyclic = False
>>> p = Permutation([3, 1, 5, 2, 4, 0])
>>> p.order()
4
>>> (p**(p.order()))
Permutation([], size=6)
parity()
Computes the parity of a permutation.
The parity of a permutation reects the parity of the number of inversions in the
permutation, i.e., the number of pairs of x and y such that x > y but p[x] < p[y].
See Also:
af parity
Examples
>>>
>>>
>>>
0
>>>
>>>
1
from sympy.combinatorics.permutations import Permutation
p = Permutation([0,1,2,3])
p.parity()
p = Permutation([3,2,0,1])
p.parity()
classmethod random(n)
Generates a random permutation of length n.
Uses the underlying Python pseudo-random number generator.
Examples
>>> from sympy.combinatorics.permutations import Permutation
>>> Permutation.random(2) in (Permutation([1, 0]), Permutation([0, 1]))
True
rank()
Returns the lexicographic rank of the permutation.
See Also:
next lex (page 207), unrank lex (page 212), cardinality (page 197), length
(page 206), order (page 208), size (page 211)
Examples
>>>
>>>
>>>
0
>>>
>>>
23
from sympy.combinatorics.permutations import Permutation
p = Permutation([0, 1, 2, 3])
p.rank()
p = Permutation([3, 2, 1, 0])
p.rank()
5.2. Combinatorics Module
209
SymPy Documentation, Release 0.7.6
rank nonlex(inv perm=None)
This is a linear time ranking algorithm that does not enforce lexicographic order [3].
See Also:
next nonlex (page 207), unrank nonlex (page 213)
Examples
>>> from sympy.combinatorics.permutations import Permutation
>>> p = Permutation([0,1,2,3])
>>> p.rank_nonlex()
23
rank trotterjohnson()
Returns the Trotter Johnson rank, which we get from the minimal change algorithm.
See [4] section 2.4.
See Also:
unrank trotterjohnson (page 213), next trotterjohnson (page 208)
Examples
>>>
>>>
>>>
0
>>>
>>>
7
from sympy.combinatorics.permutations import Permutation
p = Permutation([0,1,2,3])
p.rank_trotterjohnson()
p = Permutation([0,2,1,3])
p.rank_trotterjohnson()
static rmul(*args)
Return product of Permutations [a, b, c, ...] as the Permutation whose ith value is
a(b(c(i))).
a, b, c, ... can be Permutation objects or tuples.
Notes
All items in the sequence will be parsed by Permutation as necessary as long as the
rst item is a Permutation:
>>> Permutation.rmul(a, [0, 2, 1]) == Permutation.rmul(a, b)
True
The reverse order of arguments will raise a TypeError.
Examples
>>> from sympy.combinatorics.permutations import _af_rmul, Permutation
>>> Permutation.print_cyclic = False
210
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>>
>>>
>>>
[1,
>>>
[1,
a, b = [1, 0, 2], [0, 2, 1]
a = Permutation(a); b = Permutation(b)
list(Permutation.rmul(a, b))
2, 0]
[a(b(i)) for i in range(3)]
2, 0]
This handles the operands in reverse order compared to the * operator:
>>>
>>>
[2,
>>>
[2,
a = Permutation(a); b = Permutation(b)
list(a*b)
0, 1]
[b(a(i)) for i in range(3)]
0, 1]
static rmul with af(*args)
same as rmul, but the elements of args are Permutation objects which have
array form
runs()
Returns the runs of a permutation.
An ascending sequence in a permutation is called a run [5].
Examples
>>> from sympy.combinatorics.permutations import Permutation
>>> p = Permutation([2,5,7,3,6,0,1,4,8])
>>> p.runs()
[[2, 5, 7], [3, 6], [0, 1, 4, 8]]
>>> q = Permutation([1,3,2,0])
>>> q.runs()
[[1, 3], [2], [0]]
signature()
Gives the signature of the permutation needed to place the elements of the permutation in canonical order.
The signature is calculated as (-1)<number of inversions>
See Also:
inversions (page 203)
Examples
>>>
>>>
>>>
0
>>>
1
>>>
>>>
1
>>>
-1
from sympy.combinatorics.permutations import Permutation
p = Permutation([0,1,2])
p.inversions()
p.signature()
q = Permutation([0,2,1])
q.inversions()
q.signature()
5.2. Combinatorics Module
211
SymPy Documentation, Release 0.7.6
size
Returns the number of elements in the permutation.
See Also:
cardinality (page 197), length (page 206), order (page 208), rank (page 209)
Examples
>>> from sympy.combinatorics import Permutation
>>> Permutation([[3, 2], [0, 1]]).size
4
support()
Return the elements in permutation, P, for which P[i] != i.
Examples
>>>
>>>
>>>
[1,
>>>
[0,
from sympy.combinatorics import Permutation
p = Permutation([[3, 2], [0, 1], [4]])
p.array_form
0, 3, 2, 4]
p.support()
1, 2, 3]
transpositions()
Return the permutation decomposed into a list of transpositions.
It is always possible to express a permutation as the product of transpositions, see
[1]
References
1.http://en.wikipedia.org/wiki/Transposition %28mathematics%29#Properties
Examples
>>> from sympy.combinatorics.permutations import Permutation
>>> p = Permutation([[1, 2, 3], [0, 4, 5, 6, 7]])
>>> t = p.transpositions()
>>> t
[(0, 7), (0, 6), (0, 5), (0, 4), (1, 3), (1, 2)]
>>> print(.join(str(c) for c in t))
(0, 7)(0, 6)(0, 5)(0, 4)(1, 3)(1, 2)
>>> Permutation.rmul(*[Permutation([ti], size=p.size) for ti in t]) == p
True
classmethod unrank lex(size, rank)
Lexicographic permutation unranking.
See Also:
rank (page 209), next lex (page 207)
212
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.combinatorics.permutations import Permutation
>>> Permutation.print_cyclic = False
>>> a = Permutation.unrank_lex(5, 10)
>>> a.rank()
10
>>> a
Permutation([0, 2, 4, 1, 3])
classmethod unrank nonlex(n, r)
This is a linear time unranking algorithm that does not respect lexicographic order
[3].
See Also:
next nonlex (page 207), rank nonlex (page 209)
Examples
>>> from sympy.combinatorics.permutations import Permutation
>>> Permutation.print_cyclic = False
>>> Permutation.unrank_nonlex(4, 5)
Permutation([2, 0, 3, 1])
>>> Permutation.unrank_nonlex(4, -1)
Permutation([0, 1, 2, 3])
classmethod unrank trotterjohnson(size, rank)
Trotter Johnson permutation unranking. See [4] section 2.4.
See Also:
rank trotterjohnson (page 210), next trotterjohnson (page 208)
Examples
>>> from sympy.combinatorics.permutations import Permutation
>>> Permutation.unrank_trotterjohnson(5, 10)
Permutation([0, 3, 1, 2, 4])
class sympy.combinatorics.permutations.Cycle(*args)
Wrapper around dict which provides the functionality of a disjoint cycle.
A cycle shows the rule to use to move subsets of elements to obtain a permutation. The
Cycle class is more exible that Permutation in that 1) all elements need not be present in
order to investigate how multiple cycles act in sequence and 2) it can contain singletons:
>>> from sympy.combinatorics.permutations import Perm, Cycle
A Cycle will automatically parse a cycle given as a tuple on the rhs:
>>> Cycle(1, 2)(2, 3)
Cycle(1, 3, 2)
The identity cycle, Cycle(), can be used to start a product:
5.2. Combinatorics Module
213
SymPy Documentation, Release 0.7.6
>>> Cycle()(1, 2)(2,3)
Cycle(1, 3, 2)
The array form of a Cycle can be obtained by calling the list method (or passing it to the
list function) and all elements from 0 will be shown:
>>>
>>>
[0,
>>>
[0,
a = Cycle(1, 2)
a.list()
2, 1]
list(a)
2, 1]
If a larger (or smaller) range is desired use the list method and provide the desired size
but the Cycle cannot be truncated to a size smaller than the largest element that is out
of place:
>>>
>>>
[0,
>>>
[0,
>>>
[0,
b = Cycle(2,4)(1,2)(3,1,4)(1,3)
b.list()
2, 1, 3, 4]
b.list(b.size + 1)
2, 1, 3, 4, 5]
b.list(-1)
2, 1]
Singletons are not shown when printing with one exception: the largest element is always shown as a singleton if necessary:
>>> Cycle(1, 4, 10)(4, 5)
Cycle(1, 5, 4, 10)
>>> Cycle(1, 2)(4)(5)(10)
Cycle(1, 2)(10)
The array form can be used to instantiate a Permutation so other properties of the permutation can be investigated:
>>> Perm(Cycle(1,2)(3,4).list()).transpositions()
[(1, 2), (3, 4)]
See Also:
Permutation (page 192)
Notes
The underlying structure of the Cycle is a dictionary and although the iter method
has been redened to give the array form of the cycle, the underlying dictionary items
are still available with the such methods as items():
>>> list(Cycle(1, 2).items())
[(1, 2), (2, 1)]
list(size=None)
Return the cycles as an explicit list starting from 0 up to the greater of the largest
value in the cycles and size.
Truncation of trailing unmoved items will occur when size is less than the maximum
element in the cycle; if this is desired, setting size=-1 will guarantee such trimming.
214
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>>
>>>
>>>
>>>
>>>
[0,
>>>
[0,
from sympy.combinatorics.permutations import Cycle
from sympy.combinatorics.permutations import Permutation
Permutation.print_cyclic = False
p = Cycle(2, 3)(4, 5)
p.list()
1, 3, 2, 5, 4]
p.list(10)
1, 3, 2, 5, 4, 6, 7, 8, 9]
Passing a length too small will trim trailing, unchanged elements in the permutation:
>>> Cycle(2, 4)(1, 2, 4).list(-1)
[0, 2, 1]
Generators
static generators.symmetric(n)
Generates the symmetric group of order n, Sn.
Examples
>>> from sympy.combinatorics.permutations import Permutation
>>> Permutation.print_cyclic = True
>>> from sympy.combinatorics.generators import symmetric
>>> list(symmetric(3))
[Permutation(2), Permutation(1, 2), Permutation(2)(0, 1),
Permutation(0, 1, 2), Permutation(0, 2, 1), Permutation(0, 2)]
static generators.cyclic(n)
Generates the cyclic group of order n, Cn.
See Also:
dihedral (page 216)
Examples
>>> from sympy.combinatorics.permutations import Permutation
>>> Permutation.print_cyclic = True
>>> from sympy.combinatorics.generators import cyclic
>>> list(cyclic(5))
[Permutation(4), Permutation(0, 1, 2, 3, 4), Permutation(0, 2, 4, 1, 3),
Permutation(0, 3, 1, 4, 2), Permutation(0, 4, 3, 2, 1)]
static generators.alternating(n)
Generates the alternating group of order n, An.
5.2. Combinatorics Module
215
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.combinatorics.permutations import Permutation
>>> Permutation.print_cyclic = True
>>> from sympy.combinatorics.generators import alternating
>>> list(alternating(3))
[Permutation(2), Permutation(0, 1, 2), Permutation(0, 2, 1)]
static generators.dihedral(n)
Generates the dihedral group of order 2n, Dn.
The result is given as a subgroup of Sn, except for the special cases n=1 (the group S2)
and n=2 (the Klein 4-group) where thats not possible and embeddings in S2 and S4
respectively are given.
See Also:
cyclic (page 215)
Examples
>>> from sympy.combinatorics.permutations import Permutation
>>> Permutation.print_cyclic = True
>>> from sympy.combinatorics.generators import dihedral
>>> list(dihedral(3))
[Permutation(2), Permutation(0, 2), Permutation(0, 1, 2),
Permutation(1, 2), Permutation(0, 2, 1), Permutation(2)(0, 1)]
Permutation Groups
class sympy.combinatorics.perm groups.PermutationGroup
The class dening a Permutation group.
PermutationGroup([p1, p2, ..., pn]) returns the permutation group generated by the list
of permutations. This group can be supplied to Polyhedron if one desires to decorate the
elements to which the indices of the permutation refer.
See Also:
sympy.combinatorics.polyhedron.Polyhedron
(page
sympy.combinatorics.permutations.Permutation (page 192)
243),
References
[1] Holt, D., Eick, B., OBrien, E. Handbook of Computational Group Theory
[2] Seress, A. Permutation Group Algorithms
[3] http://en.wikipedia.org/wiki/Schreier vector
[4]
http://en.wikipedia.org/wiki/Nielsen transformation
uct replacement algorithm
#Prod-
[5] Frank Celler, Charles R.Leedham-Green, Scott H.Murray, Alice C.Niemeyer, and
E.A.OBrien. Generating Random Elements of a Finite Group
[6] http://en.wikipedia.org/wiki/Block %28permutation group theory%29
216
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
[7] http://www.algorithmist.com/index.php/Union Find
[8] http://en.wikipedia.org/wiki/Multiply transitive group#Multiply transitive groups
[9] http://en.wikipedia.org/wiki/Center %28group theory%29
[10] http://en.wikipedia.org/wiki/Centralizer and normalizer
[11] http://groupprops.subwiki.org/wiki/Derived subgroup
[12] http://en.wikipedia.org/wiki/Nilpotent group
[13] http://www.math.colostate.edu/hulpke/CGT/cgtnotes.pdf
Examples
>>>
>>>
>>>
>>>
>>>
from sympy.combinatorics import Permutation
Permutation.print_cyclic = True
from sympy.combinatorics.permutations import Cycle
from sympy.combinatorics.polyhedron import Polyhedron
from sympy.combinatorics.perm_groups import PermutationGroup
The permutations corresponding to motion of the front, right and bottom face of a 2x2
Rubiks cube are dened:
>>> F = Permutation(2, 19, 21, 8)(3, 17, 20, 10)(4, 6, 7, 5)
>>> R = Permutation(1, 5, 21, 14)(3, 7, 23, 12)(8, 10, 11, 9)
>>> D = Permutation(6, 18, 14, 10)(7, 19, 15, 11)(20, 22, 23, 21)
These are passed as permutations to PermutationGroup:
>>> G = PermutationGroup(F, R, D)
>>> G.order()
3674160
The group can be supplied to a Polyhedron in order to track the objects being moved.
An example involving the 2x2 Rubiks cube is given there, but here is a simple demonstration:
>>>
>>>
>>>
>>>
>>>
(A,
>>>
>>>
(A,
>>>
>>>
(A,
a = Permutation(2, 1)
b = Permutation(1, 0)
G = PermutationGroup(a, b)
P = Polyhedron(list(ABC), pgroup=G)
P.corners
B, C)
P.rotate(0) # apply permutation 0
P.corners
C, B)
P.reset()
P.corners
B, C)
Or one can make a permutation as a product of selected permutations and apply them
to an iterable directly:
>>> P10 = G.make_perm([0, 1])
>>> P10(ABC)
[C, A, B]
base
Return a base from the Schreier-Sims algorithm.
5.2. Combinatorics Module
217
SymPy Documentation, Release 0.7.6
For a permutation group G, a base is a sequence of points B = (b 1, b 2, ...,
b k) such that no element of G apart from the identity xes all the points in B. The
concepts of a base and strong generating set and their applications are discussed in
depth in [1], pp. 87-89 and [2], pp. 55-57.
An alternative way to think of B is that it gives the indices of the stabilizer cosets
that contain more than the identity permutation.
See Also:
strong gens (page 241), basic transversals (page 220), basic orbits (page 219),
basic stabilizers (page 219)
Examples
>>>
>>>
>>>
[0,
from sympy.combinatorics import Permutation, PermutationGroup
G = PermutationGroup([Permutation(0, 1, 3)(2, 4)])
G.base
2]
baseswap(base, strong gens, pos, randomized=False, transversals=None, basic orbits=None, strong gens distr=None)
Swap two consecutive base points in base and strong generating set.
If a base for a group G is given by (b 1, b 2, ..., b k), this function returns a
base (b 1, b 2, ..., b {i+1}, b i, ..., b k), where i is given by pos, and a
strong generating set relative to that base. The original base and strong generating
set are not modied.
The randomized version (default) is of Las Vegas type.
Parameters base, strong gens :
The base and strong generating set.
pos :
The position at which swapping is performed.
randomized :
A switch between randomized and deterministic version.
transversals :
The transversals for the basic orbits, if known.
basic orbits :
The basic orbits, if known.
strong gens distr :
The strong generators distributed by basic stabilizers, if known.
Returns (base, strong gens) :
base is the new base, and strong gens is a generating set relative to
it.
See Also:
schreier sims (page 238)
218
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Notes
The deterministic version of the algorithm is discussed in [1], pp. 102-103; the
randomized version is discussed in [1], p.103, and [2], p.98. It is of Las Vegas type.
Notice that [1] contains a mistake in the pseudocode and discussion of BASESWAP:
on line 3 of the pseudocode, |\beta {i+1}{\left\langle T\right\rangle}| should
be replaced by |\beta {i}{\left\langle T\right\rangle}|, and the same for the
discussion of the algorithm.
Examples
>>> from sympy.combinatorics.named_groups import SymmetricGroup
>>> from sympy.combinatorics.testutil import _verify_bsgs
>>> from sympy.combinatorics.perm_groups import PermutationGroup
>>> S = SymmetricGroup(4)
>>> S.schreier_sims()
>>> S.base
[0, 1, 2]
>>> base, gens = S.baseswap(S.base, S.strong_gens, 1, randomized=False)
>>> base, gens
([0, 2, 1],
[Permutation(0, 1, 2, 3), Permutation(3)(0, 1), Permutation(1, 3, 2),
Permutation(2, 3), Permutation(1, 3)])
check that base, gens is a BSGS
>>> S1 = PermutationGroup(gens)
>>> _verify_bsgs(S1, base, gens)
True
basic orbits
Return the basic orbits relative to a base and strong generating set.
If (b 1, b 2, ..., b k) is a base for a group G, and G{(i)} = G {b 1, b 2,
..., b {i-1}} is the i-th basic stabilizer (so that G{(1)} = G), the i-th basic orbit
relative to this base is the orbit of b i under G{(i)}. See [1], pp. 87-89 for more
information.
See Also:
base (page 217), strong gens (page 241), basic transversals (page 220), basic stabilizers (page 219)
Examples
>>> from sympy.combinatorics.named_groups import SymmetricGroup
>>> S = SymmetricGroup(4)
>>> S.basic_orbits
[[0, 1, 2, 3], [1, 2, 3], [2, 3]]
basic stabilizers
Return a chain of stabilizers relative to a base and strong generating set.
The i-th basic stabilizer G{(i)} relative to a base (b 1, b 2, ..., b k) is G {b 1,
b 2, ..., b {i-1}}. For more information, see [1], pp. 87-89.
See Also:
5.2. Combinatorics Module
219
SymPy Documentation, Release 0.7.6
base (page 217), strong gens (page 241),
sic transversals (page 220)
basic orbits (page 219),
ba-
Examples
>>> from sympy.combinatorics.named_groups import AlternatingGroup
>>> A = AlternatingGroup(4)
>>> A.schreier_sims()
>>> A.base
[0, 1]
>>> for g in A.basic_stabilizers:
...
print(g)
...
PermutationGroup([
Permutation(3)(0, 1, 2),
Permutation(1, 2, 3)])
PermutationGroup([
Permutation(1, 2, 3)])
basic transversals
Return basic transversals relative to a base and strong generating set.
The basic transversals are transversals of the basic orbits. They are provided as a
list of dictionaries, each dictionary having keys - the elements of one of the basic
orbits, and values - the corresponding transversal elements. See [1], pp. 87-89 for
more information.
See Also:
strong gens (page 241), base (page 217),
sic stabilizers (page 219)
basic orbits (page 219),
ba-
Examples
>>> from sympy.combinatorics.named_groups import AlternatingGroup
>>> A = AlternatingGroup(4)
>>> A.basic_transversals
[{0: Permutation(3),
1: Permutation(3)(0, 1, 2),
2: Permutation(3)(0, 2, 1),
3: Permutation(0, 3, 1)},
{1: Permutation(3),
2: Permutation(1, 2, 3),
3: Permutation(1, 3, 2)}]
center()
Return the center of a permutation group.
The center for a group G is dened as Z(G) = \{z\in G | \forall g\in G, zg =
gz \}, the set of elements of G that commute with all elements of G. It is equal to the
centralizer of G inside G, and is naturally a subgroup of G ([9]).
See Also:
centralizer (page 221)
220
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Notes
This is a naive implementation that is a straightforward application of .centralizer()
Examples
>>>
>>>
>>>
>>>
>>>
2
from sympy.combinatorics.perm_groups import PermutationGroup
from sympy.combinatorics.named_groups import DihedralGroup
D = DihedralGroup(4)
G = D.center()
G.order()
centralizer(other)
Return the centralizer of a group/set/element.
The centralizer of a set of permutations S inside a group G is the set of elements of
G that commute with all elements of S:
C G(S) = \{ g \in G | gs = sg \forall s \in S\} ([10])
Usually, S is a subset of G, but if G is a proper subgroup of the full symmetric group,
we allow for S to have elements outside G.
It is naturally a subgroup of G; the centralizer of a permutation group is equal to the
centralizer of any set of generators for that group, since any element commuting
with the generators commutes with any product of the generators.
Parameters other :
a permutation group/list of permutations/single permutation
See Also:
subgroup search (page 241)
Notes
The implementation is an application of .subgroup search() with tests using a specic base for the group G.
Examples
>>> from sympy.combinatorics.named_groups import (SymmetricGroup,
... CyclicGroup)
>>> S = SymmetricGroup(6)
>>> C = CyclicGroup(6)
>>> H = S.centralizer(C)
>>> H.is_subgroup(C)
True
commutator(G, H)
Return the commutator of two subgroups.
5.2. Combinatorics Module
221
SymPy Documentation, Release 0.7.6
For a permutation group K and subgroups G, H, the commutator of G and H is dened
as the group generated by all the commutators [g, h] = hgh{-1}g{-1} for g in G
and h in H. It is naturally a subgroup of K ([1], p.27).
See Also:
derived subgroup (page 225)
Notes
The commutator of two subgroups H, G is equal to the normal closure of the commutators of all the generators, i.e. hgh{-1}g{-1} for h a generator of H and g a
generator of G ([1], p.28)
Examples
>>> from sympy.combinatorics.named_groups import (SymmetricGroup,
... AlternatingGroup)
>>> S = SymmetricGroup(5)
>>> A = AlternatingGroup(5)
>>> G = S.commutator(S, A)
>>> G.is_subgroup(A)
True
contains(g, strict=True)
Test if permutation g belong to self, G.
If g is an element of G it can be written as a product of factors drawn from the cosets
of Gs stabilizers. To see if g is one of the actual generators dening the group use
G.has(g).
If strict is not True, g will be resized, if necessary, to match the size of permutations
in self.
See Also:
coset factor (page 223), has, in
Examples
>>> from sympy.combinatorics import Permutation
>>> Permutation.print_cyclic = True
>>> from sympy.combinatorics.perm_groups import PermutationGroup
>>> a = Permutation(1, 2)
>>> b = Permutation(2, 3, 1)
>>> G = PermutationGroup(a, b, degree=5)
>>> G.contains(G[0]) # trivial check
True
>>> elem = Permutation([[2, 3]], size=5)
>>> G.contains(elem)
True
>>> G.contains(Permutation(4)(0, 1, 2, 3))
False
If strict is False, a permutation will be resized, if necessary:
222
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> H = PermutationGroup(Permutation(5))
>>> H.contains(Permutation(3))
False
>>> H.contains(Permutation(3), strict=False)
True
To test if a given permutation is present in the group:
>>> elem in G.generators
False
>>> G.has(elem)
False
coset factor(g, factor index=False)
Return Gs (selfs) coset factorization of g
If g is an element of G then it can be written as the product of permutations drawn
from the Schreier-Sims coset decomposition,
The permutations returned in f are those for which the product gives g: g =
f[n]*...f[1]*f[0] where n = len(B) and B = G.base. f[i] is one of the permutations in self. basic orbits[i].
If factor index==True, returns a tuple [b[0],..,b[n]], where b[i] belongs to
self. basic orbits[i]
Examples
>>>
>>>
>>>
>>>
>>>
from sympy.combinatorics
Permutation.print_cyclic
a = Permutation(0, 1, 3,
b = Permutation(0, 1, 3,
G = PermutationGroup([a,
import Permutation, PermutationGroup
= True
7, 6, 4)(2, 5)
2)(4, 5, 7, 6)
b])
Dene g:
>>> g = Permutation(7)(1, 2, 4)(3, 6, 5)
Conrm that it is an element of G:
>>> G.contains(g)
True
Thus, it can be written as a product of factors (up to 3) drawn from u. See below
that a factor from u1 and u2 and the Identity permutation have been used:
>>> f = G.coset_factor(g)
>>> f[2]*f[1]*f[0] == g
True
>>> f1 = G.coset_factor(g, True); f1
[0, 4, 4]
>>> tr = G.basic_transversals
>>> f[0] == tr[0][f1[0]]
True
If g is not an element of G then [] is returned:
5.2. Combinatorics Module
223
SymPy Documentation, Release 0.7.6
>>> c = Permutation(5, 6, 7)
>>> G.coset_factor(c)
[]
see util. strip
coset rank(g)
rank using Schreier-Sims representation
The coset rank of g is the ordering number in which it appears in the lexicographic
listing according to the coset decomposition
The ordering is the same as in G.generate(method=coset). If g does not belong to
the group it returns None.
See Also:
coset factor (page 223)
Examples
>>> from sympy.combinatorics import Permutation
>>> Permutation.print_cyclic = True
>>> from sympy.combinatorics.perm_groups import PermutationGroup
>>> a = Permutation(0, 1, 3, 7, 6, 4)(2, 5)
>>> b = Permutation(0, 1, 3, 2)(4, 5, 7, 6)
>>> G = PermutationGroup([a, b])
>>> c = Permutation(7)(2, 4)(3, 5)
>>> G.coset_rank(c)
16
>>> G.coset_unrank(16)
Permutation(7)(2, 4)(3, 5)
coset unrank(rank, af=False)
unrank using Schreier-Sims representation
coset unrank is the inverse operation of coset rank if 0 <= rank < order; otherwise
it returns None.
degree
Returns the size of the permutations in the group.
The number of permutations comprising the group is given by len(group); the number of permutations that can be generated by the group is given by group.order().
See Also:
order (page 237)
Examples
>>>
>>>
>>>
>>>
>>>
>>>
3
>>>
224
from sympy.combinatorics import Permutation
Permutation.print_cyclic = True
from sympy.combinatorics.perm_groups import PermutationGroup
a = Permutation([1, 0, 2])
G = PermutationGroup([a])
G.degree
len(G)
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
1
>>> G.order()
2
>>> list(G.generate())
[Permutation(2), Permutation(2)(0, 1)]
derived series()
Return the derived series for the group.
The derived series for a group G is dened as G = G 0 > G 1 > G 2 > \ldots
where G i = [G {i-1}, G {i-1}], i.e. G i is the derived subgroup of G {i-1}, for
i\in\mathbb{N}. When we have G k = G {k-1} for some k\in\mathbb{N}, the series
terminates.
Returns A list of permutation groups containing the members of the
derived :
series in the order G = G 0, G 1, G 2, ldots. :
See Also:
derived subgroup (page 225)
Examples
>>> from sympy.combinatorics.named_groups import (SymmetricGroup,
... AlternatingGroup, DihedralGroup)
>>> A = AlternatingGroup(5)
>>> len(A.derived_series())
1
>>> S = SymmetricGroup(4)
>>> len(S.derived_series())
4
>>> S.derived_series()[1].is_subgroup(AlternatingGroup(4))
True
>>> S.derived_series()[2].is_subgroup(DihedralGroup(2))
True
derived subgroup()
Compute the derived subgroup.
The derived subgroup, or commutator subgroup is the subgroup generated by all
commutators [g, h] = hgh{-1}g{-1} for g, h\in G ; it is equal to the normal
closure of the set of commutators of the generators ([1], p.28, [11]).
See Also:
derived series (page 225)
Examples
>>>
>>>
>>>
>>>
>>>
>>>
>>>
from sympy.combinatorics import Permutation
Permutation.print_cyclic = True
from sympy.combinatorics.perm_groups import PermutationGroup
a = Permutation([1, 0, 2, 4, 3])
b = Permutation([0, 1, 3, 2, 4])
G = PermutationGroup([a, b])
C = G.derived_subgroup()
5.2. Combinatorics Module
225
SymPy Documentation, Release 0.7.6
>>> list(C.generate(af=True))
[[0, 1, 2, 3, 4], [0, 1, 3, 4, 2], [0, 1, 4, 2, 3]]
generate(method=coset, af=False)
Return iterator to generate the elements of the group
Iteration is done with one of these methods:
method=coset using the Schreier-Sims coset representation
method=dimino using the Dimino method
If af = True it yields the array form of the permutations
Examples
>>>
>>>
>>>
>>>
from sympy.combinatorics import Permutation
Permutation.print_cyclic = True
from sympy.combinatorics import PermutationGroup
from sympy.combinatorics.polyhedron import tetrahedron
The permutation group given in the tetrahedron object is not true groups:
>>> G = tetrahedron.pgroup
>>> G.is_group()
False
But the group generated by the permutations in the tetrahedron pgroup even the
rst two is a proper group:
>>> H = PermutationGroup(G[0], G[1])
>>> J = PermutationGroup(list(H.generate())); J
PermutationGroup([
Permutation(0, 1)(2, 3),
Permutation(3),
Permutation(1, 2, 3),
Permutation(1, 3, 2),
Permutation(0, 3, 1),
Permutation(0, 2, 3),
Permutation(0, 3)(1, 2),
Permutation(0, 1, 3),
Permutation(3)(0, 2, 1),
Permutation(0, 3, 2),
Permutation(3)(0, 1, 2),
Permutation(0, 2)(1, 3)])
>>> _.is_group()
True
generate dimino(af=False)
Yield group elements using Diminos algorithm
If af == True it yields the array form of the permutations
References
[1] The Implementation of Various Algorithms for Permutation Groups in the Computer Algebra System: AXIOM, N.J. Doye, M.Sc. Thesis
226
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.combinatorics import Permutation
>>> Permutation.print_cyclic = True
>>> from sympy.combinatorics.perm_groups import PermutationGroup
>>> a = Permutation([0, 2, 1, 3])
>>> b = Permutation([0, 2, 3, 1])
>>> g = PermutationGroup([a, b])
>>> list(g.generate_dimino(af=True))
[[0, 1, 2, 3], [0, 2, 1, 3], [0, 2, 3, 1],
[0, 1, 3, 2], [0, 3, 2, 1], [0, 3, 1, 2]]
generate schreier sims(af=False)
Yield group elements using the Schreier-Sims representation in coset rank order
If af = True it yields the array form of the permutations
Examples
>>> from sympy.combinatorics import Permutation
>>> Permutation.print_cyclic = True
>>> from sympy.combinatorics.perm_groups import PermutationGroup
>>> a = Permutation([0, 2, 1, 3])
>>> b = Permutation([0, 2, 3, 1])
>>> g = PermutationGroup([a, b])
>>> list(g.generate_schreier_sims(af=True))
[[0, 1, 2, 3], [0, 2, 1, 3], [0, 3, 2, 1],
[0, 1, 3, 2], [0, 2, 3, 1], [0, 3, 1, 2]]
generators
Returns the generators of the group.
Examples
>>> from sympy.combinatorics import Permutation
>>> Permutation.print_cyclic = True
>>> from sympy.combinatorics.perm_groups import PermutationGroup
>>> a = Permutation([0, 2, 1])
>>> b = Permutation([1, 0, 2])
>>> G = PermutationGroup([a, b])
>>> G.generators
[Permutation(1, 2), Permutation(2)(0, 1)]
is abelian
Test if the group is Abelian.
Examples
>>>
>>>
>>>
>>>
>>>
from sympy.combinatorics import Permutation
Permutation.print_cyclic = True
from sympy.combinatorics.perm_groups import PermutationGroup
a = Permutation([0, 2, 1])
b = Permutation([1, 0, 2])
5.2. Combinatorics Module
227
SymPy Documentation, Release 0.7.6
>>> G = PermutationGroup([a, b])
>>> G.is_abelian
False
>>> a = Permutation([0, 2, 1])
>>> G = PermutationGroup([a])
>>> G.is_abelian
True
is alt sym(eps=0.05, random prec=None)
Monte Carlo test for the symmetric/alternating group for degrees >= 8.
More specically, it is one-sided Monte Carlo with the answer True (i.e., G is symmetric/alternating) guaranteed to be correct, and the answer False being incorrect
with probability eps.
See Also:
check cycles alt sym
Notes
The algorithm itself uses some nontrivial results from group theory and number
theory: 1) If a transitive group G of degree n contains an element with a cycle of
length n/2 < p < n-2 for p a prime, G is the symmetric or alternating group ([1],
pp. 81-82) 2) The proportion of elements in the symmetric/alternating group having
the property described in 1) is approximately \log(2)/\log(n) ([1], p.82; [2], pp.
226-227). The helper function check cycles alt sym is used to go over the cycles
in a permutation and look for ones satisfying 1).
Examples
>>> from sympy.combinatorics.perm_groups import PermutationGroup
>>> from sympy.combinatorics.named_groups import DihedralGroup
>>> D = DihedralGroup(10)
>>> D.is_alt_sym()
False
is group()
Return True if the group meets three criteria: identity is present, the inverse of
every element is also an element, and the product of any two elements is also an
element. If any of the tests fail, False is returned.
Examples
>>>
>>>
>>>
>>>
from sympy.combinatorics import Permutation
Permutation.print_cyclic = True
from sympy.combinatorics import PermutationGroup
from sympy.combinatorics.polyhedron import tetrahedron
The permutation group given in the tetrahedron object is not a true group:
>>> G = tetrahedron.pgroup
>>> G.is_group()
False
228
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
But the group generated by the permutations in the tetrahedron pgroup is a proper
group:
>>> H = PermutationGroup(list(G.generate()))
>>> H.is_group()
True
The identity permutation is present:
>>> H.has(Permutation(G.degree - 1))
True
The product of any two elements from the group is also in the group:
>>>
>>>
>>>
>>>
>>>
...
...
>>>
from sympy import TableForm
g = list(H)
n = len(g)
m = []
for i in g:
m.append([g.index(i*H) for H in g])
TableForm(m, headings=[range(n), range(n)], wipe_zeros=False)
| 0 1 2 3 4 5 6 7 8 9 10 11
---------------------------------------0 | 11 0 8 10 6 2 7 4 5 3 9 1
1 | 0 1 2 3 4 5 6 7 8 9 10 11
2 | 6 2 7 4 5 3 9 1 11 0 8 10
3 | 5 3 9 1 11 0 8 10 6 2 7 4
4 | 3 4 0 2 10 6 11 8 9 7 1 5
5 | 4 5 6 7 8 9 10 11 0 1 2 3
6 | 10 6 11 8 9 7 1 5 3 4 0 2
7 | 9 7 1 5 3 4 0 2 10 6 11 8
8 | 7 8 4 6 2 10 3 0 1 11 5 9
9 | 8 9 10 11 0 1 2 3 4 5 6 7
10 | 2 10 3 0 1 11 5 9 7 8 4 6
11 | 1 11 5 9 7 8 4 6 2 10 3 0
>>>
The entries in the table give the element in the group corresponding
to the product of a given column element and row element:
>>> g[3]*g[2] == g[9]
True
The inverse of every element is also in the group:
>>> TableForm([[g.index(~gi) for gi in g]], headings=[[], range(n)],
...
wipe_zeros=False)
0 1 2 3 4 5 6 7 8 9 10 11
--------------------------11 1 7 3 10 9 6 2 8 5 4 0
So we see that g[1] and g[3] are equivalent to their inverse while g[7] == g[2].
is nilpotent
Test if the group is nilpotent.
A group G is nilpotent if it has a central series of nite length. Alternatively, G is
nilpotent if its lower central series terminates with the trivial group. Every nilpotent
group is also solvable ([1], p.29, [12]).
See Also:
5.2. Combinatorics Module
229
SymPy Documentation, Release 0.7.6
lower central series (page 232), is solvable (page 231)
Examples
>>> from sympy.combinatorics.named_groups import (SymmetricGroup,
... CyclicGroup)
>>> C = CyclicGroup(6)
>>> C.is_nilpotent
True
>>> S = SymmetricGroup(5)
>>> S.is_nilpotent
False
is normal(gr)
Test if G=self is a normal subgroup of gr.
G is normal in gr if for each g2 in G, g1 in gr, g = g1*g2*g1**-1 belongs to G It is
sucient to check this for each g1 in gr.generator and g2 g2 in G.generator
Examples
>>> from sympy.combinatorics import Permutation
>>> Permutation.print_cyclic = True
>>> from sympy.combinatorics.perm_groups import PermutationGroup
>>> a = Permutation([1, 2, 0])
>>> b = Permutation([1, 0, 2])
>>> G = PermutationGroup([a, b])
>>> G1 = PermutationGroup([a, Permutation([2, 0, 1])])
>>> G1.is_normal(G)
True
is primitive(randomized=True)
Test if a group is primitive.
A permutation group G acting on a set S is called primitive if S contains no nontrivial
block under the action of G (a block is nontrivial if its cardinality is more than 1).
See Also:
minimal block (page 234), random stab (page 238)
Notes
The algorithm is described in [1], p.83, and uses the function minimal block to search
for blocks of the form \{0, k\} for k ranging over representatives for the orbits of
G 0, the stabilizer of 0. This algorithm has complexity O(n2) where n is the degree
of the group, and will perform badly if G 0 is small.
There are two implementations oered: one nds G 0 deterministically using the
function stabilizer, and the other (default) produces random elements of G 0 using random stab, hoping that they generate a subgroup of G 0 with not too many
more orbits than G 0 (this is suggested in [1], p.83). Behavior is changed by the
randomized ag.
230
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.combinatorics.perm_groups import PermutationGroup
>>> from sympy.combinatorics.named_groups import DihedralGroup
>>> D = DihedralGroup(10)
>>> D.is_primitive()
False
is solvable
Test if the group is solvable.
G is solvable if its derived series terminates with the trivial group ([1], p.29).
See Also:
is nilpotent (page 229), derived series (page 225)
Examples
>>> from sympy.combinatorics.named_groups import SymmetricGroup
>>> S = SymmetricGroup(3)
>>> S.is_solvable
True
is subgroup(G, strict=True)
Return True if all elements of self belong to G.
If strict is False then if selfs degree is smaller than Gs, the elements will be
resized to have the same degree.
Examples
>>> from sympy.combinatorics import Permutation, PermutationGroup
>>> from sympy.combinatorics.named_groups import (SymmetricGroup,
...
CyclicGroup)
Testing is strict by default: the degree of each group must be the same:
>>> p = Permutation(0, 1, 2, 3, 4, 5)
>>> G1 = PermutationGroup([Permutation(0, 1, 2), Permutation(0, 1)])
>>> G2 = PermutationGroup([Permutation(0, 2), Permutation(0, 1, 2)])
>>> G3 = PermutationGroup([p, p**2])
>>> assert G1.order() == G2.order() == G3.order() == 6
>>> G1.is_subgroup(G2)
True
>>> G1.is_subgroup(G3)
False
>>> G3.is_subgroup(PermutationGroup(G3[1]))
False
>>> G3.is_subgroup(PermutationGroup(G3[0]))
True
To ignore the size, set strict to False:
>>> S3 = SymmetricGroup(3)
>>> S5 = SymmetricGroup(5)
>>> S3.is_subgroup(S5, strict=False)
5.2. Combinatorics Module
231
SymPy Documentation, Release 0.7.6
True
>>> C7 = CyclicGroup(7)
>>> G = S5*C7
>>> S5.is_subgroup(G, False)
True
>>> C7.is_subgroup(G, 0)
False
is transitive(strict=True)
Test if the group is transitive.
A group is transitive if it has a single orbit.
If strict is False the group is transitive if it has a single orbit of length dierent
from 1.
Examples
>>> from sympy.combinatorics.permutations import Permutation
>>> from sympy.combinatorics.perm_groups import PermutationGroup
>>> a = Permutation([0, 2, 1, 3])
>>> b = Permutation([2, 0, 1, 3])
>>> G1 = PermutationGroup([a, b])
>>> G1.is_transitive()
False
>>> G1.is_transitive(strict=False)
True
>>> c = Permutation([2, 3, 0, 1])
>>> G2 = PermutationGroup([a, c])
>>> G2.is_transitive()
True
>>> d = Permutation([1,0,2,3])
>>> e = Permutation([0,1,3,2])
>>> G3 = PermutationGroup([d, e])
>>> G3.is_transitive() or G3.is_transitive(strict=False)
False
is trivial
Test if the group is the trivial group.
This is true if the group contains only the identity permutation.
Examples
>>> from sympy.combinatorics import Permutation
>>> from sympy.combinatorics.perm_groups import PermutationGroup
>>> G = PermutationGroup([Permutation([0, 1, 2])])
>>> G.is_trivial
True
lower central series()
Return the lower central series for the group.
The lower central series for a group G is the series G = G 0 > G 1 > G 2 > \ldots
where G k = [G, G {k-1}], i.e. every term after the rst is equal to the commutator
of G and the previous term in G1 ([1], p.29).
232
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Returns A list of permutation groups in the order :
G = G 0, G 1, G 2, ldots :
See Also:
commutator (page 221), derived series (page 225)
Examples
>>> from sympy.combinatorics.named_groups import (AlternatingGroup,
... DihedralGroup)
>>> A = AlternatingGroup(4)
>>> len(A.lower_central_series())
2
>>> A.lower_central_series()[1].is_subgroup(DihedralGroup(2))
True
make perm(n, seed=None)
Multiply n randomly selected permutations from pgroup together, starting with the
identity permutation. If n is a list of integers, those integers will be used to select
the permutations and they will be applied in L to R order: make perm((A, B, C)) will
give CBA(I) where I is the identity permutation.
seed is used to set the seed for the random selection of permutations from pgroup. If
this is a list of integers, the corresponding permutations from pgroup will be selected
in the order give. This is mainly used for testing purposes.
See Also:
random (page 238)
Examples
>>> from sympy.combinatorics import Permutation
>>> Permutation.print_cyclic = True
>>> from sympy.combinatorics.perm_groups import PermutationGroup
>>> a, b = [Permutation([1, 0, 3, 2]), Permutation([1, 3, 0, 2])]
>>> G = PermutationGroup([a, b])
>>> G.make_perm(1, [0])
Permutation(0, 1)(2, 3)
>>> G.make_perm(3, [0, 1, 0])
Permutation(0, 2, 3, 1)
>>> G.make_perm([0, 1, 0])
Permutation(0, 2, 3, 1)
max div
Maximum proper divisor of the degree of a permutation group.
See Also:
minimal block (page 234), union find merge
Notes
Obviously, this is the degree divided by its minimal proper divisor (larger than 1, if
one exists). As it is guaranteed to be prime, the sieve from sympy.ntheory is used.
5.2. Combinatorics Module
233
SymPy Documentation, Release 0.7.6
This function is also used as an optimization tool for the functions minimal block
and union find merge.
Examples
>>>
>>>
>>>
>>>
2
from sympy.combinatorics import Permutation
from sympy.combinatorics.perm_groups import PermutationGroup
G = PermutationGroup([Permutation([0,2,1,3])])
G.max_div
minimal block(points)
For a transitive group, nds the block system generated by points.
If a group G acts on a set S, a nonempty subset B of S is called a block under the
action of G if for all g in G we have gB = B (g xes B) or gB and B have no common
points (g moves B entirely). ([1], p.23; [6]).
The distinct translates gB of a block B for g in G partition the set S and this set of
translates is known as a block system. Moreover, we obviously have that all blocks
in the partition have the same size, hence the block size divides |S| ([1], p.23). A Gcongruence is an equivalence relation on the set S such that a b implies g(a)
g(b) for all g in G. For a transitive group, the equivalence classes of a G-congruence
and the blocks of a block system are the same thing ([1], p.23).
The algorithm below checks the group for transitivity, and then nds the Gcongruence generated by the pairs (p 0, p 1), (p 0, p 2), ..., (p 0,p {k-1})
which is the same as nding the maximal block system (i.e., the one with minimum
block size) such that p 0, ..., p {k-1} are in the same block ([1], p.83).
It is an implementation of Atkinsons algorithm, as suggested in [1], and manipulates
an equivalence relation on the set S using a union-nd data structure. The running
time is just above O(|points||S|). ([1], pp. 83-87; [7]).
See Also:
union find rep,
(page 230)
union find merge, is transitive (page 232), is primitive
Examples
>>>
>>>
>>>
>>>
[0,
>>>
[0,
from sympy.combinatorics.perm_groups import PermutationGroup
from sympy.combinatorics.named_groups import DihedralGroup
D = DihedralGroup(10)
D.minimal_block([0,5])
6, 2, 8, 4, 0, 6, 2, 8, 4]
D.minimal_block([0,1])
0, 0, 0, 0, 0, 0, 0, 0, 0]
normal closure(other, k=10)
Return the normal closure of a subgroup/set of permutations.
If S is a subset of a group G, the normal closure of A in G is dened as the intersection of all normal subgroups of G that contain A ([1], p.14). Alternatively, it is the
group generated by the conjugates x{-1}yx for x a generator of G and y a generator of the subgroup \left\langle S\right\rangle generated by S (for some chosen
generating set for \left\langle S\right\rangle) ([1], p.73).
234
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Parameters other :
a subgroup/list of permutations/single permutation
k:
an implementation-specic parameter that determines the number of
conjugates that are adjoined to other at once
See Also:
commutator (page 221), derived subgroup (page 225), random pr (page 238)
Notes
The algorithm is described in [1], pp. 73-74; it makes use of the generation of random elements for permutation groups by the product replacement algorithm.
Examples
>>> from sympy.combinatorics.named_groups import (SymmetricGroup,
... CyclicGroup, AlternatingGroup)
>>> S = SymmetricGroup(5)
>>> C = CyclicGroup(5)
>>> G = S.normal_closure(C)
>>> G.order()
60
>>> G.is_subgroup(AlternatingGroup(5))
True
orbit(alpha, action=tuples)
Compute the orbit of alpha \{g(\alpha) | g \in G\} as a set.
The time complexity of the algorithm used here is O(|Orb|*r) where |Orb| is the size
of the orbit and r is the number of generators of the group. For a more detailed
analysis, see [1], p.78, [2], pp. 19-21. Here alpha can be a single point, or a list of
points.
If alpha is a single point, the ordinary orbit is computed. if alpha is a list of points,
there are three available options:
union - computes the union of the orbits of the points in the list tuples - computes
the orbit of the list interpreted as an ordered tuple under the group action ( i.e.,
g((1,2,3)) = (g(1), g(2), g(3)) ) sets - computes the orbit of the list interpreted as a
sets
See Also:
orbit transversal (page 236)
Examples
>>>
>>>
>>>
>>>
>>>
from sympy.combinatorics import Permutation
from sympy.combinatorics.perm_groups import PermutationGroup
a = Permutation([1,2,0,4,5,6,3])
G = PermutationGroup([a])
G.orbit(0)
5.2. Combinatorics Module
235
SymPy Documentation, Release 0.7.6
set([0, 1, 2])
>>> G.orbit([0,4], union)
set([0, 1, 2, 3, 4, 5, 6])
orbit rep(alpha, beta, schreier vector=None)
Return a group element which sends alpha to beta.
If beta is not in the orbit of alpha, the function returns False. This implementation
makes use of the schreier vector. For a proof of correctness, see [1], p.80
See Also:
schreier vector (page 240)
Examples
>>> from sympy.combinatorics import Permutation
>>> Permutation.print_cyclic = True
>>> from sympy.combinatorics.perm_groups import PermutationGroup
>>> from sympy.combinatorics.named_groups import AlternatingGroup
>>> G = AlternatingGroup(5)
>>> G.orbit_rep(0, 4)
Permutation(0, 4, 1, 2, 3)
orbit transversal(alpha, pairs=False)
Computes a transversal for the orbit of alpha as a set.
For a permutation group G, a transversal for the orbit Orb = \{g(\alpha) | g \in
G\} is a set \{g \beta | g \beta(\alpha) = \beta\} for \beta \in Orb. Note that
there may be more than one possible transversal. If pairs is set to True, it returns
the list of pairs (\beta, g \beta). For a proof of correctness, see [1], p.79
See Also:
orbit (page 235)
Examples
>>> from sympy.combinatorics import Permutation
>>> Permutation.print_cyclic = True
>>> from sympy.combinatorics.perm_groups import PermutationGroup
>>> from sympy.combinatorics.named_groups import DihedralGroup
>>> G = DihedralGroup(6)
>>> G.orbit_transversal(0)
[Permutation(5),
Permutation(0, 1, 2, 3, 4, 5),
Permutation(0, 5)(1, 4)(2, 3),
Permutation(0, 2, 4)(1, 3, 5),
Permutation(5)(0, 4)(1, 3),
Permutation(0, 3)(1, 4)(2, 5)]
orbits(rep=False)
Return the orbits of self, ordered according to lowest element in each orbit.
236
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.combinatorics.permutations import Permutation
>>> from sympy.combinatorics.perm_groups import PermutationGroup
>>> a = Permutation(1,5)(2,3)(4,0,6)
>>> b = Permutation(1,5)(3,4)(2,6,0)
>>> G = PermutationGroup([a, b])
>>> G.orbits()
[set([0, 2, 3, 4, 6]), set([1, 5])]
order()
Return the order of the group: the number of permutations that can be generated
from elements of the group.
The number of permutations comprising the group is given by len(group); the length
of each permutation in the group is given by group.size.
See Also:
degree (page 224)
Examples
>>> from sympy.combinatorics.permutations import Permutation
>>> from sympy.combinatorics.perm_groups import PermutationGroup
>>> a = Permutation([1, 0, 2])
>>> G = PermutationGroup([a])
>>> G.degree
3
>>> len(G)
1
>>> G.order()
2
>>> list(G.generate())
[Permutation(2), Permutation(2)(0, 1)]
>>>
>>>
>>>
>>>
6
a = Permutation([0, 2, 1])
b = Permutation([1, 0, 2])
G = PermutationGroup([a, b])
G.order()
pointwise stabilizer(points, incremental=True)
Return the pointwise stabilizer for a set of points.
For a permutation group G and a set of points \{p 1, p 2,\ldots, p k\}, the pointwise stabilizer of p 1, p 2, \ldots, p k is dened as G {p 1,\ldots, p k} = \{g\in
G | g(p i) = p i \forall i\in\{1, 2,\ldots,k\}\} ([1],p20). It is a subgroup
of G.
See Also:
stabilizer (page 241), schreier sims incremental (page 238)
5.2. Combinatorics Module
237
SymPy Documentation, Release 0.7.6
Notes
When incremental == True, rather than the obvious implementation using successive calls to .stabilizer(), this uses the incremental Schreier-Sims algorithm to obtain
a base with starting segment - the given points.
Examples
>>> from sympy.combinatorics.named_groups import SymmetricGroup
>>> S = SymmetricGroup(7)
>>> Stab = S.pointwise_stabilizer([2, 3, 5])
>>> Stab.is_subgroup(S.stabilizer(2).stabilizer(3).stabilizer(5))
True
random(af=False)
Return a random group element
random pr(gen count=11, iterations=50, random prec=None)
Return a random group element using product replacement.
For the details of the product replacement algorithm, see random pr init In random pr the actual product replacement is performed. Notice that if the attribute
random gens is empty, it needs to be initialized by random pr init.
See Also:
random pr init
random stab(alpha, schreier vector=None, random prec=None)
Random element from the stabilizer of alpha.
The schreier vector for alpha is an optional argument used for speeding up repeated
calls. The algorithm is described in [1], p.81
See Also:
random pr (page 238), orbit rep (page 236)
schreier sims()
Schreier-Sims algorithm.
It computes the generators of the chain of stabilizers G > G {b 1} > .. > G {b1,..,b r}
> 1 in which G {b 1,..,b i} stabilizes b 1,..,b i, and the corresponding s cosets. An
element of the group can be written as the product h 1*..*h s.
We use the incremental Schreier-Sims algorithm.
Examples
>>> from sympy.combinatorics.permutations import Permutation
>>> from sympy.combinatorics.perm_groups import PermutationGroup
>>> a = Permutation([0, 2, 1])
>>> b = Permutation([1, 0, 2])
>>> G = PermutationGroup([a, b])
>>> G.schreier_sims()
>>> G.basic_transversals
[{0: Permutation(2)(0, 1), 1: Permutation(2), 2: Permutation(1, 2)},
{0: Permutation(2), 2: Permutation(0, 2)}]
238
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
schreier sims incremental(base=None, gens=None)
Extend a sequence of points and generating set to a base and strong generating set.
Parameters base :
The sequence of points to be extended to a base. Optional parameter
with default value [].
gens :
The generating set to be extended to a strong generating set relative to the base obtained. Optional parameter with default value
self.generators.
Returns (base, strong gens) :
base is the base obtained, and strong gens is the strong generating set relative to it. The original parameters base, gens remain unchanged.
See Also:
schreier sims (page 238), schreier sims random (page 239)
Notes
This version of the Schreier-Sims algorithm runs in polynomial time. There are certain assumptions in the implementation - if the trivial group is provided, base and
gens are returned immediately, as any sequence of points is a base for the trivial
group. If the identity is present in the generators gens, it is removed as it is a redundant generator. The implementation is described in [1], pp. 90-93.
Examples
>>> from sympy.combinatorics.named_groups import AlternatingGroup
>>> from sympy.combinatorics.perm_groups import PermutationGroup
>>> from sympy.combinatorics.testutil import _verify_bsgs
>>> A = AlternatingGroup(7)
>>> base = [2, 3]
>>> seq = [2, 3]
>>> base, strong_gens = A.schreier_sims_incremental(base=seq)
>>> _verify_bsgs(A, base, strong_gens)
True
>>> base[:2]
[2, 3]
schreier sims random(base=None,
gens=None,
random prec=None)
Randomized Schreier-Sims algorithm.
consec succ=10,
The randomized Schreier-Sims algorithm takes the sequence base and the generating set gens, and extends base to a base, and gens to a strong generating set
relative to that base with probability of a wrong answer at most 2{-consec\ succ},
provided the random generators are suciently random.
Parameters base :
The sequence to be extended to a base.
gens :
5.2. Combinatorics Module
239
SymPy Documentation, Release 0.7.6
The generating set to be extended to a strong generating set.
consec succ :
The parameter dening the probability of a wrong answer.
random prec :
An internal parameter used for testing purposes.
Returns (base, strong gens) :
base is the base and strong gens is the strong generating set relative
to it.
See Also:
schreier sims (page 238)
Notes
The algorithm is described in detail in [1], pp. 97-98. It extends the orbits orbs and
the permutation groups stabs to basic orbits and basic stabilizers for the base and
strong generating set produced in the end. The idea of the extension process is to
sift random group elements through the stabilizer chain and amend the stabilizers/orbits along the way when a sift is not successful. The helper function strip
is used to attempt to decompose a random group element according to the current
state of the stabilizer chain and report whether the element was fully decomposed
(successful sift) or not (unsuccessful sift). In the latter case, the level at which the
sift failed is reported and used to amend stabs, base, gens and orbs accordingly.
The halting condition is for consec succ consecutive successful sifts to pass. This
makes sure that the current base and gens form a BSGS with probability at least 1
- 1/\text{consec\ succ}.
Examples
>>> from sympy.combinatorics.perm_groups import PermutationGroup
>>> from sympy.combinatorics.testutil import _verify_bsgs
>>> from sympy.combinatorics.named_groups import SymmetricGroup
>>> S = SymmetricGroup(5)
>>> base, strong_gens = S.schreier_sims_random(consec_succ=5)
>>> _verify_bsgs(S, base, strong_gens)
True
schreier vector(alpha)
Computes the schreier vector for alpha.
The Schreier vector eciently stores information about the orbit of alpha. It can
later be used to quickly obtain elements of the group that send alpha to a particular
element in the orbit. Notice that the Schreier vector depends on the order in which
the group generators are listed. For a denition, see [3]. Since list indices start from
zero, we adopt the convention to use None instead of 0 to signify that an element
doesnt belong to the orbit. For the algorithm and its correctness, see [2], pp.78-80.
See Also:
orbit (page 235)
240
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.combinatorics.perm_groups import PermutationGroup
>>> from sympy.combinatorics.permutations import Permutation
>>> a = Permutation([2,4,6,3,1,5,0])
>>> b = Permutation([0,1,3,5,4,6,2])
>>> G = PermutationGroup([a,b])
>>> G.schreier_vector(0)
[-1, None, 0, 1, None, 1, 0]
stabilizer(alpha)
Return the stabilizer subgroup of alpha.
The stabilizer of \alpha is the group G \alpha = \{g \in G | g(\alpha) = \alpha\}.
For a proof of correctness, see [1], p.79.
See Also:
orbit (page 235)
Examples
>>> from sympy.combinatorics import Permutation
>>> Permutation.print_cyclic = True
>>> from sympy.combinatorics.perm_groups import PermutationGroup
>>> from sympy.combinatorics.named_groups import DihedralGroup
>>> G = DihedralGroup(6)
>>> G.stabilizer(5)
PermutationGroup([
Permutation(5)(0, 4)(1, 3),
Permutation(5)])
strong gens
Return a strong generating set from the Schreier-Sims algorithm.
A generating set S = \{g 1, g 2, ..., g t\} for a permutation group G is a strong
generating set relative to the sequence of points (referred to as a base) (b 1, b 2,
..., b k) if, for 1 \leq i \leq k we have that the intersection of the pointwise
stabilizer G{(i+1)} := G {b 1, b 2, ..., b i} with S generates the pointwise
stabilizer G{(i+1)}. The concepts of a base and strong generating set and their
applications are discussed in depth in [1], pp. 87-89 and [2], pp. 55-57.
See Also:
base (page 217), basic transversals (page 220), basic orbits (page 219), basic stabilizers (page 219)
Examples
>>> from sympy.combinatorics.named_groups import DihedralGroup
>>> D = DihedralGroup(4)
>>> D.strong_gens
[Permutation(0, 1, 2, 3), Permutation(0, 3)(1, 2), Permutation(1, 3)]
>>> D.base
[0, 1]
5.2. Combinatorics Module
241
SymPy Documentation, Release 0.7.6
subgroup search(prop,
base=None,
strong gens=None,
init subgroup=None)
Find the subgroup of all elements satisfying the property prop.
tests=None,
This is done by a depth-rst search with respect to base images that uses several
tests to prune the search tree.
Parameters prop :
The property to be used. Has to be callable on group elements and
always return True or False. It is assumed that all group elements
satisfying prop indeed form a subgroup.
base :
A base for the supergroup.
strong gens :
A strong generating set for the supergroup.
tests :
A list of callables of length equal to the length of base. These are
used to rule out group elements by partial base images, so that
tests[l](g) returns False if the element g is known not to satisfy
prop base on where g sends the rst l + 1 base points.
init subgroup :
if a subgroup of the sought group is known in advance, it can be
passed to the function as this parameter.
Returns res :
The subgroup of all elements satisfying prop. The generating set for
this group is guaranteed to be a strong generating set relative to the
base base.
Notes
This function is extremely lenghty and complicated and will require some careful
attention. The implementation is described in [1], pp. 114-117, and the comments
for the code here follow the lines of the pseudocode in the book for clarity.
The complexity is exponential in general, since the search process by itself visits
all members of the supergroup. However, there are a lot of tests which are used to
prune the search tree, and users can dene their own tests via the tests parameter,
so in practice, and for some computations, its not terrible.
A crucial part in the procedure is the frequent base change performed (this is line
11 in the pseudocode) in order to obtain a new basic stabilizer. The book mentiones
that this can be done by using .baseswap(...), however the current imlementation
uses a more straightforward way to nd the next basic stabilizer - calling the function
.stabilizer(...) on the previous basic stabilizer.
Examples
242
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> from sympy.combinatorics.named_groups import (SymmetricGroup,
... AlternatingGroup)
>>> from sympy.combinatorics.perm_groups import PermutationGroup
>>> from sympy.combinatorics.testutil import _verify_bsgs
>>> S = SymmetricGroup(7)
>>> prop_even = lambda x: x.is_even
>>> base, strong_gens = S.schreier_sims_incremental()
>>> G = S.subgroup_search(prop_even, base=base, strong_gens=strong_gens)
>>> G.is_subgroup(AlternatingGroup(7))
True
>>> _verify_bsgs(G, base, G.generators)
True
transitivity degree
Compute the degree of transitivity of the group.
A permutation group G acting on \Omega = \{0, 1, ..., n-1\} is k-fold transitive,
if, for any k points (a 1, a 2, ..., a k)\in\Omega and any k points (b 1, b 2,
..., b k)\in\Omega there exists g\in G such that g(a 1)=b 1, g(a 2)=b 2, ...,
g(a k)=b k The degree of transitivity of G is the maximum k such that G is k-fold
transitive. ([8])
See Also:
is transitive (page 232), orbit (page 235)
Examples
>>>
>>>
>>>
>>>
>>>
>>>
3
from sympy.combinatorics.perm_groups import PermutationGroup
from sympy.combinatorics.permutations import Permutation
a = Permutation([1, 2, 0])
b = Permutation([1, 0, 2])
G = PermutationGroup([a,b])
G.transitivity_degree
Polyhedron
class sympy.combinatorics.polyhedron.Polyhedron
Represents the polyhedral symmetry group (PSG).
The PSG is one of the symmetry groups of the Platonic solids. There are three polyhedral
groups: the tetrahedral group of order 12, the octahedral group of order 24, and the
icosahedral group of order 60.
All doctests have been given in the docstring of the constructor of the object.
References
http://mathworld.wolfram.com/PolyhedralGroup.html
array form
Return the indices of the corners.
The indices are given relative to the original position of corners.
See Also:
5.2. Combinatorics Module
243
SymPy Documentation, Release 0.7.6
corners (page 244), cyclic form (page 244)
Examples
>>>
>>>
>>>
[0,
from sympy.combinatorics import Permutation, Cycle
from sympy.combinatorics.polyhedron import tetrahedron
tetrahedron.array_form
1, 2, 3]
>>>
>>>
[0,
>>>
[0,
tetrahedron.rotate(0)
tetrahedron.array_form
2, 3, 1]
tetrahedron.pgroup[0].array_form
2, 3, 1]
corners
Get the corners of the Polyhedron.
The method vertices is an alias for corners.
See Also:
array form (page 243), cyclic form (page 244)
Examples
>>> from sympy.combinatorics import Polyhedron
>>> from sympy.abc import a, b, c, d
>>> p = Polyhedron(list(abcd))
>>> p.corners == p.vertices == (a, b, c, d)
True
cyclic form
Return the indices of the corners in cyclic notation.
The indices are given relative to the original position of corners.
See Also:
corners (page 244), array form (page 243)
edges
Given the faces of the polyhedra we can get the edges.
Examples
>>> from sympy.combinatorics import Polyhedron
>>> from sympy.abc import a, b, c
>>> corners = (a, b, c)
>>> faces = [(0, 1, 2)]
>>> Polyhedron(corners, faces).edges
{(0, 1), (0, 2), (1, 2)}
faces
Get the faces of the Polyhedron.
244
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
pgroup
Get the permutations of the Polyhedron.
reset()
Return corners to their original positions.
Examples
>>>
>>>
(0,
>>>
>>>
(0,
>>>
>>>
(0,
from sympy.combinatorics.polyhedron import tetrahedron as T
T.corners
1, 2, 3)
T.rotate(0)
T.corners
2, 3, 1)
T.reset()
T.corners
1, 2, 3)
rotate(perm)
Apply a permutation to the polyhedron in place. The permutation may be given as
a Permutation instance or an integer indicating which permutation from pgroup of
the Polyhedron should be applied.
This is an operation that is analogous to rotation about an axis by a xed increment.
Notes
When a Permutation is applied, no check is done to see if that is a valid permutation for the Polyhedron. For example, a cube could be given a permutation which
eectively swaps only 2 vertices. A valid permutation (that rotates the object in a
physical way) will be obtained if one only uses permutations from the pgroup of the
Polyhedron. On the other hand, allowing arbitrary rotations (applications of permutations) gives a way to follow named elements rather than indices since Polyhedron
allows vertices to be named while Permutation works only with indices.
Examples
>>>
>>>
>>>
(0,
>>>
>>>
(1,
from sympy.combinatorics import Polyhedron, Permutation
from sympy.combinatorics.polyhedron import cube
cube.corners
1, 2, 3, 4, 5, 6, 7)
cube.rotate(0)
cube.corners
2, 3, 0, 5, 6, 7, 4)
A non-physical rotation that is not prohibited by this method:
>>>
>>>
>>>
(0,
cube.reset()
cube.rotate(Permutation([[1,2]], size=8))
cube.corners
2, 1, 3, 4, 5, 6, 7)
Polyhedron can be used to follow elements of set that are identied by letters instead
of integers:
5.2. Combinatorics Module
245
SymPy Documentation, Release 0.7.6
>>> shadow = h5 = Polyhedron(list(abcde))
>>> p = Permutation([3, 0, 1, 2, 4])
>>> h5.rotate(p)
>>> h5.corners
(d, a, b, c, e)
>>> _ == shadow.corners
True
>>> copy = h5.copy()
>>> h5.rotate(p)
>>> h5.corners == copy.corners
False
size
Get the number of corners of the Polyhedron.
vertices
Get the corners of the Polyhedron.
The method vertices is an alias for corners.
See Also:
array form (page 243), cyclic form (page 244)
Examples
>>> from sympy.combinatorics import Polyhedron
>>> from sympy.abc import a, b, c, d
>>> p = Polyhedron(list(abcd))
>>> p.corners == p.vertices == (a, b, c, d)
True
Prufer Sequences
class sympy.combinatorics.prufer.Prufer
The Prufer correspondence is an algorithm that describes the bijection between labeled
trees and the Prufer code. A Prufer code of a labeled tree is unique up to isomorphism
and has a length of n - 2.
Prufer sequences were rst used by Heinz Prufer to give a proof of Cayleys formula.
References
[R14] (page 1902)
static edges(*runs)
Return a list of edges and the number of nodes from the given runs that connect
nodes in an integer-labelled tree.
All node numbers will be shifted so that the minimum node is 0. It is not a problem
if edges are repeated in the runs; only unique edges are returned. There is no
assumption made about what the range of the node labels should be, but all nodes
from the smallest through the largest must be present.
246
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.combinatorics.prufer import Prufer
>>> Prufer.edges([1, 2, 3], [2, 4, 5]) # a T
([[0, 1], [1, 2], [1, 3], [3, 4]], 5)
Duplicate edges are removed:
>>> Prufer.edges([0, 1, 2, 3], [1, 4, 5], [1, 4, 6]) # a K
([[0, 1], [1, 2], [1, 4], [2, 3], [4, 5], [4, 6]], 7)
next(delta=1)
Generates the Prufer sequence that is delta beyond the current one.
See Also:
prufer rank (page 248), rank (page 248), prev (page 247), size (page 248)
Examples
>>> from sympy.combinatorics.prufer import Prufer
>>> a = Prufer([[0, 1], [0, 2], [0, 3]])
>>> b = a.next(1) # == a.next()
>>> b.tree_repr
[[0, 2], [0, 1], [1, 3]]
>>> b.rank
1
nodes
Returns the number of nodes in the tree.
Examples
>>> from sympy.combinatorics.prufer import Prufer
>>> Prufer([[0, 3], [1, 3], [2, 3], [3, 4], [4, 5]]).nodes
6
>>> Prufer([1, 0, 0]).nodes
5
prev(delta=1)
Generates the Prufer sequence that is -delta before the current one.
See Also:
prufer rank (page 248), rank (page 248), next (page 247), size (page 248)
Examples
>>> from sympy.combinatorics.prufer import Prufer
>>> a = Prufer([[0, 1], [1, 2], [2, 3], [1, 4]])
>>> a.rank
36
>>> b = a.prev()
>>> b
Prufer([1, 2, 0])
5.2. Combinatorics Module
247
SymPy Documentation, Release 0.7.6
>>> b.rank
35
prufer rank()
Computes the rank of a Prufer sequence.
See Also:
rank (page 248), next (page 247), prev (page 247), size (page 248)
Examples
>>> from sympy.combinatorics.prufer import Prufer
>>> a = Prufer([[0, 1], [0, 2], [0, 3]])
>>> a.prufer_rank()
0
prufer repr
Returns Prufer sequence for the Prufer object.
This sequence is found by removing the highest numbered vertex, recording the
node it was attached to, and continuuing until only two verices remain. The Prufer
sequence is the list of recorded nodes.
See Also:
to prufer (page 249)
Examples
>>>
>>>
[3,
>>>
[1,
from sympy.combinatorics.prufer import Prufer
Prufer([[0, 3], [1, 3], [2, 3], [3, 4], [4, 5]]).prufer_repr
3, 3, 4]
Prufer([1, 0, 0]).prufer_repr
0, 0]
rank
Returns the rank of the Prufer sequence.
See Also:
prufer rank (page 248), next (page 247), prev (page 247), size (page 248)
Examples
>>>
>>>
>>>
778
>>>
779
>>>
777
from sympy.combinatorics.prufer import Prufer
p = Prufer([[0, 3], [1, 3], [2, 3], [3, 4], [4, 5]])
p.rank
p.next(1).rank
p.prev().rank
size
Return the number of possible trees of this Prufer object.
248
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
See Also:
prufer rank (page 248), rank (page 248), next (page 247), prev (page 247)
Examples
>>> from sympy.combinatorics.prufer import Prufer
>>> Prufer([0]*4).size == Prufer([6]*4).size == 1296
True
static to prufer(tree, n)
Return the Prufer sequence for a tree given as a list of edges where n is the number
of nodes in the tree.
See Also:
prufer repr (page 248) returns Prufer sequence of a Prufer object.
Examples
>>>
>>>
>>>
[0,
>>>
[0,
from sympy.combinatorics.prufer import Prufer
a = Prufer([[0, 1], [0, 2], [0, 3]])
a.prufer_repr
0]
Prufer.to_prufer([[0, 1], [0, 2], [0, 3]], 4)
0]
static to tree(prufer)
Return the tree (as a list of edges) of the given Prufer sequence.
See Also:
tree repr (page 249) returns tree representation of a Prufer object.
References
http://hamberg.no/erlend/posts/2010-11-06-prufer-sequence-compact-treerepresentation.html
Examples
>>> from sympy.combinatorics.prufer import Prufer
>>> a = Prufer([0, 2], 4)
>>> a.tree_repr
[[0, 1], [0, 2], [2, 3]]
>>> Prufer.to_tree([0, 2])
[[0, 1], [0, 2], [2, 3]]
tree repr
Returns the tree representation of the Prufer object.
See Also:
to tree (page 249)
5.2. Combinatorics Module
249
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.combinatorics.prufer import Prufer
>>> Prufer([[0, 3], [1, 3], [2, 3], [3, 4], [4, 5]]).tree_repr
[[0, 3], [1, 3], [2, 3], [3, 4], [4, 5]]
>>> Prufer([1, 0, 0]).tree_repr
[[1, 2], [0, 1], [0, 3], [0, 4]]
classmethod unrank(rank, n)
Finds the unranked Prufer sequence.
Examples
>>> from sympy.combinatorics.prufer import Prufer
>>> Prufer.unrank(0, 4)
Prufer([0, 0])
Subsets
class sympy.combinatorics.subsets.Subset
Represents a basic subset object.
We generate subsets using essentially two techniques, binary enumeration and lexicographic enumeration. The Subset class takes two arguments, the rst one describes the
initial subset to consider and the second describes the superset.
Examples
>>> from sympy.combinatorics.subsets import Subset
>>> a = Subset([c,d], [a,b,c,d])
>>> a.next_binary().subset
[b]
>>> a.prev_binary().subset
[c]
classmethod bitlist from subset(subset, superset)
Gets the bitlist corresponding to a subset.
See Also:
subset from bitlist (page 254)
Examples
>>> from sympy.combinatorics.subsets import Subset
>>> Subset.bitlist_from_subset([c,d], [a,b,c,d])
0011
cardinality
Returns the number of all possible subsets.
See Also:
250
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
subset (page 254), superset (page 255), size (page 254), superset size
(page 255)
Examples
>>> from sympy.combinatorics.subsets import Subset
>>> a = Subset([c,d], [a,b,c,d])
>>> a.cardinality
16
iterate binary(k)
This is a helper function. It iterates over the binary subsets by k steps. This variable
can be both positive or negative.
See Also:
next binary (page 251), prev binary (page 252)
Examples
>>> from sympy.combinatorics.subsets import Subset
>>> a = Subset([c,d], [a,b,c,d])
>>> a.iterate_binary(-2).subset
[d]
>>> a = Subset([a,b,c], [a,b,c,d])
>>> a.iterate_binary(2).subset
[]
iterate graycode(k)
Helper function used for prev gray and next gray. It performs k step overs to get
the respective Gray codes.
See Also:
next gray (page 252), prev gray (page 253)
Examples
>>>
>>>
>>>
[1,
>>>
[1,
from sympy.combinatorics.subsets import Subset
a = Subset([1,2,3], [1,2,3,4])
a.iterate_graycode(3).subset
4]
a.iterate_graycode(-2).subset
2, 4]
next binary()
Generates the next binary ordered subset.
See Also:
prev binary (page 252), iterate binary (page 251)
5.2. Combinatorics Module
251
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.combinatorics.subsets import Subset
>>> a = Subset([c,d], [a,b,c,d])
>>> a.next_binary().subset
[b]
>>> a = Subset([a,b,c,d], [a,b,c,d])
>>> a.next_binary().subset
[]
next gray()
Generates the next Gray code ordered subset.
See Also:
iterate graycode (page 251), prev gray (page 253)
Examples
>>>
>>>
>>>
[1,
from sympy.combinatorics.subsets import Subset
a = Subset([1,2,3], [1,2,3,4])
a.next_gray().subset
3]
next lexicographic()
Generates the next lexicographically ordered subset.
See Also:
prev lexicographic (page 253)
Examples
>>> from sympy.combinatorics.subsets import Subset
>>> a = Subset([c,d], [a,b,c,d])
>>> a.next_lexicographic().subset
[d]
>>> a = Subset([d], [a,b,c,d])
>>> a.next_lexicographic().subset
[]
prev binary()
Generates the previous binary ordered subset.
See Also:
next binary (page 251), iterate binary (page 251)
Examples
>>> from sympy.combinatorics.subsets import Subset
>>> a = Subset([], [a,b,c,d])
>>> a.prev_binary().subset
[a, b, c, d]
>>> a = Subset([c,d], [a,b,c,d])
252
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> a.prev_binary().subset
[c]
prev gray()
Generates the previous Gray code ordered subset.
See Also:
iterate graycode (page 251), next gray (page 252)
Examples
>>>
>>>
>>>
[2,
from sympy.combinatorics.subsets import Subset
a = Subset([2,3,4], [1,2,3,4,5])
a.prev_gray().subset
3, 4, 5]
prev lexicographic()
Generates the previous lexicographically ordered subset.
See Also:
next lexicographic (page 252)
Examples
>>> from sympy.combinatorics.subsets import Subset
>>> a = Subset([], [a,b,c,d])
>>> a.prev_lexicographic().subset
[d]
>>> a = Subset([c,d], [a,b,c,d])
>>> a.prev_lexicographic().subset
[c]
rank binary
Computes the binary ordered rank.
See Also:
iterate binary (page 251), unrank binary (page 255)
Examples
>>>
>>>
>>>
0
>>>
>>>
3
from sympy.combinatorics.subsets import Subset
a = Subset([], [a,b,c,d])
a.rank_binary
a = Subset([c,d], [a,b,c,d])
a.rank_binary
rank gray
Computes the Gray code ranking of the subset.
See Also:
iterate graycode (page 251), unrank gray (page 256)
5.2. Combinatorics Module
253
SymPy Documentation, Release 0.7.6
Examples
>>>
>>>
>>>
2
>>>
>>>
27
from sympy.combinatorics.subsets import Subset
a = Subset([c,d], [a,b,c,d])
a.rank_gray
a = Subset([2,4,5], [1,2,3,4,5,6])
a.rank_gray
rank lexicographic
Computes the lexicographic ranking of the subset.
Examples
>>>
>>>
>>>
14
>>>
>>>
43
from sympy.combinatorics.subsets import Subset
a = Subset([c,d], [a,b,c,d])
a.rank_lexicographic
a = Subset([2,4,5], [1,2,3,4,5,6])
a.rank_lexicographic
size
Gets the size of the subset.
See Also:
subset (page 254), superset (page 255), superset size (page 255), cardinality
(page 250)
Examples
>>> from sympy.combinatorics.subsets import Subset
>>> a = Subset([c,d], [a,b,c,d])
>>> a.size
2
subset
Gets the subset represented by the current instance.
See Also:
superset (page 255), size (page 254), superset size (page 255), cardinality
(page 250)
Examples
>>> from sympy.combinatorics.subsets import Subset
>>> a = Subset([c,d], [a,b,c,d])
>>> a.subset
[c, d]
254
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
classmethod subset from bitlist(super set, bitlist)
Gets the subset dened by the bitlist.
See Also:
bitlist from subset (page 250)
Examples
>>> from sympy.combinatorics.subsets import Subset
>>> Subset.subset_from_bitlist([a,b,c,d], 0011).subset
[c, d]
classmethod subset indices(subset, superset)
Return indices of subset in superset in a list; the list is empty if all elements of subset
are not in superset.
Examples:
>>>
>>>
>>>
[1,
>>>
[]
>>>
[]
from sympy.combinatorics import Subset
superset = [1, 3, 2, 5, 4]
Subset.subset_indices([3, 2, 1], superset)
2, 0]
Subset.subset_indices([1, 6], superset)
Subset.subset_indices([], superset)
superset
Gets the superset of the subset.
See Also:
subset (page 254), size (page 254), superset size (page 255), cardinality
(page 250)
Examples
>>> from sympy.combinatorics.subsets import Subset
>>> a = Subset([c,d], [a,b,c,d])
>>> a.superset
[a, b, c, d]
superset size
Returns the size of the superset.
See Also:
subset (page 254), superset (page 255), size (page 254), cardinality (page 250)
Examples
>>> from sympy.combinatorics.subsets import Subset
>>> a = Subset([c,d], [a,b,c,d])
>>> a.superset_size
4
5.2. Combinatorics Module
255
SymPy Documentation, Release 0.7.6
classmethod unrank binary(rank, superset)
Gets the binary ordered subset of the specied rank.
See Also:
iterate binary (page 251), rank binary (page 253)
Examples
>>> from sympy.combinatorics.subsets import Subset
>>> Subset.unrank_binary(4, [a,b,c,d]).subset
[b]
classmethod unrank gray(rank, superset)
Gets the Gray code ordered subset of the specied rank.
See Also:
iterate graycode (page 251), rank gray (page 253)
Examples
>>> from sympy.combinatorics.subsets import Subset
>>> Subset.unrank_gray(4, [a,b,c]).subset
[a, b]
>>> Subset.unrank_gray(0, [a,b,c]).subset
[]
static subsets.ksubsets(superset, k)
Finds the subsets of size k in lexicographic order.
This uses the itertools generator.
See Also:
class Subset
Examples
>>> from sympy.combinatorics.subsets import ksubsets
>>> list(ksubsets([1,2,3], 2))
[(1, 2), (1, 3), (2, 3)]
>>> list(ksubsets([1,2,3,4,5], 2))
[(1, 2), (1, 3), (1, 4), (1, 5), (2, 3), (2, 4),
(2, 5), (3, 4), (3, 5), (4, 5)]
Gray Code
class sympy.combinatorics.graycode.GrayCode
A Gray code is essentially a Hamiltonian walk on a n-dimensional cube with edge length
of one. The vertices of the cube are represented by vectors whose values are binary.
The Hamilton walk visits each vertex exactly once. The Gray code for a 3d cube is
[000,100,110,010,011,111,101, 001].
A Gray code solves the problem of sequentially generating all possible subsets of n objects in such a way that each subset is obtained from the previous one by either deleting
256
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
or adding a single object. In the above example, 1 indicates that the object is present,
and 0 indicates that its absent.
Gray codes have applications in statistics as well when we want to compute various
statistics related to subsets in an ecient manner.
References: [1] Nijenhuis,A. and Wilf,H.S.(1978). Combinatorial Algorithms. Academic
Press. [2] Knuth, D. (2011). The Art of Computer Programming, Vol 4 Addison Wesley
Examples
>>> from sympy.combinatorics.graycode import GrayCode
>>> a = GrayCode(3)
>>> list(a.generate_gray())
[000, 001, 011, 010, 110, 111, 101, 100]
>>> a = GrayCode(4)
>>> list(a.generate_gray())
[0000, 0001, 0011, 0010, 0110, 0111, 0101, 0100,
1100, 1101, 1111, 11
current
Returns the currently referenced Gray code as a bit string.
Examples
>>> from sympy.combinatorics.graycode import GrayCode
>>> GrayCode(3, start=100).current
100
generate gray(**hints)
Generates the sequence of bit vectors of a Gray Code.
[1] Knuth, D. (2011). The Art of Computer Programming, Vol 4, Addison Wesley
See Also:
skip (page 258)
Examples
>>> from sympy.combinatorics.graycode import GrayCode
>>> a = GrayCode(3)
>>> list(a.generate_gray())
[000, 001, 011, 010, 110, 111, 101, 100]
>>> list(a.generate_gray(start=011))
[011, 010, 110, 111, 101, 100]
>>> list(a.generate_gray(rank=4))
[110, 111, 101, 100]
n
Returns the dimension of the Gray code.
5.2. Combinatorics Module
257
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.combinatorics.graycode import GrayCode
>>> a = GrayCode(5)
>>> a.n
5
next(delta=1)
Returns the Gray code a distance delta (default = 1) from the current value in
canonical order.
Examples
>>> from sympy.combinatorics.graycode import GrayCode
>>> a = GrayCode(3, start=110)
>>> a.next().current
111
>>> a.next(-1).current
010
rank
Ranks the Gray code.
A ranking algorithm determines the position (or rank) of a combinatorial object
among all the objects w.r.t. a given order. For example, the 4 bit binary reected
Gray code (BRGC) 0101 has a rank of 6 as it appears in the 6th position in the
canonical ordering of the family of 4 bit Gray codes.
References: [1] http://statweb.stanford.edu/susan/courses/s208/node12.html
See Also:
unrank (page 259)
Examples
>>> from sympy.combinatorics.graycode import GrayCode
>>> a = GrayCode(3)
>>> list(a.generate_gray())
[000, 001, 011, 010, 110, 111, 101, 100]
>>> GrayCode(3, start=100).rank
7
>>> GrayCode(3, rank=7).current
100
selections
Returns the number of bit vectors in the Gray code.
Examples
>>> from sympy.combinatorics.graycode import GrayCode
>>> a = GrayCode(3)
>>> a.selections
8
258
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
skip()
Skips the bit generation.
See Also:
generate gray (page 257)
Examples
>>> from sympy.combinatorics.graycode import GrayCode
>>> a = GrayCode(3)
>>> for i in a.generate_gray():
...
if i == 010:
...
a.skip()
...
print(i)
...
000
001
011
010
111
101
100
classmethod unrank(n, rank)
Unranks an n-bit sized Gray code of rank k. This method exists so that a derivative
GrayCode class can dene its own code of a given rank.
The string here is generated in reverse order to allow for tail-call optimization.
See Also:
rank (page 258)
Examples
>>> from sympy.combinatorics.graycode import GrayCode
>>> GrayCode(5, rank=3).current
00010
>>> GrayCode.unrank(5, 3)
00010
static graycode.random bitstring(n)
Generates a random bitlist of length n.
Examples
>>> from sympy.combinatorics.graycode import random_bitstring
>>> random_bitstring(3)
100
static graycode.gray to bin(bin list)
Convert from Gray coding to binary coding.
We assume big endian encoding.
See Also:
5.2. Combinatorics Module
259
SymPy Documentation, Release 0.7.6
bin to gray (page 260)
Examples
>>> from sympy.combinatorics.graycode import gray_to_bin
>>> gray_to_bin(100)
111
static graycode.bin to gray(bin list)
Convert from binary coding to gray coding.
We assume big endian encoding.
See Also:
gray to bin (page 259)
Examples
>>> from sympy.combinatorics.graycode import bin_to_gray
>>> bin_to_gray(111)
100
static graycode.get subset from bitstring(super set, bitstring)
Gets the subset dened by the bitstring.
See Also:
graycode subsets (page 260)
Examples
>>> from sympy.combinatorics.graycode import get_subset_from_bitstring
>>> get_subset_from_bitstring([a,b,c,d], 0011)
[c, d]
>>> get_subset_from_bitstring([c,a,c,c], 1100)
[c, a]
static graycode.graycode subsets(gray code set)
Generates the subsets as enumerated by a Gray code.
See Also:
get subset from bitstring (page 260)
Examples
>>> from sympy.combinatorics.graycode import graycode_subsets
>>> list(graycode_subsets([a,b,c]))
[[], [c], [b, c], [b], [a, b], [a, b, c],
>>> list(graycode_subsets([a,b,c,c]))
[[], [c], [c, c], [c], [b, c], [b, c, c],
260
[a, c], [a]]
[b, c], [b], [a, b], [
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Named Groups
sympy.combinatorics.named groups.SymmetricGroup(n)
Generates the symmetric group on n elements as a permutation group.
The generators taken are the n-cycle (0 1 2 ... n-1) and the transposition (0 1) (in
cycle notation). (See [1]). After the group is generated, some of its basic properties are
set.
See Also:
CyclicGroup (page 261), DihedralGroup (page 261), AlternatingGroup (page 262)
References
[1] http://en.wikipedia.org/wiki/Symmetric group#Generators and relations
Examples
>>> from sympy.combinatorics.named_groups import SymmetricGroup
>>> G = SymmetricGroup(4)
>>> G.is_group()
False
>>> G.order()
24
>>> list(G.generate_schreier_sims(af=True))
[[0, 1, 2, 3], [1, 2, 3, 0], [2, 3, 0, 1], [3, 1, 2, 0], [0, 2, 3, 1],
[1, 3, 0, 2], [2, 0, 1, 3], [3, 2, 0, 1], [0, 3, 1, 2], [1, 0, 2, 3],
[2, 1, 3, 0], [3, 0, 1, 2], [0, 1, 3, 2], [1, 2, 0, 3], [2, 3, 1, 0],
[3, 1, 0, 2], [0, 2, 1, 3], [1, 3, 2, 0], [2, 0, 3, 1], [3, 2, 1, 0],
[0, 3, 2, 1], [1, 0, 3, 2], [2, 1, 0, 3], [3, 0, 2, 1]]
sympy.combinatorics.named groups.CyclicGroup(n)
Generates the cyclic group of order n as a permutation group.
The generator taken is the n-cycle (0 1 2 ... n-1) (in cycle notation). After the group
is generated, some of its basic properties are set.
See Also:
SymmetricGroup (page 261), DihedralGroup (page 261), AlternatingGroup (page 262)
Examples
>>> from sympy.combinatorics.named_groups import CyclicGroup
>>> G = CyclicGroup(6)
>>> G.is_group()
False
>>> G.order()
6
>>> list(G.generate_schreier_sims(af=True))
[[0, 1, 2, 3, 4, 5], [1, 2, 3, 4, 5, 0], [2, 3, 4, 5, 0, 1],
[3, 4, 5, 0, 1, 2], [4, 5, 0, 1, 2, 3], [5, 0, 1, 2, 3, 4]]
5.2. Combinatorics Module
261
SymPy Documentation, Release 0.7.6
sympy.combinatorics.named groups.DihedralGroup(n)
Generates the dihedral group Dn as a permutation group.
The dihedral group Dn is the group of symmetries of the regular n-gon. The generators
taken are the n-cycle a = (0 1 2 ... n-1) (a rotation of the n-gon) and b = (0 n-1)(1
n-2)... (a reection of the n-gon) in cycle rotation. It is easy to see that these satisfy
a**n = b**2 = 1 and bab = a so they indeed generate Dn (See [1]). After the group is
generated, some of its basic properties are set.
See Also:
SymmetricGroup (page 261), CyclicGroup (page 261), AlternatingGroup (page 262)
References
[1] http://en.wikipedia.org/wiki/Dihedral group
Examples
>>> from sympy.combinatorics.named_groups import DihedralGroup
>>> G = DihedralGroup(5)
>>> G.is_group()
False
>>> a = list(G.generate_dimino())
>>> [perm.cyclic_form for perm in a]
[[], [[0, 1, 2, 3, 4]], [[0, 2, 4, 1, 3]],
[[0, 3, 1, 4, 2]], [[0, 4, 3, 2, 1]], [[0, 4], [1, 3]],
[[1, 4], [2, 3]], [[0, 1], [2, 4]], [[0, 2], [3, 4]],
[[0, 3], [1, 2]]]
sympy.combinatorics.named groups.AlternatingGroup(n)
Generates the alternating group on n elements as a permutation group.
For n > 2, the generators taken are (0 1 2), (0 1 2 ... n-1) for n odd and (0 1 2),
(1 2 ... n-1) for n even (See [1], p.31, ex.6.9.). After the group is generated, some of
its basic properties are set. The cases n = 1, 2 are handled separately.
See Also:
SymmetricGroup (page 261), CyclicGroup (page 261), DihedralGroup (page 261)
References
[1] Armstrong, M. Groups and Symmetry
Examples
>>> from sympy.combinatorics.named_groups import AlternatingGroup
>>> G = AlternatingGroup(4)
>>> G.is_group()
False
>>> a = list(G.generate_dimino())
>>> len(a)
12
262
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> all(perm.is_even for perm in a)
True
sympy.combinatorics.named groups.AbelianGroup(*cyclic orders)
Returns the direct product of cyclic groups with the given orders.
According to the structure theorem for nite abelian groups ([1]), every nite abelian
group can be written as the direct product of nitely many cyclic groups.
[1]
http://groupprops.subwiki.org/wiki/Structure theorem for nitely generated abelian groups
See Also:
DirectProduct
Examples
>>> from sympy.combinatorics import Permutation
>>> Permutation.print_cyclic = True
>>> from sympy.combinatorics.named_groups import AbelianGroup
>>> AbelianGroup(3, 4)
PermutationGroup([
Permutation(6)(0, 1, 2),
Permutation(3, 4, 5, 6)])
>>> _.is_group()
False
Utilities
sympy.combinatorics.util. base ordering(base, degree)
Order {0, 1, ..., n 1} so that base points come rst and in order.
Parameters base - the base :
degree - the degree of the associated permutation group :
Returns A list base ordering such that base ordering[point] is the
:
number of point in the ordering. :
Examples :
======== :
>>> from sympy.combinatorics.named groups import SymmetricGroup :
>>> from sympy.combinatorics.util import base ordering :
>>> S = SymmetricGroup(4) :
>>> S.schreier sims() :
>>> base ordering(S.base, S.degree) :
[0, 1, 2, 3] :
5.2. Combinatorics Module
263
SymPy Documentation, Release 0.7.6
Notes
This is used in backtrack searches, when we dene a relation << on the underlying set
for a permutation group of degree n, {0, 1, ..., n 1}, so that if (b1 , b2 , ..., bk ) is a base we
have bi << bj whenever i < j and bi << a for all i {1, 2, ..., k} and a is not in the base. The
idea is developed and applied to backtracking algorithms in [1], pp.108-132. The points
that are not in the base are taken in increasing order.
References
[1] Holt, D., Eick, B., OBrien, E. Handbook of computational group theory
sympy.combinatorics.util. check cycles alt sym(perm)
Checks for cycles of prime length p with n/2 < p < n-2.
Here n is the degree of the permutation. This is a helper function for the function
is alt sym from sympy.combinatorics.perm groups.
See Also:
sympy.combinatorics.perm groups.PermutationGroup.is alt sym (page 228)
Examples
>>> from sympy.combinatorics.util import _check_cycles_alt_sym
>>> from sympy.combinatorics.permutations import Permutation
>>> a = Permutation([[0,1,2,3,4,5,6,7,8,9,10], [11, 12]])
>>> _check_cycles_alt_sym(a)
False
>>> b = Permutation([[0,1,2,3,4,5,6], [7,8,9,10]])
>>> _check_cycles_alt_sym(b)
True
sympy.combinatorics.util. distribute gens by base(base, gens)
Distribute the group elements gens by membership in basic stabilizers.
Notice that for a base (b1 , b2 , ..., bk ), the basic stabilizers are dened as G(i) = Gb1 ,...,bi1 for
i {1, 2, ..., k}.
Parameters base - a sequence of points in {0, 1, ..., n-1} :
gens - a list of elements of a permutation group of degree n. :
Returns List of length k, where k is :
the length of base. The i-th entry contains those elements in :
gens which x the rst i elements of base (so that the :
0-th entry is equal to gens itself). If no element xes the rst :
i elements of base, the i-th element is set to a list containing :
the identity element. :
See Also:
strong gens from distr (page 268),
handle precomputed bsgs (page 265)
264
orbits transversals from bsgs (page 266),
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.combinatorics import Permutation
>>> Permutation.print_cyclic = True
>>> from sympy.combinatorics.named_groups import DihedralGroup
>>> from sympy.combinatorics.util import _distribute_gens_by_base
>>> D = DihedralGroup(3)
>>> D.schreier_sims()
>>> D.strong_gens
[Permutation(0, 1, 2), Permutation(0, 2), Permutation(1, 2)]
>>> D.base
[0, 1]
>>> _distribute_gens_by_base(D.base, D.strong_gens)
[[Permutation(0, 1, 2), Permutation(0, 2), Permutation(1, 2)],
[Permutation(1, 2)]]
sympy.combinatorics.util. handle precomputed bsgs(base,
strong gens,
transversals=None,
basic orbits=None,
strong gens distr=None)
Calculate BSGS-related structures from those present.
The base and strong generating set must be provided; if any of the transversals, basic
orbits or distributed strong generators are not provided, they will be calculated from the
base and strong generating set.
Parameters base - the base :
strong gens - the strong generators :
transversals - basic transversals :
basic orbits - basic orbits :
strong gens distr - strong generators distributed by membership
in basic :
stabilizers :
Returns (transversals,
transversals :
basic orbits,
strong gens distr)
where
are the basic transversals, basic orbits are the basic orbits, and :
strong gens distr are the strong generators distributed by membership :
in basic stabilizers. :
See Also:
orbits transversals from bsgs (page 266), distribute gens by base
Examples
>>>
>>>
>>>
>>>
>>>
>>>
from sympy.combinatorics import Permutation
Permutation.print_cyclic = True
from sympy.combinatorics.named_groups import DihedralGroup
from sympy.combinatorics.util import _handle_precomputed_bsgs
D = DihedralGroup(3)
D.schreier_sims()
5.2. Combinatorics Module
265
SymPy Documentation, Release 0.7.6
>>> _handle_precomputed_bsgs(D.base, D.strong_gens,
... basic_orbits=D.basic_orbits)
([{0: Permutation(2), 1: Permutation(0, 1, 2), 2: Permutation(0, 2)},
{1: Permutation(2), 2: Permutation(1, 2)}],
[[0, 1, 2], [1, 2]], [[Permutation(0, 1, 2),
Permutation(0, 2),
Permutation(1, 2)],
[Permutation(1, 2)]])
sympy.combinatorics.util. orbits transversals from bsgs(base,
strong gens distr,
transversals only=False)
Compute basic orbits and transversals from a base and strong generating set.
The generators are provided as distributed across the basic stabilizers. If the optional
argument transversals only is set to True, only the transversals are returned.
Parameters base - the base :
strong gens distr - strong generators distributed by membership
in basic :
stabilizers :
transversals only - a ag switching between returning only the :
transversals/ both orbits and transversals :
See Also:
distribute gens by base (page 264), handle precomputed bsgs (page 265)
Examples
>>> from sympy.combinatorics import Permutation
>>> Permutation.print_cyclic = True
>>> from sympy.combinatorics.named_groups import SymmetricGroup
>>> from sympy.combinatorics.util import _orbits_transversals_from_bsgs
>>> from sympy.combinatorics.util import (_orbits_transversals_from_bsgs,
... _distribute_gens_by_base)
>>> S = SymmetricGroup(3)
>>> S.schreier_sims()
>>> strong_gens_distr = _distribute_gens_by_base(S.base, S.strong_gens)
>>> _orbits_transversals_from_bsgs(S.base, strong_gens_distr)
([[0, 1, 2], [1, 2]],
[{0: Permutation(2), 1: Permutation(0, 1, 2), 2: Permutation(0, 2, 1)},
{1: Permutation(2), 2: Permutation(1, 2)}])
sympy.combinatorics.util. remove gens(base,
strong gens,
basic orbits=None,
strong gens distr=None)
Remove redundant generators from a strong generating set.
Parameters base - a base :
strong gens - a strong generating set relative to base :
basic orbits - basic orbits :
strong gens distr - strong generators distributed by membership
in basic :
266
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
stabilizers :
Returns A strong generating set with respect to base which is a subset of :
strong gens. :
Notes
This procedure is outlined in [1],p.95.
References
[1] Holt, D., Eick, B., OBrien, E. Handbook of computational group theory
Examples
>>> from sympy.combinatorics.named_groups import SymmetricGroup
>>> from sympy.combinatorics.perm_groups import PermutationGroup
>>> from sympy.combinatorics.util import _remove_gens
>>> from sympy.combinatorics.testutil import _verify_bsgs
>>> S = SymmetricGroup(15)
>>> base, strong_gens = S.schreier_sims_incremental()
>>> len(strong_gens)
26
>>> new_gens = _remove_gens(base, strong_gens)
>>> len(new_gens)
14
>>> _verify_bsgs(S, base, new_gens)
True
sympy.combinatorics.util. strip(g, base, orbits, transversals)
Attempt to decompose a permutation using a (possibly partial) BSGS structure.
This is done by treating the sequence base as an actual base, and the orbits orbits and
transversals transversals as basic orbits and transversals relative to it.
This process is called sifting. A sift is unsuccessful when a certain orbit element is not
found or when after the sift the decomposition doesnt end with the identity element.
The argument transversals is a list of dictionaries that provides transversal elements
for the orbits orbits.
Parameters g - permutation to be decomposed :
base - sequence of points :
orbits - a list in which the i-th entry is an orbit of base[i] :
under some subgroup of the pointwise stabilizer of :
base[0], base[1], ..., base[i - 1]. The groups themselves are implicit
:
in this function since the only information we need is encoded in the
orbits :
and transversals :
5.2. Combinatorics Module
267
SymPy Documentation, Release 0.7.6
transversals - a list of orbit transversals associated with the orbits
:
orbits. :
See Also:
sympy.combinatorics.perm groups.PermutationGroup.schreier sims (page
sympy.combinatorics.perm groups.PermutationGroup.schreier sims random
(page 239)
238),
Notes
The algorithm is described in [1],pp.89-90. The reason for returning both the current
state of the element being decomposed and the level at which the sifting ends is that
they provide important information for the randomized version of the Schreier-Sims algorithm.
References
[1] Holt, D., Eick, B., OBrien, E. Handbook of computational group theory
Examples
>>> from sympy.combinatorics import Permutation
>>> Permutation.print_cyclic = True
>>> from sympy.combinatorics.named_groups import SymmetricGroup
>>> from sympy.combinatorics.permutations import Permutation
>>> from sympy.combinatorics.util import _strip
>>> S = SymmetricGroup(5)
>>> S.schreier_sims()
>>> g = Permutation([0, 2, 3, 1, 4])
>>> _strip(g, S.base, S.basic_orbits, S.basic_transversals)
(Permutation(4), 5)
sympy.combinatorics.util. strong gens from distr(strong gens distr)
Retrieve strong generating set from generators of basic stabilizers.
This is just the union of the generators of the rst and second basic stabilizers.
Parameters strong gens distr - strong generators distributed by
membership in basic :
stabilizers :
See Also:
distribute gens by base (page 264)
Examples
>>>
>>>
>>>
>>>
268
from sympy.combinatorics import Permutation
Permutation.print_cyclic = True
from sympy.combinatorics.named_groups import SymmetricGroup
from sympy.combinatorics.util import (_strong_gens_from_distr,
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
... _distribute_gens_by_base)
>>> S = SymmetricGroup(3)
>>> S.schreier_sims()
>>> S.strong_gens
[Permutation(0, 1, 2), Permutation(2)(0, 1), Permutation(1, 2)]
>>> strong_gens_distr = _distribute_gens_by_base(S.base, S.strong_gens)
>>> _strong_gens_from_distr(strong_gens_distr)
[Permutation(0, 1, 2), Permutation(2)(0, 1), Permutation(1, 2)]
Group constructors
sympy.combinatorics.group constructs.DirectProduct(*groups)
Returns the direct product of several groups as a permutation group.
This is implemented much like the mul procedure for taking the direct product of two
permutation groups, but the idea of shifting the generators is realized in the case of an
arbitrary number of groups. A call to DirectProduct(G1, G2, ..., Gn) is generally expected
to be faster than a call to G1*G2*...*Gn (and thus the need for this algorithm).
See Also:
mul
Examples
>>>
>>>
>>>
>>>
>>>
64
from sympy.combinatorics.group_constructs import DirectProduct
from sympy.combinatorics.named_groups import CyclicGroup
C = CyclicGroup(4)
G = DirectProduct(C,C,C)
G.order()
Test Utilities
sympy.combinatorics.testutil. cmp perm lists(rst, second)
Compare two lists of permutations as sets.
This is used for testing purposes. Since the array form of a permutation is currently a
list, Permutation is not hashable and cannot be put into a set.
Examples
>>> from sympy.combinatorics.permutations import Permutation
>>> from sympy.combinatorics.testutil import _cmp_perm_lists
>>> a = Permutation([0, 2, 3, 4, 1])
>>> b = Permutation([1, 2, 0, 4, 3])
>>> c = Permutation([3, 4, 0, 1, 2])
>>> ls1 = [a, b, c]
>>> ls2 = [b, c, a]
>>> _cmp_perm_lists(ls1, ls2)
True
sympy.combinatorics.testutil. naive list centralizer(self, other, af=False)
5.2. Combinatorics Module
269
SymPy Documentation, Release 0.7.6
sympy.combinatorics.testutil. verify bsgs(group, base, gens)
Verify the correctness of a base and strong generating set.
This is a naive implementation using the denition of a base and a strong generating set
relative to it. There are other procedures for verifying a base and strong generating set,
but this one will serve for more robust testing.
See Also:
sympy.combinatorics.perm groups.PermutationGroup.schreier sims (page 238)
Examples
>>> from sympy.combinatorics.named_groups import AlternatingGroup
>>> from sympy.combinatorics.testutil import _verify_bsgs
>>> A = AlternatingGroup(4)
>>> A.schreier_sims()
>>> _verify_bsgs(A, A.base, A.strong_gens)
True
sympy.combinatorics.testutil. verify centralizer(group, arg, centr=None)
Verify the centralizer of a group/set/element inside another group.
This is used for testing .centralizer() from sympy.combinatorics.perm groups
See Also:
naive list centralizer (page 269), sympy.combinatorics.perm groups.PermutationGroup.centra
(page 221), cmp perm lists (page 269)
Examples
>>> from sympy.combinatorics.named_groups import (SymmetricGroup,
... AlternatingGroup)
>>> from sympy.combinatorics.perm_groups import PermutationGroup
>>> from sympy.combinatorics.permutations import Permutation
>>> from sympy.combinatorics.testutil import _verify_centralizer
>>> S = SymmetricGroup(5)
>>> A = AlternatingGroup(5)
>>> centr = PermutationGroup([Permutation([0, 1, 2, 3, 4])])
>>> _verify_centralizer(S, A, centr)
True
sympy.combinatorics.testutil. verify normal closure(group, arg, closure=None)
Tensor Canonicalization
sympy.combinatorics.tensor can.canonicalize(g, dummies, msym, *v)
canonicalize tensor formed by tensors
Parameters g : permutation representing the tensor
dummies : list representing the dummy indices
it can be a list of dummy indices of the same type or a list of lists of
dummy indices, one list for each type of index; the dummy indices
must come after the free indices, and put in order contravariant, covariant [d0, -d0, d1,-d1,...]
270
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
msym : symmetry of the metric(s)
it can be an integer or a list; in the rst case it is the symmetry of the
dummy index metric; in the second case it is the list of the symmetries
of the index metric for each type
v : list, (base i, gens i, n i, sym i) for tensors of type i
base i, gens i : BSGS for tensors of this type.
The BSGS should have minimal base under lexicographic ordering; if
not, an attempt is made do get the minimal BSGS; in case of failure,
canonicalize naive is used, which is much slower.
n i : number of tensors of type i.
sym i : symmetry under exchange of component tensors of type i.
Both for msym and sym i the cases are
None no symmetry
0 commuting
1 anticommuting
Returns 0 if the tensor is zero, else return the array form of :
the permutation representing the canonical form of the tensor. :
Examples
one type of index with commuting metric;
Aab and Bab antisymmetric and commuting
T = Ad0d1 B d0 d2 B d2d1
ord = [d0, d0, d1, d1, d2, d2] order of the indices
g = [1,3,0,5,4,2,6,7]
Tc = 0
>>>
>>>
>>>
>>>
>>>
>>>
>>>
0
from sympy.combinatorics.tensor_can import get_symmetric_group_sgs, canonicalize, bsgs_direc
from sympy.combinatorics import Permutation
base2a, gens2a = get_symmetric_group_sgs(2, 1)
t0 = (base2a, gens2a, 1, 0)
t1 = (base2a, gens2a, 2, 0)
g = Permutation([1,3,0,5,4,2,6,7])
canonicalize(g, range(6), 0, t0, t1)
same as above, but with Bab anticommuting
Tc = Ad0d1 Bd0 d2 Bd1d2
can = [0,2,1,4,3,5,7,6]
>>> t1 = (base2a, gens2a, 2, 1)
>>> canonicalize(g, range(6), 0, t0, t1)
[0, 2, 1, 4, 3, 5, 7, 6]
two types of indices [a, b, c, d, e, f ] and [m, n], in this order, both with commuting metric
f abc antisymmetric, commuting
5.2. Combinatorics Module
271
SymPy Documentation, Release 0.7.6
Ama no symmetry, commuting
T = f c da f f eb Am d Amb An a Ane
ord = [c,f,a,-a,b,-b,d,-d,e,-e,m,-m,n,-n]
g = [0,7,3, 1,9,5, 11,6, 10,4, 13,2, 12,8, 14,15]
The canonical tensor is Tc = f cab f f de Am a Amd An b Ane
can = [0,2,4, 1,6,8, 10,3, 11,7, 12,5, 13,9, 15,14]
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
[0,
base_f, gens_f = get_symmetric_group_sgs(3, 1)
base1, gens1 = get_symmetric_group_sgs(1)
base_A, gens_A = bsgs_direct_product(base1, gens1, base1, gens1)
t0 = (base_f, gens_f, 2, 0)
t1 = (base_A, gens_A, 4, 0)
dummies = [range(2, 10), range(10, 14)]
g = Permutation([0,7,3,1,9,5,11,6,10,4,13,2,12,8,14,15])
canonicalize(g, dummies, [0, 0], t0, t1)
2, 4, 1, 6, 8, 10, 3, 11, 7, 12, 5, 13, 9, 15, 14]
Algorithm
First one uses canonical free to get the minimum tensor under lexicographic order, using only the slot symmetries. If the component tensors have not minimal BSGS, it is
attempted to nd it; if the attempt fails canonicalize naive is used instead.
Compute the residual slot symmetry keeping xed the free indices using tensor gens(base, gens, list free indices, sym).
Reduce the problem eliminating the free indices.
Then use double coset can rep and lift back the result reintroducing the free indices.
sympy.combinatorics.tensor can.double coset can rep(dummies, sym, b S, sgens,
S transversals, g)
Butler-Portugal algorithm for tensor canonicalization with dummy indices
dummies list of lists of dummy indices, one list for each type of index; the
dummy indices are put in order contravariant, covariant [d0, -d0, d1, -d1,
...].
sym list of the symmetries of the index metric for each type.
possible symmetries of the metrics
0 symmetric
1 antisymmetric
None no symmetry
b S base of a minimal slot symmetry BSGS.
sgens generators of the slot symmetry BSGS.
S transversals transversals for the slot BSGS.
g permutation representing the tensor.
Return 0 if the tensor is zero, else return the array form of the permutation representing
the canonical form of the tensor.
A tensor with dummy indices can be represented in a number of equivalent ways which
typically grows exponentially with the number of indices. To be able to establish if two
272
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
tensors with many indices are equal becomes computationally very slow in absence of
an ecient algorithm.
The Butler-Portugal algorithm [3] is an ecient algorithm to put tensors in canonical
form, solving the above problem.
Portugal observed that a tensor can be represented by a permutation, and that the class
of tensors equivalent to it under slot and dummy symmetries is equivalent to the double
coset D g S (Note: in this documentation we use the conventions for multiplication
of permutations p, q with (p*q)(i) = p[q[i]] which is opposite to the one used in the
Permutation class)
Using the algorithm by Butler to nd a representative of the double coset one can nd a
canonical form for the tensor.
To see this correspondence, let g be a permutation in array form; a tensor with indices
ind (the indices including both the contravariant and the covariant ones) can be written
as
t = T (ind[g[0], ..., ind[g[n 1]]),
where n = len(ind); g has size n + 2, the last two indices for the sign of the tensor (trick
introduced in [4]).
A slot symmetry transformation s is a permutation acting on the slots t > T (ind[(g
s)[0]], ..., ind[(g s)[n 1]])
A dummy symmetry transformation acts on ind t > T (ind[(d g)[0]], ..., ind[(d g)[n 1]])
Being interested only in the transformations of the tensor under these symmetries, one
can represent the tensor by g, which transforms as
g > d g s, so it belongs to the coset D g S.
Let us explain the conventions by an example.
Given a tensor T d3d2d1 d1d2d3 with the slot symmetries T a0a1a2a3a4a5 = T a2a1a0a3a4a5
T a0a1a2a3a4a5 = T a4a1a2a3a0a5
and symmetric metric, nd the tensor equivalent to it which is the lowest under the
ordering of indices: lexicographic ordering d1, d2, d3 then and contravariant index before
covariant index; that is the canonical form of the tensor.
The canonical form is T d1d2d3 d1d2d3 obtained using T a0a1a2a3a4a5 = T a2a1a0a3a4a5 .
To convert this problem in the input for this function, use the following labelling of the
index names (- for covariant for short) d1, d1, d2, d2, d3, d3
T d3d2d1 d1d2d3 corresponds to g = [4, 2, 0, 1, 3, 5, 6, 7] where the last two indices are for the sign
sgens = [P ermutation(0, 2)(6, 7), P ermutation(0, 4)(6, 7)]
sgens[0] is the slot symmetry (0, 2) T a0a1a2a3a4a5 = T a2a1a0a3a4a5
sgens[1] is the slot symmetry (0, 4) T a0a1a2a3a4a5 = T a4a1a2a3a0a5
The dummy symmetry group D is generated by the strong base generators
[(0, 1), (2, 3), (4, 5), (0, 1)(2, 3), (2, 3)(4, 5)]
The dummy symmetry acts from the left d = [1, 0, 2, 3, 4, 5, 6, 7] exchange d1 > d1
T d3d2d1 d1d2d3 == T d3d2 d1 d1 d2d3
g = [4, 2, 0, 1, 3, 5, 6, 7] > [4, 2, 1, 0, 3, 5, 6, 7] =a fr mul(d, g) which diers from a fr mul(g, d).
The slot symmetry acts from the right s = [2, 1, 0, 3, 4, 5, 7, 6] exchanges slots 0 and 2 and
changes sign T d3d2d1 d1d2d3 == T d1d2d3 d1d2d3
5.2. Combinatorics Module
273
SymPy Documentation, Release 0.7.6
g = [4, 2, 0, 1, 3, 5, 6, 7] > [0, 2, 4, 1, 3, 5, 7, 6] =a fr mul(g, s)
Example in which the tensor is zero, same slot symmetries as above: T d3 d1,d2 d1 d3 d2
= T d3 d1,d3 d1 d2 d2 under slot symmetry (2, 4);
= Td3d1 d3d1 d2 d2 under slot symmetry (0, 2);
= T d3 d1d3 d1 d2 d2 symmetric metric;
= 0 since two of these lines have tensors dier only for the sign.
The double coset D*g*S consists of permutations h = d g s corresponding to equivalent
tensors; if there are two h which are the same apart from the sign, return zero; otherwise
choose as representative the tensor with indices ordered lexicographically according to
[d1, d1, d2, d2, d3, d3] that is rep = min(D g S) = min([d g sf ordinDf orsinS])
The indices are xed one by one; rst choose the lowest index for slot 0, then the lowest
remaining index for slot 1, etc. Doing this one obtains a chain of stabilizers
S > Sb0 > Sb0,b1 > ... and D > Dp0 > Dp0,p1 > ...
where [b0, b1, ...] = range(b) is a base of the symmetric group; the strong base bS of S is an
ordered sublist of it; therefore it is sucient to compute once the strong base generators
of S using the Schreier-Sims algorithm; the stabilizers of the strong base generators are
the strong base generators of the stabilizer subgroup.
dbase = [p0, p1, ...] is not in general in lexicographic order, so that one must recompute
the strong base generators each time; however this is trivial, there is no need to use the
Schreier-Sims algorithm for D.
The algorithm keeps a TAB of elements (si , di , hi ) where hi = di g si satisfying hi [j] = pj
for 0 <= j < i starting from s0 = id, d0 = id, h0 = g.
The equations h0 [0] = p0 , h1 [1] = p1 , ... are solved in this order, choosing each time the
lowest possible value of p i
For j < i di g si Sb0 ,...,bi1 bj = Dp0 ,...,pi1 pj so that for dx in Dp0 ,...,pi1 and sx in
Sbase[0],...,base[i1] one has dx di g si sx bj = pj
Search for dx, sx such that this equation holds for j = i; it can be written as si sx bj =
J, dx di g J = pj sx bj = si 1 J; sx = trace(si 1, Sb0 ,...,bi1 ) dx 1 pj = di g J; dx =
trace(di g J, Dp0 ,...,pi1 )
si+1 = si trace(si 1 J, Sb0 ,...,bi1 ) di+1 = trace(di g J, Dp0 ,...,pi1 ) 1 di hi+1 bi =
di+1 g si+1 bi = pi
hn bj = pj for all j, so that hn is the solution.
Add the found (s, d, h) to TAB1.
At the end of the iteration sort TAB1 with respect to the h; if there are two consecutive
h in TAB1 which dier only for the sign, the tensor is zero, so return 0; if there are two
consecutive h which are equal, keep only one.
Then stabilize the slot generators under i and the dummy generators under pi .
Assign T AB = T AB1 at the end of the iteration step.
At the end T AB contains a unique (s, d, h), since all the slots of the tensor h have been
xed to have the minimum value according to the symmetries. The algorithm returns h.
It is important that the slot BSGS has lexicographic minimal base, otherwise there is an
i which does not belong to the slot base for which pi is xed by the dummy symmetry
only, while i is not invariant from the slot stabilizer, so pi is not in general the minimal
value.
274
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
This algorithm diers slightly from the original algorithm [3]: the
canonical
form is minimal lexicographically, and the BSGS has minimal base under lexicographic order. Equal tensors h are eliminated from TAB.
Examples
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
[0,
from sympy.combinatorics.permutations import Permutation
from sympy.combinatorics.perm_groups import PermutationGroup
from sympy.combinatorics.tensor_can import double_coset_can_rep, get_transversals
gens = [Permutation(x) for x in [[2,1,0,3,4,5,7,6], [4,1,2,3,0,5,7,6]]]
base = [0, 2]
g = Permutation([4,2,0,1,3,5,6,7])
transversals = get_transversals(base, gens)
double_coset_can_rep([list(range(6))], [0], base, gens, transversals, g)
1, 2, 3, 4, 5, 7, 6]
>>> g = Permutation([4,1,3,0,5,2,6,7])
>>> double_coset_can_rep([list(range(6))], [0], base, gens, transversals, g)
0
sympy.combinatorics.tensor can.get symmetric group sgs(n, antisym=False)
Return base, gens of the minimal BSGS for (anti)symmetric tensor
n rank of the tensor
antisym = False symmetric tensor antisym = True antisymmetric tensor
Examples
>>> from sympy.combinatorics import Permutation
>>> from sympy.combinatorics.tensor_can import get_symmetric_group_sgs
>>> Permutation.print_cyclic = True
>>> get_symmetric_group_sgs(3)
([0, 1], [Permutation(4)(0, 1), Permutation(4)(1, 2)])
sympy.combinatorics.tensor can.bsgs direct product(base1, gens1, base2, gens2,
signed=True)
direct product of two BSGS
base1 base of the rst BSGS.
gens1 strong generating sequence of the rst BSGS.
base2, gens2 similarly for the second BSGS.
signed ag for signed permutations.
Examples
>>> from sympy.combinatorics import Permutation
>>> from sympy.combinatorics.tensor_can import (get_symmetric_group_sgs, bsgs_direct_product)
>>> Permutation.print_cyclic = True
>>> base1, gens1 = get_symmetric_group_sgs(1)
>>> base2, gens2 = get_symmetric_group_sgs(2)
>>> bsgs_direct_product(base1, gens1, base2, gens2)
([1], [Permutation(4)(1, 2)])
5.2. Combinatorics Module
275
SymPy Documentation, Release 0.7.6
5.3 Number Theory
5.3.1 Ntheory Class Reference
class sympy.ntheory.generate.Sieve
An innite list of prime numbers, implemented as a dynamically growing sieve of Eratosthenes. When a lookup is requested involving an odd number that has not been sieved,
the sieve is automatically extended up to that number.
>>> from sympy import sieve
>>> from array import array # this line and next for doctest only
>>> sieve._list = array(l, [2, 3, 5, 7, 11, 13])
>>> 25 in sieve
False
>>> sieve._list
array(l, [2, 3, 5, 7, 11, 13, 17, 19, 23])
extend(n)
Grow the sieve to cover all primes <= n (a real number).
Examples
>>> from sympy import sieve
>>> from array import array # this line and next for doctest only
>>> sieve._list = array(l, [2, 3, 5, 7, 11, 13])
>>> sieve.extend(30)
>>> sieve[10] == 29
True
extend to no(i)
Extend to include the ith prime number.
i must be an integer.
The list is extended by 50% if it is too short, so it is likely that it will be longer than
requested.
Examples
>>> from sympy import sieve
>>> from array import array # this line and next for doctest only
>>> sieve._list = array(l, [2, 3, 5, 7, 11, 13])
>>> sieve.extend_to_no(9)
>>> sieve._list
array(l, [2, 3, 5, 7, 11, 13, 17, 19, 23])
primerange(a, b)
Generate all prime numbers in the range [a, b).
276
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import sieve
>>> print([i for i in sieve.primerange(7, 18)])
[7, 11, 13, 17]
search(n)
Return the indices i, j of the primes that bound n.
If n is prime then i == j.
Although n can be an expression, if ceiling cannot convert it to an integer then an n
error will be raised.
Examples
>>>
>>>
(9,
>>>
(9,
from sympy import sieve
sieve.search(25)
10)
sieve.search(23)
9)
5.3.2 Ntheory Functions Reference
sympy.ntheory.generate.prime(nth)
Return the nth prime, with the primes indexed as prime(1) = 2, prime(2) = 3, etc.... The
nth prime is approximately n*log(n) and can never be larger than 2**n.
See Also:
sympy.ntheory.primetest.isprime (page 294) Test if n is prime
primerange (page 278) Generate all primes in a given range
primepi (page 277) Return the number of primes less than or equal to n
References
http://primes.utm.edu/glossary/xpage/BertrandsPostulate.html
Examples
>>> from sympy import prime
>>> prime(10)
29
>>> prime(1)
2
sympy.ntheory.generate.primepi(n)
Return the value of the prime counting function pi(n) = the number of prime numbers
less than or equal to n.
See Also:
5.3. Number Theory
277
SymPy Documentation, Release 0.7.6
sympy.ntheory.primetest.isprime (page 294) Test if n is prime
primerange (page 278) Generate all primes in a given range
prime (page 277) Return the nth prime
Examples
>>> from sympy import primepi
>>> primepi(25)
9
sympy.ntheory.generate.nextprime(n, ith=1)
Return the ith prime greater than n.
i must be an integer.
See Also:
prevprime (page 278) Return the largest prime smaller than n
primerange (page 278) Generate all primes in a given range
Notes
Potential primes are located at 6*j +/- 1. This property is used during searching.
>>> from sympy import nextprime
>>> [(i, nextprime(i)) for i in range(10, 15)]
[(10, 11), (11, 13), (12, 13), (13, 17), (14, 17)]
>>> nextprime(2, ith=2) # the 2nd prime after 2
5
sympy.ntheory.generate.prevprime(n)
Return the largest prime smaller than n.
See Also:
nextprime (page 278) Return the ith prime greater than n
primerange (page 278) Generates all primes in a given range
Notes
Potential primes are located at 6*j +/- 1. This property is used during searching.
>>> from sympy import prevprime
>>> [(i, prevprime(i)) for i in range(10, 15)]
[(10, 7), (11, 7), (12, 11), (13, 11), (14, 13)]
sympy.ntheory.generate.primerange(a, b)
Generate a list of all prime numbers in the range [a, b).
If the range exists in the default sieve, the values will be returned from there; otherwise
values will be returned but will not modify the sieve.
See Also:
nextprime (page 278) Return the ith prime greater than n
278
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
prevprime (page 278) Return the largest prime smaller than n
randprime (page 279) Returns a random prime in a given range
primorial (page 280) Returns the product of primes based on condition
Sieve.primerange (page 276) return range from already computed primes or extend
the sieve to contain the requested range.
Notes
Some famous conjectures about the occurence of primes in a given range are [1]:
Twin primes: though often not, the following will give 2 primes
an innite number of times: primerange(6*n - 1, 6*n + 2)
Legendres: the following always yields at least one prime primerange(n**2,
(n+1)**2+1)
Bertrands (proven): there is always a prime in the range primerange(n,
2*n)
Brocards: there are at least four primes in the range
primerange(prime(n)**2, prime(n+1)**2)
The average gap between primes is log(n) [2]; the gap between primes can be arbitrarily
large since sequences of composite numbers are arbitrarily large, e.g. the numbers in
the sequence n! + 2, n! + 3 ... n! + n are all composite.
References
1.http://en.wikipedia.org/wiki/Prime number
2.http://primes.utm.edu/notes/gaps.html
Examples
>>> from sympy import primerange, sieve
>>> print([i for i in primerange(1, 30)])
[2, 3, 5, 7, 11, 13, 17, 19, 23, 29]
The Sieve method, primerange, is generally faster but it will occupy more memory as
the sieve stores values. The default instance of Sieve, named sieve, can be used:
>>> list(sieve.primerange(1, 30))
[2, 3, 5, 7, 11, 13, 17, 19, 23, 29]
sympy.ntheory.generate.randprime(a, b)
Return a random prime number in the range [a, b).
Bertrands postulate assures that randprime(a, 2*a) will always succeed for a > 1.
See Also:
primerange (page 278) Generate all primes in a given range
5.3. Number Theory
279
SymPy Documentation, Release 0.7.6
References
http://en.wikipedia.org/wiki/Bertrands postulate
Examples
>>> from sympy import randprime, isprime
>>> randprime(1, 30)
13
>>> isprime(randprime(1, 30))
True
sympy.ntheory.generate.primorial(n, nth=True)
Returns the product of the rst n primes (default) or the primes less than or equal to n
(when nth=False).
>>>
>>>
>>>
210
>>>
6
>>>
2
>>>
1
>>>
210
from sympy.ntheory.generate import primorial, randprime, primerange
from sympy import factorint, Mul, primefactors, sqrt
primorial(4) # the first 4 primes are 2, 3, 5, 7
primorial(4, nth=False) # primes <= 4 are 2 and 3
primorial(1)
primorial(1, nth=False)
primorial(sqrt(101), nth=False)
One can argue that the primes are innite since if you take a set of primes and multiply
them together (e.g. the primorial) and then add or subtract 1, the result cannot be
divided by any of the original factors, hence either 1 or more new primes must divide
this product of primes.
In this case, the number itself is a new prime:
>>> factorint(primorial(4) + 1)
{211: 1}
In this case two new primes are the factors:
>>> factorint(primorial(4) - 1)
{11: 1, 19: 1}
Here, some primes smaller and larger than the primes multiplied together are obtained:
>>> p = list(primerange(10, 20))
>>> sorted(set(primefactors(Mul(*p) + 1)).difference(set(p)))
[2, 5, 31, 149]
See Also:
primerange (page 278) Generate all primes in a given range
sympy.ntheory.generate.cycle length(f, x0, nmax=None, values=False)
For a given iterated sequence, return a generator that gives the length of the iterated
cycle (lambda) and the length of terms before the cycle begins (mu); if values is True
280
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
then the terms of the sequence will be returned instead. The sequence is started with
value x0.
Note: more than the rst lambda + mu terms may be returned and this is the cost of
cycle detection with Brents method; there are, however, generally less terms calculated
than would have been calculated if the proper ending point were determined, e.g. by
using Floyds method.
>>> from sympy.ntheory.generate import cycle_length
This will yield successive values of i < func(i):
>>> def iter(func, i):
...
while 1:
...
ii = func(i)
...
yield ii
...
i = ii
...
A function is dened:
>>> func = lambda i: (i**2 + 1) % 51
and given a seed of 4 and the mu and lambda terms calculated:
>>> next(cycle_length(func, 4))
(6, 2)
We can see what is meant by looking at the output:
>>> n = cycle_length(func, 4, values=True)
>>> list(ni for ni in n)
[17, 35, 2, 5, 26, 14, 44, 50, 2, 5, 26, 14]
There are 6 repeating values after the rst 2.
If a sequence is suspected of being longer than you might wish, nmax can be used to exit
early (and mu will be returned as None):
>>> next(cycle_length(func, 4, nmax = 4))
(4, None)
>>> [ni for ni in cycle_length(func, 4, nmax = 4, values=True)]
[17, 35, 2, 5]
Code modied from: http://en.wikipedia.org/wiki/Cycle detection.
sympy.ntheory.factor .smoothness(n)
Return the B-smooth and B-power smooth values of n.
The smoothness of n is the largest prime factor of n; the power- smoothness is the largest
divisor raised to its multiplicity.
>>> from sympy.ntheory.factor_ import smoothness
>>> smoothness(2**7*3**2)
(3, 128)
>>> smoothness(2**4*13)
(13, 16)
>>> smoothness(2)
(2, 2)
5.3. Number Theory
281
SymPy Documentation, Release 0.7.6
See Also:
factorint (page 287), smoothness p (page 282)
sympy.ntheory.factor .smoothness p(n, m=-1, power=0, visual=None)
Return a list of [m, (p, (M, sm(p + m), psm(p + m)))...] where:
1.p**M is the base-p divisor of n
2.sm(p + m) is the smoothness of p + m (m = -1 by default)
3.psm(p + m) is the power smoothness of p + m
The list is sorted according to smoothness (default) or by power smoothness if power=1.
The smoothness of the numbers to the left (m = -1) or right (m = 1) of a factor govern
the results that are obtained from the p +/- 1 type factoring methods.
>>> from sympy.ntheory.factor_ import smoothness_p, factorint
>>> smoothness_p(10431, m=1)
(1, [(3, (2, 2, 4)), (19, (1, 5, 5)), (61, (1, 31, 31))])
>>> smoothness_p(10431)
(-1, [(3, (2, 2, 2)), (19, (1, 3, 9)), (61, (1, 5, 5))])
>>> smoothness_p(10431, power=1)
(-1, [(3, (2, 2, 2)), (61, (1, 5, 5)), (19, (1, 3, 9))])
If visual=True then an annotated string will be returned:
>>> print(smoothness_p(21477639576571, visual=1))
p**i=4410317**1 has p-1 B=1787, B-pow=1787
p**i=4869863**1 has p-1 B=2434931, B-pow=2434931
This string can also be generated directly from a factorization dictionary and vice versa:
>>> factorint(17*9)
{3: 2, 17: 1}
>>> smoothness_p(_)
p**i=3**2 has p-1 B=2, B-pow=2\np**i=17**1 has p-1 B=2, B-pow=16
>>> smoothness_p(_)
{3: 2, 17: 1}
The table of the output logic is:
Visual
Input
dict
str
tuple
n
mul
True
str
str
str
str
str
False
tuple
tuple
tuple
tuple
tuple
other
str
dict
str
tuple
tuple
See Also:
factorint (page 287), smoothness (page 281)
sympy.ntheory.factor .trailing(n)
Count the number of trailing zero digits in the binary representation of n, i.e. determine
the largest power of 2 that divides n.
282
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import trailing
>>> trailing(128)
7
>>> trailing(63)
0
sympy.ntheory.factor .multiplicity(p, n)
Find the greatest integer m such that p**m divides n.
Examples
>>>
>>>
>>>
[0,
>>>
-2
from sympy.ntheory import multiplicity
from sympy.core.numbers import Rational as R
[multiplicity(5, n) for n in [8, 5, 25, 125, 250]]
1, 2, 3, 3]
multiplicity(3, R(1, 9))
sympy.ntheory.factor .perfect power(n, candidates=None, big=True, factor=True)
Return (b, e) such that n == b**e if n is a perfect power; otherwise return False.
By default, the base is recursively decomposed and the exponents collected so the largest
possible e is sought. If big=False then the smallest possible e (thus prime) will be
chosen.
If candidates for exponents are given, they are assumed to be sorted and the rst one
that is larger than the computed maximum will signal failure for the routine.
If factor=True then simultaneous factorization of n is attempted since nding a factor
indicates the only possible root for n. This is True by default since only a few small
factors will be tested in the course of searching for the perfect power.
Examples
>>>
>>>
(2,
>>>
(4,
from sympy import perfect_power
perfect_power(16)
4)
perfect_power(16, big = False)
2)
sympy.ntheory.factor .pollard rho(n,
s=2,
a=1,
retries=5,
seed=1234,
max steps=None, F=None)
Use Pollards rho method to try to extract a nontrivial factor of n. The returned factor
may be a composite number. If no factor is found, None is returned.
The algorithm generates pseudo-random values of x with a generator function, replacing
x with F(x). If F is not supplied then the function x**2 + a is used. The rst value supplied
to F(x) is s. Upon failure (if retries is > 0) a new a and s will be supplied; the a will be
ignored if F was supplied.
The sequence of numbers generated by such functions generally have a a lead-up to some
number and then loop around back to that number and begin to repeat the sequence,
e.g. 1, 2, 3, 4, 5, 3, 4, 5 this leader and loop look a bit like the Greek letter rho, and
thus the name, rho.
5.3. Number Theory
283
SymPy Documentation, Release 0.7.6
For a given function, very dierent leader-loop values can be obtained so it is a good
idea to allow for retries:
>>> from sympy.ntheory.generate import cycle_length
>>> n = 16843009
>>> F = lambda x:(2048*pow(x, 2, n) + 32767) % n
>>> for s in range(5):
...
print(loop length = %4i; leader length = %3i % next(cycle_length(F, s)))
...
loop length = 2489; leader length = 42
loop length =
78; leader length = 120
loop length = 1482; leader length = 99
loop length = 1482; leader length = 285
loop length = 1482; leader length = 100
Here is an explicit example where there is a two element leadup to a sequence of 3
numbers (11, 14, 4) that then repeat:
>>> x=2
>>> for i in range(9):
...
x=(x**2+12)%17
...
print(x)
...
16
13
11
14
4
11
14
4
11
>>> next(cycle_length(lambda x: (x**2+12)%17, 2))
(3, 2)
>>> list(cycle_length(lambda x: (x**2+12)%17, 2, values=True))
[16, 13, 11, 14, 4]
Instead of checking the dierences of all generated values for a gcd with n, only the kth
and 2*kth numbers are checked, e.g. 1st and 2nd, 2nd and 4th, 3rd and 6th until it has
been detected that the loop has been traversed. Loops may be many thousands of steps
long before rho nds a factor or reports failure. If max steps is specied, the iteration
is cancelled with a failure after the specied number of steps.
References
Richard Crandall & Carl Pomerance (2005), Prime Numbers: A Computational Perspective, Springer, 2nd edition, 229-231
Examples
>>>
>>>
>>>
>>>
257
284
from sympy import pollard_rho
n=16843009
F=lambda x:(2048*pow(x,2,n) + 32767) % n
pollard_rho(n, F=F)
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Use the default setting with a bad value of a and no retries:
>>> pollard_rho(n, a=n-2, retries=0)
If retries is > 0 then perhaps the problem will correct itself when new values are generated for a:
>>> pollard_rho(n, a=n-2, retries=1)
257
sympy.ntheory.factor .pollard pm1(n, B=10, a=2, retries=0, seed=1234)
Use Pollards p-1 method to try to extract a nontrivial factor of n. Either a divisor (perhaps composite) or None is returned.
The value of a is the base that is used in the test gcd(a**M - 1, n). The default is 2. If
retries > 0 then if no factor is found after the rst attempt, a new a will be generated
randomly (using the seed) and the process repeated.
Note: the value of M is lcm(1..B) = reduce(ilcm, range(2, B + 1)).
A search is made for factors next to even numbers having a power smoothness less than
B. Choosing a larger B increases the likelihood of nding a larger factor but takes longer.
Whether a factor of n is found or not depends on a and the power smoothness of the even
mumber just less than the factor p (hence the name p - 1).
Although some discussion of what constitutes a good a some descriptions are hard to
interpret. At the modular.math site referenced below it is stated that if gcd(a**M - 1, n)
= N then a**M % q**r is 1 for every prime power divisor of N. But consider the following:
>>> from sympy.ntheory.factor_ import smoothness_p, pollard_pm1
>>> n=257*1009
>>> smoothness_p(n)
(-1, [(257, (1, 2, 256)), (1009, (1, 7, 16))])
So we should (and can) nd a root with B=16:
>>> pollard_pm1(n, B=16, a=3)
1009
If we attempt to increase B to 256 we nd that it doesnt work:
>>> pollard_pm1(n, B=256)
>>>
But if the value of a is changed we nd that only multiples of 257 work, e.g.:
>>> pollard_pm1(n, B=256, a=257)
1009
Checking dierent a values shows that all the ones that didnt work had a gcd value not
equal to n but equal to one of the factors:
>>> from sympy.core.numbers import ilcm, igcd
>>> from sympy import factorint, Pow
>>> M = 1
>>> for i in range(2, 256):
...
M = ilcm(M, i)
...
>>> set([igcd(pow(a, M, n) - 1, n) for a in range(2, 256) if
...
igcd(pow(a, M, n) - 1, n) != n])
set([1009])
5.3. Number Theory
285
SymPy Documentation, Release 0.7.6
But does aM % d for every divisor of n give 1?
>>> aM = pow(255, M, n)
>>> [(d, aM%Pow(*d.args)) for d in factorint(n, visual=True).args]
[(257**1, 1), (1009**1, 1)]
No, only one of them. So perhaps the principle is that a root will be found for a given
value of B provided that:
1.the power smoothness of the p - 1 value next to the root does not exceed B
2.a**M % p != 1 for any of the divisors of n.
By trying more than one a it is possible that one of them will yield a factor.
References
Richard Crandall & Carl Pomerance (2005), Prime Numbers: A Computational Perspective, Springer, 2nd edition, 236-238
http://modular.math.washington.edu/edu/2007/spring/ent/ent-html/node81.html
http://www.cs.toronto.edu/yuvalf/Factorization.pdf
Examples
With the default smoothness bound, this number cant be cracked:
>>> from sympy.ntheory import pollard_pm1, primefactors
>>> pollard_pm1(21477639576571)
Increasing the smoothness bound helps:
>>> pollard_pm1(21477639576571, B=2000)
4410317
Looking at the smoothness of the factors of this number we nd:
>>> from sympy.utilities import flatten
>>> from sympy.ntheory.factor_ import smoothness_p, factorint
>>> print(smoothness_p(21477639576571, visual=1))
p**i=4410317**1 has p-1 B=1787, B-pow=1787
p**i=4869863**1 has p-1 B=2434931, B-pow=2434931
The B and B-pow are the same for the p - 1 factorizations of the divisors because those
factorizations had a very large prime factor:
>>>
{2:
>>>
{2:
factorint(4410317 - 1)
2, 617: 1, 1787: 1}
factorint(4869863-1)
1, 2434931: 1}
Note that until B reaches the B-pow value of 1787, the number is not cracked;
>>> pollard_pm1(21477639576571, B=1786)
>>> pollard_pm1(21477639576571, B=1787)
4410317
286
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
The B value has to do with the factors of the number next to the divisor, not the divisors themselves. A worst case scenario is that the number next to the factor p has a
large prime divisisor or is a perfect power. If these conditions apply then the powersmoothness will be about p/2 or p. The more realistic is that there will be a large prime
factor next to p requiring a B value on the order of p/2. Although primes may have been
searched for up to this level, the p/2 is a factor of p - 1, something that we dont know.
The modular.math reference below states that 15% of numbers in the range of 10**15
to 15**15 + 10**4 are 10**6 power smooth so a B of 10**6 will fail 85% of the time in
that range. From 10**8 to 10**8 + 10**3 the percentages are nearly reversed...but in
that range the simple trial division is quite fast.
sympy.ntheory.factor .factorint(n, limit=None, use trial=True, use rho=True,
use pm1=True, verbose=False, visual=None)
Given a positive integer n, factorint(n) returns a dict containing the prime factors of
n as keys and their respective multiplicities as values. For example:
>>> from sympy.ntheory import factorint
>>> factorint(2000)
# 2000 = (2**4) * (5**3)
{2: 4, 5: 3}
>>> factorint(65537)
# This number is prime
{65537: 1}
For input less than 2, factorint behaves as follows:
factorint(1) returns the empty factorization, {}
factorint(0) returns {0:1}
factorint(-n) adds -1:1 to the factors and then factors n
Partial Factorization:
If limit (> 3) is specied, the search is stopped after performing trial division up to (and
including) the limit (or taking a corresponding number of rho/p-1 steps). This is useful if
one has a large number and only is interested in nding small factors (if any). Note that
setting a limit does not prevent larger factors from being found early; it simply means
that the largest factor may be composite. Since checking for perfect power is relatively
cheap, it is done regardless of the limit setting.
This number, for example, has two small factors and a huge semi-prime factor that cannot
be reduced easily:
>>> from sympy.ntheory import isprime
>>> from sympy.core.compatibility import long
>>> a = 1407633717262338957430697921446883
>>> f = factorint(a, limit=10000)
>>> f == {991: 1, long(202916782076162456022877024859): 1, 7: 1}
True
>>> isprime(max(f))
False
This number has a small factor and a residual perfect power whose base is greater than
the limit:
>>> factorint(3*101**7, limit=5)
{3: 1, 101: 7}
Visual Factorization:
If visual is set to True, then it will return a visual factorization of the integer. For
example:
5.3. Number Theory
287
SymPy Documentation, Release 0.7.6
>>> from sympy import pprint
>>> pprint(factorint(4200, visual=True))
3 1 2 1
2 *3 *5 *7
Note that this is achieved by using the evaluate=False ag in Mul and Pow. If you do other
manipulations with an expression where evaluate=False, it may evaluate. Therefore, you
should use the visual option only for visualization, and use the normal dictionary returned
by visual=False if you want to perform operations on the factors.
You can easily switch between the two forms by sending them back to factorint:
>>> from sympy import Mul, Pow
>>> regular = factorint(1764); regular
{2: 2, 3: 2, 7: 2}
>>> pprint(factorint(regular))
2 2 2
2 *3 *7
>>> visual = factorint(1764, visual=True); pprint(visual)
2 2 2
2 *3 *7
>>> print(factorint(visual))
{2: 2, 3: 2, 7: 2}
If you want to send a number to be factored in a partially factored form you can do so
with a dictionary or unevaluated expression:
>>>
{2:
>>>
{2:
factorint(factorint({4: 2, 12: 3})) # twice to toggle to dict form
10, 3: 3}
factorint(Mul(4, 12, evaluate=False))
4, 3: 1}
The table of the output logic is:
Input
dict
n
mul
True
mul
mul
mul
False
dict
dict
dict
other
mul
dict
dict
See Also:
smoothness (page 281), smoothness p (page 282), divisors (page 289)
Notes
Algorithm:
The function switches between multiple algorithms. Trial division quickly nds small
factors (of the order 1-5 digits), and nds all large factors if given enough time. The
Pollard rho and p-1 algorithms are used to nd large factors ahead of time; they will
often nd factors of the order of 10 digits within a few seconds:
>>> factors = factorint(12345678910111213141516)
>>> for base, exp in sorted(factors.items()):
...
print(%s %s % (base, exp))
...
2 2
288
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
2507191691 1
1231026625769 1
Any of these methods can optionally be disabled with the following boolean parameters:
use trial: Toggle use of trial division
use rho: Toggle use of Pollards rho method
use pm1: Toggle use of Pollards p-1 method
factorint also periodically checks if the remaining part is a prime number or a perfect
power, and in those cases stops.
If verbose is set to True, detailed progress is printed.
sympy.ntheory.factor .primefactors(n, limit=None, verbose=False)
Return a sorted list of ns prime factors, ignoring multiplicity and any composite factor
that remains if the limit was set too low for complete factorization. Unlike factorint(),
primefactors() does not return -1 or 0.
See Also:
divisors (page 289)
Examples
>>>
>>>
[2,
>>>
[5]
from sympy.ntheory import primefactors, factorint, isprime
primefactors(6)
3]
primefactors(-5)
>>> sorted(factorint(123456).items())
[(2, 6), (3, 1), (643, 1)]
>>> primefactors(123456)
[2, 3, 643]
>>> sorted(factorint(10000000001, limit=200).items())
[(101, 1), (99009901, 1)]
>>> isprime(99009901)
False
>>> primefactors(10000000001, limit=300)
[101]
sympy.ntheory.factor .divisors(n, generator=False)
Return all divisors of n sorted from 1..n by default. If generator is True an unordered
generator is returned.
The number of divisors of n can be quite large if there are many prime factors (counting
repeated factors). If only the number of factors is desired use divisor count(n).
See Also:
primefactors (page 289), factorint (page 287), divisor count (page 290)
5.3. Number Theory
289
SymPy Documentation, Release 0.7.6
Examples
>>>
>>>
[1,
>>>
8
from sympy import divisors, divisor_count
divisors(24)
2, 3, 4, 6, 8, 12, 24]
divisor_count(24)
>>> list(divisors(120, generator=True))
[1, 2, 4, 8, 3, 6, 12, 24, 5, 10, 20, 40, 15, 30, 60, 120]
This
is
a
slightly
modied
version
of
Tim
Peters
http://stackoverow.com/questions/1010381/python-factorization
referenced
at:
sympy.ntheory.factor .divisor count(n, modulus=1)
Return the number of divisors of n. If modulus is not 1 then only those that are divisible
by modulus are counted.
See Also:
factorint (page 287), divisors (page 289), totient (page 290)
References
http://www.mayer.dial.pipex.com/maths/formulae.htm
>>> from sympy import divisor_count
>>> divisor_count(6)
4
sympy.ntheory.factor .totient()
Calculate the Euler totient function phi(n)
>>> from sympy.ntheory import totient
>>> totient(1)
1
>>> totient(25)
20
See Also:
divisor count (page 290)
sympy.ntheory.modular.symmetric residue(a, m)
Return the residual mod m such that it is within half of the modulus.
>>> from sympy.ntheory.modular import symmetric_residue
>>> symmetric_residue(1, 6)
1
>>> symmetric_residue(4, 6)
-2
sympy.ntheory.modular.crt(m, v, symmetric=False, check=True)
Chinese Remainder Theorem.
The moduli in m are assumed to be pairwise coprime. The output is then an integer f,
such that f = v i mod m i for each pair out of v and m. If symmetric is False a positive
integer will be returned, else |f| will be less than or equal to the LCM of the moduli, and
thus f may be negative.
290
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
If the moduli are not co-prime the correct result will be returned if/when the test of the
result is found to be incorrect. This result will be None if there is no solution.
The keyword check can be set to False if it is known that the moduli are coprime.
As an example consider a set of residues U = [49, 76, 65] and a set of moduli M =
[99, 97, 95]. Then we have:
>>> from sympy.ntheory.modular import crt, solve_congruence
>>> crt([99, 97, 95], [49, 76, 65])
(639985, 912285)
This is the correct result because:
>>> [639985 % m for m in [99, 97, 95]]
[49, 76, 65]
If the moduli are not co-prime, you may receive an incorrect result if you use
check=False:
>>> crt([12, 6, 17], [3, 4, 2], check=False)
(954, 1224)
>>> [954 % m for m in [12, 6, 17]]
[6, 0, 2]
>>> crt([12, 6, 17], [3, 4, 2]) is None
True
>>> crt([3, 6], [2, 5])
(5, 6)
Note:
the order of gf crts arguments is reversed relative to crt,
solve congruence takes residue, modulus pairs.
and that
Programmers note: rather than checking that all pairs of moduli share no GCD (an
O(n**2) test) and rather than factoring all moduli and seeing that there is no factor in
common, a check that the result gives the indicated residuals is performed an O(n)
operation.
See Also:
solve congruence (page 292)
sympy.polys.galoistools.gf crt (page 1197) low level crt routine used by this routine
sympy.ntheory.modular.crt1(m)
First part of Chinese Remainder Theorem, for multiple application.
Examples
>>> from sympy.ntheory.modular import crt1
>>> crt1([18, 42, 6])
(4536, [252, 108, 756], [0, 2, 0])
sympy.ntheory.modular.crt2(m, v, mm, e, s, symmetric=False)
Second part of Chinese Remainder Theorem, for multiple application.
5.3. Number Theory
291
SymPy Documentation, Release 0.7.6
Examples
>>>
>>>
>>>
(0,
from sympy.ntheory.modular import crt1, crt2
mm, e, s = crt1([18, 42, 6])
crt2([18, 42, 6], [0, 0, 0], mm, e, s)
4536)
sympy.ntheory.modular.solve congruence(*remainder modulus pairs, **hint)
Compute the integer n that has the residual ai when it is divided by mi where the ai
and mi are given as pairs to this function: ((a1, m1), (a2, m2), ...). If there is no solution,
return None. Otherwise return n and its modulus.
The mi values need not be co-prime. If it is known that the moduli are not co-prime then
the hint check can be set to False (default=True) and the check for a quicker solution
via crt() (valid when the moduli are co-prime) will be skipped.
If the hint symmetric is True (default is False), the value of n will be within 1/2 of the
modulus, possibly negative.
See Also:
crt (page 290) high level routine implementing the Chinese Remainder Theorem
Examples
>>> from sympy.ntheory.modular import solve_congruence
What number is 2 mod 3, 3 mod 5 and 2 mod 7?
>>> solve_congruence((2, 3), (3, 5), (2, 7))
(23, 105)
>>> [23 % m for m in [3, 5, 7]]
[2, 3, 2]
If you prefer to work with all remainder in one list and all moduli in another, send the
arguments like this:
>>> solve_congruence(*zip((2, 3, 2), (3, 5, 7)))
(23, 105)
The moduli need not be co-prime; in this case there may or may not be a solution:
>>> solve_congruence((2, 3), (4, 6)) is None
True
>>> solve_congruence((2, 3), (5, 6))
(5, 6)
The symmetric ag will make the result be within 1/2 of the modulus:
>>> solve_congruence((2, 3), (5, 6), symmetric=True)
(-1, 6)
sympy.ntheory.multinomial.binomial coefficients(n)
Return a dictionary containing pairs (k1, k2) : Ck n where Ck n are binomial coecients and
n = k1 + k2. Examples ========
292
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> from sympy.ntheory import binomial_coefficients
>>> binomial_coefficients(9)
{(0, 9): 1, (1, 8): 9, (2, 7): 36, (3, 6): 84,
(4, 5): 126, (5, 4): 126, (6, 3): 84, (7, 2): 36, (8, 1): 9, (9, 0): 1}
See Also:
binomial coefficients list (page 293), multinomial coefficients (page 293)
sympy.ntheory.multinomial.binomial coefficients list(n)
Return a list of binomial coecients as rows of the Pascals triangle.
See Also:
binomial coefficients (page 292), multinomial coefficients (page 293)
Examples
>>> from sympy.ntheory import binomial_coefficients_list
>>> binomial_coefficients_list(9)
[1, 9, 36, 84, 126, 126, 84, 36, 9, 1]
sympy.ntheory.multinomial.multinomial coefficients(m, n)
Return a dictionary containing pairs {(k1,k2,..,km) : C kn} where C kn are multinomial coecients such that n=k1+k2+..+km.
For example:
>>> from sympy.ntheory import multinomial_coefficients
>>> multinomial_coefficients(2, 5) # indirect doctest
{(0, 5): 1, (1, 4): 5, (2, 3): 10, (3, 2): 10, (4, 1): 5, (5, 0): 1}
The algorithm is based on the following result:
(
n
k1 , . . . , k m
)
m (
k1 + 1
n
=
n k1 i=2 k1 + 1, . . . , ki 1, . . .
Code contributed to Sage by Yann Laigle-Chapuy, copied with permission of the author.
See Also:
binomial coefficients list (page 293), binomial coefficients (page 292)
sympy.ntheory.multinomial.multinomial coefficients iterator(m,
tuple=<type
tuple>)
multinomial coecient iterator
n,
This routine has been optimized for m large with respect to n by taking advantage of the
fact that when the monomial tuples t are stripped of zeros, their coecient is the same
as that of the monomial tuples from multinomial coefficients(n, n). Therefore, the
latter coecients are precomputed to save memory and time.
>>> from sympy.ntheory.multinomial import multinomial_coefficients
>>> m53, m33 = multinomial_coefficients(5,3), multinomial_coefficients(3,3)
>>> m53[(0,0,0,1,2)] == m53[(0,0,1,0,2)] == m53[(1,0,2,0,0)] == m33[(0,1,2)]
True
5.3. Number Theory
293
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.ntheory.multinomial import multinomial_coefficients_iterator
>>> it = multinomial_coefficients_iterator(20,3)
>>> next(it)
((3, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0), 1)
sympy.ntheory.partitions .npartitions(n, verbose=False)
Calculate the partition function P(n), i.e. the number of ways that n can be written as a
sum of positive integers.
P(n) is computed using the Hardy-Ramanujan-Rademacher formula, described e.g. at
http://mathworld.wolfram.com/PartitionFunctionP.html
The correctness of this implementation has been tested for 10**n up to n = 8.
Examples
>>> from sympy.ntheory import npartitions
>>> npartitions(25)
1958
sympy.ntheory.primetest.mr(n, bases)
Perform a Miller-Rabin strong pseudoprime test on n using a given list of
bases/witnesses.
References
Richard Crandall & Carl Pomerance (2005), Prime Numbers: A Computational Perspective, Springer, 2nd edition, 135-138
A
list
of
thresholds
and
the
bases
they
require
are
here:
http://en.wikipedia.org/wiki/Miller%E2%80%93Rabin primality test#Deterministic variants of the tes
Examples
>>> from sympy.ntheory.primetest import mr
>>> mr(1373651, [2, 3])
False
>>> mr(479001599, [31, 73])
True
sympy.ntheory.primetest.isprime(n)
Test if n is a prime number (True) or not (False). For n < 10**16 the answer is accurate;
greater n values have a small probability of actually being pseudoprimes.
Negative primes (e.g. -2) are not considered prime.
The function rst looks for trivial factors, and if none is found, performs a safe MillerRabin strong pseudoprime test with bases that are known to prove a number prime.
Finally, a general Miller-Rabin test is done with the rst k bases which will report a
pseudoprime as a prime with an error of about 4**-k. The current value of k is 46 so the
error is about 2 x 10**-28.
See Also:
294
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
sympy.ntheory.generate.primerange (page 278) Generates all primes in a given
range
sympy.ntheory.generate.primepi (page 277) Return the number of primes less than
or equal to n
sympy.ntheory.generate.prime (page 277) Return the nth prime
Examples
>>> from sympy.ntheory import isprime
>>> isprime(13)
True
>>> isprime(15)
False
sympy.ntheory.residue ntheory.n order(a, n)
Returns the order of a modulo n.
The order of a modulo n is the smallest integer k such that a**k leaves a remainder of 1
with n.
Examples
>>> from sympy.ntheory import n_order
>>> n_order(3, 7)
6
>>> n_order(4, 7)
3
sympy.ntheory.residue ntheory.is primitive root(a, p)
Returns True if a is a primitive root of p
a is said to be the primitive root of p if gcd(a, p) == 1 and totient(p) is the smallest
positive number s.t.
a**totient(p) cong 1 mod(p)
Examples
>>> from sympy.ntheory import is_primitive_root, n_order, totient
>>> is_primitive_root(3, 10)
True
>>> is_primitive_root(9, 10)
False
>>> n_order(3, 10) == totient(10)
True
>>> n_order(9, 10) == totient(10)
False
sympy.ntheory.residue ntheory.primitive root(p)
Returns the smallest primitive root or None
Parameters p : positive integer
5.3. Number Theory
295
SymPy Documentation, Release 0.7.6
References
[1] W. Stein Elementary Number Theory (2011), page 44 [2] P. Hackman Elementary
Number Theory (2009), Chapter C
Examples
>>> from sympy.ntheory.residue_ntheory import primitive_root
>>> primitive_root(19)
2
sympy.ntheory.residue ntheory.sqrt mod(a, p, all roots=False)
nd a root of x**2 = a mod p
Parameters a : integer
p : positive integer
all roots : if True the list of roots is returned or None
Notes
If there is no root it is returned None; else the returned root is less or equal to p // 2;
in general is not the smallest one. It is returned p // 2 only if it is the only root.
Use all roots only when it is expected that all the roots t in memory; otherwise use
sqrt mod iter.
Examples
>>>
>>>
21
>>>
[7,
from sympy.ntheory import sqrt_mod
sqrt_mod(11, 43)
sqrt_mod(17, 32, True)
9, 23, 25]
sympy.ntheory.residue ntheory.quadratic residues(p)
Returns the list of quadratic residues.
Examples
>>> from sympy.ntheory.residue_ntheory import quadratic_residues
>>> quadratic_residues(7)
[0, 1, 2, 4]
sympy.ntheory.residue ntheory.nthroot mod(a, n, p, all roots=False)
nd the solutions to x**n = a mod p
Parameters a : integer
n : positive integer
p : positive integer
all roots : if False returns the smallest root, else the list of roots
296
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>>
>>>
8
>>>
[8,
>>>
23
from sympy.ntheory.residue_ntheory import nthroot_mod
nthroot_mod(11, 4, 19)
nthroot_mod(11, 4, 19, True)
11]
nthroot_mod(68, 3, 109)
sympy.ntheory.residue ntheory.is nthpow residue(a, n, m)
Returns True if x**n == a (mod m) has solutions.
References
16.Hackman Elementary Number Theory (2009), page 76
sympy.ntheory.residue ntheory.is quad residue(a, p)
Returns True if a (mod p) is in the set of squares mod p, i.e a % p in set([i**2 % p for i in
range(p)]). If p is an odd prime, an iterative method is used to make the determination:
>>>
>>>
[0,
>>>
[0,
from sympy.ntheory import is_quad_residue
list(set([i**2 % 7 for i in range(7)]))
1, 2, 4]
[j for j in range(7) if is_quad_residue(j, 7)]
1, 2, 4]
See Also:
legendre symbol (page 297), jacobi symbol (page 297)
sympy.ntheory.residue ntheory.legendre symbol(a, p)
Returns 1. 0 if a is multiple of p :
2. 1 if a is a quadratic residue of p :
3. -1 otherwise :
p should be an odd prime by denition :
See Also:
is quad residue (page 297), jacobi symbol (page 297)
Examples
>>>
>>>
[0,
>>>
[0,
from sympy.ntheory import legendre_symbol
[legendre_symbol(i, 7) for i in range(7)]
1, 1, -1, 1, -1, -1]
list(set([i**2 % 7 for i in range(7)]))
1, 2, 4]
sympy.ntheory.residue ntheory.jacobi symbol(m, n)
Returns the product of the legendre symbol(m, p) for all the prime factors, p, of n.
Returns 1. 0 if m cong 0 mod(n) :
2. 1 if x**2 cong m mod(n) has a solution :
5.3. Number Theory
297
SymPy Documentation, Release 0.7.6
3. -1 otherwise :
See Also:
is quad residue (page 297), legendre symbol (page 297)
Examples
>>>
>>>
>>>
-1
>>>
1
from sympy.ntheory import jacobi_symbol, legendre_symbol
from sympy import Mul, S
jacobi_symbol(45, 77)
jacobi_symbol(60, 121)
The relationship between the jacobi symbol and legendre symbol can be demonstrated
as follows:
>>> L = legendre_symbol
>>> S(45).factors()
{3: 2, 5: 1}
>>> jacobi_symbol(7, 45) == L(7, 3)**2 * L(7, 5)**1
True
sympy.ntheory.continued fraction.continued fraction convergents(cf)
Return an iterator over the convergents of a continued fraction (cf).
The parameter should be an iterable returning successive partial quotients of the continued fraction, such as might be returned by continued fraction iterator. In computing
the convergents, the continued fraction need not be strictly in canonical form (all integers, all but the rst positive). Rational and negative elements may be present in the
expansion.
See Also:
continued fraction iterator (page 299)
Examples
>>> from sympy.core import Rational, pi
>>> from sympy import S
>>> from sympy.ntheory.continued_fraction import
continued_fraction_convergents, con
>>> list(continued_fraction_convergents([0, 2, 1, 2]))
[0, 1/2, 1/3, 3/8]
>>> list(continued_fraction_convergents([1, S(1/2), -7, S(1/4)]))
[1, 3, 19/5, 7]
>>> it = continued_fraction_convergents(continued_fraction_iterator(pi))
>>> for n in range(7):
...
print(next(it))
3
22/7
333/106
355/113
103993/33102
298
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
104348/33215
208341/66317
sympy.ntheory.continued fraction.continued fraction iterator(x)
Return continued fraction expansion of x as iterator.
References
[R311] (page 1902)
Examples
>>> from sympy.core import Rational, pi
>>> from sympy.ntheory.continued_fraction import continued_fraction_iterator
>>> list(continued_fraction_iterator(Rational(3, 8)))
[0, 2, 1, 2]
>>> list(continued_fraction_iterator(Rational(-3, 8)))
[-1, 1, 1, 1, 2]
>>> for i, v in enumerate(continued_fraction_iterator(pi)):
...
if i > 7:
...
break
...
print(v)
3
7
15
1
292
1
1
1
sympy.ntheory.continued fraction.continued fraction periodic(p, q, d=0)
Find the periodic continued fraction expansion of a quadratic irrational.
Compute
the continued fraction expansion of a rational or a quadratic irrational number,
p+ d
i.e. q , where p, q and d 0 are integers.
Returns the continued fraction representation (canonical form) as a list of integers, optionally ending (for quadratic irrationals) with repeating block as the last term of this
list.
Parameters p : int
the rational part of the numbers numerator
q : int
the denominator of the number
d : int, optional
the irrational part (discriminator) of the numbers numerator
See Also:
continued fraction iterator (page 299), continued fraction reduce (page 300)
5.3. Number Theory
299
SymPy Documentation, Release 0.7.6
References
[R312] (page 1902), [R313] (page 1902)
Examples
>>> from sympy.ntheory.continued_fraction import continued_fraction_periodic
>>> continued_fraction_periodic(3, 2, 7)
[2, [1, 4, 1, 1]]
Golden ratio has the simplest continued fraction expansion:
>>> continued_fraction_periodic(1, 2, 5)
[[1]]
If the discriminator is zero or a perfect square then the number will be a rational number:
>>>
[1,
>>>
[3,
continued_fraction_periodic(4, 3, 0)
3]
continued_fraction_periodic(4, 3, 49)
1, 2]
sympy.ntheory.continued fraction.continued fraction reduce(cf)
Reduce a continued fraction to a rational or quadratic irrational.
Compute the rational or quadratic irrational number from its terminating or periodic
continued fraction expansion. The continued fraction expansion (cf) should be supplied
as a terminating iterator supplying the terms of the expansion. For terminating continued fractions, this is equivalent to list(continued fraction convergents(cf))[-1],
only a little more ecient. If the expansion has a repeating part, a list of the repeating terms should be returned as the last element from the iterator. This is the format
returned by continued fraction periodic.
For quadratic irrationals, returns the largest solution found, which is generally the one
sought, if the fraction is in canonical form (all terms positive except possibly the rst).
See Also:
continued fraction periodic (page 299)
Examples
>>> from sympy.ntheory.continued_fraction import continued_fraction_reduce
>>> continued_fraction_reduce([1, 2, 3, 4, 5])
225/157
>>> continued_fraction_reduce([-2, 1, 9, 7, 1, 2])
-256/233
>>> continued_fraction_reduce([2, 1, 2, 1, 1, 4, 1, 1, 6, 1, 1, 8]).n(10)
2.718281835
>>> continued_fraction_reduce([1, 4, 2, [3, 1]])
(sqrt(21) + 287)/238
>>> continued_fraction_reduce([[1]])
1/2 + sqrt(5)/2
>>> from sympy.ntheory.continued_fraction import continued_fraction_periodic
>>> continued_fraction_reduce(continued_fraction_periodic(8, 5, 13))
(sqrt(13) + 8)/5
300
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
class sympy.ntheory.mobius
Mbius function maps natural number to {-1, 0, 1}
It is dened as follows:
1. 1 if n = 1.
2. 0 if n has a squared prime factor.
3. (1)k if n is a square-free positive integer with k number of prime factors.
It is an important multiplicative function in number theory and combinatorics. It has
applications in mathematical series, algebraic number theory and also physics (Fermion
operator has very concrete realization with Mbius Function model).
Parameters n : positive integer
References
[R314] (page 1902), [R315] (page 1902)
Examples
>>>
>>>
1
>>>
1
>>>
-1
>>>
0
from sympy.ntheory import mobius
mobius(13*7)
mobius(1)
mobius(13*7*5)
mobius(13**2)
sympy.ntheory.egyptian fraction.egyptian fraction(r, algorithm=Greedy)
Return the list of denominators of an Egyptian fraction expansion [R316] (page 1902) of
the said rational r.
Parameters r : Rational
a positive rational number.
algorithm : { Greedy, Graham Jewett, Takenouchi, Golomb }, optional
Denotes the algorithm to be used (the default is Greedy).
See Also:
sympy.core.numbers.Rational (page 129)
Notes
Currently the following algorithms are supported:
1.Greedy Algorithm
Also called the Fibonacci-Sylvester algorithm [R317] (page 1902). At each step,
extract the largest unit fraction less than the target and replace the target with the
remainder.
It has some distinct properties:
5.3. Number Theory
301
SymPy Documentation, Release 0.7.6
(a)Given p/q in lowest terms, generates an expansion of maximum length p. Even
as the numerators get large, the number of terms is seldom more than a handful.
(b)Uses minimal memory.
(c)The terms can blow up (standard examples of this are 5/121 and 31/311). The
denominator is at most squared at each step (doubly-exponential growth) and
typically exhibits singly-exponential growth.
2.Graham Jewett Algorithm
The algorithm suggested by the result of Graham and Jewett. Note that this has a
tendency to blow up: the length of the resulting expansion is always 2**(x/gcd(x,
y)) - 1. See [R318] (page 1902).
3.Takenouchi Algorithm
The algorithm suggested by Takenouchi (1921). Diers from the Graham-Jewett
algorithm only in the handling of duplicates. See [R318] (page 1902).
4.Golombs Algorithm
A method given by Golumb (1962), using modular arithmetic and inverses. It yields
the same results as a method using continued fractions proposed by Bleicher (1972).
See [R319] (page 1902).
If the given rational is greater than or equal to 1, a greedy algorithm of summing the
harmonic sequence 1/1 + 1/2 + 1/3 + ... is used, taking all the unit fractions of this
sequence until adding one more would be greater than the given number. This list of
denominators is prexed to the result from the requested algorithm used on the remainder. For example, if r is 8/3, using the Greedy algorithm, we get [1, 2, 3, 4, 5, 6, 7, 14,
420], where the beginning of the sequence, [1, 2, 3, 4, 5, 6, 7] is part of the harmonic
sequence summing to 363/140, leaving a remainder of 31/420, which yields [14, 420] by
the Greedy algorithm. The result of egyptian fraction(Rational(8, 3), Golomb) is [1, 2,
3, 4, 5, 6, 7, 14, 574, 2788, 6460, 11590, 33062, 113820], and so on.
References
[R316] (page 1902), [R317] (page 1902), [R318] (page 1902), [R319] (page 1902)
Examples
>>>
>>>
>>>
[3,
>>>
[7,
>>>
[4,
>>>
[3,
>>>
[1,
302
from sympy import Rational
from sympy.ntheory.egyptian_fraction import egyptian_fraction
egyptian_fraction(Rational(3, 7))
11, 231]
egyptian_fraction(Rational(3, 7), Graham Jewett)
8, 9, 56, 57, 72, 3192]
egyptian_fraction(Rational(3, 7), Takenouchi)
7, 28]
egyptian_fraction(Rational(3, 7), Golomb)
15, 35]
egyptian_fraction(Rational(11, 5), Golomb)
2, 3, 4, 9, 234, 1118, 2580]
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
5.4 Basic Cryptography Module
Included in this module are both block ciphers and stream ciphers.
Shift cipher
Ane cipher
Bid ciphers
Vigeneres cipher
substitution ciphers
Hills cipher
RSA
Kid RSA
linear feedback shift registers (a stream cipher)
ElGamal encryption
sympy.crypto.crypto.alphabet of cipher(symbols=ABCDEFGHIJKLMNOPQRSTUVWXYZ)
Returns the list of characters in the string input dening the alphabet.
Notes
First, some basic denitions.
A substitution cipher is a method of encryption by which units (not necessarily characters) of plaintext are replaced with ciphertext according to a regular system. The units
may be characters (ie, words of length 1), words of length 2, and so forth.
A transposition cipher is a method of encryption by which the positions held by units
of plaintext are replaced by a permutation of the plaintext. That is, the order of the
units is changed using a bijective function on the characters positions to perform the
encryption.
A monoalphabetic cipher uses xed substitution over the entire message, whereas a
polyalphabetic cipher uses a number of substitutions at dierent times in the message.
Each of these ciphers require an alphabet for the messages to be constructed from.
Examples
>>> from sympy.crypto.crypto import
>>> alphabet_of_cipher()
[A, B, C, D, E, F, G,
N, O, P, Q, R, S, T,
>>> L = [str(i) for i in range(10)]
[0, 1, 2, 3, 4, 5, 6,
>>> A = .join(L); A
0123456789abc
>>> alphabet_of_cipher(A)
[0, 1, 2, 3, 4, 5, 6,
>>> alphabet_of_cipher()
[A, B, C, D, E, F, G,
N, O, P, Q, R, S, T,
5.4. Basic Cryptography Module
alphabet_of_cipher
H, I, J, K, L, M,
U, V, W, X, Y, Z]
+ [a, b, c]; L
7, 8, 9, a, b, c]
7, 8, 9, a, b, c]
H, I, J, K, L, M,
U, V, W, X, Y, Z]
303
SymPy Documentation, Release 0.7.6
sympy.crypto.crypto.cycle list(k, n)
Returns the cyclic shift of the list range(n) by k.
Examples
>>> from sympy.crypto.crypto import cycle_list, alphabet_of_cipher
>>> L = cycle_list(3,26); L
[3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16,
17, 18, 19, 20, 21, 22, 23, 24, 25, 0, 1, 2]
>>> A = alphabet_of_cipher(); A
[A, B, C, D, E, F, G, H, I, J, K, L, M,
N, O, P, Q, R, S, T, U, V, W, X, Y, Z]
>>> [A[i] for i in L]
[D, E, F, G, H, I, J, K, L, M, N, O, P,
Q, R, S, T, U, V, W, X, Y, Z, A, B, C]
sympy.crypto.crypto.encipher shift(pt, key, symbols=ABCDEFGHIJKLMNOPQRSTUVWXYZ)
Performs shift cipher encryption on plaintext pt, and returns the ciphertext.
Notes
The shift cipher is also called the Caesar cipher, after Julius Caesar, who, according
to Suetonius, used it with a shift of three to protect messages of military signicance.
Caesars nephew Augustus reportedtly used a similar cipher, but with a right shift of 1.
ALGORITHM:
INPUT:
k: an integer from 0 to 25 (the secret key)
m: string of upper-case letters (the plaintext message)
OUTPUT:
c: string of upper-case letters (the ciphertext message)
STEPS:
0. Identify the alphabet A, ..., Z with the integers 0, ..., 25.
1. Compute from the string m a list L1 of corresponding integers.
2. Compute from the list L1 a new list L2, given by adding (k mod 26) to
each element in L1.
3. Compute from the list L2 a string c of corresponding letters.
Examples
>>> from sympy.crypto.crypto import encipher_shift
>>> pt = GONAVYBEATARMY
>>> encipher_shift(pt, 1)
HPOBWZCFBUBSNZ
>>> encipher_shift(pt, 0)
GONAVYBEATARMY
>>> encipher_shift(pt, -1)
FNMZUXADZSZQLX
304
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
sympy.crypto.crypto.encipher affine(pt, key, symbols=ABCDEFGHIJKLMNOPQRSTUVWXYZ)
Performs the ane cipher encryption on plaintext pt, and returns the ciphertext.
Encryption is based on the map x ax + b (mod 26). Decryption is based on the map
x cx + d (mod 26), where c = a1 (mod 26) and d = a1 c (mod 26). (In particular, for the
map to be invertible, we need gcd(a, 26) = 1.)
Notes
This is a straightforward generalization of the shift cipher.
ALGORITHM:
INPUT:
a, b: a pair integers, where gcd(a, 26) = 1 (the secret key)
m: string of upper-case letters (the plaintext message)
OUTPUT:
c: string of upper-case letters (the ciphertext message)
STEPS:
0. Identify the alphabet A, ..., Z with the integers 0, ..., 25.
1. Compute from the string m a list L1 of corresponding integers.
2. Compute from the list L1 a new list L2, given by replacing x by a*x + b
(mod 26), for each element x in L1.
3. Compute from the list L2 a string c of corresponding letters.
Examples
>>> from sympy.crypto.crypto import encipher_affine
>>> pt = GONAVYBEATARMY
>>> encipher_affine(pt, (1, 1))
HPOBWZCFBUBSNZ
>>> encipher_affine(pt, (1, 0))
GONAVYBEATARMY
>>> pt = GONAVYBEATARMY
>>> encipher_affine(pt, (3, 1))
TROBMVENBGBALV
>>> ct = TROBMVENBGBALV
>>> encipher_affine(ct, (9, 17))
GONAVYBEATARMY
sympy.crypto.crypto.encipher substitution(pt,
key,
symbols=ABCDEFGHIJKLMNOPQRSTUVWXYZ)
Performs the substitution cipher encryption on plaintext pt, and returns the ciphertext.
Assumes the pt has only letters taken from symbols. Assumes key is a permutation of
the symbols. This function permutes the letters of the plaintext using the permutation
given in key. The decription uses the inverse permutation. Note that if the permutation
in key is order 2 (eg, a transposition) then the encryption permutation and the decryption
permutation are the same.
5.4. Basic Cryptography Module
305
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.crypto.crypto import alphabet_of_cipher, encipher_substitution
>>> symbols = ABCDEFGHIJKLMNOPQRSTUVWXYZ
>>> A = alphabet_of_cipher(symbols)
>>> key = BACDEFGHIJKLMNOPQRSTUVWXYZ
>>> pt = go navy! beat army!
>>> encipher_substitution(pt, key)
GONBVYAEBTBRMY
>>> ct = GONBVYAEBTBRMY
>>> encipher_substitution(ct, key)
GONAVYBEATARMY
sympy.crypto.crypto.encipher vigenere(pt, key, symbols=ABCDEFGHIJKLMNOPQRSTUVWXYZ)
Performs the Vigenre cipher encryption on plaintext pt, and returns the ciphertext.
Notes
The Vigenre cipher is named after Blaise de Vigenre, a sixteenth century diplomat
and cryptographer, by a historical accident. Vigenre actually invented a dierent and
more complicated cipher. The so-called Vigenre cipher was actually invented by Giovan
Batista Belaso in 1553.
This cipher was used in the 1800s, for example, during the American Civil War. The
Confederacy used a brass cipher disk to implement the Vigenre cipher (now on display
in the NSA Museum in Fort Meade) [R56] (page 1902).
The Vigenre cipher is a generalization of the shift cipher. Whereas the shift cipher
shifts each letter by the same amount (that amount being the key of the shift cipher) the
Vigenre cipher shifts a letter by an amount determined by the key (which is a word or
phrase known only to the sender and receiver).
For example, if the key was a single letter, such as C, then the so-called Vigenere cipher
is actually a shift cipher with a shift of 2 (since C is the 2nd letter of the alphabet, if
you start counting at 0). If the key was a word with two letters, such as CA, then the socalled Vigenre cipher will shift letters in even positions by 2 and letters in odd positions
are left alone (shifted by 0, since A is the 0th letter, if you start counting at 0).
ALGORITHM:
INPUT:
key: a string of upper-case letters (the secret key)
m: string of upper-case letters (the plaintext message)
OUTPUT:
c: string of upper-case letters (the ciphertext message)
STEPS:
0. Identify the alphabet A, ..., Z with the integers 0, ..., 25.
1. Compute from the string key a list L1 of corresponding integers. Let n1
= len(L1).
2. Compute from the string m a list L2 of corresponding integers. Let n2 =
len(L2).
306
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
3. Break L2 up sequencially into sublists of size n1, and one sublist at the
end of size smaller or equal to n1.
4. For each of these sublists L of L2, compute a new list C given by C[i] =
L[i] + L1[i] (mod 26) to the i-th element in the sublist, for each i.
5. Assemble these lists C by concatenation into a new list of length n2.
6. Compute from the new list a string c of corresponding letters.
Once it is known that the key is, say, n characters long, frequency analysis can be applied
to every n-th letter of the ciphertext to determine the plaintext. This method is called
Kasiski examination (although it was rst discovered by Babbage).
The cipher Vigenre actually discovered is an auto-key cipher described as follows.
ALGORITHM:
INPUT:
key: a string of upper-case letters (the secret key)
m: string of upper-case letters (the plaintext message)
OUTPUT:
c: string of upper-case letters (the ciphertext message)
STEPS:
0. Identify the alphabet A, ..., Z with the integers 0, ..., 25.
1. Compute from the string m a list L2 of corresponding integers. Let n2 =
len(L2).
2. Let n1 be the length of the key. Concatenate the string key with the rst
n2 - n1 characters of the plaintext message. Compute from this string
of length n2 a list L1 of corresponding integers. Note n2 = len(L1).
3. Compute a new list C given by C[i] = L1[i] + L2[i] (mod 26), for
each i. Note n2 = len(C).
4. Compute from the new list a string c of corresponding letters.
References
[R56] (page 1902)
Examples
>>> from sympy.crypto.crypto import encipher_vigenere
>>> key = encrypt
>>> pt = meet me on monday
>>> encipher_vigenere(pt, key)
QRGKKTHRZQEBPR
sympy.crypto.crypto.decipher vigenere(ct, key, symbols=ABCDEFGHIJKLMNOPQRSTUVWXYZ)
Decode using the Vigenre cipher.
5.4. Basic Cryptography Module
307
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.crypto.crypto import decipher_vigenere
>>> key = encrypt
>>> ct = QRGK kt HRZQE BPR
>>> decipher_vigenere(ct, key)
MEETMEONMONDAY
sympy.crypto.crypto.encipher hill(pt, key, symbols=ABCDEFGHIJKLMNOPQRSTUVWXYZ)
Performs the Hill cipher encryption on plaintext pt, and returns the ciphertext.
Notes
The Hill cipher [R57] (page 1902), invented by Lester S. Hill in the 1920s [R58]
(page 1902), was the rst polygraphic cipher in which it was practical (though barely)
to operate on more than three symbols at once. The following discussion assumes an
elementary knowledge of matrices.
First, each letter is rst encoded as a number. We assume here that A 0, B 1, ...,
Z 25. We denote the integers {0, 1, ..., 25} by Z26 . Suppose your message m consists
of n capital letters, with no spaces. This may be regarded an n-tuple M of elements of
Z26 . A key in the Hill cipher is a kxk matrix K, all of whose entries are in Z26 , such that
k
k
is one-to-one).
Z26
the matrix K is invertible (ie, that the linear transformation K : Z26
ALGORITHM:
INPUT:
key: a kxk invertible matrix K, all of whose entries are in Z26
m: string of n upper-case letters (the plaintext message) (Note: Sage
assumes that n is a multiple of k.)
OUTPUT:
c: string of upper-case letters (the ciphertext message)
STEPS:
0. Identify the alphabet A, ..., Z with the integers 0, ..., 25.
1. Compute from the string m a list L of corresponding integers. Let n =
len(L).
2. Break the list L up into t = ceiling(n/k) sublists L 1, ..., L t of size k
(where the last list might be padded by 0s to ensure it is size k).
3. Compute new list C 1, ..., C t given by C[i] = K*L i (arithmetic is done
mod 26), for each i.
4. Concatenate these into a list C = C 1 + ... + C t.
5. Compute from C a string c of corresponding letters. This has length k*t.
References
[R57] (page 1902), [R58] (page 1902)
308
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.crypto.crypto import encipher_hill
>>> from sympy import Matrix
>>> pt = meet me on monday
>>> key = Matrix([[1, 2], [3, 5]])
>>> encipher_hill(pt, key)
UEQDUEODOCTCWQ
>>> pt = meet me on tuesday
>>> encipher_hill(pt, key)
UEQDUEODHBOYDJYU
>>> pt = GONAVYBEATARMY
>>> key = Matrix([[1, 0, 1], [0, 1, 1], [2, 2, 3]])
>>> encipher_hill(pt, key)
TBBYTKBEKKRLMYU
sympy.crypto.crypto.decipher hill(ct, key, symbols=ABCDEFGHIJKLMNOPQRSTUVWXYZ)
Deciphering is the same as enciphering but using the inverse of the key matrix.
Examples
>>> from sympy.crypto.crypto import decipher_hill
>>> from sympy import Matrix
>>> ct = UEQDUEODOCTCWQ
>>> key = Matrix([[1, 2], [3, 5]])
>>> decipher_hill(ct, key)
MEETMEONMONDAY
>>> ct = UEQDUEODHBOYDJYU
>>> decipher_hill(ct, key)
MEETMEONTUESDAYA
sympy.crypto.crypto.encipher bifid5(pt, key)
Performs the Bid cipher encryption on plaintext pt, and returns the ciphertext.
This is the version of the Bid cipher that uses the 5 5 Polybius square.
Notes
The Bid cipher was invented around 1901 by Felix Delastelle. It is a fractional substitution cipher, where letters are replaced by pairs of symbols from a smaller alphabet.
The cipher uses a 5 5 square lled with some ordering of the alphabet, except that is
and js are identied (this is a so-called Polybius square; there is a 6 6 analog if you
add back in j and also append onto the usual 26 letter alphabet, the digits 0, 1, ..., 9).
According to Helen Gaines book Cryptanalysis, this type of cipher was used in the eld
by the German Army during World War I.
ALGORITHM: (5x5 case)
INPUT:
pt: plaintext string (no js)
key: short string for key (no repetitions, no js)
OUTPUT:
ciphertext (using Bid5 cipher in all caps, no spaces, no Js)
5.4. Basic Cryptography Module
309
SymPy Documentation, Release 0.7.6
STEPS:
1. Create the 5 5 Polybius square S associated to the k as follows:
(a) starting top left, moving left-to-right, top-to-bottom, place the letters
of the key into a 5x5 matrix,
(b) when nished, add the letters of the alphabet not in the key until the
5x5 square is lled
2. Create a list P of pairs of numbers which are the coordinates in the Polybius square of the letters in pt.
3. Let L1 be the list of all rst coordinates of P (length of L1 = n), let L2 be
the list of all second coordinates of P (so the length of L2 is also n).
4. Let L be the concatenation of L1 and L2 (length L = 2*n), except that
consecutive numbers are paired (L[2*i], L[2*i + 1]). You can regard
L as a list of pairs of length n.
5. Let C be the list of all letters which are of the form S[i, j], for all (i,
j) in L. As a string, this is the ciphertext ct.
Examples
>>> from sympy.crypto.crypto import encipher_bifid5
>>> pt = meet me on monday
>>> key = encrypt
>>> encipher_bifid5(pt, key)
LNLLQNPPNPGADK
>>> pt = meet me on friday
>>> encipher_bifid5(pt, key)
LNLLFGPPNPGRSK
sympy.crypto.crypto.decipher bifid5(ct, key)
Performs the Bid cipher decryption on ciphertext ct, and returns the plaintext.
This is the version of the Bid cipher that uses the 5 5 Polybius square.
INPUT:
ct: ciphertext string (digits okay)
key: short string for key (no repetitions, digits okay)
OUTPUT:
plaintext from Bid5 cipher (all caps, no spaces, no Js)
Examples
>>> from sympy.crypto.crypto import encipher_bifid5, decipher_bifid5
>>> key = encrypt
>>> pt = meet me on monday
>>> encipher_bifid5(pt, key)
LNLLQNPPNPGADK
>>> ct = LNLLQNPPNPGADK
>>> decipher_bifid5(ct, key)
MEETMEONMONDAY
310
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
sympy.crypto.crypto.bifid5 square(key)
5x5 Polybius square.
Produce the Polybius square for the 5 5 Bid cipher.
Examples
>>> from sympy.crypto.crypto import bifid5_square
>>> bifid5_square(gold bug)
Matrix([
[G, O, L, D, B],
[U, A, C, E, F],
[H, I, K, M, N],
[P, Q, R, S, T],
[V, W, X, Y, Z]])
sympy.crypto.crypto.encipher bifid6(pt, key)
Performs the Bid cipher encryption on plaintext pt, and returns the ciphertext.
This is the version of the Bid cipher that uses the 6 6 Polybius square. Assumes
alphabet of symbols is A, ..., Z, 0, ..., 9.
INPUT:
pt: plaintext string (digits okay)
key: short string for key (no repetitions, digits okay)
OUTPUT:
ciphertext from Bid cipher (all caps, no spaces)
Examples
>>> from sympy.crypto.crypto import encipher_bifid6
>>> key = encrypt
>>> pt = meet me on monday at 8am
>>> encipher_bifid6(pt, key)
HNHOKNTA5MEPEGNQZYG
>>> encipher_bifid6(pt, key)
HNHOKNTA5MEPEGNQZYG
sympy.crypto.crypto.decipher bifid6(ct, key)
Performs the Bid cipher decryption on ciphertext ct, and returns the plaintext.
This is the version of the Bid cipher that uses the 6 6 Polybius square. Assumes
alphabet of symbols is A, ..., Z, 0, ..., 9.
INPUT:
ct: ciphertext string (digits okay)
key: short string for key (no repetitions, digits okay)
OUTPUT:
plaintext from Bid cipher (all caps, no spaces)
5.4. Basic Cryptography Module
311
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.crypto.crypto import encipher_bifid6, decipher_bifid6
>>> key = encrypt
>>> pt = meet me on monday at 8am
>>> encipher_bifid6(pt, key)
HNHOKNTA5MEPEGNQZYG
>>> ct = HNHOKNTA5MEPEGNQZYG
>>> decipher_bifid6(ct, key)
MEETMEONMONDAYAT8AM
sympy.crypto.crypto.bifid6 square(key)
6x6 Polybius square.
Produces the Polybius square for the 6 6 Bid cipher. Assumes alphabet of symbols is
A, ..., Z, 0, ..., 9.
Examples
>>> from sympy.crypto.crypto import bifid6_square
>>> key = encrypt
>>> bifid6_square(key)
Matrix([
[E, N, C, R, Y, P],
[T, A, B, D, F, G],
[H, I, J, K, L, M],
[O, Q, S, U, V, W],
[X, Z, 0, 1, 2, 3],
[4, 5, 6, 7, 8, 9]])
sympy.crypto.crypto.encipher bifid7(pt, key)
Performs the Bid cipher encryption on plaintext pt, and returns the ciphertext.
This is the version of the Bid cipher that uses the 7 7 Polybius square. Assumes
alphabet of symbols is A, ..., Z, 0, ..., 22. (Also, assumes you have some way of
distinguishing 22 from 2, 2 juxtaposed together for deciphering...)
INPUT:
pt: plaintext string (digits okay)
key: short string for key (no repetitions, digits okay)
OUTPUT:
ciphertext from Bid7 cipher (all caps, no spaces)
Examples
>>> from sympy.crypto.crypto import encipher_bifid7
>>> key = encrypt
>>> pt = meet me on monday at 8am
>>> encipher_bifid7(pt, key)
JEJJLNAA3ME19YF3J222R
sympy.crypto.crypto.bifid7 square(key)
7x7 Polybius square.
312
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Produce the Polybius square for the 7 7 Bid cipher. Assumes alphabet of symbols is
A, ..., Z, 0, ..., 22. (Also, assumes you have some way of distinguishing 22 from
2, 2 juxtaposed together for deciphering...)
Examples
>>> from sympy.crypto.crypto import bifid7_square
>>> bifid7_square(gold bug)
Matrix([
[ G, O, L, D, B, U, A],
[ C, E, F, H, I, J, K],
[ M, N, P, Q, R, S, T],
[ V, W, X, Y, Z, 0, 1],
[ 2, 3, 4, 5, 6, 7, 8],
[ 9, 10, 11, 12, 13, 14, 15],
[16, 17, 18, 19, 20, 21, 22]])
sympy.crypto.crypto.rsa public key(p, q, e)
The RSA public key is the pair (n, e), where n is a product of two primes and e is relatively
prime (coprime) to the Euler totient (n).
Examples
>>>
>>>
>>>
>>>
15
>>>
7
from sympy.crypto.crypto import rsa_public_key
p, q, e = 3, 5, 7
n, e = rsa_public_key(p, q, e)
n
e
sympy.crypto.crypto.rsa private key(p, q, e)
The RSA private key is the pair (n, d), where n is a product of two primes and d is the
inverse of e (mod (n)).
Examples
>>> from sympy.crypto.crypto import rsa_private_key
>>> p, q, e = 3, 5, 7
>>> rsa_private_key(p, q, e)
(15, 7)
sympy.crypto.crypto.encipher rsa(pt, puk)
In RSA, a message m is encrypted by computing me (mod n), where puk is the public key
(n, e).
Examples
>>> from sympy.crypto.crypto import encipher_rsa, rsa_public_key
>>> p, q, e = 3, 5, 7
>>> puk = rsa_public_key(p, q, e)
5.4. Basic Cryptography Module
313
SymPy Documentation, Release 0.7.6
>>> pt = 12
>>> encipher_rsa(pt, puk)
3
sympy.crypto.crypto.decipher rsa(ct, prk)
In RSA, a ciphertext c is decrypted by computing cd (mod n), where prk is the private key
(n, d).
Examples
>>>
>>>
>>>
>>>
>>>
12
from sympy.crypto.crypto import decipher_rsa, rsa_private_key
p, q, e = 3, 5, 7
prk = rsa_private_key(p, q, e)
ct = 3
decipher_rsa(ct, prk)
sympy.crypto.crypto.kid rsa public key(a, b, A, B)
Kid RSA is a version of RSA useful to teach grade school children since it does not involve
exponentiation.
Alice wants to talk to Bob. Bob generates keys as follows. Key generation:
Select positive integers a, b, A, B at random.
Compute M = ab 1, e = AM + a, d = BM + b, n = (ed 1)/M .
The public key is (n, e). Bob sends these to Alice.
The private key is d, which Bob keeps secret.
Encryption: If m is the plaintext message then the ciphertext is c = me (mod n).
Decryption: If c is the ciphertext message then the plaintext is m = cd (mod n).
Examples
>>> from sympy.crypto.crypto import kid_rsa_public_key
>>> a, b, A, B = 3, 4, 5, 6
>>> kid_rsa_public_key(a, b, A, B)
(369, 58)
sympy.crypto.crypto.kid rsa private key(a, b, A, B)
Compute M = ab 1, e = AM + a, d = BM + b, n = (ed 1)/M . The private key is d, which
Bob keeps secret.
Examples
>>> from sympy.crypto.crypto import kid_rsa_private_key
>>> a, b, A, B = 3, 4, 5, 6
>>> kid_rsa_private_key(a, b, A, B)
(369, 70)
sympy.crypto.crypto.encipher kid rsa(pt, puk)
Here pt is the plaintext and puk is the public key.
314
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>>
>>>
>>>
>>>
>>>
161
from sympy.crypto.crypto import encipher_kid_rsa, kid_rsa_public_key
pt = 200
a, b, A, B = 3, 4, 5, 6
pk = kid_rsa_public_key(a, b, A, B)
encipher_kid_rsa(pt, pk)
sympy.crypto.crypto.decipher kid rsa(ct, prk)
Here pt is the plaintext and prk is the private key.
Examples
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
200
from sympy.crypto.crypto import kid_rsa_public_key, kid_rsa_private_key, decipher_kid_rsa, e
a, b, A, B = 3, 4, 5, 6
d = kid_rsa_private_key(a, b, A, B)
pt = 200
pk = kid_rsa_public_key(a, b, A, B)
prk = kid_rsa_private_key(a, b, A, B)
ct = encipher_kid_rsa(pt, pk)
decipher_kid_rsa(ct, prk)
sympy.crypto.crypto.encode morse(pt)
Encodes a plaintext into popular Morse Code with letters separated by | and words by
||.
References
[R59] (page 1902)
Examples
>>> from sympy.crypto.crypto import encode_morse
>>> pt = ATTACK THE RIGHT FLANK
>>> encode_morse(pt)
.-|-|-|.-|-.-.|-.-||-|....|.||.-.|..|--.|....|-||..-.|.-..|.-|-.|-.-
sympy.crypto.crypto.decode morse(mc)
Decodes a Morse Code with letters separated by | and words by || into plaintext.
References
[R60] (page 1902)
Examples
5.4. Basic Cryptography Module
315
SymPy Documentation, Release 0.7.6
>>> from sympy.crypto.crypto import decode_morse
>>> mc = --|---|...-|.||.|.-|...|-
>>> decode_morse(mc)
MOVE EAST
sympy.crypto.crypto.lfsr sequence(key, ll, n)
This function creates an lfsr sequence.
INPUT:
key: a list of nite eld elements, [c0 , c1 , . . . , ck ].
fill: the list of the initial terms of the lfsr sequence, [x0 , x1 , . . . , xk ].
n: number of terms of the sequence that the function returns.
OUTPUT:
The lfsr sequence dened by xn+1 = ck xn + . . . + c0 xnk , for n k.
Notes
S. Golomb [G60] (page 1902) gives a list of three statistical properties a sequence of
numbers a = {an }
n=1 , an {0, 1}, should display to be considered random. Dene the
autocorrelation of a to be
N
1
(1)an +an+k .
N N
n=1
C(k) = C(k, a) = lim
In the case where a is periodic with period P then this reduces to
C(k) =
P
1
(1)an +an+k .
P n=1
Assume a is periodic with period P .
balance:
P
an
(1) 1.
n=1
low autocorrelation:
{
C(k) =
1,
,
k = 0,
k 6= 0.
(For sequences satisfying these rst two properties, it is known that = 1/P must
hold.)
proportional runs property: In each period, half the runs have length 1, one-fourth
have length 2, etc. Moreover, there are as many runs of 1s as there are of 0s.
References
[G60] (page 1902)
316
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.crypto.crypto import lfsr_sequence
>>> from sympy.polys.domains import FF
>>> F = FF(2)
>>> fill = [F(1), F(1), F(0), F(1)]
>>> key = [F(1), F(0), F(0), F(1)]
>>> lfsr_sequence(key, fill, 10)
[1 mod 2, 1 mod 2, 0 mod 2, 1 mod 2, 0 mod 2, 1 mod 2, 1 mod 2, 0 mod 2, 0 mod 2, 1 mod 2]
sympy.crypto.crypto.lfsr autocorrelation(L, P, k)
This function computes the lsfr autocorrelation function.
INPUT:
L: is a periodic sequence of elements of GF (2). L must have length larger than
P.
P: the period of L
k: an integer (0 < k < p)
OUTPUT:
the k-th value of the autocorrelation of the LFSR L
Examples
>>> from sympy.crypto.crypto import lfsr_sequence, lfsr_autocorrelation
>>> from sympy.polys.domains import FF
>>> F = FF(2)
>>> fill = [F(1), F(1), F(0), F(1)]
>>> key = [F(1), F(0), F(0), F(1)]
>>> s = lfsr_sequence(key, fill, 20)
>>> lfsr_autocorrelation(s, 15, 7)
-1/15
>>> lfsr_autocorrelation(s, 15, 0)
1
sympy.crypto.crypto.lfsr connection polynomial(s)
This function computes the lsfr connection polynomial.
INPUT:
s: a sequence of elements of even length, with entries in a nite eld
OUTPUT:
C(x): the connection polynomial of a minimal LFSR yielding s.
This implements the algorithm in section 3 of J. L. Masseys article [M61] (page 1902).
References
[M61] (page 1902)
5.4. Basic Cryptography Module
317
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.crypto.crypto import lfsr_sequence, lfsr_connection_polynomial
>>> from sympy.polys.domains import FF
>>> F = FF(2)
>>> fill = [F(1), F(1), F(0), F(1)]
>>> key = [F(1), F(0), F(0), F(1)]
>>> s = lfsr_sequence(key, fill, 20)
>>> lfsr_connection_polynomial(s)
x**4 + x + 1
>>> fill = [F(1), F(0), F(0), F(1)]
>>> key = [F(1), F(1), F(0), F(1)]
>>> s = lfsr_sequence(key, fill, 20)
>>> lfsr_connection_polynomial(s)
x**3 + 1
>>> fill = [F(1), F(0), F(1)]
>>> key = [F(1), F(1), F(0)]
>>> s = lfsr_sequence(key, fill, 20)
>>> lfsr_connection_polynomial(s)
x**3 + x**2 + 1
>>> fill = [F(1), F(0), F(1)]
>>> key = [F(1), F(0), F(1)]
>>> s = lfsr_sequence(key, fill, 20)
>>> lfsr_connection_polynomial(s)
x**3 + x + 1
sympy.crypto.crypto.elgamal public key(prk)
Return three number tuple as public key.
Parameters prk : Tuple (p, r, e) generated by elgamal private key
Returns (p, r, e = r**d mod p) : d is a random number in private key.
Examples
>>> from sympy.crypto.crypto import elgamal_public_key
>>> elgamal_public_key((1031, 14, 636))
(1031, 14, 212)
sympy.crypto.crypto.elgamal private key(digit=10)
Return three number tuple as private key.
Elgamal encryption is based on the mathmatical problem called the Discrete Logarithm
Problem (DLP). For example,
ab c (mod p)
In general, if a and b are known, c is easily calculated. If b is unknown, it is hard to use
a and c to get b.
Parameters digit : Key length in binary
Returns (p, r, d) : p = prime number, r = primitive root, d = random number
Examples
318
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> from sympy.crypto.crypto import elgamal_private_key
>>> from sympy.ntheory import is_primitive_root, isprime
>>> a, b, _ = elgamal_private_key()
>>> isprime(a)
True
>>> is_primitive_root(b, a)
True
sympy.crypto.crypto.encipher elgamal(m, puk)
Encrypt message with public key
m is plain text message in int. puk is public key (p, r, e). In order to encrypt a message,
a random number a between 2 and p, encryped message is c1 and c2
c1 ra (mod p)
c2 mea (mod p)
Parameters m : int of encoded message
puk : public key
Returns (c1, c2) : Encipher into two number
Examples
>>> from sympy.crypto.crypto import encipher_elgamal
>>> encipher_elgamal(100, (1031, 14, 212))
(835, 271)
sympy.crypto.crypto.decipher elgamal(ct, prk)
Decrypt message with private key
ct = (c1 , c2 )
prk = (p, r, d)
According to extended Eucliden theorem, ucd1 + pn = 1
u 1/c1 d (mod p)
uc2
1
c2
cd
1
1
mea
r ad
1
c
r ad 2
1
mrda
r ad
(mod p)
m (mod p)
Examples
>>> from sympy.crypto.crypto import decipher_elgamal
>>> decipher_elgamal((835, 271), (1031, 14, 636))
100
5.5 Concrete Mathematics
5.5.1 Hypergeometric terms
The center stage, in recurrence solving and summations, play hypergeometric terms. Formally these are sequences annihilated by rst order linear recurrence operators. In simple
5.5. Concrete Mathematics
319
SymPy Documentation, Release 0.7.6
words if we are given term a(n) then it is hypergeometric if its consecutive term ratio is a
rational function in n.
To check if a sequence is of this type you can use the is hypergeometric method which is
available in Basic class. Here is simple example involving a polynomial:
>>> from sympy import *
>>> n, k = symbols(n,k)
>>> (n**2 + 1).is_hypergeometric(n)
True
Of course polynomials are hypergeometric but are there any more complicated sequences of
this type? Here are some trivial examples:
>>> factorial(n).is_hypergeometric(n)
True
>>> binomial(n, k).is_hypergeometric(n)
True
>>> rf(n, k).is_hypergeometric(n)
True
>>> ff(n, k).is_hypergeometric(n)
True
>>> gamma(n).is_hypergeometric(n)
True
>>> (2**n).is_hypergeometric(n)
True
We see that all species used in summations and other parts of concrete mathematics are
hypergeometric. Note also that binomial coecients and both rising and falling factorials
are hypergeometric in both their arguments:
>>> binomial(n, k).is_hypergeometric(k)
True
>>> rf(n, k).is_hypergeometric(k)
True
>>> ff(n, k).is_hypergeometric(k)
True
To say more, all previously shown examples are valid for integer linear arguments:
>>> factorial(2*n).is_hypergeometric(n)
True
>>> binomial(3*n+1, k).is_hypergeometric(n)
True
>>> rf(n+1, k-1).is_hypergeometric(n)
True
>>> ff(n-1, k+1).is_hypergeometric(n)
True
>>> gamma(5*n).is_hypergeometric(n)
True
>>> (2**(n-7)).is_hypergeometric(n)
True
However nonlinear arguments make those sequences fail to be hypergeometric:
>>> factorial(n**2).is_hypergeometric(n)
False
>>> (2**(n**3 + 1)).is_hypergeometric(n)
False
320
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
If not only the knowledge of being hypergeometric or not is needed, you can use hypersimp()
function. It will try to simplify combinatorial expression and if the term given is hypergeometric it will return a quotient of polynomials of minimal degree. Otherwise is will return
N one to say that sequence is not hypergeometric:
>>> hypersimp(factorial(2*n), n)
2*(n + 1)*(2*n + 1)
>>> hypersimp(factorial(n**2), n)
5.5.2 Concrete Class Reference
class sympy.concrete.summations.Sum
Represents unevaluated summation.
Sum represents a nite or innite series, with the rst argument being the general form
of terms in the series, and the second argument being (dummy variable, start, end),
with dummy variable taking all integer values from start through end. In accordance
with long-standing mathematical convention, the end term is included in the summation.
See Also:
summation, Product, product
References
[R15] (page 1902), [R16] (page 1902), [R17] (page 1902)
Examples
>>> from sympy.abc import i, k, m, n, x
>>> from sympy import Sum, factorial, oo
>>> Sum(k,(k,1,m))
Sum(k, (k, 1, m))
>>> Sum(k,(k,1,m)).doit()
m**2/2 + m/2
>>> Sum(k**2,(k,1,m))
Sum(k**2, (k, 1, m))
>>> Sum(k**2,(k,1,m)).doit()
m**3/3 + m**2/2 + m/6
>>> Sum(x**k,(k,0,oo))
Sum(x**k, (k, 0, oo))
>>> Sum(x**k,(k,0,oo)).doit()
Piecewise((1/(-x + 1), Abs(x) < 1), (Sum(x**k, (k, 0, oo)), True))
>>> Sum(x**k/factorial(k),(k,0,oo)).doit()
exp(x)
An example showing that the symbolic result of a summation is still valid for seemingly
nonsensical values of the limits. Then the Karr convention allows us to give a perfectly
valid interpretation to those sums by interchanging the limits according to the above
rules:
>>> S = Sum(i, (i,1,n)).doit()
>>> S
n**2/2 + n/2
5.5. Concrete Mathematics
321
SymPy Documentation, Release 0.7.6
>>> S.subs(n, -4)
6
>>> Sum(i, (i, 1, -4)).doit()
6
>>> Sum(-i, (i, -3, 0)).doit()
6
An explicit example of the Karr summation convention:
>>> S1 = Sum(i**2, (i, m, m+n-1)).doit()
>>> S1
m**2*n + m*n**2 - m*n + n**3/3 - n**2/2 + n/6
>>> S2 = Sum(i**2, (i, m+n, m-1)).doit()
>>> S2
-m**2*n - m*n**2 + m*n - n**3/3 + n**2/2 - n/6
>>> S1 + S2
0
>>> S3 = Sum(i, (i, m, m-1)).doit()
>>> S3
0
Finite Sums
For nite sums (and sums with symbolic limits assumed to be nite) we follow the summation convention described by Karr [1], especially denition 3 of section 1.4. The sum:
f (i)
mi<n
has the obvious meaning for m < n, namely:
f (i) = f (m) + f (m + 1) + . . . + f (n 2) + f (n 1)
mi<n
with the upper limit value f (n) excluded. The sum over an empty set is zero if and only
if m = n:
f (i) = 0 for m = n
mi<n
Finally, for all other sums over empty sets we assume the following denition:
f (i) for m > n
f (i) =
mi<n
ni<m
It is important to note that Karr denes all sums with the upper limit being exclusive.
This is in contrast to the usual mathematical notation, but does not aect the summation
convention. Indeed we have:
mi<n
f (i) =
n1
f (i)
i=m
where the dierence in notation is intentional to emphasize the meaning, with limits
typeset on the top being inclusive.
322
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
euler maclaurin(m=0, n=0, eps=0, eval integral=True)
Return an Euler-Maclaurin approximation of self, where m is the number of leading
terms to sum directly and n is the number of terms in the tail.
With m = n = 0, this is simply the corresponding integral plus a rst-order endpoint
correction.
Returns (s, e) where s is the Euler-Maclaurin approximation and e is the estimated
error (taken to be the magnitude of the rst omitted term in the tail):
>>> from sympy.abc import k, a, b
>>> from sympy import Sum
>>> Sum(1/k, (k, 2, 5)).doit().evalf()
1.28333333333333
>>> s, e = Sum(1/k, (k, 2, 5)).euler_maclaurin()
>>> s
-log(2) + 7/20 + log(5)
>>> from sympy import sstr
>>> print(sstr((s.evalf(), e.evalf()), full_prec=True))
(1.26629073187415, 0.0175000000000000)
The endpoints may be symbolic:
>>> s, e = Sum(1/k, (k, a, b)).euler_maclaurin()
>>> s
-log(a) + log(b) + 1/(2*b) + 1/(2*a)
>>> e
Abs(1/(12*b**2) - 1/(12*a**2))
If the function is a polynomial of degree at most 2n+1, the Euler-Maclaurin formula
becomes exact (and e = 0 is returned):
>>> Sum(k, (k, 2, b)).euler_maclaurin()
(b**2/2 + b/2 - 1, 0)
>>> Sum(k, (k, 2, b)).doit()
b**2/2 + b/2 - 1
With a nonzero eps specied, the summation is ended as soon as the remainder term
is less than the epsilon.
reverse order(*indices)
Reverse the order of a limit in a Sum.
See Also:
index, reorder limit, reorder
References
[R18] (page 1902)
Examples
>>> from sympy import Sum
>>> from sympy.abc import x, y, a, b, c, d
5.5. Concrete Mathematics
323
SymPy Documentation, Release 0.7.6
>>> Sum(x, (x, 0, 3)).reverse_order(x)
Sum(-x, (x, 4, -1))
>>> Sum(x*y, (x, 1, 5), (y, 0, 6)).reverse_order(x, y)
Sum(x*y, (x, 6, 0), (y, 7, -1))
>>> Sum(x, (x, a, b)).reverse_order(x)
Sum(-x, (x, b + 1, a - 1))
>>> Sum(x, (x, a, b)).reverse_order(0)
Sum(-x, (x, b + 1, a - 1))
While one should prefer variable names when specifying which limits to reverse, the
index counting notation comes in handy in case there are several symbols with the
same name.
>>> S = Sum(x**2, (x, a, b), (x, c, d))
>>> S
Sum(x**2, (x, a, b), (x, c, d))
>>> S0 = S.reverse_order( 0)
>>> S0
Sum(-x**2, (x, b + 1, a - 1), (x, c, d))
>>> S1 = S0.reverse_order( 1)
>>> S1
Sum(x**2, (x, b + 1, a - 1), (x, d + 1, c - 1))
Of course we can mix both notations:
>>> Sum(x*y,
Sum(x*y, (x,
>>> Sum(x*y,
Sum(x*y, (x,
(x,
b +
(x,
b +
a,
1,
a,
1,
b),
a b),
a -
(y,
1),
(y,
1),
2, 5)).reverse_order( x, 1)
(y, 6, 1))
2, 5)).reverse_order( y, x)
(y, 6, 1))
Usage
reverse order(self, *indices) reverses some limits in the expression self which
can be either a Sum or a Product. The selectors in the argument indices specify
some indices whose limits get reversed. These selectors are either variable names
or numerical indices counted starting from the inner-most limit tuple.
class sympy.concrete.products.Product
Represents unevaluated products.
Product represents a nite or innite product, with the rst argument being the general
form of terms in the series, and the second argument being (dummy variable, start,
end), with dummy variable taking all integer values from start through end. In accordance with long-standing mathematical convention, the end term is included in the
product.
See Also:
Sum, summation, product
References
[R19] (page 1903), [R20] (page 1903), [R21] (page 1903)
324
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.abc import a, b, i, k, m, n, x
>>> from sympy import Product, factorial, oo
>>> Product(k,(k,1,m))
Product(k, (k, 1, m))
>>> Product(k,(k,1,m)).doit()
factorial(m)
>>> Product(k**2,(k,1,m))
Product(k**2, (k, 1, m))
>>> Product(k**2,(k,1,m)).doit()
(factorial(m))**2
Wallis product for pi:
>>> W = Product(2*i/(2*i-1) * 2*i/(2*i+1), (i, 1, oo))
>>> W
Product(4*i**2/((2*i - 1)*(2*i + 1)), (i, 1, oo))
Direct computation currently fails:
>>> W.doit()
nan
But we can approach the innite product by a limit of nite products:
>>> from sympy import limit
>>> W2 = Product(2*i/(2*i-1)*2*i/(2*i+1), (i, 1, n))
>>> W2
Product(4*i**2/((2*i - 1)*(2*i + 1)), (i, 1, n))
>>> W2e = W2.doit()
>>> W2e
2**(-2*n)*4**n*(factorial(n))**2/(RisingFactorial(1/2, n)*RisingFactorial(3/2, n))
>>> limit(W2e, n, oo)
pi/2
By the same formula we can compute sin(pi/2):
>>> from sympy import pi, gamma, simplify
>>> P = pi * x * Product(1 - x**2/k**2,(k,1,n))
>>> P = P.subs(x, pi/2)
>>> P
pi**2*Product(1 - pi**2/(4*k**2), (k, 1, n))/2
>>> Pe = P.doit()
>>> Pe
pi**2*RisingFactorial(1 + pi/2, n)*RisingFactorial(-pi/2 + 1, n)/(2*(factorial(n))**2)
>>> Pe = Pe.rewrite(gamma)
>>> Pe
pi**2*gamma(n + 1 + pi/2)*gamma(n - pi/2 + 1)/(2*gamma(1 + pi/2)*gamma(-pi/2 + 1)*gamma(n + 1)**
>>> Pe = simplify(Pe)
>>> Pe
sin(pi**2/2)*gamma(n + 1 + pi/2)*gamma(n - pi/2 + 1)/gamma(n + 1)**2
>>> limit(Pe, n, oo)
sin(pi**2/2)
Products with the lower limit being larger than the upper one:
>>> Product(1/i, (i, 6, 1)).doit()
120
5.5. Concrete Mathematics
325
SymPy Documentation, Release 0.7.6
>>> Product(i, (i, 2, 5)).doit()
120
The empty product:
>>> Product(i, (i, n, n-1)).doit()
1
An example showing that the symbolic result of a product is still valid for seemingly
nonsensical values of the limits. Then the Karr convention allows us to give a perfectly
valid interpretation to those products by interchanging the limits according to the above
rules:
>>> P = Product(2, (i, 10, n)).doit()
>>> P
2**(n - 9)
>>> P.subs(n, 5)
1/16
>>> Product(2, (i, 10, 5)).doit()
1/16
>>> 1/Product(2, (i, 6, 9)).doit()
1/16
An explicit example of the Karr summation convention applied to products:
>>> P1 = Product(x, (i, a, b)).doit()
>>> P1
x**(-a + b + 1)
>>> P2 = Product(x, (i, b+1, a-1)).doit()
>>> P2
x**(a - b - 1)
>>> simplify(P1 * P2)
1
And another one:
>>> P1 = Product(i, (i, b, a)).doit()
>>> P1
RisingFactorial(b, a - b + 1)
>>> P2 = Product(i, (i, a+1, b-1)).doit()
>>> P2
RisingFactorial(a + 1, -a + b - 1)
>>> P1 * P2
RisingFactorial(b, a - b + 1)*RisingFactorial(a + 1, -a + b - 1)
>>> simplify(P1 * P2)
1
Finite Products
For nite products (and products with symbolic limits assumed to be nite) we follow
the analogue of the summation convention described by Karr [1], especially denition 3
of section 1.4. The product:
f (i)
mi<n
326
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
has the obvious meaning for m < n, namely:
f (i) = f (m)f (m + 1) . . . f (n 2)f (n 1)
mi<n
with the upper limit value f (n) excluded. The product over an empty set is one if and
only if m = n:
f (i) = 1 for m = n
mi<n
Finally, for all other products over empty sets we assume the following denition:
f (i) =
ni<m
mi<n
f (i)
for
m>n
It is important to note that above we dene all products with the upper limit being exclusive. This is in contrast to the usual mathematical notation, but does not aect the
product convention. Indeed we have:
f (i) =
mi<n
n1
f (i)
i=m
where the dierence in notation is intentional to emphasize the meaning, with limits
typeset on the top being inclusive.
reverse order(expr, *indices)
Reverse the order of a limit in a Product.
See Also:
index, reorder limit, reorder
References
[R22] (page 1903)
Examples
>>> from sympy import Product, simplify, RisingFactorial, gamma, Sum
>>> from sympy.abc import x, y, a, b, c, d
>>> P = Product(x, (x, a, b))
>>> Pr = P.reverse_order(x)
>>> Pr
Product(1/x, (x, b + 1, a - 1))
>>> Pr = Pr.doit()
>>> Pr
1/RisingFactorial(b + 1, a - b - 1)
>>> simplify(Pr)
gamma(b + 1)/gamma(a)
>>> P = P.doit()
>>> P
RisingFactorial(a, -a + b + 1)
>>> simplify(P)
gamma(b + 1)/gamma(a)
5.5. Concrete Mathematics
327
SymPy Documentation, Release 0.7.6
While one should prefer variable names when specifying which limits to reverse, the
index counting notation comes in handy in case there are several symbols with the
same name.
>>> S = Sum(x*y, (x, a, b), (y, c, d))
>>> S
Sum(x*y, (x, a, b), (y, c, d))
>>> S0 = S.reverse_order( 0)
>>> S0
Sum(-x*y, (x, b + 1, a - 1), (y, c, d))
>>> S1 = S0.reverse_order( 1)
>>> S1
Sum(x*y, (x, b + 1, a - 1), (y, d + 1, c - 1))
Of course we can mix both notations:
>>> Sum(x*y,
Sum(x*y, (x,
>>> Sum(x*y,
Sum(x*y, (x,
(x,
b +
(x,
b +
a,
1,
a,
1,
b),
a b),
a -
(y,
1),
(y,
1),
2, 5)).reverse_order( x, 1)
(y, 6, 1))
2, 5)).reverse_order( y, x)
(y, 6, 1))
Usage
reverse order(expr, *indices) reverses some limits in the expression expr which
can be either a Sum or a Product. The selectors in the argument indices specify
some indices whose limits get reversed. These selectors are either variable names
or numerical indices counted starting from the inner-most limit tuple.
5.5.3 Concrete Functions Reference
sympy.concrete.summations.summation(f, *symbols, **kwargs)
Compute the summation of f with respect to symbols.
The notation for symbols is similar to the notation used in Integral. summation(f, (i, a,
b)) computes the sum of f with respect to i from a to b, i.e.,
b
summation(f, (i, a, b)) =
)
f
/ ,
i = a
If it cannot compute the sum, it returns an unevaluated Sum object. Repeated sums can
be computed by introducing additional symbols tuples:
>>> from sympy import summation, oo, symbols, log
>>> i, n, m = symbols(i n m, integer=True)
>>> summation(2*i - 1, (i, 1, n))
n**2
>>> summation(1/2**i, (i, 0, oo))
2
>>> summation(1/log(n)**n, (n, 2, oo))
Sum(log(n)**(-n), (n, 2, oo))
328
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> summation(i, (i, 0, n), (n, 0, m))
m**3/6 + m**2/2 + m/3
>>> from sympy.abc import x
>>> from sympy import factorial
>>> summation(x**n/factorial(n), (n, 0, oo))
exp(x)
See Also:
Sum, Product, product
sympy.concrete.products.product(*args, **kwargs)
Compute the product.
The notation for symbols is similiar to the notation used in Sum or Integral. product(f,
(i, a, b)) computes the product of f with respect to i from a to b, i.e.,
b
product(f(n), (i, a, b)) = |
| f(n)
|
|
i = a
If it cannot compute the product, it returns an unevaluated Product object. Repeated
products can be computed by introducing additional symbols tuples:
>>> from sympy import product, symbols
>>> i, n, m, k = symbols(i n m k, integer=True)
>>> product(i, (i, 1,
factorial(k)
>>> product(m, (i, 1,
m**k
>>> product(i, (i, 1,
Product(factorial(k),
k))
k))
k), (k, 1, n))
(k, 1, n))
sympy.concrete.gosper.gosper normal(f, g, n, polys=True)
Compute the Gospers normal form of f and g.
Given relatively prime univariate polynomials f and g, rewrite their quotient to a normal
form dened as follows:
f (n)
A(n)C(n + 1)
=Z
g(n)
B(n)C(n)
where Z is an arbitrary constant and A, B, C are monic polynomials in n with the following
properties:
1.gcd(A(n), B(n + h)) = 1h N
2.gcd(B(n), C(n + 1)) = 1
3.gcd(A(n), C(n)) = 1
This normal form, or rational factorization in other words, is a crucial step in Gospers
algorithm and in solving of dierence equations. It can be also used to decide if two
hypergeometric terms are similar or not.
This procedure will return a tuple containing elements of this factorization in the form
(Z*A, B, C).
5.5. Concrete Mathematics
329
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.concrete.gosper import gosper_normal
>>> from sympy.abc import n
>>> gosper_normal(4*n+5, 2*(4*n+1)*(2*n+3), n, polys=False)
(1/4, n + 3/2, n + 1/4)
sympy.concrete.gosper.gosper term(f, n)
Compute Gospers hypergeometric term for f.
Suppose f is a hypergeometric term such that:
sn =
n1
fk
k=0
and fk doesnt depend on n. Returns a hypergeometric term gn such that gn+1 gn = fn .
Examples
>>> from sympy.concrete.gosper import gosper_term
>>> from sympy.functions import factorial
>>> from sympy.abc import n
>>> gosper_term((4*n + 1)*factorial(n)/factorial(2*n + 1), n)
(-n - 1/2)/(n + 1/4)
sympy.concrete.gosper.gosper sum(f, k)
Gospers hypergeometric summation algorithm.
Given a hypergeometric term f such that:
sn =
n1
fk
k=0
and f (n) doesnt depend on n, returns gn g(0) where gn+1 gn = fn , or None if sn can not
be expressed in closed form as a sum of hypergeometric terms.
References
[R23] (page 1903)
Examples
>>> from sympy.concrete.gosper import gosper_sum
>>> from sympy.functions import factorial
>>> from sympy.abc import i, n, k
330
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> f = (4*k + 1)*factorial(k)/factorial(2*k + 1)
>>> gosper_sum(f, (k, 0, n))
(-factorial(n) + 2*factorial(2*n + 1))/factorial(2*n + 1)
>>> _.subs(n, 2) == sum(f.subs(k, i) for i in [0, 1, 2])
True
>>> gosper_sum(f, (k, 3, n))
(-60*factorial(n) + factorial(2*n + 1))/(60*factorial(2*n + 1))
>>> _.subs(n, 5) == sum(f.subs(k, i) for i in [3, 4, 5])
True
5.6 Numerical evaluation
5.6.1 Basics
Exact SymPy expressions can be converted to oating-point approximations (decimal numbers) using either the .evalf() method or the N() function. N(expr, <args>) is equivalent
to sympify(expr).evalf(<args>).
>>> from sympy import *
>>> N(sqrt(2)*pi)
4.44288293815837
>>> (sqrt(2)*pi).evalf()
4.44288293815837
By default, numerical evaluation is performed to an accuracy of 15 decimal digits. You can
optionally pass a desired accuracy (which should be a positive integer) as an argument to
evalf or N:
>>> N(sqrt(2)*pi, 5)
4.4429
>>> N(sqrt(2)*pi, 50)
4.4428829381583662470158809900606936986146216893757
Complex numbers are supported:
>>> N(1/(pi + I), 20)
0.28902548222223624241 - 0.091999668350375232456*I
If the expression contains symbols or for some other reason cannot be evaluated numerically,
calling .evalf() or N() returns the original expression, or in some cases a partially evaluated expression. For example, when the expression is a polynomial in expanded form, the
coecients are evaluated:
>>> x = Symbol(x)
>>> (pi*x**2 + x/3).evalf()
3.14159265358979*x**2 + 0.333333333333333*x
You can also use the standard Python functions float(), complex() to convert SymPy expressions to regular Python numbers:
>>> float(pi)
3.1415926535...
>>> complex(pi+E*I)
(3.1415926535...+2.7182818284...j)
5.6. Numerical evaluation
331
SymPy Documentation, Release 0.7.6
If these functions are used, failure to evaluate the expression to an explicit number (for example if the expression contains symbols) will raise an exception.
There is essentially no upper precision limit. The following command, for example, computes
the rst 100,000 digits of /e:
>>> N(pi/E, 100000)
...
This shows digits 999,951 through 1,000,000 of pi:
>>> str(N(pi, 10**6))[-50:]
95678796130331164628399634646042209010610577945815
High-precision calculations can be slow. It is recommended (but entirely optional) to install
gmpy (http://code.google.com/p/gmpy/), which will signicantly speed up computations such
as the one above.
5.6.2 Floating-point numbers
Floating-point numbers in SymPy are instances of the class Float. A Float can be created
with a custom precision as second argument:
>>> Float(0.1)
0.100000000000000
>>> Float(0.1, 10)
0.1000000000
>>> Float(0.125, 30)
0.125000000000000000000000000000
>>> Float(0.1, 30)
0.100000000000000005551115123126
As the last example shows, some Python oats are only accurate to about 15 digits as inputs,
while others (those that have a denominator that is a power of 2, like .125 = 1/4) are exact. To
create a Float from a high-precision decimal number, it is better to pass a string, Rational,
or evalf a Rational:
>>> Float(0.1, 30)
0.100000000000000000000000000000
>>> Float(Rational(1, 10), 30)
0.100000000000000000000000000000
>>> Rational(1, 10).evalf(30)
0.100000000000000000000000000000
The precision of a number determines 1) the precision to use when performing arithmetic
with the number, and 2) the number of digits to display when printing the number. When two
numbers with dierent precision are used together in an arithmetic operation, the higher of
the precisions is used for the result. The product of 0.1 +/- 0.001 and 3.1415 +/- 0.0001 has
an uncertainty of about 0.003 and yet 5 digits of precision are shown.
>>> Float(0.1, 3)*Float(3.1415, 5)
0.31417
So the displayed precision should not be used as a model of error propagation or signicance
arithmetic; rather, this scheme is employed to ensure stability of numerical algorithms.
N and evalf can be used to change the precision of existing oating-point numbers:
332
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> N(3.5)
3.50000000000000
>>> N(3.5, 5)
3.5000
>>> N(3.5, 30)
3.50000000000000000000000000000
5.6.3 Accuracy and error handling
When the input to N or evalf is a complicated expression, numerical error propagation becomes a concern. As an example, consider
the 100th Fibonacci number and the excellent
(but not exact) approximation 100 / 5 where is the golden ratio. With ordinary oatingpoint arithmetic, subtracting these numbers from each other erroneously results in a complete cancellation:
>>> a, b = GoldenRatio**1000/sqrt(5), fibonacci(1000)
>>> float(a)
4.34665576869e+208
>>> float(b)
4.34665576869e+208
>>> float(a) - float(b)
0.0
N and evalf keep track of errors and automatically increase the precision used internally in
order to obtain a correct result:
>>> N(fibonacci(100) - GoldenRatio**100/sqrt(5))
-5.64613129282185e-22
Unfortunately, numerical evaluation cannot tell an expression that is exactly zero apart from
one that is merely very small. The working precision is therefore capped, by default to around
100 digits. If we try with the 1000th Fibonacci number, the following happens:
>>> N(fibonacci(1000) - (GoldenRatio)**1000/sqrt(5))
0.e+85
The lack of digits in the returned number indicates that N failed to achieve full accuracy. The
result indicates that the magnitude of the expression is something less than 1084, but that
is not a particularly good answer. To force a higher working precision, the maxn keyword
argument can be used:
>>> N(fibonacci(1000) - (GoldenRatio)**1000/sqrt(5), maxn=500)
-4.60123853010113e-210
Normally, maxn can be set very high (thousands of digits), but be aware that this may cause
signicant slowdown in extreme cases. Alternatively, the strict=True option can be set to
force an exception instead of silently returning a value with less than the requested accuracy:
>>> N(fibonacci(1000) - (GoldenRatio)**1000/sqrt(5), strict=True)
Traceback (most recent call last):
...
PrecisionExhausted: Failed to distinguish the expression:
-sqrt(5)*GoldenRatio**1000/5 + 4346655768693745643568852767504062580256466051737178040248172908953655
from zero. Try simplifying the input, using chop=True, or providing a higher maxn for evalf
Traceback (most recent call last):
5.6. Numerical evaluation
333
SymPy Documentation, Release 0.7.6
...
PrecisionExhausted: Failed to distinguish the expression:
If we add a term so that the Fibonacci approximation becomes exact (the full form of Binets
formula), we get an expression that is exactly zero, but N does not know this:
>>> f = fibonacci(100) - (GoldenRatio**100 - (GoldenRatio-1)**100)/sqrt(5)
>>> N(f)
0.e-104
>>> N(f, maxn=1000)
0.e-1336
In situations where such cancellations are known to occur, the chop options is useful. This
basically replaces very small numbers in the real or imaginary portions of a number with
exact zeros:
>>> N(f, chop=True)
0
>>> N(3 + I*f, chop=True)
3.00000000000000
In situations where you wish to remove meaningless digits, re-evaluation or the use of the
round method are useful:
>>> Float(.1, )*Float(.12345, )
0.012297
>>> ans = _
>>> N(ans, 1)
0.01
>>> ans.round(2)
0.01
If you are dealing with a numeric expression that contains no oats, it can be evaluated to
arbitrary precision. To round the result relative to a given decimal, the round method is
useful:
>>> v = 10*pi + cos(1)
>>> N(v)
31.9562288417661
>>> v.round(3)
31.956
5.6.4 Sums and integrals
Sums (in particular, innite series) and integrals can be used like regular closed-form expressions, and support arbitrary-precision evaluation:
>>> var(n x)
(n, x)
>>> Sum(1/n**n, (n, 1, oo)).evalf()
1.29128599706266
>>> Integral(x**(-x), (x, 0, 1)).evalf()
1.29128599706266
>>> Sum(1/n**n, (n, 1, oo)).evalf(50)
1.2912859970626635404072825905956005414986193682745
>>> Integral(x**(-x), (x, 0, 1)).evalf(50)
1.2912859970626635404072825905956005414986193682745
334
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> (Integral(exp(-x**2), (x, -oo, oo)) ** 2).evalf(30)
3.14159265358979323846264338328
By default, the tanh-sinh quadrature algorithm is used to evaluate integrals. This algorithm
is very ecient and robust for smooth integrands (and even integrals with endpoint singularities), but may struggle with integrals that are highly oscillatory or have mid-interval discontinuities. In many cases, evalf/N will correctly estimate the error. With the following integral,
the result is accurate but only good to four digits:
>>> f = abs(sin(x))
>>> Integral(abs(sin(x)), (x, 0, 4)).evalf()
2.346
It is better to split this integral into two pieces:
>>> (Integral(f, (x, 0, pi)) + Integral(f, (x, pi, 4))).evalf()
2.34635637913639
A similar example is the following oscillatory integral:
>>> Integral(sin(x)/x**2, (x, 1, oo)).evalf(maxn=20)
0.5
It can be dealt with much more eciently by telling evalf or N to use an oscillatory quadrature
algorithm:
>>> Integral(sin(x)/x**2, (x, 1, oo)).evalf(quad=osc)
0.504067061906928
>>> Integral(sin(x)/x**2, (x, 1, oo)).evalf(20, quad=osc)
0.50406706190692837199
Oscillatory quadrature requires an integrand containing a factor cos(ax+b) or sin(ax+b).
Note that many other oscillatory integrals can be transformed to this form with a change
of variables:
>>> init_printing(use_unicode=False, wrap_line=False, no_global=True)
>>> intgrl = Integral(sin(1/x), (x, 0, 1)).transform(x, 1/x)
>>> intgrl
oo
/
|
| sin(x)
| ------ dx
|
2
|
x
|
/
1
>>> N(intgrl, quad=osc)
0.504067061906928
Innite series use direct summation if the series converges quickly enough. Otherwise, extrapolation methods (generally the Euler-Maclaurin formula but also Richardson extrapolation) are used to speed up convergence. This allows high-precision evaluation of slowly convergent series:
>>> var(k)
k
>>> Sum(1/k**2, (k, 1, oo)).evalf()
1.64493406684823
5.6. Numerical evaluation
335
SymPy Documentation, Release 0.7.6
>>> zeta(2).evalf()
1.64493406684823
>>> Sum(1/k-log(1+1/k), (k, 1, oo)).evalf()
0.577215664901533
>>> Sum(1/k-log(1+1/k), (k, 1, oo)).evalf(50)
0.57721566490153286060651209008240243104215933593992
>>> EulerGamma.evalf(50)
0.57721566490153286060651209008240243104215933593992
The Euler-Maclaurin formula is also used for nite series, allowing them to be approximated
quickly without evaluating all terms:
>>> Sum(1/k, (k, 10000000, 20000000)).evalf()
0.693147255559946
Note that evalf makes some assumptions that are not always optimal. For ne-tuned
control over numerical summation, it might be worthwhile to manually use the method
Sum.euler maclaurin.
Special optimizations are used for rational hypergeometric series (where the term is a product
of polynomials, powers, factorials, binomial coecients and the like). N/evalf sum series of
this type very rapidly to high precision. For example, this Ramanujan formula for pi can be
summed to 10,000 digits in a fraction of a second with a simple command:
>>> f = factorial
>>> n = Symbol(n, integer=True)
>>> R = 9801/sqrt(8)/Sum(f(4*n)*(1103+26390*n)/f(n)**4/396**(4*n),
...
(n, 0, oo))
>>> N(R, 10000)
3.141592653589793238462643383279502884197169399375105820974944592307816406286208
99862803482534211706798214808651328230664709384460955058223172535940812848111745
02841027019385211055596446229489549303819644288109756659334461284756482337867831
...
5.6.5 Numerical simplication
The function nsimplify attempts to nd a formula that is numerically equal to the given
input. This feature can be used to guess an exact formula for an approximate oating-point
input, or to guess a simpler formula for a complicated symbolic input. The algorithm used
by nsimplify is capable of identifying simple fractions, simple algebraic expressions, linear
combinations of given constants, and certain elementary functional transformations of any of
the preceding.
Optionally, nsimplify can be passed a list of constants to include (e.g. pi) and a minimum
numerical tolerance. Here are some elementary examples:
>>> nsimplify(0.1)
1/10
>>> nsimplify(6.28, [pi], tolerance=0.01)
2*pi
>>> nsimplify(pi, tolerance=0.01)
22/7
>>> nsimplify(pi, tolerance=0.001)
355
--113
>>> nsimplify(0.33333, tolerance=1e-4)
336
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
1/3
>>> nsimplify(2.0**(1/3.), tolerance=0.001)
635
--504
>>> nsimplify(2.0**(1/3.), tolerance=0.001, full=True)
3 ___
\/ 2
Here are several more advanced examples:
>>> nsimplify(Float(0.130198866629986772369127970337,30), [pi, E])
1
---------5*pi
---- + 2*E
7
>>> nsimplify(cos(atan(1/3)))
____
3*\/ 10
-------10
>>> nsimplify(4/(1+sqrt(5)), [GoldenRatio])
-2 + 2*GoldenRatio
>>> nsimplify(2 + exp(2*atan(1/4)*I))
49
8*I
-- + --17
17
>>> nsimplify((1/(exp(3*pi*I/5)+1)))
___________
/
___
1
/ \/ 5
1
- - I* /
----- + 2
\/
10
4
>>> nsimplify(I**I, [pi])
-pi
---2
e
>>> n = Symbol(n)
>>> nsimplify(Sum(1/n**2, (n, 1, oo)), [pi])
2
pi
--6
>>> nsimplify(gamma(1/4)*gamma(3/4), [pi])
___
\/ 2 *pi
5.7 Numeric Computation
Symbolic computer algebra systems like SymPy facilitate the construction and manipulation
of mathematical expressions. Unfortunately when it comes time to evaluate these expressions
on numerical data, symbolic systems often have poor performance.
Fortunately SymPy oers a number of easy-to-use hooks into other numeric systems, allowing
5.7. Numeric Computation
337
SymPy Documentation, Release 0.7.6
you to create mathematical expressions in SymPy and then ship them o to the numeric
system of your choice. This page documents many of the options available including the math
library, the popular array computing package numpy, code generation in Fortran or C, and
the use of the array compiler Theano.
5.7.1 Subs/evalf
Subs is the slowest but simplest option. It runs at SymPy speeds. The .subs(...).evalf()
method can substitute a numeric value for a symbolic one and then evaluate the result within
SymPy.
>>> from sympy import *
>>> from sympy.abc import x
>>> expr = sin(x)/x
>>> expr.evalf(subs={x: 3.14})
0.000507214304613640
This method is slow. You should use this method production only if performance is not an
issue. You can expect .subs to take tens of microseconds. It can be useful while prototyping
or if you just want to see a value once.
5.7.2 Lambdify
The lambdify function translates SymPy expressions into Python functions, leveraging a variety of numerical libraries. It is used as follows:
>>> from sympy import *
>>> from sympy.abc import x
>>> expr = sin(x)/x
>>> f = lambdify(x, expr)
>>> f(3.14)
0.000507214304614
Here lambdify makes a function that computes f(x) = sin(x)/x. By default lambdify relies
on implementations in the math standard library. This numerical evaluation takes on the order
of hundreds of nanoseconds, roughly two orders of magnitude faster than the .subs method.
This is the speed dierence between SymPy and raw Python.
Lambdify can leverage a variety of numerical backends. By default it uses the math library.
However it also supports mpmath and most notably, numpy. Using the numpy library gives the
generated function access to powerful vectorized ufuncs that are backed by compiled C code.
>>>
>>>
>>>
>>>
from sympy import *
from sympy.abc import x
expr = sin(x)/x
f = lambdify(x, expr, numpy)
>>> import numpy
>>> data = numpy.linspace(1, 10, 10000)
>>> f(data)
[ 0.84147098 0.84119981 0.84092844 ..., -0.05426074 -0.05433146
-0.05440211]
If you have array-based data this can confer a considerable speedup, on the order of 10 nanoseconds per element. Unfortunately numpy incurs some start-up time and introduces an
overhead of a few microseconds.
338
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
5.7.3 uFuncify
While NumPy operations are very ecient for vectorized data they sometimes incur unnecessary costs when chained together. Consider the following operation
>>> x = get_numpy_array(...)
>>> y = sin(x) / x
The operators sin and / call routines that execute tight for loops in C. The resulting computation looks something like this
for(int i = 0; i < n; i++)
{
temp[i] = sin(x[i]);
}
for(int i = i; i < n; i++)
{
y[i] = temp[i] / x[i];
}
This is slightly sub-optimal because
1. We allocate an extra temp array
2. We walk over x memory twice when once would have been sucient
A better solution would fuse both element-wise operations into a single for loop
for(int i = i; i < n; i++)
{
y[i] = sin(x[i]) / x[i];
}
Statically compiled projects like NumPy are unable to take advantage of such optimizations.
Fortunately, SymPy is able to generate ecient low-level C or Fortran code. It can then
depend on projects like Cython or f2py to compile and reconnect that code back up to Python.
Fortunately this process is well automated and a SymPy user wishing to make use of this code
generation should call the ufuncify function
>>> from sympy import *
>>> from sympy.abc import x
>>> expr = sin(x)/x
>>> from sympy.utilities.autowrap import ufuncify
>>> f = ufuncify([x], expr)
This function f consumes and returns a NumPy array. Generally ufuncify performs at least
as well as lambdify. If the expression is complicated then ufuncify often signicantly outperforms the NumPy backed solution. Jensen has a good blog post on this topic.
5.7.4 Theano
SymPy has a strong connection with Theano, a mathematical array compiler. SymPy expressions can be easily translated to Theano graphs and then compiled using the Theano compiler
chain.
5.7. Numeric Computation
339
SymPy Documentation, Release 0.7.6
>>> from sympy import *
>>> from sympy.abc import x
>>> expr = sin(x)/x
>>> from sympy.printing.theanocode import theano_function
>>> f = theano_function([x], [expr])
If array broadcasting or types are desired then Theano requires this extra information
>>> f = theano_function([x], [expr], dims={x: 1}, dtypes={x: float64})
Theano has a more sophisticated code generation system than SymPys C/Fortran code printers. Among other things it handles common sub-expressions and compilation onto the GPU.
Theano also supports SymPy Matrix and Matrix Expression objects.
5.7.5 So Which Should I Use?
The options here were listed in order from slowest and least dependencies to fastest and
most dependencies. For example, if you have Theano installed then that will often be the best
choice. If you dont have Theano but do have f2py then you should use ufuncify.
Tool
subs/evalf
lambdify
lambdify-numpy
ufuncify
Theano
Speed
50us
1us
10ns
10ns
10ns
Qualities
Simple
Scalar functions
Vector functions
Complex vector expressions
Many outputs, CSE, GPUs
Dependencies
None
math
numpy
f2py, Cython
Theano
5.8 Functions Module
All
functions
support
the
methods
sympy.core.function.Function (page 169).
documented
below,
inherited
from
class sympy.core.function.Function
Base class for applied mathematical functions.
It also serves as a constructor for undened function classes.
Examples
First example shows how to use Function as a constructor for undened function classes:
>>> from sympy import Function, Symbol
>>> x = Symbol(x)
>>> f = Function(f)
>>> g = Function(g)(x)
>>> f
f
>>> f(x)
f(x)
>>> g
g(x)
>>> f(x).diff(x)
340
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Derivative(f(x), x)
>>> g.diff(x)
Derivative(g(x), x)
In the following example Function is used as a base class for my func that represents a
mathematical function my func. Suppose that it is well known, that my func(0) is 1 and
my func at innity goes to 0, so we want those two simplications to occur automatically.
Suppose also that my func(x) is real exactly when x is real. Here is an implementation
that honours those requirements:
>>> from sympy import Function, S, oo, I, sin
>>> class my_func(Function):
...
...
@classmethod
...
def eval(cls, x):
...
if x.is_Number:
...
if x is S.Zero:
...
return S.One
...
elif x is S.Infinity:
...
return S.Zero
...
...
def _eval_is_real(self):
...
return self.args[0].is_real
...
>>> x = S(x)
>>> my_func(0) + sin(0)
1
>>> my_func(oo)
0
>>> my_func(3.54).n() # Not yet implemented for my_func.
my_func(3.54)
>>> my_func(I).is_real
False
In order for my func to become useful, several other methods would need to be implemented. See source code of some of the already implemented functions for more complete examples.
Also, if the function can take more than one argument, then nargs must be dened, e.g.
if my func can take one or two arguments then,
>>> class my_func(Function):
...
nargs = (1, 2)
...
>>>
Function.as base exp()
Returns the method as the 2-tuple (base, exponent).
Function.fdiff(argindex=1)
Returns the rst derivative of the function.
Function.is commutative
Returns whether the functon is commutative.
5.8. Functions Module
341
SymPy Documentation, Release 0.7.6
5.8.1 Contents
Elementary
This module implements elementary functions, as well as functions like Abs, Max, etc.
Abs
Returns the absolute value of the argument.
Examples:
>>> from sympy.functions import Abs
>>> Abs(-1)
1
class sympy.functions.elementary.complexes.Abs
Return the absolute value of the argument.
This is an extension of the built-in function abs() to accept symbolic values. If you pass
a SymPy expression to the built-in abs(), it will pass it automatically to Abs().
See Also:
sign, conjugate
Examples
>>> from sympy import Abs, Symbol, S
>>> Abs(-1)
1
>>> x = Symbol(x, real=True)
>>> Abs(-x)
Abs(x)
>>> Abs(x**2)
x**2
>>> abs(-x) # The Python built-in
Abs(x)
Note that the Python built-in will return either an Expr or int depending on the argument:
>>> type(abs(-1))
<... int>
>>> type(abs(S.NegativeOne))
<class sympy.core.numbers.One>
Abs will always return a sympy object.
fdiff(argindex=1)
Get the rst derivative of the argument to Abs().
Examples
342
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> from sympy.abc import x
>>> from sympy.functions import Abs
>>> Abs(-x).fdiff()
sign(x)
acos
class sympy.functions.elementary.trigonometric.acos
The inverse cosine function.
Returns the arc cosine of x (measured in radians).
See Also:
sin, csc, cos, sec, tan, cot, asin, acsc, asec, atan, acot, atan2
Notes
acos(x) will evaluate automatically in the cases oo, -oo, 0, 1, -1.
References
[R87] (page 1903), [R88] (page 1903), [R89] (page 1903)
Examples
>>> from sympy import acos, oo, pi
>>> acos(1)
0
>>> acos(0)
pi/2
>>> acos(oo)
oo*I
inverse(argindex=1)
Returns the inverse of this function.
acosh
class sympy.functions.elementary.hyperbolic.acosh
The inverse hyperbolic cosine function.
acosh(x) -> Returns the inverse hyperbolic cosine of x
See Also:
asinh, atanh, cosh
inverse(argindex=1)
Returns the inverse of this function.
5.8. Functions Module
343
SymPy Documentation, Release 0.7.6
acot
class sympy.functions.elementary.trigonometric.acot
The inverse cotangent function.
Returns the arc cotangent of x (measured in radians).
See Also:
sin, csc, cos, sec, tan, cot, asin, acsc, acos, asec, atan, atan2
References
[R90] (page 1903), [R91] (page 1903), [R92] (page 1903)
inverse(argindex=1)
Returns the inverse of this function.
acoth
class sympy.functions.elementary.hyperbolic.acoth
The inverse hyperbolic cotangent function.
acoth(x) -> Returns the inverse hyperbolic cotangent of x
inverse(argindex=1)
Returns the inverse of this function.
arg
Returns the argument (in radians) of a complex number. For a real number, the argument is
always 0.
Examples:
>>> from sympy.functions import arg
>>> from sympy import I, sqrt
>>> arg(2.0)
0
>>> arg(I)
pi/2
>>> arg(sqrt(2) + I*sqrt(2))
pi/4
class sympy.functions.elementary.complexes.arg
Returns the argument (in radians) of a complex number
asin
class sympy.functions.elementary.trigonometric.asin
The inverse sine function.
Returns the arcsine of x in radians.
See Also:
sin, csc, cos, sec, tan, cot, acsc, acos, asec, atan, acot, atan2
344
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Notes
asin(x) will evaluate automatically in the cases oo, -oo, 0, 1, -1 and for some instances
when the result is a rational multiple of pi (see the eval class method).
References
[R93] (page 1903), [R94] (page 1903), [R95] (page 1903)
Examples
>>> from sympy import asin, oo, pi
>>> asin(1)
pi/2
>>> asin(-1)
-pi/2
inverse(argindex=1)
Returns the inverse of this function.
asinh
class sympy.functions.elementary.hyperbolic.asinh
The inverse hyperbolic sine function.
asinh(x) -> Returns the inverse hyperbolic sine of x
See Also:
acosh, atanh, sinh
inverse(argindex=1)
Returns the inverse of this function.
atan
class sympy.functions.elementary.trigonometric.atan
The inverse tangent function.
Returns the arc tangent of x (measured in radians).
See Also:
sin, csc, cos, sec, tan, cot, asin, acsc, acos, asec, acot, atan2
Notes
atan(x) will evaluate automatically in the cases oo, -oo, 0, 1, -1.
References
[R96] (page 1903), [R97] (page 1903), [R98] (page 1903)
5.8. Functions Module
345
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import atan, oo, pi
>>> atan(0)
0
>>> atan(1)
pi/4
>>> atan(oo)
pi/2
inverse(argindex=1)
Returns the inverse of this function.
atan2
This function is like atan, but considers the sign of both arguments in order to correctly
determine the quadrant of its result.
class sympy.functions.elementary.trigonometric.atan2
The function atan2(y, x) computes atan(y/x) taking two arguments y and x. Signs of
both y and x are considered to determine the appropriate quadrant of atan(y/x). The
range is (, ]. The complete denition reads as follows:
( )
x>0
arctan xy
(y)
arctan x +
y 0, x < 0
arctan ( y )
y < 0, x < 0
x
atan2(y, x) =
+
y > 0, x = 0
2
y < 0, x = 0
undened
y = 0, x = 0
Attention: Note the role reversal of both arguments. The y-coordinate is the rst argument and the x-coordinate the second.
See Also:
sin, csc, cos, sec, tan, cot, asin, acsc, acos, asec, atan, acot
References
[R99] (page 1903), [R100] (page 1903), [R101] (page 1903)
Examples
Going counter-clock wise around the origin we nd the following angles:
>>> from sympy import atan2
>>> atan2(0, 1)
0
>>> atan2(1, 1)
pi/4
>>> atan2(1, 0)
pi/2
>>> atan2(1, -1)
346
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
3*pi/4
>>> atan2(0, -1)
pi
>>> atan2(-1, -1)
-3*pi/4
>>> atan2(-1, 0)
-pi/2
>>> atan2(-1, 1)
-pi/4
which are all correct. Compare this to the results of the ordinary atan function for the
point (x, y) = (1, 1)
>>> from sympy import atan, S
>>> atan(S(1) / -1)
-pi/4
>>> atan2(1, -1)
3*pi/4
where only the atan2 function reurns what we expect. We can dierentiate the function
with respect to both arguments:
>>> from sympy import diff
>>> from sympy.abc import x, y
>>> diff(atan2(y, x), x)
-y/(x**2 + y**2)
>>> diff(atan2(y, x), y)
x/(x**2 + y**2)
We can express the atan2 function in terms of complex logarithms:
>>> from sympy import log
>>> atan2(y, x).rewrite(log)
-I*log((x + I*y)/sqrt(x**2 + y**2))
and in terms of ( atan):
>>> from sympy import atan
>>> atan2(y, x).rewrite(atan)
2*atan(y/(x + sqrt(x**2 + y**2)))
but note that this form is undened on the negative real axis.
atanh
class sympy.functions.elementary.hyperbolic.atanh
The inverse hyperbolic tangent function.
atanh(x) -> Returns the inverse hyperbolic tangent of x
See Also:
asinh, acosh, tanh
inverse(argindex=1)
Returns the inverse of this function.
5.8. Functions Module
347
SymPy Documentation, Release 0.7.6
ceiling
class sympy.functions.elementary.integers.ceiling
Ceiling is a univariate function which returns the smallest integer value not less than its
argument. Ceiling function is generalized in this implementation to complex numbers.
More information can be found in Concrete mathematics by Graham, pp. 87 or visit
http://mathworld.wolfram.com/CeilingFunction.html.
>>>
>>>
17
>>>
3
>>>
6
>>>
0
>>>
I
from sympy import ceiling, E, I, Float, Rational
ceiling(17)
ceiling(Rational(23, 10))
ceiling(2*E)
ceiling(-Float(0.567))
ceiling(I/2)
See Also:
floor
conjugate
Returns the complex conjugate of an argument. In mathematics, the complex conjugate of a
complex number is given by changing the sign of the imaginary part. Thus, the conjugate of
the complex number
a + ib
(where a and b are real numbers) is
a ib
Examples:
>>>
>>>
>>>
2
>>>
-I
from sympy.functions import conjugate
from sympy import I
conjugate(2)
conjugate(I)
class sympy.functions.elementary.complexes.conjugate
Changes the sign of the imaginary part of a complex number.
See Also:
sign, Abs
Examples
>>> from sympy import conjugate, I
348
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> conjugate(1 + I)
1 - I
cos
class sympy.functions.elementary.trigonometric.cos
The cosine function.
Returns the cosine of x (measured in radians).
See Also:
sin, csc, sec, tan, cot, asin, acsc, acos, asec, atan, acot, atan2
Notes
See sin() for notes about automatic evaluation.
References
[R102] (page 1903), [R103] (page 1903), [R104] (page 1903)
Examples
>>> from sympy import cos, pi
>>> from sympy.abc import x
>>> cos(x**2).diff(x)
-2*x*sin(x**2)
>>> cos(1).diff(x)
0
>>> cos(pi)
-1
>>> cos(pi/2)
0
>>> cos(2*pi/3)
-1/2
cosh
class sympy.functions.elementary.hyperbolic.cosh
x
x
The hyperbolic cosine function, e +e
.
2
cosh(x) -> Returns the hyperbolic cosine of x
See Also:
sinh, tanh, acosh
5.8. Functions Module
349
SymPy Documentation, Release 0.7.6
cot
class sympy.functions.elementary.trigonometric.cot
The cotangent function.
Returns the cotangent of x (measured in radians).
See Also:
sin, csc, cos, sec, tan, asin, acsc, acos, asec, atan, acot, atan2
Notes
See sin() for notes about automatic evaluation.
References
[R105] (page 1903), [R106] (page 1903), [R107] (page 1903)
Examples
>>> from sympy import cot
>>> from sympy.abc import x
>>> cot(x**2).diff(x)
2*x*(-cot(x**2)**2 - 1)
>>> cot(1).diff(x)
0
inverse(argindex=1)
Returns the inverse of this function.
coth
class sympy.functions.elementary.hyperbolic.coth
The hyperbolic cotangent function, cosh(x)
sinh(x) .
coth(x) -> Returns the hyperbolic cotangent of x
inverse(argindex=1)
Returns the inverse of this function.
exp
class sympy.functions.elementary.exponential.exp
The exponential function, ex .
See Also:
log
as real imag(deep=True, **hints)
Returns this function as a 2-tuple representing a complex number.
See Also:
350
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
sympy.functions.elementary.complexes.re
sympy.functions.elementary.complexes.im
(page
357),
Examples
>>> from sympy import I
>>> from sympy.abc import x
>>> from sympy.functions import exp
>>> exp(x).as_real_imag()
(exp(re(x))*cos(im(x)), exp(re(x))*sin(im(x)))
>>> exp(1).as_real_imag()
(E, 0)
>>> exp(I).as_real_imag()
(cos(1), sin(1))
>>> exp(1+I).as_real_imag()
(E*cos(1), E*sin(1))
base
Returns the base of the exponential function.
fdiff(argindex=1)
Returns the rst derivative of this function.
static taylor term(*args, **kwargs)
Calculates the next term in the Taylor series expansion.
See Also:
classes sympy.functions.elementary.exponential.log (page 353)
ExprCondPair
class sympy.functions.elementary.piecewise.ExprCondPair
Represents an expression, condition pair.
cond
Returns the condition of this pair.
expr
Returns the expression of this pair.
free symbols
Return the free symbols of this pair.
oor
class sympy.functions.elementary.integers.floor
Floor is a univariate function which returns the largest integer value not greater than
its argument. However this implementation generalizes oor to complex numbers.
More information can be found in Concrete mathematics by Graham, pp. 87 or visit
http://mathworld.wolfram.com/FloorFunction.html.
>>> from sympy import floor, E, I, Float, Rational
>>> floor(17)
17
>>> floor(Rational(23, 10))
5.8. Functions Module
351
SymPy Documentation, Release 0.7.6
2
>>> floor(2*E)
5
>>> floor(-Float(0.567))
-1
>>> floor(-I/2)
-I
See Also:
ceiling
HyperbolicFunction
class sympy.functions.elementary.hyperbolic.HyperbolicFunction
Base class for hyperbolic functions.
See Also:
sinh, cosh, tanh, coth
IdentityFunction
class sympy.functions.elementary.miscellaneous.IdentityFunction
The identity function
Examples
>>> from sympy import Id, Symbol
>>> x = Symbol(x)
>>> Id(x)
x
im
Returns the imaginary part of an expression.
Examples:
>>> from sympy.functions import im
>>> from sympy import I
>>> im(2+3*I)
3
See Also:
sympy.functions.elementary.complexes.re (page 357)
LambertW
class sympy.functions.elementary.exponential.LambertW
The Lambert W function W (z) is dened as the inverse function of w exp(w) [R108]
(page 1903).
352
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
In other words, the value of W (z) is such that z = W (z) exp(W (z)) for any complex number
z. The Lambert W function is a multivalued function with innitely many branches Wk (z),
indexed by k Z. Each branch gives a dierent solution w of the equation z = w exp(w).
The Lambert W function has two partially real branches: the principal branch (k = 0) is
real for real z > 1/e, and the k = 1 branch is real for 1/e < z < 0. All branches except
k = 0 have a logarithmic singularity at z = 0.
References
[R108] (page 1903)
Examples
>>> from sympy import LambertW
>>> LambertW(1.2)
0.635564016364870
>>> LambertW(1.2, -1).n()
-1.34747534407696 - 4.41624341514535*I
>>> LambertW(-1).is_real
False
fdiff(argindex=1)
Return the rst derivative of this function.
log
class sympy.functions.elementary.exponential.log
The natural logarithm function ln(x) or log(x). Logarithms are taken with the natural
base, e. To get a logarithm of a dierent base b, use log(x, b), which is essentially
short-hand for log(x)/log(b).
See Also:
exp
as base exp()
Returns this function in the form (base, exponent).
as real imag(deep=True, **hints)
Returns this function as a complex coordinate.
Examples
>>> from sympy import I
>>> from sympy.abc import x
>>> from sympy.functions import log
>>> log(x).as_real_imag()
(log(Abs(x)), arg(x))
>>> log(I).as_real_imag()
(0, pi/2)
>>> log(1 + I).as_real_imag()
(log(sqrt(2)), pi/4)
5.8. Functions Module
353
SymPy Documentation, Release 0.7.6
>>> log(I*x).as_real_imag()
(log(Abs(x)), arg(I*x))
fdiff(argindex=1)
Returns the rst derivative of the function.
inverse(argindex=1)
Returns ex , the inverse function of log(x).
static taylor term(*args, **kwargs)
Returns the next term in the Taylor series expansion of log(1 + x).
See Also:
classes sympy.functions.elementary.exponential.exp (page 350)
Min
Returns the minimum of two (comparable) expressions.
Examples:
>>> from sympy.functions import Min
>>> Min(1,2)
1
>>> from sympy.abc import x
>>> Min(1, x)
Min(1, x)
It is named Min and not min to avoid conicts with the built-in function min.
class sympy.functions.elementary.miscellaneous.Min
Return, if possible, the minimum value of the list.
See Also:
Max nd maximum values
Examples
>>>
>>>
>>>
>>>
from sympy import Min, Symbol, oo
from sympy.abc import x, y
p = Symbol(p, positive=True)
n = Symbol(n, negative=True)
>>> Min(x, -2)
Min(x, -2)
>>> Min(x, -2).subs(x, 3)
-2
>>> Min(p, -3)
-3
>>> Min(x, y)
Min(x, y)
354
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> Min(n, 8, p, -7, p, oo)
Min(n, -7)
Max
Returns the maximum of two (comparable) expressions
It is named Max and not max to avoid conicts with the built-in function max.
class sympy.functions.elementary.miscellaneous.Max
Return, if possible, the maximum value of the list.
When number of arguments is equal one, then return this argument.
When number of arguments is equal two, then return, if possible, the value from (a, b)
that is >= the other.
In common case, when the length of list greater than 2, the task is more complicated.
Return only the arguments, which are greater than others, if it is possible to determine
directional relation.
If is not possible to determine such a relation, return a partially evaluated result.
Assumptions are used to make the decision too.
Also, only comparable arguments are permitted.
See Also:
Min nd minimum values
References
[R109] (page 1903), [R110] (page 1903)
Examples
>>>
>>>
>>>
>>>
from sympy import Max, Symbol, oo
from sympy.abc import x, y
p = Symbol(p, positive=True)
n = Symbol(n, negative=True)
>>> Max(x, -2)
Max(x, -2)
>>> Max(x, -2).subs(x, 3)
3
>>> Max(p, -2)
p
>>> Max(x, y)
Max(x, y)
>>> Max(x, y) == Max(y, x)
True
5.8. Functions Module
355
SymPy Documentation, Release 0.7.6
>>> Max(x, Max(y, z))
Max(x, y, z)
>>> Max(n, 8, p, 7, -oo)
Max(8, p)
>>> Max (1, x, oo)
oo
Algorithm
The task can be considered as searching of supremums in the directed complete partial
orders [R109] (page 1903).
The source values are sequentially allocated by the isolated subsets in which supremums
are searched and result as Max arguments.
If the resulted supremum is single, then it is returned.
The isolated subsets are the sets of values which are only the comparable with each
other in the current set. E.g. natural numbers are comparable with each other, but not
comparable with the x symbol. Another example: the symbol x with negative assumption
is comparable with a natural number.
Also there are least elements, which are comparable with all others, and have a zero
property (maximum or minimum for all elements). E.g. oo. In case of it the allocation
operation is terminated and only this value is returned.
Assumption:
if A > B > C then A > C
if A==B then B can be removed
Piecewise
class sympy.functions.elementary.piecewise.Piecewise
Represents a piecewise function.
Usage:
Piecewise( (expr,cond), (expr,cond), ... )
Each argument is a 2-tuple dening an expression and condition
The conds are evaluated in turn returning the rst that is True. If any of
the evaluated conds are not determined explicitly False, e.g. x < 1, the
function is returned in symbolic form.
If the function is evaluated at a place where all conditions are False, a
ValueError exception will be raised.
Pairs where the cond is explicitly False, will be removed.
See Also:
piecewise fold
356
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import Piecewise, log
>>> from sympy.abc import x
>>> f = x**2
>>> g = log(x)
>>> p = Piecewise( (0, x<-1), (f, x<=1), (g, True))
>>> p.subs(x,1)
1
>>> p.subs(x,5)
log(5)
doit(**hints)
Evaluate this piecewise function.
sympy.functions.elementary.piecewise.piecewise fold(expr)
Takes an expression containing a piecewise function and returns the expression in piecewise form.
See Also:
Piecewise
Examples
>>> from sympy import Piecewise, piecewise_fold, sympify as S
>>> from sympy.abc import x
>>> p = Piecewise((x, x < 1), (1, S(1) <= x))
>>> piecewise_fold(x*p)
Piecewise((x**2, x < 1), (x, 1 <= x))
re
Return the real part of an expression.
Examples:
>>> from sympy.functions import re
>>> from sympy import I
>>> re(2+3*I)
2
class sympy.functions.elementary.complexes.re
Returns real part of expression. This function performs only elementary analysis and so
it will fail to decompose properly more complicated expressions. If completely simplied
result is needed then use Basic.as real imag() or perform complex expansion on instance
of this function.
>>> from sympy import re, im, I, E
>>> from sympy.abc import x, y
>>> re(2*E)
2*E
>>> re(2*I + 17)
17
5.8. Functions Module
357
SymPy Documentation, Release 0.7.6
>>> re(2*I)
0
>>> re(im(x) + x*I + 2)
2
See Also:
im
as real imag(deep=True, **hints)
Returns the real number with a zero complex part.
See Also:
sympy.functions.elementary.complexes.im
root
sympy.functions.elementary.miscellaneous.root(arg, n)
The n-th root function (a shortcut for arg**(1/n))
root(x, n) -> Returns the principal n-th root of x.
See Also:
sympy.polys.rootoftools.RootOf (page 1119), sympy.core.power.integer nthroot
(page 141), sqrt, real root
References
http://en.wikipedia.org/wiki/Square root
http://en.wikipedia.org/wiki/Real root
http://en.wikipedia.org/wiki/Root of unity
http://en.wikipedia.org/wiki/Principal value
http://mathworld.wolfram.com/CubeRoot.html
Examples
>>> from sympy import root, Rational
>>> from sympy.abc import x, n
>>> root(x, 2)
sqrt(x)
>>> root(x, 3)
x**(1/3)
>>> root(x, n)
x**(1/n)
>>> root(x, -Rational(2, 3))
x**(-3/2)
358
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
To get all n n-th roots you can use the RootOf function. The following examples show the
roots of unity for n equal 2, 3 and 4:
>>> from sympy import RootOf, I
>>> [ RootOf(x**2-1,i) for i in (0,1) ]
[-1, 1]
>>> [ RootOf(x**3-1,i) for i in (0,1,2) ]
[1, -1/2 - sqrt(3)*I/2, -1/2 + sqrt(3)*I/2]
>>> [ RootOf(x**4-1,i) for i in (0,1,2,3) ]
[-1, 1, -I, I]
SymPy, like other symbolic algebra systems, returns the complex root of negative numbers. This is the principal root and diers from the text-book result that one might be
expecting. For example, the cube root of -8 does not come back as -2:
>>> root(-8, 3)
2*(-1)**(1/3)
The real root function can be used to either make such a result real or simply return the
real root in the rst place:
>>> from sympy import real_root
>>> real_root(_)
-2
>>> real_root(-32, 5)
-2
RoundFunction
class sympy.functions.elementary.integers.RoundFunction
The base class for rounding functions.
sin
class sympy.functions.elementary.trigonometric.sin
The sine function.
Returns the sine of x (measured in radians).
See Also:
csc, cos, sec, tan, cot, asin, acsc, acos, asec, atan, acot, atan2
Notes
This function will evaluate automatically in the case x/pi is some rational number [R114]
(page 1903). For example, if x is a multiple of pi, pi/2, pi/3, pi/4 and pi/6.
References
[R111] (page 1903), [R112] (page 1903), [R113] (page 1903), [R114] (page 1903)
5.8. Functions Module
359
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import sin, pi
>>> from sympy.abc import x
>>> sin(x**2).diff(x)
2*x*cos(x**2)
>>> sin(1).diff(x)
0
>>> sin(pi)
0
>>> sin(pi/2)
1
>>> sin(pi/6)
1/2
sinh
class sympy.functions.elementary.hyperbolic.sinh
x
x
.
The hyperbolic sine function, e e
2
sinh(x) -> Returns the hyperbolic sine of x
See Also:
cosh, tanh, asinh
as real imag(deep=True, **hints)
Returns this function as a complex coordinate.
fdiff(argindex=1)
Returns the rst derivative of this function.
inverse(argindex=1)
Returns the inverse of this function.
static taylor term(*args, **kwargs)
Returns the next term in the Taylor series expansion.
sqrt
Returns the square root of an expression. It is equivalent to raise to Rational(1,2).
>>> from sympy.functions import sqrt
>>> from sympy import Rational
>>> sqrt(2) == 2**Rational(1,2)
True
class sympy.functions.elementary.miscellaneous.sqrt
The square root function
sqrt(x) -> Returns the principal square root of x.
See Also:
sympy.polys.rootoftools.RootOf (page 1119), root, real root
360
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
References
http://en.wikipedia.org/wiki/Square root
http://en.wikipedia.org/wiki/Principal value
Examples
>>> from sympy import sqrt, Symbol
>>> x = Symbol(x)
>>> sqrt(x)
sqrt(x)
>>> sqrt(x)**2
x
Note that sqrt(x**2) does not simplify to x.
>>> sqrt(x**2)
sqrt(x**2)
This is because the two are not equal to each other in general. For example, consider x
== -1:
>>> from sympy import Eq
>>> Eq(sqrt(x**2), x).subs(x, -1)
False
This is because sqrt computes the principal square root, so the square may put the argument in a dierent branch. This identity does hold if x is positive:
>>> y = Symbol(y, positive=True)
>>> sqrt(y**2)
y
You can force this simplication by using the powdenest() function with the force option
set to True:
>>> from sympy import powdenest
>>> sqrt(x**2)
sqrt(x**2)
>>> powdenest(sqrt(x**2), force=True)
x
To get both branches of the square root you can use the RootOf function:
>>> from sympy import RootOf
>>> [ RootOf(x**2-3,i) for i in (0,1) ]
[-sqrt(3), sqrt(3)]
sign
class sympy.functions.elementary.complexes.sign
Returns the complex sign of an expression:
5.8. Functions Module
361
SymPy Documentation, Release 0.7.6
If the expresssion is real the sign will be:
1 if expression is positive
0 if expression is equal to zero
-1 if expression is negative
If the expresssion is imaginary the sign will be:
I if im(expression) is positive
-I if im(expression) is negative
Otherwise an unevaluated expression will be returned. When evaluated, the result (in
general) will be cos(arg(expr)) + I*sin(arg(expr)).
See Also:
Abs, conjugate
Examples
>>> from sympy.functions import sign
>>> from sympy.core.numbers import I
>>> sign(-1)
-1
>>> sign(0)
0
>>> sign(-3*I)
-I
>>> sign(1 + I)
sign(1 + I)
>>> _.evalf()
0.707106781186548 + 0.707106781186548*I
tan
class sympy.functions.elementary.trigonometric.tan
The tangent function.
Returns the tangent of x (measured in radians).
See Also:
sin, csc, cos, sec, cot, asin, acsc, acos, asec, atan, acot, atan2
Notes
See sin() for notes about automatic evaluation.
References
[R115] (page 1903), [R116] (page 1903), [R117] (page 1903)
362
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import tan
>>> from sympy.abc import x
>>> tan(x**2).diff(x)
2*x*(tan(x**2)**2 + 1)
>>> tan(1).diff(x)
0
inverse(argindex=1)
Returns the inverse of this function.
tanh
class sympy.functions.elementary.hyperbolic.tanh
sinh(x)
The hyperbolic tangent function, cosh(x)
.
tanh(x) -> Returns the hyperbolic tangent of x
See Also:
sinh, cosh, atanh
inverse(argindex=1)
Returns the inverse of this function.
Combinatorial
This module implements various combinatorial functions.
bell
class sympy.functions.combinatorial.numbers.bell
Bell numbers / Bell polynomials
The Bell numbers satisfy B0 = 1 and
Bn =
n1
(
)
n1
Bk .
k
k=0
They are also given by:
1 kn
.
Bn =
e
k!
k=0
The Bell polynomials are given by B0 (x) = 1 and
Bn (x) = x
n1
(
k=1
5.8. Functions Module
)
n1
Bk1 (x).
k1
363
SymPy Documentation, Release 0.7.6
The second kind of Bell polynomials (are sometimes called partial Bell polynomials or
incomplete Bell polynomials) are dened as
(
)jnk+1
( x )j 1 ( x )j 2
n!
xnk+1
1
2
Bn,k (x1 , x2 , . . . xnk+1 ) =
.
j1 !j2 ! jnk+1 ! 1!
2!
(n k + 1)!
j +j +j +=k
1
2
2
j1 +2j2 +3j2 +=n
bell(n) gives the nth Bell number, Bn .
bell(n, x) gives the nth Bell polynomial, Bn (x).
bell(n, k, (x1, x2,
Bn,k (x1 , x2 , . . . , xnk+1 ).
...))
gives
Bell
polynomials
of
the
second
kind,
See Also:
bernoulli, catalan, euler, fibonacci, harmonic, lucas
Notes
Not to be confused with Bernoulli numbers and Bernoulli polynomials, which use the
same notation.
References
[R63] (page 1904), [R64] (page 1904), [R65] (page 1904)
Examples
>>> from sympy import bell, Symbol, symbols
>>> [bell(n) for n in range(11)]
[1, 1, 2, 5, 15, 52, 203, 877, 4140, 21147, 115975]
>>> bell(30)
846749014511809332450147
>>> bell(4, Symbol(t))
t**4 + 6*t**3 + 7*t**2 + t
>>> bell(6, 2, symbols(x:6)[1:])
6*x1*x5 + 15*x2*x4 + 10*x3**2
bernoulli
class sympy.functions.combinatorial.numbers.bernoulli
Bernoulli numbers / Bernoulli polynomials
The Bernoulli numbers are a sequence of rational numbers dened by B 0 = 1 and the
recursive relation (n > 0):
n
0 =
364
\
/ n + 1 \
)
|
| * B .
\
k
/
k
/
k = 0
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
They are also commonly dened by their exponential generating function, which is
x/(exp(x) - 1). For odd indices > 1, the Bernoulli numbers are zero.
The Bernoulli polynomials satisfy the analogous formula:
n
\
/ n \
n-k
B (x) = )
|
| * B * x
.
\ k /
k
n
/
k = 0
Bernoulli numbers and Bernoulli polynomials are related as B n(0) = B n.
We compute Bernoulli numbers using Ramanujans formula:
B
n
(A(n) - S(n))
/ n + 3 \
|
|
\
n
/
where A(n) = (n+3)/3 when n = 0 or 2 (mod 6), A(n) = -(n+3)/6 when n = 4 (mod 6), and:
[n/6]
S(n) =
\
/ n + 3 \
)
|
| * B
/
\ n - 6*k /
n-6*k
k = 1
This formula is similar to the sum given in the denition, but cuts 2/3 of the terms. For
Bernoulli polynomials, we use the formula in the denition.
bernoulli(n) gives the nth Bernoulli number, B n
bernoulli(n, x) gives the nth Bernoulli polynomial in x, B n(x)
See Also:
bell, catalan, euler, fibonacci, harmonic, lucas
References
[R66] (page 1904), [R67] (page 1904), [R68] (page 1904), [R69] (page 1904)
Examples
>>> from sympy import bernoulli
>>> [bernoulli(n) for n in range(11)]
[1, -1/2, 1/6, 0, -1/30, 0, 1/42, 0, -1/30, 0, 5/66]
>>> bernoulli(1000001)
0
binomial
class sympy.functions.combinatorial.factorials.binomial
Implementation of the binomial coecient. It can be dened in two ways depending on
its desired interpretation:
5.8. Functions Module
365
SymPy Documentation, Release 0.7.6
C(n,k) = n!/(k!(n-k)!) or C(n, k) = (n, k)/k!
First, in a strict combinatorial sense it denes the number of ways we can choose k
elements from a set of n elements. In this case both arguments are nonnegative integers
and binomial is computed using an ecient algorithm based on prime factorization.
The other denition is generalization for arbitrary n, however k must also be nonnegative. This case is very useful when evaluating summations.
For the sake of convenience for negative k this function will return zero no matter what
valued is the other argument.
To expand the binomial when n is a symbol, use either expand func() or expand(func=True). The former will keep the polynomial in factored form while the latter
will expand the polynomial itself. See examples for details.
Examples
>>> from sympy import Symbol, Rational, binomial, expand_func
>>> n = Symbol(n, integer=True)
>>> binomial(15, 8)
6435
>>> binomial(n, -1)
0
>>>
[1]
>>>
[1,
>>>
[1,
>>>
[1,
>>>
[1,
[ binomial(0, i) for i in range(1)]
[ binomial(1,
1]
[ binomial(2,
2, 1]
[ binomial(3,
3, 3, 1]
[ binomial(4,
4, 6, 4, 1]
i) for i in range(2)]
i) for i in range(3)]
i) for i in range(4)]
i) for i in range(5)]
>>> binomial(Rational(5,4), 3)
-5/128
>>> binomial(n, 3)
binomial(n, 3)
>>> binomial(n, 3).expand(func=True)
n**3/6 - n**2/2 + n/3
>>> expand_func(binomial(n, 3))
n*(n - 2)*(n - 1)/6
catalan
class sympy.functions.combinatorial.numbers.catalan
Catalan numbers
The n-th catalan number is given by:
366
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
1
/ 2*n \
C = ----- |
|
n
n + 1 \ n /
catalan(n) gives the n-th Catalan number, C n
See Also:
bell,
bernoulli,
euler,
fibonacci,
harmonic,
sympy.functions.combinatorial.factorials.binomial (page 365)
lucas,
References
[R70] (page 1904), [R71] (page 1904), [R72] (page 1904), [R73] (page 1904)
Examples
>>> from sympy import (Symbol, binomial, gamma, hyper, polygamma,
...
catalan, diff, combsimp, Rational, I)
>>> [ catalan(i) for i in range(1,10) ]
[1, 2, 5, 14, 42, 132, 429, 1430, 4862]
>>> n = Symbol(n, integer=True)
>>> catalan(n)
catalan(n)
Catalan numbers can be transformed into several other, identical expressions involving
other mathematical functions
>>> catalan(n).rewrite(binomial)
binomial(2*n, n)/(n + 1)
>>> catalan(n).rewrite(gamma)
4**n*gamma(n + 1/2)/(sqrt(pi)*gamma(n + 2))
>>> catalan(n).rewrite(hyper)
hyper((-n + 1, -n), (2,), 1)
For some non-integer values of n we can get closed form expressions by rewriting in
terms of gamma functions:
>>> catalan(Rational(1,2)).rewrite(gamma)
8/(3*pi)
We can dierentiate the Catalan numbers C(n) interpreted as a continuous real funtion
in n:
>>> diff(catalan(n), n)
(polygamma(0, n + 1/2) - polygamma(0, n + 2) + log(4))*catalan(n)
As a more advanced example consider the following ratio between consecutive numbers:
5.8. Functions Module
367
SymPy Documentation, Release 0.7.6
>>> combsimp((catalan(n + 1)/catalan(n)).rewrite(binomial))
2*(2*n + 1)/(n + 2)
The Catalan numbers can be generalized to complex numbers:
>>> catalan(I).rewrite(gamma)
4**I*gamma(1/2 + I)/(sqrt(pi)*gamma(2 + I))
and evaluated with arbitrary precision:
>>> catalan(I).evalf(20)
0.39764993382373624267 - 0.020884341620842555705*I
euler
class sympy.functions.combinatorial.numbers.euler
Euler numbers
The euler numbers are given by:
2*n+1
j
2*n+1
\
\
/ k \ (-1) * (k-2*j)
E
= I )
)
|
| -------------------2n
/
/
\ j /
k
k
k = 1 j = 0
2 * I * k
E
= 0
2n+1
euler(n) gives the n-th Euler number, E n
See Also:
bell, bernoulli, catalan, fibonacci, harmonic, lucas
References
[R74] (page 1904), [R75] (page 1904), [R76] (page 1904), [R77] (page 1904)
Examples
>>> from sympy import Symbol
>>> from sympy.functions import euler
>>> [euler(n) for n in range(10)]
[1, 0, -1, 0, 5, 0, -61, 0, 1385, 0]
>>> n = Symbol(n)
>>> euler(n+2*n)
euler(3*n)
368
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
factorial
class sympy.functions.combinatorial.factorials.factorial
Implementation of factorial function over nonnegative integers. By convention (consistent with the gamma function and the binomial coecients), factorial of a negative
integer is complex innity.
The factorial is very important in combinatorics where it gives the number of ways in
which n objects can be permuted. It also arises in calculus, probability, number theory,
etc.
There is strict relation of factorial with gamma function. In fact n! = gamma(n+1)
for nonnegative integers. Rewrite of this kind is very useful in case of combinatorial
simplication.
Computation of the factorial is done using two algorithms. For small arguments naive
product is evaluated. However for bigger input algorithm Prime-Swing is used. It is
the fastest algorithm known and computes n! via prime factorization of special class of
numbers, called here the Swing Numbers.
See Also:
factorial2, RisingFactorial, FallingFactorial
Examples
>>> from sympy import Symbol, factorial, S
>>> n = Symbol(n, integer=True)
>>> factorial(0)
1
>>> factorial(7)
5040
>>> factorial(-2)
zoo
>>> factorial(n)
factorial(n)
>>> factorial(2*n)
factorial(2*n)
>>> factorial(S(1)/2)
factorial(1/2)
factorial2 / double factorial
class sympy.functions.combinatorial.factorials.factorial2
The double factorial n!!, not to be confused with (n!)!
The double factorial is dened for integers >= -1 as:
5.8. Functions Module
369
SymPy Documentation, Release 0.7.6
,
| n*(n - 2)*(n - 4)* ... * 1
n!! = { n*(n - 2)*(n - 4)* ... * 2
| 1
for n odd
for n even
for n = 0, -1
See Also:
factorial, RisingFactorial, FallingFactorial
Examples
>>> from sympy import factorial2, var
>>> var(n)
n
>>> factorial2(n + 1)
factorial2(n + 1)
>>> factorial2(5)
15
>>> factorial2(-1)
1
FallingFactorial
class sympy.functions.combinatorial.factorials.FallingFactorial
Falling factorial (related to rising factorial) is a double valued function arising in concrete
mathematics, hypergeometric functions and series expansions. It is dened by
(x, k) = x * (x-1) * ... * (x - k+1)
where x can be arbitrary expression and k is an integer.
information check Concrete mathematics by Graham, pp.
http://mathworld.wolfram.com/FallingFactorial.html page.
66
For more
or visit
>>> from sympy import ff
>>> from sympy.abc import x
>>> ff(x, 0)
1
>>> ff(5, 5)
120
>>> ff(x, 5) == x*(x-1)*(x-2)*(x-3)*(x-4)
True
See Also:
factorial, factorial2, RisingFactorial
bonacci
class sympy.functions.combinatorial.numbers.fibonacci
Fibonacci numbers / Fibonacci polynomials
370
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
The Fibonacci numbers are the integer sequence dened by the initial terms F 0 = 0,
F 1 = 1 and the two-term recurrence relation F n = F {n-1} + F {n-2}.
The Fibonacci polynomials are dened by F 1(x) = 1, F 2(x) = x, and F n(x) = x*F {n-1}(x)
+ F {n-2}(x) for n > 2. For all positive integers n, F n(1) = F n.
bonacci(n) gives the nth Fibonacci number, F n
bonacci(n, x) gives the nth Fibonacci polynomial in x, F n(x)
See Also:
bell, bernoulli, catalan, euler, harmonic, lucas
References
[R78] (page 1904), [R79] (page 1904)
Examples
>>> from sympy import fibonacci, Symbol
>>> [fibonacci(x) for x in range(11)]
[0, 1, 1, 2, 3, 5, 8, 13, 21, 34, 55]
>>> fibonacci(5, Symbol(t))
t**4 + 3*t**2 + 1
harmonic
class sympy.functions.combinatorial.numbers.harmonic
Harmonic numbers
The nth harmonic number is given by Hn = 1 +
1
2
1
3
+ . . . + n1 .
More generally:
Hn,m =
1
km
k=1
As n , Hn,m (m), the Riemann zeta function.
harmonic(n) gives the nth harmonic number, Hn
harmonic(n, m) gives the nth generalized harmonic number of order m, Hn,m , where
harmonic(n) == harmonic(n, 1)
See Also:
bell, bernoulli, catalan, euler, fibonacci, lucas
References
[R80] (page 1904), [R81] (page 1904), [R82] (page 1904)
5.8. Functions Module
371
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import harmonic, oo
>>> [harmonic(n) for n in range(6)]
[0, 1, 3/2, 11/6, 25/12, 137/60]
>>> [harmonic(n, 2) for n in range(6)]
[0, 1, 5/4, 49/36, 205/144, 5269/3600]
>>> harmonic(oo, 2)
pi**2/6
>>> from sympy import Symbol, Sum
>>> n = Symbol(n)
>>> harmonic(n).rewrite(Sum)
Sum(1/_k, (_k, 1, n))
We can evaluate harmonic numbers for all integral and positive rational arguments:
>>> from sympy import S, expand_func, simplify
>>> harmonic(8)
761/280
>>> harmonic(11)
83711/27720
>>> H = harmonic(1/S(3))
>>> H
harmonic(1/3)
>>> He = expand_func(H)
>>> He
-log(6) - sqrt(3)*pi/6 + 2*Sum(log(sin(_k*pi/3))*cos(2*_k*pi/3), (_k, 1, 1))
+ 3*Sum(1/(3*_k + 1), (_k, 0, 0))
>>> He.doit()
-log(6) - sqrt(3)*pi/6 - log(sqrt(3)/2) + 3
>>> H = harmonic(25/S(7))
>>> He = simplify(expand_func(H).doit())
>>> He
log(sin(pi/7)**(-2*cos(pi/7))*sin(2*pi/7)**(2*cos(16*pi/7))*cos(pi/14)**(-2*sin(pi/14))/14)
+ pi*tan(pi/14)/2 + 30247/9900
>>> He.n(40)
1.983697455232980674869851942390639915940
>>> harmonic(25/S(7)).n(40)
1.983697455232980674869851942390639915940
We can rewrite harmonic numbers in terms of polygamma functions:
>>> from sympy import digamma, polygamma
>>> m = Symbol(m)
>>> harmonic(n).rewrite(digamma)
polygamma(0, n + 1) + EulerGamma
>>> harmonic(n).rewrite(polygamma)
polygamma(0, n + 1) + EulerGamma
>>> harmonic(n,3).rewrite(polygamma)
polygamma(2, n + 1)/2 - polygamma(2, 1)/2
372
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> harmonic(n,m).rewrite(polygamma)
(-1)**m*(polygamma(m - 1, 1) - polygamma(m - 1, n + 1))/factorial(m - 1)
Integer osets in the argument can be pulled out:
>>> from sympy import expand_func
>>> expand_func(harmonic(n+4))
harmonic(n) + 1/(n + 4) + 1/(n + 3) + 1/(n + 2) + 1/(n + 1)
>>> expand_func(harmonic(n-4))
harmonic(n) - 1/(n - 1) - 1/(n - 2) - 1/(n - 3) - 1/n
Some limits can be computed as well:
>>> from sympy import limit, oo
>>> limit(harmonic(n), n, oo)
oo
>>> limit(harmonic(n, 2), n, oo)
pi**2/6
>>> limit(harmonic(n, 3), n, oo)
-polygamma(2, 1)/2
However we can not compute the general relation yet:
>>> limit(harmonic(n, m), n, oo)
harmonic(oo, m)
which equals zeta(m) for m > 1.
lucas
class sympy.functions.combinatorial.numbers.lucas
Lucas numbers
Lucas numbers satisfy a recurrence relation similar to that of the Fibonacci sequence,
in which each term is the sum of the preceding two. They are generated by choosing the
initial values L 0 = 2 and L 1 = 1.
lucas(n) gives the nth Lucas number
See Also:
bell, bernoulli, catalan, euler, fibonacci, harmonic
References
[R83] (page 1904), [R84] (page 1904)
Examples
5.8. Functions Module
373
SymPy Documentation, Release 0.7.6
>>> from sympy import lucas
>>> [lucas(x) for x in range(11)]
[2, 1, 3, 4, 7, 11, 18, 29, 47, 76, 123]
MultiFactorial
class sympy.functions.combinatorial.factorials.MultiFactorial
RisingFactorial
class sympy.functions.combinatorial.factorials.RisingFactorial
Rising factorial (also called Pochhammer symbol) is a double valued function arising in
concrete mathematics, hypergeometric functions and series expansions. It is dened by:
rf(x, k) = x * (x+1) * ... * (x + k-1)
where x can be arbitrary expression and k is an integer.
information check Concrete mathematics by Graham, pp.
http://mathworld.wolfram.com/RisingFactorial.html page.
66
For more
or visit
See Also:
factorial, factorial2, FallingFactorial
Examples
>>> from sympy import rf
>>> from sympy.abc import x
>>> rf(x, 0)
1
>>> rf(1, 5)
120
>>> rf(x, 5) == x*(1 + x)*(2 + x)*(3 + x)*(4 + x)
True
stirling
sympy.functions.combinatorial.numbers.stirling(n,
k,
d=None,
signed=False)
Return Stirling number S(n, k) of the rst or second (default) kind.
kind=2,
The sum of all Stirling numbers of the second kind for k = 1 through n is bell(n). The
recurrence relationship for these numbers is:
{0}
{ } = 1;
{0}
374
{n}
{0}
{n + 1}
{n}
{ n }
{ } = { } = 0; {
} = j*{ } + {
}
{0}
{k}
{ k }
{k}
{k - 1}
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
where j is:: n for Stirling numbers of the rst kind -n for signed Stirling numbers of
the rst kind k for Stirling numbers of the second kind
The rst kind of Stirling number counts the number of permutations of n distinct items
that have k cycles; the second kind counts the ways in which n distinct items can be
partitioned into k parts. If d is given, the reduced Stirling number of the second kind
is returned: S{d}(n, k) = S(n - d + 1, k - d + 1) with n >= k >= d. (This counts
the ways to partition n consecutive integers into k groups with no pairwise dierence
less than d. See example below.)
To obtain the signed Stirling numbers of the rst kind, use keyword signed=True. Using
this keyword automatically sets kind to 1.
See Also:
sympy.utilities.iterables.multiset partitions (page 1563)
References
[R85] (page 1904), [R86] (page 1904)
Examples
>>> from sympy.functions.combinatorial.numbers import stirling, bell
>>> from sympy.combinatorics import Permutation
>>> from sympy.utilities.iterables import multiset_partitions, permutations
First kind (unsigned by default):
>>>
[0,
>>>
>>>
[0,
>>>
[0,
[stirling(6, i, kind=1) for i in range(7)]
120, 274, 225, 85, 15, 1]
perms = list(permutations(range(4)))
[sum(Permutation(p).cycles == i for p in perms) for i in range(5)]
6, 11, 6, 1]
[stirling(4, i, kind=1) for i in range(5)]
6, 11, 6, 1]
First kind (signed):
>>> [stirling(4, i, signed=True) for i in range(5)]
[0, -6, 11, -6, 1]
Second kind:
>>> [stirling(10, i) for i in range(12)]
[0, 1, 511, 9330, 34105, 42525, 22827, 5880, 750, 45, 1, 0]
>>> sum(_) == bell(10)
True
>>> len(list(multiset_partitions(range(4), 2))) == stirling(4, 2)
True
Reduced second kind:
>>> from sympy import subsets, oo
>>> def delta(p):
...
if len(p) == 1:
...
return oo
5.8. Functions Module
375
SymPy Documentation, Release 0.7.6
...
>>>
>>>
>>>
7
>>>
7
return min(abs(i[0] - i[1]) for i in subsets(p, 2))
parts = multiset_partitions(range(5), 3)
d = 2
sum(1 for p in parts if all(delta(i) >= d for i in p))
stirling(5, 3, 2)
Enumeration
Three functions are available. Each of them attempts to eciently compute a given combinatorial quantity for a given set or multiset which can be entered as an integer, sequence or
multiset (dictionary with elements as keys and multiplicities as values). The k parameter indicates the number of elements to pick (or the number of partitions to make). When k is None,
the sum of the enumeration for all k (from 0 through the number of items represented by n)
is returned. A replacement parameter is recognized for combinations and permutations; this
indicates that any item may appear with multiplicity as high as the number of items in the
original set.
>>> from sympy.functions.combinatorial.numbers import nC, nP, nT
>>> items = baby
nC
Calculate the number of combinations of length k.
>>> [nC(items, k) for k in range(len(items) + 1)], nC(items)
([1, 3, 4, 3, 1], 12)
>>> nC(aaa, 2)
1
>>> nC(abc, 2)
3
>>> nC(3, 2)
3
nP
Calculate the number of permutations of length k.
>>> [nP(items, k) for k in range(len(items) + 1)], nP(items)
([1, 3, 7, 12, 12], 35)
>>> nC(aaa, 2)
1
>>> nC(abc, 2)
3
>>> nC(3, 2)
3
nT
Calculate the number of partitions that have k parts.
376
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> [nT(items, k) for k in range(len(items) + 1)], nT(items)
([0, 1, 5, 4, 1], 11)
>>> nT(aaa, 2)
1
>>> nT(abc, 2)
3
>>> nT(3, 2)
1
Note that the integer for n indicates identical items for nT but indicates n dierent items for
nC and nP.
Special
DiracDelta
class sympy.functions.special.delta functions.DiracDelta
The DiracDelta function and its derivatives.
DiracDelta function has the following properties:
1.diff(Heaviside(x),x) = DiracDelta(x)
2.integrate(DiracDelta(x-a)*f(x),(x,-oo,oo))
=
integrate(DiracDelta(x-a)*f(x),(x,a-e,a+e)) = f(a)
f(a)
and
3.DiracDelta(x) = 0 for all x != 0
4.DiracDelta(g(x)) = Sum i(DiracDelta(x-x i)/abs(g(x i))) Where x i-s are
the roots of g
Derivatives of k-th order of DiracDelta have the following property:
5.DiracDelta(x,k) = 0, for all x != 0
See Also:
Heaviside, simplify, is simple, sympy.functions.special.tensor functions.KroneckerDelta
(page 446)
References
[R118] (page 1904)
is simple(self, x)
Tells whether the argument(args[0]) of DiracDelta is a linear expression in x.
x can be:
a symbol
See Also:
simplify, Directdelta
Examples
>>> from sympy import DiracDelta, cos
>>> from sympy.abc import x, y
5.8. Functions Module
377
SymPy Documentation, Release 0.7.6
>>> DiracDelta(x*y).is_simple(x)
True
>>> DiracDelta(x*y).is_simple(y)
True
>>> DiracDelta(x**2+x-2).is_simple(x)
False
>>> DiracDelta(cos(x)).is_simple(x)
False
simplify(self, x)
Compute a simplied representation of the function using property number 4.
x can be:
a symbol
See Also:
is simple, Directdelta
Examples
>>> from sympy import DiracDelta
>>> from sympy.abc import x, y
>>> DiracDelta(x*y).simplify(x)
DiracDelta(x)/Abs(y)
>>> DiracDelta(x*y).simplify(y)
DiracDelta(y)/Abs(x)
>>> DiracDelta(x**2 + x - 2).simplify(x)
DiracDelta(x - 1)/3 + DiracDelta(x + 2)/3
Heaviside
class sympy.functions.special.delta functions.Heaviside
Heaviside Piecewise function
Heaviside function has the following properties 3 :
1.diff(Heaviside(x),x) = DiracDelta(x) ( 0, if x < 0
2.Heaviside(x) = < ( 1/2 if x==0 [*] ( 1, if x > 0
I think is better to have H(0) = 1/2, due to the following:
integrate(DiracDelta(x), x) = Heaviside(x)
integrate(DiracDelta(x), (x, -oo, oo)) = 1
and since DiracDelta is a symmetric function, integrate(DiracDelta(x), (x, 0, oo))
should be 1/2 (which is what Maple returns).
3
Regarding to the value at 0, Mathematica denes H(0) = 1, but Maple uses H(0) = undefined
378
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
If we take Heaviside(0) = 1/2, we would have integrate(DiracDelta(x), (x,
0, oo)) = Heaviside(oo) - Heaviside(0) = 1 - 1/2 = 1/2 and integrate(DiracDelta(x), (x, -oo, 0)) = Heaviside(0) - Heaviside(-oo) =
1/2 - 0 = 1/2
If we consider, instead Heaviside(0) = 1, we would have integrate(DiracDelta(x),
(x, 0, oo)) = Heaviside(oo) - Heaviside(0) = 0 and integrate(DiracDelta(x),
(x, -oo, 0)) = Heaviside(0) - Heaviside(-oo) = 1
See Also:
DiracDelta
References
[R119] (page 1904)
Gamma, Beta and related Functions
class sympy.functions.special.gamma functions.gamma
The gamma function
(x) :=
tx1 et dt.
0
The gamma function implements the function which passes through the values of the
factorial function, i.e. (n) = (n 1)! when n is an integer. More general, (z) is dened
in the whole complex plane except at the negative integers where there are simple poles.
See Also:
lowergamma (page 385) Lower incomplete gamma function.
uppergamma (page 384) Upper incomplete gamma function.
polygamma (page 382) Polygamma function.
loggamma (page 380) Log Gamma function.
digamma (page 383) Digamma function.
trigamma (page 384) Trigamma function.
sympy.functions.special.beta functions.beta (page 386) Euler Beta function.
References
[R120] (page 1904), [R121] (page 1904), [R122] (page 1904), [R123] (page 1904)
Examples
>>> from sympy import S, I, pi, oo, gamma
>>> from sympy.abc import x
Several special values are known:
5.8. Functions Module
379
SymPy Documentation, Release 0.7.6
>>> gamma(1)
1
>>> gamma(4)
6
>>> gamma(S(3)/2)
sqrt(pi)/2
The Gamma function obeys the mirror symmetry:
>>> from sympy import conjugate
>>> conjugate(gamma(x))
gamma(conjugate(x))
Dierentiation with respect to x is supported:
>>> from sympy import diff
>>> diff(gamma(x), x)
gamma(x)*polygamma(0, x)
Series expansion is also supported:
>>> from sympy import series
>>> series(gamma(x), x, 0, 3)
1/x - EulerGamma + x*(EulerGamma**2/2 + pi**2/12) + x**2*(-EulerGamma*pi**2/12 + polygamma(2, 1)
We can numerically evaluate the gamma function to arbitrary precision on the whole
complex plane:
>>> gamma(pi).evalf(40)
2.288037795340032417959588909060233922890
>>> gamma(1+I).evalf(20)
0.49801566811835604271 - 0.15494982830181068512*I
class sympy.functions.special.gamma functions.loggamma
The loggamma function implements the logarithm of the gamma function i.e, log (x).
See Also:
gamma (page 379) Gamma function.
lowergamma (page 385) Lower incomplete gamma function.
uppergamma (page 384) Upper incomplete gamma function.
polygamma (page 382) Polygamma function.
digamma (page 383) Digamma function.
trigamma (page 384) Trigamma function.
sympy.functions.special.beta functions.beta (page 386) Euler Beta function.
References
[R124] (page 1904), [R125] (page 1904), [R126] (page 1904), [R127] (page 1904)
Examples
Several special values are known. For numerical integral arguments we have:
380
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> from sympy import loggamma
>>> loggamma(-2)
oo
>>> loggamma(0)
oo
>>> loggamma(1)
0
>>> loggamma(2)
0
>>> loggamma(3)
log(2)
and for symbolic values:
>>> from sympy import Symbol
>>> n = Symbol(n, integer=True, positive=True)
>>> loggamma(n)
log(gamma(n))
>>> loggamma(-n)
oo
for half-integral values:
>>> from sympy import S, pi
>>> loggamma(S(5)/2)
log(3*sqrt(pi)/4)
>>> loggamma(n/2)
log(2**(-n + 1)*sqrt(pi)*gamma(n)/gamma(n/2 + 1/2))
and general rational arguments:
>>> from sympy import expand_func
>>> L = loggamma(S(16)/3)
>>> expand_func(L).doit()
-5*log(3) + loggamma(1/3) + log(4) + log(7) + log(10) + log(13)
>>> L = loggamma(S(19)/4)
>>> expand_func(L).doit()
-4*log(4) + loggamma(3/4) + log(3) + log(7) + log(11) + log(15)
>>> L = loggamma(S(23)/7)
>>> expand_func(L).doit()
-3*log(7) + log(2) + loggamma(2/7) + log(9) + log(16)
The loggamma function has the following limits towards innity:
>>> from sympy import oo
>>> loggamma(oo)
oo
>>> loggamma(-oo)
zoo
The loggamma function obeys the mirror symmetry if x C \ {, 0}:
>>> from sympy.abc import x
>>> from sympy import conjugate
>>> conjugate(loggamma(x))
loggamma(conjugate(x))
Dierentiation with respect to x is supported:
5.8. Functions Module
381
SymPy Documentation, Release 0.7.6
>>> from sympy import diff
>>> diff(loggamma(x), x)
polygamma(0, x)
Series expansion is also supported:
>>> from sympy import series
>>> series(loggamma(x), x, 0, 4)
-log(x) - EulerGamma*x + pi**2*x**2/12 + x**3*polygamma(2, 1)/6 + O(x**4)
We can numerically evaluate the gamma function to arbitrary precision on the whole
complex plane:
>>> from sympy import I
>>> loggamma(5).evalf(30)
3.17805383034794561964694160130
>>> loggamma(I).evalf(20)
-0.65092319930185633889 - 1.8724366472624298171*I
class sympy.functions.special.gamma functions.polygamma
The function polygamma(n, z) returns log(gamma(z)).diff(n + 1).
It is a meromorphic function on C and dened as the (n+1)-th derivative of the logarithm
of the gamma function:
n+1
(n) (z) :=
d
log (z).
dz n+1
See Also:
gamma (page 379) Gamma function.
lowergamma (page 385) Lower incomplete gamma function.
uppergamma (page 384) Upper incomplete gamma function.
loggamma (page 380) Log Gamma function.
digamma (page 383) Digamma function.
trigamma (page 384) Trigamma function.
sympy.functions.special.beta functions.beta (page 386) Euler Beta function.
References
[R128] (page 1904), [R129] (page 1904), [R130] (page 1904), [R131] (page 1904)
Examples
Several special values are known:
>>> from sympy import S, polygamma
>>> polygamma(0, 1)
-EulerGamma
>>> polygamma(0, 1/S(2))
-2*log(2) - EulerGamma
382
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> polygamma(0, 1/S(3))
-3*log(3)/2 - sqrt(3)*pi/6 - EulerGamma
>>> polygamma(0, 1/S(4))
-3*log(2) - pi/2 - EulerGamma
>>> polygamma(0, 2)
-EulerGamma + 1
>>> polygamma(0, 23)
-EulerGamma + 19093197/5173168
>>>
>>>
oo
>>>
oo
>>>
oo
>>>
oo
from sympy import oo, I
polygamma(0, oo)
polygamma(0, -oo)
polygamma(0, I*oo)
polygamma(0, -I*oo)
Dierentiation with respect to x is supported:
>>> from sympy import
>>> x = Symbol(x)
>>> diff(polygamma(0,
polygamma(1, x)
>>> diff(polygamma(0,
polygamma(2, x)
>>> diff(polygamma(0,
polygamma(3, x)
>>> diff(polygamma(1,
polygamma(2, x)
>>> diff(polygamma(1,
polygamma(3, x)
>>> diff(polygamma(2,
polygamma(3, x)
>>> diff(polygamma(2,
polygamma(4, x)
Symbol, diff
x), x)
x), x, 2)
x), x, 3)
x), x)
x), x, 2)
x), x)
x), x, 2)
>>> n = Symbol(n)
>>> diff(polygamma(n, x), x)
polygamma(n + 1, x)
>>> diff(polygamma(n, x), x, 2)
polygamma(n + 2, x)
We can rewrite polygamma functions in terms of harmonic numbers:
>>> from sympy import harmonic
>>> polygamma(0, x).rewrite(harmonic)
harmonic(x - 1) - EulerGamma
>>> polygamma(2, x).rewrite(harmonic)
2*harmonic(x - 1, 3) - 2*zeta(3)
>>> ni = Symbol(n, integer=True)
>>> polygamma(ni, x).rewrite(harmonic)
(-1)**(n + 1)*(-harmonic(x - 1, n + 1) + zeta(n + 1))*factorial(n)
sympy.functions.special.gamma functions.digamma(x)
5.8. Functions Module
383
SymPy Documentation, Release 0.7.6
The digamma function is the rst derivative of the loggamma function i.e,
(x) :=
d
0 (z)
log (z) =
dz
(z)
In this case, digamma(z) = polygamma(0, z).
See Also:
gamma (page 379) Gamma function.
lowergamma (page 385) Lower incomplete gamma function.
uppergamma (page 384) Upper incomplete gamma function.
polygamma (page 382) Polygamma function.
loggamma (page 380) Log Gamma function.
trigamma (page 384) Trigamma function.
sympy.functions.special.beta functions.beta (page 386) Euler Beta function.
References
[R132] (page 1905), [R133] (page 1905), [R134] (page 1905)
sympy.functions.special.gamma functions.trigamma(x)
The trigamma function is the second derivative of the loggamma function i.e,
2
(1) (z) :=
d
log (z).
dz 2
In this case, trigamma(z) = polygamma(1, z).
See Also:
gamma (page 379) Gamma function.
lowergamma (page 385) Lower incomplete gamma function.
uppergamma (page 384) Upper incomplete gamma function.
polygamma (page 382) Polygamma function.
loggamma (page 380) Log Gamma function.
digamma (page 383) Digamma function.
sympy.functions.special.beta functions.beta (page 386) Euler Beta function.
References
[R135] (page 1905), [R136] (page 1905), [R137] (page 1905)
class sympy.functions.special.gamma functions.uppergamma
The upper incomplete gamma function.
384
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
It can be dened as the meromorphic continuation of
(s, x) :=
ts1 et dt = (s) (s, x).
x
where (s, x) is the lower incomplete gamma function, lowergamma (page 385). This can
be shown to be the same as
)
(
xs
s
(s, x) = (s) 1 F1
x ,
s
s + 1
where 1 F1 is the (conuent) hypergeometric function.
The upper incomplete gamma function is also essentially equivalent to the generalized
exponential integral:
xt
e
En (x) =
dt = xn1 (1 n, x).
tn
1
See Also:
gamma (page 379) Gamma function.
lowergamma (page 385) Lower incomplete gamma function.
polygamma (page 382) Polygamma function.
loggamma (page 380) Log Gamma function.
digamma (page 383) Digamma function.
trigamma (page 384) Trigamma function.
sympy.functions.special.beta functions.beta (page 386) Euler Beta function.
References
[R138] (page 1905), [R139] (page 1905), [R140] (page 1905), [R141] (page 1905), [R142]
(page 1905), [R143] (page 1905)
Examples
>>> from sympy import uppergamma, S
>>> from sympy.abc import s, x
>>> uppergamma(s, x)
uppergamma(s, x)
>>> uppergamma(3, x)
x**2*exp(-x) + 2*x*exp(-x) + 2*exp(-x)
>>> uppergamma(-S(1)/2, x)
-2*sqrt(pi)*(-erf(sqrt(x)) + 1) + 2*exp(-x)/sqrt(x)
>>> uppergamma(-2, x)
expint(3, x)/x**2
5.8. Functions Module
385
SymPy Documentation, Release 0.7.6
class sympy.functions.special.gamma functions.lowergamma
The lower incomplete gamma function.
It can be dened as the meromorphic continuation of
x
(s, x) :=
ts1 et dt = (s) (s, x).
0
This can be shown to be the same as
(s, x) =
xs
1 F1
s
)
s
x ,
s + 1
where 1 F1 is the (conuent) hypergeometric function.
See Also:
gamma (page 379) Gamma function.
uppergamma (page 384) Upper incomplete gamma function.
polygamma (page 382) Polygamma function.
loggamma (page 380) Log Gamma function.
digamma (page 383) Digamma function.
trigamma (page 384) Trigamma function.
sympy.functions.special.beta functions.beta (page 386) Euler Beta function.
References
[R144] (page 1905), [R145] (page 1905), [R146] (page 1905), [R147] (page 1905), [R148]
(page 1905)
Examples
>>> from sympy import lowergamma, S
>>> from sympy.abc import s, x
>>> lowergamma(s, x)
lowergamma(s, x)
>>> lowergamma(3, x)
-x**2*exp(-x) - 2*x*exp(-x) + 2 - 2*exp(-x)
>>> lowergamma(-S(1)/2, x)
-2*sqrt(pi)*erf(sqrt(x)) - 2*exp(-x)/sqrt(x)
class sympy.functions.special.beta functions.beta
The beta integral is called the Eulerian integral of the rst kind by Legendre:
tx1 (1 t)y1 dt.
B(x, y) :=
0
386
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Beta function or Eulers rst integral is closely associated with gamma function. The
Beta function often used in probability theory and mathematical statistics. It satises
properties like:
1
a
B(a, b) = B(b, a)
B(a, 1) =
B(a, b) =
(a)(b)
(a + b)
Therefore for integral values of a and b:
B=
(a 1)!(b 1)!
(a + b 1)!
See Also:
sympy.functions.special.gamma functions.gamma (page 379) Gamma function.
sympy.functions.special.gamma functions.uppergamma (page 384) Upper
plete gamma function.
incom-
sympy.functions.special.gamma functions.lowergamma (page 385) Lower
plete gamma function.
incom-
sympy.functions.special.gamma functions.polygamma (page 382) Polygamma
function.
sympy.functions.special.gamma functions.loggamma (page 380) Log Gamma function.
sympy.functions.special.gamma functions.digamma (page 383) Digamma
tion.
func-
sympy.functions.special.gamma functions.trigamma (page 384) Trigamma
tion.
func-
References
[R149] (page 1905), [R150] (page 1905), [R151] (page 1905)
Examples
>>> from sympy import I, pi
>>> from sympy.abc import x,y
The Beta function obeys the mirror symmetry:
>>> from sympy import beta
>>> from sympy import conjugate
>>> conjugate(beta(x,y))
beta(conjugate(x), conjugate(y))
Dierentiation with respect to both x and y is supported:
5.8. Functions Module
387
SymPy Documentation, Release 0.7.6
>>> from sympy import beta
>>> from sympy import diff
>>> diff(beta(x,y), x)
(polygamma(0, x) - polygamma(0, x + y))*beta(x, y)
>>> from sympy import beta
>>> from sympy import diff
>>> diff(beta(x,y), y)
(polygamma(0, y) - polygamma(0, x + y))*beta(x, y)
We can numerically evaluate the gamma function to arbitrary precision on the whole
complex plane:
>>> from sympy import beta
>>> beta(pi,pi).evalf(40)
0.02671848900111377452242355235388489324562
>>> beta(1+I,1+I).evalf(20)
-0.2112723729365330143 - 0.7655283165378005676*I
Error Functions and Fresnel Integrals
class sympy.functions.special.error functions.erf
The Gauss error function. This function is dened as:
x
2
2
erf(x) =
et dt.
0
See Also:
erfc (page 389) Complementary error function.
erfi (page 390) Imaginary error function.
erf2 (page 391) Two-argument error function.
erfinv (page 392) Inverse error function.
erfcinv (page 393) Inverse Complementary error function.
erf2inv (page 394) Inverse two-argument error function.
References
[R152] (page 1905), [R153] (page 1905), [R154] (page 1905), [R155] (page 1905)
Examples
>>> from sympy import I, oo, erf
>>> from sympy.abc import z
Several special values are known:
388
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> erf(0)
0
>>> erf(oo)
1
>>> erf(-oo)
-1
>>> erf(I*oo)
oo*I
>>> erf(-I*oo)
-oo*I
In general one can pull out factors of -1 and I from the argument:
>>> erf(-z)
-erf(z)
The error function obeys the mirror symmetry:
>>> from sympy import conjugate
>>> conjugate(erf(z))
erf(conjugate(z))
Dierentiation with respect to z is supported:
>>> from sympy import diff
>>> diff(erf(z), z)
2*exp(-z**2)/sqrt(pi)
We can numerically evaluate the error function to arbitrary precision on the whole complex plane:
>>> erf(4).evalf(30)
0.999999984582742099719981147840
>>> erf(-4*I).evalf(30)
-1296959.73071763923152794095062*I
class sympy.functions.special.error functions.erfc
Complementary Error Function. The function is dened as:
2
2
erfc(x) =
et dt
x
See Also:
erf (page 388) Gaussian error function.
erfi (page 390) Imaginary error function.
erf2 (page 391) Two-argument error function.
erfinv (page 392) Inverse error function.
erfcinv (page 393) Inverse Complementary error function.
erf2inv (page 394) Inverse two-argument error function.
References
[R156] (page 1905), [R157] (page 1905), [R158] (page 1905), [R159] (page 1905)
5.8. Functions Module
389
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import I, oo, erfc
>>> from sympy.abc import z
Several special values are known:
>>> erfc(0)
1
>>> erfc(oo)
0
>>> erfc(-oo)
2
>>> erfc(I*oo)
-oo*I
>>> erfc(-I*oo)
oo*I
The error function obeys the mirror symmetry:
>>> from sympy import conjugate
>>> conjugate(erfc(z))
erfc(conjugate(z))
Dierentiation with respect to z is supported:
>>> from sympy import diff
>>> diff(erfc(z), z)
-2*exp(-z**2)/sqrt(pi)
It also follows
>>> erfc(-z)
-erfc(z) + 2
We can numerically evaluate the complementary error function to arbitrary precision on
the whole complex plane:
>>> erfc(4).evalf(30)
0.0000000154172579002800188521596734869
>>> erfc(4*I).evalf(30)
1.0 - 1296959.73071763923152794095062*I
class sympy.functions.special.error functions.erfi
Imaginary error function. The function er is dened as:
x
2
2
erfi(x) =
et dt
0
See Also:
erf (page 388) Gaussian error function.
erfc (page 389) Complementary error function.
erf2 (page 391) Two-argument error function.
erfinv (page 392) Inverse error function.
390
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
erfcinv (page 393) Inverse Complementary error function.
erf2inv (page 394) Inverse two-argument error function.
References
[R160] (page 1905), [R161] (page 1905), [R162] (page 1905)
Examples
>>> from sympy import I, oo, erfi
>>> from sympy.abc import z
Several special values are known:
>>>
0
>>>
oo
>>>
-oo
>>>
I
>>>
-I
erfi(0)
erfi(oo)
erfi(-oo)
erfi(I*oo)
erfi(-I*oo)
In general one can pull out factors of -1 and I from the argument:
>>> erfi(-z)
-erfi(z)
>>> from sympy import conjugate
>>> conjugate(erfi(z))
erfi(conjugate(z))
Dierentiation with respect to z is supported:
>>> from sympy import diff
>>> diff(erfi(z), z)
2*exp(z**2)/sqrt(pi)
We can numerically evaluate the imaginary error function to arbitrary precision on the
whole complex plane:
>>> erfi(2).evalf(30)
18.5648024145755525987042919132
>>> erfi(-2*I).evalf(30)
-0.995322265018952734162069256367*I
class sympy.functions.special.error functions.erf2
Two-argument error function. This function is dened as:
y
2
2
et dt
erf2(x, y) =
x
5.8. Functions Module
391
SymPy Documentation, Release 0.7.6
See Also:
erf (page 388) Gaussian error function.
erfc (page 389) Complementary error function.
erfi (page 390) Imaginary error function.
erfinv (page 392) Inverse error function.
erfcinv (page 393) Inverse Complementary error function.
erf2inv (page 394) Inverse two-argument error function.
References
[R163] (page 1905)
Examples
>>> from sympy import I, oo, erf2
>>> from sympy.abc import x, y
Several special values are known:
>>> erf2(0, 0)
0
>>> erf2(x, x)
0
>>> erf2(x, oo)
-erf(x) + 1
>>> erf2(x, -oo)
-erf(x) - 1
>>> erf2(oo, y)
erf(y) - 1
>>> erf2(-oo, y)
erf(y) + 1
In general one can pull out factors of -1:
>>> erf2(-x, -y)
-erf2(x, y)
The error function obeys the mirror symmetry:
>>> from sympy import conjugate
>>> conjugate(erf2(x, y))
erf2(conjugate(x), conjugate(y))
Dierentiation with respect to x, y is supported:
>>> from sympy import diff
>>> diff(erf2(x, y), x)
-2*exp(-x**2)/sqrt(pi)
>>> diff(erf2(x, y), y)
2*exp(-y**2)/sqrt(pi)
392
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
class sympy.functions.special.error functions.erfinv
Inverse Error Function. The ernv function is dened as:
erf(x) = y
erfinv(y) = x
See Also:
erf (page 388) Gaussian error function.
erfc (page 389) Complementary error function.
erfi (page 390) Imaginary error function.
erf2 (page 391) Two-argument error function.
erfcinv (page 393) Inverse Complementary error function.
erf2inv (page 394) Inverse two-argument error function.
References
[R164] (page 1905), [R165] (page 1905)
Examples
>>> from sympy import I, oo, erfinv
>>> from sympy.abc import x
Several special values are known:
>>> erfinv(0)
0
>>> erfinv(1)
oo
Dierentiation with respect to x is supported:
>>> from sympy import diff
>>> diff(erfinv(x), x)
sqrt(pi)*exp(erfinv(x)**2)/2
We can numerically evaluate the inverse error function to arbitrary precision on [-1, 1]:
>>> erfinv(0.2).evalf(30)
0.179143454621291692285822705344
class sympy.functions.special.error functions.erfcinv
Inverse Complementary Error Function. The erfcinv function is dened as:
erfc(x) = y
erfcinv(y) = x
See Also:
erf (page 388) Gaussian error function.
erfc (page 389) Complementary error function.
5.8. Functions Module
393
SymPy Documentation, Release 0.7.6
erfi (page 390) Imaginary error function.
erf2 (page 391) Two-argument error function.
erfinv (page 392) Inverse error function.
erf2inv (page 394) Inverse two-argument error function.
References
[R166] (page 1905), [R167] (page 1905)
Examples
>>> from sympy import I, oo, erfcinv
>>> from sympy.abc import x
Several special values are known:
>>> erfcinv(1)
0
>>> erfcinv(0)
oo
Dierentiation with respect to x is supported:
>>> from sympy import diff
>>> diff(erfcinv(x), x)
-sqrt(pi)*exp(erfcinv(x)**2)/2
class sympy.functions.special.error functions.erf2inv
Two-argument Inverse error function. The erf2inv function is dened as:
erf2(x, w) = y
erf2inv(x, y) = w
See Also:
erf (page 388) Gaussian error function.
erfc (page 389) Complementary error function.
erfi (page 390) Imaginary error function.
erf2 (page 391) Two-argument error function.
erfinv (page 392) Inverse error function.
erfcinv (page 393) Inverse complementary error function.
References
[R168] (page 1905)
394
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import I, oo, erf2inv, erfinv, erfcinv
>>> from sympy.abc import x, y
Several special values are known:
>>> erf2inv(0, 0)
0
>>> erf2inv(1, 0)
1
>>> erf2inv(0, 1)
oo
>>> erf2inv(0, y)
erfinv(y)
>>> erf2inv(oo, y)
erfcinv(-y)
Dierentiation with respect to x and y is supported:
>>> from sympy import diff
>>> diff(erf2inv(x, y), x)
exp(-x**2 + erf2inv(x, y)**2)
>>> diff(erf2inv(x, y), y)
sqrt(pi)*exp(erf2inv(x, y)**2)/2
class sympy.functions.special.error functions.FresnelIntegral
Base class for the Fresnel integrals.
class sympy.functions.special.error functions.fresnels
Fresnel integral S.
This function is dened by
S(z) =
sin
0
2
t dt.
2
It is an entire function.
See Also:
fresnelc (page 396) Fresnel cosine integral.
References
[R169] (page 1906), [R170] (page 1906), [R171] (page 1906), [R172] (page 1906), [R173]
(page 1906)
Examples
>>> from sympy import I, oo, fresnels
>>> from sympy.abc import z
Several special values are known:
5.8. Functions Module
395
SymPy Documentation, Release 0.7.6
>>> fresnels(0)
0
>>> fresnels(oo)
1/2
>>> fresnels(-oo)
-1/2
>>> fresnels(I*oo)
-I/2
>>> fresnels(-I*oo)
I/2
In general one can pull out factors of -1 and i from the argument:
>>> fresnels(-z)
-fresnels(z)
>>> fresnels(I*z)
-I*fresnels(z)
The Fresnel S integral obeys the mirror symmetry S(z) = S(
z ):
>>> from sympy import conjugate
>>> conjugate(fresnels(z))
fresnels(conjugate(z))
Dierentiation with respect to z is supported:
>>> from sympy import diff
>>> diff(fresnels(z), z)
sin(pi*z**2/2)
Dening the Fresnel functions via an integral
>>> from sympy import integrate, pi, sin, gamma, expand_func
>>> integrate(sin(pi*z**2/2), z)
3*fresnels(z)*gamma(3/4)/(4*gamma(7/4))
>>> expand_func(integrate(sin(pi*z**2/2), z))
fresnels(z)
We can numerically evaluate the Fresnel integral to arbitrary precision on the whole
complex plane:
>>> fresnels(2).evalf(30)
0.343415678363698242195300815958
>>> fresnels(-2*I).evalf(30)
0.343415678363698242195300815958*I
class sympy.functions.special.error functions.fresnelc
Fresnel integral C.
This function is dened by
C(z) =
cos
0
2
t dt.
2
It is an entire function.
See Also:
fresnels (page 395) Fresnel sine integral.
396
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
References
[R174] (page 1906), [R175] (page 1906), [R176] (page 1906), [R177] (page 1906), [R178]
(page 1906)
Examples
>>> from sympy import I, oo, fresnelc
>>> from sympy.abc import z
Several special values are known:
>>> fresnelc(0)
0
>>> fresnelc(oo)
1/2
>>> fresnelc(-oo)
-1/2
>>> fresnelc(I*oo)
I/2
>>> fresnelc(-I*oo)
-I/2
In general one can pull out factors of -1 and i from the argument:
>>> fresnelc(-z)
-fresnelc(z)
>>> fresnelc(I*z)
I*fresnelc(z)
The Fresnel C integral obeys the mirror symmetry C(z) = C(
z ):
>>> from sympy import conjugate
>>> conjugate(fresnelc(z))
fresnelc(conjugate(z))
Dierentiation with respect to z is supported:
>>> from sympy import diff
>>> diff(fresnelc(z), z)
cos(pi*z**2/2)
Dening the Fresnel functions via an integral
>>> from sympy import integrate, pi, cos, gamma, expand_func
>>> integrate(cos(pi*z**2/2), z)
fresnelc(z)*gamma(1/4)/(4*gamma(5/4))
>>> expand_func(integrate(cos(pi*z**2/2), z))
fresnelc(z)
We can numerically evaluate the Fresnel integral to arbitrary precision on the whole
complex plane:
>>> fresnelc(2).evalf(30)
0.488253406075340754500223503357
>>> fresnelc(-2*I).evalf(30)
-0.488253406075340754500223503357*I
5.8. Functions Module
397
SymPy Documentation, Release 0.7.6
Exponential, Logarithmic and Trigonometric Integrals
class sympy.functions.special.error functions.Ei
The classical exponential integral.
For use in SymPy, this function is dened as
Ei(x) =
xn
+ log(x) + ,
n n!
n=1
where is the Euler-Mascheroni constant.
If x is a polar number, this denes an analytic function on the Riemann surface of the
logarithm. Otherwise this denes an analytic function in the cut plane C \ (, 0].
Background
The name exponential integral comes from the following statement:
x t
e
Ei(x) =
dt
t
If the integral is interpreted as a Cauchy principal value, this statement holds for x > 0
and Ei(x) as dened above.
Note that we carefully avoided dening Ei(x) for negative real x. This is because above
integral formula does not hold for any polar lift of such x, indeed all branches of Ei(x)
above the negative reals are imaginary.
However, the following statement holds for all x R :
(
)
(
)
x t
Ei |x|ei arg(x) + Ei |x|ei arg(x)
e
dt =
,
2
t
where the integral is again understood to be a principal value if x > 0, and |x|ei arg(x) ,
|x|ei arg(x) denote two conjugate polar lifts of x.
See Also:
expint (page 399) Generalised exponential integral.
E1 (page 401) Special case of the generalised exponential integral.
li (page 401) Logarithmic integral.
Li (page 403) Oset logarithmic integral.
Si (page 404) Sine integral.
Ci (page 405) Cosine integral.
Shi (page 406) Hyperbolic sine integral.
Chi (page 407) Hyperbolic cosine integral.
sympy.functions.special.gamma functions.uppergamma (page 384) Upper
plete gamma function.
398
incom-
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
References
[R179] (page 1906), [R180] (page 1906), [R181] (page 1906)
Examples
>>> from sympy import Ei, polar_lift, exp_polar, I, pi
>>> from sympy.abc import x
The exponential integral in SymPy is strictly undened for negative values of the argument. For convenience, exponential integrals with negative arguments are immediately
converted into an expression that agrees with the classical integral denition:
>>> Ei(-1)
-I*pi + Ei(exp_polar(I*pi))
This yields a real value:
>>> Ei(-1).n(chop=True)
-0.219383934395520
On the other hand the analytic continuation is not real:
>>> Ei(polar_lift(-1)).n(chop=True)
-0.21938393439552 + 3.14159265358979*I
The exponential integral has a logarithmic branch point at the origin:
>>> Ei(x*exp_polar(2*I*pi))
Ei(x) + 2*I*pi
Dierentiation is supported:
>>> Ei(x).diff(x)
exp(x)/x
The exponential integral is related to many other special functions. For example:
>>> from sympy import uppergamma, expint, Shi
>>> Ei(x).rewrite(expint)
-expint(1, x*exp_polar(I*pi)) - I*pi
>>> Ei(x).rewrite(Shi)
Chi(x) + Shi(x)
class sympy.functions.special.error functions.expint
Generalized exponential integral.
This function is dened as
E (z) = z 1 (1 , z),
where (1 , z) is the upper incomplete gamma function (uppergamma).
Hence for z with positive real part we have
E (z) =
1
5.8. Functions Module
ezt
dt,
z
399
SymPy Documentation, Release 0.7.6
which explains the name.
The representation as an incomplete gamma function provides an analytic continuation
for E (z). If is a non-positive integer the exponential integral is thus an unbranched
function of z, otherwise there is a branch point at the origin. Refer to the incomplete
gamma function documentation for details of the branching behavior.
See Also:
Ei (page 398) Another related function called exponential integral.
E1 (page 401) The classical case, returns expint(1, z).
li (page 401) Logarithmic integral.
Li (page 403) Oset logarithmic integral.
Si (page 404) Sine integral.
Ci (page 405) Cosine integral.
Shi (page 406) Hyperbolic sine integral.
Chi (page 407) Hyperbolic cosine integral.
sympy.functions.special.gamma functions.uppergamma (page 384)
References
[R182] (page 1906), [R183] (page 1906), [R184] (page 1906)
Examples
>>> from sympy import expint, S
>>> from sympy.abc import nu, z
Dierentiation is supported. Dierentiation with respect to z explains further the name:
for integral orders, the exponential integral is an iterated integral of the exponential
function.
>>> expint(nu, z).diff(z)
-expint(nu - 1, z)
Dierentiation with respect to nu has no classical expression:
>>> expint(nu, z).diff(nu)
-z**(nu - 1)*meijerg(((), (1, 1)), ((0, 0, -nu + 1), ()), z)
At non-postive integer orders, the exponential integral reduces to the exponential function:
>>> expint(0, z)
exp(-z)/z
>>> expint(-1, z)
exp(-z)/z + exp(-z)/z**2
At half-integers it reduces to error functions:
>>> expint(S(1)/2, z)
-sqrt(pi)*erf(sqrt(z))/sqrt(z) + sqrt(pi)/sqrt(z)
400
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
At positive integer orders it can be rewritten in terms of exponentials and expint(1, z).
Use expand func() to do this:
>>> from sympy import expand_func
>>> expand_func(expint(5, z))
z**4*expint(1, z)/24 + (-z**3 + z**2 - 2*z + 6)*exp(-z)/24
The generalised exponential integral is essentially equivalent to the incomplete gamma
function:
>>> from sympy import uppergamma
>>> expint(nu, z).rewrite(uppergamma)
z**(nu - 1)*uppergamma(-nu + 1, z)
As such it is branched at the origin:
>>> from sympy import exp_polar, pi, I
>>> expint(4, z*exp_polar(2*pi*I))
I*pi*z**3/3 + expint(4, z)
>>> expint(nu, z*exp_polar(2*pi*I))
z**(nu - 1)*(exp(2*I*pi*nu) - 1)*gamma(-nu + 1) + expint(nu, z)
sympy.functions.special.error functions.E1(z)
Classical case of the generalized exponential integral.
This is equivalent to expint(1, z).
See Also:
Ei (page 398) Exponential integral.
expint (page 399) Generalised exponential integral.
li (page 401) Logarithmic integral.
Li (page 403) Oset logarithmic integral.
Si (page 404) Sine integral.
Ci (page 405) Cosine integral.
Shi (page 406) Hyperbolic sine integral.
Chi (page 407) Hyperbolic cosine integral.
class sympy.functions.special.error functions.li
The classical logarithmic integral.
For the use in SymPy, this function is dened as
x
1
li(x) =
dt .
0 log(t)
See Also:
Li (page 403) Oset logarithmic integral.
Ei (page 398) Exponential integral.
expint (page 399) Generalised exponential integral.
E1 (page 401) Special case of the generalised exponential integral.
Si (page 404) Sine integral.
5.8. Functions Module
401
SymPy Documentation, Release 0.7.6
Ci (page 405) Cosine integral.
Shi (page 406) Hyperbolic sine integral.
Chi (page 407) Hyperbolic cosine integral.
References
[R185] (page 1906), [R186] (page 1906), [R187] (page 1906), [R188] (page 1906)
Examples
>>> from sympy import I, oo, li
>>> from sympy.abc import z
Several special values are known:
>>> li(0)
0
>>> li(1)
-oo
>>> li(oo)
oo
Dierentiation with respect to z is supported:
>>> from sympy import diff
>>> diff(li(z), z)
1/log(z)
Dening the li function via an integral:
The logarithmic integral can also be dened in terms of Ei:
>>> from sympy import Ei
>>> li(z).rewrite(Ei)
Ei(log(z))
>>> diff(li(z).rewrite(Ei), z)
1/log(z)
We can numerically evaluate the logarithmic integral to arbitrary precision on the whole
complex plane (except the singular points):
>>> li(2).evalf(30)
1.04516378011749278484458888919
>>> li(2*I).evalf(30)
1.0652795784357498247001125598 + 3.08346052231061726610939702133*I
We can even compute Soldners constant by the help of mpmath:
>>> from sympy.mpmath import findroot
>>> findroot(li, 2)
1.45136923488338
Further transformations include rewriting li in terms of the trigonometric integrals Si,
Ci, Shi and Chi:
402
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> from sympy import Si, Ci, Shi,
>>> li(z).rewrite(Si)
-log(I*log(z)) - log(1/log(z))/2 +
>>> li(z).rewrite(Ci)
-log(I*log(z)) - log(1/log(z))/2 +
>>> li(z).rewrite(Shi)
-log(1/log(z))/2 + log(log(z))/2 +
>>> li(z).rewrite(Chi)
-log(1/log(z))/2 + log(log(z))/2 +
Chi
log(log(z))/2 + Ci(I*log(z)) + Shi(log(z))
log(log(z))/2 + Ci(I*log(z)) + Shi(log(z))
Chi(log(z)) - Shi(log(z))
Chi(log(z)) - Shi(log(z))
class sympy.functions.special.error functions.Li
The oset logarithmic integral.
For the use in SymPy, this function is dened as
Li(x) = li(x) li(2)
See Also:
li (page 401) Logarithmic integral.
Ei (page 398) Exponential integral.
expint (page 399) Generalised exponential integral.
E1 (page 401) Special case of the generalised exponential integral.
Si (page 404) Sine integral.
Ci (page 405) Cosine integral.
Shi (page 406) Hyperbolic sine integral.
Chi (page 407) Hyperbolic cosine integral.
References
[R189] (page 1906), [R190] (page 1906), [R191] (page 1906)
Examples
>>> from sympy import I, oo, Li
>>> from sympy.abc import z
The following special value is known:
>>> Li(2)
0
Dierentiation with respect to z is supported:
>>> from sympy import diff
>>> diff(Li(z), z)
1/log(z)
The shifted logarithmic integral can be written in terms of li(z):
5.8. Functions Module
403
SymPy Documentation, Release 0.7.6
>>> from sympy import li
>>> Li(z).rewrite(li)
li(z) - li(2)
We can numerically evaluate the logarithmic integral to arbitrary precision on the whole
complex plane (except the singular points):
>>> Li(2).evalf(30)
0
>>> Li(4).evalf(30)
1.92242131492155809316615998938
class sympy.functions.special.error functions.Si
Sine integral.
This function is dened by
Si(z) =
0
sin t
dt.
t
It is an entire function.
See Also:
Ci (page 405) Cosine integral.
Shi (page 406) Hyperbolic sine integral.
Chi (page 407) Hyperbolic cosine integral.
Ei (page 398) Exponential integral.
expint (page 399) Generalised exponential integral.
E1 (page 401) Special case of the generalised exponential integral.
li (page 401) Logarithmic integral.
Li (page 403) Oset logarithmic integral.
References
[R192] (page 1906)
Examples
>>> from sympy import Si
>>> from sympy.abc import z
The sine integral is an antiderivative of sin(z)/z:
>>> Si(z).diff(z)
sin(z)/z
It is unbranched:
404
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> from sympy import exp_polar, I, pi
>>> Si(z*exp_polar(2*I*pi))
Si(z)
Sine integral behaves much like ordinary sine under multiplication by I:
>>> Si(I*z)
I*Shi(z)
>>> Si(-z)
-Si(z)
It can also be expressed in terms of exponential integrals, but beware that the latter is
branched:
>>> from sympy import expint
>>> Si(z).rewrite(expint)
-I*(-expint(1, z*exp_polar(-I*pi/2))/2 +
expint(1, z*exp_polar(I*pi/2))/2) + pi/2
class sympy.functions.special.error functions.Ci
Cosine integral.
This function is dened for positive x by
x
cos t 1
cos t
Ci(x) = + log x +
dt =
dt,
t
t
0
x
where is the Euler-Mascheroni constant.
We have
(
)
(
)
E1 ei/2 z + E1 ei/2 z
Ci(z) =
2
which holds for all polar z and thus provides an analytic continuation to the Riemann
surface of the logarithm.
The formula also holds as stated for z C with <(z) > 0. By lifting to the principal branch
we obtain an analytic function on the cut complex plane.
See Also:
Si (page 404) Sine integral.
Shi (page 406) Hyperbolic sine integral.
Chi (page 407) Hyperbolic cosine integral.
Ei (page 398) Exponential integral.
expint (page 399) Generalised exponential integral.
E1 (page 401) Special case of the generalised exponential integral.
li (page 401) Logarithmic integral.
Li (page 403) Oset logarithmic integral.
References
[R193] (page 1906)
5.8. Functions Module
405
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import Ci
>>> from sympy.abc import z
The cosine integral is a primitive of cos(z)/z:
>>> Ci(z).diff(z)
cos(z)/z
It has a logarithmic branch point at the origin:
>>> from sympy import exp_polar, I, pi
>>> Ci(z*exp_polar(2*I*pi))
Ci(z) + 2*I*pi
The cosine integral behaves somewhat like ordinary cos under multiplication by i:
>>> from sympy import polar_lift
>>> Ci(polar_lift(I)*z)
Chi(z) + I*pi/2
>>> Ci(polar_lift(-1)*z)
Ci(z) + I*pi
It can also be expressed in terms of exponential integrals:
>>> from sympy import expint
>>> Ci(z).rewrite(expint)
-expint(1, z*exp_polar(-I*pi/2))/2 - expint(1, z*exp_polar(I*pi/2))/2
class sympy.functions.special.error functions.Shi
Sinh integral.
This function is dened by
Shi(z) =
0
sinh t
dt.
t
It is an entire function.
See Also:
Si (page 404) Sine integral.
Ci (page 405) Cosine integral.
Chi (page 407) Hyperbolic cosine integral.
Ei (page 398) Exponential integral.
expint (page 399) Generalised exponential integral.
E1 (page 401) Special case of the generalised exponential integral.
li (page 401) Logarithmic integral.
Li (page 403) Oset logarithmic integral.
References
[R194] (page 1906)
406
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import Shi
>>> from sympy.abc import z
The Sinh integral is a primitive of sinh(z)/z:
>>> Shi(z).diff(z)
sinh(z)/z
It is unbranched:
>>> from sympy import exp_polar, I, pi
>>> Shi(z*exp_polar(2*I*pi))
Shi(z)
The sinh integral behaves much like ordinary sinh under multiplication by i:
>>> Shi(I*z)
I*Si(z)
>>> Shi(-z)
-Shi(z)
It can also be expressed in terms of exponential integrals, but beware that the latter is
branched:
>>> from sympy import expint
>>> Shi(z).rewrite(expint)
expint(1, z)/2 - expint(1, z*exp_polar(I*pi))/2 - I*pi/2
class sympy.functions.special.error functions.Chi
Cosh integral.
This function is dened for positive x by
Chi(x) = + log x +
0
cosh t 1
dt,
t
where is the Euler-Mascheroni constant.
We have
(
)
Chi(z) = Ci ei/2 z i ,
2
which holds for all polar z and thus provides an analytic continuation to the Riemann
surface of the logarithm. By lifting to the principal branch we obtain an analytic function
on the cut complex plane.
See Also:
Si (page 404) Sine integral.
Ci (page 405) Cosine integral.
Shi (page 406) Hyperbolic sine integral.
Ei (page 398) Exponential integral.
expint (page 399) Generalised exponential integral.
5.8. Functions Module
407
SymPy Documentation, Release 0.7.6
E1 (page 401) Special case of the generalised exponential integral.
li (page 401) Logarithmic integral.
Li (page 403) Oset logarithmic integral.
References
[R195] (page 1906)
Examples
>>> from sympy import Chi
>>> from sympy.abc import z
The cosh integral is a primitive of cosh(z)/z:
>>> Chi(z).diff(z)
cosh(z)/z
It has a logarithmic branch point at the origin:
>>> from sympy import exp_polar, I, pi
>>> Chi(z*exp_polar(2*I*pi))
Chi(z) + 2*I*pi
The cosh integral behaves somewhat like ordinary cosh under multiplication by i:
>>> from sympy import polar_lift
>>> Chi(polar_lift(I)*z)
Ci(z) + I*pi/2
>>> Chi(polar_lift(-1)*z)
Chi(z) + I*pi
It can also be expressed in terms of exponential integrals:
>>> from sympy import expint
>>> Chi(z).rewrite(expint)
-expint(1, z)/2 - expint(1, z*exp_polar(I*pi))/2 - I*pi/2
Bessel Type Functions
class sympy.functions.special.bessel.BesselBase
Abstract base class for bessel-type functions.
This class is meant to reduce code duplication. All Bessel type functions can 1) be dierentiated, and the derivatives expressed in terms of similar functions and 2) be rewritten
in terms of other bessel-type functions.
Here bessel-type functions are assumed to have one complex parameter.
To use this base class, dene class attributes a and b such that 2*F n = - a*F {n+1}
+ b*F {n-1}.
argument
The argument of the bessel-type function.
408
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
order
The order of the bessel-type function.
class sympy.functions.special.bessel.besselj
Bessel function of the rst kind.
The Bessel J function of order is dened to be the function satisfying Bessels dierential equation
2
z2
d w
dw
+ (z 2 2 )w = 0,
+z
dz 2
dz
with Laurent expansion
(
J (z) = z
)
1
2
+ O(z ) ,
( + 1)2
if is not a negative integer. If = n Z<0 is a negative integer, then the denition is
Jn (z) = (1)n Jn (z).
See Also:
bessely (page 409), besseli (page 410), besselk (page 411)
References
[R196] (page 1906), [R197] (page 1906), [R198] (page 1906), [R199] (page 1906)
Examples
Create a Bessel function object:
>>> from sympy import besselj, jn
>>> from sympy.abc import z, n
>>> b = besselj(n, z)
Dierentiate it:
>>> b.diff(z)
besselj(n - 1, z)/2 - besselj(n + 1, z)/2
Rewrite in terms of spherical Bessel functions:
>>> b.rewrite(jn)
sqrt(2)*sqrt(z)*jn(n - 1/2, z)/sqrt(pi)
Access the parameter and argument:
>>> b.order
n
>>> b.argument
z
5.8. Functions Module
409
SymPy Documentation, Release 0.7.6
class sympy.functions.special.bessel.bessely
Bessel function of the second kind.
The Bessel Y function of order is dened as
Y (z) = lim
J (z) cos() J (z)
,
sin()
where J (z) is the Bessel function of the rst kind.
It is a solution to Bessels equation, and linearly independent from J .
See Also:
besselj (page 409), besseli (page 410), besselk (page 411)
References
[R200] (page 1906)
Examples
>>> from sympy import bessely, yn
>>> from sympy.abc import z, n
>>> b = bessely(n, z)
>>> b.diff(z)
bessely(n - 1, z)/2 - bessely(n + 1, z)/2
>>> b.rewrite(yn)
sqrt(2)*sqrt(z)*yn(n - 1/2, z)/sqrt(pi)
class sympy.functions.special.bessel.besseli
Modied Bessel function of the rst kind.
The Bessel I function is a solution to the modied Bessel equation
2
z2
d w
dw
+z
+ (z 2 + 2 )2 w = 0.
dz 2
dz
It can be dened as
I (z) = i J (iz),
where J (z) is the Bessel function of the rst kind.
See Also:
besselj (page 409), bessely (page 409), besselk (page 411)
References
[R201] (page 1906)
410
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import besseli
>>> from sympy.abc import z, n
>>> besseli(n, z).diff(z)
besseli(n - 1, z)/2 + besseli(n + 1, z)/2
class sympy.functions.special.bessel.besselk
Modied Bessel function of the second kind.
The Bessel K function of order is dened as
K (z) = lim
I (z) I (z)
,
2
sin()
where I (z) is the modied Bessel function of the rst kind.
It is a solution of the modied Bessel equation, and linearly independent from Y .
See Also:
besselj (page 409), besseli (page 410), bessely (page 409)
References
[R202] (page 1906)
Examples
>>> from sympy import besselk
>>> from sympy.abc import z, n
>>> besselk(n, z).diff(z)
-besselk(n - 1, z)/2 - besselk(n + 1, z)/2
class sympy.functions.special.bessel.hankel1
Hankel function of the rst kind.
This function is dened as
H(1) = J (z) + iY (z),
where J (z) is the Bessel function of the rst kind, and Y (z) is the Bessel function of the
second kind.
It is a solution to Bessels equation.
See Also:
hankel2 (page 412), besselj (page 409), bessely (page 409)
References
[R203] (page 1906)
5.8. Functions Module
411
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import hankel1
>>> from sympy.abc import z, n
>>> hankel1(n, z).diff(z)
hankel1(n - 1, z)/2 - hankel1(n + 1, z)/2
class sympy.functions.special.bessel.hankel2
Hankel function of the second kind.
This function is dened as
H(2) = J (z) iY (z),
where J (z) is the Bessel function of the rst kind, and Y (z) is the Bessel function of the
second kind.
(1)
It is a solution to Bessels equation, and linearly independent from H .
See Also:
hankel1 (page 411), besselj (page 409), bessely (page 409)
References
[R204] (page 1906)
Examples
>>> from sympy import hankel2
>>> from sympy.abc import z, n
>>> hankel2(n, z).diff(z)
hankel2(n - 1, z)/2 - hankel2(n + 1, z)/2
class sympy.functions.special.bessel.jn
Spherical Bessel function of the rst kind.
This function is a solution to the spherical Bessel equation
2
z2
d w
dw
+ 2z
+ (z 2 ( + 1))w = 0.
dz 2
dz
It can be dened as
j (z) =
J 1 (z),
2z + 2
where J (z) is the Bessel function of the rst kind.
See Also:
besselj (page 409), bessely (page 409), besselk (page 411), yn (page 413)
412
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import Symbol, jn, sin, cos, expand_func
>>> z = Symbol(z)
>>> print(jn(0, z).expand(func=True))
sin(z)/z
>>> jn(1, z).expand(func=True) == sin(z)/z**2 - cos(z)/z
True
>>> expand_func(jn(3, z))
(-6/z**2 + 15/z**4)*sin(z) + (1/z - 15/z**3)*cos(z)
The spherical Bessel functions of integral order are calculated using the formula:
jn (z) = fn (z) sin z + (1)n+1 fn1 (z) cos z,
where the coecients fn (z) are available as polys.orthopolys.spherical bessel fn().
class sympy.functions.special.bessel.yn
Spherical Bessel function of the second kind.
This function is another solution to the spherical Bessel equation, and linearly independent from jn . It can be dened as
j (z) =
Y 1 (z),
2z + 2
where Y (z) is the Bessel function of the second kind.
See Also:
besselj (page 409), bessely (page 409), besselk (page 411), jn (page 412)
Examples
>>> from sympy import Symbol, yn, sin, cos, expand_func
>>> z = Symbol(z)
>>> print(expand_func(yn(0, z)))
-cos(z)/z
>>> expand_func(yn(1, z)) == -cos(z)/z**2-sin(z)/z
True
For integral orders n, yn is calculated using the formula:
yn (z) = (1)n+1 jn1 (z)
sympy.functions.special.bessel.jn zeros(n, k, method=sympy, dps=15)
Zeros of the spherical Bessel function of the rst kind.
This returns an array of zeros of jn up to the k-th zero.
method = sympy: uses mpmath.besseljzero() (page 822)
method = scipy: uses the SciPys sph jn and newton to nd all roots, which is
faster than computing the zeros using a general numerical solver, but it requires
SciPy and only works with low precision oating point numbers. [The function used
with method=sympy is a recent addition to mpmath, before that a general solver
was used.]
5.8. Functions Module
413
SymPy Documentation, Release 0.7.6
See Also:
jn (page 412), yn (page 413), besselj (page 409), besselk (page 411), bessely
(page 409)
Examples
>>> from sympy import jn_zeros
>>> jn_zeros(2, 4, dps=5)
[5.7635, 9.095, 12.323, 15.515]
Airy Functions
class sympy.functions.special.bessel.AiryBase
Abstract base class for Airy functions.
This class is meant to reduce code duplication.
class sympy.functions.special.bessel.airyai
The Airy function Ai of the rst kind.
The Airy function Ai(z) is dened to be the function satisfying Airys dierential equation
2
d w(z)
zw(z) = 0.
dz 2
Equivalently, for real z
1
Ai(z) :=
cos
0
)
t3
+ zt dt.
3
See Also:
airybi (page 415) Airy function of the second kind.
airyaiprime (page 417) Derivative of the Airy function of the rst kind.
airybiprime (page 418) Derivative of the Airy function of the second kind.
References
[R205] (page 1906), [R206] (page 1907), [R207] (page 1907), [R208] (page 1907)
Examples
Create an Airy function object:
>>> from sympy import airyai
>>> from sympy.abc import z
414
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> airyai(z)
airyai(z)
Several special values are known:
>>> airyai(0)
3**(1/3)/(3*gamma(2/3))
>>> from sympy import oo
>>> airyai(oo)
0
>>> airyai(-oo)
0
The Airy function obeys the mirror symmetry:
>>> from sympy import conjugate
>>> conjugate(airyai(z))
airyai(conjugate(z))
Dierentiation with respect to z is supported:
>>> from sympy import diff
>>> diff(airyai(z), z)
airyaiprime(z)
>>> diff(airyai(z), z, 2)
z*airyai(z)
Series expansion is also supported:
>>> from sympy import series
>>> series(airyai(z), z, 0, 3)
3**(5/6)*gamma(1/3)/(6*pi) - 3**(1/6)*z*gamma(2/3)/(2*pi) + O(z**3)
We can numerically evaluate the Airy function to arbitrary precision on the whole complex plane:
>>> airyai(-2).evalf(50)
0.22740742820168557599192443603787379946077222541710
Rewrite Ai(z) in terms of hypergeometric functions:
>>> from sympy import hyper
>>> airyai(z).rewrite(hyper)
-3**(2/3)*z*hyper((), (4/3,), z**3/9)/(3*gamma(1/3)) + 3**(1/3)*hyper((), (2/3,), z**3/9)/(3*gam
class sympy.functions.special.bessel.airybi
The Airy function Bi of the second kind.
The Airy function Bi(z) is dened to be the function satisfying Airys dierential equation
2
d w(z)
zw(z) = 0.
dz 2
Equivalently, for real z
1
Bi(z) :=
5.8. Functions Module
)
( 3
)
t3
t
exp + zt + sin
+ zt dt.
3
3
415
SymPy Documentation, Release 0.7.6
See Also:
airyai (page 414) Airy function of the rst kind.
airyaiprime (page 417) Derivative of the Airy function of the rst kind.
airybiprime (page 418) Derivative of the Airy function of the second kind.
References
[R209] (page 1907), [R210] (page 1907), [R211] (page 1907), [R212] (page 1907)
Examples
Create an Airy function object:
>>> from sympy import airybi
>>> from sympy.abc import z
>>> airybi(z)
airybi(z)
Several special values are known:
>>> airybi(0)
3**(5/6)/(3*gamma(2/3))
>>> from sympy import oo
>>> airybi(oo)
oo
>>> airybi(-oo)
0
The Airy function obeys the mirror symmetry:
>>> from sympy import conjugate
>>> conjugate(airybi(z))
airybi(conjugate(z))
Dierentiation with respect to z is supported:
>>> from sympy import diff
>>> diff(airybi(z), z)
airybiprime(z)
>>> diff(airybi(z), z, 2)
z*airybi(z)
Series expansion is also supported:
>>> from sympy import series
>>> series(airybi(z), z, 0, 3)
3**(1/3)*gamma(1/3)/(2*pi) + 3**(2/3)*z*gamma(2/3)/(2*pi) + O(z**3)
We can numerically evaluate the Airy function to arbitrary precision on the whole complex plane:
>>> airybi(-2).evalf(50)
-0.41230258795639848808323405461146104203453483447240
416
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Rewrite Bi(z) in terms of hypergeometric functions:
>>> from sympy import hyper
>>> airybi(z).rewrite(hyper)
3**(1/6)*z*hyper((), (4/3,), z**3/9)/gamma(1/3) + 3**(5/6)*hyper((), (2/3,), z**3/9)/(3*gamma(2/
class sympy.functions.special.bessel.airyaiprime
The derivative Ai0 of the Airy function of the rst kind.
The Airy function Ai0 (z) is dened to be the function
Ai0 (z) :=
d Ai(z)
.
dz
See Also:
airyai (page 414) Airy function of the rst kind.
airybi (page 415) Airy function of the second kind.
airybiprime (page 418) Derivative of the Airy function of the second kind.
References
[R213] (page 1907), [R214] (page 1907), [R215] (page 1907), [R216] (page 1907)
Examples
Create an Airy function object:
>>> from sympy import airyaiprime
>>> from sympy.abc import z
>>> airyaiprime(z)
airyaiprime(z)
Several special values are known:
>>> airyaiprime(0)
-3**(2/3)/(3*gamma(1/3))
>>> from sympy import oo
>>> airyaiprime(oo)
0
The Airy function obeys the mirror symmetry:
>>> from sympy import conjugate
>>> conjugate(airyaiprime(z))
airyaiprime(conjugate(z))
Dierentiation with respect to z is supported:
>>> from sympy import diff
>>> diff(airyaiprime(z), z)
z*airyai(z)
>>> diff(airyaiprime(z), z, 2)
z*airyaiprime(z) + airyai(z)
5.8. Functions Module
417
SymPy Documentation, Release 0.7.6
Series expansion is also supported:
>>> from sympy import series
>>> series(airyaiprime(z), z, 0, 3)
-3**(2/3)/(3*gamma(1/3)) + 3**(1/3)*z**2/(6*gamma(2/3)) + O(z**3)
We can numerically evaluate the Airy function to arbitrary precision on the whole complex plane:
>>> airyaiprime(-2).evalf(50)
0.61825902074169104140626429133247528291577794512415
Rewrite Ai(z) in terms of hypergeometric functions:
>>> from sympy import hyper
>>> airyaiprime(z).rewrite(hyper)
3**(1/3)*z**2*hyper((), (5/3,), z**3/9)/(6*gamma(2/3)) - 3**(2/3)*hyper((), (1/3,), z**3/9)/(3*g
class sympy.functions.special.bessel.airybiprime
The derivative Bi0 of the Airy function of the rst kind.
The Airy function Bi0 (z) is dened to be the function
Bi0 (z) :=
d Bi(z)
.
dz
See Also:
airyai (page 414) Airy function of the rst kind.
airybi (page 415) Airy function of the second kind.
airyaiprime (page 417) Derivative of the Airy function of the rst kind.
References
[R217] (page 1907), [R218] (page 1907), [R219] (page 1907), [R220] (page 1907)
Examples
Create an Airy function object:
>>> from sympy import airybiprime
>>> from sympy.abc import z
>>> airybiprime(z)
airybiprime(z)
Several special values are known:
>>> airybiprime(0)
3**(1/6)/gamma(1/3)
>>> from sympy import oo
>>> airybiprime(oo)
oo
>>> airybiprime(-oo)
0
418
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
The Airy function obeys the mirror symmetry:
>>> from sympy import conjugate
>>> conjugate(airybiprime(z))
airybiprime(conjugate(z))
Dierentiation with respect to z is supported:
>>> from sympy import diff
>>> diff(airybiprime(z), z)
z*airybi(z)
>>> diff(airybiprime(z), z, 2)
z*airybiprime(z) + airybi(z)
Series expansion is also supported:
>>> from sympy import series
>>> series(airybiprime(z), z, 0, 3)
3**(1/6)/gamma(1/3) + 3**(5/6)*z**2/(6*gamma(2/3)) + O(z**3)
We can numerically evaluate the Airy function to arbitrary precision on the whole complex plane:
>>> airybiprime(-2).evalf(50)
0.27879516692116952268509756941098324140300059345163
Rewrite Bi(z) in terms of hypergeometric functions:
>>> from sympy import hyper
>>> airybiprime(z).rewrite(hyper)
3**(5/6)*z**2*hyper((), (5/3,), z**3/9)/(6*gamma(2/3)) + 3**(1/6)*hyper((), (1/3,), z**3/9)/gamm
B-Splines
sympy.functions.special.bsplines.bspline basis(d, knots, n, x, close=True)
The n-th B-spline at x of degree d with knots.
B-Splines are piecewise polynomials of degree d [R221] (page 1907). They are dened
on a set of knots, which is a sequence of integers or oats.
The 0th degree splines have a value of one on a single interval:
>>> from sympy import bspline_basis
>>> from sympy.abc import x
>>> d = 0
>>> knots = range(5)
>>> bspline_basis(d, knots, 0, x)
Piecewise((1, And(x <= 1, x >= 0)), (0, True))
For a given (d, knots) there are len(knots)-d-1 B-splines dened, that are indexed
by n (starting at 0).
Here is an example of a cubic B-spline:
>>> bspline_basis(3, range(5), 0, x)
Piecewise((x**3/6, And(x < 1, x >= 0)),
(-x**3/2 + 2*x**2 - 2*x + 2/3, And(x < 2, x >= 1)),
(x**3/2 - 4*x**2 + 10*x - 22/3, And(x < 3, x >= 2)),
(-x**3/6 + 2*x**2 - 8*x + 32/3, And(x <= 4, x >= 3)),
(0, True))
5.8. Functions Module
419
SymPy Documentation, Release 0.7.6
By repeating knot points, you can introduce discontinuities in the B-splines and their
derivatives:
>>> d = 1
>>> knots = [0,0,2,3,4]
>>> bspline_basis(d, knots, 0, x)
Piecewise((-x/2 + 1, And(x <= 2, x >= 0)), (0, True))
It is quite time consuming to construct and evaluate B-splines. If you need to evaluate a
B-splines many times, it is best to lambdify them rst:
>>>
>>>
>>>
>>>
>>>
>>>
from sympy import lambdify
d = 3
knots = range(10)
b0 = bspline_basis(d, knots, 0, x)
f = lambdify(x, b0)
y = f(0.5)
See Also:
bsplines basis set
References
[R221] (page 1907)
sympy.functions.special.bsplines.bspline basis set(d, knots, x)
Return the len(knots)-d-1 B-splines at x of degree d with knots.
This function returns a list of Piecewise polynomials that are the len(knots)-d-1 Bsplines of degree d for the given knots. This function calls bspline basis(d, knots,
n, x) for dierent values of n.
See Also:
bsplines basis
Examples
>>> from sympy import bspline_basis_set
>>> from sympy.abc import x
>>> d = 2
>>> knots = range(5)
>>> splines = bspline_basis_set(d, knots, x)
>>> splines
[Piecewise((x**2/2, And(x < 1, x >= 0)),
(-x**2 + 3*x - 3/2, And(x < 2, x >= 1)),
(x**2/2 - 3*x + 9/2, And(x <= 3, x >= 2)),
(0, True)),
Piecewise((x**2/2 - x + 1/2, And(x < 2, x >= 1)),
(-x**2 + 5*x - 11/2, And(x < 3, x >= 2)),
(x**2/2 - 4*x + 8, And(x <= 4, x >= 3)),
(0, True))]
420
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Riemann Zeta and Related Functions
class sympy.functions.special.zeta functions.zeta
Hurwitz zeta function (or Riemann zeta function).
For Re(a) > 0 and Re(s) > 1, this function is dened as
(s, a) =
1
,
(n
+
a)s
n=0
where the standard choice of argument for n + a is used. For xed a with Re(a) > 0 the
Hurwitz zeta function admits a meromorphic continuation to all of C, it is an unbranched
function with a simple pole at s = 1.
Analytic continuation to other a is possible under some circumstances, but this is not
typically done.
The Hurwitz zeta function is a special case of the Lerch transcendent:
(s, a) = (1, s, a).
This formula denes an analytic continuation for all possible values of s and a (also
Re(a) < 0), see the documentation of lerchphi (page 424) for a description of the branching behavior.
If no value is passed for a, by this function assumes a default value of a = 1, yielding the
Riemann zeta function.
See Also:
dirichlet eta (page 422), lerchphi (page 424), polylog (page 423)
References
[R222] (page 1907), [R223] (page 1907)
Examples
For a = 1 the Hurwitz zeta function reduces to the famous Riemann zeta function:
(s, 1) = (s) =
1
.
s
n
n=1
>>> from sympy import zeta
>>> from sympy.abc import s
>>> zeta(s, 1)
zeta(s)
>>> zeta(s)
zeta(s)
The Riemann zeta function can also be expressed using the Dirichlet eta function:
5.8. Functions Module
421
SymPy Documentation, Release 0.7.6
>>> from sympy import dirichlet_eta
>>> zeta(s).rewrite(dirichlet_eta)
dirichlet_eta(s)/(-2**(-s + 1) + 1)
The Riemann zeta function at positive even integer and negative odd integer values is
related to the Bernoulli numbers:
>>> zeta(2)
pi**2/6
>>> zeta(4)
pi**4/90
>>> zeta(-1)
-1/12
The specic formulae are:
(2n) = (1)n+1
(n) =
B2n (2)2n
2(2n)!
Bn+1
n+1
At negative even integers the Riemann zeta function is zero:
>>> zeta(-4)
0
No closed-form expressions are known at positive odd integers, but numerical evaluation
is possible:
>>> zeta(3).n()
1.20205690315959
The derivative of (s, a) with respect to a is easily computed:
>>> from sympy.abc import a
>>> zeta(s, a).diff(a)
-s*zeta(s + 1, a)
However the derivative with respect to s has no useful closed form expression:
>>> zeta(s, a).diff(s)
Derivative(zeta(s, a), s)
The Hurwitz zeta function can be expressed in terms of the Lerch transcendent,
sympy.functions.special.lerchphi:
>>> from sympy import lerchphi
>>> zeta(s, a).rewrite(lerchphi)
lerchphi(1, s, a)
class sympy.functions.special.zeta functions.dirichlet eta
Dirichlet eta function.
422
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
For Re(s) > 0, this function is dened as
(s) =
(1)n
.
ns
n=1
It admits a unique analytic continuation to all of C. It is an entire, unbranched function.
See Also:
zeta (page 421)
References
[R224] (page 1907)
Examples
The Dirichlet eta function is closely related to the Riemann zeta function:
>>> from sympy import dirichlet_eta, zeta
>>> from sympy.abc import s
>>> dirichlet_eta(s).rewrite(zeta)
(-2**(-s + 1) + 1)*zeta(s)
class sympy.functions.special.zeta functions.polylog
Polylogarithm function.
For |z| < 1 and s C, the polylogarithm is dened by
Lis (z) =
zn
,
ns
n=1
where the standard branch of the argument is used for n. It admits an analytic continuation which is branched at z = 1 (notably not on the sheet of initial denition), z = 0 and
z = .
The name polylogarithm comes from the fact that for s = 1, the polylogarithm is related
to the ordinary logarithm (see examples), and that
z
Lis (t)
Lis+1 (z) =
dt.
t
0
The polylogarithm is a special case of the Lerch transcendent:
Lis (z) = z(z, s, 1)
See Also:
zeta (page 421), lerchphi (page 424)
5.8. Functions Module
423
SymPy Documentation, Release 0.7.6
Examples
For z {0, 1, 1}, the polylogarithm is automatically expressed using other functions:
>>> from sympy import polylog
>>> from sympy.abc import s
>>> polylog(s, 0)
0
>>> polylog(s, 1)
zeta(s)
>>> polylog(s, -1)
dirichlet_eta(s)
If s is a negative integer, 0 or 1, the polylogarithm can be expressed using elementary
functions. This can be done using expand func():
>>> from sympy import expand_func
>>> from sympy.abc import z
>>> expand_func(polylog(1, z))
-log(z*exp_polar(-I*pi) + 1)
>>> expand_func(polylog(0, z))
z/(-z + 1)
The derivative with respect to z can be computed in closed form:
>>> polylog(s, z).diff(z)
polylog(s - 1, z)/z
The polylogarithm can be expressed in terms of the lerch transcendent:
>>> from sympy import lerchphi
>>> polylog(s, z).rewrite(lerchphi)
z*lerchphi(z, s, 1)
class sympy.functions.special.zeta functions.lerchphi
Lerch transcendent (Lerch phi function).
For Re(a) > 0, |z| < 1 and s C, the Lerch transcendent is dened as
(z, s, a) =
zn
,
(n + a)s
n=0
where the standard branch of the argument is used for n+a, and by analytic continuation
for other values of the parameters.
A commonly used related function is the Lerch zeta function, dened by
L(q, s, a) = (e2iq , s, a).
Analytic Continuation and Branching Behavior
It can be shown that
(z, s, a) = z(z, s, a + 1) + as .
This provides the analytic continuation to Re(a) 0.
424
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Assume now Re(a) > 0. The integral representation
s1 at
dt
t e
0 (z, s, a) =
1 zet (s)
0
provides an analytic continuation to C [1, ). Finally, for x (1, ) we nd
s1
lim+ 0 (x + i, s, a) lim+ 0 (x i, s, a) =
0
0
2i log
x
,
xa (s)
using the standard branch for both log x and log log x (a branch of log log x is needed to
evaluate log xs1 ). This concludes the analytic continuation. The Lerch transcendent is
thus branched at z {0, 1, } and a Z0 . For xed z, a outside these branch points, it is
an entire function of s.
See Also:
polylog (page 423), zeta (page 421)
References
[R225] (page 1907), [R226] (page 1907), [R227] (page 1907)
Examples
The Lerch transcendent is a fairly general function, for this reason it does not automatically evaluate to simpler functions. Use expand func() to achieve this.
If z = 1, the Lerch transcendent reduces to the Hurwitz zeta function:
>>> from sympy import lerchphi, expand_func
>>> from sympy.abc import z, s, a
>>> expand_func(lerchphi(1, s, a))
zeta(s, a)
More generally, if z is a root of unity, the Lerch transcendent reduces to a sum of Hurwitz
zeta functions:
>>> expand_func(lerchphi(-1, s, a))
2**(-s)*zeta(s, a/2) - 2**(-s)*zeta(s, a/2 + 1/2)
If a = 1, the Lerch transcendent reduces to the polylogarithm:
>>> expand_func(lerchphi(z, s, 1))
polylog(s, z)/z
More generally, if a is rational, the Lerch transcendent reduces to a sum of polylogarithms:
>>> from sympy import S
>>> expand_func(lerchphi(z, s, S(1)/2))
2**(s - 1)*(polylog(s, sqrt(z))/sqrt(z) polylog(s, sqrt(z)*exp_polar(I*pi))/sqrt(z))
>>> expand_func(lerchphi(z, s, S(3)/2))
-2**s/z + 2**(s - 1)*(polylog(s, sqrt(z))/sqrt(z) polylog(s, sqrt(z)*exp_polar(I*pi))/sqrt(z))/z
5.8. Functions Module
425
SymPy Documentation, Release 0.7.6
The derivatives with respect to z and a can be computed in closed form:
>>> lerchphi(z, s,
(-a*lerchphi(z, s,
>>> lerchphi(z, s,
-s*lerchphi(z, s +
a).diff(z)
a) + lerchphi(z, s - 1, a))/z
a).diff(a)
1, a)
Hypergeometric Functions
class sympy.functions.special.hyper.hyper
The (generalized) hypergeometric function is dened by a series where the ratios of
successive terms are a rational function of the summation index. When convergent, it is
continued analytically to the largest possible domain.
The hypergeometric function depends on two vectors of parameters, called the numerator parameters ap , and the denominator parameters bq . It also has an argument z. The
series denition is
)
(
(a1 )n . . . (ap )n z n
a1 , . . . , ap
F
z
=
,
p q
b1 , . . . , bq
(b1 )n . . . (bq )n n!
n=0
where (a)n = (a)(a + 1) . . . (a + n 1) denotes the rising factorial.
If one of the bq is a non-positive integer then the series is undened unless one of the
ap is a larger (i.e. smaller in magnitude) non-positive integer. If none of the bq is a nonpositive integer and one of the ap is a non-positive integer, then the series reduces to a
polynomial. To simplify the following discussion, we assume that none of the ap or bq is
a non-positive integer. For more details, see the references.
The series converges for all z if p q, and thus denes an entire single-valued function
in this case. If p = q + 1 the series converges for |z| < 1, and can be continued analytically
into a half-plane. If p > q + 1 the series is divergent for all z.
Note: The hypergeometric function constructor currently does not check if the parameters actually yield a well-dened function.
See Also:
sympy.simplify.hyperexpand (page 1347), sympy.functions.special.gamma functions.gamma
(page 379), meijerg
References
[R228] (page 1907), [R229] (page 1907)
Examples
The parameters ap and bq can be passed as arbitrary iterables, for example:
>>> from sympy.functions import hyper
>>> from sympy.abc import x, n, a
>>> hyper((1, 2, 3), [3, 4], x)
hyper((1, 2, 3), (3, 4), x)
426
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
There is also pretty printing (it looks better using unicode):
>>> from sympy import pprint
>>> pprint(hyper((1, 2, 3), [3, 4], x), use_unicode=False)
_
|_ /1, 2, 3 | \
|
|
| x|
3 2 \ 3, 4 | /
The parameters must always be iterables, even if they are vectors of length one or zero:
>>> hyper((1, ), [], x)
hyper((1,), (), x)
But of course they may be variables (but if they depend on x then you should not expect
much implemented functionality):
>>> hyper((n, a), (n**2,), x)
hyper((n, a), (n**2,), x)
The hypergeometric function generalizes many named special functions. The function
hyperexpand() tries to express a hypergeometric function using named special functions.
For example:
>>> from sympy import hyperexpand
>>> hyperexpand(hyper([], [], x))
exp(x)
You can also use expand func:
>>> from sympy import expand_func
>>> expand_func(x*hyper([1, 1], [2], -x))
log(x + 1)
More examples:
>>> from sympy import S
>>> hyperexpand(hyper([], [S(1)/2], -x**2/4))
cos(x)
>>> hyperexpand(x*hyper([S(1)/2, S(1)/2], [S(3)/2], x**2))
asin(x)
We can also sometimes hyperexpand parametric functions:
>>> from sympy.abc import a
>>> hyperexpand(hyper([-a], [], x))
(-x + 1)**a
ap
Numerator parameters of the hypergeometric function.
argument
Argument of the hypergeometric function.
bq
Denominator parameters of the hypergeometric function.
convergence statement
Return a condition on z under which the series converges.
eta
A quantity related to the convergence of the series.
5.8. Functions Module
427
SymPy Documentation, Release 0.7.6
radius of convergence
Compute the radius of convergence of the dening series.
Note that even if this is not oo, the function may still be evaluated outside of the
radius of convergence by analytic continuation. But if this is zero, then the function
is not actually dened anywhere else.
>>>
>>>
>>>
1
>>>
0
>>>
oo
from sympy.functions import hyper
from sympy.abc import z
hyper((1, 2), [3], z).radius_of_convergence
hyper((1, 2, 3), [4], z).radius_of_convergence
hyper((1, 2), (3, 4), z).radius_of_convergence
class sympy.functions.special.hyper.meijerg
The Meijer G-function is dened by a Mellin-Barnes type integral that resembles an inverse Mellin transform. It generalizes the hypergeometric functions.
The Meijer G-function depends on four sets of parameters. There are numerator parameters a1 , . . . , an and an+1 , . . . , ap , and there are denominator parameters b1 , . . . , bm
and bm+1 , . . . , bq . Confusingly, it is traditionally denoted as follows (note the position of
m, n, p, q, and how they relate to the lengths of the four parameter vectors):
)
(
a1 , . . . , an an+1 , . . . , ap
Gm,n
z .
p,q
b1 , . . . , bm bm+1 , . . . , bq
However, in sympy the four parameter vectors are always available separately (see examples), so that there is no need to keep track of the decorating sub- and super-scripts
on the G symbol.
The G function is dened as the following integral:
n
m
1
j=1 (1 aj + s)
j=1 (bj s)
q
z s ds,
2i L j=m+1 (1 bj + s) pj=n+1 (aj s)
where (z) is the gamma function. There are three possible contours which we will not
describe in detail here (see the references). If the integral converges along more than
one of them the denitions agree. The contours all separate the poles of (1 aj + s) from
the poles of (bk s), so in particular the G function is undened if aj bk Z>0 for some
j n and k m.
The conditions under which one of the contours yields a convergent integral are complicated and we do not state them here, see the references.
Note: Currently the Meijer G-function constructor does not check any convergence conditions.
See Also:
hyper, sympy.simplify.hyperexpand (page 1347)
References
[R230] (page 1907), [R231] (page 1907)
428
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
You can pass the parameters either as four separate vectors:
>>> from sympy.functions import meijerg
>>> from sympy.abc import x, a
>>> from sympy.core.containers import Tuple
>>> from sympy import pprint
>>> pprint(meijerg((1, 2), (a, 4), (5,), [], x), use_unicode=False)
__1, 2 /1, 2 a, 4 | \
/__
|
| x|
\_|4, 1 \ 5
| /
or as two nested vectors:
>>> pprint(meijerg([(1, 2), (3, 4)], ([5], Tuple()), x), use_unicode=False)
__1, 2 /1, 2 3, 4 | \
/__
|
| x|
\_|4, 1 \ 5
| /
As with the hypergeometric function, the parameters may be passed as arbitrary iterables. Vectors of length zero and one also have to be passed as iterables. The parameters
need not be constants, but if they depend on the argument then not much implemented
functionality should be expected.
All the subvectors of parameters are available:
>>> from sympy import pprint
>>> g = meijerg([1], [2], [3], [4], x)
>>> pprint(g, use_unicode=False)
__1, 1 /1 2 | \
/__
|
| x|
\_|2, 2 \3 4 | /
>>> g.an
(1,)
>>> g.ap
(1, 2)
>>> g.aother
(2,)
>>> g.bm
(3,)
>>> g.bq
(3, 4)
>>> g.bother
(4,)
The Meijer G-function generalizes the hypergeometric functions. In some cases it can be
expressed in terms of hypergeometric functions, using Slaters theorem. For example:
>>> from sympy import hyperexpand
>>> from sympy.abc import a, b, c
>>> hyperexpand(meijerg([a], [], [c], [b], x), allow_hyper=True)
x**c*gamma(-a + c + 1)*hyper((-a + c + 1,),
(-b + c + 1,), -x)/gamma(-b + c + 1)
Thus the Meijer G-function also subsumes many named functions as special cases. You
can use expand func or hyperexpand to (try to) rewrite a Meijer G-function in terms of
named special functions. For example:
5.8. Functions Module
429
SymPy Documentation, Release 0.7.6
>>> from sympy import expand_func, S
>>> expand_func(meijerg([[],[]], [[0],[]], -x))
exp(x)
>>> hyperexpand(meijerg([[],[]], [[S(1)/2],[0]], (x/2)**2))
sin(x)/sqrt(pi)
an
First set of numerator parameters.
aother
Second set of numerator parameters.
ap
Combined numerator parameters.
argument
Argument of the Meijer G-function.
bm
First set of denominator parameters.
bother
Second set of denominator parameters.
bq
Combined denominator parameters.
delta
A quantity related to the convergence region of the integral, c.f. references.
get period()
Return a number P such that G(x*exp(I*P)) == G(x).
>>> from sympy.functions.special.hyper import meijerg
>>> from sympy.abc import z
>>> from sympy import pi, S
>>> meijerg([1], [], [], [], z).get_period()
2*pi
>>> meijerg([pi], [], [], [], z).get_period()
oo
>>> meijerg([1, 2], [], [], [], z).get_period()
oo
>>> meijerg([1,1], [2], [1, S(1)/2, S(1)/3], [1], z).get_period()
12*pi
integrand(s)
Get the dening integrand D(s).
nu
A quantity related to the convergence region of the integral, c.f. references.
Elliptic integrals
class sympy.functions.special.elliptic integrals.elliptic k
The complete elliptic integral of the rst kind, dened by
( )
K(z) = F 2 z
430
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
where F (z|m) is the Legendre incomplete elliptic integral of the rst kind.
The function K(z) is a single-valued function on the complex plane with branch cut along
the interval (1, ).
See Also:
elliptic f (page 431)
References
[R232] (page 1907), [R233] (page 1907)
Examples
>>> from sympy import elliptic_k, I, pi
>>> from sympy.abc import z
>>> elliptic_k(0)
pi/2
>>> elliptic_k(1.0 + I)
1.50923695405127 + 0.625146415202697*I
>>> elliptic_k(z).series(z, n=3)
pi/2 + pi*z/8 + 9*pi*z**2/128 + O(z**3)
class sympy.functions.special.elliptic integrals.elliptic f
The Legendre incomplete elliptic integral of the rst kind, dened by
z
dt
F (z|m) =
0
1 m sin2 t
This function reduces to a complete elliptic integral of the rst kind, K(m), when z = /2.
See Also:
elliptic k (page 430)
References
[R234] (page 1907), [R235] (page 1907)
Examples
>>> from sympy import elliptic_f, I, O
>>> from sympy.abc import z, m
>>> elliptic_f(z, m).series(z)
z + z**5*(3*m**2/40 - m/30) + m*z**3/6 + O(z**6)
>>> elliptic_f(3.0 + I/2, 1.0 + I)
2.909449841483 + 1.74720545502474*I
class sympy.functions.special.elliptic integrals.elliptic e
Called with two arguments z and m, evaluates the incomplete elliptic integral of the
5.8. Functions Module
431
SymPy Documentation, Release 0.7.6
second kind, dened by
E (z|m) =
1 m sin2 tdt
Called with a single argument z, evaluates the Legendre complete elliptic integral of the
second kind
( )
E(z) = E 2 z
The function E(z) is a single-valued function on the complex plane with branch cut along
the interval (1, ).
References
[R236] (page 1907), [R237] (page 1907), [R238] (page 1907)
Examples
>>> from sympy import elliptic_e, I, pi, O
>>> from sympy.abc import z, m
>>> elliptic_e(z, m).series(z)
z + z**5*(-m**2/40 + m/30) - m*z**3/6 + O(z**6)
>>> elliptic_e(z).series(z, n=4)
pi/2 - pi*z/8 - 3*pi*z**2/128 - 5*pi*z**3/512 + O(z**4)
>>> elliptic_e(1 + I, 2 - I/2).n()
1.55203744279187 + 0.290764986058437*I
>>> elliptic_e(0)
pi/2
>>> elliptic_e(2.0 - I)
0.991052601328069 + 0.81879421395609*I
class sympy.functions.special.elliptic integrals.elliptic pi
Called with three arguments n, z and m, evaluates the Legendre incomplete elliptic integral of the third kind, dened by
z
dt
(
)
(n; z|m) =
2
0
1 n sin t
1 m sin2 t
Called with two arguments n and m, evaluates the complete elliptic integral of the third
kind:
)
(
(n|m) = n; 2 m
References
[R239] (page 1907), [R240] (page 1907), [R241] (page 1907)
432
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import elliptic_pi, I, pi, O, S
>>> from sympy.abc import z, n, m
>>> elliptic_pi(n, z, m).series(z, n=4)
z + z**3*(m/6 + n/3) + O(z**4)
>>> elliptic_pi(0.5 + I, 1.0 - I, 1.2)
2.50232379629182 - 0.760939574180767*I
>>> elliptic_pi(0, 0)
pi/2
>>> elliptic_pi(1.0 - I/3, 2.0 + I)
3.29136443417283 + 0.32555634906645*I
Orthogonal Polynomials
This module mainly implements special orthogonal polynomials.
See also functions.combinatorial.numbers which contains some combinatorial polynomials.
Jacobi Polynomials
class sympy.functions.special.polynomials.jacobi
(,)
(x)
Jacobi polynomial Pn
(,)
jacobi(n, alpha, beta, x) gives the nth Jacobi polynomial in x, Pn
(x).
The Jacobi polynomials are orthogonal on [1, 1] with respect to the weight
(1 x) (1 + x) .
See Also:
gegenbauer
(page
435),
chebyshevt root
(page
438),
chebyshevu
(page 437),
chebyshevu root (page 438),
legendre (page 439),
assoc legendre (page 439),
hermite (page 440),
laguerre (page 441),
assoc laguerre
(page
442),
sympy.polys.orthopolys.jacobi poly
(page
1120),
sympy.polys.orthopolys.gegenbauer poly
(page
1120),
sympy.polys.orthopolys.chebyshevt poly (page 1120), sympy.polys.orthopolys.chebyshevu poly
(page
1120),
sympy.polys.orthopolys.hermite poly
(page
1120),
sympy.polys.orthopolys.legendre poly (page 1120), sympy.polys.orthopolys.laguerre poly
(page 1120)
References
[R242] (page 1907), [R243] (page 1908), [R244] (page 1908)
Examples
>>> from sympy import jacobi, S, conjugate, diff
>>> from sympy.abc import n,a,b,x
>>> jacobi(0, a, b, x)
1
>>> jacobi(1, a, b, x)
5.8. Functions Module
433
SymPy Documentation, Release 0.7.6
a/2 - b/2 + x*(a/2 + b/2 + 1)
>>> jacobi(2, a, b, x)
(a**2/8 - a*b/4 - a/8 + b**2/8 - b/8 + x**2*(a**2/8 + a*b/4 + 7*a/8 +
b**2/8 + 7*b/8 + 3/2) + x*(a**2/4 + 3*a/4 - b**2/4 - 3*b/4) - 1/2)
>>> jacobi(n, a, b, x)
jacobi(n, a, b, x)
>>> jacobi(n, a, a, x)
RisingFactorial(a + 1, n)*gegenbauer(n,
a + 1/2, x)/RisingFactorial(2*a + 1, n)
>>> jacobi(n, 0, 0, x)
legendre(n, x)
>>> jacobi(n, S(1)/2, S(1)/2, x)
RisingFactorial(3/2, n)*chebyshevu(n, x)/factorial(n + 1)
>>> jacobi(n, -S(1)/2, -S(1)/2, x)
RisingFactorial(1/2, n)*chebyshevt(n, x)/factorial(n)
>>> jacobi(n, a, b, -x)
(-1)**n*jacobi(n, b, a, x)
>>> jacobi(n, a, b,
2**(-n)*gamma(a + n
>>> jacobi(n, a, b,
RisingFactorial(a +
0)
+ 1)*hyper((-b - n, -n), (a + 1,), -1)/(factorial(n)*gamma(a + 1))
1)
1, n)/factorial(n)
>>> conjugate(jacobi(n, a, b, x))
jacobi(n, conjugate(a), conjugate(b), conjugate(x))
>>> diff(jacobi(n,a,b,x), x)
(a/2 + b/2 + n/2 + 1/2)*jacobi(n - 1, a + 1, b + 1, x)
sympy.functions.special.polynomials.jacobi normalized(n, a, b, x)
(,)
(x)
Jacobi polynomial Pn
(,)
jacobi normalized(n, alpha, beta, x) gives the nth Jacobi polynomial in x, Pn
(x).
The Jacobi polynomials are orthogonal on [1, 1] with respect to the weight
(1 x) (1 + x) .
This functions returns the polynomials normilzed:
1
1
(,)
Pm
(x)Pn(,) (x)(1 x) (1 + x) dx = m,n
See Also:
gegenbauer
(page
435),
chebyshevt root
(page
438),
chebyshevu
(page 437),
chebyshevu root (page 438),
legendre (page 439),
assoc legendre (page 439),
hermite (page 440),
laguerre (page 441),
assoc laguerre
(page
442),
sympy.polys.orthopolys.jacobi poly
(page
1120),
sympy.polys.orthopolys.gegenbauer poly
(page
1120),
sympy.polys.orthopolys.chebyshevt poly (page 1120), sympy.polys.orthopolys.chebyshevu poly
(page
1120),
sympy.polys.orthopolys.hermite poly
(page
1120),
434
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
sympy.polys.orthopolys.legendre poly (page 1120), sympy.polys.orthopolys.laguerre poly
(page 1120)
References
[R245] (page 1908), [R246] (page 1908), [R247] (page 1908)
Examples
>>> from sympy import jacobi_normalized
>>> from sympy.abc import n,a,b,x
>>> jacobi_normalized(n, a, b, x)
jacobi(n, a, b, x)/sqrt(2**(a + b + 1)*gamma(a + n + 1)*gamma(b + n + 1)/((a + b + 2*n + 1)*fact
Gegenbauer Polynomials
class sympy.functions.special.polynomials.gegenbauer
()
Gegenbauer polynomial Cn (x)
()
gegenbauer(n, alpha, x) gives the nth Gegenbauer polynomial in x, Cn (x).
The Gegenbauer polynomials are orthogonal on [1, 1] with respect to the weight
(
) 21
1 x2
.
See Also:
jacobi
(page
433),
chebyshevt root
(page
438),
chebyshevu
(page 437),
chebyshevu root (page 438),
legendre (page 439),
assoc legendre (page 439),
hermite (page 440),
laguerre (page 441),
assoc laguerre
(page
442),
sympy.polys.orthopolys.jacobi poly
(page
1120),
sympy.polys.orthopolys.gegenbauer poly
(page
1120),
sympy.polys.orthopolys.chebyshevt poly (page 1120), sympy.polys.orthopolys.chebyshevu poly
(page
1120),
sympy.polys.orthopolys.hermite poly
(page
1120),
sympy.polys.orthopolys.legendre poly (page 1120), sympy.polys.orthopolys.laguerre poly
(page 1120)
References
[R248] (page 1908), [R249] (page 1908), [R250] (page 1908)
Examples
>>> from sympy import gegenbauer, conjugate, diff
>>> from sympy.abc import n,a,x
>>> gegenbauer(0, a, x)
1
>>> gegenbauer(1, a, x)
2*a*x
>>> gegenbauer(2, a, x)
-a + x**2*(2*a**2 + 2*a)
5.8. Functions Module
435
SymPy Documentation, Release 0.7.6
>>> gegenbauer(3, a, x)
x**3*(4*a**3/3 + 4*a**2 + 8*a/3) + x*(-2*a**2 - 2*a)
>>> gegenbauer(n, a, x)
gegenbauer(n, a, x)
>>> gegenbauer(n, a, -x)
(-1)**n*gegenbauer(n, a, x)
>>> gegenbauer(n, a, 0)
2**n*sqrt(pi)*gamma(a + n/2)/(gamma(a)*gamma(-n/2 + 1/2)*gamma(n + 1))
>>> gegenbauer(n, a, 1)
gamma(2*a + n)/(gamma(2*a)*gamma(n + 1))
>>> conjugate(gegenbauer(n, a, x))
gegenbauer(n, conjugate(a), conjugate(x))
>>> diff(gegenbauer(n, a, x), x)
2*a*gegenbauer(n - 1, a + 1, x)
Chebyshev Polynomials
class sympy.functions.special.polynomials.chebyshevt
Chebyshev polynomial of the rst kind, Tn (x)
chebyshevt(n, x) gives the nth Chebyshev polynomial (of the rst kind) in x, Tn (x).
The Chebyshev polynomials of the rst kind are orthogonal on [1, 1] with respect to the
1
weight 1x
.
2
See Also:
jacobi (page 433), gegenbauer (page 435), chebyshevt root (page 438),
chebyshevu (page 437), chebyshevu root (page 438), legendre (page 439),
assoc legendre (page 439),
hermite (page 440),
laguerre (page 441),
assoc laguerre
(page
442),
sympy.polys.orthopolys.jacobi poly
(page
1120),
sympy.polys.orthopolys.gegenbauer poly
(page
1120),
sympy.polys.orthopolys.chebyshevt poly (page 1120), sympy.polys.orthopolys.chebyshevu poly
(page
1120),
sympy.polys.orthopolys.hermite poly
(page
1120),
sympy.polys.orthopolys.legendre poly (page 1120), sympy.polys.orthopolys.laguerre poly
(page 1120)
References
[R251] (page 1908), [R252] (page 1908), [R253] (page 1908), [R254] (page 1908), [R255]
(page 1908)
Examples
>>>
>>>
>>>
1
>>>
x
436
from sympy import chebyshevt, chebyshevu, diff
from sympy.abc import n,x
chebyshevt(0, x)
chebyshevt(1, x)
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> chebyshevt(2, x)
2*x**2 - 1
>>> chebyshevt(n, x)
chebyshevt(n, x)
>>> chebyshevt(n, -x)
(-1)**n*chebyshevt(n, x)
>>> chebyshevt(-n, x)
chebyshevt(n, x)
>>> chebyshevt(n, 0)
cos(pi*n/2)
>>> chebyshevt(n, -1)
(-1)**n
>>> diff(chebyshevt(n, x), x)
n*chebyshevu(n - 1, x)
class sympy.functions.special.polynomials.chebyshevu
Chebyshev polynomial of the second kind, Un (x)
chebyshevu(n, x) gives the nth Chebyshev polynomial of the second kind in x, Un (x).
The Chebyshev
polynomials of the second kind are orthogonal on [1, 1] with respect to
the weight 1 x2 .
See Also:
jacobi (page 433),
gegenbauer (page 435),
chebyshevt (page 436),
chebyshevt root
(page
438),
chebyshevu root
(page
438),
legendre
(page 439),
assoc legendre (page 439),
hermite (page 440),
laguerre
(page 441), assoc laguerre (page 442), sympy.polys.orthopolys.jacobi poly
(page
1120),
sympy.polys.orthopolys.gegenbauer poly
(page
1120),
sympy.polys.orthopolys.chebyshevt poly (page 1120), sympy.polys.orthopolys.chebyshevu poly
(page
1120),
sympy.polys.orthopolys.hermite poly
(page
1120),
sympy.polys.orthopolys.legendre poly (page 1120), sympy.polys.orthopolys.laguerre poly
(page 1120)
References
[R256] (page 1908), [R257] (page 1908), [R258] (page 1908), [R259] (page 1908), [R260]
(page 1908)
Examples
>>> from sympy import chebyshevt, chebyshevu, diff
>>> from sympy.abc import n,x
>>> chebyshevu(0, x)
1
>>> chebyshevu(1, x)
2*x
>>> chebyshevu(2, x)
4*x**2 - 1
5.8. Functions Module
437
SymPy Documentation, Release 0.7.6
>>> chebyshevu(n, x)
chebyshevu(n, x)
>>> chebyshevu(n, -x)
(-1)**n*chebyshevu(n, x)
>>> chebyshevu(-n, x)
-chebyshevu(n - 2, x)
>>> chebyshevu(n, 0)
cos(pi*n/2)
>>> chebyshevu(n, 1)
n + 1
>>> diff(chebyshevu(n, x), x)
(-x*chebyshevu(n, x) + (n + 1)*chebyshevt(n + 1, x))/(x**2 - 1)
class sympy.functions.special.polynomials.chebyshevt root
chebyshev root(n, k) returns the kth root (indexed from zero) of the nth Chebyshev polynomial of the rst kind; that is, if 0 <= k < n, chebyshevt(n, chebyshevt root(n, k)) ==
0.
See Also:
jacobi (page 433), gegenbauer (page 435), chebyshevt (page 436), chebyshevu (page 437), chebyshevu root (page 438), legendre (page 439), assoc legendre (page 439),
hermite (page 440),
laguerre (page 441),
assoc laguerre
(page
442),
sympy.polys.orthopolys.jacobi poly
(page
1120),
sympy.polys.orthopolys.gegenbauer poly
(page
1120),
sympy.polys.orthopolys.chebyshevt poly (page 1120), sympy.polys.orthopolys.chebyshevu poly
(page
1120),
sympy.polys.orthopolys.hermite poly
(page
1120),
sympy.polys.orthopolys.legendre poly (page 1120), sympy.polys.orthopolys.laguerre poly
(page 1120)
Examples
>>> from sympy import chebyshevt, chebyshevt_root
>>> chebyshevt_root(3, 2)
-sqrt(3)/2
>>> chebyshevt(3, chebyshevt_root(3, 2))
0
class sympy.functions.special.polynomials.chebyshevu root
chebyshevu root(n, k) returns the kth root (indexed from zero) of the nth Chebyshev
polynomial of the second kind; that is, if 0 <= k < n, chebyshevu(n, chebyshevu root(n,
k)) == 0.
See Also:
chebyshevt (page 436), chebyshevt root (page 438), chebyshevu (page 437),
legendre (page 439), assoc legendre (page 439), hermite (page 440), laguerre
(page 441), assoc laguerre (page 442), sympy.polys.orthopolys.jacobi poly
(page
1120),
sympy.polys.orthopolys.gegenbauer poly
(page
1120),
sympy.polys.orthopolys.chebyshevt poly (page 1120), sympy.polys.orthopolys.chebyshevu poly
(page
1120),
sympy.polys.orthopolys.hermite poly
(page
1120),
sympy.polys.orthopolys.legendre poly (page 1120), sympy.polys.orthopolys.laguerre poly
(page 1120)
438
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import chebyshevu, chebyshevu_root
>>> chebyshevu_root(3, 2)
-sqrt(2)/2
>>> chebyshevu(3, chebyshevu_root(3, 2))
0
Legendre Polynomials
class sympy.functions.special.polynomials.legendre
legendre(n, x) gives the nth Legendre polynomial of x, Pn (x)
The Legendre polynomials are orthogonal on [-1, 1] with respect to the constant weight
1. They satisfy Pn (1) = 1 for all n; further, Pn is odd for odd n and even for even n.
See Also:
jacobi (page 433),
gegenbauer (page 435),
chebyshevt (page 436),
chebyshevt root
(page
438),
chebyshevu
(page
437),
chebyshevu root
(page 438),
assoc legendre (page 439),
hermite (page 440),
laguerre
(page 441), assoc laguerre (page 442), sympy.polys.orthopolys.jacobi poly
(page
1120),
sympy.polys.orthopolys.gegenbauer poly
(page
1120),
sympy.polys.orthopolys.chebyshevt poly (page 1120), sympy.polys.orthopolys.chebyshevu poly
(page
1120),
sympy.polys.orthopolys.hermite poly
(page
1120),
sympy.polys.orthopolys.legendre poly (page 1120), sympy.polys.orthopolys.laguerre poly
(page 1120)
References
[R261] (page 1908), [R262] (page 1908), [R263] (page 1908), [R264] (page 1908)
Examples
>>> from sympy import legendre, diff
>>> from sympy.abc import x, n
>>> legendre(0, x)
1
>>> legendre(1, x)
x
>>> legendre(2, x)
3*x**2/2 - 1/2
>>> legendre(n, x)
legendre(n, x)
>>> diff(legendre(n,x), x)
n*(x*legendre(n, x) - legendre(n - 1, x))/(x**2 - 1)
class sympy.functions.special.polynomials.assoc legendre
assoc legendre(n,m, x) gives Pnm (x), where n and m are the degree and order or an expression which is related to the nth order Legendre polynomial, Pn (x) in the following
manner:
m
Pnm (x) = (1)m (1 x2 ) 2
5.8. Functions Module
d Pn (x)
dxm
439
SymPy Documentation, Release 0.7.6
Associated Legendre polynomial are orthogonal on [-1, 1] with:
weight = 1 for the same m, and dierent n.
weight = 1/(1-x**2) for the same n, and dierent m.
See Also:
jacobi (page 433),
gegenbauer (page 435),
chebyshevt (page 436),
chebyshevt root
(page
438),
chebyshevu
(page
437),
chebyshevu root
(page
438),
legendre
(page
439),
hermite
(page
440),
laguerre
(page 441), assoc laguerre (page 442), sympy.polys.orthopolys.jacobi poly
(page
1120),
sympy.polys.orthopolys.gegenbauer poly
(page
1120),
sympy.polys.orthopolys.chebyshevt poly (page 1120), sympy.polys.orthopolys.chebyshevu poly
(page
1120),
sympy.polys.orthopolys.hermite poly
(page
1120),
sympy.polys.orthopolys.legendre poly (page 1120), sympy.polys.orthopolys.laguerre poly
(page 1120)
References
[R265] (page 1908), [R266] (page 1908), [R267] (page 1908), [R268] (page 1908)
Examples
>>> from sympy import assoc_legendre
>>> from sympy.abc import x, m, n
>>> assoc_legendre(0,0, x)
1
>>> assoc_legendre(1,0, x)
x
>>> assoc_legendre(1,1, x)
-sqrt(-x**2 + 1)
>>> assoc_legendre(n,m,x)
assoc_legendre(n, m, x)
Hermite Polynomials
class sympy.functions.special.polynomials.hermite
hermite(n, x) gives the nth Hermite polynomial in x, Hn (x)
The ( Hermite
polynomials are orthogonal on (, ) with respect to the weight
)
x2
exp 2 .
See Also:
jacobi (page 433),
gegenbauer (page 435),
chebyshevt (page 436),
chebyshevt root
(page
438),
chebyshevu
(page
437),
chebyshevu root
(page 438), legendre (page 439), assoc legendre (page 439), laguerre
(page 441), assoc laguerre (page 442), sympy.polys.orthopolys.jacobi poly
(page
1120),
sympy.polys.orthopolys.gegenbauer poly
(page
1120),
sympy.polys.orthopolys.chebyshevt poly (page 1120), sympy.polys.orthopolys.chebyshevu poly
(page
1120),
sympy.polys.orthopolys.hermite poly
(page
1120),
sympy.polys.orthopolys.legendre poly (page 1120), sympy.polys.orthopolys.laguerre poly
(page 1120)
440
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
References
[R269] (page 1908), [R270] (page 1908), [R271] (page 1908)
Examples
>>> from sympy import hermite, diff
>>> from sympy.abc import x, n
>>> hermite(0, x)
1
>>> hermite(1, x)
2*x
>>> hermite(2, x)
4*x**2 - 2
>>> hermite(n, x)
hermite(n, x)
>>> diff(hermite(n,x), x)
2*n*hermite(n - 1, x)
>>> diff(hermite(n,x), x)
2*n*hermite(n - 1, x)
>>> hermite(n, -x)
(-1)**n*hermite(n, x)
Laguerre Polynomials
class sympy.functions.special.polynomials.laguerre
Returns the nth Laguerre polynomial in x, Ln (x).
Parameters n : int
Degree of Laguerre polynomial. Must be n >= 0.
See Also:
jacobi (page 433),
gegenbauer (page 435),
chebyshevt (page 436),
chebyshevt root
(page
438),
chebyshevu
(page
437),
chebyshevu root
(page 438),
legendre (page 439),
assoc legendre (page 439),
hermite
(page 440), assoc laguerre (page 442), sympy.polys.orthopolys.jacobi poly
(page
1120),
sympy.polys.orthopolys.gegenbauer poly
(page
1120),
sympy.polys.orthopolys.chebyshevt poly (page 1120), sympy.polys.orthopolys.chebyshevu poly
(page
1120),
sympy.polys.orthopolys.hermite poly
(page
1120),
sympy.polys.orthopolys.legendre poly (page 1120), sympy.polys.orthopolys.laguerre poly
(page 1120)
References
[R272] (page 1908), [R273] (page 1908), [R274] (page 1908), [R275] (page 1908)
Examples
>>> from sympy import laguerre, diff
>>> from sympy.abc import x, n
>>> laguerre(0, x)
1
5.8. Functions Module
441
SymPy Documentation, Release 0.7.6
>>> laguerre(1, x)
-x + 1
>>> laguerre(2, x)
x**2/2 - 2*x + 1
>>> laguerre(3, x)
-x**3/6 + 3*x**2/2 - 3*x + 1
>>> laguerre(n, x)
laguerre(n, x)
>>> diff(laguerre(n, x), x)
-assoc_laguerre(n - 1, 1, x)
class sympy.functions.special.polynomials.assoc laguerre
Returns the nth generalized Laguerre polynomial in x, Ln (x).
Parameters n : int
Degree of Laguerre polynomial. Must be n >= 0.
alpha : Expr
Arbitrary expression. For alpha=0 regular Laguerre polynomials will
be generated.
See Also:
jacobi (page 433),
gegenbauer (page 435),
chebyshevt (page 436),
chebyshevt root
(page
438),
chebyshevu
(page
437),
chebyshevu root
(page 438),
legendre (page 439),
assoc legendre (page 439),
hermite
(page
440),
laguerre
(page
441),
sympy.polys.orthopolys.jacobi poly
(page
1120),
sympy.polys.orthopolys.gegenbauer poly
(page
1120),
sympy.polys.orthopolys.chebyshevt poly (page 1120), sympy.polys.orthopolys.chebyshevu poly
(page
1120),
sympy.polys.orthopolys.hermite poly
(page
1120),
sympy.polys.orthopolys.legendre poly (page 1120), sympy.polys.orthopolys.laguerre poly
(page 1120)
References
[R276] (page 1908), [R277] (page 1908), [R278] (page 1908), [R279] (page 1908)
Examples
>>> from sympy import laguerre,
>>> from sympy.abc import x, n,
>>> assoc_laguerre(0, a, x)
1
>>> assoc_laguerre(1, a, x)
a - x + 1
>>> assoc_laguerre(2, a, x)
a**2/2 + 3*a/2 + x**2/2 + x*(-a
>>> assoc_laguerre(3, a, x)
a**3/6 + a**2 + 11*a/6 - x**3/6
x*(-a**2/2 - 5*a/2 - 3) + 1
assoc_laguerre, diff
a
- 2) + 1
+ x**2*(a/2 + 3/2) +
>>> assoc_laguerre(n, a, 0)
binomial(a + n, a)
442
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> assoc_laguerre(n, a, x)
assoc_laguerre(n, a, x)
>>> assoc_laguerre(n, 0, x)
laguerre(n, x)
>>> diff(assoc_laguerre(n, a, x), x)
-assoc_laguerre(n - 1, a + 1, x)
>>> diff(assoc_laguerre(n, a, x), a)
Sum(assoc_laguerre(_k, a, x)/(-a + n), (_k, 0, n - 1))
Spherical Harmonics
class sympy.functions.special.spherical harmonics.Ynm
Spherical harmonics dened as
(2n + 1)(n m)!
m
Yn (, ) :=
exp(im)Pm
n (cos())
4(n + m)!
Ynm() gives the spherical harmonic function of order n and m in and , Ynm (, ). The
four parameters are as follows: n 0 an integer and m an integer such that n m n
holds. The two angles are real-valued with [0, ] and [0, 2].
See Also:
Ynm c, Znm
References
[R280] (page 1908), [R281] (page 1909), [R282] (page 1909), [R283] (page 1909)
Examples
>>>
>>>
>>>
>>>
from sympy import Ynm, Symbol
from sympy.abc import n,m
theta = Symbol(theta)
phi = Symbol(phi)
>>> Ynm(n, m, theta, phi)
Ynm(n, m, theta, phi)
Several symmetries are known, for the order
>>>
>>>
>>>
>>>
from sympy import Ynm, Symbol
from sympy.abc import n,m
theta = Symbol(theta)
phi = Symbol(phi)
>>> Ynm(n, -m, theta, phi)
(-1)**m*exp(-2*I*m*phi)*Ynm(n, m, theta, phi)
5.8. Functions Module
443
SymPy Documentation, Release 0.7.6
as well as for the angles
>>>
>>>
>>>
>>>
from sympy import Ynm, Symbol, simplify
from sympy.abc import n,m
theta = Symbol(theta)
phi = Symbol(phi)
>>> Ynm(n, m, -theta, phi)
Ynm(n, m, theta, phi)
>>> Ynm(n, m, theta, -phi)
exp(-2*I*m*phi)*Ynm(n, m, theta, phi)
For specic integers n and m we can evalute the harmonics to more useful expressions
>>> simplify(Ynm(0, 0, theta, phi).expand(func=True))
1/(2*sqrt(pi))
>>> simplify(Ynm(1, -1, theta, phi).expand(func=True))
sqrt(6)*exp(-I*phi)*sin(theta)/(4*sqrt(pi))
>>> simplify(Ynm(1, 0, theta, phi).expand(func=True))
sqrt(3)*cos(theta)/(2*sqrt(pi))
>>> simplify(Ynm(1, 1, theta, phi).expand(func=True))
-sqrt(6)*exp(I*phi)*sin(theta)/(4*sqrt(pi))
>>> simplify(Ynm(2, -2, theta, phi).expand(func=True))
sqrt(30)*exp(-2*I*phi)*sin(theta)**2/(8*sqrt(pi))
>>> simplify(Ynm(2, -1, theta, phi).expand(func=True))
sqrt(30)*exp(-I*phi)*sin(2*theta)/(8*sqrt(pi))
>>> simplify(Ynm(2, 0, theta, phi).expand(func=True))
sqrt(5)*(3*cos(theta)**2 - 1)/(4*sqrt(pi))
>>> simplify(Ynm(2, 1, theta, phi).expand(func=True))
-sqrt(30)*exp(I*phi)*sin(2*theta)/(8*sqrt(pi))
>>> simplify(Ynm(2, 2, theta, phi).expand(func=True))
sqrt(30)*exp(2*I*phi)*sin(theta)**2/(8*sqrt(pi))
We can dierentiate the functions with respect to both angles
>>>
>>>
>>>
>>>
from sympy import Ynm, Symbol, diff
from sympy.abc import n,m
theta = Symbol(theta)
phi = Symbol(phi)
>>> diff(Ynm(n, m, theta, phi), theta)
m*cot(theta)*Ynm(n, m, theta, phi) + sqrt((-m + n)*(m + n + 1))*exp(-I*phi)*Ynm(n, m + 1, theta,
>>> diff(Ynm(n, m, theta, phi), phi)
I*m*Ynm(n, m, theta, phi)
Further we can compute the complex conjugation
444
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>>
>>>
>>>
>>>
from sympy import Ynm, Symbol, conjugate
from sympy.abc import n,m
theta = Symbol(theta)
phi = Symbol(phi)
>>> conjugate(Ynm(n, m, theta, phi))
(-1)**(2*m)*exp(-2*I*m*phi)*Ynm(n, m, theta, phi)
To get back the well known expressions in spherical coordinates we use full expansion
>>>
>>>
>>>
>>>
from sympy import Ynm, Symbol, expand_func
from sympy.abc import n,m
theta = Symbol(theta)
phi = Symbol(phi)
>>> expand_func(Ynm(n, m, theta, phi))
sqrt((2*n + 1)*factorial(-m + n)/factorial(m + n))*exp(I*m*phi)*assoc_legendre(n, m, cos(theta))
sympy.functions.special.spherical harmonics.Ynm c(n, m, theta, phi)
Conjugate spherical harmonics dened as
Ynm (, ) := (1)m Ynm (, )
See Also:
Ynm, Znm
References
[R284] (page 1909), [R285] (page 1909), [R286] (page 1909)
class sympy.functions.special.spherical harmonics.Znm
Real spherical harmonics dened as
m
Yn (,)+Ynm (,)
m>0
2
m
m
Zn (, ) := Yn (, )
m=0
m (,)
Ynm (,)Y
n
m<0
i 2
which gives in simplied form
m
Y (,)+(1)m Ynm (,)
2
m
m
Zn (, ) = Yn (, )
m m
Yn (,)
Ynm (,)(1)
i 2
m>0
m=0
m<0
See Also:
Ynm, Ynm c
References
[R287] (page 1909), [R288] (page 1909), [R289] (page 1909)
5.8. Functions Module
445
SymPy Documentation, Release 0.7.6
Tensor Functions
sympy.functions.special.tensor functions.Eijk(*args, **kwargs)
Represent the Levi-Civita symbol.
This is just compatibility wrapper to LeviCivita().
See Also:
LeviCivita
sympy.functions.special.tensor functions.eval levicivita(*args)
Evaluate Levi-Civita symbol.
class sympy.functions.special.tensor functions.LeviCivita
Represent the Levi-Civita symbol.
For even permutations of indices it returns 1, for odd permutations -1, and for everything
else (a repeated index) it returns 0.
Thus it represents an alternating pseudotensor.
See Also:
Eijk
Examples
>>> from sympy import LeviCivita
>>> from sympy.abc import i, j, k
>>> LeviCivita(1, 2, 3)
1
>>> LeviCivita(1, 3, 2)
-1
>>> LeviCivita(1, 2, 2)
0
>>> LeviCivita(i, j, k)
LeviCivita(i, j, k)
>>> LeviCivita(i, j, i)
0
class sympy.functions.special.tensor functions.KroneckerDelta
The discrete, or Kronecker, delta function.
A function that takes in two integers i and j. It returns 0 if i and j are not equal or it
returns 1 if i and j are equal.
Parameters i : Number, Symbol
The rst index of the delta function.
j : Number, Symbol
The second index of the delta function.
See Also:
eval, sympy.functions.special.delta functions.DiracDelta (page 377)
References
[R290] (page 1909)
446
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
A simple example with integer indices:
>>> from sympy.functions.special.tensor_functions import KroneckerDelta
>>> KroneckerDelta(1, 2)
0
>>> KroneckerDelta(3, 3)
1
Symbolic indices:
>>> from sympy.abc import
>>> KroneckerDelta(i, j)
KroneckerDelta(i, j)
>>> KroneckerDelta(i, i)
1
>>> KroneckerDelta(i, i +
0
>>> KroneckerDelta(i, i +
KroneckerDelta(i, i + k +
i, j, k
1)
1 + k)
1)
classmethod eval(i, j)
Evaluates the discrete delta function.
Examples
>>> from sympy.functions.special.tensor_functions import KroneckerDelta
>>> from sympy.abc import i, j, k
>>> KroneckerDelta(i,
KroneckerDelta(i, j)
>>> KroneckerDelta(i,
1
>>> KroneckerDelta(i,
0
>>> KroneckerDelta(i,
KroneckerDelta(i, i +
j)
i)
i + 1)
i + 1 + k)
k + 1)
# indirect doctest
indices contain equal information
Returns True if indices are either both above or below fermi.
Examples
>>> from sympy.functions.special.tensor_functions import KroneckerDelta
>>> from sympy import Symbol
>>> a = Symbol(a, above_fermi=True)
>>> i = Symbol(i, below_fermi=True)
>>> p = Symbol(p)
>>> q = Symbol(q)
>>> KroneckerDelta(p, q).indices_contain_equal_information
True
>>> KroneckerDelta(p, q+1).indices_contain_equal_information
5.8. Functions Module
447
SymPy Documentation, Release 0.7.6
True
>>> KroneckerDelta(i, p).indices_contain_equal_information
False
is above fermi
True if Delta can be non-zero above fermi
See Also:
is below fermi, is only below fermi, is only above fermi
Examples
>>> from sympy.functions.special.tensor_functions import KroneckerDelta
>>> from sympy import Symbol
>>> a = Symbol(a, above_fermi=True)
>>> i = Symbol(i, below_fermi=True)
>>> p = Symbol(p)
>>> q = Symbol(q)
>>> KroneckerDelta(p, a).is_above_fermi
True
>>> KroneckerDelta(p, i).is_above_fermi
False
>>> KroneckerDelta(p, q).is_above_fermi
True
is below fermi
True if Delta can be non-zero below fermi
See Also:
is above fermi, is only above fermi, is only below fermi
Examples
>>> from sympy.functions.special.tensor_functions import KroneckerDelta
>>> from sympy import Symbol
>>> a = Symbol(a, above_fermi=True)
>>> i = Symbol(i, below_fermi=True)
>>> p = Symbol(p)
>>> q = Symbol(q)
>>> KroneckerDelta(p, a).is_below_fermi
False
>>> KroneckerDelta(p, i).is_below_fermi
True
>>> KroneckerDelta(p, q).is_below_fermi
True
is only above fermi
True if Delta is restricted to above fermi
See Also:
is above fermi, is below fermi, is only below fermi
448
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.functions.special.tensor_functions import KroneckerDelta
>>> from sympy import Symbol
>>> a = Symbol(a, above_fermi=True)
>>> i = Symbol(i, below_fermi=True)
>>> p = Symbol(p)
>>> q = Symbol(q)
>>> KroneckerDelta(p, a).is_only_above_fermi
True
>>> KroneckerDelta(p, q).is_only_above_fermi
False
>>> KroneckerDelta(p, i).is_only_above_fermi
False
is only below fermi
True if Delta is restricted to below fermi
See Also:
is above fermi, is below fermi, is only above fermi
Examples
>>> from sympy.functions.special.tensor_functions import KroneckerDelta
>>> from sympy import Symbol
>>> a = Symbol(a, above_fermi=True)
>>> i = Symbol(i, below_fermi=True)
>>> p = Symbol(p)
>>> q = Symbol(q)
>>> KroneckerDelta(p, i).is_only_below_fermi
True
>>> KroneckerDelta(p, q).is_only_below_fermi
False
>>> KroneckerDelta(p, a).is_only_below_fermi
False
killable index
Returns the index which is preferred to substitute in the nal expression.
The index to substitute is the index with less information regarding fermi level. If
indices contain same information, a is preferred before b.
See Also:
preferred index
Examples
>>>
>>>
>>>
>>>
>>>
>>>
>>>
p
from sympy.functions.special.tensor_functions import KroneckerDelta
from sympy import Symbol
a = Symbol(a, above_fermi=True)
i = Symbol(i, below_fermi=True)
j = Symbol(j, below_fermi=True)
p = Symbol(p)
KroneckerDelta(p, i).killable_index
5.8. Functions Module
449
SymPy Documentation, Release 0.7.6
>>> KroneckerDelta(p, a).killable_index
p
>>> KroneckerDelta(i, j).killable_index
j
preferred index
Returns the index which is preferred to keep in the nal expression.
The preferred index is the index with more information regarding fermi level. If
indices contain same information, a is preferred before b.
See Also:
killable index
Examples
>>>
>>>
>>>
>>>
>>>
>>>
>>>
i
>>>
a
>>>
i
from sympy.functions.special.tensor_functions import KroneckerDelta
from sympy import Symbol
a = Symbol(a, above_fermi=True)
i = Symbol(i, below_fermi=True)
j = Symbol(j, below_fermi=True)
p = Symbol(p)
KroneckerDelta(p, i).preferred_index
KroneckerDelta(p, a).preferred_index
KroneckerDelta(i, j).preferred_index
5.9 Geometric Algebra Module
A module for geometric algebra
Documentation for Geometric Algebra module.
Contents:
5.9.1 Geometric Algebra
Author Alan Bromborsky
450
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Abstract
This document describes the implementation, installation and use of a geometric algebra
module written in python that utilizes the sympy (page 704) symbolic algebra library. The
python module galgebra has been developed for coordinate free calculations using the
operations (geometric, outer, and inner products etc.) of geometric algebra. The operations can be dened using a completely arbitrary metric dened by the inner products of
a set of arbitrary vectors or the metric can be restricted to enforce orthogonality and signature constraints on the set of vectors. In addition the module includes the geometric,
outer (curl) and inner (div) derivatives and the ability to dene a curvilinear coordinate
system. The module requires the sympy module and the numpy module for numerical
linear algebra calculations. For latex output a latex distribution must be installed.
What is Geometric Algebra?
Geometric algebra is the Cliord algebra of a real nite dimensional vector space or the
algebra that results when a real nite dimensional vector space is extended with a product
of vectors (geometric product) that is associative, left and right distributive, and yields a real
number for the square (geometric product) of any vector [Hestenes] (page 1909), [Doran]
(page 1909). The elements of the geometric algebra are called multivectors and consist of
the linear combination of scalars, vectors, and the geometric product of two or more vectors.
The additional axioms for the geometric algebra are that for any vectors a, b, and c in the base
vector space ([Doran] (page 1909),p85):
a (bc) = (ab) c
a (b + c) = ab + ac
(a + b) c = ac + bc
aa = a2 <
By induction the rst three axioms also apply to any multivectors. The dot product of two
vectors is dened by ([Doran] (page 1909),p86)
a b (ab + ba)/2
Then consider
c=a+b
2
c = (a + b)2
c2 = a2 + ab + ba + b2
a b = (c2 a2 b2 )/2 <
Thus a b is real. The objects generated from linear combinations of the geometric products of
vectors are called multivectors. If a basis for the underlying vector space is the set of vectors
formed from e1 , . . . , en a complete basis for the geometric algebra is given by the scalar 1, the
vectors e1 , . . . , en and all geometric products of vectors
ei1 ei2 . . . eir where 0 r n, 0 ij n and 0 < i1 < i2 < < ir n
Each base of the complete basis is represented by a noncommutative symbol (except for the
scalar 1) with name ei1 . . . eir so that the general multivector A is represented by (A is the
scalar part of the multivector and the Ai1 ,...,ir are scalars)
A=A+
Ai1 ,...,ir ei1 ei2 . . . er
r=1 i1 ,...,ir , 0ij n
5.9. Geometric Algebra Module
451
SymPy Documentation, Release 0.7.6
The critical operation in setting up the geometric algebra is reducing the geometric product
of any two bases to a linear combination of bases so that we can calculate a multiplication
table for the bases. Since the geometric product is associative we can use the operation (by
denition for two vectors a b (ab + ba)/2 which is a scalar)
eij+1 eij = 2eij+1 eij eij eij+1
(5.1)
These processes are repeated untill every basis list in A is in normal (ascending) order with
no repeated elements. As an example consider the following
e3 e2 e1 = (2(e2 e3 ) e2 e3 )e1
= 2(e2 e3 )e1 e2 e3 e1
= 2(e2 e3 )e1 e2 (2(e1 e3 ) e1 e3 )
= 2((e2 e3 )e1 (e1 e3 )e2 ) + e2 e1 e3
= 2((e2 e3 )e1 (e1 e3 )e2 + (e1 e2 )e3 ) e1 e2 e3
which results from repeated application of equation 5.1 (page 452). If the product of basis
vectors contains repeated factors equation 5.1 (page 452) can be used to bring the repeated
factors next to one another so that if eij = eij+1 then eij eij+1 = eij eij+1 which is a scalar that
commutes with all the terms in the product and can be brought to the front of the product.
Since every repeated pair of vectors in a geometric product of r factors reduces the number
of noncommutative factors in the product by r 2. (The
) number of bases in the multivector
algebra is 2n and the number containing r factors is nr which is the number of combinations
or n things taken r at a time (binominal coecient).
The other construction required for formulating the geometric algebra is the outer or wedge
product (symbol ) of r vectors denoted by a1 ar . The wedge product of r vectors is
called an r-blade and is dened by ([Doran] (page 1909),p86)
a1 ar
ij1 ...ijr aij1 . . . aij1
ij1 ...ijr
where ij1 ...ijr is the contravariant permutation symbol which is +1 for an even permutation
of the superscripts, 0 if any superscripts are repeated, and 1 for an odd permutation of the
superscripts. From the denition a1 ar is antisymmetric in all its arguments and the
following relation for the wedge product of a vector a and an r-blade Br can be derived
a Br = (aBr + (1)r Br a)/2
(5.2)
Using equation 5.2 (page 452) one can represent the wedge product of all the basis vectors
in terms of the geometric product of all the basis vectors so that one can solve (the system of
equations is lower diagonal) for the geometric product of all the basis vectors in terms of the
wedge product of all the basis vectors. Thus a general multivector B can be represented as
a linear combination of a scalar and the basis blades.
B=B+
B i1 ,...,ir ei1 ei2 er
r=1 i1 ,...,ir , 0ij n
Using the blades ei1 ei2 er creates a graded algebra where r is the grade of the basis
blades. The grade-r part of B is the linear combination of all terms with grade r basis blades.
The scalar part of B is dened to be grade-0. Now that the blade expansion of B is dened
we can also dene the grade projection operator hBir by
hBir =
B i1 ,...,ir ei1 ei2 er
i1 ,...,ir , 0ij n
452
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
and
hBi hBi0 = B
Then if Ar is an r-grade multivector and B s is an s-grade multivector we have
Ar B s = hAr B s i|rs| + hAr B s i|rs|+2 + hAr B s ir+s
and dene ([Hestenes] (page 1909),p6)
Ar B s hAr B s ir+s
{
r or s 6= 0 :
Ar B s
r or s = 0 :
hAr B s i|rs|
0
where Ar B s is called the dot or inner product of two pure grade multivectors. For the case
of two non-pure grade multivectors
AB =
hAir hBis
r,s
AB =
hAir hBis
r,s6=0
Two other products, the right (c) and left (b) contractions, are dened by
{ hAr B s i
rs
AbB
0
r,s
{ hAr B s i
sr
AcB
0
r,s
rs
r<s
sr
s<r
}
}
A nal operation for multivectors is the reverse. If a multivector A is the geometric product
of r vectors (versor) so that A = a1 . . . ar the reverse is dened by
A ar . . . a 1
where for a general multivector we have (the the sum of the reverse of versors)
A = A +
(1)r(r1)/2
r=1
Ai1 ,...,ir ei1 ei2 er
i1 ,...,ir , 0ij n
note that if A is a versor then AA < and (AA 6= 0)
A1 =
A
AA
Representation of Multivectors in Sympy
The sympy python module oers a simple way of representing multivectors using linear combinations of commutative expressions (expressions consisting only of commuting sympy objects) and noncommutative symbols. We start by dening n noncommutative sympy symbols
5.9. Geometric Algebra Module
453
SymPy Documentation, Release 0.7.6
(e_1,...,e_n) = symbols(e_1,...,e_n,commutative=False)
Several software packages for numerical geometric algebra calculations are available from
Doran-Lasenby group and the Dorst group. Symbolic packages for Cliord algebra using
orthongonal bases such as ei ej + ej ei = 2ij , where ij is a numeric array are available in
Maple and Mathematica. The symbolic algebra module, galgebra, developed for python does
not depend on an orthogonal basis representation, but rather is generated from a set of n
arbitrary symbolic vectors, e1 , e2 , . . . , en and a symbolic metric tensor gij = ei ej .
In order not to reinvent the wheel all scalar symbolic algebra is handled by the python module sympy (page 704) and the abstract basis vectors are encoded as noncommuting sympy
symbols.
The basic geometic algebra operations will be implemented in python by dening a multivector class, MV, and overloading the python operators in Table 5.1 (page 454) where A and B
are any two multivectors (In the case of +, -, *, , and | the operation is also dened if A or B
is a sympy symbol or a sympy real number).
Operation
A+B
A-B
A*B
AB
A|B
A<B
A>B
Result
sum of multivectors
dierence of multivectors
geometric product
outer product of multivectors
inner product of multivectors
left contraction of multivectors
right contraction of multivectors
Table 5.1 (page 454). Multivector operations for galgebra
Since < and > have no r-forms (in python for the < and > operators there are no rlt () and
rlt () member functions to overload) we can only have mixed modes (scalars and multivectors) if the rst operand is a multivector.
Note: Except for < and > all the multivector operators have r-forms so that as long as one of
the operands, left or right, is a multivector the other can be a multivector or a scalar (sympy
symbol or integer).
Warning: Note that the operator order precedence is determined by python and is not
necessarily that used by geometric algebra. It is absolutely essential to use parenthesis
in multivector expressions containing , |, <, and/or >. As an example let A and B be any
two multivectors. Then A + A*B = A +(A*B), but A+AB = (2*A)B since in python the
operator has a lower precedence than the + operator. In geometric algebra the outer
and inner products and the left and right contractions have a higher precedence than the
geometric product and the geometric product has a higher precedence than addition and
subtraction. In python the , |, <, and > all have a lower precedence than + and - while *
has a higher precedence than + and -.
For those users who wish to dene a default operator precedence the functions dene precedence() and GAeval() are available in the module galgebra/precedence.
sympy.galgebra.precedence.define precedence(gd, op ord=<>|, , *)
Dene the precedence of the multivector operations. The function dene precedence()
must be called from the main program and the rst argument gd must be set to globals().
The second argument op ord determines the operator precedence for expressions input
to the function GAeval(). The default value of op ord is <>|,,*. For the default value
the <, >, and | operations have equal precedence followed by , and is followed by *.
454
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
sympy.galgebra.precedence.GAeval(s, pstr=False)
The function GAeval() returns a multivector expression dened by the string s where
the operations in the string are parsed according to the precedences dened by dene precedence(). pstr is a ag to print the input and output of GAeval() for debugging
purposes. GAeval() works by adding parenthesis to the input string s with the precedence dened by op ord=<>|,,*. Then the parsed string is converted to a sympy expression using the python eval() function. For example consider where X, Y, Z, and W
are multivectors
define_precedence(globals())
V = GAeval(X|Y^Z*W)
The sympy variable V would evaluate to ((X|Y)Z)*W.
Vector Basis and Metric
The two structures that dene the MV (multivector) class are the symbolic basis vectors and
the symbolic metric. The symbolic basis vectors are input as a string with the symbol name
separated by spaces. For example if we are calculating the geometric algebra of a system with
three vectors that we wish to denote as a0, a1, and a2 we would dene the string variable:
basis = a0 a1 a2
that would be input into the multivector setup function. The next step would be to dene the
symbolic metric for the geometric algebra of the basis we have dened. The default metric
is the most general and is the matrix of the following symbols
(a0.a0) (a0.a1) (a0.a2)
(5.3)
g = (a0.a1) (a1.a1) (a1.a2)
(a0.a2) (a1.a2) (a2.a2)
where each of the gij is a symbol representing all of the dot products of the basis vectors. Note
that the symbols are named so that gij = gji since for the symbol function (a0.a1) 6= (a1.a0).
Note that the strings shown in equation 5.3 (page 455) are only used when the values of gij
are output (printed). In the galgebra module (library) the gij symbols are stored in a static
member of the multivector class MV as the sympy matrix MV.metric (gij = MV.metric[i,j]).
The default denition of g can be overwritten by specifying a string that will dene g. As an
example consider a symbolic representation for conformal geometry. Dene for a basis
basis = a0 a1 a2 n nbar
and for a metric
metric = # # # 0 0, # # # 0 0, # # # 0 0, 0 0 0 0 2, 0 0 0 2 0
then calling MV.setup(basis,metric) would initialize the metric
(a0.a0) (a0.a1) (a0.a2) 0 0
(a0.a1) (a1.a1) (a1.a2) 0 0
g=
(a0.a2) (a1.a2) (a2.a2) 0 0
0
0
0
0 2
0
0
0
2 0
tensor
Here we have specied that n and nbar are orthonal to all the as, (n.n) = (nbar.nbar) = 0, and
(n.nbar) = 2. Using # in the metric denition string just tells the program to use the default
symbol for that value.
5.9. Geometric Algebra Module
455
SymPy Documentation, Release 0.7.6
When MV.setup is called multivector representations of the basis local to the program are
instantiated. For our rst example that means that the symbolic vectors named a0, a1, and
a2 are created and returned from MV.setup via a tuple as in (a_1,a_2,a3) = MV.setup(a_1 a_2 a_3,metric=metric)
Note that the python variable name for a basis vector does not have to correspond to the
name give in MV.setup(), one may wish to use a shorted python variable name to reduce
programming (typing) errors, for example one could use (a1,a2,a3) = MV.setup(a_1 a_2 a_3,metric=metric)
or
(g1,g2,g3) = MV.setup(gamma_1 gamma_2 gamma_3,metric=metric)
so that if the latex printer is used e1 would print as e1 and g1 as 1 .
Note: Additionally MV.setup has simpied options for naming a set of basis vectors and for
inputing an othogonal basis.
If one wishes to name the basis vectors ex , ey , and ez then set basis=e*x|y|z or to name t ,
x , y , and z then set basis=gamma*t|x|y|z.
For the case of an othogonal basis if the signature of the vector space is (1, 1, 1) (Euclidian 3space) set metric=[1,1,1] or if it is (1, 1, 1, 1) (Minkowsi 4-space) set metric=[1,-1,-1,-1].
Representation and Reduction of Multivector Bases
In our symbolic geometric algebra all multivectors can be obtained from the symbolic basis
vectors we have input, via the dierent operations available to geometric algebra. The rst
problem we have is representing the general multivector in terms terms of the basis vectors.
To do this we form the ordered geometric products of the basis vectors and develop an internal representation of these products in terms of python classes. The ordered geometric
products are all multivectors of the form ai1 ai2 . . . air where i1 < i2 < < ir and r n. We call
these multivectors bases and represent them internally with noncommutative symbols so for
example a1 a2 a3 is represented by
Symbol(a_1*a_2*a_3,commutative=False)
In the simplist case of two basis vectors a 1 and a 2 we have a list of bases
MV.bases = [[Symbol(ONE,commutative=False)],[Symbol(a_1,commutative=False),\
Symbol(a_2,commutative=False)],[Symbol(a_1*a_2,commutative=False)]]
Note: The reason that the base for the scalar component of the multivector is dened as
Symbol(ONE,commutative=False), a noncommutative symbol is because of the properties
of the left and right contraction operators which are non commutative if one is contracting a
multivector with a scalar.
For the case of the basis blades we have
MV.blades = [[Symbol(ONE,commutative=False)],[Symbol(a_1,commutative=False),\
Symbol(a_2,commutative=False)],[Symbol(a_1^a_2,commutative=False)]]
456
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Note: For all grades/pseudo-grades greater than one (vectors) the * in the name of the
base symbol is replaced with a in the name of the blade symbol so that for all basis bases
and blades of grade/pseudo-grade greater than one there are dierent symbols for the corresponding bases and blades.
The function that builds all the required arrays and dictionaries upto the base multiplication
table is shown below. MV.dim is the number of basis vectors and the combinations functions
from itertools constructs the index tupels for the bases of each pseudo grade. Then the noncommutative symbol representing each base is constructed from each index tuple. MV.ONE
is the noncommutative symbol for the scalar base. For example if MV.dim = 3 then
MV.index = ((),((0,),(1,),(2,)),((0,1),(0,2),(1,2)),((0,1,2)))
Note: In the case that the metric tensor is diagonal (orthogonal basis vectors) both base
and blade bases are identical and fewer arrays and dictionaries need to be constructed.
@staticmethod
def build_base_blade_arrays(debug):
indexes = tuple(range(MV.dim))
MV.index = [()]
for i in indexes:
MV.index.append(tuple(combinations(indexes,i+1)))
MV.index = tuple(MV.index)
#Set up base and blade and index arrays
if not MV.is_orthogonal:
MV.bases_flat = []
MV.bases = [MV.ONE]
MV.base_to_index = {MV.ONE:()}
MV.index_to_base = {():MV.ONE}
MV.base_grades
= {MV.ONE:0}
MV.base_grades[ONE] = 0
MV.blades = [MV.ONE]
MV.blades_flat = []
MV.blade_grades
= {MV.ONE:0}
MV.blade_grades[ONE] = 0
MV.blade_to_index = {MV.ONE:()}
MV.index_to_blade = {():MV.ONE}
ig = 1 #pseudo grade and grade index
for igrade in MV.index[1:]:
if not MV.is_orthogonal:
= [] #base symbol array within pseudo grade
bases
blades
= [] #blade symbol array within grade
ib = 0 #base index within grade
for ibase in igrade:
#build base name string
(base_sym,base_str,blade_sym,blade_str) = MV.make_base_blade_symbol(ibase)
if not MV.is_orthogonal:
bases.append(base_sym)
MV.bases_flat.append(base_sym)
5.9. Geometric Algebra Module
457
SymPy Documentation, Release 0.7.6
blades.append(blade_sym)
MV.blades_flat.append(blade_sym)
base_index = MV.index[ig][ib]
#Add to dictionarys relating symbols and indexes
if not MV.is_orthogonal:
MV.base_to_index[base_sym]
= base_index
MV.index_to_base[base_index] = base_sym
MV.base_grades[base_sym]
= ig
MV.blade_to_index[blade_sym] = base_index
MV.index_to_blade[base_index] = blade_sym
MV.blade_grades[blade_sym] = ig
ib += 1
ig += 1
if not MV.is_orthogonal:
MV.bases.append(tuple(bases))
MV.blades.append(tuple(blades))
if not MV.is_orthogonal:
= tuple(MV.bases)
MV.bases
MV.bases_flat = tuple(MV.bases_flat)
MV.bases_flat1 = (MV.ONE,)+MV.bases_flat
MV.bases_set
= set(MV.bases_flat[MV.dim:])
MV.blades
MV.blades_flat
MV.blades_flat1
MV.blades_set
=
=
=
=
tuple(MV.blades)
tuple(MV.blades_flat)
(MV.ONE,)+MV.blades_flat
set(MV.blades_flat[MV.dim:])
return
Base Representation of Multivectors
In terms of the bases dened as noncommutative sympy symbols the general multivector is a
linear combination (scalar sympy coecients) of bases so that for the case of two bases the
most general multivector is given by A = A_0*MV.bases[0][0]+A__1*MV.bases[1][0]+A__2*MV.bases[1][1]+A__12*MV.bases[2][0]
If we have another multivector B to multiply with A we can calculate the product in terms of
a linear combination of bases if we have a multiplication table for the bases.
Blade Representation of Multivectors
Since we can now calculate the symbolic geometric product of any two multivectors we can
also calculate the blades corresponding to the product of the symbolic basis vectors using the
formula
1
r
Ar b = (Ar b (1) bAr ) ,
2
where Ar is a multivector of grade r and b is a vector. For our example basis the result is
shown in Table 5.3 (page 458).
458
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
1 = 1
a0 = a0
a1 = a1
a2 = a2
a0a1 = {-(a0.a1)}1+a0a1
a0a2 = {-(a0.a2)}1+a0a2
a1a2 = {-(a1.a2)}1+a1a2
a0a1a2 = {-(a1.a2)}a0+{(a0.a2)}a1+{-(a0.a1)}a2+a0a1a2
Table 5.3 (page 458). Bases blades in terms of bases.
The important thing to notice about Table 5.3 (page 458) is that it is a triagonal (lower triangular) system of equations so that using a simple back substitution algorithm we can solve
for the pseudo bases in terms of the blades giving Table 5.4 (page 459).
1 = 1
a0 = a0
a1 = a1
a2 = a2
a0a1 = {(a0.a1)}1+a0a1
a0a2 = {(a0.a2)}1+a0a2
a1a2 = {(a1.a2)}1+a1a2
a0a1a2 = {(a1.a2)}a0+{-(a0.a2)}a1+{(a0.a1)}a2+a0a1a2
Table 5.4 (page 459). Bases in terms of basis blades.
Using Table 5.4 (page 459) and simple substitution we can convert from a base multivector
representation to a blade representation. Likewise, using Table 5.3 (page 458) we can convert
from blades to bases.
Using the blade representation it becomes simple to program functions that will calculate the
grade projection, reverse, even, and odd multivector functions.
Note that in the multivector class MV there is a class variable for each instantiation,
self.bladeg, that is set to False for a base representation and True for a blade representation. One needs to keep track of which representation is in use since various multivector
operations require conversion from one representation to the other.
Warning: When the geometric product of two multivectors is calculated the module looks
to see if either multivector is in blade representation. If either is the result of the geometric
product is converted to a blade representation. One result of this is that if either of the
multivectors is a simple vector (which is automatically a blade) the result will be in a blade
representation. If a and b are vectors then the result a*b will be (a.b)+ab or simply ab
if (a.b) = 0.
Outer and Inner Products, Left and Right Contractions
In geometric algebra any general multivector A can be decomposed into pure grade multivectors (a linear combination of blades of all the same order) so that in a n-dimensional vector
space
n
A=
Ar
r=0
The geometric product of two pure grade multivectors Ar and Bs has the form
Ar Bs = hAr Bs i|rs| + hAr Bs i|rs|+2 + + hAr Bs ir+s
5.9. Geometric Algebra Module
459
SymPy Documentation, Release 0.7.6
where hit projects the t grade components of the multivector argument. The inner and outer
products of Ar and Bs are then dened to be
Ar Bs = hAr Bs i|rs|
Ar Bs = hAr Bs ir+s
and
AB =
Ar Bs
r,s
AB =
Ar Bs
r,s
Likewise the right (b) and left (c) contractions are dened as
{
}
hAr Bs irs r s
Ar bBs =
0
r<s
{
}
hAr Bs isr s r
Ar cBs =
0
s<r
and
AbB =
Ar bBs
r,s
AcB =
Ar cBs
r,s
Warning: In the MV class we have overloaded the operator to represent the outer
product so that instead of calling the outer product function we can write mv1 mv2. Due
to the precedence rules for python it is absolutely essential to enclose outer products in
parenthesis.
Warning: In the MV class we have overloaded the | operator for the inner product, >
operator for the right contraction, and < operator for the left contraction. Instead of calling
the inner product function we can write mv1|mv2, mv1>mv2, or mv1<mv2 respectively
for the inner product, right contraction, or left contraction. Again, due to the precedence
rules for python it is absolutely essential to enclose inner products and/or contractions
in parenthesis.
Reverse of Multivector
If A is the geometric product of r vectors
A = a1 . . . ar
then the reverse of A designated A is dened by
A ar . . . a 1 .
The reverse is simply the product with the order of terms reversed. The reverse of a sum of
products is dened as the sum of the reverses so that for a general multivector A we have
A =
hAii
i=0
460
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
but
hAii = (1)
i(i1)
2
hAii
(5.4)
which is proved by expanding the blade bases in terms of orthogonal vectors and showing
that equation 5.4 (page 461) holds for the geometric product of orthogonal vectors.
The reverse is important in the theory of rotations in n-dimensions. If R is the product of an
even number of vectors and RR = 1 then RaR is a composition of rotations of the vector a.
If R is the product of two vectors then the plane that R denes is the plane of the rotation.
That is to say that RaR rotates the component of a that is projected into the plane dened by
2U
a and b where
( 2 R = )ab. R may be written R = e , where is the angle of rotation and u is a
unit blade u = 1 that denes the plane of rotation.
Reciprocal Frames
If we have M linearly independent vectors (a frame), a1 , . . . , aM , then the reciprocal frame is
a1 , . . . , aM where ai aj = ij , ij is the Kronecker delta (zero if i 6= j and one if i = j). The
reciprocal frame is constructed as follows:
EM = a1 aM
1
EM
=
Then
i1
ai = (1)
EM
2
EM
1
(a1 a
i aM ) EM
where a
i indicates that ai is to be deleted from the product. In the standard notation if a
vector is denoted with a subscript the reciprocal vector is denoted with a superscript. The
multivector setup function MV.setup(basis,metric,rframe) has the argument rframe with a
default value of False. If it is set to True the reciprocal frame of the basis vectors is calculated.
Additionally there is the function reciprocal frame(vlst,names=) external to the MV class
that will calculate the reciprocal frame of a list, vlst, of vectors. If the argument names is set
to a space delimited string of names for the vectors the reciprocal vectors will be given these
names.
Geometric Derivative
If F is a multivector eld that is a function of a vector x = xi ei (we are using the summation
convention that pairs of subscripts and superscripts are summed over the dimension of the
vector space) then the geometric derivative F is given by (in this section the summation
convention is used):
F
F = ei i
x
If FR is a grade-R multivector and FR = FRi1 ...iR ei1 eiR then
FR =
FRi1 ...iR j
e (ei1 eiR )
xj
Note that ej (ei1 eiR ) can only contain grades R 1 and R + 1 so that FR also can only
contain those grades. For a grade-R multivector FR the inner (div) and outer (curl) derivatives
are dened as
FR = hFR iR1
and
FR = hFR iR+1
5.9. Geometric Algebra Module
461
SymPy Documentation, Release 0.7.6
For a general multivector function F the inner and outer derivatives are just the sum of the
inner and outer dervatives of each grade of the multivector function.
Curvilinear coordinates are derived from a vector function x() where = (1 , . . . , N ) where
the number of coordinates is equal to the dimension of the vector space. In the case of 3dimensional spherical coordinates = (r, , ) and the coordinate generating function x()
is
x = r cos () sin () ex + r sin () sin () ey + r cos () ez
x
A coordinate frame is derived from x by ei = i . The following show the frame for spherical
coordinates.
er = cos () sin () ex + sin () sin () ey + cos () ez
e = cos () cos () ex + r cos () sin () ey r sin () ez
e = r sin () sin () ex + r cos () sin () ey
The coordinate frame generated in this manner is not necessarily normalized so dene a
normalized frame by
ei
ei
e
i = 2 =
|ei |
|ei |
2
This works for all ei 6= 0 since we have dened |ei | = |e2i |. For spherical coordinates the
normalized frame vectors are
e
r = cos () sin () ex + sin () sin () ey + cos () ez
e
= cos () cos () ex + cos () sin () ey sin () ez
e
= sin () ex + cos () ey
The geometric derivative in curvilinear coordinates is given by
)
(
FR = ei i FRi1 ...iR e
i1 e
iR
x
( i1 ...iR
)
j
=e
FR
e
i1 e
iR
j
(
)
i1 ...iR
=
F
ej (
e i1 e
iR ) + FRi1 ...iR ej j (
ei1 e
iR )
j R
(
)
i1 ...iR
=
F
ej (
e i1 e
iR ) + FRi1 ...iR C {
ei1 e
iR }
j R
where
(
e i1 e
iR )
j
are the connection multivectors for the curvilinear coordinate system. For a spherical coordinate system they are
2
C {
er } =
r
cos ()
1
C {
e } =
+ e
r e
r sin () r
1
cos ()
C {
e } = e
r e
+
e
e
r
r sin ()
cos ()
1
C {
er e } =
e
r + e
r sin ()
r
1
cos ()
C {
er e
} = e
e
r e
e
r
r sin ()
2
C {
e e
} = e
r e
e
r
C {
er e
e
} = 0
C {
e i1 e
iR } = ej
462
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Numpy, LaTeX, and Ansicon Installation
To install the geometric algebra module on windows,linux, or OSX perform the following operations
1. Install sympy. galgebra is included in sympy.
2. To install texlive in linux or windows
(a) Go to <http://www.tug.org/texlive/acquire-netinstall.html> and click on installtl.zip o download
(b) Unzip install-tl.zip anywhere on your machine
(c) Open the le readme.en.html in the readme-html.dir directory. This le contains
the information needed to install texlive.
(d) Open a terminal (console) in the install-tl-XXXXXX directory
(e) Follow the instructions in readme.en.html le to run the install-tl.bat le in windows or the install-tl script le in linux.
3. For OSX install mactex from <http://tug.org/mactex/>.
4. Install python-nympy if you want to calculate numerical matrix functons (determinant,
inverse,
eigenvalues,
etc.).
For windows go to
<http://sourceforge.net/projects/numpy/les/NumPy/1.6.2/>
and
install
the
distribution of numpy appropriate for your system.
For OSX go to
<http://sourceforge.net/projects/numpy/les/NumPy/1.6.1/>.
5. It is strongly suggested that you go to <http://www.geany.org/Download/Releases> and
install the version of the geany editor appropriate for your system.
6. If you wish to use enhance print on windows (a) Go to <https://github.com/adoxa/ansicon/downloads> and download ansicon
(b) In the Edit -> Preferences -> Tools menu of geany enter into the Terminal input
the full path of ansicon.exe
After installation if you are doing you code development in the galgebra directory you need
only include
from sympy.galgebra.printing import xdvi,enhance_print
from sympy.galgebra.ga import *
to use the galgebra module.
In addition to the code shown in the examples section of this document there are more examples in the Examples directory under the galgebra directory.
Module Components
Initializing Multivector Class
The multivector class is initialized with:
sympy.galgebra.ga.MV.setup(basis, metric=None, coords=None, rframe=False, debug=False, curv=(None, None))
The basis and metric parameters were described in section Vector Basis and Metric
(page 455). If rframe=True the reciprocal frame of the symbolic bases vectors is calculated. If debug=True the data structure required to initialize the MV class are printer
out. coords is a tuple of sympy (page 704) symbols equal in length to the number of basis
5.9. Geometric Algebra Module
463
SymPy Documentation, Release 0.7.6
vectors. These symbols are used as the arguments of a multivector eld as a function
of position and for calculating the derivatives of a multivector eld (if coords is dened
then rframe is automatically set equal to True). Additionally, MV.setup() calculates the
pseudo scalar, I and makes them available to the programmer as MV.I and MV.Iinv.
MV.setup() always returns a tuple containing the basis vectors (as multivectors) so that
if we have the code
(e1,e2,e3) = MV.setup(e_1 e_2 e_3)
then we can dene a multivector by the expression
(a1,a2,a3) = symbols(a__1 a__2 a__3)
A = a1*e1+a2*e2+a3*e3
Another option is
(e1,e2,e3) = MV.setup(e*1|2|3)
which produce the same results as the previous method. Note that if we had used
(e1,e2,e3) = MV.setup(e*x|y|z)
then the basis vectors would have been labeled e x, e y, and e z. If coords is dened then
MV.setup() returns the tuple
X = (x,y.z) = symbols(x y z)
(ex,ey,ez,grad) = MV.setup(e,coords=X)
the basis vectros are again labeled e x, e y, and e z and the additional vector grad is
returned. grad acts as the gradient operator (geometric derivative) so that if F() is a
multivector function of (x,y,z) then
DFl = grad*F
DFr = F*grad
are the left and right geometric derivatives of F().
The nal parameter in MV.setup() is curv which denes a curvilinear coordinate system.
If 3-dimensional spherical coordinates are required we would dene X = (r,th,phi) = symbols(r theta phi)
curv = [[r*cos(phi)*sin(th),r*sin(phi)*sin(th),r*cos(th)],[1,r,r*sin(th)]]
(er,eth,ephi,grad) = MV.setup(e_r e_theta e_phi,metric=[1,1,1],coords=X,curv=curv)
The rst component of curv is
[r*cos(phi)*sin(th),r*sin(phi)*sin(th),r*cos(th)]
This is the position vector for the spherical coordinate system expressed in terms of the
rectangular coordinate components given in terms of the spherical coordinates r, th, and
phi. The second component of curv is
[1,r,r*sin(th)]
The components of curv[1] are the normalizing factors for the basis vectors of the spherical coordinate system that are calculated from the derivatives of curv[0] with respect to
the coordinates r, th, and phi. In theory the normalizing factors can be calculated from
the derivatives of curv[0]. In practice one cannot currently specify in sympy that the
square of a function is always positive which leads to problems when the normalizing
464
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
factor is the square root of a squared function. To avoid these problems the normalizing
factors are explicitly dened in curv[1].
Note: In the case of curvlinear coordinates debug also prints the connection multivectors.
Instantiating a Multivector
Now that grades and bases have been described we can show all the ways that a multivector
can be instantiated. As an example assume that the multivector space is initialized with
(e1,e2,e3) = MV.setup(e_1 e_2 e_3)
then multivectors could be instantiated with
(a1,a2,a3) = symbols(a__1 a__2 a__3)
x = a1*e1+a2*e2+a3*e3
y = x*e1*e2
z = x|y
w = x^y
or with the multivector class constructor:
class sympy.galgebra.ga.MV(base=None, mvtype=None, fct=False, blade rep=True)
base is a string that denes the name of the multivector for output purposes. base and
mvtype are dened by the following table and fct is a switch that will convert the symbolic
coecients of a multivector to functions if coordinate variables have been dened when
MV.setup() is called:
mvtype
default
scalar
vector
grade2 or bivector
grade
pseudo
spinor
mv
default
default
base
default
string s
string s
string s
string s,n
string s
string s
string s
sympy scalar c
multivector
result
Zero multivector
symbolic scalar of value Symbol(s)
symbolic vector
symbolic bivector
symbolic n-grade multivector
symbolic pseudoscalar
symbolic even multivector
symbolic general multivector
zero grade multivector with coecient c
copy constructor for multivector
If the base argument is a string s then the coecients of the resulting multivector are
named as follows:
The grade r coecients consist of the base string, s, followed by a double underscore, , and an index string of r symbols. If coords is dened the index string
will consist of coordinate names in a normal order dened by the coords tuple.
If coords is not dened the index string will be integers in normal (ascending)
order (for an n dimensional vector space the indices will be 1 to n). The double
underscore is used because the latex printer interprets it as a superscript and
superscripts in the coecients will balance subscripts in the bases.
For example if If coords=(x,y,z) and the base is A, the list of all possible coecients for the most general multivector would be A, A x, A y, A z, A xy, A xz,
A yz, and A xyz. If the latex printer is used and e is the base for the basis vectors then the pseudo scalar would print as Axyz ex ey ez . If coordinates are
5.9. Geometric Algebra Module
465
SymPy Documentation, Release 0.7.6
not dened it would print as A123 e1 e2 e3 . For printed output all multivectors
are represented in terms of products of the basis vectors, either as geometric
products or wedge products. This is also true for the output of expressions
containing reciprocal basis vectors.
If the fct argument of MV() is set to True and the coords argument in MV.setup() is
dened the symbolic coecients of the multivector are functions of the coordinates.
Basic Multivector Class Functions
sympy.galgebra.ga.MV.convert to blades(self)
Convert multivector from the base representation to the blade representation. If multivector is already in blade representation nothing is done.
sympy.galgebra.ga.MV.convert from blades(self)
Convert multivector from the blade representation to the base representation. If multivector is already in base representation nothing is done.
sympy.galgebra.ga.MV.dd(self, v)
For a mutivector function F and a vector v then F.dd(v) is the directional derivate of F in
the direction v, (v )F .
sympy.galgebra.ga.MV.diff(self, var)
Calculate derivative of each multivector coecient with resepect to variable var and
form new multivector from coecients.
sympy.galgebra.ga.MV.dual(self)
Return dual of multivector which is multivector left multiplied by pseudoscalar MV.I
([Hestenes] (page 1909),p22).
sympy.galgebra.ga.MV.even(self)
Return the even grade components of the multivector.
sympy.galgebra.ga.MV.exp(self, alpha=1, norm=0, mode=T)
Return exponential of a blade (if self is not a blade error message is generated). If A is
the blade then eA is returned where the default mode, T, assumes AA < 0 so that
(
)
(
) A
eA = cos A2 + sin A2
.
A2
If the mode is not T then AA > 0 is assumed so that
( )
( ) A
eA = cosh A2 + sinh A2 .
A2
If norm = N > 0 then
eA = cos (N ) + sin (N )
A
N
or
eA = cosh (N ) + sinh (N )
A
N
depending on the value of mode.
sympy.galgebra.ga.MV.expand(self)
Return multivector in which each coecient has been expanded using sympy expand()
function.
sympy.galgebra.ga.MV.factor(self)
Apply the sympy factor function to each coecient of the multivector.
466
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
sympy.galgebra.ga.MV.func(self, fct)
Apply the sympy scalar function fct to each coecient of the multivector.
sympy.galgebra.ga.MV.grade(self, igrade=0)
Return a multivector that consists of the part of the multivector of grade equal to igrade.
If the multivector has no igrade part return a zero multivector.
sympy.galgebra.ga.MV.inv(self)
Return the inverse of the multivector if self*se.rev() is a nonzero ctor.
sympy.galgebra.ga.MV.norm(self)
Return the norm of the multvector M (M.norm()) dened by M M . If M M is a scalar
(a sympy scalar is returned). If M M in not a scalar the program exits with an error
message.
sympy.galgebra.ga.MV.norm(self)
Return the square of norm of the multvector M (M.norm2()) dened by M M . If M M is
a scalar (a sympy scalar is returned). If M M in not a scalar the program exits with an
error message.
sympy.galgebra.ga.MV.scalar(self)
Return the coecient (sympy scalar) of the scalar part of a multivector.
sympy.galgebra.ga.MV.simplify(self)
Return multivector where sympy simplify function has been applied to each coecient
of the multivector.
sympy.galgebra.ga.MV.subs(self, x)
Return multivector where sympy subs function has been applied to each coecient of
multivector for argument dictionary/list x.
sympy.galgebra.ga.MV.rev(self)
Return the reverse of the multivector. See section Reverse of Multivector (page 460).
sympy.galgebra.ga.MV.set coef(self, grade, base, value)
Set the multivector coecient of index (grade,base) to value.
sympy.galgebra.ga.MV.trigsimp(self, **kwargs)
Apply the sympy trignometric simplication fuction trigsimp to each coecient of the
multivector. **kwargs are the arguments of trigsimp. See sympy documentation on
trigsimp for more information.
Basic Multivector Functions
sympy.galgebra.ga.diagpq(p, q=0)
Returns string equivalent metric tensor for signature (p, q).
sympy.galgebra.ga.arbitrary metric(n)
Returns string equivalent metric tensor for arbitrary signature.
sympy.galgebra.ga.arbitrary metric conformal(n)
Returns string equivalent metric tensor for arbitrary signature (n+1,1).
sympy.galgebra.ga.Com(A, B)
Calulate commutator of multivectors A and B. Returns (AB BA)/2.
sympy.galgebra.ga.DD(v, f)
Calculate directional derivative of multivector function f in direction of vector v. Returns
f.dd(v).
sympy.galgebra.ga.Format(Fmode=True, Dmode=True, ipy=False)
See latex printing.
5.9. Geometric Algebra Module
467
SymPy Documentation, Release 0.7.6
sympy.galgebra.precedence.GAeval(s, pstr=False)
Returns multivector expression for string s with operator precedence for string s dened
by inputs to function dene precedence(). if pstr=True s and s with parenthesis added
to enforce operator precedence are printed.
sympy.galgebra.ga.Nga(x, prec=5)
If x is a multivector with coecients that contain oating point numbers, Nga() rounds
all these numbers to a precision of prec and returns the rounded multivector.
sympy.galgebra.ga.ReciprocalFrame(basis, mode=norm)
If basis is a list/tuple of vectors, ReciprocalFrame() returns a tuple of reciprocal vectors.
If mode=norm the vectors are normalized. If mode is anything other than norm the
vectors are unnormalized and the normalization coecient is added to the end of the
tuple. One must divide by the coecient to normalize the vectors.
sympy.galgebra.ga.ScalarFunction(TheFunction)
If TheFuction is a real sympy fuction a scalar multivector function is returned.
sympy.galgebra.ga.cross(M1, M2)
If M1 and M2 are 3-dimensional euclidian vectors the vector cross product is returned,
v1 v2 = I (v1 v2 ).
sympy.galgebra.precedence.define precedence(gd, op ord=<>|, , *)
This is used with the GAeval() fuction to evaluate a string representing a multivector
expression with a revised operator precedence. dene precedence() redenes the operator precedence for multivectors. dene precedence() must be called in the main program an the argument gd must be globals(). The argument op ord denes the order of
operator precedence from high to low with groups of equal precedence separated by
commas. the default precedence op ord=<>|,,* is that used by Hestenes ([Hestenes]
(page 1909),p7,[Doran] ,p38).
sympy.galgebra.ga.dual(M)
Return the dual of the multivector M, M I 1 .
sympy.galgebra.ga.inv(B)
If for the multivector B, BB is a nonzero scalar, return B 1 = B /(BB ).
sympy.galgebra.ga.proj(B, A)
Project blade A on blade B returning (AbB) B 1 .
sympy.galgebra.ga.refl(B, A)
Reect blade A in blade B. If r is grade of A and s is grade of B returns (1)s(r+1) BAB 1 .
sympy.galgebra.ga.rot(itheta, A)
Rotate blade A by 2-blade itheta. Is is assumed that itheta*itheta > 0 so that the rotation
is Euclidian and not hyperbolic so that the angle of rotation is theta = itheta.norm().
Ther in 3-dimensional Euclidian space. theta is the angle of rotation (scalar in radians)
and n is the vector axis of rotation. Returned is the rotor cos(theta)+sin(theta)*N where
N is the normalized dual of n.
Multivector Derivatives
The various derivatives of a multivector function is accomplished by multiplying the gradient
operator vector with the function. The gradiant operation vector is returned by the MV.setup()
function if coordinates are dened. For example if we have for a 3-D vector space
X = (x,y,z) = symbols(x y z)
(ex,ey,ez,grad) = MV.setup(e*x|y|z,metric=[1,1,1],coords=X)
468
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Then the gradient operator vector is grad (actually the user can give it any name he wants
to). Then the derivatives of the multivector function F are given by
F = MV(F,mv,fct=True)
F = grad F
F = F grad
F = grad F
F = F grad
F = grad|F
F F = F |grad
bF = grad < F
F b = F < grad
cF = grad > F
F c = F > grad
The preceding code block gives examples of all possible multivector derivatives of the multivector function F where * give the left and right geometric derivatives, gives the left and
right exterior (curl) derivatives, | gives the left and right interior (div) derivatives, < give the
left and right derivatives for the left contraction, and > give the left and right derivatives
for the right contraction. To understand the left and right derivatives see a reference on
geometric calculus ([Doran] (page 1909),chapter6).
If one is taking the derivative of a complex expression that expression should be in parenthesis. Additionally, whether or not one is taking the derivative of a complex expression the grad
vector and the expression it is operating on should always be in parenthesis unless the grad
operator and the expression it is operating on are the only objects in the expression.
Vector Manifolds
In addtition to the galgebra module there is a manifold module that allows for the denition of a geometric algebra and calculus on a vector manifold. The vector mainfold is
dened by a vector function of some coordinates in an embedding vector space ([Doran]
(page 1909),p202,[Hestenes] ,p139). For example the unit 2-sphere would be the collection
of vectors on the unit shpere in 3-dimensions with possible coordinates of and the angles
of elevation and azimuth. A vector function X (, ) that denes the manifold would be given
by
X (, ) = cos () ez + cos () (cos () ex + sin () ey )
The module manifold.py is transitionary in that all calculation are performed in the embedding vector space (geometric algebra). Thus due to the limitations on sympys simplify() and
trigsimp(), simple expressions may appear to be very complicated since they are expressed
in terms of the basis vectors (bases/blades) of the embedding space and not in terms of the
vector space (geometric algebra) formed from the manifolds basis vectors. A future implementation of Manifold.py will correct this diciency. The member functions of the vector
manifold follow.
sympy.galgebra.ga.Manifold(x, coords, debug=False, I=None)
Initializer for vector manifold where x is the vector function of the coords that denes
the manifold and coords is the list/tuple of sympy symbols that are the coordinates. The
basis vectors of the manifold as a fuction of the coordinates are returned as a tuple. I
is the pseudo scalar for the manifold. The default is for the initializer to calculate I,
however for complicated x functions (especially where trigonometric functions of the
5.9. Geometric Algebra Module
469
SymPy Documentation, Release 0.7.6
coordinates are involved) it is sometimes a good idea to calculate I separately and input
it to Manifold().
sympy.galgebra.ga.Basis(self)
Return the basis vectors of the manifold as a tuple.
sympy.galgebra.ga.DD(self, v, F, opstr=False)
Return the manifold directional derivative of a multivector function F dened on the
manifold in the vector direction v.
sympy.galgebra.ga.Grad(self, F)
Return the manifold multivector derivative of the multivector function F dened on the
manifold.
sympy.galgebra.ga.Proj(self, F)
Return the projection of the multivector F onto the manifold tangent space.
An example of a simple vector manifold is shown below which demonstrates the instanciation
of a manifold, the dening of vector and scalar functions on the manifold and the calculation
of the geometric derivative of those functions.
Standard Printing
Printing of multivectors is handled by the module galgebra/printing which contains a string
printer class, GA Printer derived from the sympy string printer class and a latex printer class,
470
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
GA LatexPrinter, derived from the sympy latex printer class. Additionally, there is an enhanced print class that enhances the console output of sympy to make the printed output
multivectors, functions, and derivatives more readable. enhanced print requires an ansi console such as is supplied in linux or the program ansicon (github.com/adoxa/ansicon) for windows which replaces cmd.exe.
For a windows user the simplest way to implement ansicon is to use the geany editor and in
the Edit->Preferences->Tools menu replace cmd.exe with ansicon.exe (be sure to supply the
path to ansicon).
If enhanced print is called in a program (linux) when multivectors are printed the basis blades
or bases are printed in bold text, functions are printed in red, and derivative operators in
green.
For formatting the multivector output there is the member function
sympy.galgebra.ga.Fmt(self, fmt=1, title=None)
Fmt is used to control how the multivector is printed with the argument fmt. If fmt=1 the
entire multivector is printed on one line. If fmt=2 each grade of the multivector is printed
on one line. If fmt=3 each component (base) of the multivector is printed on one line. If a
title is given then title = multivector is printed. If the usual print command is used the entire
multivector is printed on one line.
sympy.galgebra.ga.ga print on()
Redirects printer output from standard sympy print handler. Needed if one wishes to use
compact forms of function and derivative output strings for interactive use.
sympy.galgebra.ga.ga print off()
Restores standard sympy print handler.
sympy.galgebra.printing.GA Printer()
Context handler for use inside Sympy code. The code indented under the with GA Printer():
statement will run with the compact forms of function and derivative output strings, and have
the printing restored to standard sympy printing after it has nished, even if it is aborted with
an exception.
Latex Printing
For latex printing one uses one functions from the galgebra module and one function from
the galgebra/printing module. The functions are
sympy.galgebra.printing.Format(Fmode=True, Dmode=True, ipy=False)
This function from the galgebra module turns on latex printing with the following options
argument
Fmode
Dmode
ipy
value
True
False
True
False
False
True
result
Print functions without argument list, f
Print functions with standard sympy latex formatting, f (x, y, z)
Print partial derivatives with condensed notatation, x f
Print partial derivatives with standard sympy latex formatting f
x
Redirect print output to le for post-processing by latex
Do not redirect print output (This is used for Ipython with MathJax)
sympy.galgebra.printing.xdvi(lename=None, pdf=, debug=False, paper=(14,
11))
This function from the galgebra/printing module post-processes the output captured
from print statements. Write the resulting latex strings to the le lename, processes
the le with pdatex, and displays the resulting pdf le. pdf is the name of the pdf viewer
5.9. Geometric Algebra Module
471
SymPy Documentation, Release 0.7.6
on your computer. If you are running ubuntu the evince viewer is automatically used.
On other operating systems if pdf = the name of the pdf le is executed. If the pdf le
type is associated with a viewer this will launch the viewer with the associated le. All
latex les except the pdf le are deleted. If debug = True the le lename is printed
to standard output for debugging purposes and lename (the tex le) is saved. If lename is not entered the default lename is the root name of the python program being
executed with .tex appended. The format for the paper is
paper=(w,h)
paper=letter
w is paper width in inches and,*h* is paper height in inches
paper is standard leter size 8.5 in 11 in
The default of paper=(14,11) was chosen so that long multivector expressions would not
be truncated on the display.
The xdvi function requires that latex and a pdf viewer be installed on the computer.
As an example of using the latex printing options when the following code is executed
from sympy.galgebra.printing import xdvi
from sympy.galgebra.ga import *
Format()
(ex,ey,ez) = MV.setup(e*x|y|z)
A = MV(A,mv)
print r\bm{A} =,A
A.Fmt(2,r\bm{A})
A.Fmt(3,r\bm{A})
xdvi()
The following is displayed
A =A + Ax ex + Ay ey + Az ez + Axy ex ey + Axz ex ez + Ayz ey ez + Axyz ex ey ez
A =A
+ Ax ex + Ay ey + Az ez
+ Axy ex ey + Axz ex ez + Ayz ey ez
+ Axyz ex ey ez
A =A
+ Ax ex
+ Ay ey
+ Az ez
+ Axy ex ey
+ Axz ex ez
+ Ayz ey ez
+ Axyz ex ey ez
For the cases of derivatives the code is
from sympy.galgebra.printing import xdvi
from sympy.galgebra.ga import *
Format()
X = (x,y,z) = symbols(x y z)
(ex,ey,ez,grad) = MV.setup(e_x e_y e_z,metric=[1,1,1],coords=X)
472
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
f = MV(f,scalar,fct=True)
A = MV(A,vector,fct=True)
B = MV(B,grade2,fct=True)
print r\bm{A} =,A
print r\bm{B} =,B
print grad*f =,grad*f
print rgrad|\bm{A} =,grad|A
print rgrad*\bm{A} =,grad*A
print
print
print
print
r-I*(grad^\bm{A}) =,-MV.I*(grad^A)
rgrad*\bm{B} =,grad*B
rgrad^\bm{B} =,grad^B
rgrad|\bm{B} =,grad|B
xdvi()
and the latex displayed output is (f is a scalar function)
A =Ax ex + Ay ey + Az ez
B =B xy ex ey + B xz ex ez + B yz ey ez
f =x f ex + y f ey + z f ez
A =x Ax + y Ay + z Az
A =x Ax + y Ay + z Az + (y Ax + x Ay ) ex ey + (z Ax + x Az ) ex ez
+ (z Ay + y Az ) ey ez
I( A) = (z Ay + y Az ) ex + (z Ax x Az ) ey + (y Ax + x Ay ) ez
B = (y B xy z B xz ) ex + (x B xy z B yz ) ey + (x B xz + y B yz ) ez
+ (z B xy y B xz + x B yz ) ex ey ez
B = (z B xy y B xz + x B yz ) ex ey ez
B = (y B xy z B xz ) ex + (x B xy z B yz ) ey + (x B xz + y B yz ) ez
This example also demonstrates several other features of the latex printer.
In the
case that strings are input into the latex printer such as rgrad*\boldsymbol{A},
rgrad\boldsymbol{A}, or rgrad*\boldsymbol{A}. The text symbols grad, , |, and *
are mapped by the xdvi() post-processor as follows if the string contains an =.
original
grad*A
AB
A|B
A*B
A<B
A>B
replacement
\boldsymbol{\nabla}A
A\wedge B
A\cdot B
AB
A\lfloor B
A\rfloor B
displayed latex
A
AB
AB
AB
AbB
AcB
If the string to be printed contains a % none of the above substitutions are made before the
latex processor is applied. In general for the latex printer strings are assumed to be in a math
environment (equation* or align*) unless the string contains a #.
Note: Except where noted the conventions for latex printing follow those of the latex printing
module of sympy. This includes translating sympy variables with Greek name (such as alpha)
to the equivalent Greek symbol () for the purpose of latex printing. Also a single underscore
5.9. Geometric Algebra Module
473
SymPy Documentation, Release 0.7.6
in the variable name (such as X j) indicates a subscript (Xj ), and a double underscore (such
as X k) a superscript (X k ). The only other change with regard to the sympy latex printer is
that matrices are printed full size (equation displaystyle).
Printer Redirection
In order to print transparently, that is to simply use the print statement with both text and
LaTeX printing the printer output is redirected. In the case of text printing the reason for
redirecting the printer output is because the sympy printing functions print Derivative and
print Function are redened to make the output more compact. If one does not wish to use
the compact notation redirection is not required for the text printer. If one wishes to use
the redened print Derivative and print Function the printer should be redirected with the
function ga print on() and restored with the function ga print o(). Both functions can be
imported from sympy.galgebra.ga (see examples/galgebra/terminal check.py for usage).
SymPy provides a context handler that will allow you to rewrite any ga print on(); do something; ga print o(); sequence like this:
with GA_printer:
do something
This has the advantage that even in the case of an exception inside do something, the
ga print o(); call will be made.
For LaTeX printing the Format() (import from sympy.galgebra.ga) redirects the printer output to a string. After all printing requests one must call the function xdvi() (import from
sympy.galgebra.printing) tp process the string to a LaTeX format, compile with pdatex, and
displayed the resulting pdf le. The function xdvi() also restores the printer output to normal
for standard sympy printing. If Format(ipy=True) is used there is no printer redirection and
the LaTeX output is simply sent to sys.stdout for use in Ipython (Ipython LaTeX interface for
galgebra not yet implemented).
Other Printing Functions
These functions are used together if one wishes to print both code and output in a single le.
They work for text printing and for latex printing.
For these functions to work properly the last function dened must not contain a
Print Function() call (the last function dened is usually a dummy() function that does nothing).
Additionally, to work properly none of the functions containing Print Function() can contain
function denintions (local functions).
sympy.galgebra.printing.Get Program(o=False)
Tells program to print both code and output from functions that have been properly
tagged with Print Function(). Get Program() must be in main program before the functions that you wish code printing from are executed. the o argument in Get Program()
allows one to turn o the printing of the code by changing one line in the entire program
(o=True).
sympy.galgebra.printing.Print Function()
Print Function() is included in those functions where one wishes to print the code block
along with (before) the usual printed output. The Print Function() statement should
474
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
be included immediately after the function def statement. For proper usage of both
Print Function() and Get Program() see the following example.
As an example consider the following code
from sympy.galgebra.printing import xdvi,Get_Program,Print_Function
from sympy.galgebra.ga import *
Format()
def basic_multivector_operations_3D():
Print_Function()
(ex,ey,ez) = MV.setup(e*x|y|z)
A = MV(A,mv)
A.Fmt(1,A)
A.Fmt(2,A)
A.Fmt(3,A)
A.even().Fmt(1,%A_{+})
A.odd().Fmt(1,%A_{-})
X = MV(X,vector)
Y = MV(Y,vector)
print g_{ij} =,MV.metric
X.Fmt(1,X)
Y.Fmt(1,Y)
(X*Y).Fmt(2,X*Y)
(X^Y).Fmt(2,X^Y)
(X|Y).Fmt(2,X|Y)
return
def basic_multivector_operations_2D():
Print_Function()
(ex,ey) = MV.setup(e*x|y)
print g_{ij} =,MV.metric
X = MV(X,vector)
A = MV(A,spinor)
X.Fmt(1,X)
A.Fmt(1,A)
(X|A).Fmt(2,X|A)
(X<A).Fmt(2,X<A)
(A>X).Fmt(2,A>X)
return
def dummy():
return
Get_Program()
basic_multivector_operations_3D()
5.9. Geometric Algebra Module
475
SymPy Documentation, Release 0.7.6
basic_multivector_operations_2D()
xdvi()
The latex output of the code is
476
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
Algebra
BAC-CAB Formulas This example demonstrates the most general metric tensor
(a a) (a b) (a c) (a d)
(a b) (b b) (b c) (b d)
gij =
(a c) (b c) (c c) (c d)
(a d) (b d) (c d) (d d)
and how the galgebra module can be used to verify and expand geometric algebra identities
consisting of relations between the abstract vectors a, b, c, and d.
from sympy.galgebra.printing import xdvi
from sympy.galgebra.ga import *
Format()
(a,b,c,d) = MV.setup(a b c d)
print \\bm{a|(b*c)} =,a|(b*c)
print \\bm{a|(b^c)} =,a|(b^c)
print \\bm{a|(b^c^d)} =,a|(b^c^d)
print \\bm{a|(b^c)+c|(a^b)+b|(c^a)} =,(a|(b^c))+(c|(a^b))+(b|(c^a))
print \\bm{a*(b^c)-b*(a^c)+c*(a^b)} =,a*(b^c)-b*(a^c)+c*(a^b)
print \\bm{a*(b^c^d)-b*(a^c^d)+c*(a^b^d)-d*(a^b^c)} =,a*(b^c^d)-b*(a^c^d)+c*(a^b^d)-d*(a^b^c)
print \\bm{(a^b)|(c^d)} =,(a^b)|(c^d)
print \\bm{((a^b)|c)|d} =,((a^b)|c)|d
print \\bm{(a^b)\\times (c^d)} =,Com(a^b,c^d)
xdvi()
The preceeding code block also demonstrates the mapping of *, , and | to appropriate latex
symbols.
Note: The symbol is the commutator product of two multivectors, A B = (AB BA)/2.
a (bc) = (a c) b + (a b) c
a (b c) = (a c) b + (a b) c
a (b c d) = (a d) b c (a c) b d + (a b) c d
a (b c) + c (a b) + b (c a) =0
a(b c) b(a c) + c(a b) =3a b c
a(b c d) b(a c d) + c(a b d) d(a b c) =4a b c d
(a b) (c d) = (a c) (b d) + (a d) (b c)
((a b) c) d = (a c) (b d) + (a d) (b c)
(a b) (c d) = (b d) a c + (b c) a d + (a d) b c (a c) b d
Reciprocal Frame The reciprocal frame of vectors with respect to the basis vectors is required for the evaluation of the geometric dervative. The following example demonstrates
that for the case of an arbitrary 3-dimensional Euclidian basis the reciprocal basis vectors
are correctly calculated.
5.9. Geometric Algebra Module
477
SymPy Documentation, Release 0.7.6
from sympy.galgebra.printing import xdvi
from sympy.galgebra.ga import *
Format()
metric = 1 # #,+ \
# 1 #,+ \
# # 1,
(e1,e2,e3) = MV.setup(e1 e2 e3,metric)
E = e1^e2^e3
Esq = (E*E).scalar()
print E =,E
print %E^{2} =,Esq
Esq_inv = 1/Esq
E1 = (e2^e3)*E
E2 = (-1)*(e1^e3)*E
E3 = (e1^e2)*E
print E1 = (e2^e3)*E =,E1
print E2 =-(e1^e3)*E =,E2
print E3 = (e1^e2)*E =,E3
print E1|e2 =,(E1|e2).expand()
print E1|e3 =,(E1|e3).expand()
print E2|e1 =,(E2|e1).expand()
print E2|e3 =,(E2|e3).expand()
print E3|e1 =,(E3|e1).expand()
print E3|e2 =,(E3|e2).expand()
w = ((E1|e1).expand()).scalar()
Esq = expand(Esq)
print %(E1\\cdot e1)/E^{2} =,simplify(w/Esq)
w = ((E2|e2).expand()).scalar()
print %(E2\\cdot e2)/E^{2} =,simplify(w/Esq)
w = ((E3|e3).expand()).scalar()
print %(E3\\cdot e3)/E^{2} =,simplify(w/Esq)
xdvi()
The preceeding code also demonstrated the use of the % directive in printing a string so that
is treated literally and not translated to \wedge. Note that %E{2} = is printed as E 2 =
478
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
and not as E 2 =.
E =e1 e2 e3
2
E 2 = (e1 e2 ) 2 (e1 e2 ) (e1 e3 ) (e2 e3 ) + (e1 e3 ) + (e2 e3 ) 1
(
)
2
E1 =(e2 e3)E = (e2 e3 ) 1 e1 + ((e1 e2 ) (e1 e3 ) (e2 e3 )) e2 + ( (e1 e2 ) (e2 e3 ) + (e1 e3 )) e3
(
)
2
E2 = (e1 e3)E = ((e1 e2 ) (e1 e3 ) (e2 e3 )) e1 + (e1 e3 ) 1 e2 + ( (e1 e2 ) (e1 e3 ) + (e2 e3 )) e3
(
)
2
E3 =(e1 e2)E = ( (e1 e2 ) (e2 e3 ) + (e1 e3 )) e1 + ( (e1 e2 ) (e1 e3 ) + (e2 e3 )) e2 + (e1 e2 ) 1 e3
E1 e2 =0
E1 e3 =0
E2 e1 =0
E2 e3 =0
E3 e1 =0
E3 e2 =0
(E1 e1)/E 2 =1
(E2 e2)/E 2 =1
(E3 e3)/E 2 =1
The formulas derived for E1, E2, E3, and E 2 could also be applied to the numerical calculations
of crystal properties.
Lorentz-Transformation A simple physics demonstation of geometric algebra is the
derivation of the Lorentz-Transformation. In this demonstration a 2-dimensional Minkowski
space is dened and the Lorentz-Transformation is generated from a rotation of a vector in
the Minkowski space using the rotor R.
from sympy import symbols,sinh,cosh
from sympy.galgebra.printing import xdvi
from sympy.galgebra.ga import *
Format()
(alpha,beta,gamma) = symbols(alpha beta gamma)
(x,t,xp,tp) = symbols(x t x t)
(g0,g1) = MV.setup(gamma*t|x,metric=[1,-1])
R = cosh(alpha/2)+sinh(alpha/2)*(g0^g1)
X = t*g0+x*g1
Xp = tp*g0+xp*g1
print R =,R
print r#%t\bm{\gamma_{t}}+x\bm{\gamma_{x}} = t\bm{\gamma_{t}}+x\bm{\gamma_{x}} +
r= R\left ( t\bm{\gamma_{t}}+x\bm{\gamma_{x}}\right ) R^{\dagger}
Xpp = R*Xp*R.rev()
Xpp = Xpp.collect([xp,tp])
Xpp = Xpp.subs({2*sinh(alpha/2)*cosh(alpha/2):sinh(alpha),\
sinh(alpha/2)**2+cosh(alpha/2)**2:cosh(alpha)})
print r%t\bm{\gamma_{t}}+x\bm{\gamma_{x}} =,Xpp
Xpp = Xpp.subs({sinh(alpha):gamma*beta,cosh(alpha):gamma})
print r%\f{\sinh}{\alpha} = \gamma\beta
5.9. Geometric Algebra Module
479
SymPy Documentation, Release 0.7.6
print r%\f{\cosh}{\alpha} = \gamma
print r%t\bm{\gamma_{t}}+x\bm{\gamma_{x}} =,Xpp.collect(gamma)
xdvi()
The preceeding code also demonstrates how to use the sympy subs functions to perform the hyperbolic half angle transformation.
The code
also shows the use of both the # and % directives in the text string
r#%t\bm{\gamma {t}}+x\bm{\gamma {x}}
=
t\bm{\gamma {t}}+x\bm{\gamma {x}}
=
R\left ( t\bm{\gamma {t}}+x\bm{\gamma {x}}\right ) R{\dagger}. Both the # and %
are needed in this text string for two reasons. First, the text string contains an = sign. The
latex preprocessor uses this a key to combine the text string with a sympy expression to be
printed after the text string. The # is required to inform the preprocessor that there is no
sympy expression to follow. Second, the % is requires to inform the preprocessor that the
text string is to be displayed in latex math mode and not in text mode (if # is present the
default latex mode is text mode unless overridden by the % directive).
( )
( )
1
1
+ sinh
t x
R = cosh
2
2
tt + xx =t0 t0 + x0 x0 = R (t0 t + x0 x ) R
tt + xx = (t0 cosh () x0 sinh ()) t + (t0 sinh () + x0 cosh ()) x
sinh () =
cosh () =
tt + xx = ( (x0 + t0 )) t + ( (t0 + x0 )) x
Calculus
Derivatives in Spherical Coordinates The following code shows how to use galgebra to
use spherical coordinates. The gradient of a scalar function, f , the divergence and curl of a
vector function, A, and the exterior derivative (curl) of a bivector function, B are calculated.
Note that to get the standard curl of a 3-dimension function the result is multiplied by I the
negative of the pseudoscalar.
Note: In geometric calculus the operator 2 is well dened on its own as the geometic
derivative of the geometric derivative. However, if needed we have for the vector function
A the relations (since A is a scalar its curl is equal to its geometric derivative and its
divergence is zero) A = A + A
2 A = ( A) + ( A)
2 A = ( A) + ( A) + ( A) + ( A)
2 A = ( A) + ( ) A ( A) + ( A)
2 A = A + ( ) A
In the derivation we have used that ( A) = ( ) A ( A) which is implicit in the
second BAC-CAB formula. No parenthesis is needed for the geometric curl of the curl (exterior
derivative of exterior derivative) since the operation is associative unlike the vector curl
operator and is the usual Laplacian operator.
480
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
from sympy import sin,cos
from sympy.galgebra.printing import xdvi
from sympy.galgebra.ga import *
Format()
X = (r,th,phi) = symbols(r theta phi)
curv = [[r*cos(phi)*sin(th),r*sin(phi)*sin(th),r*cos(th)],[1,r,r*sin(th)]]
(er,eth,ephi,grad) = MV.setup(e_r e_theta e_phi,metric=[1,1,1],coords=X,curv=curv)
f = MV(f,scalar,fct=True)
A = MV(A,vector,fct=True)
B = MV(B,grade2,fct=True)
print A =,A
print B =,B
grad*f =,grad*f
grad|A =,grad|A
-I*(grad^A) =,-MV.I*(grad^A)
grad^B =,grad^B
xdvi()
Results of code
A =Ar er + A e + A e
B =B r er e + B r er e + B e e
f
f
f =r f er +
e +
e
r
r sin ()
A
Ar
A
A
A =r Ar +
+2
+
+
r tan ()
r
r
r sin ()
(
)
(
)
(
)
A cos () + sin () A A
A
Ar
rr A + A Ar
I( A) =
er + r A
+
e +
e
r sin ()
r
r sin ()
r
(
)
B
B r
B r
B r
B = r B + 2
+
er e e
r
r tan ()
r
r sin ()
Maxwells Equations The geometric algebra formulation of Maxwells equations is deomonstrated with the formalism developed in Geometric Algebra for Physicists [Doran]
(page 1909). In this formalism the signature of the metric is (1, 1, 1, 1) and the basis vectors are t , x , y , and z . The if E and B are the normal electric and magnetic eld vectors
the electric and magnetic bivectors are given by E = Et and B = Bt . The electromagnetic
bivector is then F = E + IB where I = t x y z is the pesudo-scalar for the Minkowski space.
Note that the electromagnetic bivector is isomorphic to the electromagnetic tensor. Then if J
is the 4-current all of Maxwells equations are given by F = J. For more details see [Doran]
(page 1909) chapter 7.
from sympy import symbols,sin,cos
from sympy.galgebra.printing import xdvi
from sympy.galgebra.ga import *
Format()
vars = symbols(t x y z)
(g0,g1,g2,g3,grad) = MV.setup(gamma*t|x|y|z,metric=[1,-1,-1,-1],coords=vars)
5.9. Geometric Algebra Module
481
SymPy Documentation, Release 0.7.6
I = MV.I
B = MV(B,vector,fct=True)
E = MV(E,vector,fct=True)
B.set_coef(1,0,0)
E.set_coef(1,0,0)
B *= g0
E *= g0
J = MV(J,vector,fct=True)
F = E+I*B
print B = \\bm{B\\gamma_{t}} =,B
print E = \\bm{E\\gamma_{t}} =,E
print F = E+IB =,F
print J =,J
gradF = grad*F
gradF.Fmt(3,grad*F)
print grad*F = J
(gradF.grade(1)-J).Fmt(3,%\\grade{\\nabla F}_{1} -J = 0)
(gradF.grade(3)).Fmt(3,%\\grade{\\nabla F}_{3} = 0)
xdvi()
B =Bt = B x t x B y t y B z t z
E =Et = E x t x E y t y E z t z
F =E + IB = E x t x E y t y E z t z B z x y + B y x z B x y z
J =J t t + J x x + J y y + J z z
F = (x E x + y E y + z E z ) t
+ (z B y + y B z t E x ) x
+ (z B x x B z t E y ) y
+ (y B x + x B y t E z ) z
+ (t B z + y E x x E y ) t x y
+ (t B y + z E x x E z ) t x z
+ (t B x + z E y y E z ) t y z
+ (x B x + y B y + z B z ) x y z
F =J
(
)
hF i1 J = 0 = J t + x E x + y E y + z E z t
+ (J x z B y + y B z t E x ) x
+ (J y + z B x x B z t E y ) y
+ (J z y B x + x B y t E z ) z
hF i3 = 0 = (t B z + y E x x E y ) t x y
+ (t B y + z E x x E z ) t x z
+ (t B x + z E y y E z ) t y z
+ (x B x + y B y + z B z ) x y z
Dirac Equation In [Doran] (page 1909) equation 8.89 (page 283) is the geometric algebra
formulation of the Dirac equation. In this equation is an 8-component real spinor which
482
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
is to say that it is a multivector with sacalar, bivector, and pseudo-vector components in the
space-time geometric algebra (it consists only of even grade components).
from sympy import symbols,sin,cos
from sympy.galgebra.printing import xdvi
from sympy.galgebra.ga import *
Format()
vars = symbols(t x y z)
(g0,g1,g2,g3,grad) = MV.setup(gamma*t|x|y|z,metric=[1,-1,-1,-1],coords=vars)
I = MV.I
(m,e) = symbols(m e)
psi = MV(psi,spinor,fct=True)
A = MV(A,vector,fct=True)
sig_z = g3*g0
print \\bm{A} =,A
print \\bm{\\psi} =,psi
dirac_eq = (grad*psi)*I*sig_z-e*A*psi-m*psi*g0
dirac_eq.simplify()
dirac_eq.Fmt(3,r\nabla \bm{\psi} I \sigma_{z}-e\bm{A}\bm{\psi}-m\bm{\psi}\gamma_{t} = 0)
xdvi()
The equations displayed are the partial dierential equations for each component of the Dirac
equation in rectangular coordinates we the driver for the equations is the 4-potential A. One
utility of these equations is to setup a numerical solver for the Dirac equation.
A =At t + Ax x + Ay y + Az z
= + tx t x + ty t y + tz t z + xy x y + xz x z + yz y z
+ txyz t x y z
(
)
Iz eA mt = 0 = eAt eAx tx eAy ty eAz tz m y tx z txyz + x ty + t xy t
(
)
+ eAt tx eAx eAy xy eAz xz + m tx + y t ty x xy + z yz x
(
)
+ eAt ty + eAx xy eAy eAz yz + m ty x + t tx y xy z xz y
(
)
+ eAt tz + eAx xz + eAy yz eAz + m tz + t txyz z xy + y xz x yz z
(
)
+ eAt xy + eAx ty eAy tx eAz txyz m xy t + x tx + y ty + z tz t x
(
)
+ eAt xz + eAx tz + eAy txyz eAz tx m xz + x txyz + z ty y tz t yz t
(
)
+ eAt yz eAx txyz + eAy tz eAz ty m yz z tx + y txyz + x tz + t xz t
(
)
+ eAt txyz eAx yz + eAy xz eAz xy + m txyz + z t tz x xz y yz x
5.9.2 Manifold for Geometric Algebra
Class Reference
class sympy.galgebra.manifold.Manifold(x, coords, debug=False, I=None)
5.9. Geometric Algebra Module
483
SymPy Documentation, Release 0.7.6
5.9.3 Vector for Geometric Algebra
Class Reference
class sympy.galgebra.vector.Vector(basis str=None)
Vector class.
Setup is done by dening a set of basis vectors in static function Bases. The linear combination of scalar (commutative) sympy quatities and the basis vectors form the vector
space. If the number of basis vectors is n the metric tensor is formed as an n by n sympy
matrix of scalar symbols and represents the dot products of pairs of basis vectors.
static basic dot(v1, v2)
Dot product of two basis vectors returns a Symbol
static setup(base, n=None, metric=None, coords=None, curv=(None, None), debug=False)
Generate basis of vector space as tuple of vectors and associated metric tensor as
Matrix. See str array(base,n) for usage of base and n and str array(metric) for usage
of metric.
To overide elements in the default metric use the character # in the metric string.
For example if one wishes the diagonal elements of the metric tensor to be zero
enter metric = 0 #,# 0.
If the basis vectors are e1 and e2 then the default metric Vector.metric = ((dot(e1,e1),dot(e1,e2)),dot(e2,e1),dot(e2,e2))
becomes Vector.metric = ((0,dot(e1,e2)),(dot(e2,e1),0)).
The function dot returns a Symbol and is symmetric.
The functions Bases calculates the global quantities: Vector.basis tuple of basis vectors
Vector.base to index dictionary to convert base to base inded
Vector.metric metric tensor represented as a matrix of symbols and numbers
Function Reference
sympy.galgebra.vector.flatten(lst)
sympy.galgebra.vector.TrigSimp(x)
5.9.4 Precedence for Geometric Algebra
Function Reference
sympy.galgebra.precedence.add paren(line, re exprs)
sympy.galgebra.precedence.contains interval(interval1, interval2)
sympy.galgebra.precedence.define precedence(gd, op ord=<>|, , *)
sympy.galgebra.precedence.GAeval(s, pstr=False)
484
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
sympy.galgebra.precedence.parse line(line)
sympy.galgebra.precedence.parse paren(line)
sympy.galgebra.precedence.sub paren(s)
sympy.galgebra.precedence.unparse paren(level lst)
5.9.5 Printing for Geometric Algebra
Class Reference
class sympy.galgebra.printing.enhance print(base=None, fct=None, deriv=None,
on=True, keys=False)
A class for color coding the string printing going to a terminal.
class sympy.galgebra.printing.GA Printer(settings=None)
An enhanced string printer that is galgebra-aware.
class sympy.galgebra.printing.GA LatexPrinter(settings=None)
An enhanced LaTeX printer that is galgebra-aware.
The latex printer is turned on with the function (in ga.py) Format(Fmode=True,Dmode=True,ipy=False)
where Fmode is the function printing mode that surpresses printing arguments, Dmode
is the derivative printing mode that does not use fractions, and ipy=True is the Ipython
notebook mode that does not redirect the print output.
The latex output is post processed and displayed with the function (in GAPrint.py) xdvi(lename=tmplatex.tex,debug=False)
where lename is the name of the tex le one would keep for future inclusion in documents and debug=True would display the tex le immediately.
There are three options for printing multivectors in latex. They are acessed with the
multivector member function A.Fmt(self,fmt=1,title=None)
where fmt=1, 2, or 3 determines whether the entire multivector A is printed entirely on
one line, or one grade is printed per line, or one base is printed per line. If title is not
None then the latex string generated is of the form title+ = +str(A)
where it is assumed that title is a latex math mode string. If title contains % it is treated
as a pure latex math mode string. If it does not contain % then the following character
mappings are applied grad replaced by \bm{\nabla}
* replaced by
replaced by \W
| replaced by \cdot
> replaced by \loor
< replaced by \roor
In the case of a print statement of the form print(title,A)
5.9. Geometric Algebra Module
485
SymPy Documentation, Release 0.7.6
everthing in the title processing still applies except that the multivector formatting is
one multivector per line.
For print statements of the form print(title)
where no program variables are printed if title contains # then title is printed as regular
latex line. If title does not contain # then title is printed in equation mode. % has the
same eect in title as in the Fmt() member function.
Function Reference
sympy.galgebra.printing.find executable(executable, path=None)
Try to nd executable in the directories listed in path (a string listing directories separated by os.pathsep; defaults to os.environ[PATH]). Returns the complete lename
or None if not found
sympy.galgebra.printing.Get Program(o=False)
sympy.galgebra.printing.latex(expr, **settings)
Return the LaTeX representation of the given expression.
sympy.galgebra.printing.Print Function()
sympy.galgebra.printing.print latex(expr, **settings)
Print the LaTeX representation of the given expression.
sympy.galgebra.printing.xdvi(lename=None, debug=False, paper=(14, 11))
Post processes LaTeX output (see comments below), adds preamble and postscript, generates tex le, inputs le to latex, displays resulting pdf le.
5.9.6 Noncommutative utilities for Geometric Algebra
Function Reference
sympy.galgebra.ncutil.bilinear function(expr, fct)
If a sympy Expr is of the form:
expr = expr 0 + expr 1*a 1 + ... + expr n*a n + expr 11*a 1*a 1
... + expr ij*a i*a j + ... + expr nn*a n*a n
where all the a j are noncommuting symbols in basis then
bilinear function(expr) = bilinear product(expr 0) + bilinear product(expr 1*a 1) + ... + bilin
bilinear + product(expr 11*a 1*a 1) + ... + bilinear product(expr nn*a n*a n)
sympy.galgebra.ncutil.bilinear product(expr, fct)
If a sympy Expr is of the form:
expr = expr ij*a i*a j or expr 0 or expr i*a i
where all the a i are noncommuting symbols in basis and the exprs
are commuting expressions then
bilinear product(expr) = expr ij*fct(a i, a j)
bilinear product(expr 0) = expr 0
bilinear product(expr i*a i) = expr i*a i
486
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
sympy.galgebra.ncutil.coef function(expr, fct)
If a sympy Expr is of the form:
expr = expr 0 + expr 1*a 1 + ... + expr n*a n
where all the a j are noncommuting symbols in basis then
f(expr) = fct(expr 0) + fct(expr 1)*a 1 + ... + fct(expr n)*a n
is returned
sympy.galgebra.ncutil.linear derivation(expr, fct, x)
If a sympy Expr is of the form:
expr = expr 0 + expr 1*a 1 + ... + expr n*a n
where all the a j are noncommuting symbols in basis then
linear drivation(expr) = di(expr 0, x) + di(expr 1, x)*a 1 + ...
di(expr n, x)*a n + expr 1*fct(a 1, x) + ...
expr n*fct(a n, x)
sympy.galgebra.ncutil.linear expand(expr)
If a sympy Expr is of the form:
expr = expr 0 + expr 1*a 1 + ... + expr n*a n
where all the a j are noncommuting symbols in basis then
(expr 0, ..., expr n) and (1, a 1, ..., a n) are returned. Note that expr j*a j does not have
to be of that form, but rather can be any Mul with a j as a factor (it doen not have to be
a postmultiplier). expr 0 is the scalar part of the expression.
sympy.galgebra.ncutil.linear function(expr, fct)
If a sympy Expr is of the form:
expr = expr 0 + expr 1*a 1 + ... + expr n*a n
where all the a j are noncommuting symbols in basis then
f(expr) = expr 0 + expr 1*f(a 1) + ... + expr n*f(a n)
is returned
sympy.galgebra.ncutil.linear projection(expr, plist=None)
If a sympy Expr is of the form:
expr = expr 0 + expr 1*a 1 + ... + expr n*a n
where all the a j are noncommuting symbols in basis then
proj(expr) returns the sum of those terms where a j is in plist
sympy.galgebra.ncutil.multilinear derivation(F, fct, x)
If a sympy Expr is of the form (summation convention):
expr = expr 0 + expr i1i2...ir*a i1*...*a ir
where all the a j are noncommuting symbols in basis then
dexpr = di(expr 0, x) + d(expr i1i2...ir*a i1*...*a ir)
is returned where d() is the product derivation
sympy.galgebra.ncutil.multilinear function(expr, fct)
If a sympy Expr is of the form summation convention):
expr = expr 0 + Sum{0 < r <= n}{expr i1i2...ir*a i1*a i2*...*a ir}
5.9. Geometric Algebra Module
487
SymPy Documentation, Release 0.7.6
where all the a j are noncommuting symbols in basis then and the dimension of the basis
in n then
bilinear function(expr) = multilinear product(expr 0)
Sum{0<r<=n}multilinear product(expr i1i2...ir*a i1*a i2*...*a ir)
sympy.galgebra.ncutil.multilinear product(expr, fct)
If a sympy Expr is of the form:
expr = expr i1i2...irj*a i1*a i2*...*a ir or expr 0
where all the a i are noncommuting symbols in basis and the exprs
are commuting expressions then
multilinear product(expr) = expr i1i2...ir*fct(a i1, a i2, ..., a ir)
bilinear product(expr 0) = expr 0
where fct() is dened for r <= n the total number of bases
sympy.galgebra.ncutil.non scalar projection(expr)
If a sympy Expr is of the form:
expr = expr 0*S.One + expr 1*a 1 + ... + expr n*a n
where all the a j are noncommuting symbols in basis then
proj(expr) returns the sum of those terms where a j is in plist
sympy.galgebra.ncutil.product derivation(F, fct, x)
If a sympy Expr is of the form:
expr = expr 0*a 1*...*a n
where all the a j are noncommuting symbols in basis then
product derivation(expr) = di(expr 0, x)*a 1*...*a n
expr 0*(fct(a 1, x)*a 2*...*a n + ...
a 1*...*a (i-1)*fct(a i, x)*a (i + 1)*...*a n + ...
a 1*...*a (n-1)*fct(a n, x))
5.9.7 String manipulation for Geometric Algebra
Function Reference
sympy.galgebra.stringarrays.fct sym array(str lst, coords=None)
Construct list of symbols or functions with names in str lst. If coords are given (tuple
of symbols) function list constructed, otherwise a symbol list is constructed.
sympy.galgebra.stringarrays.str array(base, n=None)
Generate one dimensional (list of strings) or two dimensional (list of list of strings) string
array.
For one dimensional arrays: base is string of variable names separated by blanks such as base = a b c
which produces the string list [a,b,c] or it is a string with no blanks than in
conjunction with the integer n generates str array(v,n=-3)
[v 1,v 2,v 3].
488
[v 1,v 2,v 3]
str array(v,n=3)
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
In the case of LaTeX printing the would give a subscript and the a super
script.
For two dimensional arrays: base is string where elements are separated by spaces and rows by commas so
that str array(a b,c d) = [[a,b],[c,d]]
sympy.galgebra.stringarrays.str combinations(base, lst, rank=1, mode= )
Construct a list of strings of the form base+mode+indexes where the indexes are
formed by converting lst to a list of strings and then forming the indexes by concatenating combinations of elements from lst taken rank at a time.
sympy.galgebra.stringarrays.symbol array(base, n=None)
Generates a string arrary with str array and replaces each string in array with Symbol
of same name.
5.9.8 Debug code for Geometric Algebra
Function Reference
sympy.galgebra.debug.ostr(obj, dict mode=False)
Recursively convert iterated object (list/tuple/dict/set) to string.
sympy.galgebra.debug.oprint(*args, **kwargs)
Debug printing for iterated (list/tuple/dict/set) objects.
tle1,object1,title2,object2,...) and prints:
args is of form (ti-
title1 = object1 title2 = object2 ...
If you only wish to print a title set object = None.
sympy.galgebra.debug.print product table(title,
keys,
pdict,
blade rep=True)
Print product dictionary, pdict, according to order of keys in keys
op=*,
sympy.galgebra.debug.print sub table(title, keys, sdict, blade rep=True)
Print substitution dictionary, sdict, according to order of keys in keys
5.10 Geometry Module
5.10.1 Introduction
The geometry module for SymPy allows one to create two-dimensional geometrical entities,
such as lines and circles, and query for information about these entities. This could include
asking the area of an ellipse, checking for collinearity of a set of points, or nding the intersection between two lines. The primary use case of the module involves entities with numerical
values, but it is possible to also use symbolic representations.
5.10.2 Available Entities
The following entities are currently available in the geometry module:
Point
5.10. Geometry Module
489
SymPy Documentation, Release 0.7.6
Line, Ray, Segment
Ellipse, Circle
Polygon, RegularPolygon, Triangle
Most of the work one will do will be through the properties and methods of these entities, but
several global methods exist:
intersection(entity1, entity2)
are similar(entity1, entity2)
convex hull(points)
For a full API listing and an explanation of the methods and their return values please see the
list of classes at the end of this document.
5.10.3 Example Usage
The following Python session gives one an idea of how to work with some of the geometry
module.
>>> from sympy import *
>>> from sympy.geometry import *
>>> x = Point(0, 0)
>>> y = Point(1, 1)
>>> z = Point(2, 2)
>>> zp = Point(1, 0)
>>> Point.is_collinear(x, y, z)
True
>>> Point.is_collinear(x, y, zp)
False
>>> t = Triangle(zp, y, x)
>>> t.area
1/2
>>> t.medians[x]
Segment(Point(0, 0), Point(1, 1/2))
>>> Segment(Point(1, S(1)/2), Point(0, 0))
Segment(Point(0, 0), Point(1, 1/2))
>>> m = t.medians
>>> intersection(m[x], m[y], m[zp])
[Point(2/3, 1/3)]
>>> c = Circle(x, 5)
>>> l = Line(Point(5, -5), Point(5, 5))
>>> c.is_tangent(l) # is l tangent to c?
True
>>> l = Line(x, y)
>>> c.is_tangent(l) # is l tangent to c?
False
>>> intersection(c, l)
[Point(-5*sqrt(2)/2, -5*sqrt(2)/2), Point(5*sqrt(2)/2, 5*sqrt(2)/2)]
5.10.4 Intersection of medians
>>> from sympy import symbols
>>> from sympy.geometry import Point, Triangle, intersection
490
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> a, b = symbols(a,b, positive=True)
>>>
>>>
>>>
>>>
x
y
z
t
=
=
=
=
Point(0, 0)
Point(a, 0)
Point(2*a, b)
Triangle(x, y, z)
>>> t.area
a*b/2
>>> t.medians[x]
Segment(Point(0, 0), Point(3*a/2, b/2))
>>> intersection(t.medians[x], t.medians[y], t.medians[z])
[Point(a, b/3)]
5.10.5 An in-depth example: Pappus Hexagon Theorem
From Wikipedia ([WikiPappus] (page 1909)):
Given one set of collinear points A, B, C, and another set of collinear points a, b, c,
then the intersection points X, Y , Z of line pairs Ab and aB, Ac and aC, Bc and bC
are collinear.
>>> from sympy import *
>>> from sympy.geometry import *
>>>
>>> l1 = Line(Point(0, 0), Point(5, 6))
>>> l2 = Line(Point(0, 0), Point(2, -2))
>>>
>>> def subs_point(l, val):
...
Take an arbitrary point and make it a fixed point.
...
t = Symbol(t, real=True)
...
ap = l.arbitrary_point()
...
return Point(ap.x.subs(t, val), ap.y.subs(t, val))
...
>>> p11 = subs_point(l1, 5)
>>> p12 = subs_point(l1, 6)
>>> p13 = subs_point(l1, 11)
>>>
>>> p21 = subs_point(l2, -1)
>>> p22 = subs_point(l2, 2)
>>> p23 = subs_point(l2, 13)
>>>
>>> ll1 = Line(p11, p22)
>>> ll2 = Line(p11, p23)
>>> ll3 = Line(p12, p21)
>>> ll4 = Line(p12, p23)
>>> ll5 = Line(p13, p21)
>>> ll6 = Line(p13, p22)
>>>
>>> pp1 = intersection(ll1, ll3)[0]
>>> pp2 = intersection(ll2, ll5)[0]
>>> pp3 = intersection(ll4, ll6)[0]
>>>
>>> Point.is_collinear(pp1, pp2, pp3)
True
5.10. Geometry Module
491
SymPy Documentation, Release 0.7.6
References
5.10.6 Miscellaneous Notes
The area property of Polygon and Triangle may return a positive or negative value,
depending on whether or not the points are oriented counter-clockwise or clockwise,
respectively. If you always want a positive value be sure to use the abs function.
Although Polygon can refer to any type of polygon, the code has been written for simple
polygons. Hence, expect potential problems if dealing with complex polygons (overlapping sides).
Since SymPy is still in its infancy some things may not simplify properly and hence some
things that should return True (e.g., Point.is collinear) may not actually do so. Similarly, attempting to nd the intersection of entities that do intersect may result in an
empty result.
5.10.7 Future Work
Truth Setting Expressions
When one deals with symbolic entities, it often happens that an assertion cannot be guaranteed. For example, consider the following code:
>>> from sympy import *
>>> from sympy.geometry import *
>>> x,y,z = map(Symbol, xyz)
>>> p1,p2,p3 = Point(x, y), Point(y, z), Point(2*x*y, y)
>>> Point.is_collinear(p1, p2, p3)
False
Even though the result is currently False, this is not always true. If the quantity z y 2
y z + 2 y 2 == 0 then the points will be collinear. It would be really nice to inform the
user of this because such a quantity may be useful to a user for further calculation and, at
the very least, being nice to know. This could be potentially done by returning an object (e.g.,
GeometryResult) that the user could use. This actually would not involve an extensive amount
of work.
Three Dimensions and Beyond
Currently there are no plans for extending the module to three dimensions, but it certainly
would be a good addition. This would probably involve a fair amount of work since many of
the algorithms used are specic to two dimensions.
Geometry Visualization
The plotting module is capable of plotting geometric entities. See Plotting Geometric Entities
(page 1286) in the plotting module entry.
492
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Submodules
Entities
class sympy.geometry.entity.GeometryEntity
The base class for all geometrical entities.
This class doesnt represent any particular geometric entity, it only provides the implementation of some methods common to all subclasses.
encloses(o)
Return True if o is inside (not on or outside) the boundaries of self.
The object will be decomposed into Points and individual Entities need only dene
an encloses point method for their class.
See Also:
sympy.geometry.ellipse.Ellipse.encloses point
(page
sympy.geometry.polygon.Polygon.encloses point (page 560)
545),
Examples
>>> from sympy import RegularPolygon, Point, Polygon
>>> t = Polygon(*RegularPolygon(Point(0, 0), 1, 3).vertices)
>>> t2 = Polygon(*RegularPolygon(Point(0, 0), 2, 3).vertices)
>>> t2.encloses(t)
True
>>> t.encloses(t2)
False
intersection(o)
Returns a list of all of the intersections of self with o.
See Also:
sympy.geometry.util.intersection (page 495)
Notes
An entity is not required to implement this method.
If two dierent types of entities can intersect, the item with higher index in ordering of classes should implement intersections with anything having a lower index.
is similar(other)
Is this geometrical entity similar to another geometrical entity?
Two entities are similar if a uniform scaling (enlarging or shrinking) of one of the
entities will allow one to obtain the other.
See Also:
scale (page 494)
5.10. Geometry Module
493
SymPy Documentation, Release 0.7.6
Notes
This method is not intended to be used directly but rather through the ares imilar
function found in util.py. An entity is not required to implement this method. If two
dierent types of entities can be similar, it is only required that one of them be able
to determine this.
rotate(angle, pt=None)
Rotate angle radians counterclockwise about Point pt.
The default pt is the origin, Point(0, 0)
See Also:
scale (page 494), translate (page 494)
Examples
>>> from sympy import Point, RegularPolygon, Polygon, pi
>>> t = Polygon(*RegularPolygon(Point(0, 0), 1, 3).vertices)
>>> t # vertex on x axis
Triangle(Point(1, 0), Point(-1/2, sqrt(3)/2), Point(-1/2, -sqrt(3)/2))
>>> t.rotate(pi/2) # vertex on y axis now
Triangle(Point(0, 1), Point(-sqrt(3)/2, -1/2), Point(sqrt(3)/2, -1/2))
scale(x=1, y=1, pt=None)
Scale the object by multiplying the x,y-coordinates by x and y.
If pt is given, the scaling is done relative to that point; the object is shifted by -pt,
scaled, and shifted by pt.
See Also:
rotate (page 494), translate (page 494)
Examples
>>> from sympy import RegularPolygon, Point, Polygon
>>> t = Polygon(*RegularPolygon(Point(0, 0), 1, 3).vertices)
>>> t
Triangle(Point(1, 0), Point(-1/2, sqrt(3)/2), Point(-1/2, -sqrt(3)/2))
>>> t.scale(2)
Triangle(Point(2, 0), Point(-1, sqrt(3)/2), Point(-1, -sqrt(3)/2))
>>> t.scale(2,2)
Triangle(Point(2, 0), Point(-1, sqrt(3)), Point(-1, -sqrt(3)))
translate(x=0, y=0)
Shift the object by adding to the x,y-coordinates the values x and y.
See Also:
rotate (page 494), scale (page 494)
Examples
494
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> from sympy import RegularPolygon, Point, Polygon
>>> t = Polygon(*RegularPolygon(Point(0, 0), 1, 3).vertices)
>>> t
Triangle(Point(1, 0), Point(-1/2, sqrt(3)/2), Point(-1/2, -sqrt(3)/2))
>>> t.translate(2)
Triangle(Point(3, 0), Point(3/2, sqrt(3)/2), Point(3/2, -sqrt(3)/2))
>>> t.translate(2, 2)
Triangle(Point(3, 2), Point(3/2, sqrt(3)/2 + 2),
Point(3/2, -sqrt(3)/2 + 2))
Utils
sympy.geometry.util.intersection(*entities)
The intersection of a collection of GeometryEntity instances.
Parameters entities : sequence of GeometryEntity
Returns intersection : list of GeometryEntity
Raises NotImplementedError :
When unable to calculate intersection.
See Also:
sympy.geometry.entity.GeometryEntity.intersection (page 493)
Notes
The intersection of any geometrical entity with itself should return a list with one item:
the entity in question. An intersection requires two or more entities. If only a single entity is given then the function will return an empty list. It is possible for intersection to miss
intersections that one knows exists because the required quantities were not fully simplied internally. Reals should be converted to Rationals, e.g. Rational(str(real num))
or else failures due to oating point issues may result.
Examples
>>> from sympy.geometry import Point, Line, Circle, intersection
>>> p1, p2, p3 = Point(0, 0), Point(1, 1), Point(-1, 5)
>>> l1, l2 = Line(p1, p2), Line(p3, p2)
>>> c = Circle(p2, 1)
>>> intersection(l1, p2)
[Point(1, 1)]
>>> intersection(l1, l2)
[Point(1, 1)]
>>> intersection(c, p2)
[]
>>> intersection(c, Point(1, 0))
[Point(1, 0)]
>>> intersection(c, l2)
[Point(-sqrt(5)/5 + 1, 2*sqrt(5)/5 + 1),
Point(sqrt(5)/5 + 1, -2*sqrt(5)/5 + 1)]
sympy.geometry.util.convex hull(*args)
The convex hull surrounding the Points contained in the list of entities.
5.10. Geometry Module
495
SymPy Documentation, Release 0.7.6
Parameters args : a collection of Points, Segments and/or Polygons
Returns convex hull : Polygon
See Also:
sympy.geometry.point.Point
(page 556)
(page
497),
sympy.geometry.polygon.Polygon
Notes
This can only be performed on a set of non-symbolic points.
References
[1] http://en.wikipedia.org/wiki/Graham scan
[2] Andrews Monotone Chain Algorithm (A.M. Andrew, Another Ecient Algorithm for
Convex Hulls in Two Dimensions, 1979) http://geomalgorithms.com/a10- hull-1.html
Examples
>>> from sympy.geometry import Point, convex_hull
>>> points = [(1,1), (1,2), (3,1), (-5,2), (15,4)]
>>> convex_hull(*points)
Polygon(Point(-5, 2), Point(1, 1), Point(3, 1), Point(15, 4))
sympy.geometry.util.are similar(e1, e2)
Are two geometrical entities similar.
Can one geometrical entity be uniformly scaled to the other?
Parameters e1 : GeometryEntity
e2 : GeometryEntity
Returns are similar : boolean
Raises GeometryError :
When e1 and e2 cannot be compared.
See Also:
sympy.geometry.entity.GeometryEntity.is similar (page 493)
Notes
If the two objects are equal then they are similar.
Examples
496
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> from sympy import Point, Circle, Triangle, are_similar
>>> c1, c2 = Circle(Point(0, 0), 4), Circle(Point(1, 4), 3)
>>> t1 = Triangle(Point(0, 0), Point(1, 0), Point(0, 1))
>>> t2 = Triangle(Point(0, 0), Point(2, 0), Point(0, 2))
>>> t3 = Triangle(Point(0, 0), Point(3, 0), Point(0, 1))
>>> are_similar(t1, t2)
True
>>> are_similar(t1, t3)
False
sympy.geometry.util.centroid(*args)
Find the centroid (center of mass) of the collection containing only Points, Segments
or Polygons. The centroid is the weighted average of the individual centroid where the
weights are the lengths (of segments) or areas (of polygons). Overlapping regions will
add to the weight of that region.
If there are no objects (or a mixture of objects) then None is returned.
See Also:
sympy.geometry.point.Point (page 497), sympy.geometry.line.Segment (page 522),
sympy.geometry.polygon.Polygon (page 556)
Examples
>>> from sympy import Point, Segment, Polygon
>>> from sympy.geometry.util import centroid
>>> p = Polygon((0, 0), (10, 0), (10, 10))
>>> q = p.translate(0, 20)
>>> p.centroid, q.centroid
(Point(20/3, 10/3), Point(20/3, 70/3))
>>> centroid(p, q)
Point(20/3, 40/3)
>>> p, q = Segment((0, 0), (2, 0)), Segment((0, 0), (2, 2))
>>> centroid(p, q)
Point(1, -sqrt(2) + 2)
>>> centroid(Point(0, 0), Point(2, 0))
Point(1, 0)
Stacking 3 polygons on top of each other eectively triples the weight of that polygon:
>>> p = Polygon((0, 0), (1, 0), (1, 1), (0, 1))
>>> q = Polygon((1, 0), (3, 0), (3, 1), (1, 1))
>>> centroid(p, q)
Point(3/2, 1/2)
>>> centroid(p, p, p, q) # centroid x-coord shifts left
Point(11/10, 1/2)
Stacking the squares vertically above and below p has the same eect:
>>> centroid(p, p.translate(0, 1), p.translate(0, -1), q)
Point(11/10, 1/2)
Points
class sympy.geometry.point.Point
A point in a 2-dimensional Euclidean space.
5.10. Geometry Module
497
SymPy Documentation, Release 0.7.6
Parameters coords : sequence of 2 coordinate values.
Raises TypeError :
When trying to add or subtract points with dierent dimensions.
When trying to create a point with more than two dimensions. When
intersection is called with object other than a Point.
See Also:
sympy.geometry.line.Segment (page 522) Connects two Points
Examples
>>> from sympy.geometry import Point
>>> from sympy.abc import x
>>> Point(1, 2)
Point(1, 2)
>>> Point([1, 2])
Point(1, 2)
>>> Point(0, x)
Point(0, x)
Floats are automatically converted to Rational unless the evaluate ag is False:
>>> Point(0.5, 0.25)
Point(1/2, 1/4)
>>> Point(0.5, 0.25, evaluate=False)
Point(0.5, 0.25)
Attributes
x
y
length
distance(p)
The Euclidean distance from self to point p.
Parameters p : Point
Returns distance : number or symbolic expression.
See Also:
sympy.geometry.line.Segment.length (page 523)
Examples
>>> from sympy.geometry import Point
>>> p1, p2 = Point(1, 1), Point(4, 5)
>>> p1.distance(p2)
5
498
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> from sympy.abc import x, y
>>> p3 = Point(x, y)
>>> p3.distance(Point(0, 0))
sqrt(x**2 + y**2)
dot(p2)
Return dot product of self with another Point.
evalf(prec=None, **options)
Evaluate the coordinates of the point.
This method will, where possible, create and return a new Point where the coordinates are evaluated as oating point numbers to the precision indicated (default=15).
Returns point : Point
Examples
>>> from sympy import Point, Rational
>>> p1 = Point(Rational(1, 2), Rational(3, 2))
>>> p1
Point(1/2, 3/2)
>>> p1.evalf()
Point(0.5, 1.5)
intersection(o)
The intersection between this point and another point.
Parameters other : Point
Returns intersection : list of Points
Notes
The return value will either be an empty list if there is no intersection, otherwise it
will contain this point.
Examples
>>> from sympy import Point
>>> p1, p2, p3 = Point(0, 0), Point(1, 1), Point(0, 0)
>>> p1.intersection(p2)
[]
>>> p1.intersection(p3)
[Point(0, 0)]
is collinear(*points)
Is a sequence of points collinear?
Test whether or not a set of points are collinear. Returns True if the set of points
are collinear, or False otherwise.
Parameters points : sequence of Point
Returns is collinear : boolean
5.10. Geometry Module
499
SymPy Documentation, Release 0.7.6
See Also:
sympy.geometry.line.Line (page 517)
Notes
Slope is preserved everywhere on a line, so the slope between any two points on
the line should be the same. Take the rst two points, p1 and p2, and create a
translated point v1 with p1 as the origin. Now for every other point we create a
translated point, vi with p1 also as the origin. Note that these translations preserve
slope since everything is consistently translated to a new origin of p1. Since slope
is preserved then we have the following equality:
v1 slope = vi slope
v1.y/v1.x = vi.y/vi.x (due to translation)
v1.y*vi.x = vi.y*v1.x
v1.y*vi.x - vi.y*v1.x = 0 (*)
Hence, if we have a vi such that the equality in (*) is False then the points are not
collinear. We do this test for every point in the list, and if all pass then they are
collinear.
Examples
>>> from sympy import Point
>>> from sympy.abc import x
>>> p1, p2 = Point(0, 0), Point(1, 1)
>>> p3, p4, p5 = Point(2, 2), Point(x, x), Point(1, 2)
>>> Point.is_collinear(p1, p2, p3, p4)
True
>>> Point.is_collinear(p1, p2, p3, p5)
False
is concyclic(*points)
Is a sequence of points concyclic?
Test whether or not a sequence of points are concyclic (i.e., they lie on a circle).
Parameters points : sequence of Points
Returns is concyclic : boolean
True if points are concyclic, False otherwise.
See Also:
sympy.geometry.ellipse.Circle (page 554)
Notes
No points are not considered to be concyclic. One or two points are denitely concyclic and three points are conyclic i they are not collinear.
For more than three points, create a circle from the rst three points. If the circle
cannot be created (i.e., they are collinear) then all of the points cannot be concyclic.
500
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
If the circle is created successfully then simply check the remaining points for containment in the circle.
Examples
>>> from sympy.geometry import Point
>>> p1, p2 = Point(-1, 0), Point(1, 0)
>>> p3, p4 = Point(0, 1), Point(-1, 2)
>>> Point.is_concyclic(p1, p2, p3)
True
>>> Point.is_concyclic(p1, p2, p3, p4)
False
length
Treating a Point as a Line, this returns 0 for the length of a Point.
Examples
>>> from sympy import Point
>>> p = Point(0, 1)
>>> p.length
0
midpoint(p)
The midpoint between self and point p.
Parameters p : Point
Returns midpoint : Point
See Also:
sympy.geometry.line.Segment.midpoint (page 524)
Examples
>>> from sympy.geometry import Point
>>> p1, p2 = Point(1, 1), Point(13, 5)
>>> p1.midpoint(p2)
Point(7, 3)
n(prec=None, **options)
Evaluate the coordinates of the point.
This method will, where possible, create and return a new Point where the coordinates are evaluated as oating point numbers to the precision indicated (default=15).
Returns point : Point
Examples
5.10. Geometry Module
501
SymPy Documentation, Release 0.7.6
>>> from sympy import Point, Rational
>>> p1 = Point(Rational(1, 2), Rational(3, 2))
>>> p1
Point(1/2, 3/2)
>>> p1.evalf()
Point(0.5, 1.5)
rotate(angle, pt=None)
Rotate angle radians counterclockwise about Point pt.
See Also:
rotate (page 502), scale (page 502)
Examples
>>> from sympy import Point, pi
>>> t = Point(1, 0)
>>> t.rotate(pi/2)
Point(0, 1)
>>> t.rotate(pi/2, (2, 0))
Point(2, -1)
scale(x=1, y=1, pt=None)
Scale the coordinates of the Point by multiplying by x and y after subtracting pt
default is (0, 0) and then adding pt back again (i.e. pt is the point of reference for
the scaling).
See Also:
rotate (page 502), translate (page 502)
Examples
>>> from sympy import Point
>>> t = Point(1, 1)
>>> t.scale(2)
Point(2, 1)
>>> t.scale(2, 2)
Point(2, 2)
transform(matrix)
Return the point after applying the transformation described by the 3x3 Matrix,
matrix.
See Also:
geometry.entity.rotate, geometry.entity.scale, geometry.entity.translate
translate(x=0, y=0)
Shift the Point by adding x and y to the coordinates of the Point.
See Also:
rotate (page 502), scale (page 502)
502
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import Point
>>> t = Point(0, 1)
>>> t.translate(2)
Point(2, 1)
>>> t.translate(2, 2)
Point(2, 3)
>>> t + Point(2, 2)
Point(2, 3)
x
Returns the X coordinate of the Point.
Examples
>>> from sympy import Point
>>> p = Point(0, 1)
>>> p.x
0
y
Returns the Y coordinate of the Point.
Examples
>>> from sympy import Point
>>> p = Point(0, 1)
>>> p.y
1
3D Point
Geometrical Points.
Contains Point3D
class sympy.geometry.point3d.Point3D
A point in a 3-dimensional Euclidean space.
Parameters coords : sequence of 3 coordinate values.
Raises NotImplementedError :
When trying to create a point other than 2 or 3 dimensions. When
intersection is called with object other than a Point.
TypeError :
When trying to add or subtract points with dierent dimensions.
Notes
Currently only 2-dimensional and 3-dimensional points are supported.
5.10. Geometry Module
503
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import Point3D
>>> from sympy.abc import x
>>> Point3D(1, 2, 3)
Point3D(1, 2, 3)
>>> Point3D([1, 2, 3])
Point3D(1, 2, 3)
>>> Point3D(0, x, 3)
Point3D(0, x, 3)
Floats are automatically converted to Rational unless the evaluate ag is False:
>>> Point3D(0.5, 0.25, 2)
Point3D(1/2, 1/4, 2)
>>> Point3D(0.5, 0.25, 3, evaluate=False)
Point3D(0.5, 0.25, 3)
Attributes
x
y
z
length
static are collinear(*points)
Is a sequence of points collinear?
Test whether or not a set of points are collinear. Returns True if the set of points
are collinear, or False otherwise.
Parameters points : sequence of Point
Returns are collinear : boolean
See Also:
sympy.geometry.line3d.Line3D (page 525)
Examples
>>> from sympy import Point3D, Matrix
>>> from sympy.abc import x
>>> p1, p2 = Point3D(0, 0, 0), Point3D(1, 1, 1)
>>> p3, p4, p5 = Point3D(2, 2, 2), Point3D(x, x, x), Point3D(1, 2, 6)
>>> Point3D.are_collinear(p1, p2, p3, p4)
True
>>> Point3D.are_collinear(p1, p2, p3, p5)
False
static are coplanar(*points)
This function tests whether passed points are coplanar or not. It uses the fact that
the triple scalar product of three vectors vanishes if the vectors are coplanar. Which
means that the volume of the solid described by them will have to be zero for coplanarity.
Parameters A set of points 3D points :
504
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Returns boolean :
Examples
>>> from sympy import Point3D
>>> p1 = Point3D(1, 2, 2)
>>> p2 = Point3D(2, 7, 2)
>>> p3 = Point3D(0, 0, 2)
>>> p4 = Point3D(1, 1, 2)
>>> Point3D.are_coplanar(p1, p2, p3, p4)
True
>>> p5 = Point3D(0, 1, 3)
>>> Point3D.are_coplanar(p1, p2, p3, p5)
False
direction cosine(point)
Gives the direction cosine between 2 points
Parameters p : Point3D
Returns list :
Examples
>>> from sympy import Point3D
>>> p1 = Point3D(1, 2, 3)
>>> p1.direction_cosine(Point3D(2, 3, 5))
[sqrt(6)/6, sqrt(6)/6, sqrt(6)/3]
direction ratio(point)
Gives the direction ratio between 2 points
Parameters p : Point3D
Returns list :
Examples
>>>
>>>
>>>
[1,
from sympy import Point3D
p1 = Point3D(1, 2, 3)
p1.direction_ratio(Point3D(2, 3, 5))
1, 2]
distance(p)
The Euclidean distance from self to point p.
Parameters p : Point
Returns distance : number or symbolic expression.
See Also:
sympy.geometry.line.Segment.length (page 523)
5.10. Geometry Module
505
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import Point3D
>>> p1, p2 = Point3D(1, 1, 1), Point3D(4, 5, 0)
>>> p1.distance(p2)
sqrt(26)
>>> from sympy.abc import x, y, z
>>> p3 = Point3D(x, y, z)
>>> p3.distance(Point3D(0, 0, 0))
sqrt(x**2 + y**2 + z**2)
dot(p2)
Return dot product of self with another Point.
evalf(prec=None, **options)
Evaluate the coordinates of the point.
This method will, where possible, create and return a new Point where the coordinates are evaluated as oating point numbers to the precision indicated (default=15).
Returns point : Point
Examples
>>> from sympy import Point3D, Rational
>>> p1 = Point3D(Rational(1, 2), Rational(3, 2), Rational(5, 2))
>>> p1
Point3D(1/2, 3/2, 5/2)
>>> p1.evalf()
Point3D(0.5, 1.5, 2.5)
intersection(o)
The intersection between this point and another point.
Parameters other : Point
Returns intersection : list of Points
Notes
The return value will either be an empty list if there is no intersection, otherwise it
will contain this point.
Examples
>>> from sympy import Point3D
>>> p1, p2, p3 = Point3D(0, 0, 0), Point3D(1, 1, 1), Point3D(0, 0, 0)
>>> p1.intersection(p2)
[]
>>> p1.intersection(p3)
[Point3D(0, 0, 0)]
506
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
length
Treating a Point as a Line, this returns 0 for the length of a Point.
Examples
>>> from sympy import Point3D
>>> p = Point3D(0, 1, 1)
>>> p.length
0
midpoint(p)
The midpoint between self and point p.
Parameters p : Point
Returns midpoint : Point
See Also:
sympy.geometry.line.Segment.midpoint (page 524)
Examples
>>> from sympy import Point3D
>>> p1, p2 = Point3D(1, 1, 1), Point3D(13, 5, 1)
>>> p1.midpoint(p2)
Point3D(7, 3, 1)
n(prec=None, **options)
Evaluate the coordinates of the point.
This method will, where possible, create and return a new Point where the coordinates are evaluated as oating point numbers to the precision indicated (default=15).
Returns point : Point
Examples
>>> from sympy import Point3D, Rational
>>> p1 = Point3D(Rational(1, 2), Rational(3, 2), Rational(5, 2))
>>> p1
Point3D(1/2, 3/2, 5/2)
>>> p1.evalf()
Point3D(0.5, 1.5, 2.5)
scale(x=1, y=1, z=1, pt=None)
Scale the coordinates of the Point by multiplying by x and y after subtracting pt
default is (0, 0) and then adding pt back again (i.e. pt is the point of reference for
the scaling).
See Also:
translate (page 508)
5.10. Geometry Module
507
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import Point3D
>>> t = Point3D(1, 1, 1)
>>> t.scale(2)
Point3D(2, 1, 1)
>>> t.scale(2, 2)
Point3D(2, 2, 1)
transform(matrix)
Return the point after applying the transformation described by the 3x3 Matrix,
matrix.
See Also:
geometry.entity.rotate, geometry.entity.scale, geometry.entity.translate
translate(x=0, y=0, z=0)
Shift the Point by adding x and y to the coordinates of the Point.
See Also:
rotate, scale (page 507)
Examples
>>> from sympy import Point3D
>>> t = Point3D(0, 1, 1)
>>> t.translate(2)
Point3D(2, 1, 1)
>>> t.translate(2, 2)
Point3D(2, 3, 1)
>>> t + Point3D(2, 2, 2)
Point3D(2, 3, 3)
x
Returns the X coordinate of the Point.
Examples
>>> from sympy import Point3D
>>> p = Point3D(0, 1, 3)
>>> p.x
0
y
Returns the Y coordinate of the Point.
Examples
>>> from sympy import Point3D
>>> p = Point3D(0, 1, 2)
>>> p.y
1
508
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
z
Returns the Z coordinate of the Point.
Examples
>>> from sympy import Point3D
>>> p = Point3D(0, 1, 1)
>>> p.z
1
Lines
class sympy.geometry.line.LinearEntity
A base class for all linear entities (line, ray and segment) in a 2-dimensional Euclidean
space.
See Also:
sympy.geometry.entity.GeometryEntity (page 493)
Notes
This is an abstract class and is not meant to be instantiated.
Attributes
p1
p2
coecients
slope
points
angle between(l1, l2)
The angle formed between the two linear entities.
Parameters l1 : LinearEntity
l2 : LinearEntity
Returns angle : angle in radians
See Also:
is perpendicular (page 512)
Notes
From the dot product of vectors v1 and v2 it is known that:
dot(v1, v2) = |v1|*|v2|*cos(A)
where A is the angle formed between the two vectors. We can get the directional
vectors of the two lines and readily nd the angle between the two using the above
formula.
5.10. Geometry Module
509
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import Point, Line
>>> p1, p2, p3 = Point(0, 0), Point(0, 4), Point(2, 0)
>>> l1, l2 = Line(p1, p2), Line(p1, p3)
>>> l1.angle_between(l2)
pi/2
arbitrary point(parameter=t)
A parameterized point on the Line.
Parameters parameter : str, optional
The name of the parameter which will be used for the parametric
point. The default value is t. When this parameter is 0, the rst
point used to dene the line will be returned, and when it is 1 the
second point will be returned.
Returns point : Point
Raises ValueError :
When parameter already appears in the Lines denition.
See Also:
sympy.geometry.point.Point (page 497)
Examples
>>> from sympy import Point, Line
>>> p1, p2 = Point(1, 0), Point(5, 3)
>>> l1 = Line(p1, p2)
>>> l1.arbitrary_point()
Point(4*t + 1, 3*t)
static are concurrent(*lines)
Is a sequence of linear entities concurrent?
Two or more linear entities are concurrent if they all intersect at a single point.
Parameters lines : a sequence of linear entities.
Returns True : if the set of linear entities are concurrent,
False : otherwise.
See Also:
sympy.geometry.util.intersection (page 495)
Notes
Simply take the rst two lines and nd their intersection. If there is no intersection,
then the rst two lines were parallel and had no intersection so concurrency is impossible amongst the whole set. Otherwise, check to see if the intersection point of
the rst two lines is a member on the rest of the lines. If so, the lines are concurrent.
510
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import Point, Line, Line3D
>>> p1, p2 = Point(0, 0), Point(3, 5)
>>> p3, p4 = Point(-2, -2), Point(0, 2)
>>> l1, l2, l3 = Line(p1, p2), Line(p1, p3), Line(p1, p4)
>>> Line.are_concurrent(l1, l2, l3)
True
>>> l4 = Line(p2, p3)
>>> Line.are_concurrent(l2, l3, l4)
False
coefficients
The coecients (a, b, c) for ax + by + c = 0.
See Also:
sympy.geometry.line.Line.equation (page 518)
Examples
>>> from sympy import Point, Line
>>> from sympy.abc import x, y
>>> p1, p2 = Point(0, 0), Point(5, 3)
>>> l = Line(p1, p2)
>>> l.coefficients
(-3, 5, 0)
>>> p3 = Point(x, y)
>>> l2 = Line(p1, p3)
>>> l2.coefficients
(-y, x, 0)
contains(other)
Subclasses should implement this method and should return True if other is on the
boundaries of self; False if not on the boundaries of self; None if a determination
cannot be made.
intersection(o)
The intersection with another geometrical entity.
Parameters o : Point or LinearEntity
Returns intersection : list of geometrical entities
See Also:
sympy.geometry.point.Point (page 497)
Examples
>>> from sympy import Point, Line, Segment
>>> p1, p2, p3 = Point(0, 0), Point(1, 1), Point(7, 7)
>>> l1 = Line(p1, p2)
>>> l1.intersection(p3)
[Point(7, 7)]
5.10. Geometry Module
511
SymPy Documentation, Release 0.7.6
>>> p4, p5 = Point(5, 0), Point(0, 3)
>>> l2 = Line(p4, p5)
>>> l1.intersection(l2)
[Point(15/8, 15/8)]
>>> p6, p7 = Point(0, 5), Point(2, 6)
>>> s1 = Segment(p6, p7)
>>> l1.intersection(s1)
[]
is parallel(l1, l2)
Are two linear entities parallel?
Parameters l1 : LinearEntity
l2 : LinearEntity
Returns True : if l1 and l2 are parallel,
False : otherwise.
See Also:
coefficients (page 511)
Examples
>>> from sympy import Point, Line
>>> p1, p2 = Point(0, 0), Point(1, 1)
>>> p3, p4 = Point(3, 4), Point(6, 7)
>>> l1, l2 = Line(p1, p2), Line(p3, p4)
>>> Line.is_parallel(l1, l2)
True
>>> p5 = Point(6, 6)
>>> l3 = Line(p3, p5)
>>> Line.is_parallel(l1, l3)
False
is perpendicular(l1, l2)
Are two linear entities perpendicular?
Parameters l1 : LinearEntity
l2 : LinearEntity
Returns True : if l1 and l2 are perpendicular,
False : otherwise.
See Also:
coefficients (page 511)
Examples
>>> from sympy import Point, Line
>>> p1, p2, p3 = Point(0, 0), Point(1, 1), Point(-1, 1)
>>> l1, l2 = Line(p1, p2), Line(p1, p3)
512
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> l1.is_perpendicular(l2)
True
>>> p4 = Point(5, 3)
>>> l3 = Line(p1, p4)
>>> l1.is_perpendicular(l3)
False
is similar(other)
Return True if self and other are contained in the same line.
Examples
>>> from sympy import Point, Line
>>> p1, p2, p3 = Point(0, 1), Point(3, 4), Point(2, 3)
>>> l1 = Line(p1, p2)
>>> l2 = Line(p1, p3)
>>> l1.is_similar(l2)
True
length
The length of the line.
Examples
>>>
>>>
>>>
>>>
oo
from sympy import Point, Line
p1, p2 = Point(0, 0), Point(3, 5)
l1 = Line(p1, p2)
l1.length
p1
The rst dening point of a linear entity.
See Also:
sympy.geometry.point.Point (page 497)
Examples
>>> from sympy import Point, Line
>>> p1, p2 = Point(0, 0), Point(5, 3)
>>> l = Line(p1, p2)
>>> l.p1
Point(0, 0)
p2
The second dening point of a linear entity.
See Also:
sympy.geometry.point.Point (page 497)
5.10. Geometry Module
513
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import Point, Line
>>> p1, p2 = Point(0, 0), Point(5, 3)
>>> l = Line(p1, p2)
>>> l.p2
Point(5, 3)
parallel line(p)
Create a new Line parallel to this linear entity which passes through the point p.
Parameters p : Point
Returns line : Line
See Also:
is parallel (page 512)
Examples
>>> from sympy import Point, Line
>>> p1, p2, p3 = Point(0, 0), Point(2, 3), Point(-2, 2)
>>> l1 = Line(p1, p2)
>>> l2 = l1.parallel_line(p3)
>>> p3 in l2
True
>>> l1.is_parallel(l2)
True
perpendicular line(p)
Create a new Line perpendicular to this linear entity which passes through the point
p.
Parameters p : Point
Returns line : Line
See Also:
is perpendicular (page 512), perpendicular segment (page 514)
Examples
>>> from sympy import Point, Line
>>> p1, p2, p3 = Point(0, 0), Point(2, 3), Point(-2, 2)
>>> l1 = Line(p1, p2)
>>> l2 = l1.perpendicular_line(p3)
>>> p3 in l2
True
>>> l1.is_perpendicular(l2)
True
perpendicular segment(p)
Create a perpendicular line segment from p to this line.
The enpoints of the segment are p and the closest point in the line containing self.
(If self is not a line, the point might not be in self.)
514
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Parameters p : Point
Returns segment : Segment
See Also:
perpendicular line (page 514)
Notes
Returns p itself if p is on this linear entity.
Examples
>>> from sympy import Point, Line
>>> p1, p2, p3 = Point(0, 0), Point(1, 1), Point(0, 2)
>>> l1 = Line(p1, p2)
>>> s1 = l1.perpendicular_segment(p3)
>>> l1.is_perpendicular(s1)
True
>>> p3 in s1
True
>>> l1.perpendicular_segment(Point(4, 0))
Segment(Point(2, 2), Point(4, 0))
points
The two points used to dene this linear entity.
Returns points : tuple of Points
See Also:
sympy.geometry.point.Point (page 497)
Examples
>>> from sympy import Point, Line
>>> p1, p2 = Point(0, 0), Point(5, 11)
>>> l1 = Line(p1, p2)
>>> l1.points
(Point(0, 0), Point(5, 11))
projection(o)
Project a point, line, ray, or segment onto this linear entity.
Parameters other : Point or LinearEntity (Line, Ray, Segment)
Returns projection : Point or LinearEntity (Line, Ray, Segment)
The return type matches the type of the parameter other.
Raises GeometryError :
When method is unable to perform projection.
See Also:
sympy.geometry.point.Point (page 497), perpendicular line (page 514)
5.10. Geometry Module
515
SymPy Documentation, Release 0.7.6
Notes
A projection involves taking the two points that dene the linear entity and projecting those points onto a Line and then reforming the linear entity using these
projections. A point P is projected onto a line L by nding the point on L that is
closest to P. This point is the intersection of L and the line perpendicular to L that
passes through P.
Examples
>>> from sympy import Point, Line, Segment, Rational
>>> p1, p2, p3 = Point(0, 0), Point(1, 1), Point(Rational(1, 2), 0)
>>> l1 = Line(p1, p2)
>>> l1.projection(p3)
Point(1/4, 1/4)
>>> p4, p5 = Point(10, 0), Point(12, 1)
>>> s1 = Segment(p4, p5)
>>> l1.projection(s1)
Segment(Point(5, 5), Point(13/2, 13/2))
random point()
A random point on a LinearEntity.
Returns point : Point
See Also:
sympy.geometry.point.Point (page 497)
Examples
>>> from sympy import Point, Line
>>> p1, p2 = Point(0, 0), Point(5, 3)
>>> l1 = Line(p1, p2)
>>> p3 = l1.random_point()
>>> # random point - dont know its coords in advance
>>> p3
Point(...)
>>> # point should belong to the line
>>> p3 in l1
True
slope
The slope of this linear entity, or innity if vertical.
Returns slope : number or sympy expression
See Also:
coefficients (page 511)
516
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>>
>>>
>>>
>>>
5/3
from sympy import Point, Line
p1, p2 = Point(0, 0), Point(3, 5)
l1 = Line(p1, p2)
l1.slope
>>> p3 = Point(0, 4)
>>> l2 = Line(p1, p3)
>>> l2.slope
oo
class sympy.geometry.line.Line
An innite line in space.
A line is declared with two distinct points or a point and slope as dened using keyword
slope.
Parameters p1 : Point
pt : Point
slope : sympy expression
See Also:
sympy.geometry.point.Point (page 497)
Notes
At the moment only lines in a 2D space can be declared, because Points can be dened
only for 2D spaces.
Examples
>>> import sympy
>>> from sympy import Point
>>> from sympy.abc import L
>>> from sympy.geometry import Line, Segment
>>> L = Line(Point(2,3), Point(3,5))
>>> L
Line(Point(2, 3), Point(3, 5))
>>> L.points
(Point(2, 3), Point(3, 5))
>>> L.equation()
-2*x + y + 1
>>> L.coefficients
(-2, 1, 1)
Instantiate with keyword slope:
>>> Line(Point(0, 0), slope=0)
Line(Point(0, 0), Point(1, 0))
Instantiate with another linear object
5.10. Geometry Module
517
SymPy Documentation, Release 0.7.6
>>> s = Segment((0, 0), (0, 1))
>>> Line(s).equation()
x
contains(o)
Return True if o is on this Line, or False otherwise.
Examples
>>> from sympy import Line,Point
>>> p1, p2 = Point(0, 1), Point(3, 4)
>>> l = Line(p1, p2)
>>> l.contains(p1)
True
>>> l.contains((0, 1))
True
>>> l.contains((0, 0))
False
distance(o)
Finds the shortest distance between a line and a point.
Raises NotImplementedError is raised if o is not a Point :
Examples
>>> from sympy import Point, Line
>>> p1, p2 = Point(0, 0), Point(1, 1)
>>> s = Line(p1, p2)
>>> s.distance(Point(-1, 1))
sqrt(2)
>>> s.distance((-1, 2))
3*sqrt(2)/2
equal(other)
Returns True if self and other are the same mathematical entities
equation(x=x, y=y)
The equation of the line: ax + by + c.
Parameters x : str, optional
The name to use for the x-axis, default value is x.
y : str, optional
The name to use for the y-axis, default value is y.
Returns equation : sympy expression
See Also:
LinearEntity.coefficients (page 511)
518
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import Point, Line
>>> p1, p2 = Point(1, 0), Point(5, 3)
>>> l1 = Line(p1, p2)
>>> l1.equation()
-3*x + 4*y + 3
plot interval(parameter=t)
The plot interval for the default geometric plot of line. Gives values that will produce
a line that is +/- 5 units long (where a unit is the distance between the two points
that dene the line).
Parameters parameter : str, optional
Default value is t.
Returns plot interval : list (plot interval)
[parameter, lower bound, upper bound]
Examples
>>>
>>>
>>>
>>>
[t,
from sympy import Point, Line
p1, p2 = Point(0, 0), Point(5, 3)
l1 = Line(p1, p2)
l1.plot_interval()
-5, 5]
class sympy.geometry.line.Ray
A Ray is a semi-line in the space with a source point and a direction.
Parameters p1 : Point
The source of the Ray
p2 : Point or radian value
This point determines the direction in which the Ray propagates. If
given as an angle it is interpreted in radians with the positive direction being ccw.
See Also:
sympy.geometry.point.Point (page 497), Line (page 517)
Notes
At the moment only rays in a 2D space can be declared, because Points can be dened
only for 2D spaces.
Examples
>>>
>>>
>>>
>>>
import sympy
from sympy import Point, pi
from sympy.abc import r
from sympy.geometry import Ray
5.10. Geometry Module
519
SymPy Documentation, Release 0.7.6
>>> r = Ray(Point(2, 3), Point(3, 5))
>>> r = Ray(Point(2, 3), Point(3, 5))
>>> r
Ray(Point(2, 3), Point(3, 5))
>>> r.points
(Point(2, 3), Point(3, 5))
>>> r.source
Point(2, 3)
>>> r.xdirection
oo
>>> r.ydirection
oo
>>> r.slope
2
>>> Ray(Point(0, 0), angle=pi/4).slope
1
Attributes
source
xdirection
ydirection
contains(o)
Is other GeometryEntity contained in this Ray?
Examples
>>> from sympy import Ray,Point,Segment
>>> p1, p2 = Point(0, 0), Point(4, 4)
>>> r = Ray(p1, p2)
>>> r.contains(p1)
True
>>> r.contains((1, 1))
True
>>> r.contains((1, 3))
False
>>> s = Segment((1, 1), (2, 2))
>>> r.contains(s)
True
>>> s = Segment((1, 2), (2, 5))
>>> r.contains(s)
False
>>> r1 = Ray((2, 2), (3, 3))
>>> r.contains(r1)
True
>>> r1 = Ray((2, 2), (3, 5))
>>> r.contains(r1)
False
distance(o)
Finds the shortest distance between the ray and a point.
Raises NotImplementedError is raised if o is not a Point :
520
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import Point, Ray
>>> p1, p2 = Point(0, 0), Point(1, 1)
>>> s = Ray(p1, p2)
>>> s.distance(Point(-1, -1))
sqrt(2)
>>> s.distance((-1, 2))
3*sqrt(2)/2
equals(other)
Returns True if self and other are the same mathematical entities
plot interval(parameter=t)
The plot interval for the default geometric plot of the Ray. Gives values that will
produce a ray that is 10 units long (where a unit is the distance between the two
points that dene the ray).
Parameters parameter : str, optional
Default value is t.
Returns plot interval : list
[parameter, lower bound, upper bound]
Examples
>>>
>>>
>>>
[t,
from sympy import Point, Ray, pi
r = Ray((0, 0), angle=pi/4)
r.plot_interval()
0, 10]
source
The point from which the ray emanates.
See Also:
sympy.geometry.point.Point (page 497)
Examples
>>> from sympy import Point, Ray
>>> p1, p2 = Point(0, 0), Point(4, 1)
>>> r1 = Ray(p1, p2)
>>> r1.source
Point(0, 0)
xdirection
The x direction of the ray.
Positive innity if the ray points in the positive x direction, negative innity if the
ray points in the negative x direction, or 0 if the ray is vertical.
See Also:
ydirection (page 522)
5.10. Geometry Module
521
SymPy Documentation, Release 0.7.6
Examples
>>>
>>>
>>>
>>>
oo
>>>
0
from sympy import Point, Ray
p1, p2, p3 = Point(0, 0), Point(1, 1), Point(0, -1)
r1, r2 = Ray(p1, p2), Ray(p1, p3)
r1.xdirection
r2.xdirection
ydirection
The y direction of the ray.
Positive innity if the ray points in the positive y direction, negative innity if the
ray points in the negative y direction, or 0 if the ray is horizontal.
See Also:
xdirection (page 521)
Examples
>>>
>>>
>>>
>>>
-oo
>>>
0
from sympy import Point, Ray
p1, p2, p3 = Point(0, 0), Point(-1, -1), Point(-1, 0)
r1, r2 = Ray(p1, p2), Ray(p1, p3)
r1.ydirection
r2.ydirection
class sympy.geometry.line.Segment
A undirected line segment in space.
Parameters p1 : Point
p2 : Point
See Also:
sympy.geometry.point.Point (page 497), Line (page 517)
Notes
At the moment only segments in a 2D space can be declared, because Points can be
dened only for 2D spaces.
Examples
>>> import sympy
>>> from sympy import Point
>>> from sympy.abc import s
>>> from sympy.geometry import Segment
>>> Segment((1, 0), (1, 1)) # tuples are interpreted as pts
Segment(Point(1, 0), Point(1, 1))
>>> s = Segment(Point(4, 3), Point(1, 1))
>>> s
522
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Segment(Point(1, 1), Point(4, 3))
>>> s.points
(Point(1, 1), Point(4, 3))
>>> s.slope
2/3
>>> s.length
sqrt(13)
>>> s.midpoint
Point(5/2, 2)
Attributes
length
midpoint
number or sympy expression
Point
contains(other)
Is the other GeometryEntity contained within this Segment?
Examples
>>> from sympy import Point, Segment
>>> p1, p2 = Point(0, 1), Point(3, 4)
>>> s = Segment(p1, p2)
>>> s2 = Segment(p2, p1)
>>> s.contains(s2)
True
distance(o)
Finds the shortest distance between a line segment and a point.
Raises NotImplementedError is raised if o is not a Point :
Examples
>>> from sympy import Point, Segment
>>> p1, p2 = Point(0, 1), Point(3, 4)
>>> s = Segment(p1, p2)
>>> s.distance(Point(10, 15))
sqrt(170)
>>> s.distance((0, 12))
sqrt(73)
length
The length of the line segment.
See Also:
sympy.geometry.point.Point.distance (page 498)
Examples
5.10. Geometry Module
523
SymPy Documentation, Release 0.7.6
>>>
>>>
>>>
>>>
5
from sympy import Point, Segment
p1, p2 = Point(0, 0), Point(4, 3)
s1 = Segment(p1, p2)
s1.length
midpoint
The midpoint of the line segment.
See Also:
sympy.geometry.point.Point.midpoint (page 501)
Examples
>>> from sympy import Point, Segment
>>> p1, p2 = Point(0, 0), Point(4, 3)
>>> s1 = Segment(p1, p2)
>>> s1.midpoint
Point(2, 3/2)
perpendicular bisector(p=None)
The perpendicular bisector of this segment.
If no point is specied or the point specied is not on the bisector then the bisector
is returned as a Line. Otherwise a Segment is returned that joins the point specied
and the intersection of the bisector and the segment.
Parameters p : Point
Returns bisector : Line or Segment
See Also:
LinearEntity.perpendicular segment (page 514)
Examples
>>> from sympy import Point, Segment
>>> p1, p2, p3 = Point(0, 0), Point(6, 6), Point(5, 1)
>>> s1 = Segment(p1, p2)
>>> s1.perpendicular_bisector()
Line(Point(3, 3), Point(9, -3))
>>> s1.perpendicular_bisector(p3)
Segment(Point(3, 3), Point(5, 1))
plot interval(parameter=t)
The plot interval for the default geometric plot of the Segment gives values that will
produce the full segment in a plot.
Parameters parameter : str, optional
Default value is t.
Returns plot interval : list
[parameter, lower bound, upper bound]
524
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>>
>>>
>>>
>>>
[t,
from sympy import Point, Segment
p1, p2 = Point(0, 0), Point(5, 3)
s1 = Segment(p1, p2)
s1.plot_interval()
0, 1]
3D Line
Line-like geometrical entities.
Contains LinearEntity3D Line3D Ray3D Segment3D
class sympy.geometry.line3d.Line3D
An innite 3D line in space.
A line is declared with two distinct points or a point and direction ratio as dened using
keyword directionr atio.
Parameters p1 : Point3D
pt : Point3D
direction ratio : list
See Also:
sympy.geometry.point3d.Point3D (page 503)
Examples
>>> import sympy
>>> from sympy import Point3D
>>> from sympy.abc import L
>>> from sympy.geometry import Line3D, Segment3D
>>> L = Line3D(Point3D(2, 3, 4), Point3D(3, 5, 1))
>>> L
Line3D(Point3D(2, 3, 4), Point3D(3, 5, 1))
>>> L.points
(Point3D(2, 3, 4), Point3D(3, 5, 1))
contains(o)
Return True if o is on this Line, or False otherwise.
Examples
>>>
>>>
>>>
>>>
>>>
>>>
>>>
from sympy import Line3D
a = (0, 0, 0)
b = (1, 1, 1)
c = (2, 2, 2)
l1 = Line3D(a, b)
l2 = Line3D(b, a)
l1 == l2
5.10. Geometry Module
525
SymPy Documentation, Release 0.7.6
False
>>> l1 in l2
True
distance(o)
Finds the shortest distance between a line and a point.
Raises NotImplementedError is raised if o is not an instance of
Point3D :
Examples
>>> from sympy import Point3D, Line3D
>>> p1, p2 = Point3D(0, 0, 0), Point3D(1, 1, 1)
>>> s = Line3D(p1, p2)
>>> s.distance(Point3D(-1, 1, 1))
2*sqrt(6)/3
>>> s.distance((-1, 1, 1))
2*sqrt(6)/3
equals(other)
Returns True if self and other are the same mathematical entities
equation(x=x, y=y, z=z, k=k)
The equation of the line in 3D
Parameters x : str, optional
The name to use for the x-axis, default value is x.
y : str, optional
The name to use for the y-axis, default value is y.
z : str, optional
The name to use for the x-axis, default value is z.
Returns equation : tuple
Examples
>>> from sympy import Point3D, Line3D
>>> p1, p2 = Point3D(1, 0, 0), Point3D(5, 3, 0)
>>> l1 = Line3D(p1, p2)
>>> l1.equation()
(x/4 - 1/4, y/3, zoo*z, k)
plot interval(parameter=t)
The plot interval for the default geometric plot of line. Gives values that will produce
a line that is +/- 5 units long (where a unit is the distance between the two points
that dene the line).
Parameters parameter : str, optional
Default value is t.
Returns plot interval : list (plot interval)
[parameter, lower bound, upper bound]
526
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>>
>>>
>>>
>>>
[t,
from sympy import Point3D, Line3D
p1, p2 = Point3D(0, 0, 0), Point3D(5, 3, 1)
l1 = Line3D(p1, p2)
l1.plot_interval()
-5, 5]
class sympy.geometry.line3d.LinearEntity3D
An base class for all linear entities (line, ray and segment) in a 3-dimensional Euclidean
space.
Notes
This is a base class and is not meant to be instantiated.
Attributes
p1
p2
direction ratio
direction cosine
points
angle between(l1, l2)
The angle formed between the two linear entities.
Parameters l1 : LinearEntity
l2 : LinearEntity
Returns angle : angle in radians
See Also:
is perpendicular (page 530)
Notes
From the dot product of vectors v1 and v2 it is known that:
dot(v1, v2) = |v1|*|v2|*cos(A)
where A is the angle formed between the two vectors. We can get the directional
vectors of the two lines and readily nd the angle between the two using the above
formula.
Examples
>>> from sympy import Point3D, Line3D
>>> p1, p2, p3 = Point3D(0, 0, 0), Point3D(1, 1, 1), Point3D(-1, 2, 0)
>>> l1, l2 = Line3D(p1, p2), Line3D(p2, p3)
>>> l1.angle_between(l2)
acos(-sqrt(2)/3)
5.10. Geometry Module
527
SymPy Documentation, Release 0.7.6
arbitrary point(parameter=t)
A parameterized point on the Line.
Parameters parameter : str, optional
The name of the parameter which will be used for the parametric
point. The default value is t. When this parameter is 0, the rst
point used to dene the line will be returned, and when it is 1 the
second point will be returned.
Returns point : Point3D
Raises ValueError :
When parameter already appears in the Lines denition.
See Also:
sympy.geometry.point3d.Point3D (page 503)
Examples
>>> from sympy import Point3D, Line3D
>>> p1, p2 = Point3D(1, 0, 0), Point3D(5, 3, 1)
>>> l1 = Line3D(p1, p2)
>>> l1.arbitrary_point()
Point3D(4*t + 1, 3*t, t)
static are concurrent(*lines)
Is a sequence of linear entities concurrent?
Two or more linear entities are concurrent if they all intersect at a single point.
Parameters lines : a sequence of linear entities.
Returns True : if the set of linear entities are concurrent,
False : otherwise.
See Also:
sympy.geometry.util.intersection (page 495)
Notes
Simply take the rst two lines and nd their intersection. If there is no intersection,
then the rst two lines were parallel and had no intersection so concurrency is impossible amongst the whole set. Otherwise, check to see if the intersection point of
the rst two lines is a member on the rest of the lines. If so, the lines are concurrent.
Examples
>>> from sympy import Point3D, Line3D
>>> p1, p2 = Point3D(0, 0, 0), Point3D(3, 5, 2)
>>> p3, p4 = Point3D(-2, -2, -2), Point3D(0, 2, 1)
>>> l1, l2, l3 = Line3D(p1, p2), Line3D(p1, p3), Line3D(p1, p4)
>>> Line3D.are_concurrent(l1, l2, l3)
True
528
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> l4 = Line3D(p2, p3)
>>> Line3D.are_concurrent(l2, l3, l4)
False
contains(other)
Subclasses should implement this method and should return True if other is on the
boundaries of self; False if not on the boundaries of self; None if a determination
cannot be made.
direction cosine
The normalized direction ratio of a given line in 3D.
See Also:
sympy.geometry.line.Line.equation (page 518)
Examples
>>> from sympy import Point3D, Line3D
>>> p1, p2 = Point3D(0, 0, 0), Point3D(5, 3, 1)
>>> l = Line3D(p1, p2)
>>> l.direction_cosine
[sqrt(35)/7, 3*sqrt(35)/35, sqrt(35)/35]
>>> sum(i**2 for i in _)
1
direction ratio
The direction ratio of a given line in 3D.
See Also:
sympy.geometry.line.Line.equation (page 518)
Examples
>>>
>>>
>>>
>>>
[5,
from sympy import Point3D, Line3D
p1, p2 = Point3D(0, 0, 0), Point3D(5, 3, 1)
l = Line3D(p1, p2)
l.direction_ratio
3, 1]
intersection(o)
The intersection with another geometrical entity.
Parameters o : Point or LinearEntity3D
Returns intersection : list of geometrical entities
See Also:
sympy.geometry.point3d.Point3D (page 503)
Examples
>>> from sympy import Point3D, Line3D, Segment3D
>>> p1, p2, p3 = Point3D(0, 0, 0), Point3D(1, 1, 1), Point3D(7, 7, 7)
>>> l1 = Line3D(p1, p2)
5.10. Geometry Module
529
SymPy Documentation, Release 0.7.6
>>> l1.intersection(p3)
[Point3D(7, 7, 7)]
>>> l1 = Line3D(Point3D(4,19,12), Point3D(5,25,17))
>>> l2 = Line3D(Point3D(-3, -15, -19), direction_ratio=[2,8,8])
>>> l1.intersection(l2)
[Point3D(1, 1, -3)]
>>> p6, p7 = Point3D(0, 5, 2), Point3D(2, 6, 3)
>>> s1 = Segment3D(p6, p7)
>>> l1.intersection(s1)
[]
is parallel(l1, l2)
Are two linear entities parallel?
Parameters l1 : LinearEntity
l2 : LinearEntity
Returns True : if l1 and l2 are parallel,
False : otherwise.
Examples
>>> from sympy import Point3D, Line3D
>>> p1, p2 = Point3D(0, 0, 0), Point3D(3, 4, 5)
>>> p3, p4 = Point3D(2, 1, 1), Point3D(8, 9, 11)
>>> l1, l2 = Line3D(p1, p2), Line3D(p3, p4)
>>> Line3D.is_parallel(l1, l2)
True
>>> p5 = Point3D(6, 6, 6)
>>> l3 = Line3D(p3, p5)
>>> Line3D.is_parallel(l1, l3)
False
is perpendicular(l1, l2)
Are two linear entities perpendicular?
Parameters l1 : LinearEntity
l2 : LinearEntity
Returns True : if l1 and l2 are perpendicular,
False : otherwise.
See Also:
direction ratio (page 529)
Examples
>>> from sympy import Point3D, Line3D
>>> p1, p2, p3 = Point3D(0, 0, 0), Point3D(1, 1, 1), Point3D(-1, 2, 0)
>>> l1, l2 = Line3D(p1, p2), Line3D(p2, p3)
530
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> l1.is_perpendicular(l2)
False
>>> p4 = Point3D(5, 3, 7)
>>> l3 = Line3D(p1, p4)
>>> l1.is_perpendicular(l3)
False
is similar(other)
Return True if self and other are contained in the same line.
Examples
>>> from sympy import Point3D, Line3D
>>> p1, p2, p3 = Point3D(0, 0, 0), Point3D(1, 1, 1), Point3D(2, 2, 2)
>>> l1 = Line3D(p1, p2)
>>> l2 = Line3D(p1, p3)
>>> l1.is_similar(l2)
True
length
The length of the line.
Examples
>>>
>>>
>>>
>>>
oo
from sympy import Point3D, Line3D
p1, p2 = Point3D(0, 0, 0), Point3D(3, 5, 1)
l1 = Line3D(p1, p2)
l1.length
p1
The rst dening point of a linear entity.
See Also:
sympy.geometry.point3d.Point3D (page 503)
Examples
>>> from sympy import Point3D, Line3D
>>> p1, p2 = Point3D(0, 0, 0), Point3D(5, 3, 1)
>>> l = Line3D(p1, p2)
>>> l.p1
Point3D(0, 0, 0)
p2
The second dening point of a linear entity.
See Also:
sympy.geometry.point3d.Point3D (page 503)
5.10. Geometry Module
531
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import Point3D, Line3D
>>> p1, p2 = Point3D(0, 0, 0), Point3D(5, 3, 1)
>>> l = Line3D(p1, p2)
>>> l.p2
Point3D(5, 3, 1)
parallel line(p)
Create a new Line parallel to this linear entity which passes through the point p.
Parameters p : Point3D
Returns line : Line3D
See Also:
is parallel (page 530)
Examples
>>> from sympy import Point3D, Line3D
>>> p1, p2, p3 = Point3D(0, 0, 0), Point3D(2, 3, 4), Point3D(-2, 2, 0)
>>> l1 = Line3D(p1, p2)
>>> l2 = l1.parallel_line(p3)
>>> p3 in l2
True
>>> l1.is_parallel(l2)
True
perpendicular line(p)
Create a new Line perpendicular to this linear entity which passes through the point
p.
Parameters p : Point3D
Returns line : Line3D
See Also:
is perpendicular (page 530), perpendicular segment (page 532)
Examples
>>> from sympy import Point3D, Line3D
>>> p1, p2, p3 = Point3D(0, 0, 0), Point3D(2, 3, 4), Point3D(-2, 2, 0)
>>> l1 = Line3D(p1, p2)
>>> l2 = l1.perpendicular_line(p3)
>>> p3 in l2
True
>>> l1.is_perpendicular(l2)
True
perpendicular segment(p)
Create a perpendicular line segment from p to this line.
The enpoints of the segment are p and the closest point in the line containing self.
(If self is not a line, the point might not be in self.)
532
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Parameters p : Point3D
Returns segment : Segment3D
See Also:
perpendicular line (page 532)
Notes
Returns p itself if p is on this linear entity.
Examples
>>> from sympy import Point3D, Line3D
>>> p1, p2, p3 = Point3D(0, 0, 0), Point3D(1, 1, 1), Point3D(0, 2, 0)
>>> l1 = Line3D(p1, p2)
>>> s1 = l1.perpendicular_segment(p3)
>>> l1.is_perpendicular(s1)
True
>>> p3 in s1
True
>>> l1.perpendicular_segment(Point3D(4, 0, 0))
Segment3D(Point3D(4/3, 4/3, 4/3), Point3D(4, 0, 0))
points
The two points used to dene this linear entity.
Returns points : tuple of Points
See Also:
sympy.geometry.point3d.Point3D (page 503)
Examples
>>> from sympy import Point3D, Line3D
>>> p1, p2 = Point3D(0, 0, 0), Point3D(5, 11, 1)
>>> l1 = Line3D(p1, p2)
>>> l1.points
(Point3D(0, 0, 0), Point3D(5, 11, 1))
projection(o)
Project a point, line, ray, or segment onto this linear entity.
Parameters other : Point or LinearEntity (Line, Ray, Segment)
Returns projection : Point or LinearEntity (Line, Ray, Segment)
The return type matches the type of the parameter other.
Raises GeometryError :
When method is unable to perform projection.
See Also:
sympy.geometry.point3d.Point3D (page 503), perpendicular line (page 532)
5.10. Geometry Module
533
SymPy Documentation, Release 0.7.6
Notes
A projection involves taking the two points that dene the linear entity and projecting those points onto a Line and then reforming the linear entity using these
projections. A point P is projected onto a line L by nding the point on L that is
closest to P. This point is the intersection of L and the line perpendicular to L that
passes through P.
Examples
>>> from sympy import Point3D, Line3D, Segment3D, Rational
>>> p1, p2, p3 = Point3D(0, 0, 1), Point3D(1, 1, 2), Point3D(2, 0, 1)
>>> l1 = Line3D(p1, p2)
>>> l1.projection(p3)
Point3D(2/3, 2/3, 5/3)
>>> p4, p5 = Point3D(10, 0, 1), Point3D(12, 1, 3)
>>> s1 = Segment3D(p4, p5)
>>> l1.projection(s1)
[Segment3D(Point3D(10/3, 10/3, 13/3), Point3D(5, 5, 6))]
class sympy.geometry.line3d.Ray3D
A Ray is a semi-line in the space with a source point and a direction.
Parameters p1 : Point3D
The source of the Ray
p2 : Point or a direction vector
direction ratio: Determines the direction in which the Ray propagates. :
See Also:
sympy.geometry.point3d.Point3D (page 503), Line3D (page 525)
Examples
>>> import sympy
>>> from sympy import Point3D, pi
>>> from sympy.abc import r
>>> from sympy.geometry import Ray3D
>>> r = Ray3D(Point3D(2, 3, 4), Point3D(3, 5, 0))
>>> r
Ray3D(Point3D(2, 3, 4), Point3D(3, 5, 0))
>>> r.points
(Point3D(2, 3, 4), Point3D(3, 5, 0))
>>> r.source
Point3D(2, 3, 4)
>>> r.xdirection
oo
>>> r.ydirection
oo
>>> r.direction_ratio
[1, 2, -4]
534
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Attributes
source
xdirection
ydirection
zdirection
contains(o)
Is other GeometryEntity contained in this Ray?
distance(o)
Finds the shortest distance between the ray and a point.
Raises NotImplementedError is raised if o is not a Point :
Examples
>>> from sympy import Point3D, Ray3D
>>> p1, p2 = Point3D(0, 0, 0), Point3D(1, 1, 2)
>>> s = Ray3D(p1, p2)
>>> s.distance(Point3D(-1, -1, 2))
sqrt(6)
>>> s.distance((-1, -1, 2))
sqrt(6)
equals(other)
Returns True if self and other are the same mathematical entities
plot interval(parameter=t)
The plot interval for the default geometric plot of the Ray. Gives values that will
produce a ray that is 10 units long (where a unit is the distance between the two
points that dene the ray).
Parameters parameter : str, optional
Default value is t.
Returns plot interval : list
[parameter, lower bound, upper bound]
Examples
>>>
>>>
>>>
[t,
from sympy import Point3D, Ray3D, pi
r = Ray3D(Point3D(0, 0, 0), Point3D(1, 1, 1))
r.plot_interval()
0, 10]
source
The point from which the ray emanates.
See Also:
sympy.geometry.point3d.Point3D (page 503)
5.10. Geometry Module
535
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import Point3D, Ray3D
>>> p1, p2 = Point3D(0, 0, 0), Point3D(4, 1, 5)
>>> r1 = Ray3D(p1, p2)
>>> r1.source
Point3D(0, 0, 0)
xdirection
The x direction of the ray.
Positive innity if the ray points in the positive x direction, negative innity if the
ray points in the negative x direction, or 0 if the ray is vertical.
See Also:
ydirection (page 536)
Examples
>>>
>>>
>>>
>>>
oo
>>>
0
from sympy import Point3D, Ray3D
p1, p2, p3 = Point3D(0, 0, 0), Point3D(1, 1, 1), Point3D(0, -1, 0)
r1, r2 = Ray3D(p1, p2), Ray3D(p1, p3)
r1.xdirection
r2.xdirection
ydirection
The y direction of the ray.
Positive innity if the ray points in the positive y direction, negative innity if the
ray points in the negative y direction, or 0 if the ray is horizontal.
See Also:
xdirection (page 536)
Examples
>>>
>>>
>>>
>>>
-oo
>>>
0
from sympy import Point3D, Ray3D
p1, p2, p3 = Point3D(0, 0, 0), Point3D(-1, -1, -1), Point3D(-1, 0, 0)
r1, r2 = Ray3D(p1, p2), Ray3D(p1, p3)
r1.ydirection
r2.ydirection
zdirection
The z direction of the ray.
Positive innity if the ray points in the positive z direction, negative innity if the
ray points in the negative z direction, or 0 if the ray is horizontal.
See Also:
xdirection (page 536)
536
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>>
>>>
>>>
>>>
-oo
>>>
0
>>>
0
from sympy import Point3D, Ray3D
p1, p2, p3 = Point3D(0, 0, 0), Point3D(-1, -1, -1), Point3D(-1, 0, 0)
r1, r2 = Ray3D(p1, p2), Ray3D(p1, p3)
r1.ydirection
r2.ydirection
r2.zdirection
class sympy.geometry.line3d.Segment3D
A undirected line segment in a 3D space.
Parameters p1 : Point3D
p2 : Point3D
See Also:
sympy.geometry.point.Point3D, Line3D (page 525)
Examples
>>> import sympy
>>> from sympy import Point3D
>>> from sympy.abc import s
>>> from sympy.geometry import Segment3D
>>> Segment3D((1, 0, 0), (1, 1, 1)) # tuples are interpreted as pts
Segment3D(Point3D(1, 0, 0), Point3D(1, 1, 1))
>>> s = Segment3D(Point3D(4, 3, 9), Point3D(1, 1, 7))
>>> s
Segment3D(Point3D(1, 1, 7), Point3D(4, 3, 9))
>>> s.points
(Point3D(1, 1, 7), Point3D(4, 3, 9))
>>> s.length
sqrt(17)
>>> s.midpoint
Point3D(5/2, 2, 8)
Attributes
length
midpoint
number or sympy expression
Point3D
contains(other)
Is the other GeometryEntity contained within this Segment?
Examples
>>> from sympy import Point3D, Segment3D
>>> p1, p2 = Point3D(0, 1, 1), Point3D(3, 4, 5)
>>> s = Segment3D(p1, p2)
5.10. Geometry Module
537
SymPy Documentation, Release 0.7.6
>>> s2 = Segment3D(p2, p1)
>>> s.contains(s2)
True
distance(o)
Finds the shortest distance between a line segment and a point.
Raises NotImplementedError is raised if o is not a Point3D :
Examples
>>> from sympy import Point3D, Segment3D
>>> p1, p2 = Point3D(0, 0, 3), Point3D(1, 1, 4)
>>> s = Segment3D(p1, p2)
>>> s.distance(Point3D(10, 15, 12))
sqrt(341)
>>> s.distance((10, 15, 12))
sqrt(341)
length
The length of the line segment.
See Also:
sympy.geometry.point3d.Point3D.distance (page 505)
Examples
>>> from sympy import Point3D, Segment3D
>>> p1, p2 = Point3D(0, 0, 0), Point3D(4, 3, 3)
>>> s1 = Segment3D(p1, p2)
>>> s1.length
sqrt(34)
midpoint
The midpoint of the line segment.
See Also:
sympy.geometry.point3d.Point3D.midpoint (page 507)
Examples
>>> from sympy import Point3D, Segment3D
>>> p1, p2 = Point3D(0, 0, 0), Point3D(4, 3, 3)
>>> s1 = Segment3D(p1, p2)
>>> s1.midpoint
Point3D(2, 3/2, 3/2)
plot interval(parameter=t)
The plot interval for the default geometric plot of the Segment gives values that will
produce the full segment in a plot.
Parameters parameter : str, optional
Default value is t.
538
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Returns plot interval : list
[parameter, lower bound, upper bound]
Examples
>>>
>>>
>>>
>>>
[t,
from sympy import Point3D, Segment3D
p1, p2 = Point3D(0, 0, 0), Point3D(5, 3, 0)
s1 = Segment3D(p1, p2)
s1.plot_interval()
0, 1]
Curves
class sympy.geometry.curve.Curve
A curve in space.
A curve is dened by parametric functions for the coordinates, a parameter and the lower
and upper bounds for the parameter value.
Parameters function : list of functions
limits : 3-tuple
Function parameter and lower and upper bounds.
Raises ValueError :
When f unctions are specied incorrectly. When limits are specied
incorrectly.
See Also:
sympy.core.function.Function (page 169), sympy.polys.polyfuncs.interpolate
(page 1114)
Examples
>>> from sympy import sin, cos, Symbol, interpolate
>>> from sympy.abc import t, a
>>> from sympy.geometry import Curve
>>> C = Curve((sin(t), cos(t)), (t, 0, 2))
>>> C.functions
(sin(t), cos(t))
>>> C.limits
(t, 0, 2)
>>> C.parameter
t
>>> C = Curve((t, interpolate([1, 4, 9, 16], t)), (t, 0, 1)); C
Curve((t, t**2), (t, 0, 1))
>>> C.subs(t, 4)
Point(4, 16)
>>> C.arbitrary_point(a)
Point(a, a**2)
5.10. Geometry Module
539
SymPy Documentation, Release 0.7.6
Attributes
functions
parameter
limits
arbitrary point(parameter=t)
A parameterized point on the curve.
Parameters parameter : str or Symbol, optional
Default value is t; the Curves parameter is selected with None or
self.parameter otherwise the provided symbol is used.
Returns arbitrary point : Point
Raises ValueError :
When parameter already appears in the functions.
See Also:
sympy.geometry.point.Point (page 497)
Examples
>>> from sympy import Symbol
>>> from sympy.abc import s
>>> from sympy.geometry import Curve
>>> C = Curve([2*s, s**2], (s, 0, 2))
>>> C.arbitrary_point()
Point(2*t, t**2)
>>> C.arbitrary_point(C.parameter)
Point(2*s, s**2)
>>> C.arbitrary_point(None)
Point(2*s, s**2)
>>> C.arbitrary_point(Symbol(a))
Point(2*a, a**2)
free symbols
Return a set of symbols other than the bound symbols used to parametrically dene
the Curve.
Examples
>>> from sympy.abc import t, a
>>> from sympy.geometry import Curve
>>> Curve((t, t**2), (t, 0, 2)).free_symbols
set()
>>> Curve((t, t**2), (t, a, 2)).free_symbols
set([a])
functions
The functions specifying the curve.
Returns functions : list of parameterized coordinate functions.
540
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
See Also:
parameter (page 541)
Examples
>>>
>>>
>>>
>>>
(t,
from sympy.abc import t
from sympy.geometry import Curve
C = Curve((t, t**2), (t, 0, 2))
C.functions
t**2)
limits
The limits for the curve.
Returns limits : tuple
Contains parameter and lower and upper limits.
See Also:
plot interval (page 541)
Examples
>>>
>>>
>>>
>>>
(t,
from sympy.abc import t
from sympy.geometry import Curve
C = Curve([t, t**3], (t, -2, 2))
C.limits
-2, 2)
parameter
The curve function variable.
Returns parameter : SymPy symbol
See Also:
functions (page 540)
Examples
>>>
>>>
>>>
>>>
t
from sympy.abc import t
from sympy.geometry import Curve
C = Curve([t, t**2], (t, 0, 2))
C.parameter
plot interval(parameter=t)
The plot interval for the default geometric plot of the curve.
Parameters parameter : str or Symbol, optional
Default value is t; otherwise the provided symbol is used.
Returns plot interval : list (plot interval)
[parameter, lower bound, upper bound]
See Also:
5.10. Geometry Module
541
SymPy Documentation, Release 0.7.6
limits (page 541) Returns limits of the parameter interval
Examples
>>>
>>>
>>>
[t,
>>>
[s,
from sympy import Curve, sin
from sympy.abc import x, t, s
Curve((x, sin(x)), (x, 1, 2)).plot_interval()
1, 2]
Curve((x, sin(x)), (x, 1, 2)).plot_interval(s)
1, 2]
rotate(angle=0, pt=None)
Rotate angle radians counterclockwise about Point pt.
The default pt is the origin, Point(0, 0).
Examples
>>> from sympy.geometry.curve import Curve
>>> from sympy.abc import x
>>> from sympy import pi
>>> Curve((x, x), (x, 0, 1)).rotate(pi/2)
Curve((-x, x), (x, 0, 1))
scale(x=1, y=1, pt=None)
Override GeometryEntity.scale since Curve is not made up of Points.
Examples
>>> from sympy.geometry.curve import Curve
>>> from sympy import pi
>>> from sympy.abc import x
>>> Curve((x, x), (x, 0, 1)).scale(2)
Curve((2*x, x), (x, 0, 1))
translate(x=0, y=0)
Translate the Curve by (x, y).
Examples
>>> from sympy.geometry.curve import Curve
>>> from sympy import pi
>>> from sympy.abc import x
>>> Curve((x, x), (x, 0, 1)).translate(1, 2)
Curve((x + 1, x + 2), (x, 0, 1))
Ellipses
class sympy.geometry.ellipse.Ellipse
An elliptical GeometryEntity.
542
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Parameters center : Point, optional
Default value is Point(0, 0)
hradius : number or SymPy expression, optional
vradius : number or SymPy expression, optional
eccentricity : number or SymPy expression, optional
Two of hradius, vradius and eccentricity must be supplied to create an
Ellipse. The third is derived from the two supplied.
Raises GeometryError :
When hradius, vradius and eccentricity are incorrectly supplied as parameters.
TypeError :
When center is not a Point.
See Also:
Circle (page 554)
Notes
Constructed from a center and two radii, the rst being the horizontal radius (along the
x-axis) and the second being the vertical radius (along the y-axis).
When symbolic value for hradius and vradius are used, any calculation that refers to the
foci or the major or minor axis will assume that the ellipse has its major radius on the
x-axis. If this is not true then a manual rotation is necessary.
Examples
>>> from sympy import Ellipse, Point, Rational
>>> e1 = Ellipse(Point(0, 0), 5, 1)
>>> e1.hradius, e1.vradius
(5, 1)
>>> e2 = Ellipse(Point(3, 1), hradius=3, eccentricity=Rational(4, 5))
>>> e2
Ellipse(Point(3, 1), 3, 9/5)
Plotting:
>>> from sympy.plotting.pygletplot import PygletPlot as Plot
>>> from sympy import Circle, Segment
>>> c1 = Circle(Point(0,0), 1)
>>> Plot(c1)
[0]: cos(t), sin(t), mode=parametric
>>> p = Plot()
>>> p[0] = c1
>>> radius = Segment(c1.center, c1.random_point())
>>> p[1] = radius
>>> p
[0]: cos(t), sin(t), mode=parametric
[1]: t*cos(1.546086215036205357975518382),
t*sin(1.546086215036205357975518382), mode=parametric
5.10. Geometry Module
543
SymPy Documentation, Release 0.7.6
Attributes
center
hradius
vradius
area
circumference
eccentricity
periapsis
apoapsis
focus distance
foci
apoapsis
The apoapsis of the ellipse.
The greatest distance between the focus and the contour.
Returns apoapsis : number
See Also:
periapsis (page 550) Returns shortest distance between foci and contour
Examples
>>> from sympy import Point, Ellipse
>>> p1 = Point(0, 0)
>>> e1 = Ellipse(p1, 3, 1)
>>> e1.apoapsis
2*sqrt(2) + 3
arbitrary point(parameter=t)
A parameterized point on the ellipse.
Parameters parameter : str, optional
Default value is t.
Returns arbitrary point : Point
Raises ValueError :
When parameter already appears in the functions.
See Also:
sympy.geometry.point.Point (page 497)
Examples
>>> from sympy import Point, Ellipse
>>> e1 = Ellipse(Point(0, 0), 3, 2)
>>> e1.arbitrary_point()
Point(3*cos(t), 2*sin(t))
area
The area of the ellipse.
544
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Returns area : number
Examples
>>> from sympy import Point, Ellipse
>>> p1 = Point(0, 0)
>>> e1 = Ellipse(p1, 3, 1)
>>> e1.area
3*pi
center
The center of the ellipse.
Returns center : number
See Also:
sympy.geometry.point.Point (page 497)
Examples
>>> from sympy import Point, Ellipse
>>> p1 = Point(0, 0)
>>> e1 = Ellipse(p1, 3, 1)
>>> e1.center
Point(0, 0)
circumference
The circumference of the ellipse.
Examples
>>> from sympy import Point, Ellipse
>>> p1 = Point(0, 0)
>>> e1 = Ellipse(p1, 3, 1)
>>> e1.circumference
12*Integral(sqrt((-8*_x**2/9 + 1)/(-_x**2 + 1)), (_x, 0, 1))
eccentricity
The eccentricity of the ellipse.
Returns eccentricity : number
Examples
>>> from sympy import Point, Ellipse, sqrt
>>> p1 = Point(0, 0)
>>> e1 = Ellipse(p1, 3, sqrt(2))
>>> e1.eccentricity
sqrt(7)/3
encloses point(p)
Return True if p is enclosed by (is inside of) self.
5.10. Geometry Module
545
SymPy Documentation, Release 0.7.6
Parameters p : Point
Returns encloses point : True, False or None
See Also:
sympy.geometry.point.Point (page 497)
Notes
Being on the border of self is considered False.
Examples
>>> from sympy import Ellipse, S
>>> from sympy.abc import t
>>> e = Ellipse((0, 0), 3, 2)
>>> e.encloses_point((0, 0))
True
>>> e.encloses_point(e.arbitrary_point(t).subs(t, S.Half))
False
>>> e.encloses_point((4, 0))
False
equation(x=x, y=y)
The equation of the ellipse.
Parameters x : str, optional
Label for the x-axis. Default value is x.
y : str, optional
Label for the y-axis. Default value is y.
Returns equation : sympy expression
See Also:
arbitrary point (page 544) Returns parameterized point on ellipse
Examples
>>> from sympy import Point, Ellipse
>>> e1 = Ellipse(Point(1, 0), 3, 2)
>>> e1.equation()
y**2/4 + (x/3 - 1/3)**2 - 1
evolute(x=x, y=y)
The equation of evolute of the ellipse.
Parameters x : str, optional
Label for the x-axis. Default value is x.
y : str, optional
Label for the y-axis. Default value is y.
Returns equation : sympy expression
546
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import Point, Ellipse
>>> e1 = Ellipse(Point(1, 0), 3, 2)
>>> e1.evolute()
2**(2/3)*y**(2/3) + (3*x - 3)**(2/3) - 5**(2/3)
foci
The foci of the ellipse.
Raises ValueError :
When the major and minor axis cannot be determined.
See Also:
sympy.geometry.point.Point (page 497)
focus distance (page 547) Returns the distance between focus and center
Notes
The foci can only be calculated if the major/minor axes are known.
Examples
>>> from sympy import Point, Ellipse
>>> p1 = Point(0, 0)
>>> e1 = Ellipse(p1, 3, 1)
>>> e1.foci
(Point(-2*sqrt(2), 0), Point(2*sqrt(2), 0))
focus distance
The focale distance of the ellipse.
The distance between the center and one focus.
Returns focus distance : number
See Also:
foci (page 547)
Examples
>>> from sympy import Point, Ellipse
>>> p1 = Point(0, 0)
>>> e1 = Ellipse(p1, 3, 1)
>>> e1.focus_distance
2*sqrt(2)
hradius
The horizontal radius of the ellipse.
Returns hradius : number
See Also:
vradius (page 553), major (page 549), minor (page 549)
5.10. Geometry Module
547
SymPy Documentation, Release 0.7.6
Examples
>>>
>>>
>>>
>>>
3
from sympy import Point, Ellipse
p1 = Point(0, 0)
e1 = Ellipse(p1, 3, 1)
e1.hradius
intersection(o)
The intersection of this ellipse and another geometrical entity o.
Parameters o : GeometryEntity
Returns intersection : list of GeometryEntity objects
See Also:
sympy.geometry.entity.GeometryEntity (page 493)
Notes
Currently supports intersections with Point, Line, Segment, Ray, Circle and Ellipse
types.
Examples
>>> from sympy import Ellipse, Point, Line, sqrt
>>> e = Ellipse(Point(0, 0), 5, 7)
>>> e.intersection(Point(0, 0))
[]
>>> e.intersection(Point(5, 0))
[Point(5, 0)]
>>> e.intersection(Line(Point(0,0), Point(0, 1)))
[Point(0, -7), Point(0, 7)]
>>> e.intersection(Line(Point(5,0), Point(5, 1)))
[Point(5, 0)]
>>> e.intersection(Line(Point(6,0), Point(6, 1)))
[]
>>> e = Ellipse(Point(-1, 0), 4, 3)
>>> e.intersection(Ellipse(Point(1, 0), 4, 3))
[Point(0, -3*sqrt(15)/4), Point(0, 3*sqrt(15)/4)]
>>> e.intersection(Ellipse(Point(5, 0), 4, 3))
[Point(2, -3*sqrt(7)/4), Point(2, 3*sqrt(7)/4)]
>>> e.intersection(Ellipse(Point(100500, 0), 4, 3))
[]
>>> e.intersection(Ellipse(Point(0, 0), 3, 4))
[Point(-363/175, -48*sqrt(111)/175), Point(-363/175, 48*sqrt(111)/175), Point(3, 0)]
>>> e.intersection(Ellipse(Point(-1, 0), 3, 4))
[Point(-17/5, -12/5), Point(-17/5, 12/5), Point(7/5, -12/5), Point(7/5, 12/5)]
is tangent(o)
Is o tangent to the ellipse?
Parameters o : GeometryEntity
An Ellipse, LinearEntity or Polygon
548
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Returns is tangent: boolean :
True if o is tangent to the ellipse, False otherwise.
Raises NotImplementedError :
When the wrong type of argument is supplied.
See Also:
tangent lines (page 553)
Examples
>>> from sympy import Point, Ellipse, Line
>>> p0, p1, p2 = Point(0, 0), Point(3, 0), Point(3, 3)
>>> e1 = Ellipse(p0, 3, 2)
>>> l1 = Line(p1, p2)
>>> e1.is_tangent(l1)
True
major
Longer axis of the ellipse (if it can be determined) else hradius.
Returns major : number or expression
See Also:
hradius (page 547), vradius (page 553), minor (page 549)
Examples
>>>
>>>
>>>
>>>
3
from sympy import Point, Ellipse, Symbol
p1 = Point(0, 0)
e1 = Ellipse(p1, 3, 1)
e1.major
>>>
>>>
>>>
a
>>>
b
a = Symbol(a)
b = Symbol(b)
Ellipse(p1, a, b).major
>>>
>>>
>>>
m +
m = Symbol(m)
M = m + 1
Ellipse(p1, m, M).major
1
Ellipse(p1, b, a).major
minor
Shorter axis of the ellipse (if it can be determined) else vradius.
Returns minor : number or expression
See Also:
hradius (page 547), vradius (page 553), major (page 549)
5.10. Geometry Module
549
SymPy Documentation, Release 0.7.6
Examples
>>>
>>>
>>>
>>>
1
from sympy import Point, Ellipse, Symbol
p1 = Point(0, 0)
e1 = Ellipse(p1, 3, 1)
e1.minor
>>>
>>>
>>>
b
>>>
a
a = Symbol(a)
b = Symbol(b)
Ellipse(p1, a, b).minor
Ellipse(p1, b, a).minor
>>> m = Symbol(m)
>>> M = m + 1
>>> Ellipse(p1, m, M).minor
m
normal lines(p, prec=None)
Normal lines between p and the ellipse.
Parameters p : Point
Returns normal lines : list with 1, 2 or 4 Lines
Examples
>>> from sympy import Line, Point, Ellipse
>>> e = Ellipse((0, 0), 2, 3)
>>> c = e.center
>>> e.normal_lines(c + Point(1, 0))
[Line(Point(0, 0), Point(1, 0))]
>>> e.normal_lines(c)
[Line(Point(0, 0), Point(0, 1)), Line(Point(0, 0), Point(1, 0))]
O-axis points require the solution of a quartic equation. This often leads to very
large expressions that may be of little practical use. An approximate solution of prec
digits can be obtained by passing in the desired value:
>>> e.normal_lines((3, 3), prec=2)
[Line(Point(-38/47, -85/31), Point(9/47, -21/17)),
Line(Point(19/13, -43/21), Point(32/13, -8/3))]
Whereas the above solution has an operation count of 12, the exact solution has an
operation count of 2020.
periapsis
The periapsis of the ellipse.
The shortest distance between the focus and the contour.
Returns periapsis : number
See Also:
apoapsis (page 544) Returns greatest distance between focus and contour
550
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import Point, Ellipse
>>> p1 = Point(0, 0)
>>> e1 = Ellipse(p1, 3, 1)
>>> e1.periapsis
-2*sqrt(2) + 3
plot interval(parameter=t)
The plot interval for the default geometric plot of the Ellipse.
Parameters parameter : str, optional
Default value is t.
Returns plot interval : list
[parameter, lower bound, upper bound]
Examples
>>>
>>>
>>>
[t,
from sympy import Point, Ellipse
e1 = Ellipse(Point(0, 0), 3, 2)
e1.plot_interval()
-pi, pi]
random point(seed=None)
A random point on the ellipse.
Returns point : Point
See Also:
sympy.geometry.point.Point (page 497)
arbitrary point (page 544) Returns parameterized point on ellipse
Notes
An arbitrary point with a random value of t substituted into it may not test as being
on the ellipse because the expression tested that a point is on the ellipse doesnt
simplify to zero and doesnt evaluate exactly to zero:
>>> from sympy.abc import t
>>> e1.arbitrary_point(t)
Point(3*cos(t), 2*sin(t))
>>> p2 = _.subs(t, 0.1)
>>> p2 in e1
False
Note that arbitrary point routine does not take this approach. A value for cos(t) and
sin(t) (not t) is substituted into the arbitrary point. There is a small chance that this
will give a point that will not test as being in the ellipse, so the process is repeated
(up to 10 times) until a valid point is obtained.
5.10. Geometry Module
551
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import Point, Ellipse, Segment
>>> e1 = Ellipse(Point(0, 0), 3, 2)
>>> e1.random_point() # gives some random point
Point(...)
>>> p1 = e1.random_point(seed=0); p1.n(2)
Point(2.1, 1.4)
The random point method assures that the point will test as being in the ellipse:
>>> p1 in e1
True
reflect(line)
Override GeometryEntity.reect since the radius is not a GeometryEntity.
Notes
Until the general ellipse (with no axis parallel to the x-axis) is supported a NotImplemented error is raised and the equation whose zeros dene the rotated ellipse is
given.
Examples
>>> from sympy import Circle, Line
>>> Circle((0, 1), 1).reflect(Line((0, 0), (1, 1)))
Circle(Point(1, 0), -1)
>>> from sympy import Ellipse, Line, Point
>>> Ellipse(Point(3, 4), 1, 3).reflect(Line(Point(0, -4), Point(5, 0)))
Traceback (most recent call last):
...
NotImplementedError:
General Ellipse is not supported but the equation of the reflected
Ellipse is given by the zeros of: f(x, y) = (9*x/41 + 40*y/41 +
37/41)**2 + (40*x/123 - 3*y/41 - 364/123)**2 - 1
Traceback (most recent call last):
...
NotImplementedError:
rotate(angle=0, pt=None)
Rotate angle radians counterclockwise about Point pt.
Note: since the general ellipse is not supported, only rotations that are integer
multiples of pi/2 are allowed.
Examples
>>> from sympy import Ellipse, pi
>>> Ellipse((1, 0), 2, 1).rotate(pi/2)
Ellipse(Point(0, 1), 1, 2)
>>> Ellipse((1, 0), 2, 1).rotate(pi)
Ellipse(Point(-1, 0), 2, 1)
552
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
scale(x=1, y=1, pt=None)
Override GeometryEntity.scale since it is the major and minor axes which must be
scaled and they are not GeometryEntities.
Examples
>>> from sympy import Ellipse
>>> Ellipse((0, 0), 2, 1).scale(2, 4)
Circle(Point(0, 0), 4)
>>> Ellipse((0, 0), 2, 1).scale(2)
Ellipse(Point(0, 0), 4, 1)
tangent lines(p)
Tangent lines between p and the ellipse.
If p is on the ellipse, returns the tangent line through point p. Otherwise, returns
the tangent line(s) from p to the ellipse, or None if no tangent line is possible (e.g.,
p inside ellipse).
Parameters p : Point
Returns tangent lines : list with 1 or 2 Lines
Raises NotImplementedError :
Can only nd tangent lines for a point, p, on the ellipse.
See Also:
sympy.geometry.point.Point (page 497), sympy.geometry.line.Line (page 517)
Examples
>>> from sympy import Point, Ellipse
>>> e1 = Ellipse(Point(0, 0), 3, 2)
>>> e1.tangent_lines(Point(3, 0))
[Line(Point(3, 0), Point(3, -12))]
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
# This will plot an ellipse together with a tangent line.
from sympy.plotting.pygletplot import PygletPlot as Plot
from sympy import Point, Ellipse
e = Ellipse(Point(0,0), 3, 2)
t = e.tangent_lines(e.random_point())
p = Plot()
p[0] = e
p[1] = t
vradius
The vertical radius of the ellipse.
Returns vradius : number
See Also:
hradius (page 547), major (page 549), minor (page 549)
5.10. Geometry Module
553
SymPy Documentation, Release 0.7.6
Examples
>>>
>>>
>>>
>>>
1
from sympy import Point, Ellipse
p1 = Point(0, 0)
e1 = Ellipse(p1, 3, 1)
e1.vradius
class sympy.geometry.ellipse.Circle
A circle in space.
Constructed simply from a center and a radius, or from three non-collinear points.
Parameters center : Point
radius : number or sympy expression
points : sequence of three Points
Raises GeometryError :
When trying to construct circle from three collinear points. When
trying to construct circle from incorrect parameters.
See Also:
Ellipse (page 542), sympy.geometry.point.Point (page 497)
Examples
>>>
>>>
>>>
>>>
(5,
from sympy.geometry import Point, Circle
# a circle constructed from a center and radius
c1 = Circle(Point(0, 0), 5)
c1.hradius, c1.vradius, c1.radius
5, 5)
>>> # a circle costructed from three points
>>> c2 = Circle(Point(0, 0), Point(1, 1), Point(1, 0))
>>> c2.hradius, c2.vradius, c2.radius, c2.center
(sqrt(2)/2, sqrt(2)/2, sqrt(2)/2, Point(1/2, 1/2))
Attributes
radius (synonymous with hradius, vradius, major and minor)
circumference
equation
circumference
The circumference of the circle.
Returns circumference : number or SymPy expression
Examples
554
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> from sympy import Point, Circle
>>> c1 = Circle(Point(3, 4), 6)
>>> c1.circumference
12*pi
equation(x=x, y=y)
The equation of the circle.
Parameters x : str or Symbol, optional
Default value is x.
y : str or Symbol, optional
Default value is y.
Returns equation : SymPy expression
Examples
>>> from sympy import Point, Circle
>>> c1 = Circle(Point(0, 0), 5)
>>> c1.equation()
x**2 + y**2 - 25
intersection(o)
The intersection of this circle with another geometrical entity.
Parameters o : GeometryEntity
Returns intersection : list of GeometryEntities
Examples
>>> from sympy import Point, Circle, Line, Ray
>>> p1, p2, p3 = Point(0, 0), Point(5, 5), Point(6, 0)
>>> p4 = Point(5, 0)
>>> c1 = Circle(p1, 5)
>>> c1.intersection(p2)
[]
>>> c1.intersection(p4)
[Point(5, 0)]
>>> c1.intersection(Ray(p1, p2))
[Point(5*sqrt(2)/2, 5*sqrt(2)/2)]
>>> c1.intersection(Line(p2, p3))
[]
radius
The radius of the circle.
Returns radius : number or sympy expression
See Also:
Ellipse.major (page 549), Ellipse.minor
(page 547), Ellipse.vradius (page 553)
5.10. Geometry Module
(page
549),
Ellipse.hradius
555
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import Point, Circle
>>> c1 = Circle(Point(3, 4), 6)
>>> c1.radius
6
reflect(line)
Override GeometryEntity.reect since the radius is not a GeometryEntity.
Examples
>>> from sympy import Circle, Line
>>> Circle((0, 1), 1).reflect(Line((0, 0), (1, 1)))
Circle(Point(1, 0), -1)
scale(x=1, y=1, pt=None)
Override GeometryEntity.scale since the radius is not a GeometryEntity.
Examples
>>> from sympy import Circle
>>> Circle((0, 0), 1).scale(2, 2)
Circle(Point(0, 0), 2)
>>> Circle((0, 0), 1).scale(2, 4)
Ellipse(Point(0, 0), 2, 4)
vradius
This Ellipse property is an alias for the Circles radius.
Whereas hradius, major and minor can use Ellipses conventions, the vradius does
not exist for a circle. It is always a positive value in order that the Circle, like
Polygons, will have an area that can be positive or negative as determined by the
sign of the hradius.
Examples
>>> from sympy import Point, Circle
>>> c1 = Circle(Point(3, 4), 6)
>>> c1.vradius
6
Polygons
class sympy.geometry.polygon.Polygon
A two-dimensional polygon.
A simple polygon in space. Can be constructed from a sequence of points or from a
center, radius, number of sides and rotation angle.
Parameters vertices : sequence of Points
Raises GeometryError :
556
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
If all parameters are not Points.
If the Polygon has intersecting sides.
See Also:
sympy.geometry.point.Point (page 497), sympy.geometry.line.Segment (page 522),
Triangle (page 570)
Notes
Polygons are treated as closed paths rather than 2D areas so some calculations can be
be negative or positive (e.g., area) based on the orientation of the points.
Any consecutive identical points are reduced to a single point and any points collinear
and between two points will be removed unless they are needed to dene an explicit
intersection (see examples).
A Triangle, Segment or Point will be returned when there are 3 or fewer points provided.
Examples
>>> from sympy import Point, Polygon, pi
>>> p1, p2, p3, p4, p5 = [(0, 0), (1, 0), (5, 1), (0, 1), (3, 0)]
>>> Polygon(p1, p2, p3, p4)
Polygon(Point(0, 0), Point(1, 0), Point(5, 1), Point(0, 1))
>>> Polygon(p1, p2)
Segment(Point(0, 0), Point(1, 0))
>>> Polygon(p1, p2, p5)
Segment(Point(0, 0), Point(3, 0))
While the sides of a polygon are not allowed to cross implicitly, they can do so explicitly.
For example, a polygon shaped like a Z with the top left connecting to the bottom right
of the Z must have the point in the middle of the Z explicitly given:
>>> mid = Point(1, 1)
>>> Polygon((0, 2), (2, 2), mid, (0, 0), (2, 0), mid).area
0
>>> Polygon((0, 2), (2, 2), mid, (2, 0), (0, 0), mid).area
-2
When the the keyword n is used to dene the number of sides of the Polygon then a
RegularPolygon is created and the other arguments are interpreted as center, radius and
rotation. The unrotated RegularPolygon will always have a vertex at Point(r, 0) where r
is the radius of the circle that circumscribes the RegularPolygon. Its method spin can be
used to increment that angle.
>>> p = Polygon((0,0), 1, n=3)
>>> p
RegularPolygon(Point(0, 0), 1, 3, 0)
>>> p.vertices[0]
Point(1, 0)
>>> p.args[0]
Point(0, 0)
>>> p.spin(pi/2)
>>> p.vertices[0]
Point(0, 1)
5.10. Geometry Module
557
SymPy Documentation, Release 0.7.6
Attributes
area
angles
perimeter
vertices
centroid
sides
angles
The internal angle at each vertex.
Returns angles : dict
A dictionary where each key is a vertex and each value is the internal
angle at that vertex. The vertices are represented as Points.
See Also:
sympy.geometry.point.Point (page 497), sympy.geometry.line.LinearEntity.angle between
(page 509)
Examples
>>> from sympy import Point, Polygon
>>> p1, p2, p3, p4 = map(Point, [(0, 0), (1, 0), (5, 1), (0, 1)])
>>> poly = Polygon(p1, p2, p3, p4)
>>> poly.angles[p1]
pi/2
>>> poly.angles[p2]
acos(-4*sqrt(17)/17)
arbitrary point(parameter=t)
A parameterized point on the polygon.
The parameter, varying from 0 to 1, assigns points to the position on the perimeter
that is that fraction of the total perimeter. So the point evaluated at t=1/2 would
return the point from the rst vertex that is 1/2 way around the polygon.
Parameters parameter : str, optional
Default value is t.
Returns arbitrary point : Point
Raises ValueError :
When parameter already appears in the Polygons denition.
See Also:
sympy.geometry.point.Point (page 497)
Examples
>>>
>>>
>>>
>>>
558
from sympy import Polygon, S, Symbol
t = Symbol(t, real=True)
tri = Polygon((0, 0), (1, 0), (1, 1))
p = tri.arbitrary_point(t)
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> perimeter = tri.perimeter
>>> s1, s2 = [s.length for s in tri.sides[:2]]
>>> p.subs(t, (s1 + s2/2)/perimeter)
Point(1, 1/2)
area
The area of the polygon.
See Also:
sympy.geometry.ellipse.Ellipse.area (page 544)
Notes
The area calculation can be positive or negative based on the orientation of the
points.
Examples
>>>
>>>
>>>
>>>
3
from sympy import Point, Polygon
p1, p2, p3, p4 = map(Point, [(0, 0), (1, 0), (5, 1), (0, 1)])
poly = Polygon(p1, p2, p3, p4)
poly.area
centroid
The centroid of the polygon.
Returns centroid : Point
See Also:
sympy.geometry.point.Point
(page 497)
(page
497),
sympy.geometry.util.centroid
Examples
>>> from sympy import Point, Polygon
>>> p1, p2, p3, p4 = map(Point, [(0, 0), (1, 0), (5, 1), (0, 1)])
>>> poly = Polygon(p1, p2, p3, p4)
>>> poly.centroid
Point(31/18, 11/18)
distance(o)
Returns the shortest distance between self and o.
If o is a point, then self does not need to be convex. If o is another polygon self and
o must be complex.
Examples
5.10. Geometry Module
559
SymPy Documentation, Release 0.7.6
>>> from sympy import Point, Polygon, RegularPolygon
>>> p1, p2 = map(Point, [(0, 0), (7, 5)])
>>> poly = Polygon(*RegularPolygon(p1, 1, 3).vertices)
>>> poly.distance(p2)
sqrt(61)
encloses point(p)
Return True if p is enclosed by (is inside of) self.
Parameters p : Point
Returns encloses point : True, False or None
See Also:
sympy.geometry.point.Point (page 497), sympy.geometry.ellipse.Ellipse.encloses point
(page 545)
Notes
Being on the border of self is considered False.
References
[1] http://www.ariel.com.au/a/python-point-int-poly.html
Examples
>>> from sympy import Polygon, Point
>>> from sympy.abc import t
>>> p = Polygon((0, 0), (4, 0), (4, 4))
>>> p.encloses_point(Point(2, 1))
True
>>> p.encloses_point(Point(2, 2))
False
>>> p.encloses_point(Point(5, 5))
False
intersection(o)
The intersection of two polygons.
The intersection may be empty and can contain individual Points and complete Line
Segments.
Parameters other: Polygon :
Returns intersection : list
The list of Segments and Points
See Also:
sympy.geometry.point.Point
(page 522)
560
(page
497),
sympy.geometry.line.Segment
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import Point, Polygon
>>> p1, p2, p3, p4 = map(Point, [(0, 0), (1, 0), (5, 1), (0, 1)])
>>> poly1 = Polygon(p1, p2, p3, p4)
>>> p5, p6, p7 = map(Point, [(3, 2), (1, -1), (0, 2)])
>>> poly2 = Polygon(p5, p6, p7)
>>> poly1.intersection(poly2)
[Point(2/3, 0), Point(9/5, 1/5), Point(7/3, 1), Point(1/3, 1)]
is convex()
Is the polygon convex?
A polygon is convex if all its interior angles are less than 180 degrees.
Returns is convex : boolean
True if this polygon is convex, False otherwise.
See Also:
sympy.geometry.util.convex hull (page 495)
Examples
>>> from sympy import Point, Polygon
>>> p1, p2, p3, p4 = map(Point, [(0, 0), (1, 0), (5, 1), (0, 1)])
>>> poly = Polygon(p1, p2, p3, p4)
>>> poly.is_convex()
True
perimeter
The perimeter of the polygon.
Returns perimeter : number or Basic instance
See Also:
sympy.geometry.line.Segment.length (page 523)
Examples
>>> from sympy import Point, Polygon
>>> p1, p2, p3, p4 = map(Point, [(0, 0), (1, 0), (5, 1), (0, 1)])
>>> poly = Polygon(p1, p2, p3, p4)
>>> poly.perimeter
sqrt(17) + 7
plot interval(parameter=t)
The plot interval for the default geometric plot of the polygon.
Parameters parameter : str, optional
Default value is t.
Returns plot interval : list (plot interval)
[parameter, lower bound, upper bound]
5.10. Geometry Module
561
SymPy Documentation, Release 0.7.6
Examples
>>>
>>>
>>>
[t,
from sympy import Polygon
p = Polygon((0, 0), (1, 0), (1, 1))
p.plot_interval()
0, 1]
sides
The line segments that form the sides of the polygon.
Returns sides : list of sides
Each side is a Segment.
See Also:
sympy.geometry.point.Point
(page 522)
(page
497),
sympy.geometry.line.Segment
Notes
The Segments that represent the sides are an undirected line segment so cannot be
used to tell the orientation of the polygon.
Examples
>>> from sympy import Point, Polygon
>>> p1, p2, p3, p4 = map(Point, [(0, 0), (1, 0), (5, 1), (0, 1)])
>>> poly = Polygon(p1, p2, p3, p4)
>>> poly.sides
[Segment(Point(0, 0), Point(1, 0)),
Segment(Point(1, 0), Point(5, 1)),
Segment(Point(0, 1), Point(5, 1)), Segment(Point(0, 0), Point(0, 1))]
vertices
The vertices of the polygon.
Returns vertices : tuple of Points
See Also:
sympy.geometry.point.Point (page 497)
Notes
When iterating over the vertices, it is more ecient to index self rather than to
request the vertices and index them. Only use the vertices when you want to process
all of them at once. This is even more important with RegularPolygons that calculate
each vertex.
Examples
562
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> from sympy import Point, Polygon
>>> p1, p2, p3, p4 = map(Point, [(0, 0), (1, 0), (5, 1), (0, 1)])
>>> poly = Polygon(p1, p2, p3, p4)
>>> poly.vertices
(Point(0, 0), Point(1, 0), Point(5, 1), Point(0, 1))
>>> poly.args[0]
Point(0, 0)
class sympy.geometry.polygon.RegularPolygon
A regular polygon.
Such a polygon has all internal angles equal and all sides the same length.
Parameters center : Point
radius : number or Basic instance
The distance from the center to a vertex
n : int
The number of sides
Raises GeometryError :
If the center is not a Point, or the radius is not a number or Basic
instance, or the number of sides, n, is less than three.
See Also:
sympy.geometry.point.Point (page 497), Polygon (page 556)
Notes
A RegularPolygon can be instantiated with Polygon with the kwarg n.
Regular polygons are instantiated with a center, radius, number of sides and a rotation
angle. Whereas the arguments of a Polygon are vertices, the vertices of the RegularPolygon must be obtained with the vertices method.
Examples
>>> from sympy.geometry import RegularPolygon, Point
>>> r = RegularPolygon(Point(0, 0), 5, 3)
>>> r
RegularPolygon(Point(0, 0), 5, 3, 0)
>>> r.vertices[0]
Point(5, 0)
5.10. Geometry Module
563
SymPy Documentation, Release 0.7.6
Attributes
vertices
center
radius
rotation
apothem
interior angle
exterior angle
circumcircle
incircle
angles
angles
Returns a dictionary with keys, the vertices of the Polygon, and values, the interior
angle at each vertex.
Examples
>>> from sympy import RegularPolygon, Point
>>> r = RegularPolygon(Point(0, 0), 5, 3)
>>> r.angles
{Point(-5/2, -5*sqrt(3)/2): pi/3,
Point(-5/2, 5*sqrt(3)/2): pi/3,
Point(5, 0): pi/3}
apothem
The inradius of the RegularPolygon.
The apothem/inradius is the radius of the inscribed circle.
Returns apothem : number or instance of Basic
See Also:
sympy.geometry.line.Segment.length (page 523), sympy.geometry.ellipse.Circle.radius
(page 555)
Examples
>>> from sympy import Symbol
>>> from sympy.geometry import RegularPolygon, Point
>>> radius = Symbol(r)
>>> rp = RegularPolygon(Point(0, 0), radius, 4)
>>> rp.apothem
sqrt(2)*r/2
area
Returns the area.
Examples
564
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> from sympy.geometry import RegularPolygon
>>> square = RegularPolygon((0, 0), 1, 4)
>>> square.area
2
>>> _ == square.length**2
True
args
Returns the center point, the radius, the number of sides, and the orientation angle.
Examples
>>> from sympy import RegularPolygon, Point
>>> r = RegularPolygon(Point(0, 0), 5, 3)
>>> r.args
(Point(0, 0), 5, 3, 0)
center
The center of the RegularPolygon
This is also the center of the circumscribing circle.
Returns center : Point
See Also:
sympy.geometry.point.Point (page 497), sympy.geometry.ellipse.Ellipse.center
(page 545)
Examples
>>> from sympy.geometry import RegularPolygon, Point
>>> rp = RegularPolygon(Point(0, 0), 5, 4)
>>> rp.center
Point(0, 0)
centroid
The center of the RegularPolygon
This is also the center of the circumscribing circle.
Returns center : Point
See Also:
sympy.geometry.point.Point (page 497), sympy.geometry.ellipse.Ellipse.center
(page 545)
Examples
>>> from sympy.geometry import RegularPolygon, Point
>>> rp = RegularPolygon(Point(0, 0), 5, 4)
>>> rp.center
Point(0, 0)
circumcenter
Alias for center.
5.10. Geometry Module
565
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.geometry import RegularPolygon, Point
>>> rp = RegularPolygon(Point(0, 0), 5, 4)
>>> rp.circumcenter
Point(0, 0)
circumcircle
The circumcircle of the RegularPolygon.
Returns circumcircle : Circle
See Also:
circumcenter (page 565), sympy.geometry.ellipse.Circle (page 554)
Examples
>>> from sympy.geometry import RegularPolygon, Point
>>> rp = RegularPolygon(Point(0, 0), 4, 8)
>>> rp.circumcircle
Circle(Point(0, 0), 4)
circumradius
Alias for radius.
Examples
>>>
>>>
>>>
>>>
>>>
r
from sympy import Symbol
from sympy.geometry import RegularPolygon, Point
radius = Symbol(r)
rp = RegularPolygon(Point(0, 0), radius, 4)
rp.circumradius
encloses point(p)
Return True if p is enclosed by (is inside of) self.
Parameters p : Point
Returns encloses point : True, False or None
See Also:
sympy.geometry.ellipse.Ellipse.encloses point (page 545)
Notes
Being on the border of self is considered False.
The general Polygon.encloses point method is called only if a point is not within or
beyond the incircle or circumcircle, respectively.
566
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import RegularPolygon, S, Point, Symbol
>>> p = RegularPolygon((0, 0), 3, 4)
>>> p.encloses_point(Point(0, 0))
True
>>> r, R = p.inradius, p.circumradius
>>> p.encloses_point(Point((r + R)/2, 0))
True
>>> p.encloses_point(Point(R/2, R/2 + (R - r)/10))
False
>>> t = Symbol(t, real=True)
>>> p.encloses_point(p.arbitrary_point().subs(t, S.Half))
False
>>> p.encloses_point(Point(5, 5))
False
exterior angle
Measure of the exterior angles.
Returns exterior angle : number
See Also:
sympy.geometry.line.LinearEntity.angle between (page 509)
Examples
>>> from sympy.geometry import RegularPolygon, Point
>>> rp = RegularPolygon(Point(0, 0), 4, 8)
>>> rp.exterior_angle
pi/4
incircle
The incircle of the RegularPolygon.
Returns incircle : Circle
See Also:
inradius (page 567), sympy.geometry.ellipse.Circle (page 554)
Examples
>>> from sympy.geometry import RegularPolygon, Point
>>> rp = RegularPolygon(Point(0, 0), 4, 7)
>>> rp.incircle
Circle(Point(0, 0), 4*cos(pi/7))
inradius
Alias for apothem.
Examples
5.10. Geometry Module
567
SymPy Documentation, Release 0.7.6
>>> from sympy import Symbol
>>> from sympy.geometry import RegularPolygon, Point
>>> radius = Symbol(r)
>>> rp = RegularPolygon(Point(0, 0), radius, 4)
>>> rp.inradius
sqrt(2)*r/2
interior angle
Measure of the interior angles.
Returns interior angle : number
See Also:
sympy.geometry.line.LinearEntity.angle between (page 509)
Examples
>>> from sympy.geometry import RegularPolygon, Point
>>> rp = RegularPolygon(Point(0, 0), 4, 8)
>>> rp.interior_angle
3*pi/4
length
Returns the length of the sides.
The half-length of the side and the apothem form two legs of a right triangle whose
hypotenuse is the radius of the regular polygon.
Examples
>>> from sympy.geometry import RegularPolygon
>>> from sympy import sqrt
>>> s = square_in_unit_circle = RegularPolygon((0, 0), 1, 4)
>>> s.length
sqrt(2)
>>> sqrt((_/2)**2 + s.apothem**2) == s.radius
True
radius
Radius of the RegularPolygon
This is also the radius of the circumscribing circle.
Returns radius : number or instance of Basic
See Also:
sympy.geometry.line.Segment.length (page 523), sympy.geometry.ellipse.Circle.radius
(page 555)
Examples
>>> from sympy import Symbol
>>> from sympy.geometry import RegularPolygon, Point
>>> radius = Symbol(r)
568
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> rp = RegularPolygon(Point(0, 0), radius, 4)
>>> rp.radius
r
reflect(line)
Override GeometryEntity.reect since this is not made of only points.
>>> from sympy import RegularPolygon, Line
>>> RegularPolygon((0, 0), 1, 4).reflect(Line((0, 1), slope=-2))
RegularPolygon(Point(4/5, 2/5), -1, 4, acos(3/5))
rotate(angle, pt=None)
Override GeometryEntity.rotate to rst rotate the RegularPolygon about its center.
>>> from sympy import Point, RegularPolygon, Polygon, pi
>>> t = RegularPolygon(Point(1, 0), 1, 3)
>>> t.vertices[0] # vertex on x-axis
Point(2, 0)
>>> t.rotate(pi/2).vertices[0] # vertex on y axis now
Point(0, 2)
See Also:
rotation (page 569)
spin (page 569) Rotates a RegularPolygon in place
rotation
CCW angle by which the RegularPolygon is rotated
Returns rotation : number or instance of Basic
Examples
>>> from sympy import pi
>>> from sympy.geometry import RegularPolygon, Point
>>> RegularPolygon(Point(0, 0), 3, 4, pi).rotation
pi
scale(x=1, y=1, pt=None)
Override GeometryEntity.scale since it is the radius that must be scaled (if x == y)
or else a new Polygon must be returned.
>>> from sympy import RegularPolygon
Symmetric scaling returns a RegularPolygon:
>>> RegularPolygon((0, 0), 1, 4).scale(2, 2)
RegularPolygon(Point(0, 0), 2, 4, 0)
Asymmetric scaling returns a kite as a Polygon:
>>> RegularPolygon((0, 0), 1, 4).scale(2, 1)
Polygon(Point(2, 0), Point(0, 1), Point(-2, 0), Point(0, -1))
spin(angle)
Increment in place the virtual Polygons rotation by ccw angle.
See also: rotate method which moves the center.
5.10. Geometry Module
569
SymPy Documentation, Release 0.7.6
>>> from sympy import Polygon, Point, pi
>>> r = Polygon(Point(0,0), 1, n=3)
>>> r.vertices[0]
Point(1, 0)
>>> r.spin(pi/6)
>>> r.vertices[0]
Point(sqrt(3)/2, 1/2)
See Also:
rotation (page 569)
rotate (page 569) Creates a copy of the RegularPolygon rotated about a Point
vertices
The vertices of the RegularPolygon.
Returns vertices : list
Each vertex is a Point.
See Also:
sympy.geometry.point.Point (page 497)
Examples
>>> from sympy.geometry import RegularPolygon, Point
>>> rp = RegularPolygon(Point(0, 0), 5, 4)
>>> rp.vertices
[Point(5, 0), Point(0, 5), Point(-5, 0), Point(0, -5)]
class sympy.geometry.polygon.Triangle
A polygon with three vertices and three sides.
Parameters points : sequence of Points
keyword: asa, sas, or sss to specify sides/angles of the triangle :
Raises GeometryError :
If the number of vertices is not equal to three, or one of the vertices
is not a Point, or a valid keyword is not given.
See Also:
sympy.geometry.point.Point (page 497), Polygon (page 556)
Examples
>>> from sympy.geometry import Triangle, Point
>>> Triangle(Point(0, 0), Point(4, 0), Point(4, 3))
Triangle(Point(0, 0), Point(4, 0), Point(4, 3))
Keywords sss, sas, or asa can be used to give the desired side lengths (in order) and
interior angles (in degrees) that dene the triangle:
>>> Triangle(sss=(3, 4, 5))
Triangle(Point(0, 0), Point(3, 0), Point(3, 4))
>>> Triangle(asa=(30, 1, 30))
570
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Triangle(Point(0, 0), Point(1, 0), Point(1/2, sqrt(3)/6))
>>> Triangle(sas=(1, 45, 2))
Triangle(Point(0, 0), Point(2, 0), Point(sqrt(2)/2, sqrt(2)/2))
Attributes
vertices
altitudes
orthocenter
circumcenter
circumradius
circumcircle
inradius
incircle
medians
medial
altitudes
The altitudes of the triangle.
An altitude of a triangle is a segment through a vertex, perpendicular to the opposite
side, with length being the height of the vertex measured from the line containing
the side.
Returns altitudes : dict
The dictionary consists of keys which are vertices and values which
are Segments.
See Also:
sympy.geometry.point.Point (page 497), sympy.geometry.line.Segment.length
(page 523)
Examples
>>> from sympy.geometry import Point, Triangle
>>> p1, p2, p3 = Point(0, 0), Point(1, 0), Point(0, 1)
>>> t = Triangle(p1, p2, p3)
>>> t.altitudes[p1]
Segment(Point(0, 0), Point(1/2, 1/2))
bisectors()
The angle bisectors of the triangle.
An angle bisector of a triangle is a straight line through a vertex which cuts the
corresponding angle in half.
Returns bisectors : dict
Each key is a vertex (Point) and each value is the corresponding bisector (Segment).
See Also:
sympy.geometry.point.Point
(page 522)
5.10. Geometry Module
(page
497),
sympy.geometry.line.Segment
571
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.geometry import Point, Triangle, Segment
>>> p1, p2, p3 = Point(0, 0), Point(1, 0), Point(0, 1)
>>> t = Triangle(p1, p2, p3)
>>> from sympy import sqrt
>>> t.bisectors()[p2] == Segment(Point(0, sqrt(2) - 1), Point(1, 0))
True
circumcenter
The circumcenter of the triangle
The circumcenter is the center of the circumcircle.
Returns circumcenter : Point
See Also:
sympy.geometry.point.Point (page 497)
Examples
>>> from sympy.geometry import Point, Triangle
>>> p1, p2, p3 = Point(0, 0), Point(1, 0), Point(0, 1)
>>> t = Triangle(p1, p2, p3)
>>> t.circumcenter
Point(1/2, 1/2)
circumcircle
The circle which passes through the three vertices of the triangle.
Returns circumcircle : Circle
See Also:
sympy.geometry.ellipse.Circle (page 554)
Examples
>>> from sympy.geometry import Point, Triangle
>>> p1, p2, p3 = Point(0, 0), Point(1, 0), Point(0, 1)
>>> t = Triangle(p1, p2, p3)
>>> t.circumcircle
Circle(Point(1/2, 1/2), sqrt(2)/2)
circumradius
The radius of the circumcircle of the triangle.
Returns circumradius : number of Basic instance
See Also:
sympy.geometry.ellipse.Circle.radius (page 555)
572
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import Symbol
>>> from sympy.geometry import Point, Triangle
>>> a = Symbol(a)
>>> p1, p2, p3 = Point(0, 0), Point(1, 0), Point(0, a)
>>> t = Triangle(p1, p2, p3)
>>> t.circumradius
sqrt(a**2/4 + 1/4)
incenter
The center of the incircle.
The incircle is the circle which lies inside the triangle and touches all three sides.
Returns incenter : Point
See Also:
incircle (page 573), sympy.geometry.point.Point (page 497)
Examples
>>> from sympy.geometry import Point, Triangle
>>> p1, p2, p3 = Point(0, 0), Point(1, 0), Point(0, 1)
>>> t = Triangle(p1, p2, p3)
>>> t.incenter
Point(-sqrt(2)/2 + 1, -sqrt(2)/2 + 1)
incircle
The incircle of the triangle.
The incircle is the circle which lies inside the triangle and touches all three sides.
Returns incircle : Circle
See Also:
sympy.geometry.ellipse.Circle (page 554)
Examples
>>> from sympy.geometry import Point, Triangle
>>> p1, p2, p3 = Point(0, 0), Point(2, 0), Point(0, 2)
>>> t = Triangle(p1, p2, p3)
>>> t.incircle
Circle(Point(-sqrt(2) + 2, -sqrt(2) + 2), -sqrt(2) + 2)
inradius
The radius of the incircle.
Returns inradius : number of Basic instance
See Also:
incircle (page 573), sympy.geometry.ellipse.Circle.radius (page 555)
5.10. Geometry Module
573
SymPy Documentation, Release 0.7.6
Examples
>>>
>>>
>>>
>>>
1
from sympy.geometry import Point, Triangle
p1, p2, p3 = Point(0, 0), Point(4, 0), Point(0, 3)
t = Triangle(p1, p2, p3)
t.inradius
is equilateral()
Are all the sides the same length?
Returns is equilateral : boolean
See Also:
sympy.geometry.entity.GeometryEntity.is similar (page 493), RegularPolygon (page 563), is isosceles (page 574), is right (page 574), is scalene
(page 575)
Examples
>>> from sympy.geometry import Triangle, Point
>>> t1 = Triangle(Point(0, 0), Point(4, 0), Point(4, 3))
>>> t1.is_equilateral()
False
>>> from sympy import sqrt
>>> t2 = Triangle(Point(0, 0), Point(10, 0), Point(5, 5*sqrt(3)))
>>> t2.is_equilateral()
True
is isosceles()
Are two or more of the sides the same length?
Returns is isosceles : boolean
See Also:
is equilateral (page 574), is right (page 574), is scalene (page 575)
Examples
>>> from sympy.geometry import Triangle, Point
>>> t1 = Triangle(Point(0, 0), Point(4, 0), Point(2, 4))
>>> t1.is_isosceles()
True
is right()
Is the triangle right-angled.
Returns is right : boolean
See Also:
sympy.geometry.line.LinearEntity.is perpendicular
(page
512),
is equilateral (page 574), is isosceles (page 574), is scalene (page 575)
574
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.geometry import Triangle, Point
>>> t1 = Triangle(Point(0, 0), Point(4, 0), Point(4, 3))
>>> t1.is_right()
True
is scalene()
Are all the sides of the triangle of dierent lengths?
Returns is scalene : boolean
See Also:
is equilateral (page 574), is isosceles (page 574), is right (page 574)
Examples
>>> from sympy.geometry import Triangle, Point
>>> t1 = Triangle(Point(0, 0), Point(4, 0), Point(1, 4))
>>> t1.is_scalene()
True
is similar(t1, t2)
Is another triangle similar to this one.
Two triangles are similar if one can be uniformly scaled to the other.
Parameters other: Triangle :
Returns is similar : boolean
See Also:
sympy.geometry.entity.GeometryEntity.is similar (page 493)
Examples
>>> from sympy.geometry import Triangle, Point
>>> t1 = Triangle(Point(0, 0), Point(4, 0), Point(4, 3))
>>> t2 = Triangle(Point(0, 0), Point(-4, 0), Point(-4, -3))
>>> t1.is_similar(t2)
True
>>> t2 = Triangle(Point(0, 0), Point(-4, 0), Point(-4, -4))
>>> t1.is_similar(t2)
False
medial
The medial triangle of the triangle.
The triangle which is formed from the midpoints of the three sides.
Returns medial : Triangle
See Also:
sympy.geometry.line.Segment.midpoint (page 524)
5.10. Geometry Module
575
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.geometry import Point, Triangle
>>> p1, p2, p3 = Point(0, 0), Point(1, 0), Point(0, 1)
>>> t = Triangle(p1, p2, p3)
>>> t.medial
Triangle(Point(1/2, 0), Point(1/2, 1/2), Point(0, 1/2))
medians
The medians of the triangle.
A median of a triangle is a straight line through a vertex and the midpoint of the
opposite side, and divides the triangle into two equal areas.
Returns medians : dict
Each key is a vertex (Point) and each value is the median (Segment)
at that point.
See Also:
sympy.geometry.point.Point.midpoint (page 501), sympy.geometry.line.Segment.midpoint
(page 524)
Examples
>>> from sympy.geometry import Point, Triangle
>>> p1, p2, p3 = Point(0, 0), Point(1, 0), Point(0, 1)
>>> t = Triangle(p1, p2, p3)
>>> t.medians[p1]
Segment(Point(0, 0), Point(1/2, 1/2))
orthocenter
The orthocenter of the triangle.
The orthocenter is the intersection of the altitudes of a triangle. It may lie inside,
outside or on the triangle.
Returns orthocenter : Point
See Also:
sympy.geometry.point.Point (page 497)
Examples
>>> from sympy.geometry import Point, Triangle
>>> p1, p2, p3 = Point(0, 0), Point(1, 0), Point(0, 1)
>>> t = Triangle(p1, p2, p3)
>>> t.orthocenter
Point(0, 0)
vertices
The triangles vertices
Returns vertices : tuple
Each element in the tuple is a Point
576
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
See Also:
sympy.geometry.point.Point (page 497)
Examples
>>> from sympy.geometry import Triangle, Point
>>> t = Triangle(Point(0, 0), Point(4, 0), Point(4, 3))
>>> t.vertices
(Point(0, 0), Point(4, 0), Point(4, 3))
Plane
Geometrical Planes.
Contains Plane
class sympy.geometry.plane.Plane
A plane is a at, two-dimensional surface. A plane is the two-dimensional analogue of a
point (zero-dimensions), a line (one-dimension) and a solid (three-dimensions). A plane
can generally be constructed by two types of inputs. They are three non-collinear points
and a point and the planes normal vector.
Examples
>>> from sympy import Plane, Point3D
>>> from sympy.abc import x
>>> Plane(Point3D(1, 1, 1), Point3D(2, 3, 4), Point3D(2, 2, 2))
Plane(Point3D(1, 1, 1), (-1, 2, -1))
>>> Plane((1, 1, 1), (2, 3, 4), (2, 2, 2))
Plane(Point3D(1, 1, 1), (-1, 2, -1))
>>> Plane(Point3D(1, 1, 1), normal_vector=(1,4,7))
Plane(Point3D(1, 1, 1), (1, 4, 7))
Attributes
p1
normal vector
angle between(o)
Angle between the plane and other geometric entity.
Parameters LinearEntity3D, Plane. :
Returns angle : angle in radians
Notes
This method accepts only 3D entities as its parameter, but if you want to calculate
the angle between a 2D entity and a plane you should rst convert to a 3D entity by
projecting onto a desired plane and then proceed to calculate the angle.
5.10. Geometry Module
577
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import Point3D, Line3D, Plane
>>> a = Plane(Point3D(1, 2, 2), normal_vector=(1, 2, 3))
>>> b = Line3D(Point3D(1, 3, 4), Point3D(2, 2, 2))
>>> a.angle_between(b)
-asin(sqrt(21)/6)
arbitrary point(t=None)
Returns an arbitrary point on the Plane; varying t from 0 to 2*pi will move the point
in a circle of radius 1 about p1 of the Plane.
Returns Point3D :
Examples
>>> from sympy.geometry.plane import Plane
>>> from sympy.abc import t
>>> p = Plane((0, 0, 0), (0, 0, 1), (0, 1, 0))
>>> p.arbitrary_point(t)
Point3D(0, cos(t), sin(t))
>>> _.distance(p.p1).simplify()
1
static are concurrent(*planes)
Is a sequence of Planes concurrent?
Two or more Planes are concurrent if their intersections are a common line.
Parameters planes: list :
Returns Boolean :
Examples
>>> from sympy import Plane, Point3D
>>> a = Plane(Point3D(5, 0, 0), normal_vector=(1, -1, 1))
>>> b = Plane(Point3D(0, -2, 0), normal_vector=(3, 1, 1))
>>> c = Plane(Point3D(0, -1, 0), normal_vector=(5, -1, 9))
>>> Plane.are_concurrent(a, b)
True
>>> Plane.are_concurrent(a, b, c)
False
distance(o)
Distance beteen the plane and another geometric entity.
Parameters Point3D, LinearEntity3D, Plane. :
Returns distance :
Notes
This method accepts only 3D entities as its parameter, but if you want to calculate
the distance between a 2D entity and a plane you should rst convert to a 3D entity
by projecting onto a desired plane and then proceed to calculate the distance.
578
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import Point, Point3D, Line, Line3D, Plane
>>> a = Plane(Point3D(1, 1, 1), normal_vector=(1, 1, 1))
>>> b = Point3D(1, 2, 3)
>>> a.distance(b)
sqrt(3)
>>> c = Line3D(Point3D(2, 3, 1), Point3D(1, 2, 2))
>>> a.distance(c)
0
equation(x=None, y=None, z=None)
The equation of the Plane.
Examples
>>> from sympy import Point3D, Plane
>>> a = Plane(Point3D(1, 1, 2), Point3D(2, 4, 7), Point3D(3, 5, 1))
>>> a.equation()
-23*x + 11*y - 2*z + 16
>>> a = Plane(Point3D(1, 4, 2), normal_vector=(6, 6, 6))
>>> a.equation()
6*x + 6*y + 6*z - 42
intersection(o)
The intersection with other geometrical entity.
Parameters Point, Point3D, LinearEntity, LinearEntity3D, Plane :
Returns List :
Examples
>>> from sympy import Point, Point3D, Line, Line3D, Plane
>>> a = Plane(Point3D(1, 2, 3), normal_vector=(1, 1, 1))
>>> b = Point3D(1, 2, 3)
>>> a.intersection(b)
[Point3D(1, 2, 3)]
>>> c = Line3D(Point3D(1, 4, 7), Point3D(2, 2, 2))
>>> a.intersection(c)
[Point3D(2, 2, 2)]
>>> d = Plane(Point3D(6, 0, 0), normal_vector=(2, -5, 3))
>>> e = Plane(Point3D(2, 0, 0), normal_vector=(3, 4, -3))
>>> d.intersection(e)
[Line3D(Point3D(78/23, -24/23, 0), Point3D(147/23, 321/23, 23))]
is coplanar(o)
Returns True if o is coplanar with self, else False.
Examples
>>> from sympy import Plane, Point3D
>>> o = (0, 0, 0)
>>> p = Plane(o, (1, 1, 1))
5.10. Geometry Module
579
SymPy Documentation, Release 0.7.6
>>> p2 = Plane(o, (2, 2, 2))
>>> p == p2
False
>>> p.is_coplanar(p2)
True
is parallel(l)
Is the given geometric entity parallel to the plane?
Parameters LinearEntity3D or Plane :
Returns Boolean :
Examples
>>> from sympy import Plane, Point3D
>>> a = Plane(Point3D(1,4,6), normal_vector=(2, 4, 6))
>>> b = Plane(Point3D(3,1,3), normal_vector=(4, 8, 12))
>>> a.is_parallel(b)
True
is perpendicular(l)
is the given geometric entity perpendicualar to the given plane?
Parameters LinearEntity3D or Plane :
Returns Boolean :
Examples
>>> from sympy import Plane, Point3D
>>> a = Plane(Point3D(1,4,6), normal_vector=(2, 4, 6))
>>> b = Plane(Point3D(2, 2, 2), normal_vector=(-1, 2, -1))
>>> a.is_perpendicular(b)
True
normal vector
Normal vector of the given plane.
Examples
>>> from sympy import Point3D, Plane
>>> a = Plane(Point3D(1, 1, 1), Point3D(2, 3, 4), Point3D(2, 2, 2))
>>> a.normal_vector
(-1, 2, -1)
>>> a = Plane(Point3D(1, 1, 1), normal_vector=(1, 4, 7))
>>> a.normal_vector
(1, 4, 7)
p1
The only dening point of the plane. Others can be obtained from the arbitrary point
method.
See Also:
sympy.geometry.point3d.Point3D (page 503)
580
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import Point3D, Plane
>>> a = Plane(Point3D(1, 1, 1), Point3D(2, 3, 4), Point3D(2, 2, 2))
>>> a.p1
Point3D(1, 1, 1)
parallel plane(pt)
Plane parallel to the given plane and passing through the point pt.
Parameters pt: Point3D :
Returns Plane :
Examples
>>> from sympy import Plane, Point3D
>>> a = Plane(Point3D(1, 4, 6), normal_vector=(2, 4, 6))
>>> a.parallel_plane(Point3D(2, 3, 5))
Plane(Point3D(2, 3, 5), (2, 4, 6))
perpendicular line(pt)
A line perpendicular to the given plane.
Parameters pt: Point3D :
Returns Line3D :
Examples
>>> from sympy import Plane, Point3D, Line3D
>>> a = Plane(Point3D(1,4,6), normal_vector=(2, 4, 6))
>>> a.perpendicular_line(Point3D(9, 8, 7))
Line3D(Point3D(9, 8, 7), Point3D(11, 12, 13))
perpendicular plane(*pts)
Return a perpendicular passing through the given points. If the direction ratio between the points is the same as the Planes normal vector then, to select from the
innite number of possible planes, a third point will be chosen on the z-axis (or the
y-axis if the normal vector is already parallel to the z-axis). If less than two points
are given they will be supplied as follows: if no point is given then pt1 will be self.p1;
if a second point is not given it will be a point through pt1 on a line parallel to the
z-axis (if the normal is not already the z-axis, otherwise on the line parallel to the
y-axis).
Parameters pts: 0, 1 or 2 Point3D :
Returns Plane :
Examples
>>>
>>>
>>>
>>>
from sympy import Plane, Point3D, Line3D
a, b = Point3D(0, 0, 0), Point3D(0, 1, 0)
Z = (0, 0, 1)
p = Plane(a, normal_vector=Z)
5.10. Geometry Module
581
SymPy Documentation, Release 0.7.6
>>> p.perpendicular_plane(a, b)
Plane(Point3D(0, 0, 0), (1, 0, 0))
projection(pt)
Project the given point onto the plane along the plane normal.
Parameters Point or Point3D :
Returns Point3D :
Examples
>>> from sympy import Plane, Point, Point3D
>>> A = Plane(Point3D(1, 1, 2), normal_vector=(1, 1, 1))
The projection is along the normal vector direction, not the z axis, so (1, 1) does not
project to (1, 1, 2) on the plane A:
>>> b = Point(1, 1)
>>> A.projection(b)
Point3D(5/3, 5/3, 2/3)
>>> _ in A
True
But the point (1, 1, 2) projects to (1, 1) on the XY-plane:
>>> XY = Plane((0, 0, 0), (0, 0, 1))
>>> XY.projection((1, 1, 2))
Point3D(1, 1, 0)
projection line(line)
Project the given line onto the plane through the normal plane containing the line.
Parameters LinearEntity or LinearEntity3D :
Returns Point3D, Line3D, Ray3D or Segment3D :
Notes
For the interaction between 2D and 3D lines(segments, rays), you should convert the
line to 3D by using this method. For example for nding the intersection between
a 2D and a 3D line, convert the 2D line to a 3D line by projecting it on a required
plane and then proceed to nd the intersection between those lines.
Examples
>>> from sympy import Plane, Line, Line3D, Point, Point3D
>>> a = Plane(Point3D(1, 1, 1), normal_vector=(1, 1, 1))
>>> b = Line(Point(1, 1), Point(2, 2))
>>> a.projection_line(b)
Line3D(Point3D(4/3, 4/3, 1/3), Point3D(5/3, 5/3, -1/3))
>>> c = Line3D(Point3D(1, 1, 1), Point3D(2, 2, 2))
>>> a.projection_line(c)
Point3D(1, 1, 1)
582
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
random point(seed=None)
Returns a random point on the Plane.
Returns Point3D :
5.11 Symbolic Integrals
The integrals module in SymPy implements methods to calculate denite and indenite
integrals of expressions.
Principal method in this module is integrate() (page 608)
integrate(f, x) returns the indenite integral f dx
integrate(f, (x, a, b)) returns the denite integral
b
a
f dx
5.11.1 Examples
SymPy can integrate a vast array of functions. It can integrate polynomial functions:
>>> from sympy import *
>>> init_printing(use_unicode=False, wrap_line=False, no_global=True)
>>> x = Symbol(x)
>>> integrate(x**2 + x + 1, x)
3
2
x
x
-- + -- + x
3
2
Rational functions:
>>> integrate(x/(x**2+2*x+1), x)
1
log(x + 1) + ----x + 1
Exponential-polynomial functions. These multiplicative combinations of polynomials and the
functions exp, cos and sin can be integrated by hand using repeated integration by parts,
which is an extremely tedious process. Happily, SymPy will deal with these integrals.
>>> integrate(x**2 * exp(x) * cos(x), x)
2 x
2 x
x
x
x *e *sin(x)
x *e *cos(x)
x
e *sin(x)
e *cos(x)
------------ + ------------ - x*e *sin(x) + --------- - --------2
2
2
2
even a few nonelementary integrals (in particular, some integrals involving the error function)
can be evaluated:
>>> integrate(exp(-x**2)*erf(x), x)
____
2
\/ pi *erf (x)
-------------4
5.11. Symbolic Integrals
583
SymPy Documentation, Release 0.7.6
5.11.2 Integral Transforms
SymPy has special support for denite integrals, and integral transforms.
sympy.integrals.transforms.mellin transform(f, x, s, **hints)
Compute the Mellin transform F (s) of f (x),
F (s) =
xs1 f (x)dx.
0
For all sensible functions, this converges absolutely in a strip a < Re(s) < b.
The Mellin transform is related via change of variables to the Fourier transform, and
also to the (bilateral) Laplace transform.
This function returns (F, (a, b), cond) where F is the Mellin transform of f, (a, b)
is the fundamental strip (as above), and cond are auxiliary convergence conditions.
If the integral cannot be computed in closed form, this function returns an unevaluated
MellinTransform object.
For
a
description
of
possible
hints,
refer
to
the
docstring
of
sympy.integrals.transforms.IntegralTransform.doit(). If noconds=False, then
only F will be returned (i.e. not cond, and also not the strip (a, b)).
>>> from sympy.integrals.transforms import mellin_transform
>>> from sympy import exp
>>> from sympy.abc import x, s
>>> mellin_transform(exp(-x), x, s)
(gamma(s), (0, oo), True)
See Also:
inverse mellin transform
(page
fourier transform
(page
586),
verse hankel transform (page 590)
584),
laplace transform
hankel transform
(page
(page
589),
585),
in-
sympy.integrals.transforms.inverse mellin transform(F, s, x, strip, **hints)
Compute the inverse Mellin transform of F (s) over the fundamental strip given by
strip=(a, b).
This can be dened as
c+i
f (x) =
xs F (s)ds,
ci
for any c in the fundamental strip. Under certain regularity conditions on F and/or f ,
this recovers f from its Mellin transform F (and vice versa), for positive real x.
One of a or b may be passed as None; a suitable c will be inferred.
If the integral cannot be computed in closed form, this function returns an unevaluated
InverseMellinTransform object.
Note that this function will assume x to be positive and real, regardless of the sympy
assumptions!
For
a
description
of
possible
hints,
refer
to
sympy.integrals.transforms.IntegralTransform.doit().
584
the
docstring
of
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> from sympy.integrals.transforms import inverse_mellin_transform
>>> from sympy import oo, gamma
>>> from sympy.abc import x, s
>>> inverse_mellin_transform(gamma(s), s, x, (0, oo))
exp(-x)
The fundamental strip matters:
>>> f = 1/(s**2 - 1)
>>> inverse_mellin_transform(f, s, x, (-oo, -1))
(x/2 - 1/(2*x))*Heaviside(x - 1)
>>> inverse_mellin_transform(f, s, x, (-1, 1))
-x*Heaviside(-x + 1)/2 - Heaviside(x - 1)/(2*x)
>>> inverse_mellin_transform(f, s, x, (1, oo))
(-x/2 + 1/(2*x))*Heaviside(-x + 1)
See Also:
mellin transform
(page
584),
verse hankel transform (page 590)
hankel transform
(page
589),
in-
sympy.integrals.transforms.laplace transform(f, t, s, **hints)
Compute the Laplace Transform F (s) of f (t),
F (s) =
est f (t)dt.
0
For all sensible functions, this converges absolutely in a half plane a < Re(s).
This function returns (F, a, cond) where F is the Laplace transform of f, Re(s) > a is
the half-plane of convergence, and cond are auxiliary convergence conditions.
If the integral cannot be computed in closed form, this function returns an unevaluated
LaplaceTransform object.
For
a
description
of
possible
hints,
refer
to
the
docstring
of
sympy.integrals.transforms.IntegralTransform.doit(). If noconds=True, only F
will be returned (i.e. not cond, and also not the plane a).
>>> from sympy.integrals import laplace_transform
>>> from sympy.abc import t, s, a
>>> laplace_transform(t**a, t, s)
(s**(-a)*gamma(a + 1)/s, 0, -re(a) < 1)
See Also:
inverse laplace transform
(page
fourier transform
(page
586),
verse hankel transform (page 590)
585),
mellin transform
hankel transform
(page
sympy.integrals.transforms.inverse laplace transform(F, s, t,
**hints)
Compute the inverse Laplace transform of F (s), dened as
(page
589),
584),
in-
plane=None,
c+i
est F (s)ds,
f (t) =
ci
for c so large that F (s) has no singularites in the half-plane Re(s) > c .
5.11. Symbolic Integrals
585
SymPy Documentation, Release 0.7.6
The plane can be specied by argument plane, but will be inferred if passed as None.
Under certain regularity conditions, this recovers f (t) from its Laplace Transform F (s),
for non-negative t, and vice versa.
If the integral cannot be computed in closed form, this function returns an unevaluated
InverseLaplaceTransform object.
Note that this function will always assume t to be real, regardless of the sympy assumption on t.
For
a
description
of
possible
hints,
refer
to
sympy.integrals.transforms.IntegralTransform.doit().
the
docstring
of
>>> from sympy.integrals.transforms import inverse_laplace_transform
>>> from sympy import exp, Symbol
>>> from sympy.abc import s, t
>>> a = Symbol(a, positive=True)
>>> inverse_laplace_transform(exp(-a*s)/s, s, t)
Heaviside(-a + t)
See Also:
laplace transform
(page
585),
verse hankel transform (page 590)
hankel transform
(page
589),
in-
sympy.integrals.transforms.fourier transform(f, x, k, **hints)
Compute the unitary, ordinary-frequency Fourier transform of f , dened as
F (k) =
f (x)e2ixk dx.
If the transform cannot be computed in closed form, this function returns an unevaluated
FourierTransform object.
For
other
Fourier
transform
conventions,
sympy.integrals.transforms. fourier transform().
see
the
function
For
a
description
of
possible
hints,
refer
to
the
docstring
of
sympy.integrals.transforms.IntegralTransform.doit(). Note that for this transform, by default noconds=True.
>>> from sympy import fourier_transform, exp
>>> from sympy.abc import x, k
>>> fourier_transform(exp(-x**2), x, k)
sqrt(pi)*exp(-pi**2*k**2)
>>> fourier_transform(exp(-x**2), x, k, noconds=False)
(sqrt(pi)*exp(-pi**2*k**2), True)
See Also:
inverse fourier transform (page 586),
sine transform (page 587),
in(page
587),
cosine transform
(page
588),
inverse sine transform
verse cosine transform (page 588),
hankel transform (page 589),
inverse hankel transform (page 590), mellin transform (page 584), laplace transform
(page 585)
sympy.integrals.transforms.inverse fourier transform(F, k, x, **hints)
586
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Compute the unitary, ordinary-frequency inverse Fourier transform of F , dened as
f (x) =
F (k)e2ixk dk.
If the transform cannot be computed in closed form, this function returns an unevaluated
InverseFourierTransform object.
For
other
Fourier
transform
conventions,
sympy.integrals.transforms. fourier transform().
see
the
function
For
a
description
of
possible
hints,
refer
to
the
docstring
of
sympy.integrals.transforms.IntegralTransform.doit(). Note that for this transform, by default noconds=True.
>>> from sympy import inverse_fourier_transform, exp, sqrt, pi
>>> from sympy.abc import x, k
>>> inverse_fourier_transform(sqrt(pi)*exp(-(pi*k)**2), k, x)
exp(-x**2)
>>> inverse_fourier_transform(sqrt(pi)*exp(-(pi*k)**2), k, x, noconds=False)
(exp(-x**2), True)
See Also:
fourier transform (page 586), sine transform (page 587), inverse sine transform
(page 587), cosine transform (page 588), inverse cosine transform (page 588), hankel transform (page 589), inverse hankel transform (page 590), mellin transform
(page 584), laplace transform (page 585)
sympy.integrals.transforms.sine transform(f, x, k, **hints)
Compute the unitary, ordinary-frequency sine transform of f , dened as
2
F (k) =
f (x) sin(2xk)dx.
0
If the transform cannot be computed in closed form, this function returns an unevaluated
SineTransform object.
For
a
description
of
possible
hints,
refer
to
the
docstring
of
sympy.integrals.transforms.IntegralTransform.doit(). Note that for this transform, by default noconds=True.
>>> from sympy import sine_transform, exp
>>> from sympy.abc import x, k, a
>>> sine_transform(x*exp(-a*x**2), x, k)
sqrt(2)*k*exp(-k**2/(4*a))/(4*a**(3/2))
>>> sine_transform(x**(-a), x, k)
2**(-a + 1/2)*k**(a - 1)*gamma(-a/2 + 1)/gamma(a/2 + 1/2)
See Also:
fourier transform
(page
586),
inverse fourier transform
(page
586),
inverse sine transform (page 587),
cosine transform (page 588),
inhankel transform (page 589),
inverse cosine transform (page 588),
verse hankel transform (page 590), mellin transform (page 584), laplace transform
(page 585)
5.11. Symbolic Integrals
587
SymPy Documentation, Release 0.7.6
sympy.integrals.transforms.inverse sine transform(F, k, x, **hints)
Compute the unitary, ordinary-frequency inverse sine transform of F , dened as
2
f (x) =
F (k) sin(2xk)dk.
0
If the transform cannot be computed in closed form, this function returns an unevaluated
InverseSineTransform object.
For
a
description
of
possible
hints,
refer
to
the
docstring
of
sympy.integrals.transforms.IntegralTransform.doit(). Note that for this transform, by default noconds=True.
>>> from sympy import inverse_sine_transform, exp, sqrt, gamma, pi
>>> from sympy.abc import x, k, a
>>> inverse_sine_transform(2**((1-2*a)/2)*k**(a - 1)*
...
gamma(-a/2 + 1)/gamma((a+1)/2), k, x)
x**(-a)
>>> inverse_sine_transform(sqrt(2)*k*exp(-k**2/(4*a))/(4*sqrt(a)**3), k, x)
x*exp(-a*x**2)
See Also:
fourier transform
(page
586),
inverse fourier transform
(page
586),
sine transform (page 587), cosine transform (page 588), inverse cosine transform
(page 588), hankel transform (page 589), inverse hankel transform (page 590),
mellin transform (page 584), laplace transform (page 585)
sympy.integrals.transforms.cosine transform(f, x, k, **hints)
Compute the unitary, ordinary-frequency cosine transform of f , dened as
2
F (k) =
f (x) cos(2xk)dx.
0
If the transform cannot be computed in closed form, this function returns an unevaluated
CosineTransform object.
For
a
description
of
possible
hints,
refer
to
the
docstring
of
sympy.integrals.transforms.IntegralTransform.doit(). Note that for this transform, by default noconds=True.
>>> from sympy import cosine_transform, exp, sqrt, cos
>>> from sympy.abc import x, k, a
>>> cosine_transform(exp(-a*x), x, k)
sqrt(2)*a/(sqrt(pi)*(a**2 + k**2))
>>> cosine_transform(exp(-a*sqrt(x))*cos(a*sqrt(x)), x, k)
a*exp(-a**2/(2*k))/(2*k**(3/2))
See Also:
fourier transform
(page
586),
inverse fourier transform
(page
586),
sine transform
(page
587),
inverse sine transform
(page
587),
inverse cosine transform (page 588),
hankel transform (page 589),
inverse hankel transform (page 590), mellin transform (page 584), laplace transform
(page 585)
588
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
sympy.integrals.transforms.inverse cosine transform(F, k, x, **hints)
Compute the unitary, ordinary-frequency inverse cosine transform of F , dened as
2
f (x) =
F (k) cos(2xk)dk.
0
If the transform cannot be computed in closed form, this function returns an unevaluated
InverseCosineTransform object.
For
a
description
of
possible
hints,
refer
to
the
docstring
of
sympy.integrals.transforms.IntegralTransform.doit(). Note that for this transform, by default noconds=True.
>>> from sympy import inverse_cosine_transform, exp, sqrt, pi
>>> from sympy.abc import x, k, a
>>> inverse_cosine_transform(sqrt(2)*a/(sqrt(pi)*(a**2 + k**2)), k, x)
exp(-a*x)
>>> inverse_cosine_transform(1/sqrt(k), k, x)
1/sqrt(x)
See Also:
fourier transform
(page
586),
inverse fourier transform
(page
586),
sine transform (page 587), inverse sine transform (page 587), cosine transform
(page 588), hankel transform (page 589), inverse hankel transform (page 590),
mellin transform (page 584), laplace transform (page 585)
sympy.integrals.transforms.hankel transform(f, r, k, nu, **hints)
Compute the Hankel transform of f , dened as
F (k) =
f (r)J (kr)rdr.
0
If the transform cannot be computed in closed form, this function returns an unevaluated
HankelTransform object.
For
a
description
of
possible
hints,
refer
to
the
docstring
of
sympy.integrals.transforms.IntegralTransform.doit(). Note that for this transform, by default noconds=True.
>>> from sympy import hankel_transform, inverse_hankel_transform
>>> from sympy import gamma, exp, sinh, cosh
>>> from sympy.abc import r, k, m, nu, a
>>> ht = hankel_transform(1/r**m, r, k, nu)
>>> ht
2*2**(-m)*k**(m - 2)*gamma(-m/2 + nu/2 + 1)/gamma(m/2 + nu/2)
>>> inverse_hankel_transform(ht, k, r, nu)
r**(-m)
>>> ht = hankel_transform(exp(-a*r), r, k, 0)
>>> ht
a/(k**3*(a**2/k**2 + 1)**(3/2))
5.11. Symbolic Integrals
589
SymPy Documentation, Release 0.7.6
>>> inverse_hankel_transform(ht, k, r, 0)
exp(-a*r)
See Also:
fourier transform
(page
586),
inverse fourier transform
(page
586),
sine transform (page 587), inverse sine transform (page 587), cosine transform
(page 588), inverse cosine transform (page 588), inverse hankel transform
(page 590), mellin transform (page 584), laplace transform (page 585)
sympy.integrals.transforms.inverse hankel transform(F, k, r, nu, **hints)
Compute the inverse Hankel transform of F dened as
F (k)J (kr)kdk.
f (r) =
0
If the transform cannot be computed in closed form, this function returns an unevaluated
InverseHankelTransform object.
For
a
description
of
possible
hints,
refer
to
the
docstring
of
sympy.integrals.transforms.IntegralTransform.doit(). Note that for this transform, by default noconds=True.
>>> from sympy import hankel_transform, inverse_hankel_transform, gamma
>>> from sympy import gamma, exp, sinh, cosh
>>> from sympy.abc import r, k, m, nu, a
>>> ht = hankel_transform(1/r**m, r, k, nu)
>>> ht
2*2**(-m)*k**(m - 2)*gamma(-m/2 + nu/2 + 1)/gamma(m/2 + nu/2)
>>> inverse_hankel_transform(ht, k, r, nu)
r**(-m)
>>> ht = hankel_transform(exp(-a*r), r, k, 0)
>>> ht
a/(k**3*(a**2/k**2 + 1)**(3/2))
>>> inverse_hankel_transform(ht, k, r, 0)
exp(-a*r)
See Also:
fourier transform
(page
586),
inverse fourier transform
(page
586),
sine transform (page 587), inverse sine transform (page 587), cosine transform
(page 588), inverse cosine transform (page 588), hankel transform (page 589),
mellin transform (page 584), laplace transform (page 585)
5.11.3 Internals
There is a general method for calculating antiderivatives of elementary functions, called the
Risch algorithm. The Risch algorithm is a decision procedure that can determine whether an
elementary solution exists, and in that case calculate it. It can be extended to handle many
nonelementary functions in addition to the elementary ones.
SymPy currently uses a simplied version of the Risch algorithm, called the Risch-Norman
algorithm. This algorithm is much faster, but may fail to nd an antiderivative, although it is
590
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
still very powerful. SymPy also uses pattern matching and heuristics to speed up evaluation
of some types of integrals, e.g. polynomials.
For non-elementary denite integrals, SymPy uses so-called Meijer G-functions. Details are
described here:
Computing Integrals using Meijer G-Functions
This text aims do describe in some detail the steps (and subtleties) involved in using Meijer
G-functions for computing denite and indenite integrals. We shall ignore proofs completely.
Overview
The algorithm to compute
f (x)dx or
f (x)dx generally consists of three steps:
1. Rewrite the integrand using Meijer G-functions (one or sometimes two).
2. Apply an integration theorem, to get the answer (usually expressed as another Gfunction).
3. Expand the result in named special functions.
Step (3) is implemented in the function hyperexpand (q.v.). Steps (1) and (2) are described
below. Morever, G-functions are usually branched. Thus our treatment of branched functions
is described rst.
Some other integrals (e.g. ) can also be computed by rst recasting them into one of the
above forms. There is a lot of choice involved here, and the algorithm is heuristic at best.
Polar Numbers and Branched Functions
Both Meijer G-Functions and Hypergeometric functions are typically branched (possible
branchpoints being 0, 1, ). This is not very important when e.g. expanding a single hypergeometric function into named special functions, since sorting out the branches can be left
to the human user. However this algorithm manipulates and transforms G-functions, and to
do this correctly it needs at least some crude understanding of the branchings involved.
To begin, we consider the set S = {(r, ) : r > 0, R}. We have a map p : S : C {0}, (r, ) 7
rei . Decreeing this to be a local biholomorphism gives S both a topology and a complex
structure. This Riemann Surface is usually referred to as the Riemann Surface of the logarithm, for the following reason: We can dene maps Exp : C S, (x + iy) 7 (exp(x), y) and
Log : S C, (ex , y) 7 x + iy. These can both be shown to be holomorphic, and are indeed
mutual inverses.
We also sometimes formally attach a point zero (0) to S and denote the resulting object
S0 . Notably there is no complex structure dened near 0. A fundamental system of neighbourhoods is given by {Exp(z) : <(z) < k}, which at least denes a topology. Elements of
S0 shall be called polar numbers. We further dene functions Arg : S R, (r, ) 7 and
|.| : S0 R>0 , (r, ) 7 r. These have evident meaning and are both continuous everywhere.
Using these maps many operations can be extended from C to S. We dene Exp(a) Exp(b) =
Exp(a + b) for a, b C, also for a S and b C we dene ab = Exp(b Log(a)). It can be checked
easily that using these denitions, many algebraic properties holding for positive reals (e.g.
(ab)c = ac bc ) which hold in C only for some numbers (because of branch cuts) hold indeed for
all polar numbers.
5.11. Symbolic Integrals
591
SymPy Documentation, Release 0.7.6
As one peculiarity it should be mentioned that addition of polar numbers is not usually dened.
However, formal sums of polar numbers
can be used to express
branching behaviour. For
example, consider the functions F (z) = 1 + z and G(a, b) = a + b, where a, b, z are polar
numbers. The general rule is that functions of a single polar variable are dened in such a
way that they are continuous on circles, and agree with the usual denition for positive reals.
Thus if S(z) denotes the standard branch of the square root function on C, we are forced to
dene
: |z| < 1
S(p(z))
F (z) = S(p(z))
: < Arg(z) + 4n for some n Z .
S(p(z)) : else
(We
are omitting |z| = 1 here, this does not matter for integration.) Finally we dene G(a, b) =
aF (b/a).
Representing Branched Functions on the Argand Plane
Suppose f : S C is a holomorphic function. We wish to dene a function F on (part of)
the complex numbers C that represents f as closely as possible. This process is knows as
introducing branch cuts. In our situation, there is actually a canonical way of doing this
(which is adhered to in all of SymPy), as follows: Introduce the cut complex plane C = C\R0 .
Dene a function l : C S via rei 7 r Exp(i). Here r > 0 and < . Then l is
holomorphic, and we dene G = f l. This called lifting to the principal branch throughout
the SymPy documentation.
Table Lookups and Inverse Mellin Transforms
Suppose we are given an integrand f (x) and are trying to rewrite it as a single G-function.
To do this, we rst split f (x) into the form xs g(x) (where g(x) is supposed to be simpler than
f (x)). This is because multiplicative powers can be absorbed into the G-function later. This
splitting is done by split mul(f, x). Then we assemble a tuple of functions that occur in f
(e.g. if f (x) = ex cos x, we would assemble the tuple (cos, exp)). This is done by the function
mytype(f, x). Next we index a lookup table (created using create lookup table()) with
this tuple. This (hopefully) yields a list of Meijer G-function formulae involving these functions, we then pattern-match all of them. If one ts, we were successful, otherwise not and
we have to try something else.
Suppose now we want to rewrite as a product of two G-functions. To do this, we (try to)
nd all inequivalent ways of splitting f (x) into a product f1 (x)f2 (x). We could try these splittings in any order, but it is often a good idea to minimise (a) the number of powers occuring in fi (x) and (b) the number of dierent functions occuring in fi (x). Thus given e.g.
f (x) = sin x ex sin 2x we should try f1 (x) = sin x sin 2x, f2 (x) = ex rst. All of this is done by the
function mul as two parts(f).
Finally, we can try a recursive Mellin transform technique. Since the Meijer G-function is
dened essentially as a certain inverse mellin transform, if we want to write a function f (x)
as a G-function, we can compute its mellin transform F (s). If F (s) is in the right form, the
G-function expression can be read o. This technique generalises many standard rewritings,
e.g. eax ebx = e(a+b)x .
One twist is that some functions dont have mellin transforms, even though they can be written
as G-functions. This is true for example for f (x) = ex sin x (the function grows too rapidly to
have a mellin transform). However if the function is recognised to be analytic, then we can
592
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
try to compute the mellin-transform of f (ax) for a parameter a, and deduce the G-function
expression by analytic continuation. (Checking for analyticity is easy. Since we can only
deal with a certain subset of functions anyway, we only have to lter out those which are not
analyitc.)
The function rewrite single does the table lookup and recursive mellin transform.
The functions rewrite1 and rewrite2 respectively use above-mentioned helpers and
rewrite single to rewrite their argument as respectively one or two G-functions.
Applying the Integral Theorems
If the integrand has been recast into G-functions, evaluating the integral is relatively easy. We
rst do some substitutions to reduce e.g. the exponent of the argument of the G-function to
unity (see rewrite saxena 1 and rewrite saxena, respectively, for one or two G-functions).
Next we go through a list of conditions under which the integral theorem applies. It can fail
for basically two reasons: either the integral does not exist, or the manipulations in deriving
the theorem may not be allowed (for more details, see this [BlogPost] (page 1909)).
Sometimes this can be remedied by reducing the argument of the G-functions involved. For
example it is clear that the G-function representing ez is satises G(Exp(2i)z) = G(z) for all
z S. The function meijerg.get period() can be used to discover this, and the function
principal branch(z, period) in functions/elementary/complexes.py can be used to exploit the information. This is done transparently by the integration code.
The G-Function Integration Theorems
This section intends to display in detail the denite integration theorems used in the code.
The following two formulae go back to Meijer (In fact he proved more general formulae;
indeed in the literature formulae are usually staded in more general form. However it is very
easy to deduce the general formulae from the ones we give here. It seemed best to keep the
theorems as simple as possible, since they are very complicated anyway.):
1.
(
Gm,n
p,q
2.
n
m
)
(aj )
a1 , . . . , ap
j=1 (bj + 1)
p j=1
x dx = q
b1 , . . . , bq
j=m+1 (bj ) j=n+1 (aj + 1)
)
)
)
(
(
c1 , . . . , cu
a1 , . . . , ap
a1 , . . . , an , d1 , . . . , dv , an+1 , . . . , ap
m+t,n+s
m,n
x Gp,q
x dx = Gv+p,u+q
d1 , . . . , dv
b1 , . . . , b q
b1 , . . . , bm , c1 , . . . , cu , bm+1 , . . . , bq
(
Gs,t
u,v
The more interesting question is under what conditions these formulae are valid. Below we
detail the conditions implemented in SymPy. They are an amalgamation of conditions found
in [Prudnikov1990] (page 1912) and [Luke1969] (page 1912); please let us know if you nd
any errors.
Conditions of Convergence for Integral (1)
We can without loss of generality assume p q, since the G-functions of indices m, n, p, q and
of indices n, m, q, p can be related easily (see e.g. [Luke1969] (page 1912), section 5.3). We
5.11. Symbolic Integrals
593
SymPy Documentation, Release 0.7.6
introduce the following notation:
=m+np
p+q
=m+n
2
C3 : <(bj ) < 1 for j = 1, . . . , m
0 < <(aj ) for j = 1, . . . , n
C3 : <(bj ) < 1 for j = 1, . . . , q
0 < <(aj ) for j = 1, . . . , p
C4 : <() +
q+1p
>qp
2
The convergence conditions will be detailed in several cases, numbered one to ve. For
later use it will be helpful to separate conditions at innity from conditions at zero. By
conditions at innity we mean conditions that only depend on the behaviour of the integrand
for large, positive values of x, whereas by conditions at zero we mean conditions that only
depend on the behaviour of the integrand on (0, ) for any > 0. Since all our conditions are
specied in terms of parameters of the G-functions, this distinction is not immediately visible.
They are, however, of very distinct character mathematically; the conditions at innity being
in particular much harder to control.
In order for the integral theorem to be valid, conditions n at zero and at innity both have
to be fullled, for some n.
These are the conditions at innity:
1.
> 0 | arg()| < (A B C),
where
A=1np<q1m
B = 1 p 1 m q = p + 1 (n = 0 m = p + 1)
C = 1 n q = p | arg()| 6= ( 2k) for k = 0, 1, . . .
594
.
2
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
2.
n = 0 p + 1 m | arg()| <
3.
(p < q 1 m > 0 | arg()| = ) (p q 2 = 0 arg() = 0)
4.
p = q = 0 arg() = 0 =
6 0 <
bj aj < 0
j=1
5.
> 0 | arg()| <
And these are the conditions at zero:
1.
6= 0 C3
2.
C3
3.
C3 C4
4.
C3
5.
C3
5.11. Symbolic Integrals
595
SymPy Documentation, Release 0.7.6
Conditions of Convergence for Integral (2)
We introduce the following notation:
b = s + t
u+v
2
c = m + n
dj
j=1
j=1
p+q
2
cj +
uv
+1
2
aj +
pq
+1
2
j=1
bj
j=1
=qp
uv
+1
2
= 1 (v u)
(q m n) + | arg()|
qp
(v s t) + | arg()|)
vu
c = (q p)||1/(qp) cos + (v u)||1/(vu) cos
s0 (c1 , c2 ) = c1 (q p)||1/(qp) sin + c2 (v u)||1/(vu) sin
s0 (1, 1) s0 (1, 1)
(sign (arg ()) , 1) (sign (arg ()) , 1)
s0
s0
s =
s0 (1, sign (arg ())) s0 (1, sign (arg ()))
s0 (sign (arg ()) , sign (arg ()))
596
for arg() = 0 arg() = 0
for arg() 6= 0 arg() = 0
for arg() = 0 arg() 6= 0)
otherwise
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
z0 =
i(b +c )
e
z1 =
i(b +c )
e
The following conditions will be helpful:
C1 : (ai bj
/ Z>0 for i = 1, . . . , n, j = 1, . . . , m)
(ci dj
/ Z>0 for i = 1, . . . , t, j = 1, . . . , s)
C2 : <(1 + bi + dj ) > 0 for i = 1, . . . , m, j = 1, . . . , s
C3 : <(ai + cj ) < 1 for i = 1, . . . , n, j = 1, . . . , t
C4 : (p q)<(ci ) <() >
3
for i = 1, . . . , t
2
C5 : (p q)<(1 + di ) <() >
C6 : (u v)<(ai ) <() >
3
for i = 1, . . . , s
2
3
for i = 1, . . . , n
2
C7 : (u v)<(1 + bi ) <() >
3
for i = 1, . . . , m
2
C8 : 0 < || + 2< (( 1) (u + v) + (p + q) ( 1) + (p + q) (u + v))
C9 : 0 < || 2< (( 1) (u + v) + (p + q) ( 1) + (p + q) (u + v))
C10 : |arg ()| < b
C11 : |arg ()| = b
5.11. Symbolic Integrals
597
SymPy Documentation, Release 0.7.6
C12 : | arg()| < c
C13 : | arg()| = c
1
C14
: (z0 6= 1 | arg(1 z0 )| < ) (z0 = 1 <( + u + v) < 1)
2
C14
: (z1 6= 1 | arg(1 z1 )| < ) (z1 = 1 <( + p + q) < 1)
1
2
C14 : = 0 b + c 1 (C14
C14
)
C15 : c > 0 (c = 0 s 6= 0 <() > 1) (c = 0 s = 0 <() > 0)
C16 :
Gs,t
u,v (x)dx converges at innity
C17 :
Gm,n
p,q (x)dx converges at innity
Note that C16 and C17 are the reason we split the convergence conditions for integral (1).
With this notation established, the implemented convergence conditions can be enumerated
as follows:
1.
mnst 6= 0 0 < b 0 < c C1 C2 C3 C10 C12
2.
u = v b = 0 0 < c 0 < < < 1 C1 C2 C3 C12
3.
p = q u = v b = 0 c = 0 0 < 0 < < < 1 < < 1 6= C1 C2 C3
4.
p = q u = v b = 0 c = 0 0 < 0 < < ( + ) < 1 6= C1 C2 C3
598
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
5.
p = q u = v b = 0 c = 0 0 < 0 < < ( + ) < 1 6= C1 C2 C3
6.
q < p 0 < s 0 < b 0 c C1 C2 C3 C5 C10 C13
7.
p < q 0 < t 0 < b 0 c C1 C2 C3 C4 C10 C13
8.
v < u 0 < m 0 < c 0 b C1 C2 C3 C7 C11 C12
9.
u < v 0 < n 0 < c 0 b C1 C2 C3 C6 C11 C12
10.
q < p u = v b = 0 0 c 0 < < < 1 C1 C2 C3 C5 C13
11.
p < q u = v b = 0 0 c 0 < < < 1 C1 C2 C3 C4 C13
12.
p = q v < u 0 b c = 0 0 < < < 1 C1 C2 C3 C7 C11
13.
p = q u < v 0 b c = 0 0 < < < 1 C1 C2 C3 C6 C11
14.
p < q v < u 0 b 0 c C1 C2 C3 C4 C7 C11 C13
15.
q < p u < v 0 b 0 c C1 C2 C3 C5 C6 C11 C13
5.11. Symbolic Integrals
599
SymPy Documentation, Release 0.7.6
16.
q < p v < u 0 b 0 c C1 C2 C3 C5 C7 C8 C11 C13 C14
17.
p < q u < v 0 b 0 c C1 C2 C3 C4 C6 C9 C11 C13 C14
18.
t = 0 0 < s 0 < b 0 < C1 C2 C10
19.
s = 0 0 < t 0 < b < 0 C1 C3 C10
20.
n = 0 0 < m 0 < c < 0 C1 C2 C12
21.
m = 0 0 < n 0 < c 0 < C1 C3 C12
22.
st = 0 0 < b 0 < c C1 C2 C3 C10 C12
23.
mn = 0 0 < b 0 < c C1 C2 C3 C10 C12
24.
p < m + n t = 0 = 0 0 < s 0 < b c < 0 |arg ()| < (m + n p + 1) C1 C2 C10 C14 C15
25.
q < m + n s = 0 = 0 0 < t 0 < b c < 0 |arg ()| < (m + n q + 1) C1 C3 C10 C14 C15
26.
p = q 1 t = 0 = 0 0 < s 0 < b 0 c c < |arg ()| C1 C2 C10 C14 C15
600
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
27.
p = q + 1 s = 0 = 0 0 < t 0 < b 0 c c < |arg ()| C1 C3 C10 C14 C15
28.
p < q 1 t = 0 = 0 0 < s 0 < b 0 c c < |arg ()| |arg ()| < (m + n p + 1) C1 C2 C10
29.
q + 1 < p s = 0 = 0 0 < t 0 < b 0 c c < |arg ()| |arg ()| < (m + n q + 1) C1 C3 C10
30.
n = 0 = 0 0 < s + t 0 < m 0 < c b < 0 |arg ()| < (s + t u + 1) C1 C2 C12 C14 C15
31.
m = 0 = 0 v < s + t 0 < n 0 < c b < 0 |arg ()| < (s + t v + 1) C1 C3 C12 C14 C15
32.
n = 0 = 0 u = v 1 0 < m 0 < c 0 b b < |arg ()| |arg ()| < (b + 1) C1 C2 C12 C14
33.
m = 0 = 0 u = v + 1 0 < n 0 < c 0 b b < |arg ()| |arg ()| < (b + 1) C1 C3 C12 C14
34.
n = 0 = 0 u < v 1 0 < m 0 < c 0 b b < |arg ()| |arg ()| < (s + t u + 1) C1 C2 C12
35.
m = 0 = 0 v + 1 < u 0 < n 0 < c 0 b b < |arg ()| |arg ()| < (s + t v + 1) C1 C3 C12
36.
C17 t = 0 u < s 0 < b C10 C1 C2 C3
37.
C17 s = 0 v < t 0 < b C10 C1 C2 C3
5.11. Symbolic Integrals
601
SymPy Documentation, Release 0.7.6
38.
C16 n = 0 p < m 0 < c C12 C1 C2 C3
39.
C16 m = 0 q < n 0 < c C12 C1 C2 C3
The Inverse Laplace Transform of a G-function
The inverse laplace transform of a Meijer G-function can be expressed as another G-function.
This is a fairly versatile method for computing this transform. However, I could not nd
the details in the literature, so I work them out here. In [Luke1969] (page 1912), section
5.6.3, there is a formula for the inverse Laplace transform of a G-function of argument bz,
and convergence conditions are also given. However, we need a formula for argument bz a for
rational a.
We are asked to compute
f (t) =
1
2i
c+i
ezt G(bz a )dz,
ci
for positive real t. Three questions arise:
1. When does this integral converge?
2. How can we compute the integral?
3. When is our computation valid?
How to compute the integral
We shall work formally for now. Denote by (s) the product of gamma functions appearing in
the denition of G, so that
1
G(z) =
(s)z s ds.
2i L
Thus
1
f (t) =
(2i)2
c+i
ezt (s)bs z as dsdz.
ci
We interchange the order of integration to get
f (t) =
602
1
2i
c+i
ezt z as
bs (s)
L
ci
dz
ds.
2i
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
1
1
The inner integral is easily seen to be (as)
t1+as . (Using Cauchys theorem and Jordans
lemma deform the contour to run from to , encircling 0 once in the negative sense.
For as real and greater than one, this contour can be pushed onto the negative real axis
and the integral is recognised as a product of a sine and a gamma function. The formula is
then proved using the functional equation of the gamma function, and extended to the entire
domain of convergence of the original integral by appealing to analytic continuation.) Hence
we nd
( )s
1 1
1
b
f (t) =
(s)
ds,
t 2i L
(as) ta
which is a so-called Fox H function (of argument tba ). For rational a, this can be expressed as
a Meijer G-function using the gamma function multiplication theorem.
When this computation is valid
There are a number of obstacles in this computation. Interchange of integrals is only valid
if all integrals involved are absolutely convergent. In particular the inner integral has to
converge. Also, for our identication of the nal integral as a Fox H / Meijer G-function to be
correct, the poles of the newly obtained gamma function must be separated properly.
It is easy to check that the inner integal converges absolutely for <(as) < 1. Thus the contour
L has to run left of the line <(as) = 1. Under this condition, the poles of the newly-introduced
gamma function are separated properly.
It remains to observe that the Meijer G-function is an analytic, unbranched function of its
parameters, and of the coecient b. Hence so is f (t). Thus the nal computation remains
valid as long as the initial integral converges, and if there exists a changed set of parameters
where the computation is valid. If we assume w.l.o.g. that a > 0, then the latter condition is
fullled if G converges along contours (2) or (3) of [Luke1969] (page 1912), section 5.2, i.e.
either >= a2 or p 1, p q.
When the integral exists
Using [Luke1969] (page 1912), section 5.10, for any given meijer G-function we can nd a
c
dominant term of the form z a ebz (although this expression might not be the best possible,
because of cancellation).
We must thus investigate
c+iT
ezt z a ebz dz.
lim
ciT
(This principal value integral is the exact statement used in the Laplace inversion theorem.)
We write z = c + i . Then arg(z) 2 , and so ezt eit (where shall always mean asymptot
ically equivalent up to a positive real multiplicative constant). Also z x+iy | |x eiy log | | exi 2 .
Set = bei<(c) 2 . We have three cases:
1. b = 0 or <(c) 0. In this case the integral converges if <(a) 1.
2. b 6= 0, =(c) = 0, <(c) > 0. In this case the integral converges if <( ) < 0.
3. b 6= 0, =(c) = 0, <(c) > 0, <( ) 0, and at least one of <( ) = 0. Here the same condition
as in (1) applies.
5.11. Symbolic Integrals
603
SymPy Documentation, Release 0.7.6
Implemented G-Function Formulae
An important part of the algorithm is a table expressing various functions as Meijer Gfunctions. This is essentially a table of Mellin Transforms in disguise. The following automatically generated table shows the formulae currently implemented in SymPy. An entry
generated means that the corresponding G-function has a variable number of parameters.
This table is intended to shrink in future, when the algorithms capabilities of deriving new
formulae improve. Of course it has to grow whenever a new class of special functions is to
be dealt with. Elementary functions:
)
)
(
(
1
1,0
0,1 1
z
z + aG1,1
a = aG1,1
0
0
(z p + b)
q )
z p
sin (a)
b
(
a
1 a1 2,2 0, a
ba + (z q p)
=
b
G
2,2 0, a
zq p b
+ 12 , 2b + 1
0
q )
z p
b a2
(
)b
ab b 1,2 2b + 12 , 2b + 1
q
2
a + z p + a
= G2,2
b
2
q )
z p
0 a2
(
a+
604
zq p
a2
)b
ab b
= G1,2
2 2,2
)b
z q p + a2
z q p + a2
1
= ab1 G1,2
2,2
)b
a2 b
p + z q p + a = G2,1
2 2,2
b
(b
)b
(
a + z q p + a2
1 b1 1,2 2b + 12 , 2b
=
a G2,2
0
z q p + a2
a +
z2
q )
z p
b
(
ba 1,1 a + 1
=
G1,1
0
(a)
(b
2
+ 12 , 2b
b
(b
+1
0, 12
( b
b
)b
a2 b
2 + 1
z
p + z q p + a = G2,1
0, 21
2 2,2
q
2
q )
z p
b a2
q )
z p
0 a2
)
2b + 1 z q p
a
b
2
)
+ 1 z q p
a
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
(
z2
)b
(b 1
p + zq p + a
1 b 1
2 + 2
q
= a 2 2 G2,1
2,2
0, 21
z p+a
( q
)b
(
z 2 p + z q p + a
1 b 1 2,1 2b +
q
= a 2 2 G2,2
0, 21
z p+a
)
2b + 12 z q p
a
1
2
b
2
)
+ 12 z q p
a
Functions involving |z q p b|:
|z p b|
q
(
a
|b|
1,1 a + 1
( )
=
G2,2
0
cos a
2 (a)
)
a2 + 12 z q p
, if <a < 1
a2 + 12 b
Functions involving Chi (z q p):
(
3
2 2,0
Chi (z p) = G2,4
0, 0
2
q
1
p2 2q
2 , 1
z
1 1
2, 2 4
Functions involving Ci (z q p):
Ci (z q p) =
2,0
G1,3
0, 0
2
)
1 p2 2q
z
1
2 4
Functions involving Ei (z q p):
(
q
Ei (z p) =
iG1,0
1,1
)
)
(
)
(
1
1 q i
2,0
0,1 1
z
z G1,2
z pe
iG1,1
0
0, 0
0
Functions involving (z q p b):
q
a1
a1
(z p b)
(
(z p b) = b
q
a1
G0,1
1,1
(
(z p + b)
a1
(z p + b) = b
G1,0
1,1
q )
z p
(a), if b > 0
0 b
)
a z q p
b (a), if b > 0
0
a1
(
( ) q1 )
b
a
= ba1 G0,1
z
1,1
p
a1
(
)
( ) q1 )
b
a z q p
= ba1 G1,0
z +
1,1 0
b (a), if b > 0
p
(z p b)
q )
z p
(a), if b > 0
0 b
(z p + b)
5.11. Symbolic Integrals
605
SymPy Documentation, Release 0.7.6
Functions involving (z q p + 1), log (z q p):
n
log (z q p) (z q p + 1) = generated
log (z q p) (z q p 1) = generated
Functions involving Shi (z q p):
2
)
p 2q i
z e
21 , 12 4
p q 1,1 21
z G1,3
Shi (z p) =
0
4
q
Functions involving Si (z q p):
1,1
G
2 1,3
Si (z q p) =
2
)
p 2q
z
0, 0 4
1
1
2
Functions involving Ia (z q p):
(
q
Ia (z p) =
G1,0
1,3
1
a
p2 2q
2 + 2
z
a a
1
2, 2 + 2 4
a
2
Functions involving Ja (z q p):
2
)
p 2q
z
a2 4
(
q
Ja (z p) =
G1,0
0,2
a
2
Functions involving Ka (z q p):
Ka (z q p) =
1 2,0
G
2 0,2
2
)
p 2q
z
4
(
a
a
2,2
Functions involving Ya (z q p):
)
a2 12 p2 2q
z
a2 12 4
(
q
Ya (z p) =
G2,0
1,3
a
a
2,2
Functions involving cos (z q p):
cos (z p) = G1,0
0,2
2
)
p 2q
z
1
2 4
Functions involving cosh (z q p):
(
606
2
)
p 2q
z
1 1
2, 2 4
1
2
cosh (z q p) = 2 G1,0
1,3
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Functions involving E (z q p):
)
q
z p
0
(1
3
2, 2
1
E (z p) = G1,2
4 2,2
q
Functions involving K (z q p):
1
K (z p) = G1,2
2 2,2
q
)
q
z p
0
(1
1
2, 2
Functions involving erf (z q p):
1
erf (z p) = G1,1
1,2
)
2q 2
z p
0
1
1
2
Functions involving erfc (z q p):
)
1 2q 2
z p
(
1 2,0
erfc (z p) = G1,2
0, 12
Functions involving erfi (z q p):
zq p
erfi (z p) = G1,1
1,2
q
Functions involving ez
pei
)
2q 2
1 z p
(1
2
:
ez
pei
(
= G1,0
0,1
)
q
z p
Functions involving Ea (z q p):
(
q
Ea (z p) =
G2,0
1,2
a 1, 0
)
a q
z p
Functions involving C (z q p):
1
C (z p) = G1,1
2 1,3
1
1
4
2 4
)
p 4q
z
0, 34 16
Functions involving S (z q p):
1
S (z p) = G1,1
2 1,3
q
5.11. Symbolic Integrals
1
3
4
2 4
)
p 4q
z
0, 14 16
607
SymPy Documentation, Release 0.7.6
Functions involving log (z q p):
n
log (z q p) = generated
(
q
log (z p + a) =
G1,0
1,1
(
log (|z q p a|) = G1,0
1,1
)
)
(
(
1
0,1 1
1,2 1, 1
z log (a) + G1,1
z log (a) + G2,2
0
0
1
q )
z p
0 a
)
)
(
(
)
1 zq p
1
0,1 1
2
z log (|a|) + G1,2 1, 1
z
log
(|a|)
+
G
3,3
1,1
0
0
1 0, 12 a
Functions involving sin (z q p):
sin (z q p) =
1,0
G0,2
(
1
2
2
)
p 2q
z
0 4
Functions involving sinh (z q p):
(
3
sinh (z q p) = 2 G1,0
1,3
1
2
)
1 p2 2q
z
1, 0 4
5.11.4 API reference
static integrals.integrate(f, var, ...)
Compute denite or indenite integral of one or more variables using Risch-Norman
algorithm and table lookup. This procedure is able to handle elementary algebraic
and transcendental functions and also a huge class of special functions, including Airy,
Bessel, Whittaker and Lambert.
var can be:
a symbol indenite integration
a tuple (symbol, a) indenite integration with result given with a replacing
symbol
a tuple (symbol, a, b) denite integration
Several variables can be specied, in which case the result is multiple integration. (If
var is omitted and the integrand is univariate, the indenite integral in that variable will
be performed.)
Indenite integrals are returned without terms that are independent of the integration
variables. (see examples)
Denite improper integrals often entail delicate convergence conditions.
Pass
conds=piecewise, separate or none to have these returned, respectively, as a Piecewise function, as a separate result (i.e. result will be a tuple), or not at all (default is
piecewise).
Strategy
608
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
SymPy uses various approaches to denite integration. One method is to nd an antiderivative for the integrand, and then use the fundamental theorem of calculus. Various
functions are implemented to integrate polynomial, rational and trigonometric functions,
and integrands containing DiracDelta terms.
SymPy also implements the part of the Risch algorithm, which is a decision procedure
for integrating elementary functions, i.e., the algorithm can either nd an elementary
antiderivative, or prove that one does not exist. There is also a (very successful, albeit
somewhat slow) general implementation of the heuristic Risch algorithm. This algorithm
will eventually be phased out as more of the full Risch algorithm is implemented. See
the docstring of Integral. eval integral() for more details on computing the antiderivative
using algebraic methods.
The option risch=True can be used to use only the (full) Risch algorithm. This is useful
if you want to know if an elementary function has an elementary antiderivative. If the
indenite Integral returned by this function is an instance of NonElementaryIntegral,
that means that the Risch algorithm has proven that integral to be non-elementary. Note
that by default, additional methods (such as the Meijer G method outlined below) are
tried on these integrals, as they may be expressible in terms of special functions, so if
you only care about elementary answers, use risch=True. Also note that an unevaluated
Integral returned by this function is not necessarily a NonElementaryIntegral, even with
risch=True, as it may just be an indication that the particular part of the Risch algorithm
needed to integrate that function is not yet implemented.
Another family of strategies comes from re-writing the integrand in terms of so-called
Meijer G-functions. Indenite integrals of a single G-function can always be computed,
and the denite integral of a product of two G-functions can be computed from zero to
innity. Various strategies are implemented to rewrite integrands as G-functions, and
use this information to compute integrals (see the meijerint module).
The option manual=True can be used to use only an algorithm that tries to mimic integration by hand. This algorithm does not handle as many integrands as the other
algorithms implemented but may return results in a more familiar form. The manualintegrate module has functions that return the steps used (see the module docstring for
more information).
In general, the algebraic methods work best for computing antiderivatives of (possibly
complicated) combinations of elementary functions. The G-function methods work best
for computing denite integrals from zero to innity of moderately complicated combinations of special functions, or indenite integrals of very simple combinations of special
functions.
The strategy employed by the integration code is as follows:
If computing a denite integral, and both limits are real, and at least one limit is +oo, try the G-function method of denite integration rst.
Try to nd an antiderivative, using all available methods, ordered by performance
(that is try fastest method rst, slowest last; in particular polynomial integration is
tried rst, meijer g-functions second to last, and heuristic risch last).
If still not successful, try G-functions irrespective of the limits.
The option meijerg=True, False, None can be used to, respectively: always use Gfunction methods and no others, never use G-function methods, or use all available methods (in order as described above). It defaults to None.
See Also:
Integral, Integral.doit
5.11. Symbolic Integrals
609
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import integrate, log, exp, oo
>>> from sympy.abc import a, x, y
>>> integrate(x*y, x)
x**2*y/2
>>> integrate(log(x), x)
x*log(x) - x
>>> integrate(log(x), (x, 1, a))
a*log(a) - a + 1
>>> integrate(x)
x**2/2
Terms that are independent of x are dropped by indenite integration:
>>> from sympy import sqrt
>>> integrate(sqrt(1 + x), (x, 0, x))
2*(x + 1)**(3/2)/3 - 2/3
>>> integrate(sqrt(1 + x), x)
2*(x + 1)**(3/2)/3
>>> integrate(x*y)
Traceback (most recent call last):
...
ValueError: specify integration variables to integrate x*y
Traceback (most recent call last):
...
ValueError: specify integration variables to integrate x*y
Note that integrate(x) syntax is meant only for convenience in interactive sessions and
should be avoided in library code.
>>> integrate(x**a*exp(-x), (x, 0, oo)) # same as conds=piecewise
Piecewise((gamma(a + 1), -re(a) < 1),
(Integral(x**a*exp(-x), (x, 0, oo)), True))
>>> integrate(x**a*exp(-x), (x, 0, oo), conds=none)
gamma(a + 1)
>>> integrate(x**a*exp(-x), (x, 0, oo), conds=separate)
(gamma(a + 1), -re(a) < 1)
static integrals.line integrate(eld, Curve, variables)
Compute the line integral.
See Also:
integrate, Integral
610
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import Curve, line_integrate, E, ln
>>> from sympy.abc import x, y, t
>>> C = Curve([E**t + 1, E**t - 1], (t, 0, ln(2)))
>>> line_integrate(x + y, C, [x, y])
3*sqrt(2)
static integrals.deltaintegrate(f, x)
The idea for integration is the following:
If we are dealing with a DiracDelta expression, i.e. DiracDelta(g(x)), we try to simplify it.
If we could simplify it, then we integrate the resulting expression. We already know
we can integrate a simplied expression, because only simple DiracDelta expressions are involved.
If we couldnt simplify it, there are two cases:
1.The expression is a simple expression: we return the integral, taking care if we
are dealing with a Derivative or with a proper DiracDelta.
2.The expression is not simple (i.e. DiracDelta(cos(x))): we can do nothing at all.
If the node is a multiplication node having a DiracDelta term:
First we expand it.
If the expansion did work, the we try to integrate the expansion.
If not, we try to extract a simple DiracDelta term, then we have two cases:
1.We have a simple DiracDelta term, so we return the integral.
2.We didnt have a simple term, but we do have an expression with simplied
DiracDelta terms, so we integrate this expression.
See Also:
sympy.functions.special.delta functions.DiracDelta
sympy.integrals.integrals.Integral
(page
377),
Examples
>>> from sympy.abc import x, y, z
>>> from sympy.integrals.deltafunctions import deltaintegrate
>>> from sympy import sin, cos, DiracDelta, Heaviside
>>> deltaintegrate(x*sin(x)*cos(x)*DiracDelta(x - 1), x)
sin(1)*cos(1)*Heaviside(x - 1)
>>> deltaintegrate(y**2*DiracDelta(x - z)*DiracDelta(y - z), y)
z**2*DiracDelta(x - z)*Heaviside(y - z)
static integrals.ratint(f, x, **ags)
Performs indenite integration of rational functions.
Given a eld K and a rational function f = p/q, where p and q are polynomials in K[x],
returns a function g such that f = g 0 .
>>> from sympy.integrals.rationaltools import ratint
>>> from sympy.abc import x
5.11. Symbolic Integrals
611
SymPy Documentation, Release 0.7.6
>>> ratint(36/(x**5 - 2*x**4 - 2*x**3 + 4*x**2 + x - 2), x)
(12*x + 6)/(x**2 - 1) + 4*log(x - 2) - 4*log(x + 1)
See Also:
sympy.integrals.integrals.Integral.doit, ratint logpart, ratint ratpart
References
[Bro05] (page 1909)
static integrals.heurisch(f,
x,
rewrite=False,
hints=None,
mappings=None,
retries=3,
degree oset=0,
unnecessary permutations=None)
Compute indenite integral using heuristic Risch algorithm.
This is a heuristic approach to indenite integration in nite terms using the extended
heuristic (parallel) Risch algorithm, based on Manuel Bronsteins Poor Mans Integrator.
The algorithm supports various classes of functions including transcendental elementary
or special functions like Airy, Bessel, Whittaker and Lambert.
Note that this algorithm is not a decision procedure. If it isnt able to compute the
antiderivative for a given function, then this is not a proof that such a functions does not
exist. One should use recursive Risch algorithm in such case. Its an open question if
this algorithm can be made a full decision procedure.
This is an internal integrator procedure. You should use toplevel integrate function in
most cases, as this procedure needs some preprocessing steps and otherwise may fail.
See Also:
sympy.integrals.integrals.Integral.doit, sympy.integrals.integrals.Integral,
components
Examples
>>> from sympy import tan
>>> from sympy.integrals.heurisch import heurisch
>>> from sympy.abc import x, y
>>> heurisch(y*tan(x), x)
y*log(tan(x)**2 + 1)/2
See Manuel Bronsteins Poor Mans Integrator:
[1] http://www-sop.inria.fr/cafe/Manuel.Bronstein/pmint/index.html
For more information on the implemented algorithm refer to:
[2] K. Geddes, L. Stefanus, On the Risch-Norman Integration Method and its Implementation in Maple, Proceedings of ISSAC89, ACM Press, 212-217.
[3] J. H. Davenport, On the Parallel Risch Algorithm (I), Proceedings of EUROCAM82, LNCS 144, Springer, 144-157.
[4] J. H. Davenport, On the Parallel Risch Algorithm (III): Use
SIGSAM Bulletin 16 (1982), 3-6.
612
of
Tangents,
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
[5] J. H. Davenport, B. M. Trager, On the Parallel Risch Algorithm
Transactions on Mathematical Software 11 (1985), 356-362.
(II),
ACM
Specication
heurisch(f, x, rewrite=False, hints=None)
where f : expression x : symbol
rewrite -> force rewrite f in terms of tan and tanh hints -> a list of functions that may appear in anti-derivate
hints = None > no suggestions at all
hints = [ ] > try to gure out
hints = [f1, ..., fn] > we know better
static integrals.heurisch wrapper(f,
x,
rewrite=False,
hints=None,
mappings=None, retries=3, degree oset=0, unnecessary permutations=None)
A wrapper around the heurisch integration algorithm.
This method takes the result from heurisch and checks for poles in the denominator. For
each of these poles, the integral is reevaluated, and the nal integration result is given
in terms of a Piecewise.
See Also:
heurisch
Examples
>>> from sympy.core import symbols
>>> from sympy.functions import cos
>>> from sympy.integrals.heurisch import heurisch, heurisch_wrapper
>>> n, x = symbols(n x)
>>> heurisch(cos(n*x), x)
sin(n*x)/n
>>> heurisch_wrapper(cos(n*x), x)
Piecewise((x, n == 0), (sin(n*x)/n, True))
static integrals.trigintegrate(f, x, conds=piecewise)
Integrate f = Mul(trig) over x
>>> from sympy import Symbol, sin, cos, tan, sec, csc, cot
>>> from sympy.integrals.trigonometry import trigintegrate
>>> from sympy.abc import x
>>> trigintegrate(sin(x)*cos(x), x)
sin(x)**2/2
>>> trigintegrate(sin(x)**2, x)
x/2 - sin(x)*cos(x)/2
>>> trigintegrate(tan(x)*sec(x), x)
1/cos(x)
5.11. Symbolic Integrals
613
SymPy Documentation, Release 0.7.6
>>> trigintegrate(sin(x)*tan(x), x)
-log(sin(x) - 1)/2 + log(sin(x) + 1)/2 - sin(x)
http://en.wikibooks.org/wiki/Calculus/Integration techniques
See Also:
sympy.integrals.integrals.Integral.doit, sympy.integrals.integrals.Integral
static integrals.manualintegrate(f, var)
Compute indenite integral of a single variable using an algorithm that resembles what
a student would do by hand.
Unlike integrate, var can only be a single symbol.
See Also:
sympy.integrals.integrals.integrate, sympy.integrals.integrals.Integral.doit,
sympy.integrals.integrals.Integral
Examples
>>> from sympy import sin, cos, tan, exp, log, integrate
>>> from sympy.integrals.manualintegrate import manualintegrate
>>> from sympy.abc import x
>>> manualintegrate(1 / x, x)
log(x)
>>> integrate(1/x)
log(x)
>>> manualintegrate(log(x), x)
x*log(x) - x
>>> integrate(log(x))
x*log(x) - x
>>> manualintegrate(exp(x) / (1 + exp(2 * x)), x)
atan(exp(x))
>>> integrate(exp(x) / (1 + exp(2 * x)))
RootSum(4*_z**2 + 1, Lambda(_i, _i*log(2*_i + exp(x))))
>>> manualintegrate(cos(x)**4 * sin(x), x)
-cos(x)**5/5
>>> integrate(cos(x)**4 * sin(x), x)
-cos(x)**5/5
>>> manualintegrate(cos(x)**4 * sin(x)**3, x)
cos(x)**7/7 - cos(x)**5/5
>>> integrate(cos(x)**4 * sin(x)**3, x)
cos(x)**7/7 - cos(x)**5/5
>>> manualintegrate(tan(x), x)
-log(cos(x))
>>> integrate(tan(x), x)
-log(sin(x)**2 - 1)/2
static manualintegrate.integral steps(integrand, symbol, **options)
Returns the steps needed to compute an integral.
This function attempts to mirror what a student would do by hand as closely as possible.
SymPy Gamma uses this to provide a step-by-step explanation of an integral.
The code it uses to format the results of this function can be found at
https://github.com/sympy/sympy gamma/blob/master/app/logic/intsteps.py.
Returns rule : namedtuple
614
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
The rst step; most rules have substeps that must also be considered.
These substeps can be evaluated using manualintegrate to obtain a
result.
Examples
>>> from sympy import exp, sin, cos
>>> from sympy.integrals.manualintegrate import integral_steps
>>> from sympy.abc import x
>>> print(repr(integral_steps(exp(x) / (1 + exp(2 * x)), x)))
URule(u_var=_u, u_func=exp(x), constant=1,
substep=ArctanRule(context=1/(_u**2 + 1), symbol=_u),
context=exp(x)/(exp(2*x) + 1), symbol=x)
>>> print(repr(integral_steps(sin(x), x)))
TrigRule(func=sin, arg=x, context=sin(x), symbol=x)
>>> print(repr(integral_steps((x**2 + 3)**2 , x)))
RewriteRule(rewritten=x**4 + 6*x**2 + 9,
substep=AddRule(substeps=[PowerRule(base=x, exp=4, context=x**4, symbol=x),
ConstantTimesRule(constant=6, other=x**2,
substep=PowerRule(base=x, exp=2, context=x**2, symbol=x),
context=6*x**2, symbol=x),
ConstantRule(constant=9, context=9, symbol=x)],
context=x**4 + 6*x**2 + 9, symbol=x), context=(x**2 + 3)**2, symbol=x)
The class Integral represents an unevaluated integral and has some methods that help in the
integration of an expression.
class sympy.integrals.Integral
Represents unevaluated integral.
is commutative
Returns whether all the free symbols in the integral are commutative.
as sum(n, method=midpoint)
Approximates the denite integral by a sum.
method ... one of: left, right, midpoint, trapezoid
These are all basically the rectangle method [1], the only dierence is where the
function value is taken in each interval to dene the rectangle.
[1] http://en.wikipedia.org/wiki/Rectangle method
See Also:
Integral.doit Perform the integration using any hints
Examples
>>> from sympy import sin, sqrt
>>> from sympy.abc import x
>>> from sympy.integrals import Integral
>>> e = Integral(sin(x), (x, 3, 7))
>>> e
Integral(sin(x), (x, 3, 7))
For demonstration purposes, this interval will only be split into 2 regions, bounded
by [3, 5] and [5, 7].
5.11. Symbolic Integrals
615
SymPy Documentation, Release 0.7.6
The left-hand rule uses function evaluations at the left of each interval:
>>> e.as_sum(2, left)
2*sin(5) + 2*sin(3)
The midpoint rule uses evaluations at the center of each interval:
>>> e.as_sum(2, midpoint)
2*sin(4) + 2*sin(6)
The right-hand rule uses function evaluations at the right of each interval:
>>> e.as_sum(2, right)
2*sin(5) + 2*sin(7)
The trapezoid rule uses function evaluations on both sides of the intervals. This is
equivalent to taking the average of the left and right hand rule results:
>>> e.as_sum(2, trapezoid)
2*sin(5) + sin(3) + sin(7)
>>> (e.as_sum(2, left) + e.as_sum(2, right))/2 == _
True
All but the trapexoid method may be used when dealing with a function with a discontinuity. Here, the discontinuity at x = 0 can be avoided by using the midpoint or
right-hand method:
>>> e = Integral(1/sqrt(x), (x, 0, 1))
>>> e.as_sum(5).n(4)
1.730
>>> e.as_sum(10).n(4)
1.809
>>> e.doit().n(4) # the actual value is 2
2.000
The left- or trapezoid method will encounter the discontinuity and return oo:
>>> e.as_sum(5, left)
oo
>>> e.as_sum(5, trapezoid)
oo
doit(**hints)
Perform the integration using any hints given.
See Also:
sympy.integrals.trigonometry.trigintegrate, sympy.integrals.risch.heurisch,
sympy.integrals.rationaltools.ratint
as sum Approximate the integral using a sum
Examples
>>> from sympy import Integral
>>> from sympy.abc import x, i
>>> Integral(x**i, (i, 1, 3)).doit()
Piecewise((2, log(x) == 0), (x**3/log(x) - x/log(x), True))
616
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
free symbols
This method returns the symbols that will exist when the integral is evaluated. This
is useful if one is trying to determine whether an integral depends on a certain
symbol or not.
See Also:
function, limits, variables
Examples
>>> from sympy import Integral
>>> from sympy.abc import x, y
>>> Integral(x, (x, y, 1)).free_symbols
set([y])
transform(x, u)
Performs a change of variables from x to u using the relationship given by x and u
which will dene the transformations f and F (which are inverses of each other) as
follows:
1.If x is a Symbol (which is a variable of integration) then u will be interpreted as
some function, f(u), with inverse F(u). This, in eect, just makes the substitution
of x with f(x).
2.If u is a Symbol then x will be interpreted as some function, F(x), with inverse
f(u). This is commonly referred to as u-substitution.
Once f and F have been identied, the transformation is made as follows:
F (b)
xdx
a
f (x)
F (a)
d
dx
where F (x) is the inverse of f (x) and the limits and integrand have been corrected
so as to retain the same value after integration.
See Also:
variables Lists the integration variables
as dummy Replace integration variables with dummy ones
Notes
The mappings, F(x) or f(u), must lead to a unique integral. Linear or rational linear
expression, 2 x, 1/x and sqrt(x), will always work; quadratic expressions like x 2 1
are acceptable as long as the resulting integrand does not depend on the sign of the
solutions (see examples).
The integral will be returned unchanged if x is not a variable of integration.
x must be (or contain) only one of of the integration variables. If u has more than
one free symbol then it should be sent as a tuple (u, uvar) where uvar identies which
variable is replacing the integration variable. XXX can it contain another integration
variable?
5.11. Symbolic Integrals
617
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.abc import a, b, c, d, x, u, y
>>> from sympy import Integral, S, cos, sqrt
>>> i = Integral(x*cos(x**2 - 1), (x, 0, 1))
transform can change the variable of integration
>>> i.transform(x, u)
Integral(u*cos(u**2 - 1), (u, 0, 1))
transform can perform u-substitution as long as a unique integrand is obtained:
>>> i.transform(x**2 - 1, u)
Integral(cos(u)/2, (u, -1, 0))
This attempt fails because x = +/-sqrt(u + 1) and the sign does not cancel out of the
integrand:
>>> Integral(cos(x**2 - 1), (x, 0, 1)).transform(x**2 - 1, u)
Traceback (most recent call last):
...
ValueError:
The mapping between F(x) and f(u) did not give a unique integrand.
Traceback (most recent call last):
...
ValueError:
transform can do a substitution. Here, the previous result is transformed back into
the original expression using u-substitution:
>>> ui = _
>>> _.transform(sqrt(u + 1), x) == i
True
We can accomplish the same with a regular substitution:
>>> ui.transform(u, x**2 - 1) == i
True
If the x does not contain a symbol of integration then the integral will be returned
unchanged. Integral i does not have an integration variable a so no change is made:
>>> i.transform(a, x) == i
True
When u has more than one free symbol the symbol that is replacing x must be identied by passing u as a tuple:
>>> Integral(x,
Integral(a + u,
>>> Integral(x,
Integral(a + u,
618
(x,
(u,
(x,
(a,
0, 1)).transform(x, (u + a, u))
-a, -a + 1))
0, 1)).transform(x, (u + a, a))
-u, -u + 1))
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
5.11.5 TODO and Bugs
There are still lots of functions that SymPy does not know how to integrate. For bugs related
to this module, see https://github.com/sympy/sympy/issues?labels=Integration
5.12 Numeric Integrals
SymPy has functions to calculate points and weights for Gaussian quadrature of any order
and any precision:
sympy.integrals.quadrature.gauss legendre(n, n digits)
Computes the Gauss-Legendre quadrature [R291] (page 1909) points and weights.
The Gauss-Legendre quadrature approximates the integral:
1
1
f (x) dx
wi f (xi )
i=1
The nodes xi of an order n quadrature rule are the roots of Pn and the weights wi are
given by:
wi =
2
2
(1 x2i ) (Pn0 (xi ))
Parameters n : the order of quadrature
n digits : number of signicant digits of the points and weights to return
Returns (x, w) : the x and w are lists of points and weights as Floats.
The points xi and weights wi are returned as (x, w) tuple of lists.
See Also:
gauss laguerre,
gauss gen laguerre,
gauss chebyshev u, gauss jacobi
gauss hermite,
gauss chebyshev t,
References
[R291] (page 1909), [R292] (page 1909)
Examples
>>> from sympy.integrals.quadrature import gauss_legendre
>>> x, w = gauss_legendre(3, 5)
>>> x
[-0.7746, 0, 0.7746]
>>> w
[0.55556, 0.88889, 0.55556]
>>> x, w = gauss_legendre(4, 5)
>>> x
[-0.86114, -0.33998, 0.33998, 0.86114]
5.12. Numeric Integrals
619
SymPy Documentation, Release 0.7.6
>>> w
[0.34786, 0.65215, 0.65215, 0.34786]
sympy.integrals.quadrature.gauss laguerre(n, n digits)
Computes the Gauss-Laguerre quadrature [R293] (page 1909) points and weights.
The Gauss-Laguerre quadrature approximates the integral:
n
ex f (x) dx
wi f (xi )
0
i=1
The nodes xi of an order n quadrature rule are the roots of Ln and the weights wi are
given by:
xi
wi =
2
(n + 1)2 (Ln+1 (xi ))
Parameters n : the order of quadrature
n digits : number of signicant digits of the points and weights to return
Returns (x, w) : the x and w are lists of points and weights as Floats.
The points xi and weights wi are returned as (x, w) tuple of lists.
See Also:
gauss legendre,
gauss gen laguerre,
gauss chebyshev u, gauss jacobi
gauss hermite,
gauss chebyshev t,
References
[R293] (page 1909), [R294] (page 1909)
Examples
>>> from sympy.integrals.quadrature import gauss_laguerre
>>> x, w = gauss_laguerre(3, 5)
>>> x
[0.41577, 2.2943, 6.2899]
>>> w
[0.71109, 0.27852, 0.010389]
>>> x, w = gauss_laguerre(6, 5)
>>> x
[0.22285, 1.1889, 2.9927, 5.7751, 9.8375, 15.983]
>>> w
[0.45896, 0.417, 0.11337, 0.010399, 0.00026102, 8.9855e-7]
sympy.integrals.quadrature.gauss hermite(n, n digits)
Computes the Gauss-Hermite quadrature [R295] (page 1909) points and weights.
The Gauss-Hermite quadrature approximates the integral:
n
2
ex f (x) dx
wi f (xi )
620
i=1
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
The nodes xi of an order n quadrature rule are the roots of Hn and the weights wi are
given by:
2n1 n!
wi =
2
n2 (Hn1 (xi ))
Parameters n : the order of quadrature
n digits : number of signicant digits of the points and weights to return
Returns (x, w) : the x and w are lists of points and weights as Floats.
The points xi and weights wi are returned as (x, w) tuple of lists.
See Also:
gauss laguerre,
gauss legendre,
gauss chebyshev u, gauss jacobi
gauss gen laguerre,
gauss chebyshev t,
References
[R295] (page 1909), [R296] (page 1909), [R297] (page 1909)
Examples
>>> from sympy.integrals.quadrature import gauss_hermite
>>> x, w = gauss_hermite(3, 5)
>>> x
[-1.2247, 0, 1.2247]
>>> w
[0.29541, 1.1816, 0.29541]
>>> x, w = gauss_hermite(6, 5)
>>> x
[-2.3506, -1.3358, -0.43608, 0.43608, 1.3358, 2.3506]
>>> w
[0.00453, 0.15707, 0.72463, 0.72463, 0.15707, 0.00453]
sympy.integrals.quadrature.gauss gen laguerre(n, alpha, n digits)
Computes the generalized Gauss-Laguerre quadrature [R298] (page 1909) points and
weights.
The generalized Gauss-Laguerre quadrature approximates the integral:
x e
0
f (x) dx
wi f (xi )
i=1
The nodes xi of an order n quadrature rule are the roots of L
n and the weights wi are
given by:
wi =
5.12. Numeric Integrals
( + n)
+1
n(n)L
n1 (xi )Ln1 (xi )
621
SymPy Documentation, Release 0.7.6
Parameters n : the order of quadrature
alpha : the exponent of the singularity, > 1
n digits : number of signicant digits of the points and weights to return
Returns (x, w) : the x and w are lists of points and weights as Floats.
The points xi and weights wi are returned as (x, w) tuple of lists.
See Also:
gauss laguerre,
gauss legendre,
gauss chebyshev u, gauss jacobi
gauss hermite,
gauss chebyshev t,
References
[R298] (page 1909), [R299] (page 1909)
Examples
>>> from sympy import S
>>> from sympy.integrals.quadrature import gauss_gen_laguerre
>>> x, w = gauss_gen_laguerre(3, -S.Half, 5)
>>> x
[0.19016, 1.7845, 5.5253]
>>> w
[1.4493, 0.31413, 0.00906]
>>> x, w = gauss_gen_laguerre(4, 3*S.Half, 5)
>>> x
[0.97851, 2.9904, 6.3193, 11.712]
>>> w
[0.53087, 0.67721, 0.11895, 0.0023152]
sympy.integrals.quadrature.gauss chebyshev t(n, n digits)
Computes the Gauss-Chebyshev quadrature [R300] (page 1909) points and weights of
the rst kind.
The Gauss-Chebyshev quadrature of the rst kind approximates the integral:
1
f (x) dx
wi f (xi )
1 x2
i=1
n
The nodes xi of an order n quadrature rule are the roots of Tn and the weights wi are
given by:
wi =
Parameters n : the order of quadrature
n digits : number of signicant digits of the points and weights to return
Returns (x, w) : the x and w are lists of points and weights as Floats.
The points xi and weights wi are returned as (x, w) tuple of lists.
622
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
See Also:
gauss legendre,
gauss laguerre,
gauss chebyshev u, gauss jacobi
gauss hermite,
gauss gen laguerre,
References
[R300] (page 1909), [R301] (page 1909)
Examples
>>> from sympy import S
>>> from sympy.integrals.quadrature import gauss_chebyshev_t
>>> x, w = gauss_chebyshev_t(3, 5)
>>> x
[0.86602, 0, -0.86602]
>>> w
[1.0472, 1.0472, 1.0472]
>>> x, w = gauss_chebyshev_t(6, 5)
>>> x
[0.96593, 0.70711, 0.25882, -0.25882, -0.70711, -0.96593]
>>> w
[0.5236, 0.5236, 0.5236, 0.5236, 0.5236, 0.5236]
sympy.integrals.quadrature.gauss chebyshev u(n, n digits)
Computes the Gauss-Chebyshev quadrature [R302] (page 1909) points and weights of
the second kind.
The Gauss-Chebyshev quadrature of the second kind approximates the integral:
1
1
1 x2 f (x) dx
wi f (xi )
i=1
The nodes xi of an order n quadrature rule are the roots of Un and the weights wi are
given by:
(
)
i
2
sin
wi =
n+1
n+1
Parameters n : the order of quadrature
n digits : number of signicant digits of the points and weights to return
Returns (x, w) : the x and w are lists of points and weights as Floats.
The points xi and weights wi are returned as (x, w) tuple of lists.
See Also:
gauss legendre,
gauss laguerre,
gauss chebyshev t, gauss jacobi
5.12. Numeric Integrals
gauss hermite,
gauss gen laguerre,
623
SymPy Documentation, Release 0.7.6
References
[R302] (page 1909), [R303] (page 1909)
Examples
>>> from sympy import S
>>> from sympy.integrals.quadrature import gauss_chebyshev_u
>>> x, w = gauss_chebyshev_u(3, 5)
>>> x
[0.70711, 0, -0.70711]
>>> w
[0.3927, 0.7854, 0.3927]
>>> x, w = gauss_chebyshev_u(6, 5)
>>> x
[0.90097, 0.62349, 0.22252, -0.22252, -0.62349, -0.90097]
>>> w
[0.084489, 0.27433, 0.42658, 0.42658, 0.27433, 0.084489]
sympy.integrals.quadrature.gauss jacobi(n, alpha, beta, n digits)
Computes the Gauss-Jacobi quadrature [R304] (page 1909) points and weights.
The Gauss-Jacobi quadrature of the rst kind approximates the integral:
(1 x) (1 + x) f (x) dx
wi f (xi )
i=1
(,)
The nodes xi of an order n quadrature rule are the roots of Pn
given by:
wi =
and the weights wi are
2n + + + 2 (n + + 1)(n + + 1)
2+
n + + + 1 (n + + + 1)(n + 1)! Pn0 (xi )P (,) (xi )
n+1
Parameters n : the order of quadrature
alpha : the rst parameter of the Jacobi Polynomial, > 1
beta : the second parameter of the Jacobi Polynomial, > 1
n digits : number of signicant digits of the points and weights to return
Returns (x, w) : the x and w are lists of points and weights as Floats.
The points xi and weights wi are returned as (x, w) tuple of lists.
See Also:
gauss legendre,
gauss laguerre,
gauss chebyshev t, gauss chebyshev u
gauss hermite,
gauss gen laguerre,
References
[R304] (page 1909), [R305] (page 1909), [R306] (page 1909)
624
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import S
>>> from sympy.integrals.quadrature import gauss_jacobi
>>> x, w = gauss_jacobi(3, S.Half, -S.Half, 5)
>>> x
[-0.90097, -0.22252, 0.62349]
>>> w
[1.7063, 1.0973, 0.33795]
>>> x, w = gauss_jacobi(6, 1, 1, 5)
>>> x
[-0.87174, -0.5917, -0.2093, 0.2093, 0.5917, 0.87174]
>>> w
[0.050584, 0.22169, 0.39439, 0.39439, 0.22169, 0.050584]
5.13 Logic Module
5.13.1 Introduction
The logic module for SymPy allows to form and manipulate logic expressions using symbolic
and Boolean values.
5.13.2 Forming logical expressions
You can build Boolean expressions with the standard python operators & (And), | (Or), (Not):
>>> from sympy import *
>>> x, y = symbols(x,y)
>>> y | (x & y)
Or(And(x, y), y)
>>> x | y
Or(x, y)
>>> ~x
Not(x)
You can also form implications with >> and <<:
>>> x >> y
Implies(x, y)
>>> x << y
Implies(y, x)
Like most types in SymPy, Boolean expressions inherit from Basic:
>>> (y & x).subs({x: True, y: True})
True
>>> (x | y).atoms()
set([x, y])
The logic module also includes the following functions to derive boolean expressions from
their truth tables-
5.13. Logic Module
625
SymPy Documentation, Release 0.7.6
sympy.logic.boolalg.SOPform(variables, minterms, dontcares=None)
The SOPform function uses simplied pairs and a redundant group- eliminating algorithm to convert the list of all input combos that generate 1 (the minterms) into the
smallest Sum of Products form.
The variables must be given as the rst argument.
Return a logical Or function (i.e., the sum of products or SOP form) that gives the
desired outcome. If there are inputs that can be ignored, pass them as a list, too.
The result will be one of the (perhaps many) functions that satisfy the conditions.
References
[R307] (page 1909)
Examples
>>> from sympy.logic import SOPform
>>> minterms = [[0, 0, 0, 1], [0, 0, 1, 1],
...
[0, 1, 1, 1], [1, 0, 1, 1], [1, 1, 1, 1]]
>>> dontcares = [[0, 0, 0, 0], [0, 0, 1, 0], [0, 1, 0, 1]]
>>> SOPform([w,x,y,z], minterms, dontcares)
Or(And(Not(w), z), And(y, z))
sympy.logic.boolalg.POSform(variables, minterms, dontcares=None)
The POSform function uses simplied pairs and a redundant-group eliminating algorithm to convert the list of all input combinations that generate 1 (the minterms) into
the smallest Product of Sums form.
The variables must be given as the rst argument.
Return a logical And function (i.e., the product of sums or POS form) that gives the
desired outcome. If there are inputs that can be ignored, pass them as a list, too.
The result will be one of the (perhaps many) functions that satisfy the conditions.
References
[R308] (page 1909)
Examples
>>> from sympy.logic import POSform
>>> minterms = [[0, 0, 0, 1], [0, 0, 1, 1], [0, 1, 1, 1],
...
[1, 0, 1, 1], [1, 1, 1, 1]]
>>> dontcares = [[0, 0, 0, 0], [0, 0, 1, 0], [0, 1, 0, 1]]
>>> POSform([w,x,y,z], minterms, dontcares)
And(Or(Not(w), y), z)
626
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
5.13.3 Boolean functions
class sympy.logic.boolalg.BooleanTrue
SymPy version of True.
The instances of this class are singletonized and can be accessed via S.true.
This is the SymPy version of True, for use in the logic module. The primary advantage of
using true instead of True is that shorthand boolean operations like and >> will work
as expected on this class, whereas with True they act bitwise on 1. Functions in the logic
module will return this class when they evaluate to true.
See Also:
sympy.logic.boolalg.BooleanFalse (page 627)
Examples
>>> from sympy import sympify, true, Or
>>> sympify(True)
True
>>> ~true
False
>>> ~True
-2
>>> Or(True, False)
True
class sympy.logic.boolalg.BooleanFalse
SymPy version of False.
The instances of this class are singletonized and can be accessed via S.false.
This is the SymPy version of False, for use in the logic module. The primary advantage
of using false instead of False is that shorthand boolean operations like and >> will
work as expected on this class, whereas with False they act bitwise on 0. Functions in
the logic module will return this class when they evaluate to false.
See Also:
sympy.logic.boolalg.BooleanTrue (page 627)
Examples
>>> from sympy import sympify, false, Or, true
>>> sympify(False)
False
>>> false >> false
True
>>> False >> False
0
>>> Or(True, False)
True
class sympy.logic.boolalg.And
Logical AND function.
It evaluates its arguments in order, giving False immediately if any of them are False,
and True if they are all True.
5.13. Logic Module
627
SymPy Documentation, Release 0.7.6
Notes
The & operator is provided as a convenience, but note that its use here is dierent from
its normal use in Python, which is bitwise and. Hence, And(a, b) and a & b will return
dierent things if a and b are integers.
>>> And(x, y).subs(x, 1)
y
Examples
>>> from sympy.core import symbols
>>> from sympy.abc import x, y
>>> from sympy.logic.boolalg import And
>>> x & y
And(x, y)
class sympy.logic.boolalg.Or
Logical OR function
It evaluates its arguments in order, giving True immediately if any of them are True, and
False if they are all False.
Notes
The | operator is provided as a convenience, but note that its use here is dierent from
its normal use in Python, which is bitwise or. Hence, Or(a, b) and a | b will return
dierent things if a and b are integers.
>>> Or(x, y).subs(x, 0)
y
Examples
>>> from sympy.core import symbols
>>> from sympy.abc import x, y
>>> from sympy.logic.boolalg import Or
>>> x | y
Or(x, y)
class sympy.logic.boolalg.Not
Logical Not function (negation)
Returns True if the statement is False Returns False if the statement is True
Notes
The operator is provided as a convenience, but note that its use here is dierent
from its normal use in Python, which is bitwise not. In particular, a and Not(a) will
be dierent if a is an integer. Furthermore, since bools in Python subclass from int,
True is the same as 1 which is -2, which has a boolean value of True. To avoid this
issue, use the SymPy boolean types true and false.
628
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> from sympy import true
>>> ~True
-2
>>> ~true
False
Examples
>>> from sympy.logic.boolalg import Not, And, Or
>>> from sympy.abc import x, A, B
>>> Not(True)
False
>>> Not(False)
True
>>> Not(And(True, False))
True
>>> Not(Or(True, False))
False
>>> Not(And(And(True, x), Or(x, False)))
Not(x)
>>> ~x
Not(x)
>>> Not(And(Or(A, B), Or(~A, ~B)))
Not(And(Or(A, B), Or(Not(A), Not(B))))
class sympy.logic.boolalg.Xor
Logical XOR (exclusive OR) function.
Returns True if an odd number of the arguments are True and the rest are False.
Returns False if an even number of the arguments are True and the rest are False.
Notes
The operator is provided as a convenience, but note that its use here is dierent from
its normal use in Python, which is bitwise xor. In particular, a b and Xor(a, b) will
be dierent if a and b are integers.
>>> Xor(x, y).subs(y, 0)
x
Examples
>>> from sympy.logic.boolalg import Xor
>>> from sympy import symbols
>>> x, y = symbols(x y)
>>> Xor(True, False)
True
>>> Xor(True, True)
False
>>> Xor(True, False, True, True, False)
True
>>> Xor(True, False, True, False)
False
5.13. Logic Module
629
SymPy Documentation, Release 0.7.6
>>> x ^ y
Xor(x, y)
class sympy.logic.boolalg.Nand
Logical NAND function.
It evaluates its arguments in order, giving True immediately if any of them are False, and
False if they are all True.
Returns True if any of the arguments are False Returns False if all arguments are True
Examples
>>> from sympy.logic.boolalg import Nand
>>> from sympy import symbols
>>> x, y = symbols(x y)
>>> Nand(False, True)
True
>>> Nand(True, True)
False
>>> Nand(x, y)
Not(And(x, y))
class sympy.logic.boolalg.Nor
Logical NOR function.
It evaluates its arguments in order, giving False immediately if any of them are True, and
True if they are all False.
Returns False if any argument is True Returns True if all arguments are False
Examples
>>> from sympy.logic.boolalg import Nor
>>> from sympy import symbols
>>> x, y = symbols(x y)
>>> Nor(True, False)
False
>>> Nor(True, True)
False
>>> Nor(False, True)
False
>>> Nor(False, False)
True
>>> Nor(x, y)
Not(Or(x, y))
class sympy.logic.boolalg.Implies
Logical implication.
A implies B is equivalent to !A v B
Accepts two Boolean arguments; A and B. Returns False if A is True and B is False Returns
True otherwise.
630
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Notes
The >> and << operators are provided as a convenience, but note that their use here is
dierent from their normal use in Python, which is bit shifts. Hence, Implies(a, b) and
a >> b will return dierent things if a and b are integers. In particular, since Python
considers True and False to be integers, True >> True will be the same as 1 >> 1,
i.e., 0, which has a truth value of False. To avoid this issue, use the SymPy objects true
and false.
>>> from sympy import true, false
>>> True >> False
1
>>> true >> false
False
Examples
>>> from sympy.logic.boolalg import Implies
>>> from sympy import symbols
>>> x, y = symbols(x y)
>>> Implies(True, False)
False
>>> Implies(False, False)
True
>>> Implies(True, True)
True
>>> Implies(False, True)
True
>>> x >> y
Implies(x, y)
>>> y << x
Implies(x, y)
class sympy.logic.boolalg.Equivalent
Equivalence relation.
Equivalent(A, B) is True i A and B are both True or both False
Returns True if all of the arguments are logically equivalent. Returns False otherwise.
Examples
>>> from sympy.logic.boolalg import Equivalent, And
>>> from sympy.abc import x, y
>>> Equivalent(False, False, False)
True
>>> Equivalent(True, False, False)
False
>>> Equivalent(x, And(x, True))
True
class sympy.logic.boolalg.ITE
If then else clause.
5.13. Logic Module
631
SymPy Documentation, Release 0.7.6
ITE(A, B, C) evaluates and returns the result of B if A is true else it returns the result of
C
Examples
>>> from sympy.logic.boolalg import ITE, And, Xor, Or
>>> from sympy.abc import x, y, z
>>> ITE(True, False, True)
False
>>> ITE(Or(True, False), And(True, True), Xor(True, True))
True
>>> ITE(x, y, z)
ITE(x, y, z)
>>> ITE(True, x, y)
x
>>> ITE(False, x, y)
y
>>> ITE(x, y, y)
y
The following functions can be used to handle Conjunctive and Disjunctive Normal formssympy.logic.boolalg.to cnf(expr, simplify=False)
Convert a propositional logical sentence s to conjunctive normal form. That is, of the
form ((A | B | ...) & (B | C | ...) & ...) If simplify is True, the expr is evaluated to its
simplest CNF form.
Examples
>>> from sympy.logic.boolalg import to_cnf
>>> from sympy.abc import A, B, D
>>> to_cnf(~(A | B) | D)
And(Or(D, Not(A)), Or(D, Not(B)))
>>> to_cnf((A | B) & (A | ~A), True)
Or(A, B)
sympy.logic.boolalg.to dnf(expr, simplify=False)
Convert a propositional logical sentence s to disjunctive normal form. That is, of the
form ((A & B & ...) | (B & C & ...) | ...) If simplify is True, the expr is evaluated to its
simplest DNF form.
Examples
>>> from sympy.logic.boolalg import to_dnf
>>> from sympy.abc import A, B, C
>>> to_dnf(B & (A | C))
Or(And(A, B), And(B, C))
>>> to_dnf((A & B) | (A & ~B) | (B & C) | (~B & C), True)
Or(A, C)
sympy.logic.boolalg.is cnf(expr)
Test whether or not an expression is in conjunctive normal form.
632
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.logic.boolalg import is_cnf
>>> from sympy.abc import A, B, C
>>> is_cnf(A | B | C)
True
>>> is_cnf(A & B & C)
True
>>> is_cnf((A & B) | C)
False
sympy.logic.boolalg.is dnf(expr)
Test whether or not an expression is in disjunctive normal form.
Examples
>>> from sympy.logic.boolalg import is_dnf
>>> from sympy.abc import A, B, C
>>> is_dnf(A | B | C)
True
>>> is_dnf(A & B & C)
True
>>> is_dnf((A & B) | C)
True
>>> is_dnf(A & (B | C))
False
5.13.4 Simplication and equivalence-testing
sympy.logic.boolalg.simplify logic(expr, form=None, deep=True)
This function simplies a boolean function to its simplied version in SOP or POS form.
The return type is an Or or And object in SymPy.
Parameters expr : string or boolean expression
form : string (cnf or dnf) or None (default).
If cnf or dnf, the simplest expression in the corresponding normal
form is returned; if None, the answer is returned according to the
form with fewest args (in CNF by default).
deep : boolean (default True)
indicates whether to recursively simplify any non-boolean functions
contained within the input.
Examples
>>> from sympy.logic import simplify_logic
>>> from sympy.abc import x, y, z
>>> from sympy import S
>>> b = (~x & ~y & ~z) | ( ~x & ~y & z)
>>> simplify_logic(b)
And(Not(x), Not(y))
5.13. Logic Module
633
SymPy Documentation, Release 0.7.6
>>> S(b)
Or(And(Not(x), Not(y), Not(z)), And(Not(x), Not(y), z))
>>> simplify_logic(_)
And(Not(x), Not(y))
SymPys simplify() function can also be used to simplify logic expressions to their simplest
forms.
sympy.logic.boolalg.bool map(bool1, bool2)
Return the simplied version of bool1, and the mapping of variables that makes the two
expressions bool1 and bool2 represent the same logical behaviour for some correspondence between the variables of each. If more than one mappings of this sort exist, one
of them is returned. For example, And(x, y) is logically equivalent to And(a, b) for the
mapping {x: a, y:b} or {x: b, y:a}. If no such mapping exists, return False.
Examples
>>> from sympy import SOPform, bool_map, Or, And, Not, Xor
>>> from sympy.abc import w, x, y, z, a, b, c, d
>>> function1 = SOPform([x,z,y],[[1, 0, 1], [0, 0, 1]])
>>> function2 = SOPform([a,b,c],[[1, 0, 1], [1, 0, 0]])
>>> bool_map(function1, function2)
(And(Not(z), y), {y: a, z: b})
The results are not necessarily unique, but they are canonical. Here, (w, z) could be
(a, d) or (d, a):
>>> eq = Or(And(Not(y), w), And(Not(y), z), And(x,
>>> eq2 = Or(And(Not(c), a), And(Not(c), d), And(b,
>>> bool_map(eq, eq2)
(Or(And(Not(y), w), And(Not(y), z), And(x, y)), {w:
>>> eq = And(Xor(a, b), c, And(c,d))
>>> bool_map(eq, eq.subs(c, x))
(And(Or(Not(a), Not(b)), Or(a, b), c, d), {a: a, b:
y))
c))
a, x: b, y: c, z: d})
b, c: d, d: x})
5.13.5 Inference
This module implements some inference routines in propositional logic.
The function satisable will test that a given Boolean expression is satisable, that is, you can
assign values to the variables to make the sentence T rue.
For example, the expression x & x is not satisable, since there are no values for x that
make this sentence True. On the other hand, (x | y) & (x | y) & (x | y) is satisable
with both x and y being True.
>>> from sympy.logic.inference import satisfiable
>>> from sympy import Symbol
>>> x = Symbol(x)
>>> y = Symbol(y)
>>> satisfiable(x & ~x)
False
>>> satisfiable((x | y) & (x | ~y) & (~x | y))
{x: True, y: True}
634
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
As you see, when a sentence is satisable, it returns a model that makes that sentence True.
If it is not satisable it will return False.
sympy.logic.inference.satisfiable(expr, algorithm=dpll2, all models=False)
Check satisability of a propositional sentence. Returns a model when it succeeds. Returns {true: true} for trivially true expressions.
On setting all models to True, if given expr is satisable then returns a generator of
models. However, if expr is unsatisable then returns a generator containing the single
element False.
Examples
>>> from sympy.abc import A, B
>>> from sympy.logic.inference import satisfiable
>>> satisfiable(A & ~B)
{A: True, B: False}
>>> satisfiable(A & ~A)
False
>>> satisfiable(True)
{True: True}
>>> next(satisfiable(A & ~A, all_models=True))
False
>>> models = satisfiable((A >> B) & B, all_models=True)
>>> next(models)
{A: False, B: True}
>>> next(models)
{A: True, B: True}
>>> def use_models(models):
...
for model in models:
...
if model:
...
# Do something with the model.
...
print(model)
...
else:
...
# Given expr is unsatisfiable.
...
print(UNSAT)
>>> use_models(satisfiable(A >> ~A, all_models=True))
{A: False}
>>> use_models(satisfiable(A ^ A, all_models=True))
UNSAT
5.14 Matrices
A module that handles matrices.
Includes functions for fast creating matrices like zero, one/eye, random matrix, etc.
Contents:
5.14. Matrices
635
SymPy Documentation, Release 0.7.6
5.14.1 Matrices (linear algebra)
Creating Matrices
The linear algebra module is designed to be as simple as possible. First, we import and
declare our rst Matrix object:
>>> from sympy.interactive.printing import init_printing
>>> init_printing(use_unicode=False, wrap_line=False, no_global=True)
>>> from sympy.matrices import Matrix, eye, zeros, ones, diag, GramSchmidt
>>> M = Matrix([[1,0,0], [0,0,0]]); M
[1 0 0]
[
]
[0 0 0]
>>> Matrix([M, (0, 0, -1)])
[1 0 0 ]
[
]
[0 0 0 ]
[
]
[0 0 -1]
>>> Matrix([[1, 2, 3]])
[1 2 3]
>>> Matrix([1, 2, 3])
[1]
[ ]
[2]
[ ]
[3]
In addition to creating a matrix from a list of appropriately-sized lists and/or matrices, SymPy
also supports more advanced methods of matrix creation including a single list of values and
dimension inputs:
>>> Matrix(2, 3, [1, 2, 3, 4, 5, 6])
[1 2 3]
[
]
[4 5 6]
More interesting (and useful), is the ability to use a 2-variable function (or lambda) to create
a matrix. Here we create an indicator function which is 1 on the diagonal and then use it to
make the identity matrix:
>>>
...
...
...
...
...
>>>
[1
[
[0
[
[0
[
[0
def f(i,j):
if i == j:
return 1
else:
return 0
Matrix(4, 4, f)
0 0 0]
]
1 0 0]
]
0 1 0]
]
0 0 1]
Finally lets use lambda to create a 1-line matrix with 1s in the even permutation entries:
636
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>>
[1
[
[0
[
[1
Matrix(3, 4, lambda i,j: 1 - (i+j) % 2)
0 1 0]
]
1 0 1]
]
0 1 0]
There are also a couple of special constructors for quick matrix construction: eye is the
identity matrix, zeros and ones for matrices of all zeros and ones, respectively, and diag to
put matrices or elements along the diagonal:
>>>
[1
[
[0
[
[0
[
[0
>>>
[0
[
[0
>>>
[0
[
[0
>>>
[1
[
[1
[
[1
>>>
[1
>>>
[1
[
[0
[
[0
eye(4)
0 0 0]
]
1 0 0]
]
0 1 0]
]
0 0 1]
zeros(2)
0]
]
0]
zeros(2, 5)
0 0 0 0]
]
0 0 0 0]
ones(3)
1 1]
]
1 1]
]
1 1]
ones(1, 3)
1 1]
diag(1, Matrix([[1, 2], [3, 4]]))
0 0]
]
1 2]
]
3 4]
Basic Manipulation
While learning to work with matrices, lets choose one where the entries are readily identiable. One useful thing to know is that while matrices are 2-dimensional, the storage is not
and so it is allowable - though one should be careful - to access the entries as if they were a
1-d list.
>>> M = Matrix(2, 3, [1, 2, 3, 4, 5, 6])
>>> M[4]
5
Now, the more standard entry access is a pair of indices which will always return the value
at the corresponding row and column of the matrix:
5.14. Matrices
637
SymPy Documentation, Release 0.7.6
>>> M[1, 2]
6
>>> M[0, 0]
1
>>> M[1, 1]
5
Since this is Python were also able to slice submatrices; slices always give a matrix in return,
even if the dimension is 1 x 1:
>>>
[1
[
[4
>>>
[]
>>>
[3]
[ ]
[6]
>>>
[3]
M[0:2, 0:2]
2]
]
5]
M[2:2, 2]
M[:, 2]
M[:1, 2]
In the second example above notice that the slice 2:2 gives an empty range. Note also (in
keeping with 0-based indexing of Python) the rst row/column is 0.
You cannot access rows or columns that are not present unless they are in a slice:
>>> M[:, 10] # the 10-th column (not there)
Traceback (most recent call last):
...
IndexError: Index out of range: a[[0, 10]]
>>> M[:, 10:11] # the 10-th column (if there)
[]
>>> M[:, :10] # all columns up to the 10-th
[1 2 3]
[
]
[4 5 6]
Traceback (most recent call last):
...
IndexError: Index out of range: a[[0, 10]]
Slicing an empty matrix works as long as you use a slice for the coordinate that has no size:
>>> Matrix(0, 3, [])[:, 1]
[]
Slicing gives a copy of what is sliced, so modications of one object do not aect the other:
>>> M2 = M[:, :]
>>> M2[0, 0] = 100
>>> M[0, 0] == 100
False
Notice that changing M2 didnt change M. Since we can slice, we can also assign entries:
>>> M = Matrix(([1,2,3,4],[5,6,7,8],[9,10,11,12],[13,14,15,16]))
>>> M
[1
2
3
4 ]
[
]
638
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
[5
6
7
[
[9
10 11
[
[13 14 15
>>> M[2,2] =
>>> M
[1
2
3
[
[5
6
7
[
[9
10 0
[
[13 14 15
8 ]
]
12]
]
16]
M[0,3] = 0
0 ]
]
8 ]
]
12]
]
16]
as well as assign slices:
>>> M = Matrix(([1,2,3,4],[5,6,7,8],[9,10,11,12],[13,14,15,16]))
>>> M[2:,2:] = Matrix(2,2,lambda i,j: 0)
>>> M
[1
2
3 4]
[
]
[5
6
7 8]
[
]
[9
10 0 0]
[
]
[13 14 0 0]
All the standard arithmetic operations are supported:
>>> M = Matrix(([1,2,3],[4,5,6],[7,8,9]))
>>> M - M
[0 0 0]
[
]
[0 0 0]
[
]
[0 0 0]
>>> M + M
[2
4
6 ]
[
]
[8
10 12]
[
]
[14 16 18]
>>> M * M
[30
36
42 ]
[
]
[66
81
96 ]
[
]
[102 126 150]
>>> M2 = Matrix(3,1,[1,5,0])
>>> M*M2
[11]
[ ]
[29]
[ ]
[47]
>>> M**2
[30
36
42 ]
[
]
5.14. Matrices
639
SymPy Documentation, Release 0.7.6
[66
[
[102
81
126
96 ]
]
150]
As well as some useful vector operations:
>>>
>>>
[4
[
[7
>>>
>>>
[4
[
[7
>>>
>>>
>>>
>>>
32
>>>
0
>>>
0
M.row_del(0)
M
5 6]
]
8 9]
M.col_del(1)
M
6]
]
9]
v1 = Matrix([1,2,3])
v2 = Matrix([4,5,6])
v3 = v1.cross(v2)
v1.dot(v2)
v2.dot(v3)
v1.dot(v3)
Recall that the row del() and col del() operations dont return a value - they simply change
the matrix object. We can also glue together matrices of the appropriate size:
>>>
>>>
>>>
[1
[
[0
[
[0
>>>
>>>
[1
[
[0
[
[0
[
[0
[
[0
[
[0
[
[0
M1 = eye(3)
M2 = zeros(3, 4)
M1.row_join(M2)
0 0 0 0 0 0]
]
1 0 0 0 0 0]
]
0 1 0 0 0 0]
M3 = zeros(4, 3)
M1.col_join(M3)
0 0]
]
1 0]
]
0 1]
]
0 0]
]
0 0]
]
0 0]
]
0 0]
Operations on entries
We are not restricted to having multiplication between two matrices:
640
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>>
>>>
[2
[
[0
[
[0
>>>
[3
[
[0
[
[0
M = eye(3)
2*M
0 0]
]
2 0]
]
0 2]
3*M
0 0]
]
3 0]
]
0 3]
but we can also apply functions to our matrix entries using applyfunc(). Here well declare
a function that double any input number. Then we apply it to the 3x3 identity matrix:
>>>
>>>
[2
[
[0
[
[0
f = lambda x: 2*x
eye(3).applyfunc(f)
0 0]
]
2 0]
]
0 2]
One more useful matrix-wide entry application function is the substitution function. Lets
declare a matrix with symbolic entries then substitute a value. Remember we can substitute
anything - even another symbol!:
>>>
>>>
>>>
>>>
[x
[
[0
[
[0
>>>
[4
[
[0
[
[0
>>>
>>>
[y
[
[0
[
[0
from sympy import Symbol
x = Symbol(x)
M = eye(3) * x
M
0 0]
]
x 0]
]
0 x]
M.subs(x, 4)
0 0]
]
4 0]
]
0 4]
y = Symbol(y)
M.subs(x, y)
0 0]
]
y 0]
]
0 y]
Linear algebra
Now that we have the basics out of the way, lets see what we can do with the actual matrices.
Of course, one of the rst things that comes to mind is the determinant:
5.14. Matrices
641
SymPy Documentation, Release 0.7.6
>>>
>>>
-28
>>>
>>>
1
>>>
>>>
0
M = Matrix(( [1, 2, 3], [3, 6, 2], [2, 0, 1] ))
M.det()
M2 = eye(3)
M2.det()
M3 = Matrix(( [1, 0, 0], [1, 0, 0], [1, 0, 0] ))
M3.det()
Another common operation is the inverse: In SymPy, this is computed by Gaussian elimination
by default (for dense matrices) but we can specify it be done by LU decomposition as well:
>>> M2.inv()
[1 0 0]
[
]
[0 1 0]
[
]
[0 0 1]
>>> M2.inv(method=LU)
[1 0 0]
[
]
[0 1 0]
[
]
[0 0 1]
>>> M.inv(method=LU)
[-3/14 1/14 1/2 ]
[
]
[-1/28 5/28 -1/4]
[
]
[ 3/7
-1/7
0 ]
>>> M * M.inv(method=LU)
[1 0 0]
[
]
[0 1 0]
[
]
[0 0 1]
We can perform a QR factorization which is handy for solving systems:
>>> A = Matrix([[1,1,1],[1,1,3],[2,3,4]])
>>> Q, R = A.QRdecomposition()
>>> Q
[ ___
___
___ ]
[\/ 6
-\/ 3
-\/ 2 ]
[----- ------- -------]
[ 6
3
2
]
[
]
[ ___
___
___ ]
[\/ 6
-\/ 3
\/ 2 ]
[----- ----------- ]
[ 6
3
2
]
[
]
[ ___
___
]
[\/ 6
\/ 3
]
[--------0
]
[ 3
3
]
>>> R
[
___
]
642
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
[ ___ 4*\/ 6
[\/ 6
------[
3
[
[
___
[
\/ 3
[ 0
----[
3
[
[
[ 0
0
>>> Q*R
[1 1 1]
[
]
[1 1 3]
[
]
[2 3 4]
___]
2*\/ 6 ]
]
]
]
]
0
]
]
]
___ ]
\/ 2 ]
In addition to the solvers in the solver.py le, we can solve the system Ax=b by passing
the b vector to the matrix As LUsolve function. Here well cheat a little choose A and x then
multiply to get b. Then we can solve for x and check that its correct:
>>>
>>>
>>>
>>>
>>>
[3]
[ ]
[7]
[ ]
[5]
A = Matrix([ [2, 3, 5], [3, 6, 2], [8, 3, 6] ])
x = Matrix(3,1,[3,7,5])
b = A*x
soln = A.LUsolve(b)
soln
Theres also a nice Gram-Schmidt orthogonalizer which will take a set of vectors and orthogonalize then with respect to another another. There is an optional argument which species
whether or not the output should also be normalized, it defaults to False. Lets take some
vectors and orthogonalize them - one normalized and one not:
>>> L = [Matrix([2,3,5]), Matrix([3,6,2]), Matrix([8,3,6])]
>>> out1 = GramSchmidt(L)
>>> out2 = GramSchmidt(L, True)
Lets take a look at the vectors:
>>> for i in out1:
...
print(i)
...
Matrix([[2], [3], [5]])
Matrix([[23/19], [63/19], [-47/19]])
Matrix([[1692/353], [-1551/706], [-423/706]])
>>> for i in out2:
...
print(i)
...
Matrix([[sqrt(38)/19], [3*sqrt(38)/38], [5*sqrt(38)/38]])
Matrix([[23*sqrt(6707)/6707], [63*sqrt(6707)/6707], [-47*sqrt(6707)/6707]])
Matrix([[12*sqrt(706)/353], [-11*sqrt(706)/706], [-3*sqrt(706)/706]])
We can spot-check their orthogonality with dot() and their normality with norm():
5.14. Matrices
643
SymPy Documentation, Release 0.7.6
>>>
0
>>>
0
>>>
0
>>>
1
>>>
1
>>>
1
out1[0].dot(out1[1])
out1[0].dot(out1[2])
out1[1].dot(out1[2])
out2[0].norm()
out2[1].norm()
out2[2].norm()
So there is quite a bit that can be done with the module including eigenvalues, eigenvectors,
nullspace calculation, cofactor expansion tools, and so on. From here one might want to look
over the matrices.py le for all functionality.
MatrixBase Class Reference
class sympy.matrices.matrices.MatrixBase
Attributes
is Identity
C
By-element conjugation.
D
Return Dirac conjugate (if self.rows == 4).
See Also:
conjugate By-element conjugation
H (page 645) Hermite conjugation
Examples
>>> from sympy import Matrix, I, eye
>>> m = Matrix((0, 1 + I, 2, 3))
>>> m.D
Matrix([[0, 1 - I, -2, -3]])
>>> m = (eye(4) + I*eye(4))
>>> m[0, 3] = 2
>>> m.D
Matrix([
[1 - I,
0,
0,
0],
[
0, 1 - I,
0,
0],
[
0,
0, -1 + I,
0],
[
2,
0,
0, -1 + I]])
If the matrix does not have 4 rows an AttributeError will be raised because this
property is only dened for matrices with 4 rows.
644
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> Matrix(eye(2)).D
Traceback (most recent
...
AttributeError: Matrix
Traceback (most recent
...
AttributeError: Matrix
call last):
has no attribute D.
call last):
has no attribute D.
H
Return Hermite conjugate.
See Also:
conjugate By-element conjugation
D (page 644) Dirac conjugation
Examples
>>> from sympy import Matrix, I
>>> m = Matrix((0, 1 + I, 2, 3))
>>> m
Matrix([
[
0],
[1 + I],
[
2],
[
3]])
>>> m.H
Matrix([[0, 1 - I, 2, 3]])
LDLdecomposition()
Returns the LDL Decomposition (L, D) of matrix A, such that L * D * L.T == A This
method eliminates the use of square root. Further this ensures that all the diagonal
entries of L are 1. A must be a square, symmetric, positive-denite and non-singular
matrix.
See Also:
cholesky (page 651), LUdecomposition (page 646), QRdecomposition (page 647)
Examples
>>> from sympy.matrices import Matrix, eye
>>> A = Matrix(((25, 15, -5), (15, 18, 0), (-5, 0, 11)))
>>> L, D = A.LDLdecomposition()
>>> L
Matrix([
[
1,
0, 0],
[ 3/5,
1, 0],
[-1/5, 1/3, 1]])
>>> D
Matrix([
[25, 0, 0],
[ 0, 9, 0],
[ 0, 0, 9]])
5.14. Matrices
645
SymPy Documentation, Release 0.7.6
>>> L * D * L.T * A.inv() == eye(A.rows)
True
LDLsolve(rhs)
Solves Ax = B using LDL decomposition, for a general square and non-singular matrix.
For a non-square matrix with rows > cols, the least squares solution is returned.
See Also:
LDLdecomposition (page 645), lower triangular solve (page 668), upper triangular solve (page 677), cholesky solve (page 651), diagonal solve
(page 653), LUsolve (page 647), QRsolve (page 648), pinv solve (page 671)
Examples
>>> from sympy.matrices import Matrix, eye
>>> A = eye(2)*2
>>> B = Matrix([[1, 2], [3, 4]])
>>> A.LDLsolve(B) == B/2
True
LUdecomposition(iszerofunc=<function iszero at 0xa2e4aac>)
Returns the decomposition LU and the row swaps p.
See Also:
cholesky (page 651), LDLdecomposition (page 645), QRdecomposition (page 647),
LUdecomposition Simple (page 647), LUdecompositionFF (page 646), LUsolve
(page 647)
Examples
>>> from sympy import Matrix
>>> a = Matrix([[4, 3], [6, 3]])
>>> L, U, _ = a.LUdecomposition()
>>> L
Matrix([
[ 1, 0],
[3/2, 1]])
>>> U
Matrix([
[4,
3],
[0, -3/2]])
LUdecompositionFF()
Compute a fraction-free LU decomposition.
Returns 4 matrices P, L, D, U such that PA = L D**-1 U. If the elements of the matrix
belong to some integral domain I, then all elements of L, D and U are guaranteed to
belong to I.
Reference
W. Zhou & D.J. Jerey, Fraction-free matrix factors: new forms for LU and
QR factors. Frontiers in Computer Science in China, Vol 2, no. 1, pp. 67-80,
2008.
646
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
See Also:
LUdecomposition (page 646), LUdecomposition Simple (page 647), LUsolve
(page 647)
LUdecomposition Simple(iszerofunc=<function iszero at 0xa2e4aac>)
Returns A comprised of L, U (Ls diag entries are 1) and p which is the list of the row
swaps (in order).
See Also:
LUdecomposition (page 646), LUdecompositionFF (page 646), LUsolve (page 647)
LUsolve(rhs, iszerofunc=<function iszero at 0xa2e4aac>)
Solve the linear system Ax = rhs for x where A = self.
This is for symbolic matrices, for real or complex ones use sympy.mpmath.lu solve
or sympy.mpmath.qr solve.
See Also:
lower triangular solve (page 668), upper triangular solve (page 677),
cholesky solve (page 651), diagonal solve (page 653), LDLsolve (page 646),
QRsolve (page 648), pinv solve (page 671), LUdecomposition (page 646)
QRdecomposition()
Return Q, R where A = Q*R, Q is orthogonal and R is upper triangular.
See Also:
cholesky (page 651), LDLdecomposition (page 645), LUdecomposition (page 646),
QRsolve (page 648)
Examples
This is the example from wikipedia:
>>> from sympy import Matrix
>>> A = Matrix([[12, -51, 4], [6, 167, -68], [-4, 24, -41]])
>>> Q, R = A.QRdecomposition()
>>> Q
Matrix([
[ 6/7, -69/175, -58/175],
[ 3/7, 158/175,
6/175],
[-2/7,
6/35, -33/35]])
>>> R
Matrix([
[14, 21, -14],
[ 0, 175, -70],
[ 0,
0, 35]])
>>> A == Q*R
True
QR factorization of an identity matrix:
>>> A = Matrix([[1, 0, 0], [0, 1, 0], [0, 0, 1]])
>>> Q, R = A.QRdecomposition()
>>> Q
Matrix([
[1, 0, 0],
[0, 1, 0],
[0, 0, 1]])
5.14. Matrices
647
SymPy Documentation, Release 0.7.6
>>> R
Matrix([
[1, 0, 0],
[0, 1, 0],
[0, 0, 1]])
QRsolve(b)
Solve the linear system Ax = b.
self is the matrix A, the method argument is the vector b. The method returns
the solution vector x. If b is a matrix, the system is solved for each column of b
and the return value is a matrix of the same shape as b.
This method is slower (approximately by a factor of 2) but more stable for oatingpoint arithmetic than the LUsolve method. However, LUsolve usually uses an exact
arithmetic, so you dont need to use QRsolve.
This is mainly for educational purposes and symbolic matrices, for real (or complex)
matrices use sympy.mpmath.qr solve.
See Also:
lower triangular solve (page 668), upper triangular solve (page 677),
cholesky solve (page 651), diagonal solve (page 653), LDLsolve (page 646),
LUsolve (page 647), pinv solve (page 671), QRdecomposition (page 647)
T
Matrix transposition.
add(b)
Return self + b
adjoint()
Conjugate transpose or Hermitian conjugation.
adjugate(method=berkowitz)
Returns the adjugate matrix.
Adjugate matrix is the transpose of the cofactor matrix.
http://en.wikipedia.org/wiki/Adjugate
See Also:
cofactorMatrix (page 651), transpose, berkowitz (page 648)
atoms(*types)
Returns the atoms that form the current object.
Examples
>>> from sympy.abc import x, y
>>> from sympy.matrices import Matrix
>>> Matrix([[x]])
Matrix([[x]])
>>> _.atoms()
set([x])
berkowitz()
The Berkowitz algorithm.
648
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Given N x N matrix with symbolic content, compute eciently coecients of
characteristic polynomials of self and all its square sub-matrices composed
by removing both i-th row and column, without division in the ground domain.
This method is particularly useful for computing determinant, principal minors and characteristic polynomial, when self has complicated coecients
e.g. polynomials. Semi-direct usage of this algorithm is also important in
computing eciently sub-resultant PRS.
Assuming that M is a square matrix of dimension N x N and I is N x N identity
matrix, then the following following denition of characteristic polynomial is
begin used:
charpoly(M) = det(t*I - M)
As a consequence, all polynomials generated by Berkowitz algorithm are
monic.
>>> from sympy import Matrix
>>> from sympy.abc import x, y, z
>>> M = Matrix([[x, y, z], [1, 0, 0], [y, z, x]])
>>> p, q, r = M.berkowitz()
>>> p # 1 x 1 Ms sub-matrix
(1, -x)
>>> q # 2 x 2 Ms sub-matrix
(1, -x, -y)
>>> r # 3 x 3 Ms sub-matrix
(1, -2*x, x**2 - y*z - y, x*y - z**2)
For more information on the implemented algorithm refer to:
[1] S.J. Berkowitz, On computing the determinant in small parallel
time using a small number of processors, ACM, Information Processing
Letters 18, 1984, pp. 147-150
[2] M. Keber, Division-Free computation of sub-resultants using
Bezout matrices, Tech. Report MPI-I-2006-1-006, Saarbrucken, 2006
See Also:
berkowitz det (page 650), berkowitz minors (page 650), berkowitz charpoly
(page 649), berkowitz eigenvals (page 650)
berkowitz charpoly(x= lambda, simplify=<function simplify at 0xa29b8ec>)
Computes characteristic polynomial minors using Berkowitz method.
A PurePoly is returned so using dierent variables for x does not aect the comparison or the polynomials:
See Also:
berkowitz (page 648)
5.14. Matrices
649
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import Matrix
>>> from sympy.abc import x, y
>>> A = Matrix([[1, 3], [2, 0]])
>>> A.berkowitz_charpoly(x) == A.berkowitz_charpoly(y)
True
Specifying x is optional; a Dummy with name lambda is used by default (which looks
good when pretty-printed in unicode):
>>> A.berkowitz_charpoly().as_expr()
_lambda**2 - _lambda - 6
No test is done to see that x doesnt clash with an existing symbol, so using the
default (lambda) or your own Dummy symbol is the safest option:
>>> A = Matrix([[1, 2], [x, 0]])
>>> A.charpoly().as_expr()
_lambda**2 - _lambda - 2*x
>>> A.charpoly(x).as_expr()
x**2 - 3*x
berkowitz det()
Computes determinant using Berkowitz method.
See Also:
det (page 653), berkowitz (page 648)
berkowitz eigenvals(**ags)
Computes eigenvalues of a Matrix using Berkowitz method.
See Also:
berkowitz (page 648)
berkowitz minors()
Computes principal minors using Berkowitz method.
See Also:
berkowitz (page 648)
charpoly(x= lambda, simplify=<function simplify at 0xa29b8ec>)
Computes characteristic polynomial minors using Berkowitz method.
A PurePoly is returned so using dierent variables for x does not aect the comparison or the polynomials:
See Also:
berkowitz (page 648)
Examples
>>> from sympy import Matrix
>>> from sympy.abc import x, y
>>> A = Matrix([[1, 3], [2, 0]])
>>> A.berkowitz_charpoly(x) == A.berkowitz_charpoly(y)
True
650
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Specifying x is optional; a Dummy with name lambda is used by default (which looks
good when pretty-printed in unicode):
>>> A.berkowitz_charpoly().as_expr()
_lambda**2 - _lambda - 6
No test is done to see that x doesnt clash with an existing symbol, so using the
default (lambda) or your own Dummy symbol is the safest option:
>>> A = Matrix([[1, 2], [x, 0]])
>>> A.charpoly().as_expr()
_lambda**2 - _lambda - 2*x
>>> A.charpoly(x).as_expr()
x**2 - 3*x
cholesky()
Returns the Cholesky decomposition L of a matrix A such that L * L.T = A
A must be a square, symmetric, positive-denite and non-singular matrix.
See Also:
LDLdecomposition (page 645), LUdecomposition (page 646), QRdecomposition
(page 647)
Examples
>>> from sympy.matrices import Matrix
>>> A = Matrix(((25, 15, -5), (15, 18, 0), (-5, 0, 11)))
>>> A.cholesky()
Matrix([
[ 5, 0, 0],
[ 3, 3, 0],
[-1, 1, 3]])
>>> A.cholesky() * A.cholesky().T
Matrix([
[25, 15, -5],
[15, 18, 0],
[-5, 0, 11]])
cholesky solve(rhs)
Solves Ax = B using Cholesky decomposition, for a general square non-singular matrix. For a non-square matrix with rows > cols, the least squares solution is returned.
See Also:
lower triangular solve (page 668), upper triangular solve (page 677), diagonal solve (page 653), LDLsolve (page 646), LUsolve (page 647), QRsolve
(page 648), pinv solve (page 671)
cofactor(i, j, method=berkowitz)
Calculate the cofactor of an element.
See Also:
cofactorMatrix (page 651), minorEntry (page 668), minorMatrix (page 668)
cofactorMatrix(method=berkowitz)
Return a matrix containing the cofactor of each element.
See Also:
5.14. Matrices
651
SymPy Documentation, Release 0.7.6
cofactor (page 651), minorEntry (page 668), minorMatrix (page 668), adjugate
(page 648)
col insert(pos, mti)
Insert one or more columns at the given column position.
See Also:
col, row insert (page 673)
Examples
>>> from sympy import zeros, ones
>>> M = zeros(3)
>>> V = ones(3, 1)
>>> M.col_insert(1, V)
Matrix([
[0, 1, 0, 0],
[0, 1, 0, 0],
[0, 1, 0, 0]])
col join(bott)
Concatenates two matrices along selfs last and botts rst row
See Also:
col, row join (page 673)
Examples
>>> from sympy import zeros, ones
>>> M = zeros(3)
>>> V = ones(1, 3)
>>> M.col_join(V)
Matrix([
[0, 0, 0],
[0, 0, 0],
[0, 0, 0],
[1, 1, 1]])
condition number()
Returns the condition number of a matrix.
This is the maximum singular value divided by the minimum singular value
See Also:
singular values (page 674)
Examples
>>> from sympy import Matrix, S
>>> A = Matrix([[1, 0, 0], [0, 10, 0], [0, 0, S.One/10]])
>>> A.condition_number()
100
652
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
cross(b)
Return the cross product of self and b relaxing the condition of compatible dimensions: if each has 3 elements, a matrix of the same type and shape as self will be
returned. If b has the same shape as self then common identities for the cross product (like axb = bxa) will hold.
See Also:
dot (page 654), multiply (page 668), multiply elementwise (page 668)
det(method=bareis)
Computes the matrix determinant using the method method.
Possible values for method: bareis ... det bareis berkowitz ... berkowitz det
det LU ... det LU decomposition
See Also:
det bareis (page 653), berkowitz det (page 650), det LU
det LU decomposition()
Compute matrix determinant using LU decomposition
Note that this method fails if the LU decomposition itself fails. In particular, if the
matrix has no inverse this method will fail.
TODO:
Implement
algorithm
for
sparse
http://www.eecis.udel.edu/saunders/papers/sge/it5.ps.
matrices
(SFF),
See Also:
det (page 653), det bareis (page 653), berkowitz det (page 650)
det bareis()
Compute matrix determinant using Bareis fraction-free algorithm which is an extension of the well known Gaussian elimination method. This approach is best suited
for dense symbolic matrices and will result in a determinant with minimal number
of fractions. It means that less term rewriting is needed on resulting formulae.
TODO:
Implement
algorithm
for
sparse
http://www.eecis.udel.edu/saunders/papers/sge/it5.ps.
matrices
(SFF),
See Also:
det (page 653), berkowitz det (page 650)
diagonal solve(rhs)
Solves Ax = B eciently, where A is a diagonal Matrix, with non-zero diagonal entries.
See Also:
lower triangular solve (page 668), upper triangular solve (page 677),
cholesky solve (page 651), LDLsolve (page 646), LUsolve (page 647), QRsolve
(page 648), pinv solve (page 671)
Examples
>>> from sympy.matrices import Matrix, eye
>>> A = eye(2)*2
>>> B = Matrix([[1, 2], [3, 4]])
>>> A.diagonal_solve(B) == B/2
True
5.14. Matrices
653
SymPy Documentation, Release 0.7.6
diagonalize(reals only=False, sort=False, normalize=False)
Return (P, D), where D is diagonal and
D = P-1 * M * P
where M is current matrix.
See Also:
is diagonal (page 659), is diagonalizable (page 660)
Examples
>>> from sympy import Matrix
>>> m = Matrix(3, 3, [1, 2, 0, 0, 3, 0, 2, -4, 2])
>>> m
Matrix([
[1, 2, 0],
[0, 3, 0],
[2, -4, 2]])
>>> (P, D) = m.diagonalize()
>>> D
Matrix([
[1, 0, 0],
[0, 2, 0],
[0, 0, 3]])
>>> P
Matrix([
[-1, 0, -1],
[ 0, 0, -1],
[ 2, 1, 2]])
>>> P.inv() * m * P
Matrix([
[1, 0, 0],
[0, 2, 0],
[0, 0, 3]])
diff(*args)
Calculate the derivative of each element in the matrix.
See Also:
integrate (page 657), limit (page 668)
Examples
>>> from sympy.matrices import Matrix
>>> from sympy.abc import x, y
>>> M = Matrix([[x, y], [1, 0]])
>>> M.diff(x)
Matrix([
[1, 0],
[0, 0]])
dot(b)
Return the dot product of Matrix self and b relaxing the condition of compatible
dimensions: if either the number of rows or columns are the same as the length
of b then the dot product is returned. If self is a row or column vector, a scalar is
654
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
returned. Otherwise, a list of results is returned (and in that case the number of
columns in self must match the length of b).
See Also:
cross (page 652), multiply (page 668), multiply elementwise (page 668)
Examples
>>>
>>>
>>>
>>>
6
>>>
12
>>>
[6,
from sympy import Matrix
M = Matrix([[1, 2, 3], [4, 5, 6], [7, 8, 9]])
v = [1, 1, 1]
M.row(0).dot(v)
M.col(0).dot(v)
M.dot(v)
15, 24]
dual()
Returns the dual of a matrix, which is:
(1/2) levicivita(i, j, k, l) M (k, l) summed over indices k and l
Since the levicivita method is anti symmetric for any pairwise exchange of indices,
the dual of a symmetric matrix is the zero matrix. Strictly speaking the dual dened
here assumes that the matrix M is a contravariant anti symmetric second rank
tensor, so that the dual is a covariant second rank tensor.
eigenvals(**ags)
Return eigen values using the berkowitz eigenvals routine.
Since the roots routine doesnt always work well with Floats, they will be replaced
with Rationals before calling that routine. If this is not desired, set ag rational to
False.
eigenvects(**ags)
Return list of triples (eigenval, multiplicity, basis).
The ag simplify has two eects: 1)
if
bool(simplify)
is
True,
as content primitive() will be used to tidy up normalization artifacts; 2) if
nullspace needs simplication to compute the basis, the simplify ag will be
passed on to the nullspace routine which will interpret it there.
If the matrix contains any Floats, they will be changed to Rationals for computation
purposes, but the answers will be returned after being evaluated with evalf. If it is
desired to removed small imaginary portions during the evalf step, pass a value for
the chop ag.
evalf(prec=None, **options)
Apply evalf() to each element of self.
exp()
Return the exponentiation of a square matrix.
expand(deep=True,
modulus=None,
power base=True,
power exp=True,
mul=True, log=True, multinomial=True, basic=True, **hints)
Apply core.function.expand to each entry of the matrix.
5.14. Matrices
655
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.abc import x
>>> from sympy.matrices import Matrix
>>> Matrix(1, 1, [x*(x+1)])
Matrix([[x*(x + 1)]])
>>> _.expand()
Matrix([[x**2 + x]])
extract(rowsList, colsList)
Return a submatrix by specifying a list of rows and columns. Negative indices can
be given. All indices must be in the range -n <= i < n where n is the number of rows
or columns.
Examples
>>> from sympy import Matrix
>>> m = Matrix(4, 3, range(12))
>>> m
Matrix([
[0, 1, 2],
[3, 4, 5],
[6, 7, 8],
[9, 10, 11]])
>>> m.extract([0, 1, 3], [0, 1])
Matrix([
[0, 1],
[3, 4],
[9, 10]])
Rows or columns can be repeated:
>>> m.extract([0, 0, 1], [-1])
Matrix([
[2],
[2],
[5]])
Every other row can be taken by using range to provide the indices:
>>> m.extract(range(0, m.rows, 2), [-1])
Matrix([
[2],
[8]])
free symbols
Returns the free symbols within the matrix.
Examples
>>> from sympy.abc import x
>>> from sympy.matrices import Matrix
>>> Matrix([[x], [1]]).free_symbols
set([x])
656
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
get diag blocks()
Obtains the square sub-matrices on the main diagonal of a square matrix.
Useful for inverting symbolic matrices or solving systems of linear equations which
may be decoupled by having a block diagonal structure.
Examples
>>> from sympy import Matrix
>>> from sympy.abc import x, y, z
>>> A = Matrix([[1, 3, 0, 0], [y, z*z, 0, 0], [0, 0, x, 0], [0, 0, 0, 0]])
>>> a1, a2, a3 = A.get_diag_blocks()
>>> a1
Matrix([
[1,
3],
[y, z**2]])
>>> a2
Matrix([[x]])
>>> a3
Matrix([[0]])
has(*patterns)
Test whether any subexpression matches any of the patterns.
Examples
>>> from sympy import Matrix, Float
>>> from sympy.abc import x, y
>>> A = Matrix(((1, x), (0.2, 3)))
>>> A.has(x)
True
>>> A.has(y)
False
>>> A.has(Float)
True
classmethod hstack(*args)
Return a matrix formed by joining args horizontally (i.e. by repeated application of
row join).
Examples
>>> from sympy.matrices import Matrix, eye
>>> Matrix.hstack(eye(2), 2*eye(2))
Matrix([
[1, 0, 2, 0],
[0, 1, 0, 2]])
integrate(*args)
Integrate each element of the matrix.
See Also:
limit (page 668), diff (page 654)
5.14. Matrices
657
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.matrices import Matrix
>>> from sympy.abc import x, y
>>> M = Matrix([[x, y], [1, 0]])
>>> M.integrate((x, ))
Matrix([
[x**2/2, x*y],
[
x,
0]])
>>> M.integrate((x, 0, 2))
Matrix([
[2, 2*y],
[2,
0]])
inv mod(m)
Returns the inverse of the matrix K (mod m), if it exists.
Method to nd the matrix inverse of K (mod m) implemented in this function:
Compute adj(K) = cof(K)t , the adjoint matrix of K.
Compute r = 1/det(K) (mod m).
K 1 = r adj(K) (mod m).
Examples
>>> from sympy import Matrix
>>> A = Matrix(2, 2, [1, 2, 3, 4])
>>> A.inv_mod(5)
Matrix([
[3, 1],
[4, 2]])
>>> A.inv_mod(3)
Matrix([
[1, 1],
[0, 1]])
inverse ADJ(iszerofunc=<function iszero at 0xa2e4aac>)
Calculates the inverse using the adjugate matrix and a determinant.
See Also:
inv, inverse LU (page 658), inverse GE (page 658)
inverse GE(iszerofunc=<function iszero at 0xa2e4aac>)
Calculates the inverse using Gaussian elimination.
See Also:
inv, inverse LU (page 658), inverse ADJ (page 658)
inverse LU(iszerofunc=<function iszero at 0xa2e4aac>)
Calculates the inverse using LU decomposition.
See Also:
inv, inverse GE (page 658), inverse ADJ (page 658)
is anti symmetric(simplify=True)
Check if matrix M is an antisymmetric matrix, that is, M is a square matrix with all
M[i, j] == -M[j, i].
658
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
When simplify=True (default), the sum M[i, j] + M[j, i] is simplied before testing
to see if it is zero. By default, the SymPy simplify function is used. To use a custom
function set simplify to a function that accepts a single argument which returns
a simplied expression. To skip simplication, set simplify to False but note that
although this will be faster, it may induce false negatives.
Examples
>>> from sympy import Matrix, symbols
>>> m = Matrix(2, 2, [0, 1, -1, 0])
>>> m
Matrix([
[ 0, 1],
[-1, 0]])
>>> m.is_anti_symmetric()
True
>>> x, y = symbols(x y)
>>> m = Matrix(2, 3, [0, 0, x, -y, 0, 0])
>>> m
Matrix([
[ 0, 0, x],
[-y, 0, 0]])
>>> m.is_anti_symmetric()
False
>>> from sympy.abc import x, y
>>> m = Matrix(3, 3, [0, x**2 + 2*x + 1, y,
...
-(x + 1)**2 , 0, x*y,
...
-y, -x*y, 0])
Simplication of matrix elements is done by default so even though two elements
which should be equal and opposite wouldnt pass an equality test, the matrix is still
reported as anti-symmetric:
>>> m[0, 1] == -m[1, 0]
False
>>> m.is_anti_symmetric()
True
If simplify=False is used for the case when a Matrix is already simplied, this will
speed things up. Here, we see that without simplication the matrix does not appear
anti-symmetric:
>>> m.is_anti_symmetric(simplify=False)
False
But if the matrix were already expanded, then it would appear anti-symmetric and
simplication in the is anti symmetric routine is not needed:
>>> m = m.expand()
>>> m.is_anti_symmetric(simplify=False)
True
is diagonal()
Check if matrix is diagonal, that is matrix in which the entries outside the main
diagonal are all zero.
See Also:
5.14. Matrices
659
SymPy Documentation, Release 0.7.6
is lower (page 661), is upper (page 664), is diagonalizable (page 660), diagonalize (page 653)
Examples
>>> from sympy import Matrix, diag
>>> m = Matrix(2, 2, [1, 0, 0, 2])
>>> m
Matrix([
[1, 0],
[0, 2]])
>>> m.is_diagonal()
True
>>> m = Matrix(2, 2, [1, 1, 0, 2])
>>> m
Matrix([
[1, 1],
[0, 2]])
>>> m.is_diagonal()
False
>>> m = diag(1, 2, 3)
>>> m
Matrix([
[1, 0, 0],
[0, 2, 0],
[0, 0, 3]])
>>> m.is_diagonal()
True
is diagonalizable(reals only=False, clear subproducts=True)
Check if matrix is diagonalizable.
If reals only==True then check that diagonalized matrix consists of the only not
complex values.
Some subproducts could be used further in other methods to avoid double calculations, By default (if clear subproducts==True) they will be deleted.
See Also:
is diagonal (page 659), diagonalize (page 653)
Examples
>>> from sympy import Matrix
>>> m = Matrix(3, 3, [1, 2, 0, 0, 3, 0, 2, -4, 2])
>>> m
Matrix([
[1, 2, 0],
[0, 3, 0],
[2, -4, 2]])
>>> m.is_diagonalizable()
True
>>> m = Matrix(2, 2, [0, 1, 0, 0])
>>> m
660
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Matrix([
[0, 1],
[0, 0]])
>>> m.is_diagonalizable()
False
>>> m = Matrix(2, 2, [0, 1, -1, 0])
>>> m
Matrix([
[ 0, 1],
[-1, 0]])
>>> m.is_diagonalizable()
True
>>> m.is_diagonalizable(True)
False
is hermitian
Checks if the matrix is Hermitian.
In a Hermitian matrix element i,j is the complex conjugate of element j,i.
Examples
>>> from sympy.matrices import Matrix
>>> from sympy import I
>>> from sympy.abc import x
>>> a = Matrix([[1, I], [-I, 1]])
>>> a
Matrix([
[ 1, I],
[-I, 1]])
>>> a.is_hermitian
True
>>> a[0, 0] = 2*I
>>> a.is_hermitian
False
>>> a[0, 0] = x
>>> a.is_hermitian
>>> a[0, 1] = a[1, 0]*I
>>> a.is_hermitian
False
is lower
Check if matrix is a lower triangular matrix. True can be returned even if the matrix
is not square.
See Also:
is upper (page 664), is diagonal (page 659), is lower hessenberg (page 662)
Examples
>>> from sympy import Matrix
>>> m = Matrix(2, 2, [1, 0, 0, 1])
>>> m
Matrix([
[1, 0],
5.14. Matrices
661
SymPy Documentation, Release 0.7.6
[0, 1]])
>>> m.is_lower
True
>>> m = Matrix(4, 3, [0, 0, 0, 2, 0, 0, 1, 4 , 0, 6, 6, 5])
>>> m
Matrix([
[0, 0, 0],
[2, 0, 0],
[1, 4, 0],
[6, 6, 5]])
>>> m.is_lower
True
>>> from sympy.abc import x, y
>>> m = Matrix(2, 2, [x**2 + y, y**2 + x, 0, x + y])
>>> m
Matrix([
[x**2 + y, x + y**2],
[
0,
x + y]])
>>> m.is_lower
False
is lower hessenberg
Checks if the matrix is in the lower-Hessenberg form.
The lower hessenberg matrix has zero entries above the rst superdiagonal.
See Also:
is upper hessenberg (page 665), is lower (page 661)
Examples
>>> from sympy.matrices import Matrix
>>> a = Matrix([[1, 2, 0, 0], [5, 2, 3, 0], [3, 4, 3, 7], [5, 6, 1, 1]])
>>> a
Matrix([
[1, 2, 0, 0],
[5, 2, 3, 0],
[3, 4, 3, 7],
[5, 6, 1, 1]])
>>> a.is_lower_hessenberg
True
is nilpotent()
Checks if a matrix is nilpotent.
A matrix B is nilpotent if for some integer k, B**k is a zero matrix.
Examples
>>> from sympy import Matrix
>>> a = Matrix([[0, 0, 0], [1, 0, 0], [1, 1, 0]])
>>> a.is_nilpotent()
True
662
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> a = Matrix([[1, 0, 1], [1, 0, 0], [1, 1, 0]])
>>> a.is_nilpotent()
False
is square
Checks if a matrix is square.
A matrix is square if the number of rows equals the number of columns. The empty
matrix is square by denition, since the number of rows and the number of columns
are both zero.
Examples
>>> from sympy import Matrix
>>> a = Matrix([[1, 2, 3], [4, 5, 6]])
>>> b = Matrix([[1, 2, 3], [4, 5, 6], [7, 8, 9]])
>>> c = Matrix([])
>>> a.is_square
False
>>> b.is_square
True
>>> c.is_square
True
is symbolic()
Checks if any elements contain Symbols.
Examples
>>> from sympy.matrices import Matrix
>>> from sympy.abc import x, y
>>> M = Matrix([[x, y], [1, 0]])
>>> M.is_symbolic()
True
is symmetric(simplify=True)
Check if matrix is symmetric matrix, that is square matrix and is equal to its transpose.
By default, simplications occur before testing symmetry. They can be skipped using simplify=False; while speeding things a bit, this may however induce false
negatives.
Examples
>>> from sympy import Matrix
>>> m = Matrix(2, 2, [0, 1, 1, 2])
>>> m
Matrix([
[0, 1],
[1, 2]])
>>> m.is_symmetric()
True
5.14. Matrices
663
SymPy Documentation, Release 0.7.6
>>> m = Matrix(2, 2, [0, 1, 2, 0])
>>> m
Matrix([
[0, 1],
[2, 0]])
>>> m.is_symmetric()
False
>>> m = Matrix(2, 3, [0, 0, 0, 0, 0, 0])
>>> m
Matrix([
[0, 0, 0],
[0, 0, 0]])
>>> m.is_symmetric()
False
>>> from sympy.abc import x, y
>>> m = Matrix(3, 3, [1, x**2 + 2*x + 1, y, (x + 1)**2 , 2, 0, y, 0, 3])
>>> m
Matrix([
[
1, x**2 + 2*x + 1, y],
[(x + 1)**2,
2, 0],
[
y,
0, 3]])
>>> m.is_symmetric()
True
If the matrix is already simplied, you may speed-up is symmetric() test by using
simplify=False.
>>> m.is_symmetric(simplify=False)
False
>>> m1 = m.expand()
>>> m1.is_symmetric(simplify=False)
True
is upper
Check if matrix is an upper triangular matrix. True can be returned even if the
matrix is not square.
See Also:
is lower (page 661), is diagonal (page 659), is upper hessenberg (page 665)
Examples
>>> from sympy import Matrix
>>> m = Matrix(2, 2, [1, 0, 0, 1])
>>> m
Matrix([
[1, 0],
[0, 1]])
>>> m.is_upper
True
>>> m = Matrix(4, 3, [5, 1, 9, 0, 4 , 6, 0, 0, 5, 0, 0, 0])
>>> m
Matrix([
664
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
[5, 1, 9],
[0, 4, 6],
[0, 0, 5],
[0, 0, 0]])
>>> m.is_upper
True
>>> m = Matrix(2, 3, [4, 2, 5, 6, 1, 1])
>>> m
Matrix([
[4, 2, 5],
[6, 1, 1]])
>>> m.is_upper
False
is upper hessenberg
Checks if the matrix is the upper-Hessenberg form.
The upper hessenberg matrix has zero entries below the rst subdiagonal.
See Also:
is lower hessenberg (page 662), is upper (page 664)
Examples
>>> from sympy.matrices import Matrix
>>> a = Matrix([[1, 4, 2, 3], [3, 4, 1, 7], [0, 2, 3, 4], [0, 0, 1, 3]])
>>> a
Matrix([
[1, 4, 2, 3],
[3, 4, 1, 7],
[0, 2, 3, 4],
[0, 0, 1, 3]])
>>> a.is_upper_hessenberg
True
is zero
Checks if a matrix is a zero matrix.
A matrix is zero if every element is zero. A matrix need not be square to be considered zero. The empty matrix is zero by the principle of vacuous truth. For a matrix
that may or may not be zero (e.g. contains a symbol), this will be None
Examples
>>> from sympy import Matrix, zeros
>>> from sympy.abc import x
>>> a = Matrix([[0, 0], [0, 0]])
>>> b = zeros(3, 4)
>>> c = Matrix([[0, 1], [0, 0]])
>>> d = Matrix([])
>>> e = Matrix([[x, 0], [0, 0]])
>>> a.is_zero
True
>>> b.is_zero
5.14. Matrices
665
SymPy Documentation, Release 0.7.6
True
>>> c.is_zero
False
>>> d.is_zero
True
>>> e.is_zero
jacobian(X)
Calculates the Jacobian matrix (derivative of a vectorial function).
Parameters self : vector of expressions representing functions f i(x 1, ...,
x n).
X : set of x is in order, it can be a list or a Matrix
Both self and X can be a row or a column matrix in any order :
(i.e., jacobian() should always work). :
See Also:
hessian, wronskian
Examples
>>> from sympy import sin, cos, Matrix
>>> from sympy.abc import rho, phi
>>> X = Matrix([rho*cos(phi), rho*sin(phi), rho**2])
>>> Y = Matrix([rho, phi])
>>> X.jacobian(Y)
Matrix([
[cos(phi), -rho*sin(phi)],
[sin(phi), rho*cos(phi)],
[
2*rho,
0]])
>>> X = Matrix([rho*cos(phi), rho*sin(phi)])
>>> X.jacobian(Y)
Matrix([
[cos(phi), -rho*sin(phi)],
[sin(phi), rho*cos(phi)]])
jordan cells(calc transformation=True)
Return a list of Jordan cells of current matrix. This list shape Jordan matrix J.
If calc transformation is specied as False, then transformation P such that
J = P 1 M P
will not be calculated.
See Also:
jordan form (page 667)
Notes
Calculation of transformation P is not implemented yet.
666
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>>
>>>
...
...
...
...
from sympy import Matrix
m = Matrix(4, 4, [
6, 5, -2, -3,
-3, -1, 3, 3,
2, 1, -2, -3,
-1, 1, 5, 5])
>>> P, Jcells = m.jordan_cells()
>>> Jcells[0]
Matrix([
[2, 1],
[0, 2]])
>>> Jcells[1]
Matrix([
[2, 1],
[0, 2]])
jordan form(calc transformation=True)
Return Jordan form J of current matrix.
Also the transformation P such that
J = P 1 M P
and the jordan blocks forming J will be calculated.
See Also:
jordan cells (page 666)
Examples
>>> from sympy import Matrix
>>> m = Matrix([
...
[ 6, 5, -2, -3],
...
[-3, -1, 3, 3],
...
[ 2, 1, -2, -3],
...
[-1, 1, 5, 5]])
>>> P, J = m.jordan_form()
>>> J
Matrix([
[2, 1, 0, 0],
[0, 2, 0, 0],
[0, 0, 2, 1],
[0, 0, 0, 2]])
key2bounds(keys)
Converts a key with potentially mixed types of keys (integer and slice) into a tuple
of ranges and raises an error if any index is out of selfs range.
See Also:
key2ij (page 667)
key2ij(key)
Converts key into canonical form, converting integers or indexable items into valid
integers for selfs range or returning slices unchanged.
5.14. Matrices
667
SymPy Documentation, Release 0.7.6
See Also:
key2bounds (page 667)
limit(*args)
Calculate the limit of each element in the matrix.
See Also:
integrate (page 657), diff (page 654)
Examples
>>> from sympy.matrices import Matrix
>>> from sympy.abc import x, y
>>> M = Matrix([[x, y], [1, 0]])
>>> M.limit(x, 2)
Matrix([
[2, y],
[1, 0]])
lower triangular solve(rhs)
Solves Ax = B, where A is a lower triangular matrix.
See Also:
upper triangular solve (page 677), cholesky solve (page 651), diagonal solve
(page 653), LDLsolve (page 646), LUsolve (page 647), QRsolve (page 648),
pinv solve (page 671)
minorEntry(i, j, method=berkowitz)
Calculate the minor of an element.
See Also:
minorMatrix (page 668), cofactor (page 651), cofactorMatrix (page 651)
minorMatrix(i, j)
Creates the minor matrix of a given element.
See Also:
minorEntry (page 668), cofactor (page 651), cofactorMatrix (page 651)
multiply(b)
Returns self*b
See Also:
dot (page 654), cross (page 652), multiply elementwise (page 668)
multiply elementwise(b)
Return the Hadamard product (elementwise product) of A and B
See Also:
cross (page 652), dot (page 654), multiply (page 668)
Examples
668
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> from sympy.matrices import Matrix
>>> A = Matrix([[0, 1, 2], [3, 4, 5]])
>>> B = Matrix([[1, 10, 100], [100, 10, 1]])
>>> A.multiply_elementwise(B)
Matrix([
[ 0, 10, 200],
[300, 40,
5]])
n(prec=None, **options)
Apply evalf() to each element of self.
norm(ord=None)
Return the Norm of a Matrix or Vector. In the simplest case this is the geometric
size of the vector Other norms can be specied by the ord parameter
ord
None
fro
norm for matrices
Frobenius norm
Frobenius norm
norm for vectors
2-norm
does not exist
inf
-inf
1
-1
2
-2
other
2-norm
(largest
sing.
value)
smallest singular value
max(abs(x))
min(abs(x))
as below
as below
as below
as below
sum(abs(x)**ord)**(1./ord)
does not exist
See Also:
normalized (page 669)
Examples
>>> from sympy import Matrix, Symbol, trigsimp, cos, sin, oo
>>> x = Symbol(x, real=True)
>>> v = Matrix([cos(x), sin(x)])
>>> trigsimp( v.norm() )
1
>>> v.norm(10)
(sin(x)**10 + cos(x)**10)**(1/10)
>>> A = Matrix([[1, 1], [1, 1]])
>>> A.norm(2)# Spectral norm (max of |Ax|/|x| under 2-vector-norm)
2
>>> A.norm(-2) # Inverse spectral norm (smallest singular value)
0
>>> A.norm() # Frobenius Norm
2
>>> Matrix([1, -2]).norm(oo)
2
>>> Matrix([-1, 2]).norm(-oo)
1
5.14. Matrices
669
SymPy Documentation, Release 0.7.6
normalized()
Return the normalized version of self.
See Also:
norm (page 669)
nullspace(simplify=False)
Returns list of vectors (Matrix objects) that span nullspace of self
permuteBkwd(perm)
Permute the rows of the matrix with the given permutation in reverse.
See Also:
permuteFwd (page 670)
Examples
>>> from sympy.matrices import eye
>>> M = eye(3)
>>> M.permuteBkwd([[0, 1], [0, 2]])
Matrix([
[0, 1, 0],
[0, 0, 1],
[1, 0, 0]])
permuteFwd(perm)
Permute the rows of the matrix with the given permutation.
See Also:
permuteBkwd (page 670)
Examples
>>> from sympy.matrices import eye
>>> M = eye(3)
>>> M.permuteFwd([[0, 1], [0, 2]])
Matrix([
[0, 0, 1],
[1, 0, 0],
[0, 1, 0]])
pinv()
Calculate the Moore-Penrose pseudoinverse of the matrix.
The Moore-Penrose pseudoinverse exists and is unique for any matrix. If the matrix
is invertible, the pseudoinverse is the same as the inverse.
See Also:
inv, pinv solve (page 671)
References
[R309] (page 1909)
670
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import Matrix
>>> Matrix([[1, 2, 3], [4, 5, 6]]).pinv()
Matrix([
[-17/18, 4/9],
[ -1/9, 1/9],
[ 13/18, -2/9]])
pinv solve(B, arbitrary matrix=None)
Solve Ax = B using the Moore-Penrose pseudoinverse.
There may be zero, one, or innite solutions. If one solution exists, it will be returned. If innite solutions exist, one will be returned based on the value of arbitrary matrix. If no solutions exist, the least-squares solution is returned.
Parameters B : Matrix
The right hand side of the equation to be solved for. Must have the
same number of rows as matrix A.
arbitrary matrix : Matrix
If the system is underdetermined (e.g. A has more columns than
rows), innite solutions are possible, in terms of an arbitrary matrix.
This parameter may be set to a specic matrix to use for that purpose;
if so, it must be the same shape as x, with as many rows as matrix A
has columns, and as many columns as matrix B. If left as None, an appropriate matrix containing dummy symbols in the form of wn m will
be used, with n and m being row and column position of each symbol.
Returns x : Matrix
The matrix that will satisfy Ax = B. Will have as many rows as matrix
A has columns, and as many columns as matrix B.
See Also:
lower triangular solve (page 668), upper triangular solve (page 677),
cholesky solve (page 651), diagonal solve (page 653), LDLsolve (page 646),
LUsolve (page 647), QRsolve (page 648), pinv (page 670)
Notes
This may return either exact solutions or least squares solutions. To determine
which, check A * A.pinv() * B == B. It will be True if exact solutions exist,
and False if only a least-squares solution exists. Be aware that the left hand side of
that equation may need to be simplied to correctly compare to the right hand side.
References
[R310] (page 1910)
5.14. Matrices
671
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import Matrix
>>> A = Matrix([[1, 2, 3], [4, 5, 6]])
>>> B = Matrix([7, 8])
>>> A.pinv_solve(B)
Matrix([
[ _w0_0/6 - _w1_0/3 + _w2_0/6 - 55/18],
[-_w0_0/3 + 2*_w1_0/3 - _w2_0/3 + 1/9],
[ _w0_0/6 - _w1_0/3 + _w2_0/6 + 59/18]])
>>> A.pinv_solve(B, arbitrary_matrix=Matrix([0, 0, 0]))
Matrix([
[-55/18],
[
1/9],
[ 59/18]])
print nonzero(symb=X)
Shows location of non-zero entries for fast shape lookup.
Examples
>>> from sympy.matrices import Matrix, eye
>>> m = Matrix(2, 3, lambda i, j: i*3+j)
>>> m
Matrix([
[0, 1, 2],
[3, 4, 5]])
>>> m.print_nonzero()
[ XX]
[XXX]
>>> m = eye(4)
>>> m.print_nonzero(x)
[x
]
[ x ]
[ x ]
[
x]
project(v)
Return the projection of self onto the line containing v.
Examples
>>> from sympy import Matrix, S, sqrt
>>> V = Matrix([sqrt(3)/2, S.Half])
>>> x = Matrix([[1, 0]])
>>> V.project(x)
Matrix([[sqrt(3)/2, 0]])
>>> V.project(-x)
Matrix([[sqrt(3)/2, 0]])
rank(iszerofunc=<function iszero at 0xa2e4aac>, simplify=False)
Returns the rank of a matrix
672
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>>
>>>
>>>
>>>
2
>>>
>>>
2
from sympy import Matrix
from sympy.abc import x
m = Matrix([[1, 2], [x, 1 - 1/x]])
m.rank()
n = Matrix(3, 3, range(1, 10))
n.rank()
replace(F, G, map=False)
Replaces Function F in Matrix entries with Function G.
Examples
>>> from sympy import symbols, Function, Matrix
>>> F, G = symbols(F, G, cls=Function)
>>> M = Matrix(2, 2, lambda i, j: F(i+j)) ; M
Matrix([
[F(0), F(1)],
[F(1), F(2)]])
>>> N = M.replace(F,G)
>>> N
Matrix([
[G(0), G(1)],
[G(1), G(2)]])
row insert(pos, mti)
Insert one or more rows at the given row position.
See Also:
row, col insert (page 652)
Examples
>>> from sympy import zeros, ones
>>> M = zeros(3)
>>> V = ones(1, 3)
>>> M.row_insert(1, V)
Matrix([
[0, 0, 0],
[1, 1, 1],
[0, 0, 0],
[0, 0, 0]])
row join(rhs)
Concatenates two matrices along selfs last and rhss rst column
See Also:
row, col join (page 652)
5.14. Matrices
673
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import zeros, ones
>>> M = zeros(3)
>>> V = ones(3, 1)
>>> M.row_join(V)
Matrix([
[0, 0, 0, 1],
[0, 0, 0, 1],
[0, 0, 0, 1]])
rref(iszerofunc=<function iszero at 0xa2e4aac>, simplify=False)
Return reduced row-echelon form of matrix and indices of pivot vars.
To simplify elements before nding nonzero pivots set simplify=True (to use the
default SymPy simplify function) or pass a custom simplify function.
Examples
>>> from sympy import Matrix
>>> from sympy.abc import x
>>> m = Matrix([[1, 2], [x, 1 - 1/x]])
>>> m.rref()
(Matrix([
[1, 0],
[0, 1]]), [0, 1])
shape
The shape (dimensions) of the matrix as the 2-tuple (rows, cols).
Examples
>>>
>>>
>>>
(2,
>>>
2
>>>
3
from sympy.matrices import zeros
M = zeros(2, 3)
M.shape
3)
M.rows
M.cols
simplify(ratio=1.7, measure=<function count ops at 0x9db2b54>)
Apply simplify to each element of the matrix.
Examples
>>> from sympy.abc import x, y
>>> from sympy import sin, cos
>>> from sympy.matrices import SparseMatrix
>>> SparseMatrix(1, 1, [x*sin(y)**2 + x*cos(y)**2])
Matrix([[x*sin(y)**2 + x*cos(y)**2]])
>>> _.simplify()
Matrix([[x]])
674
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
singular values()
Compute the singular values of a Matrix
See Also:
condition number (page 652)
Examples
>>> from sympy import Matrix, Symbol
>>> x = Symbol(x, real=True)
>>> A = Matrix([[0, 1, 0], [0, x, 0], [-1, 0, 0]])
>>> A.singular_values()
[sqrt(x**2 + 1), 1, 0]
solve(rhs, method=GE)
Return solution to self*soln = rhs using given inversion method.
For a list of possible inversion methods, see the .inv() docstring.
solve least squares(rhs, method=CH)
Return the least-square t to the data.
By default the cholesky solve routine is used (method=CH); other methods of matrix inversion can be used. To nd out which are available, see the docstring of the
.inv() method.
Examples
>>> from sympy.matrices import Matrix, ones
>>> A = Matrix([1, 2, 3])
>>> B = Matrix([2, 3, 4])
>>> S = Matrix(A.row_join(B))
>>> S
Matrix([
[1, 2],
[2, 3],
[3, 4]])
If each line of S represent coecients of Ax + By and x and y are [2, 3] then S*xy is:
>>> r = S*Matrix([2, 3]); r
Matrix([
[ 8],
[13],
[18]])
But lets add 1 to the middle value and then solve for the least-squares value of xy:
>>> xy = S.solve_least_squares(Matrix([8, 14, 18])); xy
Matrix([
[ 5/3],
[10/3]])
The error is given by S*xy - r:
5.14. Matrices
675
SymPy Documentation, Release 0.7.6
>>> S*xy - r
Matrix([
[1/3],
[1/3],
[1/3]])
>>> _.norm().n(2)
0.58
If a dierent xy is used, the norm will be higher:
>>> xy += ones(2, 1)/10
>>> (S*xy - r).norm().n(2)
1.5
subs(*args, **kwargs)
Return a new matrix with subs applied to each entry.
Examples
>>> from sympy.abc import x, y
>>> from sympy.matrices import SparseMatrix, Matrix
>>> SparseMatrix(1, 1, [x])
Matrix([[x]])
>>> _.subs(x, y)
Matrix([[y]])
>>> Matrix(_).subs(y, x)
Matrix([[x]])
table(printer, rowstart=[, rowend=], rowsep=n, colsep=, , align=right)
String form of Matrix as a table.
printer is the printer to use for on the elements (generally something like StrPrinter())
rowstart is the string used to start each row (by default [).
rowend is the string used to end each row (by default ]).
rowsep is the string used to separate rows (by default a newline).
colsep is the string used to separate columns (by default , ).
align denes how the elements are aligned. Must be one of left, right, or center.
You can also use <, >, and to mean the same thing, respectively.
This is used by the string printer for Matrix.
Examples
>>> from sympy import Matrix
>>> from sympy.printing.str import StrPrinter
>>> M = Matrix([[1, 2], [-33, 4]])
>>> printer = StrPrinter()
>>> M.table(printer)
[ 1, 2]\n[-33, 4]
>>> print(M.table(printer))
[ 1, 2]
[-33, 4]
676
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> print(M.table(printer, rowsep=,\n))
[ 1, 2],
[-33, 4]
>>> print([%s] % M.table(printer, rowsep=,\n))
[[ 1, 2],
[-33, 4]]
>>> print(M.table(printer, colsep= ))
[ 1 2]
[-33 4]
>>> print(M.table(printer, align=center))
[ 1 , 2]
[-33, 4]
>>> print(M.table(printer, rowstart={, rowend=}))
{ 1, 2}
{-33, 4}
upper triangular solve(rhs)
Solves Ax = B, where A is an upper triangular matrix.
See Also:
lower triangular solve (page 668), cholesky solve (page 651), diagonal solve
(page 653), LDLsolve (page 646), LUsolve (page 647), QRsolve (page 648),
pinv solve (page 671)
values()
Return non-zero values of self.
vec()
Return the Matrix converted into a one column matrix by stacking columns
See Also:
vech (page 677)
Examples
>>> from sympy import Matrix
>>> m=Matrix([[1, 3], [2, 4]])
>>> m
Matrix([
[1, 3],
[2, 4]])
>>> m.vec()
Matrix([
[1],
[2],
[3],
[4]])
vech(diagonal=True, check symmetry=True)
Return the unique elements of a symmetric Matrix as a one column matrix by stacking the elements in the lower triangle.
Arguments: diagonal include the diagonal cells of self or not check symmetry
checks symmetry of self but not completely reliably
See Also:
vec (page 677)
5.14. Matrices
677
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import Matrix
>>> m=Matrix([[1, 2], [2, 3]])
>>> m
Matrix([
[1, 2],
[2, 3]])
>>> m.vech()
Matrix([
[1],
[2],
[3]])
>>> m.vech(diagonal=False)
Matrix([[2]])
classmethod vstack(*args)
Return a matrix formed by joining args vertically (i.e. by repeated application of
col join).
Examples
>>> from sympy.matrices import Matrix, eye
>>> Matrix.vstack(eye(2), 2*eye(2))
Matrix([
[1, 0],
[0, 1],
[2, 0],
[0, 2]])
Matrix Exceptions Reference
class sympy.matrices.matrices.MatrixError
class sympy.matrices.matrices.ShapeError
Wrong matrix shape
class sympy.matrices.matrices.NonSquareMatrixError
Matrix Functions Reference
sympy.matrices.matrices.classof(A, B)
Get the type of the result when combining matrices of dierent types.
Currently the strategy is that immutability is contagious.
Examples
>>>
>>>
>>>
>>>
678
from sympy import Matrix, ImmutableMatrix
from sympy.matrices.matrices import classof
M = Matrix([[1, 2], [3, 4]]) # a Mutable Matrix
IM = ImmutableMatrix([[1, 2], [3, 4]])
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> classof(M, IM)
<class sympy.matrices.immutable.ImmutableMatrix>
sympy.matrices.dense.matrix multiply elementwise(A, B)
Return the Hadamard product (elementwise product) of A and B
>>> from sympy.matrices import matrix_multiply_elementwise
>>> from sympy.matrices import Matrix
>>> A = Matrix([[0, 1, 2], [3, 4, 5]])
>>> B = Matrix([[1, 10, 100], [100, 10, 1]])
>>> matrix_multiply_elementwise(A, B)
Matrix([
[ 0, 10, 200],
[300, 40,
5]])
See Also:
mul
sympy.matrices.dense.zeros(r, c=None, cls=None)
Returns a matrix of zeros with r rows and c columns; if c is omitted a square matrix will
be returned.
See Also:
ones, eye, diag
sympy.matrices.dense.ones(r, c=None)
Returns a matrix of ones with r rows and c columns; if c is omitted a square matrix will
be returned.
See Also:
zeros, eye, diag
sympy.matrices.dense.eye(n, cls=None)
Create square identity matrix n x n
See Also:
diag, zeros, ones
sympy.matrices.dense.diag(*values, **kwargs)
Create a sparse, diagonal matrix from a list of diagonal values.
See Also:
eye
Notes
When arguments are matrices they are tted in resultant matrix.
The returned matrix is a mutable, dense matrix. To make it a dierent type, send the
desired class for keyword cls.
Examples
5.14. Matrices
679
SymPy Documentation, Release 0.7.6
>>> from sympy.matrices import diag, Matrix, ones
>>> diag(1, 2, 3)
Matrix([
[1, 0, 0],
[0, 2, 0],
[0, 0, 3]])
>>> diag(*[1, 2, 3])
Matrix([
[1, 0, 0],
[0, 2, 0],
[0, 0, 3]])
The diagonal elements can be matrices; diagonal lling will continue on the diagonal
from the last element of the matrix:
>>> from sympy.abc import x, y, z
>>> a = Matrix([x, y, z])
>>> b = Matrix([[1, 2], [3, 4]])
>>> c = Matrix([[5, 6]])
>>> diag(a, 7, b, c)
Matrix([
[x, 0, 0, 0, 0, 0],
[y, 0, 0, 0, 0, 0],
[z, 0, 0, 0, 0, 0],
[0, 7, 0, 0, 0, 0],
[0, 0, 1, 2, 0, 0],
[0, 0, 3, 4, 0, 0],
[0, 0, 0, 0, 5, 6]])
When diagonal elements are lists, they will be treated as arguments to Matrix:
>>> diag([1, 2, 3], 4)
Matrix([
[1, 0],
[2, 0],
[3, 0],
[0, 4]])
>>> diag([[1, 2, 3]], 4)
Matrix([
[1, 2, 3, 0],
[0, 0, 0, 4]])
A given band o the diagonal can be made by padding with a vertical or horizontal kerning vector:
>>> hpad = ones(0, 2)
>>> vpad = ones(2, 0)
>>> diag(vpad, 1, 2, 3, hpad) + diag(hpad, 4, 5, 6, vpad)
Matrix([
[0, 0, 4, 0, 0],
[0, 0, 0, 5, 0],
[1, 0, 0, 0, 6],
[0, 2, 0, 0, 0],
[0, 0, 3, 0, 0]])
The type is mutable by default but can be made immutable by setting the mutable ag
to False:
680
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> type(diag(1))
<class sympy.matrices.dense.MutableDenseMatrix>
>>> from sympy.matrices import ImmutableMatrix
>>> type(diag(1, cls=ImmutableMatrix))
<class sympy.matrices.immutable.ImmutableMatrix>
sympy.matrices.dense.jordan cell(eigenval, n)
Create matrix of Jordan cell kind:
Examples
>>> from sympy.matrices import jordan_cell
>>> from sympy.abc import x
>>> jordan_cell(x, 4)
Matrix([
[x, 1, 0, 0],
[0, x, 1, 0],
[0, 0, x, 1],
[0, 0, 0, x]])
sympy.matrices.dense.hessian(f, varlist, constraints=[])
Compute Hessian matrix for a function f wrt parameters in varlist which may be given
as a sequence or a row/column vector. A list of constraints may optionally be given.
See Also:
sympy.matrices.mutable.Matrix.jacobian, wronskian
References
http://en.wikipedia.org/wiki/Hessian matrix
Examples
>>> from sympy import Function, hessian, pprint
>>> from sympy.abc import x, y
>>> f = Function(f)(x, y)
>>> g1 = Function(g)(x, y)
>>> g2 = x**2 + 3*y
>>> pprint(hessian(f, (x, y), [g1, g2]))
[
d
d
]
[
0
0
--(g(x, y))
--(g(x, y)) ]
[
dx
dy
]
[
]
[
0
0
2*x
3
]
[
]
[
2
2
]
[d
d
d
]
[--(g(x, y)) 2*x
---(f(x, y))
-----(f(x, y))]
[dx
2
dy dx
]
[
dx
]
[
]
[
2
2
]
[d
d
d
]
5.14. Matrices
681
SymPy Documentation, Release 0.7.6
[--(g(x, y))
[dy
[
-----(f(x, y))
dy dx
---(f(x, y)) ]
2
]
dy
]
sympy.matrices.dense.GramSchmidt(vlist, orthog=False)
Apply the Gram-Schmidt process to a set of vectors.
see: http://en.wikipedia.org/wiki/Gram%E2%80%93Schmidt process
sympy.matrices.dense.wronskian(functions, var, method=bareis)
Compute Wronskian for [] of functions
| f1
f2
| f1
f2
| .
.
W(f1, ..., fn) = | .
.
| .
.
| (n)
(n)
| D
(f1) D
(f2)
...
...
.
.
.
...
fn
fn
.
.
.
(n)
D
(fn)
|
|
|
|
|
|
|
see: http://en.wikipedia.org/wiki/Wronskian
See Also:
sympy.matrices.mutable.Matrix.jacobian, hessian
sympy.matrices.dense.casoratian(seqs, n, zero=True)
Given linear dierence operator L of order k and homogeneous equation Ly = 0 we
want to compute kernel of L, which is a set of k sequences: a(n), b(n), ... z(n).
Solutions of L are linearly independent i their Casoratian, denoted as C(a, b, ..., z), do
not vanish for n = 0.
Casoratian is dened by k x k determinant:
+ a(n)
b(n)
. . . z(n)
+
| a(n+1)
b(n+1)
. . . z(n+1)
|
|
.
.
.
.
|
|
.
.
.
.
|
|
.
.
.
.
|
+ a(n+k-1) b(n+k-1) . . . z(n+k-1) +
It proves very useful in rsolve hyper() where it is applied to a generating set of a recurrence to factor out linearly dependent solutions and return a basis:
>>> from sympy import Symbol, casoratian, factorial
>>> n = Symbol(n, integer=True)
Exponential and factorial are linearly independent:
>>> casoratian([2**n, factorial(n)], n) != 0
True
sympy.matrices.dense.randMatrix(r, c=None, min=0, max=99, seed=None, symmetric=False, percent=100)
Create random matrix with dimensions r x c. If c is omitted the matrix will be square.
If symmetric is True the matrix must be square. If percent is less than 100 then only
approximately the given percentage of elements will be non-zero.
682
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.matrices import randMatrix
>>> randMatrix(3)
[25, 45, 27]
[44, 54, 9]
[23, 96, 46]
>>> randMatrix(3, 2)
[87, 29]
[23, 37]
[90, 26]
>>> randMatrix(3, 3, 0, 2)
[0, 2, 0]
[2, 0, 1]
[0, 0, 1]
>>> randMatrix(3, symmetric=True)
[85, 26, 29]
[26, 71, 43]
[29, 43, 57]
>>> A = randMatrix(3, seed=1)
>>> B = randMatrix(3, seed=2)
>>> A == B
False
>>> A == randMatrix(3, seed=1)
True
>>> randMatrix(3, symmetric=True, percent=50)
[0, 68, 43]
[0, 68, 0]
[0, 91, 34]
Numpy Utility Functions Reference
sympy.matrices.dense.list2numpy(l, dtype=<type object>)
Converts python list of SymPy expressions to a NumPy array.
See Also:
matrix2numpy
sympy.matrices.dense.matrix2numpy(m, dtype=<type object>)
Converts SymPys matrix to a NumPy array.
See Also:
list2numpy
sympy.matrices.dense.symarray(prex, shape)
Create a numpy ndarray of symbols (as an object array).
The created symbols are named prefix i1 i2 ... You should thus provide a non-empty
prex if you want your symbols to be unique for dierent output arrays, as SymPy symbols with identical names are the same object.
Parameters prex : string
A prex prepended to the name of every symbol.
shape : int or tuple
Shape of the created array. If an int, the array is one-dimensional; for
more than one dimension the shape must be a tuple.
5.14. Matrices
683
SymPy Documentation, Release 0.7.6
Examples
These doctests require numpy.
>>> from sympy import symarray
>>> symarray(, 3)
[_0 _1 _2]
If you want multiple symarrays to contain distinct symbols, you must provide unique
prexes:
>>> a = symarray(, 3)
>>> b = symarray(, 3)
>>> a[0] == b[0]
True
>>> a = symarray(a, 3)
>>> b = symarray(b, 3)
>>> a[0] == b[0]
False
Creating symarrays with a prex:
>>> symarray(a, 3)
[a_0 a_1 a_2]
For more than one dimension, the shape must be given as a tuple:
>>> symarray(a, (2, 3))
[[a_0_0 a_0_1 a_0_2]
[a_1_0 a_1_1 a_1_2]]
>>> symarray(a, (2, 3, 2))
[[[a_0_0_0 a_0_0_1]
[a_0_1_0 a_0_1_1]
[a_0_2_0 a_0_2_1]]
[[a_1_0_0 a_1_0_1]
[a_1_1_0 a_1_1_1]
[a_1_2_0 a_1_2_1]]]
sympy.matrices.dense.rot axis1(theta)
Returns a rotation matrix for a rotation of theta (in radians) about the 1-axis.
See Also:
rot axis2 Returns a rotation matrix for a rotation of theta (in radians) about the 2-axis
rot axis3 Returns a rotation matrix for a rotation of theta (in radians) about the 3-axis
Examples
>>> from sympy import pi
>>> from sympy.matrices import rot_axis1
A rotation of pi/3 (60 degrees):
>>> theta = pi/3
>>> rot_axis1(theta)
Matrix([
684
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
[1,
0,
0],
[0,
1/2, sqrt(3)/2],
[0, -sqrt(3)/2,
1/2]])
If we rotate by pi/2 (90 degrees):
>>> rot_axis1(pi/2)
Matrix([
[1, 0, 0],
[0, 0, 1],
[0, -1, 0]])
sympy.matrices.dense.rot axis2(theta)
Returns a rotation matrix for a rotation of theta (in radians) about the 2-axis.
See Also:
rot axis1 Returns a rotation matrix for a rotation of theta (in radians) about the 1-axis
rot axis3 Returns a rotation matrix for a rotation of theta (in radians) about the 3-axis
Examples
>>> from sympy import pi
>>> from sympy.matrices import rot_axis2
A rotation of pi/3 (60 degrees):
>>> theta = pi/3
>>> rot_axis2(theta)
Matrix([
[
1/2, 0, -sqrt(3)/2],
[
0, 1,
0],
[sqrt(3)/2, 0,
1/2]])
If we rotate by pi/2 (90 degrees):
>>> rot_axis2(pi/2)
Matrix([
[0, 0, -1],
[0, 1, 0],
[1, 0, 0]])
sympy.matrices.dense.rot axis3(theta)
Returns a rotation matrix for a rotation of theta (in radians) about the 3-axis.
See Also:
rot axis1 Returns a rotation matrix for a rotation of theta (in radians) about the 1-axis
rot axis2 Returns a rotation matrix for a rotation of theta (in radians) about the 2-axis
Examples
>>> from sympy import pi
>>> from sympy.matrices import rot_axis3
5.14. Matrices
685
SymPy Documentation, Release 0.7.6
A rotation of pi/3 (60 degrees):
>>> theta = pi/3
>>> rot_axis3(theta)
Matrix([
[
1/2, sqrt(3)/2, 0],
[-sqrt(3)/2,
1/2, 0],
[
0,
0, 1]])
If we rotate by pi/2 (90 degrees):
>>> rot_axis3(pi/2)
Matrix([
[ 0, 1, 0],
[-1, 0, 0],
[ 0, 0, 1]])
sympy.matrices.matrices.a2idx(j, n=None)
Return integer after making positive and validating against n.
5.14.2 Dense Matrices
Matrix Class Reference
class sympy.matrices.dense.MutableDenseMatrix
col del(i)
Delete the given column.
See Also:
col, row del
Examples
>>> from sympy.matrices import eye
>>> M = eye(3)
>>> M.col_del(1)
>>> M
Matrix([
[1, 0],
[0, 0],
[0, 1]])
col op(j, f)
In-place operation on col j using two-arg functor whose args are interpreted as
(self[i, j], i).
See Also:
col, row op
686
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.matrices import eye
>>> M = eye(3)
>>> M.col_op(1, lambda v, i: v + 2*M[i, 0]); M
Matrix([
[1, 2, 0],
[0, 1, 0],
[0, 0, 1]])
col swap(i, j)
Swap the two given columns of the matrix in-place.
See Also:
col, row swap
Examples
>>> from sympy.matrices import Matrix
>>> M = Matrix([[1, 0], [1, 0]])
>>> M
Matrix([
[1, 0],
[1, 0]])
>>> M.col_swap(0, 1)
>>> M
Matrix([
[0, 1],
[0, 1]])
copyin list(key, value)
Copy in elements from a list.
Parameters key : slice
The section of this matrix to replace.
value : iterable
The iterable to copy values from.
See Also:
copyin matrix
Examples
>>> from sympy.matrices import eye
>>> I = eye(3)
>>> I[:2, 0] = [1, 2] # col
>>> I
Matrix([
[1, 0, 0],
[2, 1, 0],
[0, 0, 1]])
>>> I[1, :2] = [[3, 4]]
>>> I
5.14. Matrices
687
SymPy Documentation, Release 0.7.6
Matrix([
[1, 0, 0],
[3, 4, 0],
[0, 0, 1]])
copyin matrix(key, value)
Copy in values from a matrix into the given bounds.
Parameters key : slice
The section of this matrix to replace.
value : Matrix
The matrix to copy values from.
See Also:
copyin list
Examples
>>> from sympy.matrices import Matrix, eye
>>> M = Matrix([[0, 1], [2, 3], [4, 5]])
>>> I = eye(3)
>>> I[:3, :2] = M
>>> I
Matrix([
[0, 1, 0],
[2, 3, 0],
[4, 5, 1]])
>>> I[0, 1] = M
>>> I
Matrix([
[0, 0, 1],
[2, 2, 3],
[4, 4, 5]])
fill(value)
Fill the matrix with the scalar value.
See Also:
zeros, ones
row del(i)
Delete the given row.
See Also:
row, col del
Examples
>>> from sympy.matrices import eye
>>> M = eye(3)
>>> M.row_del(1)
>>> M
Matrix([
688
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
[1, 0, 0],
[0, 0, 1]])
row op(i, f)
In-place operation on row i using two-arg functor whose args are interpreted as
(self[i, j], j).
See Also:
row, zip row op, col op
Examples
>>> from sympy.matrices import eye
>>> M = eye(3)
>>> M.row_op(1, lambda v, j: v + 2*M[0, j]); M
Matrix([
[1, 0, 0],
[2, 1, 0],
[0, 0, 1]])
row swap(i, j)
Swap the two given rows of the matrix in-place.
See Also:
row, col swap
Examples
>>> from sympy.matrices import Matrix
>>> M = Matrix([[0, 1], [1, 0]])
>>> M
Matrix([
[0, 1],
[1, 0]])
>>> M.row_swap(0, 1)
>>> M
Matrix([
[1, 0],
[0, 1]])
simplify(ratio=1.7, measure=<function count ops at 0x9db2b54>)
Applies simplify to the elements of a matrix in place.
This is a shortcut for M.applyfunc(lambda x: simplify(x, ratio, measure))
See Also:
sympy.simplify.simplify.simplify (page 1327)
zip row op(i, k, f)
In-place operation on row i using two-arg functor whose args are interpreted as
(self[i, j], self[k, j]).
See Also:
row, row op, col op
5.14. Matrices
689
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.matrices import eye
>>> M = eye(3)
>>> M.zip_row_op(1, 0, lambda v, u: v + 2*u); M
Matrix([
[1, 0, 0],
[2, 1, 0],
[0, 0, 1]])
ImmutableMatrix Class Reference
class sympy.matrices.immutable.ImmutableMatrix
Create an immutable version of a matrix.
Examples
>>> from sympy import eye
>>> from sympy.matrices import ImmutableMatrix
>>> ImmutableMatrix(eye(3))
Matrix([
[1, 0, 0],
[0, 1, 0],
[0, 0, 1]])
>>> _[0, 0] = 42
Traceback (most recent call last):
...
TypeError: Cannot set values of ImmutableDenseMatrix
Traceback (most recent call last):
...
TypeError: Cannot set values of ImmutableDenseMatrix
ImmutableMatrix.C
By-element conjugation.
ImmutableMatrix.adjoint()
Conjugate transpose or Hermitian conjugation.
ImmutableMatrix.as mutable()
Returns a mutable version of this matrix
Examples
>>> from sympy import ImmutableMatrix
>>> X = ImmutableMatrix([[1, 2], [3, 4]])
>>> Y = X.as_mutable()
>>> Y[1, 1] = 5 # Can set values in Y
>>> Y
Matrix([
[1, 2],
[3, 5]])
ImmutableMatrix.equals(other, failing expression=False)
Applies equals to corresponding elements of the matrices, trying to prove that the
690
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
elements are equivalent, returning True if they are, False if any pair is not, and None
(or the rst failing expression if failing expression is True) if it cannot be decided if
the expressions are equivalent or not. This is, in general, an expensive operation.
See Also:
sympy.core.expr.equals
Examples
>>> from sympy.matrices import Matrix
>>> from sympy.abc import x
>>> from sympy import cos
>>> A = Matrix([x*(x - 1), 0])
>>> B = Matrix([x**2 - x, 0])
>>> A == B
False
>>> A.simplify() == B.simplify()
True
>>> A.equals(B)
True
>>> A.equals(2)
False
ImmutableMatrix.is zero
Checks if a matrix is a zero matrix.
A matrix is zero if every element is zero. A matrix need not be square to be considered zero. The empty matrix is zero by the principle of vacuous truth. For a matrix
that may or may not be zero (e.g. contains a symbol), this will be None
Examples
>>> from sympy import Matrix, zeros
>>> from sympy.abc import x
>>> a = Matrix([[0, 0], [0, 0]])
>>> b = zeros(3, 4)
>>> c = Matrix([[0, 1], [0, 0]])
>>> d = Matrix([])
>>> e = Matrix([[x, 0], [0, 0]])
>>> a.is_zero
True
>>> b.is_zero
True
>>> c.is_zero
False
>>> d.is_zero
True
>>> e.is_zero
5.14. Matrices
691
SymPy Documentation, Release 0.7.6
5.14.3 Sparse Matrices
SparseMatrix Class Reference
class sympy.matrices.sparse.SparseMatrix(*args)
A sparse matrix (a matrix with a large number of zero elements).
See Also:
sympy.matrices.dense.Matrix
Examples
>>> from sympy.matrices import SparseMatrix
>>> SparseMatrix(2, 2, range(4))
Matrix([
[0, 1],
[2, 3]])
>>> SparseMatrix(2, 2, {(1, 1): 2})
Matrix([
[0, 0],
[0, 2]])
CL
Alternate faster representation
LDLdecomposition()
Returns the LDL Decomposition (matrices L and D) of matrix A, such that L * D *
L.T == A. A must be a square, symmetric, positive-denite and non-singular.
This method eliminates the use of square root and ensures that all the diagonal
entries of L are 1.
Examples
>>> from sympy.matrices import SparseMatrix
>>> A = SparseMatrix(((25, 15, -5), (15, 18, 0), (-5, 0, 11)))
>>> L, D = A.LDLdecomposition()
>>> L
Matrix([
[
1,
0, 0],
[ 3/5,
1, 0],
[-1/5, 1/3, 1]])
>>> D
Matrix([
[25, 0, 0],
[ 0, 9, 0],
[ 0, 0, 9]])
>>> L * D * L.T == A
True
RL
Alternate faster representation
add(other)
Add two sparse matrices with dictionary representation.
692
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
See Also:
multiply (page 696)
Examples
>>> from sympy.matrices import SparseMatrix, eye, ones
>>> SparseMatrix(eye(3)).add(SparseMatrix(ones(3)))
Matrix([
[2, 1, 1],
[1, 2, 1],
[1, 1, 2]])
>>> SparseMatrix(eye(3)).add(-SparseMatrix(eye(3)))
Matrix([
[0, 0, 0],
[0, 0, 0],
[0, 0, 0]])
Only the non-zero elements are stored, so the resulting dictionary that is used to
represent the sparse matrix is empty: >>> . smat {}
applyfunc(f)
Apply a function to each element of the matrix.
Examples
>>> from sympy.matrices import SparseMatrix
>>> m = SparseMatrix(2, 2, lambda i, j: i*2+j)
>>> m
Matrix([
[0, 1],
[2, 3]])
>>> m.applyfunc(lambda i: 2*i)
Matrix([
[0, 2],
[4, 6]])
as immutable()
Returns an Immutable version of this Matrix.
as mutable()
Returns a mutable version of this matrix.
Examples
>>> from sympy import ImmutableMatrix
>>> X = ImmutableMatrix([[1, 2], [3, 4]])
>>> Y = X.as_mutable()
>>> Y[1, 1] = 5 # Can set values in Y
>>> Y
Matrix([
[1, 2],
[3, 5]])
5.14. Matrices
693
SymPy Documentation, Release 0.7.6
cholesky()
Returns the Cholesky decomposition L of a matrix A such that L * L.T = A
A must be a square, symmetric, positive-denite and non-singular matrix
Examples
>>> from sympy.matrices import SparseMatrix
>>> A = SparseMatrix(((25,15,-5),(15,18,0),(-5,0,11)))
>>> A.cholesky()
Matrix([
[ 5, 0, 0],
[ 3, 3, 0],
[-1, 1, 3]])
>>> A.cholesky() * A.cholesky().T == A
True
col(j)
Returns column j from self as a column vector.
See Also:
row (page 697), col list (page 694)
Examples
>>> from sympy.matrices import SparseMatrix
>>> a = SparseMatrix(((1, 2), (3, 4)))
>>> a.col(0)
Matrix([
[1],
[3]])
col list()
Returns a column-sorted list of non-zero elements of the matrix.
See Also:
col op, row list (page 697)
Examples
>>> from sympy.matrices import SparseMatrix
>>> a=SparseMatrix(((1, 2), (3, 4)))
>>> a
Matrix([
[1, 2],
[3, 4]])
>>> a.CL
[(0, 0, 1), (1, 0, 3), (0, 1, 2), (1, 1, 4)]
extract(rowsList, colsList)
Return a submatrix by specifying a list of rows and columns. Negative indices can
be given. All indices must be in the range -n <= i < n where n is the number of rows
or columns.
694
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import Matrix
>>> m = Matrix(4, 3, range(12))
>>> m
Matrix([
[0, 1, 2],
[3, 4, 5],
[6, 7, 8],
[9, 10, 11]])
>>> m.extract([0, 1, 3], [0, 1])
Matrix([
[0, 1],
[3, 4],
[9, 10]])
Rows or columns can be repeated:
>>> m.extract([0, 0, 1], [-1])
Matrix([
[2],
[2],
[5]])
Every other row can be taken by using range to provide the indices:
>>> m.extract(range(0, m.rows, 2), [-1])
Matrix([
[2],
[8]])
classmethod eye(n)
Return an n x n identity matrix.
has(*patterns)
Test whether any subexpression matches any of the patterns.
Examples
>>> from sympy import SparseMatrix, Float
>>> from sympy.abc import x, y
>>> A = SparseMatrix(((1, x), (0.2, 3)))
>>> A.has(x)
True
>>> A.has(y)
False
>>> A.has(Float)
True
is hermitian
Checks if the matrix is Hermitian.
In a Hermitian matrix element i,j is the complex conjugate of element j,i.
5.14. Matrices
695
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.matrices import SparseMatrix
>>> from sympy import I
>>> from sympy.abc import x
>>> a = SparseMatrix([[1, I], [-I, 1]])
>>> a
Matrix([
[ 1, I],
[-I, 1]])
>>> a.is_hermitian
True
>>> a[0, 0] = 2*I
>>> a.is_hermitian
False
>>> a[0, 0] = x
>>> a.is_hermitian
>>> a[0, 1] = a[1, 0]*I
>>> a.is_hermitian
False
is symmetric(simplify=True)
Return True if self is symmetric.
Examples
>>> from sympy.matrices import SparseMatrix, eye
>>> M = SparseMatrix(eye(3))
>>> M.is_symmetric()
True
>>> M[0, 2] = 1
>>> M.is_symmetric()
False
liupc()
Lius algorithm, for pre-determination of the Elimination Tree of the given matrix,
used in row-based symbolic Cholesky factorization.
References
Symbolic Sparse Cholesky Factorization using Elimination Trees, Jeroen Van Grondelle (1999) http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.39.7582
Examples
>>> from sympy.matrices import SparseMatrix
>>> S = SparseMatrix([
... [1, 0, 3, 2],
... [0, 0, 1, 0],
... [4, 0, 0, 5],
... [0, 6, 7, 0]])
>>> S.liupc()
([[0], [], [0], [1, 2]], [4, 3, 4, 4])
696
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
multiply(other)
Fast multiplication exploiting the sparsity of the matrix.
See Also:
add (page 692)
Examples
>>> from sympy.matrices import SparseMatrix, ones
>>> A, B = SparseMatrix(ones(4, 3)), SparseMatrix(ones(3, 4))
>>> A.multiply(B) == 3*ones(4)
True
nnz()
Returns the number of non-zero elements in Matrix.
reshape(rows, cols)
Reshape matrix while retaining original size.
Examples
>>> from sympy.matrices import SparseMatrix
>>> S = SparseMatrix(4, 2, range(8))
>>> S.reshape(2, 4)
Matrix([
[0, 1, 2, 3],
[4, 5, 6, 7]])
row(i)
Returns column i from self as a row vector.
See Also:
col (page 694), row list (page 697)
Examples
>>> from sympy.matrices import SparseMatrix
>>> a = SparseMatrix(((1, 2), (3, 4)))
>>> a.row(0)
Matrix([[1, 2]])
row list()
Returns a row-sorted list of non-zero elements of the matrix.
See Also:
row op, col list (page 694)
Examples
5.14. Matrices
697
SymPy Documentation, Release 0.7.6
>>> from sympy.matrices import SparseMatrix
>>> a = SparseMatrix(((1, 2), (3, 4)))
>>> a
Matrix([
[1, 2],
[3, 4]])
>>> a.RL
[(0, 0, 1), (0, 1, 2), (1, 0, 3), (1, 1, 4)]
row structure symbolic cholesky()
Symbolic cholesky factorization, for pre-determination of the non-zero structure of
the Cholesky factororization.
References
Symbolic Sparse Cholesky Factorization using Elimination Trees, Jeroen Van Grondelle (1999) http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.39.7582
Examples
>>> from sympy.matrices import SparseMatrix
>>> S = SparseMatrix([
... [1, 0, 3, 2],
... [0, 0, 1, 0],
... [4, 0, 0, 5],
... [0, 6, 7, 0]])
>>> S.row_structure_symbolic_cholesky()
[[0], [], [0], [1, 2]]
scalar multiply(scalar)
Scalar element-wise multiplication
solve(rhs, method=LDL)
Return solution to self*soln = rhs using given inversion method.
For a list of possible inversion methods, see the .inv() docstring.
solve least squares(rhs, method=LDL)
Return the least-square t to the data.
By default the cholesky solve routine is used (method=CH); other methods of matrix inversion can be used. To nd out which are available, see the docstring of the
.inv() method.
Examples
>>> from sympy.matrices import SparseMatrix, Matrix, ones
>>> A = Matrix([1, 2, 3])
>>> B = Matrix([2, 3, 4])
>>> S = SparseMatrix(A.row_join(B))
>>> S
Matrix([
[1, 2],
[2, 3],
[3, 4]])
698
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
If each line of S represent coecients of Ax + By and x and y are [2, 3] then S*xy is:
>>> r = S*Matrix([2, 3]); r
Matrix([
[ 8],
[13],
[18]])
But lets add 1 to the middle value and then solve for the least-squares value of xy:
>>> xy = S.solve_least_squares(Matrix([8, 14, 18])); xy
Matrix([
[ 5/3],
[10/3]])
The error is given by S*xy - r:
>>> S*xy - r
Matrix([
[1/3],
[1/3],
[1/3]])
>>> _.norm().n(2)
0.58
If a dierent xy is used, the norm will be higher:
>>> xy += ones(2, 1)/10
>>> (S*xy - r).norm().n(2)
1.5
tolist()
Convert this sparse matrix into a list of nested Python lists.
Examples
>>> from sympy.matrices import SparseMatrix, ones
>>> a = SparseMatrix(((1, 2), (3, 4)))
>>> a.tolist()
[[1, 2], [3, 4]]
When there are no rows then it will not be possible to tell how many columns were
in the original matrix:
>>> SparseMatrix(ones(0, 3)).tolist()
[]
classmethod zeros(r, c=None)
Return an r x c matrix of zeros, square if c is omitted.
class sympy.matrices.sparse.MutableSparseMatrix(*args)
col del(k)
Delete the given column of the matrix.
See Also:
row del (page 701)
5.14. Matrices
699
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.matrices import SparseMatrix
>>> M = SparseMatrix([[0, 0], [0, 1]])
>>> M
Matrix([
[0, 0],
[0, 1]])
>>> M.col_del(0)
>>> M
Matrix([
[0],
[1]])
col join(other)
Returns B augmented beneath A (row-wise joining):
[A]
[B]
Examples
>>> from sympy import SparseMatrix, Matrix, ones
>>> A = SparseMatrix(ones(3))
>>> A
Matrix([
[1, 1, 1],
[1, 1, 1],
[1, 1, 1]])
>>> B = SparseMatrix.eye(3)
>>> B
Matrix([
[1, 0, 0],
[0, 1, 0],
[0, 0, 1]])
>>> C = A.col_join(B); C
Matrix([
[1, 1, 1],
[1, 1, 1],
[1, 1, 1],
[1, 0, 0],
[0, 1, 0],
[0, 0, 1]])
>>> C == A.col_join(Matrix(B))
True
Joining along columns is the same as appending rows at the end of the matrix:
>>> C == A.row_insert(A.rows, Matrix(B))
True
col op(j, f)
In-place operation on col j using two-arg functor whose args are interpreted as
(self[i, j], i) for i in range(self.rows).
700
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.matrices import SparseMatrix
>>> M = SparseMatrix.eye(3)*2
>>> M[1, 0] = -1
>>> M.col_op(1, lambda v, i: v + 2*M[i, 0]); M
Matrix([
[ 2, 4, 0],
[-1, 0, 0],
[ 0, 0, 2]])
col swap(i, j)
Swap, in place, columns i and j.
Examples
>>> from sympy.matrices import SparseMatrix
>>> S = SparseMatrix.eye(3); S[2, 1] = 2
>>> S.col_swap(1, 0); S
Matrix([
[0, 1, 0],
[1, 0, 0],
[2, 0, 1]])
fill(value)
Fill self with the given value.
Notes
Unless many values are going to be deleted (i.e. set to zero) this will create a matrix
that is slower than a dense matrix in operations.
Examples
>>> from sympy.matrices import SparseMatrix
>>> M = SparseMatrix.zeros(3); M
Matrix([
[0, 0, 0],
[0, 0, 0],
[0, 0, 0]])
>>> M.fill(1); M
Matrix([
[1, 1, 1],
[1, 1, 1],
[1, 1, 1]])
row del(k)
Delete the given row of the matrix.
See Also:
col del (page 699)
5.14. Matrices
701
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.matrices import SparseMatrix
>>> M = SparseMatrix([[0, 0], [0, 1]])
>>> M
Matrix([
[0, 0],
[0, 1]])
>>> M.row_del(0)
>>> M
Matrix([[0, 1]])
row join(other)
Returns B appended after A (column-wise augmenting):
[A B]
Examples
>>> from sympy import SparseMatrix, Matrix
>>> A = SparseMatrix(((1, 0, 1), (0, 1, 0), (1, 1, 0)))
>>> A
Matrix([
[1, 0, 1],
[0, 1, 0],
[1, 1, 0]])
>>> B = SparseMatrix(((1, 0, 0), (0, 1, 0), (0, 0, 1)))
>>> B
Matrix([
[1, 0, 0],
[0, 1, 0],
[0, 0, 1]])
>>> C = A.row_join(B); C
Matrix([
[1, 0, 1, 1, 0, 0],
[0, 1, 0, 0, 1, 0],
[1, 1, 0, 0, 0, 1]])
>>> C == A.row_join(Matrix(B))
True
Joining at row ends is the same as appending columns at the end of the matrix:
>>> C == A.col_insert(A.cols, B)
True
row op(i, f)
In-place operation on row i using two-arg functor whose args are interpreted as
(self[i, j], j).
See Also:
row, zip row op (page 703), col op (page 700)
702
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.matrices import SparseMatrix
>>> M = SparseMatrix.eye(3)*2
>>> M[0, 1] = -1
>>> M.row_op(1, lambda v, j: v + 2*M[0, j]); M
Matrix([
[2, -1, 0],
[4, 0, 0],
[0, 0, 2]])
row swap(i, j)
Swap, in place, columns i and j.
Examples
>>> from sympy.matrices import SparseMatrix
>>> S = SparseMatrix.eye(3); S[2, 1] = 2
>>> S.row_swap(1, 0); S
Matrix([
[0, 1, 0],
[1, 0, 0],
[0, 2, 1]])
zip row op(i, k, f)
In-place operation on row i using two-arg functor whose args are interpreted as
(self[i, j], self[k, j]).
See Also:
row, row op (page 702), col op (page 700)
Examples
>>> from sympy.matrices import SparseMatrix
>>> M = SparseMatrix.eye(3)*2
>>> M[0, 1] = -1
>>> M.zip_row_op(1, 0, lambda v, u: v + 2*u); M
Matrix([
[2, -1, 0],
[4, 0, 0],
[0, 0, 2]])
ImmutableSparseMatrix Class Reference
class sympy.matrices.immutable.ImmutableSparseMatrix(*args)
Create an immutable version of a sparse matrix.
Examples
5.14. Matrices
703
SymPy Documentation, Release 0.7.6
>>> from sympy import eye
>>> from sympy.matrices.immutable import ImmutableSparseMatrix
>>> ImmutableSparseMatrix(1, 1, {})
Matrix([[0]])
>>> ImmutableSparseMatrix(eye(3))
Matrix([
[1, 0, 0],
[0, 1, 0],
[0, 0, 1]])
>>> _[0, 0] = 42
Traceback (most recent call last):
...
TypeError: Cannot set values of ImmutableSparseMatrix
>>> _.shape
(3, 3)
Traceback (most recent call last):
...
TypeError: Cannot set values of ImmutableSparseMatrix
subs(*args, **kwargs)
Return a new matrix with subs applied to each entry.
Examples
>>> from sympy.abc import x, y
>>> from sympy.matrices import SparseMatrix, Matrix
>>> SparseMatrix(1, 1, [x])
Matrix([[x]])
>>> _.subs(x, y)
Matrix([[y]])
>>> Matrix(_).subs(y, x)
Matrix([[x]])
5.14.4 Immutable Matrices
The standard Matrix class in SymPy is mutable. This is important for performance reasons
but means that standard matrices can not interact well with the rest of SymPy. This is because
the Basic object, from which most SymPy classes inherit, is immutable.
The mission of the ImmutableMatrix class is to bridge the tension between performance/mutability and safety/immutability. Immutable matrices can do almost everything that
normal matrices can do but they inherit from Basic and can thus interact more naturally with
the rest of SymPy. ImmutableMatrix also inherits from MatrixExpr, allowing it to interact
freely with SymPys Matrix Expression module.
You can turn any Matrix-like object into an ImmutableMatrix by calling the constructor
>>> from sympy import Matrix, ImmutableMatrix
>>> M = Matrix([[1, 2, 3], [4, 5, 6], [7, 8, 9]])
>>> M[1, 1] = 0
>>> IM = ImmutableMatrix(M)
>>> IM
Matrix([
[1, 2, 3],
[4, 0, 6],
704
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
[7, 8, 9]])
>>> IM[1, 1] = 5
Traceback (most recent
...
TypeError: Can not set
Traceback (most recent
...
TypeError: Can not set
call last):
values in Immutable Matrix. Use Matrix instead.
call last):
values in Immutable Matrix. Use Matrix instead.
ImmutableMatrix Class Reference
class sympy.matrices.immutable.ImmutableMatrix
Create an immutable version of a matrix.
Examples
>>> from sympy import eye
>>> from sympy.matrices import ImmutableMatrix
>>> ImmutableMatrix(eye(3))
Matrix([
[1, 0, 0],
[0, 1, 0],
[0, 0, 1]])
>>> _[0, 0] = 42
Traceback (most recent call last):
...
TypeError: Cannot set values of ImmutableDenseMatrix
Traceback (most recent call last):
...
TypeError: Cannot set values of ImmutableDenseMatrix
C
By-element conjugation.
adjoint()
Conjugate transpose or Hermitian conjugation.
as mutable()
Returns a mutable version of this matrix
Examples
>>> from sympy import ImmutableMatrix
>>> X = ImmutableMatrix([[1, 2], [3, 4]])
>>> Y = X.as_mutable()
>>> Y[1, 1] = 5 # Can set values in Y
>>> Y
Matrix([
[1, 2],
[3, 5]])
equals(other, failing expression=False)
Applies equals to corresponding elements of the matrices, trying to prove that the
elements are equivalent, returning True if they are, False if any pair is not, and None
5.14. Matrices
705
SymPy Documentation, Release 0.7.6
(or the rst failing expression if failing expression is True) if it cannot be decided if
the expressions are equivalent or not. This is, in general, an expensive operation.
See Also:
sympy.core.expr.equals
Examples
>>> from sympy.matrices import Matrix
>>> from sympy.abc import x
>>> from sympy import cos
>>> A = Matrix([x*(x - 1), 0])
>>> B = Matrix([x**2 - x, 0])
>>> A == B
False
>>> A.simplify() == B.simplify()
True
>>> A.equals(B)
True
>>> A.equals(2)
False
is zero
Checks if a matrix is a zero matrix.
A matrix is zero if every element is zero. A matrix need not be square to be considered zero. The empty matrix is zero by the principle of vacuous truth. For a matrix
that may or may not be zero (e.g. contains a symbol), this will be None
Examples
>>> from sympy import Matrix, zeros
>>> from sympy.abc import x
>>> a = Matrix([[0, 0], [0, 0]])
>>> b = zeros(3, 4)
>>> c = Matrix([[0, 1], [0, 0]])
>>> d = Matrix([])
>>> e = Matrix([[x, 0], [0, 0]])
>>> a.is_zero
True
>>> b.is_zero
True
>>> c.is_zero
False
>>> d.is_zero
True
>>> e.is_zero
5.14.5 Matrix Expressions
The Matrix expression module allows users to write down statements like
706
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> from sympy import MatrixSymbol, Matrix
>>> X = MatrixSymbol(X, 3, 3)
>>> Y = MatrixSymbol(Y, 3, 3)
>>> (X.T*X).I*Y
X^-1*X^-1*Y
>>> Matrix(X)
Matrix([
[X[0, 0], X[0, 1], X[0, 2]],
[X[1, 0], X[1, 1], X[1, 2]],
[X[2, 0], X[2, 1], X[2, 2]]])
>>> (X*Y)[1, 2]
X[1, 0]*Y[0, 2] + X[1, 1]*Y[1, 2] + X[1, 2]*Y[2, 2]
where X and Y are MatrixSymbol (page 708)s rather than scalar symbols.
Matrix Expressions Core Reference
class sympy.matrices.expressions.MatrixExpr
Superclass for Matrix Expressions
MatrixExprs represent abstract matrices, linear transformations represented within a
particular basis.
Examples
>>>
>>>
>>>
>>>
from sympy import MatrixSymbol
A = MatrixSymbol(A, 3, 3)
y = MatrixSymbol(y, 3, 1)
x = (A.T*A).I * A * y
Attributes
is Identity
T
Matrix transposition.
as explicit()
Returns a dense Matrix with elements represented explicitly
Returns an object of type ImmutableMatrix.
See Also:
as mutable (page 708) returns mutable Matrix type
Examples
5.14. Matrices
707
SymPy Documentation, Release 0.7.6
>>> from sympy import Identity
>>> I = Identity(3)
>>> I
I
>>> I.as_explicit()
Matrix([
[1, 0, 0],
[0, 1, 0],
[0, 0, 1]])
as mutable()
Returns a dense, mutable matrix with elements represented explicitly
See Also:
as explicit (page 707) returns ImmutableMatrix
Examples
>>> from sympy import Identity
>>> I = Identity(3)
>>> I
I
>>> I.shape
(3, 3)
>>> I.as_mutable()
Matrix([
[1, 0, 0],
[0, 1, 0],
[0, 0, 1]])
equals(other)
Test elementwise equality between matrices, potentially of dierent types
>>> from sympy import Identity, eye
>>> Identity(3).equals(eye(3))
True
class sympy.matrices.expressions.MatrixSymbol
Symbolic representation of a Matrix object
Creates a SymPy Symbol to represent a Matrix. This matrix has a shape and can be
included in Matrix Expressions
>>>
>>>
>>>
>>>
(3,
>>>
I +
from sympy import MatrixSymbol, Identity
A = MatrixSymbol(A, 3, 4) # A 3 by 4 Matrix
B = MatrixSymbol(B, 4, 3) # A 4 by 3 Matrix
A.shape
4)
2*A*B + Identity(3)
2*A*B
Attributes
is Identity
708
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
class sympy.matrices.expressions.MatAdd
A Sum of Matrix Expressions
MatAdd inherits from and operates like SymPy Add
>>>
>>>
>>>
>>>
>>>
A +
from sympy import MatAdd, MatrixSymbol
A = MatrixSymbol(A, 5, 5)
B = MatrixSymbol(B, 5, 5)
C = MatrixSymbol(C, 5, 5)
MatAdd(A, B, C)
B + C
Attributes
is Identity
class sympy.matrices.expressions.MatMul
A product of matrix expressions
Examples
>>> from sympy import MatMul, MatrixSymbol
>>> A = MatrixSymbol(A, 5, 4)
>>> B = MatrixSymbol(B, 4, 3)
>>> C = MatrixSymbol(C, 3, 6)
>>> MatMul(A, B, C)
A*B*C
Attributes
is Identity
class sympy.matrices.expressions.MatPow
Attributes
is Identity
class sympy.matrices.expressions.Inverse
The multiplicative inverse of a matrix expression
This is a symbolic object that simply stores its argument without evaluating it. To actually
compute the inverse, use the .inverse() method of matrices.
Examples
>>> from sympy import MatrixSymbol, Inverse
>>> A = MatrixSymbol(A, 3, 3)
>>> B = MatrixSymbol(B, 3, 3)
>>> Inverse(A)
A^-1
>>> A.inverse() == Inverse(A)
5.14. Matrices
709
SymPy Documentation, Release 0.7.6
True
>>> (A*B).inverse()
B^-1*A^-1
>>> Inverse(A*B)
(A*B)^-1
Attributes
is Identity
class sympy.matrices.expressions.Transpose
The transpose of a matrix expression.
This is a symbolic object that simply stores its argument without evaluating it. To actually
compute the transpose, use the transpose() function, or the .T attribute of matrices.
Examples
>>> from sympy.matrices import MatrixSymbol, Transpose
>>> from sympy.functions import transpose
>>> A = MatrixSymbol(A, 3, 5)
>>> B = MatrixSymbol(B, 5, 3)
>>> Transpose(A)
A
>>> A.T == transpose(A) == Transpose(A)
True
>>> Transpose(A*B)
(A*B)
>>> transpose(A*B)
B*A
Attributes
is Identity
class sympy.matrices.expressions.Trace
Matrix Trace
Represents the trace of a matrix expression.
>>> from sympy import MatrixSymbol, Trace, eye
>>> A = MatrixSymbol(A, 3, 3)
>>> Trace(A)
Trace(A)
See Also: trace
class sympy.matrices.expressions.FunctionMatrix
Represents a Matrix using a function (Lambda)
This class is an alternative to SparseMatrix
710
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> from sympy import FunctionMatrix, symbols, Lambda, MatMul, Matrix
>>> i, j = symbols(i,j)
>>> X = FunctionMatrix(3, 3, Lambda((i, j), i + j))
>>> Matrix(X)
Matrix([
[0, 1, 2],
[1, 2, 3],
[2, 3, 4]])
>>> Y = FunctionMatrix(1000, 1000, Lambda((i, j), i + j))
>>> isinstance(Y*Y, MatMul) # this is an expression object
True
>>> (Y**2)[10,10] # So this is evaluated lazily
342923500
Attributes
is Identity
class sympy.matrices.expressions.Identity
The Matrix Identity I - multiplicative identity
>>>
>>>
>>>
>>>
A
from sympy.matrices import Identity, MatrixSymbol
A = MatrixSymbol(A, 3, 5)
I = Identity(3)
I*A
class sympy.matrices.expressions.ZeroMatrix
The Matrix Zero 0 - additive identity
>>>
>>>
>>>
>>>
A
>>>
0
from sympy import MatrixSymbol, ZeroMatrix
A = MatrixSymbol(A, 3, 5)
Z = ZeroMatrix(3, 5)
A+Z
Z*A.T
Attributes
is Identity
Block Matrices
Block matrices allow you to construct larger matrices out of smaller sub-blocks. They can
work with MatrixExpr (page 707) or ImmutableMatrix objects.
class sympy.matrices.expressions.blockmatrix.BlockMatrix
A BlockMatrix is a Matrix composed of other smaller, submatrices
5.14. Matrices
711
SymPy Documentation, Release 0.7.6
The submatrices are stored in a SymPy Matrix object but accessed as part of a Matrix
Expression
>>> from sympy import (MatrixSymbol, BlockMatrix, symbols,
...
Identity, ZeroMatrix, block_collapse)
>>> n,m,l = symbols(n m l)
>>> X = MatrixSymbol(X, n, n)
>>> Y = MatrixSymbol(Y, m ,m)
>>> Z = MatrixSymbol(Z, n, m)
>>> B = BlockMatrix([[X, Z], [ZeroMatrix(m,n), Y]])
>>> print(B)
Matrix([
[X, Z],
[0, Y]])
>>> C = BlockMatrix([[Identity(n), Z]])
>>> print(C)
Matrix([[I, Z]])
>>> print(block_collapse(C*B))
Matrix([[X, Z*Y + Z]])
transpose()
Return transpose of matrix.
Examples
>>> from sympy import MatrixSymbol, BlockMatrix, ZeroMatrix
>>> from sympy.abc import l, m, n
>>> X = MatrixSymbol(X, n, n)
>>> Y = MatrixSymbol(Y, m ,m)
>>> Z = MatrixSymbol(Z, n, m)
>>> B = BlockMatrix([[X, Z], [ZeroMatrix(m,n), Y]])
>>> B.transpose()
Matrix([
[X, 0],
[Z, Y]])
>>> _.transpose()
Matrix([
[X, Z],
[0, Y]])
class sympy.matrices.expressions.blockmatrix.BlockDiagMatrix
A BlockDiagMatrix is a BlockMatrix with matrices only along the diagonal
>>> from sympy import MatrixSymbol, BlockDiagMatrix, symbols, Identity
>>> n,m,l = symbols(n m l)
>>> X = MatrixSymbol(X, n, n)
>>> Y = MatrixSymbol(Y, m ,m)
>>> BlockDiagMatrix(X, Y)
Matrix([
[X, 0],
[0, Y]])
sympy.matrices.expressions.blockmatrix.block collapse(expr)
Evaluates a block matrix expression
712
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> from sympy import MatrixSymbol, BlockMatrix, symbols,
>>> n,m,l = symbols(n m l)
>>> X = MatrixSymbol(X, n, n)
>>> Y = MatrixSymbol(Y, m ,m)
>>> Z = MatrixSymbol(Z, n, m)
>>> B = BlockMatrix([[X, Z], [ZeroMatrix(m, n), Y]])
>>> print(B)
Matrix([
[X, Z],
[0, Y]])
Identity, Ma
>>> C = BlockMatrix([[Identity(n), Z]])
>>> print(C)
Matrix([[I, Z]])
>>> print(block_collapse(C*B))
Matrix([[X, Z*Y + Z]])
5.15 Welcome to mpmaths documentation!
Mpmath is a Python library for arbitrary-precision oating-point arithmetic. For general information about mpmath, see the project website http://code.google.com/p/mpmath/
These documentation pages include general information as well as docstring listing with
extensive use of examples that can be run in the interactive Python interpreter. For
quick access to the docstrings of individual functions, use the index listing, or type
help(mpmath.function name) in the Python interactive prompt.
5.15.1 Introduction
Setting up mpmath
Download and installation
Installer The mpmath setup les can be downloaded from the mpmath download page or
the Python Package Index. Download the source package (available as both .zip and .tar.gz),
extract it, open the extracted directory, and run
python setup.py install
If you are using Windows, you can download the binary installer
mpmath-(version).win32.exe
from the mpmath website or the Python Package Index. Run the installer and follow the
instructions.
Using setuptools If you have setuptools installed, you can download and install mpmath in
one step by running:
easy install mpmath
or
python -m easy install mpmath
5.15. Welcome to mpmaths documentation!
713
SymPy Documentation, Release 0.7.6
If you have an old version of mpmath installed already, you may have to pass easy install
the -U ag to force an upgrade.
Debian/Ubuntu
Debian and Ubuntu users can install mpmath with
sudo apt-get install python-mpmath
See debian and ubuntu package information; please verify that you are getting the latest
version.
OpenSUSE Mpmath is provided in the Science repository for all recent versions
of openSUSE. To add this repository to the YAST software management tool, see
http://en.opensuse.org/SDB:Add package repositories
Look up http://download.opensuse.org/repositories/science/ for a list of supported OpenSUSE
versions and use http://download.opensuse.org/repositories/science/openSUSE 11.1/ (or accordingly for your OpenSUSE version) as the repository URL for YAST.
Current development version See http://code.google.com/p/mpmath/source/checkout for
instructions on how to check out the mpmath Subversion repository. The source code can
also be browsed online from the Google Code page.
Checking that it works After the setup has completed, you should be able to re up the
interactive Python interpreter and do the following:
>>> from mpmath import *
>>> mp.dps = 50
>>> print mpf(2) ** mpf(0.5)
1.4142135623730950488016887242096980785696718753769
>>> print 2*pi
6.2831853071795864769252867665590057683943387987502
Note: if you have are upgrading mpmath from an earlier version, you may have to manually
uninstall the old version or remove the old les.
Using gmpy (optional)
By default, mpmath uses Python integers internally. If gmpy version 1.03 or later is installed
on your system, mpmath will automatically detect it and transparently use gmpy integers
intead. This makes mpmath much faster, especially at high precision (approximately above
100 digits).
To verify that mpmath uses gmpy, check the internal variable BACKEND is not equal to python:
>>> import mpmath.libmp
>>> mpmath.libmp.BACKEND
gmpy
The gmpy mode can be disabled by setting the MPMATH NOGMPY environment variable.
Note that the mode cannot be switched during runtime; mpmath must be re-imported for this
change to take eect.
714
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Running tests
It is recommended that you run mpmaths full set of unit tests to make sure everything works.
The tests are located in the tests subdirectory of the main mpmath directory. They can be
run in the interactive interpreter using the runtests() function:
import mpmath
mpmath.runtests()
Alternatively, they can be run from the tests directory via
python runtests.py
The tests should nish in about a minute. If you have psyco installed, the tests can also be
run with
python runtests.py -psyco
which will cut the running time in half.
If any test fails, please send a detailed bug report to the mpmath issue tracker. The tests can
also be run with py.test. This will sometimes generate more useful information in case of a
failure.
To run the tests with support for gmpy disabled, use
python runtests.py -nogmpy
To enable extra diagnostics, use
python runtests.py -strict
Compiling the documentation
If you downloaded the source package, the text source for these documentation pages is
included in the doc directory. The documentation can be compiled to pretty HTML using
Sphinx. Go to the doc directory and run
python build.py
You can also test that all the interactive examples in the documentation work by running
python run doctest.py
and by running the individual .py les in the mpmath source.
(The doctests may take several minutes.)
Finally, some additional demo scripts are available in the demo directory included in the source
package.
Mpmath under SymPy
Mpmath is available as a subpackage of SymPy. With SymPy installed, you can just do
import sympy.mpmath as mpmath
instead of import mpmath. Note that the SymPy version of mpmath might not be the most
recent. You can make a separate mpmath installation even if SymPy is installed; the two
mpmath packages will not interfere with each other.
5.15. Welcome to mpmaths documentation!
715
SymPy Documentation, Release 0.7.6
Mpmath under Sage
Mpmath is a standard package in Sage, in version 4.1 or later of Sage. Mpmath is preinstalled
a regular Python module, and can be imported as usual within Sage:
---------------------------------------------------------------------| Sage Version 4.1, Release Date: 2009-07-09
|
| Type notebook() for the GUI, and license() for information.
|
---------------------------------------------------------------------sage: import mpmath
sage: mpmath.mp.dps = 50
sage: print mpmath.mpf(2) ** 0.5
1.4142135623730950488016887242096980785696718753769
The mpmath installation under Sage automatically use Sage integers for asymptotically fast
arithmetic, so there is no need to install GMPY:
sage: mpmath.libmp.BACKEND
sage
In Sage,
mpmath can alternatively
sage.libs.mpmath.all. For example:
be
imported
via
the
interface
library
sage: import sage.libs.mpmath.all as mpmath
This module provides a few extra conversion functions, including call() which permits calling any mpmath function with Sage numbers as input, and getting Sage RealNumber or ComplexNumber instances with the appropriate precision back:
sage: w = mpmath.call(mpmath.erf, 2+3*I, prec=100)
sage: w
-20.829461427614568389103088452 + 8.6873182714701631444280787545*I
sage: type(w)
<type sage.rings.complex number.ComplexNumber>
sage: w.prec()
100
See the help for sage.libs.mpmath.all for further information.
Basic usage
In interactive code examples that follow, it will be assumed that all items in the mpmath namespace have been imported:
>>> from mpmath import *
Importing everything can be convenient, especially when using mpmath interactively, but be
careful when mixing mpmath with other libraries! To avoid inadvertently overriding other
functions or objects, explicitly import only the needed objects, or use the mpmath. or mp.
namespaces:
from mpmath import sin, cos
sin(1), cos(1)
import mpmath
mpmath.sin(1), mpmath.cos(1)
716
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
from mpmath import mp
mp.sin(1), mp.cos(1)
# mp context object -- to be explained
Number types
Mpmath provides the following numerical types:
Class
mpf
mpc
matrix
Description
Real oat
Complex oat
Matrix
The following section will provide a very short introduction to the types mpf and mpc. Intervals
and matrices are described further in the documentation chapters on interval arithmetic and
matrices / linear algebra.
The mpf type is analogous to Pythons built-in float. It holds a real number or one of the
special values inf (positive innity), -inf (negative innity) and nan (not-a-number, indicating
an indeterminate result). You can create mpf instances from strings, integers, oats, and other
mpf instances:
>>> mpf(4)
mpf(4.0)
>>> mpf(2.5)
mpf(2.5)
>>> mpf(1.25e6)
mpf(1250000.0)
>>> mpf(mpf(2))
mpf(2.0)
>>> mpf(inf)
mpf(+inf)
The mpc type represents a complex number in rectangular form as a pair of mpf instances. It
can be constructed from a Python complex, a real number, or a pair of real numbers:
>>> mpc(2,3)
mpc(real=2.0, imag=3.0)
>>> mpc(complex(2,3)).imag
mpf(3.0)
You can mix mpf and mpc instances with each other and with Python numbers:
>>> mpf(3) + 2*mpf(2.5) + 1.0
mpf(9.0)
>>> mp.dps = 15
# Set precision (see below)
>>> mpc(1j)**0.5
mpc(real=0.70710678118654757, imag=0.70710678118654757)
Setting the precision
Mpmath uses a global working precision; it does not keep track of the precision or accuracy of
individual numbers. Performing an arithmetic operation or calling mpf() rounds the result to
the current working precision. The working precision is controlled by a context object called
mp, which has the following default state:
5.15. Welcome to mpmaths documentation!
717
SymPy Documentation, Release 0.7.6
>>> print mp
Mpmath settings:
mp.prec = 53
mp.dps = 15
mp.trap_complex = False
[default: 53]
[default: 15]
[default: False]
The term prec denotes the binary precision (measured in bits) while dps (short for decimal
places) is the decimal precision. Binary and decimal precision are related roughly according
to the formula prec = 3.33*dps. For example, it takes a precision of roughly 333 bits to hold
an approximation of pi that is accurate to 100 decimal places (actually slightly more than 333
bits is used).
Changing either precision property of the mp object automatically updates the other; usually
you just want to change the dps value:
>>> mp.dps = 100
>>> mp.dps
100
>>> mp.prec
336
When the precision has been set, all mpf operations are carried out at that precision:
>>> mp.dps = 50
>>> mpf(1) / 6
mpf(0.16666666666666666666666666666666666666666666666666656)
>>> mp.dps = 25
>>> mpf(2) ** mpf(0.5)
mpf(1.414213562373095048801688713)
The precision of complex arithmetic is also controlled by the mp object:
>>> mp.dps = 10
>>> mpc(1,2) / 3
mpc(real=0.3333333333321, imag=0.6666666666642)
There is no restriction on the magnitude of numbers. An mpf can for example hold an approximation of a large Mersenne prime:
>>> mp.dps = 15
>>> print mpf(2)**32582657 - 1
1.24575026015369e+9808357
Or why not 1 googolplex:
>>> print mpf(10) ** (10**100)
1.0e+100000000000000000000000000000000000000000000000000...
The (binary) exponent is stored exactly and is independent of the precision.
Temporarily changing the precision It is often useful to change the precision during
only part of a calculation. A way to temporarily increase the precision and then restore it is
as follows:
>>> mp.prec += 2
>>> # do_something()
>>> mp.prec -= 2
718
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
In Python 2.5, the with statement along with the mpmath functions workprec, workdps, extraprec and extradps can be used to temporarily change precision in a more safe manner:
>>> from __future__ import with_statement
>>> with workdps(20):
...
print mpf(1)/7
...
with extradps(10):
...
print mpf(1)/7
...
0.14285714285714285714
0.142857142857142857142857142857
>>> mp.dps
15
The with statement ensures that the precision gets reset when exiting the block, even in the
case that an exception is raised. (The eect of the with statement can be emulated in Python
2.4 by using a try/finally block.)
The workprec family of functions can also be used as function decorators:
>>> @workdps(6)
... def f():
...
return mpf(1)/3
...
>>> f()
mpf(0.33333331346511841)
Some functions accept the prec and dps keyword arguments and this will override the global
working precision. Note that this will not aect the precision at which the result is printed,
so to get all digits, you must either use increase precision afterward when printing or use
nstr/nprint:
>>> mp.dps = 15
>>> print exp(1)
2.71828182845905
>>> print exp(1, dps=50)
# Extra digits wont be printed
2.71828182845905
>>> nprint(exp(1, dps=50), 50)
2.7182818284590452353602874713526624977572470937
Finally, instead of using the global context object mp, you can create custom contexts and
work with methods of those instances instead of global functions. The working precision will
be local to each context object:
>>> mp2 = mp.clone()
>>> mp.dps = 10
>>> mp2.dps = 20
>>> print mp.mpf(1) / 3
0.3333333333
>>> print mp2.mpf(1) / 3
0.33333333333333333333
Note: the ability to create multiple contexts is a new feature that is only partially implemented. Not all mpmath functions are yet available as context-local methods. In the present
version, you are likely to encounter bugs if you try mixing dierent contexts.
5.15. Welcome to mpmaths documentation!
719
SymPy Documentation, Release 0.7.6
Providing correct input
Note that when creating a new mpf, the value will at most be as accurate as the input. Be
careful when mixing mpmath numbers with Python oats. When working at high precision,
fractional mpf values should be created from strings or integers:
>>> mp.dps = 30
>>> mpf(10.9)
# bad
mpf(10.9000000000000003552713678800501)
>>> mpf(10.9) # good
mpf(10.8999999999999999999999999999997)
>>> mpf(109) / mpf(10)
# also good
mpf(10.8999999999999999999999999999997)
>>> mp.dps = 15
(Binary fractions such as 0.5, 1.5, 0.75, 0.125, etc, are generally safe as input, however, since
those can be represented exactly by Python oats.)
Printing
By default, the repr() of a number includes its type signature. This way eval can be used to
recreate a number from its string representation:
>>> eval(repr(mpf(2.5)))
mpf(2.5)
Prettier output can be obtained by using str() or print, which hide the mpf and mpc signatures and also suppress rounding artifacts in the last few digits:
>>> mpf(3.14159)
mpf(3.1415899999999999)
>>> print mpf(3.14159)
3.14159
>>> print mpc(1j)**0.5
(0.707106781186548 + 0.707106781186548j)
Setting the mp.pretty option will use the str()-style output for repr() as well:
>>> mp.pretty = True
>>> mpf(0.6)
0.6
>>> mp.pretty = False
>>> mpf(0.6)
mpf(0.59999999999999998)
The number of digits with which numbers are printed by default is determined by the working
precision. To specify the number of digits to show without changing the working precision,
use mpmath.nstr() (page 727) and mpmath.nprint() (page 727):
>>> a = mpf(1) / 6
>>> a
mpf(0.16666666666666666)
>>> nstr(a, 8)
0.16666667
>>> nprint(a, 8)
0.16666667
720
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> nstr(a, 50)
0.16666666666666665741480812812369549646973609924316
5.15.2 Basic features
Contexts
High-level code in mpmath is implemented as methods on a context object. The context
implements arithmetic, type conversions and other fundamental operations. The context also
holds settings such as precision, and stores cache data. A few dierent contexts (with a mostly
compatible interface) are provided so that the high-level algorithms can be used with dierent
implementations of the underlying arithmetic, allowing dierent features and speed-accuracy
tradeos. Currently, mpmath provides the following contexts:
Arbitrary-precision arithmetic (mp)
A faster Cython-based version of mp (used by default in Sage, and currently only available
there)
Arbitrary-precision interval arithmetic (iv)
Double-precision arithmetic using Pythons builtin float and complex types (fp)
Most global functions in the global mpmath namespace are actually methods of the mp context.
This fact is usually transparent to the user, but sometimes shows up in the form of an initial
parameter called ctx visible in the help for the function:
>>> import mpmath
>>> help(mpmath.fsum)
Help on method fsum in module mpmath.ctx_mp_python:
fsum(ctx, terms, absolute=False, squared=False) method of mpmath.ctx_mp.MPContext instance
Calculates a sum containing a finite number of terms (for infinite
series, see :func:~mpmath.nsum). The terms will be converted to
...
The following operations are equivalent:
>>> mpmath.mp.dps = 15; mpmath.mp.pretty = False
>>> mpmath.fsum([1,2,3])
mpf(6.0)
>>> mpmath.mp.fsum([1,2,3])
mpf(6.0)
The corresponding operation using the fp context:
>>> mpmath.fp.fsum([1,2,3])
6.0
Common interface
ctx.mpf creates a real number:
>>> from mpmath import mp, fp
>>> mp.mpf(3)
mpf(3.0)
5.15. Welcome to mpmaths documentation!
721
SymPy Documentation, Release 0.7.6
>>> fp.mpf(3)
3.0
ctx.mpc creates a complex number:
>>> mp.mpc(2,3)
mpc(real=2.0, imag=3.0)
>>> fp.mpc(2,3)
(2+3j)
ctx.matrix creates a matrix:
>>> mp.matrix([[1,0],[0,1]])
matrix(
[[1.0, 0.0],
[0.0, 1.0]])
>>> _[0,0]
mpf(1.0)
>>> fp.matrix([[1,0],[0,1]])
matrix(
[[1.0, 0.0],
[0.0, 1.0]])
>>> _[0,0]
1.0
ctx.prec holds the current precision (in bits):
>>> mp.prec
53
>>> fp.prec
53
ctx.dps holds the current precision (in digits):
>>> mp.dps
15
>>> fp.dps
15
ctx.pretty controls whether objects should be pretty-printed automatically by repr().
Pretty-printing for mp numbers is disabled by default so that they can clearly be distinguished
from Python numbers and so that eval(repr(x)) == x works:
>>> mp.mpf(3)
mpf(3.0)
>>> mpf = mp.mpf
>>> eval(repr(mp.mpf(3)))
mpf(3.0)
>>> mp.pretty = True
>>> mp.mpf(3)
3.0
>>> fp.matrix([[1,0],[0,1]])
matrix(
[[1.0, 0.0],
[0.0, 1.0]])
>>> fp.pretty = True
>>> fp.matrix([[1,0],[0,1]])
[1.0 0.0]
[0.0 1.0]
722
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> fp.pretty = False
>>> mp.pretty = False
Arbitrary-precision oating-point (mp)
The mp context is what most users probably want to use most of the time, as it supports the
most functions, is most well-tested, and is implemented with a high level of optimization.
Nearly all examples in this documentation use mp functions.
See Basic usage (page 716) for a description of basic usage.
Arbitrary-precision interval arithmetic (iv)
The iv.mpf type represents a closed interval [a, b]; that is, the set {x : a x b}, where a
and b are arbitrary-precision oating-point values, possibly . The iv.mpc type represents
a rectangular complex interval [a, b] + [c, d]i; that is, the set {z = x + iy : a x b c y d}.
Interval arithmetic provides rigorous error tracking. If f is a mathematical function and f is its
interval arithmetic version, then the basic guarantee of interval arithmetic is that f (v) f(v)
for any input interval v. Put dierently, if an interval represents the known uncertainty for a
xed number, any sequence of interval operations will produce an interval that contains what
would be the result of applying the same sequence of operations to the exact number. The
principal drawbacks of interval arithmetic are speed (iv arithmetic is typically at least two
times slower than mp arithmetic) and that it sometimes provides far too pessimistic bounds.
Note: The support for interval arithmetic in mpmath is still experimental, and many functions do not yet properly support intervals. Please use this feature with caution.
Intervals can be created from single numbers (treated as zero-width intervals) or pairs of
endpoint numbers. Strings are treated as exact decimal numbers. Note that a Python oat
like 0.1 generally does not represent the same number as its literal; use 0.1 instead:
>>> from mpmath import iv
>>> iv.dps = 15; iv.pretty = False
>>> iv.mpf(3)
mpi(3.0, 3.0)
>>> print iv.mpf(3)
[3.0, 3.0]
>>> iv.pretty = True
>>> iv.mpf([2,3])
[2.0, 3.0]
>>> iv.mpf(0.1)
# probably not intended
[0.10000000000000000555, 0.10000000000000000555]
>>> iv.mpf(0.1)
# good, gives a containing interval
[0.099999999999999991673, 0.10000000000000000555]
>>> iv.mpf([0.1, 0.2])
[0.099999999999999991673, 0.2000000000000000111]
The fact that 0.1 results in an interval of nonzero width indicates that 1/10 cannot be
represented using binary oating-point numbers at this precision level (in fact, it cannot be
represented exactly at any precision).
Intervals may be innite or half-innite:
5.15. Welcome to mpmaths documentation!
723
SymPy Documentation, Release 0.7.6
>>> print 1 / iv.mpf([2, inf])
[0.0, 0.5]
The equality testing operators == and != check whether their operands are identical as intervals; that is, have the same endpoints. The ordering operators < <= > >= permit inequality
testing using triple-valued logic: a guaranteed inequality returns True or False while an
indeterminate inequality returns None:
>>> iv.mpf([1,2])
True
>>> iv.mpf([1,2])
False
>>> iv.mpf([1,2])
True
>>> iv.mpf([1,2])
True
>>> iv.mpf([1,2])
False
>>> iv.mpf([1,2])
>>> iv.mpf([2,2])
False
>>> iv.mpf([1,2])
True
>>> iv.mpf([1,2])
>>> iv.mpf([1,2])
False
== iv.mpf([1,2])
!= iv.mpf([1,2])
<= 2
> 0
< 1
< 2
< 2
# returns None
<= iv.mpf([2,3])
< iv.mpf([2,3]) # returns None
< iv.mpf([-1,0])
The in operator tests whether a number or interval is contained in another interval:
>>> iv.mpf([0,2]) in iv.mpf([0,10])
True
>>> 3 in iv.mpf([-inf, 0])
False
Intervals have the properties .a, .b (endpoints), .mid, and .delta (width):
>>> x = iv.mpf([2, 5])
>>> x.a
[2.0, 2.0]
>>> x.b
[5.0, 5.0]
>>> x.mid
[3.5, 3.5]
>>> x.delta
[3.0, 3.0]
Some transcendental functions are supported:
>>> iv.dps = 15
>>> mp.dps = 15
>>> iv.mpf([0.5,1.5]) ** iv.mpf([0.5, 1.5])
[0.35355339059327373086, 1.837117307087383633]
>>> iv.exp(0)
[1.0, 1.0]
>>> iv.exp([-inf,inf])
[0.0, +inf]
>>>
>>> iv.exp([-inf,0])
[0.0, 1.0]
724
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> iv.exp([0,inf])
[1.0, +inf]
>>> iv.exp([0,1])
[1.0, 2.7182818284590455349]
>>>
>>> iv.log(1)
[0.0, 0.0]
>>> iv.log([0,1])
[-inf, 0.0]
>>> iv.log([0,inf])
[-inf, +inf]
>>> iv.log(2)
[0.69314718055994528623, 0.69314718055994539725]
>>>
>>> iv.sin([100,inf])
[-1.0, 1.0]
>>> iv.cos([-0.1,0.1])
[0.99500416527802570954, 1.0]
Interval arithmetic is useful for proving inequalities involving irrational numbers. Naive use
of mp arithmetic may result in wrong conclusions, such as the following:
>>> mp.dps = 25
>>> x = mp.exp(mp.pi*mp.sqrt(163))
>>> y = mp.mpf(640320**3+744)
>>> print x
262537412640768744.0000001
>>> print y
262537412640768744.0
>>> x > y
True
But the correct result is e
163
< 262537412640768744, as can be seen by increasing the precision:
>>> mp.dps = 50
>>> print mp.exp(mp.pi*mp.sqrt(163))
262537412640768743.99999999999925007259719818568888
With interval arithmetic, the comparison returns None until the precision is large enough for
x y to have a denite sign:
>>> iv.dps = 15
>>> iv.exp(iv.pi*iv.sqrt(163)) > (640320**3+744)
>>> iv.dps = 30
>>> iv.exp(iv.pi*iv.sqrt(163)) > (640320**3+744)
>>> iv.dps = 60
>>> iv.exp(iv.pi*iv.sqrt(163)) > (640320**3+744)
False
>>> iv.dps = 15
Fast low-precision arithmetic (fp)
Although mpmath is generally designed for arbitrary-precision arithmetic, many of the highlevel algorithms work perfectly well with ordinary Python float and complex numbers, which
use hardware double precision (on most systems, this corresponds to 53 bits of precision).
Whereas the global functions (which are methods of the mp object) always convert inputs to
mpmath numbers, the fp object instead converts them to float or complex, and in some
5.15. Welcome to mpmaths documentation!
725
SymPy Documentation, Release 0.7.6
cases employs basic functions optimized for double precision. When large amounts of function evaluations (numerical integration, plotting, etc) are required, and when fp arithmetic
provides sucient accuracy, this can give a signicant speedup over mp arithmetic.
To take advantage of this feature, simply use the fp prex, i.e. write fp.func instead of func
or mp.func:
>>> u = fp.erfc(2.5)
>>> print u
0.000406952017445
>>> type(u)
<type float>
>>> mp.dps = 15
>>> print mp.erfc(2.5)
0.000406952017444959
>>> fp.matrix([[1,2],[3,4]]) ** 2
matrix(
[[7.0, 10.0],
[15.0, 22.0]])
>>>
>>> type(_[0,0])
<type float>
>>> print fp.quad(fp.sin, [0, fp.pi])
2.0
# numerical integration
The fp context wraps Pythons math and cmath modules for elementary functions. It supports
both real and complex numbers and automatically generates complex results for real inputs
(math raises an exception):
>>> fp.sqrt(5)
2.23606797749979
>>> fp.sqrt(-5)
2.23606797749979j
>>> fp.sin(10)
-0.5440211108893698
>>> fp.power(-1, 0.25)
(0.7071067811865476+0.7071067811865475j)
>>> (-1) ** 0.25
Traceback (most recent call last):
...
ValueError: negative number cannot be raised to a fractional power
Traceback (most recent call last):
...
ValueError: negative number cannot be raised to a fractional power
The prec and dps attributes can be changed (for interface compatibility with the mp context)
but this has no eect:
>>>
53
>>>
15
>>>
>>>
53
>>>
15
fp.prec
fp.dps
fp.prec = 80
fp.prec
fp.dps
Due to intermediate rounding and cancellation errors, results computed with fp arithmetic
may be much less accurate than those computed with mp using an equivalent precision
726
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
(mp.prec = 53), since the latter often uses increased internal precision. The accuracy is
highly problem-dependent: for some functions, fp almost always gives 14-15 correct digits;
for others, results can be accurate to only 2-3 digits or even completely wrong. The recommended use for fp is therefore to speed up large-scale computations where accuracy can be
veried in advance on a subset of the input set, or where results can be veried afterwards.
Utility functions
This page lists functions that perform basic operations on numbers or aid general programming.
Conversion and printing
mpmathify() / convert()
mpmath.mpmathify(x, strings=True)
Converts x to an mpf or mpc. If x is of type mpf, mpc, int, float, complex, the conversion
will be performed losslessly.
If x is a string, the result will be rounded to the present working precision. Strings
representing fractions or complex numbers are permitted.
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = False
>>> mpmathify(3.5)
mpf(3.5)
>>> mpmathify(2.1)
mpf(2.1000000000000001)
>>> mpmathify(3/4)
mpf(0.75)
>>> mpmathify(2+3j)
mpc(real=2.0, imag=3.0)
nstr()
mpmath.nstr(x, n=6, **kwargs)
Convert an mpf or mpc to a decimal string literal with n signicant digits. The small default value for n is chosen to make this function useful for printing collections of numbers
(lists, matrices, etc).
If x is a list or tuple, nstr() (page 727) is applied recursively to each element. For
unrecognized classes, nstr() (page 727) simply returns str(x).
The companion function nprint() (page 727) prints the result instead of returning it.
>>> from sympy.mpmath import *
>>> nstr([+pi, ldexp(1,-500)])
[3.14159, 3.05494e-151]
>>> nprint([+pi, ldexp(1,-500)])
[3.14159, 3.05494e-151]
nprint()
mpmath.nprint(x, n=6, **kwargs)
Equivalent to print(nstr(x, n)).
5.15. Welcome to mpmaths documentation!
727
SymPy Documentation, Release 0.7.6
Arithmetic operations
See also mpmath.sqrt() (page 752), mpmath.exp() (page 756) etc., listed in Powers and logarithms (page 751)
fadd()
mpmath.fadd(ctx, x, y, **kwargs)
Adds the numbers x and y, giving a oating-point result, optionally using a custom precision and rounding mode.
The default precision is the working precision of the context. You can specify a custom
precision in bits by passing the prec keyword argument, or by providing an equivalent
decimal precision with the dps keyword argument. If the precision is set to +inf, or if
the ag exact=True is passed, an exact addition with no rounding is performed.
When the precision is nite, the optional rounding keyword argument species the direction of rounding. Valid options are n for nearest (default), f for oor, c for
ceiling, d for down, u for up.
Examples
Using fadd() (page 728) with precision and rounding control:
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = False
>>> fadd(2, 1e-20)
mpf(2.0)
>>> fadd(2, 1e-20, rounding=u)
mpf(2.0000000000000004)
>>> nprint(fadd(2, 1e-20, prec=100), 25)
2.00000000000000000001
>>> nprint(fadd(2, 1e-20, dps=15), 25)
2.0
>>> nprint(fadd(2, 1e-20, dps=25), 25)
2.00000000000000000001
>>> nprint(fadd(2, 1e-20, exact=True), 25)
2.00000000000000000001
Exact addition avoids cancellation errors, enforcing familiar laws of numbers such as
x + y x = y, which dont hold in oating-point arithmetic with nite precision:
>>> x, y = mpf(2), mpf(1e-1000)
>>> print(x + y - x)
0.0
>>> print(fadd(x, y, prec=inf) - x)
1.0e-1000
>>> print(fadd(x, y, exact=True) - x)
1.0e-1000
Exact addition can be inecient and may be impossible to perform with large magnitude
dierences:
>>> fadd(1, 1e-100000000000000000000, prec=inf)
Traceback (most recent call last):
...
OverflowError: the exact result does not fit in memory
Traceback (most recent call last):
...
OverflowError: the exact result does not fit in memory
728
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
fsub()
mpmath.fsub(ctx, x, y, **kwargs)
Subtracts the numbers x and y, giving a oating-point result, optionally using a custom
precision and rounding mode.
See the documentation of fadd() (page 728) for a detailed description of how to specify
precision and rounding.
Examples
Using fsub() (page 729) with precision and rounding control:
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = False
>>> fsub(2, 1e-20)
mpf(2.0)
>>> fsub(2, 1e-20, rounding=d)
mpf(1.9999999999999998)
>>> nprint(fsub(2, 1e-20, prec=100), 25)
1.99999999999999999999
>>> nprint(fsub(2, 1e-20, dps=15), 25)
2.0
>>> nprint(fsub(2, 1e-20, dps=25), 25)
1.99999999999999999999
>>> nprint(fsub(2, 1e-20, exact=True), 25)
1.99999999999999999999
Exact subtraction avoids cancellation errors, enforcing familiar laws of numbers such as
x y + y = x, which dont hold in oating-point arithmetic with nite precision:
>>>
>>>
0.0
>>>
2.0
>>>
2.0
x, y = mpf(2), mpf(1e1000)
print(x - y + y)
print(fsub(x, y, prec=inf) + y)
print(fsub(x, y, exact=True) + y)
Exact addition can be inecient and may be impossible to perform with large magnitude
dierences:
>>> fsub(1, 1e-100000000000000000000, prec=inf)
Traceback (most recent call last):
...
OverflowError: the exact result does not fit in memory
Traceback (most recent call last):
...
OverflowError: the exact result does not fit in memory
fneg()
mpmath.fneg(ctx, x, **kwargs)
Negates the number x, giving a oating-point result, optionally using a custom precision
and rounding mode.
See the documentation of fadd() (page 728) for a detailed description of how to specify
precision and rounding.
Examples
An mpmath number is returned:
5.15. Welcome to mpmaths documentation!
729
SymPy Documentation, Release 0.7.6
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = False
>>> fneg(2.5)
mpf(-2.5)
>>> fneg(-5+2j)
mpc(real=5.0, imag=-2.0)
Precise control over rounding is possible:
>>> x = fadd(2, 1e-100, exact=True)
>>> fneg(x)
mpf(-2.0)
>>> fneg(x, rounding=f)
mpf(-2.0000000000000004)
Negating with and without roundo:
>>> n = 200000000000000000000001
>>> print(int(-mpf(n)))
-200000000000000016777216
>>> print(int(fneg(n)))
-200000000000000016777216
>>> print(int(fneg(n, prec=log(n,2)+1)))
-200000000000000000000001
>>> print(int(fneg(n, dps=log(n,10)+1)))
-200000000000000000000001
>>> print(int(fneg(n, prec=inf)))
-200000000000000000000001
>>> print(int(fneg(n, dps=inf)))
-200000000000000000000001
>>> print(int(fneg(n, exact=True)))
-200000000000000000000001
fmul()
mpmath.fmul(ctx, x, y, **kwargs)
Multiplies the numbers x and y, giving a oating-point result, optionally using a custom
precision and rounding mode.
See the documentation of fadd() (page 728) for a detailed description of how to specify
precision and rounding.
Examples
The result is an mpmath number:
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = False
>>> fmul(2, 5.0)
mpf(10.0)
>>> fmul(0.5j, 0.5)
mpc(real=0.0, imag=0.25)
Avoiding roundo:
>>> x, y = 10**10+1, 10**15+1
>>> print(x*y)
10000000001000010000000001
>>> print(mpf(x) * mpf(y))
1.0000000001e+25
730
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> print(int(mpf(x) * mpf(y)))
10000000001000011026399232
>>> print(int(fmul(x, y)))
10000000001000011026399232
>>> print(int(fmul(x, y, dps=25)))
10000000001000010000000001
>>> print(int(fmul(x, y, exact=True)))
10000000001000010000000001
Exact multiplication with complex numbers can be inecient and may be impossible to
perform with large magnitude dierences between real and imaginary parts:
>>> x = 1+2j
>>> y = mpc(2, 1e-100000000000000000000)
>>> fmul(x, y)
mpc(real=2.0, imag=4.0)
>>> fmul(x, y, rounding=u)
mpc(real=2.0, imag=4.0000000000000009)
>>> fmul(x, y, exact=True)
Traceback (most recent call last):
...
OverflowError: the exact result does not fit in memory
Traceback (most recent call last):
...
OverflowError: the exact result does not fit in memory
fdiv()
mpmath.fdiv(ctx, x, y, **kwargs)
Divides the numbers x and y, giving a oating-point result, optionally using a custom
precision and rounding mode.
See the documentation of fadd() (page 728) for a detailed description of how to specify
precision and rounding.
Examples
The result is an mpmath number:
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = False
>>> fdiv(3, 2)
mpf(1.5)
>>> fdiv(2, 3)
mpf(0.66666666666666663)
>>> fdiv(2+4j, 0.5)
mpc(real=4.0, imag=8.0)
The rounding direction and precision can be controlled:
>>> fdiv(2, 3, dps=3)
# Should be accurate to at least 3 digits
mpf(0.6666259765625)
>>> fdiv(2, 3, rounding=d)
mpf(0.66666666666666663)
>>> fdiv(2, 3, prec=60)
mpf(0.66666666666666667)
>>> fdiv(2, 3, rounding=u)
mpf(0.66666666666666674)
5.15. Welcome to mpmaths documentation!
731
SymPy Documentation, Release 0.7.6
Checking the error of a division by performing it at higher precision:
>>> fdiv(2, 3) - fdiv(2, 3, prec=100)
mpf(-3.7007434154172148e-17)
Unlike fadd() (page 728), fmul() (page 730), etc., exact division is not allowed since the
quotient of two oating-point numbers generally does not have an exact oating-point
representation. (In the future this might be changed to allow the case where the division
is actually exact.)
>>> fdiv(2, 3, exact=True)
Traceback (most recent call
...
ValueError: division is not
Traceback (most recent call
...
ValueError: division is not
last):
an exact operation
last):
an exact operation
fmod()
mpmath.fmod(x, y)
Converts x and y to mpmath numbers and returns x mod y. For mpmath numbers, this
is equivalent to x % y.
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = True
>>> fmod(100, pi)
2.61062773871641
You can use fmod() (page 732) to compute fractional parts of numbers:
>>> fmod(10.25, 1)
0.25
fsum()
mpmath.fsum(terms, absolute=False, squared=False)
Calculates a sum containing a nite number of terms (for innite series, see nsum()
(page 979)). The terms will be converted to mpmath numbers. For len(terms) > 2, this
function is generally faster and produces more accurate results than the builtin Python
function sum().
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = False
>>> fsum([1, 2, 0.5, 7])
mpf(10.5)
With squared=True each term is squared, and with absolute=True the absolute value of
each term is used.
fprod()
mpmath.fprod(factors)
Calculates a product containing a nite number of factors (for innite products, see
nprod() (page 988)). The factors will be converted to mpmath numbers.
732
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = False
>>> fprod([1, 2, 0.5, 7])
mpf(7.0)
fdot()
mpmath.fdot(A, B=None, conjugate=False)
Computes the dot product of the iterables A and B,
Ak Bk .
k=0
Alternatively, fdot() (page 733) accepts a single iterable of pairs. In other words,
fdot(A,B) and fdot(zip(A,B)) are equivalent. The elements are automatically converted to mpmath numbers.
With conjugate=True, the elements in the second vector will be conjugated:
Ak Bk
k=0
Examples
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = False
>>> A = [2, 1.5, 3]
>>> B = [1, -1, 2]
>>> fdot(A, B)
mpf(6.5)
>>> list(zip(A, B))
[(2, 1), (1.5, -1), (3, 2)]
>>> fdot(_)
mpf(6.5)
>>> A = [2, 1.5, 3j]
>>> B = [1+j, 3, -1-j]
>>> fdot(A, B)
mpc(real=9.5, imag=-1.0)
>>> fdot(A, B, conjugate=True)
mpc(real=3.5, imag=-5.0)
Complex components
fabs()
mpmath.fabs(x)
Returns the absolute value of x, |x|. Unlike abs(), fabs() (page 733) converts nonmpmath numbers (such as int) into mpmath numbers:
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = False
>>> fabs(3)
mpf(3.0)
>>> fabs(-3)
mpf(3.0)
>>> fabs(3+4j)
mpf(5.0)
5.15. Welcome to mpmaths documentation!
733
SymPy Documentation, Release 0.7.6
sign()
mpmath.sign(x)
Returns the sign of x, dened as sign(x) = x/|x| (with the special case sign(0) = 0):
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = False
>>> sign(10)
mpf(1.0)
>>> sign(-10)
mpf(-1.0)
>>> sign(0)
mpf(0.0)
Note that the sign function is also dened for complex numbers, for which it gives the
projection onto the unit circle:
>>> mp.dps = 15; mp.pretty = True
>>> sign(1+j)
(0.707106781186547 + 0.707106781186547j)
re()
mpmath.re(x)
Returns the real part of x, <(x). Unlike x.real, re() (page 734) converts x to a mpmath
number:
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = False
>>> re(3)
mpf(3.0)
>>> re(-1+4j)
mpf(-1.0)
im()
mpmath.im(x)
Returns the imaginary part of x, =(x). Unlike x.imag, im() (page 734) converts x to a
mpmath number:
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = False
>>> im(3)
mpf(0.0)
>>> im(-1+4j)
mpf(4.0)
arg()
mpmath.arg(x)
Computes the complex argument (phase) of x, dened as the signed angle between the
positive real axis and x in the complex plane:
>>>
>>>
>>>
0.0
>>>
734
from sympy.mpmath import *
mp.dps = 15; mp.pretty = True
arg(3)
arg(3+3j)
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
0.785398163397448
>>> arg(3j)
1.5707963267949
>>> arg(-3)
3.14159265358979
>>> arg(-3j)
-1.5707963267949
The angle is dened to satisfy < arg(x) and with the sign convention that a
nonnegative imaginary part results in a nonnegative argument.
The value returned by arg() (page 734) is an mpf instance.
conj()
mpmath.conj(x)
Returns the complex conjugate of x, x. Unlike x.conjugate(), im() (page 734) converts
x to a mpmath number:
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = False
>>> conj(3)
mpf(3.0)
>>> conj(-1+4j)
mpc(real=-1.0, imag=-4.0)
polar()
mpmath.polar(x)
Returns the polar representation of the complex number z as a pair (r, ) such that z =
rei :
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = True
>>> polar(-2)
(2.0, 3.14159265358979)
>>> polar(3-4j)
(5.0, -0.927295218001612)
rect()
mpmath.rect(x)
Returns the complex number represented by polar coordinates (r, ):
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = True
>>> chop(rect(2, pi))
-2.0
>>> rect(sqrt(2), -pi/4)
(1.0 - 1.0j)
Integer and fractional parts
floor()
5.15. Welcome to mpmaths documentation!
735
SymPy Documentation, Release 0.7.6
mpmath.floor(x)
Computes the oor of x, bxc, dened as the largest integer less than or equal to x:
>>> from sympy.mpmath import *
>>> mp.pretty = False
>>> floor(3.5)
mpf(3.0)
Note: floor() (page 735), ceil() (page 736) and nint() (page 736) return a oatingpoint number, not a Python int. If bxc is too large to be represented exactly at the present
working precision, the result will be rounded, not necessarily in the direction implied by
the mathematical denition of the function.
To avoid rounding, use prec=0:
>>> mp.dps = 15
>>> print(int(floor(10**30+1)))
1000000000000000019884624838656
>>> print(int(floor(10**30+1, prec=0)))
1000000000000000000000000000001
The oor function is dened for complex numbers and acts on the real and imaginary
parts separately:
>>> floor(3.25+4.75j)
mpc(real=3.0, imag=4.0)
ceil()
mpmath.ceil(x)
Computes the ceiling of x, dxe, dened as the smallest integer greater than or equal to
x:
>>> from sympy.mpmath import *
>>> mp.pretty = False
>>> ceil(3.5)
mpf(4.0)
The ceiling function is dened for complex numbers and acts on the real and imaginary
parts separately:
>>> ceil(3.25+4.75j)
mpc(real=4.0, imag=5.0)
See notes about rounding for floor() (page 735).
nint()
mpmath.nint(x)
Evaluates the nearest integer function, nint(x). This gives the nearest integer to x; on a
tie, it gives the nearest even integer:
>>> from sympy.mpmath import *
>>> mp.pretty = False
>>> nint(3.2)
mpf(3.0)
>>> nint(3.8)
736
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
mpf(4.0)
>>> nint(3.5)
mpf(4.0)
>>> nint(4.5)
mpf(4.0)
The nearest integer function is dened for complex numbers and acts on the real and
imaginary parts separately:
>>> nint(3.25+4.75j)
mpc(real=3.0, imag=5.0)
See notes about rounding for floor() (page 735).
frac()
mpmath.frac(x)
Gives the fractional part of x, dened as frac(x) = x bxc (see floor() (page 735)). In
eect, this computes x modulo 1, or x + n where n Z is such that x + n [0, 1):
>>> from sympy.mpmath import *
>>> mp.pretty = False
>>> frac(1.25)
mpf(0.25)
>>> frac(3)
mpf(0.0)
>>> frac(-1.25)
mpf(0.75)
For a complex number, the fractional part function applies to the real and imaginary
parts separately:
>>> frac(2.25+3.75j)
mpc(real=0.25, imag=0.75)
Plotted, the fractional part function gives a sawtooth wave. The Fourier series coecients have a simple form:
>>> mp.dps = 15
>>> nprint(fourier(lambda x: frac(x)-0.5, [0,1], 4))
([0.0, 0.0, 0.0, 0.0, 0.0], [0.0, -0.31831, -0.159155, -0.106103, -0.0795775])
>>> nprint([-1/(pi*k) for k in range(1,5)])
[-0.31831, -0.159155, -0.106103, -0.0795775]
Note: The fractional part is sometimes dened as a symmetric function, i.e. returning
frac(x) if x < 0. This convention is used, for instance, by Mathematicas FractionalPart.
Tolerances and approximate comparisons
chop()
mpmath.chop(x, tol=None)
Chops o small real or imaginary parts, or converts numbers close to zero to exact zeros.
The input can be a single number or an iterable:
5.15. Welcome to mpmaths documentation!
737
SymPy Documentation, Release 0.7.6
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = False
>>> chop(5+1e-10j, tol=1e-9)
mpf(5.0)
>>> nprint(chop([1.0, 1e-20, 3+1e-18j, -4, 2]))
[1.0, 0.0, 3.0, -4.0, 2.0]
The tolerance defaults to 100*eps.
almosteq()
mpmath.almosteq(s, t, rel eps=None, abs eps=None)
Determine whether the dierence between s and t is smaller than a given epsilon, either
relatively or absolutely.
Both a maximum relative dierence and a maximum dierence (epsilons) may be specied. The absolute dierence is dened as |s t| and the relative dierence is dened as
|s t|/ max(|s|, |t|).
If only one epsilon is given, both are set to the same value. If none is given, both epsilons
are set to 2p+m where p is the current working precision and m is a small integer. The
default setting typically allows almosteq() (page 738) to be used to check for mathematical equality in the presence of small rounding errors.
Examples
>>> from sympy.mpmath import *
>>> mp.dps = 15
>>> almosteq(3.141592653589793, 3.141592653589790)
True
>>> almosteq(3.141592653589793, 3.141592653589700)
False
>>> almosteq(3.141592653589793, 3.141592653589700, 1e-10)
True
>>> almosteq(1e-20, 2e-20)
True
>>> almosteq(1e-20, 2e-20, rel_eps=0, abs_eps=0)
False
Properties of numbers
isinf()
mpmath.isinf(x)
Return True if the absolute value of x is innite; otherwise return False:
>>> from sympy.mpmath import *
>>> isinf(inf)
True
>>> isinf(-inf)
True
>>> isinf(3)
False
>>> isinf(3+4j)
False
>>> isinf(mpc(3,inf))
True
738
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> isinf(mpc(inf,3))
True
isnan()
mpmath.isnan(x)
Return True if x is a NaN (not-a-number), or for a complex number, whether either the
real or complex part is NaN; otherwise return False:
>>> from sympy.mpmath import *
>>> isnan(3.14)
False
>>> isnan(nan)
True
>>> isnan(mpc(3.14,2.72))
False
>>> isnan(mpc(3.14,nan))
True
isnormal()
mpmath.isnormal(x)
Determine whether x is normal in the sense of oating-point representation; that is,
return False if x is zero, an innity or NaN; otherwise return True. By extension, a
complex number x is considered normal if its magnitude is normal:
>>> from sympy.mpmath import *
>>> isnormal(3)
True
>>> isnormal(0)
False
>>> isnormal(inf); isnormal(-inf); isnormal(nan)
False
False
False
>>> isnormal(0+0j)
False
>>> isnormal(0+3j)
True
>>> isnormal(mpc(2,nan))
False
isfinite()
mpmath.isfinite(x)
Return True if x is a nite number, i.e. neither an innity or a NaN.
>>> from sympy.mpmath import *
>>> isfinite(inf)
False
>>> isfinite(-inf)
False
>>> isfinite(3)
True
>>> isfinite(nan)
False
>>> isfinite(3+4j)
5.15. Welcome to mpmaths documentation!
739
SymPy Documentation, Release 0.7.6
True
>>> isfinite(mpc(3,inf))
False
>>> isfinite(mpc(nan,3))
False
isint()
mpmath.isint(x, gaussian=False)
Return True if x is integer-valued; otherwise return False:
>>> from sympy.mpmath import *
>>> isint(3)
True
>>> isint(mpf(3))
True
>>> isint(3.2)
False
>>> isint(inf)
False
Optionally, Gaussian integers can be checked for:
>>> isint(3+0j)
True
>>> isint(3+2j)
False
>>> isint(3+2j, gaussian=True)
True
ldexp()
mpmath.ldexp(x, n)
Computes x2n eciently. No rounding is performed. The argument x must be a real
oating-point number (or possible to convert into one) and n must be a Python int.
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = False
>>> ldexp(1, 10)
mpf(1024.0)
>>> ldexp(1, -3)
mpf(0.125)
frexp()
mpmath.frexp(x, n)
Given a real number x, returns (y, n) with y [0.5, 1), n a Python integer, and such that
x = y2n . No rounding is performed.
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = False
>>> frexp(7.5)
(mpf(0.9375), 3)
mag()
740
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
mpmath.mag(x)
Quick logarithmic magnitude estimate of a number. Returns an integer or innity m such
that |x| <= 2m . It is not guaranteed that m is an optimal bound, but it will never be too
large by more than 2 (and probably not more than 1).
Examples
>>> from sympy.mpmath import *
>>> mp.pretty = True
>>> mag(10), mag(10.0), mag(mpf(10)), int(ceil(log(10,2)))
(4, 4, 4, 4)
>>> mag(10j), mag(10+10j)
(4, 5)
>>> mag(0.01), int(ceil(log(0.01,2)))
(-6, -6)
>>> mag(0), mag(inf), mag(-inf), mag(nan)
(-inf, +inf, +inf, nan)
nint distance()
mpmath.nint distance(x)
Return (n, d) where n is the nearest integer to x and d is an estimate of log2 (|x n|). If
d < 0, d gives the precision (measured in bits) lost to cancellation when computing xn.
>>> from sympy.mpmath import *
>>> n, d = nint_distance(5)
>>> print(n); print(d)
5
-inf
>>> n, d = nint_distance(mpf(5))
>>> print(n); print(d)
5
-inf
>>> n, d = nint_distance(mpf(5.00000001))
>>> print(n); print(d)
5
-26
>>> n, d = nint_distance(mpf(4.99999999))
>>> print(n); print(d)
5
-26
>>> n, d = nint_distance(mpc(5,10))
>>> print(n); print(d)
5
4
>>> n, d = nint_distance(mpc(5,0.000001))
>>> print(n); print(d)
5
-19
Number generation
fraction()
mpmath.fraction(p, q)
Given Python integers (p, q), returns a lazy mpf representing the fraction p/q. The value
is updated with the precision.
5.15. Welcome to mpmaths documentation!
741
SymPy Documentation, Release 0.7.6
>>> from sympy.mpmath import *
>>> mp.dps = 15
>>> a = fraction(1,100)
>>> b = mpf(1)/100
>>> print(a); print(b)
0.01
0.01
>>> mp.dps = 30
>>> print(a); print(b)
# a will be accurate
0.01
0.0100000000000000002081668171172
>>> mp.dps = 15
rand()
mpmath.rand()
Returns an mpf with value chosen randomly from [0, 1). The number of randomly generated bits in the mantissa is equal to the working precision.
arange()
mpmath.arange(*args)
This is a generalized version of Pythons range() function that accepts fractional endpoints and step sizes and returns a list of mpf instances. Like range(), arange()
(page 742) can be called with 1, 2 or 3 arguments:
arange(b) [0, 1, 2, . . . , x]
arange(a, b) [a, a + 1, a + 2, . . . , x]
arange(a, b, h) [a, a + h, a + h, . . . , x]
where b 1 x < b (in the third case, b h x < b).
Like Pythons range(), the endpoint is not included. To produce ranges where the endpoint is included, linspace() (page 742) is more convenient.
Examples
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = False
>>> arange(4)
[mpf(0.0), mpf(1.0), mpf(2.0), mpf(3.0)]
>>> arange(1, 2, 0.25)
[mpf(1.0), mpf(1.25), mpf(1.5), mpf(1.75)]
>>> arange(1, -1, -0.75)
[mpf(1.0), mpf(0.25), mpf(-0.5)]
linspace()
mpmath.linspace(*args, **kwargs)
linspace(a, b, n) returns a list of n evenly spaced samples from a to b. The syntax
linspace(mpi(a,b), n) is also valid.
This function is often more convenient than arange() (page 742) for partitioning an
interval into subintervals, since the endpoint is included:
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = False
742
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> linspace(1, 4, 4)
[mpf(1.0), mpf(2.0), mpf(3.0), mpf(4.0)]
You may also provide the keyword argument endpoint=False:
>>> linspace(1, 4, 4, endpoint=False)
[mpf(1.0), mpf(1.75), mpf(2.5), mpf(3.25)]
Precision management
autoprec()
mpmath.autoprec(ctx, f, maxprec=None, catch=(), verbose=False)
Return a wrapped copy of f that repeatedly evaluates f with increasing precision until
the result converges to the full precision used at the point of the call.
This heuristically protects against rounding errors, at the cost of roughly a 2x slowdown
compared to manually setting the optimal precision. This method can, however, easily
be fooled if the results from f depend discontinuously on the precision, for instance if
catastrophic cancellation can occur. Therefore, autoprec() (page 743) should be used
judiciously.
Examples
Many functions are sensitive to perturbations of the input arguments. If the arguments
are decimal numbers, they may have to be converted to binary at a much higher precision. If the amount of required extra precision is unknown, autoprec() (page 743) is
convenient:
>>> from sympy.mpmath import *
>>> mp.dps = 15
>>> mp.pretty = True
>>> besselj(5, 125 * 10**28)
# Exact input
-8.03284785591801e-17
>>> besselj(5, 1.25e30)
# Bad
7.12954868316652e-16
>>> autoprec(besselj)(5, 1.25e30)
# Good
-8.03284785591801e-17
The following fails to converge because sin() = 0 whereas all nite-precision approximations of give nonzero values:
>>> autoprec(sin)(pi)
Traceback (most recent call last):
...
NoConvergence: autoprec: prec increased to 2910 without convergence
Traceback (most recent call last):
...
NoConvergence: autoprec: prec increased to 2910 without convergence
As the following example shows, autoprec() (page 743) can protect against cancellation,
but is fooled by too severe cancellation:
>>> x = 1e-10
>>> exp(x)-1; expm1(x); autoprec(lambda t: exp(t)-1)(x)
1.00000008274037e-10
1.00000000005e-10
1.00000000005e-10
>>> x = 1e-50
5.15. Welcome to mpmaths documentation!
743
SymPy Documentation, Release 0.7.6
>>> exp(x)-1; expm1(x); autoprec(lambda t: exp(t)-1)(x)
0.0
1.0e-50
0.0
With catch, an exception or list of exceptions to intercept may be specied. The raised
exception is interpreted as signaling insucient precision. This permits, for example,
evaluating a function where a too low precision results in a division by zero:
>>> f = lambda x: 1/(exp(x)-1)
>>> f(1e-30)
Traceback (most recent call last):
...
ZeroDivisionError
>>> autoprec(f, catch=ZeroDivisionError)(1e-30)
1.0e+30
Traceback (most recent call last):
...
ZeroDivisionError
workprec()
mpmath.workprec(ctx, n, normalize output=False)
The block
with workprec(n): <code>
sets the precision to n bits, executes <code>, and then restores the precision.
workprec(n)(f) returns a decorated version of the function f that sets the precision
to n bits before execution, and restores the precision afterwards. With normalize output=True, it rounds the return value to the parent precision.
workdps()
mpmath.workdps(ctx, n, normalize output=False)
This function is analogous to workprec (see documentation) but changes the decimal
precision instead of the number of bits.
extraprec()
mpmath.extraprec(ctx, n, normalize output=False)
The block
with extraprec(n): <code>
increases the precision n bits, executes <code>, and then restores the precision.
extraprec(n)(f) returns a decorated version of the function f that increases the working
precision by n bits before execution, and restores the parent precision afterwards. With
normalize output=True, it rounds the return value to the parent precision.
extradps()
mpmath.extradps(ctx, n, normalize output=False)
This function is analogous to extraprec (see documentation) but changes the decimal
precision instead of the number of bits.
744
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Performance and debugging
memoize()
mpmath.memoize(ctx, f)
Return a wrapped copy of f that caches computed values, i.e. a memoized copy of f.
Values are only reused if the cached precision is equal to or higher than the working
precision:
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = True
>>> f = memoize(maxcalls(sin, 1))
>>> f(2)
0.909297426825682
>>> f(2)
0.909297426825682
>>> mp.dps = 25
>>> f(2)
Traceback (most recent call last):
...
NoConvergence: maxcalls: function evaluated 1 times
Traceback (most recent call last):
...
NoConvergence: maxcalls: function evaluated 1 times
maxcalls()
mpmath.maxcalls(ctx, f, N)
Return a wrapped copy of f that raises NoConvergence when f has been called more than
N times:
>>> from sympy.mpmath import *
>>> mp.dps = 15
>>> f = maxcalls(sin, 10)
>>> print(sum(f(n) for n in range(10)))
1.95520948210738
>>> f(10)
Traceback (most recent call last):
...
NoConvergence: maxcalls: function evaluated 10 times
Traceback (most recent call last):
...
NoConvergence: maxcalls: function evaluated 10 times
monitor()
mpmath.monitor(f, input=print, output=print)
Returns a wrapped copy of f that monitors evaluation by calling input with every input
(args, kwargs) passed to f and output with every value returned from f. The default
action (specify using the special string value print) is to print inputs and outputs to
stdout, along with the total evaluation count:
>>>
>>>
>>>
in
out
in
from sympy.mpmath import *
mp.dps = 5; mp.pretty = False
diff(monitor(exp), 1)
# diff will eval f(x-h) and f(x+h)
0 (mpf(0.99999999906867742538452148),) {}
0 mpf(2.7182818259274480055282064)
1 (mpf(1.0000000009313225746154785),) {}
5.15. Welcome to mpmaths documentation!
745
SymPy Documentation, Release 0.7.6
out 1 mpf(2.7182818309906424675501024)
mpf(2.7182808)
To disable either the input or the output handler, you may pass None as argument.
Custom input and output handlers may be used e.g. to store results for later analysis:
>>> mp.dps = 15
>>> input = []
>>> output = []
>>> findroot(monitor(sin, input.append, output.append), 3.0)
mpf(3.1415926535897932)
>>> len(input) # Count number of evaluations
9
>>> print(input[3]); print(output[3])
((mpf(3.1415076583334066),), {})
8.49952562843408e-5
>>> print(input[4]); print(output[4])
((mpf(3.1415928201669122),), {})
-1.66577118985331e-7
timing()
mpmath.timing(f, *args, **kwargs)
Returns time elapsed for evaluating f(). Optionally arguments may be passed to time
the execution of f(*args, **kwargs).
If the rst call is very quick, f is called repeatedly and the best time is returned.
Plotting
If matplotlib is available, the functions plot and cplot in mpmath can be used to plot functions
respectively as x-y graphs and in the complex plane. Also, splot can be used to produce 3D
surface plots.
746
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Function curve plots
Output of plot([cos, sin], [-4, 4])
mpmath.plot(ctx, f, xlim=[-5, 5], ylim=None, points=200, le=None, dpi=None, singularities=[], axes=None)
Shows a simple 2D plot of a function f (x) or list of functions [f0 (x), f1 (x), . . . , fn (x)] over a
given interval specied by xlim. Some examples:
plot(lambda x: exp(x)*li(x), [1, 4])
plot([cos, sin], [-4, 4])
plot([fresnels, fresnelc], [-4, 4])
plot([sqrt, cbrt], [-4, 4])
plot(lambda t: zeta(0.5+t*j), [-20, 20])
plot([floor, ceil, abs, sign], [-5, 5])
Points where the function raises a numerical exception or returns an innite value are
removed from the graph. Singularities can also be excluded explicitly as follows (useful
for removing erroneous vertical lines):
plot(cot, ylim=[-5, 5])
# bad
plot(cot, ylim=[-5, 5], singularities=[-pi, 0, pi])
# good
For parts where the function assumes complex values, the real part is plotted with dashes
and the imaginary part is plotted with dots.
5.15. Welcome to mpmaths documentation!
747
SymPy Documentation, Release 0.7.6
Note: This function requires matplotlib (pylab).
Complex function plots
Output of fp.cplot(fp.gamma, points=100000)
mpmath.cplot(ctx, f, re=[-5, 5], im=[-5, 5], points=2000, color=None, verbose=False,
le=None, dpi=None, axes=None)
Plots the given complex-valued function f over a rectangular part of the complex plane
specied by the pairs of intervals re and im. For example:
cplot(lambda z: z, [-2, 2], [-10, 10])
cplot(exp)
cplot(zeta, [0, 1], [0, 50])
By default, the complex argument (phase) is shown as color (hue) and the magnitude is
show as brightness. You can also supply a custom color function (color). This function
should take a complex number as input and return an RGB 3-tuple containing oats in
the range 0.0-1.0.
To obtain a sharp image, the number of points may need to be increased to 100,000
or thereabout. Since evaluating the function that many times is likely to be slow, the
verbose option is useful to display progress.
748
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Note: This function requires matplotlib (pylab).
3D surface plots
Output of splot for the donut example.
mpmath.splot(ctx, f, u=[-5, 5], v=[-5, 5], points=100, keep aspect=True, wireframe=False, le=None, dpi=None, axes=None)
Plots the surface dened by f .
If f returns a single component, then this plots the surface dened by z = f (x, y) over the
rectangular domain with x = u and y = v.
If f returns three components, then this plots the parametric surface x, y, z = f (u, v) over
the pairs of intervals u and v.
For example, to plot a simple function:
>>> from sympy.mpmath import *
>>> f = lambda x, y: sin(x+y)*cos(y)
>>> splot(f, [-pi,pi], [-pi,pi])
Plotting a donut:
5.15. Welcome to mpmaths documentation!
749
SymPy Documentation, Release 0.7.6
>>> r, R = 1, 2.5
>>> f = lambda u, v: [r*cos(u), (R+r*sin(u))*cos(v), (R+r*sin(u))*sin(v)]
>>> splot(f, [0, 2*pi], [0, 2*pi])
Note: This function requires matplotlib (pylab) 0.98.5.3 or higher.
5.15.3 Advanced mathematics
On top of its support for arbitrary-precision arithmetic, mpmath provides extensive support
for transcendental functions, evaluation of sums, integrals, limits, roots, and so on.
Mathematical functions
Mpmath implements the standard functions from Pythons math and cmath modules, for both
real and complex numbers and with arbitrary precision. Many other functions are also available in mpmath, including commonly-used variants of standard functions (such as the alternative trigonometric functions sec, csc, cot), but also a large number of special functions
such as the gamma function, the Riemann zeta function, error functions, Bessel functions,
etc.
Mathematical constants
Mpmath supports arbitrary-precision computation of various common (and less common)
mathematical constants. These constants are implemented as lazy objects that can evaluate to any precision. Whenever the objects are used as function arguments or as operands in
arithmetic operations, they automagically evaluate to the current working precision. A lazy
number can be converted to a regular mpf using the unary + operator, or by calling it as a
function:
>>> from mpmath import *
>>> mp.dps = 15
>>> pi
<pi: 3.14159~>
>>> 2*pi
mpf(6.2831853071795862)
>>> +pi
mpf(3.1415926535897931)
>>> pi()
mpf(3.1415926535897931)
>>> mp.dps = 40
>>> pi
<pi: 3.14159~>
>>> 2*pi
mpf(6.283185307179586476925286766559005768394338)
>>> +pi
mpf(3.141592653589793238462643383279502884197169)
>>> pi()
mpf(3.141592653589793238462643383279502884197169)
750
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Exact constants The predened objects j (imaginary unit), inf (positive innity) and nan
(not-a-number) are shortcuts to mpc and mpf instances with these xed values.
Pi (pi)
mp.pi = <pi: 3.14159>
Degree (degree)
mp.degree = <1 deg = pi / 180: 0.0174533>
Base of the natural logarithm (e)
mp.e = <e = exp(1): 2.71828>
Golden ratio (phi)
mp.phi = <Golden ratio phi: 1.61803>
Eulers constant (euler)
mp.euler = <Eulers constant: 0.577216>
Catalans constant (catalan)
mp.catalan = <Catalans constant: 0.915966>
Aperys constant (apery)
mp.apery = <Aperys constant: 1.20206>
Khinchins constant (khinchin)
mp.khinchin = <Khinchins constant: 2.68545>
Glaishers constant (glaisher)
mp.glaisher = <Glaishers constant: 1.28243>
Mertens constant (mertens)
mp.mertens = <Mertens constant: 0.261497>
Twin prime constant (twinprime)
mp.twinprime = <Twin prime constant: 0.660162>
Powers and logarithms
Nth roots
5.15. Welcome to mpmaths documentation!
751
SymPy Documentation, Release 0.7.6
sqrt()
mpmath.sqrt(x, **kwargs)
sqrt(x) gives the principal square root of x, x. For positive real numbers, the principal
root is simply the positive square
root. For arbitrary complex numbers, the principal
square root is dened to satisfy x = exp(log(x)/2). The function thus has a branch cut
along the negative half real axis.
For all mpmath numbers x, calling sqrt(x) is equivalent to performing x**0.5.
Examples
Basic examples and limits:
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = True
>>> sqrt(10)
3.16227766016838
>>> sqrt(100)
10.0
>>> sqrt(-4)
(0.0 + 2.0j)
>>> sqrt(1+1j)
(1.09868411346781 + 0.455089860562227j)
>>> sqrt(inf)
+inf
Square root evaluation is fast at huge precision:
>>> mp.dps = 50000
>>> a = sqrt(3)
>>> str(a)[-10:]
9329332814
mpmath.iv.sqrt() supports interval arguments:
>>> iv.dps = 15; iv.pretty = True
>>> iv.sqrt([16,100])
[4.0, 10.0]
>>> iv.sqrt(2)
[1.4142135623730949234, 1.4142135623730951455]
>>> iv.sqrt(2) ** 2
[1.9999999999999995559, 2.0000000000000004441]
hypot()
mpmath.hypot(x, y)
Computes the Euclidean norm of the vector (x, y), equal to x2 + y 2 . Both x and y must
be real.
cbrt()
mpmath.cbrt(x, **kwargs)
cbrt(x) computes the cube root of x, x1/3 . This function is faster and more accurate
than raising to a oating-point fraction:
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = False
>>> 125**(mpf(1)/3)
mpf(4.9999999999999991)
752
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> cbrt(125)
mpf(5.0)
Every nonzero complex number has three cube roots. This function returns the cube
root dened by exp(log(x)/3) where the principal branch of the natural logarithm is used.
Note that this does not give a real cube root for negative real numbers:
>>> mp.pretty = True
>>> cbrt(-1)
(0.5 + 0.866025403784439j)
root()
mpmath.root(z, n, k=0)
root(z, n, k=0) computes an n-th root of z, i.e. returns a number r that (up to possible
approximation error) satises rn = z. (nthroot is available as an alias for root.)
Every complex number z 6= 0 has n distinct n-th roots, which are equidistant points on
a circle with radius |z|1/n , centered around the origin. A specic root may be selected
using the optional index k. The roots are indexed counterclockwise, starting with k = 0
for the root closest to the positive real half-axis.
The k = 0 root is the so-called principal n-th root, often denoted by n z or z 1/n , and also
given by exp(log(z)/n). If z is a positive real number, the principal root is just the unique
positive n-th root of z. Under some circumstances, non-principal real roots exist: for
positive real z, n even, there is a negative root given by k = n/2; for negative real z, n
odd, there is a negative root given by k = (n 1)/2.
To obtain all roots with a simple expression, use [root(z,n,k) for k in range(n)].
An important special case, root(1, n, k) returns the k-th n-th root of unity, k = e2ik/n .
Alternatively, unitroots() (page 754) provides a slightly more convenient way to obtain
the roots of unity, including the option to compute only the primitive roots of unity.
Both k and n should be integers; k outside of range(n) will be reduced modulo n. If n is
negative, x1/n = 1/x1/n (or the equivalent reciprocal for a non-principal root with k 6= 0)
is computed.
root() (page 753) is implemented to use Newtons method for small n. At high precision,
this makes x1/n not much more expensive than the regular exponentiation, xn . For very
large n, nthroot() falls back to use the exponential function.
Examples
nthroot()/root() (page 753) is faster and more accurate than raising to a oating-point
fraction:
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = False
>>> 16807 ** (mpf(1)/5)
mpf(7.0000000000000009)
>>> root(16807, 5)
mpf(7.0)
>>> nthroot(16807, 5)
# Alias
mpf(7.0)
A high-precision root:
>>> mp.dps = 50; mp.pretty = True
>>> nthroot(10, 5)
5.15. Welcome to mpmaths documentation!
753
SymPy Documentation, Release 0.7.6
1.584893192461113485202101373391507013269442133825
>>> nthroot(10, 5) ** 5
10.0
Computing principal and non-principal square and cube roots:
>>> mp.dps = 15
>>> root(10, 2)
3.16227766016838
>>> root(10, 2, 1)
-3.16227766016838
>>> root(-10, 3)
(1.07721734501594 + 1.86579517236206j)
>>> root(-10, 3, 1)
-2.15443469003188
>>> root(-10, 3, 2)
(1.07721734501594 - 1.86579517236206j)
All the 7th roots of a complex number:
>>> for r in [root(3+4j, 7, k) for k in range(7)]:
...
print(%s %s % (r, r**7))
...
(1.24747270589553 + 0.166227124177353j) (3.0 + 4.0j)
(0.647824911301003 + 1.07895435170559j) (3.0 + 4.0j)
(-0.439648254723098 + 1.17920694574172j) (3.0 + 4.0j)
(-1.19605731775069 + 0.391492658196305j) (3.0 + 4.0j)
(-1.05181082538903 - 0.691023585965793j) (3.0 + 4.0j)
(-0.115529328478668 - 1.25318497558335j) (3.0 + 4.0j)
(0.907748109144957 - 0.871672518271819j) (3.0 + 4.0j)
Cube roots of unity:
>>> for k in range(3): print(root(1, 3, k))
...
1.0
(-0.5 + 0.866025403784439j)
(-0.5 - 0.866025403784439j)
Some exact high order roots:
>>> root(75**210, 105)
5625.0
>>> root(1, 128, 96)
(0.0 - 1.0j)
>>> root(4**128, 128, 96)
(0.0 - 4.0j)
unitroots()
mpmath.unitroots(n, primitive=False)
unitroots(n) returns 0 , 1 , . . . , n1 , all the distinct n-th roots of unity, as a list. If the
option primitive=True is passed, only the primitive roots are returned.
Every n-th root of unity satises (k )n = 1. There are n distinct roots for each n (k and
j are the same when k = j (mod n)), which form a regular polygon with vertices on the
unit circle. They are ordered counterclockwise with increasing k, starting with 0 = 1.
Examples
754
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
The roots of unity up to n = 4:
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = True
>>> nprint(unitroots(1))
[1.0]
>>> nprint(unitroots(2))
[1.0, -1.0]
>>> nprint(unitroots(3))
[1.0, (-0.5 + 0.866025j), (-0.5 - 0.866025j)]
>>> nprint(unitroots(4))
[1.0, (0.0 + 1.0j), -1.0, (0.0 - 1.0j)]
Roots of unity form a geometric series that sums to 0:
>>> mp.dps = 50
>>> chop(fsum(unitroots(25)))
0.0
Primitive roots up to n = 4:
>>> mp.dps = 15
>>> nprint(unitroots(1, primitive=True))
[1.0]
>>> nprint(unitroots(2, primitive=True))
[-1.0]
>>> nprint(unitroots(3, primitive=True))
[(-0.5 + 0.866025j), (-0.5 - 0.866025j)]
>>> nprint(unitroots(4, primitive=True))
[(0.0 + 1.0j), (0.0 - 1.0j)]
There are only four primitive 12th roots:
>>> nprint(unitroots(12, primitive=True))
[(0.866025 + 0.5j), (-0.866025 + 0.5j), (-0.866025 - 0.5j), (0.866025 - 0.5j)]
The n-th roots of unity form a group, the cyclic group of order n. Any primitive root r is
a generator for this group, meaning that r0 , r1 , . . . , rn1 gives the whole set of unit roots
(in some permuted order):
>>> for r in unitroots(6): print(r)
...
1.0
(0.5 + 0.866025403784439j)
(-0.5 + 0.866025403784439j)
-1.0
(-0.5 - 0.866025403784439j)
(0.5 - 0.866025403784439j)
>>> r = unitroots(6, primitive=True)[1]
>>> for k in range(6): print(chop(r**k))
...
1.0
(0.5 - 0.866025403784439j)
(-0.5 - 0.866025403784439j)
-1.0
(-0.5 + 0.866025403784438j)
(0.5 + 0.866025403784438j)
The number of primitive roots equals the Euler totient function (n):
5.15. Welcome to mpmaths documentation!
755
SymPy Documentation, Release 0.7.6
>>> [len(unitroots(n, primitive=True)) for n in range(1,20)]
[1, 1, 2, 2, 4, 2, 6, 4, 6, 4, 10, 4, 12, 6, 8, 8, 16, 6, 18]
Exponentiation
exp()
mpmath.exp(x, **kwargs)
Computes the exponential function,
exp(x) = ex =
xk
k=0
k!
For complex numbers, the exponential function also satises
exp(x + yi) = ex (cos y + i sin y).
Basic examples
Some values of the exponential function:
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> exp(0)
1.0
>>> exp(1)
2.718281828459045235360287
>>> exp(-1)
0.3678794411714423215955238
>>> exp(inf)
+inf
>>> exp(-inf)
0.0
Arguments can be arbitrarily large:
>>> exp(10000)
8.806818225662921587261496e+4342
>>> exp(-10000)
1.135483865314736098540939e-4343
Evaluation is supported for interval arguments via mpmath.iv.exp():
>>> iv.dps = 25; iv.pretty = True
>>> iv.exp([-inf,0])
[0.0, 1.0]
>>> iv.exp([0,1])
[1.0, 2.71828182845904523536028749558]
The exponential function can be evaluated eciently to arbitrary precision:
>>> mp.dps = 10000
>>> exp(pi)
23.140692632779269005729...8984304016040616
Functional properties
Numerical verication of Eulers identity for the complex exponential function:
756
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> mp.dps = 15
>>> exp(j*pi)+1
(0.0 + 1.22464679914735e-16j)
>>> chop(exp(j*pi)+1)
0.0
This recovers the coecients (reciprocal factorials) in the Maclaurin series expansion
of exp:
>>> nprint(taylor(exp, 0, 5))
[1.0, 1.0, 0.5, 0.166667, 0.0416667, 0.00833333]
The exponential function is its own derivative and antiderivative:
>>> exp(pi)
23.1406926327793
>>> diff(exp, pi)
23.1406926327793
>>> quad(exp, [-inf, pi])
23.1406926327793
The exponential function can be evaluated using various methods, including direct summation of the series, limits, and solving the dening dierential equation:
>>> nsum(lambda k: pi**k/fac(k), [0,inf])
23.1406926327793
>>> limit(lambda k: (1+pi/k)**k, inf)
23.1406926327793
>>> odefun(lambda t, x: x, 0, 1)(pi)
23.1406926327793
power()
mpmath.power(x, y)
Converts x and y to mpmath numbers and evaluates xy = exp(y log(x)):
>>> from sympy.mpmath import *
>>> mp.dps = 30; mp.pretty = True
>>> power(2, 0.5)
1.41421356237309504880168872421
This shows the leading few digits of a large Mersenne prime (performing the exact calculation 2**43112609-1 and displaying the result in Python would be very slow):
>>> power(2, 43112609)-1
3.16470269330255923143453723949e+12978188
expj()
mpmath.expj(x, **kwargs)
Convenience function for computing eix :
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> expj(0)
(1.0 + 0.0j)
>>> expj(-1)
(0.5403023058681397174009366 - 0.8414709848078965066525023j)
>>> expj(j)
5.15. Welcome to mpmaths documentation!
757
SymPy Documentation, Release 0.7.6
(0.3678794411714423215955238 + 0.0j)
>>> expj(1+j)
(0.1987661103464129406288032 + 0.3095598756531121984439128j)
expjpi()
mpmath.expjpi(x, **kwargs)
Convenience function for computing eix . Evaluation is accurate near zeros (see also
cospi() (page 770), sinpi() (page 770)):
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> expjpi(0)
(1.0 + 0.0j)
>>> expjpi(1)
(-1.0 + 0.0j)
>>> expjpi(0.5)
(0.0 + 1.0j)
>>> expjpi(-1)
(-1.0 + 0.0j)
>>> expjpi(j)
(0.04321391826377224977441774 + 0.0j)
>>> expjpi(1+j)
(-0.04321391826377224977441774 + 0.0j)
expm1()
mpmath.expm1(x)
Computes ex 1, accurately for small x.
Unlike the expression exp(x) - 1, expm1(x) does not suer from potentially catastrophic
cancellation:
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = True
>>> exp(1e-10)-1; print(expm1(1e-10))
1.00000008274037e-10
1.00000000005e-10
>>> exp(1e-20)-1; print(expm1(1e-20))
0.0
1.0e-20
>>> 1/(exp(1e-20)-1)
Traceback (most recent call last):
...
ZeroDivisionError
>>> 1/expm1(1e-20)
1.0e+20
Traceback (most recent call last):
...
ZeroDivisionError
Evaluation works for extremely tiny values:
>>> expm1(0)
0.0
>>> expm1(1e-10000000)
1.0e-10000000
758
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
powm1()
mpmath.powm1(x, y)
Computes xy 1, accurately when xy is very close to 1.
This avoids potentially catastrophic cancellation:
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = True
>>> power(0.99999995, 1e-10) - 1
0.0
>>> powm1(0.99999995, 1e-10)
-5.00000012791934e-18
Powers exactly equal to 1, and only those powers, yield 0 exactly:
>>> powm1(-j, 4)
(0.0 + 0.0j)
>>> powm1(3, 0)
0.0
>>> powm1(fadd(-1, 1e-100, exact=True), 4)
-4.0e-100
Evaluation works for extremely tiny y:
>>> powm1(2, 1e-100000)
6.93147180559945e-100001
>>> powm1(j, 1e-1000)
(-1.23370055013617e-2000 + 1.5707963267949e-1000j)
Logarithms
log()
mpmath.log(x, b=None)
Computes the base-b logarithm of x, logb (x). If b is unspecied, log() (page 759) computes the natural (base e) logarithm and is equivalent to ln() (page 760). In general,
the base b logarithm is dened in terms of the natural logarithm as logb (x) = ln(x)/ ln(b).
By convention, we take log(0) = .
The natural logarithm is real if x > 0 and complex if x < 0 or if x is complex. The principal
branch of the complex logarithm is used, meaning that =(ln(x)) = < arg(x) .
Examples
Some basic values and limits:
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = True
>>> log(1)
0.0
>>> log(2)
0.693147180559945
>>> log(1000,10)
3.0
>>> log(4, 16)
0.5
>>> log(j)
(0.0 + 1.5707963267949j)
>>> log(-1)
5.15. Welcome to mpmaths documentation!
759
SymPy Documentation, Release 0.7.6
(0.0 + 3.14159265358979j)
>>> log(0)
-inf
>>> log(inf)
+inf
The natural logarithm is the antiderivative of 1/x:
>>> quad(lambda x: 1/x, [1, 5])
1.6094379124341
>>> log(5)
1.6094379124341
>>> diff(log, 10)
0.1
The Taylor series expansion of the natural logarithm around x = 1 has coecients
(1)n+1 /n:
>>> nprint(taylor(log, 1, 7))
[0.0, 1.0, -0.5, 0.333333, -0.25, 0.2, -0.166667, 0.142857]
log() (page 759) supports arbitrary precision evaluation:
>>> mp.dps = 50
>>> log(pi)
1.1447298858494001741434273513530587116472948129153
>>> log(pi, pi**3)
0.33333333333333333333333333333333333333333333333333
>>> mp.dps = 25
>>> log(3+4j)
(1.609437912434100374600759 + 0.9272952180016122324285125j)
ln()
mpmath.ln(x, **kwargs)
Computes the base-b logarithm of x, logb (x). If b is unspecied, log() (page 759) computes the natural (base e) logarithm and is equivalent to ln() (page 760). In general,
the base b logarithm is dened in terms of the natural logarithm as logb (x) = ln(x)/ ln(b).
By convention, we take log(0) = .
The natural logarithm is real if x > 0 and complex if x < 0 or if x is complex. The principal
branch of the complex logarithm is used, meaning that =(ln(x)) = < arg(x) .
Examples
Some basic values and limits:
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = True
>>> log(1)
0.0
>>> log(2)
0.693147180559945
>>> log(1000,10)
3.0
>>> log(4, 16)
0.5
>>> log(j)
(0.0 + 1.5707963267949j)
760
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> log(-1)
(0.0 + 3.14159265358979j)
>>> log(0)
-inf
>>> log(inf)
+inf
The natural logarithm is the antiderivative of 1/x:
>>> quad(lambda x: 1/x, [1, 5])
1.6094379124341
>>> log(5)
1.6094379124341
>>> diff(log, 10)
0.1
The Taylor series expansion of the natural logarithm around x = 1 has coecients
(1)n+1 /n:
>>> nprint(taylor(log, 1, 7))
[0.0, 1.0, -0.5, 0.333333, -0.25, 0.2, -0.166667, 0.142857]
log() (page 759) supports arbitrary precision evaluation:
>>> mp.dps = 50
>>> log(pi)
1.1447298858494001741434273513530587116472948129153
>>> log(pi, pi**3)
0.33333333333333333333333333333333333333333333333333
>>> mp.dps = 25
>>> log(3+4j)
(1.609437912434100374600759 + 0.9272952180016122324285125j)
log10()
mpmath.log10(x)
Computes the base-10 logarithm of x, log10 (x). log10(x) is equivalent to log(x, 10).
Lambert W function
lambertw()
mpmath.lambertw(z, k=0)
The Lambert W function W (z) is dened as the inverse function of w exp(w). In other
words, the value of W (z) is such that z = W (z) exp(W (z)) for any complex number z.
The Lambert W function is a multivalued function with innitely many branches Wk (z),
indexed by k Z. Each branch gives a dierent solution w of the equation z = w exp(w).
All branches are supported by lambertw() (page 761):
lambertw(z) gives the principal solution (branch 0)
lambertw(z, k) gives the solution on branch k
The Lambert W function has two partially real branches: the principal branch (k = 0) is
real for real z > 1/e, and the k = 1 branch is real for 1/e < z < 0. All branches except
k = 0 have a logarithmic singularity at z = 0.
The denition, implementation and choice of branches is based on [Corless] (page 1910).
5.15. Welcome to mpmaths documentation!
761
SymPy Documentation, Release 0.7.6
Plots
# Branches 0 and -1 of the Lambert W function
plot([lambertw, lambda x: lambertw(x,-1)], [-2,2], [-5,2], points=2000)
# Principal branch of the Lambert W function W(z)
cplot(lambertw, [-1,1], [-1,1], points=50000)
762
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Basic examples
The Lambert W function is the inverse of w exp(w):
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> w = lambertw(1)
>>> w
0.5671432904097838729999687
>>> w*exp(w)
1.0
Any branch gives a valid inverse:
>>> w = lambertw(1, k=3)
>>> w
(-2.853581755409037807206819 + 17.11353553941214591260783j)
>>> w = lambertw(1, k=25)
>>> w
(-5.047020464221569709378686 + 155.4763860949415867162066j)
>>> chop(w*exp(w))
1.0
Applications to equation-solving
The Lambert W function may be used to ...
solve various kinds of equations, such as nding
z
the value of the innite power tower z z :
5.15. Welcome to mpmaths documentation!
763
SymPy Documentation, Release 0.7.6
>>> def tower(z, n):
...
if n == 0:
...
return z
...
return z ** tower(z, n-1)
...
>>> tower(mpf(0.5), 100)
0.6411857445049859844862005
>>> -lambertw(-log(0.5))/log(0.5)
0.6411857445049859844862005
Properties
The Lambert W function grows roughly like the natural logarithm for large arguments:
>>> lambertw(1000); log(1000)
5.249602852401596227126056
6.907755278982137052053974
>>> lambertw(10**100); log(10**100)
224.8431064451185015393731
230.2585092994045684017991
The principal branch of the Lambert W function has a rational Taylor series expansion
around z = 0:
>>> nprint(taylor(lambertw, 0, 6), 10)
[0.0, 1.0, -1.0, 1.5, -2.666666667, 5.208333333, -10.8]
Some special values and limits are:
>>> lambertw(0)
0.0
>>> lambertw(1)
0.5671432904097838729999687
>>> lambertw(e)
1.0
>>> lambertw(inf)
+inf
>>> lambertw(0, k=-1)
-inf
>>> lambertw(0, k=3)
-inf
>>> lambertw(inf, k=2)
(+inf + 12.56637061435917295385057j)
>>> lambertw(inf, k=3)
(+inf + 18.84955592153875943077586j)
>>> lambertw(-inf, k=3)
(+inf + 21.9911485751285526692385j)
The k = 0 and k = 1 branches join at z = 1/e where W (z) = 1 for both branches.
Since 1/e can only be represented approximately with binary oating-point numbers,
evaluating the Lambert W function at this point only gives 1 approximately:
>>> lambertw(-1/e, 0)
-0.9999999999998371330228251
>>> lambertw(-1/e, -1)
-1.000000000000162866977175
If 1/e happens to round in the negative direction, there might be a small imaginary
part:
764
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> mp.dps = 15
>>> lambertw(-1/e)
(-1.0 + 8.22007971483662e-9j)
>>> lambertw(-1/e+eps)
-0.999999966242188
References
1.[Corless] (page 1910)
Arithmetic-geometric mean
agm()
mpmath.agm(a, b=1)
agm(a, b) computes the arithmetic-geometric mean of a and b, dened as the limit of
the following iteration:
a0 = a
b0 = b
an + bn
an+1 =
2
bn+1 = an bn
This function can be called with a single argument, computing agm(a, 1) = agm(1, a).
Examples
It is a well-known theorem that the geometric mean of two distinct positive numbers
is less than the arithmetic mean. It follows that the arithmetic-geometric mean lies
between the two means:
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = True
>>> a = mpf(3)
>>> b = mpf(4)
>>> sqrt(a*b)
3.46410161513775
>>> agm(a,b)
3.48202767635957
>>> (a+b)/2
3.5
The arithmetic-geometric mean is scale-invariant:
>>> agm(10*e, 10*pi)
29.261085515723
>>> 10*agm(e, pi)
29.261085515723
As an order-of-magnitude estimate, agm(1, x) x for large x:
>>> agm(10**10)
643448704.760133
>>> agm(10**50)
1.34814309345871e+48
For tiny x, agm(1, x) /(2 log(x/4)):
5.15. Welcome to mpmaths documentation!
765
SymPy Documentation, Release 0.7.6
>>> agm(0.01)
0.262166887202249
>>> -pi/2/log(0.0025)
0.262172347753122
The arithmetic-geometric mean can also be computed for complex numbers:
>>> agm(3, 2+j)
(2.51055133276184 + 0.547394054060638j)
The AGM iteration converges very quickly (each step doubles the number of correct
digits), so agm() (page 765) supports ecient high-precision evaluation:
>>> mp.dps = 10000
>>> a = agm(1,2)
>>> str(a)[-10:]
1679581912
Mathematical relations
The arithmetic-geometric mean may be used to evaluate the following two parametric
denite integrals:
1
I1 =
dx
2
2
(x + a )(x2 + b2 )
0
/2
1
I2 =
dx
0
a2 cos2 (x) + b2 sin2 (x)
We have:
>>> mp.dps = 15
>>> a = 3
>>> b = 4
>>> f1 = lambda x: ((x**2+a**2)*(x**2+b**2))**-0.5
>>> f2 = lambda x: ((a*cos(x))**2 + (b*sin(x))**2)**-0.5
>>> quad(f1, [0, inf])
0.451115405388492
>>> quad(f2, [0, pi/2])
0.451115405388492
>>> pi/(2*agm(a,b))
0.451115405388492
A formula for (1/4):
>>> gamma(0.25)
3.62560990822191
>>> sqrt(2*sqrt(2*pi**3)/agm(1,sqrt(2)))
3.62560990822191
Possible issues
The branch cut chosen for complex a and b is somewhat arbitrary.
Trigonometric functions
Except where otherwise noted, the trigonometric functions take a radian angle as input and
the inverse trigonometric functions return radian angles.
766
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
The ordinary trigonometric functions are single-valued functions dened everywhere in the
complex plane (except at the poles of tan, sec, csc, and cot). They are dened generally via
the exponential function, e.g.
cos(x) =
eix + eix
.
2
The inverse trigonometric functions are multivalued, thus requiring branch cuts, and are generally real-valued only on a part of the real line. Denitions and branch cuts are given in the
documentation of each function. The branch cut conventions used by mpmath are essentially
the same as those found in most standard mathematical software, such as Mathematica and
Pythons own cmath libary (as of Python 2.6; earlier Python versions implement some functions erroneously).
Degree-radian conversion
degrees()
mpmath.degrees(x)
Converts the radian angle x to a degree angle:
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = True
>>> degrees(pi/3)
60.0
radians()
mpmath.radians(x)
Converts the degree angle x to radians:
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = True
>>> radians(60)
1.0471975511966
Trigonometric functions
cos()
mpmath.cos(x, **kwargs)
Computes the cosine of x, cos(x).
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> cos(pi/3)
0.5
>>> cos(100000001)
-0.9802850113244713353133243
>>> cos(2+3j)
(-4.189625690968807230132555 - 9.109227893755336597979197j)
>>> cos(inf)
nan
>>> nprint(chop(taylor(cos, 0, 6)))
[1.0, 0.0, -0.5, 0.0, 0.0416667, 0.0, -0.00138889]
5.15. Welcome to mpmaths documentation!
767
SymPy Documentation, Release 0.7.6
Intervals are supported via mpmath.iv.cos():
>>> iv.dps = 25; iv.pretty = True
>>> iv.cos([0,1])
[0.540302305868139717400936602301, 1.0]
>>> iv.cos([0,2])
[-0.41614683654714238699756823214, 1.0]
sin()
mpmath.sin(x, **kwargs)
Computes the sine of x, sin(x).
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> sin(pi/3)
0.8660254037844386467637232
>>> sin(100000001)
0.1975887055794968911438743
>>> sin(2+3j)
(9.1544991469114295734673 - 4.168906959966564350754813j)
>>> sin(inf)
nan
>>> nprint(chop(taylor(sin, 0, 6)))
[0.0, 1.0, 0.0, -0.166667, 0.0, 0.00833333, 0.0]
Intervals are supported via mpmath.iv.sin():
>>> iv.dps = 25; iv.pretty = True
>>> iv.sin([0,1])
[0.0, 0.841470984807896506652502331201]
>>> iv.sin([0,2])
[0.0, 1.0]
tan()
mpmath.tan(x, **kwargs)
sin(x)
. The tangent function is singular at x = (n +
Computes the tangent of x, tan(x) = cos(x)
1/2), but tan(x) always returns a nite result since (n + 1/2) cannot be represented
exactly using oating-point arithmetic.
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> tan(pi/3)
1.732050807568877293527446
>>> tan(100000001)
-0.2015625081449864533091058
>>> tan(2+3j)
(-0.003764025641504248292751221 + 1.003238627353609801446359j)
>>> tan(inf)
nan
>>> nprint(chop(taylor(tan, 0, 6)))
[0.0, 1.0, 0.0, 0.333333, 0.0, 0.133333, 0.0]
Intervals are supported via mpmath.iv.tan():
>>> iv.dps = 25; iv.pretty = True
>>> iv.tan([0,1])
768
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
[0.0, 1.55740772465490223050697482944]
>>> iv.tan([0,2]) # Interval includes a singularity
[-inf, +inf]
sec()
mpmath.sec(x)
1
Computes the secant of x, sec(x) = cos(x)
. The secant function is singular at x = (n + 1/2),
but sec(x) always returns a nite result since (n + 1/2) cannot be represented exactly
using oating-point arithmetic.
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> sec(pi/3)
2.0
>>> sec(10000001)
-1.184723164360392819100265
>>> sec(2+3j)
(-0.04167496441114427004834991 + 0.0906111371962375965296612j)
>>> sec(inf)
nan
>>> nprint(chop(taylor(sec, 0, 6)))
[1.0, 0.0, 0.5, 0.0, 0.208333, 0.0, 0.0847222]
Intervals are supported via mpmath.iv.sec():
>>> iv.dps = 25; iv.pretty = True
>>> iv.sec([0,1])
[1.0, 1.85081571768092561791175326276]
>>> iv.sec([0,2]) # Interval includes a singularity
[-inf, +inf]
csc()
mpmath.csc(x)
1
Computes the cosecant of x, csc(x) = sin(x)
. This cosecant function is singular at x = n,
but with the exception of the point x = 0, csc(x) returns a nite result since n cannot
be represented exactly using oating-point arithmetic.
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> csc(pi/3)
1.154700538379251529018298
>>> csc(10000001)
-1.864910497503629858938891
>>> csc(2+3j)
(0.09047320975320743980579048 + 0.04120098628857412646300981j)
>>> csc(inf)
nan
Intervals are supported via mpmath.iv.csc():
>>> iv.dps = 25; iv.pretty = True
>>> iv.csc([0,1]) # Interval includes a singularity
[1.18839510577812121626159943988, +inf]
>>> iv.csc([0,2])
[1.0, +inf]
5.15. Welcome to mpmaths documentation!
769
SymPy Documentation, Release 0.7.6
cot()
mpmath.cot(x)
1
Computes the cotangent of x, cot(x) = tan(x)
= cos(x)
sin(x) . This cotangent function is singular
at x = n, but with the exception of the point x = 0, cot(x) returns a nite result since
n cannot be represented exactly using oating-point arithmetic.
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> cot(pi/3)
0.5773502691896257645091488
>>> cot(10000001)
1.574131876209625656003562
>>> cot(2+3j)
(-0.003739710376336956660117409 - 0.9967577965693583104609688j)
>>> cot(inf)
nan
Intervals are supported via mpmath.iv.cot():
>>> iv.dps = 25; iv.pretty = True
>>> iv.cot([0,1]) # Interval includes a singularity
[0.642092615934330703006419974862, +inf]
>>> iv.cot([1,2])
[-inf, +inf]
Trigonometric functions with modied argument
cospi()
mpmath.cospi(x, **kwargs)
Computes cos(x), more accurately than the expression cos(pi*x):
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = True
>>> cospi(10**10), cos(pi*(10**10))
(1.0, 0.999999999997493)
>>> cospi(10**10+0.5), cos(pi*(10**10+0.5))
(0.0, 1.59960492420134e-6)
sinpi()
mpmath.sinpi(x, **kwargs)
Computes sin(x), more accurately than the expression sin(pi*x):
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = True
>>> sinpi(10**10), sin(pi*(10**10))
(0.0, -2.23936276195592e-6)
>>> sinpi(10**10+0.5), sin(pi*(10**10+0.5))
(1.0, 0.999999999998721)
Inverse trigonometric functions
770
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
acos()
mpmath.acos(x, **kwargs)
Computes the inverse cosine or arccosine of x, cos1 (x). Since 1 cos(x) 1 for real x,
the inverse cosine is real-valued only for 1 x 1. On this interval, acos() (page 771)
is dened to be a monotonically decreasing function assuming values between + and 0.
Basic values are:
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> acos(-1)
3.141592653589793238462643
>>> acos(0)
1.570796326794896619231322
>>> acos(1)
0.0
>>> nprint(chop(taylor(acos, 0, 6)))
[1.5708, -1.0, 0.0, -0.166667, 0.0, -0.075, 0.0]
acos() (page 771) is dened so as to be a proper inverse function of cos() for 0 < .
We have cos(cos1 (x)) = x for all x, but cos1 (cos(x)) = x only for 0 <[x] < :
>>> for x in [1, 10, -1, 2+3j, 10+3j]:
...
print(%s %s % (cos(acos(x)), acos(cos(x))))
...
1.0 1.0
(10.0 + 0.0j) 2.566370614359172953850574
-1.0 1.0
(2.0 + 3.0j) (2.0 + 3.0j)
(10.0 + 3.0j) (2.566370614359172953850574 - 3.0j)
The inverse cosine has two branch points: x = 1. acos() (page 771) places the branch
cuts along the line segments (, 1) and (+1, +). In general,
)
(
cos1 (x) = + i log ix + 1 x2
2
where the principal-branch log and square root are implied.
asin()
mpmath.asin(x, **kwargs)
Computes the inverse sine or arcsine of x, sin1 (x). Since 1 sin(x) 1 for real x, the
inverse sine is real-valued only for 1 x 1. On this interval, it is dened to be a
monotonically increasing function assuming values between /2 and /2.
Basic values are:
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> asin(-1)
-1.570796326794896619231322
>>> asin(0)
0.0
>>> asin(1)
1.570796326794896619231322
>>> nprint(chop(taylor(asin, 0, 6)))
[0.0, 1.0, 0.0, 0.166667, 0.0, 0.075, 0.0]
5.15. Welcome to mpmaths documentation!
771
SymPy Documentation, Release 0.7.6
asin() (page 771) is dened so as to be a proper inverse function of sin() for /2 < <
/2. We have sin(sin1 (x)) = x for all x, but sin1 (sin(x)) = x only for /2 < <[x] < /2:
>>> for x in [1, 10, -1, 1+3j, -2+3j]:
...
print(%s %s % (chop(sin(asin(x))), asin(sin(x))))
...
1.0 1.0
10.0 -0.5752220392306202846120698
-1.0 -1.0
(1.0 + 3.0j) (1.0 + 3.0j)
(-2.0 + 3.0j) (-1.141592653589793238462643 - 3.0j)
The inverse sine has two branch points: x = 1. asin() (page 771) places the branch
cuts along the line segments (, 1) and (+1, +). In general,
)
(
sin1 (x) = i log ix + 1 x2
where the principal-branch log and square root are implied.
atan()
mpmath.atan(x, **kwargs)
Computes the inverse tangent or arctangent of x, tan1 (x). This is a real-valued function
for all real x, with range (/2, /2).
Basic values are:
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> atan(-inf)
-1.570796326794896619231322
>>> atan(-1)
-0.7853981633974483096156609
>>> atan(0)
0.0
>>> atan(1)
0.7853981633974483096156609
>>> atan(inf)
1.570796326794896619231322
>>> nprint(chop(taylor(atan, 0, 6)))
[0.0, 1.0, 0.0, -0.333333, 0.0, 0.2, 0.0]
The inverse tangent is often used to compute angles. However, the atan2 function is
often better for this as it preserves sign (see atan2() (page 773)).
atan() (page 772) is dened so as to be a proper inverse function of tan() for /2 < <
/2. We have tan(tan1 (x)) = x for all x, but tan1 (tan(x)) = x only for /2 < <[x] < /2:
>>> mp.dps = 25
>>> for x in [1, 10, -1, 1+3j, -2+3j]:
...
print(%s %s % (tan(atan(x)), atan(tan(x))))
...
1.0 1.0
10.0 0.5752220392306202846120698
-1.0 -1.0
(1.0 + 3.0j) (1.000000000000000000000001 + 3.0j)
(-2.0 + 3.0j) (1.141592653589793238462644 + 3.0j)
772
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
The inverse tangent has two branch points: x = i. atan() (page 772) places the branch
cuts along the line segments (i, i) and (+i, +i). In general,
tan1 (x) =
i
(log(1 ix) log(1 + ix))
2
where the principal-branch log is implied.
atan2()
mpmath.atan2(y, x)
Computes the two-argument arctangent, atan2(y, x), giving the signed angle between
the positive x-axis and the point (x, y) in the 2D plane. This function is dened for real x
and y only.
The two-argument arctangent essentially computes atan(y/x), but accounts for the signs
of both x and y to give the angle for the correct quadrant. The following examples illustrate the dierence:
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = True
>>> atan2(1,1), atan(1/1.)
(0.785398163397448, 0.785398163397448)
>>> atan2(1,-1), atan(1/-1.)
(2.35619449019234, -0.785398163397448)
>>> atan2(-1,1), atan(-1/1.)
(-0.785398163397448, -0.785398163397448)
>>> atan2(-1,-1), atan(-1/-1.)
(-2.35619449019234, 0.785398163397448)
The angle convention is the same as that used for the complex argument; see arg()
(page 734).
asec()
mpmath.asec(x)
Computes the inverse secant of x, sec1 (x) = cos1 (1/x).
acsc()
mpmath.acsc(x)
Computes the inverse cosecant of x, csc1 (x) = sin1 (1/x).
acot()
mpmath.acot(x)
Computes the inverse cotangent of x, cot1 (x) = tan1 (1/x).
Sinc function
sinc()
mpmath.sinc(x)
sinc(x) computes the unnormalized sinc function, dened as
{
sin(x)/x, if x 6= 0
sinc(x) =
1,
if x = 0.
5.15. Welcome to mpmaths documentation!
773
SymPy Documentation, Release 0.7.6
See sincpi() (page 774) for the normalized sinc function.
Simple values and limits include:
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = True
>>> sinc(0)
1.0
>>> sinc(1)
0.841470984807897
>>> sinc(inf)
0.0
The integral of the sinc function is the sine integral Si:
>>> quad(sinc, [0, 1])
0.946083070367183
>>> si(1)
0.946083070367183
sincpi()
mpmath.sincpi(x)
sincpi(x) computes the normalized sinc function, dened as
{
sin(x)/(x), if x 6= 0
sinc (x) =
1,
if x = 0.
Equivalently, we have sinc (x) = sinc(x).
The normalization entails that the function integrates to unity over the entire real line:
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = True
>>> quadosc(sincpi, [-inf, inf], period=2.0)
1.0
Like, sinpi() (page 770), sincpi() (page 774) is evaluated accurately at its roots:
>>> sincpi(10)
0.0
Hyperbolic functions
Hyperbolic functions
cosh()
mpmath.cosh(x, **kwargs)
Computes the hyperbolic cosine of x, cosh(x) = (ex + ex )/2. Values and limits include:
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> cosh(0)
1.0
>>> cosh(1)
1.543080634815243778477906
>>> cosh(-inf), cosh(+inf)
(+inf, +inf)
774
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
The hyperbolic cosine is an even, convex function with a global minimum at x = 0, having
a Maclaurin series that starts:
>>> nprint(chop(taylor(cosh, 0, 5)))
[1.0, 0.0, 0.5, 0.0, 0.0416667, 0.0]
Generalized to complex numbers, the hyperbolic cosine is equivalent to a cosine with
the argument rotated in the imaginary direction, or cosh x = cos ix:
>>> cosh(2+3j)
(-3.724545504915322565473971 + 0.5118225699873846088344638j)
>>> cos(3-2j)
(-3.724545504915322565473971 + 0.5118225699873846088344638j)
sinh()
mpmath.sinh(x, **kwargs)
Computes the hyperbolic sine of x, sinh(x) = (ex ex )/2. Values and limits include:
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> sinh(0)
0.0
>>> sinh(1)
1.175201193643801456882382
>>> sinh(-inf), sinh(+inf)
(-inf, +inf)
The hyperbolic sine is an odd function, with a Maclaurin series that starts:
>>> nprint(chop(taylor(sinh, 0, 5)))
[0.0, 1.0, 0.0, 0.166667, 0.0, 0.00833333]
Generalized to complex numbers, the hyperbolic sine is essentially a sine with a rotation
i applied to the argument; more precisely, sinh x = i sin ix:
>>> sinh(2+3j)
(-3.590564589985779952012565 + 0.5309210862485198052670401j)
>>> j*sin(3-2j)
(-3.590564589985779952012565 + 0.5309210862485198052670401j)
tanh()
mpmath.tanh(x, **kwargs)
Computes the hyperbolic tangent of x, tanh(x) = sinh(x)/ cosh(x). Values and limits include:
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> tanh(0)
0.0
>>> tanh(1)
0.7615941559557648881194583
>>> tanh(-inf), tanh(inf)
(-1.0, 1.0)
The hyperbolic tangent is an odd, sigmoidal function, similar to the inverse tangent and
error function. Its Maclaurin series is:
5.15. Welcome to mpmaths documentation!
775
SymPy Documentation, Release 0.7.6
>>> nprint(chop(taylor(tanh, 0, 5)))
[0.0, 1.0, 0.0, -0.333333, 0.0, 0.133333]
Generalized to complex numbers, the hyperbolic tangent is essentially a tangent with a
rotation i applied to the argument; more precisely, tanh x = i tan ix:
>>> tanh(2+3j)
(0.9653858790221331242784803 - 0.009884375038322493720314034j)
>>> j*tan(3-2j)
(0.9653858790221331242784803 - 0.009884375038322493720314034j)
sech()
mpmath.sech(x)
Computes the hyperbolic secant of x, sech(x) =
1
cosh(x) .
csch()
mpmath.csch(x)
Computes the hyperbolic cosecant of x, csch(x) =
1
sinh(x) .
coth()
mpmath.coth(x)
Computes the hyperbolic cotangent of x, coth(x) =
cosh(x)
sinh(x) .
Inverse hyperbolic functions
acosh()
mpmath.acosh(x, **kwargs)
1
Computes the inverse hyperbolic cosine of x, cosh (x) = log(x + x + 1 x 1).
asinh()
mpmath.asinh(x, **kwargs)
1
Computes the inverse hyperbolic sine of x, sinh (x) = log(x + 1 + x2 ).
atanh()
mpmath.atanh(x, **kwargs)
1
Computes the inverse hyperbolic tangent of x, tanh (x) =
1
2
(log(1 + x) log(1 x)).
asech()
mpmath.asech(x)
1
1
Computes the inverse hyperbolic secant of x, sech (x) = cosh (1/x).
acsch()
mpmath.acsch(x)
1
1
Computes the inverse hyperbolic cosecant of x, csch (x) = sinh (1/x).
776
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
acoth()
mpmath.acoth(x)
1
1
Computes the inverse hyperbolic cotangent of x, coth (x) = tanh (1/x).
Factorials and gamma functions
Factorials and factorial-like sums and products are basic tools of combinatorics and number theory. Much like the exponential function is fundamental to dierential equations and
analysis in general, the factorial function (and its extension to complex numbers, the gamma
function) is fundamental to dierence equations and functional equations.
A large selection of factorial-like functions is implemented in mpmath. All functions support
complex arguments, and arguments may be arbitrarily large. Results are numerical approximations, so to compute exact values a high enough precision must be set manually:
>>> mp.dps = 15; mp.pretty = True
>>> fac(100)
9.33262154439442e+157
>>> print int(_)
# most digits are wrong
93326215443944150965646704795953882578400970373184098831012889540582227238570431
295066113089288327277825849664006524270554535976289719382852181865895959724032
>>> mp.dps = 160
>>> fac(100)
93326215443944152681699238856266700490715968264381621468592963895217599993229915
608941463976156518286253697920827223758251185210916864000000000000000000000000.0
The gamma and polygamma functions are closely related to Zeta functions, L-series and polylogarithms (page 931). See also q-functions (page 968) for q-analogs of factorial-like functions.
Factorials
factorial()/fac()
mpmath.factorial(x, **kwargs)
Computes the factorial, x!. For integers n 0, we have n! = 1 2 (n 1) n and more
generally the factorial is dened for real or complex x by x! = (x + 1).
Examples
Basic values and limits:
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = True
>>> for k in range(6):
...
print(%s %s % (k, fac(k)))
...
0 1.0
1 1.0
2 2.0
3 6.0
4 24.0
5 120.0
>>> fac(inf)
+inf
>>> fac(0.5), sqrt(pi)/2
(0.886226925452758, 0.886226925452758)
5.15. Welcome to mpmaths documentation!
777
SymPy Documentation, Release 0.7.6
For large positive x, x! can be approximated by Stirlings formula:
>>> x = 10**10
>>> fac(x)
2.32579620567308e+95657055186
>>> sqrt(2*pi*x)*(x/e)**x
2.32579597597705e+95657055186
fac() supports evaluation for astronomically large values:
>>> fac(10**30)
6.22311232304258e+29565705518096748172348871081098
Reciprocal factorials appear in the Taylor series of the exponential function (among many
other contexts):
>>> nsum(lambda k:
(2.71828182845905,
>>> nsum(lambda k:
(23.1406926327793,
1/fac(k), [0, inf]), exp(1)
2.71828182845905)
pi**k/fac(k), [0, inf]), exp(pi)
23.1406926327793)
fac2()
mpmath.fac2(x)
Computes the double factorial x!!, dened for integers x > 0 by
{
1 3 (x 2) x x odd
x!! =
2 4 (x 2) x x even
and more generally by [1]
x!! = 2x/2
( )(cos(x)1)/4
2
(x
2
)
+1 .
Examples
The integer sequence of double factorials begins:
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = True
>>> nprint([fac2(n) for n in range(10)])
[1.0, 1.0, 2.0, 3.0, 8.0, 15.0, 48.0, 105.0, 384.0, 945.0]
For large x, double factorials follow a Stirling-like asymptotic approximation:
>>> x = mpf(10000)
>>> fac2(x)
5.97272691416282e+17830
>>> sqrt(pi)*x**((x+1)/2)*exp(-x/2)
5.97262736954392e+17830
The recurrence formula x!! = x(x 2)!! can be reversed to dene the double factorial of
negative odd integers (but not negative even integers):
>>> fac2(-1), fac2(-3), fac2(-5), fac2(-7)
(1.0, -1.0, 0.333333333333333, -0.0666666666666667)
>>> fac2(-2)
Traceback (most recent call last):
...
ValueError: gamma function pole
778
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Traceback (most recent call last):
...
ValueError: gamma function pole
With the exception of the poles at negative even integers, fac2() (page 778) supports
evaluation for arbitrary complex arguments. The recurrence formula is valid generally:
>>> fac2(pi+2j)
(-1.3697207890154e-12 + 3.93665300979176e-12j)
>>> (pi+2j)*fac2(pi-2+2j)
(-1.3697207890154e-12 + 3.93665300979176e-12j)
Double factorials should not be confused with nested factorials, which are immensely
larger:
>>> fac(fac(20))
5.13805976125208e+43675043585825292774
>>> fac2(20)
3715891200.0
Double factorials appear, among other things, in series expansions of Gaussian functions
and the error function. Innite series include:
>>> nsum(lambda k: 1/fac2(k), [0, inf])
3.05940740534258
>>> sqrt(e)*(1+sqrt(pi/2)*erf(sqrt(2)/2))
3.05940740534258
>>> nsum(lambda k: 2**k/fac2(2*k-1), [1, inf])
4.06015693855741
>>> e * erf(1) * sqrt(pi)
4.06015693855741
A beautiful Ramanujan sum:
>>> nsum(lambda k: (-1)**k*(fac2(2*k-1)/fac2(2*k))**3, [0,inf])
0.90917279454693
>>> (gamma(9/8)/gamma(5/4)/gamma(7/8))**2
0.90917279454693
References
1.http://functions.wolfram.com/GammaBetaErf/Factorial2/27/01/0002/
2.http://mathworld.wolfram.com/DoubleFactorial.html
Binomial coecients
binomial()
mpmath.binomial(n, k)
Computes the binomial coecient
( )
n
n!
=
.
k
k!(n k)!
The binomial coecient gives the number of ways that k items can be chosen from a set
of n items. More generally, the binomial coecient is a well-dened function of arbitrary
real or complex n and k, via the gamma function.
Examples
5.15. Welcome to mpmaths documentation!
779
SymPy Documentation, Release 0.7.6
Generate Pascals triangle:
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = True
>>> for n in range(5):
...
nprint([binomial(n,k) for k in range(n+1)])
...
[1.0]
[1.0, 1.0]
[1.0, 2.0, 1.0]
[1.0, 3.0, 3.0, 1.0]
[1.0, 4.0, 6.0, 4.0, 1.0]
There is 1 way to select 0 items from the empty set, and 0 ways to select 1 item from the
empty set:
>>> binomial(0, 0)
1.0
>>> binomial(0, 1)
0.0
binomial() (page 779) supports large arguments:
>>> binomial(10**20, 10**20-5)
8.33333333333333e+97
>>> binomial(10**20, 10**10)
2.60784095465201e+104342944813
Nonintegral binomial coecients nd use in series expansions:
>>> nprint(taylor(lambda x: (1+x)**0.25, 0, 4))
[1.0, 0.25, -0.09375, 0.0546875, -0.0375977]
>>> nprint([binomial(0.25, k) for k in range(5)])
[1.0, 0.25, -0.09375, 0.0546875, -0.0375977]
An integral representation:
>>> n, k = 5, 3
>>> f = lambda t: exp(-j*k*t)*(1+exp(j*t))**n
>>> chop(quad(f, [-pi,pi])/(2*pi))
10.0
>>> binomial(n,k)
10.0
Gamma function
gamma()
mpmath.gamma(x, **kwargs)
Computes the gamma function, (x). The gamma function is a shifted version of the
ordinary factorial, satisfying (n) = (n 1)! for integers n > 0. More generally, it is
dened by
(x) =
tx1 et dt
0
for any real or complex x with <(x) > 0 and for <(x) < 0 by analytic continuation.
Examples
780
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Basic values and limits:
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = True
>>> for k in range(1, 6):
...
print(%s %s % (k, gamma(k)))
...
1 1.0
2 1.0
3 2.0
4 6.0
5 24.0
>>> gamma(inf)
+inf
>>> gamma(0)
Traceback (most recent call last):
...
ValueError: gamma function pole
Traceback (most recent call last):
...
ValueError: gamma function pole
The gamma function of a half-integer is a rational multiple of
>>> gamma(0.5), sqrt(pi)
(1.77245385090552, 1.77245385090552)
>>> gamma(1.5), sqrt(pi)/2
(0.886226925452758, 0.886226925452758)
We can check the integral denition:
>>> gamma(3.5)
3.32335097044784
>>> quad(lambda t: t**2.5*exp(-t), [0,inf])
3.32335097044784
gamma() (page 780) supports arbitrary-precision evaluation and complex arguments:
>>> mp.dps = 50
>>> gamma(sqrt(3))
0.91510229697308632046045539308226554038315280564184
>>> mp.dps = 25
>>> gamma(2j)
(0.009902440080927490985955066 - 0.07595200133501806872408048j)
Arguments can also be large. Note that the gamma function grows very quickly:
>>> mp.dps = 15
>>> gamma(10**20)
1.9328495143101e+1956570551809674817225
rgamma()
mpmath.rgamma(x, **kwargs)
Computes the reciprocal of the gamma function, 1/(z). This function evaluates to zero
at the poles of the gamma function, z = 0, 1, 2, . . ..
Examples
Basic examples:
5.15. Welcome to mpmaths documentation!
781
SymPy Documentation, Release 0.7.6
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> rgamma(1)
1.0
>>> rgamma(4)
0.1666666666666666666666667
>>> rgamma(0); rgamma(-1)
0.0
0.0
>>> rgamma(1000)
2.485168143266784862783596e-2565
>>> rgamma(inf)
0.0
A denite integral that can be evaluated in terms of elementary integrals:
>>> quad(rgamma, [0,inf])
2.807770242028519365221501
>>> e + quad(lambda t: exp(-t)/(pi**2+log(t)**2), [0,inf])
2.807770242028519365221501
gammaprod()
mpmath.gammaprod(a, b)
Given iterables a and b, gammaprod(a, b) computes the product / quotient of gamma
functions:
(a0 )(a1 ) (ap )
(b0 )(b1 ) (bq )
Unlike direct calls to gamma() (page 780), gammaprod() (page 782) considers the entire
product as a limit and evaluates this limit properly if any of the numerator or denominator arguments are nonpositive integers such that poles of the gamma function are
encountered. That is, gammaprod() (page 782) evaluates
lim
0
(a0 + )(a1 + ) (ap + )
(b0 + )(b1 + ) (bq + )
In particular:
If there are equally many poles in the numerator and the denominator, the limit is a
rational number times the remaining, regular part of the product.
If there are more poles in the numerator, gammaprod() (page 782) returns +inf.
If there are more poles in the denominator, gammaprod() (page 782) returns 0.
Examples
The reciprocal gamma function 1/(x) evaluated at x = 0:
>>> from sympy.mpmath import *
>>> mp.dps = 15
>>> gammaprod([], [0])
0.0
A limit:
>>> gammaprod([-4], [-3])
-0.25
>>> limit(lambda x: gamma(x-1)/gamma(x), -3, direction=1)
782
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
-0.25
>>> limit(lambda x: gamma(x-1)/gamma(x), -3, direction=-1)
-0.25
loggamma()
mpmath.loggamma(x)
Computes the principal branch of the log-gamma function, ln (z). Unlike ln((z)), which
has innitely many complex branch cuts, the principal log-gamma function only has a single branch cut along the negative half-axis. The principal branch continuously matches
the asymptotic Stirling expansion
(
)
ln(2)
1
ln (z)
+ z
ln(z) z + O(z 1 ).
2
2
The real parts of both functions agree, but their imaginary parts generally dier by 2n
for some n Z. They coincide for z R, z > 0.
Computationally, it is advantageous to use loggamma() (page 783) instead of gamma()
(page 780) for extremely large arguments.
Examples
Comparing with ln((z)):
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> loggamma(13.2); log(gamma(13.2))
20.49400419456603678498394
20.49400419456603678498394
>>> loggamma(3+4j)
(-1.756626784603784110530604 + 4.742664438034657928194889j)
>>> log(gamma(3+4j))
(-1.756626784603784110530604 - 1.540520869144928548730397j)
>>> log(gamma(3+4j)) + 2*pi*j
(-1.756626784603784110530604 + 4.742664438034657928194889j)
Note the imaginary parts for negative arguments:
>>> loggamma(-0.5); loggamma(-1.5); loggamma(-2.5)
(1.265512123484645396488946 - 3.141592653589793238462643j)
(0.8600470153764810145109327 - 6.283185307179586476925287j)
(-0.05624371649767405067259453 - 9.42477796076937971538793j)
Some special values:
>>> loggamma(1); loggamma(2)
0.0
0.0
>>> loggamma(3); +ln2
0.6931471805599453094172321
0.6931471805599453094172321
>>> loggamma(3.5); log(15*sqrt(pi)/8)
1.200973602347074224816022
1.200973602347074224816022
>>> loggamma(inf)
+inf
Huge arguments are permitted:
5.15. Welcome to mpmaths documentation!
783
SymPy Documentation, Release 0.7.6
>>> loggamma(1e30)
6.807755278982137052053974e+31
>>> loggamma(1e300)
6.897755278982137052053974e+302
>>> loggamma(1e3000)
6.906755278982137052053974e+3003
>>> loggamma(1e100000000000000000000)
2.302585092994045684007991e+100000000000000000020
>>> loggamma(1e30j)
(-1.570796326794896619231322e+30 + 6.807755278982137052053974e+31j)
>>> loggamma(1e300j)
(-1.570796326794896619231322e+300 + 6.897755278982137052053974e+302j)
>>> loggamma(1e3000j)
(-1.570796326794896619231322e+3000 + 6.906755278982137052053974e+3003j)
The log-gamma function can be integrated analytically on any interval of unit length:
>>> z = 0
>>> quad(loggamma, [z,z+1]); log(2*pi)/2
0.9189385332046727417803297
0.9189385332046727417803297
>>> z = 3+4j
>>> quad(loggamma, [z,z+1]); (log(z)-1)*z + log(2*pi)/2
(-0.9619286014994750641314421 + 5.219637303741238195688575j)
(-0.9619286014994750641314421 + 5.219637303741238195688575j)
The derivatives of the log-gamma function are given by the polygamma function (psi()
(page 792)):
>>> diff(loggamma, -4+3j); psi(0, -4+3j)
(1.688493531222971393607153 + 2.554898911356806978892748j)
(1.688493531222971393607153 + 2.554898911356806978892748j)
>>> diff(loggamma, -4+3j, 2); psi(1, -4+3j)
(-0.1539414829219882371561038 - 0.1020485197430267719746479j)
(-0.1539414829219882371561038 - 0.1020485197430267719746479j)
The log-gamma function satises an additive form of the recurrence relation for the
ordinary gamma function:
>>> z = 2+3j
>>> loggamma(z); loggamma(z+1) - log(z)
(-2.092851753092733349564189 + 2.302396543466867626153708j)
(-2.092851753092733349564189 + 2.302396543466867626153708j)
Rising and falling factorials
rf()
mpmath.rf(x, n)
Computes the rising factorial or Pochhammer symbol,
x(n) = x(x + 1) (x + n 1) =
(x + n)
(x)
where the rightmost expression is valid for nonintegral n.
Examples
For integral n, the rising factorial is a polynomial:
784
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = True
>>> for n in range(5):
...
nprint(taylor(lambda x: rf(x,n), 0, n))
...
[1.0]
[0.0, 1.0]
[0.0, 1.0, 1.0]
[0.0, 2.0, 3.0, 1.0]
[0.0, 6.0, 11.0, 6.0, 1.0]
Evaluation is supported for arbitrary arguments:
>>> rf(2+3j, 5.5)
(-7202.03920483347 - 3777.58810701527j)
ff()
mpmath.ff(x, n)
Computes the falling factorial,
(x)n = x(x 1) (x n + 1) =
(x + 1)
(x n + 1)
where the rightmost expression is valid for nonintegral n.
Examples
For integral n, the falling factorial is a polynomial:
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = True
>>> for n in range(5):
...
nprint(taylor(lambda x: ff(x,n), 0, n))
...
[1.0]
[0.0, 1.0]
[0.0, -1.0, 1.0]
[0.0, 2.0, -3.0, 1.0]
[0.0, -6.0, 11.0, -6.0, 1.0]
Evaluation is supported for arbitrary arguments:
>>> ff(2+3j, 5.5)
(-720.41085888203 + 316.101124983878j)
Beta function
beta()
mpmath.beta(x, y)
Computes the beta function, B(x, y) = (x)(y)/(x + y). The beta function is also commonly dened by the integral representation
1
B(x, y) =
tx1 (1 t)y1 dt
0
Examples
5.15. Welcome to mpmaths documentation!
785
SymPy Documentation, Release 0.7.6
For integer and half-integer arguments where all three gamma functions are nite, the
beta function becomes either rational number or a rational multiple of :
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = True
>>> beta(5, 2)
0.0333333333333333
>>> beta(1.5, 2)
0.266666666666667
>>> 16*beta(2.5, 1.5)
3.14159265358979
Where appropriate, beta() (page 785) evaluates limits. A pole of the beta function is
taken to result in +inf:
>>> beta(-0.5, 0.5)
0.0
>>> beta(-3, 3)
-0.333333333333333
>>> beta(-2, 3)
+inf
>>> beta(inf, 1)
0.0
>>> beta(inf, 0)
nan
beta() (page 785) supports complex numbers and arbitrary precision evaluation:
>>> beta(1, 2+j)
(0.4 - 0.2j)
>>> mp.dps = 25
>>> beta(j,0.5)
(1.079424249270925780135675 - 1.410032405664160838288752j)
>>> mp.dps = 50
>>> beta(pi, e)
0.037890298781212201348153837138927165984170287886464
Various integrals can be computed by means of the beta function:
>>> mp.dps = 15
>>> quad(lambda t: t**2.5*(1-t)**2, [0, 1])
0.0230880230880231
>>> beta(3.5, 3)
0.0230880230880231
>>> quad(lambda t: sin(t)**4 * sqrt(cos(t)), [0, pi/2])
0.319504062596158
>>> beta(2.5, 0.75)/2
0.319504062596158
betainc()
mpmath.betainc(a, b, x1=0, x2=1, regularized=False)
betainc(a, b, x1=0, x2=1, regularized=False) gives the generalized incomplete
beta function,
x2
x2
Ix1 (a, b) =
ta1 (1 t)b1 dt.
x1
When x1 = 0, x2 = 1, this reduces to the ordinary (complete) beta function B(a, b); see
beta() (page 785).
786
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
With the keyword argument regularized=True, betainc() (page 786) computes the
regularized incomplete beta function Ixx12 (a, b)/B(a, b). This is the cumulative distribution
of the beta distribution with parameters a, b.
Examples
Verifying that betainc() (page 786) computes the integral in the denition:
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> x,y,a,b = 3, 4, 0, 6
>>> betainc(x, y, a, b)
-4010.4
>>> quad(lambda t: t**(x-1) * (1-t)**(y-1), [a, b])
-4010.4
The arguments may be arbitrary complex numbers:
>>> betainc(0.75, 1-4j, 0, 2+3j)
(0.2241657956955709603655887 + 0.3619619242700451992411724j)
With regularization:
>>> betainc(1, 2, 0, 0.25, regularized=True)
0.4375
>>> betainc(pi, e, 0, 1, regularized=True)
# Complete
1.0
The beta integral satises some simple argument transformation symmetries:
>>> mp.dps = 15
>>> betainc(2,3,4,5), -betainc(2,3,5,4), betainc(3,2,1-5,1-4)
(56.0833333333333, 56.0833333333333, 56.0833333333333)
The beta integral can often be evaluated analytically. For integer and rational arguments, the incomplete beta function typically reduces to a simple algebraic-logarithmic
expression:
>>> mp.dps = 25
>>> identify(chop(betainc(0, 0, 3, 4)))
-(log((9/8)))
>>> identify(betainc(2, 3, 4, 5))
(673/12)
>>> identify(betainc(1.5, 1, 1, 2))
((-12+sqrt(1152))/18)
Super- and hyperfactorials
superfac()
mpmath.superfac(z)
Computes the superfactorial, dened as the product of consecutive factorials
sf(n) =
k!
k=1
For general complex z, sf(z) is dened in terms of the Barnes G-function (see barnesg()
(page 790)).
5.15. Welcome to mpmaths documentation!
787
SymPy Documentation, Release 0.7.6
Examples
The rst few superfactorials are (OEIS A000178):
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = True
>>> for n in range(10):
...
print(%s %s % (n, superfac(n)))
...
0 1.0
1 1.0
2 2.0
3 12.0
4 288.0
5 34560.0
6 24883200.0
7 125411328000.0
8 5.05658474496e+15
9 1.83493347225108e+21
Superfactorials grow very rapidly:
>>> superfac(1000)
3.24570818422368e+1177245
>>> superfac(10**10)
2.61398543581249e+467427913956904067453
Evaluation is supported for arbitrary arguments:
>>> mp.dps = 25
>>> superfac(pi)
17.20051550121297985285333
>>> superfac(2+3j)
(-0.005915485633199789627466468 + 0.008156449464604044948738263j)
>>> diff(superfac, 1)
0.2645072034016070205673056
References
1.http://oeis.org/A000178
hyperfac()
mpmath.hyperfac(z)
Computes the hyperfactorial, dened for integers as the product
H(n) =
kk .
k=1
The hyperfactorial satises the recurrence formula H(z) = z z H(z 1). It can be dened
more generally in terms of the Barnes G-function (see barnesg() (page 790)) and the
gamma function by the formula
H(z) =
(z + 1)z
.
G(z)
The extension to complex numbers can also be done via the integral representation
]
[(
) z
z+1
log(t!) dt .
H(z) = (2)z/2 exp
+
2
0
788
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
The rapidly-growing sequence of hyperfactorials begins (OEIS A002109):
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = True
>>> for n in range(10):
...
print(%s %s % (n, hyperfac(n)))
...
0 1.0
1 1.0
2 4.0
3 108.0
4 27648.0
5 86400000.0
6 4031078400000.0
7 3.3197663987712e+18
8 5.56964379417266e+25
9 2.15779412229419e+34
Some even larger hyperfactorials are:
>>> hyperfac(1000)
5.46458120882585e+1392926
>>> hyperfac(10**10)
4.60408207642219e+489142638002418704309
The hyperfactorial can be evaluated for arbitrary arguments:
>>> hyperfac(0.5)
0.880449235173423
>>> diff(hyperfac, 1)
0.581061466795327
>>> hyperfac(pi)
205.211134637462
>>> hyperfac(-10+1j)
(3.01144471378225e+46 - 2.45285242480185e+46j)
The recurrence property of the hyperfactorial holds generally:
>>> z = 3-4*j
>>> hyperfac(z)
(-4.49795891462086e-7 - 6.33262283196162e-7j)
>>> z**z * hyperfac(z-1)
(-4.49795891462086e-7 - 6.33262283196162e-7j)
>>> z = mpf(-0.6)
>>> chop(z**z * hyperfac(z-1))
1.28170142849352
>>> hyperfac(z)
1.28170142849352
The hyperfactorial may also be computed using the integral denition:
>>> z = 2.5
>>> hyperfac(z)
15.9842119922237
>>> (2*pi)**(-z/2)*exp(binomial(z+1,2) +
...
quad(lambda t: loggamma(t+1), [0, z]))
15.9842119922237
hyperfac() (page 788) supports arbitrary-precision evaluation:
5.15. Welcome to mpmaths documentation!
789
SymPy Documentation, Release 0.7.6
>>> mp.dps = 50
>>> hyperfac(10)
215779412229418562091680268288000000000000000.0
>>> hyperfac(1/sqrt(2))
0.89404818005227001975423476035729076375705084390942
References
1.http://oeis.org/A002109
2.http://mathworld.wolfram.com/Hyperfactorial.html
barnesg()
mpmath.barnesg(z)
Evaluates the Barnes G-function, which generalizes the superfactorial (superfac()
(page 787)) and by extension also the hyperfactorial (hyperfac() (page 788)) to the
complex numbers in an analogous way to how the gamma function generalizes the ordinary factorial.
The Barnes G-function may be dened in terms of a Weierstrass product:
z/2 [z(z+1)+z 2 ]/2
G(z + 1) = (2)
[(
z )n z+z2 /(2n) ]
e
1+
n
n=1
For positive integers n, we have have relation to superfactorials G(n) = sf(n 2) = 0!
1! (n 2)!.
Examples
Some elementary values and limits of the Barnes G-function:
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = True
>>> barnesg(1), barnesg(2), barnesg(3)
(1.0, 1.0, 1.0)
>>> barnesg(4)
2.0
>>> barnesg(5)
12.0
>>> barnesg(6)
288.0
>>> barnesg(7)
34560.0
>>> barnesg(8)
24883200.0
>>> barnesg(inf)
+inf
>>> barnesg(0), barnesg(-1), barnesg(-2)
(0.0, 0.0, 0.0)
Closed-form values are known for some rational arguments:
>>> barnesg(1/2)
0.603244281209446
>>> sqrt(exp(0.25+log(2)/12)/sqrt(pi)/glaisher**3)
0.603244281209446
>>> barnesg(1/4)
0.29375596533861
>>> nthroot(exp(3/8)/exp(catalan/pi)/
790
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
...
gamma(0.25)**3/sqrt(glaisher)**9, 4)
0.29375596533861
The Barnes G-function satises the functional equation G(z + 1) = (z)G(z):
>>> z = pi
>>> barnesg(z+1)
2.39292119327948
>>> gamma(z)*barnesg(z)
2.39292119327948
The asymptotic growth rate of the Barnes G-function is related to the Glaisher-Kinkelin
constant:
>>> limit(lambda n: barnesg(n+1)/(n**(n**2/2-mpf(1)/12)*
...
(2*pi)**(n/2)*exp(-3*n**2/4)), inf)
0.847536694177301
>>> exp(1/12)/glaisher
0.847536694177301
The Barnes G-function can be dierentiated in closed form:
>>> z = 3
>>> diff(barnesg, z)
0.264507203401607
>>> barnesg(z)*((z-1)*psi(0,z)-z+(log(2*pi)+1)/2)
0.264507203401607
Evaluation is supported for arbitrary arguments and at arbitrary precision:
>>> barnesg(6.5)
2548.7457695685
>>> barnesg(-pi)
0.00535976768353037
>>> barnesg(3+4j)
(-0.000676375932234244 - 4.42236140124728e-5j)
>>> mp.dps = 50
>>> barnesg(1/sqrt(2))
0.81305501090451340843586085064413533788206204124732
>>> q = barnesg(10j)
>>> q.real
0.000000000021852360840356557241543036724799812371995850552234
>>> q.imag
-0.00000000000070035335320062304849020654215545839053210041457588
>>> mp.dps = 15
>>> barnesg(100)
3.10361006263698e+6626
>>> barnesg(-101)
0.0
>>> barnesg(-10.5)
5.94463017605008e+25
>>> barnesg(-10000.5)
-6.14322868174828e+167480422
>>> barnesg(1000j)
(5.21133054865546e-1173597 + 4.27461836811016e-1173597j)
>>> barnesg(-1000+1000j)
(2.43114569750291e+1026623 + 2.24851410674842e+1026623j)
References
5.15. Welcome to mpmaths documentation!
791
SymPy Documentation, Release 0.7.6
1.Whittaker & Watson, A Course of Modern Analysis, Cambridge University Press, 4th
edition (1927), p.264
2.http://en.wikipedia.org/wiki/Barnes G-function
3.http://mathworld.wolfram.com/BarnesG-Function.html
Polygamma functions and harmonic numbers
psi()/digamma()
mpmath.psi(m, z)
Gives the polygamma function of order m of z, (m) (z). Special cases are known as the
digamma function ( (0) (z)), the trigamma function ( (1) (z)), etc. The polygamma functions are dened as the logarithmic derivatives of the gamma function:
(
(m)
(z) =
d
dz
)m+1
log (z)
In particular, (0) (z) = 0 (z)/(z). In the present implementation of psi() (page 792),
the order m must be a nonnegative integer, while the argument z may be an arbitrary
complex number (with exception for the polygamma functions poles at z = 0, 1, 2, . . .).
Examples
For various rational arguments, the polygamma function reduces to a combination of
standard mathematical constants:
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> psi(0, 1), -euler
(-0.5772156649015328606065121, -0.5772156649015328606065121)
>>> psi(1, 1/4), pi**2+8*catalan
(17.19732915450711073927132, 17.19732915450711073927132)
>>> psi(2, 1/2), -14*apery
(-16.82879664423431999559633, -16.82879664423431999559633)
The polygamma functions are derivatives of each other:
>>> diff(lambda x: psi(3, x), pi), psi(4, pi)
(-0.1105749312578862734526952, -0.1105749312578862734526952)
>>> quad(lambda x: psi(4, x), [2, 3]), psi(3,3)-psi(3,2)
(-0.375, -0.375)
The digamma function diverges logarithmically as z , while higher orders tend to
zero:
>>> psi(0,inf), psi(1,inf), psi(2,inf)
(+inf, 0.0, 0.0)
Evaluation for a complex argument:
>>> psi(2, -1-2j)
(0.03902435405364952654838445 + 0.1574325240413029954685366j)
Evaluation is supported for large orders m and/or large arguments z:
792
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> psi(3, 10**100)
2.0e-300
>>> psi(250, 10**30+10**20*j)
(-1.293142504363642687204865e-7010 + 3.232856260909107391513108e-7018j)
Application to innite series
Any innite series where the summand is a rational function of the index k can be evaluated in closed form in terms of polygamma functions of the roots and poles of the summand:
>>> a = sqrt(2)
>>> b = sqrt(3)
>>> nsum(lambda k: 1/((k+a)**2*(k+b)), [0, inf])
0.4049668927517857061917531
>>> (psi(0,a)-psi(0,b)-a*psi(1,a)+b*psi(1,a))/(a-b)**2
0.4049668927517857061917531
This follows from the series representation (m > 0)
(m) (z) = (1)m+1 m!
k=0
1
.
(z + k)m+1
Since the roots of a polynomial may be complex, it is sometimes necessary to use the
complex polygamma function to evaluate an entirely real-valued sum:
>>> nsum(lambda k: 1/(k**2-2*k+3), [0, inf])
1.694361433907061256154665
>>> nprint(polyroots([1,-2,3]))
[(1.0 - 1.41421j), (1.0 + 1.41421j)]
>>> r1 = 1-sqrt(2)*j
>>> r2 = r1.conjugate()
>>> (psi(0,-r2)-psi(0,-r1))/(r1-r2)
(1.694361433907061256154665 + 0.0j)
mpmath.digamma(z)
Shortcut for psi(0,z).
harmonic()
mpmath.harmonic(z)
If n is an integer, harmonic(n) gives a oating-point approximation of the n-th harmonic
number H(n), dened as
H(n) = 1 +
1 1
1
+ + ... +
2 3
n
The rst few harmonic numbers are:
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = True
>>> for n in range(8):
...
print(%s %s % (n, harmonic(n)))
...
0 0.0
1 1.0
2 1.5
3 1.83333333333333
4 2.08333333333333
5.15. Welcome to mpmaths documentation!
793
SymPy Documentation, Release 0.7.6
5 2.28333333333333
6 2.45
7 2.59285714285714
The innite harmonic series 1 + 1/2 + 1/3 + . . . diverges:
>>> harmonic(inf)
+inf
harmonic() (page 793) is evaluated using the digamma function rather than by summing
the harmonic series term by term. It can therefore be computed quickly for arbitrarily
large n, and even for nonintegral arguments:
>>> harmonic(10**100)
230.835724964306
>>> harmonic(0.5)
0.613705638880109
>>> harmonic(3+4j)
(2.24757548223494 + 0.850502209186044j)
harmonic() (page 793) supports arbitrary precision evaluation:
>>> mp.dps = 50
>>> harmonic(11)
3.0198773448773448773448773448773448773448773448773
>>> harmonic(pi)
1.8727388590273302654363491032336134987519132374152
The harmonic series diverges, but at a glacial pace. It is possible to calculate the exact
number of terms required before the sum exceeds a given amount, say 100:
>>> mp.dps = 50
>>> v = 10**findroot(lambda x: harmonic(10**x) - 100, 10)
>>> v
15092688622113788323693563264538101449859496.864101
>>> v = int(ceil(v))
>>> print(v)
15092688622113788323693563264538101449859497
>>> harmonic(v-1)
99.999999999999999999999999999999999999999999942747
>>> harmonic(v)
100.000000000000000000000000000000000000000000009
Exponential integrals and error functions
Exponential integrals give closed-form solutions to a large class of commonly occurring transcendental integrals that cannot be evaluated using elementary functions. Integrals of this
2
type include those with an integrand of the form ta et or ex , the latter giving rise to the
Gaussian (or normal) probability distribution.
The most general function in this section is the incomplete gamma function, to which all
others can be reduced. The incomplete gamma function, in turn, can be expressed using
hypergeometric functions (see Hypergeometric functions (page 888)).
Incomplete gamma functions
794
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
gammainc()
mpmath.gammainc(z, a=0, b=inf, regularized=False)
gammainc(z, a=0, b=inf) computes the (generalized) incomplete gamma function with
integration limits [a, b]:
(z, a, b) =
tz1 et dt
The generalized incomplete gamma function reduces to the following special cases when
one or both endpoints are xed:
(z, 0, ) is the standard (complete) gamma function, (z) (available directly as the
mpmath function gamma() (page 780))
(z, a, ) is the upper incomplete gamma function, (z, a)
(z, 0, b) is the lower incomplete gamma function, (z, b).
Of course, we have (z, 0, x) + (z, x, ) = (z) for all z and x.
Note however that some authors reverse the order of the arguments when dening the
lower and upper incomplete gamma function, so one should be careful to get the correct
denition.
If also given the keyword argument regularized=True, gammainc() (page 795) computes the regularized incomplete gamma function
P (z, a, b) =
(z, a, b)
.
(z)
Examples
We can compare with numerical quadrature to verify that gammainc() (page 795) computes the integral in the denition:
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> gammainc(2+3j, 4, 10)
(0.00977212668627705160602312 - 0.0770637306312989892451977j)
>>> quad(lambda t: t**(2+3j-1) * exp(-t), [4, 10])
(0.00977212668627705160602312 - 0.0770637306312989892451977j)
Argument symmetries follow directly from the integral denition:
>>> gammainc(3, 4, 5) + gammainc(3, 5, 4)
0.0
>>> gammainc(3,0,2) + gammainc(3,2,4); gammainc(3,0,4)
1.523793388892911312363331
1.523793388892911312363331
>>> findroot(lambda z: gammainc(2,z,3), 1)
3.0
Evaluation for arbitrarily large arguments:
>>> gammainc(10, 100)
4.083660630910611272288592e-26
>>> gammainc(10, 10000000000000000)
5.290402449901174752972486e-4342944819032375
>>> gammainc(3+4j, 1000000+1000000j)
(-1.257913707524362408877881e-434284 + 2.556691003883483531962095e-434284j)
5.15. Welcome to mpmaths documentation!
795
SymPy Documentation, Release 0.7.6
Evaluation of a generalized incomplete gamma function automatically chooses the representation that gives a more accurate result, depending on which parameter is larger:
>>> gammainc(10000000, 3) - gammainc(10000000, 2)
# Bad
0.0
>>> gammainc(10000000, 2, 3)
# Good
1.755146243738946045873491e+4771204
>>> gammainc(2, 0, 100000001) - gammainc(2, 0, 100000000)
0.0
>>> gammainc(2, 100000000, 100000001)
# Good
4.078258353474186729184421e-43429441
# Bad
The incomplete gamma functions satisfy simple recurrence relations:
>>> mp.dps = 25
>>> z, a = mpf(3.5), mpf(2)
>>> gammainc(z+1, a); z*gammainc(z,a) + a**z*exp(-a)
10.60130296933533459267329
10.60130296933533459267329
>>> gammainc(z+1,0,a); z*gammainc(z,0,a) - a**z*exp(-a)
1.030425427232114336470932
1.030425427232114336470932
Evaluation at integers and poles:
>>> gammainc(-3, -4, -5)
(-0.2214577048967798566234192 + 0.0j)
>>> gammainc(-3, 0, 5)
+inf
If z is an integer, the recurrence reduces the incomplete gamma function to P (a) exp(a)+
Q(b) exp(b) where P and Q are polynomials:
>>> gammainc(1, 2); exp(-2)
0.1353352832366126918939995
0.1353352832366126918939995
>>> mp.dps = 50
>>> identify(gammainc(6, 1, 2), [exp(-1), exp(-2)])
(326*exp(-1) + (-872)*exp(-2))
The incomplete gamma functions reduce to functions such as the exponential integral Ei
and the error function for special arguments:
>>> mp.dps = 25
>>> gammainc(0, 4); -ei(-4)
0.00377935240984890647887486
0.00377935240984890647887486
>>> gammainc(0.5, 0, 2); sqrt(pi)*erf(sqrt(2))
1.691806732945198336509541
1.691806732945198336509541
Exponential integrals
ei()
mpmath.ei(x, **kwargs)
Computes the exponential integral or Ei-function, Ei(x). The exponential integral is de-
796
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
ned as
Ei(x) =
et
dt.
t
When the integration range includes t = 0, the exponential integral is interpreted as
providing the Cauchy principal value.
For real x, the Ei-function behaves roughly like Ei(x) exp(x) + log(|x|).
The Ei-function is related to the more general family of exponential integral functions
denoted by En , which are available as expint() (page 798).
Basic examples
Some basic values and limits are:
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = True
>>> ei(0)
-inf
>>> ei(1)
1.89511781635594
>>> ei(inf)
+inf
>>> ei(-inf)
0.0
For x < 0, the dening integral can be evaluated numerically as a reference:
>>> ei(-4)
-0.00377935240984891
>>> quad(lambda t: exp(t)/t, [-inf, -4])
-0.00377935240984891
ei() (page 796) supports complex arguments and arbitrary precision evaluation:
>>> mp.dps = 50
>>> ei(pi)
10.928374389331410348638445906907535171566338835056
>>> mp.dps = 25
>>> ei(3+4j)
(-4.154091651642689822535359 + 4.294418620024357476985535j)
Related functions
The exponential integral is closely related to the logarithmic integral.
(page 799) for additional information.
See li()
The exponential integral is related to the hyperbolic and trigonometric integrals (see
chi() (page 803), shi() (page 803), ci() (page 801), si() (page 802)) similarly to
how the ordinary exponential function is related to the hyperbolic and trigonometric
functions:
>>> mp.dps = 15
>>> ei(3)
9.93383257062542
>>> chi(3) + shi(3)
9.93383257062542
>>> chop(ci(3j) - j*si(3j) - pi*j/2)
9.93383257062542
5.15. Welcome to mpmaths documentation!
797
SymPy Documentation, Release 0.7.6
Beware that logarithmic corrections, as in the last example above, are required to obtain
the correct branch in general. For details, see [1].
The exponential integral is also a special case of the hypergeometric function 2 F2 :
>>> z = 0.6
>>> z*hyper([1,1],[2,2],z) + (ln(z)-ln(1/z))/2 + euler
0.769881289937359
>>> ei(z)
0.769881289937359
References
1.Relations between Ei and other functions: http://functions.wolfram.com/GammaBetaErf/ExpIntegra
2.Abramowitz & Stegun, section 5: http://people.math.sfu.ca/cbm/aands/page 228.htm
3.Asymptotic expansion for Ei: http://mathworld.wolfram.com/En-Function.html
e1()
mpmath.e1(x, **kwargs)
Computes the exponential integral E1 (z), given by
t
e
E1 (z) =
dt.
t
z
This is equivalent to expint() (page 798) with n = 1.
Examples
Two ways to evaluate this function:
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> e1(6.25)
0.0002704758872637179088496194
>>> expint(1,6.25)
0.0002704758872637179088496194
The E1-function is essentially the same as the Ei-function (ei() (page 796)) with negated
argument, except for an imaginary branch cut term:
>>> e1(2.5)
0.02491491787026973549562801
>>> -ei(-2.5)
0.02491491787026973549562801
>>> e1(-2.5)
(-7.073765894578600711923552 - 3.141592653589793238462643j)
>>> -ei(2.5)
-7.073765894578600711923552
expint()
mpmath.expint(*args)
expint(n,z)() gives the generalized exponential integral or En-function,
zt
e
En (z) =
dt,
tn
1
where n and z may both be complex numbers. The case with n = 1 is also given by e1()
(page 798).
798
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
Evaluation at real and complex arguments:
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> expint(1, 6.25)
0.0002704758872637179088496194
>>> expint(-3, 2+3j)
(0.00299658467335472929656159 + 0.06100816202125885450319632j)
>>> expint(2+3j, 4-5j)
(0.001803529474663565056945248 - 0.002235061547756185403349091j)
At negative integer values of n, En (z) reduces to a rational-exponential function:
>>> f = lambda n, z: fac(n)*sum(z**k/fac(k-1) for k in range(1,n+2))/\
...
exp(z)/z**(n+2)
>>> n = 3
>>> z = 1/pi
>>> expint(-n,z)
584.2604820613019908668219
>>> f(n,z)
584.2604820613019908668219
>>> n = 5
>>> expint(-n,z)
115366.5762594725451811138
>>> f(n,z)
115366.5762594725451811138
Logarithmic integral
li()
mpmath.li(x, **kwargs)
Computes the logarithmic integral or li-function li(x), dened by
x
1
li(x) =
dt
0 log t
The logarithmic integral has a singularity at x = 1.
Alternatively, li(x, offset=True) computes the oset logarithmic integral (used in
number theory)
x
1
Li(x) =
dt.
2 log t
These two functions are related via the simple identity Li(x) = li(x) li(2).
The logarithmic integral should also not be confused with the polylogarithm (also denoted by Li), which is implemented as polylog() (page 944).
Examples
Some basic values and limits:
>>> from sympy.mpmath import *
>>> mp.dps = 30; mp.pretty = True
>>> li(0)
0.0
5.15. Welcome to mpmaths documentation!
799
SymPy Documentation, Release 0.7.6
>>> li(1)
-inf
>>> li(1)
-inf
>>> li(2)
1.04516378011749278484458888919
>>> findroot(li, 2)
1.45136923488338105028396848589
>>> li(inf)
+inf
>>> li(2, offset=True)
0.0
>>> li(1, offset=True)
-inf
>>> li(0, offset=True)
-1.04516378011749278484458888919
>>> li(10, offset=True)
5.12043572466980515267839286347
The logarithmic integral can be evaluated for arbitrary complex arguments:
>>> mp.dps = 20
>>> li(3+4j)
(3.1343755504645775265 + 2.6769247817778742392j)
The logarithmic integral is related to the exponential integral:
>>> ei(log(3))
2.1635885946671919729
>>> li(3)
2.1635885946671919729
The logarithmic integral grows like O(x/ log(x)):
>>> mp.dps = 15
>>> x = 10**100
>>> x/log(x)
4.34294481903252e+97
>>> li(x)
4.3619719871407e+97
The prime number theorem states that the number of primes less than x is asymptotic to Li(x) (equivalently li(x)). For example, it is known that there are exactly
1,925,320,391,606,803,968,923 prime numbers less than 1023 [1]. The logarithmic integral provides a very accurate estimate:
>>> li(10**23, offset=True)
1.92532039161405e+21
A denite integral is:
>>> quad(li, [0, 1])
-0.693147180559945
>>> -ln(2)
-0.693147180559945
References
1.http://mathworld.wolfram.com/PrimeCountingFunction.html
2.http://mathworld.wolfram.com/LogarithmicIntegral.html
800
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Trigonometric integrals
ci()
mpmath.ci(x, **kwargs)
Computes the cosine integral,
Ci(x) =
cos t
dt = + log x +
t
x
0
cos t 1
dt
t
Examples
Some values and limits:
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> ci(0)
-inf
>>> ci(1)
0.3374039229009681346626462
>>> ci(pi)
0.07366791204642548599010096
>>> ci(inf)
0.0
>>> ci(-inf)
(0.0 + 3.141592653589793238462643j)
>>> ci(2+3j)
(1.408292501520849518759125 - 2.983617742029605093121118j)
The cosine integral behaves roughly like the sinc function (see sinc() (page 773)) for
large real x:
>>> ci(10**10)
-4.875060251748226537857298e-11
>>> sinc(10**10)
-4.875060250875106915277943e-11
>>> chop(limit(ci, inf))
0.0
It has innitely many roots on the positive real axis:
>>> findroot(ci, 1)
0.6165054856207162337971104
>>> findroot(ci, 2)
3.384180422551186426397851
Evaluation is supported for z anywhere in the complex plane:
>>> ci(10**6*(1+j))
(4.449410587611035724984376e+434287 + 9.75744874290013526417059e+434287j)
We can evaluate the dening integral as a reference:
>>> mp.dps = 15
>>> -quadosc(lambda t: cos(t)/t, [5, inf], omega=1)
-0.190029749656644
>>> ci(5)
-0.190029749656644
Some innite series can be evaluated using the cosine integral:
5.15. Welcome to mpmaths documentation!
801
SymPy Documentation, Release 0.7.6
>>> nsum(lambda k: (-1)**k/(fac(2*k)*(2*k)), [1,inf])
-0.239811742000565
>>> ci(1) - euler
-0.239811742000565
si()
mpmath.si(x, **kwargs)
Computes the sine integral,
Si(x) =
0
sin t
dt.
t
The sine integral is thus the antiderivative of the sinc function (see sinc() (page 773)).
Examples
Some values and limits:
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> si(0)
0.0
>>> si(1)
0.9460830703671830149413533
>>> si(-1)
-0.9460830703671830149413533
>>> si(pi)
1.851937051982466170361053
>>> si(inf)
1.570796326794896619231322
>>> si(-inf)
-1.570796326794896619231322
>>> si(2+3j)
(4.547513889562289219853204 + 1.399196580646054789459839j)
The sine integral approaches /2 for large real x:
>>> si(10**10)
1.570796326707584656968511
>>> pi/2
1.570796326794896619231322
Evaluation is supported for z anywhere in the complex plane:
>>> si(10**6*(1+j))
(-9.75744874290013526417059e+434287 + 4.449410587611035724984376e+434287j)
We can evaluate the dening integral as a reference:
>>> mp.dps = 15
>>> quad(sinc, [0, 5])
1.54993124494467
>>> si(5)
1.54993124494467
Some innite series can be evaluated using the sine integral:
>>> nsum(lambda k: (-1)**k/(fac(2*k+1)*(2*k+1)), [0,inf])
0.946083070367183
802
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> si(1)
0.946083070367183
Hyperbolic integrals
chi()
mpmath.chi(x, **kwargs)
Computes the hyperbolic cosine integral, dened in analogy with the cosine integral (see
ci() (page 801)) as
x
cosh t
cosh t 1
dt = + log x +
dt
Chi(x) =
t
t
x
0
Some values and limits:
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> chi(0)
-inf
>>> chi(1)
0.8378669409802082408946786
>>> chi(inf)
+inf
>>> findroot(chi, 0.5)
0.5238225713898644064509583
>>> chi(2+3j)
(-0.1683628683277204662429321 + 2.625115880451325002151688j)
Evaluation is supported for z anywhere in the complex plane:
>>> chi(10**6*(1+j))
(4.449410587611035724984376e+434287 - 9.75744874290013526417059e+434287j)
shi()
mpmath.shi(x, **kwargs)
Computes the hyperbolic sine integral, dened in analogy with the sine integral (see
si() (page 802)) as
x
sinh t
Shi(x) =
dt.
t
0
Some values and limits:
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> shi(0)
0.0
>>> shi(1)
1.057250875375728514571842
>>> shi(-1)
-1.057250875375728514571842
>>> shi(inf)
+inf
>>> shi(2+3j)
(-0.1931890762719198291678095 + 2.645432555362369624818525j)
5.15. Welcome to mpmaths documentation!
803
SymPy Documentation, Release 0.7.6
Evaluation is supported for z anywhere in the complex plane:
>>> shi(10**6*(1+j))
(4.449410587611035724984376e+434287 - 9.75744874290013526417059e+434287j)
Error functions
erf()
mpmath.erf(x, **kwargs)
Computes the error function, erf(x). The error function is the normalized antiderivative
of the Gaussian function exp(t2 ). More precisely,
x
2
erf(x) =
exp(t2 ) dt
0
Basic examples
Simple values and limits include:
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = True
>>> erf(0)
0.0
>>> erf(1)
0.842700792949715
>>> erf(-1)
-0.842700792949715
>>> erf(inf)
1.0
>>> erf(-inf)
-1.0
For large real x, erf(x) approaches 1 very rapidly:
>>> erf(3)
0.999977909503001
>>> erf(5)
0.999999999998463
The error function is an odd function:
>>> nprint(chop(taylor(erf, 0, 5)))
[0.0, 1.12838, 0.0, -0.376126, 0.0, 0.112838]
erf() (page 804) implements arbitrary-precision evaluation and supports complex numbers:
>>> mp.dps = 50
>>> erf(0.5)
0.52049987781304653768274665389196452873645157575796
>>> mp.dps = 25
>>> erf(1+j)
(1.316151281697947644880271 + 0.1904534692378346862841089j)
Evaluation is supported for large arguments:
804
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> mp.dps = 25
>>> erf(1e1000)
1.0
>>> erf(-1e1000)
-1.0
>>> erf(1e-1000)
1.128379167095512573896159e-1000
>>> erf(1e7j)
(0.0 + 8.593897639029319267398803e+43429448190317j)
>>> erf(1e7+1e7j)
(0.9999999858172446172631323 + 3.728805278735270407053139e-8j)
Related functions
See also erfc() (page 805), which is more accurate for large x, and erfi() (page 805)
which gives the antiderivative of exp(t2 ).
The Fresnel integrals fresnels() (page 808) and fresnelc() (page 809) are also related
to the error function.
erfc()
mpmath.erfc(x, **kwargs)
Computes the complementary error function, erfc(x) = 1 erf(x). This function avoids
cancellation that occurs when naively computing the complementary error function as
1-erf(x):
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = True
>>> 1 - erf(10)
0.0
>>> erfc(10)
2.08848758376254e-45
erfc() (page 805) works accurately even for ludicrously large arguments:
>>> erfc(10**10)
4.3504398860243e-43429448190325182776
Complex arguments are supported:
>>> erfc(500+50j)
(1.19739830969552e-107492 + 1.46072418957528e-107491j)
erfi()
mpmath.erfi(x)
Computes the imaginary error function, erfi(x). The imaginary error function is dened
in analogy with the error function, but with a positive sign in the integrand:
x
2
erfi(x) =
exp(t2 ) dt
0
Whereas the error function rapidly converges to 1 as x grows, the imaginary error function rapidly diverges to innity. The functions are related as erfi(x) = i erf(ix) for all
complex numbers x.
Examples
Basic values and limits:
5.15. Welcome to mpmaths documentation!
805
SymPy Documentation, Release 0.7.6
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = True
>>> erfi(0)
0.0
>>> erfi(1)
1.65042575879754
>>> erfi(-1)
-1.65042575879754
>>> erfi(inf)
+inf
>>> erfi(-inf)
-inf
Note the symmetry between erf and er:
>>> erfi(3j)
(0.0 + 0.999977909503001j)
>>> erf(3)
0.999977909503001
>>> erf(1+2j)
(-0.536643565778565 - 5.04914370344703j)
>>> erfi(2+1j)
(-5.04914370344703 - 0.536643565778565j)
Large arguments are supported:
>>> erfi(1000)
1.71130938718796e+434291
>>> erfi(10**10)
7.3167287567024e+43429448190325182754
>>> erfi(-10**10)
-7.3167287567024e+43429448190325182754
>>> erfi(1000-500j)
(2.49895233563961e+325717 + 2.6846779342253e+325717j)
>>> erfi(100000j)
(0.0 + 1.0j)
>>> erfi(-100000j)
(0.0 - 1.0j)
erfinv()
mpmath.erfinv(x)
Computes the inverse error function, satisfying
erf(erfinv(x)) = erfinv(erf(x)) = x.
This function is dened only for 1 x 1.
Examples
Special values include:
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = True
>>> erfinv(0)
0.0
>>> erfinv(1)
+inf
>>> erfinv(-1)
-inf
806
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
The domain is limited to the standard interval:
>>> erfinv(2)
Traceback (most recent call last):
...
ValueError: erfinv(x) is defined only for -1 <= x <= 1
Traceback (most recent call last):
...
ValueError: erfinv(x) is defined only for -1 <= x <= 1
It is simple to check that erfinv() (page 806) computes inverse values of erf()
(page 804) as promised:
>>> erf(erfinv(0.75))
0.75
>>> erf(erfinv(-0.995))
-0.995
erfinv() (page 806) supports arbitrary-precision evaluation:
>>> mp.dps = 50
>>> x = erf(2)
>>> x
0.99532226501895273416206925636725292861089179704006
>>> erfinv(x)
2.0
A denite integral involving the inverse error function:
>>> mp.dps = 15
>>> quad(erfinv, [0, 1])
0.564189583547756
>>> 1/sqrt(pi)
0.564189583547756
The inverse error function can be used to generate random numbers with a Gaussian
distribution (although this is a relatively inecient algorithm):
>>> nprint([erfinv(2*rand()-1) for n in range(6)])
[-0.586747, 1.10233, -0.376796, 0.926037, -0.708142, -0.732012]
The normal distribution
npdf()
mpmath.npdf(x, mu=0, sigma=1)
npdf(x, mu=0, sigma=1) evaluates the probability density function of a normal distribution with mean value and variance 2 .
Elementary properties of the probability distribution can be veried using numerical
integration:
>>>
>>>
>>>
1.0
>>>
0.5
from sympy.mpmath import *
mp.dps = 15; mp.pretty = True
quad(npdf, [-inf, inf])
quad(lambda x: npdf(x, 3), [3, inf])
5.15. Welcome to mpmaths documentation!
807
SymPy Documentation, Release 0.7.6
>>> quad(lambda x: npdf(x, 3, 2), [3, inf])
0.5
See also ncdf() (page 808), which gives the cumulative distribution.
ncdf()
mpmath.ncdf(x, mu=0, sigma=1)
ncdf(x, mu=0, sigma=1) evaluates the cumulative distribution function of a normal
distribution with mean value and variance 2 .
See also npdf() (page 807), which gives the probability density.
Elementary properties include:
>>>
>>>
>>>
0.5
>>>
0.0
>>>
1.0
from sympy.mpmath import *
mp.dps = 15; mp.pretty = True
ncdf(pi, mu=pi)
ncdf(-inf)
ncdf(+inf)
The cumulative distribution is the integral of the density function having identical mu
and sigma:
>>> mp.dps = 15
>>> diff(ncdf, 2)
0.053990966513188
>>> npdf(2)
0.053990966513188
>>> diff(lambda x: ncdf(x, 1, 0.5), 0)
0.107981933026376
>>> npdf(0, 1, 0.5)
0.107981933026376
Fresnel integrals
fresnels()
mpmath.fresnels(x)
Computes the Fresnel sine integral
sin
S(x) =
0
t2
2
)
dt
Note that some sources dene this function without the normalization factor /2.
Examples
Some basic values and limits:
>>>
>>>
>>>
0.0
>>>
0.5
808
from sympy.mpmath import *
mp.dps = 25; mp.pretty = True
fresnels(0)
fresnels(inf)
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> fresnels(-inf)
-0.5
>>> fresnels(1)
0.4382591473903547660767567
>>> fresnels(1+2j)
(36.72546488399143842838788 + 15.58775110440458732748279j)
Comparing with the denition:
>>> fresnels(3)
0.4963129989673750360976123
>>> quad(lambda t: sin(pi*t**2/2), [0,3])
0.4963129989673750360976123
fresnelc()
mpmath.fresnelc(x)
Computes the Fresnel cosine integral
cos
C(x) =
0
t2
2
)
dt
Note that some sources dene this function without the normalization factor /2.
Examples
Some basic values and limits:
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> fresnelc(0)
0.0
>>> fresnelc(inf)
0.5
>>> fresnelc(-inf)
-0.5
>>> fresnelc(1)
0.7798934003768228294742064
>>> fresnelc(1+2j)
(16.08787137412548041729489 - 36.22568799288165021578758j)
Comparing with the denition:
>>> fresnelc(3)
0.6057207892976856295561611
>>> quad(lambda t: cos(pi*t**2/2), [0,3])
0.6057207892976856295561611
Bessel functions and related functions
The functions in this section arise as solutions to various dierential equations in physics, typically describing wavelike oscillatory behavior or a combination of oscillation and exponential
decay or growth. Mathematically, they are special cases of the conuent hypergeometric
functions 0 F1 , 1 F1 and 1 F2 (see Hypergeometric functions (page 888)).
Bessel functions
5.15. Welcome to mpmaths documentation!
809
SymPy Documentation, Release 0.7.6
besselj()
mpmath.besselj(n, x, derivative=0)
besselj(n, x, derivative=0) gives the Bessel function of the rst kind Jn (x). Bessel
functions of the rst kind are dened as solutions of the dierential equation
x2 y 00 + xy 0 + (x2 n2 )y = 0
which appears, among other things, when solving the radial part of Laplaces equation
in cylindrical coordinates. This equation has two solutions for given n, where the Jn function is the solution that is nonsingular at x = 0. For positive integer n, Jn (x) behaves
roughly like a sine (odd n) or cosine (even n) multiplied by a magnitude factor that decays
slowly as x .
Generally, Jn is a special case of the hypergeometric function 0 F1 :
(
)
xn
x2
Jn (x) = n
0 F1 n + 1,
2 (n + 1)
4
With derivative = m 6= 0, the m-th derivative
dm
Jn (x)
dxm
is computed.
Plots
# Bessel function J_n(x) on the real line for n=0,1,2,3
j0 = lambda x: besselj(0,x)
j1 = lambda x: besselj(1,x)
j2 = lambda x: besselj(2,x)
j3 = lambda x: besselj(3,x)
plot([j0,j1,j2,j3],[0,14])
810
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
# Bessel function J_n(z) in the complex plane
cplot(lambda z: besselj(1,z), [-8,8], [-8,8], points=50000)
5.15. Welcome to mpmaths documentation!
811
SymPy Documentation, Release 0.7.6
Examples
Evaluation is supported for arbitrary arguments, and at arbitrary precision:
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = True
>>> besselj(2, 1000)
-0.024777229528606
>>> besselj(4, 0.75)
0.000801070086542314
>>> besselj(2, 1000j)
(-2.48071721019185e+432 + 6.41567059811949e-437j)
>>> mp.dps = 25
>>> besselj(0.75j, 3+4j)
(-2.778118364828153309919653 - 1.5863603889018621585533j)
>>> mp.dps = 50
>>> besselj(1, pi)
0.28461534317975275734531059968613140570981118184947
Arguments may be large:
>>> mp.dps = 25
>>> besselj(0, 10000)
-0.007096160353388801477265164
>>> besselj(0, 10**10)
0.000002175591750246891726859055
>>> besselj(2, 10**100)
7.337048736538615712436929e-51
>>> besselj(2, 10**5*j)
812
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
(-3.540725411970948860173735e+43426 + 4.4949812409615803110051e-43433j)
The Bessel functions of the rst kind satisfy simple symmetries around x = 0:
>>> mp.dps = 15
>>> nprint([besselj(n,0) for n in range(5)])
[1.0, 0.0, 0.0, 0.0, 0.0]
>>> nprint([besselj(n,pi) for n in range(5)])
[-0.304242, 0.284615, 0.485434, 0.333458, 0.151425]
>>> nprint([besselj(n,-pi) for n in range(5)])
[-0.304242, -0.284615, 0.485434, -0.333458, 0.151425]
Roots of Bessel functions are often used:
>>> nprint([findroot(j0, k)
[2.40483, 5.52008, 8.65373,
>>> nprint([findroot(j1, k)
[3.83171, 7.01559, 10.1735,
for k in
11.7915,
for k in
13.3237,
[2, 5, 8, 11, 14]])
14.9309]
[3, 7, 10, 13, 16]])
16.4706]
The roots are not periodic, but the distance between successive roots asymptotically
approaches 2. Bessel functions of the rst kind have the following normalization:
>>> quadosc(j0, [0, inf], period=2*pi)
1.0
>>> quadosc(j1, [0, inf], period=2*pi)
1.0
For n = 1/2 or n = 1/2, the Bessel function reduces to a trigonometric function:
>>> x = 10
>>> besselj(0.5, x), sqrt(2/(pi*x))*sin(x)
(-0.13726373575505, -0.13726373575505)
>>> besselj(-0.5, x), sqrt(2/(pi*x))*cos(x)
(-0.211708866331398, -0.211708866331398)
Derivatives of any order can be computed (negative orders correspond to integration):
>>> mp.dps = 25
>>> besselj(0, 7.5, 1)
-0.1352484275797055051822405
>>> diff(lambda x: besselj(0,x), 7.5)
-0.1352484275797055051822405
>>> besselj(0, 7.5, 10)
-0.1377811164763244890135677
>>> diff(lambda x: besselj(0,x), 7.5, 10)
-0.1377811164763244890135677
>>> besselj(0,7.5,-1) - besselj(0,3.5,-1)
-0.1241343240399987693521378
>>> quad(j0, [3.5, 7.5])
-0.1241343240399987693521378
Dierentiation with a noninteger order gives the fractional derivative in the sense of the
Riemann-Liouville dierintegral, as computed by differint() (page 1002):
>>> mp.dps = 15
>>> besselj(1, 3.5, 0.75)
-0.385977722939384
>>> differint(lambda x: besselj(1, x), 3.5, 0.75)
-0.385977722939384
5.15. Welcome to mpmaths documentation!
813
SymPy Documentation, Release 0.7.6
mpmath.j0(x)
Computes the Bessel function J0 (x). See besselj() (page 810).
mpmath.j1(x)
Computes the Bessel function J1 (x). See besselj() (page 810).
bessely()
mpmath.bessely(n, x, derivative=0)
bessely(n, x, derivative=0) gives the Bessel function of the second kind,
Yn (x) =
Jn (x) cos(n) Jn (x)
.
sin(n)
For n an integer, this formula should be understood as a limit. With derivative = m 6= 0,
the m-th derivative
dm
Yn (x)
dxm
is computed.
Plots
# Bessel function of 2nd kind Y_n(x) on the real line for n=0,1,2,3
y0 = lambda x: bessely(0,x)
y1 = lambda x: bessely(1,x)
y2 = lambda x: bessely(2,x)
y3 = lambda x: bessely(3,x)
plot([y0,y1,y2,y3],[0,10],[-4,1])
814
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
# Bessel function of 2nd kind Y_n(z) in the complex plane
cplot(lambda z: bessely(1,z), [-8,8], [-8,8], points=50000)
Examples
Some values of Yn (x):
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> bessely(0,0), bessely(1,0), bessely(2,0)
(-inf, -inf, -inf)
>>> bessely(1, pi)
0.3588729167767189594679827
>>> bessely(0.5, 3+4j)
(9.242861436961450520325216 - 3.085042824915332562522402j)
Arguments may be large:
>>> bessely(0, 10000)
0.00364780555898660588668872
>>> bessely(2.5, 10**50)
-4.8952500412050989295774e-26
>>> bessely(2.5, -10**50)
(0.0 + 4.8952500412050989295774e-26j)
Derivatives and antiderivatives of any order can be computed:
>>> bessely(2, 3.5, 1)
0.3842618820422660066089231
5.15. Welcome to mpmaths documentation!
815
SymPy Documentation, Release 0.7.6
>>> diff(lambda x: bessely(2, x), 3.5)
0.3842618820422660066089231
>>> bessely(0.5, 3.5, 1)
-0.2066598304156764337900417
>>> diff(lambda x: bessely(0.5, x), 3.5)
-0.2066598304156764337900417
>>> diff(lambda x: bessely(2, x), 0.5, 10)
-208173867409.5547350101511
>>> bessely(2, 0.5, 10)
-208173867409.5547350101511
>>> bessely(2, 100.5, 100)
0.02668487547301372334849043
>>> quad(lambda x: bessely(2,x), [1,3])
-1.377046859093181969213262
>>> bessely(2,3,-1) - bessely(2,1,-1)
-1.377046859093181969213262
besseli()
mpmath.besseli(n, x, derivative=0)
besseli(n, x, derivative=0) gives the modied Bessel function of the rst kind,
In (x) = in Jn (ix).
With derivative = m 6= 0, the m-th derivative
dm
In (x)
dxm
is computed.
Plots
# Modified Bessel function I_n(x) on the real line for n=0,1,2,3
i0 = lambda x: besseli(0,x)
i1 = lambda x: besseli(1,x)
i2 = lambda x: besseli(2,x)
i3 = lambda x: besseli(3,x)
plot([i0,i1,i2,i3],[0,5],[0,5])
816
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
# Modified Bessel function I_n(z) in the complex plane
cplot(lambda z: besseli(1,z), [-8,8], [-8,8], points=50000)
5.15. Welcome to mpmaths documentation!
817
SymPy Documentation, Release 0.7.6
Examples
Some values of In (x):
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> besseli(0,0)
1.0
>>> besseli(1,0)
0.0
>>> besseli(0,1)
1.266065877752008335598245
>>> besseli(3.5, 2+3j)
(-0.2904369752642538144289025 - 0.4469098397654815837307006j)
Arguments may be large:
>>> besseli(2, 1000)
2.480717210191852440616782e+432
>>> besseli(2, 10**10)
4.299602851624027900335391e+4342944813
>>> besseli(2, 6000+10000j)
(-2.114650753239580827144204e+2603 + 4.385040221241629041351886e+2602j)
For integers n, the following integral representation holds:
>>> mp.dps = 15
>>> n = 3
>>> x = 2.3
818
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> quad(lambda t: exp(x*cos(t))*cos(n*t), [0,pi])/pi
0.349223221159309
>>> besseli(n,x)
0.349223221159309
Derivatives and antiderivatives of any order can be computed:
>>> mp.dps = 25
>>> besseli(2, 7.5, 1)
195.8229038931399062565883
>>> diff(lambda x: besseli(2,x), 7.5)
195.8229038931399062565883
>>> besseli(2, 7.5, 10)
153.3296508971734525525176
>>> diff(lambda x: besseli(2,x), 7.5, 10)
153.3296508971734525525176
>>> besseli(2,7.5,-1) - besseli(2,3.5,-1)
202.5043900051930141956876
>>> quad(lambda x: besseli(2,x), [3.5, 7.5])
202.5043900051930141956876
besselk()
mpmath.besselk(n, x)
besselk(n, x) gives the modied Bessel function of the second kind,
Kn (x) =
In (x) In (x)
2
sin(n)
For n an integer, this formula should be understood as a limit.
Plots
# Modified Bessel function of 2nd kind K_n(x) on the real line for n=0,1,2,3
k0 = lambda x: besselk(0,x)
k1 = lambda x: besselk(1,x)
k2 = lambda x: besselk(2,x)
k3 = lambda x: besselk(3,x)
plot([k0,k1,k2,k3],[0,8],[0,5])
5.15. Welcome to mpmaths documentation!
819
SymPy Documentation, Release 0.7.6
# Modified Bessel function of 2nd kind K_n(z) in the complex plane
cplot(lambda z: besselk(1,z), [-8,8], [-8,8], points=50000)
820
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
Evaluation is supported for arbitrary complex arguments:
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> besselk(0,1)
0.4210244382407083333356274
>>> besselk(0, -1)
(0.4210244382407083333356274 - 3.97746326050642263725661j)
>>> besselk(3.5, 2+3j)
(-0.02090732889633760668464128 + 0.2464022641351420167819697j)
>>> besselk(2+3j, 0.5)
(0.9615816021726349402626083 + 0.1918250181801757416908224j)
Arguments may be large:
>>> besselk(0, 100)
4.656628229175902018939005e-45
>>> besselk(1, 10**6)
4.131967049321725588398296e-434298
>>> besselk(1, 10**6*j)
(0.001140348428252385844876706 - 0.0005200017201681152909000961j)
>>> besselk(4.5, fmul(10**50, j, exact=True))
(1.561034538142413947789221e-26 + 1.243554598118700063281496e-25j)
The point x = 0 is a singularity (logarithmic if n = 0):
5.15. Welcome to mpmaths documentation!
821
SymPy Documentation, Release 0.7.6
>>> besselk(0,0)
+inf
>>> besselk(1,0)
+inf
>>> for n in range(-4, 5):
...
print(besselk(n, 1e-1000))
...
4.8e+4001
8.0e+3000
2.0e+2000
1.0e+1000
2302.701024509704096466802
1.0e+1000
2.0e+2000
8.0e+3000
4.8e+4001
Bessel function zeros
besseljzero()
mpmath.besseljzero(v, m, derivative=0)
For a real order 0 and a positive integer m, returns j,m , the m-th positive zero of
the Bessel function of the rst kind J (z) (see besselj() (page 810)). Alternatively, with
0
derivative=1, gives the rst nonnegative simple zero j,m
of J0 (z).
The indexing convention is that used by Abramowitz & Stegun and the DLMF. Note the
0
= 0, while all other zeros are positive. In eect, only simple zeros are
special case j0,1
counted (all zeros of Bessel functions are simple except possibly z = 0) and j,m becomes
a monotonic function of both and m.
The zeros are interlaced according to the inequalities
0
0
j,k
< j,k < j,k+1
j,1 < j+1,2 < j,2 < j+1,2 < j,3 <
Examples
Initial zeros of the Bessel functions J0 (z), J1 (z), J2 (z):
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> besseljzero(0,1); besseljzero(0,2); besseljzero(0,3)
2.404825557695772768621632
5.520078110286310649596604
8.653727912911012216954199
>>> besseljzero(1,1); besseljzero(1,2); besseljzero(1,3)
3.831705970207512315614436
7.01558666981561875353705
10.17346813506272207718571
>>> besseljzero(2,1); besseljzero(2,2); besseljzero(2,3)
5.135622301840682556301402
8.417244140399864857783614
11.61984117214905942709415
Initial zeros of J00 (z), J10 (z), J20 (z):
822
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
0.0
3.831705970207512315614436
7.01558666981561875353705
>>> besseljzero(1,1,1); besseljzero(1,2,1); besseljzero(1,3,1)
1.84118378134065930264363
5.331442773525032636884016
8.536316366346285834358961
>>> besseljzero(2,1,1); besseljzero(2,2,1); besseljzero(2,3,1)
3.054236928227140322755932
6.706133194158459146634394
9.969467823087595793179143
Zeros with large index:
>>> besseljzero(0,100); besseljzero(0,1000); besseljzero(0,10000)
313.3742660775278447196902
3140.807295225078628895545
31415.14114171350798533666
>>> besseljzero(5,100); besseljzero(5,1000); besseljzero(5,10000)
321.1893195676003157339222
3148.657306813047523500494
31422.9947255486291798943
>>> besseljzero(0,100,1); besseljzero(0,1000,1); besseljzero(0,10000,1)
311.8018681873704508125112
3139.236339643802482833973
31413.57032947022399485808
Zeros of functions with large order:
>>> besseljzero(50,1)
57.11689916011917411936228
>>> besseljzero(50,2)
62.80769876483536093435393
>>> besseljzero(50,100)
388.6936600656058834640981
>>> besseljzero(50,1,1)
52.99764038731665010944037
>>> besseljzero(50,2,1)
60.02631933279942589882363
>>> besseljzero(50,100,1)
387.1083151608726181086283
Zeros of functions with fractional order:
>>> besseljzero(0.5,1); besseljzero(1.5,1); besseljzero(2.25,4)
3.141592653589793238462643
4.493409457909064175307881
15.15657692957458622921634
Both J (z) and J0 (z) can be expressed as innite products over their zeros:
>>> v,z = 2, mpf(1)
>>> (z/2)**v/gamma(v+1) * \
...
nprod(lambda k: 1-(z/besseljzero(v,k))**2, [1,inf])
...
0.1149034849319004804696469
>>> besselj(v,z)
0.1149034849319004804696469
>>> (z/2)**(v-1)/2/gamma(v) * \
...
nprod(lambda k: 1-(z/besseljzero(v,k,1))**2, [1,inf])
5.15. Welcome to mpmaths documentation!
823
SymPy Documentation, Release 0.7.6
...
0.2102436158811325550203884
>>> besselj(v,z,1)
0.2102436158811325550203884
besselyzero()
mpmath.besselyzero(v, m, derivative=0)
For a real order 0 and a positive integer m, returns y,m , the m-th positive zero of the
Bessel function of the second kind Y (z) (see bessely() (page 814)). Alternatively, with
0
derivative=1, gives the rst positive zero y,m
of Y0 (z).
The zeros are interlaced according to the inequalities
0
< y,k+1
y,k < y,k
y,1 < y+1,2 < y,2 < y+1,2 < y,3 <
Examples
Initial zeros of the Bessel functions Y0 (z), Y1 (z), Y2 (z):
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> besselyzero(0,1); besselyzero(0,2); besselyzero(0,3)
0.8935769662791675215848871
3.957678419314857868375677
7.086051060301772697623625
>>> besselyzero(1,1); besselyzero(1,2); besselyzero(1,3)
2.197141326031017035149034
5.429681040794135132772005
8.596005868331168926429606
>>> besselyzero(2,1); besselyzero(2,2); besselyzero(2,3)
3.384241767149593472701426
6.793807513268267538291167
10.02347797936003797850539
Initial zeros of Y00 (z), Y10 (z), Y20 (z):
>>> besselyzero(0,1,1); besselyzero(0,2,1); besselyzero(0,3,1)
2.197141326031017035149034
5.429681040794135132772005
8.596005868331168926429606
>>> besselyzero(1,1,1); besselyzero(1,2,1); besselyzero(1,3,1)
3.683022856585177699898967
6.941499953654175655751944
10.12340465543661307978775
>>> besselyzero(2,1,1); besselyzero(2,2,1); besselyzero(2,3,1)
5.002582931446063945200176
8.350724701413079526349714
11.57419546521764654624265
Zeros with large index:
>>> besselyzero(0,100); besselyzero(0,1000); besselyzero(0,10000)
311.8034717601871549333419
3139.236498918198006794026
31413.57034538691205229188
>>> besselyzero(5,100); besselyzero(5,1000); besselyzero(5,10000)
319.6183338562782156235062
824
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
3147.086508524556404473186
31421.42392920214673402828
>>> besselyzero(0,100,1); besselyzero(0,1000,1); besselyzero(0,10000,1)
313.3726705426359345050449
3140.807136030340213610065
31415.14112579761578220175
Zeros of functions with large order:
>>> besselyzero(50,1)
53.50285882040036394680237
>>> besselyzero(50,2)
60.11244442774058114686022
>>> besselyzero(50,100)
387.1096509824943957706835
>>> besselyzero(50,1,1)
56.96290427516751320063605
>>> besselyzero(50,2,1)
62.74888166945933944036623
>>> besselyzero(50,100,1)
388.6923300548309258355475
Zeros of functions with fractional order:
>>> besselyzero(0.5,1); besselyzero(1.5,1); besselyzero(2.25,4)
1.570796326794896619231322
2.798386045783887136720249
13.56721208770735123376018
Hankel functions
hankel1()
mpmath.hankel1(n, x)
hankel1(n,x) computes the Hankel function of the rst kind, which is the complex combination of Bessel functions given by
Hn(1) (x) = Jn (x) + iYn (x).
Plots
# Hankel function H1_n(x) on the real line for n=0,1,2,3
h0 = lambda x: hankel1(0,x)
h1 = lambda x: hankel1(1,x)
h2 = lambda x: hankel1(2,x)
h3 = lambda x: hankel1(3,x)
plot([h0,h1,h2,h3],[0,6],[-2,1])
5.15. Welcome to mpmaths documentation!
825
SymPy Documentation, Release 0.7.6
# Hankel function H1_n(z) in the complex plane
cplot(lambda z: hankel1(1,z), [-8,8], [-8,8], points=50000)
826
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
The Hankel function is generally complex-valued:
>>> from sympy.mpmath import
>>> mp.dps = 25; mp.pretty =
>>> hankel1(2, pi)
(0.4854339326315091097054957
>>> hankel1(3.5, pi)
(0.2340002029630507922628888
*
True
- 0.0999007139290278787734903j)
- 0.6419643823412927142424049j)
hankel2()
mpmath.hankel2(n, x)
hankel2(n,x) computes the Hankel function of the second kind, which is the complex
combination of Bessel functions given by
Hn(2) (x) = Jn (x) iYn (x).
Plots
# Hankel function H2_n(x) on the real line for n=0,1,2,3
h0 = lambda x: hankel2(0,x)
h1 = lambda x: hankel2(1,x)
h2 = lambda x: hankel2(2,x)
h3 = lambda x: hankel2(3,x)
plot([h0,h1,h2,h3],[0,6],[-1,2])
5.15. Welcome to mpmaths documentation!
827
SymPy Documentation, Release 0.7.6
# Hankel function H2_n(z) in the complex plane
cplot(lambda z: hankel2(1,z), [-8,8], [-8,8], points=50000)
828
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
The Hankel function is generally complex-valued:
>>> from sympy.mpmath import
>>> mp.dps = 25; mp.pretty =
>>> hankel2(2, pi)
(0.4854339326315091097054957
>>> hankel2(3.5, pi)
(0.2340002029630507922628888
*
True
+ 0.0999007139290278787734903j)
+ 0.6419643823412927142424049j)
Kelvin functions
ber()
mpmath.ber(ctx, n, z, **kwargs)
Computes the Kelvin function ber, which for real arguments gives the real part of the
Bessel J function of a rotated argument
)
(
Jn xe3i/4 = bern (x) + ibein (x).
The imaginary part is given by bei() (page 831).
Plots
5.15. Welcome to mpmaths documentation!
829
SymPy Documentation, Release 0.7.6
# Kelvin functions ber_n(x) and bei_n(x) on the real line for n=0,2
f0 = lambda x: ber(0,x)
f1 = lambda x: bei(0,x)
f2 = lambda x: ber(2,x)
f3 = lambda x: bei(2,x)
plot([f0,f1,f2,f3],[0,10],[-10,10])
Examples
Verifying the dening relation:
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> n, x = 2, 3.5
>>> ber(n,x)
1.442338852571888752631129
>>> bei(n,x)
-0.948359035324558320217678
>>> besselj(n, x*root(1,8,3))
(1.442338852571888752631129 - 0.948359035324558320217678j)
The ber and bei functions are also dened by analytic continuation for complex arguments:
>>> ber(1+j, 2+3j)
(4.675445984756614424069563 - 15.84901771719130765656316j)
>>> bei(1+j, 2+3j)
(15.83886679193707699364398 + 4.684053288183046528703611j)
830
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
bei()
mpmath.bei(ctx, n, z, **kwargs)
Computes the Kelvin function bei, which for real arguments gives the imaginary part of
the Bessel J function of a rotated argument. See ber() (page 829).
ker()
mpmath.ker(ctx, n, z, **kwargs)
Computes the Kelvin function ker, which for real arguments gives the real part of the
(rescaled) Bessel K function of a rotated argument
(
)
ei/2 Kn xe3i/4 = kern (x) + ikein (x).
The imaginary part is given by kei() (page 832).
Plots
# Kelvin functions ker_n(x) and kei_n(x) on the real line for n=0,2
f0 = lambda x: ker(0,x)
f1 = lambda x: kei(0,x)
f2 = lambda x: ker(2,x)
f3 = lambda x: kei(2,x)
plot([f0,f1,f2,f3],[0,5],[-1,4])
Examples
Verifying the dening relation:
5.15. Welcome to mpmaths documentation!
831
SymPy Documentation, Release 0.7.6
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> n, x = 2, 4.5
>>> ker(n,x)
0.02542895201906369640249801
>>> kei(n,x)
-0.02074960467222823237055351
>>> exp(-n*pi*j/2) * besselk(n, x*root(1,8,1))
(0.02542895201906369640249801 - 0.02074960467222823237055351j)
The ker and kei functions are also dened by analytic continuation for complex arguments:
>>> ker(1+j, 3+4j)
(1.586084268115490421090533 - 2.939717517906339193598719j)
>>> kei(1+j, 3+4j)
(-2.940403256319453402690132 - 1.585621643835618941044855j)
kei()
mpmath.kei(ctx, n, z, **kwargs)
Computes the Kelvin function kei, which for real arguments gives the imaginary part of
the (rescaled) Bessel K function of a rotated argument. See ker() (page 831).
Struve functions
struveh()
mpmath.struveh(ctx, n, z, **kwargs)
Gives the Struve function
Hn (z) =
( z )2k+n+1
(1)k
3
(k +
+ n + 2) 2
k=0
3
2 )(k
which is a solution to the Struve dierential equation
z 2 f 00 (z) + zf 0 (z) + (z 2 n2 )f (z) =
2z n+1
.
(2n 1)!!
Examples
Evaluation for arbitrary real and complex arguments:
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> struveh(0, 3.5)
0.3608207733778295024977797
>>> struveh(-1, 10)
-0.255212719726956768034732
>>> struveh(1, -100.5)
0.5819566816797362287502246
>>> struveh(2.5, 10000000000000)
3153915652525200060.308937
>>> struveh(2.5, -10000000000000)
(0.0 - 3153915652525200060.308937j)
>>> struveh(1+j, 1000000+4000000j)
(-3.066421087689197632388731e+1737173 - 1.596619701076529803290973e+1737173j)
832
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
A Struve function of half-integer order is elementary; for example:
>>> z = 3
>>> struveh(0.5, 3)
0.9167076867564138178671595
>>> sqrt(2/(pi*z))*(1-cos(z))
0.9167076867564138178671595
Numerically verifying the dierential equation:
>>> z = mpf(4.5)
>>> n = 3
>>> f = lambda z: struveh(n,z)
>>> lhs = z**2*diff(f,z,2) + z*diff(f,z) + (z**2-n**2)*f(z)
>>> rhs = 2*z**(n+1)/fac2(2*n-1)/pi
>>> lhs
17.40359302709875496632744
>>> rhs
17.40359302709875496632744
struvel()
mpmath.struvel(ctx, n, z, **kwargs)
Gives the modied Struve function
Ln (z) = ieni/2 Hn (iz)
which solves to the modied Struve dierential equation
z 2 f 00 (z) + zf 0 (z) (z 2 + n2 )f (z) =
2z n+1
.
(2n 1)!!
Examples
Evaluation for arbitrary real and complex arguments:
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> struvel(0, 3.5)
7.180846515103737996249972
>>> struvel(-1, 10)
2670.994904980850550721511
>>> struvel(1, -100.5)
1.757089288053346261497686e+42
>>> struvel(2.5, 10000000000000)
4.160893281017115450519948e+4342944819025
>>> struvel(2.5, -10000000000000)
(0.0 - 4.160893281017115450519948e+4342944819025j)
>>> struvel(1+j, 700j)
(-0.1721150049480079451246076 + 0.1240770953126831093464055j)
>>> struvel(1+j, 1000000+4000000j)
(-2.973341637511505389128708e+434290 - 5.164633059729968297147448e+434290j)
Numerically verifying the dierential equation:
>>>
>>>
>>>
>>>
>>>
z =
n =
f =
lhs
rhs
mpf(3.5)
3
lambda z: struvel(n,z)
= z**2*diff(f,z,2) + z*diff(f,z) - (z**2+n**2)*f(z)
= 2*z**(n+1)/fac2(2*n-1)/pi
5.15. Welcome to mpmaths documentation!
833
SymPy Documentation, Release 0.7.6
>>> lhs
6.368850306060678353018165
>>> rhs
6.368850306060678353018165
Anger-Weber functions
angerj()
mpmath.angerj(ctx, v, z, **kwargs)
Gives the Anger function
1
J (z) =
cos(t z sin t)dt
0
which is an entire function of both the parameter and the argument z. It solves the
inhomogeneous Bessel dierential equation
(
)
2
(z )
1 0
00
f (z) + f (z) + 1 2 f (z) =
sin().
z
z
z 2
Examples
Evaluation for real and complex parameter and argument:
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> angerj(2,3)
0.4860912605858910769078311
>>> angerj(-3+4j, 2+5j)
(-5033.358320403384472395612 + 585.8011892476145118551756j)
>>> angerj(3.25, 1e6j)
(4.630743639715893346570743e+434290 - 1.117960409887505906848456e+434291j)
>>> angerj(-1.5, 1e6)
0.0002795719747073879393087011
The Anger function coincides with the Bessel J-function when is an integer:
>>> angerj(1,3); besselj(1,3)
0.3390589585259364589255146
0.3390589585259364589255146
>>> angerj(1.5,3); besselj(1.5,3)
0.4088969848691080859328847
0.4777182150870917715515015
Verifying the dierential equation:
>>> v,z = mpf(2.25), 0.75
>>> f = lambda z: angerj(v,z)
>>> diff(f,z,2) + diff(f,z)/z + (1-(v/z)**2)*f(z)
-0.6002108774380707130367995
>>> (z-v)/(pi*z**2) * sinpi(v)
-0.6002108774380707130367995
Verifying the integral representation:
>>> angerj(v,z)
0.1145380759919333180900501
>>> quad(lambda t: cos(v*t-z*sin(t))/pi, [0,pi])
0.1145380759919333180900501
834
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
References
1.[DLMF] (page 1910) section 11.10: Anger-Weber Functions
webere()
mpmath.webere(ctx, v, z, **kwargs)
Gives the Weber function
1
E (z) =
sin(t z sin t)dt
0
which is an entire function of both the parameter and the argument z. It solves the
inhomogeneous Bessel dierential equation
(
)
1 0
2
1
00
f (z) + f (z) + 1 2 f (z) = 2 (z + + (z ) cos()).
z
z
z
Examples
Evaluation for real and complex parameter and argument:
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> webere(2,3)
-0.1057668973099018425662646
>>> webere(-3+4j, 2+5j)
(-585.8081418209852019290498 - 5033.314488899926921597203j)
>>> webere(3.25, 1e6j)
(-1.117960409887505906848456e+434291 - 4.630743639715893346570743e+434290j)
>>> webere(3.25, 1e6)
-0.00002812518265894315604914453
Up to addition of a rational function of z, the Weber function coincides with the Struve
H-function when is an integer:
>>> webere(1,3); 2/pi-struveh(1,3)
-0.3834897968188690177372881
-0.3834897968188690177372881
>>> webere(5,3); 26/(35*pi)-struveh(5,3)
0.2009680659308154011878075
0.2009680659308154011878075
Verifying the dierential equation:
>>> v,z = mpf(2.25), 0.75
>>> f = lambda z: webere(v,z)
>>> diff(f,z,2) + diff(f,z)/z + (1-(v/z)**2)*f(z)
-1.097441848875479535164627
>>> -(z+v+(z-v)*cospi(v))/(pi*z**2)
-1.097441848875479535164627
Verifying the integral representation:
>>> webere(v,z)
0.1486507351534283744485421
>>> quad(lambda t: sin(v*t-z*sin(t))/pi, [0,pi])
0.1486507351534283744485421
References
1.[DLMF] (page 1910) section 11.10: Anger-Weber Functions
5.15. Welcome to mpmaths documentation!
835
SymPy Documentation, Release 0.7.6
Lommel functions
lommels1()
mpmath.lommels1(ctx, u, v, z, **kwargs)
(1)
Gives the Lommel function s, or s,
(
)
z +1
+ 3 + + 3 z2
s, (z) =
,
;
1 F2 1;
( + 1)( + + 1)
2
2
4
which solves the inhomogeneous Bessel equation
z 2 f 00 (z) + zf 0 (z) + (z 2 2 )f (z) = z +1 .
A second solution is given by lommels2() (page 837).
Plots
# Lommel function s_(u,v)(x) on the real line for a few different u,v
f1 = lambda x: lommels1(-1,2.5,x)
f2 = lambda x: lommels1(0,0.5,x)
f3 = lambda x: lommels1(0,6,x)
f4 = lambda x: lommels1(0.5,3,x)
plot([f1,f2,f3,f4], [0,20])
Examples
An integral representation:
836
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> u,v,z = 0.25, 0.125, mpf(0.75)
>>> lommels1(u,v,z)
0.4276243877565150372999126
>>> (bessely(v,z)*quad(lambda t: t**u*besselj(v,t), [0,z]) - \
... besselj(v,z)*quad(lambda t: t**u*bessely(v,t), [0,z]))*(pi/2)
0.4276243877565150372999126
A special value:
>>> lommels1(v,v,z)
0.5461221367746048054932553
>>> gamma(v+0.5)*sqrt(pi)*power(2,v-1)*struveh(v,z)
0.5461221367746048054932553
Verifying the dierential equation:
>>> f = lambda z: lommels1(u,v,z)
>>> z**2*diff(f,z,2) + z*diff(f,z) + (z**2-v**2)*f(z)
0.6979536443265746992059141
>>> z**(u+1)
0.6979536443265746992059141
References
1.[GradshteynRyzhik] (page 1910)
2.[Weisstein] (page 1910) http://mathworld.wolfram.com/LommelFunction.html
lommels2()
mpmath.lommels2(ctx, u, v, z, **kwargs)
(2)
Gives the second Lommel function S, or s,
(
) (
)
S, (z) = s, (z) + 21 12 ( + 1) 12 ( + + 1)
[
]
sin( 21 ( ))J (z) cos( 21 ( ))Y (z)
which solves the same dierential equation as lommels1() (page 836).
Plots
# Lommel function S_(u,v)(x) on the real line for a few different u,v
f1 = lambda x: lommels2(-1,2.5,x)
f2 = lambda x: lommels2(1.5,2,x)
f3 = lambda x: lommels2(2.5,1,x)
f4 = lambda x: lommels2(3.5,-0.5,x)
plot([f1,f2,f3,f4], [0,8], [-8,8])
5.15. Welcome to mpmaths documentation!
837
SymPy Documentation, Release 0.7.6
Examples
For large |z|, S, z 1 :
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> lommels2(10,2,30000)
1.968299831601008419949804e+40
>>> power(30000,9)
1.9683e+40
A special value:
>>> u,v,z = 0.5, 0.125, mpf(0.75)
>>> lommels2(v,v,z)
0.9589683199624672099969765
>>> (struveh(v,z)-bessely(v,z))*power(2,v-1)*sqrt(pi)*gamma(v+0.5)
0.9589683199624672099969765
Verifying the dierential equation:
>>> f = lambda z: lommels2(u,v,z)
>>> z**2*diff(f,z,2) + z*diff(f,z) + (z**2-v**2)*f(z)
0.6495190528383289850727924
>>> z**(u+1)
0.6495190528383289850727924
References
1.[GradshteynRyzhik] (page 1910)
838
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
2.[Weisstein] (page 1910) http://mathworld.wolfram.com/LommelFunction.html
Airy and Scorer functions
airyai()
mpmath.airyai(z, derivative=0, **kwargs)
Computes the Airy function Ai(z), which is the solution of the Airy dierential equation
f 00 (z) zf (z) = 0 with initial conditions
1
( )
32/3 23
1
Ai0 (0) = 1/3 ( 1 ) .
3 3
Ai(0) =
Other common ways of dening the Ai-function include integrals such as
(
)
1
1 3
Ai(x) =
cos
t + xt dt
xR
0
3
( 3
)
t
z3
3
Ai(z) =
exp 3 dt.
2 0
3
3t
The Ai-function is an entire function with a turning point, behaving roughly like a slowly
decaying sine wave for z < 0 and like a rapidly decreasing exponential for z > 0. A second
solution of the Airy dierential equation is given by Bi(z) (see airybi() (page 844)).
Optionally, with derivative=alpha, airyai() can compute the -th order fractional
derivative with respect to z. For = n = 1, 2, 3, . . . this gives the derivative Ai(n) (z), and
for = n = 1, 2, 3, . . . this gives the n-fold iterated integral
f0 (z) = Ai(z)
z
fn (z) =
fn1 (t)dt.
0
The Ai-function has innitely many zeros, all located along the negative half of the real
axis. They can be computed with airyaizero() (page 848).
Plots
# Airy function Ai(x), Ai(x) and int_0^x Ai(t) dt on the real line
f = airyai
f_diff = lambda z: airyai(z, derivative=1)
f_int = lambda z: airyai(z, derivative=-1)
plot([f, f_diff, f_int], [-10,5])
5.15. Welcome to mpmaths documentation!
839
SymPy Documentation, Release 0.7.6
# Airy function Ai(z) in the complex plane
cplot(airyai, [-8,8], [-8,8], points=50000)
840
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Basic examples
Limits and values include:
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> airyai(0); 1/(power(3,2/3)*gamma(2/3))
0.3550280538878172392600632
0.3550280538878172392600632
>>> airyai(1)
0.1352924163128814155241474
>>> airyai(-1)
0.5355608832923521187995166
>>> airyai(inf); airyai(-inf)
0.0
0.0
Evaluation is supported for large magnitudes of the argument:
>>> airyai(-100)
0.1767533932395528780908311
>>> airyai(100)
2.634482152088184489550553e-291
>>> airyai(50+50j)
(-5.31790195707456404099817e-68 - 1.163588003770709748720107e-67j)
>>> airyai(-50+50j)
(1.041242537363167632587245e+158 + 3.347525544923600321838281e+157j)
Huge arguments are also ne:
5.15. Welcome to mpmaths documentation!
841
SymPy Documentation, Release 0.7.6
>>> airyai(10**10)
1.162235978298741779953693e-289529654602171
>>> airyai(-10**10)
0.0001736206448152818510510181
>>> w = airyai(10**10*(1+j))
>>> w.real
5.711508683721355528322567e-186339621747698
>>> w.imag
1.867245506962312577848166e-186339621747697
The rst root of the Ai-function is:
>>> findroot(airyai, -2)
-2.338107410459767038489197
>>> airyaizero(1)
-2.338107410459767038489197
Properties and relations
Verifying the Airy dierential equation:
>>> for z in [-3.4, 0, 2.5, 1+2j]:
...
chop(airyai(z,2) - z*airyai(z))
...
0.0
0.0
0.0
0.0
The rst few terms of the Taylor series expansion around z = 0 (every third term is zero):
>>> nprint(taylor(airyai, 0, 5))
[0.355028, -0.258819, 0.0, 0.0591713, -0.0215683, 0.0]
The Airy functions satisfy the Wronskian relation Ai(z) Bi0 (z) Ai0 (z) Bi(z) = 1/:
>>> z = -0.5
>>> airyai(z)*airybi(z,1) - airyai(z,1)*airybi(z)
0.3183098861837906715377675
>>> 1/pi
0.3183098861837906715377675
The Airy functions can be expressed in terms of Bessel functions of order 1/3. For
<[z] 0, we have:
>>> z = -3
>>> airyai(z)
-0.3788142936776580743472439
>>> y = 2*power(-z,3/2)/3
>>> (sqrt(-z) * (besselj(1/3,y) + besselj(-1/3,y)))/3
-0.3788142936776580743472439
Derivatives and integrals
Derivatives of the Ai-function (directly and using diff() (page 999)):
>>> airyai(-3,1); diff(airyai,-3)
0.3145837692165988136507873
0.3145837692165988136507873
>>> airyai(-3,2); diff(airyai,-3,2)
1.136442881032974223041732
842
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
1.136442881032974223041732
>>> airyai(1000,1); diff(airyai,1000)
-2.943133917910336090459748e-9156
-2.943133917910336090459748e-9156
Several derivatives at z = 0:
>>> airyai(0,0); airyai(0,1); airyai(0,2)
0.3550280538878172392600632
-0.2588194037928067984051836
0.0
>>> airyai(0,3); airyai(0,4); airyai(0,5)
0.3550280538878172392600632
-0.5176388075856135968103671
0.0
>>> airyai(0,15); airyai(0,16); airyai(0,17)
1292.30211615165475090663
-3188.655054727379756351861
0.0
The integral of the Ai-function:
>>> airyai(3,-1); quad(airyai, [0,3])
0.3299203760070217725002701
0.3299203760070217725002701
>>> airyai(-10,-1); quad(airyai, [0,-10])
-0.765698403134212917425148
-0.765698403134212917425148
Integrals of high or fractional order:
>>> airyai(-2,0.5); differint(airyai,-2,0.5,0)
(0.0 + 0.2453596101351438273844725j)
(0.0 + 0.2453596101351438273844725j)
>>> airyai(-2,-4); differint(airyai,-2,-4,0)
0.2939176441636809580339365
0.2939176441636809580339365
>>> airyai(0,-1); airyai(0,-2); airyai(0,-3)
0.0
0.0
0.0
Integrals of the Ai-function can be evaluated at limit points:
>>> airyai(-1000000,-1); airyai(-inf,-1)
-0.6666843728311539978751512
-0.6666666666666666666666667
>>> airyai(10,-1); airyai(+inf,-1)
0.3333333332991690159427932
0.3333333333333333333333333
>>> airyai(+inf,-2); airyai(+inf,-3)
+inf
+inf
>>> airyai(-1000000,-2); airyai(-inf,-2)
666666.4078472650651209742
+inf
>>> airyai(-1000000,-3); airyai(-inf,-3)
-333333074513.7520264995733
-inf
5.15. Welcome to mpmaths documentation!
843
SymPy Documentation, Release 0.7.6
References
1.[DLMF] (page 1910) Chapter 9: Airy and Related Functions
2.[WolframFunctions] (page 1910) section: Bessel-Type Functions
airybi()
mpmath.airybi(z, derivative=0, **kwargs)
Computes the Airy function Bi(z), which is the solution of the Airy dierential equation
f 00 (z) zf (z) = 0 with initial conditions
Bi(0) =
1
( )
31/6 23
Bi0 (0) =
31/6
( ).
13
Like the Ai-function (see airyai() (page 839)), the Bi-function is oscillatory for z < 0,
but it grows rather than decreases for z > 0.
Optionally, as for airyai() (page 839), derivatives, integrals and fractional derivatives
can be computed with the derivative parameter.
The Bi-function has innitely many zeros along the negative half-axis, as well as complex
zeros, which can all be computed with airybizero() (page 849).
Plots
# Airy function Bi(x), Bi(x) and int_0^x Bi(t) dt on the real line
f = airybi
f_diff = lambda z: airybi(z, derivative=1)
f_int = lambda z: airybi(z, derivative=-1)
plot([f, f_diff, f_int], [-10,2], [-1,2])
844
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
# Airy function Bi(z) in the complex plane
cplot(airybi, [-8,8], [-8,8], points=50000)
5.15. Welcome to mpmaths documentation!
845
SymPy Documentation, Release 0.7.6
Basic examples
Limits and values include:
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> airybi(0); 1/(power(3,1/6)*gamma(2/3))
0.6149266274460007351509224
0.6149266274460007351509224
>>> airybi(1)
1.207423594952871259436379
>>> airybi(-1)
0.10399738949694461188869
>>> airybi(inf); airybi(-inf)
+inf
0.0
Evaluation is supported for large magnitudes of the argument:
>>> airybi(-100)
0.02427388768016013160566747
>>> airybi(100)
6.041223996670201399005265e+288
>>> airybi(50+50j)
(-5.322076267321435669290334e+63 + 1.478450291165243789749427e+65j)
>>> airybi(-50+50j)
(-3.347525544923600321838281e+157 + 1.041242537363167632587245e+158j)
Huge arguments:
846
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> airybi(10**10)
1.369385787943539818688433e+289529654602165
>>> airybi(-10**10)
0.001775656141692932747610973
>>> w = airybi(10**10*(1+j))
>>> w.real
-6.559955931096196875845858e+186339621747689
>>> w.imag
-6.822462726981357180929024e+186339621747690
The rst real root of the Bi-function is:
>>> findroot(airybi, -1); airybizero(1)
-1.17371322270912792491998
-1.17371322270912792491998
Properties and relations
Verifying the Airy dierential equation:
>>> for z in [-3.4, 0, 2.5, 1+2j]:
...
chop(airybi(z,2) - z*airybi(z))
...
0.0
0.0
0.0
0.0
The rst few terms of the Taylor series expansion around z = 0 (every third term is zero):
>>> nprint(taylor(airybi, 0, 5))
[0.614927, 0.448288, 0.0, 0.102488, 0.0373574, 0.0]
The Airy functions can be expressed in terms of Bessel functions of order 1/3. For
<[z] 0, we have:
>>> z = -3
>>> airybi(z)
-0.1982896263749265432206449
>>> p = 2*power(-z,3/2)/3
>>> sqrt(-mpf(z)/3)*(besselj(-1/3,p) - besselj(1/3,p))
-0.1982896263749265432206449
Derivatives and integrals
Derivatives of the Bi-function (directly and using diff() (page 999)):
>>> airybi(-3,1); diff(airybi,-3)
-0.675611222685258537668032
-0.675611222685258537668032
>>> airybi(-3,2); diff(airybi,-3,2)
0.5948688791247796296619346
0.5948688791247796296619346
>>> airybi(1000,1); diff(airybi,1000)
1.710055114624614989262335e+9156
1.710055114624614989262335e+9156
Several derivatives at z = 0:
>>> airybi(0,0); airybi(0,1); airybi(0,2)
0.6149266274460007351509224
5.15. Welcome to mpmaths documentation!
847
SymPy Documentation, Release 0.7.6
0.4482883573538263579148237
0.0
>>> airybi(0,3); airybi(0,4); airybi(0,5)
0.6149266274460007351509224
0.8965767147076527158296474
0.0
>>> airybi(0,15); airybi(0,16); airybi(0,17)
2238.332923903442675949357
5522.912562599140729510628
0.0
The integral of the Bi-function:
>>> airybi(3,-1); quad(airybi, [0,3])
10.06200303130620056316655
10.06200303130620056316655
>>> airybi(-10,-1); quad(airybi, [0,-10])
-0.01504042480614002045135483
-0.01504042480614002045135483
Integrals of high or fractional order:
>>> airybi(-2,0.5); differint(airybi, -2, 0.5, 0)
(0.0 + 0.5019859055341699223453257j)
(0.0 + 0.5019859055341699223453257j)
>>> airybi(-2,-4); differint(airybi,-2,-4,0)
0.2809314599922447252139092
0.2809314599922447252139092
>>> airybi(0,-1); airybi(0,-2); airybi(0,-3)
0.0
0.0
0.0
Integrals of the Bi-function can be evaluated at limit points:
>>> airybi(-1000000,-1); airybi(-inf,-1)
0.000002191261128063434047966873
0.0
>>> airybi(10,-1); airybi(+inf,-1)
147809803.1074067161675853
+inf
>>> airybi(+inf,-2); airybi(+inf,-3)
+inf
+inf
>>> airybi(-1000000,-2); airybi(-inf,-2)
0.4482883750599908479851085
0.4482883573538263579148237
>>> gamma(2/3)*power(3,2/3)/(2*pi)
0.4482883573538263579148237
>>> airybi(-100000,-3); airybi(-inf,-3)
-44828.52827206932872493133
-inf
>>> airybi(-100000,-4); airybi(-inf,-4)
2241411040.437759489540248
+inf
airyaizero()
848
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
mpmath.airyaizero(k, derivative=0)
Gives the k-th zero of the Airy Ai-function, i.e. the k-th number ak ordered by magnitude
for which Ai(ak ) = 0.
Optionally, with derivative=1, the corresponding zero a0k of the derivative function, i.e.
Ai0 (a0k ) = 0, is computed.
Examples
Some values of ak :
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> airyaizero(1)
-2.338107410459767038489197
>>> airyaizero(2)
-4.087949444130970616636989
>>> airyaizero(3)
-5.520559828095551059129856
>>> airyaizero(1000)
-281.0315196125215528353364
Some values of a0k :
>>> airyaizero(1,1)
-1.018792971647471089017325
>>> airyaizero(2,1)
-3.248197582179836537875424
>>> airyaizero(3,1)
-4.820099211178735639400616
>>> airyaizero(1000,1)
-280.9378080358935070607097
Verication:
>>> chop(airyai(airyaizero(1)))
0.0
>>> chop(airyai(airyaizero(1,1),1))
0.0
airybizero()
mpmath.airybizero(k, derivative=0, complex=0)
With complex=False, gives the k-th real zero of the Airy Bi-function, i.e. the k-th number
bk ordered by magnitude for which Bi(bk ) = 0.
With complex=True, gives the k-th complex zero in the upper half plane k . Also the
conjugate k is a zero.
Optionally, with derivative=1, the corresponding zero b0k or k0 of the derivative function,
i.e. Bi0 (b0k ) = 0 or Bi0 (k0 ) = 0, is computed.
Examples
Some values of bk :
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> airybizero(1)
-1.17371322270912792491998
>>> airybizero(2)
-3.271093302836352715680228
5.15. Welcome to mpmaths documentation!
849
SymPy Documentation, Release 0.7.6
>>> airybizero(3)
-4.830737841662015932667709
>>> airybizero(1000)
-280.9378112034152401578834
Some values of bk :
>>> airybizero(1,1)
-2.294439682614123246622459
>>> airybizero(2,1)
-4.073155089071828215552369
>>> airybizero(3,1)
-5.512395729663599496259593
>>> airybizero(1000,1)
-281.0315164471118527161362
Some values of k :
>>> airybizero(1,complex=True)
(0.9775448867316206859469927 + 2.141290706038744575749139j)
>>> airybizero(2,complex=True)
(1.896775013895336346627217 + 3.627291764358919410440499j)
>>> airybizero(3,complex=True)
(2.633157739354946595708019 + 4.855468179979844983174628j)
>>> airybizero(1000,complex=True)
(140.4978560578493018899793 + 243.3907724215792121244867j)
Some values of k0 :
>>> airybizero(1,1,complex=True)
(0.2149470745374305676088329 + 1.100600143302797880647194j)
>>> airybizero(2,1,complex=True)
(1.458168309223507392028211 + 2.912249367458445419235083j)
>>> airybizero(3,1,complex=True)
(2.273760763013482299792362 + 4.254528549217097862167015j)
>>> airybizero(1000,1,complex=True)
(140.4509972835270559730423 + 243.3096175398562811896208j)
Verication:
>>>
0.0
>>>
0.0
>>>
>>>
0.0
>>>
0.0
chop(airybi(airybizero(1)))
chop(airybi(airybizero(1,1),1))
u = airybizero(1,complex=True)
chop(airybi(u))
chop(airybi(conj(u)))
The complex zeros (in the upper and lower half-planes respectively) asymptotically approach the rays z = R exp(i/3):
>>> arg(airybizero(1,complex=True))
1.142532510286334022305364
>>> arg(airybizero(1000,complex=True))
1.047271114786212061583917
>>> arg(airybizero(1000000,complex=True))
1.047197624741816183341355
850
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> pi/3
1.047197551196597746154214
scorergi()
mpmath.scorergi(z, **kwargs)
Evaluates the Scorer function
Gi(z) = Ai(z)
Bi(t)dt + Bi(z)
0
Ai(t)dt
z
which gives a particular solution to the inhomogeneous Airy dierential equation f 00 (z)
zf (z) = 1/. Another particular solution is given by the Scorer Hi-function (scorerhi()
(page 853)). The two functions are related as Gi(z) + Hi(z) = Bi(z).
Plots
# Scorer function Gi(x) and Gi(x) on the real line
plot([scorergi, diffun(scorergi)], [-10,10])
# Scorer function Gi(z) in the complex plane
cplot(scorergi, [-8,8], [-8,8], points=50000)
5.15. Welcome to mpmaths documentation!
851
SymPy Documentation, Release 0.7.6
Examples
Some values and limits:
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> scorergi(0); 1/(power(3,7/6)*gamma(2/3))
0.2049755424820002450503075
0.2049755424820002450503075
>>> diff(scorergi, 0); 1/(power(3,5/6)*gamma(1/3))
0.1494294524512754526382746
0.1494294524512754526382746
>>> scorergi(+inf); scorergi(-inf)
0.0
0.0
>>> scorergi(1)
0.2352184398104379375986902
>>> scorergi(-1)
-0.1166722172960152826494198
Evaluation for large arguments:
>>> scorergi(10)
0.03189600510067958798062034
>>> scorergi(100)
0.003183105228162961476590531
>>> scorergi(1000000)
0.0000003183098861837906721743873
>>> 1/(pi*1000000)
852
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
0.0000003183098861837906715377675
>>> scorergi(-1000)
-0.08358288400262780392338014
>>> scorergi(-100000)
0.02886866118619660226809581
>>> scorergi(50+10j)
(0.0061214102799778578790984 - 0.001224335676457532180747917j)
>>> scorergi(-50-10j)
(5.236047850352252236372551e+29 - 3.08254224233701381482228e+29j)
>>> scorergi(100000j)
(-8.806659285336231052679025e+6474077 + 8.684731303500835514850962e+6474077j)
Verifying the connection between Gi and Hi:
>>> z = 0.25
>>> scorergi(z) + scorerhi(z)
0.7287469039362150078694543
>>> airybi(z)
0.7287469039362150078694543
Verifying the dierential equation:
>>> for z in [-3.4, 0, 2.5, 1+2j]:
...
chop(diff(scorergi,z,2) - z*scorergi(z))
...
-0.3183098861837906715377675
-0.3183098861837906715377675
-0.3183098861837906715377675
-0.3183098861837906715377675
Verifying the integral representation:
>>> z = 0.5
>>> scorergi(z)
0.2447210432765581976910539
>>> Ai,Bi = airyai,airybi
>>> Bi(z)*(Ai(inf,-1)-Ai(z,-1)) + Ai(z)*(Bi(z,-1)-Bi(0,-1))
0.2447210432765581976910539
References
1.[DLMF] (page 1910) section 9.12: Scorer Functions
scorerhi()
mpmath.scorerhi(z, **kwargs)
Evaluates the second Scorer function
z
Hi(z) = Bi(z)
Ai(t)dt Ai(z)
Bi(t)dt
which gives a particular solution to the inhomogeneous Airy dierential equation f 00 (z)
zf (z) = 1/. See also scorergi() (page 851).
Plots
# Scorer function Hi(x) and Hi(x) on the real line
plot([scorerhi, diffun(scorerhi)], [-10,2], [0,2])
5.15. Welcome to mpmaths documentation!
853
SymPy Documentation, Release 0.7.6
# Scorer function Hi(z) in the complex plane
cplot(scorerhi, [-8,8], [-8,8], points=50000)
854
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
Some values and limits:
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> scorerhi(0); 2/(power(3,7/6)*gamma(2/3))
0.4099510849640004901006149
0.4099510849640004901006149
>>> diff(scorerhi,0); 2/(power(3,5/6)*gamma(1/3))
0.2988589049025509052765491
0.2988589049025509052765491
>>> scorerhi(+inf); scorerhi(-inf)
+inf
0.0
>>> scorerhi(1)
0.9722051551424333218376886
>>> scorerhi(-1)
0.2206696067929598945381098
Evaluation for large arguments:
>>> scorerhi(10)
455641153.5163291358991077
>>> scorerhi(100)
6.041223996670201399005265e+288
>>> scorerhi(1000000)
7.138269638197858094311122e+289529652
>>> scorerhi(-10)
5.15. Welcome to mpmaths documentation!
855
SymPy Documentation, Release 0.7.6
0.0317685352825022727415011
>>> scorerhi(-100)
0.003183092495767499864680483
>>> scorerhi(100j)
(-6.366197716545672122983857e-9 + 0.003183098861710582761688475j)
>>> scorerhi(50+50j)
(-5.322076267321435669290334e+63 + 1.478450291165243789749427e+65j)
>>> scorerhi(-1000-1000j)
(0.0001591549432510502796565538 - 0.000159154943091895334973109j)
Verifying the dierential equation:
>>> for z in [-3.4, 0, 2, 1+2j]:
...
chop(diff(scorerhi,z,2) - z*scorerhi(z))
...
0.3183098861837906715377675
0.3183098861837906715377675
0.3183098861837906715377675
0.3183098861837906715377675
Verifying the integral representation:
>>> z = 0.5
>>> scorerhi(z)
0.6095559998265972956089949
>>> Ai,Bi = airyai,airybi
>>> Bi(z)*(Ai(z,-1)-Ai(-inf,-1)) - Ai(z)*(Bi(z,-1)-Bi(-inf,-1))
0.6095559998265972956089949
Coulomb wave functions
coulombf()
mpmath.coulombf(l, eta, z)
Calculates the regular Coulomb wave function
Fl (, z) = Cl ()z l+1 eiz 1 F1 (l + 1 i, 2l + 2, 2iz)
where the normalization constant Cl () is as calculated by coulombc() (page 862). This
function solves the dierential equation
(
)
2 l(l + 1)
00
f (z) + 1
f (z) = 0.
z
z2
A second linearly independent solution is given by the irregular Coulomb wave function
Gl (, z) (see coulombg() (page 860)) and thus the general solution is f (z) = C1 Fl (, z) +
C2 Gl (, z) for arbitrary constants C1 , C2 . Physically, the Coulomb wave functions give the
radial solution to the Schrodinger equation for a point particle in a 1/z potential; z is
then the radius and l, are quantum numbers.
The Coulomb wave functions with real parameters are dened in Abramowitz & Stegun,
section 14. However, all parameters are permitted to be complex in this implementation
(see references).
Plots
856
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
# Regular Coulomb wave functions -- equivalent to figure 14.3 in A&S
F1 = lambda x: coulombf(0,0,x)
F2 = lambda x: coulombf(0,1,x)
F3 = lambda x: coulombf(0,5,x)
F4 = lambda x: coulombf(0,10,x)
F5 = lambda x: coulombf(0,x/2,x)
plot([F1,F2,F3,F4,F5], [0,25], [-1.2,1.6])
# Regular Coulomb wave function in the complex plane
cplot(lambda z: coulombf(1,1,z), points=50000)
5.15. Welcome to mpmaths documentation!
857
SymPy Documentation, Release 0.7.6
Examples
Evaluation is supported for arbitrary magnitudes of z:
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> coulombf(2, 1.5, 3.5)
0.4080998961088761187426445
>>> coulombf(-2, 1.5, 3.5)
0.7103040849492536747533465
>>> coulombf(2, 1.5, 1e-10)
4.143324917492256448770769e-33
>>> coulombf(2, 1.5, 1000)
0.4482623140325567050716179
>>> coulombf(2, 1.5, 10**10)
-0.066804196437694360046619
Verifying the dierential equation:
>>>
>>>
>>>
>>>
0.0
l, eta, z = 2, 3, mpf(2.75)
A, B = 1, 2
f = lambda z: A*coulombf(l,eta,z) + B*coulombg(l,eta,z)
chop(diff(f,z,2) + (1-2*eta/z - l*(l+1)/z**2)*f(z))
A Wronskian relation satised by the Coulomb wave functions:
>>> l = 2
>>> eta = 1.5
858
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> F = lambda z: coulombf(l,eta,z)
>>> G = lambda z: coulombg(l,eta,z)
>>> for z in [3.5, -1, 2+3j]:
...
chop(diff(F,z)*G(z) - F(z)*diff(G,z))
...
1.0
1.0
1.0
Another Wronskian relation:
>>> F = coulombf
>>> G = coulombg
>>> for z in [3.5, -1, 2+3j]:
...
chop(F(l-1,eta,z)*G(l,eta,z)-F(l,eta,z)*G(l-1,eta,z) - l/sqrt(l**2+eta**2))
...
0.0
0.0
0.0
An integral identity connecting the regular and irregular wave functions:
>>> l, eta, z = 4+j, 2-j, 5+2j
>>> coulombf(l,eta,z) + j*coulombg(l,eta,z)
(0.7997977752284033239714479 + 0.9294486669502295512503127j)
>>> g = lambda t: exp(-t)*t**(l-j*eta)*(t+2*j*z)**(l+j*eta)
>>> j*exp(-j*z)*z**(-l)/fac(2*l+1)/coulombc(l,eta)*quad(g, [0,inf])
(0.7997977752284033239714479 + 0.9294486669502295512503127j)
Some test case with complex parameters, taken from Michel [2]:
>>> mp.dps = 15
>>> coulombf(1+0.1j, 50+50j, 100.156)
(-1.02107292320897e+15 - 2.83675545731519e+15j)
>>> coulombg(1+0.1j, 50+50j, 100.156)
(2.83675545731519e+15 - 1.02107292320897e+15j)
>>> coulombf(1e-5j, 10+1e-5j, 0.1+1e-6j)
(4.30566371247811e-14 - 9.03347835361657e-19j)
>>> coulombg(1e-5j, 10+1e-5j, 0.1+1e-6j)
(778709182061.134 + 18418936.2660553j)
The following reproduces a table in Abramowitz & Stegun, at twice the precision:
>>> mp.dps = 10
>>> eta = 2; z = 5
>>> for l in [5, 4, 3, 2, 1, 0]:
...
print(%s %s %s % (l, coulombf(l,eta,z),
...
diff(lambda z: coulombf(l,eta,z), z)))
...
5 0.09079533488 0.1042553261
4 0.2148205331 0.2029591779
3 0.4313159311 0.320534053
2 0.7212774133 0.3952408216
1 0.9935056752 0.3708676452
0 1.143337392 0.2937960375
References
1.I.J. Thompson & A.R. Barnett, Coulomb and Bessel Functions of Complex Arguments and Order, J. Comp. Phys., vol 64, no. 2, June 1986.
5.15. Welcome to mpmaths documentation!
859
SymPy Documentation, Release 0.7.6
2.N. Michel, Precise Coulomb wave functions for a wide range of complex l, and z,
http://arxiv.org/abs/physics/0702051v1
coulombg()
mpmath.coulombg(l, eta, z)
Calculates the irregular Coulomb wave function
Gl (, z) =
Fl (, z) cos() Fl1 (, z)
sin()
where = l l1 (l + 1/2) and l () = (ln (1 + l + i) ln (1 + l i))/(2i).
See coulombf() (page 856) for additional information.
Plots
# Irregular Coulomb wave functions -- equivalent to figure 14.5 in A&S
F1 = lambda x: coulombg(0,0,x)
F2 = lambda x: coulombg(0,1,x)
F3 = lambda x: coulombg(0,5,x)
F4 = lambda x: coulombg(0,10,x)
F5 = lambda x: coulombg(0,x/2,x)
plot([F1,F2,F3,F4,F5], [0,30], [-2,2])
# Irregular Coulomb wave function in the complex plane
cplot(lambda z: coulombg(1,1,z), points=50000)
860
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
Evaluation is supported for arbitrary magnitudes of z:
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> coulombg(-2, 1.5, 3.5)
1.380011900612186346255524
>>> coulombg(2, 1.5, 3.5)
1.919153700722748795245926
>>> coulombg(-2, 1.5, 1e-10)
201126715824.7329115106793
>>> coulombg(-2, 1.5, 1000)
0.1802071520691149410425512
>>> coulombg(-2, 1.5, 10**10)
0.652103020061678070929794
The following reproduces a table in Abramowitz & Stegun, at twice the precision:
>>> mp.dps = 10
>>> eta = 2; z = 5
>>> for l in [1, 2, 3, 4, 5]:
...
print(%s %s %s % (l, coulombg(l,eta,z),
...
-diff(lambda z: coulombg(l,eta,z), z)))
...
1 1.08148276 0.6028279961
2 1.496877075 0.5661803178
3 2.048694714 0.7959909551
4 3.09408669 1.731802374
5.15. Welcome to mpmaths documentation!
861
SymPy Documentation, Release 0.7.6
5 5.629840456 4.549343289
Evaluation close to the singularity at z = 0:
>>> mp.dps = 15
>>> coulombg(0,10,1)
3088184933.67358
>>> coulombg(0,10,1e-10)
5554866000719.8
>>> coulombg(0,10,1e-100)
5554866221524.1
Evaluation with a half-integer value for l:
>>> coulombg(1.5, 1, 10)
0.852320038297334
coulombc()
mpmath.coulombc(l, eta)
Gives the normalizing Gamow constant for Coulomb wave functions,
Cl () = 2l exp (/2 + [ln (1 + l + i) + ln (1 + l i)]/2 ln (2l + 2)) ,
where the log gamma function with continuous imaginary part away from the negative
half axis (see loggamma() (page 783)) is implied.
This function is used internally for the calculation of Coulomb wave functions, and automatically cached to make multiple evaluations with xed l, fast.
Conuent U and Whittaker functions
hyperu()
mpmath.hyperu(a, b, z)
Gives the Tricomi conuent hypergeometric function U , also known as the Kummer or
conuent hypergeometric function of the second kind. This function gives a second linearly independent solution to the conuent hypergeometric dierential equation (the
rst is provided by 1 F1 see hyp1f1() (page 889)).
Examples
Evaluation for arbitrary complex arguments:
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> hyperu(2,3,4)
0.0625
>>> hyperu(0.25, 5, 1000)
0.1779949416140579573763523
>>> hyperu(0.25, 5, -1000)
(0.1256256609322773150118907 - 0.1256256609322773150118907j)
The U function may be singular at z = 0:
>>> hyperu(1.5, 2, 0)
+inf
>>> hyperu(1.5, -2, 0)
0.1719434921288400112603671
862
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Verifying the dierential equation:
>>> a, b = 1.5, 2
>>> f = lambda z: hyperu(a,b,z)
>>> for z in [-10, 3, 3+4j]:
...
chop(z*diff(f,z,2) + (b-z)*diff(f,z) - a*f(z))
...
0.0
0.0
0.0
An integral representation:
>>> a,b,z = 2, 3.5, 4.25
>>> hyperu(a,b,z)
0.06674960718150520648014567
>>> quad(lambda t: exp(-z*t)*t**(a-1)*(1+t)**(b-a-1),[0,inf]) / gamma(a)
0.06674960718150520648014567
[1] http://people.math.sfu.ca/cbm/aands/page 504.htm
whitm()
mpmath.whitm(k, m, z)
Evaluates the Whittaker function M (k, m, z), which gives a solution to the Whittaker differential equation
(
)
d2 f
1 k ( 41 m2 )
+ + +
f = 0.
dz 2
4 z
z2
A second solution is given by whitw() (page 864).
The Whittaker functions are dened in Abramowitz & Stegun, section 13.1. They are
alternate forms of the conuent hypergeometric functions 1 F1 and U :
M (k, m, z) = e 2 z z 2 +m 1 F1 ( 12 + m k, 1 + 2m, z)
1
W (k, m, z) = e 2 z z 2 +m U ( 12 + m k, 1 + 2m, z).
1
Examples
Evaluation for arbitrary real and complex arguments is supported:
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> whitm(1, 1, 1)
0.7302596799460411820509668
>>> whitm(1, 1, -1)
(0.0 - 1.417977827655098025684246j)
>>> whitm(j, j/2, 2+3j)
(3.245477713363581112736478 - 0.822879187542699127327782j)
>>> whitm(2, 3, 100000)
4.303985255686378497193063e+21707
Evaluation at zero:
>>> whitm(1,-1,0); whitm(1,-0.5,0); whitm(1,0,0)
+inf
nan
0.0
5.15. Welcome to mpmaths documentation!
863
SymPy Documentation, Release 0.7.6
We can verify that whitm() (page 863) numerically satises the dierential equation for
arbitrarily chosen values:
>>>
>>>
>>>
>>>
...
...
0.0
0.0
0.0
0.0
k =
m =
f =
for
mpf(0.25)
mpf(1.5)
lambda z: whitm(k,m,z)
z in [-1, 2.5, 3, 1+2j]:
chop(diff(f,z,2) + (-0.25 + k/z + (0.25-m**2)/z**2)*f(z))
An integral involving both whitm() (page 863) and whitw() (page 864), verifying evaluation along the real axis:
>>> quad(lambda x: exp(-x)*whitm(3,2,x)*whitw(1,-2,x), [0,inf])
3.438869842576800225207341
>>> 128/(21*sqrt(pi))
3.438869842576800225207341
whitw()
mpmath.whitw(k, m, z)
Evaluates the Whittaker function W (k, m, z), which gives a second solution to the Whittaker dierential equation. (See whitm() (page 863).)
Examples
Evaluation for arbitrary real and complex arguments is supported:
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> whitw(1, 1, 1)
1.19532063107581155661012
>>> whitw(1, 1, -1)
(-0.9424875979222187313924639 - 0.2607738054097702293308689j)
>>> whitw(j, j/2, 2+3j)
(0.1782899315111033879430369 - 0.01609578360403649340169406j)
>>> whitw(2, 3, 100000)
1.887705114889527446891274e-21705
>>> whitw(-1, -1, 100)
1.905250692824046162462058e-24
Evaluation at zero:
>>> for m in [-1, -0.5, 0, 0.5, 1]:
...
whitw(1, m, 0)
...
+inf
nan
0.0
nan
+inf
We can verify that whitw() (page 864) numerically satises the dierential equation for
arbitrarily chosen values:
864
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>>
>>>
>>>
>>>
...
...
0.0
0.0
0.0
0.0
k =
m =
f =
for
mpf(0.25)
mpf(1.5)
lambda z: whitw(k,m,z)
z in [-1, 2.5, 3, 1+2j]:
chop(diff(f,z,2) + (-0.25 + k/z + (0.25-m**2)/z**2)*f(z))
Parabolic cylinder functions
pcfd()
mpmath.pcfd(n, z, **kwargs)
Gives the parabolic cylinder function in Whittakers notation Dn (z) = U (n 1/2, z) (see
pcfu() (page 867)). It solves the dierential equation
(
)
1 1
y 00 + n + z 2 y = 0.
2 4
and can be represented in terms of Hermite polynomials (see hermite() (page 878)) as
(
)
2
z
Dn (z) = 2n/2 ez /4 Hn
.
2
Plots
# Parabolic cylinder function D_n(x) on the real line for n=0,1,2,3,4
d0 = lambda x: pcfd(0,x)
d1 = lambda x: pcfd(1,x)
d2 = lambda x: pcfd(2,x)
d3 = lambda x: pcfd(3,x)
d4 = lambda x: pcfd(4,x)
plot([d0,d1,d2,d3,d4],[-7,7])
5.15. Welcome to mpmaths documentation!
865
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> pcfd(0,0); pcfd(1,0); pcfd(2,0); pcfd(3,0)
1.0
0.0
-1.0
0.0
>>> pcfd(4,0); pcfd(-3,0)
3.0
0.6266570686577501256039413
>>> pcfd(1/2, 2+3j)
(-5.363331161232920734849056 - 3.858877821790010714163487j)
>>> pcfd(2, -10)
1.374906442631438038871515e-9
Verifying the dierential equation:
>>>
>>>
>>>
>>>
0.0
n = mpf(2.5)
y = lambda z: pcfd(n,z)
z = 1.75
chop(diff(y,z,2) + (n+0.5-0.25*z**2)*y(z))
Rational Taylor series expansion when n is an integer:
866
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> taylor(lambda z: pcfd(5,z), 0, 7)
[0.0, 15.0, 0.0, -13.75, 0.0, 3.96875, 0.0, -0.6015625]
pcfu()
mpmath.pcfu(a, z, **kwargs)
Gives the parabolic cylinder function U (a, z), which may be dened for <(z) > 0 in terms
of the conuent U-function (see hyperu() (page 862)) by
(
)
a
1 2
1
a 1 1 1 2
U (a, z) = 2 4 2 e 4 z U
+ , , z
2 4 2 2
or, for arbitrary z,
(
)
)
(
1 2
e 4 z U (a, z) = U (a, 0) 1 F1 a2 + 14 ; 12 ; 12 z 2 + U 0 (a, 0)z 1 F1 a2 + 34 ; 23 ; 12 z 2 .
Examples
Connection to other functions:
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> z = mpf(3)
>>> pcfu(0.5,z)
0.03210358129311151450551963
>>> sqrt(pi/2)*exp(z**2/4)*erfc(z/sqrt(2))
0.03210358129311151450551963
>>> pcfu(0.5,-z)
23.75012332835297233711255
>>> sqrt(pi/2)*exp(z**2/4)*erfc(-z/sqrt(2))
23.75012332835297233711255
>>> pcfu(0.5,-z)
23.75012332835297233711255
>>> sqrt(pi/2)*exp(z**2/4)*erfc(-z/sqrt(2))
23.75012332835297233711255
pcfv()
mpmath.pcfv(a, z, **kwargs)
Gives the parabolic cylinder function V (a, z), which can be represented in terms of pcfu()
(page 867) as
V (a, z) =
(a + 12 )(U (a, z) sin(a)U (a, z)
.
Examples
Wronskian relation between U and V :
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> a, z = 2, 3
>>> pcfu(a,z)*diff(pcfv,(a,z),(0,1))-diff(pcfu,(a,z),(0,1))*pcfv(a,z)
0.7978845608028653558798921
>>> sqrt(2/pi)
0.7978845608028653558798921
>>> a, z = 2.5, 3
>>> pcfu(a,z)*diff(pcfv,(a,z),(0,1))-diff(pcfu,(a,z),(0,1))*pcfv(a,z)
5.15. Welcome to mpmaths documentation!
867
SymPy Documentation, Release 0.7.6
0.7978845608028653558798921
>>> a, z = 0.25, -1
>>> pcfu(a,z)*diff(pcfv,(a,z),(0,1))-diff(pcfu,(a,z),(0,1))*pcfv(a,z)
0.7978845608028653558798921
>>> a, z = 2+1j, 2+3j
>>> chop(pcfu(a,z)*diff(pcfv,(a,z),(0,1))-diff(pcfu,(a,z),(0,1))*pcfv(a,z))
0.7978845608028653558798921
pcfw()
mpmath.pcfw(a, z, **kwargs)
Gives the parabolic cylinder function W (a, z) dened in (DLMF 12.14).
Examples
Value at the origin:
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> a = mpf(0.25)
>>> pcfw(a,0)
0.9722833245718180765617104
>>> power(2,-0.75)*sqrt(abs(gamma(0.25+0.5j*a)/gamma(0.75+0.5j*a)))
0.9722833245718180765617104
>>> diff(pcfw,(a,0),(0,1))
-0.5142533944210078966003624
>>> -power(2,-0.25)*sqrt(abs(gamma(0.75+0.5j*a)/gamma(0.25+0.5j*a)))
-0.5142533944210078966003624
Orthogonal polynomials
An orthogonal polynomial sequence is a sequence of polynomials P0 (x), P1 (x), . . . of degree
0, 1, . . ., which are mutually orthogonal in the sense that
{
cn 6= 0 if m = n
Pn (x)Pm (x)w(x)dx =
0
if m 6= n
S
where S is some domain (e.g. an interval [a, b] R) and w(x) is a xed weight function. A
sequence of orthogonal polynomials is determined completely by w, S, and a normalization
convention (e.g. cn = 1). Applications of orthogonal polynomials include function approximation and solution of dierential equations.
Orthogonal polynomials are sometimes dened using the dierential equations they satisfy
(as functions of x) or the recurrence relations they satisfy with respect to the order n. Other
ways of dening orthogonal polynomials include dierentiation formulas and generating functions. The standard orthogonal polynomials can also be represented as hypergeometric series
(see Hypergeometric functions (page 888)), more specically using the Gauss hypergeometric function 2 F1 in most cases. The following functions are generally implemented using
hypergeometric functions since this is computationally ecient and easily generalizes.
For more information, see the Wikipedia article on orthogonal polynomials.
Legendre functions
868
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
legendre()
mpmath.legendre(n, x)
legendre(n, x) evaluates the Legendre polynomial Pn (x). The Legendre polynomials
are given by the formula
Pn (x) =
1 dn 2
(x 1)n .
2n n! dxn
Alternatively, they can be computed recursively using
P0 (x) = 1
P1 (x) = x
(n + 1)Pn+1 (x) = (2n + 1)xPn (x) nPn1 (x).
A third denition is in terms of the hypergeometric function 2 F1 , whereby they can be
generalized to arbitrary n:
(
)
1x
Pn (x) = 2 F1 n, n + 1, 1,
2
Plots
# Legendre polynomials P_n(x) on [-1,1] for n=0,1,2,3,4
f0 = lambda x: legendre(0,x)
f1 = lambda x: legendre(1,x)
f2 = lambda x: legendre(2,x)
f3 = lambda x: legendre(3,x)
f4 = lambda x: legendre(4,x)
plot([f0,f1,f2,f3,f4],[-1,1])
5.15. Welcome to mpmaths documentation!
869
SymPy Documentation, Release 0.7.6
Basic evaluation
The Legendre polynomials assume xed values at the points x = 1 and x = 1:
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = True
>>> nprint([legendre(n, 1) for n in range(6)])
[1.0, 1.0, 1.0, 1.0, 1.0, 1.0]
>>> nprint([legendre(n, -1) for n in range(6)])
[1.0, -1.0, 1.0, -1.0, 1.0, -1.0]
The coecients of Legendre polynomials can be recovered using degree-n Taylor expansion:
>>> for n in range(5):
...
nprint(chop(taylor(lambda x: legendre(n, x), 0, n)))
...
[1.0]
[0.0, 1.0]
[-0.5, 0.0, 1.5]
[0.0, -1.5, 0.0, 2.5]
[0.375, 0.0, -3.75, 0.0, 4.375]
The roots of Legendre polynomials are located symmetrically on the interval [1, 1]:
>>> for n in range(5):
...
nprint(polyroots(taylor(lambda x: legendre(n, x), 0, n)[::-1]))
...
[]
[0.0]
[-0.57735, 0.57735]
[-0.774597, 0.0, 0.774597]
[-0.861136, -0.339981, 0.339981, 0.861136]
An example of an evaluation for arbitrary n:
>>> legendre(0.75, 2+4j)
(1.94952805264875 + 2.1071073099422j)
Orthogonality
The Legendre polynomials are orthogonal on [1, 1] with respect to the trivial weight
w(x) = 1. That is, Pm (x)Pn (x) integrates to zero if m 6= n and to 2/(2n + 1) if m = n:
>>> m, n = 3, 4
>>> quad(lambda x: legendre(m,x)*legendre(n,x), [-1, 1])
0.0
>>> m, n = 4, 4
>>> quad(lambda x: legendre(m,x)*legendre(n,x), [-1, 1])
0.222222222222222
Dierential equation
The Legendre polynomials satisfy the dierential equation
((1 x2 )y 0 )0 + n(n + 1)y 0 = 0.
We can verify this numerically:
>>> n = 3.6
>>> x = 0.73
>>> P = legendre
870
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> A = diff(lambda t: (1-t**2)*diff(lambda u: P(n,u), t), x)
>>> B = n*(n+1)*P(n,x)
>>> nprint(A+B,1)
9.0e-16
legenp()
mpmath.legenp(n, m, z, type=2)
Calculates the (associated) Legendre function of the rst kind of degree n and order m,
Pnm (z). Taking m = 0 gives the ordinary Legendre function of the rst kind, Pn (z). The
parameters may be complex numbers.
In terms of the Gauss hypergeometric function, the (associated) Legendre function is
dened as
)
(
1
(1 + z)m/2
1z
Pnm (z) =
F
n,
n
+
1,
1
m,
.
2 1
(1 m) (1 z)m/2
2
With type=3 instead of type=2, the alternative denition
(
)
1
(z + 1)m/2
1z
m
Pn (z) =
.
2 F1 n, n + 1, 1 m,
(1 m) (z 1)m/2
2
is used. These functions correspond respectively to LegendreP[n,m,2,z] and LegendreP[n,m,3,z] in Mathematica.
The general solution of the (associated) Legendre dierential equation
(
)
m2
2 00
0
(1 z )f (z) 2zf (z) + n(n + 1)
f (z) = 0
1 z2
m
is given by C1 Pnm (z) + C2 Qm
n (z) for arbitrary constants C1 , C2 , where Qn (z) is a Legendre
function of the second kind as implemented by legenq() (page 872).
Examples
Evaluation for arbitrary parameters and arguments:
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> legenp(2, 0, 10); legendre(2, 10)
149.5
149.5
>>> legenp(-2, 0.5, 2.5)
(1.972260393822275434196053 - 1.972260393822275434196053j)
>>> legenp(2+3j, 1-j, -0.5+4j)
(-3.335677248386698208736542 - 5.663270217461022307645625j)
>>> chop(legenp(3, 2, -1.5, type=2))
28.125
>>> chop(legenp(3, 2, -1.5, type=3))
-28.125
Verifying the associated Legendre dierential equation:
>>>
>>>
>>>
>>>
...
>>>
n, m = 2, -0.5
C1, C2 = 1, -3
f = lambda z: C1*legenp(n,m,z) + C2*legenq(n,m,z)
deq = lambda z: (1-z**2)*diff(f,z,2) - 2*z*diff(f,z) + \
(n*(n+1)-m**2/(1-z**2))*f(z)
for z in [0, 2, -1.5, 0.5+2j]:
5.15. Welcome to mpmaths documentation!
871
SymPy Documentation, Release 0.7.6
...
...
0.0
0.0
0.0
0.0
chop(deq(mpmathify(z)))
legenq()
mpmath.legenq(n, m, z, type=2)
Calculates the (associated) Legendre function of the second kind of degree n and order
m, Qm
n (z). Taking m = 0 gives the ordinary Legendre function of the second kind, Qn (z).
The parameters may complex numbers.
The Legendre functions of the second kind give a second set of solutions to the (associated) Legendre dierential equation. (See legenp() (page 871).) Unlike the Legendre
functions of the rst kind, they are not polynomials of z for integer n, m but rational or
logarithmic functions with poles at z = 1.
There are various ways to dene Legendre functions of the second kind, giving rise to
dierent complex structure. A version can be selected using the type keyword argument.
The type=2 and type=3 functions are given respectively by
(
)
(1 + m + n) m
m
Qm
(z)
=
cos(m)P
(z)
P
(z)
n
n
2 sin(m)
(1 m + n) n
)
(
(1 + m + n) m
m
im
m
(z)
(z)
(z)
=
P
P
Q
e
n
n
2 sin(m)
(1 m + n) n
where P and P are the type=2 and type=3 Legendre functions of the rst kind. The
formulas above should be understood as limits when m is an integer.
These functions correspond to LegendreQ[n,m,2,z] (or LegendreQ[n,m,z]) and LegendreQ[n,m,3,z] in Mathematica. The type=3 function is essentially the same as the
function dened in Abramowitz & Stegun (eq. 8.1.3) but with (z + 1)m/2 (z 1)m/2 instead
of (z 2 1)m/2 , giving slightly dierent branches.
Examples
Evaluation for arbitrary parameters and arguments:
>>> from sympy.mpmath import
>>> mp.dps = 25; mp.pretty =
>>> legenq(2, 0, 0.5)
-0.8186632680417568557122028
>>> legenq(-1.5, -2, 2.5)
(0.6655964618250228714288277
>>> legenq(2-j, 3+4j, -6+5j)
(-10001.95256487468541686564
*
True
+ 0.3937692045497259717762649j)
- 6011.691337610097577791134j)
Dierent versions of the function:
>>> legenq(2, 1, 0.5)
0.7298060598018049369381857
>>> legenq(2, 1, 1.5)
(-7.902916572420817192300921 + 0.1998650072605976600724502j)
>>> legenq(2, 1, 0.5, type=3)
(2.040524284763495081918338 - 0.7298060598018049369381857j)
>>> chop(legenq(2, 1, 1.5, type=3))
-0.1998650072605976600724502
872
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Chebyshev polynomials
chebyt()
mpmath.chebyt(n, x)
chebyt(n, x) evaluates the Chebyshev polynomial of the rst kind Tn (x), dened by the
identity
Tn (cos x) = cos(nx).
The Chebyshev polynomials of the rst kind are a special case of the Jacobi polynomials,
and by extension of the hypergeometric function 2 F1 . They can thus also be evaluated
for nonintegral n.
Plots
# Chebyshev polynomials T_n(x) on [-1,1] for n=0,1,2,3,4
f0 = lambda x: chebyt(0,x)
f1 = lambda x: chebyt(1,x)
f2 = lambda x: chebyt(2,x)
f3 = lambda x: chebyt(3,x)
f4 = lambda x: chebyt(4,x)
plot([f0,f1,f2,f3,f4],[-1,1])
Basic evaluation
The coecients of the n-th polynomial can be recovered using using degree-n Taylor
expansion:
5.15. Welcome to mpmaths documentation!
873
SymPy Documentation, Release 0.7.6
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = True
>>> for n in range(5):
...
nprint(chop(taylor(lambda x: chebyt(n, x), 0, n)))
...
[1.0]
[0.0, 1.0]
[-1.0, 0.0, 2.0]
[0.0, -3.0, 0.0, 4.0]
[1.0, 0.0, -8.0, 0.0, 8.0]
Orthogonality
The Chebyshev polynomials of the rstkind are orthogonal on the interval [1, 1] with
respect to the weight function w(x) = 1/ 1 x2 :
>>> f = lambda x: chebyt(m,x)*chebyt(n,x)/sqrt(1-x**2)
>>> m, n = 3, 4
>>> nprint(quad(f, [-1, 1]),1)
0.0
>>> m, n = 4, 4
>>> quad(f, [-1, 1])
1.57079632596448
chebyu()
mpmath.chebyu(n, x)
chebyu(n, x) evaluates the Chebyshev polynomial of the second kind Un (x), dened by
the identity
Un (cos x) =
sin((n + 1)x)
.
sin(x)
The Chebyshev polynomials of the second kind are a special case of the Jacobi polynomials, and by extension of the hypergeometric function 2 F1 . They can thus also be
evaluated for nonintegral n.
Plots
# Chebyshev polynomials U_n(x) on [-1,1] for n=0,1,2,3,4
f0 = lambda x: chebyu(0,x)
f1 = lambda x: chebyu(1,x)
f2 = lambda x: chebyu(2,x)
f3 = lambda x: chebyu(3,x)
f4 = lambda x: chebyu(4,x)
plot([f0,f1,f2,f3,f4],[-1,1])
874
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Basic evaluation
The coecients of the n-th polynomial can be recovered using using degree-n Taylor
expansion:
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = True
>>> for n in range(5):
...
nprint(chop(taylor(lambda x: chebyu(n, x), 0, n)))
...
[1.0]
[0.0, 2.0]
[-1.0, 0.0, 4.0]
[0.0, -4.0, 0.0, 8.0]
[1.0, 0.0, -12.0, 0.0, 16.0]
Orthogonality
The Chebyshev polynomials of the second
kind are orthogonal on the interval [1, 1] with
respect to the weight function w(x) = 1 x2 :
>>> f = lambda x: chebyu(m,x)*chebyu(n,x)*sqrt(1-x**2)
>>> m, n = 3, 4
>>> quad(f, [-1, 1])
0.0
>>> m, n = 4, 4
>>> quad(f, [-1, 1])
1.5707963267949
5.15. Welcome to mpmaths documentation!
875
SymPy Documentation, Release 0.7.6
Jacobi polynomials
jacobi()
mpmath.jacobi(n, a, b, z)
(a,b)
jacobi(n, a, b, x) evaluates the Jacobi polynomial Pn (x). The Jacobi polynomials
are a special case of the hypergeometric function 2 F1 given by:
(
)
(
)
n+a
1x
(a,b)
Pn (x) =
.
2 F1 n, 1 + a + b + n, a + 1,
n
2
Note that this denition generalizes to nonintegral values of n. When n is an integer, the
hypergeometric series terminates after a nite number of terms, giving a polynomial in
x.
Evaluation of Jacobi polynomials
(
)
(a,b)
A special evaluation is Pn (1) = n+a
n :
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = True
>>> jacobi(4, 0.5, 0.25, 1)
2.4609375
>>> binomial(4+0.5, 4)
2.4609375
A Jacobi polynomial of degree n is equal to its Taylor polynomial of degree n. The explicit coecients of Jacobi polynomials can therefore be recovered easily using taylor()
(page 1015):
>>> for n in range(5):
...
nprint(taylor(lambda x: jacobi(n,1,2,x), 0, n))
...
[1.0]
[-0.5, 2.5]
[-0.75, -1.5, 5.25]
[0.5, -3.5, -3.5, 10.5]
[0.625, 2.5, -11.25, -7.5, 20.625]
For nonintegral n, the Jacobi polynomial is no longer a polynomial:
>>> nprint(taylor(lambda x: jacobi(0.5,1,2,x), 0, 4))
[0.309983, 1.84119, -1.26933, 1.26699, -1.34808]
Orthogonality
The Jacobi polynomials are orthogonal on the interval [1, 1] with respect to the weight
(a,b)
(a,b)
function w(x) = (1 x)a (1 + x)b . That is, w(x)Pn (x)Pm (x) integrates to zero if m 6= n and
to a nonzero number if m = n.
The orthogonality is easy to verify using numerical quadrature:
>>>
>>>
>>>
>>>
>>>
>>>
0.0
>>>
876
P = jacobi
f = lambda x: (1-x)**a * (1+x)**b * P(m,a,b,x) * P(n,a,b,x)
a = 2
b = 3
m, n = 3, 4
chop(quad(f, [-1, 1]), 1)
m, n = 4, 4
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> quad(f, [-1, 1])
1.9047619047619
Dierential equation
The Jacobi polynomials are solutions of the dierential equation
(1 x2 )y 00 + (b a (a + b + 2)x)y 0 + n(n + a + b + 1)y = 0.
We can verify that jacobi() (page 876) approximately satises this equation:
>>> from sympy.mpmath import *
>>> mp.dps = 15
>>> a = 2.5
>>> b = 4
>>> n = 3
>>> y = lambda x: jacobi(n,a,b,x)
>>> x = pi
>>> A0 = n*(n+a+b+1)*y(x)
>>> A1 = (b-a-(a+b+2)*x)*diff(y,x)
>>> A2 = (1-x**2)*diff(y,x,2)
>>> nprint(A2 + A1 + A0, 1)
4.0e-12
The dierence of order 1012 is as close to zero as it could be at 15-digit working precision, since the terms are large:
>>> A0, A1, A2
(26560.2328981879, -21503.7641037294, -5056.46879445852)
Gegenbauer polynomials
gegenbauer()
mpmath.gegenbauer(n, a, z)
Evaluates the Gegenbauer polynomial, or ultraspherical polynomial,
(
)
(
)
n + 2a 1
1 1
(a)
Cn (z) =
2 F1 n, n + 2a; a + ; (1 z) .
n
2 2
When n is a nonnegative integer, this formula gives a polynomial in z of degree n, but
all parameters are permitted to be complex numbers. With a = 1/2, the Gegenbauer
polynomial reduces to a Legendre polynomial.
Examples
Evaluation for arbitrary arguments:
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> gegenbauer(3, 0.5, -10)
-2485.0
>>> gegenbauer(1000, 10, 100)
3.012757178975667428359374e+2322
>>> gegenbauer(2+3j, -0.75, -1000j)
(-5038991.358609026523401901 + 9414549.285447104177860806j)
Evaluation at negative integer orders:
5.15. Welcome to mpmaths documentation!
877
SymPy Documentation, Release 0.7.6
>>> gegenbauer(-4,
-1.0
>>> gegenbauer(-4,
0.0
>>> gegenbauer(-4,
0.0
>>> gegenbauer(-7,
8989.0
2, 1.75)
3, 1.75)
2j, 1.75)
0.5, 3)
The Gegenbauer polynomials solve the dierential equation:
>>> n, a = 4.5, 1+2j
>>> f = lambda z: gegenbauer(n, a, z)
>>> for z in [0, 0.75, -0.5j]:
...
chop((1-z**2)*diff(f,z,2) - (2*a+1)*z*diff(f,z) + n*(n+2*a)*f(z))
...
0.0
0.0
0.0
The Gegenbauer polynomials have generating function (1 2zt + t2 )a :
>>> a, z = 2.5, 1
>>> taylor(lambda t: (1-2*z*t+t**2)**(-a), 0, 3)
[1.0, 5.0, 15.0, 35.0]
>>> [gegenbauer(n,a,z) for n in range(4)]
[1.0, 5.0, 15.0, 35.0]
The Gegenbauer polynomials are orthogonal on [1, 1] with respect to the weight (1
1
z 2 )a 2 :
>>>
>>>
>>>
>>>
0.0
a, n, m = 2.5, 4, 5
Cn = lambda z: gegenbauer(n, a, z, zeroprec=1000)
Cm = lambda z: gegenbauer(m, a, z, zeroprec=1000)
chop(quad(lambda z: Cn(z)*Cm(z)*(1-z**2)*(a-0.5), [-1, 1]))
Hermite polynomials
hermite()
mpmath.hermite(n, z)
Evaluates the Hermite polynomial Hn (z), which may be dened using the recurrence
H0 (z) = 1
H1 (z) = 2z
Hn+1 = 2zHn (z) 2nHn1 (z).
The Hermite polynomials are orthogonal on (, ) with respect to the weight ez .
More generally, allowing arbitrary complex values of n, the Hermite function Hn (z) is
dened as
)
(
1
n 1n
, 2
Hn (z) = (2z)n 2 F0 ,
2
2
z
2
878
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
for <z > 0, or generally
Hn (z) = 2
(
)
(
))
n 1 2
2z
1n 3 2
(
) 1 F1 , , z ( n ) 1 F1
, ,z
.
2 2
2
2
1n
2
2
1
Plots
# Hermite polynomials H_n(x) on the real line for n=0,1,2,3,4
f0 = lambda x: hermite(0,x)
f1 = lambda x: hermite(1,x)
f2 = lambda x: hermite(2,x)
f3 = lambda x: hermite(3,x)
f4 = lambda x: hermite(4,x)
plot([f0,f1,f2,f3,f4],[-2,2],[-25,25])
Examples
Evaluation for arbitrary arguments:
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> hermite(0, 10)
1.0
>>> hermite(1, 10); hermite(2, 10)
20.0
398.0
>>> hermite(10000, 2)
4.950440066552087387515653e+19334
>>> hermite(3, -10**8)
5.15. Welcome to mpmaths documentation!
879
SymPy Documentation, Release 0.7.6
-7999999999999998800000000.0
>>> hermite(-3, -10**8)
1.675159751729877682920301e+4342944819032534
>>> hermite(2+3j, -1+2j)
(-0.07652130602993513389421901 - 0.1084662449961914580276007j)
Coecients of the rst few Hermite polynomials are:
>>> for n in range(7):
...
chop(taylor(lambda z: hermite(n, z), 0, n))
...
[1.0]
[0.0, 2.0]
[-2.0, 0.0, 4.0]
[0.0, -12.0, 0.0, 8.0]
[12.0, 0.0, -48.0, 0.0, 16.0]
[0.0, 120.0, 0.0, -160.0, 0.0, 32.0]
[-120.0, 0.0, 720.0, 0.0, -480.0, 0.0, 64.0]
Values at z = 0:
>>> for n in range(-5, 9):
...
hermite(n, 0)
...
0.02769459142039868792653387
0.08333333333333333333333333
0.2215567313631895034122709
0.5
0.8862269254527580136490837
1.0
0.0
-2.0
0.0
12.0
0.0
-120.0
0.0
1680.0
Hermite functions satisfy the dierential equation:
>>>
>>>
>>>
>>>
0.0
n = 4
f = lambda z: hermite(n, z)
z = 1.5
chop(diff(f,z,2) - 2*z*diff(f,z) + 2*n*f(z))
Verifying orthogonality:
>>> chop(quad(lambda t: hermite(2,t)*hermite(4,t)*exp(-t**2), [-inf,inf]))
0.0
Laguerre polynomials
laguerre()
880
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
mpmath.laguerre(n, a, z)
Gives the generalized (associated) Laguerre polynomial, dened by
Lan (z) =
(n + b + 1)
1 F1 (n, a + 1, z).
(b + 1)(n + 1)
With a = 0 and n a nonnegative integer, this reduces to an ordinary Laguerre polynomial,
the sequence of which begins L0 (z) = 1, L1 (z) = 1 z, L2 (z) = z 2 2z + 1, . . ..
The Laguerre polynomials are orthogonal with respect to the weight z a ez on [0, ).
Plots
# Hermite polynomials L_n(x) on the real line for n=0,1,2,3,4
f0 = lambda x: laguerre(0,0,x)
f1 = lambda x: laguerre(1,0,x)
f2 = lambda x: laguerre(2,0,x)
f3 = lambda x: laguerre(3,0,x)
f4 = lambda x: laguerre(4,0,x)
plot([f0,f1,f2,f3,f4],[0,10],[-10,10])
Examples
Evaluation for arbitrary arguments:
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> laguerre(5, 0, 0.25)
0.03726399739583333333333333
>>> laguerre(1+j, 0.5, 2+3j)
5.15. Welcome to mpmaths documentation!
881
SymPy Documentation, Release 0.7.6
(4.474921610704496808379097 - 11.02058050372068958069241j)
>>> laguerre(2, 0, 10000)
49980001.0
>>> laguerre(2.5, 0, 10000)
-9.327764910194842158583189e+4328
The rst few Laguerre polynomials, normalized to have integer coecients:
>>> for n in range(7):
...
chop(taylor(lambda z: fac(n)*laguerre(n, 0, z), 0, n))
...
[1.0]
[1.0, -1.0]
[2.0, -4.0, 1.0]
[6.0, -18.0, 9.0, -1.0]
[24.0, -96.0, 72.0, -16.0, 1.0]
[120.0, -600.0, 600.0, -200.0, 25.0, -1.0]
[720.0, -4320.0, 5400.0, -2400.0, 450.0, -36.0, 1.0]
Verifying orthogonality:
>>>
>>>
>>>
>>>
0.0
Lm = lambda t: laguerre(m,a,t)
Ln = lambda t: laguerre(n,a,t)
a, n, m = 2.5, 2, 3
chop(quad(lambda t: exp(-t)*t**a*Lm(t)*Ln(t), [0,inf]))
Spherical harmonics
spherharm()
mpmath.spherharm(l, m, theta, phi)
Evaluates the spherical harmonic Ylm (, ),
2l + 1 (l m)! m
m
Yl (, ) =
P (cos )eim
4 (l + m)! l
where Plm is an associated Legendre function (see legenp() (page 871)).
Here [0, ] denotes the polar coordinate (ranging from the north pole to the south
pole) and [0, 2] denotes the azimuthal coordinate on a sphere. Care should be used
since many dierent conventions for spherical coordinate variables are used.
Usually spherical harmonics are considered for l N, m Z, |m| l. More generally,
l, m, , are permitted to be complex numbers.
Note: spherharm() (page 882) returns a complex number, even the value is purely real.
Plots
# Real part of spherical harmonic Y_(4,0)(theta,phi)
def Y(l,m):
def g(theta,phi):
R = abs(fp.re(fp.spherharm(l,m,theta,phi)))
x = R*fp.cos(phi)*fp.sin(theta)
y = R*fp.sin(phi)*fp.sin(theta)
z = R*fp.cos(theta)
882
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
return [x,y,z]
return g
fp.splot(Y(4,0), [0,fp.pi], [0,2*fp.pi], points=300)
# fp.splot(Y(4,0), [0,fp.pi], [0,2*fp.pi], points=300)
# fp.splot(Y(4,1), [0,fp.pi], [0,2*fp.pi], points=300)
# fp.splot(Y(4,2), [0,fp.pi], [0,2*fp.pi], points=300)
# fp.splot(Y(4,3), [0,fp.pi], [0,2*fp.pi], points=300)
Y4,0 :
Y4,1 :
5.15. Welcome to mpmaths documentation!
883
SymPy Documentation, Release 0.7.6
Y4,2 :
884
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Y4,3 :
5.15. Welcome to mpmaths documentation!
885
SymPy Documentation, Release 0.7.6
Y4,4 :
886
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
Some low-order spherical harmonics with reference values:
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> theta = pi/4
>>> phi = pi/3
>>> spherharm(0,0,theta,phi); 0.5*sqrt(1/pi)*expj(0)
(0.2820947917738781434740397 + 0.0j)
(0.2820947917738781434740397 + 0.0j)
>>> spherharm(1,-1,theta,phi); 0.5*sqrt(3/(2*pi))*expj(-phi)*sin(theta)
(0.1221506279757299803965962 - 0.2115710938304086076055298j)
(0.1221506279757299803965962 - 0.2115710938304086076055298j)
>>> spherharm(1,0,theta,phi); 0.5*sqrt(3/pi)*cos(theta)*expj(0)
(0.3454941494713354792652446 + 0.0j)
(0.3454941494713354792652446 + 0.0j)
>>> spherharm(1,1,theta,phi); -0.5*sqrt(3/(2*pi))*expj(phi)*sin(theta)
(-0.1221506279757299803965962 - 0.2115710938304086076055298j)
(-0.1221506279757299803965962 - 0.2115710938304086076055298j)
With the normalization convention used, the spherical harmonics are orthonormal on the
unit sphere:
>>>
>>>
>>>
>>>
sphere = [0,pi],
dS = lambda t,p:
Y1 = lambda t,p:
Y2 = lambda t,p:
[0,2*pi]
fp.sin(t)
# differential element
fp.spherharm(l1,m1,t,p)
fp.conj(fp.spherharm(l2,m2,t,p))
5.15. Welcome to mpmaths documentation!
887
SymPy Documentation, Release 0.7.6
>>> l1 = l2 = 3; m1 = m2 = 2
>>> print(fp.quad(lambda t,p: Y1(t,p)*Y2(t,p)*dS(t,p), *sphere))
(1+0j)
>>> m2 = 1
# m1 != m2
>>> print(fp.chop(fp.quad(lambda t,p: Y1(t,p)*Y2(t,p)*dS(t,p), *sphere)))
0.0
Evaluation is accurate for large orders:
>>> spherharm(1000,750,0.5,0.25)
(3.776445785304252879026585e-102 - 5.82441278771834794493484e-102j)
Evaluation works with complex parameter values:
>>> spherharm(1+j, 2j, 2+3j, -0.5j)
(64.44922331113759992154992 + 1981.693919841408089681743j)
Hypergeometric functions
The functions listed in Exponential integrals and error functions (page 794), Bessel functions and related functions (page 809) and Orthogonal polynomials (page 868), and many
other functions as well, are merely particular instances of the generalized hypergeometric
function p Fq . The functions listed in the following section enable ecient direct evaluation
of the underlying hypergeometric series, as well as linear combinations, limits with respect
to parameters, and analytic continuations thereof. Extensions to twodimensional series are
also provided. See also the basic or q-analog of the hypergeometric series in q-functions
(page 968).
For convenience, most of the hypergeometric series of low order are provided as standalone
functions. They can equivalently be evaluated using hyper() (page 895). As will be demonstrated in the respective docstrings, all the hyp#f# functions implement analytic continuations
and/or asymptotic expansions with respect to the argument z, thereby permitting evaluation
for z anywhere in the complex plane. Functions of higher degree can be computed via hyper()
(page 895), but generally only in rapidly convergent instances.
Most hypergeometric and hypergeometric-derived functions accept optional keyword arguments to specify options for hypercomb() or hyper(). Some useful options are maxprec,
maxterms, zeroprec, accurate small, hmag, force series, asymp tol and eliminate. These options give control over what to do in case of slow convergence, extreme loss of accuracy
or evaluation at zeros (these two cases cannot generally be distinguished from each other
automatically), and singular parameter combinations.
Common hypergeometric series
hyp0f1()
mpmath.hyp0f1(a, z)
Gives the hypergeometric function 0 F1 , sometimes known as the conuent limit function,
dened as
0 F1 (a, z)
1 zk
.
(a)k k!
k=0
This function satises the dierential equation zf 00 (z) + af 0 (z) = f (z), and is related to the
Bessel function of the rst kind (see besselj() (page 810)).
888
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
hyp0f1(a,z) is equivalent to hyper([],[a],z); see documentation for hyper()
(page 895) for more information.
Examples
Evaluation for arbitrary arguments:
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> hyp0f1(2, 0.25)
1.130318207984970054415392
>>> hyp0f1((1,2), 1234567)
6.27287187546220705604627e+964
>>> hyp0f1(3+4j, 1000000j)
(3.905169561300910030267132e+606 + 3.807708544441684513934213e+606j)
Evaluation is supported for arbitrarily large values of z, using asymptotic expansions:
>>> hyp0f1(1, 10**50)
2.131705322874965310390701e+8685889638065036553022565
>>> hyp0f1(1, -10**50)
1.115945364792025420300208e-13
Verifying the dierential equation:
>>> a = 2.5
>>> f = lambda z: hyp0f1(a,z)
>>> for z in [0, 10, 3+4j]:
...
chop(z*diff(f,z,2) + a*diff(f,z) - f(z))
...
0.0
0.0
0.0
hyp1f1()
mpmath.hyp1f1(a, b, z)
Gives the conuent hypergeometric function of the rst kind,
1 F1 (a, b, z)
(a)k z k
,
(b)k k!
k=0
also known as Kummers function and sometimes denoted by M (a, b, z). This function
gives one solution to the conuent (Kummers) dierential equation
zf 00 (z) + (b z)f 0 (z) af (z) = 0.
A second solution is given by the U function; see hyperu() (page 862). Solutions are
also given in an alternate form by the Whittaker functions (whitm() (page 863), whitw()
(page 864)).
hyp1f1(a,b,z) is equivalent to hyper([a],[b],z); see documentation for hyper()
(page 895) for more information.
Examples
Evaluation for real and complex values of the argument z, with xed parameters a =
2, b = 1/3:
5.15. Welcome to mpmaths documentation!
889
SymPy Documentation, Release 0.7.6
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> hyp1f1(2, (-1,3), 3.25)
-2815.956856924817275640248
>>> hyp1f1(2, (-1,3), -3.25)
-1.145036502407444445553107
>>> hyp1f1(2, (-1,3), 1000)
-8.021799872770764149793693e+441
>>> hyp1f1(2, (-1,3), -1000)
0.000003131987633006813594535331
>>> hyp1f1(2, (-1,3), 100+100j)
(-3.189190365227034385898282e+48 - 1.106169926814270418999315e+49j)
Parameters may be complex:
>>> hyp1f1(2+3j, -1+j, 10j)
(261.8977905181045142673351 + 160.8930312845682213562172j)
Arbitrarily large values of z are supported:
>>> hyp1f1(3, 4, 10**20)
3.890569218254486878220752e+43429448190325182745
>>> hyp1f1(3, 4, -10**20)
6.0e-60
>>> hyp1f1(3, 4, 10**20*j)
(-1.935753855797342532571597e-20 - 2.291911213325184901239155e-20j)
Verifying the dierential equation:
>>> a, b = 1.5, 2
>>> f = lambda z: hyp1f1(a,b,z)
>>> for z in [0, -10, 3, 3+4j]:
...
chop(z*diff(f,z,2) + (b-z)*diff(f,z) - a*f(z))
...
0.0
0.0
0.0
0.0
An integral representation:
>>> a, b = 1.5, 3
>>> z = 1.5
>>> hyp1f1(a,b,z)
2.269381460919952778587441
>>> g = lambda t: exp(z*t)*t**(a-1)*(1-t)**(b-a-1)
>>> gammaprod([b],[a,b-a])*quad(g, [0,1])
2.269381460919952778587441
hyp1f2()
mpmath.hyp1f2(a1, b1, b2, z)
Gives the hypergeometric function 1 F2 (a1 , a2 ; b1 , b2 ; z). The call hyp1f2(a1,b1,b2,z) is
equivalent to hyper([a1],[b1,b2],z).
Evaluation works for complex and arbitrarily large arguments:
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
890
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> a, b, c = 1.5, (-1,3), 2.25
>>> hyp1f2(a, b, c, 10**20)
-1.159388148811981535941434e+8685889639
>>> hyp1f2(a, b, c, -10**20)
-12.60262607892655945795907
>>> hyp1f2(a, b, c, 10**20*j)
(4.237220401382240876065501e+6141851464 - 2.950930337531768015892987e+6141851464j)
>>> hyp1f2(2+3j, -2j, 0.5j, 10-20j)
(135881.9905586966432662004 - 86681.95885418079535738828j)
hyp2f0()
mpmath.hyp2f0(a, b, z)
Gives the hypergeometric function 2 F0 , dened formally by the series
2 F0 (a, b; ; z)
n=0
(a)n (b)n
zn
.
n!
This series usually does not converge. For small enough z, it can be viewed as an asymptotic series that may be summed directly with an appropriate truncation. When this is
not the case, hyp2f0() (page 891) gives a regularized sum, or equivalently, it uses a
representation in terms of the hypergeometric U function [1]. The series also converges
when either a or b is a nonpositive integer, as it then terminates into a polynomial after
a or b terms.
Examples
Evaluation is supported for arbitrary complex arguments:
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> hyp2f0((2,3), 1.25, -100)
0.07095851870980052763312791
>>> hyp2f0((2,3), 1.25, 100)
(-0.03254379032170590665041131 + 0.07269254613282301012735797j)
>>> hyp2f0(-0.75, 1-j, 4j)
(-0.3579987031082732264862155 - 3.052951783922142735255881j)
Even with real arguments, the regularized value of 2F0 is often complex-valued, but the
imaginary part decreases exponentially as z 0. In the following example, the rst call
uses complex evaluation while the second has a small enough z to evaluate using the
direct series and thus the returned value is strictly real (this should be taken to indicate
that the imaginary part is less than eps):
>>> mp.dps = 15
>>> hyp2f0(1.5, 0.5, 0.05)
(1.04166637647907 + 8.34584913683906e-8j)
>>> hyp2f0(1.5, 0.5, 0.0005)
1.00037535207621
The imaginary part can be retrieved by increasing the working precision:
>>> mp.dps = 80
>>> nprint(hyp2f0(1.5, 0.5, 0.009).imag)
1.23828e-46
In the polynomial case (the series terminating), 2F0 can evaluate exactly:
5.15. Welcome to mpmaths documentation!
891
SymPy Documentation, Release 0.7.6
>>> mp.dps = 15
>>> hyp2f0(-6,-6,2)
291793.0
>>> identify(hyp2f0(-2,1,0.25))
(5/8)
The coecients of the polynomials can be recovered using Taylor expansion:
>>> nprint(taylor(lambda x: hyp2f0(-3,0.5,x),
[1.0, -1.5, 2.25, -1.875, 0.0, 0.0, 0.0, 0.0,
>>> nprint(taylor(lambda x: hyp2f0(-4,0.5,x),
[1.0, -2.0, 4.5, -7.5, 6.5625, 0.0, 0.0, 0.0,
0, 10))
0.0, 0.0, 0.0]
0, 10))
0.0, 0.0, 0.0]
[1] http://people.math.sfu.ca/cbm/aands/page 504.htm
hyp2f1()
mpmath.hyp2f1(a, b, c, z)
Gives the Gauss hypergeometric function 2 F1 (often simply referred to as the hypergeometric function), dened for |z| < 1 as
2 F1 (a, b, c, z)
(a)k (b)k z k
.
(c)k k!
k=0
and for |z| 1 by analytic continuation, with a branch cut on (1, ) when necessary.
Special cases of this function include many of the orthogonal polynomials as well as the
incomplete beta function and other functions. Properties of the Gauss hypergeometric
function are documented comprehensively in many references, for example Abramowitz
& Stegun, section 15.
The implementation supports the analytic continuation as well as evaluation close
to the unit circle where |z| 1. The syntax hyp2f1(a,b,c,z) is equivalent to hyper([a,b],[c],z).
Examples
Evaluation with z inside, outside and on the unit circle, for xed parameters:
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> hyp2f1(2, (1,2), 4, 0.75)
1.303703703703703703703704
>>> hyp2f1(2, (1,2), 4, -1.75)
0.7431290566046919177853916
>>> hyp2f1(2, (1,2), 4, 1.75)
(1.418075801749271137026239 - 1.114976146679907015775102j)
>>> hyp2f1(2, (1,2), 4, 1)
1.6
>>> hyp2f1(2, (1,2), 4, -1)
0.8235498012182875315037882
>>> hyp2f1(2, (1,2), 4, j)
(0.9144026291433065674259078 + 0.2050415770437884900574923j)
>>> hyp2f1(2, (1,2), 4, 2+j)
(0.9274013540258103029011549 + 0.7455257875808100868984496j)
>>> hyp2f1(2, (1,2), 4, 0.25j)
(0.9931169055799728251931672 + 0.06154836525312066938147793j)
Evaluation with complex parameter values:
892
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> hyp2f1(1+j, 0.75, 10j, 1+5j)
(0.8834833319713479923389638 + 0.7053886880648105068343509j)
Evaluation with z = 1:
>>> hyp2f1(-2.5, 3.5, 1.5, 1)
0.0
>>> hyp2f1(-2.5, 3, 4, 1)
0.06926406926406926406926407
>>> hyp2f1(2, 3, 4, 1)
+inf
Evaluation for huge arguments:
>>> hyp2f1((-1,3), 1.75, 4, 1e100)
(7.883714220959876246415651e+32 + 1.365499358305579597618785e+33j)
>>> hyp2f1((-1,3), 1.75, 4, 1e1000000)
(7.883714220959876246415651e+333332 + 1.365499358305579597618785e+333333j)
>>> hyp2f1((-1,3), 1.75, 4, 1e1000000j)
(1.365499358305579597618785e+333333 - 7.883714220959876246415651e+333332j)
An integral representation:
>>> a,b,c,z = -0.5, 1, 2.5, 0.25
>>> g = lambda t: t**(b-1) * (1-t)**(c-b-1) * (1-t*z)**(-a)
>>> gammaprod([c],[b,c-b]) * quad(g, [0,1])
0.9480458814362824478852618
>>> hyp2f1(a,b,c,z)
0.9480458814362824478852618
Verifying the hypergeometric dierential equation:
>>> f = lambda z: hyp2f1(a,b,c,z)
>>> chop(z*(1-z)*diff(f,z,2) + (c-(a+b+1)*z)*diff(f,z) - a*b*f(z))
0.0
hyp2f2()
mpmath.hyp2f2(a1, a2, b1, b2, z)
Gives the hypergeometric function 2 F2 (a1 , a2 ; b1 , b2 ; z). The call hyp2f2(a1,a2,b1,b2,z)
is equivalent to hyper([a1,a2],[b1,b2],z).
Evaluation works for complex and arbitrarily large arguments:
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> a, b, c, d = 1.5, (-1,3), 2.25, 4
>>> hyp2f2(a, b, c, d, 10**20)
-5.275758229007902299823821e+43429448190325182663
>>> hyp2f2(a, b, c, d, -10**20)
2561445.079983207701073448
>>> hyp2f2(a, b, c, d, 10**20*j)
(2218276.509664121194836667 - 1280722.539991603850462856j)
>>> hyp2f2(2+3j, -2j, 0.5j, 4j, 10-20j)
(80500.68321405666957342788 - 20346.82752982813540993502j)
hyp2f3()
5.15. Welcome to mpmaths documentation!
893
SymPy Documentation, Release 0.7.6
mpmath.hyp2f3(a1, a2, b1, b2, b3, z)
Gives
the
hypergeometric
function
The
2 F3 (a1 , a2 ; b1 , b2 , b3 ; z).
hyp2f3(a1,a2,b1,b2,b3,z) is equivalent to hyper([a1,a2],[b1,b2,b3],z).
call
Evaluation works for arbitrarily large arguments:
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> a1,a2,b1,b2,b3 = 1.5, (-1,3), 2.25, 4, (1,5)
>>> hyp2f3(a1,a2,b1,b2,b3,10**20)
-4.169178177065714963568963e+8685889590
>>> hyp2f3(a1,a2,b1,b2,b3,-10**20)
7064472.587757755088178629
>>> hyp2f3(a1,a2,b1,b2,b3,10**20*j)
(-5.163368465314934589818543e+6141851415 + 1.783578125755972803440364e+6141851416j)
>>> hyp2f3(2+3j, -2j, 0.5j, 4j, -1-j, 10-20j)
(-2280.938956687033150740228 + 13620.97336609573659199632j)
>>> hyp2f3(2+3j, -2j, 0.5j, 4j, -1-j, 10000000-20000000j)
(4.849835186175096516193e+3504 - 3.365981529122220091353633e+3504j)
hyp3f2()
mpmath.hyp3f2(a1, a2, a3, b1, b2, z)
Gives the generalized hypergeometric function 3 F2 , dened for |z| < 1 as
3 F2 (a1 , a2 , a3 , b1 , b2 , z) =
(a1 )k (a2 )k (a3 )k z k
.
(b1 )k (b2 )k
k!
k=0
and for |z| 1 by analytic continuation. The analytic structure of this function is similar
to that of 2 F1 , generally with a singularity at z = 1 and a branch cut on (1, ).
Evaluation is supported inside, on, and outside the circle of convergence |z| = 1:
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> hyp3f2(1,2,3,4,5,0.25)
1.083533123380934241548707
>>> hyp3f2(1,2+2j,3,4,5,-10+10j)
(0.1574651066006004632914361 - 0.03194209021885226400892963j)
>>> hyp3f2(1,2,3,4,5,-10)
0.3071141169208772603266489
>>> hyp3f2(1,2,3,4,5,10)
(-0.4857045320523947050581423 - 0.5988311440454888436888028j)
>>> hyp3f2(0.25,1,1,2,1.5,1)
1.157370995096772047567631
>>> (8-pi-2*ln2)/3
1.157370995096772047567631
>>> hyp3f2(1+j,0.5j,2,1,-2j,-1)
(1.74518490615029486475959 + 0.1454701525056682297614029j)
>>> hyp3f2(1+j,0.5j,2,1,-2j,sqrt(j))
(0.9829816481834277511138055 - 0.4059040020276937085081127j)
>>> hyp3f2(-3,2,1,-5,4,1)
1.41
>>> hyp3f2(-3,2,1,-5,4,2)
2.12
Evaluation very close to the unit circle:
894
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> hyp3f2(1,2,3,4,5,1.0001)
(1.564877796743282766872279 - 3.76821518787438186031973e-11j)
>>> hyp3f2(1,2,3,4,5,1+0.0001j)
(1.564747153061671573212831 + 0.0001305757570366084557648482j)
>>> hyp3f2(1,2,3,4,5,0.9999)
1.564616644881686134983664
>>> hyp3f2(1,2,3,4,5,-0.9999)
0.7823896253461678060196207
Note: Evaluation for |z1| small can currently be inaccurate or slow for some parameter
combinations.
For various parameter combinations, 3 F2 admits representation in terms of hypergeometric functions of lower degree, or in terms of simpler functions:
>>> for a, b, z in [(1,2,-1), (2,0.5,1)]:
...
hyp2f1(a,b,a+b+0.5,z)**2
...
hyp3f2(2*a,a+b,2*b,a+b+0.5,2*a+2*b,z)
...
0.4246104461966439006086308
0.4246104461966439006086308
7.111111111111111111111111
7.111111111111111111111111
>>> z = 2+3j
>>> hyp3f2(0.5,1,1.5,2,2,z)
(0.7621440939243342419729144 + 0.4249117735058037649915723j)
>>> 4*(pi-2*ellipe(z))/(pi*z)
(0.7621440939243342419729144 + 0.4249117735058037649915723j)
Generalized hypergeometric functions
hyper()
mpmath.hyper(a s, b s, z)
Evaluates the generalized hypergeometric function
p Fq (a1 , . . . , ap ; b1 , . . . , bq ; z) =
(a1 )n (a2 )n . . . (ap )n z n
(b1 )n (b2 )n . . . (bq )n n!
n=0
where (x)n denotes the rising factorial (see rf() (page 784)).
The parameters lists a s and b s may contain integers, real numbers, complex numbers,
as well as exact fractions given in the form of tuples (p, q). hyper() (page 895) is optimized to handle integers and fractions more eciently than arbitrary oating-point
parameters (since rational parameters are by far the most common).
Examples
Verifying that hyper() (page 895) gives the sum in the denition, by comparison with
nsum() (page 979):
>>>
>>>
>>>
>>>
from sympy.mpmath import *
mp.dps = 25; mp.pretty = True
a,b,c,d = 2,3,4,5
x = 0.25
5.15. Welcome to mpmaths documentation!
895
SymPy Documentation, Release 0.7.6
>>> hyper([a,b],[c,d],x)
1.078903941164934876086237
>>> fn = lambda n: rf(a,n)*rf(b,n)/rf(c,n)/rf(d,n)*x**n/fac(n)
>>> nsum(fn, [0, inf])
1.078903941164934876086237
The parameters can be any combination of integers, fractions, oats and complex numbers:
>>> a, b, c, d, e = 1, (-1,2), pi, 3+4j, (2,3)
>>> x = 0.2j
>>> hyper([a,b],[c,d,e],x)
(0.9923571616434024810831887 - 0.005753848733883879742993122j)
>>> b, e = -0.5, mpf(2)/3
>>> fn = lambda n: rf(a,n)*rf(b,n)/rf(c,n)/rf(d,n)/rf(e,n)*x**n/fac(n)
>>> nsum(fn, [0, inf])
(0.9923571616434024810831887 - 0.005753848733883879742993122j)
The 0 F0 and 1 F0 series are just elementary functions:
>>> a, z = sqrt(2), +pi
>>> hyper([],[],z)
23.14069263277926900572909
>>> exp(z)
23.14069263277926900572909
>>> hyper([a],[],z)
(-0.09069132879922920160334114 + 0.3283224323946162083579656j)
>>> (1-z)**(-a)
(-0.09069132879922920160334114 + 0.3283224323946162083579656j)
If any ak coecient is a nonpositive integer, the series terminates into a nite polynomial:
>>> hyper([1,1,1,-3],[2,5],1)
0.7904761904761904761904762
>>> identify(_)
(83/105)
If any bk is a nonpositive integer, the function is undened (unless the series terminates
before the division by zero occurs):
>>> hyper([1,1,1,-3],[-2,5],1)
Traceback (most recent call last):
...
ZeroDivisionError: pole in hypergeometric series
>>> hyper([1,1,1,-1],[-2,5],1)
1.1
Traceback (most recent call last):
...
ZeroDivisionError: pole in hypergeometric series
Except for polynomial cases, the radius of convergence R of the hypergeometric series
is either R = (if p q), R = 1 (if p = q + 1), or R = 0 (if p > q + 1).
The analytic continuations of the functions with p = q + 1, i.e. 2 F1 , 3 F2 , 4 F3 , etc, are all
implemented and therefore these functions can be evaluated for |z| 1. The shortcuts
hyp2f1() (page 892), hyp3f2() (page 894) are available to handle the most common
cases (see their documentation), but functions of higher degree are also supported via
hyper() (page 895):
896
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> hyper([1,2,3,4], [5,6,7], 1)
# 4F3 at finite-valued branch point
1.141783505526870731311423
>>> hyper([4,5,6,7], [1,2,3], 1)
# 4F3 at pole
+inf
>>> hyper([1,2,3,4,5], [6,7,8,9], 10)
# 5F4
(1.543998916527972259717257 - 0.5876309929580408028816365j)
>>> hyper([1,2,3,4,5,6], [7,8,9,10,11], 1j)
# 6F5
(0.9996565821853579063502466 + 0.0129721075905630604445669j)
Near z = 1 with noninteger parameters:
>>> hyper([1/3,1,3/2,2], [1/5,11/6,41/8], 1)
2.219433352235586121250027
>>> hyper([1/3,1,3/2,2], [1/5,11/6,5/4], 1)
+inf
>>> eps1 = extradps(6)(lambda: 1 - mpf(1e-6))()
>>> hyper([1/3,1,3/2,2], [1/5,11/6,5/4], eps1)
2923978034.412973409330956
Please note that, as currently implemented, evaluation of p Fp1 with p 3 may be slow
or inaccurate when |z 1| is small, for some parameter values.
When p > q + 1, hyper computes the (iterated) Borel sum of the divergent series. For 2 F0
the Borel sum has an analytic solution and can be computed eciently (see hyp2f0()
(page 891)). For higher degrees, the functions is evaluated rst by attempting to sum
it directly as an asymptotic series (this only works for tiny |z|), and then by evaluating
the Borel regularized sum using numerical integration. Except for special parameter
combinations, this can be extremely slow.
>>> hyper([1,1], [], 0.5)
# regularization of 2F0
(1.340965419580146562086448 + 0.8503366631752726568782447j)
>>> hyper([1,1,1,1], [1], 0.5)
# regularization of 4F1
(1.108287213689475145830699 + 0.5327107430640678181200491j)
With the following magnitude of argument, the asymptotic series for 3 F1 gives only a few
digits. Using Borel summation, hyper can produce a value with full accuracy:
>>> mp.dps = 15
>>> hyper([2,0.5,4], [5.25], 0.08, force_series=True)
Traceback (most recent call last):
...
NoConvergence: Hypergeometric series converges too slowly. Try increasing maxterms.
>>> hyper([2,0.5,4], [5.25], 0.08, asymp_tol=1e-4)
1.0725535790737
>>> hyper([2,0.5,4], [5.25], 0.08)
(1.07269542893559 + 5.54668863216891e-5j)
>>> hyper([2,0.5,4], [5.25], -0.08, asymp_tol=1e-4)
0.946344925484879
>>> hyper([2,0.5,4], [5.25], -0.08)
0.946312503737771
>>> mp.dps = 25
>>> hyper([2,0.5,4], [5.25], -0.08)
0.9463125037377662296700858
Traceback (most recent call last):
...
NoConvergence: Hypergeometric series converges too slowly. Try increasing maxterms.
Note that with the positive z value, there is a complex part in the correct result, which
falls below the tolerance of the asymptotic series.
5.15. Welcome to mpmaths documentation!
897
SymPy Documentation, Release 0.7.6
hypercomb()
mpmath.hypercomb(ctx, function, params=[], discard known zeros=True, **kwargs)
Computes a weighted combination of hypergeometric functions
[l
]
mr
N
r
(r,k )
cr,k
k=1
n r
wr,k
pr Fqr (ar,1 , . . . , ar,p ; br,1 , . . . , br,q ; zr ) .
k=1 (r,k )
r=1
k=1
Typically the parameters are linear combinations of a small set of base parameters; hypercomb() (page 898) permits computing a correct value in the case that some of the ,
, b turn out to be nonpositive integers, or if division by zero occurs for some wc , assuming
that there are opposing singularities that cancel out. The limit is computed by evaluating
the function with the base parameters perturbed, at a higher working precision.
The rst argument should be a function that takes the perturbable base parameters
params as input and returns N tuples (w, c, alpha, beta, a, b, z), where the
coecients w, c, gamma factors alpha, beta, and hypergeometric coecients a, b each
should be lists of numbers, and z should be a single number.
Examples
The following evaluates
(a 1)
(a 3)
z
1 F1 (a, a 1, z) = e (a 4)(a + z 1)
(a 4)
with a = 1, z = 3. There is a zero factor, two gamma function poles, and the 1F1 function
is singular; all singularities cancel out to give a nite value:
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = True
>>> hypercomb(lambda a: [([a-1],[1],[a-3],[a-4],[a],[a-1],3)], [1])
-180.769832308689
>>> -9*exp(3)
-180.769832308689
Meijer G-function
meijerg()
mpmath.meijerg(a s, b s, z, r=1, **kwargs)
Evaluates the Meijer G-function, dened as
(
Gm,n
p,q
m
n
)
(1 aj s)
1
a1 , . . . , an ; an+1 . . . ap
j=1 (bj + s)
qj=1
p
z s/r ds
z; r =
b1 , . . . , bm ; bm+1 . . . bq
2i L j=n+1 (aj + s) j=m+1 (1 bj s)
for an appropriate choice of the contour L (see references).
There are p elements aj . The argument a s should be a pair of lists, the rst containing
the n elements a1 , . . . , an and the second containing the p n elements an+1 , . . . ap .
There are q elements bj . The argument b s should be a pair of lists, the rst containing
the m elements b1 , . . . , bm and the second containing the q m elements bm+1 , . . . bq .
The implicit tuple (m, n, p, q) constitutes the order or degree of the Meijer G-function, and
is determined by the lengths of the coecient vectors. Confusingly, the indices in this
tuple appear in a dierent order from the coecients, but this notation is standard. The
many examples given below should hopefully clear up any potential confusion.
898
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Algorithm
The Meijer G-function is evaluated as a combination of hypergeometric series. There are
two versions of the function, which can be selected with the optional series argument.
series=1 uses a sum of m p Fq1 functions of z
series=2 uses a sum of n q Fp1 functions of 1/z
The default series is chosen based on the degree and |z| in order to be consistent with
Mathematicas. This denition of the Meijer G-function has a discontinuity at |z| = 1 for
some orders, which can be avoided by explicitly specifying a series.
Keyword arguments are forwarded to hypercomb() (page 898).
Examples
Many standard functions are special cases of the Meijer G-function (possibly rescaled
and/or with branch cut corrections). We dene some test parameters:
>>>
>>>
>>>
>>>
>>>
from sympy.mpmath import *
mp.dps = 25; mp.pretty = True
a = mpf(0.75)
b = mpf(1.5)
z = mpf(2.25)
(
z
The exponential function: e =
G1,0
0,1
)
z
0
>>> meijerg([[],[]], [[0],[]], -z)
9.487735836358525720550369
>>> exp(z)
9.487735836358525720550369
The natural logarithm: log(1 + z) =
G1,2
2,2
)
1, 1
z
1, 0
>>> meijerg([[1,1],[]], [[1],[0]], z)
1.178654996341646117219023
>>> log(1+z)
1.178654996341646117219023
A rational function:
z
z+1
G1,2
2,2
)
1, 1
z
1, 1
>>> meijerg([[1,1],[]], [[1],[1]], z)
0.6923076923076923076923077
>>> z/(z+1)
0.6923076923076923076923077
The sine and cosine functions:
)
(
1,0
1 sin(2 z) = G
1
0,2
z
2, 0
)
(
1,0
1
z
cos(2 z) = G
0,2
0, 12
>>> meijerg([[],[]], [[0.5],[0]], (z/2)**2)
0.4389807929218676682296453
>>> sin(z)/sqrt(pi)
0.4389807929218676682296453
>>> meijerg([[],[]], [[0],[0.5]], (z/2)**2)
-0.3544090145996275423331762
5.15. Welcome to mpmaths documentation!
899
SymPy Documentation, Release 0.7.6
>>> cos(z)/sqrt(pi)
-0.3544090145996275423331762
Bessel functions:
)
(
1,0
Ja (2 z) = G0,2 a a z
2,2
)
(
a1
2,0
2
Ya (2 z) = G1,3 a a a1 z
2,2,
2
(
1,0
a/2 a/2
(z) z
Ia (2 z) = G0,2 a a
,
2 2
)
(
2Ka (2 z) = G2,0
a
a z
0,2
2,2
)
z
As the example with the Bessel I function shows, a branch factor is required for some
arguments when inverting the square root.
>>> meijerg([[],[]], [[a/2],[-a/2]], (z/2)**2)
0.5059425789597154858527264
>>> besselj(a,z)
0.5059425789597154858527264
>>> meijerg([[],[(-a-1)/2]], [[a/2,-a/2],[(-a-1)/2]], (z/2)**2)
0.1853868950066556941442559
>>> bessely(a, z)
0.1853868950066556941442559
>>> meijerg([[],[]], [[a/2],[-a/2]], -(z/2)**2)
(0.8685913322427653875717476 + 2.096964974460199200551738j)
>>> (-z)**(a/2) / z**(a/2) * besseli(a, z)
(0.8685913322427653875717476 + 2.096964974460199200551738j)
>>> 0.5*meijerg([[],[]], [[a/2,-a/2],[]], (z/2)**2)
0.09334163695597828403796071
>>> besselk(a,z)
0.09334163695597828403796071
Error functions:
(
2(a1)
z
erfc(z) = G2,0
1,2
a
a 1, a
1
2
)
z, 1
2
>>> meijerg([[],[a]], [[a-1,a-0.5],[]], z, 0.5)
0.00172839843123091957468712
>>> sqrt(pi) * z**(2*a-2) * erfc(z)
0.00172839843123091957468712
A Meijer G-function of higher degree, (1,1,2,3):
>>> meijerg([[a],[b]], [[a],[b,a-1]], z)
1.55984467443050210115617
>>> sin((b-a)*pi)/pi*(exp(z)-1)*z**(a-1)
1.55984467443050210115617
A Meijer G-function of still higher degree, (4,1,2,4), that can be expanded as a messy
combination of exponential integrals:
>>> meijerg([[a],[2*b-a]], [[b,a,b-0.5,-1-a+2*b],[]], z)
0.3323667133658557271898061
>>> chop(4**(a-b+1)*sqrt(pi)*gamma(2*b-2*a)*z**a*\
...
expint(2*b-2*a, -2*sqrt(-z))*expint(2*b-2*a, 2*sqrt(-z)))
0.3323667133658557271898061
900
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
In the following case, dierent series give dierent values:
>>> chop(meijerg([[1],[0.25]],[[3],[0.5]],-2))
-0.06417628097442437076207337
>>> meijerg([[1],[0.25]],[[3],[0.5]],-2,series=1)
0.1428699426155117511873047
>>> chop(meijerg([[1],[0.25]],[[3],[0.5]],-2,series=2))
-0.06417628097442437076207337
References
1.http://en.wikipedia.org/wiki/Meijer G-function
2.http://mathworld.wolfram.com/MeijerG-Function.html
3.http://functions.wolfram.com/HypergeometricFunctions/MeijerG/
4.http://functions.wolfram.com/HypergeometricFunctions/MeijerG1/
Bilateral hypergeometric series
bihyper()
mpmath.bihyper(a s, b s, z, **kwargs)
Evaluates the bilateral hypergeometric series
A HB (a1 , . . . , ak ; b1 , . . . , bB ; z)
(a1 )n . . . (aA )n n
z
(b1 )n . . . (bB )n
n=
where, for direct convergence, A = B and |z| = 1, although a regularized sum exists more
generally by considering the bilateral series as a sum of two ordinary hypergeometric
functions. In order for the series to make sense, none of the parameters may be integers.
Examples
The value of 2 H2 at z = 1 is given by Dougalls formula:
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> a,b,c,d = 0.5, 1.5, 2.25, 3.25
>>> bihyper([a,b],[c,d],1)
-14.49118026212345786148847
>>> gammaprod([c,d,1-a,1-b,c+d-a-b-1],[c-a,d-a,c-b,d-b])
-14.49118026212345786148847
The regularized function 1 H0 can be expressed as the sum of one 2 F0 function and one
1 F1 function:
>>> a = mpf(0.25)
>>> z = mpf(0.75)
>>> bihyper([a], [], z)
(0.2454393389657273841385582 + 0.2454393389657273841385582j)
>>> hyper([a,1],[],z) + (hyper([1],[1-a],-1/z)-1)
(0.2454393389657273841385582 + 0.2454393389657273841385582j)
>>> hyper([a,1],[],z) + hyper([1],[2-a],-1/z)/z/(a-1)
(0.2454393389657273841385582 + 0.2454393389657273841385582j)
References
1.[Slater] (page 1910) (chapter 6: Bilateral Series, pp. 180-189)
2.[Wikipedia] (page 1910) http://en.wikipedia.org/wiki/Bilateral hypergeometric series
5.15. Welcome to mpmaths documentation!
901
SymPy Documentation, Release 0.7.6
Hypergeometric functions of two variables
hyper2d()
mpmath.hyper2d(a, b, x, y, **kwargs)
Sums the generalized 2D hypergeometric series
P ((a), m, n) xm y n
Q((b), m, n) m!n!
m=0 n=0
where (a) = (a1 , . . . , ar ), (b) = (b1 , . . . , bs ) and where P and Q are products of rising factorials
such as (aj )n or (aj )m+n . P and Q are specied in the form of dicts, with the m and n
dependence as keys and parameter lists as values. The supported rising factorials are
given in the following table (note that only a few are supported in Q):
Key
m
n
m+n
m-n
n-m
2m+n
2m-n
2n-m
Rising factorial
(aj )m
(aj )n
(aj )m+n
(aj )mn
(aj )nm
(aj )2m+n
(aj )2mn
(aj )2nm
Q
Yes
Yes
Yes
No
No
No
No
No
For example, the Appell F1 and F4 functions
F1 =
(a)m+n (b)m (c)n xm y n
(d)m+n
m!n!
m=0 n=0
(a)m+n (b)m+n xm y n
F4 =
(c)m (d)n
m!n!
m=0 n=0
can be represented respectively as
hyper2d({m+n:[a], m:[b], n:[c]}, {m+n:[d]}, x, y)
hyper2d({m+n:[a,b]}, {m:[c], n:[d]}, x, y)
More generally, hyper2d() (page 902) can evaluate any of the 34 distinct convergent
second-order (generalized Gaussian) hypergeometric series enumerated by Horn, as well
as the Kampe de Feriet function.
The series is computed by rewriting it so that the inner series (i.e. the series containing
n and y) has the form of an ordinary generalized hypergeometric series and thereby can
be evaluated eciently using hyper() (page 895). If possible, manually swapping x and
y and the corresponding parameters can sometimes give better results.
Examples
Two separable cases: a product of two geometric series, and a product of two Gaussian
hypergeometric functions:
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> x, y = mpf(0.25), mpf(0.5)
>>> hyper2d({m:1,n:1}, {}, x,y)
2.666666666666666666666667
>>> 1/(1-x)/(1-y)
902
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
2.666666666666666666666667
>>> hyper2d({m:[1,2],n:[3,4]}, {m:[5],n:[6]}, x,y)
4.164358531238938319669856
>>> hyp2f1(1,2,5,x)*hyp2f1(3,4,6,y)
4.164358531238938319669856
Some more series that can be done in closed form:
>>> hyper2d({m:1,n:1},{m+n:1},x,y)
2.013417124712514809623881
>>> (exp(x)*x-exp(y)*y)/(x-y)
2.013417124712514809623881
Six of the 34 Horn functions, G1-G3 and H1-H3:
>>> from sympy.mpmath import *
>>> mp.dps = 10; mp.pretty = True
>>> x, y = 0.0625, 0.125
>>> a1,a2,b1,b2,c1,c2,d = 1.1,-1.2,-1.3,-1.4,1.5,-1.6,1.7
>>> hyper2d({m+n:a1,n-m:b1,m-n:b2},{},x,y) # G1
1.139090746
>>> nsum(lambda m,n: rf(a1,m+n)*rf(b1,n-m)*rf(b2,m-n)*\
...
x**m*y**n/fac(m)/fac(n), [0,inf], [0,inf])
1.139090746
>>> hyper2d({m:a1,n:a2,n-m:b1,m-n:b2},{},x,y) # G2
0.9503682696
>>> nsum(lambda m,n: rf(a1,m)*rf(a2,n)*rf(b1,n-m)*rf(b2,m-n)*\
...
x**m*y**n/fac(m)/fac(n), [0,inf], [0,inf])
0.9503682696
>>> hyper2d({2n-m:a1,2m-n:a2},{},x,y) # G3
1.029372029
>>> nsum(lambda m,n: rf(a1,2*n-m)*rf(a2,2*m-n)*\
...
x**m*y**n/fac(m)/fac(n), [0,inf], [0,inf])
1.029372029
>>> hyper2d({m-n:a1,m+n:b1,n:c1},{m:d},x,y) # H1
-1.605331256
>>> nsum(lambda m,n: rf(a1,m-n)*rf(b1,m+n)*rf(c1,n)/rf(d,m)*\
...
x**m*y**n/fac(m)/fac(n), [0,inf], [0,inf])
-1.605331256
>>> hyper2d({m-n:a1,m:b1,n:[c1,c2]},{m:d},x,y) # H2
-2.35405404
>>> nsum(lambda m,n: rf(a1,m-n)*rf(b1,m)*rf(c1,n)*rf(c2,n)/rf(d,m)*\
...
x**m*y**n/fac(m)/fac(n), [0,inf], [0,inf])
-2.35405404
>>> hyper2d({2m+n:a1,n:b1},{m+n:c1},x,y) # H3
0.974479074
>>> nsum(lambda m,n: rf(a1,2*m+n)*rf(b1,n)/rf(c1,m+n)*\
...
x**m*y**n/fac(m)/fac(n), [0,inf], [0,inf])
0.974479074
References
1.[SrivastavaKarlsson] (page 1910)
2.[Weisstein] (page 1910) http://mathworld.wolfram.com/HornFunction.html
3.[Weisstein] (page 1910) http://mathworld.wolfram.com/AppellHypergeometricFunction.html
appellf1()
5.15. Welcome to mpmaths documentation!
903
SymPy Documentation, Release 0.7.6
mpmath.appellf1(a, b1, b2, c, x, y, **kwargs)
Gives the Appell F1 hypergeometric function of two variables,
F1 (a, b1 , b2 , c, x, y) =
(a)m+n (b1 )m (b2 )n xm y n
.
(c)m+n
m!n!
m=0 n=0
This series is only generally convergent when |x| < 1 and |y| < 1, although appellf1()
(page 903) can evaluate an analytic continuation with respecto to either variable, and
sometimes both.
Examples
Evaluation is supported for real and complex parameters:
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> appellf1(1,0,0.5,1,0.5,0.25)
1.154700538379251529018298
>>> appellf1(1,1+j,0.5,1,0.5,0.5j)
(1.138403860350148085179415 + 1.510544741058517621110615j)
For some integer parameters, the F1 series reduces to a polynomial:
>>> appellf1(2,-4,-3,1,2,5)
-816.0
>>> appellf1(-5,1,2,1,4,5)
-20528.0
The analytic continuation with respect to either x or y, and sometimes with respect to
both, can be evaluated:
>>> appellf1(2,3,4,5,100,0.5)
(0.0006231042714165329279738662 + 0.0000005769149277148425774499857j)
>>> appellf1(1.1, 0.3, 0.2+2j, 0.4, 0.2, 1.5+3j)
(-0.1782604566893954897128702 + 0.002472407104546216117161499j)
>>> appellf1(1,2,3,4,10,12)
-0.07122993830066776374929313
For certain arguments, F1 reduces to an ordinary hypergeometric function:
>>> appellf1(1,2,3,5,0.5,0.25)
1.547902270302684019335555
>>> 4*hyp2f1(1,2,5,1/3)/3
1.547902270302684019335555
>>> appellf1(1,2,3,4,0,1.5)
(-1.717202506168937502740238 - 2.792526803190927323077905j)
>>> hyp2f1(1,3,4,1.5)
(-1.717202506168937502740238 - 2.792526803190927323077905j)
The F1 function satises a system of partial dierential equations:
>>> a,b1,b2,c,x,y = map(mpf, [1,0.5,0.25,1.125,0.25,-0.25])
>>> F = lambda x,y: appellf1(a,b1,b2,c,x,y)
>>> chop(x*(1-x)*diff(F,(x,y),(2,0)) +
...
y*(1-x)*diff(F,(x,y),(1,1)) +
...
(c-(a+b1+1)*x)*diff(F,(x,y),(1,0)) ...
b1*y*diff(F,(x,y),(0,1)) ...
a*b1*F(x,y))
0.0
>>>
904
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> chop(y*(1-y)*diff(F,(x,y),(0,2)) +
...
x*(1-y)*diff(F,(x,y),(1,1)) +
...
(c-(a+b2+1)*y)*diff(F,(x,y),(0,1)) ...
b2*x*diff(F,(x,y),(1,0)) ...
a*b2*F(x,y))
0.0
The Appell F1 function allows
for closed-form evaluation of various integrals, such as
any integral of the form xr (x + a)p (x + b)q dx:
>>> def integral(a,b,p,q,r,x1,x2):
...
a,b,p,q,r,x1,x2 = map(mpmathify, [a,b,p,q,r,x1,x2])
...
f = lambda x: x**r * (x+a)**p * (x+b)**q
...
def F(x):
...
v = x**(r+1)/(r+1) * (a+x)**p * (b+x)**q
...
v *= (1+x/a)**(-p)
...
v *= (1+x/b)**(-q)
...
v *= appellf1(r+1,-p,-q,2+r,-x/a,-x/b)
...
return v
...
print(Num. quad: %s % quad(f, [x1,x2]))
...
print(Appell F1: %s % (F(x2)-F(x1)))
...
>>> integral(1/5,4/3,-2,3,1/2,0,1)
Num. quad: 9.073335358785776206576981
Appell F1: 9.073335358785776206576981
>>> integral(3/2,4/3,-2,3,1/2,0,1)
Num. quad: 1.092829171999626454344678
Appell F1: 1.092829171999626454344678
>>> integral(3/2,4/3,-2,3,1/2,12,25)
Num. quad: 1106.323225040235116498927
Appell F1: 1106.323225040235116498927
Also incomplete elliptic integrals fall into this category [1]:
>>> def E(z, m):
...
if (pi/2).ae(z):
...
return ellipe(m)
...
return 2*round(re(z)/pi)*ellipe(m) + mpf(-1)**round(re(z)/pi)*\
...
sin(z)*appellf1(0.5,0.5,-0.5,1.5,sin(z)**2,m*sin(z)**2)
...
>>> z, m = 1, 0.5
>>> E(z,m); quad(lambda t: sqrt(1-m*sin(t)**2), [0,pi/4,3*pi/4,z])
0.9273298836244400669659042
0.9273298836244400669659042
>>> z, m = 3, 2
>>> E(z,m); quad(lambda t: sqrt(1-m*sin(t)**2), [0,pi/4,3*pi/4,z])
(1.057495752337234229715836 + 1.198140234735592207439922j)
(1.057495752337234229715836 + 1.198140234735592207439922j)
References
1.[WolframFunctions] (page 1910) http://functions.wolfram.com/EllipticIntegrals/EllipticE2/26/01/
2.[SrivastavaKarlsson] (page 1910)
3.[CabralRosetti] (page 1910)
4.[Vidunas] (page 1910)
5.[Slater] (page 1910)
5.15. Welcome to mpmaths documentation!
905
SymPy Documentation, Release 0.7.6
appellf2()
mpmath.appellf2(a, b1, b2, c1, c2, x, y, **kwargs)
Gives the Appell F2 hypergeometric function of two variables
F2 (a, b1 , b2 , c1 , c2 , x, y) =
(a)m+n (b1 )m (b2 )n xm y n
.
(c1 )m (c2 )n
m!n!
m=0 n=0
The series is generally absolutely convergent for |x| + |y| < 1.
Examples
Evaluation for real and complex arguments:
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> appellf2(1,2,3,4,5,0.25,0.125)
1.257417193533135344785602
>>> appellf2(1,-3,-4,2,3,2,3)
-42.8
>>> appellf2(0.5,0.25,-0.25,2,3,0.25j,0.25)
(0.9880539519421899867041719 + 0.01497616165031102661476978j)
>>> chop(appellf2(1,1+j,1-j,3j,-3j,0.25,0.25))
1.201311219287411337955192
>>> appellf2(1,1,1,4,6,0.125,16)
(-0.09455532250274744282125152 - 0.7647282253046207836769297j)
A transformation formula:
>>> a,b1,b2,c1,c2,x,y = map(mpf, [1,2,0.5,0.25,1.625,-0.125,0.125])
>>> appellf2(a,b1,b2,c1,c2,x,y)
0.2299211717841180783309688
>>> (1-x)**(-a)*appellf2(a,c1-b1,b2,c1,c2,x/(x-1),y/(1-x))
0.2299211717841180783309688
A system of partial dierential equations satised by F2:
>>>
>>>
>>>
...
...
...
...
0.0
>>>
...
...
...
...
0.0
a,b1,b2,c1,c2,x,y = map(mpf, [1,0.5,0.25,1.125,1.5,0.0625,-0.0625])
F = lambda x,y: appellf2(a,b1,b2,c1,c2,x,y)
chop(x*(1-x)*diff(F,(x,y),(2,0)) x*y*diff(F,(x,y),(1,1)) +
(c1-(a+b1+1)*x)*diff(F,(x,y),(1,0)) b1*y*diff(F,(x,y),(0,1)) a*b1*F(x,y))
chop(y*(1-y)*diff(F,(x,y),(0,2)) x*y*diff(F,(x,y),(1,1)) +
(c2-(a+b2+1)*y)*diff(F,(x,y),(0,1)) b2*x*diff(F,(x,y),(1,0)) a*b2*F(x,y))
References
See references for appellf1() (page 903).
appellf3()
906
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
mpmath.appellf3(a1, a2, b1, b2, c, x, y, **kwargs)
Gives the Appell F3 hypergeometric function of two variables
F3 (a1 , a2 , b1 , b2 , c, x, y) =
(a1 )m (a2 )n (b1 )m (b2 )n xm y n
.
(c)m+n
m!n!
m=0 n=0
The series is generally absolutely convergent for |x| < 1, |y| < 1.
Examples
Evaluation for various parameters and variables:
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> appellf3(1,2,3,4,5,0.5,0.25)
2.221557778107438938158705
>>> appellf3(1,2,3,4,5,6,0); hyp2f1(1,3,5,6)
(-0.5189554589089861284537389 - 0.1454441043328607980769742j)
(-0.5189554589089861284537389 - 0.1454441043328607980769742j)
>>> appellf3(1,-2,-3,1,1,4,6)
-17.4
>>> appellf3(1,2,-3,1,1,4,6)
(17.7876136773677356641825 + 19.54768762233649126154534j)
>>> appellf3(1,2,-3,1,1,6,4)
(85.02054175067929402953645 + 148.4402528821177305173599j)
>>> chop(appellf3(1+j,2,1-j,2,3,0.25,0.25))
1.719992169545200286696007
Many transformations and evaluations for special combinations of the parameters are
possible, e.g.:
>>> a,b,c,x,y = map(mpf, [0.5,0.25,0.125,0.125,-0.125])
>>> appellf3(a,c-a,b,c-b,c,x,y)
1.093432340896087107444363
>>> (1-y)**(a+b-c)*hyp2f1(a,b,c,x+y-x*y)
1.093432340896087107444363
>>> x**2*appellf3(1,1,1,1,3,x,-x)
0.01568646277445385390945083
>>> polylog(2,x**2)
0.01568646277445385390945083
>>> a1,a2,b1,b2,c,x = map(mpf, [0.5,0.25,0.125,0.5,4.25,0.125])
>>> appellf3(a1,a2,b1,b2,c,x,1)
1.03947361709111140096947
>>> gammaprod([c,c-a2-b2],[c-a2,c-b2])*hyp3f2(a1,b1,c-a2-b2,c-a2,c-b2,x)
1.03947361709111140096947
The Appell F3 function satises a pair of partial dierential equations:
>>>
>>>
>>>
...
...
...
0.0
>>>
...
...
...
0.0
a1,a2,b1,b2,c,x,y = map(mpf, [0.5,0.25,0.125,0.5,0.625,0.0625,-0.0625])
F = lambda x,y: appellf3(a1,a2,b1,b2,c,x,y)
chop(x*(1-x)*diff(F,(x,y),(2,0)) +
y*diff(F,(x,y),(1,1)) +
(c-(a1+b1+1)*x)*diff(F,(x,y),(1,0)) a1*b1*F(x,y))
chop(y*(1-y)*diff(F,(x,y),(0,2)) +
x*diff(F,(x,y),(1,1)) +
(c-(a2+b2+1)*y)*diff(F,(x,y),(0,1)) a2*b2*F(x,y))
5.15. Welcome to mpmaths documentation!
907
SymPy Documentation, Release 0.7.6
References
See references for appellf1() (page 903).
appellf4()
mpmath.appellf4(a, b, c1, c2, x, y, **kwargs)
Gives the Appell F4 hypergeometric function of two variables
F4 (a, b, c1 , c2 , x, y) =
(a)m+n (b)m+n xm y n
.
(c1 )m (c2 )n m!n!
m=0 n=0
The series is generally absolutely convergent for
|x| + |y| < 1.
Examples
Evaluation for various parameters and arguments:
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> appellf4(1,1,2,2,0.25,0.125)
1.286182069079718313546608
>>> appellf4(-2,-3,4,5,4,5)
34.8
>>> appellf4(5,4,2,3,0.25j,-0.125j)
(-0.2585967215437846642163352 + 2.436102233553582711818743j)
Reduction to 2 F1 in a special case:
>>> a,b,c,x,y = map(mpf, [0.5,0.25,0.125,0.125,-0.125])
>>> appellf4(a,b,c,a+b-c+1,x*(1-y),y*(1-x))
1.129143488466850868248364
>>> hyp2f1(a,b,c,x)*hyp2f1(a,b,a+b-c+1,y)
1.129143488466850868248364
A system of partial dierential equations satised by F4:
>>>
>>>
>>>
...
...
...
...
...
0.0
>>>
...
...
...
...
...
0.0
a,b,c1,c2,x,y = map(mpf, [1,0.5,0.25,1.125,0.0625,-0.0625])
F = lambda x,y: appellf4(a,b,c1,c2,x,y)
chop(x*(1-x)*diff(F,(x,y),(2,0)) y**2*diff(F,(x,y),(0,2)) 2*x*y*diff(F,(x,y),(1,1)) +
(c1-(a+b+1)*x)*diff(F,(x,y),(1,0)) ((a+b+1)*y)*diff(F,(x,y),(0,1)) a*b*F(x,y))
chop(y*(1-y)*diff(F,(x,y),(0,2)) x**2*diff(F,(x,y),(2,0)) 2*x*y*diff(F,(x,y),(1,1)) +
(c2-(a+b+1)*y)*diff(F,(x,y),(0,1)) ((a+b+1)*x)*diff(F,(x,y),(1,0)) a*b*F(x,y))
References
See references for appellf1() (page 903).
908
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Elliptic functions
Elliptic functions historically comprise the elliptic integrals and their inverses, and originate
from the problem of computing the arc length of an ellipse. From a more modern point of
view, an elliptic function is dened as a doubly periodic function, i.e. a function which satises
f (z + 21 ) = f (z + 22 ) = f (z)
for some half-periods 1 , 2 with Im[1 /2 ] > 0. The canonical elliptic functions are the Jacobi
elliptic functions. More broadly, this section includes quasi-doubly periodic functions (such
as the Jacobi theta functions) and other functions useful in the study of elliptic functions.
Many dierent conventions for the arguments of elliptic functions are in use. It is even standard to use dierent parameterizations for dierent functions in the same text or software
(and mpmath is no exception). The usual parameters are the elliptic nome q, which usually
must satisfy |q| < 1; the elliptic parameter m (an arbitrary complex number); the elliptic modulus k (an arbitrary complex number); and the half-period ratio , which usually must satisfy
Im[ ] > 0. These quantities can be expressed in terms of each other using the following relations:
m = k2
= i
K(1 m)
K(m)
q = ei
k=
42 (q)
43 (q)
In addition, an alternative denition is used for the nome in number theory, which we here
denote by q-bar:
q = q 2 = e2i
For convenience, mpmath provides functions to convert between the various parameters
(qfrom() (page 909), mfrom() (page 910), kfrom() (page 911), taufrom() (page 911), qbarfrom() (page 910)).
References
1. [AbramowitzStegun] (page 1910)
2. [WhittakerWatson] (page 1910)
Elliptic arguments
qfrom()
mpmath.qfrom(**kwargs)
Returns the elliptic nome q, given any of q, m, k, , q:
5.15. Welcome to mpmaths documentation!
909
SymPy Documentation, Release 0.7.6
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> qfrom(q=0.25)
0.25
>>> qfrom(m=mfrom(q=0.25))
0.25
>>> qfrom(k=kfrom(q=0.25))
0.25
>>> qfrom(tau=taufrom(q=0.25))
(0.25 + 0.0j)
>>> qfrom(qbar=qbarfrom(q=0.25))
0.25
qbarfrom()
mpmath.qbarfrom(**kwargs)
Returns the number-theoretic nome q, given any of q, m, k, , q:
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> qbarfrom(qbar=0.25)
0.25
>>> qbarfrom(q=qfrom(qbar=0.25))
0.25
>>> qbarfrom(m=extraprec(20)(mfrom)(qbar=0.25))
0.25
>>> qbarfrom(k=extraprec(20)(kfrom)(qbar=0.25))
0.25
>>> qbarfrom(tau=taufrom(qbar=0.25))
(0.25 + 0.0j)
# ill-conditioned
# ill-conditioned
mfrom()
mpmath.mfrom(**kwargs)
Returns the elliptic parameter m, given any of q, m, k, , q:
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> mfrom(m=0.25)
0.25
>>> mfrom(q=qfrom(m=0.25))
0.25
>>> mfrom(k=kfrom(m=0.25))
0.25
>>> mfrom(tau=taufrom(m=0.25))
(0.25 + 0.0j)
>>> mfrom(qbar=qbarfrom(m=0.25))
0.25
As q 1 and q 1, m rapidly approaches 1 and respectively:
>>> mfrom(q=0.75)
0.9999999999999798332943533
>>> mfrom(q=-0.75)
-49586681013729.32611558353
>>> mfrom(q=1)
1.0
910
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> mfrom(q=-1)
-inf
The inverse nome as a function of q has an integer Taylor series expansion:
>>> taylor(lambda q: mfrom(q), 0, 7)
[0.0, 16.0, -128.0, 704.0, -3072.0, 11488.0, -38400.0, 117632.0]
kfrom()
mpmath.kfrom(**kwargs)
Returns the elliptic modulus k, given any of q, m, k, , q:
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> kfrom(k=0.25)
0.25
>>> kfrom(m=mfrom(k=0.25))
0.25
>>> kfrom(q=qfrom(k=0.25))
0.25
>>> kfrom(tau=taufrom(k=0.25))
(0.25 + 0.0j)
>>> kfrom(qbar=qbarfrom(k=0.25))
0.25
As q 1 and q 1, k rapidly approaches 1 and i respectively:
>>> kfrom(q=0.75)
0.9999999999999899166471767
>>> kfrom(q=-0.75)
(0.0 + 7041781.096692038332790615j)
>>> kfrom(q=1)
1
>>> kfrom(q=-1)
(0.0 + +infj)
taufrom()
mpmath.taufrom(**kwargs)
Returns the elliptic half-period ratio , given any of q, m, k, , q:
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> taufrom(tau=0.5j)
(0.0 + 0.5j)
>>> taufrom(q=qfrom(tau=0.5j))
(0.0 + 0.5j)
>>> taufrom(m=mfrom(tau=0.5j))
(0.0 + 0.5j)
>>> taufrom(k=kfrom(tau=0.5j))
(0.0 + 0.5j)
>>> taufrom(qbar=qbarfrom(tau=0.5j))
(0.0 + 0.5j)
Legendre elliptic integrals
5.15. Welcome to mpmaths documentation!
911
SymPy Documentation, Release 0.7.6
ellipk()
mpmath.ellipk(m, **kwargs)
Evaluates the complete elliptic integral of the rst kind, K(m), dened by
K(m) =
0
/2
dt
= 2 F1
2
2
1 m sin t
)
1 1
, , 1, m .
2 2
Note that the argument is the parameter m = k 2 , not the modulus k which is sometimes
used.
Plots
# Complete elliptic integrals K(m) and E(m)
plot([ellipk, ellipe], [-2,1], [0,3], points=600)
Examples
Values and limits include:
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> ellipk(0)
1.570796326794896619231322
>>> ellipk(inf)
(0.0 + 0.0j)
>>> ellipk(-inf)
0.0
>>> ellipk(1)
+inf
912
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> ellipk(-1)
1.31102877714605990523242
>>> ellipk(2)
(1.31102877714605990523242 - 1.31102877714605990523242j)
Verifying the dening integral and hypergeometric representation:
>>> ellipk(0.5)
1.85407467730137191843385
>>> quad(lambda t: (1-0.5*sin(t)**2)**-0.5, [0, pi/2])
1.85407467730137191843385
>>> pi/2*hyp2f1(0.5,0.5,1,0.5)
1.85407467730137191843385
Evaluation is supported for arbitrary complex m:
>>> ellipk(3+4j)
(0.9111955638049650086562171 + 0.6313342832413452438845091j)
A denite integral:
>>> quad(ellipk, [0, 1])
2.0
ellipf()
mpmath.ellipf(phi, m)
Evaluates the Legendre incomplete elliptic integral of the rst kind
F (, m) =
0
dt
1 m sin2 t
or equivalently
sin
F (, m) =
0
dt
(
) (
).
1 t2
1 mt2
The function reduces to a complete elliptic integral of the rst kind (see ellipk()
(page 912)) when = 2 ; that is,
( )
F
, m = K(m).
2
In the dening integral, it is assumed that the principal branch of the square root is
taken and that the path of integration avoids crossing any branch cuts. Outside /2
<() /2, the function extends quasi-periodically as
F ( + n, m) = 2nK(m) + F (, m), n Z.
Plots
# Elliptic integral F(z,m) for some different m
f1 = lambda z: ellipf(z,-1)
f2 = lambda z: ellipf(z,-0.5)
f3 = lambda z: ellipf(z,0)
f4 = lambda z: ellipf(z,0.5)
f5 = lambda z: ellipf(z,1)
plot([f1,f2,f3,f4,f5], [0,pi], [0,4])
5.15. Welcome to mpmaths documentation!
913
SymPy Documentation, Release 0.7.6
Examples
Basic values and limits:
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> ellipf(0,1)
0.0
>>> ellipf(0,0)
0.0
>>> ellipf(1,0); ellipf(2+3j,0)
1.0
(2.0 + 3.0j)
>>> ellipf(1,1); log(sec(1)+tan(1))
1.226191170883517070813061
1.226191170883517070813061
>>> ellipf(pi/2, -0.5); ellipk(-0.5)
1.415737208425956198892166
1.415737208425956198892166
>>> ellipf(pi/2+eps, 1); ellipf(-pi/2-eps, 1)
+inf
+inf
>>> ellipf(1.5, 1)
3.340677542798311003320813
Comparing with numerical integration:
914
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> z,m = 0.5, 1.25
>>> ellipf(z,m)
0.5287219202206327872978255
>>> quad(lambda t: (1-m*sin(t)**2)**(-0.5), [0,z])
0.5287219202206327872978255
The arguments may be complex numbers:
>>> ellipf(3j, 0.5)
(0.0 + 1.713602407841590234804143j)
>>> ellipf(3+4j, 5-6j)
(1.269131241950351323305741 - 0.3561052815014558335412538j)
>>> z,m = 2+3j, 1.25
>>> k = 1011
>>> ellipf(z+pi*k,m); ellipf(z,m) + 2*k*ellipk(m)
(4086.184383622179764082821 - 3003.003538923749396546871j)
(4086.184383622179764082821 - 3003.003538923749396546871j)
For |<(z)| < /2, the function can be expressed as a hypergeometric series of two variables (see appellf1() (page 903)):
>>> z,m = 0.5, 0.25
>>> ellipf(z,m)
0.5050887275786480788831083
>>> sin(z)*appellf1(0.5,0.5,0.5,1.5,sin(z)**2,m*sin(z)**2)
0.5050887275786480788831083
ellipe()
mpmath.ellipe(*args)
Called with a single argument m, evaluates the Legendre complete elliptic integral of
the second kind, E(m), dened by
/2
E(m) =
0
(
)
1 1
2
1 m sin t dt = 2 F1
, , 1, m .
2
2 2
Called with two arguments , m, evaluates the incomplete elliptic integral of the second
kind
E(, m) =
2
1 m sin t dt =
sin z
1 mt2
dt.
1 t2
The incomplete integral reduces to a complete integral when =
( )
E
, m = E(m).
2
2;
that is,
In the dening integral, it is assumed that the principal branch of the square root is
taken and that the path of integration avoids crossing any branch cuts. Outside /2
<(z) /2, the function extends quasi-periodically as
E( + n, m) = 2nE(m) + F (, m), n Z.
Plots
5.15. Welcome to mpmaths documentation!
915
SymPy Documentation, Release 0.7.6
# Elliptic integral E(z,m) for some different m
f1 = lambda z: ellipe(z,-2)
f2 = lambda z: ellipe(z,-1)
f3 = lambda z: ellipe(z,0)
f4 = lambda z: ellipe(z,1)
f5 = lambda z: ellipe(z,2)
plot([f1,f2,f3,f4,f5], [0,pi], [0,4])
Examples for the complete integral
Basic values and limits:
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> ellipe(0)
1.570796326794896619231322
>>> ellipe(1)
1.0
>>> ellipe(-1)
1.910098894513856008952381
>>> ellipe(2)
(0.5990701173677961037199612 + 0.5990701173677961037199612j)
>>> ellipe(inf)
(0.0 + +infj)
>>> ellipe(-inf)
+inf
Verifying the dening integral and hypergeometric representation:
916
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> ellipe(0.5)
1.350643881047675502520175
>>> quad(lambda t: sqrt(1-0.5*sin(t)**2), [0, pi/2])
1.350643881047675502520175
>>> pi/2*hyp2f1(0.5,-0.5,1,0.5)
1.350643881047675502520175
Evaluation is supported for arbitrary complex m:
>>> ellipe(0.5+0.25j)
(1.360868682163129682716687 - 0.1238733442561786843557315j)
>>> ellipe(3+4j)
(1.499553520933346954333612 - 1.577879007912758274533309j)
A denite integral:
>>> quad(ellipe, [0,1])
1.333333333333333333333333
Examples for the incomplete integral
Basic values and limits:
>>> ellipe(0,1)
0.0
>>> ellipe(0,0)
0.0
>>> ellipe(1,0)
1.0
>>> ellipe(2+3j,0)
(2.0 + 3.0j)
>>> ellipe(1,1); sin(1)
0.8414709848078965066525023
0.8414709848078965066525023
>>> ellipe(pi/2, -0.5); ellipe(-0.5)
1.751771275694817862026502
1.751771275694817862026502
>>> ellipe(pi/2, 1); ellipe(-pi/2, 1)
1.0
-1.0
>>> ellipe(1.5, 1)
0.9974949866040544309417234
Comparing with numerical integration:
>>> z,m = 0.5, 1.25
>>> ellipe(z,m)
0.4740152182652628394264449
>>> quad(lambda t: sqrt(1-m*sin(t)**2), [0,z])
0.4740152182652628394264449
The arguments may be complex numbers:
>>> ellipe(3j, 0.5)
(0.0 + 7.551991234890371873502105j)
>>> ellipe(3+4j, 5-6j)
(24.15299022574220502424466 + 75.2503670480325997418156j)
>>> k = 35
>>> z,m = 2+3j, 1.25
>>> ellipe(z+pi*k,m); ellipe(z,m) + 2*k*ellipe(m)
5.15. Welcome to mpmaths documentation!
917
SymPy Documentation, Release 0.7.6
(48.30138799412005235090766 + 17.47255216721987688224357j)
(48.30138799412005235090766 + 17.47255216721987688224357j)
For |<(z)| < /2, the function can be expressed as a hypergeometric series of two variables (see appellf1() (page 903)):
>>> z,m = 0.5, 0.25
>>> ellipe(z,m)
0.4950017030164151928870375
>>> sin(z)*appellf1(0.5,0.5,-0.5,1.5,sin(z)**2,m*sin(z)**2)
0.4950017030164151928870376
ellippi()
mpmath.ellippi(*args)
Called with three arguments n, , m, evaluates the Legendre incomplete elliptic integral
of the third kind
sin
dt
dt
(n; , m) =
=
.
2
2
2
(1 nt ) 1 t2 1 mt2
0 (1 n sin t) 1 m sin t
0
Called with two arguments n, m, evaluates the complete elliptic integral of the third kind
(n, m) = (n; 2 , m).
In the dening integral, it is assumed that the principal branch of the square root is
taken and that the path of integration avoids crossing any branch cuts. Outside /2
<() /2, the function extends quasi-periodically as
(n, + k, m) = 2k(n, m) + (n, , m), k Z.
Plots
# Elliptic integral Pi(n,z,m) for some different n, m
f1 = lambda z: ellippi(0.9,z,0.9)
f2 = lambda z: ellippi(0.5,z,0.5)
f3 = lambda z: ellippi(-2,z,-0.9)
f4 = lambda z: ellippi(-0.5,z,0.5)
f5 = lambda z: ellippi(-1,z,0.5)
plot([f1,f2,f3,f4,f5], [0,pi], [0,4])
918
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples for the complete integral
Some basic values and limits:
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> ellippi(0,-5); ellipk(-5)
0.9555039270640439337379334
0.9555039270640439337379334
>>> ellippi(inf,2)
0.0
>>> ellippi(2,inf)
0.0
>>> abs(ellippi(1,5))
+inf
>>> abs(ellippi(0.25,1))
+inf
Evaluation in terms of simpler functions:
>>> ellippi(0.25,0.25); ellipe(0.25)/(1-0.25)
1.956616279119236207279727
1.956616279119236207279727
>>> ellippi(3,0); pi/(2*sqrt(-2))
(0.0 - 1.11072073453959156175397j)
(0.0 - 1.11072073453959156175397j)
>>> ellippi(-3,0); pi/(2*sqrt(4))
0.7853981633974483096156609
0.7853981633974483096156609
5.15. Welcome to mpmaths documentation!
919
SymPy Documentation, Release 0.7.6
Examples for the incomplete integral
Basic values and limits:
>>> ellippi(0.25,-0.5); ellippi(0.25,pi/2,-0.5)
1.622944760954741603710555
1.622944760954741603710555
>>> ellippi(1,0,1)
0.0
>>> ellippi(inf,0,1)
0.0
>>> ellippi(0,0.25,0.5); ellipf(0.25,0.5)
0.2513040086544925794134591
0.2513040086544925794134591
>>> ellippi(1,1,1); (log(sec(1)+tan(1))+sec(1)*tan(1))/2
2.054332933256248668692452
2.054332933256248668692452
>>> ellippi(0.25, 53*pi/2, 0.75); 53*ellippi(0.25,0.75)
135.240868757890840755058
135.240868757890840755058
>>> ellippi(0.5,pi/4,0.5); 2*ellipe(pi/4,0.5)-1/sqrt(3)
0.9190227391656969903987269
0.9190227391656969903987269
Complex arguments are supported:
>>> ellippi(0.5, 5+6j-2*pi, -7-8j)
(-0.3612856620076747660410167 + 0.5217735339984807829755815j)
Some degenerate cases:
>>> ellippi(1,1)
+inf
>>> ellippi(1,0)
+inf
>>> ellippi(1,2,0)
+inf
>>> ellippi(1,2,1)
+inf
>>> ellippi(1,0,1)
0.0
Carlson symmetric elliptic integrals
elliprf()
mpmath.elliprf(x, y, z)
Evaluates the Carlson symmetric elliptic integral of the rst kind
1
dt
RF (x, y, z) =
2 0
(t + x)(t + y)(t + z)
which is dened for x, y, z
/ (, 0), and with at most one of x, y, z being zero.
For real x, y, z 0, the principal square root is taken in the integrand. For complex x, y, z,
the principal square root is taken as t and as t 0 non-principal branches are
chosen as necessary so as to make the integrand continuous.
Examples
920
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Some basic values and limits:
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> elliprf(0,1,1); pi/2
1.570796326794896619231322
1.570796326794896619231322
>>> elliprf(0,1,inf)
0.0
>>> elliprf(1,1,1)
1.0
>>> elliprf(2,2,2)**2
0.5
>>> elliprf(1,0,0); elliprf(0,0,1); elliprf(0,1,0); elliprf(0,0,0)
+inf
+inf
+inf
+inf
Representing complete elliptic integrals in terms of RF :
>>> m = mpf(0.75)
>>> ellipk(m); elliprf(0,1-m,1)
2.156515647499643235438675
2.156515647499643235438675
>>> ellipe(m); elliprf(0,1-m,1)-m*elliprd(0,1-m,1)/3
1.211056027568459524803563
1.211056027568459524803563
Some symmetries and argument transformations:
>>> x,y,z = 2,3,4
>>> elliprf(x,y,z); elliprf(y,x,z); elliprf(z,y,x)
0.5840828416771517066928492
0.5840828416771517066928492
0.5840828416771517066928492
>>> k = mpf(100000)
>>> elliprf(k*x,k*y,k*z); k**(-0.5) * elliprf(x,y,z)
0.001847032121923321253219284
0.001847032121923321253219284
>>> l = sqrt(x*y) + sqrt(y*z) + sqrt(z*x)
>>> elliprf(x,y,z); 2*elliprf(x+l,y+l,z+l)
0.5840828416771517066928492
0.5840828416771517066928492
>>> elliprf((x+l)/4,(y+l)/4,(z+l)/4)
0.5840828416771517066928492
Comparing with numerical integration:
>>> x,y,z = 2,3,4
>>> elliprf(x,y,z)
0.5840828416771517066928492
>>> f = lambda t: 0.5*((t+x)*(t+y)*(t+z))**(-0.5)
>>> q = extradps(25)(quad)
>>> q(f, [0,inf])
0.5840828416771517066928492
With the following arguments, the square root in the integrand becomes
discontinuous
at t = 1/2 ifthe principal branch is used. To obtain the right value, r must be taken
instead of r on t (0, 1/2):
5.15. Welcome to mpmaths documentation!
921
SymPy Documentation, Release 0.7.6
>>> x,y,z = j-1,j,0
>>> elliprf(x,y,z)
(0.7961258658423391329305694 - 1.213856669836495986430094j)
>>> -q(f, [0,0.5]) + q(f, [0.5,inf])
(0.7961258658423391329305694 - 1.213856669836495986430094j)
The so-called rst lemniscate constant, a transcendental number:
>>> elliprf(0,1,2)
1.31102877714605990523242
>>> extradps(25)(quad)(lambda t: 1/sqrt(1-t**4), [0,1])
1.31102877714605990523242
>>> gamma(1/4)**2/(4*sqrt(2*pi))
1.31102877714605990523242
References
1.[Carlson] (page 1910)
2.[DLMF] (page 1910) Chapter 19. Elliptic Integrals
elliprc()
mpmath.elliprc(x, y, pv=True)
Evaluates the degenerate Carlson symmetric elliptic integral of the rst kind
1
dt
RC (x, y) = RF (x, y, y) =
.
2 0 (t + y) (t + x)
If y (, 0), either a value dened by continuity, or with pv=True the Cauchy principal
value, can be computed.
If x 0, y > 0, the value can be expressed in terms of elementary functions as
( )
x
1
cos
,
x<y
x
y
1
x=y.
RC (x, y) = ,
y
(
)
cosh
, x>y
xy
y
Examples
Some special values and limits:
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> elliprc(1,2)*4; elliprc(0,1)*2; +pi
3.141592653589793238462643
3.141592653589793238462643
3.141592653589793238462643
>>> elliprc(1,0)
+inf
>>> elliprc(5,5)**2
0.2
>>> elliprc(1,inf); elliprc(inf,1); elliprc(inf,inf)
0.0
0.0
0.0
922
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Comparing with the elementary closed-form solution:
>>> elliprc(1/3, 1/5); sqrt(7.5)*acosh(sqrt(5/3))
2.041630778983498390751238
2.041630778983498390751238
>>> elliprc(1/5, 1/3); sqrt(7.5)*acos(sqrt(3/5))
1.875180765206547065111085
1.875180765206547065111085
Comparing with numerical integration:
>>> q = extradps(25)(quad)
>>> elliprc(2, -3, pv=True)
0.3333969101113672670749334
>>> elliprc(2, -3, pv=False)
(0.3333969101113672670749334 + 0.7024814731040726393156375j)
>>> 0.5*q(lambda t: 1/(sqrt(t+2)*(t-3)), [0,3-j,6,inf])
(0.3333969101113672670749334 + 0.7024814731040726393156375j)
elliprj()
mpmath.elliprj(x, y, z, p)
Evaluates the Carlson symmetric elliptic integral of the third kind
3
dt
RJ (x, y, z, p) =
.
2 0 (t + p) (t + x)(t + y)(t + z)
Like elliprf() (page 920), the branch of the square root in the integrand is dened so
as to be continuous along the path of integration for complex values of the arguments.
Examples
Some values and limits:
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> elliprj(1,1,1,1)
1.0
>>> elliprj(2,2,2,2); 1/(2*sqrt(2))
0.3535533905932737622004222
0.3535533905932737622004222
>>> elliprj(0,1,2,2)
1.067937989667395702268688
>>> 3*(2*gamma(5/4)**2-pi**2/gamma(1/4)**2)/(sqrt(2*pi))
1.067937989667395702268688
>>> elliprj(0,1,1,2); 3*pi*(2-sqrt(2))/4
1.380226776765915172432054
1.380226776765915172432054
>>> elliprj(1,3,2,0); elliprj(0,1,1,0); elliprj(0,0,0,0)
+inf
+inf
+inf
>>> elliprj(1,inf,1,0); elliprj(1,1,1,inf)
0.0
0.0
>>> chop(elliprj(1+j, 1-j, 1, 1))
0.8505007163686739432927844
Scale transformation:
5.15. Welcome to mpmaths documentation!
923
SymPy Documentation, Release 0.7.6
>>> x,y,z,p = 2,3,4,5
>>> k = mpf(100000)
>>> elliprj(k*x,k*y,k*z,k*p); k**(-1.5)*elliprj(x,y,z,p)
4.521291677592745527851168e-9
4.521291677592745527851168e-9
Comparing with numerical integration:
>>> elliprj(1,2,3,4)
0.2398480997495677621758617
>>> f = lambda t: 1/((t+4)*sqrt((t+1)*(t+2)*(t+3)))
>>> 1.5*quad(f, [0,inf])
0.2398480997495677621758617
>>> elliprj(1,2+1j,3,4-2j)
(0.216888906014633498739952 + 0.04081912627366673332369512j)
>>> f = lambda t: 1/((t+4-2j)*sqrt((t+1)*(t+2+1j)*(t+3)))
>>> 1.5*quad(f, [0,inf])
(0.216888906014633498739952 + 0.04081912627366673332369511j)
elliprd()
mpmath.elliprd(x, y, z)
Evaluates the degenerate Carlson symmetric elliptic integral of the third kind or Carlson
elliptic integral of the second kind RD (x, y, z) = RJ (x, y, z, z).
See elliprj() (page 923) for additional information.
Examples
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> elliprd(1,2,3)
0.2904602810289906442326534
>>> elliprj(1,2,3,3)
0.2904602810289906442326534
The so-called second lemniscate constant, a transcendental number:
>>> elliprd(0,2,1)/3
0.5990701173677961037199612
>>> extradps(25)(quad)(lambda t: t**2/sqrt(1-t**4), [0,1])
0.5990701173677961037199612
>>> gamma(3/4)**2/sqrt(2*pi)
0.5990701173677961037199612
elliprg()
mpmath.elliprg(x, y, z)
Evaluates the Carlson completely symmetric elliptic integral of the second kind
)
(
1
t
y
z
x
RG (x, y, z) =
+
+
dt.
4 0
(t + x)(t + y)(t + z) t + x t + y t + z
Examples
Evaluation for real and complex arguments:
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
924
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> elliprg(0,1,1)*4; +pi
3.141592653589793238462643
3.141592653589793238462643
>>> elliprg(0,0.5,1)
0.6753219405238377512600874
>>> chop(elliprg(1+j, 1-j, 2))
1.172431327676416604532822
A double integral that can be evaluated in terms of RG :
>>> x,y,z = 2,3,4
>>> def f(t,u):
...
st = fp.sin(t); ct = fp.cos(t)
...
su = fp.sin(u); cu = fp.cos(u)
...
return (x*(st*cu)**2 + y*(st*su)**2 + z*ct**2)**0.5 * st
...
>>> nprint(mpf(fp.quad(f, [0,fp.pi], [0,2*fp.pi])/(4*fp.pi)), 13)
1.725503028069
>>> nprint(elliprg(x,y,z), 13)
1.725503028069
Jacobi theta functions
jtheta()
mpmath.jtheta(n, z, q, derivative=0)
Computes the Jacobi theta function n (z, q), where n = 1, 2, 3, 4, dened by the innite
series:
1 (z, q) = 2q 1/4
2 (z, q) = 2q
(1)n q n
n=0
1/4
n2 +n
+n
sin((2n + 1)z)
cos((2n + 1)z)
n=0
3 (z, q) = 1 + 2
q n cos(2nz)
n=1
4 (z, q) = 1 + 2
(q)n cos(2nz)
n=1
The theta functions are functions of two variables:
z is the argument, an arbitrary real or complex number
q is the nome, which must be a real or complex number in the unit disk (i.e. |q| <
1). For |q| 1, the series converge very quickly, so the Jacobi theta functions can
eciently be evaluated to high precision.
The compact notations n (q) = n (0, q) and n = n (0, q) are also frequently encountered.
Finally, Jacobi theta functions are frequently considered as functions of the half-period
ratio and then usually denoted by n (z| ).
Optionally, jtheta(n, z, q, derivative=d) with d > 0 computes a d-th derivative with
respect to z.
Examples and basic properties
Considered as functions of z, the Jacobi theta functions may be viewed as generalizations
of the ordinary trigonometric functions cos and sin. They are periodic functions:
5.15. Welcome to mpmaths documentation!
925
SymPy Documentation, Release 0.7.6
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> jtheta(1, 0.25, 0.2)
0.2945120798627300045053104
>>> jtheta(1, 0.25 + 2*pi, 0.2)
0.2945120798627300045053104
Indeed, the series dening the theta functions are essentially trigonometric Fourier series. The coecients can be retrieved using fourier() (page 1017):
>>> mp.dps = 10
>>> nprint(fourier(lambda x: jtheta(2, x, 0.5), [-pi, pi], 4))
([0.0, 1.68179, 0.0, 0.420448, 0.0], [0.0, 0.0, 0.0, 0.0, 0.0])
The Jacobi theta functions are also so-called quasiperiodic functions of z and , meaning
that for xed , n (z, q) and n (z + , q) are the same except for an exponential factor:
>>> mp.dps = 25
>>> tau = 3*j/10
>>> q = exp(pi*j*tau)
>>> z = 10
>>> jtheta(4, z+tau*pi, q)
(-0.682420280786034687520568 + 1.526683999721399103332021j)
>>> -exp(-2*j*z)/q * jtheta(4, z, q)
(-0.682420280786034687520568 + 1.526683999721399103332021j)
The Jacobi theta functions satisfy a huge number of other functional equations, such as
the following identity (valid for any q):
>>> q = mpf(3)/10
>>> jtheta(3,0,q)**4
6.823744089352763305137427
>>> jtheta(2,0,q)**4 + jtheta(4,0,q)**4
6.823744089352763305137427
Extensive listings of identities satised by the Jacobi theta functions can be found in
standard reference works.
The Jacobi theta functions are related to the gamma function for special arguments:
>>> jtheta(3, 0, exp(-pi))
1.086434811213308014575316
>>> pi**(1/4.) / gamma(3/4.)
1.086434811213308014575316
jtheta() (page 925) supports arbitrary precision evaluation and complex arguments:
>>> mp.dps = 50
>>> jtheta(4, sqrt(2), 0.5)
2.0549510717571539127004115835148878097035750653737
>>> mp.dps = 25
>>> jtheta(4, 1+2j, (1+j)/5)
(7.180331760146805926356634 - 1.634292858119162417301683j)
Evaluation of derivatives:
>>> mp.dps = 25
>>> jtheta(1, 7, 0.25, 1); diff(lambda z: jtheta(1, z, 0.25), 7)
1.209857192844475388637236
1.209857192844475388637236
926
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> jtheta(1, 7, 0.25, 2); diff(lambda
-0.2598718791650217206533052
-0.2598718791650217206533052
>>> jtheta(2, 7, 0.25, 1); diff(lambda
-1.150231437070259644461474
-1.150231437070259644461474
>>> jtheta(2, 7, 0.25, 2); diff(lambda
-0.6226636990043777445898114
-0.6226636990043777445898114
>>> jtheta(3, 7, 0.25, 1); diff(lambda
-0.9990312046096634316587882
-0.9990312046096634316587882
>>> jtheta(3, 7, 0.25, 2); diff(lambda
-0.1530388693066334936151174
-0.1530388693066334936151174
>>> jtheta(4, 7, 0.25, 1); diff(lambda
0.9820995967262793943571139
0.9820995967262793943571139
>>> jtheta(4, 7, 0.25, 2); diff(lambda
0.3936902850291437081667755
0.3936902850291437081667755
z: jtheta(1, z, 0.25), 7, 2)
z: jtheta(2, z, 0.25), 7)
z: jtheta(2, z, 0.25), 7, 2)
z: jtheta(3, z, 0.25), 7)
z: jtheta(3, z, 0.25), 7, 2)
z: jtheta(4, z, 0.25), 7)
z: jtheta(4, z, 0.25), 7, 2)
Possible issues
For |q| 1 or =( ) 0, jtheta() (page 925) raises ValueError. This exception is also
raised for |q| extremely close to 1 (or equivalently very close to 0), since the series
would converge too slowly:
>>> jtheta(1, 10, 0.99999999 * exp(0.5*j))
Traceback (most recent call last):
...
ValueError: abs(q) > THETA_Q_LIM = 1.000000
Traceback (most recent call last):
...
ValueError: abs(q) > THETA_Q_LIM = 1.000000
Jacobi elliptic functions
ellipfun()
mpmath.ellipfun(kind, u=None, m=None, q=None, k=None, tau=None)
Computes any of the Jacobi elliptic functions, dened in terms of Jacobi theta functions
as
3 (0, q) 1 (t, q)
2 (0, q) 4 (t, q)
4 (0, q) 2 (t, q)
cn(u, m) =
2 (0, q) 4 (t, q)
4 (0, q) 3 (t, q)
,
dn(u, m) =
3 (0, q) 4 (t, q)
sn(u, m) =
or more generally computes a ratio of two such functions. Here t = u/3 (0, q)2 , and
q = q(m) denotes the nome (see nome()). Optionally, you can specify the nome directly
instead of m by passing q=<value>, or you can directly specify the elliptic parameter k
with k=<value>.
5.15. Welcome to mpmaths documentation!
927
SymPy Documentation, Release 0.7.6
The rst argument should be a two-character string specifying the function using any
combination of s, c, d, n. These letters respectively denote the basic functions
sn(u, m), cn(u, m), dn(u, m), and 1. The identier species the ratio of two such functions.
For example, ns identies the function
ns(u, m) =
1
sn(u, m)
cd(u, m) =
cn(u, m)
.
dn(u, m)
and cd identies the function
If called with only the rst argument, a function object evaluating the chosen function
for given arguments is returned.
Examples
Basic evaluation:
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> ellipfun(cd, 3.5, 0.5)
-0.9891101840595543931308394
>>> ellipfun(cd, 3.5, q=0.25)
0.07111979240214668158441418
The sn-function is doubly periodic in the complex plane with periods 4K(m) and 2iK(1m)
(see ellipk() (page 912)):
>>> sn = ellipfun(sn)
>>> sn(2, 0.25)
0.9628981775982774425751399
>>> sn(2+4*ellipk(0.25), 0.25)
0.9628981775982774425751399
>>> chop(sn(2+2*j*ellipk(1-0.25), 0.25))
0.9628981775982774425751399
The cn-function is doubly periodic with periods 4K(m) and 4iK(1 m):
>>> cn = ellipfun(cn)
>>> cn(2, 0.25)
-0.2698649654510865792581416
>>> cn(2+4*ellipk(0.25), 0.25)
-0.2698649654510865792581416
>>> chop(cn(2+4*j*ellipk(1-0.25), 0.25))
-0.2698649654510865792581416
The dn-function is doubly periodic with periods 2K(m) and 4iK(1 m):
>>> dn = ellipfun(dn)
>>> dn(2, 0.25)
0.8764740583123262286931578
>>> dn(2+2*ellipk(0.25), 0.25)
0.8764740583123262286931578
>>> chop(dn(2+4*j*ellipk(1-0.25), 0.25))
0.8764740583123262286931578
Modular functions
928
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
kleinj()
mpmath.kleinj(tau=None, **kwargs)
Evaluates the Klein j-invariant, which is a modular function dened for in the upper
half-plane as
J( ) =
g23 ( )
g23 ( ) 27g32 ( )
where g2 and g3 are the modular invariants of the Weierstrass elliptic function,
(m + n)4
g2 ( ) = 60
(m,n)Z2 \(0,0)
g3 ( ) = 140
(m + n)6 .
(m,n)Z2 \(0,0)
An alternative, common notation is that of the j-function j( ) = 1728J( ).
Plots
# Klein J-function as function of the number-theoretic nome
fp.cplot(lambda q: fp.kleinj(qbar=q), [-1,1], [-1,1], points=50000)
# Klein J-function as function of the half-period ratio
fp.cplot(lambda t: fp.kleinj(tau=t), [-1,2], [0,1.5], points=50000)
5.15. Welcome to mpmaths documentation!
929
SymPy Documentation, Release 0.7.6
Examples
Verifying the functional equation J( ) = J( + 1) = J( 1 ):
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> tau = 0.625+0.75*j
>>> tau = 0.625+0.75*j
>>> kleinj(tau)
(-0.1507492166511182267125242 + 0.07595948379084571927228948j)
>>> kleinj(tau+1)
(-0.1507492166511182267125242 + 0.07595948379084571927228948j)
>>> kleinj(-1/tau)
(-0.1507492166511182267125242 + 0.07595948379084571927228946j)
The j-function has a famous Laurent series expansion in terms of the nome q, j( ) =
q1 + 744 + 196884
q + . . .:
>>> mp.dps = 15
>>> taylor(lambda q: 1728*q*kleinj(qbar=q), 0, 5, singular=True)
[1.0, 744.0, 196884.0, 21493760.0, 864299970.0, 20245856256.0]
The j-function admits exact evaluation at special algebraic points related to the Heegner
numbers 1, 2, 3, 7, 11, 19, 43, 67, 163:
>>> @extraprec(10)
... def h(n):
...
v = (1+sqrt(n)*j)
...
if n > 2:
930
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
...
v *= 0.5
...
return v
...
>>> mp.dps = 25
>>> for n in [1,2,3,7,11,19,43,67,163]:
...
n, chop(1728*kleinj(h(n)))
...
(1, 1728.0)
(2, 8000.0)
(3, 0.0)
(7, -3375.0)
(11, -32768.0)
(19, -884736.0)
(43, -884736000.0)
(67, -147197952000.0)
(163, -262537412640768000.0)
Also at other special points, the j-function assumes explicit algebraic values, e.g.:
>>> chop(1728*kleinj(j*sqrt(5)))
1264538.909475140509320227
>>> identify(cbrt(_))
# note: not simplified
((100+sqrt(13520))/2)
>>> (50+26*sqrt(5))**3
1264538.909475140509320227
Zeta functions, L-series and polylogarithms
This section includes the Riemann zeta functions and associated functions pertaining to analytic number theory.
Riemann and Hurwitz zeta functions
zeta()
mpmath.zeta(s, a=1, derivative=0)
Computes the Riemann zeta function
(s) = 1 +
1
1
1
+ s + s + ...
2s
3
4
or, with a 6= 1, the more general Hurwitz zeta function
(s, a) =
k=0
1
.
(a + k)s
Optionally, zeta(s, a, n) computes the n-th derivative with respect to s,
(n) (s, a) = (1)n
log (a + k)
k=0
(a + k)s
Although these series only converge for <(s) > 1, the Riemann and Hurwitz zeta functions
are dened through analytic continuation for arbitrary complex s 6= 1 (s = 1 is a pole).
The implementation uses three algorithms: the Borwein algorithm for the Riemann zeta
function when s is close to the real line; the Riemann-Siegel formula for the Riemann zeta
5.15. Welcome to mpmaths documentation!
931
SymPy Documentation, Release 0.7.6
function when s is large imaginary, and Euler-Maclaurin summation in all other cases.
The reection formula for <(s) < 0 is implemented in some cases. The algorithm can
be chosen with method = borwein, method=riemann-siegel or method = eulermaclaurin.
The parameter a is usually a rational number a = p/q, and may be specied as such by
passing an integer tuple (p, q). Evaluation is supported for arbitrary complex a, but may
be slow and/or inaccurate when <(s) < 0 for nonrational a or when computing derivatives.
Examples
Some values of the Riemann zeta function:
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> zeta(2); pi**2 / 6
1.644934066848226436472415
1.644934066848226436472415
>>> zeta(0)
-0.5
>>> zeta(-1)
-0.08333333333333333333333333
>>> zeta(-2)
0.0
For large positive s, (s) rapidly approaches 1:
>>> zeta(50)
1.000000000000000888178421
>>> zeta(100)
1.0
>>> zeta(inf)
1.0
>>> 1-sum((zeta(k)-1)/k for k in range(2,85)); +euler
0.5772156649015328606065121
0.5772156649015328606065121
>>> nsum(lambda k: zeta(k)-1, [2, inf])
1.0
Evaluation is supported for complex s and a:
>>> zeta(-3+4j)
(-0.03373057338827757067584698 + 0.2774499251557093745297677j)
>>> zeta(2+3j, -1+j)
(389.6841230140842816370741 + 295.2674610150305334025962j)
The Riemann zeta function has so-called nontrivial zeros on the critical line s = 1/2 + it:
>>> findroot(zeta, 0.5+14j); zetazero(1)
(0.5 + 14.13472514173469379045725j)
(0.5 + 14.13472514173469379045725j)
>>> findroot(zeta, 0.5+21j); zetazero(2)
(0.5 + 21.02203963877155499262848j)
(0.5 + 21.02203963877155499262848j)
>>> findroot(zeta, 0.5+25j); zetazero(3)
(0.5 + 25.01085758014568876321379j)
(0.5 + 25.01085758014568876321379j)
>>> chop(zeta(zetazero(10)))
0.0
932
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Evaluation on and near the critical line is supported for large heights t by means of the
Riemann-Siegel formula (currently for a = 1, n 4):
>>> zeta(0.5+100000j)
(1.073032014857753132114076 + 5.780848544363503984261041j)
>>> zeta(0.75+1000000j)
(0.9535316058375145020351559 + 0.9525945894834273060175651j)
>>> zeta(0.5+10000000j)
(11.45804061057709254500227 - 8.643437226836021723818215j)
>>> zeta(0.5+100000000j, derivative=1)
(51.12433106710194942681869 + 43.87221167872304520599418j)
>>> zeta(0.5+100000000j, derivative=2)
(-444.2760822795430400549229 - 896.3789978119185981665403j)
>>> zeta(0.5+100000000j, derivative=3)
(3230.72682687670422215339 + 14374.36950073615897616781j)
>>> zeta(0.5+100000000j, derivative=4)
(-11967.35573095046402130602 - 218945.7817789262839266148j)
>>> zeta(1+10000000j)
# off the line
(2.859846483332530337008882 + 0.491808047480981808903986j)
>>> zeta(1+10000000j, derivative=1)
(-4.333835494679647915673205 - 0.08405337962602933636096103j)
>>> zeta(1+10000000j, derivative=4)
(453.2764822702057701894278 - 581.963625832768189140995j)
For investigation of the zeta function zeros, the Riemann-Siegel Z-function is often
more convenient than working with the Riemann zeta function directly (see siegelz()
(page 939)).
Some values of the Hurwitz zeta function:
>>> zeta(2, 3); -5./4 + pi**2/6
0.3949340668482264364724152
0.3949340668482264364724152
>>> zeta(2, (3,4)); pi**2 - 8*catalan
2.541879647671606498397663
2.541879647671606498397663
For positive integer values of s, the Hurwitz zeta function is equivalent to a polygamma
function (except for a normalizing factor):
>>> zeta(4, (1,5)); psi(3, 1/5)/6
625.5408324774542966919938
625.5408324774542966919938
Evaluation of derivatives:
>>> zeta(0, 3+4j, 1); loggamma(3+4j) - ln(2*pi)/2
(-2.675565317808456852310934 + 4.742664438034657928194889j)
(-2.675565317808456852310934 + 4.742664438034657928194889j)
>>> zeta(2, 1, 20)
2432902008176640000.000242
>>> zeta(3+4j, 5.5+2j, 4)
(-0.140075548947797130681075 - 0.3109263360275413251313634j)
>>> zeta(0.5+100000j, 1, 4)
(-10407.16081931495861539236 + 13777.78669862804508537384j)
>>> zeta(-100+0.5j, (1,3), derivative=4)
(4.007180821099823942702249e+79 + 4.916117957092593868321778e+78j)
Generating a Taylor series at s = 2 using derivatives:
5.15. Welcome to mpmaths documentation!
933
SymPy Documentation, Release 0.7.6
>>> for k in range(11): print(%s * (s-2)^%i % (zeta(2,1,k)/fac(k), k))
...
1.644934066848226436472415 * (s-2)^0
-0.9375482543158437537025741 * (s-2)^1
0.9946401171494505117104293 * (s-2)^2
-1.000024300473840810940657 * (s-2)^3
1.000061933072352565457512 * (s-2)^4
-1.000006869443931806408941 * (s-2)^5
1.000000173233769531820592 * (s-2)^6
-0.9999999569989868493432399 * (s-2)^7
0.9999999937218844508684206 * (s-2)^8
-0.9999999996355013916608284 * (s-2)^9
1.000000000004610645020747 * (s-2)^10
Evaluation at zero and for negative integer s:
>>> zeta(0, 10)
-9.5
>>> zeta(-2, (2,3)); mpf(1)/81
0.01234567901234567901234568
0.01234567901234567901234568
>>> zeta(-3+4j, (5,4))
(0.2899236037682695182085988 + 0.06561206166091757973112783j)
>>> zeta(-3.25, 1/pi)
-0.0005117269627574430494396877
>>> zeta(-3.5, pi, 1)
11.156360390440003294709
>>> zeta(-100.5, (8,3))
-4.68162300487989766727122e+77
>>> zeta(-10.5, (-8,3))
(-0.01521913704446246609237979 + 29907.72510874248161608216j)
>>> zeta(-1000.5, (-8,3))
(1.031911949062334538202567e+1770 + 1.519555750556794218804724e+426j)
>>> zeta(-1+j, 3+4j)
(-16.32988355630802510888631 - 22.17706465801374033261383j)
>>> zeta(-1+j, 3+4j, 2)
(32.48985276392056641594055 - 51.11604466157397267043655j)
>>> diff(lambda s: zeta(s, 3+4j), -1+j, 2)
(32.48985276392056641594055 - 51.11604466157397267043655j)
References
1.http://mathworld.wolfram.com/RiemannZetaFunction.html
2.http://mathworld.wolfram.com/HurwitzZetaFunction.html
3.http://www.cecm.sfu.ca/personal/pborwein/PAPERS/P155.pdf
Dirichlet L-series
altzeta()
mpmath.altzeta(s)
Gives the Dirichlet eta function, (s), also known as the alternating zeta function. This
function is dened in analogy with the Riemann zeta function as providing the sum of
934
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
the alternating series
(s) =
(1)k
k=0
ks
=1
1
1
1
+ s s + ...
2s
3
4
The eta function, unlike the Riemann zeta function, is an entire function, having a nite
value for all complex s. The special case (1) = log(2) gives the value of the alternating
harmonic series.
The alternating zeta function may expressed using the Riemann zeta function as (s) =
(1 21s )(s). It can also be expressed in terms of the Hurwitz zeta function, for example
using dirichlet() (page 936) (see documentation for that function).
Examples
Some special values are:
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = True
>>> altzeta(1)
0.693147180559945
>>> altzeta(0)
0.5
>>> altzeta(-1)
0.25
>>> altzeta(-2)
0.0
An example of a sum that can be computed more accurately and eciently via altzeta()
(page 934) than via numerical summation:
>>> sum(-(-1)**n / mpf(n)**2.5 for n in range(1, 100))
0.867204951503984
>>> altzeta(2.5)
0.867199889012184
At positive even integers, the Dirichlet eta function evaluates to a rational multiple of a
power of :
>>> altzeta(2)
0.822467033424113
>>> pi**2/12
0.822467033424113
Like the Riemann zeta function, (s), approaches 1 as s approaches positive innity,
although it does so from below rather than from above:
>>> altzeta(30)
0.999999999068682
>>> altzeta(inf)
1.0
>>> mp.pretty = False
>>> altzeta(1000, rounding=d)
mpf(0.99999999999999989)
>>> altzeta(1000, rounding=u)
mpf(1.0)
References
1.http://mathworld.wolfram.com/DirichletEtaFunction.html
2.http://en.wikipedia.org/wiki/Dirichlet eta function
5.15. Welcome to mpmaths documentation!
935
SymPy Documentation, Release 0.7.6
dirichlet()
mpmath.dirichlet(s, chi, derivative=0)
Evaluates the Dirichlet L-function
L(s, ) =
(k)
k=1
ks
where is a periodic sequence of length q which should be supplied in the form of a
list [(0), (1), . . . , (q 1)]. Strictly, should be a Dirichlet character, but any periodic
sequence will work.
For example, dirichlet(s, [1]) gives the ordinary Riemann zeta function and dirichlet(s, [-1,1]) gives the alternating zeta function (Dirichlet eta function).
Also the derivative with respect to s (currently only a rst derivative) can be evaluated.
Examples
The ordinary Riemann zeta function:
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> dirichlet(3, [1]); zeta(3)
1.202056903159594285399738
1.202056903159594285399738
>>> dirichlet(1, [1])
+inf
The alternating zeta function:
>>> dirichlet(1, [-1,1]); ln(2)
0.6931471805599453094172321
0.6931471805599453094172321
The following denes the Dirichlet beta function (s) =
values of this function:
(1)k
k=0 (2k+1)s
and veries several
>>> B = lambda s, d=0: dirichlet(s, [0, 1, 0, -1], d)
>>> B(0); 1./2
0.5
0.5
>>> B(1); pi/4
0.7853981633974483096156609
0.7853981633974483096156609
>>> B(2); +catalan
0.9159655941772190150546035
0.9159655941772190150546035
>>> B(2,1); diff(B, 2)
0.08158073611659279510291217
0.08158073611659279510291217
>>> B(-1,1); 2*catalan/pi
0.5831218080616375602767689
0.5831218080616375602767689
>>> B(0,1); log(gamma(0.25)**2/(2*pi*sqrt(2)))
0.3915943927068367764719453
0.3915943927068367764719454
>>> B(1,1); 0.25*pi*(euler+2*ln2+3*ln(pi)-4*ln(gamma(0.25)))
0.1929013167969124293631898
0.1929013167969124293631898
A custom L-series of period 3:
936
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> dirichlet(2, [2,0,1])
0.7059715047839078092146831
>>> 2*nsum(lambda k: (3*k)**-2, [1,inf]) + \
...
nsum(lambda k: (3*k+2)**-2, [0,inf])
0.7059715047839078092146831
Stieltjes constants
stieltjes()
mpmath.stieltjes(n, a=1)
For a nonnegative integer n, stieltjes(n) computes the n-th Stieltjes constant n , dened as the n-th coecient in the Laurent series expansion of the Riemann zeta function
around the pole at s = 1. That is, we have:
(s) =
1 (1)n
n (s 1)n
s 1 n=0 n!
More generally, stieltjes(n, a) gives the corresponding coecient n (a) for the Hurwitz zeta function (s, a) (with n = n (1)).
Examples
The zeroth Stieltjes constant is just Eulers constant :
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = True
>>> stieltjes(0)
0.577215664901533
Some more values are:
>>> stieltjes(1)
-0.0728158454836767
>>> stieltjes(10)
0.000205332814909065
>>> stieltjes(30)
0.00355772885557316
>>> stieltjes(1000)
-1.57095384420474e+486
>>> stieltjes(2000)
2.680424678918e+1109
>>> stieltjes(1, 2.5)
-0.23747539175716
An alternative way to compute 1 :
>>> diff(extradps(15)(lambda x: 1/(x-1) - zeta(x)), 1)
-0.0728158454836767
stieltjes() (page 937) supports arbitrary precision evaluation:
>>> mp.dps = 50
>>> stieltjes(2)
-0.0096903631928723184845303860352125293590658061013408
Algorithm
5.15. Welcome to mpmaths documentation!
937
SymPy Documentation, Release 0.7.6
stieltjes() (page 937) numerically evaluates the integral in the following representation due to Ainsworth, Howell and Coey [1], [2]:
n
n (a) =
n+1
log a log
(a) 2
+ <
2a
n+1
a
(x/a i) log (a ix)
dx.
(1 + x2 /a2 )(e2x 1)
For some reference values with a = 1, see e.g. [4].
References
1.O. R. Ainsworth & L. W. Howell, An integral representation of the generalized Euler-Mascheroni constants, NASA Technical Paper 2456 (1985),
http://ntrs.nasa.gov/archive/nasa/casi.ntrs.nasa.gov/19850014994 1985014994.pdf
2.M. W. Coey, The Stieltjes constants, their relation to the j coecients, and representation of the Hurwitz zeta function, arXiv:0706.0343v1
http://arxiv.org/abs/0706.0343
3.http://mathworld.wolfram.com/StieltjesConstants.html
4.http://pi.lacim.uqam.ca/piDATA/stieltjesgamma.txt
Zeta function zeros
in the critical strip.
These functions are used for the study of the Riemann zeta function
zetazero()
mpmath.zetazero(n, verbose=False)
Computes the n-th nontrivial zero of (s) on the critical line, i.e. returns an approximation of the n-th largest complex number s = 12 + ti for which (s) = 0. Equivalently, the
imaginary part t is a zero of the Z-function (siegelz() (page 939)).
Examples
The rst few zeros:
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> zetazero(1)
(0.5 + 14.13472514173469379045725j)
>>> zetazero(2)
(0.5 + 21.02203963877155499262848j)
>>> zetazero(20)
(0.5 + 77.14484006887480537268266j)
Verifying that the values are zeros:
>>> for n in range(1,5):
...
s = zetazero(n)
...
chop(zeta(s)), chop(siegelz(s.imag))
...
(0.0, 0.0)
(0.0, 0.0)
(0.0, 0.0)
(0.0, 0.0)
Negative indices give the conjugate zeros (n = 0 is undened):
>>> zetazero(-1)
(0.5 - 14.13472514173469379045725j)
938
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
zetazero() (page 938) supports arbitrarily large n and arbitrary precision:
>>> mp.dps = 15
>>> zetazero(1234567)
(0.5 + 727690.906948208j)
>>> mp.dps = 50
>>> zetazero(1234567)
(0.5 + 727690.9069482075392389420041147142092708393819935j)
>>> chop(zeta(_)/_)
0.0
with info=True, zetazero() (page 938) gives additional information:
>>> mp.dps = 15
>>> zetazero(542964976,info=True)
((0.5 + 209039046.578535j), [542964969, 542964978], 6, (013111110))
This means that the zero is between Gram points 542964969 and 542964978; it is the
6-th zero between them. Finally (01311110) is the pattern of zeros in this interval. The
numbers indicate the number of zeros in each Gram interval (Rosser blocks between
parenthesis). In this case there is only one Rosser block of length nine.
nzeros()
mpmath.nzeros(t)
Computes the number of zeros of the Riemann zeta function in (0, 1) (0, t], usually denoted by N (t).
Examples
The rst zero has imaginary part between 14 and 15:
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = True
>>> nzeros(14)
0
>>> nzeros(15)
1
>>> zetazero(1)
(0.5 + 14.1347251417347j)
Some closely spaced zeros:
>>> nzeros(10**7)
21136125
>>> zetazero(21136125)
(0.5 + 9999999.32718175j)
>>> zetazero(21136126)
(0.5 + 10000000.2400236j)
>>> nzeros(545439823.215)
1500000001
>>> zetazero(1500000001)
(0.5 + 545439823.201985j)
>>> zetazero(1500000002)
(0.5 + 545439823.325697j)
This conrms the data given by J. van de Lune, H. J. J. te Riele and D. T. Winter in 1986.
siegelz()
5.15. Welcome to mpmaths documentation!
939
SymPy Documentation, Release 0.7.6
mpmath.siegelz(t)
Computes the Z-function, also known as the Riemann-Siegel Z function,
Z(t) = ei(t) (1/2 + it)
where (s) is the Riemann zeta function (zeta() (page 931)) and where (t) denotes the
Riemann-Siegel theta function (see siegeltheta() (page 941)).
Evaluation is supported for real and complex arguments:
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> siegelz(1)
-0.7363054628673177346778998
>>> siegelz(3+4j)
(-0.1852895764366314976003936 - 0.2773099198055652246992479j)
The rst four derivatives are supported, using the optional derivative keyword argument:
>>> siegelz(1234567, derivative=3)
56.89689348495089294249178
>>> diff(siegelz, 1234567, n=3)
56.89689348495089294249178
The Z-function has a Maclaurin expansion:
>>> nprint(chop(taylor(siegelz, 0, 4)))
[-1.46035, 0.0, 2.73588, 0.0, -8.39357]
The Z-function Z(t) is equal to |(s)| on the critical line s = 1/2+it (i.e. for real arguments
t to Z). Its zeros coincide with those of the Riemann zeta function:
>>> findroot(siegelz, 14)
14.13472514173469379045725
>>> findroot(siegelz, 20)
21.02203963877155499262848
>>> findroot(zeta, 0.5+14j)
(0.5 + 14.13472514173469379045725j)
>>> findroot(zeta, 0.5+20j)
(0.5 + 21.02203963877155499262848j)
Since the Z-function is real-valued on the critical line (and unlike |(s)| analytic), it is
useful for investigating the zeros of the Riemann zeta function. For example, one can
use a root-nding algorithm based on sign changes:
>>> findroot(siegelz, [100, 200], solver=bisect)
176.4414342977104188888926
To locate roots, Gram points gn which can be computed by grampoint() (page 941) are
useful. If (1)n Z(gn ) is positive for two consecutive n, then Z(t) must have a zero between
those points:
>>> g10 = grampoint(10)
>>> g11 = grampoint(11)
>>> (-1)**10 * siegelz(g10) > 0
True
>>> (-1)**11 * siegelz(g11) > 0
True
>>> findroot(siegelz, [g10, g11], solver=bisect)
56.44624769706339480436776
940
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> g10, g11
(54.67523744685325626632663, 57.54516517954725443703014)
siegeltheta()
mpmath.siegeltheta(t)
Computes the Riemann-Siegel theta function,
(
)
(
)
log 1+2it
log 12it
log
4
4
(t) =
t.
2i
2
The Riemann-Siegel theta function is important in providing the phase factor for the
Z-function (see siegelz() (page 939)). Evaluation is supported for real and complex
arguments:
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> siegeltheta(0)
0.0
>>> siegeltheta(inf)
+inf
>>> siegeltheta(-inf)
-inf
>>> siegeltheta(1)
-1.767547952812290388302216
>>> siegeltheta(10+0.25j)
(-3.068638039426838572528867 + 0.05804937947429712998395177j)
Arbitrary derivatives may be computed with derivative = k
>>> siegeltheta(1234, derivative=2)
0.0004051864079114053109473741
>>> diff(siegeltheta, 1234, n=2)
0.0004051864079114053109473741
The Riemann-Siegel theta function has odd symmetry around t = 0, two local extreme
points and three real roots including 0 (located symmetrically):
>>> nprint(chop(taylor(siegeltheta, 0, 5)))
[0.0, -2.68609, 0.0, 2.69433, 0.0, -6.40218]
>>> findroot(diffun(siegeltheta), 7)
6.28983598883690277966509
>>> findroot(siegeltheta, 20)
17.84559954041086081682634
For large t, there is a famous asymptotic formula for (t), to rst order given by:
>>> t = mpf(10**6)
>>> siegeltheta(t)
5488816.353078403444882823
>>> -t*log(2*pi/t)/2-t/2
5488816.745777464310273645
grampoint()
mpmath.grampoint(n)
Gives the n-th Gram point gn , dened as the solution to the equation (gn ) = n where
(t) is the Riemann-Siegel theta function (siegeltheta() (page 941)).
5.15. Welcome to mpmaths documentation!
941
SymPy Documentation, Release 0.7.6
The rst few Gram points are:
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> grampoint(0)
17.84559954041086081682634
>>> grampoint(1)
23.17028270124630927899664
>>> grampoint(2)
27.67018221781633796093849
>>> grampoint(3)
31.71797995476405317955149
Checking the denition:
>>> siegeltheta(grampoint(3))
9.42477796076937971538793
>>> 3*pi
9.42477796076937971538793
A large Gram point:
>>> grampoint(10**10)
3293531632.728335454561153
Gram points are useful when studying the Z-function (siegelz() (page 939)). See the
documentation of that function for additional examples.
grampoint() (page 941) can solve the dening equation for nonintegral n. There is a
xed point where g(x) = x:
>>> findroot(lambda x: grampoint(x) - x, 10000)
9146.698193171459265866198
References
1.http://mathworld.wolfram.com/GramPoint.html
backlunds()
mpmath.backlunds(t)
Computes the function S(t) = arg ( 12 + it)/.
See Titchmarsh Section 9.3 for details of the denition.
Examples
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = True
>>> backlunds(217.3)
0.16302205431184
Generally, the value is a small number. At Gram points it is an integer, frequently equal
to 0:
>>> chop(backlunds(grampoint(200)))
0.0
>>> backlunds(extraprec(10)(grampoint)(211))
1.0
>>> backlunds(extraprec(10)(grampoint)(232))
-1.0
942
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
The number of zeros of the Riemann zeta function up to height t satises N (t) = (t)/ +
1 + S(t) (see :func:nzeros and siegeltheta()):
>>> t = 1234.55
>>> nzeros(t)
842
>>> siegeltheta(t)/pi+1+backlunds(t)
842.0
Lerch transcendent
lerchphi()
mpmath.lerchphi(z, s, a)
Gives the Lerch transcendent, dened for |z| < 1 and <a > 0 by
(z, s, a) =
k=0
zk
(a + k)s
and generally by the recurrence (z, s, a) = z(z, s, a + 1) + as along with the integral
representation valid for <a > 0
1
zt
sin(t log z s arctan(t/a)
(z, s, a) = s +
dt 2
dt.
s
2a
(a
+
t)
(a2 + t2 )s/2 (e2t 1)
0
0
The Lerch transcendent generalizes the Hurwitz zeta function zeta() (z = 1) and the
polylogarithm polylog() (a = 1).
Examples
Several evaluations in terms of simpler functions:
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> lerchphi(-1,2,0.5); 4*catalan
3.663862376708876060218414
3.663862376708876060218414
>>> diff(lerchphi, (-1,-2,1), (0,1,0)); 7*zeta(3)/(4*pi**2)
0.2131391994087528954617607
0.2131391994087528954617607
>>> lerchphi(-4,1,1); log(5)/4
0.4023594781085250936501898
0.4023594781085250936501898
>>> lerchphi(-3+2j,1,0.5); 2*atanh(sqrt(-3+2j))/sqrt(-3+2j)
(1.142423447120257137774002 + 0.2118232380980201350495795j)
(1.142423447120257137774002 + 0.2118232380980201350495795j)
Evaluation works for complex arguments and |z| 1:
>>> lerchphi(1+2j, 3-j, 4+2j)
(0.002025009957009908600539469 + 0.003327897536813558807438089j)
>>> lerchphi(-2,2,-2.5)
-12.28676272353094275265944
>>> lerchphi(10,10,10)
(-4.462130727102185701817349e-11 + 1.575172198981096218823481e-12j)
>>> lerchphi(10,10,-10.5)
(112658784011940.5605789002 + 498113185.5756221777743631j)
5.15. Welcome to mpmaths documentation!
943
SymPy Documentation, Release 0.7.6
Some degenerate cases:
>>> lerchphi(0,1,2)
0.5
>>> lerchphi(0,1,-2)
-0.5
Reduction to simpler functions:
>>> lerchphi(1, 4.25+1j, 1)
(1.044674457556746668033975 - 0.04674508654012658932271226j)
>>> zeta(4.25+1j)
(1.044674457556746668033975 - 0.04674508654012658932271226j)
>>> lerchphi(1 - 0.5**10, 4.25+1j, 1)
(1.044629338021507546737197 - 0.04667768813963388181708101j)
>>> lerchphi(3, 4, 1)
(1.249503297023366545192592 - 0.2314252413375664776474462j)
>>> polylog(4, 3) / 3
(1.249503297023366545192592 - 0.2314252413375664776474462j)
>>> lerchphi(3, 4, 1 - 0.5**10)
(1.253978063946663945672674 + 0.2316736622836535468765376j)
References
1.[DLMF] (page 1910) section 25.14
Polylogarithms and Clausen functions
polylog()
mpmath.polylog(s, z)
Computes the polylogarithm, dened by the sum
Lis (z) =
zk
k=1
ks
This series is convergent only for |z| < 1, so elsewhere the analytic continuation is implied.
The polylogarithm should not be confused with the logarithmic integral (also denoted by
Li or li), which is implemented as li() (page 799).
Examples
The polylogarithm satises a huge number of functional identities. A sample of polylogarithm evaluations is shown below:
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = True
>>> polylog(1,0.5), log(2)
(0.693147180559945, 0.693147180559945)
>>> polylog(2,0.5), (pi**2-6*log(2)**2)/12
(0.582240526465012, 0.582240526465012)
>>> polylog(2,-phi), -log(phi)**2-pi**2/10
(-1.21852526068613, -1.21852526068613)
>>> polylog(3,0.5), 7*zeta(3)/8-pi**2*log(2)/12+log(2)**3/6
(0.53721319360804, 0.53721319360804)
polylog() (page 944) can evaluate the analytic continuation of the polylogarithm when
s is an integer:
944
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> polylog(2, 10)
(0.536301287357863 - 7.23378441241546j)
>>> polylog(2, -10)
-4.1982778868581
>>> polylog(2, 10j)
(-3.05968879432873 + 3.71678149306807j)
>>> polylog(-2, 10)
-0.150891632373114
>>> polylog(-2, -10)
0.067618332081142
>>> polylog(-2, 10j)
(0.0384353698579347 + 0.0912451798066779j)
Some more examples, with arguments on the unit circle (note that the series denition
cannot be used for computation here):
>>> polylog(2,j)
(-0.205616758356028 + 0.915965594177219j)
>>> j*catalan-pi**2/48
(-0.205616758356028 + 0.915965594177219j)
>>> polylog(3,exp(2*pi*j/3))
(-0.534247512515375 + 0.765587078525922j)
>>> -4*zeta(3)/9 + 2*j*pi**3/81
(-0.534247512515375 + 0.765587078525921j)
Polylogarithms of dierent order are related by integration and dierentiation:
>>> s, z = 3, 0.5
>>> polylog(s+1, z)
0.517479061673899
>>> quad(lambda t: polylog(s,t)/t, [0, z])
0.517479061673899
>>> z*diff(lambda t: polylog(s+2,t), z)
0.517479061673899
Taylor series expansions around z = 0 are:
>>> for n in range(-3, 4):
...
nprint(taylor(lambda x: polylog(n,x), 0, 5))
...
[0.0, 1.0, 8.0, 27.0, 64.0, 125.0]
[0.0, 1.0, 4.0, 9.0, 16.0, 25.0]
[0.0, 1.0, 2.0, 3.0, 4.0, 5.0]
[0.0, 1.0, 1.0, 1.0, 1.0, 1.0]
[0.0, 1.0, 0.5, 0.333333, 0.25, 0.2]
[0.0, 1.0, 0.25, 0.111111, 0.0625, 0.04]
[0.0, 1.0, 0.125, 0.037037, 0.015625, 0.008]
The series dening the polylogarithm is simultaneously a Taylor series and an L-series.
For certain values of z, the polylogarithm reduces to a pure zeta function:
>>> polylog(pi, 1), zeta(pi)
(1.17624173838258, 1.17624173838258)
>>> polylog(pi, -1), -altzeta(pi)
(-0.909670702980385, -0.909670702980385)
Evaluation for arbitrary, nonintegral s is supported for z within the unit circle:
5.15. Welcome to mpmaths documentation!
945
SymPy Documentation, Release 0.7.6
>>> polylog(3+4j, 0.25)
(0.24258605789446 - 0.00222938275488344j)
>>> nsum(lambda k: 0.25**k / k**(3+4j), [1,inf])
(0.24258605789446 - 0.00222938275488344j)
It is also currently supported outside of the unit circle for z not too large in magnitude:
>>> polylog(1+j, 20+40j)
(-7.1421172179728 - 3.92726697721369j)
>>> polylog(1+j, 200+400j)
Traceback (most recent call last):
...
NotImplementedError: polylog for arbitrary s and z
Traceback (most recent call last):
...
NotImplementedError: polylog for arbitrary s and z
References
1.Richard
Crandall,
Note
on
fast
polylogarithm
http://people.reed.edu/crandall/papers/Polylog.pdf
computation
2.http://en.wikipedia.org/wiki/Polylogarithm
3.http://mathworld.wolfram.com/Polylogarithm.html
clsin()
mpmath.clsin(s, z)
Computes the Clausen sine function, dened formally by the series
Cls (z) =
sin(kz)
k=1
ks
The special case Cl2 (z) (i.e. clsin(2,z)) is the classical Clausen function. More generally, the Clausen function is dened for complex s and z, even when the series does not
converge. The Clausen function is related to the polylogarithm (polylog() (page 944))
as
( )
(
))
1 (
Lis eiz Lis eiz
2i
[
]
= Im Lis (eiz )
(s, z R),
Cls (z) =
and this representation can be taken to provide the analytic continuation of the series.
The complementary function clcos() (page 948) gives the corresponding cosine sum.
Examples
Evaluation for arbitrarily chosen s and z:
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> s, z = 3, 4
>>> clsin(s, z); nsum(lambda k: sin(z*k)/k**s, [1,inf])
-0.6533010136329338746275795
-0.6533010136329338746275795
Using z + instead of z gives an alternating series:
946
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> clsin(s, z+pi)
0.8860032351260589402871624
>>> nsum(lambda k: (-1)**k*sin(z*k)/k**s, [1,inf])
0.8860032351260589402871624
With s = 1, the sum can be expressed in closed form using elementary functions:
>>> z = 1 + sqrt(3)
>>> clsin(1, z)
0.2047709230104579724675985
>>> chop((log(1-exp(-j*z)) - log(1-exp(j*z)))/(2*j))
0.2047709230104579724675985
>>> nsum(lambda k: sin(k*z)/k, [1,inf])
0.2047709230104579724675985
The classical Clausen function Cl2 () gives the value of the integral
for 0 < < 2:
ln(2 sin(x/2))dx
>>> cl2 = lambda t: clsin(2, t)
>>> cl2(3.5)
-0.2465045302347694216534255
>>> -quad(lambda x: ln(2*sin(0.5*x)), [0, 3.5])
-0.2465045302347694216534255
This function is symmetric about = with zeros and extreme points:
>>> cl2(0); cl2(pi/3); chop(cl2(pi)); cl2(5*pi/3); chop(cl2(2*pi))
0.0
1.014941606409653625021203
0.0
-1.014941606409653625021203
0.0
Catalans constant is a special value:
>>> cl2(pi/2)
0.9159655941772190150546035
>>> +catalan
0.9159655941772190150546035
The Clausen sine function can be expressed in closed form when s is an odd integer
(becoming zero when s < 0):
>>> z = 1 + sqrt(2)
>>> clsin(1, z); (pi-z)/2
0.3636895456083490948304773
0.3636895456083490948304773
>>> clsin(3, z); pi**2/6*z - pi*z**2/4 + z**3/12
0.5661751584451144991707161
0.5661751584451144991707161
>>> clsin(-1, z)
0.0
>>> clsin(-3, z)
0.0
It can also be expressed in closed form for even integer s 0, providing a nite sum for
series such as sin(z) + sin(2z) + sin(3z) + . . .:
>>> z = 1 + sqrt(2)
>>> clsin(0, z)
5.15. Welcome to mpmaths documentation!
947
SymPy Documentation, Release 0.7.6
0.1903105029507513881275865
>>> cot(z/2)/2
0.1903105029507513881275865
>>> clsin(-2, z)
-0.1089406163841548817581392
>>> -cot(z/2)*csc(z/2)**2/4
-0.1089406163841548817581392
Call with pi=True to multiply z by exactly:
>>> clsin(3, 3*pi)
-8.892316224968072424732898e-26
>>> clsin(3, 3, pi=True)
0.0
Evaluation for complex s, z in a nonconvergent case:
>>> s, z = -1-j, 1+2j
>>> clsin(s, z)
(-0.593079480117379002516034 + 0.9038644233367868273362446j)
>>> extraprec(20)(nsum)(lambda k: sin(k*z)/k**s, [1,inf])
(-0.593079480117379002516034 + 0.9038644233367868273362446j)
clcos()
mpmath.clcos(s, z)
Computes the Clausen cosine function, dened formally by the series
f s (z) =
Cl
cos(kz)
k=1
ks
This function is complementary to the Clausen sine function clsin() (page 946). In
terms of the polylogarithm,
f s (z) = 1 (Lis (eiz ) + Lis (eiz ))
Cl
[2
]
= Re Lis (eiz )
(s, z R).
Examples
Evaluation for arbitrarily chosen s and z:
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> s, z = 3, 4
>>> clcos(s, z); nsum(lambda k: cos(z*k)/k**s, [1,inf])
-0.6518926267198991308332759
-0.6518926267198991308332759
Using z + instead of z gives an alternating series:
>>> s, z = 3, 0.5
>>> clcos(s, z+pi)
-0.8155530586502260817855618
>>> nsum(lambda k: (-1)**k*cos(z*k)/k**s, [1,inf])
-0.8155530586502260817855618
With s = 1, the sum can be expressed in closed form using elementary functions:
948
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> z = 1 + sqrt(3)
>>> clcos(1, z)
-0.6720334373369714849797918
>>> chop(-0.5*(log(1-exp(j*z))+log(1-exp(-j*z))))
-0.6720334373369714849797918
>>> -log(abs(2*sin(0.5*z)))
# Equivalent to above when z is real
-0.6720334373369714849797918
>>> nsum(lambda k: cos(k*z)/k, [1,inf])
-0.6720334373369714849797918
It can also be expressed in closed form when s is an even integer. For example,
>>> clcos(2,z)
-0.7805359025135583118863007
>>> pi**2/6 - pi*z/2 + z**2/4
-0.7805359025135583118863007
The case s = 0 gives the renormalized sum of cos(z)+cos(2z)+cos(3z)+. . . (which happens
to be the same for any value of z):
>>> clcos(0, z)
-0.5
>>> nsum(lambda k: cos(k*z), [1,inf])
-0.5
Also the sums
cos(z) + 2 cos(2z) + 3 cos(3z) + . . .
and
cos(z) + 2n cos(2z) + 3n cos(3z) + . . .
for higher integer powers n = s can be done in closed form. They are zero when n is
positive and even (s negative and even):
>>> clcos(-1, z); 1/(2*cos(z)-2)
-0.2607829375240542480694126
-0.2607829375240542480694126
>>> clcos(-3, z); (2+cos(z))*csc(z/2)**4/8
0.1472635054979944390848006
0.1472635054979944390848006
>>> clcos(-2, z); clcos(-4, z); clcos(-6, z)
0.0
0.0
0.0
With z = , the series reduces to that of the Riemann zeta function (more generally, if
z = p/q, it is a nite sum over Hurwitz zeta function values):
>>> clcos(2.5, 0); zeta(2.5)
1.34148725725091717975677
1.34148725725091717975677
>>> clcos(2.5, pi); -altzeta(2.5)
-0.8671998890121841381913472
-0.8671998890121841381913472
Call with pi=True to multiply z by exactly:
5.15. Welcome to mpmaths documentation!
949
SymPy Documentation, Release 0.7.6
>>> clcos(-3, 2*pi)
2.997921055881167659267063e+102
>>> clcos(-3, 2, pi=True)
0.008333333333333333333333333
Evaluation for complex s, z in a nonconvergent case:
>>> s, z = -1-j, 1+2j
>>> clcos(s, z)
(0.9407430121562251476136807 + 0.715826296033590204557054j)
>>> extraprec(20)(nsum)(lambda k: cos(k*z)/k**s, [1,inf])
(0.9407430121562251476136807 + 0.715826296033590204557054j)
polyexp()
mpmath.polyexp(s, z)
Evaluates the polyexponential function, dened for arbitrary complex s, z by the series
Es (z) =
ks
k=1
k!
zk .
Es (z) is constructed from the exponential function analogously to how the polylogarithm
is constructed from the ordinary logarithm; as a function of s (with z xed), Es is an
L-series It is an entire function of both s and z.
The polyexponential function provides a generalization of the Bell polynomials Bn (x) (see
bell() (page 960)) to noninteger orders n. In terms of the Bell polynomials,
Es (z) = ez Bs (z) sinc(s).
Note that Bn (x) and ex En (x) are identical if n is a nonzero integer, but not otherwise. In
particular, they dier at n = 0.
Examples
Evaluating a series:
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> nsum(lambda k: sqrt(k)/fac(k), [1,inf])
2.101755547733791780315904
>>> polyexp(0.5,1)
2.101755547733791780315904
Evaluation for arbitrary arguments:
>>> polyexp(-3-4j, 2.5+2j)
(2.351660261190434618268706 + 1.202966666673054671364215j)
Evaluation is accurate for tiny function values:
>>> polyexp(4, -100)
3.499471750566824369520223e-36
If n is a nonpositive integer, En reduces to a special instance of the hypergeometric
function p Fq :
>>> n = 3
>>> x = pi
>>> polyexp(-n,x)
950
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
4.042192318847986561771779
>>> x*hyper([1]*(n+1), [2]*(n+1), x)
4.042192318847986561771779
Zeta function variants
primezeta()
mpmath.primezeta(s)
Computes the prime zeta function, which is dened in analogy with the Riemann zeta
function (zeta() (page 931)) as
P (s) =
1
ps
p
where the sum is taken over all prime numbers p. Although this sum only converges for
Re(s) > 1, the function is dened by analytic continuation in the half-plane Re(s) > 0.
Examples
Arbitrary-precision evaluation for real and complex arguments is supported:
>>> from sympy.mpmath import *
>>> mp.dps = 30; mp.pretty = True
>>> primezeta(2)
0.452247420041065498506543364832
>>> primezeta(pi)
0.15483752698840284272036497397
>>> mp.dps = 50
>>> primezeta(3)
0.17476263929944353642311331466570670097541212192615
>>> mp.dps = 20
>>> primezeta(3+4j)
(-0.12085382601645763295 - 0.013370403397787023602j)
The prime zeta function has a logarithmic pole at s = 1, with residue equal to the dierence of the Mertens and Euler constants:
>>> primezeta(1)
+inf
>>> extradps(25)(lambda x: primezeta(1+x)+log(x))(+eps)
-0.31571845205389007685
>>> mertens-euler
-0.31571845205389007685
The analytic continuation to 0 < Re(s) 1 is implemented. In this strip the function exhibits very complex behavior; on the unit interval, it has poles at 1/n for every squarefree
integer n:
>>> primezeta(0.5)
# Pole at s = 1/2
(-inf + 3.1415926535897932385j)
>>> primezeta(0.25)
(-1.0416106801757269036 + 0.52359877559829887308j)
>>> primezeta(0.5+10j)
(0.54892423556409790529 + 0.45626803423487934264j)
Although evaluation works in principle for any Re(s) > 0, it should be noted that the
evaluation time increases exponentially as s approaches the imaginary axis.
5.15. Welcome to mpmaths documentation!
951
SymPy Documentation, Release 0.7.6
For large Re(s), P (s) is asymptotic to 2s :
>>> primezeta(inf)
0.0
>>> primezeta(10), mpf(2)**-10
(0.00099360357443698021786, 0.0009765625)
>>> primezeta(1000)
9.3326361850321887899e-302
>>> primezeta(1000+1000j)
(-3.8565440833654995949e-302 - 8.4985390447553234305e-302j)
References
Carl-Erik Froberg, On the prime zeta function, BIT 8 (1968), pp. 187-202.
secondzeta()
mpmath.secondzeta(s, a=0.015, **kwargs)
Evaluates the secondary zeta function Z(s), dened for Re(s) > 1 by
Z(s) =
where
1
2
1
s
n=1 n
+ in runs through the zeros of (s) with imaginary part positive.
Z(s) extends to a meromorphic function on C with a double pole at s = 1 and simple poles
at the points 2n for n = 0, 1, 2, ...
Examples
>>> from sympy.mpmath import *
>>> mp.pretty = True; mp.dps = 15
>>> secondzeta(2)
0.023104993115419
>>> xi = lambda s: 0.5*s*(s-1)*pi**(-0.5*s)*gamma(0.5*s)*zeta(s)
>>> Xi = lambda t: xi(0.5+t*j)
>>> -0.5*diff(Xi,0,n=2)/Xi(0)
(0.023104993115419 + 0.0j)
We may ask for an approximate error value:
>>> secondzeta(0.5+100j, error=True)
((-0.216272011276718 - 0.844952708937228j), 2.22044604925031e-16)
The function has poles at the negative odd integers, and dyadic rational values at the
negative even integers:
>>> mp.dps = 30
>>> secondzeta(-8)
-0.67236328125
>>> secondzeta(-7)
+inf
Implementation notes
The function is computed as sum of four terms Z(s) = A(s) P (s) + E(s) S(s) respectively
main, prime, exponential and singular terms. The main term A(s) is computed from the
zeros of zeta. The prime term depends on the von Mangoldt function. The singular term
is responsible for the poles of the function.
952
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
The four terms depends on a small parameter a. We may change the value of a. Theoretically this has no eect on the sum of the four terms, but in practice may be important.
A smaller value of the parameter a makes A(s) depend on a smaller number of zeros of
zeta, but P (s) uses more values of von Mangoldt function.
We may also add a verbose option to obtain data about the values of the four terms.
>>> mp.dps = 10
>>> secondzeta(0.5 + 40j, error=True, verbose=True)
main term = (-30190318549.138656312556 - 13964804384.624622876523j)
computed using 19 zeros of zeta
prime term = (132717176.89212754625045 + 188980555.17563978290601j)
computed using 9 values of the von Mangoldt function
exponential term = (542447428666.07179812536 + 362434922978.80192435203j)
singular term = (512124392939.98154322355 + 348281138038.65531023921j)
((0.059471043 + 0.3463514534j), 1.455191523e-11)
>>> secondzeta(0.5 + 40j, a=0.04, error=True, verbose=True)
main term = (-151962888.19606243907725 - 217930683.90210294051982j)
computed using 9 zeros of zeta
prime term = (2476659342.3038722372461 + 28711581821.921627163136j)
computed using 37 values of the von Mangoldt function
exponential term = (178506047114.7838188264 + 819674143244.45677330576j)
singular term = (175877424884.22441310708 + 790744630738.28669174871j)
((0.059471043 + 0.3463514534j), 1.455191523e-11)
Notice the great cancellation between the four terms. Changing a, the four terms are
very dierent numbers but the cancellation gives the good value of Z(s).
References
A. Voros, Zeta functions for the Riemann zeros, Ann.
665699.
Institute Fourier, 53, (2003)
A. Voros, Zeta functions over Zeros of Zeta Functions, Lecture Notes of the Unione
Matematica Italiana, Springer, 2009.
Number-theoretical, combinatorial and integer functions
For factorial-type functions, including binomial coecients, double factorials, etc., see the
separate section Factorials and gamma functions (page 777).
Fibonacci numbers
fibonacci()/fib()
mpmath.fibonacci(n, **kwargs)
fibonacci(n) computes the n-th Fibonacci number, F (n). The Fibonacci numbers are
dened by the recurrence F (n) = F (n1)+F (n2) with the initial values F (0) = 0, F (1) = 1.
fibonacci() (page 953) extends this denition to arbitrary real and complex arguments
using the formula
F (z) =
z cos(z)z
5.15. Welcome to mpmaths documentation!
953
SymPy Documentation, Release 0.7.6
where is the golden ratio. fibonacci() (page 953) also uses this continuous formula
to compute F (n) for extremely large n, where calculating the exact integer would be
wasteful.
For convenience, fib() is available as an alias for fibonacci() (page 953).
Basic examples
Some small Fibonacci numbers are:
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = True
>>> for i in range(10):
...
print(fibonacci(i))
...
0.0
1.0
1.0
2.0
3.0
5.0
8.0
13.0
21.0
34.0
>>> fibonacci(50)
12586269025.0
The recurrence for F (n) extends backwards to negative n:
>>> for i in range(10):
...
print(fibonacci(-i))
...
0.0
1.0
-1.0
2.0
-3.0
5.0
-8.0
13.0
-21.0
34.0
Large Fibonacci numbers will be computed approximately unless the precision is set
high enough:
>>> fib(200)
2.8057117299251e+41
>>> mp.dps = 45
>>> fib(200)
280571172992510140037611932413038677189525.0
fibonacci() (page 953) can compute approximate Fibonacci numbers of stupendous
size:
>>> mp.dps = 15
>>> fibonacci(10**25)
3.49052338550226e+2089876402499787337692720
Real and complex arguments
954
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
The extended Fibonacci function is an analytic function. The property F (z) = F (z 1) +
F (z 2) holds for arbitrary z:
>>> mp.dps = 15
>>> fib(pi)
2.1170270579161
>>> fib(pi-1) + fib(pi-2)
2.1170270579161
>>> fib(3+4j)
(-5248.51130728372 - 14195.962288353j)
>>> fib(2+4j) + fib(1+4j)
(-5248.51130728372 - 14195.962288353j)
The Fibonacci function has innitely many roots on the negative half-real axis. The rst
root is at 0, the second is close to -0.18, and then there are innitely many roots that
asymptotically approach n + 1/2:
>>> findroot(fib, -0.2)
-0.183802359692956
>>> findroot(fib, -2)
-1.57077646820395
>>> findroot(fib, -17)
-16.4999999596115
>>> findroot(fib, -24)
-23.5000000000479
Mathematical relationships
For large n, F (n + 1)/F (n) approaches the golden ratio:
>>> mp.dps = 50
>>> fibonacci(101)/fibonacci(100)
1.6180339887498948482045868343656381177203127439638
>>> +phi
1.6180339887498948482045868343656381177203091798058
The sum of reciprocal Fibonacci numbers converges to an irrational number for which
no closed form expression is known:
>>> mp.dps = 15
>>> nsum(lambda n: 1/fib(n), [1, inf])
3.35988566624318
Amazingly, however, the sum of odd-index reciprocal Fibonacci numbers can be expressed in terms of a Jacobi theta function:
>>> nsum(lambda n: 1/fib(2*n+1), [0, inf])
1.82451515740692
>>> sqrt(5)*jtheta(2,0,(3-sqrt(5))/2)**2/4
1.82451515740692
Some related sums can be done in closed form:
>>> nsum(lambda k: 1/(1+fib(2*k+1)), [0, inf])
1.11803398874989
>>> phi - 0.5
1.11803398874989
>>> f = lambda k:(-1)**(k+1) / sum(fib(n)**2 for n in range(1,int(k+1)))
>>> nsum(f, [1, inf])
0.618033988749895
5.15. Welcome to mpmaths documentation!
955
SymPy Documentation, Release 0.7.6
>>> phi-1
0.618033988749895
References
1.http://mathworld.wolfram.com/FibonacciNumber.html
Bernoulli numbers and polynomials
bernoulli()
mpmath.bernoulli(n)
Computes the nth Bernoulli number, Bn , for any integer n 0.
The Bernoulli numbers are rational numbers, but this function returns a oating-point
approximation. To obtain an exact fraction, use bernfrac() (page 957) instead.
Examples
Numerical values of the rst few Bernoulli numbers:
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = True
>>> for n in range(15):
...
print(%s %s % (n, bernoulli(n)))
...
0 1.0
1 -0.5
2 0.166666666666667
3 0.0
4 -0.0333333333333333
5 0.0
6 0.0238095238095238
7 0.0
8 -0.0333333333333333
9 0.0
10 0.0757575757575758
11 0.0
12 -0.253113553113553
13 0.0
14 1.16666666666667
Bernoulli numbers can be approximated with arbitrary precision:
>>> mp.dps = 50
>>> bernoulli(100)
-2.8382249570693706959264156336481764738284680928013e+78
Arbitrarily large n are supported:
>>> mp.dps = 15
>>> bernoulli(10**20 + 2)
3.09136296657021e+1876752564973863312327
The Bernoulli numbers are related to the Riemann zeta function at integer arguments:
>>> -bernoulli(8) * (2*pi)**8 / (2*fac(8))
1.00407735619794
>>> zeta(8)
1.00407735619794
956
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Algorithm
For small n (n < 3000) bernoulli() (page 956) uses a recurrence formula due to Ramanujan. All results in this range are cached, so sequential computation of small Bernoulli
numbers is guaranteed to be fast.
For larger n, Bn is evaluated in terms of the Riemann zeta function.
bernfrac()
mpmath.bernfrac(n)
Returns a tuple of integers (p, q) such that p/q = Bn exactly, where Bn denotes the n-th
Bernoulli number. The fraction is always reduced to lowest terms. Note that for n > 1
and n odd, Bn = 0, and (0, 1) is returned.
Examples
The rst few Bernoulli numbers are exactly:
>>> from sympy.mpmath import *
>>> for n in range(15):
...
p, q = bernfrac(n)
...
print(%s %s/%s % (n, p, q))
...
0 1/1
1 -1/2
2 1/6
3 0/1
4 -1/30
5 0/1
6 1/42
7 0/1
8 -1/30
9 0/1
10 5/66
11 0/1
12 -691/2730
13 0/1
14 7/6
This function works for arbitrarily large n:
>>> p, q = bernfrac(10**4)
>>> print(q)
2338224387510
>>> print(len(str(p)))
27692
>>> mp.dps = 15
>>> print(mpf(p) / q)
-9.04942396360948e+27677
>>> print(bernoulli(10**4))
-9.04942396360948e+27677
Note: bernoulli() (page 956) computes a oating-point approximation directly, without computing the exact fraction rst. This is much faster for large n.
Algorithm
bernfrac() (page 957) works by computing the value of Bn numerically and then using
the von Staudt-Clausen theorem [1] to reconstruct the exact fraction. For large n, this is
5.15. Welcome to mpmaths documentation!
957
SymPy Documentation, Release 0.7.6
signicantly faster than computing B1 , B2 , . . . , B2 recursively with exact arithmetic. The
implementation has been tested for n = 10m up to m = 6.
In practice, bernfrac() (page 957) appears to be about three times slower than the
specialized program calcbn.exe [2]
References
1.MathWorld, von Staudt-Clausen Theorem: http://mathworld.wolfram.com/vonStaudtClausenTheorem.html
2.The Bernoulli Number Page: http://bernoulli.org/
bernpoly()
mpmath.bernpoly(n, z)
Evaluates the Bernoulli polynomial Bn (z).
The rst few Bernoulli polynomials are:
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = True
>>> for n in range(6):
...
nprint(chop(taylor(lambda x: bernpoly(n,x), 0, n)))
...
[1.0]
[-0.5, 1.0]
[0.166667, -1.0, 1.0]
[0.0, 0.5, -1.5, 1.0]
[-0.0333333, 0.0, 1.0, -2.0, 1.0]
[0.0, -0.166667, 0.0, 1.66667, -2.5, 1.0]
At z = 0, the Bernoulli polynomial evaluates to a Bernoulli number (see bernoulli()
(page 956)):
>>> bernpoly(12, 0), bernoulli(12)
(-0.253113553113553, -0.253113553113553)
>>> bernpoly(13, 0), bernoulli(13)
(0.0, 0.0)
Evaluation is accurate for large n and small z:
>>> mp.dps = 25
>>> bernpoly(100, 0.5)
2.838224957069370695926416e+78
>>> bernpoly(1000, 10.5)
5.318704469415522036482914e+1769
Euler numbers and polynomials
eulernum()
mpmath.eulernum(n)
Gives the n-th Euler number, dened as the n-th derivative of sech(t) = 1/ cosh(t) evaluated at t = 0. Equivalently, the Euler numbers give the coecients of the Taylor series
En n
t .
sech(t) =
n!
n=0
958
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
The Euler numbers are closely related to Bernoulli numbers and Bernoulli polynomials.
They can also be evaluated in terms of Euler polynomials (see eulerpoly() (page 959))
as En = 2n En (1/2).
Examples
Computing the rst few Euler numbers and verifying that they agree with the Taylor
series:
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> [eulernum(n) for n in range(11)]
[1.0, 0.0, -1.0, 0.0, 5.0, 0.0, -61.0, 0.0, 1385.0, 0.0, -50521.0]
>>> chop(diffs(sech, 0, 10))
[1.0, 0.0, -1.0, 0.0, 5.0, 0.0, -61.0, 0.0, 1385.0, 0.0, -50521.0]
Euler numbers grow very rapidly. eulernum() (page 958) eciently computes numerical
approximations for large indices:
>>> eulernum(50)
-6.053285248188621896314384e+54
>>> eulernum(1000)
3.887561841253070615257336e+2371
>>> eulernum(10**20)
4.346791453661149089338186e+1936958564106659551331
Comparing with an asymptotic formula for the Euler numbers:
>>> n = 10**5
>>> (-1)**(n//2) * 8 * sqrt(n/(2*pi)) * (2*n/(pi*e))**n
3.69919063017432362805663e+436961
>>> eulernum(n)
3.699193712834466537941283e+436961
Pass exact=True to obtain exact values of Euler numbers as integers:
>>> print(eulernum(50, exact=True))
-6053285248188621896314383785111649088103498225146815121
>>> print(eulernum(200, exact=True) % 10**10)
1925859625
>>> eulernum(1001, exact=True)
0
eulerpoly()
mpmath.eulerpoly(n, z)
Evaluates the Euler polynomial En (z), dened by the generating function representation
tn
2ezt
=
E
(z)
.
n
et + 1 n=0
n!
The Euler polynomials may also be represented in terms of Bernoulli polynomials (see
bernpoly() (page 958)) using various formulas, for example
( z ))
2 (
En (z) =
Bn (z) 2n+1 Bn
.
n+1
2
Special values include the Euler numbers En = 2n En (1/2) (see eulernum() (page 958)).
Examples
Computing the coecients of the rst few Euler polynomials:
5.15. Welcome to mpmaths documentation!
959
SymPy Documentation, Release 0.7.6
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> for n in range(6):
...
chop(taylor(lambda z: eulerpoly(n,z), 0, n))
...
[1.0]
[-0.5, 1.0]
[0.0, -1.0, 1.0]
[0.25, 0.0, -1.5, 1.0]
[0.0, 1.0, 0.0, -2.0, 1.0]
[-0.5, 0.0, 2.5, 0.0, -2.5, 1.0]
Evaluation for arbitrary z:
>>> eulerpoly(2,3)
6.0
>>> eulerpoly(5,4)
423.5
>>> eulerpoly(35, 11111111112)
3.994957561486776072734601e+351
>>> eulerpoly(4, 10+20j)
(-47990.0 - 235980.0j)
>>> eulerpoly(2, -3.5e-5)
0.000035001225
>>> eulerpoly(3, 0.5)
0.0
>>> eulerpoly(55, -10**80)
-1.0e+4400
>>> eulerpoly(5, -inf)
-inf
>>> eulerpoly(6, -inf)
+inf
Computing Euler numbers:
>>> 2**26 * eulerpoly(26,0.5)
-4087072509293123892361.0
>>> eulernum(26)
-4087072509293123892361.0
Evaluation is accurate for large n and small z:
>>> eulerpoly(100, 0.5)
2.29047999988194114177943e+108
>>> eulerpoly(1000, 10.5)
3.628120031122876847764566e+2070
>>> eulerpoly(10000, 10.5)
1.149364285543783412210773e+30688
Bell numbers and polynomials
bell()
mpmath.bell(n, x)
For n a nonnegative integer, bell(n,x) evaluates the Bell polynomial Bn (x), the rst few
960
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
of which are
B0 (x) = 1
B1 (x) = x
B2 (x) = x2 + x
B3 (x) = x3 + 3x2 + x
If x = 1 or bell() (page 960) is called with only one argument, it gives the n-th Bell
number Bn , which is the number of partitions of a set with n elements. By setting the
precision to at least log10 Bn digits, bell() (page 960) provides fast calculation of exact
Bell numbers.
In general, bell() (page 960) computes
Bn (x) = ex (sinc(n) + En (x))
where En (x) is the generalized exponential function implemented by polyexp()
(page 950). This is an extension of Dobinskis formula [1], where the modication is
the sinc term ensuring that Bn (x) is continuous in n; bell() (page 960) can thus be
evaluated, dierentiated, etc for arbitrary complex arguments.
Examples
Simple evaluations:
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> bell(0, 2.5)
1.0
>>> bell(1, 2.5)
2.5
>>> bell(2, 2.5)
8.75
Evaluation for arbitrary complex arguments:
>>> bell(5.75+1j, 2-3j)
(-10767.71345136587098445143 - 15449.55065599872579097221j)
The rst few Bell polynomials:
>>> for k in range(7):
...
nprint(taylor(lambda x: bell(k,x), 0, k))
...
[1.0]
[0.0, 1.0]
[0.0, 1.0, 1.0]
[0.0, 1.0, 3.0, 1.0]
[0.0, 1.0, 7.0, 6.0, 1.0]
[0.0, 1.0, 15.0, 25.0, 10.0, 1.0]
[0.0, 1.0, 31.0, 90.0, 65.0, 15.0, 1.0]
The rst few Bell numbers and complementary Bell numbers:
>>>
[1,
>>>
[1,
[int(bell(k)) for k in range(10)]
1, 2, 5, 15, 52, 203, 877, 4140, 21147]
[int(bell(k,-1)) for k in range(10)]
-1, 0, 1, 1, -2, -9, -9, 50, 267]
Large Bell numbers:
5.15. Welcome to mpmaths documentation!
961
SymPy Documentation, Release 0.7.6
>>> mp.dps = 50
>>> bell(50)
185724268771078270438257767181908917499221852770.0
>>> bell(50,-1)
-29113173035759403920216141265491160286912.0
Some even larger values:
>>> mp.dps = 25
>>> bell(1000,-1)
-1.237132026969293954162816e+1869
>>> bell(1000)
2.989901335682408421480422e+1927
>>> bell(1000,2)
6.591553486811969380442171e+1987
>>> bell(1000,100.5)
9.101014101401543575679639e+2529
A determinant identity satised by Bell numbers:
>>> mp.dps = 15
>>> N = 8
>>> det([[bell(k+j) for j in range(N)] for k in range(N)])
125411328000.0
>>> superfac(N-1)
125411328000.0
References
1.http://mathworld.wolfram.com/DobinskisFormula.html
Stirling numbers
stirling1()
mpmath.stirling1(n, k, exact=False)
Gives the Stirling number of the rst kind s(n, k), dened by
x(x 1)(x 2) (x n + 1) =
s(n, k)xk .
k=0
The value is computed using an integer recurrence. The implementation is not optimized
for approximating large values quickly.
Examples
Comparing with the generating function:
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> taylor(lambda x: ff(x, 5), 0, 5)
[0.0, 24.0, -50.0, 35.0, -10.0, 1.0]
>>> [stirling1(5, k) for k in range(6)]
[0.0, 24.0, -50.0, 35.0, -10.0, 1.0]
Recurrence relation:
>>> n, k = 5, 3
>>> stirling1(n+1,k) + n*stirling1(n,k) - stirling1(n,k-1)
0.0
962
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
The matrices of Stirling numbers of rst and second kind are inverses of each other:
>>> A = matrix(5, 5); B = matrix(5, 5)
>>> for n in range(5):
...
for k in range(5):
...
A[n,k] = stirling1(n,k)
...
B[n,k] = stirling2(n,k)
...
>>> A * B
[1.0 0.0 0.0 0.0 0.0]
[0.0 1.0 0.0 0.0 0.0]
[0.0 0.0 1.0 0.0 0.0]
[0.0 0.0 0.0 1.0 0.0]
[0.0 0.0 0.0 0.0 1.0]
Pass exact=True to obtain exact values of Stirling numbers as integers:
>>> stirling1(42, 5)
-2.864498971768501633736628e+50
>>> print stirling1(42, 5, exact=True)
-286449897176850163373662803014001546235808317440000
stirling2()
mpmath.stirling2(n, k, exact=False)
Gives the Stirling number of the second kind S(n, k), dened by
xn =
S(n, k)x(x 1)(x 2) (x k + 1)
k=0
The value is computed using integer arithmetic to evaluate a power sum. The implementation is not optimized for approximating large values quickly.
Examples
Comparing with the generating function:
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> taylor(lambda x: sum(stirling2(5,k) * ff(x,k) for k in range(6)), 0, 5)
[0.0, 0.0, 0.0, 0.0, 0.0, 1.0]
Recurrence relation:
>>> n, k = 5, 3
>>> stirling2(n+1,k) - k*stirling2(n,k) - stirling2(n,k-1)
0.0
Pass exact=True to obtain exact values of Stirling numbers as integers:
>>> stirling2(52, 10)
2.641822121003543906807485e+45
>>> print stirling2(52, 10, exact=True)
2641822121003543906807485307053638921722527655
Prime counting functions
5.15. Welcome to mpmaths documentation!
963
SymPy Documentation, Release 0.7.6
primepi()
mpmath.primepi(x)
Evaluates the prime counting function, (x), which gives the number of primes less than
or equal to x. The argument x may be fractional.
The prime counting function is very expensive to evaluate precisely for large x, and the
present implementation is not optimized in any way. For numerical approximation of
the prime counting function, it is better to use primepi2() (page 964) or riemannr()
(page 965).
Some values of the prime counting function:
>>> from sympy.mpmath import *
>>> [primepi(k) for k in range(20)]
[0, 0, 1, 2, 2, 3, 3, 4, 4, 4, 4, 5, 5, 6, 6, 6, 6, 7, 7, 8]
>>> primepi(3.5)
2
>>> primepi(100000)
9592
primepi2()
mpmath.primepi2(x)
Returns an interval (as an mpi instance) providing bounds for the value of the prime
counting function (x). For small x, primepi2() (page 964) returns an exact interval
based on the output of primepi() (page 964). For x > 2656, a loose interval based on
Schoenfelds inequality
x log x
|(x) li(x)| <
8
is returned. This estimate is rigorous assuming the truth of the Riemann hypothesis, and
can be computed very quickly.
Examples
Exact values of the prime counting function for small x:
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = True
>>> iv.dps = 15; iv.pretty = True
>>> primepi2(10)
[4.0, 4.0]
>>> primepi2(100)
[25.0, 25.0]
>>> primepi2(1000)
[168.0, 168.0]
Loose intervals are generated for moderately large x:
>>> primepi2(10000), primepi(10000)
([1209.0, 1283.0], 1229)
>>> primepi2(50000), primepi(50000)
([5070.0, 5263.0], 5133)
As x increases, the absolute error gets worse while the relative error improves. The
exact value of (1023 ) is 1925320391606803968923, and primepi2() (page 964) gives 9
signicant digits:
964
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> p = primepi2(10**23)
>>> p
[1.9253203909477020467e+21, 1.925320392280406229e+21]
>>> mpf(p.delta) / mpf(p.a)
6.9219865355293e-10
A more precise, nonrigorous estimate for (x) can be obtained using the Riemann R function (riemannr() (page 965)). For large enough x, the value returned by primepi2()
(page 964) essentially amounts to a small perturbation of the value returned by riemannr() (page 965):
>>> primepi2(10**100)
[4.3619719871407024816e+97, 4.3619719871407032404e+97]
>>> riemannr(10**100)
4.3619719871407e+97
riemannr()
mpmath.riemannr(x)
Evaluates the Riemann R function, a smooth approximation of the prime counting function (x) (see primepi() (page 964)). The Riemann R function gives a fast numerical
approximation useful e.g. to roughly estimate the number of primes in a given interval.
The Riemann R function is computed using the rapidly convergent Gram series,
R(x) = 1 +
k=1
log x
.
kk!(k + 1)
From the Gram series, one sees that the Riemann R function is a well-dened analytic
function (except for a branch cut along the negative real half-axis); it can be evaluated
for arbitrary real or complex arguments.
The Riemann R function gives a very accurate approximation of the prime counting function. For example, it is wrong by at most 2 for x < 1000, and for x = 109 diers from the
exact value of (x) by 79, or less than two parts in a million. It is about 10 times more
accurate than the logarithmic integral estimate (see li() (page 799)), which however is
even faster to evaluate. It is orders of magnitude more accurate than the extremely fast
x/ log x estimate.
Examples
For small arguments, the Riemann R function almost exactly gives the prime counting
function if rounded to the nearest integer:
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = True
>>> primepi(50), riemannr(50)
(15, 14.9757023241462)
>>> max(abs(primepi(n)-int(round(riemannr(n)))) for n in range(100))
1
>>> max(abs(primepi(n)-int(round(riemannr(n)))) for n in range(300))
2
The Riemann R function can be evaluated for arguments far too large for exact determination of (x) to be computationally feasible with any presently known algorithm:
>>> riemannr(10**30)
1.46923988977204e+28
>>> riemannr(10**100)
5.15. Welcome to mpmaths documentation!
965
SymPy Documentation, Release 0.7.6
4.3619719871407e+97
>>> riemannr(10**1000)
4.3448325764012e+996
A comparison of the Riemann R function and logarithmic integral estimates for (x) using
exact values of (10n ) up to n = 9. The fractional error is shown in parentheses:
>>> exact = [4,25,168,1229,9592,78498,664579,5761455,50847534]
>>> for n, p in enumerate(exact):
...
n += 1
...
r, l = riemannr(10**n), li(10**n)
...
rerr, lerr = nstr((r-p)/p,3), nstr((l-p)/p,3)
...
print(%i %i %s(%s) %s(%s) % (n, p, r, rerr, l, lerr))
...
1 4 4.56458314100509(0.141) 6.1655995047873(0.541)
2 25 25.6616332669242(0.0265) 30.1261415840796(0.205)
3 168 168.359446281167(0.00214) 177.609657990152(0.0572)
4 1229 1226.93121834343(-0.00168) 1246.13721589939(0.0139)
5 9592 9587.43173884197(-0.000476) 9629.8090010508(0.00394)
6 78498 78527.3994291277(0.000375) 78627.5491594622(0.00165)
7 664579 664667.447564748(0.000133) 664918.405048569(0.000511)
8 5761455 5761551.86732017(1.68e-5) 5762209.37544803(0.000131)
9 50847534 50847455.4277214(-1.55e-6) 50849234.9570018(3.35e-5)
The derivative of the Riemann R function gives the approximate probability for a number
of magnitude x to be prime:
>>> diff(riemannr, 1000)
0.141903028110784
>>> mpf(primepi(1050) - primepi(950)) / 100
0.15
Evaluation is supported for arbitrary arguments and at arbitrary precision:
>>> mp.dps = 30
>>> riemannr(7.5)
3.72934743264966261918857135136
>>> riemannr(-4+2j)
(-0.551002208155486427591793957644 + 2.16966398138119450043195899746j)
Cyclotomic polynomials
cyclotomic()
mpmath.cyclotomic(n, x)
Evaluates the cyclotomic polynomial n (x), dened by
n (x) =
(x )
where ranges over all primitive n-th roots of unity (see unitroots() (page 754)). An
equivalent representation, used for computation, is
n (x) =
(xd 1)(n/d) = n (x)
d|n
966
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
where (m) denotes the Moebius function. The cyclotomic polynomials are integer polynomials, the rst of which can be written explicitly as
0 (x) = 1
1 (x) = x 1
2 (x) = x + 1
3 (x) = x3 + x2 + 1
4 (x) = x2 + 1
5 (x) = x4 + x3 + x2 + x + 1
6 (x) = x2 x + 1
Examples
The coecients of low-order cyclotomic polynomials can be recovered using Taylor expansion:
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = True
>>> for n in range(9):
...
p = chop(taylor(lambda x: cyclotomic(n,x), 0, 10))
...
print(%s %s % (n, nstr(p[:10+1-p[::-1].index(1)])))
...
0 [1.0]
1 [-1.0, 1.0]
2 [1.0, 1.0]
3 [1.0, 1.0, 1.0]
4 [1.0, 0.0, 1.0]
5 [1.0, 1.0, 1.0, 1.0, 1.0]
6 [1.0, -1.0, 1.0]
7 [1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0]
8 [1.0, 0.0, 0.0, 0.0, 1.0]
The denition as a product over primitive roots may be checked by computing the product explicitly (for a real argument, this method will generally introduce numerical noise
in the imaginary part):
>>> mp.dps = 25
>>> z = 3+4j
>>> cyclotomic(10, z)
(-419.0 - 360.0j)
>>> fprod(z-r for r in unitroots(10, primitive=True))
(-419.0 - 360.0j)
>>> z = 3
>>> cyclotomic(10, z)
61.0
>>> fprod(z-r for r in unitroots(10, primitive=True))
(61.0 - 3.146045605088568607055454e-25j)
Up to permutation, the roots of a given cyclotomic polynomial can be checked to agree
with the list of primitive roots:
>>> p = taylor(lambda x: cyclotomic(6,x), 0, 6)[:3]
>>> for r in polyroots(p[::-1]):
...
print(r)
...
(0.5 - 0.8660254037844386467637232j)
(0.5 + 0.8660254037844386467637232j)
5.15. Welcome to mpmaths documentation!
967
SymPy Documentation, Release 0.7.6
>>>
>>> for r in unitroots(6, primitive=True):
...
print(r)
...
(0.5 + 0.8660254037844386467637232j)
(0.5 - 0.8660254037844386467637232j)
Arithmetic functions
mangoldt()
mpmath.mangoldt(n)
Evaluates the von Mangoldt function (n) = log p if n = pk a power of a prime, and (n) = 0
otherwise.
Examples
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> [mangoldt(n) for n in range(-2,3)]
[0.0, 0.0, 0.0, 0.0, 0.6931471805599453094172321]
>>> mangoldt(6)
0.0
>>> mangoldt(7)
1.945910149055313305105353
>>> mangoldt(8)
0.6931471805599453094172321
>>> fsum(mangoldt(n) for n in range(101))
94.04531122935739224600493
>>> fsum(mangoldt(n) for n in range(10001))
10013.39669326311478372032
q-functions
q-Pochhammer symbol
qp()
mpmath.qp(a, q=None, n=None, **kwargs)
Evaluates the q-Pochhammer symbol (or q-rising factorial)
(a; q)n =
n1
(1 aq k )
k=0
where n = is permitted if |q| < 1. Called with two arguments, qp(a,q) computes (a; q) ;
with a single argument, qp(q) computes (q; q) . The special case
(q) = (q; q) =
(1 q ) =
k=1
(1)k q (3k
k)/2
k=
is also known as the Euler function, or (up to a factor q 1/24 ) the Dedekind eta function.
Examples
If n is a positive integer, the function amounts to a nite product:
968
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> qp(2,3,5)
-725305.0
>>> fprod(1-2*3**k for k in range(5))
-725305.0
>>> qp(2,3,0)
1.0
Complex arguments are allowed:
>>> qp(2-1j, 0.75j)
(0.4628842231660149089976379 + 4.481821753552703090628793j)
The regular Pochhammer symbol (a)n is obtained in the following limit as q 1:
>>> a, n = 4, 7
>>> limit(lambda q: qp(q**a,q,n) / (1-q)**n, 1)
604800.0
>>> rf(a,n)
604800.0
The Taylor series of the reciprocal Euler function gives the partition function P (n), i.e.
the number of ways of writing n as a sum of positive integers:
>>> taylor(lambda q: 1/qp(q), 0, 10)
[1.0, 1.0, 2.0, 3.0, 5.0, 7.0, 11.0, 15.0, 22.0, 30.0, 42.0]
Special values include:
>>> qp(0)
1.0
>>> findroot(diffun(qp), -0.4)
-0.4112484791779547734440257
>>> qp(_)
1.228348867038575112586878
# location of maximum
The q-Pochhammer symbol is related to the Jacobi theta functions. For example, the
following identity holds:
>>> q = mpf(0.5)
# arbitrary
>>> qp(q)
0.2887880950866024212788997
>>> root(3,-2)*root(q,-24)*jtheta(2,pi/6,root(q,6))
0.2887880950866024212788997
q-gamma and factorial
qgamma()
mpmath.qgamma(z, q, **kwargs)
Evaluates the q-gamma function
q (z) =
(q; q)
(1 q)1z .
(q z ; q)
Examples
Evaluation for real and complex arguments:
5.15. Welcome to mpmaths documentation!
969
SymPy Documentation, Release 0.7.6
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> qgamma(4,0.75)
4.046875
>>> qgamma(6,6)
121226245.0
>>> qgamma(3+4j, 0.5j)
(0.1663082382255199834630088 + 0.01952474576025952984418217j)
The q-gamma function satises a functional equation similar to that of the ordinary
gamma function:
>>> q = mpf(0.25)
>>> z = mpf(2.5)
>>> qgamma(z+1,q)
1.428277424823760954685912
>>> (1-q**z)/(1-q)*qgamma(z,q)
1.428277424823760954685912
qfac()
mpmath.qfac(z, q, **kwargs)
Evaluates the q-factorial,
[n]q ! = (1 + q)(1 + q + q 2 ) (1 + q + + q n1 )
or more generally
[z]q ! =
(q; q)z
.
(1 q)z
Examples
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> qfac(0,0)
1.0
>>> qfac(4,3)
2080.0
>>> qfac(5,6)
121226245.0
>>> qfac(1+1j, 2+1j)
(0.4370556551322672478613695 + 0.2609739839216039203708921j)
Hypergeometric q-series
qhyper()
mpmath.qhyper(a s, b s, q, z, **kwargs)
Evaluates the basic hypergeometric series or hypergeometric q-series
[
]
)1+sr z n
n
(a1 ; q)n , . . . , (ar ; q)n (
a1 a2 . . . ar
(1)n q ( 2 )
; q, z =
r s
b1 b2 . . . bs
(b1 ; q)n , . . . , (bs ; q)n
(q; q)n
n=0
where (a; q)n denotes the q-Pochhammer symbol (see qp() (page 968)).
Examples
Evaluation works for real and complex arguments:
970
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> qhyper([0.5], [2.25], 0.25, 4)
-0.1975849091263356009534385
>>> qhyper([0.5], [2.25], 0.25-0.25j, 4)
(2.806330244925716649839237 + 3.568997623337943121769938j)
>>> qhyper([1+j], [2,3+0.5j], 0.25, 3+4j)
(9.112885171773400017270226 - 1.272756997166375050700388j)
Comparing with a summation of the dening series, using nsum() (page 979):
>>> b, q, z = 3, 0.25, 0.5
>>> qhyper([], [b], q, z)
0.6221136748254495583228324
>>> nsum(lambda n: z**n / qp(q,q,n)/qp(b,q,n) * q**(n*(n-1)), [0,inf])
0.6221136748254495583228324
Numerical calculus
Polynomials
See also taylor() and chebyfit() for approximation of functions by polynomials.
Polynomial evaluation (polyval)
mpmath.polyval(ctx, coes, x, derivative=False)
Given coecients [cn , . . . , c2 , c1 , c0 ] and a number x, polyval() (page 971) evaluates the
polynomial
P (x) = cn xn + . . . + c2 x2 + c1 x + c0 .
If derivative=True is set, polyval() (page 971) simultaneously evaluates P (x) with the
derivative, P 0 (x), and returns the tuple (P (x), P 0 (x)).
>>> from sympy.mpmath import *
>>> mp.pretty = True
>>> polyval([3, 0, 2], 0.5)
2.75
>>> polyval([3, 0, 2], 0.5, derivative=True)
(2.75, 3.0)
The coecients and the evaluation point may be any combination of real or complex
numbers.
Polynomial roots (polyroots)
mpmath.polyroots(ctx, coes, maxsteps=50, cleanup=True,
ror=False)
Computes all roots (real or complex) of a given polynomial.
extraprec=10,
er-
The roots are returned as a sorted list, where real roots appear rst followed by complex conjugate roots as adjacent elements. The polynomial should be given as a list of
coecients, in the format used by polyval() (page 971). The leading coecient must
be nonzero.
With error=True, polyroots() (page 971) returns a tuple (roots, err) where err is an
estimate of the maximum error among the computed roots.
5.15. Welcome to mpmaths documentation!
971
SymPy Documentation, Release 0.7.6
Examples
Finding the three real roots of x3 x2 14x + 24:
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = True
>>> nprint(polyroots([1,-1,-14,24]), 4)
[-4.0, 2.0, 3.0]
Finding the two complex conjugate roots of 4x2 + 3x + 2, with an error estimate:
>>> roots, err = polyroots([4,3,2], error=True)
>>> for r in roots:
...
print(r)
...
(-0.375 + 0.59947894041409j)
(-0.375 - 0.59947894041409j)
>>>
>>> err
2.22044604925031e-16
>>>
>>> polyval([4,3,2], roots[0])
(2.22044604925031e-16 + 0.0j)
>>> polyval([4,3,2], roots[1])
(2.22044604925031e-16 + 0.0j)
The following example computes all the 5th roots of unity; that is, the roots of x5 1:
>>> mp.dps = 20
>>> for r in polyroots([1, 0, 0, 0, 0, -1]):
...
print(r)
...
1.0
(-0.8090169943749474241 + 0.58778525229247312917j)
(-0.8090169943749474241 - 0.58778525229247312917j)
(0.3090169943749474241 + 0.95105651629515357212j)
(0.3090169943749474241 - 0.95105651629515357212j)
Precision and conditioning
The roots are computed to the current working precision accuracy. If this accuracy
cannot be achieved in maxsteps steps, then a N oConvergence exception is raised. The
algorithm internally is using the current working precision extended by extraprec. If
N oConvergence was raised, that is caused either by not having enough extra precision to
achieve convergence (in which case increasing extraprec should x the problem) or too
low maxsteps (in which case increasing maxsteps should x the problem), or a combination
of both.
The user should always do a convergence study with regards to extraprec to ensure accurate results. It is possible to get convergence to a wrong answer with too low extraprec.
Provided there are no repeated roots, polyroots() (page 971) can typically compute all
roots of an arbitrary polynomial to high precision:
>>> mp.dps = 60
>>> for r in polyroots([1, 0, -10, 0, 1]):
...
print r
...
-3.14626436994197234232913506571557044551247712918732870123249
-0.317837245195782244725757617296174288373133378433432554879127
0.317837245195782244725757617296174288373133378433432554879127
972
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
3.14626436994197234232913506571557044551247712918732870123249
>>>
>>> sqrt(3) + sqrt(2)
3.14626436994197234232913506571557044551247712918732870123249
>>> sqrt(3) - sqrt(2)
0.317837245195782244725757617296174288373133378433432554879127
Algorithm
polyroots() (page 971) implements the Durand-Kerner method [1], which uses complex arithmetic to locate all roots simultaneously. The Durand-Kerner method can be
viewed as approximately performing simultaneous Newton iteration for all the roots. In
particular, the convergence to simple roots is quadratic, just like Newtons method.
Although all roots are internally calculated using complex arithmetic, any root found to
have an imaginary part smaller than the estimated numerical error is truncated to a real
number (small real parts are also chopped). Real roots are placed rst in the returned
list, sorted by value. The remaining complex roots are sorted by their real parts so that
conjugate roots end up next to each other.
References
1.http://en.wikipedia.org/wiki/Durand-Kerner method
Root-nding and optimization
Root-nding (findroot)
mpmath.findroot(f, x0, solver=Secant, tol=None, verbose=False, verify=True,
**kwargs)
Find a solution to f (x) = 0, using x0 as starting point or interval for x.
Multidimensional overdetermined systems are supported. You can specify them using a
function or a list of functions.
If the found root does not satisfy |f (x)2 < tol|, an exception is raised (this can be disabled
with verify=False).
Arguments
f one dimensional function
x0 starting point, several starting points or interval (depends on solver)
tol the returned solution has an error smaller than this
verbose print additional information for each iteration if true
verify verify the solution and raise a ValueError if |f (x) > tol|
solver a generator for f and x0 returning approximative solution and error
maxsteps after how many steps the solver will cancel
df rst derivative of f (used by some solvers)
d2f second derivative of f (used by some solvers)
multidimensional force multidimensional solving
J Jacobian matrix of f (used by multidimensional solvers)
norm used vector norm (used by multidimensional solvers)
5.15. Welcome to mpmaths documentation!
973
SymPy Documentation, Release 0.7.6
solver has to be callable with (f, x0, **kwargs) and return an generator yielding
pairs of approximative solution and estimated error (which is expected to be positive).
You can use the following string aliases: secant, mnewton, halley, muller, illinois,
pegasus, anderson, ridder, anewton, bisect
See mpmath.optimization for their documentation.
Examples
The function findroot() (page 973) locates a root of a given function using the secant
method by default. A simple example use of the secant method is to compute as the
root of sin x closest to x0 = 3:
>>> from sympy.mpmath import *
>>> mp.dps = 30; mp.pretty = True
>>> findroot(sin, 3)
3.14159265358979323846264338328
The secant method can be used to nd complex roots of analytic functions, although it
must in that case generally be given a nonreal starting value (or else it will never leave
the real line):
>>> mp.dps = 15
>>> findroot(lambda x: x**3 + 2*x + 1, j)
(0.226698825758202 + 1.46771150871022j)
A nice application is to compute nontrivial roots of the Riemann zeta function with many
digits (good initial values are needed for convergence):
>>> mp.dps = 30
>>> findroot(zeta, 0.5+14j)
(0.5 + 14.1347251417346937904572519836j)
The secant method can also be used as an optimization algorithm, by passing it a derivative of a function. The following example locates the positive minimum of the gamma
function:
>>> mp.dps = 20
>>> findroot(lambda x: diff(gamma, x), 1)
1.4616321449683623413
Finally, a useful application is to compute inverse functions, such as the Lambert W
function which is the inverse of wew , given the rst term of the solutions asymptotic
expansion as the initial value. In basic cases, this gives identical results to mpmaths
built-in lambertw function:
>>> def lambert(x):
...
return findroot(lambda w: w*exp(w) - x, log(1+x))
...
>>> mp.dps = 15
>>> lambert(1); lambertw(1)
0.567143290409784
0.567143290409784
>>> lambert(1000); lambert(1000)
5.2496028524016
5.2496028524016
Multidimensional functions are also supported:
974
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> f = [lambda x1, x2: x1**2 + x2,
...
lambda x1, x2: 5*x1**2 - 3*x1 + 2*x2 - 3]
>>> findroot(f, (0, 0))
[-0.618033988749895]
[-0.381966011250105]
>>> findroot(f, (10, 10))
[ 1.61803398874989]
[-2.61803398874989]
You can verify this by solving the system manually.
Please note that the following (more general) syntax also works:
>>> def f(x1, x2):
...
return x1**2 + x2, 5*x1**2 - 3*x1 + 2*x2 - 3
...
>>> findroot(f, (0, 0))
[-0.618033988749895]
[-0.381966011250105]
Multiple roots
For multiple roots all methods of the Newtonian family (including secant) converge
slowly. Consider this example:
>>> f = lambda x: (x - 1)**99
>>> findroot(f, 0.9, verify=False)
0.918073542444929
Even for a very close starting point the secant method converges very slowly. Use verbose=True to illustrate this.
It is possible to modify Newtons method to make it converge regardless of the roots
multiplicity:
>>> findroot(f, -10, solver=mnewton)
1.0
This variant uses the rst and second derivative of the function, which is not very ecient.
Alternatively you can use an experimental Newtonian solver that keeps track of the speed
of convergence and accelerates it using Steensens method if necessary:
>>> findroot(f, -10, solver=anewton, verbose=True)
x:
-9.88888888888888888889
error: 0.111111111111111111111
converging slowly
x:
-9.77890011223344556678
error: 0.10998877665544332211
converging slowly
x:
-9.67002233332199662166
error: 0.108877778911448945119
converging slowly
accelerating convergence
x:
-9.5622443299551077669
error: 0.107778003366888854764
converging slowly
x:
0.99999999999999999214
error: 10.562244329955107759
x:
1.0
5.15. Welcome to mpmaths documentation!
975
SymPy Documentation, Release 0.7.6
error: 7.8598304758094664213e-18
ZeroDivisionError: canceled with x = 1.0
1.0
Complex roots
For complex roots its recommended to use Mullers method as it converges even for real
starting points very fast:
>>> findroot(lambda x: x**4 + x + 1, (0, 1, 2), solver=muller)
(0.727136084491197 + 0.934099289460529j)
Intersection methods
When you need to nd a root in a known interval, its highly recommended to use an
intersection-based solver like anderson or ridder. Usually they converge faster
and more reliable. They have however problems with multiple roots and usually need a
sign change to nd a root:
>>> findroot(lambda x: x**3, (-1, 1), solver=anderson)
0.0
Be careful with symmetric functions:
>>> findroot(lambda x: x**2, (-1, 1), solver=anderson)
Traceback (most recent call last):
...
ZeroDivisionError
Traceback (most recent call last):
...
ZeroDivisionError
It fails even for better starting points, because there is no sign change:
>>> findroot(lambda x: x**2, (-1, .5), solver=anderson)
Traceback (most recent call last):
...
ValueError: Could not find root within given tolerance. (1 > 2.1684e-19)
Try another starting point or tweak arguments.
Traceback (most recent call last):
...
ValueError: Could not find root within given tolerance. (1 > 2.1684e-19)
Solvers
class mpmath.calculus.optimization.Secant(ctx, f, x0, **kwargs)
1d-solver generating pairs of approximative root and error.
Needs starting points x0 and x1 close to the root. x1 defaults to x0 + 0.25.
Pro:
converges fast
Contra:
converges slowly for multiple roots
class mpmath.calculus.optimization.Newton(ctx, f, x0, **kwargs)
1d-solver generating pairs of approximative root and error.
Needs starting points x0 close to the root.
Pro:
976
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
converges fast
sometimes more robust than secant with bad second starting point
Contra:
converges slowly for multiple roots
needs rst derivative
2 function evaluations per iteration
class mpmath.calculus.optimization.MNewton(ctx, f, x0, **kwargs)
1d-solver generating pairs of approximative root and error.
Needs starting point x0 close to the root. Uses modied Newtons method that converges
fast regardless of the multiplicity of the root.
Pro:
converges fast for multiple roots
Contra:
needs rst and second derivative of f
3 function evaluations per iteration
class mpmath.calculus.optimization.Halley(ctx, f, x0, **kwargs)
1d-solver generating pairs of approximative root and error.
Needs a starting point x0 close to the root. Uses Halleys method with cubic convergence
rate.
Pro:
converges even faster the Newtons method
useful when computing with many digits
Contra:
needs rst and second derivative of f
3 function evaluations per iteration
converges slowly for multiple roots
class mpmath.calculus.optimization.Muller(ctx, f, x0, **kwargs)
1d-solver generating pairs of approximative root and error.
Needs starting points x0, x1 and x2 close to the root. x1 defaults to x0 + 0.25; x2 to x1
+ 0.25. Uses Mullers method that converges towards complex roots.
Pro:
converges fast (somewhat faster than secant)
can nd complex roots
Contra:
converges slowly for multiple roots
may have complex values for real starting points and real roots
http://en.wikipedia.org/wiki/Mullers method
class mpmath.calculus.optimization.Bisection(ctx, f, x0, **kwargs)
1d-solver generating pairs of approximative root and error.
Uses bisection method to nd a root of f in [a, b]. Might fail for multiple roots (needs
sign change).
5.15. Welcome to mpmaths documentation!
977
SymPy Documentation, Release 0.7.6
Pro:
robust and reliable
Contra:
converges slowly
needs sign change
class mpmath.calculus.optimization.Illinois(ctx, f, x0, **kwargs)
1d-solver generating pairs of approximative root and error.
Uses Illinois method or similar to nd a root of f in [a, b]. Might fail for multiple roots
(needs sign change). Combines bisect with secant (improved regula falsi).
The only dierence between the methods is the scaling factor m, which is used to ensure
convergence (you can choose one using the method keyword):
Illinois method (illinois): m = 0.5
Pegasus method (pegasus): m = fb/(fb + fz)
Anderson-Bjoerk method (anderson): m = 1 - fz/fb if positive else 0.5
Pro:
converges very fast
Contra:
has problems with multiple roots
needs sign change
class mpmath.calculus.optimization.Pegasus
1d-solver generating pairs of approximative root and error.
Uses Pegasus method to nd a root of f in [a, b].
method=pegasus.
Wrapper for illinois to use
class mpmath.calculus.optimization.Anderson
1d-solver generating pairs of approximative root and error.
Uses Anderson-Bjoerk method to nd a root of f in [a, b]. Wrapper for illinois to use
method=pegasus.
class mpmath.calculus.optimization.Ridder(ctx, f, x0, **kwargs)
1d-solver generating pairs of approximative root and error.
Ridders method to nd a root of f in [a, b]. Is told to perform as well as Brents method
while being simpler.
Pro:
very fast
simpler than Brents method
Contra:
two function evaluations per step
has problems with multiple roots
needs sign change
http://en.wikipedia.org/wiki/Ridders method
978
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
class mpmath.calculus.optimization.ANewton(ctx, f, x0, **kwargs)
EXPERIMENTAL 1d-solver generating pairs of approximative root and error.
Uses Newtons method modied to use Steensens method when convergence is slow.
(I.e. for multiple roots.)
class mpmath.calculus.optimization.MDNewton(ctx, f, x0, **kwargs)
Find the root of a vector function numerically using Newtons method.
f is a vector function representing a nonlinear equation system.
x0 is the starting point close to the root.
J is a function returning the Jacobian matrix for a point.
Supports overdetermined systems.
Use the norm keyword to specify which norm to use. Defaults to max-norm. The function to calculate the Jacobian matrix can be given using the keyword J. Otherwise it will
be calculated numerically.
Please note that this method converges only locally. Especially for high- dimensional
systems it is not trivial to nd a good starting point being close enough to the root.
It is recommended to use a faster, low-precision solver from SciPy [1] or OpenOpt [2]
to get an initial guess. Afterwards you can use this method for root-polishing to any
precision.
[1] http://scipy.org
[2] http://openopt.org/Welcome
Sums, products, limits and extrapolation
The functions listed here permit approximation of innite sums, products, and other sequence
limits. Use mpmath.fsum() (page 732) and mpmath.fprod() (page 732) for summation and
multiplication of nite sequences.
Summation
nsum()
mpmath.nsum(ctx, f, *intervals, **options)
Computes the sum
S=
f (k)
k=a
where (a, b) = interval, and where a = and/or b = are allowed, or more generally
S=
b1
k1 =a1
bn
f (k1 , . . . , kn )
kn =an
if multiple intervals are given.
Two examples of innite series that can be summed by nsum() (page 979), where the
rst converges rapidly and the second converges slowly, are:
5.15. Welcome to mpmaths documentation!
979
SymPy Documentation, Release 0.7.6
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = True
>>> nsum(lambda n: 1/fac(n), [0, inf])
2.71828182845905
>>> nsum(lambda n: 1/n**2, [1, inf])
1.64493406684823
When appropriate, nsum() (page 979) applies convergence acceleration to accurately estimate the sums of slowly convergent series. If the series is nite, nsum() (page 979) currently does not attempt to perform any extrapolation, and simply calls fsum() (page 732).
Multidimensional innite series are reduced to a single-dimensional series over expanding hypercubes; if both innite and nite dimensions are present, the nite ranges are
moved innermost. For more advanced control over the summation order, use nested
calls to nsum() (page 979), or manually rewrite the sum as a single-dimensional series.
Options
tol Desired maximum nal error. Defaults roughly to the epsilon of the working precision.
method Which summation algorithm to use (described below). Default: richardson+shanks.
maxterms Cancel after at most this many terms. Default: 10*dps.
steps An iterable giving the number of terms to add between each extrapolation attempt. The default sequence is [10, 20, 30, 40, ...]. For example, if you know that
approximately 100 terms will be required, eciency might be improved by setting
this to [100, 10]. Then the rst extrapolation will be performed after 100 terms, the
second after 110, etc.
verbose Print details about progress.
ignore If enabled, any term that raises ArithmeticError or ValueError (e.g. through
division by zero) is replaced by a zero. This is convenient for lattice sums with a
singular term near the origin.
Methods
Unfortunately, an algorithm that can eciently sum any innite series does not exist.
nsum() (page 979) implements several dierent algorithms that each work well in different cases. The method keyword argument selects a method.
The default method is r+s, i.e. both Richardson extrapolation and Shanks transformation is attempted. A slower method that handles more cases is r+s+e. For very high
precision summation, or if the summation needs to be fast (for example if multiple sums
need to be evaluated), it is a good idea to investigate which one method works best and
only use that.
richardson / r: Uses Richardson extrapolation. Provides useful extrapolation
when f (k) P (k)/Q(k) or when f (k) (1)k P (k)/Q(k) for polynomials P and Q. See
richardson() (page 991) for additional information.
shanks / s: Uses Shanks transformation. Typically provides useful extrapolation
when f (k) ck or when successive terms alternate signs. Is able to sum some divergent series. See shanks() (page 992) for additional information.
levin / l: Uses the Levin transformation. It performs better than the Shanks
transformation for logarithmic convergent or alternating divergent series. The
levin variant-keyword selects the variant. Valid choices are u, t, v and
980
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
all whereby all uses all three u,t and v simultanously (This is good for performance comparison in conjunction with verbose=True). Instead of the Levin transform one can also use the Sidi-S transform by selecting the method sidi. See
levin() (page 994) for additional details.
alternating / a: This is the convergence acceleration of alternating series developped by Cohen, Villegras and Zagier. See cohen alt() (page 998) for additional
details.
euler-maclaurin / e: Uses the Euler-Maclaurin summation formula to approximate the remainder sum by an integral. This requires high-order numerical derivatives and numerical integration. The advantage of this algorithm is that it works
regardless of the decay rate of f , as long as f is suciently smooth. See sumem()
(page 986) for additional information.
direct / d: Does not perform any extrapolation. This can be used (and should only
be used for) rapidly convergent series. The summation automatically stops when
the terms decrease below the target tolerance.
Basic examples
A nite sum:
>>> nsum(lambda k: 1/k, [1, 6])
2.45
Summation of a series going to negative innity and a doubly innite series:
>>> nsum(lambda k: 1/k**2, [-inf, -1])
1.64493406684823
>>> nsum(lambda k: 1/(1+k**2), [-inf, inf])
3.15334809493716
nsum() (page 979) handles sums of complex numbers:
>>> nsum(lambda k: (0.5+0.25j)**k, [0, inf])
(1.6 + 0.8j)
The following sum converges very rapidly, so it is most ecient to sum it by disabling
convergence acceleration:
>>> mp.dps = 1000
>>> a = nsum(lambda k: -(-1)**k * k**2 / fac(2*k), [1, inf],
...
method=direct)
>>> b = (cos(1)+sin(1))/4
>>> abs(a-b) < mpf(1e-998)
True
Examples with Richardson extrapolation
Richardson extrapolation works well for sums over rational functions, as well as their
alternating counterparts:
>>> mp.dps = 50
>>> nsum(lambda k: 1 / k**3, [1, inf],
...
method=richardson)
1.2020569031595942853997381615114499907649862923405
>>> zeta(3)
1.2020569031595942853997381615114499907649862923405
>>> nsum(lambda n: (n + 3)/(n**3 + n**2), [1, inf],
...
method=richardson)
5.15. Welcome to mpmaths documentation!
981
SymPy Documentation, Release 0.7.6
2.9348022005446793094172454999380755676568497036204
>>> pi**2/2-2
2.9348022005446793094172454999380755676568497036204
>>> nsum(lambda k: (-1)**k / k**3, [1, inf],
...
method=richardson)
-0.90154267736969571404980362113358749307373971925537
>>> -3*zeta(3)/4
-0.90154267736969571404980362113358749307373971925538
Examples with Shanks transformation
The Shanks transformation works well for geometric series and typically provides excellent acceleration for Taylor series near the border of their disk of convergence. Here we
apply it to a series for log(2), which can be seen as the Taylor series for log(1 + x) with
x = 1:
>>> nsum(lambda k: -(-1)**k/k, [1, inf],
...
method=shanks)
0.69314718055994530941723212145817656807550013436025
>>> log(2)
0.69314718055994530941723212145817656807550013436025
Here we apply it to a slowly convergent geometric series:
>>> nsum(lambda k: mpf(0.995)**k, [0, inf],
...
method=shanks)
200.0
Finally, Shanks method works very well for alternating series where f (k) = (1)k g(k),
and often does so regardless of the exact decay rate of g(k):
>>> mp.dps = 15
>>> nsum(lambda k: (-1)**(k+1) / k**1.5, [1, inf],
...
method=shanks)
0.765147024625408
>>> (2-sqrt(2))*zeta(1.5)/2
0.765147024625408
The following slowly convergent alternating series has no known closed-form value.
Evaluating the sum a second time at higher precision indicates that the value is probably
correct:
>>> nsum(lambda k: (-1)**k / log(k), [2, inf],
...
method=shanks)
0.924299897222939
>>> mp.dps = 30
>>> nsum(lambda k: (-1)**k / log(k), [2, inf],
...
method=shanks)
0.92429989722293885595957018136
Examples with Levin transformation
The following example calculates Eulers constant as the constant term in the Laurent
expansion of zeta(s) at s=1. This sum converges extremly slow because of the logarithmic convergence behaviour of the Dirichlet series for zeta.
>>> mp.dps = 30
>>> z = mp.mpf(10) ** (-10)
>>> a = mp.nsum(lambda n: n**(-(1+z)), [1, mp.inf], method = levin) - 1 / z
982
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> print(mp.chop(a - mp.euler, tol = 1e-10))
0.0
Now we sum the zeta function outside its range of convergence (attention: This does not
work at the negative integers!):
>>> mp.dps = 15
>>> w = mp.nsum(lambda n: n ** (2 + 3j), [1, mp.inf], method = levin, levin_variant = v)
>>> print(mp.chop(w - mp.zeta(-2-3j)))
0.0
The next example resummates an asymptotic series expansion of an integral related to
the exponential integral.
>>>
>>>
>>>
>>>
>>>
>>>
0.0
mp.dps = 15
z = mp.mpf(10)
# exact = mp.quad(lambda x: mp.exp(-x)/(1+x/z),[0,mp.inf])
exact = z * mp.exp(z) * mp.expint(1,z) # this is the symbolic expression for the integral
w = mp.nsum(lambda n: (-1) ** n * mp.fac(n) * z ** (-n), [0, mp.inf], method = sidi, levin
print(mp.chop(w - exact))
Following highly divergent asymptotic expansion needs some care. Firstly we need copious amount of working precision. Secondly the stepsize must not be chosen to large,
otherwise nsum may miss the point where the Levin transform converges and reach the
point where only numerical garbage is produced due to numerical cancellation.
>>>
>>>
>>>
>>>
>>>
...
>>>
0.0
mp.dps = 15
z = mp.mpf(2)
# exact = mp.quad(lambda x: mp.exp( -x * x / 2 - z * x ** 4), [0,mp.inf]) * 2 / mp.sqrt(2 *
exact = mp.exp(mp.one / (32 * z)) * mp.besselk(mp.one / 4, mp.one / (32 * z)) / (4 * mp.sqrt
w = mp.nsum(lambda n: (-z)**n * mp.fac(4 * n) / (mp.fac(n) * mp.fac(2 * n) * (4 ** n)),
[0, mp.inf], method = levin, levin_variant = t, workprec = 8*mp.prec, steps = [2] + [1
print(mp.chop(w - exact))
The hypergeoemtric function can also be summed outside its range of convergence:
>>>
>>>
>>>
>>>
>>>
>>>
0.0
mp.dps = 15
z = 2 + 1j
exact = mp.hyp2f1(2 / mp.mpf(3), 4 / mp.mpf(3), 1 / mp.mpf(3), z)
f = lambda n: mp.rf(2 / mp.mpf(3), n) * mp.rf(4 / mp.mpf(3), n) * z**n / (mp.rf(1 / mp.mpf(3
v = mp.nsum(f, [0, mp.inf], method = levin, steps = [10 for x in xrange(1000)])
print(mp.chop(exact-v))
Examples with Cohens alternating series resummation
The next example sums the alternating zeta function:
>>> v = mp.nsum(lambda n: (-1)**(n-1) / n, [1, mp.inf], method = a)
>>> print(mp.chop(v - mp.log(2)))
0.0
The derivate of the alternating zeta function outside its range of convergence:
>>> v = mp.nsum(lambda n: (-1)**n * mp.log(n) * n, [1, mp.inf], method = a)
>>> print(mp.chop(v - mp.diff(lambda s: mp.altzeta(s), -1)))
0.0
Examples with Euler-Maclaurin summation
5.15. Welcome to mpmaths documentation!
983
SymPy Documentation, Release 0.7.6
The sum in the following example has the wrong rate of convergence for either Richardson or Shanks to be eective.
>>> f = lambda k: log(k)/k**2.5
>>> mp.dps = 15
>>> nsum(f, [1, inf], method=euler-maclaurin)
0.38734195032621
>>> -diff(zeta, 2.5)
0.38734195032621
Increasing steps improves speed at higher precision:
>>> mp.dps = 50
>>> nsum(f, [1, inf], method=euler-maclaurin, steps=[250])
0.38734195032620997271199237593105101319948228874688
>>> -diff(zeta, 2.5)
0.38734195032620997271199237593105101319948228874688
Divergent series
The Shanks transformation is able to sum some divergent series. In particular, it is often
able to sum Taylor series beyond their radius of convergence (this is due to a relation between the Shanks transformation and Pade approximations; see pade() (page 1015) for
an alternative way to evaluate divergent Taylor series). Furthermore the Levin-transform
examples above contain some divergent series resummation.
Here we apply it to log(1 + x) far outside the region of convergence:
>>> mp.dps = 50
>>> nsum(lambda k: -(-9)**k/k, [1, inf],
...
method=shanks)
2.3025850929940456840179914546843642076011014886288
>>> log(10)
2.3025850929940456840179914546843642076011014886288
A particular type of divergent series that can be summed using the Shanks transformation is geometric series. The result is the same as using the closed-form formula for an
innite geometric series:
>>> mp.dps = 15
>>> for n in range(-8, 8):
...
if n == 1:
...
continue
...
print(%s %s %s % (mpf(n), mpf(1)/(1-n),
...
nsum(lambda k: n**k, [0, inf], method=shanks)))
...
-8.0 0.111111111111111 0.111111111111111
-7.0 0.125 0.125
-6.0 0.142857142857143 0.142857142857143
-5.0 0.166666666666667 0.166666666666667
-4.0 0.2 0.2
-3.0 0.25 0.25
-2.0 0.333333333333333 0.333333333333333
-1.0 0.5 0.5
0.0 1.0 1.0
2.0 -1.0 -1.0
3.0 -0.5 -0.5
4.0 -0.333333333333333 -0.333333333333333
5.0 -0.25 -0.25
6.0 -0.2 -0.2
984
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
7.0 -0.166666666666667 -0.166666666666667
Multidimensional sums
Any combination of nite and innite ranges is allowed for the summation indices:
>>> mp.dps = 15
>>> nsum(lambda
28.0
>>> nsum(lambda
6.0
>>> nsum(lambda
6.0
>>> nsum(lambda
7.0
>>> nsum(lambda
7.0
>>> nsum(lambda
7.0
x,y: x+y, [2,3], [4,5])
x,y: x/2**y, [1,3], [1,inf])
x,y: y/2**x, [1,inf], [1,3])
x,y,z: z/(2**x*2**y), [1,inf], [1,inf], [3,4])
x,y,z: y/(2**x*2**z), [1,inf], [3,4], [1,inf])
x,y,z: x/(2**z*2**y), [3,4], [1,inf], [1,inf])
Some nice examples of double series with analytic solutions or reductions to singledimensional series (see [1]):
>>> nsum(lambda m, n: 1/2**(m*n), [1,inf], [1,inf])
1.60669515241529
>>> nsum(lambda n: 1/(2**n-1), [1,inf])
1.60669515241529
>>> nsum(lambda i,j: (-1)**(i+j)/(i**2+j**2), [1,inf], [1,inf])
0.278070510848213
>>> pi*(pi-3*ln2)/12
0.278070510848213
>>> nsum(lambda i,j: (-1)**(i+j)/(i+j)**2, [1,inf], [1,inf])
0.129319852864168
>>> altzeta(2) - altzeta(1)
0.129319852864168
>>> nsum(lambda i,j: (-1)**(i+j)/(i+j)**3, [1,inf], [1,inf])
0.0790756439455825
>>> altzeta(3) - altzeta(2)
0.0790756439455825
>>> nsum(lambda m,n: m**2*n/(3**m*(n*3**m+m*3**n)),
...
[1,inf], [1,inf])
0.28125
>>> mpf(9)/32
0.28125
>>> nsum(lambda i,j: fac(i-1)*fac(j-1)/fac(i+j),
...
[1,inf], [1,inf], workprec=400)
1.64493406684823
>>> zeta(2)
1.64493406684823
A hard example of a multidimensional sum is the Madelung constant in three dimensions
(see [2]). The dening sum converges very slowly and only conditionally, so nsum()
(page 979) is lucky to obtain an accurate value through convergence acceleration. The
second evaluation below uses a much more ecient, rapidly convergent 2D sum:
5.15. Welcome to mpmaths documentation!
985
SymPy Documentation, Release 0.7.6
>>> nsum(lambda x,y,z: (-1)**(x+y+z)/(x*x+y*y+z*z)**0.5,
...
[-inf,inf], [-inf,inf], [-inf,inf], ignore=True)
-1.74756459463318
>>> nsum(lambda x,y: -12*pi*sech(0.5*pi * \
...
sqrt((2*x+1)**2+(2*y+1)**2))**2, [0,inf], [0,inf])
-1.74756459463318
Another example of a lattice sum in 2D:
>>> nsum(lambda x,y: (-1)**(x+y) / (x**2+y**2), [-inf,inf],
...
[-inf,inf], ignore=True)
-2.1775860903036
>>> -pi*ln2
-2.1775860903036
An example of an Eisenstein series:
>>> nsum(lambda m,n: (m+n*1j)**(-4), [-inf,inf], [-inf,inf],
...
ignore=True)
(3.1512120021539 + 0.0j)
References
1.[Weisstein] (page 1910) http://mathworld.wolfram.com/DoubleSeries.html,
2.[Weisstein] (page 1910) http://mathworld.wolfram.com/MadelungConstants.html
sumem()
mpmath.sumem(ctx, f, interval, tol=None, reject=10, integral=None, adis=None, bdis=None, verbose=False, error=False, fast abort=False)
Uses the Euler-Maclaurin formula to compute an approximation accurate to within tol
(which defaults to the present epsilon) of the sum
S=
f (k)
k=a
where (a, b) are given by interval and a or b may be innite. The approximation is
f (x) dx +
a
)
f (a) + f (b) B2k ( (2k1)
+
f
(b) f (2k1) (a) .
2
(2k)!
k=1
The last sum in the Euler-Maclaurin formula is not generally convergent (a notable exception is if f is a polynomial, in which case Euler-Maclaurin actually gives an exact
result).
The summation is stopped as soon as the quotient between two consecutive terms falls
below reject. That is, by default (reject = 10), the summation is continued as long as
each term adds at least one decimal.
Although not convergent, convergence to a given tolerance can often be forced if b =
by summing up to a + N and then applying the Euler-Maclaurin formula to the sum over
the range (a + N + 1, . . . , ). This procedure is implemented by nsum() (page 979).
By default numerical quadrature and dierentiation is used. If the symbolic values of the
integral and endpoint derivatives are known, it is more ecient to pass the value of the
integral explicitly as integral and the derivatives explicitly as adiffs and bdiffs. The
derivatives should be given as iterables that yield f (a), f 0 (a), f 00 (a), . . . (and the equivalent
for b).
986
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
Summation of an innite series, with automatic and symbolic integral and derivative
values (the second should be much faster):
>>> from sympy.mpmath import *
>>> mp.dps = 50; mp.pretty = True
>>> sumem(lambda n: 1/n**2, [32, inf])
0.03174336652030209012658168043874142714132886413417
>>> I = mpf(1)/32
>>> D = adiffs=((-1)**n*fac(n+1)*32**(-2-n) for n in range(999))
>>> sumem(lambda n: 1/n**2, [32, inf], integral=I, adiffs=D)
0.03174336652030209012658168043874142714132886413417
An exact evaluation of a nite polynomial sum:
>>> sumem(lambda n: n**5-12*n**2+3*n, [-100000, 200000])
10500155000624963999742499550000.0
>>> print(sum(n**5-12*n**2+3*n for n in range(-100000, 200001)))
10500155000624963999742499550000
sumap()
mpmath.sumap(ctx, f, interval, integral=None, error=False)
Evaluates an innite series of an analytic summand f using the Abel-Plana formula
k=0
f (k) =
0
1
f (t)dt + f (0) + i
2
f (it) f (it)
dt.
e2t 1
Unlike the Euler-Maclaurin formula (see sumem() (page 986)), the Abel-Plana formula
does not require derivatives. However, it only works when |f (it)f (it)| does not increase
too rapidly with t.
Examples
The Abel-Plana formula is particularly useful when the summand decreases like a power
of k; for example when the sum is a pure zeta function:
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> sumap(lambda k: 1/k**2.5, [1,inf])
1.34148725725091717975677
>>> zeta(2.5)
1.34148725725091717975677
>>> sumap(lambda k: 1/(k+1j)**(2.5+2.5j), [1,inf])
(-3.385361068546473342286084 - 0.7432082105196321803869551j)
>>> zeta(2.5+2.5j, 1+1j)
(-3.385361068546473342286084 - 0.7432082105196321803869551j)
If the series is alternating, numerical quadrature along the real line is likely to give poor
results, so it is better to evaluate the rst term symbolically whenever possible:
>>> n=3; z=-0.75
>>> I = expint(n,-log(z))
>>> chop(sumap(lambda k: z**k / k**n, [1,inf], integral=I))
-0.6917036036904594510141448
>>> polylog(n,z)
-0.6917036036904594510141448
5.15. Welcome to mpmaths documentation!
987
SymPy Documentation, Release 0.7.6
Products
nprod()
mpmath.nprod(ctx, f, interval, nsum=False, **kwargs)
Computes the product
P =
f (k)
k=a
where (a, b) = interval, and where a = and/or b = are allowed.
By default, nprod() (page 988) uses the same extrapolation methods as nsum()
(page 979), except applied to the partial products rather than partial sums, and the same
keyword options as for nsum() (page 979) are supported. If nsum=True, the product is
instead computed via nsum() (page 979) as
( b
)
P = exp
log(f (k)) .
k=a
This is slower, but can sometimes yield better results. It is also required (and used
automatically) when Euler-Maclaurin summation is requested.
Examples
A simple nite product:
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> nprod(lambda k: k, [1, 4])
24.0
A large number of innite products have known exact values, and can therefore be used
as a reference. Most of the following examples are taken from MathWorld [1].
A few innite products with simple values are:
>>> 2*nprod(lambda k: (4*k**2)/(4*k**2-1), [1, inf])
3.141592653589793238462643
>>> nprod(lambda k: (1+1/k)**2/(1+2/k), [1, inf])
2.0
>>> nprod(lambda k: (k**3-1)/(k**3+1), [2, inf])
0.6666666666666666666666667
>>> nprod(lambda k: (1-1/k**2), [2, inf])
0.5
Next, several more innite products with more complicated values:
>>> nprod(lambda k: exp(1/k**2), [1, inf]); exp(pi**2/6)
5.180668317897115748416626
5.180668317897115748416626
>>> nprod(lambda k: (k**2-1)/(k**2+1), [2, inf]); pi*csch(pi)
0.2720290549821331629502366
0.2720290549821331629502366
>>> nprod(lambda k: (k**4-1)/(k**4+1), [2, inf])
0.8480540493529003921296502
>>> pi*sinh(pi)/(cosh(sqrt(2)*pi)-cos(sqrt(2)*pi))
988
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
0.8480540493529003921296502
>>> nprod(lambda k: (1+1/k+1/k**2)**2/(1+2/k+3/k**2), [1, inf])
1.848936182858244485224927
>>> 3*sqrt(2)*cosh(pi*sqrt(3)/2)**2*csch(pi*sqrt(2))/pi
1.848936182858244485224927
>>> nprod(lambda k: (1-1/k**4), [2, inf]); sinh(pi)/(4*pi)
0.9190194775937444301739244
0.9190194775937444301739244
>>> nprod(lambda k: (1-1/k**6), [2, inf])
0.9826842777421925183244759
>>> (1+cosh(pi*sqrt(3)))/(12*pi**2)
0.9826842777421925183244759
>>> nprod(lambda k: (1+1/k**2), [2, inf]); sinh(pi)/(2*pi)
1.838038955187488860347849
1.838038955187488860347849
>>> nprod(lambda n: (1+1/n)**n * exp(1/(2*n)-1), [1, inf])
1.447255926890365298959138
>>> exp(1+euler/2)/sqrt(2*pi)
1.447255926890365298959138
The following two products are equivalent and can be evaluated in terms of a Jacobi
theta function. Pi can be replaced by any value (as long as convergence is preserved):
>>> nprod(lambda k: (1-pi**-k)/(1+pi**-k), [1, inf])
0.3838451207481672404778686
>>> nprod(lambda k: tanh(k*log(pi)/2), [1, inf])
0.3838451207481672404778686
>>> jtheta(4,0,1/pi)
0.3838451207481672404778686
This product does not have a known closed form value:
>>> nprod(lambda k: (1-1/2**k), [1, inf])
0.2887880950866024212788997
A product taken from :
>>> nprod(lambda k: 1-k**(-3), [-inf,-2])
0.8093965973662901095786805
>>> cosh(pi*sqrt(3)/2)/(3*pi)
0.8093965973662901095786805
A doubly innite product:
>>> nprod(lambda k: exp(1/(1+k**2)), [-inf, inf])
23.41432688231864337420035
>>> exp(pi/tanh(pi))
23.41432688231864337420035
A product requiring the use of Euler-Maclaurin summation to compute an accurate value:
>>> nprod(lambda k: (1-1/k**2.5), [2, inf], method=e)
0.696155111336231052898125
References
5.15. Welcome to mpmaths documentation!
989
SymPy Documentation, Release 0.7.6
1.[Weisstein] (page 1910) http://mathworld.wolfram.com/InniteProduct.html
Limits (limit)
limit()
mpmath.limit(ctx, f, x, direction=1, exp=False, **kwargs)
Computes an estimate of the limit
lim f (t)
tx
where x may be nite or innite.
For nite x, limit() (page 990) evaluates f (x + d/n) for consecutive integer values of n,
where the approach direction d may be specied using the direction keyword argument.
For innite x, limit() (page 990) evaluates values of f (sign(x) n).
If the approach to the limit is not suciently fast to give an accurate estimate directly,
limit() (page 990) attempts to nd the limit using Richardson extrapolation or the
Shanks transformation. You can select between these methods using the method keyword (see documentation of nsum() (page 979) for more information).
Options
The following options are available with essentially the same meaning as for nsum()
(page 979): tol, method, maxterms, steps, verbose.
If the option exp=True is set, f will be sampled at exponentially spaced points n =
21 , 22 , 23 , . . . instead of the linearly spaced points n = 1, 2, 3, . . .. This can sometimes improve the rate of convergence so that limit() (page 990) may return a more accurate
answer (and faster). However, do note that this can only be used if f supports fast and accurate evaluation for arguments that are extremely close to the limit point (or if innite,
very large arguments).
Examples
A basic evaluation of a removable singularity:
>>> from sympy.mpmath import *
>>> mp.dps = 30; mp.pretty = True
>>> limit(lambda x: (x-sin(x))/x**3, 0)
0.166666666666666666666666666667
Computing the exponential function using its limit denition:
>>> limit(lambda n: (1+3/n)**n, inf)
20.0855369231876677409285296546
>>> exp(3)
20.0855369231876677409285296546
A limit for :
>>> f = lambda n: 2**(4*n+1)*fac(n)**4/(2*n+1)/fac(2*n)**2
>>> limit(f, inf)
3.14159265358979323846264338328
Calculating the coecient in Stirlings formula:
990
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> limit(lambda n: fac(n) / (sqrt(n)*(n/e)**n), inf)
2.50662827463100050241576528481
>>> sqrt(2*pi)
2.50662827463100050241576528481
Evaluating Eulers constant using the limit representation
[( n )
]
1
log(n)
= lim
n
k
k=1
(which converges notoriously slowly):
>>> f = lambda n: sum([mpf(1)/k for k in range(1,int(n)+1)]) - log(n)
>>> limit(f, inf)
0.577215664901532860606512090082
>>> +euler
0.577215664901532860606512090082
With default settings, the following limit converges too slowly to be evaluated accurately.
Changing to exponential sampling however gives a perfect result:
>>> f = lambda x: sqrt(x**3+x**2)/(sqrt(x**3)+x)
>>> limit(f, inf)
0.992831158558330281129249686491
>>> limit(f, inf, exp=True)
1.0
Extrapolation The following functions provide a direct interface to extrapolation algorithms. nsum() and limit() essentially work by calling the following functions with an increasing number of terms until the extrapolated limit is accurate enough.
The following functions may be useful to call directly if the precise number of terms needed
to achieve a desired accuracy is known in advance, or if one wishes to study the convergence
properties of the algorithms.
richardson()
mpmath.richardson(ctx, seq)
Given a list seq of the rst N elements of a slowly convergent innite sequence, richardson() (page 991) computes the N -term Richardson extrapolate for the limit.
richardson() (page 991) returns (v, c) where v is the estimated limit and c is the magnitude of the largest weight used during the computation. The weight provides an estimate
of the precision lost to cancellation. Due to cancellation eects, the sequence must be
typically be computed at a much higher precision than the target accuracy of the extrapolation.
Applicability and issues
The N -step Richardson extrapolation algorithm used by richardson() (page 991) is described in [1].
Richardson extrapolation only works for a specic type of sequence, namely one converging like partial sums of P (1)/Q(1) + P (2)/Q(2) + . . . where P and Q are polynomials. When
the sequence does not convergence at such a rate richardson() (page 991) generally
produces garbage.
5.15. Welcome to mpmaths documentation!
991
SymPy Documentation, Release 0.7.6
Richardson extrapolation has the advantage of being fast: the N -term extrapolate requires only O(N ) arithmetic operations, and usually produces an estimate that is accurate to O(N ) digits. Contrast with the Shanks transformation (see shanks() (page 992)),
which requires O(N 2 ) operations.
richardson() (page 991) is unable to produce an estimate for the approximation error.
One way to estimate the error is to perform two extrapolations with slightly dierent N
and comparing the results.
Richardson extrapolation does not work for oscillating sequences.
As a simple
workaround, richardson() (page 991) detects if the last three elements do not dier
monotonically, and in that case applies extrapolation only to the even-index elements.
Example
Applying Richardson extrapolation to the Leibniz series for :
>>> from sympy.mpmath import *
>>> mp.dps = 30; mp.pretty = True
>>> S = [4*sum(mpf(-1)**n/(2*n+1) for n in range(m))
...
for m in range(1,30)]
>>> v, c = richardson(S[:10])
>>> v
3.2126984126984126984126984127
>>> nprint([v-pi, c])
[0.0711058, 2.0]
>>> v, c = richardson(S[:30])
>>> v
3.14159265468624052829954206226
>>> nprint([v-pi, c])
[1.09645e-9, 20833.3]
References
1.[BenderOrszag] (page 1910) pp. 375-376
shanks()
mpmath.shanks(ctx, seq, table=None, randomized=False)
Given a list seq of the rst N elements of a slowly convergent innite sequence (Ak ), shanks() (page 992) computes the iterated Shanks transformation
S(A), S(S(A)), . . . , S N /2 (A). The Shanks transformation often provides strong convergence
acceleration, especially if the sequence is oscillating.
The iterated Shanks transformation is computed using the Wynn epsilon algorithm (see
[1]). shanks() (page 992) returns the full epsilon table generated by Wynns algorithm,
which can be read o as follows:
The table is a list of lists forming a lower triangular matrix, where higher row and
column indices correspond to more accurate values.
The columns with even index hold dummy entries (required for the computation)
and the columns with odd index hold the actual extrapolates.
The last element in the last row is typically the most accurate estimate of the limit.
The dierence to the third last element in the last row provides an estimate of the
approximation error.
The magnitude of the second last element provides an estimate of the numerical
accuracy lost to cancellation.
992
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
For convenience, so the extrapolation is stopped at an odd index so that shanks(seq)[1][-1] always gives an estimate of the limit.
Optionally, an existing table can be passed to shanks() (page 992). This can be used
to eciently extend a previous computation after new elements have been appended to
the sequence. The table will then be updated in-place.
The Shanks transformation
The Shanks transformation is dened as follows (see [2]): given the input sequence
(A0 , A1 , . . .), the transformed sequence is given by
S(Ak ) =
Ak+1 Ak1 A2k
Ak+1 + Ak1 2Ak
The Shanks transformation gives the exact limit A in a single step if Ak = A + aq k . Note
in particular that it extrapolates the exact sum of a geometric series in a single step.
Applying the Shanks transformation once often improves convergence substantially for
an arbitrary sequence, but the optimal eect is obtained by applying it iteratively:
S(S(Ak )), S(S(S(Ak ))), . . ..
Wynns epsilon algorithm provides an ecient way to generate the table of iterated
Shanks transformations. It reduces the computation of each element to essentially a
single division, at the cost of requiring dummy elements in the table. See [1] for details.
Precision issues
Due to cancellation eects, the sequence must be typically be computed at a much higher
precision than the target accuracy of the extrapolation.
If the Shanks transformation converges to the exact limit (such as if the sequence is
a geometric series), then a division by zero occurs. By default, shanks() (page 992)
handles this case by terminating the iteration and returning the table it has generated so
far. With randomized=True, it will instead replace the zero by a pseudorandom number
close to zero. (TODO: nd a better solution to this problem.)
Examples
We illustrate by applying Shanks transformation to the Leibniz series for :
>>> from sympy.mpmath import *
>>> mp.dps = 50
>>> S = [4*sum(mpf(-1)**n/(2*n+1) for n in range(m))
...
for m in range(1,30)]
>>>
>>> T = shanks(S[:7])
>>> for row in T:
...
nprint(row)
...
[-0.75]
[1.25, 3.16667]
[-1.75, 3.13333, -28.75]
[2.25, 3.14524, 82.25, 3.14234]
[-2.75, 3.13968, -177.75, 3.14139, -969.937]
[3.25, 3.14271, 327.25, 3.14166, 3515.06, 3.14161]
The extrapolated accuracy is about 4 digits, and about 4 digits may have been lost due
to cancellation:
>>> L = T[-1]
>>> nprint([abs(L[-1] - pi), abs(L[-1] - L[-3]), abs(L[-2])])
[2.22532e-5, 4.78309e-5, 3515.06]
5.15. Welcome to mpmaths documentation!
993
SymPy Documentation, Release 0.7.6
Now we extend the computation:
>>> T = shanks(S[:25], T)
>>> L = T[-1]
>>> nprint([abs(L[-1] - pi), abs(L[-1] - L[-3]), abs(L[-2])])
[3.75527e-19, 1.48478e-19, 2.96014e+17]
The value for pi is now accurate to 18 digits. About 18 digits may also have been lost to
cancellation.
Here is an example with a geometric series, where the convergence is immediate (the
sum is exactly 1):
>>> mp.dps = 15
>>> for row in shanks([0.5, 0.75, 0.875, 0.9375, 0.96875]):
...
nprint(row)
[4.0]
[8.0, 1.0]
References
1.[GravesMorris] (page 1910)
2.[BenderOrszag] (page 1910) pp. 368-375
levin()
mpmath.levin(ctx, method=levin, variant=u)
This interface implements Levins (nonlinear) sequence transformation for convergence acceleration and summation of divergent series. It performs better than the
Shanks/Wynn-epsilon algorithm for logarithmic convergent or alternating divergent series.
Let A be the series we want to sum:
A=
ak
k=0
Attention: all ak must be non-zero!
Let sn be the partial sums of this series:
sn =
ak .
k=0
Methods
Calling levin returns an object with the following methods.
update(...) works with the list of individual terms ak of A, and update step(...)
works with the list of partial sums sk of A:
v, e = ...update([a_0, a_1,..., a_k])
v, e = ...update_psum([s_0, s_1,..., s_k])
step(...) works with the individual terms ak and step psum(...) works with the partial sums sk :
v, e = ...step(a_k)
v, e = ...step_psum(s_k)
994
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
v is the current estimate for A, and e is an error estimate which is simply the dierence between the current estimate and the last estimate. One should not mix update,
update psum, step and step psum.
A word of caution
One can only hope for good results (i.e. convergence acceleration or resummation) if the
sn have some well dend asymptotic behavior for large n and are not erratic or random.
Furthermore one usually needs very high working precision because of the numerical
cancellation. If the working precision is insucient, levin may produce silently numerical garbage. Furthermore even if the Levin-transformation converges, in the general
case there is no proof that the result is mathematically sound. Only for very special
classes of problems one can prove that the Levin-transformation converges to the expected result (for example Stieltjes-type integrals). Furthermore the Levin-transform is
quite expensive (i.e. slow) in comparison to Shanks/Wynn-epsilon, Richardson & co. In
summary one can say that the Levin-transformation is powerful but unreliable and that
it may need a copious amount of working precision.
The Levin transform has several variants diering in the choice of weights. Some variants are better suited for the possible avours of convergence behaviour of A than other
variants:
convergence behaviour
logarithmic
linear
alternating divergent
levin-u
+
+
+
levin-t
+
+
levin-v
+
+
+
shanks/wynn-epsilon
+
+
+ means the variant is suitable,- means the variant is not suitable;
for comparison the Shanks/Wynn-epsilon transform is listed, too.
The variant is controlled though the variant keyword (i.e. variant=u, variant=t or
variant=v). Overall u is probably the best choice.
Finally it is possible to use the Sidi-S transform instead of the Levin transform by using
the keyword method=sidi. The Sidi-S transform works better than the Levin transformation for some divergent series (see the examples).
Parameters:
method
variant
levin or sidi chooses either the Levin or the Sidi-S transformation
u,t or v chooses the weight variant.
The Levin transform is also accessible through the nsum interface. method=l or
method=levin select the normal Levin transform while method=sidi selects the
Sidi-S transform. The variant is in both cases selected through the levin variant keyword. The stepsize in nsum() (page 979) must not be chosen too large, otherwise it
will miss the point where the Levin transform converges resulting in numerical overow/garbage. For highly divergent series a copious amount of working precision must
be chosen.
Examples
First we sum the zeta function:
>>>
>>>
>>>
>>>
...
from sympy.mpmath import mp
mp.prec = 53
eps = mp.mpf(mp.eps)
with mp.extraprec(2 * mp.prec): # levin needs a high working precision
L = mp.levin(method = levin, variant = u)
5.15. Welcome to mpmaths documentation!
995
SymPy Documentation, Release 0.7.6
...
S, s, n = [], 0, 1
...
while 1:
...
s += mp.one / (n * n)
...
n += 1
...
S.append(s)
...
v, e = L.update_psum(S)
...
if e < eps:
...
break
...
if n > 1000: raise RuntimeError(iteration limit exceeded)
>>> print(mp.chop(v - mp.pi ** 2 / 6))
0.0
>>> w = mp.nsum(lambda n: 1 / (n*n), [1, mp.inf], method = levin, levin_variant = u)
>>> print(mp.chop(v - w))
0.0
Now we sum the zeta function outside its range of convergence (attention: This does not
work at the negative integers!):
>>>
>>>
...
...
...
...
...
...
...
...
...
...
>>>
0.0
>>>
>>>
0.0
eps = mp.mpf(mp.eps)
with mp.extraprec(2 * mp.prec): # levin needs a high working precision
L = mp.levin(method = levin, variant = v)
A, n = [], 1
while 1:
s = mp.mpf(n) ** (2 + 3j)
n += 1
A.append(s)
v, e = L.update(A)
if e < eps:
break
if n > 1000: raise RuntimeError(iteration limit exceeded)
print(mp.chop(v - mp.zeta(-2-3j)))
w = mp.nsum(lambda n: n ** (2 + 3j), [1, mp.inf], method = levin, levin_variant = v)
print(mp.chop(v - w))
Now we sum the divergent asymptotic expansion of an integral related to the exponential
integral (see also [2] p.373). The Sidi-S transform works best here:
>>>
>>>
>>>
>>>
>>>
...
...
...
...
...
...
...
...
...
>>>
0.0
>>>
>>>
0.0
996
z = mp.mpf(10)
exact = mp.quad(lambda x: mp.exp(-x)/(1+x/z),[0,mp.inf])
# exact = z * mp.exp(z) * mp.expint(1,z) # this is the symbolic expression for the integral
eps = mp.mpf(mp.eps)
with mp.extraprec(2 * mp.prec): # high working precisions are mandatory for divergent resumm
L = mp.levin(method = sidi, variant = t)
n = 0
while 1:
s = (-1)**n * mp.fac(n) * z ** (-n)
v, e = L.step(s)
n += 1
if e < eps:
break
if n > 1000: raise RuntimeError(iteration limit exceeded)
print(mp.chop(v - exact))
w = mp.nsum(lambda n: (-1) ** n * mp.fac(n) * z ** (-n), [0, mp.inf], method = sidi, levin
print(mp.chop(v - w))
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Another highly divergent integral is also summable:
>>>
>>>
>>>
>>>
>>>
...
...
...
...
...
...
...
...
...
>>>
0.0
>>>
...
>>>
0.0
z = mp.mpf(2)
eps = mp.mpf(mp.eps)
exact = mp.quad(lambda x: mp.exp( -x * x / 2 - z * x ** 4), [0,mp.inf]) *
# exact = mp.exp(mp.one / (32 * z)) * mp.besselk(mp.one / 4, mp.one / (32
with mp.extraprec(7 * mp.prec): # we need copious amount of precision to
L = mp.levin(method = levin, variant = t)
n, s = 0, 0
while 1:
s += (-z)**n * mp.fac(4 * n) / (mp.fac(n) * mp.fac(2 * n) * (4 **
n += 1
v, e = L.step_psum(s)
if e < eps:
break
if n > 1000: raise RuntimeError(iteration limit exceeded)
print(mp.chop(v - exact))
2 / mp.sqrt(2 * mp
* z)) / (4 * mp.sq
sum this highly di
n))
w = mp.nsum(lambda n: (-z)**n * mp.fac(4 * n) / (mp.fac(n) * mp.fac(2 * n) * (4 ** n)),
[0, mp.inf], method = levin, levin_variant = t, workprec = 8*mp.prec, steps = [2] + [1
print(mp.chop(v - w))
These examples run with 15-20 decimal digits precision. For higher precision the working precision must be raised.
Examples for nsum
Here we calculate Eulers constant as the constant term in the Laurent expansion of (s)
at s = 1. This sum converges extremly slowly because of the logarithmic convergence
behaviour of the Dirichlet series for zeta:
>>>
>>>
>>>
>>>
0.0
mp.dps = 30
z = mp.mpf(10) ** (-10)
a = mp.nsum(lambda n: n**(-(1+z)), [1, mp.inf], method = l) - 1 / z
print(mp.chop(a - mp.euler, tol = 1e-10))
The Sidi-S transform performs excellently for the alternating series of log(2):
>>> a = mp.nsum(lambda n: (-1)**(n-1) / n, [1, mp.inf], method = sidi)
>>> print(mp.chop(a - mp.log(2)))
0.0
Hypergeometric series can also be summed outside their range of convergence. The
stepsize in nsum() (page 979) must not be chosen too large, otherwise it will miss the
point where the Levin transform converges resulting in numerical overow/garbage:
>>>
>>>
>>>
>>>
>>>
0.0
z = 2 + 1j
exact = mp.hyp2f1(2 / mp.mpf(3), 4 / mp.mpf(3), 1 / mp.mpf(3), z)
f = lambda n: mp.rf(2 / mp.mpf(3), n) * mp.rf(4 / mp.mpf(3), n) * z**n / (mp.rf(1 / mp.mpf(3
v = mp.nsum(f, [0, mp.inf], method = levin, steps = [10 for x in xrange(1000)])
print(mp.chop(exact-v))
References:
[1] E.J. Weniger - Nonlinear Sequence Transformations for the Acceleration of
Convergence and the Summation of Divergent Series arXiv:math/0306302
[2] A. Sidi - Pratical Extrapolation Methods
5.15. Welcome to mpmaths documentation!
997
SymPy Documentation, Release 0.7.6
[3] H.H.H. Homeier
arXiv:math/0005209
Scalar
Levin-Type
Sequence
Transformations
cohen alt()
mpmath.cohen alt(ctx)
This interface implements the convergence acceleration of alternating series as described in H. Cohen, F.R. Villegas, D. Zagier - Convergence Acceleration of Alternating
Series. This series transformation works only well if the individual terms of the series
have an alternating sign. It belongs to the class of linear series transformations (in contrast to the Shanks/Wynn-epsilon or Levin transform). This series transformation is also
able to sum some types of divergent series. See the paper under which conditions this
resummation is mathematical sound.
Let A be the series we want to sum:
A=
ak
k=0
Let sn be the partial sums of this series:
sn =
ak .
k=0
Interface
Calling cohen alt returns an object with the following methods.
Then update(...) works with the list of individual terms ak and update psum(...)
works with the list of partial sums sk :
v, e = ...update([a_0, a_1,..., a_k])
v, e = ...update_psum([s_0, s_1,..., s_k])
v is the current estimate for A, and e is an error estimate which is simply the dierence
between the current estimate and the last estimate.
Examples
Here we compute the alternating zeta function using update psum:
>>>
>>>
>>>
>>>
...
...
...
...
...
...
...
>>>
0.0
from sympy.mpmath import mp
AC = mp.cohen_alt()
S, s, n = [], 0, 1
while 1:
s += -((-1) ** n) * mp.one / (n * n)
n += 1
S.append(s)
v, e = AC.update_psum(S)
if e < mp.eps:
break
if n > 1000: raise RuntimeError(iteration limit exceeded)
print(mp.chop(v - mp.pi ** 2 / 12))
Here we compute the product
n=1
(1 + 1/(2n 1))/(1 + 1/(2n)):
>>> A = []
>>> AC = mp.cohen_alt()
>>> n = 1
998
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> while 1:
...
A.append( mp.loggamma(1 + mp.one / (2 * n - 1)))
...
A.append(-mp.loggamma(1 + mp.one / (2 * n)))
...
n += 1
...
v, e = AC.update(A)
...
if e < mp.eps:
...
break
...
if n > 1000: raise RuntimeError(iteration limit exceeded)
>>> v = mp.exp(v)
>>> print(mp.chop(v - 1.06215090557106, tol = 1e-12))
0.0
cohen alt is also accessible through the nsum() (page 979) interface:
>>>
>>>
0.0
>>>
>>>
0.0
>>>
>>>
0.0
v = mp.nsum(lambda n: (-1)**(n-1) / n, [1, mp.inf], method = a)
print(mp.chop(v - mp.log(2)))
v = mp.nsum(lambda n: (-1)**n / (2 * n + 1), [0, mp.inf], method = a)
print(mp.chop(v - mp.pi / 4))
v = mp.nsum(lambda n: (-1)**n * mp.log(n) * n, [1, mp.inf], method = a)
print(mp.chop(v - mp.diff(lambda s: mp.altzeta(s), -1)))
Dierentiation
Numerical derivatives (diff, diffs)
mpmath.diff(ctx, f, x, n=1, **options)
Numerically computes the derivative of f , f 0 (x), or generally for an integer n 0, the n-th
derivative f (n) (x). A few basic examples are:
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = True
>>> diff(lambda x: x**2 + x, 1.0)
3.0
>>> diff(lambda x: x**2 + x, 1.0, 2)
2.0
>>> diff(lambda x: x**2 + x, 1.0, 3)
0.0
>>> nprint([diff(exp, 3, n) for n in range(5)])
[20.0855, 20.0855, 20.0855, 20.0855, 20.0855]
# exp(x) = exp(x)
Even more generally, given a tuple of arguments (x1 , . . . , xk ) and order (n1 , . . . , nk ), the
partial derivative f (n1 ,...,nk ) (x1 , . . . , xk ) is evaluated. For example:
>>> diff(lambda x,y: 3*x*y + 2*y - x, (0.25, 0.5), (0,1))
2.75
>>> diff(lambda x,y: 3*x*y + 2*y - x, (0.25, 0.5), (1,1))
3.0
Options
The following optional keyword arguments are recognized:
method Supported methods are step or quad: derivatives may be computed using
either a nite dierence with a small step size h (default), or numerical quadrature.
5.15. Welcome to mpmaths documentation!
999
SymPy Documentation, Release 0.7.6
direction Direction of nite dierence: can be -1 for a left dierence, 0 for a central
dierence (default), or +1 for a right dierence; more generally can be any complex
number.
addprec Extra precision for h used to account for the functions sensitivity to perturbations (default = 10).
relative Choose h relative to the magnitude of x, rather than an absolute value; useful
for large or tiny x (default = False).
h As an alternative to addprec and relative, manually select the step size h.
singular If True, evaluation exactly at the point x is avoided; this is useful for dierentiating functions with removable singularities. Default = False.
radius Radius of integration contour (with method = quad). Default = 0.25. A larger
radius typically is faster and more accurate, but it must be chosen so that f has no
singularities within the radius from the evaluation point.
A nite dierence requires n + 1 function evaluations and must be performed at (n + 1)
times the target precision. Accordingly, f must support fast evaluation at high precision.
With integration, a larger number of function evaluations is required, but not much extra
precision is required. For high order derivatives, this method may thus be faster if f is
very expensive to evaluate at high precision.
Further examples
The direction option is useful for computing left- or right-sided derivatives of nonsmooth
functions:
>>> diff(abs, 0, direction=0)
0.0
>>> diff(abs, 0, direction=1)
1.0
>>> diff(abs, 0, direction=-1)
-1.0
More generally, if the direction is nonzero, a right dierence is computed where the step
size is multiplied by sign(direction). For example, with direction=+j, the derivative from
the positive imaginary direction will be computed:
>>> diff(abs, 0, direction=j)
(0.0 - 1.0j)
With integration, the result may have a small imaginary part even even if the result is
purely real:
>>> diff(sqrt, 1, method=quad)
(0.5 - 4.59...e-26j)
>>> chop(_)
0.5
Adding precision to obtain an accurate value:
>>> diff(cos, 1e-30)
0.0
>>> diff(cos, 1e-30, h=0.0001)
-9.99999998328279e-31
>>> diff(cos, 1e-30, addprec=100)
-1.0e-30
1000
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
mpmath.diffs(ctx, f, x, n=None, **options)
Returns a generator that yields the sequence of derivatives
f (x), f 0 (x), f 00 (x), . . . , f (k) (x), . . .
With method=step, diffs() (page 1001) uses only O(k) function evaluations to generate the rst k derivatives, rather than the roughly O(k 2 ) evaluations required if one calls
diff() (page 999) k separate times.
With n < , the generator stops as soon as the n-th derivative has been generated. If the
exact number of needed derivatives is known in advance, this is further slightly more
ecient.
Options are the same as for diff() (page 999).
Examples
>>> from sympy.mpmath import *
>>> mp.dps = 15
>>> nprint(list(diffs(cos, 1, 5)))
[0.540302, -0.841471, -0.540302, 0.841471, 0.540302, -0.841471]
>>> for i, d in zip(range(6), diffs(cos, 1)):
...
print(%s %s % (i, d))
...
0 0.54030230586814
1 -0.841470984807897
2 -0.54030230586814
3 0.841470984807897
4 0.54030230586814
5 -0.841470984807897
Composition of derivatives (diffs prod, diffs exp)
mpmath.diffs prod(ctx, factors)
Given a list of N iterables or generators yielding fk (x), fk0 (x), fk00 (x), . . . for k = 1, . . . , N ,
generate g(x), g 0 (x), g 00 (x), . . . where g(x) = f1 (x)f2 (x) fN (x).
At high precision and for large orders, this is typically more ecient than numerical
dierentiation if the derivatives of each fk (x) admit direct computation.
Note: This function does not increase the working precision internally, so guard digits
may have to be added externally for full accuracy.
Examples
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = True
>>> f = lambda x: exp(x)*cos(x)*sin(x)
>>> u = diffs(f, 1)
>>> v = mp.diffs_prod([diffs(exp,1), diffs(cos,1), diffs(sin,1)])
>>> next(u); next(v)
1.23586333600241
1.23586333600241
>>> next(u); next(v)
0.104658952245596
0.104658952245596
>>> next(u); next(v)
-5.96999877552086
-5.96999877552086
>>> next(u); next(v)
5.15. Welcome to mpmaths documentation!
1001
SymPy Documentation, Release 0.7.6
-12.4632923122697
-12.4632923122697
mpmath.diffs exp(ctx, fdis)
Given an iterable or generator yielding f (x), f 0 (x), f 00 (x), . . . generate g(x), g 0 (x), g 00 (x), . . .
where g(x) = exp(f (x)).
At high precision and for large orders, this is typically more ecient than numerical
dierentiation if the derivatives of f (x) admit direct computation.
Note: This function does not increase the working precision internally, so guard digits
may have to be added externally for full accuracy.
Examples
The derivatives of the gamma function can be computed using logarithmic dierentiation:
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = True
>>>
>>> def diffs_loggamma(x):
...
yield loggamma(x)
...
i = 0
...
while 1:
...
yield psi(i,x)
...
i += 1
...
>>> u = diffs_exp(diffs_loggamma(3))
>>> v = diffs(gamma, 3)
>>> next(u); next(v)
2.0
2.0
>>> next(u); next(v)
1.84556867019693
1.84556867019693
>>> next(u); next(v)
2.49292999190269
2.49292999190269
>>> next(u); next(v)
3.44996501352367
3.44996501352367
Fractional derivatives / dierintegration (differint)
mpmath.differint(ctx, f, x, n=1, x0=0)
Calculates the Riemann-Liouville dierintegral, or fractional derivative, dened by
x
dm
1
n
(x t)mn1 f (t)dt
x0 Dx f (x)
(m n) dxm x0
where f is a given (presumably well-behaved) function, x is the evaluation point, n is the
order, and x0 is the reference point of integration (m is an arbitrary parameter selected
automatically).
With n = 1, this is just the standard derivative f 0 (x); with n = 2,
the second
) derivative
x
x ( t
00
f (x), etc. With n = 1, it gives x0 f (t)dt, with n = 2 it gives x0 x0 f (u)du dt, etc.
As n is permitted to be any number, this operator generalizes iterated dierentiation and
iterated integration to a single operator with a continuous order parameter.
1002
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
There is an exact formula for the fractional derivative of a monomial xp , which may be
used as a reference. For example, the following gives a half-derivative (order 0.5):
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = True
>>> x = mpf(3); p = 2; n = 0.5
>>> differint(lambda t: t**p, x, n)
7.81764019044672
>>> gamma(p+1)/gamma(p-n+1) * x**(p-n)
7.81764019044672
Another useful test function is the exponential function, whose integration / dierentiation formula easy generalizes to arbitrary order. Here we rst compute a third derivative, and then a triply nested integral. (The reference point x0 is set to to avoid
nonzero endpoint terms.):
>>> differint(lambda x: exp(pi*x), -1.5, 3)
0.278538406900792
>>> exp(pi*-1.5) * pi**3
0.278538406900792
>>> differint(lambda x: exp(pi*x), 3.5, -3, -inf)
1922.50563031149
>>> exp(pi*3.5) / pi**3
1922.50563031149
However, for noninteger n, the dierentiation formula for the exponential function must
be modied to give the same result as the Riemann-Liouville dierintegral:
>>> x = mpf(3.5)
>>> c = pi
>>> n = 1+2*j
>>> differint(lambda x: exp(c*x), x, n)
(-123295.005390743 + 140955.117867654j)
>>> x**(-n) * exp(c)**x * (x*c)**n * gammainc(-n, 0, x*c) / gamma(-n)
(-123295.005390743 + 140955.117867654j)
Numerical integration (quadrature)
Standard quadrature (quad)
mpmath.quad(ctx, f, *points, **kwargs)
Computes a single, double or triple integral over a given 1D interval, 2D rectangle, or
3D cuboid. A basic example:
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = True
>>> quad(sin, [0, pi])
2.0
A basic 2D integral:
>>> f = lambda x, y: cos(x+y/2)
>>> quad(f, [-pi/2, pi/2], [0, pi])
4.0
Interval format
5.15. Welcome to mpmaths documentation!
1003
SymPy Documentation, Release 0.7.6
The integration range for each dimension may be specied using a list or tuple. Arguments are interpreted as follows:
x
quad(f, [x1, x2]) calculates x12 f (x) dx
x y
quad(f, [x1, x2], [y1, y2]) calculates x12 y12 f (x, y) dy dx
x y z
quad(f, [x1, x2], [y1, y2], [z1, z2]) calculates x12 y12 z12 f (x, y, z) dz dy dx
Endpoints may be nite or innite. An interval descriptor may also contain more than
two points. In this case, the integration is split into subintervals, between each pair
of consecutive points. This is useful for dealing with mid-interval discontinuities, or
integrating over large intervals where the function is irregular or oscillates.
Options
quad() (page 1003) recognizes the following keyword arguments:
method Chooses integration algorithm (described below).
error If set to true, quad() (page 1003) returns (v, e) where v is the integral and e is the
estimated error.
maxdegree Maximum degree of the quadrature rule to try before quitting.
verbose Print details about progress.
Algorithms
Mpmath presently implements two integration algorithms: tanh-sinh quadrature
and Gauss-Legendre quadrature.
These can be selected using method=tanhsinh or method=gauss-legendre or by passing the classes method=TanhSinh,
method=GaussLegendre. The functions quadts() and quadgl() are also available as
shortcuts.
Both algorithms have the property that doubling the number of evaluation points roughly
doubles the accuracy, so both are ideal for high precision quadrature (hundreds or thousands of digits).
At high precision, computing the nodes and weights for the integration can be expensive
(more expensive than computing the function values). To make repeated integrations
fast, nodes are automatically cached.
The advantages of the tanh-sinh algorithm are that it tends to handle endpoint singularities well, and that the nodes are cheap to compute on the rst run. For these reasons,
it is used by quad() (page 1003) as the default algorithm.
Gauss-Legendre quadrature often requires fewer function evaluations, and is therefore
often faster for repeated use, but the algorithm does not handle endpoint singularities
as well and the nodes are more expensive to compute. Gauss-Legendre quadrature can
be a better choice if the integrand is smooth and repeated integrations are required (e.g.
for multiple integrals).
See the documentation for TanhSinh and GaussLegendre for additional details.
Examples of 1D integrals
Intervals may be innite or half-innite.
The following two examples evaluate the lim
its
of
the
inverse
tangent
function
(
1/(1
+ x2 ) = tan1 x), and the Gaussian integral
2
exp(x ) dx = :
>>> mp.dps = 15
>>> quad(lambda x: 2/(x**2+1), [0, inf])
3.14159265358979
1004
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> quad(lambda x: exp(-x**2), [-inf, inf])**2
3.14159265358979
Integrals can typically be resolved to high precision. The following computes 50 digits
of by integrating the area of the half-circle dened by x2 + y 2 1, 1 x 1, y 0:
>>> mp.dps = 50
>>> 2*quad(lambda x: sqrt(1-x**2), [-1, 1])
3.1415926535897932384626433832795028841971693993751
One can just as well compute 1000 digits (output truncated):
>>> mp.dps = 1000
>>> 2*quad(lambda x: sqrt(1-x**2), [-1, 1])
3.141592653589793238462643383279502884...216420198
Complex integrals are supported. The following computes a residue at z = 0 by integrating counterclockwise along the diamond-shaped path from 1 to +i to 1 to i to 1:
>>> mp.dps = 15
>>> chop(quad(lambda z: 1/z, [1,j,-1,-j,1]))
(0.0 + 6.28318530717959j)
Examples of 2D and 3D integrals
Here are several nice examples of analytically solvable 2D integrals (taken from MathWorld [1]) that can be evaluated to high precision fairly rapidly by quad() (page 1003):
>>> mp.dps = 30
>>> f = lambda x, y: (x-1)/((1-x*y)*log(x*y))
>>> quad(f, [0, 1], [0, 1])
0.577215664901532860606512090082
>>> +euler
0.577215664901532860606512090082
>>> f = lambda x, y: 1/sqrt(1+x**2+y**2)
>>> quad(f, [-1, 1], [-1, 1])
3.17343648530607134219175646705
>>> 4*log(2+sqrt(3))-2*pi/3
3.17343648530607134219175646705
>>> f = lambda x, y: 1/(1-x**2 * y**2)
>>> quad(f, [0, 1], [0, 1])
1.23370055013616982735431137498
>>> pi**2 / 8
1.23370055013616982735431137498
>>> quad(lambda x, y: 1/(1-x*y), [0, 1], [0, 1])
1.64493406684822643647241516665
>>> pi**2 / 6
1.64493406684822643647241516665
Multiple integrals may be done over innite ranges:
>>> mp.dps = 15
>>> print(quad(lambda x,y: exp(-x-y), [0, inf], [1, inf]))
0.367879441171442
>>> print(1/e)
0.367879441171442
5.15. Welcome to mpmaths documentation!
1005
SymPy Documentation, Release 0.7.6
For nonrectangular areas, one can call quad() (page 1003) recursively. For example, we
can replicate the earlier example of calculating by integrating over the unit-circle, and
actually use double quadrature to actually measure the area circle:
>>> f = lambda x: quad(lambda y: 1, [-sqrt(1-x**2), sqrt(1-x**2)])
>>> quad(f, [-1, 1])
3.14159265358979
Here is a simple triple integral:
>>> mp.dps = 15
>>> f = lambda x,y,z: x*y/(1+z)
>>> quad(f, [0,1], [0,1], [1,2], method=gauss-legendre)
0.101366277027041
>>> (log(3)-log(2))/4
0.101366277027041
Singularities
Both tanh-sinh and Gauss-Legendre quadrature are designed to integrate smooth (innitely dierentiable) functions. Neither algorithm copes well with mid-interval singularities (such as mid-interval discontinuities in f (x) or f 0 (x)). The best solution is to split
the integral into parts:
>>> mp.dps = 15
>>> quad(lambda x: abs(sin(x)), [0, 2*pi])
# Bad
3.99900894176779
>>> quad(lambda x: abs(sin(x)), [0, pi, 2*pi]) # Good
4.0
The tanh-sinh rule often works well for integrands having a singularity at one or both
endpoints:
>>> mp.dps = 15
>>> quad(log, [0, 1], method=tanh-sinh) # Good
-1.0
>>> quad(log, [0, 1], method=gauss-legendre) # Bad
-0.999932197413801
However, the result may still be inaccurate for some functions:
>>> quad(lambda x: 1/sqrt(x), [0, 1], method=tanh-sinh)
1.99999999946942
This problem is not due to the quadrature rule per se, but to numerical amplication of
errors in the nodes. The problem can be circumvented by temporarily increasing the
precision:
>>>
>>>
>>>
>>>
2.0
mp.dps = 30
a = quad(lambda x: 1/sqrt(x), [0, 1], method=tanh-sinh)
mp.dps = 15
+a
Highly variable functions
For functions that are smooth (in the sense of being innitely dierentiable) but contain
sharp mid-interval peaks or many bumps, quad() (page 1003) may fail to provide full
accuracy. For example, with default settings, quad() (page 1003) is able to integrate
sin(x) accurately over an interval of length 100 but not over length 1000:
1006
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> quad(sin, [0, 100]); 1-cos(100)
# Good
0.137681127712316
0.137681127712316
>>> quad(sin, [0, 1000]); 1-cos(1000)
# Bad
-37.8587612408485
0.437620923709297
One solution is to break the integration into 10 intervals of length 100:
>>> quad(sin, linspace(0, 1000, 10))
0.437620923709297
# Good
Another is to increase the degree of the quadrature:
>>> quad(sin, [0, 1000], maxdegree=10)
0.437620923709297
# Also good
Whether splitting the interval or increasing the degree is more ecient diers from case
to case. Another example is the function 1/(1 + x2 ), which has a sharp peak centered
around x = 0:
>>> f = lambda x: 1/(1+x**2)
>>> quad(f, [-100, 100])
# Bad
3.64804647105268
>>> quad(f, [-100, 100], maxdegree=10)
# Good
3.12159332021646
>>> quad(f, [-100, 0, 100])
# Also good
3.12159332021646
References
1.http://mathworld.wolfram.com/DoubleIntegral.html
Oscillatory quadrature (quadosc)
mpmath.quadosc(ctx, f, interval, omega=None, period=None, zeros=None)
Calculates
b
I=
f (x)dx
a
where at least one of a and b is innite and where f (x) = g(x) cos(x + ) for some slowly
decreasing function g(x). With proper input, quadosc() (page 1007) can also handle
oscillatory integrals where the oscillation rate is dierent from a pure sine or cosine
wave.
In the standard case when |a| < , b = , quadosc() (page 1007) works by evaluating
the innite series
x1
xk+1
I=
f (x)dx +
f (x)dx
a
k=1
xk
where xk are consecutive zeros (alternatively some other periodic reference point) of
f (x). Accordingly, quadosc() (page 1007) requires information about the zeros of f (x).
For a periodic function, you can specify the zeros by either providing the angular frequency (omega) or the period 2/. In general, you can specify the n-th zero by providing the zeros arguments. Below is an example of each:
5.15. Welcome to mpmaths documentation!
1007
SymPy Documentation, Release 0.7.6
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = True
>>> f = lambda x: sin(3*x)/(x**2+1)
>>> quadosc(f, [0,inf], omega=3)
0.37833007080198
>>> quadosc(f, [0,inf], period=2*pi/3)
0.37833007080198
>>> quadosc(f, [0,inf], zeros=lambda n: pi*n/3)
0.37833007080198
>>> (ei(3)*exp(-3)-exp(3)*ei(-3))/2 # Computed by Mathematica
0.37833007080198
Note that zeros was specied to multiply n by the half-period, not the full period. In theory, it does not matter whether each partial integral is done over a half period or a full period. However, if done over half-periods, the innite series passed to nsum() (page 979)
becomes an alternating series and this typically makes the extrapolation much more
ecient.
Here is an example of an integration over the entire real line, and a half-innite integration starting at :
>>> quadosc(lambda x: cos(x)/(1+x**2), [-inf, inf], omega=1)
1.15572734979092
>>> pi/e
1.15572734979092
>>> quadosc(lambda x: cos(x)/x**2, [-inf, -1], period=2*pi)
-0.0844109505595739
>>> cos(1)+si(1)-pi/2
-0.0844109505595738
Of course, the integrand may contain a complex exponential just as well as a real sine
or cosine:
>>> quadosc(lambda x: exp(3*j*x)/(1+x**2), [-inf,inf], omega=3)
(0.156410688228254 + 0.0j)
>>> pi/e**3
0.156410688228254
>>> quadosc(lambda x: exp(3*j*x)/(2+x+x**2), [-inf,inf], omega=3)
(0.00317486988463794 - 0.0447701735209082j)
>>> 2*pi/sqrt(7)/exp(3*(j+sqrt(7))/2)
(0.00317486988463794 - 0.0447701735209082j)
Non-periodic functions
If f (x) = g(x)h(x) for some function h(x) that is not strictly periodic, omega or period
might not work, and it might be necessary to use zeros.
A notable exception can be made for Bessel functions which, though not periodic, are
asymptotically periodic in a suciently strong sense that the sum extrapolation will
work out:
>>> quadosc(j0, [0, inf], period=2*pi)
1.0
>>> quadosc(j1, [0, inf], period=2*pi)
1.0
More properly, one should provide the exact Bessel function zeros:
>>> j0zero = lambda n: findroot(j0, pi*(n-0.25))
>>> quadosc(j0, [0, inf], zeros=j0zero)
1008
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
1.0
For an example where zeros becomes necessary, consider the complete Fresnel integrals
cos x2 dx =
sin x2 dx =
.
8
0
0
Although the integrands do not decrease in magnitude as x , the integrals are convergent since the oscillation rate increases (causing consecutive periods to asymptotically cancel out). These integrals are virtually impossible to calculate to any kind of
accuracy using standard quadrature
rules. However, if one provides the correct asymp
totic distribution of zeros (xn n), quadosc() (page 1007) works:
>>> mp.dps = 30
>>> f = lambda x: cos(x**2)
>>> quadosc(f, [0,inf], zeros=lambda n:sqrt(pi*n))
0.626657068657750125603941321203
>>> f = lambda x: sin(x**2)
>>> quadosc(f, [0,inf], zeros=lambda n:sqrt(pi*n))
0.626657068657750125603941321203
>>> sqrt(pi/8)
0.626657068657750125603941321203
(Interestingly, these integrals can still be evaluated if one places some other constant
than in the square root sign.)
In general, if f (x) g(x) cos(h(x)), the zeros follow the inverse-function distribution
h1 (x):
>>> mp.dps = 15
>>> f = lambda x: sin(exp(x))
>>> quadosc(f, [1,inf], zeros=lambda n: log(n))
-0.25024394235267
>>> pi/2-si(e)
-0.250243942352671
Non-alternating functions
If the integrand oscillates around a positive value, without alternating signs, the extrapolation might fail. A simple trick that sometimes works is to multiply or divide the
frequency by 2:
>>> f = lambda x: 1/x**2+sin(x)/x**4
>>> quadosc(f, [1,inf], omega=1) # Bad
1.28642190869861
>>> quadosc(f, [1,inf], omega=0.5) # Perfect
1.28652953559617
>>> 1+(cos(1)+ci(1)+sin(1))/6
1.28652953559617
Fast decay
quadosc() (page 1007) is primarily useful for slowly decaying integrands. If the integrand decreases exponentially or faster, quad() (page 1003) will likely handle it without
trouble (and generally be much faster than quadosc() (page 1007)):
>>> quadosc(lambda x: cos(x)/exp(x), [0, inf], omega=1)
0.5
>>> quad(lambda x: cos(x)/exp(x), [0, inf])
0.5
5.15. Welcome to mpmaths documentation!
1009
SymPy Documentation, Release 0.7.6
Quadrature rules
class mpmath.calculus.quadrature.QuadratureRule(ctx)
Quadrature rules are implemented using this class, in order to simplify the code and
provide a common infrastructure for tasks such as error estimation and node caching.
You can implement a custom quadrature rule by subclassing QuadratureRule and implementing the appropriate methods. The subclass can then be used by quad() (page 1003)
by passing it as the method argument.
QuadratureRule instances are supposed to be singletons. QuadratureRule therefore
implements instance caching in new ().
calc nodes(degree, prec, verbose=False)
Compute nodes for the standard interval [1, 1]. Subclasses should probably implement only this method, and use get nodes() method to retrieve the nodes.
clear()
Delete cached node data.
estimate error(results, prec, epsilon)
Given results from integrations [I1 , I2 , . . . , Ik ] done with a quadrature of rule of degree
1, 2, . . . , k, estimate the error of Ik .
For k = 2, we estimate |I I2 | as |I2 I1 |.
For k > 2, we extrapolate |I Ik | |Ik+1 Ik | from |Ik Ik1 | and |Ik Ik2 | under the assumption that each degree increment roughly doubles the accuracy of
the quadrature rule (this is true for both TanhSinh and GaussLegendre). The extrapolation formula is given by Borwein, Bailey & Girgensohn. Although not very
conservative, this method seems to be very robust in practice.
get nodes(a, b, degree, prec, verbose=False)
Return nodes for given interval, degree and precision. The nodes are retrieved from
a cache if already computed; otherwise they are computed by calling calc nodes()
and are then cached.
Subclasses should probably not implement this method, but just implement
calc nodes() for the actual node computation.
guess degree(prec)
Given a desired precision p in bits, estimate the degree m of the quadrature required
to accomplish full accuracy for typical integrals. By default, quad() (page 1003) will
perform up to m iterations. The value of m should be a slight overestimate, so that
slightly bad integrals can be dealt with automatically using a few extra iterations.
On the other hand, it should not be too big, so quad() (page 1003) can quit within a
reasonable amount of time when it is given an unsolvable integral.
The default formula used by guess degree() is tuned for both TanhSinh and GaussLegendre. The output is roughly as follows:
p
50
100
500
3000
m
6
7
10
12
This formula is based purely on a limited amount of experimentation and will sometimes be wrong.
sum next(f, nodes, degree, prec,
previous, verbose=False)
Evaluates the step sum
wk f (xk ) where the nodes list contains the (wk , xk ) pairs.
1010
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
summation() will supply the list results of values computed by sum next() at previous degrees, in case the quadrature rule is able to reuse them.
summation(f, points, prec, epsilon, max degree, verbose=False)
Main integration function. Computes the 1D integral over the interval specied
by points. For each subinterval, performs quadrature of degree from 1 up to
max degree until estimate error() signals convergence.
summation() transforms each subintegration to the standard interval and then calls
sum next().
transform nodes(nodes, a, b, verbose=False)
Rescale standardized nodes (for [1, 1]) to a general interval [a, b]. For a nite interval,
a simple linear change of variables is used. Otherwise, the following transformations
are used:
1
+ (a 1)
x
1
[, b] : t = (b + 1)
x
x
[, ] : t =
1 x2
[a, ] : t =
Tanh-sinh rule
class mpmath.calculus.quadrature.TanhSinh(ctx)
This class implements tanh-sinh or doubly exponential quadrature. This quadrature rule is based on the Euler-Maclaurin integral formula. By performing a change
of variables involving nested exponentials / hyperbolic functions (hence the name), the
derivatives at the endpoints vanish rapidly. Since the error term in the Euler-Maclaurin
formula depends on the derivatives at the endpoints, a simple step sum becomes extremely accurate. In practice, this means that doubling the number of evaluation points
roughly doubles the number of accurate digits.
Comparison to Gauss-Legendre:
Initial computation of nodes is usually faster
Handles endpoint singularities better
Handles innite integration intervals better
Is slower for smooth integrands once nodes have been computed
The implementation of the tanh-sinh algorithm is based on the description given in Borwein, Bailey & Girgensohn, Experimentation in Mathematics - Computational Paths to
Discovery, A K Peters, 2003, pages 312-313. In the present implementation, a few
improvements have been made:
A more ecient scheme is used to compute nodes (exploiting recurrence for the
exponential function)
The nodes are computed successively instead of all at once
Various documents describing the algorithm are available online, e.g.:
http://crd-legacy.lbl.gov/dhbailey/dhbpapers/dhb-tanh-sinh.pdf
http://users.cs.dal.ca/jborwein/tanh-sinh.pdf
5.15. Welcome to mpmaths documentation!
1011
SymPy Documentation, Release 0.7.6
calc nodes(degree, prec, verbose=False)
The abscissas and weights for tanh-sinh quadrature of degree m are given by
xk = tanh(/2 sinh(tk ))
wk = /2 cosh(tk )/ cosh(/2 sinh(tk ))2
where tk = t0 + hk for a step length h 2m . The list of nodes is actually innite, but
the weights die o so rapidly that only a few are needed.
sum next(f, nodes, degree, prec, previous, verbose=False)
Step sum for tanh-sinh quadrature of degree m. We exploit the fact that half of the
abscissas at degree m are precisely the abscissas from degree m 1. Thus reusing
the result from the previous level allows a 2x speedup.
Gauss-Legendre rule
class mpmath.calculus.quadrature.GaussLegendre(ctx)
This class implements Gauss-Legendre quadrature, which is exceptionally ecient for
polynomials and polynomial-like (i.e. very smooth) integrands.
The abscissas and weights are given by roots and values of Legendre polynomials, which
are the orthogonal polynomials on [1, 1] with respect to the unit weight (see legendre()
(page 869)).
In this implementation, we take the degree m of the quadrature to denote a GaussLegendre rule of degree 3 2m (following Borwein, Bailey & Girgensohn). This way we
get quadratic, rather than linear, convergence as the degree is incremented.
Comparison to tanh-sinh quadrature:
Is faster for smooth integrands once nodes have been computed
Initial computation of nodes is usually slower
Handles endpoint singularities worse
Handles innite integration intervals worse
calc nodes(degree, prec, verbose=False)
Calculates the abscissas and weights for Gauss-Legendre quadrature of degree of
given degree (actually 3 2m ).
Ordinary dierential equations
Solving the ODE initial value problem (odefun)
mpmath.odefun(ctx, F, x0, y0, tol=None, degree=None, method=taylor, verbose=False)
Returns a function y(x) = [y0 (x), y1 (x), . . . , yn (x)] that is a numerical solution of the n + 1dimensional rst-order ordinary dierential equation (ODE) system
y00 (x) = F0 (x, [y0 (x), y1 (x), . . . , yn (x)])
y10 (x) = F1 (x, [y0 (x), y1 (x), . . . , yn (x)])
..
.
yn0 (x) = Fn (x, [y0 (x), y1 (x), . . . , yn (x)])
The derivatives are specied by the vector-valued function F that evaluates [y00 , . . . , yn0 ] =
F (x, [y0 , . . . , yn ]). The initial point x0 is specied by the scalar argument x0, and the initial
value y(x0 ) = [y0 (x0 ), . . . , yn (x0 )] is specied by the vector argument y0.
1012
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
For convenience, if the system is one-dimensional, you may optionally provide just a
scalar value for y0. In this case, F should accept a scalar y argument and return a scalar.
The solution function y will return scalar values instead of length-1 vectors.
Evaluation of the solution function y(x) is permitted for any x x0 .
A high-order ODE can be solved by transforming it into rst-order vector form. This
transformation is described in standard texts on ODEs. Examples will also be given
below.
Options, speed and accuracy
By default, odefun() (page 1012) uses a high-order Taylor series method. For reasonably
well-behaved problems, the solution will be fully accurate to within the working precision. Note that F must be possible to evaluate to very high precision for the generation
of Taylor series to work.
To get a faster but less accurate solution, you can set a large value for tol (which defaults
roughly to eps). If you just want to plot the solution or perform a basic simulation, tol =
0.01 is likely sucient.
The degree argument controls the degree of the solver (with method=taylor, this is
the degree of the Taylor series expansion). A higher degree means that a longer step
can be taken before a new local solution must be generated from F, meaning that fewer
steps are required to get from x0 to a given x1 . On the other hand, a higher degree also
means that each local solution becomes more expensive (i.e., more evaluations of F are
required per step, and at higher precision).
The optimal setting therefore involves a tradeo. Generally, decreasing the degree for
Taylor series is likely to give faster solution at low precision, while increasing is likely to
be better at higher precision.
The function object returned by odefun() (page 1012) caches the solutions at all step
points and uses polynomial interpolation between step points. Therefore, once y(x1 ) has
been evaluated for some x1 , y(x) can be evaluated very quickly for any x0 x x1 . and
continuing the evaluation up to x2 > x1 is also fast.
Examples of rst-order ODEs
We will solve the standard test problem y 0 (x) = y(x), y(0) = 1 which has explicit solution
y(x) = exp(x):
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = True
>>> f = odefun(lambda x, y: y, 0, 1)
>>> for x in [0, 1, 2.5]:
...
print((f(x), exp(x)))
...
(1.0, 1.0)
(2.71828182845905, 2.71828182845905)
(12.1824939607035, 12.1824939607035)
The solution with high precision:
>>> mp.dps = 50
>>> f = odefun(lambda x, y: y, 0, 1)
>>> f(1)
2.7182818284590452353602874713526624977572470937
>>> exp(1)
2.7182818284590452353602874713526624977572470937
5.15. Welcome to mpmaths documentation!
1013
SymPy Documentation, Release 0.7.6
Using the more general vectorized form, the test problem can be input as (note that f
returns a 1-element vector):
>>> mp.dps = 15
>>> f = odefun(lambda x, y: [y[0]], 0, [1])
>>> f(1)
[2.71828182845905]
odefun() (page 1012) can solve nonlinear ODEs, which are generally impossible (and
at best dicult) to solve analytically. As an example of a nonlinear ODE, we will solve
y 0 (x) = x sin(y(x)) for y(0) = /2.
( An(exact
)) solution happens to be known for this problem,
and is given by y(x) = 2 tan1 exp x2 /2 :
>>> f = odefun(lambda x, y: x*sin(y), 0, pi/2)
>>> for x in [2, 5, 10]:
...
print((f(x), 2*atan(exp(mpf(x)**2/2))))
...
(2.87255666284091, 2.87255666284091)
(3.14158520028345, 3.14158520028345)
(3.14159265358979, 3.14159265358979)
If F is independent of y, an ODE can be solved using direct integration. We can therefore
obtain a reference solution with quad() (page 1003):
>>> f = lambda x: (1+x**2)/(1+x**3)
>>> g = odefun(lambda x, y: f(x), pi, 0)
>>> g(2*pi)
0.72128263801696
>>> quad(f, [pi, 2*pi])
0.72128263801696
Examples of second-order ODEs
We will solve the harmonic oscillator equation y 00 (x)+y(x) = 0. To do this, we introduce the
helper functions y0 = y, y1 = y00 whereby the original equation can be written as y10 +y00 = 0.
Put together, we get the rst-order, two-dimensional vector ODE
{
y00 = y1
y10 = y0
To get a well-dened IVP, we need two initial values. With y(0) = y0 (0) = 1 and y 0 (0) =
y1 (0) = 0, the problem will of course be solved by y(x) = y0 (x) = cos(x) and y 0 (x) = y1 (x) =
sin(x). We check this:
>>> f = odefun(lambda x, y: [-y[1], y[0]], 0, [1, 0])
>>> for x in [0, 1, 2.5, 10]:
...
nprint(f(x), 15)
...
nprint([cos(x), sin(x)], 15)
...
print(---)
...
[1.0, 0.0]
[1.0, 0.0]
--[0.54030230586814, 0.841470984807897]
[0.54030230586814, 0.841470984807897]
--[-0.801143615546934, 0.598472144103957]
[-0.801143615546934, 0.598472144103957]
---
1014
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
[-0.839071529076452, -0.54402111088937]
[-0.839071529076452, -0.54402111088937]
---
Note that we get both the sine and the cosine solutions simultaneously.
TODO
Better automatic choice of degree and step size
Make determination of Taylor series convergence radius more robust
Allow solution for x < x0
Allow solution for complex x
Test for dicult (ill-conditioned) problems
Implement Runge-Kutta and other algorithms
Function approximation
Taylor series (taylor)
mpmath.taylor(ctx, f, x, n, **options)
Produces a degree-n Taylor polynomial around the point x of the given function f . The
coecients are returned as a list.
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = True
>>> nprint(chop(taylor(sin, 0, 5)))
[0.0, 1.0, 0.0, -0.166667, 0.0, 0.00833333]
The coecients are computed using high-order numerical dierentiation. The function
must be possible to evaluate to arbitrary precision. See diff() (page 999) for additional
details and supported keyword options.
Note that to evaluate the Taylor polynomial as an approximation of f , e.g. with polyval()
(page 971), the coecients must be reversed, and the point of the Taylor expansion must
be subtracted from the argument:
>>> p = taylor(exp, 2.0, 10)
>>> polyval(p[::-1], 2.5 - 2.0)
12.1824939606092
>>> exp(2.5)
12.1824939607035
Pade approximation (pade)
mpmath.pade(ctx, a, L, M)
Computes a Pade approximation of degree (L, M ) to a function. Given at least L + M + 1
Taylor coecients a approximating a function A(x), pade() (page 1015) returns coe-
5.15. Welcome to mpmaths documentation!
1015
SymPy Documentation, Release 0.7.6
cients of polynomials P, Q satisfying
P =
pk xk
k=0
Q=
qk xk
k=0
Q0 = 1
A(x)Q(x) = P (x) + O(xL+M +1 )
P (x)/Q(x) can provide a good approximation to an analytic function beyond the radius of
convergence of its Taylor series (example from G.A. Baker Essentials of Pade Approximants Academic Press, Ch.1A):
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = True
>>> one = mpf(1)
>>> def f(x):
...
return sqrt((one + 2*x)/(one + x))
...
>>> a = taylor(f, 0, 6)
>>> p, q = pade(a, 3, 3)
>>> x = 10
>>> polyval(p[::-1], x)/polyval(q[::-1], x)
1.38169105566806
>>> f(x)
1.38169855941551
Chebyshev approximation (chebyfit)
mpmath.chebyfit(ctx, f, interval, N, error=False)
Computes a polynomial of degree N 1 that approximates the given function f on the interval [a, b]. With error=True, chebyfit() (page 1016) also returns an accurate estimate
of the maximum absolute error; that is, the maximum value of |f (x) P (x)| for x [a, b].
chebyfit() (page 1016) uses the Chebyshev approximation formula, which gives a
nearly optimal solution: that is, the maximum error of the approximating polynomial
is very close to the smallest possible for any polynomial of the same degree.
Chebyshev approximation is very useful if one needs repeated evaluation of an expensive
function, such as function dened implicitly by an integral or a dierential equation. (For
example, it could be used to turn a slow mpmath function into a fast machine-precision
version of the same.)
Examples
Here we use chebyfit() (page 1016) to generate a low-degree approximation of f (x) =
cos(x), valid on the interval [1, 2]:
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = True
>>> poly, err = chebyfit(cos, [1, 2], 5, error=True)
>>> nprint(poly)
[0.00291682, 0.146166, -0.732491, 0.174141, 0.949553]
>>> nprint(err, 12)
1.61351758081e-5
The polynomial can be evaluated using polyval:
1016
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> nprint(polyval(poly, 1.6), 12)
-0.0291858904138
>>> nprint(cos(1.6), 12)
-0.0291995223013
Sampling the true error at 1000 points shows that the error estimate generated by
chebyfit is remarkably good:
>>> error = lambda x: abs(cos(x) - polyval(poly, x))
>>> nprint(max([error(1+n/1000.) for n in range(1000)]), 12)
1.61349954245e-5
Choice of degree
The degree N can be set arbitrarily high, to obtain an arbitrarily good approximation. As
a rule of thumb, an N -term Chebyshev approximation is good to N /(b a) decimal places
on a unit interval (although this depends on how well-behaved f is). The cost grows
accordingly: chebyfit evaluates the function (N 2 )/2 times to compute the coecients
and an additional N times to estimate the error.
Possible issues
One should be careful to use a suciently high working precision both when calling
chebyfit and when evaluating the resulting polynomial, as the polynomial is sometimes
ill-conditioned. It is for example dicult to reach 15-digit accuracy when evaluating
the polynomial using machine precision oats, no matter the theoretical accuracy of the
polynomial. (The option to return the coecients in Chebyshev form should be made
available in the future.)
It is important to note the Chebyshev approximation works poorly if f is not smooth. A
function containing singularities, rapid oscillation, etc can be approximated more eectively by multiplying it by a weight function that cancels out the nonsmooth features, or
by dividing the interval into several segments.
Fourier series (fourier, fourierval)
mpmath.fourier(ctx, f, interval, N)
Computes the Fourier series of degree N of the given function on the interval [a, b]. More
precisely, fourier() (page 1017) returns two lists (c, s) of coecients (the cosine series
and sine series, respectively), such that
f (x)
ck cos(kmx) + sk sin(kmx)
k=0
where m = 2/(b a).
Note that many texts dene the rst coecient as 2c0 instead of c0 . The easiest way to
evaluate the computed series correctly is to pass it to fourierval() (page 1018).
Examples
The function f (x) = x has a simple Fourier series on the standard interval [, ]. The
cosine coecients are all zero (because the function has odd symmetry), and the sine
coecients are rational numbers:
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = True
>>> c, s = fourier(lambda x: x, [-pi, pi], 5)
>>> nprint(c)
[0.0, 0.0, 0.0, 0.0, 0.0, 0.0]
5.15. Welcome to mpmaths documentation!
1017
SymPy Documentation, Release 0.7.6
>>> nprint(s)
[0.0, 2.0, -1.0, 0.666667, -0.5, 0.4]
This computes a Fourier series of a nonsymmetric function on a nonstandard interval:
>>> I = [-1, 1.5]
>>> f = lambda x: x**2 - 4*x + 1
>>> cs = fourier(f, I, 4)
>>> nprint(cs[0])
[0.583333, 1.12479, -1.27552, 0.904708, -0.441296]
>>> nprint(cs[1])
[0.0, -2.6255, 0.580905, 0.219974, -0.540057]
It is instructive to plot a function along with its truncated Fourier series:
>>> plot([f, lambda x: fourierval(cs, I, x)], I)
Fourier series generally converge slowly (and may not converge pointwise). For example,
if f (x) = cosh(x), a 10-term Fourier series gives an L2 error corresponding to 2-digit
accuracy:
>>> I = [-1, 1]
>>> cs = fourier(cosh, I, 9)
>>> g = lambda x: (cosh(x) - fourierval(cs, I, x))**2
>>> nprint(sqrt(quad(g, I)))
0.00467963
fourier() (page 1017) uses numerical quadrature. For nonsmooth functions, the accuracy (and speed) can be improved by including all singular points in the interval specication:
>>> nprint(fourier(abs, [-1, 1], 0), 10)
([0.5000441648], [0.0])
>>> nprint(fourier(abs, [-1, 0, 1], 0), 10)
([0.5], [0.0])
mpmath.fourierval(ctx, series, interval, x)
Evaluates a Fourier series (in the format computed by by fourier() (page 1017) for the
given interval) at the point x.
The series should be a pair (c, s) where c is the cosine series and s is the sine series. The
two lists need not have the same length.
Matrices
Creating matrices
Basic methods Matrices in mpmath are implemented using dictionaries. Only non-zero
values are stored, so it is cheap to represent sparse matrices.
The most basic way to create one is to use the matrix class directly. You can create an empty
matrix specifying the dimensions:
>>> from mpmath import *
>>> mp.dps = 15; mp.pretty = False
>>> matrix(2)
matrix(
[[0.0, 0.0],
1018
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
[0.0, 0.0]])
>>> matrix(2, 3)
matrix(
[[0.0, 0.0, 0.0],
[0.0, 0.0, 0.0]])
Calling matrix with one dimension will create a square matrix.
To access the dimensions of a matrix, use the rows or cols keyword:
>>> A = matrix(3, 2)
>>> A
matrix(
[[0.0, 0.0],
[0.0, 0.0],
[0.0, 0.0]])
>>> A.rows
3
>>> A.cols
2
You can also change the dimension of an existing matrix. This will set the new elements to 0.
If the new dimension is smaller than before, the concerning elements are discarded:
>>> A.rows = 2
>>> A
matrix(
[[0.0, 0.0],
[0.0, 0.0]])
Internally convert is applied every time an element is set. This is done using the syntax
A[row,column], counting from 0:
>>> A = matrix(2)
>>> A[1,1] = 1 + 1j
>>> print A
[0.0
0.0]
[0.0 (1.0 + 1.0j)]
A more comfortable way to create a matrix lets you use nested lists:
>>> matrix([[1, 2], [3, 4]])
matrix(
[[1.0, 2.0],
[3.0, 4.0]])
Advanced methods
trices:
Convenient functions are available for creating various standard ma-
>>> zeros(2)
matrix(
[[0.0, 0.0],
[0.0, 0.0]])
>>> ones(2)
matrix(
[[1.0, 1.0],
[1.0, 1.0]])
>>> diag([1, 2, 3]) # diagonal matrix
5.15. Welcome to mpmaths documentation!
1019
SymPy Documentation, Release 0.7.6
matrix(
[[1.0, 0.0, 0.0],
[0.0, 2.0, 0.0],
[0.0, 0.0, 3.0]])
>>> eye(2) # identity matrix
matrix(
[[1.0, 0.0],
[0.0, 1.0]])
You can even create random matrices:
>>> randmatrix(2)
matrix(
[[0.53491598236191806, 0.57195669543302752],
[0.85589992269513615, 0.82444367501382143]])
Vectors Vectors may also be represented by the matrix class (with rows = 1 or cols = 1).
For vectors there are some things which make life easier. A column vector can be created
using a at list, a row vectors using an almost at nested list:
>>> matrix([1, 2, 3])
matrix(
[[1.0],
[2.0],
[3.0]])
>>> matrix([[1, 2, 3]])
matrix(
[[1.0, 2.0, 3.0]])
Optionally vectors can be accessed like lists, using only a single index:
>>> x = matrix([1, 2, 3])
>>> x[1]
mpf(2.0)
>>> x[1,0]
mpf(2.0)
Other Like you probably expected, matrices can be printed:
>>> print randmatrix(3)
[ 0.782963853573023 0.802057689719883
[0.0541876859348597 0.708243266653103
[ 0.856151514955773 0.544759264818486
0.427895717335467]
0.615134039977379]
0.686210904770947]
Use nstr or nprint to specify the number of digits to print:
>>> nprint(randmatrix(5), 3)
[2.07e-1 1.66e-1 5.06e-1 1.89e-1
[6.62e-1 6.55e-1 4.47e-1 4.82e-1
[4.33e-1 7.75e-1 6.93e-2 2.86e-1
[1.01e-1 2.53e-1 6.13e-1 3.32e-1
[1.56e-1 7.27e-2 6.05e-1 6.67e-2
8.29e-1]
2.06e-2]
5.71e-1]
2.59e-1]
2.79e-1]
As matrices are mutable, you will need to copy them sometimes:
1020
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> A = matrix(2)
>>> A
matrix(
[[0.0, 0.0],
[0.0, 0.0]])
>>> B = A.copy()
>>> B[0,0] = 1
>>> B
matrix(
[[1.0, 0.0],
[0.0, 0.0]])
>>> A
matrix(
[[0.0, 0.0],
[0.0, 0.0]])
Finally, it is possible to convert a matrix to a nested list. This is very useful, as most Python
libraries involving matrices or arrays (namely NumPy or SymPy) support this format:
>>> B.tolist()
[[mpf(1.0), mpf(0.0)], [mpf(0.0), mpf(0.0)]]
Matrix operations
You can add and substract matrices of compatible dimensions:
>>> A = matrix([[1, 2], [3, 4]])
>>> B = matrix([[-2, 4], [5, 9]])
>>> A + B
matrix(
[[-1.0, 6.0],
[8.0, 13.0]])
>>> A - B
matrix(
[[3.0, -2.0],
[-2.0, -5.0]])
>>> A + ones(3)
Traceback (most recent call last):
File <stdin>, line 1, in <module>
File ..., line 238, in __add__
raise ValueError(incompatible dimensions for addition)
ValueError: incompatible dimensions for addition
Traceback (most recent call last):
File <stdin>, line 1, in <module>
File ..., line 238, in __add__
raise ValueError(incompatible dimensions for addition)
ValueError: incompatible dimensions for addition
It is possible to multiply or add matrices and scalars. In the latter case the operation will be
done element-wise:
>>> A * 2
matrix(
[[2.0, 4.0],
[6.0, 8.0]])
>>> A / 4
matrix(
5.15. Welcome to mpmaths documentation!
1021
SymPy Documentation, Release 0.7.6
[[0.25, 0.5],
[0.75, 1.0]])
>>> A - 1
matrix(
[[0.0, 1.0],
[2.0, 3.0]])
Of course you can perform matrix multiplication, if the dimensions are compatible:
>>> A * B
matrix(
[[8.0, 22.0],
[14.0, 48.0]])
>>> matrix([[1, 2, 3]]) * matrix([[-6], [7], [-2]])
matrix(
[[2.0]])
You can raise powers of square matrices:
>>> A**2
matrix(
[[7.0, 10.0],
[15.0, 22.0]])
Negative powers will calculate the inverse:
>>> A**-1
matrix(
[[-2.0, 1.0],
[1.5, -0.5]])
>>> nprint(A * A**-1, 3)
[
1.0 1.08e-19]
[-2.17e-19
1.0]
Matrix transposition is straightforward:
>>> A = ones(2, 3)
>>> A
matrix(
[[1.0, 1.0, 1.0],
[1.0, 1.0, 1.0]])
>>> A.T
matrix(
[[1.0, 1.0],
[1.0, 1.0],
[1.0, 1.0]])
Norms Sometimes you need to know how large a matrix or vector is. Due to their multidimensional nature its not possible to compare them, but there are several functions to map
a matrix or a vector to a positive real number, the so called norms.
mpmath.norm(ctx, x, p=2)
1/p
Gives the entrywise p-norm of an iterable x, i.e. the vector norm ( k |xk |p ) , for any
given 1 p .
Special cases:
If x is not iterable, this just returns absmax(x).
p=1 gives the sum of absolute values.
1022
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
p=2 is the standard Euclidean vector norm.
p=inf gives the magnitude of the largest element.
For x a matrix, p=2 is the Frobenius norm. For operator matrix norms, use mnorm()
(page 1023) instead.
You can use the string inf as well as oat(inf) or mpf(inf) to specify the innity norm.
Examples
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = False
>>> x = matrix([-10, 2, 100])
>>> norm(x, 1)
mpf(112.0)
>>> norm(x, 2)
mpf(100.5186549850325)
>>> norm(x, inf)
mpf(100.0)
mpmath.mnorm(ctx, A, p=1)
Gives the matrix (operator) p-norm of A. Currently p=1 and p=inf are supported:
p=1 gives the 1-norm (maximal column sum)
p=inf gives the -norm (maximal row sum). You can use the string inf as well as
oat(inf) or mpf(inf)
p=2 (not implemented) for a square matrix is the usual spectral matrix norm, i.e. the
largest singular value.
p=f (or F, fro, Frobenius, frobenius) gives the Frobenius norm, which is the elementwise 2-norm. The Frobenius norm is an approximation of the spectral norm and
satises
1
kAkF kAk2 kAkF
rank(A)
The Frobenius norm lacks some mathematical properties that might be expected of a
norm.
For general elementwise p-norms, use norm() (page 1022) instead.
Examples
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = False
>>> A = matrix([[1, -1000], [100, 50]])
>>> mnorm(A, 1)
mpf(1050.0)
>>> mnorm(A, inf)
mpf(1001.0)
>>> mnorm(A, F)
mpf(1006.2310867787777)
Linear algebra
Decompositions
5.15. Welcome to mpmaths documentation!
1023
SymPy Documentation, Release 0.7.6
mpmath.cholesky(ctx, A, tol=None)
Cholesky decomposition of a symmetric positive-denite matrix A. Returns a lower triangular matrix L such that A = L LT . More generally, for a complex Hermitian positivedenite matrix, a Cholesky decomposition satisfying A = L LH is returned.
The Cholesky decomposition can be used to solve linear equation systems twice as eciently as LU decomposition, or to test whether A is positive-denite.
The optional parameter tol determines the tolerance for verifying positive-deniteness.
Examples
Cholesky decomposition of a positive-denite symmetric matrix:
>>> from sympy.mpmath import *
>>> mp.dps = 25; mp.pretty = True
>>> A = eye(3) + hilbert(3)
>>> nprint(A)
[
2.0
0.5 0.333333]
[
0.5 1.33333
0.25]
[0.333333
0.25
1.2]
>>> L = cholesky(A)
>>> nprint(L)
[ 1.41421
0.0
0.0]
[0.353553 1.09924
0.0]
[0.235702 0.15162 1.05899]
>>> chop(A - L*L.T)
[0.0 0.0 0.0]
[0.0 0.0 0.0]
[0.0 0.0 0.0]
Cholesky decomposition of a Hermitian matrix:
>>> A = eye(3) + matrix([[0,0.25j,-0.5j],[-0.25j,0,0],[0.5j,0,0]])
>>> L = cholesky(A)
>>> nprint(L)
[
1.0
0.0
0.0]
[(0.0 - 0.25j) (0.968246 + 0.0j)
0.0]
[ (0.0 + 0.5j) (0.129099 + 0.0j) (0.856349 + 0.0j)]
>>> chop(A - L*L.H)
[0.0 0.0 0.0]
[0.0 0.0 0.0]
[0.0 0.0 0.0]
Attempted Cholesky decomposition of a matrix that is not positive denite:
>>> A = -eye(3) + hilbert(3)
>>> L = cholesky(A)
Traceback (most recent call last):
...
ValueError: matrix is not positive-definite
Traceback (most recent call last):
...
ValueError: matrix is not positive-definite
References
1.[Wikipedia] (page 1910) http://en.wikipedia.org/wiki/Cholesky decomposition
Linear equations
equation system:
1024
Basic linear algebra is implemented; you can for example solve the linear
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
x + 2*y = -10
3*x + 4*y = 10
using lu solve:
>>> A = matrix([[1, 2], [3, 4]])
>>> b = matrix([-10, 10])
>>> x = lu_solve(A, b)
>>> x
matrix(
[[30.0],
[-20.0]])
If you dont trust the result, use residual to calculate the residual ||A*x-b||:
>>> residual(A, x, b)
matrix(
[[3.46944695195361e-18],
[3.46944695195361e-18]])
>>> str(eps)
2.22044604925031e-16
As you can see, the solution is quite accurate. The error is caused by the inaccuracy of the
internal oating point arithmetic. Though, its even smaller than the current machine epsilon,
which basically means you can trust the result.
If you need more speed, use NumPy, or use fp instead mp matrices and methods:
>>> A = fp.matrix([[1, 2], [3, 4]])
>>> b = fp.matrix([-10, 10])
>>> fp.lu_solve(A, b)
matrix(
[[30.0],
[-20.0]])
lu solve accepts overdetermined systems. It is usually not possible to solve such systems,
so the residual is minimized instead. Internally this is done using Cholesky decomposition to
compute a least squares approximation. This means that that lu solve will square the errors.
If you cant aord this, use qr solve instead. It is twice as slow but more accurate, and it
calculates the residual automatically.
Matrix factorization The function lu computes an explicit LU factorization of a matrix:
>>> P, L, U = lu(matrix([[0,2,3],[4,5,6],[7,8,9]]))
>>> print P
[0.0 0.0 1.0]
[1.0 0.0 0.0]
[0.0 1.0 0.0]
>>> print L
[
1.0
0.0 0.0]
[
0.0
1.0 0.0]
[0.571428571428571 0.214285714285714 1.0]
>>> print U
[7.0 8.0
9.0]
[0.0 2.0
3.0]
[0.0 0.0 0.214285714285714]
>>> print P.T*L*U
[0.0 2.0 3.0]
5.15. Welcome to mpmaths documentation!
1025
SymPy Documentation, Release 0.7.6
[4.0
[7.0
5.0
8.0
6.0]
9.0]
The function qr computes a QR factorization of a matrix:
>>> A = matrix([[1, 2], [3, 4], [1, 1]])
>>> Q, R = qr(A)
>>> print Q
[-0.301511344577764
0.861640436855329
0.408248290463863]
[-0.904534033733291 -0.123091490979333 -0.408248290463863]
[-0.301511344577764 -0.492365963917331
0.816496580927726]
>>> print R
[-3.3166247903554 -4.52267016866645]
[
0.0 0.738548945875996]
[
0.0
0.0]
>>> print Q * R
[1.0 2.0]
[3.0 4.0]
[1.0 1.0]
>>> print chop(Q.T * Q)
[1.0 0.0 0.0]
[0.0 1.0 0.0]
[0.0 0.0 1.0]
The singular value decomposition The routines svd r and svd c compute the singular
value decomposition of a real or complex matrix A. svd is an unied interface calling either
svd r or svd c depending on whether A is real or complex.
Given A, two orthogonal (A real) or unitary (A complex) matrices U and V are calculated such
that
A = U SV,
U 0 U = 1,
VV0 =1
where S is a suitable shaped matrix whose o-diagonal elements are zero. Here denotes the
hermitian transpose (i.e. transposition and complex conjugation). The diagonal elements of
S are the singular values of A, i.e. the square roots of the eigenvalues of A0 A or AA0 .
Examples:
>>> from mpmath import mp
>>> A = mp.matrix([[2, -2, -1], [3, 4, -2], [-2, -2, 0]])
>>> S = mp.svd_r(A, compute_uv = False)
>>> print S
[6.0]
[3.0]
[1.0]
>>> U, S, V = mp.svd_r(A)
>>> print mp.chop(A - U * mp.diag(S) * V)
[0.0 0.0 0.0]
[0.0 0.0 0.0]
[0.0 0.0 0.0]
The Schur decomposition This routine computes the Schur decomposition of a square
matrix A. Given A, a unitary matrix Q is determined such that
Q0 AQ = R,
1026
Q0 Q = QQ0 = 1
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
where R is an upper right triangular matrix. Here denotes the hermitian transpose (i.e.
transposition and conjugation).
Examples:
>>> from mpmath import mp
>>> A = mp.matrix([[3, -1, 2], [2, 5, -5], [-2, -3, 7]])
>>> Q, R = mp.schur(A)
>>> mp.nprint(R, 3)
[2.0 0.417 -2.53]
[0.0
4.0 -4.74]
[0.0
0.0
9.0]
>>> print(mp.chop(A - Q * R * Q.transpose_conj()))
[0.0 0.0 0.0]
[0.0 0.0 0.0]
[0.0 0.0 0.0]
The eigenvalue problem The routine eig solves the (ordinary) eigenvalue problem for a
real or complex square matrix A. Given A, a vector E and matrices ER and EL are calculated
such that
A ER[:,i] =
E[i] ER[:,i]
EL[i,:] A
= EL[i,:] E[i]
E contains the eigenvalues of A. The columns of ER contain the right eigenvectors of A
whereas the rows of EL contain the left eigenvectors.
Examples:
>>> from mpmath import mp
>>> A = mp.matrix([[3, -1, 2], [2, 5, -5], [-2, -3, 7]])
>>> E, ER = mp.eig(A)
>>> print(mp.chop(A * ER[:,0] - E[0] * ER[:,0]))
[0.0]
[0.0]
[0.0]
>>> E, EL, ER = mp.eig(A,left = True, right = True)
>>> E, EL, ER = mp.eig_sort(E, EL, ER)
>>> mp.nprint(E)
[2.0, 4.0, 9.0]
>>> print(mp.chop(A * ER[:,0] - E[0] * ER[:,0]))
[0.0]
[0.0]
[0.0]
>>> print(mp.chop( EL[0,:] * A - EL[0,:] * E[0]))
[0.0 0.0 0.0]
The symmetric eigenvalue problem The routines eigsy and eighe solve the (ordinary)
eigenvalue problem for a real symmetric or complex hermitian square matrix A. eigh is an
unied interface for this two functions calling either eigsy or eighe depending on whether
A is real or complex.
Given A, an orthogonal (A real) or unitary matrix Q (A complex) is calculated which diagonalizes A:
Q0 AQ = diag(E),
QQ0 = Q0 Q = 1
5.15. Welcome to mpmaths documentation!
1027
SymPy Documentation, Release 0.7.6
Here diag(E) a is diagonal matrix whose diagonal is E. denotes the hermitian transpose (i.e.
ordinary transposition and complex conjugation).
The columns of Q are the eigenvectors of A and E contains the eigenvalues:
A Q[:,i] = E[i] Q[:,i]
Examples:
>>> from mpmath import mp
>>> A = mp.matrix([[3, 2], [2, 0]])
>>> E = mp.eigsy(A, eigvals_only = True)
>>> print E
[-1.0]
[ 4.0]
>>> A = mp.matrix([[1, 2], [2, 3]])
>>> E, Q = mp.eigsy(A)
# alternative: E, Q = mp.eigh(A)
>>> print mp.chop(A * Q[:,0] - E[0] * Q[:,0])
[0.0]
[0.0]
>>> A = mp.matrix([[1, 2 + 5j], [2 - 5j, 3]])
>>> E, Q = mp.eighe(A)
# alternative: E, Q = mp.eigh(A)
>>> print mp.chop(A * Q[:,0] - E[0] * Q[:,0])
[0.0]
[0.0]
Interval and double-precision matrices
The iv.matrix and fp.matrix classes convert inputs to intervals and Python oating-point
numbers respectively.
Interval matrices can be used to perform linear algebra operations with rigorous error tracking:
>>> a = iv.matrix([[0.1,0.3,1.0],
...
[7.1,5.5,4.8],
...
[3.2,4.4,5.6]])
>>>
>>> b = iv.matrix([4,0.6,0.5])
>>> c = iv.lu_solve(a, b)
>>> print c
[ [5.2582327113062393041, 5.2582327113062749951]]
[[-13.155049396267856583, -13.155049396267821167]]
[ [7.4206915477497212555, 7.4206915477497310922]]
>>> print a*c
[ [3.9999999999999866773, 4.0000000000000133227]]
[[0.59999999999972430942, 0.60000000000027142733]]
[[0.49999999999982236432, 0.50000000000018474111]]
Matrix functions
mpmath.expm(ctx, A, method=taylor)
Computes the matrix exponential of a square matrix A, which is dened by the power
series
exp(A) = I + A +
1028
A3
A2
+
+ ...
2!
3!
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
With method=taylor, the matrix exponential is computed using the Taylor series. With
method=pade, Pade approximants are used instead.
Examples
Basic examples:
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = True
>>> expm(zeros(3))
[1.0 0.0 0.0]
[0.0 1.0 0.0]
[0.0 0.0 1.0]
>>> expm(eye(3))
[2.71828182845905
0.0
0.0]
[
0.0 2.71828182845905
0.0]
[
0.0
0.0 2.71828182845905]
>>> expm([[1,1,0],[1,0,1],[0,1,0]])
[ 3.86814500615414 2.26812870852145 0.841130841230196]
[ 2.26812870852145 2.44114713886289
1.42699786729125]
[0.841130841230196 1.42699786729125
1.6000162976327]
>>> expm([[1,1,0],[1,0,1],[0,1,0]], method=pade)
[ 3.86814500615414 2.26812870852145 0.841130841230196]
[ 2.26812870852145 2.44114713886289
1.42699786729125]
[0.841130841230196 1.42699786729125
1.6000162976327]
>>> expm([[1+j, 0], [1+j,1]])
[(1.46869393991589 + 2.28735528717884j)
0.0]
[ (1.03776739863568 + 3.536943175722j) (2.71828182845905 + 0.0j)]
Matrices with large entries are allowed:
>>> expm(matrix([[1,2],[2,3]])**25)
[5.65024064048415e+2050488462815550
[9.14228140091932e+2050488462815550
9.14228140091932e+2050488462815550]
1.47925220414035e+2050488462815551]
The identity exp(A + B) = exp(A) exp(B) does not hold for noncommuting matrices:
>>> A = hilbert(3)
>>> B = A + eye(3)
>>> chop(mnorm(A*B - B*A))
0.0
>>> chop(mnorm(expm(A+B) - expm(A)*expm(B)))
0.0
>>> B = A + ones(3)
>>> mnorm(A*B - B*A)
1.8
>>> mnorm(expm(A+B) - expm(A)*expm(B))
42.0927851137247
mpmath.cosm(ctx, A)
Gives the cosine of a square matrix A, dened in analogy with the matrix exponential.
Examples:
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = True
>>> X = eye(3)
>>> cosm(X)
[0.54030230586814
0.0
[
0.0 0.54030230586814
[
0.0
0.0
0.0]
0.0]
0.54030230586814]
5.15. Welcome to mpmaths documentation!
1029
SymPy Documentation, Release 0.7.6
>>> X = hilbert(3)
>>> cosm(X)
[ 0.424403834569555 -0.316643413047167 -0.221474945949293]
[-0.316643413047167
0.820646708837824 -0.127183694770039]
[-0.221474945949293 -0.127183694770039
0.909236687217541]
>>> X = matrix([[1+j,-2],[0,-j]])
>>> cosm(X)
[(0.833730025131149 - 0.988897705762865j) (1.07485840848393 - 0.17192140544213j)]
[
0.0
(1.54308063481524 + 0.0j)]
mpmath.sinm(ctx, A)
Gives the sine of a square matrix A, dened in analogy with the matrix exponential.
Examples:
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = True
>>> X = eye(3)
>>> sinm(X)
[0.841470984807897
0.0
0.0]
[
0.0 0.841470984807897
0.0]
[
0.0
0.0 0.841470984807897]
>>> X = hilbert(3)
>>> sinm(X)
[0.711608512150994 0.339783913247439 0.220742837314741]
[0.339783913247439 0.244113865695532 0.187231271174372]
[0.220742837314741 0.187231271174372 0.155816730769635]
>>> X = matrix([[1+j,-2],[0,-j]])
>>> sinm(X)
[(1.29845758141598 + 0.634963914784736j) (-1.96751511930922 + 0.314700021761367j)]
[
0.0
(0.0 - 1.1752011936438j)]
mpmath.sqrtm(ctx, A, may rotate=2)
Computes a square root of the square matrix A, i.e. returns a matrix B = A1/2 such that
B 2 = A. The square root of a matrix, if it exists, is not unique.
Examples
Square roots of some simple matrices:
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = True
>>> sqrtm([[1,0], [0,1]])
[1.0 0.0]
[0.0 1.0]
>>> sqrtm([[0,0], [0,0]])
[0.0 0.0]
[0.0 0.0]
>>> sqrtm([[2,0],[0,1]])
[1.4142135623731 0.0]
[
0.0 1.0]
>>> sqrtm([[1,1],[1,0]])
[ (0.920442065259926 - 0.21728689675164j)
[(0.568864481005783 + 0.351577584254143j)
>>> sqrtm([[1,0],[0,1]])
[1.0 0.0]
[0.0 1.0]
>>> sqrtm([[-1,0],[0,1]])
[(0.0 - 1.0j)
0.0]
[
0.0 (1.0 + 0.0j)]
1030
(0.568864481005783 + 0.351577584254143j)]
(0.351577584254143 - 0.568864481005783j)]
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> sqrtm([[j,0],[0,j]])
[(0.707106781186547 + 0.707106781186547j)
[
0.0
0.0]
(0.707106781186547 + 0.707106781186547j)]
A square root of a rotation matrix, giving the corresponding half-angle rotation matrix:
>>> t1 = 0.75
>>> t2 = t1 * 0.5
>>> A1 = matrix([[cos(t1), -sin(t1)], [sin(t1), cos(t1)]])
>>> A2 = matrix([[cos(t2), -sin(t2)], [sin(t2), cos(t2)]])
>>> sqrtm(A1)
[0.930507621912314 -0.366272529086048]
[0.366272529086048
0.930507621912314]
>>> A2
[0.930507621912314 -0.366272529086048]
[0.366272529086048
0.930507621912314]
The identity (A2 )1/2 = A does not necessarily hold:
>>> A = matrix([[4,1,4],[7,8,9],[10,2,11]])
>>> sqrtm(A**2)
[ 4.0 1.0
4.0]
[ 7.0 8.0
9.0]
[10.0 2.0 11.0]
>>> sqrtm(A)**2
[ 4.0 1.0
4.0]
[ 7.0 8.0
9.0]
[10.0 2.0 11.0]
>>> A = matrix([[-4,1,4],[7,-8,9],[10,2,11]])
>>> sqrtm(A**2)
[ 7.43715112194995 -0.324127569985474
1.8481718827526]
[-0.251549715716942
9.32699765900402 2.48221180985147]
[ 4.11609388833616
0.775751877098258
13.017955697342]
>>> chop(sqrtm(A)**2)
[-4.0
1.0
4.0]
[ 7.0 -8.0
9.0]
[10.0
2.0 11.0]
For some matrices, a square root does not exist:
>>> sqrtm([[0,1], [0,0]])
Traceback (most recent call last):
...
ZeroDivisionError: matrix is numerically singular
Traceback (most recent call last):
...
ZeroDivisionError: matrix is numerically singular
Two examples from the documentation for Matlabs sqrtm:
>>> mp.dps = 15; mp.pretty = True
>>> sqrtm([[7,10],[15,22]])
[1.56669890360128 1.74077655955698]
[2.61116483933547 4.17786374293675]
>>>
>>> X = matrix(\
...
[[5,-4,1,0,0],
...
[-4,6,-4,1,0],
...
[1,-4,6,-4,1],
5.15. Welcome to mpmaths documentation!
1031
SymPy Documentation, Release 0.7.6
...
[0,1,-4,6,-4],
...
[0,0,1,-4,5]])
>>> Y = matrix(\
...
[[2,-1,-0,-0,-0],
...
[-1,2,-1,0,-0],
...
[0,-1,2,-1,0],
...
[-0,0,-1,2,-1],
...
[-0,-0,-0,-1,2]])
>>> mnorm(sqrtm(X) - Y)
4.53155328326114e-19
mpmath.logm(ctx, A)
Computes a logarithm of the square matrix A, i.e. returns a matrix B = log(A) such that
exp(B) = A. The logarithm of a matrix, if it exists, is not unique.
Examples
Logarithms of some simple matrices:
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = True
>>> X = eye(3)
>>> logm(X)
[0.0 0.0 0.0]
[0.0 0.0 0.0]
[0.0 0.0 0.0]
>>> logm(2*X)
[0.693147180559945
0.0
[
0.0 0.693147180559945
[
0.0
0.0
>>> logm(expm(X))
[1.0 0.0 0.0]
[0.0 1.0 0.0]
[0.0 0.0 1.0]
0.0]
0.0]
0.693147180559945]
A logarithm of a complex matrix:
>>> X = matrix([[2+j, 1, 3], [1-j, 1-2*j, 1], [-4, -5, j]])
>>> B = logm(X)
>>> nprint(B)
[ (0.808757 + 0.107759j)
(2.20752 + 0.202762j)
(1.07376 - 0.773874j)]
[ (0.905709 - 0.107795j) (0.0287395 - 0.824993j) (0.111619 + 0.514272j)]
[(-0.930151 + 0.399512j)
(-2.06266 - 0.674397j) (0.791552 + 0.519839j)]
>>> chop(expm(B))
[(2.0 + 1.0j)
1.0
3.0]
[(1.0 - 1.0j) (1.0 - 2.0j)
1.0]
[
-4.0
-5.0 (0.0 + 1.0j)]
A matrix X close to the identity matrix, for which log(exp(X)) = exp(log(X)) = X holds:
>>> X = eye(3) + hilbert(3)/4
>>> X
[
1.25
0.125
[
0.125 1.08333333333333
[0.0833333333333333
0.0625
>>> logm(expm(X))
[
1.25
0.125
[
0.125 1.08333333333333
[0.0833333333333333
0.0625
>>> expm(logm(X))
1032
0.0833333333333333]
0.0625]
1.05]
0.0833333333333333]
0.0625]
1.05]
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
[
1.25
[
0.125
[0.0833333333333333
0.125
1.08333333333333
0.0625
0.0833333333333333]
0.0625]
1.05]
A logarithm of a rotation matrix, giving back the angle of the rotation:
>>> t = 3.7
>>> A = matrix([[cos(t),sin(t)],[-sin(t),cos(t)]])
>>> chop(logm(A))
[
0.0 -2.58318530717959]
[2.58318530717959
0.0]
>>> (2*pi-t)
2.58318530717959
For some matrices, a logarithm does not exist:
>>> logm([[1,0], [0,0]])
Traceback (most recent call last):
...
ZeroDivisionError: matrix is numerically singular
Traceback (most recent call last):
...
ZeroDivisionError: matrix is numerically singular
Logarithm of a matrix with large entries:
>>> logm(hilbert(3)
[ 45.5597513593433
[ 1.27721006042799
[0.317662687717978
* 10**20).apply(re)
1.27721006042799 0.317662687717978]
42.5222778973542
2.24003708791604]
2.24003708791604
42.395212822267]
mpmath.powm(ctx, A, r)
Computes Ar = exp(A log r) for a matrix A and complex number r.
Examples
Powers and inverse powers of a matrix:
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = True
>>> A = matrix([[4,1,4],[7,8,9],[10,2,11]])
>>> powm(A, 2)
[ 63.0 20.0
69.0]
[174.0 89.0 199.0]
[164.0 48.0 179.0]
>>> chop(powm(powm(A, 4), 1/4.))
[ 4.0 1.0
4.0]
[ 7.0 8.0
9.0]
[10.0 2.0 11.0]
>>> powm(extraprec(20)(powm)(A, -4), -1/4.)
[ 4.0 1.0
4.0]
[ 7.0 8.0
9.0]
[10.0 2.0 11.0]
>>> chop(powm(powm(A, 1+0.5j), 1/(1+0.5j)))
[ 4.0 1.0
4.0]
[ 7.0 8.0
9.0]
[10.0 2.0 11.0]
>>> powm(extraprec(5)(powm)(A, -1.5), -1/(1.5))
[ 4.0 1.0
4.0]
5.15. Welcome to mpmaths documentation!
1033
SymPy Documentation, Release 0.7.6
[ 7.0
[10.0
8.0
2.0
9.0]
11.0]
A Fibonacci-generating matrix:
>>> powm([[1,1],[1,0]], 10)
[89.0 55.0]
[55.0 34.0]
>>> fib(10)
55.0
>>> powm([[1,1],[1,0]], 6.5)
[(16.5166626964253 - 0.0121089837381789j)
[(10.2078589271083 + 0.0195927472575932j)
>>> (phi**6.5 - (1-phi)**6.5)/sqrt(5)
(10.2078589271083 - 0.0195927472575932j)
>>> powm([[1,1],[1,0]], 6.2)
[ (14.3076953002666 - 0.008222855781077j)
[(8.81733464837593 + 0.0133048601383712j)
>>> (phi**6.2 - (1-phi)**6.2)/sqrt(5)
(8.81733464837593 - 0.0133048601383712j)
(10.2078589271083 + 0.0195927472575932j)]
(6.30880376931698 - 0.0317017309957721j)]
(8.81733464837593 + 0.0133048601383712j)]
(5.49036065189071 - 0.0215277159194482j)]
Number identication
Most function in mpmath are concerned with producing approximations from exact mathematical formulas. It is also useful to consider the inverse problem: given only a decimal
approximation for a number, such as 0.7320508075688772935274463, is it possible to nd
an exact formula?
Subject to certain restrictions, such reverse engineering is indeed possible thanks to the
existence of integer relation algorithms. Mpmath implements the PSLQ algorithm (developed
by H. Ferguson), which is one such algorithm.
Automated number recognition based on PSLQ is not a silver bullet. Any occurring transcendental constants (, e, etc) must be guessed by the user, and the relation between those
constants in the formula must be linear (such as x = 3 + 4e). More complex formulas can be
found by combining PSLQ with functional transformations; however, this is only feasible to a
limited extent since the computation time grows exponentially with the number of operations
that need to be combined.
The number identication facilities in mpmath are inspired by the Inverse Symbolic Calculator (ISC). The ISC is more powerful than mpmath, as it uses a lookup table of millions of
precomputed constants (thereby mitigating the problem with exponential complexity).
Constant recognition
identify()
mpmath.identify(ctx, x, constants=[], tol=None, maxcoe=1000, full=False, verbose=False)
Given a real number x, identify(x) attempts to nd an exact formula for x. This formula
is returned as a string. If no match is found, None is returned. With full=True, a list of
matching formulas is returned.
As a simple example, identify() (page 1034) will nd an algebraic formula for the
golden ratio:
1034
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = True
>>> identify(phi)
((1+sqrt(5))/2)
identify() (page 1034) can identify simple algebraic numbers and simple combinations
of given base constants, as well as certain basic transformations thereof. More specically, identify() (page 1034) looks for the following:
1.Fractions
2.Quadratic algebraic numbers
3.Rational linear combinations of the base constants
4.Any of the above after rst transforming x into f (x) where f (x) is 1/x, x, x2 , log x
or exp x, either directly or with x or f (x) multiplied or divided by one of the base
constants
5.Products of fractional powers of the base constants and small integers
Base constants can be given as a list of strings representing mpmath expressions (identify() (page 1034) will eval the strings to numerical values and use the original strings
for the output), or as a dict of formula:value pairs.
In order not to produce spurious results, identify() (page 1034) should be used with
high precision; preferably 50 digits or more.
Examples
Simple identications can be performed safely at standard precision. Here the default
recognition of rational, algebraic, and exp/log of algebraic numbers is demonstrated:
>>> mp.dps = 15
>>> identify(0.22222222222222222)
(2/9)
>>> identify(1.9662210973805663)
sqrt(((24+sqrt(48))/8))
>>> identify(4.1132503787829275)
exp((sqrt(8)/2))
>>> identify(0.881373587019543)
log(((2+sqrt(8))/2))
By default, identify() (page 1034) does not recognize . At standard precision it nds
a not too useful approximation. At slightly increased precision, this approximation is no
longer accurate enough and identify() (page 1034) more correctly returns None:
>>> identify(pi)
(2**(176/117)*3**(20/117)*5**(35/39))/(7**(92/117))
>>> mp.dps = 30
>>> identify(pi)
>>>
Numbers such as , and simple combinations of user-dened constants, can be identied
if they are provided explicitly:
>>> identify(3*pi-2*e, [pi, e])
(3*pi + (-2)*e)
Here is an example using a dict of constants. Note that the constants need not be
atomic; identify() (page 1034) can just as well express the given number in terms
of expressions given by formulas:
5.15. Welcome to mpmaths documentation!
1035
SymPy Documentation, Release 0.7.6
>>> identify(pi+e, {a:pi+2, b:2*e})
((-2) + 1*a + (1/2)*b)
Next, we attempt some identications with a set of base constants. It is necessary to
increase the precision a bit.
>>> mp.dps = 50
>>> base = [sqrt(2),pi,log(2)]
>>> identify(0.25, base)
(1/4)
>>> identify(3*pi + 2*sqrt(2) + 5*log(2)/7, base)
(2*sqrt(2) + 3*pi + (5/7)*log(2))
>>> identify(exp(pi+2), base)
exp((2 + 1*pi))
>>> identify(1/(3+sqrt(2)), base)
((3/7) + (-1/7)*sqrt(2))
>>> identify(sqrt(2)/(3*pi+4), base)
sqrt(2)/(4 + 3*pi)
>>> identify(5**(mpf(1)/3)*pi*log(2)**2, base)
5**(1/3)*pi*log(2)**2
An example of an erroneous solution being found when too low precision is used:
>>> mp.dps = 15
>>> identify(1/(3*pi-4*e+sqrt(8)), [pi, e, sqrt(2)])
((11/25) + (-158/75)*pi + (76/75)*e + (44/15)*sqrt(2))
>>> mp.dps = 50
>>> identify(1/(3*pi-4*e+sqrt(8)), [pi, e, sqrt(2)])
1/(3*pi + (-4)*e + 2*sqrt(2))
Finding approximate solutions
The tolerance tol defaults to 3/4 of the working precision. Lowering the tolerance is useful for nding approximate matches. We can for example try to generate approximations
for pi:
>>> mp.dps = 15
>>> identify(pi, tol=1e-2)
(22/7)
>>> identify(pi, tol=1e-3)
(355/113)
>>> identify(pi, tol=1e-10)
(5**(339/269))/(2**(64/269)*3**(13/269)*7**(92/269))
With full=True, and by supplying a few base constants, identify can generate almost
endless lists of approximations for any number (the output below has been truncated to
show only the rst few):
>>> for p in identify(pi, [e, catalan], tol=1e-5, full=True):
...
print(p)
...
e/log((6 + (-4/3)*e))
(3**3*5*e*catalan**2)/(2*7**2)
sqrt(((-13) + 1*e + 22*catalan))
log(((-6) + 24*e + 4*catalan)/e)
exp(catalan*((-1/5) + (8/15)*e))
catalan*(6 + (-6)*e + 15*catalan)
sqrt((5 + 26*e + (-3)*catalan))/e
e*sqrt(((-27) + 2*e + 25*catalan))
1036
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
log(((-1) + (-11)*e + 59*catalan))
((3/20) + (21/20)*e + (3/20)*catalan)
...
The numerical values are roughly as close to as permitted by the specied tolerance:
>>> e/log(6-4*e/3)
3.14157719846001
>>> 135*e*catalan**2/98
3.14166950419369
>>> sqrt(e-13+22*catalan)
3.14158000062992
>>> log(24*e-6+4*catalan)-1
3.14158791577159
Symbolic processing
The output formula can be evaluated as a Python expression. Note however that if fractions (like 2/3) are present in the formula, Pythons eval() may erroneously perform
integer division. Note also that the output is not necessarily in the algebraically simplest
form:
>>> identify(sqrt(2))
(sqrt(8)/2)
As a solution to both problems, consider using SymPys sympify() to convert the formula
into a symbolic expression. SymPy can be used to pretty-print or further simplify the
formula symbolically:
>>> from sympy import sympify
>>> sympify(identify(sqrt(2)))
2**(1/2)
Sometimes identify() (page 1034) can simplify an expression further than a symbolic
algorithm:
>>> from sympy import simplify
>>> x = sympify(-1/(-3/2+(1/2)*5**(1/2))*(3/2-1/2*5**(1/2))**(1/2))
>>> x
(3/2 - 5**(1/2)/2)**(-1/2)
>>> x = simplify(x)
>>> x
2/(6 - 2*5**(1/2))**(1/2)
>>> mp.dps = 30
>>> x = sympify(identify(x.evalf(30)))
>>> x
1/2 + 5**(1/2)/2
(In fact, this functionality is available directly in SymPy as the function nsimplify(),
which is essentially a wrapper for identify() (page 1034).)
Miscellaneous issues and limitations
The input x must be a real number. All base constants must be positive real numbers
and must not be rationals or rational linear combinations of each other.
The worst-case computation time grows quickly with the number of base constants. Already with 3 or 4 base constants, identify() (page 1034) may require several seconds
to nish. To search for relations among a large number of constants, you should consider
using pslq() (page 1039) directly.
5.15. Welcome to mpmaths documentation!
1037
SymPy Documentation, Release 0.7.6
The extended transformations are applied to x, not the constants separately. As a result,
identify will for example be able to recognize exp(2*pi+3) with pi given as a base
constant, but not 2*exp(pi)+3. It will be able to recognize the latter if exp(pi) is given
explicitly as a base constant.
Algebraic identication
findpoly()
mpmath.findpoly(ctx, x, n=1, **kwargs)
findpoly(x, n) returns the coecients of an integer polynomial P of degree at most
n such that P (x) 0. If no polynomial having x as a root can be found, findpoly()
(page 1038) returns None.
findpoly() (page 1038) works by successively calling pslq() (page 1039) with the vectors [1, x], [1, x, x2 ], [1, x, x2 , x3 ], ..., [1, x, x2 , .., xn ] as input. Keyword arguments given to
findpoly() (page 1038) are forwarded verbatim to pslq() (page 1039). In particular,
you can specify a tolerance for P (x) with tol and a maximum permitted coecient size
with maxcoeff.
For large values of n, it is recommended to run findpoly() (page 1038) at high precision;
preferably 50 digits or more.
Examples
By default (degree n = 1), findpoly() (page 1038) simply nds a linear polynomial with
a rational root:
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = True
>>> findpoly(0.7)
[-10, 7]
The generated coecient list is valid input to polyval and polyroots:
>>> nprint(polyval(findpoly(phi, 2), phi), 1)
-2.0e-16
>>> for r in polyroots(findpoly(phi, 2)):
...
print(r)
...
-0.618033988749895
1.61803398874989
Numbers of the form m + n p for integers (m, n, p) are solutions to quadratic equations.
As we nd here, 1 + 2 is a root of the polynomial x2 2x 1:
>>> findpoly(1+sqrt(2), 2)
[1, -2, -1]
>>> findroot(lambda x: x**2 - 2*x - 1, 1)
2.4142135623731
Despite only containing square roots, the following number results in a polynomial of
degree 4:
>>> findpoly(sqrt(2)+sqrt(3), 4)
[1, 0, -10, 0, 1]
In fact, x4 10x2 + 1 is the minimal polynomial of r = 2 + 3, meaning that a rational
polynomial of lower degree having r as a root does not exist. Given sucient preci-
1038
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
sion, findpoly() (page 1038) will usually nd the correct minimal polynomial of a given
algebraic number.
Non-algebraic numbers
If findpoly() (page 1038) fails to nd a polynomial with given coecient size and tolerance constraints, that means no such polynomial exists.
We can verify that is not an algebraic number of degree 3 with coecients less than
1000:
>>> mp.dps = 15
>>> findpoly(pi, 3)
>>>
It is always possible to nd an algebraic approximation of a number using one (or several)
of the following methods:
1.Increasing the permitted degree
2.Allowing larger coecients
3.Reducing the tolerance
One example of each method is shown below:
>>> mp.dps = 15
>>> findpoly(pi, 4)
[95, -545, 863, -183, -298]
>>> findpoly(pi, 3, maxcoeff=10000)
[836, -1734, -2658, -457]
>>> findpoly(pi, 3, tol=1e-7)
[-4, 22, -29, -2]
It is unknown whether Eulers constant is transcendental (or even irrational). We can use
findpoly() (page 1038) to check that if is an algebraic number, its minimal polynomial
must have degree at least 7 and a coecient of magnitude at least 1000000:
>>> mp.dps = 200
>>> findpoly(euler, 6, maxcoeff=10**6, tol=1e-100, maxsteps=1000)
>>>
Note that the high precision and strict tolerance is necessary for such high-degree runs,
since otherwise unwanted low-accuracy approximations will be detected. It may also be
necessary to set maxsteps high to prevent a premature exit (before the coecient bound
has been reached). Running with verbose=True to get an idea what is happening can
be useful.
Integer relations (PSLQ)
pslq()
mpmath.pslq(ctx, x, tol=None, maxcoe=1000, maxsteps=100, verbose=False)
Given a vector of real numbers x = [x0 , x1 , ..., xn ], pslq(x) uses the PSLQ algorithm to nd
a list of integers [c0 , c1 , ..., cn ] such that
|c1 x1 + c2 x2 + ... + cn xn | < tol
and such that max |ck | < maxcoeff. If no such vector exists, pslq() (page 1039) returns
None. The tolerance defaults to 3/4 of the working precision.
Examples
5.15. Welcome to mpmaths documentation!
1039
SymPy Documentation, Release 0.7.6
Find rational approximations for :
>>> from sympy.mpmath import *
>>> mp.dps = 15; mp.pretty = True
>>> pslq([-1, pi], tol=0.01)
[22, 7]
>>> pslq([-1, pi], tol=0.001)
[355, 113]
>>> mpf(22)/7; mpf(355)/113; +pi
3.14285714285714
3.14159292035398
3.14159265358979
Pi is not a rational number with denominator less than 1000:
>>> pslq([-1, pi])
>>>
To within the standard precision, it can however be approximated by at least one rational
number with denominator less than 1012 :
>>> p, q = pslq([-1, pi], maxcoeff=10**12)
>>> print(p); print(q)
238410049439
75888275702
>>> mpf(p)/q
3.14159265358979
The PSLQ algorithm can be applied to long vectors. For example, we can investigate the
rational (in)dependence of integer square roots:
>>>
>>>
>>>
>>>
>>>
>>>
[2,
mp.dps = 30
pslq([sqrt(n) for n in range(2, 5+1)])
pslq([sqrt(n) for n in range(2, 6+1)])
pslq([sqrt(n) for n in range(2, 8+1)])
0, 0, 0, 0, 0, -1]
Machin formulas
A famous formula for is Machins,
= 4 acot 5 acot 239
4
There are actually innitely many formulas of this type. Two others are
= acot 1
4
= 12 acot 49 + 32 acot 57 + 5 acot 239 + 12 acot 110443
4
We can easily verify the formulas using the PSLQ algorithm:
>>>
>>>
[1,
>>>
[1,
>>>
[1,
1040
mp.dps = 30
pslq([pi/4, acot(1)])
-1]
pslq([pi/4, acot(5), acot(239)])
-4, 1]
pslq([pi/4, acot(49), acot(57), acot(239), acot(110443)])
-12, -32, 5, -12]
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
We could try to generate a custom Machin-like formula by running the PSLQ algorithm
with a few inverse cotangent values, for example acot(2), acot(3) ... acot(10). Unfortunately, there is a linear dependence among these values, resulting in only that dependence being detected, with a zero coecient for :
>>> pslq([pi] + [acot(n) for n in range(2,11)])
[0, 1, -1, 0, 0, 0, -1, 0, 0, 0]
We get better luck by removing linearly dependent terms:
>>> pslq([pi] + [acot(n) for n in range(2,11) if n not in (3, 5)])
[1, -8, 0, 0, 4, 0, 0, 0]
In other words, we found the following formula:
>>> 8*acot(2) - 4*acot(7)
3.14159265358979323846264338328
>>> +pi
3.14159265358979323846264338328
Algorithm
This
is
a
fairly
direct
translation
to
Python
of
the
given
by
David
Bailey,
The
PSLQ
Integer
Relation
http://www.cecm.sfu.ca/organics/papers/bailey/paper/html/node3.html
pseudocode
Algorithm:
The present implementation uses xed-point instead of oating-point arithmetic, since
this is signicantly (about 7x) faster.
5.15.4 End matter
Precision and representation issues
Most of the time, using mpmath is simply a matter of setting the desired precision and entering a formula. For verication purposes, a quite (but not always!) reliable technique is to
calculate the same thing a second time at a higher precision and verifying that the results
agree.
To perform more advanced calculations, it is important to have some understanding of how
mpmath works internally and what the possible sources of error are. This section gives an
overview of arbitrary-precision binary oating-point arithmetic and some concepts from numerical analysis.
The following concepts are important to understand:
The main sources of numerical errors are rounding and cancellation, which are due to
the use of nite-precision arithmetic, and truncation or approximation errors, which are
due to approximating innite sequences or continuous functions by a nite number of
samples.
Errors propagate between calculations. A small error in the input may result in a large
error in the output.
Most numerical algorithms for complex problems (e.g. integrals, derivatives) give wrong
answers for suciently ill-behaved input. Sometimes virtually the only way to get a
wrong answer is to design some very contrived input, but at other times the chance of
accidentally obtaining a wrong result even for reasonable-looking input is quite high.
5.15. Welcome to mpmaths documentation!
1041
SymPy Documentation, Release 0.7.6
Like any complex numerical software, mpmath has implementation bugs. You should be
reasonably suspicious about any results computed by mpmath, even those it claims to be
able to compute correctly! If possible, verify results analytically, try dierent algorithms,
and cross-compare with other software.
Precision, error and tolerance
The following terms are common in this documentation:
Precision (or working precision) is the precision at which oating-point arithmetic operations are performed.
Error is the dierence between a computed approximation and the exact result.
Accuracy is the inverse of error.
Tolerance is the maximum error (or minimum accuracy) desired in a result.
Error and accuracy can be measured either directly, or logarithmically in bits or digits. Specifically, if a y is an approximation for y, then
(Direct) absolute error = |
y y|
(Direct) relative error = |
y y||y|1
(Direct) absolute accuracy = |
y y|1
(Direct) relative accuracy = |
y y|1 |y|
(Logarithmic) absolute error = logb |
y y|
(Logarithmic) relative error = logb |
y y| logb |y|
(Logarithmic) absolute accuracy = logb |
y y|
(Logarithmic) relative accuracy = logb |
y y| + logb |y|
where b = 2 and b = 10 for bits and digits respectively. Note that:
The logarithmic error roughly equals the position of the rst incorrect bit or digit
The logarithmic accuracy roughly equals the number of correct bits or digits in the result
These denitions also hold for complex numbers, using |a + bi| = a2 + b2 .
Full accuracy means that the accuracy of a result at least equals prec-1, i.e. it is correct
except possibly for the last bit.
Representation of numbers
Mpmath uses binary arithmetic. A binary oating-point number is a number of the form
man2exp where both man (the mantissa) and exp (the exponent) are integers. Some examples
of oating-point numbers are given in the following table.
Number
3
10
-16
1.25
Mantissa
3
5
-1
5
Exponent
0
1
4
-2
The representation as dened so far is not unique; one can always multiply the mantissa by 2
and subtract 1 from the exponent with no change in the numerical value. In mpmath, numbers
are always normalized so that man is an odd number, with the exception of zero which is
1042
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
always taken to have man = exp = 0. With these conventions, every representable number
has a unique representation. (Mpmath does not currently distinguish between positive and
negative zero.)
Simple mathematical operations are now easy to dene. Due to uniqueness, equality testing
of two numbers simply amounts to separately checking equality of the mantissas and the
exponents. Multiplication of nonzero numbers is straightforward: (m2e ) (n2f ) = (mn) 2e+f .
Addition is a bit more involved: we rst need to multiply the mantissa of one of the operands
by a suitable power of 2 to obtain equal exponents.
More technically, mpmath represents a oating-point number as a 4-tuple (sign, man, exp,
bc) where sign is 0 or 1 (indicating positive vs negative) and the mantissa is nonnegative; bc
(bitcount) is the size of the absolute value of the mantissa as measured in bits. Though redundant, keeping a separate sign eld and explicitly keeping track of the bitcount signicantly
speeds up arithmetic (the bitcount, especially, is frequently needed but slow to compute from
scratch due to the lack of a Python built-in function for the purpose).
Contrary to popular belief, oating-point numbers do not come with an inherent small uncertainty, although oating-point arithmetic generally is inexact. Every binary oating-point
number is an exact rational number. With arbitrary-precision integers used for the mantissa
and exponent, oating-point numbers can be added, subtracted and multiplied exactly. In
particular, integers and integer multiples of 1/2, 1/4, 1/8, 1/16, etc. can be represented,
added and multiplied exactly in binary oating-point arithmetic.
Floating-point arithmetic is generally approximate because the size of the mantissa must
be limited for eciency reasons. The maximum allowed width (bitcount) of the mantissa
is called the precision or prec for short. Sums and products of oating-point numbers are
exact as long as the absolute value of the mantissa is smaller than 2prec . As soon as the
mantissa becomes larger than this, it is truncated to contain at most prec bits (the exponent
is incremented accordingly to preserve the magnitude of the number), and this operation
introduces a rounding error. Division is also generally inexact; although we can add and
multiply exactly by setting the precision high enough, no precision is high enough to represent
for example 1/3 exactly (the same obviously applies for roots, trigonometric functions, etc).
The special numbers +inf, -inf and nan are represented internally by a zero mantissa and a
nonzero exponent.
Mpmath uses arbitrary precision integers for both the mantissa and the exponent, so numbers can be as large in magnitude as permitted by the computers memory. Some care may
be necessary when working with extremely large numbers. Although standard arithmetic
operators are safe, it is for example futile to attempt to compute the exponential function of
of 10100000 . Mpmath does not complain when asked to perform such a calculation, but instead
chugs away on the problem to the best of its ability, assuming that computer resources are
innite. In the worst case, this will be slow and allocate a huge amount of memory; if entirely impossible Python will at some point raise OverflowError: long int too large to
convert to int.
For further details on how the arithmetic is implemented, refer to the mpmath source code.
The basic arithmetic operations are found in the libmp directory; many functions there are
commented extensively.
Decimal issues
Mpmath uses binary arithmetic internally, while most interaction with the user is done via
the decimal number system. Translating between binary and decimal numbers is a somewhat
subtle matter; many Python novices run into the following bug (addressed in the General
Python FAQ):
5.15. Welcome to mpmaths documentation!
1043
SymPy Documentation, Release 0.7.6
>>> 1.2 - 1.0
0.19999999999999996
Decimal fractions fall into the category of numbers that generally cannot be represented
exactly in binary oating-point form. For example, none of the numbers 0.1, 0.01, 0.001 has
an exact representation as a binary oating-point number. Although mpmath can approximate
decimal fractions with any accuracy, it does not solve this problem for all uses; users who need
exact decimal fractions should look at the decimal module in Pythons standard library (or
perhaps use fractions, which are much faster).
With prec bits of precision, an arbitrary number can be approximated relatively to within
2prec , or within 10dps for dps decimal digits. The equivalent values for prec and dps are
therefore related proportionally via the factor C = log(10)/ log(2), or roughly 3.32. For example, the standard (binary) precision in mpmath is 53 bits, which corresponds to a decimal
precision of 15.95 digits.
More precisely, mpmath uses the following formulas to translate between prec and dps:
dps(prec) = max(1, int(round(int(prec) / C - 1)))
prec(dps) = max(1, int(round((int(dps) + 1) * C)))
Note that the dps is set 1 decimal digit lower than the corresponding binary precision. This
is done to hide minor rounding errors and artifacts resulting from binary-decimal conversion.
As a result, mpmath interprets 53 bits as giving 15 digits of decimal precision, not 16.
The dps value controls the number of digits to display when printing numbers with str(),
while the decimal precision used by repr() is set two or three digits higher. For example,
with 15 dps we have:
>>> from mpmath import *
>>> mp.dps = 15
>>> str(pi)
3.14159265358979
>>> repr(+pi)
mpf(3.1415926535897931)
The extra digits in the output from repr ensure that x == eval(repr(x)) holds, i.e. that
numbers can be converted to strings and back losslessly.
It should be noted that precision and accuracy do not always correlate when translating between binary and decimal. As a simple example, the number 0.1 has a decimal precision of 1
digit but is an innitely accurate representation of 1/10. Conversely, the number 250 has a
binary representation with 1 bit of precision that is innitely accurate; the same number can
actually be represented exactly as a decimal, but doing so requires 35 signicant digits:
0.00000000000000088817841970012523233890533447265625
All binary oating-point numbers can be represented exactly as decimals (possibly requiring
many digits), but the converse is false.
Correctness guarantees
Basic arithmetic operations (with the mp context) are always performed with correct rounding.
Results that can be represented exactly are guranteed to be exact, and results from single
inexact operations are guaranteed to be the best possible rounded values. For higher-level
operations, mpmath does not generally guarantee correct rounding. In general, mpmath only
guarantees that it will use at least the user-set precision to perform a given calculation. The
1044
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
user may have to manually set the working precision higher than the desired accuracy for the
result, possibly much higher.
Functions for evaluation of transcendental functions, linear algebra operations, numerical
integration, etc., usually automatically increase the working precision and use a stricter tolerance to give a correctly rounded result with high probability: for example, at 50 bits the
temporary precision might be set to 70 bits and the tolerance might be set to 60 bits. It can
often be assumed that such functions return values that have full accuracy, given inputs that
are exact (or suciently precise approximations of exact values), but the user must exercise
judgement about whether to trust mpmath.
The level of rigor in mpmath covers the entire spectrum from always correct by design
through nearly always correct and handling the most common errors to just computing
blindly and hoping for the best. Of course, a long-term development goal is to successively
increase the rigor where possible. The following list might give an idea of the current state.
Operations that are correctly rounded:
Addition, subtraction and multiplication of real and complex numbers.
Division and square roots of real numbers.
Powers of real numbers, assuming suciently small integer exponents (huge powers are
rounded in the right direction, but possibly farther than necessary).
Conversion from decimal to binary, for reasonably sized numbers (roughly between 10100
and 10100 ).
Typically, transcendental functions for exact input-output pairs.
Operations that should be fully accurate (however, the current implementation may be based
on a heuristic error analysis):
Radix conversion (large or small numbers).
Mathematical constants like .
Both real and imaginary parts of exp, cos, sin, cosh, sinh, log.
Other elementary functions (the largest of the real and imaginary part).
The gamma and log-gamma functions (the largest of the real and the imaginary part;
both, when close to real axis).
Some functions based on hypergeometric series (the largest of the real and imaginary
part).
Correctness of root-nding, numerical integration, etc. largely depends on the wellbehavedness of the input functions. Specic limitations are sometimes noted in the respective
sections of the documentation.
Double precision emulation
On most systems, Pythons float type represents an IEEE 754 double precision number, with
a precision of 53 bits and rounding-to-nearest. With default precision (mp.prec = 53), the
mpmath mpf type roughly emulates the behavior of the float type. Sources of incompatibility
include the following:
In hardware oating-point arithmetic, the size of the exponent is restricted to a xed
range: regular Python oats have a range between roughly 10300 and 10300 . Mpmath
does not emulate overow or underow when exponents fall outside this range.
5.15. Welcome to mpmaths documentation!
1045
SymPy Documentation, Release 0.7.6
On some systems, Python uses 80-bit (extended double) registers for oating-point operations. Due to double rounding, this makes the float type less accurate. This problem
is only known to occur with Python versions compiled with GCC on 32-bit systems.
Machine oats very close to the exponent limit round subnormally, meaning that they
lose accuracy (Python may raise an exception instead of rounding a float subnormally).
Mpmath is able to produce more accurate results for transcendental functions.
Further reading
There are many excellent textbooks on numerical analysis and oating-point arithmetic. Some
good web resources are:
David Goldberg, What Every Computer Scientist Should Know About Floating-Point
Arithmetic
The Wikipedia article about numerical analysis
References
The following is a non-comprehensive list of works used in the development of mpmath or
cited for examples or mathematical denitions used in this documentation. References not
listed here can be found in the source code.
5.16 Polynomials Manipulation Module
Computations with polynomials are at the core of computer algebra and having a fast and robust polynomials manipulation module is a key for building a powerful symbolic manipulation
system. SymPy has a dedicated module sympy.polys for computing in polynomial algebras
over various coecient domains.
There is a vast number of methods implemented, ranging from simple tools like polynomial
division, to advanced concepts including Grbner bases and multivariate factorization over
algebraic number domains.
5.16.1 Contents
Basic functionality of the module
Introduction
This tutorial tries to give an overview of the functionality concerning polynomials within
SymPy. All code examples assume:
>>> from sympy import *
>>> x, y, z = symbols(x,y,z)
>>> init_printing(use_unicode=False, wrap_line=False, no_global=True)
1046
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Basic functionality
These functions provide dierent algorithms dealing with polynomials in the form of SymPy
expression, like symbols, sums etc.
Division The function div() provides division of polynomials with remainder. That is, for
polynomials f and g, it computes q and r, such that f = g q + r and deg(r) < q. For polynomials
in one variables with coecients in a eld, say, the rational numbers, q and r are uniquely
dened this way:
>>> f = 5*x**2 + 10*x + 3
>>> g = 2*x + 2
>>> q, r = div(f, g, domain=QQ)
>>> q
5*x
5
--- + 2
2
>>> r
-2
>>> (q*g + r).expand()
2
5*x + 10*x + 3
As you can see, q has a non-integer coecient. If you want to do division only in the ring of
polynomials with integer coecients, you can specify an additional parameter:
>>> q, r = div(f, g, domain=ZZ)
>>> q
0
>>> r
2
5*x + 10*x + 3
But be warned, that this ring is no longer Euclidean and that the degree of the remainder
doesnt need to be smaller than that of f. Since 2 doesnt divide 5, 2x doesnt divide 5x2 , even
if the degree is smaller. But:
>>> g = 5*x + 1
>>> q, r = div(f, g, domain=ZZ)
>>> q
x
>>> r
9*x + 3
>>> (q*g + r).expand()
2
5*x + 10*x + 3
This also works for polynomials with multiple variables:
>>> f = x*y + y*z
>>> g = 3*x + 3*z
>>> q, r = div(f, g, domain=QQ)
>>> q
y
-
5.16. Polynomials Manipulation Module
1047
SymPy Documentation, Release 0.7.6
3
>>> r
0
In the last examples, all of the three variables x, y and z are assumed to be variables of
the polynomials. But if you have some unrelated constant as coecient, you can specify the
variables explicitly:
>>>
>>>
>>>
>>>
>>>
a*x
--3
a, b, c = symbols(a,b,c)
f = a*x**2 + b*x + c
g = 3*x + 2
q, r = div(f, g, domain=QQ)
q
2*a
b
- --- + 9
3
>>> r
4*a
2*b
--- - --- + c
9
3
GCD and LCM With division, there is also the computation of the greatest common divisor
and the least common multiple.
When the polynomials have integer coecients, the contents gcd is also considered:
>>> f = (12*x + 12)*x
>>> g = 16*x**2
>>> gcd(f, g)
4*x
But if the polynomials have rational coecients, then the returned polynomial is monic:
>>> f = 3*x**2/2
>>> g = 9*x/4
>>> gcd(f, g)
x
It also works with multiple variables. In this case, the variables are ordered alphabetically,
be default, which has inuence on the leading coecient:
>>> f = x*y/2 + y**2
>>> g = 3*x + 6*y
>>> gcd(f, g)
x + 2*y
The lcm is connected with the gcd and one can be computed using the other:
>>> f = x*y**2 + x**2*y
>>> g = x**2*y**2
>>> gcd(f, g)
x*y
>>> lcm(f, g)
3 2
2 3
x *y + x *y
>>> (f*g).expand()
4 3
3 4
1048
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
x *y + x *y
>>> (gcd(f, g, x, y)*lcm(f, g, x, y)).expand()
4 3
3 4
x *y + x *y
Square-free factorization The square-free factorization of a univariate polynomial is the
product of all factors (not necessarily irreducible) of degree 1, 2 etc.:
>>> f = 2*x**2 + 5*x**3 + 4*x**4 + x**5
>>> sqf_list(f)
(1, [(x + 2, 1), (x, 2), (x + 1, 2)])
>>> sqf(f)
2
2
x *(x + 1) *(x + 2)
Factorization This function provides factorization of univariate and multivariate polynomials with rational coecients:
>>> factor(x**4/2 + 5*x**3/12 - x**2/3)
2
x *(2*x - 1)*(3*x + 4)
---------------------12
>>> factor(x**2 + 4*x*y + 4*y**2)
2
(x + 2*y)
Groebner bases
ders:
Buchbergers algorithm is implemented, supporting various monomial or-
>>> groebner([x**2 + 1, y**4*x + x**3], x, y, order=lex)
/[ 2
4
]
\
GroebnerBasis\[x + 1, y - 1], x, y, domain=ZZ, order=lex/
>>> groebner([x**2 + 1, y**4*x + x**3, x*y*z**3], x, y, z, order=grevlex)
/[ 4
3
2
]
\
GroebnerBasis\[y - 1, z , x + 1], x, y, z, domain=ZZ, order=grevlex/
Solving Equations We have (incomplete) methods to nd the complex or even symbolic
roots of polynomials and to solve some systems of polynomial equations:
>>> from sympy import roots, solve_poly_system
>>> solve(x**3 + 2*x + 3, x)
____
____
1
\/ 11 *I 1
\/ 11 *I
[-1, - - --------, - + --------]
2
2
2
2
5.16. Polynomials Manipulation Module
1049
SymPy Documentation, Release 0.7.6
>>> p = Symbol(p)
>>> q = Symbol(q)
>>> solve(x**2 + p*x + q, x)
__________
__________
/ 2
/ 2
p
\/ p - 4*q
p
\/ p - 4*q
[- - - -------------, - - + -------------]
2
2
2
2
>>> solve_poly_system([y - x, x - 5], x, y)
[(5, 5)]
>>> solve_poly_system([y**2 - x**3 + 1, y*x], x, y)
___
___
1
\/ 3 *I
1
\/ 3 *I
[(0, -I), (0, I), (1, 0), (- - - -------, 0), (- - + -------, 0)]
2
2
2
2
Examples from Westers Article
Introduction
In this tutorial we present examples from Westers article concerning comparison and critique of mathematical abilities of several computer algebra systems (see [Wester1999]
(page 1910)). All the examples are related to polynomial and algebraic computations and
SymPy specic remarks were added to all of them.
Examples
All examples in this tutorial are computable, so one can just copy and paste them into a Python
shell and do something useful with them. All computations were done using the following
setup:
>>> from sympy import *
>>> init_printing(use_unicode=True, wrap_line=False, no_global=True)
>>> var(x,y,z,s,c,n)
(x, y, z, s, c, n)
Simple univariate polynomial factorization To obtain a factorization of a polynomial use
factor() function. By default factor() returns the result in unevaluated form, so the content
of the input polynomial is left unexpanded, as in the following example:
>>> factor(6*x - 10)
2(3x - 5)
To achieve the same eect in a more systematic way use primitive() function, which returns
the content and the primitive part of the input polynomial:
>>> primitive(6*x - 10)
(2, 3x - 5)
1050
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Note: The content and the primitive part can be computed only over a ring. To simplify
coecients of a polynomial over a eld use monic().
Univariate GCD, resultant and factorization Consider univariate polynomials f, g and h
over integers:
>>> f = 64*x**34 - 21*x**47 - 126*x**8 - 46*x**5 - 16*x**60 - 81
>>> g = 72*x**60 - 25*x**25 - 19*x**23 - 22*x**39 - 83*x**52 + 54*x**10 + 81
>>> h = 34*x**19 - 25*x**16 + 70*x**7 + 20*x**3 - 91*x - 86
We can compute the greatest common divisor (GCD) of two polynomials using gcd() function:
>>> gcd(f, g)
1
We see that f and g have no common factors. However, f*h and g*h have an obvious factor
h:
>>> gcd(expand(f*h), expand(g*h)) - h
0
The same can be veried using the resultant of univariate polynomials:
>>> resultant(expand(f*h), expand(g*h))
0
Factorization of large univariate polynomials (of degree 120 in this case) over integers is also
possible:
>>> factor(expand(f*g))
60
47
34
8
5
60
52
39
25
23
10
-16x
+ 21x
- 64x
+ 126x + 46x + 8172x
- 83x - 22x
- 25x
- 19x
+ 54x
Multivariate GCD and factorization What can be done in univariate case, can be also
done for multivariate polynomials. Consider the following polynomials f, g and h in Z[x, y, z]:
>>> f = 24*x*y**19*z**8 - 47*x**17*y**5*z**8 + 6*x**15*y**9*z**2 - 3*x**22 + 5
>>> g = 34*x**5*y**8*z**13 + 20*x**7*y**7*z**7 + 12*x**9*y**16*z**4 + 80*y**14*z
>>> h = 11*x**12*y**7*z**13 - 23*x**2*y**8*z**10 + 47*x**17*y**5*z**8
As previously, we can verify that f and g have no common factors:
>>> gcd(f, g)
1
However, f*h and g*h have an obvious factor h:
>>> gcd(expand(f*h), expand(g*h)) - h
0
Multivariate factorization of large polynomials is also possible:
>>> factor(expand(f*g))
7
9 9 3
7 6
5
12
7
22
17 5 8
15 9 2
19
-2y z6x y z + 10x z + 17x yz
+ 40y 3x
+ 47x y z - 6x y z - 24xy
5.16. Polynomials Manipulation Module
1051
SymPy Documentation, Release 0.7.6
Support for symbols in exponents Polynomial manipulation functions provided by
sympy.polys are mostly used with integer exponents. However, its perfectly valid to compute with symbolic exponents, e.g.:
>>> gcd(2*x**(n + 4) - x**(n + 2), 4*x**(n + 1) + 3*x**n)
n
x
Testing if polynomials have common zeros To test if two polynomials have a root in common we can use resultant() function. The theory says that the resultant of two polynomials
vanishes if there is a common zero of those polynomials. For example:
>>> resultant(3*x**4 + 3*x**3 + x**2 - x - 2, x**3 - 3*x**2 + x + 5)
0
We can visualize this fact by factoring the polynomials:
>>> factor(3*x**4 + 3*x**3 + x**2 - x - 2)
(x + 1)3x + x - 2
>>> factor(x**3 - 3*x**2 + x + 5)
2
(x + 1)x - 4x + 5
In both cases we obtained the factor x + 1 which tells us that the common root is x = 1.
Normalizing simple rational functions To remove common factors from the numerator
and the denominator of a rational function the elegant way, use cancel() function. For example:
>>> cancel((x**2 - 4)/(x**2 + 4*x + 4))
x - 2
----x + 2
Expanding expressions and factoring back One can work easily we expressions in both
expanded and factored forms. Consider a polynomial f in expanded form. We dierentiate it
and factor the result back:
>>> f = expand((x + 1)**20)
>>> g = diff(f, x)
>>> factor(g)
19
20(x + 1)
The same can be achieved in factored form:
>>> diff((x + 1)**20, x)
19
20(x + 1)
1052
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Factoring in terms of cyclotomic polynomials SymPy can very eciently decompose
polynomials of the form xn 1 in terms of cyclotomic polynomials:
>>> factor(x**15 - 1)
2
4
3
2
8
7
5
4
3
(x - 1)x + x + 1x + x + x + x + 1x - x + x - x + x - x + 1
The original Westers example was x100 1, but was truncated for readability purpose. Note
that this is not a big struggle for factor() to decompose polynomials of degree 1000 or
greater.
Univariate factoring over Gaussian numbers
integer coecients:
Consider a univariate polynomial f with
>>> f = 4*x**4 + 8*x**3 + 77*x**2 + 18*x + 153
We want to obtain a factorization of f over Gaussian numbers. To do this we use factor() as
previously, but this time we set gaussian keyword to True:
>>> factor(f, gaussian=True)
3i
3i
4x - ---x + ---(x + 1 - 4i)(x + 1 + 4i)
2
2
As the result we got a splitting factorization of f with monic factors (this is a general rule
when computing in a eld with SymPy). The gaussian keyword is useful for improving code
readability, however the same result can be computed using more general syntax:
>>> factor(f, extension=I)
3i
3i
4x - ---x + ---(x + 1 - 4i)(x + 1 + 4i)
2
2
Computing with automatic eld extensions
g:
Consider two univariate polynomials f and
>>> f = x**3 + (sqrt(2) - 2)*x**2 - (2*sqrt(2) + 3)*x - 3*sqrt(2)
>>> g = x**2 - 2
We would like to reduce degrees of the numerator and the denominator of a rational function
f/g. Do do this we employ cancel() function:
>>> cancel(f/g)
3
2
___ 2
___
___
x - 2x + \ 2 x - 3x - 2\ 2 x - 3\ 2
-----------------------------------------------2
x - 2
Unfortunately nothing interesting happened. This is because by default SymPy treats 2 as a
generator, obtaining a bivariate
polynomial for the numerator. To make cancel() recognize
algebraic properties of 2, one needs to use extension keyword:
>>> cancel(f/g, extension=True)
2
x - 2x - 3
------------
5.16. Polynomials Manipulation Module
1053
SymPy Documentation, Release 0.7.6
___
x - \ 2
Setting extension=True tells cancel() to nd minimal algebraic
number domain for the co
ecients of f/g. The automatically inferred domain is Q( 2). If one doesnt want to rely on
automatic inference, the same result can be obtained by setting the extension keyword with
an explicit algebraic number:
>>> cancel(f/g, extension=sqrt(2))
2
x - 2x - 3
-----------___
x - \ 2
Univariate factoring over various domains Consider a univariate polynomial f with integer coecients:
>>> f = x**4 - 3*x**2 + 1
With sympy.polys we can obtain factorizations of f over dierent domains, which includes:
rationals:
>>> factor(f)
2
2
x - x - 1x + x - 1
nite elds:
>>> factor(f, modulus=5)
2
2
(x - 2) (x + 2)
algebraic numbers:
>>> alg = AlgebraicNumber((sqrt(5) - 1)/2, alias=alpha)
>>> factor(f, extension=alg)
(x - )(x + )(x - 1 - )(x + + 1)
Factoring polynomials into linear factors Currently SymPy can factor polynomials into
irreducibles over various domains, which can result in a splitting factorization (into linear
factors). However, there is currently no systematic way to infer a splitting eld (algebraic
number eld) automatically. In future the following syntax will be implemented:
>>> factor(x**3 + x**2 - 7, split=True)
Traceback (most recent call last):
...
NotImplementedError: split option is not implemented yet
Traceback (most recent call last):
...
NotImplementedError: split option is not implemented yet
Note this is dierent from extension=True, because the later only tells how expression parsing should be done, not what should be the domain of computation. One can simulate the
split keyword for several classes of polynomials using solve() function.
1054
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Advanced factoring over nite elds
coecients:
Consider a univariate polynomial f with integer
>>> f = x**11 + x + 1
We can factor f over a large nite eld F65537 :
>>> factor(f, modulus=65537)
2
9
8
6
5
3
2
x + x + 1x - x + x - x + x - x + 1
and expand the resulting factorization back:
>>> expand(_)
11
x
+ x + 1
obtaining polynomial f. This was done using symmetric polynomial representation over nite
elds The same thing can be done using non-symmetric representation:
>>> factor(f, modulus=65537, symmetric=False)
2
9
8
6
5
3
2
x + x + 1x + 65536x + x + 65536x + x + 65536x + 1
As with symmetric representation we can expand the factorization to get the input polynomial
back. This time, however, we need to truncate coecients of the expanded polynomial modulo
65537:
>>> trunc(expand(_), 65537)
11
x
+ x + 1
Working with expressions as polynomials
Z[x, y, z]:
Consider a multivariate polynomial f in
>>> f = expand((x - 2*y**2 + 3*z**3)**20)
We want to compute factorization of f. To do this we use factor as usually, however we
note that the polynomial in consideration is already in expanded form, so we can tell the
factorization routine to skip expanding f:
>>> factor(f, expand=False)
20
2
3
x - 2y + 3z
The default in sympy.polys is to expand all expressions given as arguments to polynomial
manipulation functions and Poly class. If we know that expanding is unnecessary, then by
setting expand=False we can save quite a lot of time for complicated inputs. This can be
really important when computing with expressions like:
>>> g = expand((sin(x) - 2*cos(y)**2 + 3*tan(z)**3)**20)
>>> factor(g, expand=False)
20
2
3
-sin(x) + 2cos (y) - 3tan (z)
5.16. Polynomials Manipulation Module
1055
SymPy Documentation, Release 0.7.6
Computing reduced Grbner bases To compute a reduced Grbner basis for a set of
polynomials use groebner() function. The function accepts various monomial orderings, e.g.:
lex, grlex and grevlex, or a user dened one, via order keyword. The lex ordering is
the most interesting because it has elimination property, which means that if the system of
polynomial equations to groebner() is zero-dimensional (has nite number of solutions) the
last element of the basis is a univariate polynomial. Consider the following example:
>>> f = expand((1 - c**2)**5 * (1 - s**2)**5 * (c**2 + s**2)**10)
>>> groebner([f, c**2 + s**2 - 1])
2
2
20
18
16
14
12
10
GroebnerBasisc + s - 1, c
- 5c
+ 10c
- 10c
+ 5c
- c , s, c, domain=, order=lex
The result is an ordinary Python list, so we can easily apply a function to all its elements, for
example we can factor those elements:
>>> list(map(factor, _))
2
2
10
5
5
c + s - 1, c (c - 1) (c + 1)
From the above we can easily nd all solutions of the system of polynomial equations. Or we
can use solve() to achieve this in a more systematic way:
>>> solve([f, s**2 + c**2 - 1], c, s)
[(-1, 0), (0, -1), (0, 1), (1, 0)]
Multivariate factoring over algebraic numbers Computing with multivariate polynomials over various domains
is as simple as in univariate case. For example consider the following
factorization over Q( 3):
>>> factor(x**3 +
(x + y)x + y
y**3, extension=sqrt(-3))
___
___
1
\ 3 i
1
\ 3 i
- - -------x + y- - + -------
2
2
2
2
Note: Currently multivariate polynomials over nite elds arent supported.
Partial fraction decomposition Consider a univariate rational function f with integer coecients:
>>> f = (x**2 + 2*x + 3)/(x**3 + 4*x**2 + 5*x + 2)
To decompose f into partial fractions use apart() function:
>>> apart(f)
3
2
2
----- - ----- + -------x + 2
x + 1
2
(x + 1)
To return from partial fractions to the rational function use a composition of together() and
cancel():
1056
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> cancel(together(_))
2
x + 2x + 3
------------------3
2
x + 4x + 5x + 2
Literature
Polynomials Manipulation Module Reference
Basic polynomial manipulation functions
sympy.polys.polytools.poly(expr, *gens, **args)
Eciently transform an expression into a polynomial.
Examples
>>> from sympy import poly
>>> from sympy.abc import x
>>> poly(x*(x**2 + x - 1)**2)
Poly(x**5 + 2*x**4 - x**3 - 2*x**2 + x, x, domain=ZZ)
sympy.polys.polytools.poly from expr(expr, *gens, **args)
Construct a polynomial from an expression.
sympy.polys.polytools.parallel poly from expr(exprs, *gens, **args)
Construct polynomials from expressions.
sympy.polys.polytools.degree(f, *gens, **args)
Return the degree of f in the given variable.
The degree of 0 is negative innity.
Examples
>>> from sympy import degree
>>> from sympy.abc import x, y
>>> degree(x**2 + y*x + 1, gen=x)
2
>>> degree(x**2 + y*x + 1, gen=y)
1
>>> degree(0, x)
-oo
sympy.polys.polytools.degree list(f, *gens, **args)
Return a list of degrees of f in all variables.
5.16. Polynomials Manipulation Module
1057
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import degree_list
>>> from sympy.abc import x, y
>>> degree_list(x**2 + y*x + 1)
(2, 1)
sympy.polys.polytools.LC(f, *gens, **args)
Return the leading coecient of f.
Examples
>>> from sympy import LC
>>> from sympy.abc import x, y
>>> LC(4*x**2 + 2*x*y**2 + x*y + 3*y)
4
sympy.polys.polytools.LM(f, *gens, **args)
Return the leading monomial of f.
Examples
>>> from sympy import LM
>>> from sympy.abc import x, y
>>> LM(4*x**2 + 2*x*y**2 + x*y + 3*y)
x**2
sympy.polys.polytools.LT(f, *gens, **args)
Return the leading term of f.
Examples
>>> from sympy import LT
>>> from sympy.abc import x, y
>>> LT(4*x**2 + 2*x*y**2 + x*y + 3*y)
4*x**2
sympy.polys.polytools.pdiv(f, g, *gens, **args)
Compute polynomial pseudo-division of f and g.
Examples
>>> from sympy import pdiv
>>> from sympy.abc import x
1058
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> pdiv(x**2 + 1, 2*x - 4)
(2*x + 4, 20)
sympy.polys.polytools.prem(f, g, *gens, **args)
Compute polynomial pseudo-remainder of f and g.
Examples
>>> from sympy import prem
>>> from sympy.abc import x
>>> prem(x**2 + 1, 2*x - 4)
20
sympy.polys.polytools.pquo(f, g, *gens, **args)
Compute polynomial pseudo-quotient of f and g.
Examples
>>> from sympy import pquo
>>> from sympy.abc import x
>>>
2*x
>>>
2*x
pquo(x**2 + 1, 2*x - 4)
+ 4
pquo(x**2 - 1, 2*x - 1)
+ 1
sympy.polys.polytools.pexquo(f, g, *gens, **args)
Compute polynomial exact pseudo-quotient of f and g.
Examples
>>> from sympy import pexquo
>>> from sympy.abc import x
>>> pexquo(x**2 - 1, 2*x - 2)
2*x + 2
>>> pexquo(x**2 + 1, 2*x - 4)
Traceback (most recent call last):
...
ExactQuotientFailed: 2*x - 4 does not divide x**2 + 1
Traceback (most recent call last):
...
ExactQuotientFailed: 2*x - 4 does not divide x**2 + 1
sympy.polys.polytools.div(f, g, *gens, **args)
Compute polynomial division of f and g.
5.16. Polynomials Manipulation Module
1059
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import div, ZZ, QQ
>>> from sympy.abc import x
>>> div(x**2 + 1, 2*x - 4, domain=ZZ)
(0, x**2 + 1)
>>> div(x**2 + 1, 2*x - 4, domain=QQ)
(x/2 + 1, 5)
sympy.polys.polytools.rem(f, g, *gens, **args)
Compute polynomial remainder of f and g.
Examples
>>> from sympy import rem, ZZ, QQ
>>> from sympy.abc import x
>>> rem(x**2 + 1, 2*x - 4, domain=ZZ)
x**2 + 1
>>> rem(x**2 + 1, 2*x - 4, domain=QQ)
5
sympy.polys.polytools.quo(f, g, *gens, **args)
Compute polynomial quotient of f and g.
Examples
>>> from sympy import quo
>>> from sympy.abc import x
>>>
x/2
>>>
x +
quo(x**2 + 1, 2*x - 4)
+ 1
quo(x**2 - 1, x - 1)
1
sympy.polys.polytools.exquo(f, g, *gens, **args)
Compute polynomial exact quotient of f and g.
Examples
>>> from sympy import exquo
>>> from sympy.abc import x
>>> exquo(x**2 - 1, x - 1)
x + 1
>>> exquo(x**2 + 1, 2*x - 4)
Traceback (most recent call last):
...
ExactQuotientFailed: 2*x - 4 does not divide x**2 + 1
Traceback (most recent call last):
1060
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
...
ExactQuotientFailed: 2*x - 4 does not divide x**2 + 1
sympy.polys.polytools.half gcdex(f, g, *gens, **args)
Half extended Euclidean algorithm of f and g.
Returns (s, h) such that h = gcd(f, g) and s*f = h (mod g).
Examples
>>> from sympy import half_gcdex
>>> from sympy.abc import x
>>> half_gcdex(x**4 - 2*x**3 - 6*x**2 + 12*x + 15, x**3 + x**2 - 4*x - 4)
(-x/5 + 3/5, x + 1)
sympy.polys.polytools.gcdex(f, g, *gens, **args)
Extended Euclidean algorithm of f and g.
Returns (s, t, h) such that h = gcd(f, g) and s*f + t*g = h.
Examples
>>> from sympy import gcdex
>>> from sympy.abc import x
>>> gcdex(x**4 - 2*x**3 - 6*x**2 + 12*x + 15, x**3 + x**2 - 4*x - 4)
(-x/5 + 3/5, x**2/5 - 6*x/5 + 2, x + 1)
sympy.polys.polytools.invert(f, g, *gens, **args)
Invert f modulo g when possible.
Examples
>>> from sympy import invert
>>> from sympy.abc import x
>>> invert(x**2 - 1, 2*x - 1)
-4/3
>>> invert(x**2 - 1, x - 1)
Traceback (most recent call last):
...
NotInvertible: zero divisor
Traceback (most recent call last):
...
NotInvertible: zero divisor
sympy.polys.polytools.subresultants(f, g, *gens, **args)
Compute subresultant PRS of f and g.
5.16. Polynomials Manipulation Module
1061
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import subresultants
>>> from sympy.abc import x
>>> subresultants(x**2 + 1, x**2 - 1)
[x**2 + 1, x**2 - 1, -2]
sympy.polys.polytools.resultant(f, g, *gens, **args)
Compute resultant of f and g.
Examples
>>> from sympy import resultant
>>> from sympy.abc import x
>>> resultant(x**2 + 1, x**2 - 1)
4
sympy.polys.polytools.discriminant(f, *gens, **args)
Compute discriminant of f.
Examples
>>> from sympy import discriminant
>>> from sympy.abc import x
>>> discriminant(x**2 + 2*x + 3)
-8
sympy.polys.dispersion.dispersion(p, q=None, *gens, **args)
Compute the dispersion of polynomials.
For two polynomials f (x) and g(x) with deg f > 0 and deg g > 0 the dispersion dis(f, g) is
dened as:
dis(f, g) := max{J(f, g) {0}}
= max{{a N| gcd(f (x), g(x + a)) 6= 1} {0}}
and for a single polynomial dis(f ) := dis(f, f ). Note that we make the denition max{} :=
.
See Also:
dispersionset
References
1.[ManWright94] (page 1912)
2.[Koepf98] (page 1912)
3.[Abramov71] (page 1912)
1062
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
4.[Man93] (page 1912)
[ManWright94] (page 1912), [Koepf98] (page 1912), [Abramov71] (page 1912), [Man93]
(page 1912)
Examples
>>> from sympy import poly
>>> from sympy.polys.dispersion import dispersion, dispersionset
>>> from sympy.abc import x
Dispersion set and dispersion of a simple polynomial:
>>>
>>>
[0,
>>>
6
fp = poly((x - 3)*(x + 3), x)
sorted(dispersionset(fp))
6]
dispersion(fp)
Note that the denition of the dispersion is not symmetric:
>>>
>>>
>>>
[2,
>>>
4
>>>
[]
>>>
-oo
fp = poly(x**4 - 3*x**2 + 1, x)
gp = fp.shift(-3)
sorted(dispersionset(fp, gp))
3, 4]
dispersion(fp, gp)
sorted(dispersionset(gp, fp))
dispersion(gp, fp)
The maximum of an empty set is dened to be as seen in this example.
Computing the dispersion also works over eld extensions:
>>>
>>>
>>>
>>>
[2]
>>>
[1,
from sympy import sqrt
fp = poly(x**2 + sqrt(5)*x - 1, x, domain=QQ<sqrt(5)>)
gp = poly(x**2 + (2 + sqrt(5))*x + sqrt(5), x, domain=QQ<sqrt(5)>)
sorted(dispersionset(fp, gp))
sorted(dispersionset(gp, fp))
4]
We can even perform the computations for polynomials having symbolic coecients:
>>>
>>>
>>>
[0,
from sympy.abc import a
fp = poly(4*x**4 + (4*a + 8)*x**3 + (a**2 + 6*a + 4)*x**2 + (a**2 + 2*a)*x, x)
sorted(dispersionset(fp))
1]
sympy.polys.dispersion.dispersionset(p, q=None, *gens, **args)
Compute the dispersion set of two polynomials.
For two polynomials f (x) and g(x) with deg f > 0 and deg g > 0 the dispersion set J(f, g) is
5.16. Polynomials Manipulation Module
1063
SymPy Documentation, Release 0.7.6
dened as:
J(f, g) := {a N0 | gcd(f (x), g(x + a)) 6= 1}
= {a N0 | deg gcd(f (x), g(x + a)) 1}
For a single polynomial one denes J(f ) := J(f, f ).
See Also:
dispersion
References
1.[ManWright94] (page 1912)
2.[Koepf98] (page 1912)
3.[Abramov71] (page 1912)
4.[Man93] (page 1912)
[ManWright94] (page 1912), [Koepf98] (page 1912), [Abramov71] (page 1912), [Man93]
(page 1912)
Examples
>>> from sympy import poly
>>> from sympy.polys.dispersion import dispersion, dispersionset
>>> from sympy.abc import x
Dispersion set and dispersion of a simple polynomial:
>>>
>>>
[0,
>>>
6
fp = poly((x - 3)*(x + 3), x)
sorted(dispersionset(fp))
6]
dispersion(fp)
Note that the denition of the dispersion is not symmetric:
>>>
>>>
>>>
[2,
>>>
4
>>>
[]
>>>
-oo
fp = poly(x**4 - 3*x**2 + 1, x)
gp = fp.shift(-3)
sorted(dispersionset(fp, gp))
3, 4]
dispersion(fp, gp)
sorted(dispersionset(gp, fp))
dispersion(gp, fp)
Computing the dispersion also works over eld extensions:
>>>
>>>
>>>
>>>
1064
from sympy import sqrt
fp = poly(x**2 + sqrt(5)*x - 1, x, domain=QQ<sqrt(5)>)
gp = poly(x**2 + (2 + sqrt(5))*x + sqrt(5), x, domain=QQ<sqrt(5)>)
sorted(dispersionset(fp, gp))
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
[2]
>>> sorted(dispersionset(gp, fp))
[1, 4]
We can even perform the computations for polynomials having symbolic coecients:
>>>
>>>
>>>
[0,
from sympy.abc import a
fp = poly(4*x**4 + (4*a + 8)*x**3 + (a**2 + 6*a + 4)*x**2 + (a**2 + 2*a)*x, x)
sorted(dispersionset(fp))
1]
sympy.polys.polytools.terms gcd(f, *gens, **args)
Remove GCD of terms from f.
If the deep ag is True, then the arguments of f will have terms gcd applied to them.
If a fraction is factored out of f and f is an Add, then an unevaluated Mul will be returned
so that automatic simplication does not redistribute it. The hint clear, when set to
False, can be used to prevent such factoring when all coecients are not fractions.
See Also:
sympy.core.exprtools.gcd terms (page 185), sympy.core.exprtools.factor terms
(page 186)
Examples
>>> from sympy import terms_gcd, cos
>>> from sympy.abc import x, y
>>> terms_gcd(x**6*y**2 + x**3*y, x, y)
x**3*y*(x**3*y + 1)
The default action of polys routines is to expand the expression given to them. terms gcd
follows this behavior:
>>> terms_gcd((3+3*x)*(x+x*y))
3*x*(x*y + x + y + 1)
If this is not desired then the hint expand can be set to False. In this case the expression
will be treated as though it were comprised of one or more terms:
>>> terms_gcd((3+3*x)*(x+x*y), expand=False)
(3*x + 3)*(x*y + x)
In order to traverse factors of a Mul or the arguments of other functions, the deep hint
can be used:
>>> terms_gcd((3 + 3*x)*(x + x*y), expand=False, deep=True)
3*x*(x + 1)*(y + 1)
>>> terms_gcd(cos(x + x*y), deep=True)
cos(x*(y + 1))
Rationals are factored out by default:
>>> terms_gcd(x + y/2)
(2*x + y)/2
Only the y-term had a coecient that was a fraction; if one does not want to factor out
the 1/2 in cases like this, the ag clear can be set to False:
5.16. Polynomials Manipulation Module
1065
SymPy Documentation, Release 0.7.6
>>> terms_gcd(x + y/2, clear=False)
x + y/2
>>> terms_gcd(x*y/2 + y**2, clear=False)
y*(x/2 + y)
The clear ag is ignored if all coecients are fractions:
>>> terms_gcd(x/3 + y/2, clear=False)
(2*x + 3*y)/6
sympy.polys.polytools.cofactors(f, g, *gens, **args)
Compute GCD and cofactors of f and g.
Returns polynomials (h, cff, cfg) such that h = gcd(f, g), and cff = quo(f, h)
and cfg = quo(g, h) are, so called, cofactors of f and g.
Examples
>>> from sympy import cofactors
>>> from sympy.abc import x
>>> cofactors(x**2 - 1, x**2 - 3*x + 2)
(x - 1, x + 1, x - 2)
sympy.polys.polytools.gcd(f, g=None, *gens, **args)
Compute GCD of f and g.
Examples
>>> from sympy import gcd
>>> from sympy.abc import x
>>> gcd(x**2 - 1, x**2 - 3*x + 2)
x - 1
sympy.polys.polytools.gcd list(seq, *gens, **args)
Compute GCD of a list of polynomials.
Examples
>>> from sympy import gcd_list
>>> from sympy.abc import x
>>> gcd_list([x**3 - 1, x**2 - 1, x**2 - 3*x + 2])
x - 1
sympy.polys.polytools.lcm(f, g=None, *gens, **args)
Compute LCM of f and g.
1066
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import lcm
>>> from sympy.abc import x
>>> lcm(x**2 - 1, x**2 - 3*x + 2)
x**3 - 2*x**2 - x + 2
sympy.polys.polytools.lcm list(seq, *gens, **args)
Compute LCM of a list of polynomials.
Examples
>>> from sympy import lcm_list
>>> from sympy.abc import x
>>> lcm_list([x**3 - 1, x**2 - 1, x**2 - 3*x + 2])
x**5 - x**4 - 2*x**3 - x**2 + x + 2
sympy.polys.polytools.trunc(f, p, *gens, **args)
Reduce f modulo a constant p.
Examples
>>> from sympy import trunc
>>> from sympy.abc import x
>>> trunc(2*x**3 + 3*x**2 + 5*x + 7, 3)
-x**3 - x + 1
sympy.polys.polytools.monic(f, *gens, **args)
Divide all coecients of f by LC(f).
Examples
>>> from sympy import monic
>>> from sympy.abc import x
>>> monic(3*x**2 + 4*x + 2)
x**2 + 4*x/3 + 2/3
sympy.polys.polytools.content(f, *gens, **args)
Compute GCD of coecients of f.
Examples
>>> from sympy import content
>>> from sympy.abc import x
5.16. Polynomials Manipulation Module
1067
SymPy Documentation, Release 0.7.6
>>> content(6*x**2 + 8*x + 12)
2
sympy.polys.polytools.primitive(f, *gens, **args)
Compute content and the primitive form of f.
Examples
>>> from sympy.polys.polytools import primitive
>>> from sympy.abc import x
>>> primitive(6*x**2 + 8*x + 12)
(2, 3*x**2 + 4*x + 6)
>>> eq = (2 + 2*x)*x + 2
Expansion is performed by default:
>>> primitive(eq)
(2, x**2 + x + 1)
Set expand to False to shut this o. Note that the extraction will not be recursive; use
the as content primitive method for recursive, non-destructive Rational extraction.
>>> primitive(eq, expand=False)
(1, x*(2*x + 2) + 2)
>>> eq.as_content_primitive()
(2, x*(x + 1) + 1)
sympy.polys.polytools.compose(f, g, *gens, **args)
Compute functional composition f(g).
Examples
>>> from sympy import compose
>>> from sympy.abc import x
>>> compose(x**2 + x, x - 1)
x**2 - x
sympy.polys.polytools.decompose(f, *gens, **args)
Compute functional decomposition of f.
Examples
>>> from sympy import decompose
>>> from sympy.abc import x
>>> decompose(x**4 + 2*x**3 - x - 1)
[x**2 - x - 1, x**2 + x]
sympy.polys.polytools.sturm(f, *gens, **args)
Compute Sturm sequence of f.
1068
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import sturm
>>> from sympy.abc import x
>>> sturm(x**3 - 2*x**2 + x - 3)
[x**3 - 2*x**2 + x - 3, 3*x**2 - 4*x + 1, 2*x/9 + 25/9, -2079/4]
sympy.polys.polytools.gff list(f, *gens, **args)
Compute a list of greatest factorial factors of f.
Examples
>>> from sympy import gff_list, ff
>>> from sympy.abc import x
>>> f = x**5 + 2*x**4 - x**3 - 2*x**2
>>> gff_list(f)
[(x, 1), (x + 2, 4)]
>>> (ff(x, 1)*ff(x + 2, 4)).expand() == f
True
sympy.polys.polytools.gff(f, *gens, **args)
Compute greatest factorial factorization of f.
sympy.polys.polytools.sqf norm(f, *gens, **args)
Compute square-free norm of f.
Returns s, f, r, such that g(x) = f(x-sa) and r(x) = Norm(g(x)) is a square-free
polynomial over K, where a is the algebraic extension of the ground domain.
Examples
>>> from sympy import sqf_norm, sqrt
>>> from sympy.abc import x
>>> sqf_norm(x**2 + 1, extension=[sqrt(3)])
(1, x**2 - 2*sqrt(3)*x + 4, x**4 - 4*x**2 + 16)
sympy.polys.polytools.sqf part(f, *gens, **args)
Compute square-free part of f.
Examples
>>> from sympy import sqf_part
>>> from sympy.abc import x
>>> sqf_part(x**3 - 3*x - 2)
x**2 - x - 2
sympy.polys.polytools.sqf list(f, *gens, **args)
Compute a list of square-free factors of f.
5.16. Polynomials Manipulation Module
1069
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import sqf_list
>>> from sympy.abc import x
>>> sqf_list(2*x**5 + 16*x**4 + 50*x**3 + 76*x**2 + 56*x + 16)
(2, [(x + 1, 2), (x + 2, 3)])
sympy.polys.polytools.sqf(f, *gens, **args)
Compute square-free factorization of f.
Examples
>>> from sympy import sqf
>>> from sympy.abc import x
>>> sqf(2*x**5 + 16*x**4 + 50*x**3 + 76*x**2 + 56*x + 16)
2*(x + 1)**2*(x + 2)**3
sympy.polys.polytools.factor list(f, *gens, **args)
Compute a list of irreducible factors of f.
Examples
>>> from sympy import factor_list
>>> from sympy.abc import x, y
>>> factor_list(2*x**5 + 2*x**4*y + 4*x**3 + 4*x**2*y + 2*x + 2*y)
(2, [(x + y, 1), (x**2 + 1, 2)])
sympy.polys.polytools.factor(f, *gens, **args)
Compute the factorization of expression, f, into irreducibles. (To factor an integer into
primes, use factorint.)
There two modes implemented: symbolic and formal. If f is not an instance of Poly
(page 1075) and generators are not specied, then the former mode is used. Otherwise,
the formal mode is used.
In symbolic mode, factor() (page 1070) will traverse the expression tree and factor its
components without any prior expansion, unless an instance of Add is encountered (in
this case formal factorization is used). This way factor() (page 1070) can handle large
or symbolic exponents.
By default, the factorization is computed over the rationals. To factor over other domain,
e.g. an algebraic or nite eld, use appropriate options: extension, modulus or domain.
See Also:
sympy.ntheory.factor .factorint (page 287)
Examples
1070
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> from sympy import factor, sqrt
>>> from sympy.abc import x, y
>>> factor(2*x**5 + 2*x**4*y + 4*x**3 + 4*x**2*y + 2*x + 2*y)
2*(x + y)*(x**2 + 1)**2
>>> factor(x**2 + 1)
x**2 + 1
>>> factor(x**2 + 1, modulus=2)
(x + 1)**2
>>> factor(x**2 + 1, gaussian=True)
(x - I)*(x + I)
>>> factor(x**2 - 2, extension=sqrt(2))
(x - sqrt(2))*(x + sqrt(2))
>>> factor((x**2 - 1)/(x**2 + 4*x + 4))
(x - 1)*(x + 1)/(x + 2)**2
>>> factor((x**2 + 4*x + 4)**10000000*(x**2 + 1))
(x + 2)**20000000*(x**2 + 1)
By default, factor deals with an expression as a whole:
>>> eq = 2**(x**2 + 2*x + 1)
>>> factor(eq)
2**(x**2 + 2*x + 1)
If the deep ag is True then subexpressions will be factored:
>>> factor(eq, deep=True)
2**((x + 1)**2)
sympy.polys.polytools.intervals(F, all=False, eps=None, inf=None, sup=None,
strict=False, fast=False, sqf=False)
Compute isolating intervals for roots of f.
Examples
>>> from sympy import intervals
>>> from sympy.abc import x
>>> intervals(x**2 - 3)
[((-2, -1), 1), ((1, 2), 1)]
>>> intervals(x**2 - 3, eps=1e-2)
[((-26/15, -19/11), 1), ((19/11, 26/15), 1)]
sympy.polys.polytools.refine root(f, s, t, eps=None, steps=None, fast=False,
check sqf=False)
Rene an isolating interval of a root to the given precision.
Examples
>>> from sympy import refine_root
>>> from sympy.abc import x
5.16. Polynomials Manipulation Module
1071
SymPy Documentation, Release 0.7.6
>>> refine_root(x**2 - 3, 1, 2, eps=1e-2)
(19/11, 26/15)
sympy.polys.polytools.count roots(f, inf=None, sup=None)
Return the number of roots of f in [inf, sup] interval.
If one of inf or sup is complex, it will return the number of roots in the complex rectangle
with corners at inf and sup.
Examples
>>> from sympy import count_roots, I
>>> from sympy.abc import x
>>> count_roots(x**4 - 4, -3, 3)
2
>>> count_roots(x**4 - 4, 0, 1 + 3*I)
1
sympy.polys.polytools.real roots(f, multiple=True)
Return a list of real roots with multiplicities of f.
Examples
>>> from sympy import real_roots
>>> from sympy.abc import x
>>> real_roots(2*x**3 - 7*x**2 + 4*x + 4)
[-1/2, 2, 2]
sympy.polys.polytools.nroots(f, n=15, maxsteps=50, cleanup=True)
Compute numerical approximations of roots of f.
Examples
>>> from sympy import nroots
>>> from sympy.abc import x
>>> nroots(x**2 - 3, n=15)
[-1.73205080756888, 1.73205080756888]
>>> nroots(x**2 - 3, n=30)
[-1.73205080756887729352744634151, 1.73205080756887729352744634151]
sympy.polys.polytools.ground roots(f, *gens, **args)
Compute roots of f by factorization in the ground domain.
Examples
>>> from sympy import ground_roots
>>> from sympy.abc import x
1072
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> ground_roots(x**6 - 4*x**4 + 4*x**3 - x**2)
{0: 2, 1: 2}
sympy.polys.polytools.nth power roots poly(f, n, *gens, **args)
Construct a polynomial with n-th powers of roots of f.
Examples
>>> from sympy import nth_power_roots_poly, factor, roots
>>> from sympy.abc import x
>>> f = x**4 - x**2 + 1
>>> g = factor(nth_power_roots_poly(f, 2))
>>> g
(x**2 - x + 1)**2
>>> R_f = [ (r**2).expand() for r in roots(f) ]
>>> R_g = roots(g).keys()
>>> set(R_f) == set(R_g)
True
sympy.polys.polytools.cancel(f, *gens, **args)
Cancel common factors in a rational function f.
Examples
>>> from sympy import cancel, sqrt, Symbol
>>> from sympy.abc import x
>>> A = Symbol(A, commutative=False)
>>> cancel((2*x**2 - 2)/(x**2 - 2*x + 1))
(2*x + 2)/(x - 1)
>>> cancel((sqrt(3) + sqrt(15)*A)/(sqrt(2) + sqrt(10)*A))
sqrt(6)/2
sympy.polys.polytools.reduced(f, G, *gens, **args)
Reduces a polynomial f modulo a set of polynomials G.
Given a polynomial f and a set of polynomials G = (g 1, ..., g n), computes a set of
quotients q = (q 1, ..., q n) and the remainder r such that f = q 1*g 1 + ... +
q n*g n + r, where r vanishes or r is a completely reduced polynomial with respect to
G.
Examples
>>> from sympy import reduced
>>> from sympy.abc import x, y
>>> reduced(2*x**4 + y**2 - x**2 + y**3, [x**3 - x, y**3 - y])
([2*x, 1], x**2 + y**2 + y)
5.16. Polynomials Manipulation Module
1073
SymPy Documentation, Release 0.7.6
sympy.polys.polytools.groebner(F, *gens, **args)
Computes the reduced Groebner basis for a set of polynomials.
Use the order argument to set the monomial ordering that will be used to compute the
basis. Allowed orders are lex, grlex and grevlex. If no order is specied, it defaults to
lex.
For more information on Groebner bases, see the references and the docstring of
solvep olys ystem().
References
1.[Buchberger01] (page 1911)
2.[Cox97] (page 1911)
[Buchberger01] (page 1911), [Cox97] (page 1911)
Examples
Example taken from [1].
>>> from sympy import groebner
>>> from sympy.abc import x, y
>>> F = [x*y - 2*y, 2*y**2 - x**2]
>>> groebner(F, x, y, order=lex)
GroebnerBasis([x**2 - 2*y**2, x*y - 2*y, y**3 - 2*y], x, y,
domain=ZZ, order=lex)
>>> groebner(F, x, y, order=grlex)
GroebnerBasis([y**3 - 2*y, x**2 - 2*y**2, x*y - 2*y], x, y,
domain=ZZ, order=grlex)
>>> groebner(F, x, y, order=grevlex)
GroebnerBasis([y**3 - 2*y, x**2 - 2*y**2, x*y - 2*y], x, y,
domain=ZZ, order=grevlex)
By default, an improved implementation of the Buchberger algorithm is used. Optionally,
an implementation of the F5B algorithm can be used. The algorithm can be set using
method ag or with the setup() function from sympy.polys.polyconfig:
>>> F = [x**2 - x - 1, (2*x - 1) * y - (x**10 - (1 - x)**10)]
>>> groebner(F, x, y,
GroebnerBasis([x**2 >>> groebner(F, x, y,
GroebnerBasis([x**2 -
method=buchberger)
x - 1, y - 55], x, y, domain=ZZ, order=lex)
method=f5b)
x - 1, y - 55], x, y, domain=ZZ, order=lex)
sympy.polys.polytools.is zero dimensional(F, *gens, **args)
Checks if the ideal generated by a Groebner basis is zero-dimensional.
The algorithm checks if the set of monomials not divisible by the leading monomial of
any element of F is bounded.
1074
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
References
David A. Cox, John B. Little, Donal OShea. Ideals, Varieties and Algorithms, 3rd edition,
p. 230
class sympy.polys.polytools.Poly
Generic class for representing polynomial expressions.
EC(f, order=None)
Returns the last non-zero coecient of f.
Examples
>>> from sympy import Poly
>>> from sympy.abc import x
>>> Poly(x**3 + 2*x**2 + 3*x, x).EC()
3
EM(f, order=None)
Returns the last non-zero monomial of f.
Examples
>>> from sympy import Poly
>>> from sympy.abc import x, y
>>> Poly(4*x**2 + 2*x*y**2 + x*y + 3*y, x, y).EM()
x**0*y**1
ET(f, order=None)
Returns the last non-zero term of f.
Examples
>>> from sympy import Poly
>>> from sympy.abc import x, y
>>> Poly(4*x**2 + 2*x*y**2 + x*y + 3*y, x, y).ET()
(x**0*y**1, 3)
LC(f, order=None)
Returns the leading coecient of f.
Examples
>>> from sympy import Poly
>>> from sympy.abc import x
>>> Poly(4*x**3 + 2*x**2 + 3*x, x).LC()
4
5.16. Polynomials Manipulation Module
1075
SymPy Documentation, Release 0.7.6
LM(f, order=None)
Returns the leading monomial of f.
Examples
>>> from sympy import Poly
>>> from sympy.abc import x, y
>>> Poly(4*x**2 + 2*x*y**2 + x*y + 3*y, x, y).LM()
x**2*y**0
LT(f, order=None)
Returns the leading term of f.
Examples
>>> from sympy import Poly
>>> from sympy.abc import x, y
>>> Poly(4*x**2 + 2*x*y**2 + x*y + 3*y, x, y).LT()
(x**2*y**0, 4)
TC(f)
Returns the trailing coecient of f.
Examples
>>> from sympy import Poly
>>> from sympy.abc import x
>>> Poly(x**3 + 2*x**2 + 3*x, x).TC()
0
abs(f)
Make all coecients in f positive.
Examples
>>> from sympy import Poly
>>> from sympy.abc import x
>>> Poly(x**2 - 1, x).abs()
Poly(x**2 + 1, x, domain=ZZ)
add(f, g)
Add two polynomials f and g.
1076
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import Poly
>>> from sympy.abc import x
>>> Poly(x**2 + 1, x).add(Poly(x - 2, x))
Poly(x**2 + x - 1, x, domain=ZZ)
>>> Poly(x**2 + 1, x) + Poly(x - 2, x)
Poly(x**2 + x - 1, x, domain=ZZ)
add ground(f, coe)
Add an element of the ground domain to f.
Examples
>>> from sympy import Poly
>>> from sympy.abc import x
>>> Poly(x + 1).add_ground(2)
Poly(x + 3, x, domain=ZZ)
all coeffs(f)
Returns all coecients from a univariate polynomial f.
Examples
>>> from sympy import Poly
>>> from sympy.abc import x
>>> Poly(x**3 + 2*x - 1, x).all_coeffs()
[1, 0, 2, -1]
all monoms(f)
Returns all monomials from a univariate polynomial f.
See Also:
all terms (page 1078)
Examples
>>> from sympy import Poly
>>> from sympy.abc import x
>>> Poly(x**3 + 2*x - 1, x).all_monoms()
[(3,), (2,), (1,), (0,)]
all roots(f, multiple=True, radicals=True)
Return a list of real and complex roots with multiplicities.
5.16. Polynomials Manipulation Module
1077
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import Poly
>>> from sympy.abc import x
>>> Poly(2*x**3 - 7*x**2 + 4*x + 4).all_roots()
[-1/2, 2, 2]
>>> Poly(x**3 + x + 1).all_roots()
[RootOf(x**3 + x + 1, 0),
RootOf(x**3 + x + 1, 1),
RootOf(x**3 + x + 1, 2)]
all terms(f)
Returns all terms from a univariate polynomial f.
Examples
>>> from sympy import Poly
>>> from sympy.abc import x
>>> Poly(x**3 + 2*x - 1, x).all_terms()
[((3,), 1), ((2,), 0), ((1,), 2), ((0,), -1)]
args
Dont mess up with the core.
Examples
>>> from sympy import Poly
>>> from sympy.abc import x
>>> Poly(x**2 + 1, x).args
(x**2 + 1,)
as dict(f, native=False, zero=False)
Switch to a dict representation.
Examples
>>> from sympy import Poly
>>> from sympy.abc import x, y
>>> Poly(x**2 + 2*x*y**2 - y, x, y).as_dict()
{(0, 1): -1, (1, 2): 2, (2, 0): 1}
as expr(f, *gens)
Convert a Poly instance to an Expr instance.
1078
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import Poly
>>> from sympy.abc import x, y
>>> f = Poly(x**2 + 2*x*y**2 - y, x, y)
>>> f.as_expr()
x**2 + 2*x*y**2 - y
>>> f.as_expr({x: 5})
10*y**2 - y + 25
>>> f.as_expr(5, 6)
379
as list(f, native=False)
Switch to a list representation.
cancel(f, g, include=False)
Cancel common factors in a rational function f/g.
Examples
>>> from sympy import Poly
>>> from sympy.abc import x
>>> Poly(2*x**2 - 2, x).cancel(Poly(x**2 - 2*x + 1, x))
(1, Poly(2*x + 2, x, domain=ZZ), Poly(x - 1, x, domain=ZZ))
>>> Poly(2*x**2 - 2, x).cancel(Poly(x**2 - 2*x + 1, x), include=True)
(Poly(2*x + 2, x, domain=ZZ), Poly(x - 1, x, domain=ZZ))
clear denoms(f, convert=False)
Clear denominators, but keep the ground domain.
Examples
>>> from sympy import Poly, S, QQ
>>> from sympy.abc import x
>>> f = Poly(x/2 + S(1)/3, x, domain=QQ)
>>>
(6,
>>>
(6,
f.clear_denoms()
Poly(3*x + 2, x, domain=QQ))
f.clear_denoms(convert=True)
Poly(3*x + 2, x, domain=ZZ))
coeff monomial(f, monom)
Returns the coecient of monom in f if there, else None.
See Also:
nth (page 1100) more ecient query using exponents of the monomials generators
5.16. Polynomials Manipulation Module
1079
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import Poly, exp
>>> from sympy.abc import x, y
>>> p = Poly(24*x*y*exp(8) + 23*x, x, y)
>>> p.coeff_monomial(x)
23
>>> p.coeff_monomial(y)
0
>>> p.coeff_monomial(x*y)
24*exp(8)
Note that Expr.coeff() behaves dierently, collecting terms if possible; the Poly
must be converted to an Expr to use that method, however:
>>> p.as_expr().coeff(x)
24*y*exp(8) + 23
>>> p.as_expr().coeff(y)
24*x*exp(8)
>>> p.as_expr().coeff(x*y)
24*exp(8)
coeffs(f, order=None)
Returns all non-zero coecients from f in lex order.
See Also:
all coeffs (page 1077), coeff monomial (page 1079), nth (page 1100)
Examples
>>> from sympy import Poly
>>> from sympy.abc import x
>>> Poly(x**3 + 2*x + 3, x).coeffs()
[1, 2, 3]
cofactors(f, g)
Returns the GCD of f and g and their cofactors.
Returns polynomials (h, cff, cfg) such that h = gcd(f, g), and cff = quo(f,
h) and cfg = quo(g, h) are, so called, cofactors of f and g.
Examples
>>> from sympy import Poly
>>> from sympy.abc import x
>>> Poly(x**2 (Poly(x - 1, x,
Poly(x + 1, x,
Poly(x - 2, x,
1080
1, x).cofactors(Poly(x**2 - 3*x + 2, x))
domain=ZZ),
domain=ZZ),
domain=ZZ))
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
compose(f, g)
Computes the functional composition of f and g.
Examples
>>> from sympy import Poly
>>> from sympy.abc import x
>>> Poly(x**2 + x, x).compose(Poly(x - 1, x))
Poly(x**2 - x, x, domain=ZZ)
content(f)
Returns the GCD of polynomial coecients.
Examples
>>> from sympy import Poly
>>> from sympy.abc import x
>>> Poly(6*x**2 + 8*x + 12, x).content()
2
count roots(f, inf=None, sup=None)
Return the number of roots of f in [inf, sup] interval.
Examples
>>> from sympy import Poly, I
>>> from sympy.abc import x
>>> Poly(x**4 - 4, x).count_roots(-3, 3)
2
>>> Poly(x**4 - 4, x).count_roots(0, 1 + 3*I)
1
decompose(f)
Computes a functional decomposition of f.
Examples
>>> from sympy import Poly
>>> from sympy.abc import x
>>> Poly(x**4 + 2*x**3 - x - 1, x, domain=ZZ).decompose()
[Poly(x**2 - x - 1, x, domain=ZZ), Poly(x**2 + x, x, domain=ZZ)]
deflate(f)
Reduce degree of f by mapping x i**m to y i.
5.16. Polynomials Manipulation Module
1081
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import Poly
>>> from sympy.abc import x, y
>>> Poly(x**6*y**2 + x**3 + 1, x, y).deflate()
((3, 2), Poly(x**2*y + x + 1, x, y, domain=ZZ))
degree(f, gen=0)
Returns degree of f in x j.
The degree of 0 is negative innity.
Examples
>>> from sympy import Poly
>>> from sympy.abc import x, y
>>> Poly(x**2 + y*x + 1, x, y).degree()
2
>>> Poly(x**2 + y*x + y, x, y).degree(y)
1
>>> Poly(0, x).degree()
-oo
degree list(f)
Returns a list of degrees of f.
Examples
>>> from sympy import Poly
>>> from sympy.abc import x, y
>>> Poly(x**2 + y*x + 1, x, y).degree_list()
(2, 1)
diff(f, *specs)
Computes partial derivative of f.
Examples
>>> from sympy import Poly
>>> from sympy.abc import x, y
>>> Poly(x**2 + 2*x + 1, x).diff()
Poly(2*x + 2, x, domain=ZZ)
>>> Poly(x*y**2 + x, x, y).diff((0, 0), (1, 1))
Poly(2*x*y, x, y, domain=ZZ)
discriminant(f)
Computes the discriminant of f.
1082
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import Poly
>>> from sympy.abc import x
>>> Poly(x**2 + 2*x + 3, x).discriminant()
-8
dispersion(f, g=None)
Compute the dispersion of polynomials.
For two polynomials f (x) and g(x) with deg f > 0 and deg g > 0 the dispersion dis(f, g)
is dened as:
dis(f, g) := max{J(f, g) {0}}
= max{{a N| gcd(f (x), g(x + a)) 6= 1} {0}}
and for a single polynomial dis(f ) := dis(f, f ).
See Also:
dispersionset (page 1084)
References
1.[ManWright94] (page 1912)
2.[Koepf98] (page 1912)
3.[Abramov71] (page 1912)
4.[Man93] (page 1912)
[ManWright94] (page 1912), [Koepf98] (page 1912), [Abramov71] (page 1912),
[Man93] (page 1912)
Examples
>>> from sympy import poly
>>> from sympy.polys.dispersion import dispersion, dispersionset
>>> from sympy.abc import x
Dispersion set and dispersion of a simple polynomial:
>>>
>>>
[0,
>>>
6
fp = poly((x - 3)*(x + 3), x)
sorted(dispersionset(fp))
6]
dispersion(fp)
Note that the denition of the dispersion is not symmetric:
>>>
>>>
>>>
[2,
fp = poly(x**4 - 3*x**2 + 1, x)
gp = fp.shift(-3)
sorted(dispersionset(fp, gp))
3, 4]
5.16. Polynomials Manipulation Module
1083
SymPy Documentation, Release 0.7.6
>>> dispersion(fp, gp)
4
>>> sorted(dispersionset(gp, fp))
[]
>>> dispersion(gp, fp)
-oo
Computing the dispersion also works over eld extensions:
>>>
>>>
>>>
>>>
[2]
>>>
[1,
from sympy import sqrt
fp = poly(x**2 + sqrt(5)*x - 1, x, domain=QQ<sqrt(5)>)
gp = poly(x**2 + (2 + sqrt(5))*x + sqrt(5), x, domain=QQ<sqrt(5)>)
sorted(dispersionset(fp, gp))
sorted(dispersionset(gp, fp))
4]
We can even perform the computations for polynomials having symbolic coecients:
>>>
>>>
>>>
[0,
from sympy.abc import a
fp = poly(4*x**4 + (4*a + 8)*x**3 + (a**2 + 6*a + 4)*x**2 + (a**2 + 2*a)*x, x)
sorted(dispersionset(fp))
1]
dispersionset(f, g=None)
Compute the dispersion set of two polynomials.
For two polynomials f (x) and g(x) with deg f > 0 and deg g > 0 the dispersion set
J(f, g) is dened as:
J(f, g) := {a N0 | gcd(f (x), g(x + a)) 6= 1}
= {a N0 | deg gcd(f (x), g(x + a)) 1}
For a single polynomial one denes J(f ) := J(f, f ).
See Also:
dispersion (page 1083)
References
1.[ManWright94] (page 1912)
2.[Koepf98] (page 1912)
3.[Abramov71] (page 1912)
4.[Man93] (page 1912)
[ManWright94] (page 1912), [Koepf98] (page 1912), [Abramov71] (page 1912),
[Man93] (page 1912)
Examples
>>> from sympy import poly
>>> from sympy.polys.dispersion import dispersion, dispersionset
>>> from sympy.abc import x
1084
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Dispersion set and dispersion of a simple polynomial:
>>>
>>>
[0,
>>>
6
fp = poly((x - 3)*(x + 3), x)
sorted(dispersionset(fp))
6]
dispersion(fp)
Note that the denition of the dispersion is not symmetric:
>>>
>>>
>>>
[2,
>>>
4
>>>
[]
>>>
-oo
fp = poly(x**4 - 3*x**2 + 1, x)
gp = fp.shift(-3)
sorted(dispersionset(fp, gp))
3, 4]
dispersion(fp, gp)
sorted(dispersionset(gp, fp))
dispersion(gp, fp)
Computing the dispersion also works over eld extensions:
>>>
>>>
>>>
>>>
[2]
>>>
[1,
from sympy import sqrt
fp = poly(x**2 + sqrt(5)*x - 1, x, domain=QQ<sqrt(5)>)
gp = poly(x**2 + (2 + sqrt(5))*x + sqrt(5), x, domain=QQ<sqrt(5)>)
sorted(dispersionset(fp, gp))
sorted(dispersionset(gp, fp))
4]
We can even perform the computations for polynomials having symbolic coecients:
>>>
>>>
>>>
[0,
from sympy.abc import a
fp = poly(4*x**4 + (4*a + 8)*x**3 + (a**2 + 6*a + 4)*x**2 + (a**2 + 2*a)*x, x)
sorted(dispersionset(fp))
1]
div(f, g, auto=True)
Polynomial division with remainder of f by g.
Examples
>>> from sympy import Poly
>>> from sympy.abc import x
>>> Poly(x**2 + 1, x).div(Poly(2*x - 4, x))
(Poly(1/2*x + 1, x, domain=QQ), Poly(5, x, domain=QQ))
>>> Poly(x**2 + 1, x).div(Poly(2*x - 4, x), auto=False)
(Poly(0, x, domain=ZZ), Poly(x**2 + 1, x, domain=ZZ))
domain
Get the ground domain of self.
eject(f, *gens)
Eject selected generators into the ground domain.
5.16. Polynomials Manipulation Module
1085
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import Poly
>>> from sympy.abc import x, y
>>> f = Poly(x**2*y + x*y**3 + x*y + 1, x, y)
>>> f.eject(x)
Poly(x*y**3 + (x**2 + x)*y + 1, y, domain=ZZ[x])
>>> f.eject(y)
Poly(y*x**2 + (y**3 + y)*x + 1, x, domain=ZZ[y])
eval(f, x, a=None, auto=True)
Evaluate f at a in the given variable.
Examples
>>> from sympy import Poly
>>> from sympy.abc import x, y, z
>>> Poly(x**2 + 2*x + 3, x).eval(2)
11
>>> Poly(2*x*y + 3*x + y + 2, x, y).eval(x, 2)
Poly(5*y + 8, y, domain=ZZ)
>>> f = Poly(2*x*y + 3*x + y + 2*z, x, y, z)
>>> f.eval({x:
Poly(5*y + 2*z
>>> f.eval({x:
Poly(2*z + 31,
>>> f.eval({x:
45
2})
+ 6, y, z, domain=ZZ)
2, y: 5})
z, domain=ZZ)
2, y: 5, z: 7})
>>> f.eval((2, 5))
Poly(2*z + 31, z, domain=ZZ)
>>> f(2, 5)
Poly(2*z + 31, z, domain=ZZ)
exclude(f)
Remove unnecessary generators from f.
Examples
>>> from sympy import Poly
>>> from sympy.abc import a, b, c, d, x
>>> Poly(a + x, a, b, c, d, x).exclude()
Poly(a + x, a, x, domain=ZZ)
exquo(f, g, auto=True)
Computes polynomial exact quotient of f by g.
1086
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import Poly
>>> from sympy.abc import x
>>> Poly(x**2 - 1, x).exquo(Poly(x - 1, x))
Poly(x + 1, x, domain=ZZ)
>>> Poly(x**2 + 1, x).exquo(Poly(2*x - 4, x))
Traceback (most recent call last):
...
ExactQuotientFailed: 2*x - 4 does not divide x**2 + 1
Traceback (most recent call last):
...
ExactQuotientFailed: 2*x - 4 does not divide x**2 + 1
exquo ground(f, coe)
Exact quotient of f by a an element of the ground domain.
Examples
>>> from sympy import Poly
>>> from sympy.abc import x
>>> Poly(2*x + 4).exquo_ground(2)
Poly(x + 2, x, domain=ZZ)
>>> Poly(2*x + 3).exquo_ground(2)
Traceback (most recent call last):
...
ExactQuotientFailed: 2 does not divide 3 in ZZ
Traceback (most recent call last):
...
ExactQuotientFailed: 2 does not divide 3 in ZZ
factor list(f)
Returns a list of irreducible factors of f.
Examples
>>> from sympy import Poly
>>> from sympy.abc import x, y
>>> f = 2*x**5 + 2*x**4*y + 4*x**3 + 4*x**2*y + 2*x + 2*y
>>> Poly(f).factor_list()
(2, [(Poly(x + y, x, y, domain=ZZ), 1),
(Poly(x**2 + 1, x, y, domain=ZZ), 2)])
factor list include(f)
Returns a list of irreducible factors of f.
5.16. Polynomials Manipulation Module
1087
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import Poly
>>> from sympy.abc import x, y
>>> f = 2*x**5 + 2*x**4*y + 4*x**3 + 4*x**2*y + 2*x + 2*y
>>> Poly(f).factor_list_include()
[(Poly(2*x + 2*y, x, y, domain=ZZ), 1),
(Poly(x**2 + 1, x, y, domain=ZZ), 2)]
free symbols
Free symbols of a polynomial expression.
Examples
>>> from sympy import Poly
>>> from sympy.abc import x, y
>>> Poly(x**2 + 1).free_symbols
set([x])
>>> Poly(x**2 + y).free_symbols
set([x, y])
>>> Poly(x**2 + y, x).free_symbols
set([x, y])
free symbols in domain
Free symbols of the domain of self.
Examples
>>> from sympy import Poly
>>> from sympy.abc import x, y
>>> Poly(x**2 + 1).free_symbols_in_domain
set()
>>> Poly(x**2 + y).free_symbols_in_domain
set()
>>> Poly(x**2 + y, x).free_symbols_in_domain
set([y])
classmethod from dict(rep, *gens, **args)
Construct a polynomial from a dict.
classmethod from expr(rep, *gens, **args)
Construct a polynomial from an expression.
classmethod from list(rep, *gens, **args)
Construct a polynomial from a list.
classmethod from poly(rep, *gens, **args)
Construct a polynomial from a polynomial.
gcd(f, g)
Returns the polynomial GCD of f and g.
1088
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import Poly
>>> from sympy.abc import x
>>> Poly(x**2 - 1, x).gcd(Poly(x**2 - 3*x + 2, x))
Poly(x - 1, x, domain=ZZ)
gcdex(f, g, auto=True)
Extended Euclidean algorithm of f and g.
Returns (s, t, h) such that h = gcd(f, g) and s*f + t*g = h.
Examples
>>> from sympy import Poly
>>> from sympy.abc import x
>>> f = x**4 - 2*x**3 - 6*x**2 + 12*x + 15
>>> g = x**3 + x**2 - 4*x - 4
>>> Poly(f).gcdex(Poly(g))
(Poly(-1/5*x + 3/5, x, domain=QQ),
Poly(1/5*x**2 - 6/5*x + 2, x, domain=QQ),
Poly(x + 1, x, domain=QQ))
gen
Return the principal generator.
Examples
>>> from sympy import Poly
>>> from sympy.abc import x
>>> Poly(x**2 + 1, x).gen
x
get domain(f)
Get the ground domain of f.
get modulus(f)
Get the modulus of f.
Examples
>>> from sympy import Poly
>>> from sympy.abc import x
>>> Poly(x**2 + 1, modulus=2).get_modulus()
2
gff list(f)
Computes greatest factorial factorization of f.
5.16. Polynomials Manipulation Module
1089
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import Poly
>>> from sympy.abc import x
>>> f = x**5 + 2*x**4 - x**3 - 2*x**2
>>> Poly(f).gff_list()
[(Poly(x, x, domain=ZZ), 1), (Poly(x + 2, x, domain=ZZ), 4)]
ground roots(f)
Compute roots of f by factorization in the ground domain.
Examples
>>> from sympy import Poly
>>> from sympy.abc import x
>>> Poly(x**6 - 4*x**4 + 4*x**3 - x**2).ground_roots()
{0: 2, 1: 2}
half gcdex(f, g, auto=True)
Half extended Euclidean algorithm of f and g.
Returns (s, h) such that h = gcd(f, g) and s*f = h (mod g).
Examples
>>> from sympy import Poly
>>> from sympy.abc import x
>>> f = x**4 - 2*x**3 - 6*x**2 + 12*x + 15
>>> g = x**3 + x**2 - 4*x - 4
>>> Poly(f).half_gcdex(Poly(g))
(Poly(-1/5*x + 3/5, x, domain=QQ), Poly(x + 1, x, domain=QQ))
has only gens(f, *gens)
Return True if Poly(f, *gens) retains ground domain.
Examples
>>> from sympy import Poly
>>> from sympy.abc import x, y, z
>>> Poly(x*y + 1, x, y, z).has_only_gens(x, y)
True
>>> Poly(x*y + z, x, y, z).has_only_gens(x, y)
False
homogeneous order(f)
Returns the homogeneous order of f.
1090
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
A homogeneous polynomial is a polynomial whose all monomials with non-zero
coecients have the same total degree. This degree is the homogeneous order of f. If you only want to check if a polynomial is homogeneous, then use
Poly.is homogeneous() (page 1093).
Examples
>>> from sympy import Poly
>>> from sympy.abc import x, y
>>> f = Poly(x**5 + 2*x**3*y**2 + 9*x*y**4)
>>> f.homogeneous_order()
5
homogenize(f, s)
Returns the homogeneous polynomial of f.
A homogeneous polynomial is a polynomial whose all monomials with non-zero coecients have the same total degree. If you only want to check if a polynomial is
homogeneous, then use Poly.is homogeneous() (page 1093). If you want not only
to check if a polynomial is homogeneous but also compute its homogeneous order,
then use Poly.homogeneous order() (page 1090).
Examples
>>> from sympy import Poly
>>> from sympy.abc import x, y, z
>>> f = Poly(x**5 + 2*x**2*y**2 + 9*x*y**3)
>>> f.homogenize(z)
Poly(x**5 + 2*x**2*y**2*z + 9*x*y**3*z, x, y, z, domain=ZZ)
inject(f, front=False)
Inject ground domain generators into f.
Examples
>>> from sympy import Poly
>>> from sympy.abc import x, y
>>> f = Poly(x**2*y + x*y**3 + x*y + 1, x)
>>> f.inject()
Poly(x**2*y + x*y**3 + x*y + 1, x, y, domain=ZZ)
>>> f.inject(front=True)
Poly(y**3*x + y*x**2 + y*x + 1, y, x, domain=ZZ)
integrate(f, *specs, **args)
Computes indenite integral of f.
5.16. Polynomials Manipulation Module
1091
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import Poly
>>> from sympy.abc import x, y
>>> Poly(x**2 + 2*x + 1, x).integrate()
Poly(1/3*x**3 + x**2 + x, x, domain=QQ)
>>> Poly(x*y**2 + x, x, y).integrate((0, 1), (1, 0))
Poly(1/2*x**2*y**2 + 1/2*x**2, x, y, domain=QQ)
intervals(f, all=False, eps=None, inf=None, sup=None, fast=False, sqf=False)
Compute isolating intervals for roots of f.
For real roots the Vincent-Akritas-Strzebonski (VAS) continued fractions method is
used.
Examples
>>> from sympy import Poly
>>> from sympy.abc import x
>>> Poly(x**2 - 3, x).intervals()
[((-2, -1), 1), ((1, 2), 1)]
>>> Poly(x**2 - 3, x).intervals(eps=1e-2)
[((-26/15, -19/11), 1), ((19/11, 26/15), 1)]
References:
1. Alkiviadis G. Akritas and Adam W. Strzebonski: A Comparative Study of Two Real
Root Isolation Methods . Nonlinear Analysis: Modelling and Control, Vol. 10, No.
4, 297-304, 2005. 2. Alkiviadis G. Akritas, Adam W. Strzebonski and Panagiotis S.
Vigklas: Improving the Performance of the Continued Fractions Method Using new
Bounds of Positive Roots. Nonlinear Analysis: Modelling and Control, Vol. 13, No.
3, 265-279, 2008.
invert(f, g, auto=True)
Invert f modulo g when possible.
Examples
>>> from sympy import Poly
>>> from sympy.abc import x
>>> Poly(x**2 - 1, x).invert(Poly(2*x - 1, x))
Poly(-4/3, x, domain=QQ)
>>> Poly(x**2 - 1, x).invert(Poly(x - 1, x))
Traceback (most recent call last):
...
NotInvertible: zero divisor
Traceback (most recent call last):
1092
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
...
NotInvertible: zero divisor
is cyclotomic
Returns True if f is a cyclotomic polnomial.
Examples
>>> from sympy import Poly
>>> from sympy.abc import x
>>> f = x**16 + x**14 - x**10 + x**8 - x**6 + x**2 + 1
>>> Poly(f).is_cyclotomic
False
>>> g = x**16 + x**14 - x**10 - x**8 - x**6 + x**2 + 1
>>> Poly(g).is_cyclotomic
True
is ground
Returns True if f is an element of the ground domain.
Examples
>>> from sympy import Poly
>>> from sympy.abc import x, y
>>> Poly(x, x).is_ground
False
>>> Poly(2, x).is_ground
True
>>> Poly(y, x).is_ground
True
is homogeneous
Returns True if f is a homogeneous polynomial.
A homogeneous polynomial is a polynomial whose all monomials with non-zero
coecients have the same total degree. If you want not only to check if a
polynomial is homogeneous but also compute its homogeneous order, then use
Poly.homogeneous order() (page 1090).
Examples
>>> from sympy import Poly
>>> from sympy.abc import x, y
>>> Poly(x**2 + x*y, x, y).is_homogeneous
True
>>> Poly(x**3 + x*y, x, y).is_homogeneous
False
5.16. Polynomials Manipulation Module
1093
SymPy Documentation, Release 0.7.6
is irreducible
Returns True if f has no factors over its domain.
Examples
>>> from sympy import Poly
>>> from sympy.abc import x
>>> Poly(x**2 + x + 1, x, modulus=2).is_irreducible
True
>>> Poly(x**2 + 1, x, modulus=2).is_irreducible
False
is linear
Returns True if f is linear in all its variables.
Examples
>>> from sympy import Poly
>>> from sympy.abc import x, y
>>> Poly(x + y + 2, x, y).is_linear
True
>>> Poly(x*y + 2, x, y).is_linear
False
is monic
Returns True if the leading coecient of f is one.
Examples
>>> from sympy import Poly
>>> from sympy.abc import x
>>> Poly(x + 2, x).is_monic
True
>>> Poly(2*x + 2, x).is_monic
False
is monomial
Returns True if f is zero or has only one term.
Examples
>>> from sympy import Poly
>>> from sympy.abc import x
>>> Poly(3*x**2, x).is_monomial
True
>>> Poly(3*x**2 + 1, x).is_monomial
False
1094
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
is multivariate
Returns True if f is a multivariate polynomial.
Examples
>>> from sympy import Poly
>>> from sympy.abc import x, y
>>> Poly(x**2 +
False
>>> Poly(x*y**2
True
>>> Poly(x*y**2
False
>>> Poly(x**2 +
True
x + 1, x).is_multivariate
+ x*y + 1, x, y).is_multivariate
+ x*y + 1, x).is_multivariate
x + 1, x, y).is_multivariate
is one
Returns True if f is a unit polynomial.
Examples
>>> from sympy import Poly
>>> from sympy.abc import x
>>> Poly(0, x).is_one
False
>>> Poly(1, x).is_one
True
is primitive
Returns True if GCD of the coecients of f is one.
Examples
>>> from sympy import Poly
>>> from sympy.abc import x
>>> Poly(2*x**2 + 6*x + 12, x).is_primitive
False
>>> Poly(x**2 + 3*x + 6, x).is_primitive
True
is quadratic
Returns True if f is quadratic in all its variables.
Examples
>>> from sympy import Poly
>>> from sympy.abc import x, y
5.16. Polynomials Manipulation Module
1095
SymPy Documentation, Release 0.7.6
>>> Poly(x*y + 2, x, y).is_quadratic
True
>>> Poly(x*y**2 + 2, x, y).is_quadratic
False
is sqf
Returns True if f is a square-free polynomial.
Examples
>>> from sympy import Poly
>>> from sympy.abc import x
>>> Poly(x**2 - 2*x + 1, x).is_sqf
False
>>> Poly(x**2 - 1, x).is_sqf
True
is univariate
Returns True if f is a univariate polynomial.
Examples
>>> from sympy import Poly
>>> from sympy.abc import x, y
>>> Poly(x**2 +
True
>>> Poly(x*y**2
False
>>> Poly(x*y**2
True
>>> Poly(x**2 +
False
x + 1, x).is_univariate
+ x*y + 1, x, y).is_univariate
+ x*y + 1, x).is_univariate
x + 1, x, y).is_univariate
is zero
Returns True if f is a zero polynomial.
Examples
>>> from sympy import Poly
>>> from sympy.abc import x
>>> Poly(0, x).is_zero
True
>>> Poly(1, x).is_zero
False
l1 norm(f)
Returns l1 norm of f.
1096
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import Poly
>>> from sympy.abc import x
>>> Poly(-x**2 + 2*x - 3, x).l1_norm()
6
lcm(f, g)
Returns polynomial LCM of f and g.
Examples
>>> from sympy import Poly
>>> from sympy.abc import x
>>> Poly(x**2 - 1, x).lcm(Poly(x**2 - 3*x + 2, x))
Poly(x**3 - 2*x**2 - x + 2, x, domain=ZZ)
length(f)
Returns the number of non-zero terms in f.
Examples
>>> from sympy import Poly
>>> from sympy.abc import x
>>> Poly(x**2 + 2*x - 1).length()
3
lift(f)
Convert algebraic coecients to rationals.
Examples
>>> from sympy import Poly, I
>>> from sympy.abc import x
>>> Poly(x**2 + I*x + 1, x, extension=I).lift()
Poly(x**4 + 3*x**2 + 1, x, domain=QQ)
ltrim(f, gen)
Remove dummy generators from the left of f.
Examples
>>> from sympy import Poly
>>> from sympy.abc import x, y, z
5.16. Polynomials Manipulation Module
1097
SymPy Documentation, Release 0.7.6
>>> Poly(y**2 + y*z**2, x, y, z).ltrim(y)
Poly(y**2 + y*z**2, y, z, domain=ZZ)
max norm(f)
Returns maximum norm of f.
Examples
>>> from sympy import Poly
>>> from sympy.abc import x
>>> Poly(-x**2 + 2*x - 3, x).max_norm()
3
monic(f, auto=True)
Divides all coecients by LC(f).
Examples
>>> from sympy import Poly, ZZ
>>> from sympy.abc import x
>>> Poly(3*x**2 + 6*x + 9, x, domain=ZZ).monic()
Poly(x**2 + 2*x + 3, x, domain=QQ)
>>> Poly(3*x**2 + 4*x + 2, x, domain=ZZ).monic()
Poly(x**2 + 4/3*x + 2/3, x, domain=QQ)
monoms(f, order=None)
Returns all non-zero monomials from f in lex order.
See Also:
all monoms (page 1077)
Examples
>>> from sympy import Poly
>>> from sympy.abc import x, y
>>> Poly(x**2 + 2*x*y**2 + x*y + 3*y, x, y).monoms()
[(2, 0), (1, 2), (1, 1), (0, 1)]
mul(f, g)
Multiply two polynomials f and g.
Examples
>>> from sympy import Poly
>>> from sympy.abc import x
1098
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> Poly(x**2 + 1, x).mul(Poly(x - 2, x))
Poly(x**3 - 2*x**2 + x - 2, x, domain=ZZ)
>>> Poly(x**2 + 1, x)*Poly(x - 2, x)
Poly(x**3 - 2*x**2 + x - 2, x, domain=ZZ)
mul ground(f, coe)
Multiply f by a an element of the ground domain.
Examples
>>> from sympy import Poly
>>> from sympy.abc import x
>>> Poly(x + 1).mul_ground(2)
Poly(2*x + 2, x, domain=ZZ)
neg(f)
Negate all coecients in f.
Examples
>>> from sympy import Poly
>>> from sympy.abc import x
>>> Poly(x**2 - 1, x).neg()
Poly(-x**2 + 1, x, domain=ZZ)
>>> -Poly(x**2 - 1, x)
Poly(-x**2 + 1, x, domain=ZZ)
classmethod new(rep, *gens)
Construct Poly (page 1075) instance from raw representation.
nroots(f, n=15, maxsteps=50, cleanup=True)
Compute numerical approximations of roots of f.
Parameters n ... the number of digits to calculate :
maxsteps ... the maximum number of iterations to do :
If the accuracy n cannot be reached in maxsteps, it will raise an
:
exception. You need to rerun with higher maxsteps. :
Examples
>>> from sympy import Poly
>>> from sympy.abc import x
5.16. Polynomials Manipulation Module
1099
SymPy Documentation, Release 0.7.6
>>> Poly(x**2 - 3).nroots(n=15)
[-1.73205080756888, 1.73205080756888]
>>> Poly(x**2 - 3).nroots(n=30)
[-1.73205080756887729352744634151, 1.73205080756887729352744634151]
nth(f, *N)
Returns the n-th coecient of f where N are the exponents of the generators in the
term of interest.
See Also:
coeff monomial (page 1079)
Examples
>>> from sympy import Poly, sqrt
>>> from sympy.abc import x, y
>>> Poly(x**3 + 2*x**2 + 3*x, x).nth(2)
2
>>> Poly(x**3 + 2*x*y**2 + y**2, x, y).nth(1, 2)
2
>>> Poly(4*sqrt(x)*y)
Poly(4*y*sqrt(x), y, sqrt(x), domain=ZZ)
>>> _.nth(1, 1)
4
nth power roots poly(f, n)
Construct a polynomial with n-th powers of roots of f.
Examples
>>> from sympy import Poly
>>> from sympy.abc import x
>>> f = Poly(x**4 - x**2 + 1)
>>> f.nth_power_roots_poly(2)
Poly(x**4 - 2*x**3 + 3*x**2 - 2*x + 1, x, domain=ZZ)
>>> f.nth_power_roots_poly(3)
Poly(x**4 + 2*x**2 + 1, x, domain=ZZ)
>>> f.nth_power_roots_poly(4)
Poly(x**4 + 2*x**3 + 3*x**2 + 2*x + 1, x, domain=ZZ)
>>> f.nth_power_roots_poly(12)
Poly(x**4 - 4*x**3 + 6*x**2 - 4*x + 1, x, domain=ZZ)
one
Return one polynomial with selfs properties.
pdiv(f, g)
Polynomial pseudo-division of f by g.
1100
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import Poly
>>> from sympy.abc import x
>>> Poly(x**2 + 1, x).pdiv(Poly(2*x - 4, x))
(Poly(2*x + 4, x, domain=ZZ), Poly(20, x, domain=ZZ))
per(f, rep, gens=None, remove=None)
Create a Poly out of the given representation.
Examples
>>> from sympy import Poly, ZZ
>>> from sympy.abc import x, y
>>> from sympy.polys.polyclasses import DMP
>>> a = Poly(x**2 + 1)
>>> a.per(DMP([ZZ(1), ZZ(1)], ZZ), gens=[y])
Poly(y + 1, y, domain=ZZ)
pexquo(f, g)
Polynomial exact pseudo-quotient of f by g.
Examples
>>> from sympy import Poly
>>> from sympy.abc import x
>>> Poly(x**2 - 1, x).pexquo(Poly(2*x - 2, x))
Poly(2*x + 2, x, domain=ZZ)
>>> Poly(x**2 + 1, x).pexquo(Poly(2*x - 4, x))
Traceback (most recent call last):
...
ExactQuotientFailed: 2*x - 4 does not divide x**2 + 1
Traceback (most recent call last):
...
ExactQuotientFailed: 2*x - 4 does not divide x**2 + 1
pow(f, n)
Raise f to a non-negative power n.
Examples
>>> from sympy import Poly
>>> from sympy.abc import x
5.16. Polynomials Manipulation Module
1101
SymPy Documentation, Release 0.7.6
>>> Poly(x - 2, x).pow(3)
Poly(x**3 - 6*x**2 + 12*x - 8, x, domain=ZZ)
>>> Poly(x - 2, x)**3
Poly(x**3 - 6*x**2 + 12*x - 8, x, domain=ZZ)
pquo(f, g)
Polynomial pseudo-quotient of f by g.
Examples
>>> from sympy import Poly
>>> from sympy.abc import x
>>> Poly(x**2 + 1, x).pquo(Poly(2*x - 4, x))
Poly(2*x + 4, x, domain=ZZ)
>>> Poly(x**2 - 1, x).pquo(Poly(2*x - 2, x))
Poly(2*x + 2, x, domain=ZZ)
prem(f, g)
Polynomial pseudo-remainder of f by g.
Examples
>>> from sympy import Poly
>>> from sympy.abc import x
>>> Poly(x**2 + 1, x).prem(Poly(2*x - 4, x))
Poly(20, x, domain=ZZ)
primitive(f)
Returns the content and a primitive form of f.
Examples
>>> from sympy import Poly
>>> from sympy.abc import x
>>> Poly(2*x**2 + 8*x + 12, x).primitive()
(2, Poly(x**2 + 4*x + 6, x, domain=ZZ))
quo(f, g, auto=True)
Computes polynomial quotient of f by g.
Examples
>>> from sympy import Poly
>>> from sympy.abc import x
1102
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> Poly(x**2 + 1, x).quo(Poly(2*x - 4, x))
Poly(1/2*x + 1, x, domain=QQ)
>>> Poly(x**2 - 1, x).quo(Poly(x - 1, x))
Poly(x + 1, x, domain=ZZ)
quo ground(f, coe)
Quotient of f by a an element of the ground domain.
Examples
>>> from sympy import Poly
>>> from sympy.abc import x
>>> Poly(2*x + 4).quo_ground(2)
Poly(x + 2, x, domain=ZZ)
>>> Poly(2*x + 3).quo_ground(2)
Poly(x + 1, x, domain=ZZ)
rat clear denoms(f, g)
Clear denominators in a rational function f/g.
Examples
>>> from sympy import Poly
>>> from sympy.abc import x, y
>>> f = Poly(x**2/y + 1, x)
>>> g = Poly(x**3 + y, x)
>>> p, q = f.rat_clear_denoms(g)
>>> p
Poly(x**2 + y, x, domain=ZZ[y])
>>> q
Poly(y*x**3 + y**2, x, domain=ZZ[y])
real roots(f, multiple=True, radicals=True)
Return a list of real roots with multiplicities.
Examples
>>> from sympy import Poly
>>> from sympy.abc import x
>>> Poly(2*x**3 - 7*x**2 + 4*x + 4).real_roots()
[-1/2, 2, 2]
>>> Poly(x**3 + x + 1).real_roots()
[RootOf(x**3 + x + 1, 0)]
refine root(f, s, t, eps=None, steps=None, fast=False, check sqf=False)
Rene an isolating interval of a root to the given precision.
5.16. Polynomials Manipulation Module
1103
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import Poly
>>> from sympy.abc import x
>>> Poly(x**2 - 3, x).refine_root(1, 2, eps=1e-2)
(19/11, 26/15)
rem(f, g, auto=True)
Computes the polynomial remainder of f by g.
Examples
>>> from sympy import Poly
>>> from sympy.abc import x
>>> Poly(x**2 + 1, x).rem(Poly(2*x - 4, x))
Poly(5, x, domain=ZZ)
>>> Poly(x**2 + 1, x).rem(Poly(2*x - 4, x), auto=False)
Poly(x**2 + 1, x, domain=ZZ)
reorder(f, *gens, **args)
Eciently apply new order of generators.
Examples
>>> from sympy import Poly
>>> from sympy.abc import x, y
>>> Poly(x**2 + x*y**2, x, y).reorder(y, x)
Poly(y**2*x + x**2, y, x, domain=ZZ)
replace(f, x, y=None)
Replace x with y in generators list.
Examples
>>> from sympy import Poly
>>> from sympy.abc import x, y
>>> Poly(x**2 + 1, x).replace(x, y)
Poly(y**2 + 1, y, domain=ZZ)
resultant(f, g, includePRS=False)
Computes the resultant of f and g via PRS.
If includePRS=True, it includes the subresultant PRS in the result. Because the PRS
is used to calculate the resultant, this is more ecient than calling subresultants()
(page 1061) separately.
1104
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import Poly
>>> from sympy.abc import x
>>> f = Poly(x**2 + 1, x)
>>> f.resultant(Poly(x**2 - 1, x))
4
>>> f.resultant(Poly(x**2 - 1, x), includePRS=True)
(4, [Poly(x**2 + 1, x, domain=ZZ), Poly(x**2 - 1, x, domain=ZZ),
Poly(-2, x, domain=ZZ)])
retract(f, eld=None)
Recalculate the ground domain of a polynomial.
Examples
>>> from sympy import Poly
>>> from sympy.abc import x
>>> f = Poly(x**2 + 1, x, domain=QQ[y])
>>> f
Poly(x**2 + 1, x, domain=QQ[y])
>>> f.retract()
Poly(x**2 + 1, x, domain=ZZ)
>>> f.retract(field=True)
Poly(x**2 + 1, x, domain=QQ)
revert(f, n)
Compute f**(-1) mod x**n.
root(f, index, radicals=True)
Get an indexed root of a polynomial.
Examples
>>> from sympy import Poly
>>> from sympy.abc import x
>>> f = Poly(2*x**3 - 7*x**2 + 4*x + 4)
>>> f.root(0)
-1/2
>>> f.root(1)
2
>>> f.root(2)
2
>>> f.root(3)
Traceback (most recent call last):
...
IndexError: root index out of [-3, 2] range, got 3
5.16. Polynomials Manipulation Module
1105
SymPy Documentation, Release 0.7.6
Traceback (most recent call last):
...
IndexError: root index out of [-3, 2] range, got 3
>>> Poly(x**5 + x + 1).root(0)
RootOf(x**3 - x**2 + 1, 0)
set domain(f, domain)
Set the ground domain of f.
set modulus(f, modulus)
Set the modulus of f.
Examples
>>> from sympy import Poly
>>> from sympy.abc import x
>>> Poly(5*x**2 + 2*x - 1, x).set_modulus(2)
Poly(x**2 + 1, x, modulus=2)
shift(f, a)
Eciently compute Taylor shift f(x + a).
Examples
>>> from sympy import Poly
>>> from sympy.abc import x
>>> Poly(x**2 - 2*x + 1, x).shift(2)
Poly(x**2 + 2*x + 1, x, domain=ZZ)
slice(f, x, m, n=None)
Take a continuous subsequence of terms of f.
sqf list(f, all=False)
Returns a list of square-free factors of f.
Examples
>>> from sympy import Poly
>>> from sympy.abc import x
>>> f = 2*x**5 + 16*x**4 + 50*x**3 + 76*x**2 + 56*x + 16
>>> Poly(f).sqf_list()
(2, [(Poly(x + 1, x, domain=ZZ), 2),
(Poly(x + 2, x, domain=ZZ), 3)])
>>> Poly(f).sqf_list(all=True)
(2, [(Poly(1, x, domain=ZZ), 1),
(Poly(x + 1, x, domain=ZZ), 2),
(Poly(x + 2, x, domain=ZZ), 3)])
1106
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
sqf list include(f, all=False)
Returns a list of square-free factors of f.
Examples
>>> from sympy import Poly, expand
>>> from sympy.abc import x
>>> f = expand(2*(x + 1)**3*x**4)
>>> f
2*x**7 + 6*x**6 + 6*x**5 + 2*x**4
>>> Poly(f).sqf_list_include()
[(Poly(2, x, domain=ZZ), 1),
(Poly(x + 1, x, domain=ZZ), 3),
(Poly(x, x, domain=ZZ), 4)]
>>> Poly(f).sqf_list_include(all=True)
[(Poly(2, x, domain=ZZ), 1),
(Poly(1, x, domain=ZZ), 2),
(Poly(x + 1, x, domain=ZZ), 3),
(Poly(x, x, domain=ZZ), 4)]
sqf norm(f)
Computes square-free norm of f.
Returns s, f, r, such that g(x) = f(x-sa) and r(x) = Norm(g(x)) is a square-free
polynomial over K, where a is the algebraic extension of the ground domain.
Examples
>>> from sympy import Poly, sqrt
>>> from sympy.abc import x
>>> s, f, r = Poly(x**2 + 1, x, extension=[sqrt(3)]).sqf_norm()
>>> s
1
>>> f
Poly(x**2 - 2*sqrt(3)*x + 4, x, domain=QQ<sqrt(3)>)
>>> r
Poly(x**4 - 4*x**2 + 16, x, domain=QQ)
sqf part(f)
Computes square-free part of f.
Examples
>>> from sympy import Poly
>>> from sympy.abc import x
5.16. Polynomials Manipulation Module
1107
SymPy Documentation, Release 0.7.6
>>> Poly(x**3 - 3*x - 2, x).sqf_part()
Poly(x**2 - x - 2, x, domain=ZZ)
sqr(f)
Square a polynomial f.
Examples
>>> from sympy import Poly
>>> from sympy.abc import x
>>> Poly(x - 2, x).sqr()
Poly(x**2 - 4*x + 4, x, domain=ZZ)
>>> Poly(x - 2, x)**2
Poly(x**2 - 4*x + 4, x, domain=ZZ)
sturm(f, auto=True)
Computes the Sturm sequence of f.
Examples
>>> from sympy import Poly
>>> from sympy.abc import x
>>> Poly(x**3 - 2*x**2 + x - 3, x).sturm()
[Poly(x**3 - 2*x**2 + x - 3, x, domain=QQ),
Poly(3*x**2 - 4*x + 1, x, domain=QQ),
Poly(2/9*x + 25/9, x, domain=QQ),
Poly(-2079/4, x, domain=QQ)]
sub(f, g)
Subtract two polynomials f and g.
Examples
>>> from sympy import Poly
>>> from sympy.abc import x
>>> Poly(x**2 + 1, x).sub(Poly(x - 2, x))
Poly(x**2 - x + 3, x, domain=ZZ)
>>> Poly(x**2 + 1, x) - Poly(x - 2, x)
Poly(x**2 - x + 3, x, domain=ZZ)
sub ground(f, coe)
Subtract an element of the ground domain from f.
1108
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import Poly
>>> from sympy.abc import x
>>> Poly(x + 1).sub_ground(2)
Poly(x - 1, x, domain=ZZ)
subresultants(f, g)
Computes the subresultant PRS of f and g.
Examples
>>> from sympy import Poly
>>> from sympy.abc import x
>>> Poly(x**2 + 1, x).subresultants(Poly(x**2 - 1, x))
[Poly(x**2 + 1, x, domain=ZZ),
Poly(x**2 - 1, x, domain=ZZ),
Poly(-2, x, domain=ZZ)]
terms(f, order=None)
Returns all non-zero terms from f in lex order.
See Also:
all terms (page 1078)
Examples
>>> from sympy import Poly
>>> from sympy.abc import x, y
>>> Poly(x**2 + 2*x*y**2 + x*y + 3*y, x, y).terms()
[((2, 0), 1), ((1, 2), 2), ((1, 1), 1), ((0, 1), 3)]
terms gcd(f)
Remove GCD of terms from the polynomial f.
Examples
>>> from sympy import Poly
>>> from sympy.abc import x, y
>>> Poly(x**6*y**2 + x**3*y, x, y).terms_gcd()
((3, 1), Poly(x**3*y + 1, x, y, domain=ZZ))
termwise(f, func, *gens, **args)
Apply a function to all terms of f.
5.16. Polynomials Manipulation Module
1109
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import Poly
>>> from sympy.abc import x
>>> def func(k, coeff):
...
k = k[0]
...
return coeff//10**(2-k)
>>> Poly(x**2 + 20*x + 400).termwise(func)
Poly(x**2 + 2*x + 4, x, domain=ZZ)
to exact(f)
Make the ground domain exact.
Examples
>>> from sympy import Poly, RR
>>> from sympy.abc import x
>>> Poly(x**2 + 1.0, x, domain=RR).to_exact()
Poly(x**2 + 1, x, domain=QQ)
to field(f)
Make the ground domain a eld.
Examples
>>> from sympy import Poly, ZZ
>>> from sympy.abc import x
>>> Poly(x**2 + 1, x, domain=ZZ).to_field()
Poly(x**2 + 1, x, domain=QQ)
to ring(f)
Make the ground domain a ring.
Examples
>>> from sympy import Poly, QQ
>>> from sympy.abc import x
>>> Poly(x**2 + 1, domain=QQ).to_ring()
Poly(x**2 + 1, x, domain=ZZ)
total degree(f)
Returns the total degree of f.
1110
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import Poly
>>> from sympy.abc import x, y
>>> Poly(x**2 + y*x + 1, x, y).total_degree()
2
>>> Poly(x + y**5, x, y).total_degree()
5
trunc(f, p)
Reduce f modulo a constant p.
Examples
>>> from sympy import Poly
>>> from sympy.abc import x
>>> Poly(2*x**3 + 3*x**2 + 5*x + 7, x).trunc(3)
Poly(-x**3 - x + 1, x, domain=ZZ)
unify(f, g)
Make f and g belong to the same domain.
Examples
>>> from sympy import Poly
>>> from sympy.abc import x
>>> f, g = Poly(x/2 + 1), Poly(2*x + 1)
>>> f
Poly(1/2*x + 1, x, domain=QQ)
>>> g
Poly(2*x + 1, x, domain=ZZ)
>>> F, G = f.unify(g)
>>> F
Poly(1/2*x + 1, x, domain=QQ)
>>> G
Poly(2*x + 1, x, domain=QQ)
unit
Return unit polynomial with selfs properties.
zero
Return zero polynomial with selfs properties.
class sympy.polys.polytools.PurePoly
Class for representing pure polynomials.
free symbols
Free symbols of a polynomial.
5.16. Polynomials Manipulation Module
1111
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import PurePoly
>>> from sympy.abc import x, y
>>> PurePoly(x**2 + 1).free_symbols
set()
>>> PurePoly(x**2 + y).free_symbols
set()
>>> PurePoly(x**2 + y, x).free_symbols
set([y])
class sympy.polys.polytools.GroebnerBasis
Represents a reduced Groebner basis.
contains(poly)
Check if poly belongs the ideal generated by self.
Examples
>>> from sympy import groebner
>>> from sympy.abc import x, y
>>> f = 2*x**3 + y**3 + 3*y
>>> G = groebner([x**2 + y**2 - 1, x*y - 2])
>>> G.contains(f)
True
>>> G.contains(f + 1)
False
fglm(order)
Convert a Groebner basis from one ordering to another.
The FGLM algorithm converts reduced Groebner bases of zero-dimensional ideals
from one ordering to another. This method is often used when it is infeasible to
compute a Groebner basis with respect to a particular ordering directly.
References
J.C. Faugere, P. Gianni, D. Lazard, T. Mora (1994). Ecient Computation of Zerodimensional Groebner Bases by Change of Ordering
Examples
>>> from sympy.abc import x, y
>>> from sympy import groebner
>>> F = [x**2 - 3*y - x + 1, y**2 - 2*x + y - 1]
>>> G = groebner(F, x, y, order=grlex)
1112
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> list(G.fglm(lex))
[2*x - y**2 - y + 1, y**4 + 2*y**3 - 3*y**2 - 16*y + 7]
>>> list(groebner(F, x, y, order=lex))
[2*x - y**2 - y + 1, y**4 + 2*y**3 - 3*y**2 - 16*y + 7]
is zero dimensional
Checks if the ideal generated by a Groebner basis is zero-dimensional.
The algorithm checks if the set of monomials not divisible by the leading monomial
of any element of F is bounded.
References
David A. Cox, John B. Little, Donal OShea. Ideals, Varieties and Algorithms, 3rd
edition, p. 230
reduce(expr, auto=True)
Reduces a polynomial modulo a Groebner basis.
Given a polynomial f and a set of polynomials G = (g 1, ..., g n), computes a
set of quotients q = (q 1, ..., q n) and the remainder r such that f = q 1*f 1 +
... + q n*f n + r, where r vanishes or r is a completely reduced polynomial with
respect to G.
Examples
>>> from sympy import groebner, expand
>>> from sympy.abc import x, y
>>> f = 2*x**4 - x**2 + y**3 + y**2
>>> G = groebner([x**3 - x, y**3 - y])
>>> G.reduce(f)
([2*x, 1], x**2 + y**2 + y)
>>> Q, r = _
>>> expand(sum(q*g for q, g in zip(Q, G)) + r)
2*x**4 - x**2 + y**3 + y**2
>>> _ == f
True
Extra polynomial manipulation functions
sympy.polys.polyfuncs.symmetrize(F, *gens, **args)
Rewrite a polynomial in terms of elementary symmetric polynomials.
A symmetric polynomial is a multivariate polynomial that remains invariant under any
variable permutation, i.e., if f = f(x 1, x 2, ..., x n), then f = f(x {i 1}, x {i 2},
..., x {i n}), where (i 1, i 2, ..., i n) is a permutation of (1, 2, ..., n) (an
element of the group S n).
Returns a tuple of symmetric polynomials (f1, f2, ..., fn) such that f = f1 + f2 +
... + fn.
5.16. Polynomials Manipulation Module
1113
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.polys.polyfuncs import symmetrize
>>> from sympy.abc import x, y
>>> symmetrize(x**2 + y**2)
(-2*x*y + (x + y)**2, 0)
>>> symmetrize(x**2 + y**2, formal=True)
(s1**2 - 2*s2, 0, [(s1, x + y), (s2, x*y)])
>>> symmetrize(x**2 - y**2)
(-2*x*y + (x + y)**2, -2*y**2)
>>> symmetrize(x**2 - y**2, formal=True)
(s1**2 - 2*s2, -2*y**2, [(s1, x + y), (s2, x*y)])
sympy.polys.polyfuncs.horner(f, *gens, **args)
Rewrite a polynomial in Horner form.
Among other applications, evaluation of a polynomial at a point is optimal when it is
applied using the Horner scheme ([1]).
References
[1] - http://en.wikipedia.org/wiki/Horner scheme
Examples
>>> from sympy.polys.polyfuncs import horner
>>> from sympy.abc import x, y, a, b, c, d, e
>>> horner(9*x**4 + 8*x**3 + 7*x**2 + 6*x + 5)
x*(x*(x*(9*x + 8) + 7) + 6) + 5
>>> horner(a*x**4 + b*x**3 + c*x**2 + d*x + e)
e + x*(d + x*(c + x*(a*x + b)))
>>> f = 4*x**2*y**2 + 2*x**2*y + 2*x*y**2 + x*y
>>> horner(f, wrt=x)
x*(x*y*(4*y + 2) + y*(2*y + 1))
>>> horner(f, wrt=y)
y*(x*y*(4*x + 2) + x*(2*x + 1))
sympy.polys.polyfuncs.interpolate(data, x)
Construct an interpolating polynomial for the data points.
1114
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.polys.polyfuncs import interpolate
>>> from sympy.abc import x
A list is interpreted as though it were paired with a range starting from 1:
>>> interpolate([1, 4, 9, 16], x)
x**2
This can be made explicit by giving a list of coordinates:
>>> interpolate([(1, 1), (2, 4), (3, 9)], x)
x**2
The (x, y) coordinates can also be given as keys and values of a dictionary (and the points
need not be equispaced):
>>> interpolate([(-1, 2), (1, 2), (2, 5)], x)
x**2 + 1
>>> interpolate({-1: 2, 1: 2, 2: 5}, x)
x**2 + 1
sympy.polys.polyfuncs.viete(f, roots=None, *gens, **args)
Generate Vietes formulas for f.
Examples
>>> from sympy.polys.polyfuncs import viete
>>> from sympy import symbols
>>> x, a, b, c, r1, r2 = symbols(x,a:c,r1:3)
>>> viete(a*x**2 + b*x + c, [r1, r2], x)
[(r1 + r2, -b/a), (r1*r2, c/a)]
Domain constructors
sympy.polys.constructor.construct domain(obj, **args)
Construct a minimal domain for the list of coecients.
Algebraic number elds
sympy.polys.numberfields.minimal polynomial(ex, x=None, **args)
Computes the minimal polynomial of an algebraic element.
Parameters ex : algebraic element expression
x : independent variable of the minimal polynomial
5.16. Polynomials Manipulation Module
1115
SymPy Documentation, Release 0.7.6
Notes
By default compose=True, the minimal polynomial of the subexpressions of ex are computed, then the arithmetic operations on them are performed using the resultant and
factorization. If compose=False, a bottom-up algorithm is used with groebner. The default algorithm stalls less frequently.
If no ground domain is given, it will be generated automatically from the expression.
Examples
>>> from sympy import minimal_polynomial, sqrt, solve, QQ
>>> from sympy.abc import x, y
>>> minimal_polynomial(sqrt(2), x)
x**2 - 2
>>> minimal_polynomial(sqrt(2), x, domain=QQ.algebraic_field(sqrt(2)))
x - sqrt(2)
>>> minimal_polynomial(sqrt(2) + sqrt(3), x)
x**4 - 10*x**2 + 1
>>> minimal_polynomial(solve(x**3 + x + 3)[0], x)
x**3 + x + 3
>>> minimal_polynomial(sqrt(y), x)
x**2 - y
Options
compose : if True minpoly compose is used, if False the groebner algorithm polys : if
True returns a Poly object domain : ground domain
sympy.polys.numberfields.minpoly(ex, x=None, **args)
Computes the minimal polynomial of an algebraic element.
Parameters ex : algebraic element expression
x : independent variable of the minimal polynomial
Notes
By default compose=True, the minimal polynomial of the subexpressions of ex are computed, then the arithmetic operations on them are performed using the resultant and
factorization. If compose=False, a bottom-up algorithm is used with groebner. The default algorithm stalls less frequently.
If no ground domain is given, it will be generated automatically from the expression.
Examples
>>> from sympy import minimal_polynomial, sqrt, solve, QQ
>>> from sympy.abc import x, y
1116
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> minimal_polynomial(sqrt(2), x)
x**2 - 2
>>> minimal_polynomial(sqrt(2), x, domain=QQ.algebraic_field(sqrt(2)))
x - sqrt(2)
>>> minimal_polynomial(sqrt(2) + sqrt(3), x)
x**4 - 10*x**2 + 1
>>> minimal_polynomial(solve(x**3 + x + 3)[0], x)
x**3 + x + 3
>>> minimal_polynomial(sqrt(y), x)
x**2 - y
Options
compose : if True minpoly compose is used, if False the groebner algorithm polys : if
True returns a Poly object domain : ground domain
sympy.polys.numberfields.primitive element(extension, x=None, **args)
Construct a common number eld for all extensions.
sympy.polys.numberfields.field isomorphism(a, b, **args)
Construct an isomorphism between two number elds.
sympy.polys.numberfields.to number field(extension, theta=None, **args)
Express extension in the eld generated by theta.
sympy.polys.numberfields.isolate(alg, eps=None, fast=False)
Give a rational isolating interval for an algebraic number.
class sympy.polys.numberfields.AlgebraicNumber
Class for representing algebraic numbers in SymPy.
as expr(x=None)
Create a Basic expression from self.
as poly(x=None)
Create a Poly instance from self.
coeffs()
Returns all SymPy coecients of an algebraic number.
is aliased
Returns True if alias was set.
native coeffs()
Returns all native coecients of an algebraic number.
to algebraic integer()
Convert self to an algebraic integer.
Monomials encoded as tuples
class sympy.polys.monomials.Monomial(monom, gens=None)
Class representing a monomial, i.e. a product of powers.
sympy.polys.monomials.itermonomials(variables, degree)
Generate a set of monomials of the given total degree or less.
5.16. Polynomials Manipulation Module
1117
SymPy Documentation, Release 0.7.6
Given a set of variables V and a total degree N generate a set of monomials of degree at
most N . The total number of monomials is huge and is given by the following formula:
(#V + N )!
#V !N !
For example if we would like to generate a dense polynomial of a total degree N = 50 in
5 variables, assuming that exponents and all of coecients are 32-bit long and stored in
an array we would need almost 80 GiB of memory! Fortunately most polynomials, that
we will encounter, are sparse.
Examples
Consider monomials in variables x and y:
>>> from sympy.polys.monomials import itermonomials
>>> from sympy.polys.orderings import monomial_key
>>> from sympy.abc import x, y
>>> sorted(itermonomials([x, y], 2), key=monomial_key(grlex, [y, x]))
[1, x, y, x**2, x*y, y**2]
>>> sorted(itermonomials([x, y], 3), key=monomial_key(grlex, [y, x]))
[1, x, y, x**2, x*y, y**2, x**3, x**2*y, x*y**2, y**3]
sympy.polys.monomials.monomial count(V, N)
Computes the number of monomials.
The number of monomials is given by the following formula:
(#V + N )!
#V !N !
where N is a total degree and V is a set of variables.
Examples
>>> from sympy.polys.monomials import itermonomials, monomial_count
>>> from sympy.polys.orderings import monomial_key
>>> from sympy.abc import x, y
>>> monomial_count(2, 2)
6
>>> M = itermonomials([x, y], 2)
>>> sorted(M, key=monomial_key(grlex, [y, x]))
[1, x, y, x**2, x*y, y**2]
>>> len(M)
6
Orderings of monomials
class sympy.polys.orderings.LexOrder
Lexicographic order of monomials.
1118
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
class sympy.polys.orderings.GradedLexOrder
Graded lexicographic order of monomials.
class sympy.polys.orderings.ReversedGradedLexOrder
Reversed graded lexicographic order of monomials.
Formal manipulation of roots of polynomials
class sympy.polys.rootoftools.RootOf
Represents k-th root of a univariate polynomial.
class sympy.polys.rootoftools.RootSum
Represents a sum of all roots of a univariate polynomial.
Symbolic root-nding algorithms
sympy.polys.polyroots.roots(f, *gens, **ags)
Computes symbolic roots of a univariate polynomial.
Given a univariate polynomial f with symbolic coecients (or a list of the polynomials
coecients), returns a dictionary with its roots and their multiplicities.
Only roots expressible via radicals will be returned. To get a complete set of roots use
RootOf class or numerical methods instead. By default cubic and quartic formulas are
used in the algorithm. To disable them because of unreadable output set cubics=False
or quartics=False respectively. If cubic roots are real but are expressed in terms of
complex numbers (casus irreducibilis [1]) the trig ag can be set to True to have the
solutions returned in terms of cosine and inverse cosine functions.
To get roots from a specic domain set the filter ag with one of the following speciers: Z, Q, R, I, C. By default all roots are returned (this is equivalent to setting filter=C).
By default a dictionary is returned giving a compact result in case of multiple roots.
However to get a list containing all those roots set the multiple ag to True; the list
will have identical roots appearing next to each other in the result. (For a given Poly, the
all roots method will give the roots in sorted numerical order.)
References
1.http://en.wikipedia.org/wiki/Cubic function#Trigonometric .28and hyperbolic.29 method
Examples
>>> from sympy import Poly, roots
>>> from sympy.abc import x, y
>>> roots(x**2 - 1, x)
{-1: 1, 1: 1}
>>> p = Poly(x**2-1, x)
>>> roots(p)
{-1: 1, 1: 1}
5.16. Polynomials Manipulation Module
1119
SymPy Documentation, Release 0.7.6
>>> p = Poly(x**2-y, x, y)
>>> roots(Poly(p, x))
{-sqrt(y): 1, sqrt(y): 1}
>>> roots(x**2 - y, x)
{-sqrt(y): 1, sqrt(y): 1}
>>> roots([1, 0, -1])
{-1: 1, 1: 1}
Special polynomials
sympy.polys.specialpolys.swinnerton dyer poly(n, x=None, **args)
Generates n-th Swinnerton-Dyer polynomial in x.
sympy.polys.specialpolys.interpolating poly(n, x, X=x, Y=y)
Construct Lagrange interpolating polynomial for n data points.
sympy.polys.specialpolys.cyclotomic poly(n, x=None, **args)
Generates cyclotomic polynomial of order n in x.
sympy.polys.specialpolys.symmetric poly(n, *gens, **args)
Generates symmetric polynomial of order n.
sympy.polys.specialpolys.random poly(x, n, inf, sup, domain=ZZ, polys=False)
Return a polynomial of degree n with coecients in [inf, sup].
Orthogonal polynomials
sympy.polys.orthopolys.chebyshevt poly(n, x=None, **args)
Generates Chebyshev polynomial of the rst kind of degree n in x.
sympy.polys.orthopolys.chebyshevu poly(n, x=None, **args)
Generates Chebyshev polynomial of the second kind of degree n in x.
sympy.polys.orthopolys.gegenbauer poly(n, a, x=None, **args)
Generates Gegenbauer polynomial of degree n in x.
sympy.polys.orthopolys.hermite poly(n, x=None, **args)
Generates Hermite polynomial of degree n in x.
sympy.polys.orthopolys.jacobi poly(n, a, b, x=None, **args)
Generates Jacobi polynomial of degree n in x.
sympy.polys.orthopolys.legendre poly(n, x=None, **args)
Generates Legendre polynomial of degree n in x.
sympy.polys.orthopolys.laguerre poly(n, x=None, alpha=None, **args)
Generates Laguerre polynomial of degree n in x.
Manipulation of rational functions
sympy.polys.rationaltools.together(expr, deep=False)
Denest and combine rational expressions using symbolic methods.
1120
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
This function takes an expression or a container of expressions and puts it (them) together by denesting and combining rational subexpressions. No heroic measures are
taken to minimize degree of the resulting numerator and denominator. To obtain completely reduced expression use cancel(). However, together() (page 1120) can preserve as much as possible of the structure of the input expression in the output (no
expansion is performed).
A wide variety of objects can be put together including lists, tuples, sets, relational objects, integrals and others. It is also possible to transform interior of function applications, by setting deep ag to True.
By denition, together() (page 1120) is a complement to apart(), so
apart(together(expr)) should return expr unchanged. Note however, that together() (page 1120) uses only symbolic methods, so it might be necessary to use
cancel() to perform algebraic simplication and minimise degree of the numerator and
denominator.
Examples
>>> from sympy import together, exp
>>> from sympy.abc import x, y, z
>>> together(1/x + 1/y)
(x + y)/(x*y)
>>> together(1/x + 1/y + 1/z)
(x*y + x*z + y*z)/(x*y*z)
>>> together(1/(x*y) + 1/y**2)
(x + y)/(x*y**2)
>>> together(1/(1 + 1/x) + 1/(1 + 1/y))
(x*(y + 1) + y*(x + 1))/((x + 1)*(y + 1))
>>> together(exp(1/x + 1/y))
exp(1/y + 1/x)
>>> together(exp(1/x + 1/y), deep=True)
exp((x + y)/(x*y))
>>> together(1/exp(x) + 1/(x*exp(x)))
(x + 1)*exp(-x)/x
>>> together(1/exp(2*x) + 1/(x*exp(3*x)))
(x*exp(x) + 1)*exp(-3*x)/x
Partial fraction decomposition
sympy.polys.partfrac.apart(expr, *args, **kwargs)
Compute partial fraction decomposition of a rational function.
Given a rational function f, computes the partial fraction decomposition of f. Two algorithms are available: One is based on the undertermined coecients method, the other
is Bronsteins full partial fraction decomposition algorithm.
The undetermined coecients method (selected by full=False) uses polynomial factorization (and therefore accepts the same options as factor) for the denominator. Per
5.16. Polynomials Manipulation Module
1121
SymPy Documentation, Release 0.7.6
default it works over the rational numbers, therefore decomposition of denominators
with non-rational roots (e.g. irrational, complex roots) is not supported by default (see
options of factor).
Bronsteins algorithm can be selected by using full=True and allows a decomposition
of denominators with non-rational roots. A human-readable result can be obtained via
doit() (see examples below).
See Also:
apart list (page 1122), assemble partfrac list (page 1124)
Examples
>>> from sympy.polys.partfrac import apart
>>> from sympy.abc import x, y
By default, using the undetermined coecients method:
>>> apart(y/(x + 2)/(x + 1), x)
-y/(x + 2) + y/(x + 1)
The undetermined coecients method does not provide a result when the denominators
roots are not rational:
>>> apart(y/(x**2 + x + 1), x)
y/(x**2 + x + 1)
You can choose Bronsteins algorithm by setting full=True:
>>> apart(y/(x**2 + x + 1), x, full=True)
RootSum(_w**2 + _w + 1, Lambda(_a, (-2*_a*y/3 - y/3)/(-_a + x)))
Calling doit() yields a human-readable result:
>>> apart(y/(x**2 + x + 1), x, full=True).doit()
(-y/3 - 2*y*(-1/2 - sqrt(3)*I/2)/3)/(x + 1/2 + sqrt(3)*I/2) + (-y/3 2*y*(-1/2 + sqrt(3)*I/2)/3)/(x + 1/2 - sqrt(3)*I/2)
sympy.polys.partfrac.apart list(f, x=None, dummies=None, **options)
Compute partial fraction decomposition of a rational function and return the result in
structured form.
Given a rational function f compute the partial fraction decomposition of f. Only Bronsteins full partial fraction decomposition algorithm is supported by this method. The
return value is highly structured and perfectly suited for further algorithmic treatment
rather than being human-readable. The function returns a tuple holding three elements:
The rst item is the common coecient, free of the variable x used for decomposition. (It is an element of the base eld K.)
The second item is the polynomial part of the decomposition. This can be the zero
polynomial. (It is an element of K[x].)
The third part itself is a list of quadruples. Each quadruple has the following elements in this order:
The (not necessarily irreducible) polynomial D whose roots wi appear in the linear denominator of a bunch of related fraction terms. (This item can also be a
list of explicit roots. However, at the moment apart list never returns a result
1122
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
this way, but the related assemble partfrac list function accepts this format
as input.)
The numerator of the fraction, written as a function of the root w
The linear denominator of the fraction excluding its power exponent, written as
a function of the root w.
The power to which the denominator has to be raised.
On can always rebuild a plain expression by using the function assemble partfrac list.
See Also:
apart (page 1121), assemble partfrac list (page 1124)
References
1.[Bronstein93] (page 1911)
[Bronstein93] (page 1911)
Examples
A rst example:
>>> from sympy.polys.partfrac import apart_list, assemble_partfrac_list
>>> from sympy.abc import x, t
>>> f = (2*x**3 - 2*x) / (x**2 - 2*x + 1)
>>> pfd = apart_list(f)
>>> pfd
(1,
Poly(2*x + 4, x, domain=ZZ),
[(Poly(_w - 1, _w, domain=ZZ), Lambda(_a, 4), Lambda(_a, -_a + x), 1)])
>>> assemble_partfrac_list(pfd)
2*x + 4 + 4/(x - 1)
Second example:
>>> f = (-2*x - 2*x**2) / (3*x**2 - 6*x)
>>> pfd = apart_list(f)
>>> pfd
(-1,
Poly(2/3, x, domain=QQ),
[(Poly(_w - 2, _w, domain=ZZ), Lambda(_a, 2), Lambda(_a, -_a + x), 1)])
>>> assemble_partfrac_list(pfd)
-2/3 - 2/(x - 2)
Another example, showing symbolic parameters:
>>> pfd = apart_list(t/(x**2 + x + t), x)
>>> pfd
(1,
Poly(0, x, domain=ZZ[t]),
[(Poly(_w**2 + _w + t, _w, domain=ZZ[t]),
5.16. Polynomials Manipulation Module
1123
SymPy Documentation, Release 0.7.6
Lambda(_a, -2*_a*t/(4*t - 1) - t/(4*t - 1)),
Lambda(_a, -_a + x),
1)])
>>> assemble_partfrac_list(pfd)
RootSum(_w**2 + _w + t, Lambda(_a, (-2*_a*t/(4*t - 1) - t/(4*t - 1))/(-_a + x)))
This example is taken from Bronsteins original paper:
>>> f = 36 / (x**5 - 2*x**4 - 2*x**3 + 4*x**2 + x - 2)
>>> pfd = apart_list(f)
>>> pfd
(1,
Poly(0, x, domain=ZZ),
[(Poly(_w - 2, _w, domain=ZZ), Lambda(_a, 4), Lambda(_a, -_a + x), 1),
(Poly(_w**2 - 1, _w, domain=ZZ), Lambda(_a, -3*_a - 6), Lambda(_a, -_a + x), 2),
(Poly(_w + 1, _w, domain=ZZ), Lambda(_a, -4), Lambda(_a, -_a + x), 1)])
>>> assemble_partfrac_list(pfd)
-4/(x + 1) - 3/(x + 1)**2 - 9/(x - 1)**2 + 4/(x - 2)
sympy.polys.partfrac.assemble partfrac list(partial list)
Reassemble a full partial fraction decomposition from a structured result obtained by
the function apart list.
See Also:
apart (page 1121), apart list (page 1122)
Examples
This example is taken from Bronsteins original paper:
>>> from sympy.polys.partfrac import apart_list, assemble_partfrac_list
>>> from sympy.abc import x, y
>>> f = 36 / (x**5 - 2*x**4 - 2*x**3 + 4*x**2 + x - 2)
>>> pfd = apart_list(f)
>>> pfd
(1,
Poly(0, x, domain=ZZ),
[(Poly(_w - 2, _w, domain=ZZ), Lambda(_a, 4), Lambda(_a, -_a + x), 1),
(Poly(_w**2 - 1, _w, domain=ZZ), Lambda(_a, -3*_a - 6), Lambda(_a, -_a + x), 2),
(Poly(_w + 1, _w, domain=ZZ), Lambda(_a, -4), Lambda(_a, -_a + x), 1)])
>>> assemble_partfrac_list(pfd)
-4/(x + 1) - 3/(x + 1)**2 - 9/(x - 1)**2 + 4/(x - 2)
If we happen to know some roots we can provide them easily inside the structure:
>>> pfd = apart_list(2/(x**2-2))
>>> pfd
(1,
Poly(0, x, domain=ZZ),
[(Poly(_w**2 - 2, _w, domain=ZZ),
Lambda(_a, _a/2),
Lambda(_a, -_a + x),
1)])
1124
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> pfda = assemble_partfrac_list(pfd)
>>> pfda
RootSum(_w**2 - 2, Lambda(_a, _a/(-_a + x)))/2
>>> pfda.doit()
-sqrt(2)/(2*(x + sqrt(2))) + sqrt(2)/(2*(x - sqrt(2)))
>>> from sympy import Dummy, Poly, Lambda, sqrt
>>> a = Dummy(a)
>>> pfd = (1, Poly(0, x, domain=ZZ), [([sqrt(2),-sqrt(2)], Lambda(a, a/2), Lambda(a, -a + x),
>>> assemble_partfrac_list(pfd)
-sqrt(2)/(2*(x + sqrt(2))) + sqrt(2)/(2*(x - sqrt(2)))
Dispersion of Polynomials
sympy.polys.dispersion.dispersionset(p, q=None, *gens, **args)
Compute the dispersion set of two polynomials.
For two polynomials f (x) and g(x) with deg f > 0 and deg g > 0 the dispersion set J(f, g) is
dened as:
J(f, g) := {a N0 | gcd(f (x), g(x + a)) 6= 1}
= {a N0 | deg gcd(f (x), g(x + a)) 1}
For a single polynomial one denes J(f ) := J(f, f ).
See Also:
dispersion (page 1062)
References
1.[ManWright94] (page 1912)
2.[Koepf98] (page 1912)
3.[Abramov71] (page 1912)
4.[Man93] (page 1912)
[ManWright94] (page 1912), [Koepf98] (page 1912), [Abramov71] (page 1912), [Man93]
(page 1912)
Examples
>>> from sympy import poly
>>> from sympy.polys.dispersion import dispersion, dispersionset
>>> from sympy.abc import x
Dispersion set and dispersion of a simple polynomial:
5.16. Polynomials Manipulation Module
1125
SymPy Documentation, Release 0.7.6
>>>
>>>
[0,
>>>
6
fp = poly((x - 3)*(x + 3), x)
sorted(dispersionset(fp))
6]
dispersion(fp)
Note that the denition of the dispersion is not symmetric:
>>>
>>>
>>>
[2,
>>>
4
>>>
[]
>>>
-oo
fp = poly(x**4 - 3*x**2 + 1, x)
gp = fp.shift(-3)
sorted(dispersionset(fp, gp))
3, 4]
dispersion(fp, gp)
sorted(dispersionset(gp, fp))
dispersion(gp, fp)
Computing the dispersion also works over eld extensions:
>>>
>>>
>>>
>>>
[2]
>>>
[1,
from sympy import sqrt
fp = poly(x**2 + sqrt(5)*x - 1, x, domain=QQ<sqrt(5)>)
gp = poly(x**2 + (2 + sqrt(5))*x + sqrt(5), x, domain=QQ<sqrt(5)>)
sorted(dispersionset(fp, gp))
sorted(dispersionset(gp, fp))
4]
We can even perform the computations for polynomials having symbolic coecients:
>>>
>>>
>>>
[0,
from sympy.abc import a
fp = poly(4*x**4 + (4*a + 8)*x**3 + (a**2 + 6*a + 4)*x**2 + (a**2 + 2*a)*x, x)
sorted(dispersionset(fp))
1]
sympy.polys.dispersion.dispersion(p, q=None, *gens, **args)
Compute the dispersion of polynomials.
For two polynomials f (x) and g(x) with deg f > 0 and deg g > 0 the dispersion dis(f, g) is
dened as:
dis(f, g) := max{J(f, g) {0}}
= max{{a N| gcd(f (x), g(x + a)) 6= 1} {0}}
and for a single polynomial dis(f ) := dis(f, f ). Note that we make the denition max{} :=
.
See Also:
dispersionset (page 1063)
References
1.[ManWright94] (page 1912)
2.[Koepf98] (page 1912)
3.[Abramov71] (page 1912)
1126
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
4.[Man93] (page 1912)
[ManWright94] (page 1912), [Koepf98] (page 1912), [Abramov71] (page 1912), [Man93]
(page 1912)
Examples
>>> from sympy import poly
>>> from sympy.polys.dispersion import dispersion, dispersionset
>>> from sympy.abc import x
Dispersion set and dispersion of a simple polynomial:
>>>
>>>
[0,
>>>
6
fp = poly((x - 3)*(x + 3), x)
sorted(dispersionset(fp))
6]
dispersion(fp)
Note that the denition of the dispersion is not symmetric:
>>>
>>>
>>>
[2,
>>>
4
>>>
[]
>>>
-oo
fp = poly(x**4 - 3*x**2 + 1, x)
gp = fp.shift(-3)
sorted(dispersionset(fp, gp))
3, 4]
dispersion(fp, gp)
sorted(dispersionset(gp, fp))
dispersion(gp, fp)
The maximum of an empty set is dened to be as seen in this example.
Computing the dispersion also works over eld extensions:
>>>
>>>
>>>
>>>
[2]
>>>
[1,
from sympy import sqrt
fp = poly(x**2 + sqrt(5)*x - 1, x, domain=QQ<sqrt(5)>)
gp = poly(x**2 + (2 + sqrt(5))*x + sqrt(5), x, domain=QQ<sqrt(5)>)
sorted(dispersionset(fp, gp))
sorted(dispersionset(gp, fp))
4]
We can even perform the computations for polynomials having symbolic coecients:
>>>
>>>
>>>
[0,
from sympy.abc import a
fp = poly(4*x**4 + (4*a + 8)*x**3 + (a**2 + 6*a + 4)*x**2 + (a**2 + 2*a)*x, x)
sorted(dispersionset(fp))
1]
AGCA - Algebraic Geometry and Commutative Algebra Module
Introduction
Algebraic geometry is a mixture of the ideas of two Mediterranean cultures. It is
the superposition of the Arab science of the lightening calculation of the solutions
5.16. Polynomials Manipulation Module
1127
SymPy Documentation, Release 0.7.6
of equations over the Greek art of position and shape. This tapestry was originally
woven on European soil and is still being rened under the inuence of international
fashion. Algebraic geometry studies the delicate balance between the geometrically
plausible and the algebraically possible. Whenever one side of this mathematical
teeter-totter outweighs the other, one immediately loses interest and runs o in
search of a more exciting amusement.
George R. Kempf 1944 2002
Algebraic Geometry refers to the study of geometric problems via algebraic methods (and
sometimes vice versa). While this is a rather old topic, algebraic geometry as understood
today is very much a 20th century development. Building on ideas of e.g. Riemann and
Dedekind, it was realized that there is an intimate connection between properties of the set
of solutions of a system of polynomial equations (called an algebraic variety) and the behavior
of the set of polynomial functions on that variety (called the coordinate ring).
As in many geometric disciplines, we can distinguish between local and global questions (and
methods). Local investigations in algebraic geometry are essentially equivalent to the study of
certain rings, their ideals and modules. This latter topic is also called commutative algebra.
It is the basic local toolset of algebraic geometers, in much the same way that dierential
analysis is the local toolset of dierential geometers.
A good conceptual introduction to commutative algebra is [Atiyah69] (page 1911). An introduction more geared towards computations, and the work most of the algorithms in this
module are based on, is [Greuel2008] (page 1911).
This module aims to eventually allow expression and solution of both local and global geometric problems, both in the classical case over a eld and in the more modern arithmetic
cases. So far, however, there is no geometric functionality at all. Currently the module only
provides tools for computational commutative algebra over elds.
All code examples assume:
>>> from sympy import *
>>> x, y, z = symbols(x,y,z)
>>> init_printing(use_unicode=True, wrap_line=False, no_global=True)
Reference
In this section we document the usage of the AGCA module. For convenience of the reader,
some denitions and examples/explanations are interspersed.
Base Rings Almost all computations in commutative algebra are relative to a base ring.
(For example, when asking questions about an ideal, the base ring is the ring the ideal is
a subset of.) In principle all polys domains can be used as base rings. However, useful
functionality is only implemented for polynomial rings over elds, and various localizations
and quotients thereof.
As demonstrated in the examples below, the most convenient method to create objects you
are interested in is to build them up from the ground eld, and then use the various methods
to create new objects from old. For example, in order to create the local ring of the nodal
cubic y 2 = x3 at the origin, over Q, you do:
>>> lr = QQ.old_poly_ring(x, y, order=ilex) / [y**2 - x**3]
>>> lr
[x, y, order=ilex]
-------------------
1128
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
3
2\
\- x + y
Note how the python list notation can be used as a short cut to express ideals. You can use
the convert method to return ordinary sympy objects into objects understood by the AGCA
module (although in many cases this will be done automatically for example the list was
automatically turned into an ideal, and in the process the symbols x and y were automatically
converted into other representations). For example:
>>> X, Y = lr.convert(x), lr.convert(y) ; X
3
2\
x + \- x + y
>>> x**3 == y**2
False
>>> X**3 == Y**2
True
When no localisation is needed, a more mathematical notation can be used. For example, let
us create the coordinate ring of three-dimensional ane space A3 :
>>> ar = QQ.old_poly_ring(x, y, z); ar
[x, y, z]
For more details, refer to the following class documentation. Note that the base rings, being
domains, are the main point of overlap between the AGCA module and the rest of the polys
module. All domains are documented in detail in the polys reference, so we show here only
an abridged version, with the methods most pertinent to the AGCA module.
class sympy.polys.domains.ring.Ring
Represents a ring domain.
Attributes
alias
dtype
one
rep
zero
Ring.free module(rank)
Generate a free module of rank rank over self.
>>> from sympy.abc import x
>>> from sympy import QQ
>>> QQ.old_poly_ring(x).free_module(2)
QQ[x]**2
Ring.ideal(*gens)
Generate an ideal of self.
>>> from sympy.abc import x
>>> from sympy import QQ
>>> QQ.old_poly_ring(x).ideal(x**2)
<x**2>
5.16. Polynomials Manipulation Module
1129
SymPy Documentation, Release 0.7.6
Ring.quotient ring(e)
Form a quotient ring of self.
Here e can be an ideal or an iterable.
>>> from sympy.abc import x
>>> from sympy import QQ
>>> QQ.old_poly_ring(x).quotient_ring(QQ.old_poly_ring(x).ideal(x**2))
QQ[x]/<x**2>
>>> QQ.old_poly_ring(x).quotient_ring([x**2])
QQ[x]/<x**2>
The division operator has been overloaded for this:
>>> QQ.old_poly_ring(x)/[x**2]
QQ[x]/<x**2>
sympy.polys.domains.polynomialring.PolynomialRing(domain or ring,
symbols=None, order=None)
A class for representing multivariate polynomial rings.
Attributes
alias
domain
dtype
gens
ngens
rep
symbols
class sympy.polys.domains.quotientring.QuotientRing(ring, ideal)
Class representing (commutative) quotient rings.
You should not usually instantiate this by hand, instead use the constructor from the
base ring in the construction.
>>> from sympy.abc import x
>>> from sympy import QQ
>>> I = QQ.old_poly_ring(x).ideal(x**3 + 1)
>>> QQ.old_poly_ring(x).quotient_ring(I)
QQ[x]/<x**3 + 1>
Shorter versions are possible:
>>> QQ.old_poly_ring(x)/I
QQ[x]/<x**3 + 1>
>>> QQ.old_poly_ring(x)/[x**3 + 1]
QQ[x]/<x**3 + 1>
Attributes:
ring - the base ring
base ideal - the ideal used to form the quotient
1130
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Attributes
alias
one
rep
zero
Modules, Ideals and their Elementary Properties Let A be a ring. An A-module is a set
M , together with two binary operations + : M M M and : R M M called addition
and scalar multiplication. These are required to satisfy certain axioms, which can be found
in e.g. [Atiyah69] (page 1911). In this way modules are a direct generalisation of both vector
spaces (A being a eld) and abelian groups (A = Z). A submodule of the A-module M is a
subset N M , such that the binary operations restrict to N , and N becomes an A-module
with these operations.
The ring A itself has a natural A-module structure where addition and multiplication in the
module coincide with addition and multiplication in the ring. This A-module is also written as
A. An A-submodule of A is called an ideal of A. Ideals come up very naturally in algebraic geometry. More general modules can be seen as a technically convenient elbow room beyond
talking only about ideals.
If M , N are A-modules, then there is a natural (componentwise) A-module structure on M N .
Similarly there are A-module structures on cartesian products of more components. (For
the categorically inclined: the cartesian product of nitely many A-modules, with this Amodule structure, is the nite biproduct in the category of all A-modules. With innitely
many components, it is the direct product (but the innite direct sum has to be constructed
dierently).) As usual, repeated product of the A-module M is denoted M, M 2 , M 3 . . . , or M I
for arbitrary index sets I.
An A-module M is called free if it is isomorphic to the A-module AI for some (not necessarily
nite) index set I (refer to the next section for a denition of isomorphism). The cardinality
of I is called the rank of M ; one may prove this is well-dened. In general, the AGCA module
only works with free modules of nite rank, and other closely related modules. The easiest
way to create modules is to use member methods of the objects they are made up from. For
example, let us create a free module of rank 4 over the coordinate ring of A2 we created above,
together with a submodule:
>>> F = ar.free_module(4) ; F
4
[x, y, z]
>>> S = F.submodule([1, x, x**2, x**3], [0, 1, 0, y]) ; S
2
3
\
\1, x, x , x , [0, 1, 0, y]
Note how python lists can be used as a short-cut notation for module elements (vectors). As
usual, the convert method can be used to convert sympy/python objects into the internal
AGCA representation (see detailed reference below).
Here is the detailed documentation of the classes for modules, free modules, and submodules:
class sympy.polys.agca.modules.Module(ring)
Abstract base class for modules.
Do not instantiate - use ring explicit constructors instead:
5.16. Polynomials Manipulation Module
1131
SymPy Documentation, Release 0.7.6
>>> from sympy import QQ
>>> from sympy.abc import x
>>> QQ.old_poly_ring(x).free_module(2)
QQ[x]**2
Attributes:
dtype - type of elements
ring - containing ring
Non-implemented methods:
submodule
quotient module
is zero
is submodule
multiply ideal
The method convert likely needs to be changed in subclasses.
contains(elem)
Return True if elem is an element of this module.
convert(elem, M=None)
Convert elem into internal representation of this module.
If M is not None, it should be a module containing it.
identity hom()
Return the identity homomorphism on self.
is submodule(other)
Returns True if other is a submodule of self.
is zero()
Returns True if self is a zero module.
multiply ideal(other)
Multiply self by the ideal other.
quotient module(other)
Generate a quotient module.
submodule(*gens)
Generate a submodule.
subset(other)
Returns True if other is is a subset of self.
>>> from sympy.abc import x
>>> from sympy import QQ
>>> F = QQ.old_poly_ring(x).free_module(2)
>>> F.subset([(1, x), (x, 2)])
True
>>> F.subset([(1/x, x), (x, 2)])
False
class sympy.polys.agca.modules.FreeModule(ring, rank)
Abstract base class for free modules.
Additional attributes:
1132
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
rank - rank of the free module
Non-implemented methods:
submodule
basis()
Return a set of basis elements.
>>> from sympy.abc import x
>>> from sympy import QQ
>>> QQ.old_poly_ring(x).free_module(3).basis()
([1, 0, 0], [0, 1, 0], [0, 0, 1])
convert(elem, M=None)
Convert elem into the internal representation.
This method is called implicitly whenever computations involve elements not in the
internal representation.
>>>
>>>
>>>
>>>
[1,
from sympy.abc import x
from sympy import QQ
F = QQ.old_poly_ring(x).free_module(2)
F.convert([1, 0])
0]
dtype
alias of FreeModuleElement
identity hom()
Return the identity homomorphism on self.
>>> from sympy.abc import x
>>> from sympy import QQ
>>> QQ.old_poly_ring(x).free_module(2).identity_hom()
Matrix([
[1, 0], : QQ[x]**2 -> QQ[x]**2
[0, 1]])
is submodule(other)
Returns True if other is a submodule of self.
>>> from sympy.abc import x
>>> from sympy import QQ
>>> F = QQ.old_poly_ring(x).free_module(2)
>>> M = F.submodule([2, x])
>>> F.is_submodule(F)
True
>>> F.is_submodule(M)
True
>>> M.is_submodule(F)
False
is zero()
Returns True if self is a zero module.
(If, as this implementation assumes, the coecient ring is not the zero ring, then
this is equivalent to the rank being zero.)
>>> from sympy.abc import x
>>> from sympy import QQ
>>> QQ.old_poly_ring(x).free_module(0).is_zero()
5.16. Polynomials Manipulation Module
1133
SymPy Documentation, Release 0.7.6
True
>>> QQ.old_poly_ring(x).free_module(1).is_zero()
False
multiply ideal(other)
Multiply self by the ideal other.
>>> from sympy.abc import x
>>> from sympy import QQ
>>> I = QQ.old_poly_ring(x).ideal(x)
>>> F = QQ.old_poly_ring(x).free_module(2)
>>> F.multiply_ideal(I)
<[x, 0], [0, x]>
quotient module(submodule)
Return a quotient module.
>>> from sympy.abc import x
>>> from sympy import QQ
>>> M = QQ.old_poly_ring(x).free_module(2)
>>> M.quotient_module(M.submodule([1, x], [x, 2]))
QQ[x]**2/<[1, x], [x, 2]>
Or more conicisely, using the overloaded division operator:
>>> QQ.old_poly_ring(x).free_module(2) / [[1, x], [x, 2]]
QQ[x]**2/<[1, x], [x, 2]>
class sympy.polys.agca.modules.SubModule(gens, container)
Base class for submodules.
Attributes:
container - containing module
gens - generators (subset of containing module)
rank - rank of containing module
Non-implemented methods:
contains
syzygies
in terms of generators
intersect
module quotient
Methods that likely need change in subclasses:
reduce element
convert(elem, M=None)
Convert elem into the internal represantition.
Mostly called implicitly.
>>>
>>>
>>>
>>>
[2,
1134
from sympy.abc import x
from sympy import QQ
M = QQ.old_poly_ring(x).free_module(2).submodule([1, x])
M.convert([2, 2*x])
2*x]
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
identity hom()
Return the identity homomorphism on self.
>>> from sympy.abc import x
>>> from sympy import QQ
>>> QQ.old_poly_ring(x).free_module(2).submodule([x, x]).identity_hom()
Matrix([
[1, 0], : <[x, x]> -> <[x, x]>
[0, 1]])
in terms of generators(e)
Express element e of self in terms of the generators.
>>> from sympy.abc import x
>>> from sympy import QQ
>>> F = QQ.old_poly_ring(x).free_module(2)
>>> M = F.submodule([1, 0], [1, 1])
>>> M.in_terms_of_generators([x, x**2])
[-x**2 + x, x**2]
inclusion hom()
Return a homomorphism representing the inclusion map of self.
That is, the natural map from self to self.container.
>>> from sympy.abc import x
>>> from sympy import QQ
>>> QQ.old_poly_ring(x).free_module(2).submodule([x, x]).inclusion_hom()
Matrix([
[1, 0], : <[x, x]> -> QQ[x]**2
[0, 1]])
intersect(other, **options)
Returns the intersection of self with submodule other.
>>> from sympy.abc import x, y
>>> from sympy import QQ
>>> F = QQ.old_poly_ring(x, y).free_module(2)
>>> F.submodule([x, x]).intersect(F.submodule([y, y]))
<[x*y, x*y]>
Some implementation allow further options to be passed. Currently, to only one
implemented is relations=True, in which case the function will return a triple (res,
rela, relb), where res is the intersection module, and rela and relb are lists of
coecient vectors, expressing the generators of res in terms of the generators of
self (rela) and other (relb).
>>> F.submodule([x, x]).intersect(F.submodule([y, y]), relations=True)
(<[x*y, x*y]>, [(y,)], [(x,)])
The above result says: the intersection module is generated by the single element
(xy, xy) = y(x, x) = x(y, y), where (x, x) and (y, y) respectively are the unique
generators of the two modules being intersected.
is full module()
Return True if self is the entire free module.
>>> from sympy.abc import x
>>> from sympy import QQ
>>> F = QQ.old_poly_ring(x).free_module(2)
5.16. Polynomials Manipulation Module
1135
SymPy Documentation, Release 0.7.6
>>> F.submodule([x, 1]).is_full_module()
False
>>> F.submodule([1, 1], [1, 2]).is_full_module()
True
is submodule(other)
Returns True if other is a submodule of self.
>>> from sympy.abc import x
>>> from sympy import QQ
>>> F = QQ.old_poly_ring(x).free_module(2)
>>> M = F.submodule([2, x])
>>> N = M.submodule([2*x, x**2])
>>> M.is_submodule(M)
True
>>> M.is_submodule(N)
True
>>> N.is_submodule(M)
False
is zero()
Return True if self is a zero module.
>>> from sympy.abc import x
>>> from sympy import QQ
>>> F = QQ.old_poly_ring(x).free_module(2)
>>> F.submodule([x, 1]).is_zero()
False
>>> F.submodule([0, 0]).is_zero()
True
module quotient(other, **options)
Returns the module quotient of self by submodule other.
That is, if self is the module M and other is N , then return the ideal {f R|f N M }.
>>>
>>>
>>>
>>>
>>>
>>>
<y>
from sympy import QQ
from sympy.abc import x, y
F = QQ.old_poly_ring(x, y).free_module(2)
S = F.submodule([x*y, x*y])
T = F.submodule([x, x])
S.module_quotient(T)
Some implementations allow further options to be passed. Currently, the only one
implemented is relations=True, which may only be passed if other is prinicipal.
In this case the function will return a pair (res, rel) where res is the ideal, and
rel is a list of coecient vectors, expressing the generators of the ideal, multiplied
by the generator of other in terms of generators of self.
>>> S.module_quotient(T, relations=True)
(<y>, [[1]])
This means that the quotient ideal is generated by the single element y, and that
y(x, x) = 1(xy, xy), (x, x) and (xy, xy) being the generators of T and S, respectively.
multiply ideal(I)
Multiply self by the ideal I.
1136
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> from sympy.abc import x
>>> from sympy import QQ
>>> I = QQ.old_poly_ring(x).ideal(x**2)
>>> M = QQ.old_poly_ring(x).free_module(2).submodule([1, 1])
>>> I*M
<[x**2, x**2]>
quotient module(other, **opts)
Return a quotient module.
This is the same as taking a submodule of a quotient of the containing module.
>>> from sympy.abc import x
>>> from sympy import QQ
>>> F = QQ.old_poly_ring(x).free_module(2)
>>> S1 = F.submodule([x, 1])
>>> S2 = F.submodule([x**2, x])
>>> S1.quotient_module(S2)
<[x, 1] + <[x**2, x]>>
Or more coincisely, using the overloaded division operator:
>>> F.submodule([x, 1]) / [(x**2, x)]
<[x, 1] + <[x**2, x]>>
reduce element(x)
Reduce the element x of our ring modulo the ideal self.
Here reduce has no specic meaning, it could return a unique normal form, simplify the expression a bit, or just do nothing.
submodule(*gens)
Generate a submodule.
>>> from sympy.abc import x
>>> from sympy import QQ
>>> M = QQ.old_poly_ring(x).free_module(2).submodule([x, 1])
>>> M.submodule([x**2, x])
<[x**2, x]>
syzygy module(**opts)
Compute the syzygy module of the generators of self.
Suppose M is generated by f1 , . . . , fn over the ring R. Consider the homomorphism
: Rn M , given by sending (r1 , . . . , rn ) r1 f1 + + rn fn . The syzygy module is
dened to be the kernel of .
The syzygy module is zero i the generators generate freely a free submodule:
>>> from sympy.abc import x, y
>>> from sympy import QQ
>>> QQ.old_poly_ring(x).free_module(2).submodule([1, 0], [1, 1]).syzygy_module().is_zero()
True
A slightly more interesting example:
>>> M = QQ.old_poly_ring(x, y).free_module(2).submodule([x, 2*x], [y, 2*y])
>>> S = QQ.old_poly_ring(x, y).free_module(2).submodule([y, -x])
>>> M.syzygy_module() == S
True
5.16. Polynomials Manipulation Module
1137
SymPy Documentation, Release 0.7.6
union(other)
Returns the module generated by the union of self and other.
>>> from sympy.abc import x
>>> from sympy import QQ
>>> F = QQ.old_poly_ring(x).free_module(1)
>>> M = F.submodule([x**2 + x]) # <x(x+1)>
>>> N = F.submodule([x**2 - 1]) # <(x-1)(x+1)>
>>> M.union(N) == F.submodule([x+1])
True
Ideals are created very similarly to modules. For example, lets verify that the nodal cubic is
indeed singular at the origin:
>>> I = lr.ideal(x, y)
>>> I == lr.ideal(x)
False
>>> I == lr.ideal(y)
False
We are using here the fact that a curve is non-singular at a point if and only if the maximal ideal
of the local ring is principal, and that in this case at least one of x and y must be generators.
This is the detailed documentation of the class ideal. Please note that most of the methods
regarding properties of ideals (primality etc.) are not yet implemented.
class sympy.polys.agca.ideals.Ideal(ring)
Abstract base class for ideals.
Do not instantiate - use explicit constructors in the ring class instead:
>>> from sympy import QQ
>>> from sympy.abc import x
>>> QQ.old_poly_ring(x).ideal(x+1)
<x + 1>
Attributes
ring - the ring this ideal belongs to
Non-implemented methods:
contains elem
contains ideal
quotient
intersect
union
product
is whole ring
is zero
is prime, is maximal, is primary, is radical
is principal
height, depth
radical
Methods that likely should be overridden in subclasses:
1138
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
reduce element
contains(elem)
Return True if elem is an element of this ideal.
>>> from sympy.abc import x
>>> from sympy import QQ
>>> QQ.old_poly_ring(x).ideal(x+1, x-1).contains(3)
True
>>> QQ.old_poly_ring(x).ideal(x**2, x**3).contains(x)
False
depth()
Compute the depth of self.
height()
Compute the height of self.
intersect(J)
Compute the intersection of self with ideal J.
>>> from sympy.abc import x, y
>>> from sympy import QQ
>>> R = QQ.old_poly_ring(x, y)
>>> R.ideal(x).intersect(R.ideal(y))
<x*y>
is maximal()
Return True if self is a maximal ideal.
is primary()
Return True if self is a primary ideal.
is prime()
Return True if self is a prime ideal.
is principal()
Return True if self is a principal ideal.
is radical()
Return True if self is a radical ideal.
is whole ring()
Return True if self is the whole ring.
is zero()
Return True if self is the zero ideal.
product(J)
Compute the ideal product of self and J.
That is, compute the ideal generated by products xy, for x an element of self and
y J.
>>> from sympy.abc import x, y
>>> from sympy import QQ
>>> QQ.old_poly_ring(x, y).ideal(x).product(QQ.old_poly_ring(x, y).ideal(y))
<x*y>
quotient(J, **opts)
Compute the ideal quotient of self by J.
That is, if self is the ideal I, compute the set I : J = {x R|xJ I}.
5.16. Polynomials Manipulation Module
1139
SymPy Documentation, Release 0.7.6
>>>
>>>
>>>
>>>
<y>
from sympy.abc import x, y
from sympy import QQ
R = QQ.old_poly_ring(x, y)
R.ideal(x*y).quotient(R.ideal(x))
radical()
Compute the radical of self.
reduce element(x)
Reduce the element x of our ring modulo the ideal self.
Here reduce has no specic meaning: it could return a unique normal form, simplify the expression a bit, or just do nothing.
saturate(J)
Compute the ideal saturation of self by J.
That is, if self is the ideal I, compute the set I : J = {x R|xJ n I for some n}.
subset(other)
Returns True if other is is a subset of self.
Here other may be an ideal.
>>> from sympy.abc import x
>>> from sympy import QQ
>>> I = QQ.old_poly_ring(x).ideal(x+1)
>>> I.subset([x**2 - 1, x**2 + 2*x + 1])
True
>>> I.subset([x**2 + 1, x + 1])
False
>>> I.subset(QQ.old_poly_ring(x).ideal(x**2 - 1))
True
union(J)
Compute the ideal generated by the union of self and J.
>>> from sympy.abc import x
>>> from sympy import QQ
>>> QQ.old_poly_ring(x).ideal(x**2 - 1).union(QQ.old_poly_ring(x).ideal((x+1)**2)) == QQ.old
True
If M is an A-module and N is an A-submodule, we can dene two elements x and y of M to
be equivalent if x y N . The set of equivalence classes is written M /N , and has a natural
A-module structure. This is called the quotient module of M by N . If K is a submodule of M
containing N , then K/N is in a natural way a submodule of M /N . Such a module is called a
subquotient. Here is the documentation of quotient and subquotient modules:
class sympy.polys.agca.modules.QuotientModule(ring, base, submodule)
Class for quotient modules.
Do not instantiate this directly. For subquotients, see the SubQuotientModule class.
Attributes:
base - the base module we are a quotient of
killed module - the submodule used to form the quotient
rank of the base
convert(elem, M=None)
Convert elem into the internal representation.
1140
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
This method is called implicitly whenever computations involve elements not in the
internal representation.
>>>
>>>
>>>
>>>
[1,
from sympy.abc import x
from sympy import QQ
F = QQ.old_poly_ring(x).free_module(2) / [(1, 2), (1, x)]
F.convert([1, 0])
0] + <[1, 2], [1, x]>
dtype
alias of QuotientModuleElement
identity hom()
Return the identity homomorphism on self.
>>> from sympy.abc import x
>>> from sympy import QQ
>>> M = QQ.old_poly_ring(x).free_module(2) / [(1, 2), (1, x)]
>>> M.identity_hom()
Matrix([
[1, 0], : QQ[x]**2/<[1, 2], [1, x]> -> QQ[x]**2/<[1, 2], [1, x]>
[0, 1]])
is submodule(other)
Return True if other is a submodule of self.
>>> from sympy.abc import x
>>> from sympy import QQ
>>> Q = QQ.old_poly_ring(x).free_module(2) / [(x, x)]
>>> S = Q.submodule([1, 0])
>>> Q.is_submodule(S)
True
>>> S.is_submodule(Q)
False
is zero()
Return True if self is a zero module.
This happens if and only if the base module is the same as the submodule being
killed.
>>> from sympy.abc import x
>>> from sympy import QQ
>>> F = QQ.old_poly_ring(x).free_module(2)
>>> (F/[(1, 0)]).is_zero()
False
>>> (F/[(1, 0), (0, 1)]).is_zero()
True
quotient hom()
Return the quotient homomorphism to self.
That is, return a homomorphism representing the natural map from self.base to
self.
>>> from sympy.abc import x
>>> from sympy import QQ
>>> M = QQ.old_poly_ring(x).free_module(2) / [(1, 2), (1, x)]
>>> M.quotient_hom()
Matrix([
5.16. Polynomials Manipulation Module
1141
SymPy Documentation, Release 0.7.6
[1, 0], : QQ[x]**2 -> QQ[x]**2/<[1, 2], [1, x]>
[0, 1]])
submodule(*gens, **opts)
Generate a submodule.
This is the same as taking a quotient of a submodule of the base module.
>>> from sympy.abc import x
>>> from sympy import QQ
>>> Q = QQ.old_poly_ring(x).free_module(2) / [(x, x)]
>>> Q.submodule([x, 0])
<[x, 0] + <[x, x]>>
class sympy.polys.agca.modules.SubQuotientModule(gens, container, **opts)
Submodule of a quotient module.
Equivalently, quotient module of a submodule.
Do not instantiate this, instead use the submodule or quotient module constructing methods:
>>> from sympy.abc import x
>>> from sympy import QQ
>>> F = QQ.old_poly_ring(x).free_module(2)
>>> S = F.submodule([1, 0], [1, x])
>>> Q = F/[(1, 0)]
>>> S/[(1, 0)] == Q.submodule([5, x])
True
Attributes:
base - base module we are quotient of
killed module - submodule used to form the quotient
is full module()
Return True if self is the entire free module.
>>> from sympy.abc import x
>>> from sympy import QQ
>>> F = QQ.old_poly_ring(x).free_module(2)
>>> F.submodule([x, 1]).is_full_module()
False
>>> F.submodule([1, 1], [1, 2]).is_full_module()
True
quotient hom()
Return the quotient homomorphism to self.
That is, return the natural map from self.base to self.
>>> from sympy.abc import x
>>> from sympy import QQ
>>> M = (QQ.old_poly_ring(x).free_module(2) / [(1, x)]).submodule([1, 0])
>>> M.quotient_hom()
Matrix([
[1, 0], : <[1, 0], [1, x]> -> <[1, 0] + <[1, x]>, [1, x] + <[1, x]>>
[0, 1]])
1142
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Module Homomorphisms and Syzygies Let M and N be A-modules. A mapping f : M
N satisfying various obvious properties (see [Atiyah69] (page 1911)) is called an A-module
homomorphism. In this case M is called the domain and N the codomain. The set {x M |f (x) =
0} is called the kernel ker(f ), whereas the set {f (x)|x M } is called the image im(f ). The kernel
is a submodule of M , the image is a submodule of N . The homomorphism f is injective if and
only if ker(f ) = 0 and surjective if and only if im(f ) = N . A bijective homomorphism is called
an isomorphism. Equivalently, ker(f ) = 0 and im(f ) = N . (A related notion, which currently
has no special name in the AGCA module, is that of the cokernel, coker(f ) = N /im(f ).)
Suppose now M is an A-module. M is called nitely generated if there exists a surjective
homomorphism An M for some n. If such a morphism f is chosen, the images of the standard
basis of An are called the generators of M . The module ker(f ) is called syzygy module with
respect to the generators. A module is called nitely presented if it is nitely generated with
a nitely generated syzygy module. The class of nitely presented modules is essentially the
largest class we can hope to be able to meaningfully compute in.
It is an important theorem that, for all the rings we are considering, all submodules of nitely
generated modules are nitely generated, and hence nitely generated and nitely presented
modules are the same.
The notion of syzygies, while it may rst seem rather abstract, is actually very computational.
This is because there exist (fairly easy) algorithms for computing them, and more general
questions (kernels, intersections, ...) are often reduced to syzygy computation.
Let us say a few words about the denition of homomorphisms in the AGCA module. Suppose
rst that f : M N is an arbitrary morphism of A-modules. Then if K is a submodule of M , f
naturally denes a new homomorphism g : K N (via g(x) = f (x)), called the restriction of f to
K. If now K contained in the kernel of f , then moreover f denes in a natural homomorphism
g : M /K N (same formula as above!), and we say that f descends to M /K. Similarly, if L
is a submodule of N , there is a natural homomorphism g : M N /L, we say that g factors
through f . Finally, if now L contains the image of f , then there is a natural homomorphism
g : M L (dened, again, by the same formula), and we say g is obtained from f by restriction
of codomain. Observe also that each of these four operations is reversible, in the sense that
given g, one can always (non-uniquely) nd f such that g is obtained from f in the above way.
Note that all modules implemented in AGCA are obtained from free modules by taking a
succession of submodules and quotients. Hence, in order to explain how to dene a homomorphism between arbitrary modules, in light of the above, we need only explain how to
dene homomorphisms of free modules. But, essentially by the denition of free module, a
homomorphism from a free module An to any module M is precisely the same as giving n
elements of M (the images of the standard basis), and giving an element of a free module
Am is precisely the same as giving m elements of A. Hence a homomorphism of free modules
An Am can be specied via a matrix, entirely analogously to the case of vector spaces.
The functions restrict domain etc. of the class Homomorphism can be used to carry out
the operations described above, and homomorphisms of free modules can in principle be
instantiated by hand. Since these operations are so common, there is a convenience function
homomorphism to dene a homomorphism between arbitrary modules via the method outlined
above. It is essentially the only way homomorphisms need ever be created by the user.
sympy.polys.agca.homomorphisms.homomorphism(domain, codomain, matrix)
Create a homomorphism object.
This function tries to build a homomorphism from domain to codomain via the matrix
matrix.
5.16. Polynomials Manipulation Module
1143
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import QQ
>>> from sympy.abc import x
>>> from sympy.polys.agca import homomorphism
>>> R = QQ.old_poly_ring(x)
>>> T = R.free_module(2)
If domain is a free module generated by e1 , . . . , en , then matrix should be an n-element
iterable (b1 , . . . , bn ) where the bi are elements of codomain. The constructed homomorphism is the unique homomorphism sending ei to bi .
>>> F = R.free_module(2)
>>> h = homomorphism(F, T, [[1, x], [x**2, 0]])
>>> h
Matrix([
[1, x**2], : QQ[x]**2 -> QQ[x]**2
[x,
0]])
>>> h([1, 0])
[1, x]
>>> h([0, 1])
[x**2, 0]
>>> h([1, 1])
[x**2 + 1, x]
If domain is a submodule of a free module, them matrix determines a homomoprhism
from the containing free module to codomain, and the homomorphism returned is obtained by restriction to domain.
>>> S = F.submodule([1, 0], [0, x])
>>> homomorphism(S, T, [[1, x], [x**2, 0]])
Matrix([
[1, x**2], : <[1, 0], [0, x]> -> QQ[x]**2
[x,
0]])
If domain is a (sub)quotient N /K, then matrix determines a homomorphism from N
to codomain. If the kernel contains K, this homomorphism descends to domain and is
returned; otherwise an exception is raised.
>>> homomorphism(S/[(1, 0)], T, [0, [x**2, 0]])
Matrix([
[0, x**2], : <[1, 0] + <[1, 0]>, [0, x] + <[1, 0]>, [1, 0] + <[1, 0]>> -> QQ[x]**2
[0,
0]])
>>> homomorphism(S/[(0, x)], T, [0, [x**2, 0]])
Traceback (most recent call last):
...
ValueError: kernel <[1, 0], [0, 0]> must contain sm, got <[0,x]>
Traceback (most recent call last):
...
ValueError: kernel <[1, 0], [0, 0]> must contain sm, got <[0,x]>
Finally, here is the detailed reference of the actual homomorphism class:
class sympy.polys.agca.homomorphisms.ModuleHomomorphism(domain, codomain)
Abstract base class for module homomoprhisms. Do not instantiate.
Instead, use the homomorphism function:
1144
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> from sympy import QQ
>>> from sympy.abc import x
>>> from sympy.polys.agca import homomorphism
>>> F = QQ.old_poly_ring(x).free_module(2)
>>> homomorphism(F, F, [[1, 0], [0, 1]])
Matrix([
[1, 0], : QQ[x]**2 -> QQ[x]**2
[0, 1]])
Attributes:
ring - the ring over which we are considering modules
domain - the domain module
codomain - the codomain module
ker - cached kernel
img - cachd image
Non-implemented methods:
kernel
image
restrict domain
restrict codomain
quotient domain
quotient codomain
apply
mul scalar
compose
add
image()
Compute the image of self.
That is, if self is the homomorphism : M N , then compute im() = {(x)|x M }.
This is a submodule of N .
>>> from sympy import QQ
>>> from sympy.abc import x
>>> from sympy.polys.agca import homomorphism
>>> F = QQ.old_poly_ring(x).free_module(2)
>>> homomorphism(F, F, [[1, 0], [x, 0]]).image() == F.submodule([1, 0])
True
is injective()
Return True if self is injective.
That is, check if the elements of the domain are mapped to the same codomain
element.
>>> from sympy import QQ
>>> from sympy.abc import x
>>> from sympy.polys.agca import homomorphism
5.16. Polynomials Manipulation Module
1145
SymPy Documentation, Release 0.7.6
>>> F = QQ.old_poly_ring(x).free_module(2)
>>> h = homomorphism(F, F, [[1, 0], [x, 0]])
>>> h.is_injective()
False
>>> h.quotient_domain(h.kernel()).is_injective()
True
is isomorphism()
Return True if self is an isomorphism.
That is, check if every element of the codomain has precisely one preimage. Equivalently, self is both injective and surjective.
>>> from sympy import QQ
>>> from sympy.abc import x
>>> from sympy.polys.agca import homomorphism
>>> F = QQ.old_poly_ring(x).free_module(2)
>>> h = homomorphism(F, F, [[1, 0], [x, 0]])
>>> h = h.restrict_codomain(h.image())
>>> h.is_isomorphism()
False
>>> h.quotient_domain(h.kernel()).is_isomorphism()
True
is surjective()
Return True if self is surjective.
That is, check if every element of the codomain has at least one preimage.
>>> from sympy import QQ
>>> from sympy.abc import x
>>> from sympy.polys.agca import homomorphism
>>> F = QQ.old_poly_ring(x).free_module(2)
>>> h = homomorphism(F, F, [[1, 0], [x, 0]])
>>> h.is_surjective()
False
>>> h.restrict_codomain(h.image()).is_surjective()
True
is zero()
Return True if self is a zero morphism.
That is, check if every element of the domain is mapped to zero under self.
>>> from sympy import QQ
>>> from sympy.abc import x
>>> from sympy.polys.agca import homomorphism
>>> F = QQ.old_poly_ring(x).free_module(2)
>>> h = homomorphism(F, F, [[1, 0], [x, 0]])
>>> h.is_zero()
False
>>> h.restrict_domain(F.submodule()).is_zero()
True
>>> h.quotient_codomain(h.image()).is_zero()
True
1146
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
kernel()
Compute the kernel of self.
That is, if self is the homomorphism : M N , then compute ker() = {x M |(x) =
0}. This is a submodule of M .
>>> from sympy import QQ
>>> from sympy.abc import x
>>> from sympy.polys.agca import homomorphism
>>> F = QQ.old_poly_ring(x).free_module(2)
>>> homomorphism(F, F, [[1, 0], [x, 0]]).kernel()
<[x, -1]>
quotient codomain(sm)
Return self with codomain replaced by codomain/sm.
Here sm must be a submodule of self.codomain.
>>> from sympy import QQ
>>> from sympy.abc import x
>>> from sympy.polys.agca import homomorphism
>>> F = QQ.old_poly_ring(x).free_module(2)
>>> h = homomorphism(F, F, [[1, 0], [x, 0]])
>>> h
Matrix([
[1, x], : QQ[x]**2 -> QQ[x]**2
[0, 0]])
>>> h.quotient_codomain(F.submodule([1, 1]))
Matrix([
[1, x], : QQ[x]**2 -> QQ[x]**2/<[1, 1]>
[0, 0]])
This is the same as composing with the quotient map on the left:
>>> (F/[(1, 1)]).quotient_hom() * h
Matrix([
[1, x], : QQ[x]**2 -> QQ[x]**2/<[1, 1]>
[0, 0]])
quotient domain(sm)
Return self with domain replaced by domain/sm.
Here sm must be a submodule of self.kernel().
>>> from sympy import QQ
>>> from sympy.abc import x
>>> from sympy.polys.agca import homomorphism
>>> F = QQ.old_poly_ring(x).free_module(2)
>>> h = homomorphism(F, F, [[1, 0], [x, 0]])
>>> h
Matrix([
[1, x], : QQ[x]**2 -> QQ[x]**2
[0, 0]])
>>> h.quotient_domain(F.submodule([-x, 1]))
Matrix([
[1, x], : QQ[x]**2/<[-x, 1]> -> QQ[x]**2
[0, 0]])
5.16. Polynomials Manipulation Module
1147
SymPy Documentation, Release 0.7.6
restrict codomain(sm)
Return self, with codomain restricted to to sm.
Here sm has to be a submodule of self.codomain containing the image.
>>> from sympy import QQ
>>> from sympy.abc import x
>>> from sympy.polys.agca import homomorphism
>>> F = QQ.old_poly_ring(x).free_module(2)
>>> h = homomorphism(F, F, [[1, 0], [x, 0]])
>>> h
Matrix([
[1, x], : QQ[x]**2 -> QQ[x]**2
[0, 0]])
>>> h.restrict_codomain(F.submodule([1, 0]))
Matrix([
[1, x], : QQ[x]**2 -> <[1, 0]>
[0, 0]])
restrict domain(sm)
Return self, with the domain restricted to sm.
Here sm has to be a submodule of self.domain.
>>> from sympy import QQ
>>> from sympy.abc import x
>>> from sympy.polys.agca import homomorphism
>>> F = QQ.old_poly_ring(x).free_module(2)
>>> h = homomorphism(F, F, [[1, 0], [x, 0]])
>>> h
Matrix([
[1, x], : QQ[x]**2 -> QQ[x]**2
[0, 0]])
>>> h.restrict_domain(F.submodule([1, 0]))
Matrix([
[1, x], : <[1, 0]> -> QQ[x]**2
[0, 0]])
This is the same as just composing on the right with the submodule inclusion:
>>> h * F.submodule([1, 0]).inclusion_hom()
Matrix([
[1, x], : <[1, 0]> -> QQ[x]**2
[0, 0]])
Internals of the Polynomial Manipulation Module
The implementation of the polynomials module is structured internally in levels. There
are four levels, called L0, L1, L2 and L3. The levels three and four contain the user-facing
functionality and were described in the previous section. This section focuses on levels zero
and one.
Level zero provides core polynomial manipulation functionality with C-like, low-level interfaces. Level one wraps this low-level functionality into object oriented structures. These
are not the classes seen by the user, but rather classes used internally throughout the polys
module.
1148
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
There is one additional complication in the implementation. This comes from the fact that
all polynomial manipulations are relative to a ground domain. For example, when factoring
a polynomial like x10 1, one has to decide what ring the coecients are supposed to belong
to, or less trivially, what coecients are allowed to appear in the factorization. This choice
of coecients is called a ground domain. Typical choices include the integers Z, the rational
numbers Q or various related rings and elds. But it is perfectly legitimate (although in this
case uninteresting) to factorize over polynomial rings such as k[Y ], where k is some xed eld.
Thus the polynomial manipulation algorithms (both complicated ones like factoring, and simpler ones like addition or multiplication) have to rely on other code to manipulate the coecients. In the polynomial manipulation module, such code is encapsulated in so-called
domains. A domain is basically a factory object: it takes various representations of data,
and converts them into objects with unied interface. Every object created by a domain has
to implement the arithmetic operations +, and . Other operations are accessed through
the domain, e.g. as in ZZ.quo(ZZ(4), ZZ(2)).
Note that there is some amount of circularity: the polynomial ring domains use the level
one classes, the level one classes use the level zero functions, and level zero functions use
domains. It is possible, in principle, but not in the current implementation, to work in rings
like k[X][Y ]. This would create even more layers. For this reason, working in the isomorphic
ring k[X, Y ] is preferred.
Domains
Here we document the various implemented ground domains. There are three types: abstract
domains, concrete domains, and implementation domains. Abstract domains cannot be
(usefully) instantiated at all, and just collect together functionality shared by many other
domains. Concrete domains are those meant to be instantiated and used in the polynomial
manipulation algorithms. In some cases, there are various possible ways to implement the
data type the domain provides. For example, depending on what libraries are available on the
system, the integers are implemented either using the python built-in integers, or using gmpy.
Note that various aliases are created automatically depending on the libraries available. As
such e.g. ZZ always refers to the most ecient implementation of the integer ring available.
Abstract Domains
class sympy.polys.domains.domain.Domain
Represents an abstract domain.
Attributes
alias
dtype
one
rep
zero
abs(a)
Absolute value of a, implies
add(a, b)
Sum of a and b, implies
abs .
add .
algebraic field(*extension)
Returns an algebraic eld, i.e. K(, . . . ).
5.16. Polynomials Manipulation Module
1149
SymPy Documentation, Release 0.7.6
almosteq(a, b, tolerance=None)
Check if a and b are almost equal.
characteristic()
Return the characteristic of this domain.
cofactors(a, b)
Returns GCD and cofactors of a and b.
convert(element, base=None)
Convert element to self.dtype.
convert from(element, base)
Convert element to self.dtype given the base domain.
denom(a)
Returns denominator of a.
div(a, b)
Division of a and b, implies something.
evalf(a, prec=None, **options)
Returns numerical approximation of a.
exquo(a, b)
Exact quotient of a and b, implies something.
frac field(*symbols, **kwargs)
Returns a fraction eld, i.e. K(X).
from AlgebraicField(K1, a, K0)
Convert an algebraic number to dtype.
from ComplexField(K1, a, K0)
Convert a complex element to dtype.
from ExpressionDomain(K1, a, K0)
Convert a EX object to dtype.
from FF gmpy(K1, a, K0)
Convert ModularInteger(mpz) to dtype.
from FF python(K1, a, K0)
Convert ModularInteger(int) to dtype.
from FractionField(K1, a, K0)
Convert a rational function to dtype.
from GlobalPolynomialRing(K1, a, K0)
Convert a polynomial to dtype.
from PolynomialRing(K1, a, K0)
Convert a polynomial to dtype.
from QQ gmpy(K1, a, K0)
Convert a GMPY mpq object to dtype.
from QQ python(K1, a, K0)
Convert a Python Fraction object to dtype.
from RealField(K1, a, K0)
Convert a real element object to dtype.
from ZZ gmpy(K1, a, K0)
Convert a GMPY mpz object to dtype.
1150
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
from ZZ python(K1, a, K0)
Convert a Python int object to dtype.
from sympy(a)
Convert a SymPy object to dtype.
gcd(a, b)
Returns GCD of a and b.
gcdex(a, b)
Extended GCD of a and b.
get exact()
Returns an exact domain associated with self.
get field()
Returns a eld associated with self.
get ring()
Returns a ring associated with self.
half gcdex(a, b)
Half extended GCD of a and b.
inject(*symbols)
Inject generators into this domain.
invert(a, b)
Returns inversion of a mod b, implies something.
is negative(a)
Returns True if a is negative.
is nonnegative(a)
Returns True if a is non-negative.
is nonpositive(a)
Returns True if a is non-positive.
is one(a)
Returns True if a is one.
is positive(a)
Returns True if a is positive.
is zero(a)
Returns True if a is zero.
lcm(a, b)
Returns LCM of a and b.
log(a, b)
Returns b-base logarithm of a.
map(seq)
Rersively apply self to all elements of seq.
mul(a, b)
Product of a and b, implies
mul .
n(a, prec=None, **options)
Returns numerical approximation of a.
neg(a)
Returns a negated, implies
neg .
5.16. Polynomials Manipulation Module
1151
SymPy Documentation, Release 0.7.6
numer(a)
Returns numerator of a.
of type(element)
Check if a is of type dtype.
old frac field(*symbols, **kwargs)
Returns a fraction eld, i.e. K(X).
old poly ring(*symbols, **kwargs)
Returns a polynomial ring, i.e. K[X].
poly ring(*symbols, **kwargs)
Returns a polynomial ring, i.e. K[X].
pos(a)
Returns a positive, implies
pos .
pow(a, b)
Raise a to power b, implies
pow .
quo(a, b)
Quotient of a and b, implies something.
rem(a, b)
Remainder of a and b, implies
mod .
revert(a)
Returns a**(-1) if possible.
sqrt(a)
Returns square root of a.
sub(a, b)
Dierence of a and b, implies
sub .
to sympy(a)
Convert a to a SymPy object.
unify(K0, K1, symbols=None)
Construct a minimal domain that contains elements of K0 and K1.
Known domains (from smallest to largest):
GF(p)
ZZ
QQ
RR(prec, tol)
CC(prec, tol)
ALG(a, b, c)
K[x, y, z]
K(x, y, z)
EX
class sympy.polys.domains.field.Field
Represents a eld domain.
1152
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Attributes
alias
dtype
one
rep
zero
div(a, b)
Division of a and b, implies
div .
exquo(a, b)
Exact quotient of a and b, implies
div .
gcd(a, b)
Returns GCD of a and b.
This denition of GCD over elds allows to clear denominators in primitive().
>>> from sympy.polys.domains import QQ
>>> from sympy import S, gcd, primitive
>>> from sympy.abc import x
>>> QQ.gcd(QQ(2, 3), QQ(4, 9))
2/9
>>> gcd(S(2)/3, S(4)/9)
2/9
>>> primitive(2*x/3 + S(4)/9)
(2/9, 3*x + 2)
get field()
Returns a eld associated with self.
get ring()
Returns a ring associated with self.
lcm(a, b)
Returns LCM of a and b.
>>> from sympy.polys.domains import QQ
>>> from sympy import S, lcm
>>> QQ.lcm(QQ(2, 3), QQ(4, 9))
4/3
>>> lcm(S(2)/3, S(4)/9)
4/3
quo(a, b)
Quotient of a and b, implies
div .
rem(a, b)
Remainder of a and b, implies nothing.
revert(a)
Returns a**(-1) if possible.
class sympy.polys.domains.ring.Ring
Represents a ring domain.
5.16. Polynomials Manipulation Module
1153
SymPy Documentation, Release 0.7.6
Attributes
alias
dtype
one
rep
zero
denom(a)
Returns denominator of a.
div(a, b)
Division of a and b, implies
divmod .
exquo(a, b)
Exact quotient of a and b, implies
floordiv .
free module(rank)
Generate a free module of rank rank over self.
>>> from sympy.abc import x
>>> from sympy import QQ
>>> QQ.old_poly_ring(x).free_module(2)
QQ[x]**2
get ring()
Returns a ring associated with self.
ideal(*gens)
Generate an ideal of self.
>>> from sympy.abc import x
>>> from sympy import QQ
>>> QQ.old_poly_ring(x).ideal(x**2)
<x**2>
invert(a, b)
Returns inversion of a mod b.
numer(a)
Returns numerator of a.
quo(a, b)
Quotient of a and b, implies
floordiv .
quotient ring(e)
Form a quotient ring of self.
Here e can be an ideal or an iterable.
>>> from sympy.abc import x
>>> from sympy import QQ
>>> QQ.old_poly_ring(x).quotient_ring(QQ.old_poly_ring(x).ideal(x**2))
QQ[x]/<x**2>
>>> QQ.old_poly_ring(x).quotient_ring([x**2])
QQ[x]/<x**2>
The division operator has been overloaded for this:
>>> QQ.old_poly_ring(x)/[x**2]
QQ[x]/<x**2>
1154
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
rem(a, b)
Remainder of a and b, implies
mod .
revert(a)
Returns a**(-1) if possible.
class sympy.polys.domains.simpledomain.SimpleDomain
Base class for simple domains, e.g. ZZ, QQ.
Attributes
alias
dtype
one
rep
zero
inject(*gens)
Inject generators into this domain.
class sympy.polys.domains.compositedomain.CompositeDomain
Base class for composite domains, e.g. ZZ[x], ZZ(X).
Attributes
alias
domain
dtype
gens
ngens
one
rep
symbols
zero
inject(*symbols)
Inject generators into this domain.
Concrete Domains
class sympy.polys.domains.FiniteField(mod, dom=None, symmetric=True)
General class for nite elds.
Attributes
alias
dom
dtype
mod
one
zero
characteristic()
Return the characteristic of this domain.
5.16. Polynomials Manipulation Module
1155
SymPy Documentation, Release 0.7.6
from FF gmpy(K1, a, K0=None)
Convert ModularInteger(mpz) to dtype.
from FF python(K1, a, K0=None)
Convert ModularInteger(int) to dtype.
from QQ gmpy(K1, a, K0=None)
Convert GMPYs mpq to dtype.
from QQ python(K1, a, K0=None)
Convert Pythons Fraction to dtype.
from RealField(K1, a, K0)
Convert mpmaths mpf to dtype.
from ZZ gmpy(K1, a, K0=None)
Convert GMPYs mpz to dtype.
from ZZ python(K1, a, K0=None)
Convert Pythons int to dtype.
from sympy(a)
Convert SymPys Integer to SymPys Integer.
get field()
Returns a eld associated with self.
to sympy(a)
Convert a to a SymPy object.
class sympy.polys.domains.IntegerRing
General class for integer rings.
Attributes
alias
dtype
one
zero
algebraic field(*extension)
Returns an algebraic eld, i.e. Q(, . . . ).
from AlgebraicField(K1, a, K0)
Convert a ANP object to dtype.
get field()
Returns a eld associated with self.
log(a, b)
Returns b-base logarithm of a.
class sympy.polys.domains.PolynomialRing(domain or ring,
der=None)
A class for representing multivariate polynomial rings.
1156
symbols=None,
or-
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Attributes
alias
domain
dtype
gens
ngens
rep
symbols
factorial(a)
Returns factorial of a.
from AlgebraicField(K1, a, K0)
Convert an algebraic number to dtype.
from FractionField(K1, a, K0)
Convert a rational function to dtype.
from PolynomialRing(K1, a, K0)
Convert a polynomial to dtype.
from QQ gmpy(K1, a, K0)
Convert a GMPY mpq object to dtype.
from QQ python(K1, a, K0)
Convert a Python F raction object to dtype.
from RealField(K1, a, K0)
Convert a mpmath mpf object to dtype.
from ZZ gmpy(K1, a, K0)
Convert a GMPY mpz object to dtype.
from ZZ python(K1, a, K0)
Convert a Python int object to dtype.
from sympy(a)
Convert SymPys expression to dtype.
gcd(a, b)
Returns GCD of a and b.
gcdex(a, b)
Extended GCD of a and b.
get field()
Returns a eld associated with self .
is negative(a)
Returns True if LC(a) is negative.
is nonnegative(a)
Returns True if LC(a) is non-negative.
is nonpositive(a)
Returns True if LC(a) is non-positive.
is positive(a)
Returns True if LC(a) is positive.
lcm(a, b)
Returns LCM of a and b.
5.16. Polynomials Manipulation Module
1157
SymPy Documentation, Release 0.7.6
to sympy(a)
Convert a to a SymPy object.
class sympy.polys.domains.RationalField
General class for rational elds.
Attributes
alias
dtype
one
zero
algebraic field(*extension)
Returns an algebraic eld, i.e. Q(, . . . ).
from AlgebraicField(K1, a, K0)
Convert a ANP object to dtype.
class sympy.polys.domains.AlgebraicField(dom, *ext)
A class for representing algebraic number elds.
Attributes
alias
one
rep
zero
algebraic field(*extension)
Returns an algebraic eld, i.e. Q(, . . . ).
denom(a)
Returns denominator of a.
dtype
alias of ANP
from QQ gmpy(K1, a, K0)
Convert a GMPY mpq object to dtype.
from QQ python(K1, a, K0)
Convert a Python Fraction object to dtype.
from RealField(K1, a, K0)
Convert a mpmath mpf object to dtype.
from ZZ gmpy(K1, a, K0)
Convert a GMPY mpz object to dtype.
from ZZ python(K1, a, K0)
Convert a Python int object to dtype.
from sympy(a)
Convert SymPys expression to dtype.
get ring()
Returns a ring associated with self.
1158
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
is negative(a)
Returns True if a is negative.
is nonnegative(a)
Returns True if a is non-negative.
is nonpositive(a)
Returns True if a is non-positive.
is positive(a)
Returns True if a is positive.
numer(a)
Returns numerator of a.
to sympy(a)
Convert a to a SymPy object.
class sympy.polys.domains.FractionField(domain or eld,
symbols=None,
der=None)
A class for representing multivariate rational function elds.
or-
Attributes
alias
domain
dtype
gens
ngens
rep
symbols
denom(a)
Returns denominator of a.
factorial(a)
Returns factorial of a.
from AlgebraicField(K1, a, K0)
Convert an algebraic number to dtype.
from FractionField(K1, a, K0)
Convert a rational function to dtype.
from PolynomialRing(K1, a, K0)
Convert a polynomial to dtype.
from QQ gmpy(K1, a, K0)
Convert a GMPY mpq object to dtype.
from QQ python(K1, a, K0)
Convert a Python F raction object to dtype.
from RealField(K1, a, K0)
Convert a mpmath mpf object to dtype.
from ZZ gmpy(K1, a, K0)
Convert a GMPY mpz object to dtype.
from ZZ python(K1, a, K0)
Convert a Python int object to dtype.
5.16. Polynomials Manipulation Module
1159
SymPy Documentation, Release 0.7.6
from sympy(a)
Convert SymPys expression to dtype.
get ring()
Returns a eld associated with self .
is negative(a)
Returns True if LC(a) is negative.
is nonnegative(a)
Returns True if LC(a) is non-negative.
is nonpositive(a)
Returns True if LC(a) is non-positive.
is positive(a)
Returns True if LC(a) is positive.
numer(a)
Returns numerator of a.
to sympy(a)
Convert a to a SymPy object.
class sympy.polys.domains.RealField(prec=53, dps=None, tol=None)
Real numbers up to the given precision.
Attributes
alias
dtype
one
zero
almosteq(a, b, tolerance=None)
Check if a and b are almost equal.
from sympy(expr)
Convert SymPys number to dtype.
gcd(a, b)
Returns GCD of a and b.
get exact()
Returns an exact domain associated with self.
get ring()
Returns a ring associated with self.
lcm(a, b)
Returns LCM of a and b.
to rational(element, limit=True)
Convert a real number to rational number.
to sympy(element)
Convert element to SymPy number.
class sympy.polys.domains.ExpressionDomain
A class for arbitrary expressions.
1160
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Attributes
alias
class Expression(ex)
An arbitrary expression.
ExpressionDomain.denom(a)
Returns denominator of a.
ExpressionDomain.dtype
alias of Expression (page 1161)
ExpressionDomain.from ExpressionDomain(K1, a, K0)
Convert a EX object to dtype.
ExpressionDomain.from FractionField(K1, a, K0)
Convert a DMF object to dtype.
ExpressionDomain.from PolynomialRing(K1, a, K0)
Convert a DMP object to dtype.
ExpressionDomain.from QQ gmpy(K1, a, K0)
Convert a GMPY mpq object to dtype.
ExpressionDomain.from QQ python(K1, a, K0)
Convert a Python Fraction object to dtype.
ExpressionDomain.from RealField(K1, a, K0)
Convert a mpmath mpf object to dtype.
ExpressionDomain.from ZZ gmpy(K1, a, K0)
Convert a GMPY mpz object to dtype.
ExpressionDomain.from ZZ python(K1, a, K0)
Convert a Python int object to dtype.
ExpressionDomain.from sympy(a)
Convert SymPys expression to dtype.
ExpressionDomain.get field()
Returns a eld associated with self.
ExpressionDomain.get ring()
Returns a ring associated with self.
ExpressionDomain.is negative(a)
Returns True if a is negative.
ExpressionDomain.is nonnegative(a)
Returns True if a is non-negative.
ExpressionDomain.is nonpositive(a)
Returns True if a is non-positive.
ExpressionDomain.is positive(a)
Returns True if a is positive.
ExpressionDomain.numer(a)
Returns numerator of a.
ExpressionDomain.to sympy(a)
Convert a to a SymPy object.
5.16. Polynomials Manipulation Module
1161
SymPy Documentation, Release 0.7.6
Implementation Domains
class sympy.polys.domains.PythonFiniteField(mod, symmetric=True)
Finite eld based on Pythons integers.
Attributes
dom
dtype
mod
one
zero
class sympy.polys.domains.GMPYFiniteField(mod, symmetric=True)
Finite eld based on GMPY integers.
Attributes
dom
dtype
mod
one
zero
class sympy.polys.domains.PythonIntegerRing
Integer ring based on Pythons int type.
class sympy.polys.domains.GMPYIntegerRing
Integer ring based on GMPYs mpz type.
class sympy.polys.domains.PythonRationalField
Rational eld based on Python rational number type.
class sympy.polys.domains.GMPYRationalField
Rational eld based on GMPY mpq class.
Level One
class sympy.polys.polyclasses.DMP(rep, dom, lev=None, ring=None)
Dense Multivariate Polynomials over K.
LC(f)
Returns the leading coecient of f.
TC(f)
Returns the trailing coecient of f.
abs(f)
Make all coecients in f positive.
add(f, g)
Add two multivariate polynomials f and g.
add ground(f, c)
Add an element of the ground domain to f.
all coeffs(f)
Returns all coecients from f.
1162
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
all monoms(f)
Returns all monomials from f.
all terms(f)
Returns all terms from a f.
cancel(f, g, include=True)
Cancel common factors in a rational function f/g.
clear denoms(f)
Clear denominators, but keep the ground domain.
coeffs(f, order=None)
Returns all non-zero coecients from f in lex order.
cofactors(f, g)
Returns GCD of f and g and their cofactors.
compose(f, g)
Computes functional composition of f and g.
content(f)
Returns GCD of polynomial coecients.
convert(f, dom)
Convert the ground domain of f.
count complex roots(f, inf=None, sup=None)
Return the number of complex roots of f in [inf, sup].
count real roots(f, inf=None, sup=None)
Return the number of real roots of f in [inf, sup].
decompose(f)
Computes functional decomposition of f.
deflate(f)
Reduce degree of f by mapping xm
i to yi .
degree(f, j=0)
Returns the leading degree of f in x j.
degree list(f)
Returns a list of degrees of f.
diff(f, m=1, j=0)
Computes the m-th order derivative of f in x j.
discriminant(f)
Computes discriminant of f.
div(f, g)
Polynomial division with remainder of f and g.
eject(f, dom, front=False)
Eject selected generators into the ground domain.
eval(f, a, j=0)
Evaluates f at the given point a in x j.
exclude(f)
Remove useless generators from f.
Returns the removed generators and the new excluded f.
5.16. Polynomials Manipulation Module
1163
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.polys.polyclasses import DMP
>>> from sympy.polys.domains import ZZ
>>> DMP([[[ZZ(1)]], [[ZZ(1)], [ZZ(2)]]], ZZ).exclude()
([2], DMP([[1], [1, 2]], ZZ, None))
exquo(f, g)
Computes polynomial exact quotient of f and g.
exquo ground(f, c)
Exact quotient of f by a an element of the ground domain.
factor list(f)
Returns a list of irreducible factors of f.
factor list include(f)
Returns a list of irreducible factors of f.
classmethod from dict(rep, lev, dom)
Construct and instance of cls from a dict representation.
classmethod from list(rep, lev, dom)
Create an instance of cls given a list of native coecients.
classmethod from sympy list(rep, lev, dom)
Create an instance of cls given a list of SymPy coecients.
gcd(f, g)
Returns polynomial GCD of f and g.
gcdex(f, g)
Extended Euclidean algorithm, if univariate.
gff list(f)
Computes greatest factorial factorization of f.
half gcdex(f, g)
Half extended Euclidean algorithm, if univariate.
homogeneous order(f)
Returns the homogeneous order of f.
homogenize(f, s)
Return homogeneous polynomial of f
inject(f, front=False)
Inject ground domain generators into f.
integrate(f, m=1, j=0)
Computes the m-th order indenite integral of f in x j.
intervals(f, all=False, eps=None, inf=None, sup=None, fast=False, sqf=False)
Compute isolating intervals for roots of f.
invert(f, g)
Invert f modulo g, if possible.
is cyclotomic
Returns True if f is a cyclotomic polynomial.
is ground
Returns True if f is an element of the ground domain.
1164
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
is homogeneous
Returns True if f is a homogeneous polynomial.
is irreducible
Returns True if f has no factors over its domain.
is linear
Returns True if f is linear in all its variables.
is monic
Returns True if the leading coecient of f is one.
is monomial
Returns True if f is zero or has only one term.
is one
Returns True if f is a unit polynomial.
is primitive
Returns True if the GCD of the coecients of f is one.
is quadratic
Returns True if f is quadratic in all its variables.
is sqf
Returns True if f is a square-free polynomial.
is zero
Returns True if f is a zero polynomial.
l1 norm(f)
Returns l1 norm of f.
lcm(f, g)
Returns polynomial LCM of f and g.
lift(f)
Convert algebraic coecients to rationals.
max norm(f)
Returns maximum norm of f.
monic(f)
Divides all coecients by LC(f).
monoms(f, order=None)
Returns all non-zero monomials from f in lex order.
mul(f, g)
Multiply two multivariate polynomials f and g.
mul ground(f, c)
Multiply f by a an element of the ground domain.
neg(f)
Negate all coecients in f.
nth(f, *N)
Returns the n-th coecient of f.
pdiv(f, g)
Polynomial pseudo-division of f and g.
per(f, rep, dom=None, kill=False, ring=None)
Create a DMP out of the given representation.
5.16. Polynomials Manipulation Module
1165
SymPy Documentation, Release 0.7.6
permute(f, P)
Returns a polynomial in K[xP (1) , ..., xP (n) ].
Examples
>>> from sympy.polys.polyclasses import DMP
>>> from sympy.polys.domains import ZZ
>>> DMP([[[ZZ(2)], [ZZ(1), ZZ(0)]], [[]]], ZZ).permute([1, 0, 2])
DMP([[[2], []], [[1, 0], []]], ZZ, None)
>>> DMP([[[ZZ(2)], [ZZ(1), ZZ(0)]], [[]]], ZZ).permute([1, 2, 0])
DMP([[[1], []], [[2, 0], []]], ZZ, None)
pexquo(f, g)
Polynomial exact pseudo-quotient of f and g.
pow(f, n)
Raise f to a non-negative power n.
pquo(f, g)
Polynomial pseudo-quotient of f and g.
prem(f, g)
Polynomial pseudo-remainder of f and g.
primitive(f)
Returns content and a primitive form of f.
quo(f, g)
Computes polynomial quotient of f and g.
quo ground(f, c)
Quotient of f by a an element of the ground domain.
refine root(f, s, t, eps=None, steps=None, fast=False)
Rene an isolating interval to the given precision.
eps should be a rational number.
rem(f, g)
Computes polynomial remainder of f and g.
resultant(f, g, includePRS=False)
Computes resultant of f and g via PRS.
revert(f, n)
Compute f**(-1) mod x**n.
shift(f, a)
Eciently compute Taylor shift f(x + a).
slice(f, m, n, j=0)
Take a continuous subsequence of terms of f.
sqf list(f, all=False)
Returns a list of square-free factors of f.
sqf list include(f, all=False)
Returns a list of square-free factors of f.
1166
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
sqf norm(f)
Computes square-free norm of f.
sqf part(f)
Computes square-free part of f.
sqr(f)
Square a multivariate polynomial f.
sturm(f)
Computes the Sturm sequence of f.
sub(f, g)
Subtract two multivariate polynomials f and g.
sub ground(f, c)
Subtract an element of the ground domain from f.
subresultants(f, g)
Computes subresultant PRS sequence of f and g.
terms(f, order=None)
Returns all non-zero terms from f in lex order.
terms gcd(f)
Remove GCD of terms from the polynomial f.
to dict(f, zero=False)
Convert f to a dict representation with native coecients.
to exact(f)
Make the ground domain exact.
to field(f)
Make the ground domain a eld.
to ring(f)
Make the ground domain a ring.
to sympy dict(f, zero=False)
Convert f to a dict representation with SymPy coecients.
to tuple(f)
Convert f to a tuple representation with native coecients.
This is needed for hashing.
total degree(f)
Returns the total degree of f.
trunc(f, p)
Reduce f modulo a constant p.
unify(f, g)
Unify representations of two multivariate polynomials.
class sympy.polys.polyclasses.DMF(rep, dom, lev=None, ring=None)
Dense Multivariate Fractions over K.
add(f, g)
Add two multivariate fractions f and g.
cancel(f)
Remove common factors from f.num and f.den.
denom(f)
Returns the denominator of f.
5.16. Polynomials Manipulation Module
1167
SymPy Documentation, Release 0.7.6
exquo(f, g)
Computes quotient of fractions f and g.
frac unify(f, g)
Unify representations of two multivariate fractions.
half per(f, rep, kill=False)
Create a DMP out of the given representation.
invert(f, check=True)
Computes inverse of a fraction f.
is one
Returns True if f is a unit fraction.
is zero
Returns True if f is a zero fraction.
mul(f, g)
Multiply two multivariate fractions f and g.
neg(f)
Negate all coecients in f.
numer(f)
Returns the numerator of f.
per(f, num, den, cancel=True, kill=False, ring=None)
Create a DMF out of the given representation.
poly unify(f, g)
Unify a multivariate fraction and a polynomial.
pow(f, n)
Raise f to a non-negative power n.
quo(f, g)
Computes quotient of fractions f and g.
sub(f, g)
Subtract two multivariate fractions f and g.
class sympy.polys.polyclasses.ANP(rep, mod, dom)
Dense Algebraic Number Polynomials over a eld.
LC(f)
Returns the leading coecient of f.
TC(f)
Returns the trailing coecient of f.
is ground
Returns True if f is an element of the ground domain.
is one
Returns True if f is a unit algebraic number.
is zero
Returns True if f is a zero algebraic number.
pow(f, n)
Raise f to a non-negative power n.
to dict(f)
Convert f to a dict representation with native coecients.
1168
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
to list(f)
Convert f to a list representation with native coecients.
to sympy dict(f)
Convert f to a dict representation with SymPy coecients.
to sympy list(f)
Convert f to a list representation with SymPy coecients.
to tuple(f)
Convert f to a tuple representation with native coecients.
This is needed for hashing.
unify(f, g)
Unify representations of two algebraic numbers.
Level Zero
Level zero contains the bulk code of the polynomial manipulation module.
Manipulation of dense, multivariate polynomials These functions can be used to manipulate polynomials in K[X0 , . . . , Xu ]. Functions for manipulating multivariate polynomials in
the dense representation have the prex dmp . Functions which only apply to univariate polynomials (i.e. u = 0) have the prex dup . The ground domain K has to be passed explicitly.
For many multivariate polynomial manipulation functions also the level u, i.e. the number of
generators minus one, has to be passed. (Note that, in many cases, dup versions of functions
are available, which may be slightly more ecient.)
Basic manipulation:
sympy.polys.densebasic.dmp LC(f, K)
Return leading coecient of f.
Examples
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.densebasic import poly_LC
>>> poly_LC([], ZZ)
0
>>> poly_LC([ZZ(1), ZZ(2), ZZ(3)], ZZ)
1
sympy.polys.densebasic.dmp TC(f, K)
Return trailing coecient of f.
Examples
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.densebasic import poly_TC
5.16. Polynomials Manipulation Module
1169
SymPy Documentation, Release 0.7.6
>>> poly_TC([], ZZ)
0
>>> poly_TC([ZZ(1), ZZ(2), ZZ(3)], ZZ)
3
sympy.polys.densebasic.dmp ground LC(f, u, K)
Return the ground leading coecient.
Examples
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.densebasic import dmp_ground_LC
>>> f = ZZ.map([[[1], [2, 3]]])
>>> dmp_ground_LC(f, 2, ZZ)
1
sympy.polys.densebasic.dmp ground TC(f, u, K)
Return the ground trailing coecient.
Examples
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.densebasic import dmp_ground_TC
>>> f = ZZ.map([[[1], [2, 3]]])
>>> dmp_ground_TC(f, 2, ZZ)
3
sympy.polys.densebasic.dmp true LT(f, u, K)
Return the leading term c * x 1**n 1 ... x k**n k.
Examples
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.densebasic import dmp_true_LT
>>> f = ZZ.map([[4], [2, 0], [3, 0, 0]])
>>> dmp_true_LT(f, 1, ZZ)
((2, 0), 4)
sympy.polys.densebasic.dmp degree(f, u)
Return the leading degree of f in x 0 in K[X].
Note that the degree of 0 is negative innity (the SymPy object -oo).
1170
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.densebasic import dmp_degree
>>> dmp_degree([[[]]], 2)
-oo
>>> f = ZZ.map([[2], [1, 2, 3]])
>>> dmp_degree(f, 1)
1
sympy.polys.densebasic.dmp degree in(f, j, u)
Return the leading degree of f in x j in K[X].
Examples
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.densebasic import dmp_degree_in
>>> f = ZZ.map([[2], [1, 2, 3]])
>>> dmp_degree_in(f, 0, 1)
1
>>> dmp_degree_in(f, 1, 1)
2
sympy.polys.densebasic.dmp degree list(f, u)
Return a list of degrees of f in K[X].
Examples
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.densebasic import dmp_degree_list
>>> f = ZZ.map([[1], [1, 2, 3]])
>>> dmp_degree_list(f, 1)
(1, 2)
sympy.polys.densebasic.dmp strip(f, u)
Remove leading zeros from f in K[X].
Examples
>>> from sympy.polys.densebasic import dmp_strip
>>> dmp_strip([[], [0, 1, 2], [1]], 1)
[[0, 1, 2], [1]]
sympy.polys.densebasic.dmp validate(f, K=None)
Return the number of levels in f and recursively strip it.
5.16. Polynomials Manipulation Module
1171
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.polys.densebasic import dmp_validate
>>> dmp_validate([[], [0, 1, 2], [1]])
([[1, 2], [1]], 1)
>>> dmp_validate([[1], 1])
Traceback (most recent call last):
...
ValueError: invalid data structure for a multivariate polynomial
Traceback (most recent call last):
...
ValueError: invalid data structure for a multivariate polynomial
sympy.polys.densebasic.dup reverse(f)
Compute x**n * f(1/x), i.e.: reverse f in K[x].
Examples
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.densebasic import dup_reverse
>>> f = ZZ.map([1, 2, 3, 0])
>>> dup_reverse(f)
[3, 2, 1]
sympy.polys.densebasic.dmp copy(f, u)
Create a new copy of a polynomial f in K[X].
Examples
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.densebasic import dmp_copy
>>> f = ZZ.map([[1], [1, 2]])
>>> dmp_copy(f, 1)
[[1], [1, 2]]
sympy.polys.densebasic.dmp to tuple(f, u)
Convert f into a nested tuple of tuples.
This is needed for hashing. This is similar to dmp copy().
Examples
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.densebasic import dmp_to_tuple
1172
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> f = ZZ.map([[1], [1, 2]])
>>> dmp_to_tuple(f, 1)
((1,), (1, 2))
sympy.polys.densebasic.dmp normal(f, u, K)
Normalize a multivariate polynomial in the given domain.
Examples
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.densebasic import dmp_normal
>>> dmp_normal([[], [0, 1.5, 2]], 1, ZZ)
[[1, 2]]
sympy.polys.densebasic.dmp convert(f, u, K0, K1)
Convert the ground domain of f from K0 to K1.
Examples
>>> from sympy.polys.rings import ring
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.densebasic import dmp_convert
>>> R, x = ring(x, ZZ)
>>> dmp_convert([[R(1)], [R(2)]], 1, R.to_domain(), ZZ)
[[1], [2]]
>>> dmp_convert([[ZZ(1)], [ZZ(2)]], 1, ZZ, R.to_domain())
[[1], [2]]
sympy.polys.densebasic.dmp from sympy(f, u, K)
Convert the ground domain of f from SymPy to K.
Examples
>>> from sympy import S
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.densebasic import dmp_from_sympy
>>> dmp_from_sympy([[S(1)], [S(2)]], 1, ZZ) == [[ZZ(1)], [ZZ(2)]]
True
sympy.polys.densebasic.dmp nth(f, n, u, K)
Return the n-th coecient of f in K[x].
Examples
5.16. Polynomials Manipulation Module
1173
SymPy Documentation, Release 0.7.6
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.densebasic import dmp_nth
>>> f = ZZ.map([[1], [2], [3]])
>>> dmp_nth(f, 0, 1, ZZ)
[3]
>>> dmp_nth(f, 4, 1, ZZ)
[]
sympy.polys.densebasic.dmp ground nth(f, N, u, K)
Return the ground n-th coecient of f in K[x].
Examples
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.densebasic import dmp_ground_nth
>>> f = ZZ.map([[1], [2, 3]])
>>> dmp_ground_nth(f, (0, 1), 1, ZZ)
2
sympy.polys.densebasic.dmp zero p(f, u)
Return True if f is zero in K[X].
Examples
>>> from sympy.polys.densebasic import dmp_zero_p
>>> dmp_zero_p([[[[[]]]]], 4)
True
>>> dmp_zero_p([[[[[1]]]]], 4)
False
sympy.polys.densebasic.dmp zero(u)
Return a multivariate zero.
Examples
>>> from sympy.polys.densebasic import dmp_zero
>>> dmp_zero(4)
[[[[[]]]]]
sympy.polys.densebasic.dmp one p(f, u, K)
Return True if f is one in K[X].
1174
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.densebasic import dmp_one_p
>>> dmp_one_p([[[ZZ(1)]]], 2, ZZ)
True
sympy.polys.densebasic.dmp one(u, K)
Return a multivariate one over K.
Examples
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.densebasic import dmp_one
>>> dmp_one(2, ZZ)
[[[1]]]
sympy.polys.densebasic.dmp ground p(f, c, u)
Return True if f is constant in K[X].
Examples
>>> from sympy.polys.densebasic import dmp_ground_p
>>> dmp_ground_p([[[3]]], 3, 2)
True
>>> dmp_ground_p([[[4]]], None, 2)
True
sympy.polys.densebasic.dmp ground(c, u)
Return a multivariate constant.
Examples
>>> from sympy.polys.densebasic import dmp_ground
>>> dmp_ground(3, 5)
[[[[[[3]]]]]]
>>> dmp_ground(1, -1)
1
sympy.polys.densebasic.dmp zeros(n, u, K)
Return a list of multivariate zeros.
Examples
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.densebasic import dmp_zeros
5.16. Polynomials Manipulation Module
1175
SymPy Documentation, Release 0.7.6
>>> dmp_zeros(3, 2, ZZ)
[[[[]]], [[[]]], [[[]]]]
>>> dmp_zeros(3, -1, ZZ)
[0, 0, 0]
sympy.polys.densebasic.dmp grounds(c, n, u)
Return a list of multivariate constants.
Examples
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.densebasic import dmp_grounds
>>> dmp_grounds(ZZ(4), 3, 2)
[[[[4]]], [[[4]]], [[[4]]]]
>>> dmp_grounds(ZZ(4), 3, -1)
[4, 4, 4]
sympy.polys.densebasic.dmp negative p(f, u, K)
Return True if LC(f) is negative.
Examples
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.densebasic import dmp_negative_p
>>> dmp_negative_p([[ZZ(1)], [-ZZ(1)]], 1, ZZ)
False
>>> dmp_negative_p([[-ZZ(1)], [ZZ(1)]], 1, ZZ)
True
sympy.polys.densebasic.dmp positive p(f, u, K)
Return True if LC(f) is positive.
Examples
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.densebasic import dmp_positive_p
>>> dmp_positive_p([[ZZ(1)], [-ZZ(1)]], 1, ZZ)
True
>>> dmp_positive_p([[-ZZ(1)], [ZZ(1)]], 1, ZZ)
False
sympy.polys.densebasic.dmp from dict(f, u, K)
Create a K[X] polynomial from a dict.
Examples
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.densebasic import dmp_from_dict
1176
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> dmp_from_dict({(0, 0): ZZ(3), (0, 1): ZZ(2), (2, 1): ZZ(1)}, 1, ZZ)
[[1, 0], [], [2, 3]]
>>> dmp_from_dict({}, 0, ZZ)
[]
sympy.polys.densebasic.dmp to dict(f, u, K=None, zero=False)
Convert a K[X] polynomial to a dict.
Examples
>>> from sympy.polys.densebasic import dmp_to_dict
>>> dmp_to_dict([[1, 0], [], [2, 3]], 1)
{(0, 0): 3, (0, 1): 2, (2, 1): 1}
>>> dmp_to_dict([], 0)
{}
sympy.polys.densebasic.dmp swap(f, i, j, u, K)
Transform K[..x i..x j..] to K[..x j..x i..].
Examples
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.densebasic import dmp_swap
>>> f = ZZ.map([[[2], [1, 0]], []])
>>> dmp_swap(f, 0, 1, 2, ZZ)
[[[2], []], [[1, 0], []]]
>>> dmp_swap(f, 1, 2, 2, ZZ)
[[[1], [2, 0]], [[]]]
>>> dmp_swap(f, 0, 2, 2, ZZ)
[[[1, 0]], [[2, 0], []]]
sympy.polys.densebasic.dmp permute(f, P, u, K)
Return a polynomial in K[x {P(1)},..,x {P(n)}].
Examples
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.densebasic import dmp_permute
>>> f = ZZ.map([[[2], [1, 0]], []])
>>> dmp_permute(f, [1, 0, 2], 2, ZZ)
[[[2], []], [[1, 0], []]]
>>> dmp_permute(f, [1, 2, 0], 2, ZZ)
[[[1], []], [[2, 0], []]]
sympy.polys.densebasic.dmp nest(f, l, K)
Return a multivariate value nested l-levels.
5.16. Polynomials Manipulation Module
1177
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.densebasic import dmp_nest
>>> dmp_nest([[ZZ(1)]], 2, ZZ)
[[[[1]]]]
sympy.polys.densebasic.dmp raise(f, l, u, K)
Return a multivariate polynomial raised l-levels.
Examples
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.densebasic import dmp_raise
>>> f = ZZ.map([[], [1, 2]])
>>> dmp_raise(f, 2, 1, ZZ)
[[[[]]], [[[1]], [[2]]]]
sympy.polys.densebasic.dmp deflate(f, u, K)
Map x i**m i to y i in a polynomial in K[X].
Examples
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.densebasic import dmp_deflate
>>> f = ZZ.map([[1, 0, 0, 2], [], [3, 0, 0, 4]])
>>> dmp_deflate(f, 1, ZZ)
((2, 3), [[1, 2], [3, 4]])
sympy.polys.densebasic.dmp multi deflate(polys, u, K)
Map x i**m i to y i in a set of polynomials in K[X].
Examples
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.densebasic import dmp_multi_deflate
>>> f = ZZ.map([[1, 0, 0, 2], [], [3, 0, 0, 4]])
>>> g = ZZ.map([[1, 0, 2], [], [3, 0, 4]])
>>> dmp_multi_deflate((f, g), 1, ZZ)
((2, 1), ([[1, 0, 0, 2], [3, 0, 0, 4]], [[1, 0, 2], [3, 0, 4]]))
sympy.polys.densebasic.dmp inflate(f, M, u, K)
Map y i to x i**k i in a polynomial in K[X].
1178
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.densebasic import dmp_inflate
>>> f = ZZ.map([[1, 2], [3, 4]])
>>> dmp_inflate(f, (2, 3), 1, ZZ)
[[1, 0, 0, 2], [], [3, 0, 0, 4]]
sympy.polys.densebasic.dmp exclude(f, u, K)
Exclude useless levels from f.
Return the levels excluded, the new excluded f, and the new u.
Examples
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.densebasic import dmp_exclude
>>> f = ZZ.map([[[1]], [[1], [2]]])
>>> dmp_exclude(f, 2, ZZ)
([2], [[1], [1, 2]], 1)
sympy.polys.densebasic.dmp include(f, J, u, K)
Include useless levels in f.
Examples
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.densebasic import dmp_include
>>> f = ZZ.map([[1], [1, 2]])
>>> dmp_include(f, [2], 1, ZZ)
[[[1]], [[1], [2]]]
sympy.polys.densebasic.dmp inject(f, u, K, front=False)
Convert f from K[X][Y] to K[X,Y].
Examples
>>> from sympy.polys.rings import ring
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.densebasic import dmp_inject
>>> R, x,y = ring(x,y, ZZ)
5.16. Polynomials Manipulation Module
1179
SymPy Documentation, Release 0.7.6
>>> dmp_inject([R(1), x + 2], 0, R.to_domain())
([[[1]], [[1], [2]]], 2)
>>> dmp_inject([R(1), x + 2], 0, R.to_domain(), front=True)
([[[1]], [[1, 2]]], 2)
sympy.polys.densebasic.dmp eject(f, u, K, front=False)
Convert f from K[X,Y] to K[X][Y].
Examples
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.densebasic import dmp_eject
>>> dmp_eject([[[1]], [[1], [2]]], 2, ZZ[x, y])
[1, x + 2]
sympy.polys.densebasic.dmp terms gcd(f, u, K)
Remove GCD of terms from f in K[X].
Examples
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.densebasic import dmp_terms_gcd
>>> f = ZZ.map([[1, 0], [1, 0, 0], [], []])
>>> dmp_terms_gcd(f, 1, ZZ)
((2, 1), [[1], [1, 0]])
sympy.polys.densebasic.dmp list terms(f, u, K, order=None)
List all non-zero terms from f in the given order order.
Examples
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.densebasic import dmp_list_terms
>>> f = ZZ.map([[1, 1], [2, 3]])
>>> dmp_list_terms(f, 1, ZZ)
[((1, 1), 1), ((1, 0), 1), ((0, 1), 2), ((0, 0), 3)]
>>> dmp_list_terms(f, 1, ZZ, order=grevlex)
[((1, 1), 1), ((1, 0), 1), ((0, 1), 2), ((0, 0), 3)]
sympy.polys.densebasic.dmp apply pairs(f, g, h, args, u, K)
Apply h to pairs of coecients of f and g.
Examples
1180
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.densebasic import dmp_apply_pairs
>>> h = lambda x, y, z: 2*x + y - z
>>> dmp_apply_pairs([[1], [2, 3]], [[3], [2, 1]], h, (1,), 1, ZZ)
[[4], [5, 6]]
sympy.polys.densebasic.dmp slice(f, m, n, u, K)
Take a continuous subsequence of terms of f in K[X].
sympy.polys.densebasic.dup random(n, a, b, K)
Return a polynomial of degree n with coecients in [a, b].
Examples
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.densebasic import dup_random
>>> dup_random(3, -10, 10, ZZ)
[-2, -8, 9, -4]
Arithmetic operations:
sympy.polys.densearith.dmp add term(f, c, i, u, K)
Add c(x 2..x u)*x 0**i to f in K[X].
Examples
>>> from sympy.polys import ring, ZZ
>>> R, x,y = ring(x,y, ZZ)
>>> R.dmp_add_term(x*y + 1, 2, 2)
2*x**2 + x*y + 1
sympy.polys.densearith.dmp sub term(f, c, i, u, K)
Subtract c(x 2..x u)*x 0**i from f in K[X].
Examples
>>> from sympy.polys import ring, ZZ
>>> R, x,y = ring(x,y, ZZ)
>>> R.dmp_sub_term(2*x**2 + x*y + 1, 2, 2)
x*y + 1
sympy.polys.densearith.dmp mul term(f, c, i, u, K)
Multiply f by c(x 2..x u)*x 0**i in K[X].
5.16. Polynomials Manipulation Module
1181
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.polys import ring, ZZ
>>> R, x,y = ring(x,y, ZZ)
>>> R.dmp_mul_term(x**2*y + x, 3*y, 2)
3*x**4*y**2 + 3*x**3*y
sympy.polys.densearith.dmp add ground(f, c, u, K)
Add an element of the ground domain to f.
Examples
>>> from sympy.polys import ring, ZZ
>>> R, x,y = ring(x,y, ZZ)
>>> R.dmp_add_ground(x**3 + 2*x**2 + 3*x + 4, ZZ(4))
x**3 + 2*x**2 + 3*x + 8
sympy.polys.densearith.dmp sub ground(f, c, u, K)
Subtract an element of the ground domain from f.
Examples
>>> from sympy.polys import ring, ZZ
>>> R, x,y = ring(x,y, ZZ)
>>> R.dmp_sub_ground(x**3 + 2*x**2 + 3*x + 4, ZZ(4))
x**3 + 2*x**2 + 3*x
sympy.polys.densearith.dmp mul ground(f, c, u, K)
Multiply f by a constant value in K[X].
Examples
>>> from sympy.polys import ring, ZZ
>>> R, x,y = ring(x,y, ZZ)
>>> R.dmp_mul_ground(2*x + 2*y, ZZ(3))
6*x + 6*y
sympy.polys.densearith.dmp quo ground(f, c, u, K)
Quotient by a constant in K[X].
Examples
>>> from sympy.polys import ring, ZZ, QQ
1182
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> R, x,y = ring(x,y, ZZ)
>>> R.dmp_quo_ground(2*x**2*y + 3*x, ZZ(2))
x**2*y + x
>>> R, x,y = ring(x,y, QQ)
>>> R.dmp_quo_ground(2*x**2*y + 3*x, QQ(2))
x**2*y + 3/2*x
sympy.polys.densearith.dmp exquo ground(f, c, u, K)
Exact quotient by a constant in K[X].
Examples
>>> from sympy.polys import ring, QQ
>>> R, x,y = ring(x,y, QQ)
>>> R.dmp_exquo_ground(x**2*y + 2*x, QQ(2))
1/2*x**2*y + x
sympy.polys.densearith.dup lshift(f, n, K)
Eciently multiply f by x**n in K[x].
Examples
>>> from sympy.polys import ring, ZZ
>>> R, x = ring(x, ZZ)
>>> R.dup_lshift(x**2 + 1, 2)
x**4 + x**2
sympy.polys.densearith.dup rshift(f, n, K)
Eciently divide f by x**n in K[x].
Examples
>>> from sympy.polys import ring, ZZ
>>> R, x = ring(x, ZZ)
>>> R.dup_rshift(x**4 + x**2, 2)
x**2 + 1
>>> R.dup_rshift(x**4 + x**2 + 2, 2)
x**2 + 1
sympy.polys.densearith.dmp abs(f, u, K)
Make all coecients positive in K[X].
Examples
>>> from sympy.polys import ring, ZZ
>>> R, x,y = ring(x,y, ZZ)
5.16. Polynomials Manipulation Module
1183
SymPy Documentation, Release 0.7.6
>>> R.dmp_abs(x**2*y - x)
x**2*y + x
sympy.polys.densearith.dmp neg(f, u, K)
Negate a polynomial in K[X].
Examples
>>> from sympy.polys import ring, ZZ
>>> R, x,y = ring(x,y, ZZ)
>>> R.dmp_neg(x**2*y - x)
-x**2*y + x
sympy.polys.densearith.dmp add(f, g, u, K)
Add dense polynomials in K[X].
Examples
>>> from sympy.polys import ring, ZZ
>>> R, x,y = ring(x,y, ZZ)
>>> R.dmp_add(x**2 + y, x**2*y + x)
x**2*y + x**2 + x + y
sympy.polys.densearith.dmp sub(f, g, u, K)
Subtract dense polynomials in K[X].
Examples
>>> from sympy.polys import ring, ZZ
>>> R, x,y = ring(x,y, ZZ)
>>> R.dmp_sub(x**2 + y, x**2*y + x)
-x**2*y + x**2 - x + y
sympy.polys.densearith.dmp add mul(f, g, h, u, K)
Returns f + g*h where f, g, h are in K[X].
Examples
>>> from sympy.polys import ring, ZZ
>>> R, x,y = ring(x,y, ZZ)
>>> R.dmp_add_mul(x**2 + y, x, x + 2)
2*x**2 + 2*x + y
sympy.polys.densearith.dmp sub mul(f, g, h, u, K)
Returns f - g*h where f, g, h are in K[X].
1184
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.polys import ring, ZZ
>>> R, x,y = ring(x,y, ZZ)
>>> R.dmp_sub_mul(x**2 + y, x, x + 2)
-2*x + y
sympy.polys.densearith.dmp mul(f, g, u, K)
Multiply dense polynomials in K[X].
Examples
>>> from sympy.polys import ring, ZZ
>>> R, x,y = ring(x,y, ZZ)
>>> R.dmp_mul(x*y + 1, x)
x**2*y + x
sympy.polys.densearith.dmp sqr(f, u, K)
Square dense polynomials in K[X].
Examples
>>> from sympy.polys import ring, ZZ
>>> R, x,y = ring(x,y, ZZ)
>>> R.dmp_sqr(x**2 + x*y + y**2)
x**4 + 2*x**3*y + 3*x**2*y**2 + 2*x*y**3 + y**4
sympy.polys.densearith.dmp pow(f, n, u, K)
Raise f to the n-th power in K[X].
Examples
>>> from sympy.polys import ring, ZZ
>>> R, x,y = ring(x,y, ZZ)
>>> R.dmp_pow(x*y + 1, 3)
x**3*y**3 + 3*x**2*y**2 + 3*x*y + 1
sympy.polys.densearith.dmp pdiv(f, g, u, K)
Polynomial pseudo-division in K[X].
Examples
>>> from sympy.polys import ring, ZZ
>>> R, x,y = ring(x,y, ZZ)
5.16. Polynomials Manipulation Module
1185
SymPy Documentation, Release 0.7.6
>>> R.dmp_pdiv(x**2 + x*y, 2*x + 2)
(2*x + 2*y - 2, -4*y + 4)
sympy.polys.densearith.dmp prem(f, g, u, K)
Polynomial pseudo-remainder in K[X].
Examples
>>> from sympy.polys import ring, ZZ
>>> R, x,y = ring(x,y, ZZ)
>>> R.dmp_prem(x**2 + x*y, 2*x + 2)
-4*y + 4
sympy.polys.densearith.dmp pquo(f, g, u, K)
Polynomial exact pseudo-quotient in K[X].
Examples
>>> from sympy.polys import ring, ZZ
>>> R, x,y = ring(x,y, ZZ)
>>> f = x**2 + x*y
>>> g = 2*x + 2*y
>>> h = 2*x + 2
>>> R.dmp_pquo(f, g)
2*x
>>> R.dmp_pquo(f, h)
2*x + 2*y - 2
sympy.polys.densearith.dmp pexquo(f, g, u, K)
Polynomial pseudo-quotient in K[X].
Examples
>>> from sympy.polys import ring, ZZ
>>> R, x,y = ring(x,y, ZZ)
>>> f = x**2 + x*y
>>> g = 2*x + 2*y
>>> h = 2*x + 2
>>> R.dmp_pexquo(f, g)
2*x
>>> R.dmp_pexquo(f, h)
Traceback (most recent call last):
...
ExactQuotientFailed: [[2], [2]] does not divide [[1], [1, 0], []]
Traceback (most recent call last):
1186
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
...
ExactQuotientFailed: [[2], [2]] does not divide [[1], [1, 0], []]
sympy.polys.densearith.dmp rr div(f, g, u, K)
Multivariate division with remainder over a ring.
Examples
>>> from sympy.polys import ring, ZZ
>>> R, x,y = ring(x,y, ZZ)
>>> R.dmp_rr_div(x**2 + x*y, 2*x + 2)
(0, x**2 + x*y)
sympy.polys.densearith.dmp ff div(f, g, u, K)
Polynomial division with remainder over a eld.
Examples
>>> from sympy.polys import ring, QQ
>>> R, x,y = ring(x,y, QQ)
>>> R.dmp_ff_div(x**2 + x*y, 2*x + 2)
(1/2*x + 1/2*y - 1/2, -y + 1)
sympy.polys.densearith.dmp div(f, g, u, K)
Polynomial division with remainder in K[X].
Examples
>>> from sympy.polys import ring, ZZ, QQ
>>> R, x,y = ring(x,y, ZZ)
>>> R.dmp_div(x**2 + x*y, 2*x + 2)
(0, x**2 + x*y)
>>> R, x,y = ring(x,y, QQ)
>>> R.dmp_div(x**2 + x*y, 2*x + 2)
(1/2*x + 1/2*y - 1/2, -y + 1)
sympy.polys.densearith.dmp rem(f, g, u, K)
Returns polynomial remainder in K[X].
Examples
>>> from sympy.polys import ring, ZZ, QQ
>>> R, x,y = ring(x,y, ZZ)
>>> R.dmp_rem(x**2 + x*y, 2*x + 2)
x**2 + x*y
5.16. Polynomials Manipulation Module
1187
SymPy Documentation, Release 0.7.6
>>> R, x,y = ring(x,y, QQ)
>>> R.dmp_rem(x**2 + x*y, 2*x + 2)
-y + 1
sympy.polys.densearith.dmp quo(f, g, u, K)
Returns exact polynomial quotient in K[X].
Examples
>>> from sympy.polys import ring, ZZ, QQ
>>> R, x,y = ring(x,y, ZZ)
>>> R.dmp_quo(x**2 + x*y, 2*x + 2)
0
>>> R, x,y = ring(x,y, QQ)
>>> R.dmp_quo(x**2 + x*y, 2*x + 2)
1/2*x + 1/2*y - 1/2
sympy.polys.densearith.dmp exquo(f, g, u, K)
Returns polynomial quotient in K[X].
Examples
>>> from sympy.polys import ring, ZZ
>>> R, x,y = ring(x,y, ZZ)
>>> f = x**2 + x*y
>>> g = x + y
>>> h = 2*x + 2
>>> R.dmp_exquo(f, g)
x
>>> R.dmp_exquo(f, h)
Traceback (most recent call last):
...
ExactQuotientFailed: [[2], [2]] does not divide [[1], [1, 0], []]
Traceback (most recent call last):
...
ExactQuotientFailed: [[2], [2]] does not divide [[1], [1, 0], []]
sympy.polys.densearith.dmp max norm(f, u, K)
Returns maximum norm of a polynomial in K[X].
Examples
>>> from sympy.polys import ring, ZZ
>>> R, x,y = ring(x,y, ZZ)
>>> R.dmp_max_norm(2*x*y - x - 3)
3
1188
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
sympy.polys.densearith.dmp l1 norm(f, u, K)
Returns l1 norm of a polynomial in K[X].
Examples
>>> from sympy.polys import ring, ZZ
>>> R, x,y = ring(x,y, ZZ)
>>> R.dmp_l1_norm(2*x*y - x - 3)
6
sympy.polys.densearith.dmp expand(polys, u, K)
Multiply together several polynomials in K[X].
Examples
>>> from sympy.polys import ring, ZZ
>>> R, x,y = ring(x,y, ZZ)
>>> R.dmp_expand([x**2 + y**2, x + 1])
x**3 + x**2 + x*y**2 + y**2
Further tools:
sympy.polys.densetools.dmp integrate(f, m, u, K)
Computes the indenite integral of f in x 0 in K[X].
Examples
>>> from sympy.polys import ring, QQ
>>> R, x,y = ring(x,y, QQ)
>>> R.dmp_integrate(x + 2*y, 1)
1/2*x**2 + 2*x*y
>>> R.dmp_integrate(x + 2*y, 2)
1/6*x**3 + x**2*y
sympy.polys.densetools.dmp integrate in(f, m, j, u, K)
Computes the indenite integral of f in x j in K[X].
Examples
>>> from sympy.polys import ring, QQ
>>> R, x,y = ring(x,y, QQ)
>>> R.dmp_integrate_in(x + 2*y, 1, 0)
1/2*x**2 + 2*x*y
>>> R.dmp_integrate_in(x + 2*y, 1, 1)
x*y + y**2
sympy.polys.densetools.dmp diff(f, m, u, K)
m-th order derivative in x 0 of a polynomial in K[X].
5.16. Polynomials Manipulation Module
1189
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.polys import ring, ZZ
>>> R, x,y = ring(x,y, ZZ)
>>> f = x*y**2 + 2*x*y + 3*x + 2*y**2 + 3*y + 1
>>> R.dmp_diff(f, 1)
y**2 + 2*y + 3
>>> R.dmp_diff(f, 2)
0
sympy.polys.densetools.dmp diff in(f, m, j, u, K)
m-th order derivative in x j of a polynomial in K[X].
Examples
>>> from sympy.polys import ring, ZZ
>>> R, x,y = ring(x,y, ZZ)
>>> f = x*y**2 + 2*x*y + 3*x + 2*y**2 + 3*y + 1
>>> R.dmp_diff_in(f, 1, 0)
y**2 + 2*y + 3
>>> R.dmp_diff_in(f, 1, 1)
2*x*y + 2*x + 4*y + 3
sympy.polys.densetools.dmp eval(f, a, u, K)
Evaluate a polynomial at x 0 = a in K[X] using the Horner scheme.
Examples
>>> from sympy.polys import ring, ZZ
>>> R, x,y = ring(x,y, ZZ)
>>> R.dmp_eval(2*x*y + 3*x + y + 2, 2)
5*y + 8
sympy.polys.densetools.dmp eval in(f, a, j, u, K)
Evaluate a polynomial at x j = a in K[X] using the Horner scheme.
Examples
>>> from sympy.polys import ring, ZZ
>>> R, x,y = ring(x,y, ZZ)
>>> f = 2*x*y + 3*x + y + 2
>>>
5*y
>>>
7*x
1190
R.dmp_eval_in(f, 2, 0)
+ 8
R.dmp_eval_in(f, 2, 1)
+ 4
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
sympy.polys.densetools.dmp eval tail(f, A, u, K)
Evaluate a polynomial at x j = a j, ... in K[X].
Examples
>>> from sympy.polys import ring, ZZ
>>> R, x,y = ring(x,y, ZZ)
>>> f = 2*x*y + 3*x + y + 2
>>> R.dmp_eval_tail(f, [2])
7*x + 4
>>> R.dmp_eval_tail(f, [2, 2])
18
sympy.polys.densetools.dmp diff eval in(f, m, a, j, u, K)
Dierentiate and evaluate a polynomial in x j at a in K[X].
Examples
>>> from sympy.polys import ring, ZZ
>>> R, x,y = ring(x,y, ZZ)
>>> f = x*y**2 + 2*x*y + 3*x + 2*y**2 + 3*y + 1
>>> R.dmp_diff_eval_in(f, 1, 2, 0)
y**2 + 2*y + 3
>>> R.dmp_diff_eval_in(f, 1, 2, 1)
6*x + 11
sympy.polys.densetools.dmp trunc(f, p, u, K)
Reduce a K[X] polynomial modulo a polynomial p in K[Y].
Examples
>>> from sympy.polys import ring, ZZ
>>> R, x,y = ring(x,y, ZZ)
>>> f = 3*x**2*y + 8*x**2 + 5*x*y + 6*x + 2*y + 3
>>> g = (y - 1).drop(x)
>>> R.dmp_trunc(f, g)
11*x**2 + 11*x + 5
sympy.polys.densetools.dmp ground trunc(f, p, u, K)
Reduce a K[X] polynomial modulo a constant p in K.
Examples
5.16. Polynomials Manipulation Module
1191
SymPy Documentation, Release 0.7.6
>>> from sympy.polys import ring, ZZ
>>> R, x,y = ring(x,y, ZZ)
>>> f = 3*x**2*y + 8*x**2 + 5*x*y + 6*x + 2*y + 3
>>> R.dmp_ground_trunc(f, ZZ(3))
-x**2 - x*y - y
sympy.polys.densetools.dup monic(f, K)
Divide all coecients by LC(f) in K[x].
Examples
>>> from sympy.polys import ring, ZZ, QQ
>>> R, x = ring(x, ZZ)
>>> R.dup_monic(3*x**2 + 6*x + 9)
x**2 + 2*x + 3
>>> R, x = ring(x, QQ)
>>> R.dup_monic(3*x**2 + 4*x + 2)
x**2 + 4/3*x + 2/3
sympy.polys.densetools.dmp ground monic(f, u, K)
Divide all coecients by LC(f) in K[X].
Examples
>>> from sympy.polys import ring, ZZ, QQ
>>> R, x,y = ring(x,y, ZZ)
>>> f = 3*x**2*y + 6*x**2 + 3*x*y + 9*y + 3
>>> R.dmp_ground_monic(f)
x**2*y + 2*x**2 + x*y + 3*y + 1
>>> R, x,y = ring(x,y, QQ)
>>> f = 3*x**2*y + 8*x**2 + 5*x*y + 6*x + 2*y + 3
>>> R.dmp_ground_monic(f)
x**2*y + 8/3*x**2 + 5/3*x*y + 2*x + 2/3*y + 1
sympy.polys.densetools.dup content(f, K)
Compute the GCD of coecients of f in K[x].
Examples
>>> from sympy.polys import ring, ZZ, QQ
>>> R, x = ring(x, ZZ)
>>> f = 6*x**2 + 8*x + 12
1192
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> R.dup_content(f)
2
>>> R, x = ring(x, QQ)
>>> f = 6*x**2 + 8*x + 12
>>> R.dup_content(f)
2
sympy.polys.densetools.dmp ground content(f, u, K)
Compute the GCD of coecients of f in K[X].
Examples
>>> from sympy.polys import ring, ZZ, QQ
>>> R, x,y = ring(x,y, ZZ)
>>> f = 2*x*y + 6*x + 4*y + 12
>>> R.dmp_ground_content(f)
2
>>> R, x,y = ring(x,y, QQ)
>>> f = 2*x*y + 6*x + 4*y + 12
>>> R.dmp_ground_content(f)
2
sympy.polys.densetools.dup primitive(f, K)
Compute content and the primitive form of f in K[x].
Examples
>>> from sympy.polys import ring, ZZ, QQ
>>> R, x = ring(x, ZZ)
>>> f = 6*x**2 + 8*x + 12
>>> R.dup_primitive(f)
(2, 3*x**2 + 4*x + 6)
>>> R, x = ring(x, QQ)
>>> f = 6*x**2 + 8*x + 12
>>> R.dup_primitive(f)
(2, 3*x**2 + 4*x + 6)
sympy.polys.densetools.dmp ground primitive(f, u, K)
Compute content and the primitive form of f in K[X].
5.16. Polynomials Manipulation Module
1193
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.polys import ring, ZZ, QQ
>>> R, x,y = ring(x,y, ZZ)
>>> f = 2*x*y + 6*x + 4*y + 12
>>> R.dmp_ground_primitive(f)
(2, x*y + 3*x + 2*y + 6)
>>> R, x,y = ring(x,y, QQ)
>>> f = 2*x*y + 6*x + 4*y + 12
>>> R.dmp_ground_primitive(f)
(2, x*y + 3*x + 2*y + 6)
sympy.polys.densetools.dup extract(f, g, K)
Extract common content from a pair of polynomials in K[x].
Examples
>>> from sympy.polys import ring, ZZ
>>> R, x = ring(x, ZZ)
>>> R.dup_extract(6*x**2 + 12*x + 18, 4*x**2 + 8*x + 12)
(2, 3*x**2 + 6*x + 9, 2*x**2 + 4*x + 6)
sympy.polys.densetools.dmp ground extract(f, g, u, K)
Extract common content from a pair of polynomials in K[X].
Examples
>>> from sympy.polys import ring, ZZ
>>> R, x,y = ring(x,y, ZZ)
>>> R.dmp_ground_extract(6*x*y + 12*x + 18, 4*x*y + 8*x + 12)
(2, 3*x*y + 6*x + 9, 2*x*y + 4*x + 6)
sympy.polys.densetools.dup real imag(f, K)
Return bivariate polynomials f1 and f2, such that f = f1 + f2*I.
Examples
>>> from sympy.polys import ring, ZZ
>>> R, x,y = ring(x,y, ZZ)
>>> R.dup_real_imag(x**3 + x**2 + x + 1)
(x**3 + x**2 - 3*x*y**2 + x - y**2 + 1, 3*x**2*y + 2*x*y - y**3 + y)
sympy.polys.densetools.dup mirror(f, K)
Evaluate eciently the composition f(-x) in K[x].
1194
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.polys import ring, ZZ
>>> R, x = ring(x, ZZ)
>>> R.dup_mirror(x**3 + 2*x**2 - 4*x + 2)
-x**3 + 2*x**2 + 4*x + 2
sympy.polys.densetools.dup scale(f, a, K)
Evaluate eciently composition f(a*x) in K[x].
Examples
>>> from sympy.polys import ring, ZZ
>>> R, x = ring(x, ZZ)
>>> R.dup_scale(x**2 - 2*x + 1, ZZ(2))
4*x**2 - 4*x + 1
sympy.polys.densetools.dup shift(f, a, K)
Evaluate eciently Taylor shift f(x + a) in K[x].
Examples
>>> from sympy.polys import ring, ZZ
>>> R, x = ring(x, ZZ)
>>> R.dup_shift(x**2 - 2*x + 1, ZZ(2))
x**2 + 2*x + 1
sympy.polys.densetools.dup transform(f, p, q, K)
Evaluate functional transformation q**n * f(p/q) in K[x].
Examples
>>> from sympy.polys import ring, ZZ
>>> R, x = ring(x, ZZ)
>>> R.dup_transform(x**2 - 2*x + 1, x**2 + 1, x - 1)
x**4 - 2*x**3 + 5*x**2 - 4*x + 4
sympy.polys.densetools.dmp compose(f, g, u, K)
Evaluate functional composition f(g) in K[X].
Examples
>>> from sympy.polys import ring, ZZ
>>> R, x,y = ring(x,y, ZZ)
5.16. Polynomials Manipulation Module
1195
SymPy Documentation, Release 0.7.6
>>> R.dmp_compose(x*y + 2*x + y, y)
y**2 + 3*y
sympy.polys.densetools.dup decompose(f, K)
Computes functional decomposition of f in K[x].
Given a univariate polynomial f with coecients in a eld of characteristic zero, returns
list [f 1, f 2, ..., f n], where:
f = f 1 o f 2 o ... f n = f 1(f 2(... f n))
and f 2, ..., f n are monic and homogeneous polynomials of at least second degree.
Unlike factorization, complete functional decompositions of polynomials are not unique,
consider examples:
1.f o g = f(x + b) o (g - b)
2.x**n o x**m = x**m o x**n
3.T n o T m = T m o T n
where T n and T m are Chebyshev polynomials.
References
1.[Kozen89] (page 1911)
[Kozen89] (page 1911)
Examples
>>> from sympy.polys import ring, ZZ
>>> R, x = ring(x, ZZ)
>>> R.dup_decompose(x**4 - 2*x**3 + x**2)
[x**2, x**2 - x]
sympy.polys.densetools.dmp lift(f, u, K)
Convert algebraic coecients to integers in K[X].
Examples
>>> from sympy.polys import ring, QQ
>>> from sympy import I
>>> K = QQ.algebraic_field(I)
>>> R, x = ring(x, K)
>>> f = x**2 + K([QQ(1), QQ(0)])*x + K([QQ(2), QQ(0)])
>>> R.dmp_lift(f)
x**8 + 2*x**6 + 9*x**4 - 8*x**2 + 16
sympy.polys.densetools.dup sign variations(f, K)
Compute the number of sign variations of f in K[x].
1196
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.polys import ring, ZZ
>>> R, x = ring(x, ZZ)
>>> R.dup_sign_variations(x**4 - x**2 - x + 1)
2
sympy.polys.densetools.dmp clear denoms(f, u, K0, K1=None, convert=False)
Clear denominators, i.e. transform K 0 to K 1.
Examples
>>> from sympy.polys import ring, QQ
>>> R, x,y = ring(x,y, QQ)
>>> f = QQ(1,2)*x + QQ(1,3)*y + 1
>>>
(6,
>>>
(6,
R.dmp_clear_denoms(f, convert=False)
3*x + 2*y + 6)
R.dmp_clear_denoms(f, convert=True)
3*x + 2*y + 6)
sympy.polys.densetools.dmp revert(f, g, u, K)
Compute f**(-1) mod x**n using Newton iteration.
Examples
>>> from sympy.polys import ring, QQ
>>> R, x,y = ring(x,y, QQ)
Manipulation of dense, univariate polynomials with nite eld coecients Functions
in this module carry the prex gf , referring to the classical name Galois Fields for nite
elds. Note that many polynomial factorization algorithms work by reduction to the nite
eld case, so having special implementations for this case is justied both by performance,
and by the necessity of certain methods which do not even make sense over general elds.
sympy.polys.galoistools.gf crt(U, M, K=None)
Chinese Remainder Theorem.
Given a set of integer residues u 0,...,u n and a set of co-prime integer moduli
m 0,...,m n, returns an integer u, such that u = u i mod m i for i = 0,...,n.
As an example consider a set of residues U = [49, 76, 65] and a set of moduli M =
[99, 97, 95]. Then we have:
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.galoistools import gf_crt
>>> from sympy.ntheory.modular import solve_congruence
>>> gf_crt([49, 76, 65], [99, 97, 95], ZZ)
639985
This is the correct result because:
5.16. Polynomials Manipulation Module
1197
SymPy Documentation, Release 0.7.6
>>> [639985 % m for m in [99, 97, 95]]
[49, 76, 65]
Note: this is a low-level routine with no error checking.
See Also:
sympy.ntheory.modular.crt (page 290) a higher level crt routine
sympy.ntheory.modular.solve congruence (page 292)
sympy.polys.galoistools.gf crt1(M, K)
First part of the Chinese Remainder Theorem.
Examples
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.galoistools import gf_crt1
>>> gf_crt1([99, 97, 95], ZZ)
(912285, [9215, 9405, 9603], [62, 24, 12])
sympy.polys.galoistools.gf crt2(U, M, p, E, S, K)
Second part of the Chinese Remainder Theorem.
Examples
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.galoistools import gf_crt2
>>>
>>>
>>>
>>>
>>>
U
M
p
E
S
=
=
=
=
=
[49, 76, 65]
[99, 97, 95]
912285
[9215, 9405, 9603]
[62, 24, 12]
>>> gf_crt2(U, M, p, E, S, ZZ)
639985
sympy.polys.galoistools.gf int(a, p)
Coerce a mod p to an integer in the range [-p/2, p/2].
Examples
>>> from sympy.polys.galoistools import gf_int
>>> gf_int(2, 7)
2
>>> gf_int(5, 7)
-2
sympy.polys.galoistools.gf degree(f)
Return the leading degree of f.
1198
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.polys.galoistools import gf_degree
>>> gf_degree([1, 1, 2, 0])
3
>>> gf_degree([])
-1
sympy.polys.galoistools.gf LC(f, K)
Return the leading coecient of f.
Examples
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.galoistools import gf_LC
>>> gf_LC([3, 0, 1], ZZ)
3
sympy.polys.galoistools.gf TC(f, K)
Return the trailing coecient of f.
Examples
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.galoistools import gf_TC
>>> gf_TC([3, 0, 1], ZZ)
1
sympy.polys.galoistools.gf strip(f)
Remove leading zeros from f.
Examples
>>> from sympy.polys.galoistools import gf_strip
>>> gf_strip([0, 0, 0, 3, 0, 1])
[3, 0, 1]
sympy.polys.galoistools.gf trunc(f, p)
Reduce all coecients modulo p.
Examples
>>> from sympy.polys.galoistools import gf_trunc
>>> gf_trunc([7, -2, 3], 5)
[2, 3, 3]
5.16. Polynomials Manipulation Module
1199
SymPy Documentation, Release 0.7.6
sympy.polys.galoistools.gf normal(f, p, K)
Normalize all coecients in K.
Examples
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.galoistools import gf_normal
>>> gf_normal([5, 10, 21, -3], 5, ZZ)
[1, 2]
sympy.polys.galoistools.gf from dict(f, p, K)
Create a GF(p)[x] polynomial from a dict.
Examples
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.galoistools import gf_from_dict
>>> gf_from_dict({10: ZZ(4), 4: ZZ(33), 0: ZZ(-1)}, 5, ZZ)
[4, 0, 0, 0, 0, 0, 3, 0, 0, 0, 4]
sympy.polys.galoistools.gf to dict(f, p, symmetric=True)
Convert a GF(p)[x] polynomial to a dict.
Examples
>>> from sympy.polys.galoistools import gf_to_dict
>>>
{0:
>>>
{0:
gf_to_dict([4, 0, 0, 0, 0, 0, 3, 0, 0, 0, 4], 5)
-1, 4: -2, 10: -1}
gf_to_dict([4, 0, 0, 0, 0, 0, 3, 0, 0, 0, 4], 5, symmetric=False)
4, 4: 3, 10: 4}
sympy.polys.galoistools.gf from int poly(f, p)
Create a GF(p)[x] polynomial from Z[x].
Examples
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.galoistools import gf_from_int_poly
>>> gf_from_int_poly([7, -2, 3], 5)
[2, 3, 3]
sympy.polys.galoistools.gf to int poly(f, p, symmetric=True)
Convert a GF(p)[x] polynomial to Z[x].
1200
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.polys.galoistools import gf_to_int_poly
>>>
[2,
>>>
[2,
gf_to_int_poly([2, 3, 3], 5)
-2, -2]
gf_to_int_poly([2, 3, 3], 5, symmetric=False)
3, 3]
sympy.polys.galoistools.gf neg(f, p, K)
Negate a polynomial in GF(p)[x].
Examples
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.galoistools import gf_neg
>>> gf_neg([3, 2, 1, 0], 5, ZZ)
[2, 3, 4, 0]
sympy.polys.galoistools.gf add ground(f, a, p, K)
Compute f + a where f in GF(p)[x] and a in GF(p).
Examples
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.galoistools import gf_add_ground
>>> gf_add_ground([3, 2, 4], 2, 5, ZZ)
[3, 2, 1]
sympy.polys.galoistools.gf sub ground(f, a, p, K)
Compute f - a where f in GF(p)[x] and a in GF(p).
Examples
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.galoistools import gf_sub_ground
>>> gf_sub_ground([3, 2, 4], 2, 5, ZZ)
[3, 2, 2]
sympy.polys.galoistools.gf mul ground(f, a, p, K)
Compute f * a where f in GF(p)[x] and a in GF(p).
Examples
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.galoistools import gf_mul_ground
5.16. Polynomials Manipulation Module
1201
SymPy Documentation, Release 0.7.6
>>> gf_mul_ground([3, 2, 4], 2, 5, ZZ)
[1, 4, 3]
sympy.polys.galoistools.gf quo ground(f, a, p, K)
Compute f/a where f in GF(p)[x] and a in GF(p).
Examples
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.galoistools import gf_quo_ground
>>> gf_quo_ground(ZZ.map([3, 2, 4]), ZZ(2), 5, ZZ)
[4, 1, 2]
sympy.polys.galoistools.gf add(f, g, p, K)
Add polynomials in GF(p)[x].
Examples
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.galoistools import gf_add
>>> gf_add([3, 2, 4], [2, 2, 2], 5, ZZ)
[4, 1]
sympy.polys.galoistools.gf sub(f, g, p, K)
Subtract polynomials in GF(p)[x].
Examples
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.galoistools import gf_sub
>>> gf_sub([3, 2, 4], [2, 2, 2], 5, ZZ)
[1, 0, 2]
sympy.polys.galoistools.gf mul(f, g, p, K)
Multiply polynomials in GF(p)[x].
Examples
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.galoistools import gf_mul
>>> gf_mul([3, 2, 4], [2, 2, 2], 5, ZZ)
[1, 0, 3, 2, 3]
sympy.polys.galoistools.gf sqr(f, p, K)
Square polynomials in GF(p)[x].
1202
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.galoistools import gf_sqr
>>> gf_sqr([3, 2, 4], 5, ZZ)
[4, 2, 3, 1, 1]
sympy.polys.galoistools.gf add mul(f, g, h, p, K)
Returns f + g*h where f, g, h in GF(p)[x].
Examples
>>>
>>>
>>>
[2,
from sympy.polys.domains import ZZ
from sympy.polys.galoistools import gf_add_mul
gf_add_mul([3, 2, 4], [2, 2, 2], [1, 4], 5, ZZ)
3, 2, 2]
sympy.polys.galoistools.gf sub mul(f, g, h, p, K)
Compute f - g*h where f, g, h in GF(p)[x].
Examples
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.galoistools import gf_sub_mul
>>> gf_sub_mul([3, 2, 4], [2, 2, 2], [1, 4], 5, ZZ)
[3, 3, 2, 1]
sympy.polys.galoistools.gf expand(F, p, K)
Expand results of factor() in GF(p)[x].
Examples
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.galoistools import gf_expand
>>> gf_expand([([3, 2, 4], 1), ([2, 2], 2), ([3, 1], 3)], 5, ZZ)
[4, 3, 0, 3, 0, 1, 4, 1]
sympy.polys.galoistools.gf div(f, g, p, K)
Division with remainder in GF(p)[x].
Given univariate polynomials f and g with coecients in a nite eld with p elements,
returns polynomials q and r (quotient and remainder) such that f = q*g + r.
Consider polynomials x**3 + x + 1 and x**2 + x in GF(2):
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.galoistools import gf_div, gf_add_mul
>>> gf_div(ZZ.map([1, 0, 1, 1]), ZZ.map([1, 1, 0]), 2, ZZ)
([1, 1], [1])
5.16. Polynomials Manipulation Module
1203
SymPy Documentation, Release 0.7.6
As result we obtained quotient x + 1 and remainder 1, thus:
>>> gf_add_mul(ZZ.map([1]), ZZ.map([1, 1]), ZZ.map([1, 1, 0]), 2, ZZ)
[1, 0, 1, 1]
References
1.[Monagan93] (page 1911)
2.[Gathen99] (page 1911)
[Monagan93] (page 1911), [Gathen99] (page 1911)
sympy.polys.galoistools.gf rem(f, g, p, K)
Compute polynomial remainder in GF(p)[x].
Examples
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.galoistools import gf_rem
>>> gf_rem(ZZ.map([1, 0, 1, 1]), ZZ.map([1, 1, 0]), 2, ZZ)
[1]
sympy.polys.galoistools.gf quo(f, g, p, K)
Compute exact quotient in GF(p)[x].
Examples
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.galoistools import gf_quo
>>>
[1,
>>>
[3,
gf_quo(ZZ.map([1, 0, 1, 1]), ZZ.map([1, 1, 0]), 2, ZZ)
1]
gf_quo(ZZ.map([1, 0, 3, 2, 3]), ZZ.map([2, 2, 2]), 5, ZZ)
2, 4]
sympy.polys.galoistools.gf exquo(f, g, p, K)
Compute polynomial quotient in GF(p)[x].
Examples
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.galoistools import gf_exquo
>>> gf_exquo(ZZ.map([1, 0, 3, 2, 3]), ZZ.map([2, 2, 2]), 5, ZZ)
[3, 2, 4]
>>> gf_exquo(ZZ.map([1, 0, 1, 1]), ZZ.map([1, 1, 0]), 2, ZZ)
Traceback (most recent call last):
...
ExactQuotientFailed: [1, 1, 0] does not divide [1, 0, 1, 1]
1204
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Traceback (most recent call last):
...
ExactQuotientFailed: [1, 1, 0] does not divide [1, 0, 1, 1]
sympy.polys.galoistools.gf lshift(f, n, K)
Eciently multiply f by x**n.
Examples
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.galoistools import gf_lshift
>>> gf_lshift([3, 2, 4], 4, ZZ)
[3, 2, 4, 0, 0, 0, 0]
sympy.polys.galoistools.gf rshift(f, n, K)
Eciently divide f by x**n.
Examples
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.galoistools import gf_rshift
>>> gf_rshift([1, 2, 3, 4, 0], 3, ZZ)
([1, 2], [3, 4, 0])
sympy.polys.galoistools.gf pow(f, n, p, K)
Compute f**n in GF(p)[x] using repeated squaring.
Examples
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.galoistools import gf_pow
>>> gf_pow([3, 2, 4], 3, 5, ZZ)
[2, 4, 4, 2, 2, 1, 4]
sympy.polys.galoistools.gf pow mod(f, n, g, p, K)
Compute f**n in GF(p)[x]/(g) using repeated squaring.
Given polynomials f and g in GF(p)[x] and a non-negative integer n, eciently computes
f**n (mod g) i.e. the remainder of f**n from division by g, using the repeated squaring
algorithm.
References
1.[Gathen99] (page 1911)
[Gathen99] (page 1911)
5.16. Polynomials Manipulation Module
1205
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.galoistools import gf_pow_mod
>>> gf_pow_mod(ZZ.map([3, 2, 4]), 3, ZZ.map([1, 1]), 5, ZZ)
[]
sympy.polys.galoistools.gf gcd(f, g, p, K)
Euclidean Algorithm in GF(p)[x].
Examples
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.galoistools import gf_gcd
>>> gf_gcd(ZZ.map([3, 2, 4]), ZZ.map([2, 2, 3]), 5, ZZ)
[1, 3]
sympy.polys.galoistools.gf lcm(f, g, p, K)
Compute polynomial LCM in GF(p)[x].
Examples
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.galoistools import gf_lcm
>>> gf_lcm(ZZ.map([3, 2, 4]), ZZ.map([2, 2, 3]), 5, ZZ)
[1, 2, 0, 4]
sympy.polys.galoistools.gf cofactors(f, g, p, K)
Compute polynomial GCD and cofactors in GF(p)[x].
Examples
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.galoistools import gf_cofactors
>>> gf_cofactors(ZZ.map([3, 2, 4]), ZZ.map([2, 2, 3]), 5, ZZ)
([1, 3], [3, 3], [2, 1])
sympy.polys.galoistools.gf gcdex(f, g, p, K)
Extended Euclidean Algorithm in GF(p)[x].
Given polynomials f and g in GF(p)[x], computes polynomials s, t and h, such that h =
gcd(f, g) and s*f + t*g = h. The typical application of EEA is solving polynomial
diophantine equations.
Consider polynomials f = (x + 7) (x + 1), g = (x + 7) (x**2 + 1) in GF(11)[x].
Application of Extended Euclidean Algorithm gives:
1206
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.galoistools import gf_gcdex, gf_mul, gf_add
>>> s, t, g = gf_gcdex(ZZ.map([1, 8, 7]), ZZ.map([1, 7, 1, 7]), 11, ZZ)
>>> s, t, g
([5, 6], [6], [1, 7])
As result we obtained polynomials s = 5*x + 6 and t = 6, and additionally gcd(f, g)
= x + 7. This is correct because:
>>> S = gf_mul(s, ZZ.map([1, 8, 7]), 11, ZZ)
>>> T = gf_mul(t, ZZ.map([1, 7, 1, 7]), 11, ZZ)
>>> gf_add(S, T, 11, ZZ) == [1, 7]
True
References
1.[Gathen99] (page 1911)
[Gathen99] (page 1911)
sympy.polys.galoistools.gf monic(f, p, K)
Compute LC and a monic polynomial in GF(p)[x].
Examples
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.galoistools import gf_monic
>>> gf_monic(ZZ.map([3, 2, 4]), 5, ZZ)
(3, [1, 4, 3])
sympy.polys.galoistools.gf diff(f, p, K)
Dierentiate polynomial in GF(p)[x].
Examples
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.galoistools import gf_diff
>>> gf_diff([3, 2, 4], 5, ZZ)
[1, 2]
sympy.polys.galoistools.gf eval(f, a, p, K)
Evaluate f(a) in GF(p) using Horner scheme.
Examples
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.galoistools import gf_eval
5.16. Polynomials Manipulation Module
1207
SymPy Documentation, Release 0.7.6
>>> gf_eval([3, 2, 4], 2, 5, ZZ)
0
sympy.polys.galoistools.gf multi eval(f, A, p, K)
Evaluate f(a) for a in [a 1, ..., a n].
Examples
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.galoistools import gf_multi_eval
>>> gf_multi_eval([3, 2, 4], [0, 1, 2, 3, 4], 5, ZZ)
[4, 4, 0, 2, 0]
sympy.polys.galoistools.gf compose(f, g, p, K)
Compute polynomial composition f(g) in GF(p)[x].
Examples
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.galoistools import gf_compose
>>> gf_compose([3, 2, 4], [2, 2, 2], 5, ZZ)
[2, 4, 0, 3, 0]
sympy.polys.galoistools.gf compose mod(g, h, f, p, K)
Compute polynomial composition g(h) in GF(p)[x]/(f).
Examples
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.galoistools import gf_compose_mod
>>> gf_compose_mod(ZZ.map([3, 2, 4]), ZZ.map([2, 2, 2]), ZZ.map([4, 3]), 5, ZZ)
[4]
sympy.polys.galoistools.gf trace map(a, b, c, n, f, p, K)
Compute polynomial trace map in GF(p)[x]/(f).
Given a polynomial f in GF(p)[x], polynomials a, b, c in the quotient ring GF(p)[x]/(f)
such that b = c**t (mod f) for some positive power t of p, and a positive integer n,
returns a mapping:
a -> a**t**n, a + a**t + a**t**2 + ... + a**t**n (mod f)
In factorization context, b = x**p mod f and c = x mod f. This way we can eciently
compute trace polynomials in equal degree factorization routine, much faster than with
other methods, like iterated Frobenius algorithm, for large degrees.
References
1.[Gathen92] (page 1911)
1208
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
[Gathen92] (page 1911)
Examples
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.galoistools import gf_trace_map
>>> gf_trace_map([1, 2], [4, 4], [1, 1], 4, [3, 2, 4], 5, ZZ)
([1, 3], [1, 3])
sympy.polys.galoistools.gf random(n, p, K)
Generate a random polynomial in GF(p)[x] of degree n.
Examples
>>>
>>>
>>>
[1,
from sympy.polys.domains import ZZ
from sympy.polys.galoistools import gf_random
gf_random(10, 5, ZZ)
2, 3, 2, 1, 1, 1, 2, 0, 4, 2]
sympy.polys.galoistools.gf irreducible(n, p, K)
Generate random irreducible polynomial of degree n in GF(p)[x].
Examples
>>>
>>>
>>>
[1,
from sympy.polys.domains import ZZ
from sympy.polys.galoistools import gf_irreducible
gf_irreducible(10, 5, ZZ)
4, 2, 2, 3, 2, 4, 1, 4, 0, 4]
sympy.polys.galoistools.gf irreducible p(f, p, K)
Test irreducibility of a polynomial f in GF(p)[x].
Examples
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.galoistools import gf_irreducible_p
>>> gf_irreducible_p(ZZ.map([1, 4, 2, 2, 3, 2, 4, 1, 4, 0, 4]), 5, ZZ)
True
>>> gf_irreducible_p(ZZ.map([3, 2, 4]), 5, ZZ)
False
sympy.polys.galoistools.gf sqf p(f, p, K)
Return True if f is square-free in GF(p)[x].
Examples
5.16. Polynomials Manipulation Module
1209
SymPy Documentation, Release 0.7.6
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.galoistools import gf_sqf_p
>>> gf_sqf_p(ZZ.map([3, 2, 4]), 5, ZZ)
True
>>> gf_sqf_p(ZZ.map([2, 4, 4, 2, 2, 1, 4]), 5, ZZ)
False
sympy.polys.galoistools.gf sqf part(f, p, K)
Return square-free part of a GF(p)[x] polynomial.
Examples
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.galoistools import gf_sqf_part
>>> gf_sqf_part(ZZ.map([1, 1, 3, 0, 1, 0, 2, 2, 1]), 5, ZZ)
[1, 4, 3]
sympy.polys.galoistools.gf sqf list(f, p, K, all=False)
Return the square-free decomposition of a GF(p)[x] polynomial.
Given a polynomial f in GF(p)[x], returns the leading coecient of f and a square-free
decomposition f 1**e 1 f 2**e 2 ... f k**e k such that all f i are monic polynomials
and (f i, f j) for i != j are co-prime and e 1 ... e k are given in increasing order.
All trivial terms (i.e. f i = 1) arent included in the output.
Consider polynomial f = x**11 + 1 over GF(11)[x]:
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.galoistools import (
...
gf_from_dict, gf_diff, gf_sqf_list, gf_pow,
... )
...
>>> f = gf_from_dict({11: ZZ(1), 0: ZZ(1)}, 11, ZZ)
Note that f(x) = 0:
>>> gf_diff(f, 11, ZZ)
[]
This phenomenon doesnt happen in characteristic zero. However we can still compute
square-free decomposition of f using gf sqf():
>>> gf_sqf_list(f, 11, ZZ)
(1, [([1, 1], 11)])
We obtained factorization f = (x + 1)**11. This is correct because:
>>> gf_pow([1, 1], 11, 11, ZZ) == f
True
References
1.[Geddes92] (page 1911)
1210
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
[Geddes92] (page 1911)
sympy.polys.galoistools.gf Qmatrix(f, p, K)
Calculate Berlekamps Q matrix.
Examples
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.galoistools import gf_Qmatrix
>>> gf_Qmatrix([3, 2, 4], 5, ZZ)
[[1, 0],
[3, 4]]
>>> gf_Qmatrix([1, 0, 0, 0, 1], 5, ZZ)
[[1, 0, 0, 0],
[0, 4, 0, 0],
[0, 0, 1, 0],
[0, 0, 0, 4]]
sympy.polys.galoistools.gf Qbasis(Q, p, K)
Compute a basis of the kernel of Q.
Examples
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.galoistools import gf_Qmatrix, gf_Qbasis
>>> gf_Qbasis(gf_Qmatrix([1, 0, 0, 0, 1], 5, ZZ), 5, ZZ)
[[1, 0, 0, 0], [0, 0, 1, 0]]
>>> gf_Qbasis(gf_Qmatrix([3, 2, 4], 5, ZZ), 5, ZZ)
[[1, 0]]
sympy.polys.galoistools.gf berlekamp(f, p, K)
Factor a square-free f in GF(p)[x] for small p.
Examples
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.galoistools import gf_berlekamp
>>> gf_berlekamp([1, 0, 0, 0, 1], 5, ZZ)
[[1, 0, 2], [1, 0, 3]]
sympy.polys.galoistools.gf zassenhaus(f, p, K)
Factor a square-free f in GF(p)[x] for medium p.
Examples
5.16. Polynomials Manipulation Module
1211
SymPy Documentation, Release 0.7.6
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.galoistools import gf_zassenhaus
>>> gf_zassenhaus(ZZ.map([1, 4, 3]), 5, ZZ)
[[1, 1], [1, 3]]
sympy.polys.galoistools.gf shoup(f, p, K)
Factor a square-free f in GF(p)[x] for large p.
Examples
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.galoistools import gf_shoup
>>> gf_shoup(ZZ.map([1, 4, 3]), 5, ZZ)
[[1, 1], [1, 3]]
sympy.polys.galoistools.gf factor sqf(f, p, K, method=None)
Factor a square-free polynomial f in GF(p)[x].
Examples
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.galoistools import gf_factor_sqf
>>> gf_factor_sqf(ZZ.map([3, 2, 4]), 5, ZZ)
(3, [[1, 1], [1, 3]])
sympy.polys.galoistools.gf factor(f, p, K)
Factor (non square-free) polynomials in GF(p)[x].
Given a possibly non square-free polynomial f in GF(p)[x], returns its complete factorization into irreducibles:
f 1(x)**e 1 f 2(x)**e 2 ... f d(x)**e d
where each f i is a monic polynomial and gcd(f i, f j) == 1, for i != j. The result
is given as a tuple consisting of the leading coecient of f and a list of factors of f with
their multiplicities.
The algorithm proceeds by rst computing square-free decomposition of f and then iteratively factoring each of square-free factors.
Consider a non square-free polynomial f = (7*x + 1) (x + 2)**2 in GF(11)[x]. We
obtain its factorization into irreducibles as follows:
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.galoistools import gf_factor
>>> gf_factor(ZZ.map([5, 2, 7, 2]), 11, ZZ)
(5, [([1, 2], 1), ([1, 8], 2)])
We arrived with factorization f = 5 (x + 2) (x + 8)**2. We didnt recover the exact
form of the input polynomial because we requested to get monic factors of f and its
leading coecient separately.
1212
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Square-free factors of f can be factored into irreducibles over GF(p) using three very
dierent methods:
Berlekamp ecient for very small values of p (usually p < 25)
Cantor-Zassenhaus ecient on average input and with typical p
Shoup-Kaltofen-Gathen ecient with very large inputs and modulus
If you want to use a specic factorization method, instead of the default one, set
GF FACTOR METHOD with one of berlekamp, zassenhaus or shoup values.
References
1.[Gathen99] (page 1911)
[Gathen99] (page 1911)
sympy.polys.galoistools.gf value(f, a)
Value of polynomial f at a in eld R.
Examples
>>> from sympy.polys.galoistools import gf_value
>>> gf_value([1, 7, 2, 4], 11)
2204
sympy.polys.galoistools.gf csolve(f, n)
To solve f(x) congruent 0 mod(n).
n is divided into canonical factors and f(x) cong 0 mod(p**e) will be solved for each factor.
Applying the Chinese Remainder Theorem to the results returns the nal answers.
References
[1] An introduction to the Theory of Numbers 5th Edition by Ivan Niven,
Zuckerman and Montgomery.
Examples
Solve [1, 1, 7] congruent 0 mod(189):
>>> from sympy.polys.galoistools import gf_csolve
>>> gf_csolve([1, 1, 7], 189)
[13, 49, 76, 112, 139, 175]
Manipulation of sparse, distributed polynomials and vectors Dense representations
quickly require infeasible amounts of storage and computation time if the number of variables
increases. For this reason, there is code to manipulate polynomials in a sparse representation.
Sparse polynomials are represented as dictionaries.
5.16. Polynomials Manipulation Module
1213
SymPy Documentation, Release 0.7.6
sympy.polys.rings.ring(symbols, domain, order=LexOrder())
Construct a polynomial ring returning (ring, x 1, ..., x n).
Parameters symbols : str, Symbol/Expr or sequence of str, Symbol/Expr (nonempty)
domain : Domain or coercible
order : Order or coercible, optional, defaults to lex
Examples
>>> from sympy.polys.rings import ring
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.orderings import lex
>>> R, x, y, z = ring(x,y,z, ZZ, lex)
>>> R
Polynomial ring in x, y, z over ZZ with lex order
>>> x + y + z
x + y + z
>>> type(_)
<class sympy.polys.rings.PolyElement>
sympy.polys.rings.xring(symbols, domain, order=LexOrder())
Construct a polynomial ring returning (ring, (x 1, ..., x n)).
Parameters symbols : str, Symbol/Expr or sequence of str, Symbol/Expr (nonempty)
domain : Domain or coercible
order : Order or coercible, optional, defaults to lex
Examples
>>> from sympy.polys.rings import xring
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.orderings import lex
>>> R, (x, y, z) = xring(x,y,z, ZZ, lex)
>>> R
Polynomial ring in x, y, z over ZZ with lex order
>>> x + y + z
x + y + z
>>> type(_)
<class sympy.polys.rings.PolyElement>
sympy.polys.rings.vring(symbols, domain, order=LexOrder())
Construct a polynomial ring and inject x 1, ..., x n into the global namespace.
Parameters symbols : str, Symbol/Expr or sequence of str, Symbol/Expr (nonempty)
domain : Domain or coercible
order : Order or coercible, optional, defaults to lex
1214
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.polys.rings import vring
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.orderings import lex
>>> vring(x,y,z, ZZ, lex)
Polynomial ring in x, y, z over ZZ with lex order
>>> x + y + z
x + y + z
>>> type(_)
<class sympy.polys.rings.PolyElement>
sympy.polys.rings.sring(exprs, *symbols, **options)
Construct a ring deriving generators and domain from options and input expressions.
Parameters exprs : Expr or sequence of Expr (sympiable)
symbols : sequence of Symbol/Expr
options : keyword arguments understood by Options
Examples
>>>
>>>
>>>
>>>
from
from
from
from
sympy.core import symbols
sympy.polys.rings import sring
sympy.polys.domains import ZZ
sympy.polys.orderings import lex
>>> x, y, z = symbols(x,y,z)
>>> R, f = sring(x + 2*y + 3*z)
>>> R
Polynomial ring in x, y, z over ZZ with lex order
>>> f
x + 2*y + 3*z
>>> type(_)
<class sympy.polys.rings.PolyElement>
class sympy.polys.rings.PolyRing
Multivariate distributed polynomial ring.
Attributes
domain
gens
ngens
order
symbols
add(*objs)
Add a sequence of polynomials or containers of polynomials.
5.16. Polynomials Manipulation Module
1215
SymPy Documentation, Release 0.7.6
Example
>>> from sympy.polys.rings import ring
>>> from sympy.polys.domains import ZZ
>>> R, x = ring(x, ZZ)
>>> R.add([ x**2 + 2*i + 3 for i in range(4) ])
4*x**2 + 24
>>> _.factor_list()
(4, [(x**2 + 6, 1)])
drop(*gens)
Remove specied generators from this ring.
drop to ground(*gens)
Remove specied generators from the ring and inject them into its domain.
index(gen)
Compute index of gen in self.gens.
monomial basis(i)
Return the ith-basis element.
mul(*objs)
Multiply a sequence of polynomials or containers of polynomials.
Example
>>> from sympy.polys.rings import ring
>>> from sympy.polys.domains import ZZ
>>> R, x = ring(x, ZZ)
>>> R.mul([ x**2 + 2*i + 3 for i in range(4) ])
x**8 + 24*x**6 + 206*x**4 + 744*x**2 + 945
>>> _.factor_list()
(1, [(x**2 + 3, 1), (x**2 + 5, 1), (x**2 + 7, 1), (x**2 + 9, 1)])
class sympy.polys.rings.PolyElement
Element of multivariate distributed polynomial ring.
almosteq(p1, p2, tolerance=None)
Approximate equality test for polynomials.
cancel(f, g)
Cancel common factors in a rational function f/g.
Examples
>>> from sympy.polys import ring, ZZ
>>> R, x,y = ring(x,y, ZZ)
>>> (2*x**2 - 2).cancel(x**2 - 2*x + 1)
(2*x + 2, x - 1)
coeff(element)
Returns the coecient that stands next to the given monomial.
1216
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Parameters element : PolyElement (with is monomial = True) or 1
Examples
>>> from sympy.polys.rings import ring
>>> from sympy.polys.domains import ZZ
>>> _, x, y, z = ring(x,y,z, ZZ)
>>> f = 3*x**2*y - x*y*z + 7*z**3 + 23
>>> f.coeff(x**2*y)
3
>>> f.coeff(x*y)
0
>>> f.coeff(1)
23
coeffs(order=None)
Ordered list of polynomial coecients.
Parameters order : Order or coercible, optional
Examples
>>> from sympy.polys.rings import ring
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.orderings import lex, grlex
>>> _, x, y = ring(x, y, ZZ, lex)
>>> f = x*y**7 + 2*x**2*y**3
>>>
[2,
>>>
[1,
f.coeffs()
1]
f.coeffs(grlex)
2]
const()
Returns the constant coecient.
content(f)
Returns GCD of polynomials coecients.
copy()
Return a copy of polynomial self.
Polynomials are mutable; if one is interested in preserving a polynomial, and one
plans to use inplace operations, one can copy the polynomial. This method makes a
shallow copy.
Examples
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.rings import ring
5.16. Polynomials Manipulation Module
1217
SymPy Documentation, Release 0.7.6
>>> R, x, y = ring(x, y, ZZ)
>>> p = (x + y)**2
>>> p1 = p.copy()
>>> p2 = p
>>> p[R.zero_monom] = 3
>>> p
x**2 + 2*x*y + y**2 + 3
>>> p1
x**2 + 2*x*y + y**2
>>> p2
x**2 + 2*x*y + y**2 + 3
degree(f, x=None)
The leading degree in x or the main variable.
Note that the degree of 0 is negative innity (the SymPy object -oo).
degrees(f)
A tuple containing leading degrees in all variables.
Note that the degree of 0 is negative innity (the SymPy object -oo)
diff(f, x)
Computes partial derivative in x.
Examples
>>> from sympy.polys.rings import ring
>>> from sympy.polys.domains import ZZ
>>> _, x, y = ring(x,y, ZZ)
>>> p = x + x**2*y**3
>>> p.diff(x)
2*x*y**3 + 1
div(fv)
Division algorithm, see [CLO] p64.
fv array of polynomials return qv, r such that self = sum(fv[i]*qv[i]) + r
All polynomials are required not to be Laurent polynomials.
Examples
>>> from sympy.polys.rings import ring
>>> from sympy.polys.domains import ZZ
>>> _, x, y = ring(x, y, ZZ)
>>> f = x**3
>>> f0 = x - y**2
>>> f1 = x - y
>>> qv, r = f.div((f0, f1))
>>> qv[0]
x**2 + x*y**2 + y**4
>>> qv[1]
0
1218
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> r
y**6
imul num(p, c)
multiply inplace the polynomial p by an element in the coecient ring, provided p
is not one of the generators; else multiply not inplace
Examples
>>> from sympy.polys.rings import ring
>>> from sympy.polys.domains import ZZ
>>> _, x, y = ring(x, y, ZZ)
>>> p = x + y**2
>>> p1 = p.imul_num(3)
>>> p1
3*x + 3*y**2
>>> p1 is p
True
>>> p = x
>>> p1 = p.imul_num(3)
>>> p1
3*x
>>> p1 is p
False
itercoeffs()
Iterator over coecients of a polynomial.
itermonoms()
Iterator over monomials of a polynomial.
iterterms()
Iterator over terms of a polynomial.
leading expv()
Leading monomial tuple according to the monomial ordering.
Examples
>>> from sympy.polys.rings import ring
>>> from sympy.polys.domains import ZZ
>>>
>>>
>>>
(4,
_, x, y, z = ring(x, y, z, ZZ)
p = x**4 + x**3*y + x**2*z**2 + z**7
p.leading_expv()
0, 0)
leading monom()
Leading monomial as a polynomial element.
5.16. Polynomials Manipulation Module
1219
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.polys.rings import ring
>>> from sympy.polys.domains import ZZ
>>> _, x, y = ring(x, y, ZZ)
>>> (3*x*y + y**2).leading_monom()
x*y
leading term()
Leading term as a polynomial element.
Examples
>>> from sympy.polys.rings import ring
>>> from sympy.polys.domains import ZZ
>>> _, x, y = ring(x, y, ZZ)
>>> (3*x*y + y**2).leading_term()
3*x*y
listcoeffs()
Unordered list of polynomial coecients.
listmonoms()
Unordered list of polynomial monomials.
listterms()
Unordered list of polynomial terms.
monic(f)
Divides all coecients by the leading coecient.
monoms(order=None)
Ordered list of polynomial monomials.
Parameters order : Order or coercible, optional
Examples
>>> from sympy.polys.rings import ring
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.orderings import lex, grlex
>>> _, x, y = ring(x, y, ZZ, lex)
>>> f = x*y**7 + 2*x**2*y**3
>>> f.monoms()
[(2, 3), (1, 7)]
>>> f.monoms(grlex)
[(1, 7), (2, 3)]
primitive(f)
Returns content and a primitive polynomial.
square()
square of a polynomial
1220
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.polys.rings import ring
>>> from sympy.polys.domains import ZZ
>>> _, x, y = ring(x, y, ZZ)
>>> p = x + y**2
>>> p.square()
x**2 + 2*x*y**2 + y**4
strip zero()
Eliminate monomials with zero coecient.
tail degree(f, x=None)
The tail degree in x or the main variable.
Note that the degree of 0 is negative innity (the SymPy object -oo)
tail degrees(f)
A tuple containing tail degrees in all variables.
Note that the degree of 0 is negative innity (the SymPy object -oo)
terms(order=None)
Ordered list of polynomial terms.
Parameters order : Order or coercible, optional
Examples
>>> from sympy.polys.rings import ring
>>> from sympy.polys.domains import ZZ
>>> from sympy.polys.orderings import lex, grlex
>>> _, x, y = ring(x, y, ZZ, lex)
>>> f = x*y**7 + 2*x**2*y**3
>>> f.terms()
[((2, 3), 2), ((1, 7), 1)]
>>> f.terms(grlex)
[((1, 7), 1), ((2, 3), 2)]
In commutative algebra, one often studies not only polynomials, but also modules over polynomial rings. The polynomial manipulation module provides rudimentary low-level support
for nitely generated free modules. This is mainly used for Groebner basis computations (see
there), so manipulation functions are only provided to the extend needed. They carry the
prex sdm . Note that in examples, the generators of the free module are called f1 , f2 , . . . .
sympy.polys.distributedmodules.sdm monomial mul(M, X)
Multiply tuple X representing a monomial of K[X] into the tuple M representing a monomial of F .
Examples
Multiplying xy 3 into xf1 yields x2 y 3 f1 :
5.16. Polynomials Manipulation Module
1221
SymPy Documentation, Release 0.7.6
>>> from sympy.polys.distributedmodules import sdm_monomial_mul
>>> sdm_monomial_mul((1, 1, 0), (1, 3))
(1, 2, 3)
sympy.polys.distributedmodules.sdm monomial deg(M)
Return the total degree of M.
Examples
For example, the total degree of x2 yf5 is 3:
>>> from sympy.polys.distributedmodules import sdm_monomial_deg
>>> sdm_monomial_deg((5, 2, 1))
3
sympy.polys.distributedmodules.sdm monomial divides(A, B)
Does there exist a (polynomial) monomial X such that XA = B?
Examples
Positive examples:
In the following examples, the monomial is given in terms of x, y and the generator(s),
f 1, f 2 etc. The tuple form of that monomial is used in the call to sdm monomial divides.
Note: the generator appears last in the expression but rst in the tuple and other factors
appear in the same order that they appear in the monomial expression.
A = f1 divides B = f1
>>> from sympy.polys.distributedmodules import sdm_monomial_divides
>>> sdm_monomial_divides((1, 0, 0), (1, 0, 0))
True
A = f1 divides B = x2 yf1
>>> sdm_monomial_divides((1, 0, 0), (1, 2, 1))
True
A = xyf5 divides B = x2 yf5
>>> sdm_monomial_divides((5, 1, 1), (5, 2, 1))
True
Negative examples:
A = f1 does not divide B = f2
>>> sdm_monomial_divides((1, 0, 0), (2, 0, 0))
False
A = xf1 does not divide B = f1
>>> sdm_monomial_divides((1, 1, 0), (1, 0, 0))
False
A = xy 2 f5 does not divide B = yf5
>>> sdm_monomial_divides((5, 1, 2), (5, 0, 1))
False
1222
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
sympy.polys.distributedmodules.sdm LC(f, K)
Returns the leading coecient of f.
sympy.polys.distributedmodules.sdm to dict(f)
Make a dictionary from a distributed polynomial.
sympy.polys.distributedmodules.sdm from dict(d, O)
Create an sdm from a dictionary.
Here O is the monomial order to use.
>>> from sympy.polys.distributedmodules import sdm_from_dict
>>> from sympy.polys import QQ, lex
>>> dic = {(1, 1, 0): QQ(1), (1, 0, 0): QQ(2), (0, 1, 0): QQ(0)}
>>> sdm_from_dict(dic, lex)
[((1, 1, 0), 1), ((1, 0, 0), 2)]
sympy.polys.distributedmodules.sdm add(f, g, O, K)
Add two module elements f, g.
Addition is done over the ground eld K, monomials are ordered according to O.
Examples
All examples use lexicographic order.
(xyf1 ) + (f2 ) = f2 + xyf1
>>> from sympy.polys.distributedmodules import sdm_add
>>> from sympy.polys import lex, QQ
>>> sdm_add([((1, 1, 1), QQ(1))], [((2, 0, 0), QQ(1))], lex, QQ)
[((2, 0, 0), 1), ((1, 1, 1), 1)]
(xyf1 ) + (xyf1 ) = 0
>>> sdm_add([((1, 1, 1), QQ(1))], [((1, 1, 1), QQ(-1))], lex, QQ)
[]
(f1 ) + (2f1 ) = 3f1
>>> sdm_add([((1, 0, 0), QQ(1))], [((1, 0, 0), QQ(2))], lex, QQ)
[((1, 0, 0), 3)]
(yf1 ) + (xf1 ) = xf1 + yf1
>>> sdm_add([((1, 0, 1), QQ(1))], [((1, 1, 0), QQ(1))], lex, QQ)
[((1, 1, 0), 1), ((1, 0, 1), 1)]
sympy.polys.distributedmodules.sdm LM(f)
Returns the leading monomial of f.
Only valid if f 6= 0.
Examples
>>> from sympy.polys.distributedmodules import sdm_LM, sdm_from_dict
>>> from sympy.polys import QQ, lex
>>> dic = {(1, 2, 3): QQ(1), (4, 0, 0): QQ(1), (4, 0, 1): QQ(1)}
5.16. Polynomials Manipulation Module
1223
SymPy Documentation, Release 0.7.6
>>> sdm_LM(sdm_from_dict(dic, lex))
(4, 0, 1)
sympy.polys.distributedmodules.sdm LT(f)
Returns the leading term of f.
Only valid if f 6= 0.
Examples
>>> from sympy.polys.distributedmodules import sdm_LT, sdm_from_dict
>>> from sympy.polys import QQ, lex
>>> dic = {(1, 2, 3): QQ(1), (4, 0, 0): QQ(2), (4, 0, 1): QQ(3)}
>>> sdm_LT(sdm_from_dict(dic, lex))
((4, 0, 1), 3)
sympy.polys.distributedmodules.sdm mul term(f, term, O, K)
Multiply a distributed module element f by a (polynomial) term term.
Multiplication of coecients is done over the ground eld K, and monomials are ordered
according to O.
Examples
0f1 = 0
>>> from sympy.polys.distributedmodules import sdm_mul_term
>>> from sympy.polys import lex, QQ
>>> sdm_mul_term([((1, 0, 0), QQ(1))], ((0, 0), QQ(0)), lex, QQ)
[]
x0 = 0
>>> sdm_mul_term([], ((1, 0), QQ(1)), lex, QQ)
[]
(x)(f1 ) = xf1
>>> sdm_mul_term([((1, 0, 0), QQ(1))], ((1, 0), QQ(1)), lex, QQ)
[((1, 1, 0), 1)]
(2xy)(3xf1 + 4yf2 ) = 8xy 2 f2 + 6x2 yf1
>>> f = [((2, 0, 1), QQ(4)), ((1, 1, 0), QQ(3))]
>>> sdm_mul_term(f, ((1, 1), QQ(2)), lex, QQ)
[((2, 1, 2), 8), ((1, 2, 1), 6)]
sympy.polys.distributedmodules.sdm zero()
Return the zero module element.
sympy.polys.distributedmodules.sdm deg(f)
Degree of f.
This is the maximum of the degrees of all its monomials. Invalid if f is zero.
1224
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.polys.distributedmodules import sdm_deg
>>> sdm_deg([((1, 2, 3), 1), ((10, 0, 1), 1), ((2, 3, 4), 4)])
7
sympy.polys.distributedmodules.sdm from vector(vec, O, K, **opts)
Create an sdm from an iterable of expressions.
Coecients are created in the ground eld K, and terms are ordered according to monomial order O. Named arguments are passed on to the polys conversion code and can be
used to specify for example generators.
Examples
>>> from sympy.polys.distributedmodules import sdm_from_vector
>>> from sympy.abc import x, y, z
>>> from sympy.polys import QQ, lex
>>> sdm_from_vector([x**2+y**2, 2*z], lex, QQ)
[((1, 0, 0, 1), 2), ((0, 2, 0, 0), 1), ((0, 0, 2, 0), 1)]
sympy.polys.distributedmodules.sdm to vector(f, gens, K, n=None)
Convert sdm f into a list of polynomial expressions.
The generators for the polynomial ring are specied via gens. The rank of the module is
guessed, or passed via n. The ground eld is assumed to be K.
Examples
>>> from sympy.polys.distributedmodules import sdm_to_vector
>>> from sympy.abc import x, y, z
>>> from sympy.polys import QQ, lex
>>> f = [((1, 0, 0, 1), QQ(2)), ((0, 2, 0, 0), QQ(1)), ((0, 0, 2, 0), QQ(1))]
>>> sdm_to_vector(f, [x, y, z], QQ)
[x**2 + y**2, 2*z]
Polynomial factorization algorithms Many variants of Euclids algorithm:
sympy.polys.euclidtools.dmp half gcdex(f, g, u, K)
Half extended Euclidean algorithm in F [X].
Examples
>>> from sympy.polys import ring, ZZ
>>> R, x,y = ring(x,y, ZZ)
sympy.polys.euclidtools.dmp gcdex(f, g, u, K)
Extended Euclidean algorithm in F [X].
5.16. Polynomials Manipulation Module
1225
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.polys import ring, ZZ
>>> R, x,y = ring(x,y, ZZ)
sympy.polys.euclidtools.dmp invert(f, g, u, K)
Compute multiplicative inverse of f modulo g in F [X].
Examples
>>> from sympy.polys import ring, QQ
>>> R, x = ring(x, QQ)
sympy.polys.euclidtools.dmp euclidean prs(f, g, u, K)
Euclidean polynomial remainder sequence (PRS) in K[X].
Examples
>>> from sympy.polys import ring, ZZ
>>> R, x,y = ring(x,y, ZZ)
sympy.polys.euclidtools.dmp primitive prs(f, g, u, K)
Primitive polynomial remainder sequence (PRS) in K[X].
Examples
>>> from sympy.polys import ring, ZZ
>>> R, x,y = ring(x,y, ZZ)
sympy.polys.euclidtools.dmp inner subresultants(f, g, u, K)
Subresultant PRS algorithm in K[X].
Examples
>>> from sympy.polys import ring, ZZ
>>> R, x,y = ring(x,y, ZZ)
>>> f = 3*x**2*y - y**3 - 4
>>> g = x**2 + x*y**3 - 9
>>> a = 3*x*y**4 + y**3 - 27*y + 4
>>> b = -3*y**10 - 12*y**7 + y**6 - 54*y**4 + 8*y**3 + 729*y**2 - 216*y + 16
>>> prs = [f, g, a, b]
>>> beta = [[-1], [1], [9, 0, 0, 0, 0, 0, 0, 0, 0]]
>>> delta = [0, 1, 1]
>>> R.dmp_inner_subresultants(f, g) == (prs, beta, delta)
True
1226
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
sympy.polys.euclidtools.dmp subresultants(f, g, u, K)
Computes subresultant PRS of two polynomials in K[X].
Examples
>>> from sympy.polys import ring, ZZ
>>> R, x,y = ring(x,y, ZZ)
>>> f = 3*x**2*y - y**3 - 4
>>> g = x**2 + x*y**3 - 9
>>> a = 3*x*y**4 + y**3 - 27*y + 4
>>> b = -3*y**10 - 12*y**7 + y**6 - 54*y**4 + 8*y**3 + 729*y**2 - 216*y + 16
>>> R.dmp_subresultants(f, g) == [f, g, a, b]
True
sympy.polys.euclidtools.dmp prs resultant(f, g, u, K)
Resultant algorithm in K[X] using subresultant PRS.
Examples
>>> from sympy.polys import ring, ZZ
>>> R, x,y = ring(x,y, ZZ)
>>> f = 3*x**2*y - y**3 - 4
>>> g = x**2 + x*y**3 - 9
>>> a = 3*x*y**4 + y**3 - 27*y + 4
>>> b = -3*y**10 - 12*y**7 + y**6 - 54*y**4 + 8*y**3 + 729*y**2 - 216*y + 16
>>> res, prs = R.dmp_prs_resultant(f, g)
>>> res == b
False
>>> res == b.drop(x)
True
>>> prs == [f, g, a, b]
True
# resultant has n-1 variables
sympy.polys.euclidtools.dmp zz modular resultant(f, g, p, u, K)
Compute resultant of f and g modulo a prime p.
Examples
>>> from sympy.polys import ring, ZZ
>>> R, x,y = ring(x,y, ZZ)
>>> f = x + y + 2
>>> g = 2*x*y + x + 3
5.16. Polynomials Manipulation Module
1227
SymPy Documentation, Release 0.7.6
>>> R.dmp_zz_modular_resultant(f, g, 5)
-2*y**2 + 1
sympy.polys.euclidtools.dmp zz collins resultant(f, g, u, K)
Collinss modular resultant algorithm in Z[X].
Examples
>>> from sympy.polys import ring, ZZ
>>> R, x,y = ring(x,y, ZZ)
>>> f = x + y + 2
>>> g = 2*x*y + x + 3
>>> R.dmp_zz_collins_resultant(f, g)
-2*y**2 - 5*y + 1
sympy.polys.euclidtools.dmp qq collins resultant(f, g, u, K0)
Collinss modular resultant algorithm in Q[X].
Examples
>>> from sympy.polys import ring, QQ
>>> R, x,y = ring(x,y, QQ)
>>> f = QQ(1,2)*x + y + QQ(2,3)
>>> g = 2*x*y + x + 3
>>> R.dmp_qq_collins_resultant(f, g)
-2*y**2 - 7/3*y + 5/6
sympy.polys.euclidtools.dmp resultant(f, g, u, K, includePRS=False)
Computes resultant of two polynomials in K[X].
Examples
>>> from sympy.polys import ring, ZZ
>>> R, x,y = ring(x,y, ZZ)
>>> f = 3*x**2*y - y**3 - 4
>>> g = x**2 + x*y**3 - 9
>>> R.dmp_resultant(f, g)
-3*y**10 - 12*y**7 + y**6 - 54*y**4 + 8*y**3 + 729*y**2 - 216*y + 16
sympy.polys.euclidtools.dmp discriminant(f, u, K)
Computes discriminant of a polynomial in K[X].
1228
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.polys import ring, ZZ
>>> R, x,y,z,t = ring(x,y,z,t, ZZ)
>>> R.dmp_discriminant(x**2*y + x*z + t)
-4*y*t + z**2
sympy.polys.euclidtools.dmp rr prs gcd(f, g, u, K)
Computes polynomial GCD using subresultants over a ring.
Returns (h, cff, cfg) such that a = gcd(f, g), cff = quo(f, h), and cfg = quo(g,
h).
Examples
>>> from sympy.polys import ring, ZZ
>>> R, x,y, = ring(x,y, ZZ)
>>> f = x**2 + 2*x*y + y**2
>>> g = x**2 + x*y
>>> R.dmp_rr_prs_gcd(f, g)
(x + y, x + y, x)
sympy.polys.euclidtools.dmp ff prs gcd(f, g, u, K)
Computes polynomial GCD using subresultants over a eld.
Returns (h, cff, cfg) such that a = gcd(f, g), cff = quo(f, h), and cfg = quo(g,
h).
Examples
>>> from sympy.polys import ring, QQ
>>> R, x,y, = ring(x,y, QQ)
>>> f = QQ(1,2)*x**2 + x*y + QQ(1,2)*y**2
>>> g = x**2 + x*y
>>> R.dmp_ff_prs_gcd(f, g)
(x + y, 1/2*x + 1/2*y, x)
sympy.polys.euclidtools.dmp zz heu gcd(f, g, u, K)
Heuristic polynomial GCD in Z[X].
Given univariate polynomials f and g in Z[X], returns their GCD and cofactors, i.e. polynomials h, cff and cfg such that:
h = gcd(f, g), cff = quo(f, h) and cfg = quo(g, h)
The algorithm is purely heuristic which means it may fail to compute the GCD. This will
be signaled by raising an exception. In this case you will need to switch to another GCD
method.
5.16. Polynomials Manipulation Module
1229
SymPy Documentation, Release 0.7.6
The algorithm computes the polynomial GCD by evaluating polynomials f and g at certain
points and computing (fast) integer GCD of those evaluations. The polynomial GCD is
recovered from the integer image by interpolation. The evaluation proces reduces f and
g variable by variable into a large integer. The nal step is to verify if the interpolated
polynomial is the correct GCD. This gives cofactors of the input polynomials as a side
eect.
References
1.[Liao95] (page 1911)
[Liao95] (page 1911)
Examples
>>> from sympy.polys import ring, ZZ
>>> R, x,y, = ring(x,y, ZZ)
>>> f = x**2 + 2*x*y + y**2
>>> g = x**2 + x*y
>>> R.dmp_zz_heu_gcd(f, g)
(x + y, x + y, x)
sympy.polys.euclidtools.dmp qq heu gcd(f, g, u, K0)
Heuristic polynomial GCD in Q[X].
Returns (h, cff, cfg) such that a = gcd(f, g), cff = quo(f, h), and cfg = quo(g,
h).
Examples
>>> from sympy.polys import ring, QQ
>>> R, x,y, = ring(x,y, QQ)
>>> f = QQ(1,4)*x**2 + x*y + y**2
>>> g = QQ(1,2)*x**2 + x*y
>>> R.dmp_qq_heu_gcd(f, g)
(x + 2*y, 1/4*x + 1/2*y, 1/2*x)
sympy.polys.euclidtools.dmp inner gcd(f, g, u, K)
Computes polynomial GCD and cofactors of f and g in K[X].
Returns (h, cff, cfg) such that a = gcd(f, g), cff = quo(f, h), and cfg = quo(g,
h).
Examples
>>> from sympy.polys import ring, ZZ
>>> R, x,y, = ring(x,y, ZZ)
1230
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> f = x**2 + 2*x*y + y**2
>>> g = x**2 + x*y
>>> R.dmp_inner_gcd(f, g)
(x + y, x + y, x)
sympy.polys.euclidtools.dmp gcd(f, g, u, K)
Computes polynomial GCD of f and g in K[X].
Examples
>>> from sympy.polys import ring, ZZ
>>> R, x,y, = ring(x,y, ZZ)
>>> f = x**2 + 2*x*y + y**2
>>> g = x**2 + x*y
>>> R.dmp_gcd(f, g)
x + y
sympy.polys.euclidtools.dmp lcm(f, g, u, K)
Computes polynomial LCM of f and g in K[X].
Examples
>>> from sympy.polys import ring, ZZ
>>> R, x,y, = ring(x,y, ZZ)
>>> f = x**2 + 2*x*y + y**2
>>> g = x**2 + x*y
>>> R.dmp_lcm(f, g)
x**3 + 2*x**2*y + x*y**2
sympy.polys.euclidtools.dmp content(f, u, K)
Returns GCD of multivariate coecients.
Examples
>>> from sympy.polys import ring, ZZ
>>> R, x,y, = ring(x,y, ZZ)
>>> R.dmp_content(2*x*y + 6*x + 4*y + 12)
2*y + 6
sympy.polys.euclidtools.dmp primitive(f, u, K)
Returns multivariate content and a primitive polynomial.
5.16. Polynomials Manipulation Module
1231
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.polys import ring, ZZ
>>> R, x,y, = ring(x,y, ZZ)
>>> R.dmp_primitive(2*x*y + 6*x + 4*y + 12)
(2*y + 6, x + 2)
sympy.polys.euclidtools.dmp cancel(f, g, u, K, include=True)
Cancel common factors in a rational function f /g.
Examples
>>> from sympy.polys import ring, ZZ
>>> R, x,y = ring(x,y, ZZ)
>>> R.dmp_cancel(2*x**2 - 2, x**2 - 2*x + 1)
(2*x + 2, x - 1)
Polynomial factorization in characteristic zero:
sympy.polys.factortools.dmp trial division(f, factors, u, K)
Determine multiplicities of factors using trial division.
sympy.polys.factortools.dmp zz mignotte bound(f, u, K)
Mignotte bound for multivariate polynomials in K[X].
sympy.polys.factortools.dup zz hensel step(m, f, g, h, s, t, K)
One step in Hensel lifting in Z[x].
Given positive integer m and Z[x] polynomials f , g, h, s and t such that:
f == g*h (mod m)
s*g + t*h == 1 (mod m)
lc(f) is not a zero divisor (mod m)
lc(h) == 1
deg(f) == deg(g) + deg(h)
deg(s) < deg(h)
deg(t) < deg(g)
returns polynomials G, H, S and T , such that:
f == G*H (mod m**2)
S*G + T**H == 1 (mod m**2)
References
1.[Gathen99] (page 1911)
[Gathen99] (page 1911)
sympy.polys.factortools.dup zz hensel lift(p, f, f list, l, K)
Multifactor Hensel lifting in Z[x].
1232
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Given a prime p, polynomial f over Z[x] such that lc(f ) is a unit modulo p, monic pair-wise
coprime polynomials fi over Z[x] satisfying:
f = lc(f) f 1 ... f r (mod p)
and a positive integer l, returns a list of monic polynomials F1 , F2 , ..., Fr satisfying:
f = lc(f) F 1 ... F r (mod p**l)
F i = f i (mod p), i = 1..r
References
1.[Gathen99] (page 1911)
[Gathen99] (page 1911)
sympy.polys.factortools.dup zz zassenhaus(f, K)
Factor primitive square-free polynomials in Z[x].
sympy.polys.factortools.dup zz irreducible p(f, K)
Test irreducibility using Eisensteins criterion.
sympy.polys.factortools.dup cyclotomic p(f, K, irreducible=False)
Eciently test if f is a cyclotomic polnomial.
Examples
>>> from sympy.polys import ring, ZZ
>>> R, x = ring(x, ZZ)
>>> f = x**16 + x**14 - x**10 + x**8 - x**6 + x**2 + 1
>>> R.dup_cyclotomic_p(f)
False
>>> g = x**16 + x**14 - x**10 - x**8 - x**6 + x**2 + 1
>>> R.dup_cyclotomic_p(g)
True
sympy.polys.factortools.dup zz cyclotomic poly(n, K)
Eciently generate n-th cyclotomic polnomial.
sympy.polys.factortools.dup zz cyclotomic factor(f, K)
Eciently factor polynomials x n 1 and x n + 1 in Z[x].
Given a univariate polynomial f in Z[x] returns a list of factors of f , provided that f is in
the form x n 1 or x n + 1 for n >= 1. Otherwise returns None.
Factorization is performed using using cyclotomic decomposition of f , which makes this
method much faster that any other direct factorization approach (e.g. Zassenhauss).
References
1.[Weisstein09] (page 1911)
[Weisstein09] (page 1911)
5.16. Polynomials Manipulation Module
1233
SymPy Documentation, Release 0.7.6
sympy.polys.factortools.dup zz factor sqf(f, K)
Factor square-free (non-primitive) polyomials in Z[x].
sympy.polys.factortools.dup zz factor(f, K)
Factor (non square-free) polynomials in Z[x].
Given a univariate polynomial f in Z[x] computes its complete factorization f1 , ..., fn into
irreducibles over integers:
f = content(f) f 1**k 1 ... f n**k n
The factorization is computed by reducing the input polynomial into a primitive squarefree polynomial and factoring it using Zassenhaus algorithm. Trial division is used to
recover the multiplicities of factors.
The result is returned as a tuple consisting of:
(content(f), [(f 1, k 1), ..., (f n, k n))
Consider polynomial f = 2 x 4 2:
>>> from sympy.polys import ring, ZZ
>>> R, x = ring(x, ZZ)
>>> R.dup_zz_factor(2*x**4 - 2)
(2, [(x - 1, 1), (x + 1, 1), (x**2 + 1, 1)])
In result we got the following factorization:
f = 2 (x - 1) (x + 1) (x**2 + 1)
Note that this is a complete factorization over integers, however over Gaussian integers
we can factor the last term.
By default, polynomials xn1 and xn+1 are factored using cyclotomic decomposition
to speedup computations. To disable this behaviour set cyclotomic=False.
References
1.[Gathen99] (page 1911)
[Gathen99] (page 1911)
sympy.polys.factortools.dmp zz wang non divisors(E, cs, ct, K)
Wang/EEZ: Compute a set of valid divisors.
sympy.polys.factortools.dmp zz wang test points(f, T, ct, A, u, K)
Wang/EEZ: Test evaluation points for suitability.
sympy.polys.factortools.dmp zz wang lead coeffs(f, T, cs, E, H, A, u, K)
Wang/EEZ: Compute correct leading coecients.
sympy.polys.factortools.dmp zz diophantine(F, c, A, d, p, u, K)
Wang/EEZ: Solve multivariate Diophantine equations.
sympy.polys.factortools.dmp zz wang hensel lifting(f, H, LC, A, p, u, K)
Wang/EEZ: Parallel Hensel lifting algorithm.
sympy.polys.factortools.dmp zz wang(f, u, K, mod=None, seed=None)
Factor primitive square-free polynomials in Z[X].
1234
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Given a multivariate polynomial f in Z[x1 , ..., xn ], which is primitive and square-free in x1 ,
computes factorization of f into irreducibles over integers.
The procedure is based on Wangs Enhanced Extended Zassenhaus algorithm. The algorithm works by viewing f as a univariate polynomial in Z[x2 , ..., xn ][x1 ], for which an
evaluation mapping is computed:
x 2 -> a 2, ..., x n -> a n
where ai , for i = 2, ..., n, are carefully chosen integers. The mapping is used to transform f
into a univariate polynomial in Z[x1 ], which can be factored eciently using Zassenhaus
algorithm. The last step is to lift univariate factors to obtain true multivariate factors.
For this purpose a parallel Hensel lifting procedure is used.
The parameter seed is passed to randint and can be used to seed randint (when an
integer) or (for testing purposes) can be a sequence of numbers.
References
1.[Wang78] (page 1911)
2.[Geddes92] (page 1911)
[Wang78] (page 1911), [Geddes92] (page 1911)
sympy.polys.factortools.dmp zz factor(f, u, K)
Factor (non square-free) polynomials in Z[X].
Given a multivariate polynomial f in Z[x] computes its complete factorization f1 , ..., fn into
irreducibles over integers:
f = content(f) f 1**k 1 ... f n**k n
The factorization is computed by reducing the input polynomial into a primitive squarefree polynomial and factoring it using Enhanced Extended Zassenhaus (EEZ) algorithm.
Trial division is used to recover the multiplicities of factors.
The result is returned as a tuple consisting of:
(content(f), [(f 1, k 1), ..., (f n, k n))
Consider polynomial f = 2 (x 2 y 2):
>>> from sympy.polys import ring, ZZ
>>> R, x,y = ring(x,y, ZZ)
>>> R.dmp_zz_factor(2*x**2 - 2*y**2)
(2, [(x - y, 1), (x + y, 1)])
In result we got the following factorization:
f = 2 (x - y) (x + y)
References
1.[Gathen99] (page 1911)
[Gathen99] (page 1911)
5.16. Polynomials Manipulation Module
1235
SymPy Documentation, Release 0.7.6
sympy.polys.factortools.dmp ext factor(f, u, K)
Factor multivariate polynomials over algebraic number elds.
sympy.polys.factortools.dup gf factor(f, K)
Factor univariate polynomials over nite elds.
sympy.polys.factortools.dmp factor list(f, u, K0)
Factor polynomials into irreducibles in K[X].
sympy.polys.factortools.dmp factor list include(f, u, K)
Factor polynomials into irreducibles in K[X].
sympy.polys.factortools.dmp irreducible p(f, u, K)
Returns True if f has no factors over its domain.
Groebner basis algorithms Groebner bases can be used to answer many problems in
computational commutative algebra. Their computation in rather complicated, and very
performance-sensitive. We present here various low-level implementations of Groebner basis
computation algorithms; please see the previous section of the manual for usage.
sympy.polys.groebnertools.groebner(seq, ring, method=None)
Computes Groebner basis for a set of polynomials in K[X].
Wrapper around the (default) improved Buchberger and the other algorithms for computing Groebner bases. The choice of algorithm can be changed via method argument
or setup() from sympy.polys.polyconfig, where method can be either buchberger or
f5b.
sympy.polys.groebnertools.spoly(p1, p2, ring)
Compute LCM(LM(p1), LM(p2))/LM(p1)*p1 - LCM(LM(p1), LM(p2))/LM(p2)*p2 This is
the S-poly provided p1 and p2 are monic
sympy.polys.groebnertools.red groebner(G, ring)
Compute reduced Groebner basis, from BeckerWeispfenning93, p. 216
Selects a subset of generators, that already generate the ideal and computes a reduced
Groebner basis for them.
sympy.polys.groebnertools.is groebner(G, ring)
Check if G is a Groebner basis.
sympy.polys.groebnertools.is minimal(G, ring)
Checks if G is a minimal Groebner basis.
sympy.polys.groebnertools.is reduced(G, ring)
Checks if G is a reduced Groebner basis.
sympy.polys.fglmtools.matrix fglm(F, ring, O to)
Converts the reduced Groebner basis F of a zero-dimensional ideal w.r.t. O from to a
reduced Groebner basis w.r.t. O to.
References
J.C. Faugere, P. Gianni, D. Lazard, T. Mora (1994).
dimensional Groebner Bases by Change of Ordering
Ecient Computation of Zero-
Groebner basis algorithms for modules are also provided:
sympy.polys.distributedmodules.sdm spoly(f, g, O, K, phantom=None)
Compute the generalized s-polynomial of f and g.
1236
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
The ground eld is assumed to be K, and monomials ordered according to O.
This is invalid if either of f or g is zero.
If the leading terms of f and g involve dierent basis elements of F , their s-poly is dened
to be zero. Otherwise it is a certain linear combination of f and g in which the leading
terms cancel. See [SCA, defn 2.3.6] for details.
If phantom is not None, it should be a pair of module elements on which to perform the
same operation(s) as on f and g. The in this case both results are returned.
Examples
>>> from sympy.polys.distributedmodules import sdm_spoly
>>> from sympy.polys import QQ, lex
>>> f = [((2, 1, 1), QQ(1)), ((1, 0, 1), QQ(1))]
>>> g = [((2, 3, 0), QQ(1))]
>>> h = [((1, 2, 3), QQ(1))]
>>> sdm_spoly(f, h, lex, QQ)
[]
>>> sdm_spoly(f, g, lex, QQ)
[((1, 2, 1), 1)]
sympy.polys.distributedmodules.sdm ecart(f)
Compute the ecart of f.
This is dened to be the dierence of the total degree of f and the total degree of the
leading monomial of f [SCA, defn 2.3.7].
Invalid if f is zero.
Examples
>>> from sympy.polys.distributedmodules import sdm_ecart
>>> sdm_ecart([((1, 2, 3), 1), ((1, 0, 1), 1)])
0
>>> sdm_ecart([((2, 2, 1), 1), ((1, 5, 1), 1)])
3
sympy.polys.distributedmodules.sdm nf mora(f, G, O, K, phantom=None)
Compute a weak normal form of f with respect to G and order O.
The ground eld is assumed to be K, and monomials ordered according to O.
Weak normal forms are dened in [SCA, defn 2.3.3]. They are not unique. This function
deterministically computes a weak normal form, depending on the order of G.
The most important property of a weak normal form is the following: if R is the ring associated with the monomial ordering (if the ordering is global, we just have R = K[x1 , . . . , xn ],
otherwise it is a certain localization thereof), I any ideal of R and G a standard basis for
I, then for any f R, we have f I if and only if N F (f |G) = 0.
This is the generalized Mora algorithm for computing weak normal forms with respect
to arbitrary monomial orders [SCA, algorithm 2.3.9].
If phantom is not None, it should be a pair of phantom arguments on which to perform
the same computations as on f, G, both results are then returned.
5.16. Polynomials Manipulation Module
1237
SymPy Documentation, Release 0.7.6
sympy.polys.distributedmodules.sdm groebner(G, NF, O, K, extended=False)
Compute a minimal standard basis of G with respect to order O.
The algorithm uses a normal form NF, for example sdm nf mora. The ground eld is
assumed to be K, and monomials ordered according to O.
Let N denote the submodule generated by elements of G. A standard basis for N is a
subset S of N , such that in(S) = in(N ), where for any subset X of F , in(X) denotes the
submodule generated by the initial forms of elements of X. [SCA, defn 2.3.2]
A standard basis is called minimal if no subset of it is a standard basis.
One may show that standard bases are always generating sets.
Minimal standard bases are not unique. This algorithm computes a deterministic result,
depending on the particular order of G.
If extended=True, also compute the transition matrix from the initial generators to the
groebner basis. That is, return a list of coecient vectors, expressing the elements of
the groebner basis in terms of the elements of G.
This functions implements the sugar strategy, see
Giovini et al: One sugar cube, please OR Selection strategies in Buchberger algorithm.
Exceptions
These are exceptions dened by the polynomials module.
TODO sort and explain
class sympy.polys.polyerrors.BasePolynomialError
Base class for polynomial related exceptions.
class sympy.polys.polyerrors.ExactQuotientFailed(f, g, dom=None)
class sympy.polys.polyerrors.OperationNotSupported(poly, func)
class sympy.polys.polyerrors.HeuristicGCDFailed
class sympy.polys.polyerrors.HomomorphismFailed
class sympy.polys.polyerrors.IsomorphismFailed
class sympy.polys.polyerrors.ExtraneousFactors
class sympy.polys.polyerrors.EvaluationFailed
class sympy.polys.polyerrors.RefinementFailed
class sympy.polys.polyerrors.CoercionFailed
class sympy.polys.polyerrors.NotInvertible
class sympy.polys.polyerrors.NotReversible
class sympy.polys.polyerrors.NotAlgebraic
class sympy.polys.polyerrors.DomainError
class sympy.polys.polyerrors.PolynomialError
class sympy.polys.polyerrors.UnificationFailed
class sympy.polys.polyerrors.GeneratorsNeeded
class sympy.polys.polyerrors.ComputationFailed(func, nargs, exc)
class sympy.polys.polyerrors.GeneratorsError
1238
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
class sympy.polys.polyerrors.UnivariatePolynomialError
class sympy.polys.polyerrors.MultivariatePolynomialError
class sympy.polys.polyerrors.PolificationFailed(opt, origs, exprs, seq=False)
class sympy.polys.polyerrors.OptionError
class sympy.polys.polyerrors.FlagError
Reference
Modular GCD
class sympy.polys.modulargcd.modgcd univariate
Computes the GCD of two polynomials in Z[x] using a modular algorithm.
The algorithm computes the GCD of two univariate integer polynomials f and g by computing the GCD in Zp [x] for suitable primes p and then reconstructing the coecients
with the Chinese Remainder Theorem. Trial division is only made for candidates which
are very likely the desired GCD.
Parameters f : PolyElement
univariate integer polynomial
g : PolyElement
univariate integer polynomial
Returns h : PolyElement
GCD of the polynomials f and g
c : PolyElement
cofactor of f , i.e.
f
h
cfg : PolyElement
cofactor of g, i.e.
g
h
References
1.[Monagan00] (page 1911)
[Monagan00] (page 1911)
Examples
>>> from sympy.polys.modulargcd import modgcd_univariate
>>> from sympy.polys import ring, ZZ
>>> R, x = ring(x, ZZ)
>>> f = x**5 - 1
>>> g = x - 1
>>> h, cff, cfg = modgcd_univariate(f, g)
>>> h, cff, cfg
(x - 1, x**4 + x**3 + x**2 + x + 1, 1)
5.16. Polynomials Manipulation Module
1239
SymPy Documentation, Release 0.7.6
>>> cff * h == f
True
>>> cfg * h == g
True
>>> f = 6*x**2 - 6
>>> g = 2*x**2 + 4*x + 2
>>> h, cff, cfg = modgcd_univariate(f, g)
>>> h, cff, cfg
(2*x + 2, 3*x - 3, x + 1)
>>> cff * h == f
True
>>> cfg * h == g
True
class sympy.polys.modulargcd.modgcd bivariate
Computes the GCD of two polynomials in Z[x, y] using a modular algorithm.
The algorithm computes the GCD of two bivariate integer polynomials f and g by calculating the GCD in Zp [x, y] for suitable primes p and then reconstructing the coecients
with the Chinese Remainder Theorem. To compute the bivariate GCD over Zp , the polynomials f mod p and g mod p are evaluated at y = a for certain a Zp and then their
univariate GCD in Zp [x] is computed. Interpolating those yields the bivariate GCD in
Zp [x, y]. To verify the result in Z[x, y], trial division is done, but only for candidates which
are very likely the desired GCD.
Parameters f : PolyElement
bivariate integer polynomial
g : PolyElement
bivariate integer polynomial
Returns h : PolyElement
GCD of the polynomials f and g
c : PolyElement
cofactor of f , i.e.
f
h
cfg : PolyElement
cofactor of g, i.e.
g
h
References
1.[Monagan00] (page 1911)
[Monagan00] (page 1911)
Examples
>>> from sympy.polys.modulargcd import modgcd_bivariate
>>> from sympy.polys import ring, ZZ
1240
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> R, x, y = ring(x, y, ZZ)
>>> f = x**2 - y**2
>>> g = x**2 + 2*x*y + y**2
>>> h, cff, cfg = modgcd_bivariate(f, g)
>>> h, cff, cfg
(x + y, x - y, x + y)
>>> cff * h == f
True
>>> cfg * h == g
True
>>> f = x**2*y - x**2 - 4*y + 4
>>> g = x + 2
>>> h, cff, cfg = modgcd_bivariate(f, g)
>>> h, cff, cfg
(x + 2, x*y - x - 2*y + 2, 1)
>>> cff * h == f
True
>>> cfg * h == g
True
class sympy.polys.modulargcd.modgcd multivariate
Compute the GCD of two polynomials in Z[x0 , . . . , xk1 ] using a modular algorithm.
The algorithm computes the GCD of two multivariate integer polynomials f and g by
calculating the GCD in Zp [x0 , . . . , xk1 ] for suitable primes p and then reconstructing the
coecients with the Chinese Remainder Theorem. To compute the multivariate GCD
over Zp the recursive subroutine modgcd multivariate p is used. To verify the result
in Z[x0 , . . . , xk1 ], trial division is done, but only for candidates which are very likely the
desired GCD.
Parameters f : PolyElement
multivariate integer polynomial
g : PolyElement
multivariate integer polynomial
Returns h : PolyElement
GCD of the polynomials f and g
c : PolyElement
cofactor of f , i.e.
f
h
cfg : PolyElement
cofactor of g, i.e.
g
h
See Also:
modgcd multivariate p
5.16. Polynomials Manipulation Module
1241
SymPy Documentation, Release 0.7.6
References
1.[Monagan00] (page 1911)
2.[Brown71] (page 1912)
[Monagan00] (page 1911), [Brown71] (page 1912)
Examples
>>> from sympy.polys.modulargcd import modgcd_multivariate
>>> from sympy.polys import ring, ZZ
>>> R, x, y = ring(x, y, ZZ)
>>> f = x**2 - y**2
>>> g = x**2 + 2*x*y + y**2
>>> h, cff, cfg = modgcd_multivariate(f, g)
>>> h, cff, cfg
(x + y, x - y, x + y)
>>> cff * h == f
True
>>> cfg * h == g
True
>>> R, x, y, z = ring(x, y, z, ZZ)
>>> f = x*z**2 - y*z**2
>>> g = x**2*z + z
>>> h, cff, cfg = modgcd_multivariate(f, g)
>>> h, cff, cfg
(z, x*z - y*z, x**2 + 1)
>>> cff * h == f
True
>>> cfg * h == g
True
class sympy.polys.modulargcd.func field modgcd
Compute the GCD of two polynomials f and g in Q()[x0 , . . . , xn1 ] using a modular algorithm.
The algorithm rst computes the primitive associate m
(z) of the minimal polynomial
m in Z[z] and the primitive associates of f and g in Z[x1 , . . . , xn1 ][z]/(m
)[x0 ]. Then
it computes the GCD in Q(x1 , . . . , xn1 )[z]/(m (z))[x0 ]. This is done by calculating the
GCD in Zp (x1 , . . . , xn1 )[z]/(m
(z))[x0 ] for suitable primes p and then reconstructing the
coecients with the Chinese Remainder Theorem and Rational Reconstuction. The
GCD over Zp (x1 , . . . , xn1 )[z]/(m
(z))[x0 ] is computed with a recursive subroutine, which
evaluates the polynomials at xn1 = a for suitable evaluation points a Zp and then
calls itself recursively until the ground domain does no longer contain any parameters. For Zp [z]/(m
(z))[x0 ] the Euclidean Algorithm is used. The results of those recursive calls are then interpolated and Rational Function Reconstruction is used to
1242
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
obtain the correct coecients. The results, both in Q(x1 , . . . , xn1 )[z]/(m (z))[x0 ] and
Zp (x1 , . . . , xn1 )[z]/(m
(z))[x0 ], are veried by a fraction free trial division.
Apart from the above GCD computation some GCDs in Q()[x1 , . . . , xn1 ] have to be calculated, because treating the polynomials as univariate ones can result in a spurious
content of the GCD. For this func field modgcd is called recursively.
Parameters f, g : PolyElement
polynomials in Q()[x0 , . . . , xn1 ]
Returns h : PolyElement
monic GCD of the polynomials f and g
c : PolyElement
cofactor of f , i.e.
f
h
cfg : PolyElement
cofactor of g, i.e.
g
h
References
1.[Hoeij04] (page 1912)
[Hoeij04] (page 1912)
Examples
>>> from sympy.polys.modulargcd import func_field_modgcd
>>> from sympy.polys import AlgebraicField, QQ, ring
>>> from sympy import sqrt
>>> A = AlgebraicField(QQ, sqrt(2))
>>> R, x = ring(x, A)
>>> f = x**2 - 2
>>> g = x + sqrt(2)
>>> h, cff, cfg = func_field_modgcd(f, g)
>>> h == x + sqrt(2)
True
>>> cff * h == f
True
>>> cfg * h == g
True
>>> R, x, y = ring(x, y, A)
>>> f = x**2 + 2*sqrt(2)*x*y + 2*y**2
>>> g = x + sqrt(2)*y
>>> h, cff, cfg = func_field_modgcd(f, g)
5.16. Polynomials Manipulation Module
1243
SymPy Documentation, Release 0.7.6
>>> h == x + sqrt(2)*y
True
>>> cff * h == f
True
>>> cfg * h == g
True
>>> f = x + sqrt(2)*y
>>> g = x + y
>>> h, cff, cfg = func_field_modgcd(f, g)
>>> h == R.one
True
>>> cff * h == f
True
>>> cfg * h == g
True
Manipulation of power series Functions in this module carry the prex rs , standing for
ring series. They manipulate nite power series in the sparse representation provided by
polys.ring.ring.
sympy.polys.ring series.rs trunc(p1, x, prec)
truncate the series in the x variable with precision prec, that is modulo O(x**prec)
Examples
>>> from sympy.polys.domains import QQ
>>> from sympy.polys.rings import ring
>>> from sympy.polys.ring_series import rs_trunc
>>> R, x = ring(x, QQ)
>>> p = x**10 + x**5 + x + 1
>>> rs_trunc(p, x, 12)
x**10 + x**5 + x + 1
>>> rs_trunc(p, x, 10)
x**5 + x + 1
sympy.polys.ring series.rs mul(p1, p2, x, prec)
product of series modulo O(x**prec)
x is the series variable or its position in the generators.
Examples
>>> from sympy.polys.domains import QQ
>>> from sympy.polys.rings import ring
>>> from sympy.polys.ring_series import rs_mul
>>> R, x = ring(x, QQ)
>>> p1 = x**2 + 2*x + 1
>>> p2 = x + 1
>>> rs_mul(p1, p2, x, 3)
3*x**2 + 3*x + 1
1244
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
sympy.polys.ring series.rs square(p1, x, prec)
square modulo O(x**prec)
Examples
>>> from sympy.polys.domains import QQ
>>> from sympy.polys.rings import ring
>>> from sympy.polys.ring_series import rs_square
>>> R, x = ring(x, QQ)
>>> p = x**2 + 2*x + 1
>>> rs_square(p, x, 3)
6*x**2 + 4*x + 1
sympy.polys.ring series.rs pow(p1, n, x, prec)
return p1**n modulo O(x**prec)
Examples
>>> from sympy.polys.domains import QQ
>>> from sympy.polys.rings import ring
>>> from sympy.polys.ring_series import rs_pow
>>> R, x = ring(x, QQ)
>>> p = x + 1
>>> rs_pow(p, 4, x, 3)
6*x**2 + 4*x + 1
sympy.polys.ring series.rs series inversion(p, x, prec)
multivariate series inversion 1/p modulo O(x**prec)
Examples
>>> from sympy.polys.domains import QQ
>>> from sympy.polys.rings import ring
>>> from sympy.polys.ring_series import rs_series_inversion
>>> R, x, y = ring(x, y, QQ)
>>> rs_series_inversion(1 + x*y**2, x, 4)
-x**3*y**6 + x**2*y**4 - x*y**2 + 1
>>> rs_series_inversion(1 + x*y**2, y, 4)
-x*y**2 + 1
sympy.polys.ring series.rs series from list(p, c, x, prec, concur=1)
series sum c[n]*p**n modulo O(x**prec)
reduce the number of multiplication summing concurrently ax = [1, p, p**2, ..,
p**(J - 1)] s = sum(c[i]*ax[i] for i in range(r, (r + 1)*J))*p**((K - 1)*J)
with K >= (n + 1)/J
See Also:
sympy.polys.ring.compose
5.16. Polynomials Manipulation Module
1245
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.polys.domains import QQ
>>> from sympy.polys.rings import ring
>>> from sympy.polys.ring_series import rs_series_from_list, rs_trunc
>>> R, x = ring(x, QQ)
>>> p = x**2 + x + 1
>>> c = [1, 2, 3]
>>> rs_series_from_list(p, c, x, 4)
6*x**3 + 11*x**2 + 8*x + 6
>>> rs_trunc(1 + 2*p + 3*p**2, x, 4)
6*x**3 + 11*x**2 + 8*x + 6
>>> pc = R.from_list(list(reversed(c)))
>>> rs_trunc(pc.compose(x, p), x, 4)
6*x**3 + 11*x**2 + 8*x + 6
sympy.polys.ring series.rs integrate(self, x)
integrate p with respect to x
Examples
>>> from sympy.polys.domains import QQ
>>> from sympy.polys.rings import ring
>>> from sympy.polys.ring_series import rs_integrate
>>> R, x, y = ring(x, y, QQ)
>>> p = x + x**2*y**3
>>> rs_integrate(p, x)
1/3*x**3*y**3 + 1/2*x**2
sympy.polys.ring series.rs log(p, x, prec)
logarithm of p modulo O(x**prec)
Notes
truncation of integral dx p**-1*d p/dx is used.
Examples
>>> from sympy.polys.domains import QQ
>>> from sympy.polys.rings import ring
>>> from sympy.polys.ring_series import rs_log
>>> R, x = ring(x, QQ)
>>> rs_log(1 + x, x, 8)
1/7*x**7 - 1/6*x**6 + 1/5*x**5 - 1/4*x**4 + 1/3*x**3 - 1/2*x**2 + x
sympy.polys.ring series.rs exp(p, x, prec)
exponentiation of a series modulo O(x**prec)
Examples
1246
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> from sympy.polys.domains import QQ
>>> from sympy.polys.rings import ring
>>> from sympy.polys.ring_series import rs_exp
>>> R, x = ring(x, QQ)
>>> rs_exp(x**2, x, 7)
1/6*x**6 + 1/2*x**4 + x**2 + 1
sympy.polys.ring series.rs newton(p, x, prec)
compute the truncated Newton sum of the polynomial p
Examples
>>> from sympy.polys.domains import QQ
>>> from sympy.polys.rings import ring
>>> from sympy.polys.ring_series import rs_newton
>>> R, x = ring(x, QQ)
>>> p = x**2 - 2
>>> rs_newton(p, x, 5)
8*x**4 + 4*x**2 + 2
sympy.polys.ring series.rs hadamard exp(p1, inverse=False)
return sum f i/i!*x**i from sum f i*x**i, where x is the rst variable.
If invers=True return sum f i*i!*x**i
Examples
>>> from sympy.polys.domains import QQ
>>> from sympy.polys.rings import ring
>>> from sympy.polys.ring_series import rs_hadamard_exp
>>> R, x = ring(x, QQ)
>>> p = 1 + x + x**2 + x**3
>>> rs_hadamard_exp(p)
1/6*x**3 + 1/2*x**2 + x + 1
sympy.polys.ring series.rs compose add(p1, p2)
compute the composed sum prod(p2(x - beta) for beta root of p1)
References
A. Bostan, P. Flajolet, B. Salvy and E. Schost Fast Computation with Two Algebraic Numbers, (2002) Research Report 4579, Institut National de Recherche en Informatique et
en Automatique
Examples
>>>
>>>
>>>
>>>
>>>
>>>
from sympy.polys.domains import QQ
from sympy.polys.rings import ring
from sympy.polys.ring_series import rs_compose_add
R, x = ring(x, QQ)
f = x**2 - 2
g = x**2 - 3
5.16. Polynomials Manipulation Module
1247
SymPy Documentation, Release 0.7.6
>>> rs_compose_add(f, g)
x**4 - 10*x**2 + 1
Undocumented
Many parts of the polys module are still undocumented, and even where there is documentation it is scarce. Please contribute!
Literature
The following is a non-comprehensive list of publications that were used as a theoretical
foundation for implementing polynomials manipulation module.
5.17 Printing System
See the Printing (page 18) section in Tutorial for introduction into printing.
This guide documents the printing system in SymPy and how it works internally.
5.17.1 Printer Class
Printing subsystem driver
SymPys printing system works the following way: Any expression can be passed to a designated Printer who then is responsible to return an adequate representation of that expression.
The basic concept is the following:
1. Let the object print itself if it knows how.
2. Take the best tting method dened in the printer.
3. As fall-back use the emptyPrinter method for the printer.
Some more information how the single concepts work and who should use which:
1. The object prints itself
This was the original way of doing printing in sympy. Every class had its own
latex, mathml, str and repr methods, but it turned out that it is hard to produce
a high quality printer, if all the methods are spread out that far. Therefor all
printing code was combined into the dierent printers, which works great for
built-in sympy objects, but not that good for user dened classes where it is
inconvenient to patch the printers.
Nevertheless, to get a tting representation, the printers look for a specic
method in every object, that will be called if its available and is then responsible
for the representation. The name of that method depends on the specic printer
and is dened under Printer.printmethod.
2. Take the best tting method dened in the printer.
The printer loops through expr classes (class + its bases), and tries to dispatch
the work to print <EXPR CLASS>
e.g., suppose we have the following class hierarchy:
1248
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Basic
|
Atom
|
Number
|
Rational
then, for expr=Rational(...), in order to dispatch, we will try calling printer methods as shown in the gure below:
p. print(expr)
|
|-- p. print Rational(expr)
|
|-- p. print Number(expr)
|
|-- p. print Atom(expr)
|
-- p. print Basic(expr)
if . print Rational method exists in the printer, then it is called, and the result is
returned back.
otherwise, we proceed with trying Rational bases in the inheritance order.
3. As fall-back use the emptyPrinter method for the printer.
As fall-back self.emptyPrinter will be called with the expression. If not dened
in the Printer subclass this will be the same as str(expr).
The main class responsible for printing is Printer (see also its source code):
class sympy.printing.printer.Printer(settings=None)
Generic printer
Its job is to provide infrastructure for implementing new printers easily.
Basically, if you want to implement a printer, all you have to do is:
1.Subclass Printer.
2.Dene Printer.printmethod in your subclass. If a object has a method with that name,
this method will be used for printing.
3.In your subclass, dene print <CLASS> methods
For each class you want to provide printing to, dene an appropriate method how
to do it. For example if you want a class FOO to be printed in its own way, dene
print FOO:
def _print_FOO(self, e):
...
this should return how FOO instance e is printed
Also, if BAR is a subclass of FOO, print FOO(bar) will be called for instance of BAR, if
no print BAR is provided. Thus, usually, we dont need to provide printing routines
for every class we want to support only generic routine has to be provided for a
set of classes.
A good example for this are functions - for example PrettyPrinter only denes
print Function, and there is no print sin, print tan, etc...
5.17. Printing System
1249
SymPy Documentation, Release 0.7.6
On the other hand, a good printer will probably have to dene separate routines for
Symbol, Atom, Number, Integral, Limit, etc...
4.If convenient, override self.emptyPrinter
This callable will be called to obtain printing result as a last resort, that is when no
appropriate print method was found for an expression.
Examples of overloading StrPrinter:
from sympy import Basic, Function, Symbol
from sympy.printing.str import StrPrinter
class CustomStrPrinter(StrPrinter):
Examples of how to customize the StrPrinter for both a SymPy class and a
user defined class subclassed from the SymPy Basic class.
def _print_Derivative(self, expr):
Custom printing of the SymPy Derivative class.
Instead of:
D(x(t), t) or D(x(t), t, t)
We will print:
x
or
In this example, expr.args == (x(t), t), and expr.args[0] == x(t), and
expr.args[0].func == x
return str(expr.args[0].func) + *len(expr.args[1:])
def _print_MyClass(self, expr):
Print the characters of MyClass.s alternatively lower case and upper
case
s =
i = 0
for char in expr.s:
if i % 2 == 0:
s += char.lower()
else:
s += char.upper()
i += 1
return s
# Override the __str__ method of to use CustromStrPrinter
Basic.__str__ = lambda self: CustomStrPrinter().doprint(self)
# Demonstration of CustomStrPrinter:
t = Symbol(t)
x = Function(x)(t)
# dxdt is a Derivative instance
dxdt = x.diff(t)
d2xdt2 = dxdt.diff(t)
# dxdt2 is a Derivative instance
ex = MyClass(I like both lowercase and upper case)
1250
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
print dxdt
print d2xdt2
print ex
The output of the above code is:
x
x
i lIkE BoTh lOwErCaSe aNd uPpEr cAsE
By overriding Basic. str , we can customize the printing of anything that is subclassed
from Basic.
Attributes
printmethod
printmethod = None
doprint(expr)
Returns printers representation for expr (as a string)
print(expr, *args, **kwargs)
Internal dispatcher
Tries the following concepts to print an expression:
1. Let the object print itself if it knows how.
2. Take the best tting method dened in the printer.
3. As fall-back use the emptyPrinter method for the printer.
classmethod set global settings(**settings)
Set system-wide printing settings.
5.17.2 PrettyPrinter Class
The pretty printing subsystem is implemented in sympy.printing.pretty.pretty
by the PrettyPrinter class deriving from Printer.
It relies on the modules
sympy.printing.pretty.stringPict,
and sympy.printing.pretty.pretty symbology
for rendering nice-looking formulas.
The module stringPict provides a base class stringPict and a derived class prettyForm
that ease the creation and manipulation of formulas that span across multiple lines.
The module pretty symbology provides primitives to construct 2D shapes (hline, vline, etc)
together with a technique to use unicode automatically when possible.
class sympy.printing.pretty.pretty.PrettyPrinter(settings=None)
Printer, which converts an expression into 2D ASCII-art gure.
printmethod = pretty
sympy.printing.pretty.pretty.pretty(expr, **settings)
Returns a string containing the prettied form of expr.
For information on keyword arguments see pretty print function.
5.17. Printing System
1251
SymPy Documentation, Release 0.7.6
sympy.printing.pretty.pretty.pretty print(expr, **settings)
Prints expr in pretty form.
pprint is just a shortcut for this function.
Parameters expr : expression
the expression to print
wrap line : bool, optional
line wrapping enabled/disabled, defaults to True
num columns : int or None, optional
number of columns before line breaking (default to None which reads
the terminal width), useful when using SymPy without terminal.
use unicode : bool or None, optional
use unicode characters, such as the Greek letter pi instead of the
string pi.
full prec : bool or string, optional
use full precision. Default to auto
order : bool or string, optional
set to none for long expressions if slow; default is None
5.17.3 CCodePrinter
This class implements C code printing (i.e. it converts Python expressions to strings of C
code).
Usage:
>>> from sympy.printing import print_ccode
>>> from sympy.functions import sin, cos, Abs
>>> from sympy.abc import x
>>> print_ccode(sin(x)**2 + cos(x)**2)
pow(sin(x), 2) + pow(cos(x), 2)
>>> print_ccode(2*x + cos(x), assign_to=result)
result = 2*x + cos(x);
>>> print_ccode(Abs(x**2))
fabs(pow(x, 2))
sympy.printing.ccode.known functions = {sinh: sinh, asin: asin, cos: cos, log: log, atan
class sympy.printing.ccode.CCodePrinter(settings={})
A printer to convert python expressions to strings of c code
printmethod = ccode
indent code(code)
Accepts a string of code or a list of code lines
sympy.printing.ccode.ccode(expr, assign to=None, **settings)
Converts an expr to a string of c code
Parameters expr : Expr
A sympy expression to be converted.
assign to : optional
1252
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
When given, the argument is used as the name of the variable to
which the expression is assigned. Can be a string, Symbol, MatrixSymbol, or Indexed type. This is helpful in case of line-wrapping,
or for expressions that generate multi-line statements.
precision : integer, optional
The precision for numbers such as pi [default=15].
user functions : dict, optional
A dictionary where keys are FunctionClass instances and values are
their string representations. Alternatively, the dictionary value can
be a list of tuples i.e. [(argument test, cfunction string)]. See below
for examples.
dereference : iterable, optional
An iterable of symbols that should be dereferenced in the printed
code expression. These would be values passed by address to the
function. For example, if dereference=[a], the resulting code would
print (*a) instead of a.
human : bool, optional
If True, the result is a single string that may contain some constant
declarations for the number symbols. If False, the same information
is returned in a tuple of (symbols to declare, not supported functions,
code text). [default=True].
contract: bool, optional :
If True, Indexed instances are assumed to obey tensor contraction
rules and the corresponding nested loops over indices are generated. Setting contract=False will not generate loops, instead the
user is responsible to provide values for the indices in the code. [default=True].
Examples
>>> from sympy import ccode, symbols, Rational, sin, ceiling, Abs
>>> x, tau = symbols(x, tau)
>>> ccode((2*tau)**Rational(7, 2))
8*sqrt(2)*pow(tau, 7.0L/2.0L)
>>> ccode(sin(x), assign_to=s)
s = sin(x);
Custom printing can be dened for certain types by passing a dictionary of type :
function to the user functions kwarg. Alternatively, the dictionary value can be a list
of tuples i.e. [(argument test, cfunction string)].
>>> custom_functions = {
...
ceiling: CEIL,
...
Abs: [(lambda x: not x.is_integer, fabs),
...
(lambda x: x.is_integer, ABS)]
... }
>>> ccode(Abs(x) + ceiling(x), user_functions=custom_functions)
fabs(x) + CEIL(x)
5.17. Printing System
1253
SymPy Documentation, Release 0.7.6
Piecewise expressions are converted into conditionals. If an assign to variable is provided an if statement is created, otherwise the ternary operator is used. Note that if
the Piecewise lacks a default term, represented by (expr, True) then an error will be
thrown. This is to prevent generating an expression that may not evaluate to anything.
>>> from sympy import Piecewise
>>> expr = Piecewise((x + 1, x > 0), (x, True))
>>> print(ccode(expr, tau))
if (x > 0) {
tau = x + 1;
}
else {
tau = x;
}
Support for loops is provided through Indexed types. With contract=True these expressions will be turned into loops, whereas contract=False will just print the assignment
expression that should be looped over:
>>> from sympy import Eq, IndexedBase, Idx
>>> len_y = 5
>>> y = IndexedBase(y, shape=(len_y,))
>>> t = IndexedBase(t, shape=(len_y,))
>>> Dy = IndexedBase(Dy, shape=(len_y-1,))
>>> i = Idx(i, len_y-1)
>>> e=Eq(Dy[i], (y[i+1]-y[i])/(t[i+1]-t[i]))
>>> ccode(e.rhs, assign_to=e.lhs, contract=False)
Dy[i] = (y[i + 1] - y[i])/(t[i + 1] - t[i]);
Matrices are also supported, but a MatrixSymbol of the same dimensions must be provided to assign to. Note that any expression that can be generated normally can also
exist inside a Matrix:
>>> from sympy import Matrix, MatrixSymbol
>>> mat = Matrix([x**2, Piecewise((x + 1, x > 0), (x, True)), sin(x)])
>>> A = MatrixSymbol(A, 3, 1)
>>> print(ccode(mat, A))
A[0] = pow(x, 2);
if (x > 0) {
A[1] = x + 1;
}
else {
A[1] = x;
}
A[2] = sin(x);
sympy.printing.ccode.print ccode(expr, **settings)
Prints C representation of the given expression.
5.17.4 Fortran Printing
The fcode function translates a sympy expression into Fortran code. The main purpose is
to take away the burden of manually translating long mathematical expressions. Therefore
the resulting expression should also require no (or very little) manual tweaking to make it
compilable. The optional arguments of fcode can be used to ne-tune the behavior of fcode
in such a way that manual changes in the result are no longer needed.
1254
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
sympy.printing.fcode.fcode(expr, assign to=None, **settings)
Converts an expr to a string of c code
Parameters expr : Expr
A sympy expression to be converted.
assign to : optional
When given, the argument is used as the name of the variable to
which the expression is assigned. Can be a string, Symbol, MatrixSymbol, or Indexed type. This is helpful in case of line-wrapping,
or for expressions that generate multi-line statements.
precision : integer, optional
The precision for numbers such as pi [default=15].
user functions : dict, optional
A dictionary where keys are FunctionClass instances and values are
their string representations. Alternatively, the dictionary value can
be a list of tuples i.e. [(argument test, cfunction string)]. See below
for examples.
human : bool, optional
If True, the result is a single string that may contain some constant
declarations for the number symbols. If False, the same information
is returned in a tuple of (symbols to declare, not supported functions,
code text). [default=True].
contract: bool, optional :
If True, Indexed instances are assumed to obey tensor contraction
rules and the corresponding nested loops over indices are generated. Setting contract=False will not generate loops, instead the
user is responsible to provide values for the indices in the code. [default=True].
source format : optional
The source format can be either xed or free. [default=xed]
standard : integer, optional
The Fortran standard to be followed. This is specied as an integer.
Acceptable standards are 66, 77, 90, 95, 2003, and 2008. Default
is 77. Note that currently the only distinction internally is between
standards before 95, and those 95 and after. This may change later
as more features are added.
Examples
>>>
>>>
>>>
>>>
from sympy import fcode, symbols, Rational, sin, ceiling, floor
x, tau = symbols(x, tau)
fcode((2*tau)**Rational(7, 2))
8*sqrt(2.0d0)*tau**(7.0d0/2.0d0)
fcode(sin(x), assign_to=s)
s = sin(x)
5.17. Printing System
1255
SymPy Documentation, Release 0.7.6
Custom printing can be dened for certain types by passing a dictionary of type :
function to the user functions kwarg. Alternatively, the dictionary value can be a list
of tuples i.e. [(argument test, cfunction string)].
>>> custom_functions = {
...
ceiling: CEIL,
...
floor: [(lambda x: not x.is_integer, FLOOR1),
...
(lambda x: x.is_integer, FLOOR2)]
... }
>>> fcode(floor(x) + ceiling(x), user_functions=custom_functions)
CEIL(x) + FLOOR1(x)
Piecewise expressions are converted into conditionals. If an assign to variable is provided an if statement is created, otherwise the ternary operator is used. Note that if
the Piecewise lacks a default term, represented by (expr, True) then an error will be
thrown. This is to prevent generating an expression that may not evaluate to anything.
>>> from sympy import Piecewise
>>> expr = Piecewise((x + 1, x > 0), (x, True))
>>> print(fcode(expr, tau))
if (x > 0) then
tau = x + 1
else
tau = x
end if
Support for loops is provided through Indexed types. With contract=True these expressions will be turned into loops, whereas contract=False will just print the assignment
expression that should be looped over:
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
from sympy import Eq, IndexedBase, Idx
len_y = 5
y = IndexedBase(y, shape=(len_y,))
t = IndexedBase(t, shape=(len_y,))
Dy = IndexedBase(Dy, shape=(len_y-1,))
i = Idx(i, len_y-1)
e=Eq(Dy[i], (y[i+1]-y[i])/(t[i+1]-t[i]))
fcode(e.rhs, assign_to=e.lhs, contract=False)
Dy(i) = (y(i + 1) - y(i))/(t(i + 1) - t(i))
Matrices are also supported, but a MatrixSymbol of the same dimensions must be provided to assign to. Note that any expression that can be generated normally can also
exist inside a Matrix:
>>>
>>>
>>>
>>>
from sympy import Matrix, MatrixSymbol
mat = Matrix([x**2, Piecewise((x + 1, x > 0), (x, True)), sin(x)])
A = MatrixSymbol(A, 3, 1)
print(fcode(mat, A))
A(1, 1) = x**2
if (x > 0) then
A(2, 1) = x + 1
else
A(2, 1) = x
end if
A(3, 1) = sin(x)
sympy.printing.fcode.print fcode(expr, **settings)
Prints the Fortran representation of the given expression.
See fcode for the meaning of the optional arguments.
1256
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
class sympy.printing.fcode.FCodePrinter(settings={})
A printer to convert sympy expressions to strings of Fortran code
printmethod = fcode
indent code(code)
Accepts a string of code or a list of code lines
Two basic examples:
>>>
>>>
>>>
>>>
from sympy import *
x = symbols(x)
fcode(sqrt(1-x**2))
sqrt(-x**2 + 1)
fcode((3 + 4*I)/(1 - conjugate(x)))
(cmplx(3,4))/(-conjg(x) + 1)
An example where line wrapping is required:
>>> expr = sqrt(1-x**2).series(x,n=20).removeO()
>>> print(fcode(expr))
-715.0d0/65536.0d0*x**18 - 429.0d0/32768.0d0*x**16 - 33.0d0/
@ 2048.0d0*x**14 - 21.0d0/1024.0d0*x**12 - 7.0d0/256.0d0*x**10 @ 5.0d0/128.0d0*x**8 - 1.0d0/16.0d0*x**6 - 1.0d0/8.0d0*x**4 - 1.0d0
@ /2.0d0*x**2 + 1
In case of line wrapping, it is handy to include the assignment so that lines are wrapped
properly when the assignment part is added.
>>> print(fcode(expr, assign_to=var))
var = -715.0d0/65536.0d0*x**18 - 429.0d0/32768.0d0*x**16 - 33.0d0/
@ 2048.0d0*x**14 - 21.0d0/1024.0d0*x**12 - 7.0d0/256.0d0*x**10 @ 5.0d0/128.0d0*x**8 - 1.0d0/16.0d0*x**6 - 1.0d0/8.0d0*x**4 - 1.0d0
@ /2.0d0*x**2 + 1
For piecewise functions, the assign to option is mandatory:
>>> print(fcode(Piecewise((x,x<1),(x**2,True)), assign_to=var))
if (x < 1) then
var = x
else
var = x**2
end if
Note that by default only top-level piecewise functions are supported due to the lack of a
conditional operator in Fortran 77. Inline conditionals can be supported using the merge
function introduced in Fortran 95 by setting of the kwarg standard=95:
>>> print(fcode(Piecewise((x,x<1),(x**2,True)), standard=95))
merge(x, x**2, x < 1)
Loops are generated if there are Indexed objects in the expression. This also requires use of
the assign to option.
>>>
>>>
>>>
>>>
A, B = map(IndexedBase, [A, B])
m = Symbol(m, integer=True)
i = Idx(i, m)
print(fcode(2*B[i], assign_to=A[i]))
do i = 1, m
A(i) = 2*B(i)
end do
5.17. Printing System
1257
SymPy Documentation, Release 0.7.6
Repeated indices in an expression with Indexed objects are interpreted as summation. For
instance, code for the trace of a matrix can be generated with
>>> print(fcode(A[i, i], assign_to=x))
x = 0
do i = 1, m
x = x + A(i, i)
end do
By default, number symbols such as pi and E are detected and dened as Fortran parameters. The precision of the constants can be tuned with the precision argument. Parameter
denitions are easily avoided using the N function.
>>> print(fcode(x - pi**2 - E))
parameter (E = 2.71828182845905d0)
parameter (pi = 3.14159265358979d0)
x - pi**2 - E
>>> print(fcode(x - pi**2 - E, precision=25))
parameter (E = 2.718281828459045235360287d0)
parameter (pi = 3.141592653589793238462643d0)
x - pi**2 - E
>>> print(fcode(N(x - pi**2, 25)))
x - 9.869604401089358618834491d0
When some functions are not part of the Fortran standard, it might be desirable to introduce
the names of user-dened functions in the Fortran expression.
>>> print(fcode(1 - gamma(x)**2, user_functions={gamma: mygamma}))
-mygamma(x)**2 + 1
However, when the user functions argument is not provided, fcode attempts to use a reasonable default and adds a comment to inform the user of the issue.
>>> print(fcode(1 - gamma(x)**2))
C
Not supported in Fortran:
C
gamma
-gamma(x)**2 + 1
By default the output is human readable code, ready for copy and paste. With the option
human=False, the return value is suitable for post-processing with source code generators
that write routines with multiple instructions. The return value is a three-tuple containing:
(i) a set of number symbols that must be dened as Fortran parameters, (ii) a list functions
that can not be translated in pure Fortran and (iii) a string of Fortran code. A few examples:
>>> fcode(1 - gamma(x)**2, human=False)
(set(), set([gamma(x)]),
-gamma(x)**2 + 1)
>>> fcode(1 - sin(x)**2, human=False)
(set(), set(),
-sin(x)**2 + 1)
>>> fcode(x - pi**2, human=False)
(set([(pi, 3.14159265358979d0)]), set(),
x - pi**2)
5.17.5 Mathematica code printing
sympy.printing.mathematica.known functions = {asin: [(<function <lambda> at 0xa02bdbc>, Ar
class sympy.printing.mathematica.MCodePrinter(settings={})
A printer to convert python expressions to strings of the Wolframs Mathematica code
1258
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
printmethod = mcode
doprint(expr)
Returns printers representation for expr (as a string)
sympy.printing.mathematica.mathematica code(expr, **settings)
Converts an expr to a string of the Wolfram Mathematica code
Examples
>>> from sympy import mathematica_code as mcode, symbols, sin
>>> x = symbols(x)
>>> mcode(sin(x).series(x).removeO())
(1/120)*x^5 - 1/6*x^3 + x
5.17.6 Gtk
You can print to a grkmathview widget using the function print gtk located in
sympy.printing.gtk (it requires to have installed gtkmatmatview and libgtkmathview-bin
in some systems).
GtkMathView accepts MathML, so this rendering depends on the MathML representation of
the expression.
Usage:
from sympy import *
print_gtk(x**2 + 2*exp(x**3))
sympy.printing.gtk.print gtk(x, start viewer=True)
Print to Gtkmathview, a gtk widget capable of rendering MathML.
Needs libgtkmathview-bin
5.17.7 LambdaPrinter
This classes implements printing to strings that can
sympy.utilities.lambdify.lambdify() (page 1573) function.
be
used
by
the
class sympy.printing.lambdarepr.LambdaPrinter(settings=None)
This printer converts expressions into strings that can be used by lambdify.
printmethod = sympystr
sympy.printing.lambdarepr.lambdarepr(expr, **settings)
Returns a string usable for lambdifying.
5.17.8 LatexPrinter
This class implements LaTeX printing. See sympy.printing.latex.
sympy.printing.latex.accepted latex functions = [arcsin, arccos, arctan, sin, cos, tan, si
list() -> new empty list list(iterable) -> new list initialized from iterables items
class sympy.printing.latex.LatexPrinter(settings=None)
5.17. Printing System
1259
SymPy Documentation, Release 0.7.6
printmethod = latex
sympy.printing.latex.latex(expr, **settings)
Convert the given expression to LaTeX representation.
>>> from sympy import latex, pi, sin, asin, Integral, Matrix, Rational
>>> from sympy.abc import x, y, mu, r, tau
>>> print(latex((2*tau)**Rational(7,2)))
8 \sqrt{2} \tau^{\frac{7}{2}}
order: Any of the supported monomial orderings (currently lex, grlex, or grevlex),
old, and none. This parameter does nothing for Mul objects. Setting order to old
uses the compatibility ordering for Add dened in Printer. For very large expressions,
set the order keyword to none if speed is a concern.
mode: Species how the generated code will be delimited. mode can be one of plain,
inline, equation or equation*. If mode is set to plain, then the resulting code will
not be delimited at all (this is the default). If mode is set to inline then inline LaTeX
$ $ will be used. If mode is set to equation or equation*, the resulting code will be
enclosed in the equation or equation* environment (remember to import amsmath
for equation*), unless the itex option is set. In the latter case, the $$ $$ syntax is used.
>>> print(latex((2*mu)**Rational(7,2), mode=plain))
8 \sqrt{2} \mu^{\frac{7}{2}}
>>> print(latex((2*tau)**Rational(7,2), mode=inline))
$8 \sqrt{2} \tau^{\frac{7}{2}}$
>>> print(latex((2*mu)**Rational(7,2), mode=equation*))
\begin{equation*}8 \sqrt{2} \mu^{\frac{7}{2}}\end{equation*}
>>> print(latex((2*mu)**Rational(7,2), mode=equation))
\begin{equation}8 \sqrt{2} \mu^{\frac{7}{2}}\end{equation}
itex: Species if itex-specic syntax is used, including emitting $$ $$.
>>> print(latex((2*mu)**Rational(7,2), mode=equation, itex=True))
$$8 \sqrt{2} \mu^{\frac{7}{2}}$$
fold frac powers: Emit {p/q} instead of {frac{p}{q}} for fractional powers.
>>> print(latex((2*tau)**Rational(7,2), fold_frac_powers=True))
8 \sqrt{2} \tau^{7/2}
fold func brackets: Fold function brackets where applicable.
>>> print(latex((2*tau)**sin(Rational(7,2))))
\left(2 \tau\right)^{\sin{\left (\frac{7}{2} \right )}}
>>> print(latex((2*tau)**sin(Rational(7,2)), fold_func_brackets = True))
\left(2 \tau\right)^{\sin {\frac{7}{2}}}
fold short frac: Emit p / q instead of frac{p}{q} when the denominator is simple
enough (at most two terms and no powers). The default value is T rue for inline mode,
False otherwise.
>>> print(latex(3*x**2/y))
\frac{3 x^{2}}{y}
1260
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> print(latex(3*x**2/y, fold_short_frac=True))
3 x^{2} / y
long frac ratio: The allowed ratio of the width of the numerator to the width of the denominator before we start breaking o long fractions. The default value is 2.
>>> print(latex(Integral(r, r)/2/pi, long_frac_ratio=2))
\frac{\int r\, dr}{2 \pi}
>>> print(latex(Integral(r, r)/2/pi, long_frac_ratio=0))
\frac{1}{2 \pi} \int r\, dr
mul symbol: The symbol to use for multiplication. Can be one of None, ldot, dot, or
times.
>>> print(latex((2*tau)**sin(Rational(7,2)), mul_symbol=times))
\left(2 \times \tau\right)^{\sin{\left (\frac{7}{2} \right )}}
inv trig style: How inverse trig functions should be displayed. Can be one of abbreviated, full, or power. Defaults to abbreviated.
>>> print(latex(asin(Rational(7,2))))
\operatorname{asin}{\left (\frac{7}{2} \right )}
>>> print(latex(asin(Rational(7,2)), inv_trig_style=full))
\arcsin{\left (\frac{7}{2} \right )}
>>> print(latex(asin(Rational(7,2)), inv_trig_style=power))
\sin^{-1}{\left (\frac{7}{2} \right )}
mat str: Which matrix environment string to emit. smallmatrix, matrix, array,
etc. Defaults to smallmatrix for inline mode, matrix for matrices of no more than 10
columns, and array otherwise.
>>> print(latex(Matrix(2, 1, [x, y])))
\left[\begin{matrix}x\\y\end{matrix}\right]
>>> print(latex(Matrix(2, 1, [x, y]), mat_str = array))
\left[\begin{array}{c}x\\y\end{array}\right]
mat delim: The delimiter to wrap around matrices. Can be one of [, (, or the empty
string. Defaults to [.
>>> print(latex(Matrix(2, 1, [x, y]), mat_delim=())
\left(\begin{matrix}x\\y\end{matrix}\right)
symbol names: Dictionary of symbols and the custom strings they should be emitted as.
>>> print(latex(x**2, symbol_names={x:x_i}))
x_i^{2}
latex also supports the builtin container types list, tuple, and dictionary.
>>> print(latex([2/x, y], mode=inline))
$\left [ 2 / x, \quad y\right ]$
sympy.printing.latex.print latex(expr, **settings)
Prints LaTeX representation of the given expression.
5.17. Printing System
1261
SymPy Documentation, Release 0.7.6
5.17.9 MathMLPrinter
This class is responsible for MathML printing. See sympy.printing.mathml.
More info on mathml content: http://www.w3.org/TR/MathML2/chapter4.html
class sympy.printing.mathml.MathMLPrinter(settings=None)
Prints an expression to the MathML markup language
Whenever possible tries to use Content markup and not Presentation markup.
References: http://www.w3.org/TR/MathML2/
printmethod = mathml
doprint(expr)
Prints the expression as MathML.
mathml tag(e)
Returns the MathML tag for an expression.
sympy.printing.mathml.mathml(expr, **settings)
Returns the MathML representation of expr
sympy.printing.mathml.print mathml(expr, **settings)
Prints a pretty representation of the MathML code for expr
Examples
>>> ##
>>> from sympy.printing.mathml import print_mathml
>>> from sympy.abc import x
>>> print_mathml(x+1)
<apply>
<plus/>
<ci>x</ci>
<cn>1</cn>
</apply>
5.17.10 PythonPrinter
This class implements Python printing. Usage:
>>> from sympy import print_python, sin
>>> from sympy.abc import x
>>> print_python(5*x**3 + sin(x))
x = Symbol(x)
e = 5*x**3 + sin(x)
5.17.11 ReprPrinter
This printer generates executable code. This code satises the identity eval(srepr(expr))
== expr.
class sympy.printing.repr.ReprPrinter(settings=None)
1262
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
printmethod = sympyrepr
emptyPrinter(expr)
The fallback printer.
reprify(args, sep)
Prints each item in args and joins them with sep.
sympy.printing.repr.srepr(expr, **settings)
return expr in repr form
5.17.12 StrPrinter
This module generates readable representations of SymPy expressions.
class sympy.printing.str.StrPrinter(settings=None)
printmethod = sympystr
sympy.printing.str.sstrrepr(expr, **settings)
return expr in mixed str/repr form
i.e. strings are returned in repr form with quotes, and everything else is returned in str
form.
This function could be useful for hooking into sys.displayhook
5.17.13 Tree Printing
The functions in this module create a representation of an expression as a tree.
sympy.printing.tree.pprint nodes(subtrees)
Prettyprints systems of nodes.
Examples
>>> from sympy.printing.tree import pprint_nodes
>>> print(pprint_nodes([a, b1\nb2, c]))
+-a
+-b1
| b2
+-c
sympy.printing.tree.print node(node)
Returns information about the node.
This includes class name, string representation and assumptions.
sympy.printing.tree.tree(node)
Returns a tree representation of node as a string.
It uses print node() together with pprint nodes() on node.args recursively.
See also: print tree()
sympy.printing.tree.print tree(node)
Prints a tree representation of node.
5.17. Printing System
1263
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.printing import print_tree
>>> from sympy import Symbol
>>> x = Symbol(x, odd=True)
>>> y = Symbol(y, even=True)
>>> print_tree(y**x)
Pow: y**x
+-Symbol: y
| algebraic: True
| commutative: True
| complex: True
| even: True
| hermitian: True
| imaginary: False
| integer: True
| irrational: False
| noninteger: False
| odd: False
| rational: True
| real: True
| transcendental: False
+-Symbol: x
algebraic: True
commutative: True
complex: True
even: False
hermitian: True
imaginary: False
integer: True
irrational: False
noninteger: False
nonzero: True
odd: True
rational: True
real: True
transcendental: False
zero: False
See also: tree()
5.17.14 Preview
A useful function is preview:
sympy.printing.preview.preview(expr, output=png, viewer=None, euler=True,
packages=(), lename=None, outputbuer=None,
preamble=None, dvioptions=None, outputTexFile=None, **latex settings)
View expression or LaTeX markup in PNG, DVI, PostScript or PDF form.
If the expr argument is an expression, it will be exported to LaTeX and then compiled
using the available TeX distribution. The rst argument, expr, may also be a LaTeX
string. The function will then run the appropriate viewer for the given output format or
use the user dened one. By default png output is generated.
By default pretty Euler fonts are used for typesetting (they were used to typeset the
well known Concrete Mathematics book). For that to work, you need the eulervm.sty
1264
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
LaTeX style (in Debian/Ubuntu, install the texlive-fonts-extra package). If you prefer
default AMS fonts or your system lacks eulervm LaTeX package then unset the euler
keyword argument.
To use viewer auto-detection, lets say for png output, issue
>>> from sympy import symbols, preview, Symbol
>>> x, y = symbols(x,y)
>>> preview(x + y, output=png)
This will choose pyglet by default. To select a dierent one, do
>>> preview(x + y, output=png, viewer=gimp)
The png format is considered special. For all other formats the rules are slightly dierent. As an example we will take dvi output format. If you would run
>>> preview(x + y, output=dvi)
then view will look for available dvi viewers on your system (predened in the function,
so it will try evince, rst, then kdvi and xdvi). If nothing is found you will need to set the
viewer explicitly.
>>> preview(x + y, output=dvi, viewer=superior-dvi-viewer)
This will skip auto-detection and will run user specied superior-dvi-viewer. If view
fails to nd it on your system it will gracefully raise an exception.
You may also enter le for the viewer argument. Doing so will cause this function to
return a le object in read-only mode, if lename is unset. However, if it was set, then
preview writes the genereted le to this lename instead.
There is also support for writing to a BytesIO like object, which needs to be passed to
the outputbuer argument.
>>> from io import BytesIO
>>> obj = BytesIO()
>>> preview(x + y, output=png, viewer=BytesIO,
...
outputbuffer=obj)
The LaTeX preamble can be customized by setting the preamble keyword argument.
This can be used, e.g., to set a dierent font size, use a custom documentclass or import
certain set of LaTeX packages.
>>> preamble = \\documentclass[10pt]{article}\n \
...
\\usepackage{amsmath,amsfonts}\\begin{document}
>>> preview(x + y, output=png, preamble=preamble)
If the value of output is dierent from dvi then command line options can be set (dvioptions argument) for the execution of the dvi+output conversion tool. These options
have to be in the form of a list of strings (see subprocess.Popen).
Additional keyword args will be passed to the latex call, e.g., the symbol names ag.
>>> phidd = Symbol(phidd)
>>> preview(phidd, symbol_names={phidd:r\ddot{\varphi}})
For post-processing the generated TeX File can be written to a le by passing the desired
lename to the outputTexFile keyword argument. To write the TeX code to a le named
sample.tex and run the default png viewer to display the resulting bitmap, do
5.17. Printing System
1265
SymPy Documentation, Release 0.7.6
>>> preview(x + y, outputTexFile=sample.tex)
5.17.15 Implementation - Helper Classes/Functions
sympy.printing.conventions.split super sub(text)
Split a symbol name into a name, superscripts and subscripts
The rst part of the symbol name is considered to be its actual name, followed by
super- and subscripts. Each superscript is preceded with a character or by .
Each subscript is preceded by a character. The three return values are the actual
name, a list with superscripts and a list with subscripts.
>>> from sympy.printing.conventions import split_super_sub
>>> split_super_sub(a_x^1)
(a, [1], [x])
>>> split_super_sub(var_sub1__sup_sub2)
(var, [sup], [sub1, sub2])
CodePrinter
This class is a base class for other classes that implement code-printing functionality, and
additionally lists a number of functions that cannot be easily translated to C or Fortran.
class sympy.printing.codeprinter.CodePrinter(settings=None)
The base class for code-printing subclasses.
printmethod = sympystr
exception sympy.printing.codeprinter.AssignmentError
Raised if an assignment variable for a loop is missing.
Precedence
sympy.printing.precedence.PRECEDENCE = {And: 30, Add: 40, Pow: 60, Xor: 10, Mul: 50, No
Default precedence values for some basic types.
sympy.printing.precedence.PRECEDENCE VALUES = {And: 30, Xor: 10, Sub: 40, factorial: 60, Po
A dictionary assigning precedence values to certain classes. These values are treated
like they were inherited, so not every single class has to be named here.
sympy.printing.precedence.PRECEDENCE FUNCTIONS = {FracElement: <function precedence FracEl
Sometimes its not enough to assign a xed precedence value to a class. Then a function
can be inserted in this dictionary that takes an instance of this class as argument and
returns the appropriate precedence value.
sympy.printing.precedence.precedence(item)
Returns the precedence of a given object.
5.17.16 Pretty-Printing Implementation Helpers
sympy.printing.pretty.pretty symbology.U(name)
unicode character by name or None if not found
1266
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
sympy.printing.pretty.pretty symbology.pretty use unicode(ag=None)
Set whether pretty-printer should use unicode by default
sympy.printing.pretty.pretty symbology.pretty try use unicode()
See if unicode output is available and leverage it if possible
sympy.printing.pretty.pretty symbology.xstr(*args)
call str or unicode depending on current mode
The following two functions return the Unicode version of the inputted Greek letter.
sympy.printing.pretty.pretty symbology.g(l)
sympy.printing.pretty.pretty symbology.G(l)
sympy.printing.pretty.pretty symbology.greek letters = [alpha, beta, gamma, delta, epsilo
list() -> new empty list list(iterable) -> new list initialized from iterables items
sympy.printing.pretty.pretty symbology.digit 2txt = {1: ONE, 0: ZERO, 3: THREE, 2:
sympy.printing.pretty.pretty symbology.symb 2txt = {int: INTEGRAL, {}: CURLY BRACKET,
The following functions return the Unicode subscript/superscript version of the character.
sympy.printing.pretty.pretty symbology.sub = {): u\u208e, chi: u\u1d6a, +: u\u208a, -:
sympy.printing.pretty.pretty symbology.sup = {): u\u207e, i: u\u2071, (: u\u207d, +: u
The following functions return Unicode vertical objects.
sympy.printing.pretty.pretty symbology.xobj(symb, length)
Construct spatial object of given length.
return: [] of equal-length strings
sympy.printing.pretty.pretty symbology.vobj(symb, height)
Construct vertical object of a given height
see: xobj
sympy.printing.pretty.pretty symbology.hobj(symb, width)
Construct horizontal object of a given width
see: xobj
The following constants are for rendering roots and fractions.
sympy.printing.pretty.pretty symbology.root = {2: u\u221a, 3: u\u221b, 4: u\u221c}
sympy.printing.pretty.pretty symbology.VF(txt)
sympy.printing.pretty.pretty symbology.frac = {(1, 3): u\u2153, (5, 6): u\u215a, (1, 4): u\xb
The following constants/functions are for rendering atoms and symbols.
sympy.printing.pretty.pretty symbology.xsym(sym)
get symbology for a character
sympy.printing.pretty.pretty symbology.atoms table = {Integers: u\u2124, NegativeInnity:
sympy.printing.pretty.pretty symbology.pretty atom(atom name, default=None)
return pretty representation of an atom
sympy.printing.pretty.pretty symbology.pretty symbol(symb name)
return pretty representation of a symbol
sympy.printing.pretty.pretty symbology.annotated(letter)
Return a stylised drawing of the letter letter, together with information on how to put
annotations (super- and subscripts to the left and to the right) on it.
See pretty.py functions print meijerg, print hyper on how to use this information.
5.17. Printing System
1267
SymPy Documentation, Release 0.7.6
Prettyprinter by Jurjen Bos. (I hate spammers: mail me at pietjepuk314 at the reverse of
ku.oc.oohay). All objects have a method that create a stringPict, that can be used in the str
method for pretty printing.
Updates by Jason Gedge (email <my last name> at cs mun ca)
terminal string() method
minor xes and changes (mostly to prettyForm)
TODO:
Allow left/center/right alignment options for above/below and top/center/bottom
alignment options for left/right
class sympy.printing.pretty.stringpict.stringPict(s, baseline=0)
An ASCII picture. The pictures are represented as a list of equal length strings.
above(*args)
Put pictures above this picture. Returns string, baseline arguments for stringPict.
Baseline is baseline of bottom picture.
below(*args)
Put pictures under this picture. Returns string, baseline arguments for stringPict.
Baseline is baseline of top picture
Examples
>>> from sympy.printing.pretty.stringpict import stringPict
>>> print(stringPict(x+3).below(
...
stringPict.LINE, 3)[0])
x+3
--3
height()
The height of the picture in characters.
left(*args)
Put pictures (left to right) at left. Returns string, baseline arguments for stringPict.
leftslash()
Precede object by a slash of the proper size.
static next(*args)
Put a string of stringPicts next to each other. Returns string, baseline arguments
for stringPict.
parens(left=(, right=), ifascii nougly=False)
Put parentheses around self. Returns string, baseline arguments for stringPict.
left or right can be None or empty string which means no paren from that side
render(*args, **kwargs)
Return the string form of self.
Unless the argument line break is set to False, it will break the expression in a form
that can be printed on the terminal without being broken up.
right(*args)
Put pictures next to this one. Returns string, baseline arguments for stringPict.
(Multiline) strings are allowed, and are given a baseline of 0.
1268
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.printing.pretty.stringpict import stringPict
>>> print(stringPict(10).right( + ,stringPict(1\r-\r2,1))[0])
1
10 + 2
root(n=None)
Produce a nice root symbol. Produces ugly results for big n inserts.
static stack(*args)
Put pictures on top of each other, from top to bottom. Returns string, baseline arguments for stringPict. The baseline is the baseline of the second picture. Everything
is centered. Baseline is the baseline of the second picture. Strings are allowed. The
special value stringPict.LINE is a row of - extended to the width.
terminal width()
Return the terminal width if possible, otherwise return 0.
width()
The width of the picture in characters.
class sympy.printing.pretty.stringpict.prettyForm(s, baseline=0, binding=0,
unicode=None)
Extension of the stringPict class that knows about basic math applications, optimizing
double minus signs.
Binding is interpreted as follows:
ATOM
FUNC
DIV
POW
MUL
ADD
NEG
this is an atom: never needs to be parenthesized
this is a function application: parenthesize if added (?)
this is a division: make wider division if divided
this is a power: only parenthesize if exponent
this is a multiplication: parenthesize if powered
this is an addition: parenthesize if multiplied or powered
this is a negative number: optimize if added, parenthesize if
multiplied or powered
OPEN this is an open object: parenthesize if added, multiplied, or
powered (example: Piecewise)
static apply(function, *args)
Functions of one or more variables.
5.17.17 dotprint
sympy.printing.dot.dotprint(expr,
styles=[(<class
sympy.core.basic.Basic>,
{color:
blue,
shape:
ellipse}),
(<class
sympy.core.expr.Expr>,
{color:
black})],
atom=<function
<lambda>
at
0x1997fe9c>,
maxdepth=None,
repeat=True,
labelfunc=<type
str>, **kwargs)
DOT description of a SymPy expression tree
Options are
styles: Styles for dierent classes. The default is:
5.17. Printing System
1269
SymPy Documentation, Release 0.7.6
[(Basic, {color: blue, shape: ellipse}),
(Expr, {color: black})]
atom: Function used to determine if an arg is an atom. The default is lambda x:
not isinstance(x, Basic). Another good choice is lambda x: not x.args.
maxdepth: The maximum depth. The default is None, meaning no limit.
repeat: Whether to dierent nodes for separate common subexpressions. The
default is True. For example, for x + x*y with repeat=True, it will have two nodes
for x and with repeat=False, it will have one (warning: even if it appears twice
in the same object, like Pow(x, x), it will still only appear only once. Hence, with
repeat=False, the number of arrows out of an object might not equal the number of
args it has).
labelfunc: How to label leaf nodes. The default is str. Another good option is
srepr. For example with str, the leaf nodes of x + 1 are labeled, x and 1. With
srepr, they are labeled Symbol(x) and Integer(1).
Additional keyword arguments are included as styles for the graph.
Examples
>>> from sympy.printing.dot import dotprint
>>> from sympy.abc import x
>>> print(dotprint(x+2))
digraph{
# Graph style
ordering=out
rankdir=TD
#########
# Nodes #
#########
Add(Integer(2), Symbol(x))_() [color=black, label=Add, shape=ellipse];
Integer(2)_(0,) [color=black, label=2, shape=ellipse];
Symbol(x)_(1,) [color=black, label=x, shape=ellipse];
#########
# Edges #
#########
Add(Integer(2), Symbol(x))_() -> Integer(2)_(0,);
Add(Integer(2), Symbol(x))_() -> Symbol(x)_(1,);
}
5.18 Plotting Module
5.18.1 Introduction
The plotting module allows you to make 2-dimensional and 3-dimensional plots. Presently the
plots are rendered using matplotlib as a backend. It is also possible to plot 2-dimensional
1270
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
plots using a TextBackend if you dont have matplotlib.
The plotting module has the following functions:
plot: Plots 2D line plots.
plot parametric: Plots 2D parametric plots.
plot implicit: Plots 2D implicit and region plots.
plot3d: Plots 3D plots of functions in two variables.
plot3d parametric line: Plots 3D line plots, dened by a parameter.
plot3d parametric surface: Plots 3D parametric surface plots.
The above functions are only for convenience and ease of use. It is possible to plot any plot
by passing the corresponding Series class to Plot as argument.
5.18.2 Plot Class
class sympy.plotting.plot.Plot(*args, **kwargs)
The central class of the plotting module.
For interactive work the function plot is better suited.
This class permits the plotting of sympy expressions using numerous backends (matplotlib, textplot, the old pyglet module for sympy, Google charts api, etc).
The gure can contain an arbitrary number of plots of sympy expressions, lists of coordinates of points, etc. Plot has a private attribute series that contains all data series
to be plotted (expressions for lines or surfaces, lists of points, etc (all subclasses of
BaseSeries)). Those data series are instances of classes not imported by from sympy
import *.
The customization of the gure is on two levels. Global options that concern the gure as
a whole (eg title, xlabel, scale, etc) and per-data series options (eg name) and aesthetics
(eg. color, point shape, line type, etc.).
The dierence between options and aesthetics is that an aesthetic can be a function
of the coordinates (or parameters in a parametric plot). The supported values for an
aesthetic are: - None (the backend uses default values) - a constant - a function of one
variable (the rst coordinate or parameter) - a function of two variables (the rst and
second coordinate or parameters) - a function of three variables (only in nonparametric
3D plots) Their implementation depends on the backend so they may not work in some
backends.
If the plot is parametric and the arity of the aesthetic function permits it the aesthetic
is calculated over parameters and not over coordinates. If the arity does not permit
calculation over parameters the calculation is done over coordinates.
Only cartesian coordinates are supported for the moment, but you can use the parametric
plots to plot in polar, spherical and cylindrical coordinates.
The arguments for the constructor Plot must be subclasses of BaseSeries.
Any global option can be specied as a keyword argument.
The global options for a gure are:
title : str
xlabel : str
ylabel : str
5.18. Plotting Module
1271
SymPy Documentation, Release 0.7.6
legend : bool
xscale : {linear, log}
yscale : {linear, log}
axis : bool
axis center : tuple of two oats or {center, auto}
xlim : tuple of two oats
ylim : tuple of two oats
aspect ratio : tuple of two oats or {auto}
autoscale : bool
margin : oat in [0, 1]
The per data series options and aesthetics are: There are none in the base series. See
below for options for subclasses.
Some data series support additional aesthetics or options:
ListSeries, LineOver1DRangeSeries, Parametric2DLineSeries, Parametric3DLineSeries
support the following:
Aesthetics:
line color : function which returns a oat.
options:
label : str
steps : bool
integers only : bool
SurfaceOver2DRangeSeries, ParametricSurfaceSeries support the following:
aesthetics:
surface color : function which returns a oat.
append(arg)
Adds an element from a plots series to an existing plot.
See Also:
extend (page 1272)
Examples
Consider two Plot objects, p1 and p2. To add the second plots rst series object to
the rst, use the append method, like so:
>>> from sympy import symbols
>>> from sympy.plotting import plot
>>> x = symbols(x)
>>> p1 = plot(x*x)
>>> p2 = plot(x)
>>> p1.append(p2[0])
>>> p1
Plot object containing:
[0]: cartesian line: x**2 for x over (-10.0, 10.0)
[1]: cartesian line: x for x over (-10.0, 10.0)
1272
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
extend(arg)
Adds all series from another plot.
Examples
Consider two Plot objects, p1 and p2. To add the second plot to the rst, use the
extend method, like so:
>>> from sympy import symbols
>>> from sympy.plotting import plot
>>> x = symbols(x)
>>> p1 = plot(x*x)
>>> p2 = plot(x)
>>> p1.extend(p2)
>>> p1
Plot object containing:
[0]: cartesian line: x**2 for x over (-10.0, 10.0)
[1]: cartesian line: x for x over (-10.0, 10.0)
5.18.3 Plotting Function Reference
sympy.plotting.plot.plot(*args, **kwargs)
Plots a function of a single variable and returns an instance of the Plot class (also, see
the description of the show keyword argument below).
The plotting uses an adaptive algorithm which samples recursively to accurately plot the
plot. The adaptive algorithm uses a random point near the midpoint of two points that
has to be further sampled. Hence the same plots can appear slightly dierent.
See Also:
Plot (page 1271), LineOver1DRangeSeries.
Examples
>>> from sympy import symbols
>>> from sympy.plotting import plot
>>> x = symbols(x)
Single Plot
>>> plot(x**2, (x, -5, 5))
Plot object containing:
[0]: cartesian line: x**2 for x over (-5.0, 5.0)
Multiple plots with single range.
>>> plot(x, x**2, x**3, (x, -5, 5))
Plot object containing:
[0]: cartesian line: x for x over (-5.0, 5.0)
[1]: cartesian line: x**2 for x over (-5.0, 5.0)
[2]: cartesian line: x**3 for x over (-5.0, 5.0)
Multiple plots with dierent ranges.
5.18. Plotting Module
1273
SymPy Documentation, Release 0.7.6
>>> plot((x**2, (x, -6, 6)), (x, (x, -5, 5)))
Plot object containing:
[0]: cartesian line: x**2 for x over (-6.0, 6.0)
[1]: cartesian line: x for x over (-5.0, 5.0)
No adaptive sampling.
>>> plot(x**2, adaptive=False, nb_of_points=400)
Plot object containing:
[0]: cartesian line: x**2 for x over (-10.0, 10.0)
Usage
Single Plot
plot(expr, range, **kwargs)
If the range is not specied, then a default range of (-10, 10) is used.
Multiple plots with same range.
plot(expr1, expr2, ..., range, **kwargs)
If the range is not specied, then a default range of (-10, 10) is used.
Multiple plots with dierent ranges.
plot((expr1, range), (expr2, range), ..., **kwargs)
Range has to be specied for every expression.
Default range may change in the future if a more advanced default range detection algorithm is implemented.
Arguments
expr : Expression representing the function of single variable
range: (x, 0, 5), A 3-tuple denoting the range of the free variable.
Keyword Arguments
Arguments for plot function:
show: Boolean. The default value is set to True. Set show to False and the function will
not display the plot. The returned instance of the Plot class can then be used to save or
display the plot by calling the save() and show() methods respectively.
Arguments for LineOver1DRangeSeries class:
adaptive: Boolean. The default value is set to True. Set adaptive to False and specify
nb of points if uniform sampling is required.
depth: int Recursion depth of the adaptive algorithm. A depth of value n samples a
maximum of 2n points.
nb of points: int. Used when the adaptive is set to False. The function is uniformly
sampled at nb of points number of points.
Aesthetics options:
1274
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
line color: oat. Species the color for the plot. See Plot to see how to set color for
the plots.
If there are multiple plots, then the same series series are applied to all the plots. If you
want to set these options separately, you can index the Plot object returned and set it.
Arguments for Plot class:
title : str. Title of the plot. It is set to the latex representation of the expression, if the
plot has only one expression.
xlabel : str. Label for the x-axis.
ylabel : str. Label for the y-axis.
xscale: {linear, log} Sets the scaling of the x-axis.
yscale: {linear, log} Sets the scaling if the y-axis.
axis center: tuple of two oats denoting the coordinates of the center or {center,
auto}
xlim : tuple of two oats, denoting the x-axis limits.
ylim : tuple of two oats, denoting the y-axis limits.
sympy.plotting.plot.plot parametric(*args, **kwargs)
Plots a 2D parametric plot.
The plotting uses an adaptive algorithm which samples recursively to accurately plot the
plot. The adaptive algorithm uses a random point near the midpoint of two points that
has to be further sampled. Hence the same plots can appear slightly dierent.
See Also:
Plot (page 1271), Parametric2DLineSeries (page 1282)
Examples
>>> from sympy import symbols, cos, sin
>>> from sympy.plotting import plot_parametric
>>> u = symbols(u)
Single Parametric plot
>>> plot_parametric(cos(u), sin(u), (u, -5, 5))
Plot object containing:
[0]: parametric cartesian line: (cos(u), sin(u)) for u over (-5.0, 5.0)
Multiple parametric plot with single range.
>>> plot_parametric((cos(u), sin(u)), (u, cos(u)))
Plot object containing:
[0]: parametric cartesian line: (cos(u), sin(u)) for u over (-10.0, 10.0)
[1]: parametric cartesian line: (u, cos(u)) for u over (-10.0, 10.0)
Multiple parametric plots.
>>> plot_parametric((cos(u), sin(u), (u, -5, 5)),
...
(cos(u), u, (u, -5, 5)))
Plot object containing:
[0]: parametric cartesian line: (cos(u), sin(u)) for u over (-5.0, 5.0)
[1]: parametric cartesian line: (cos(u), u) for u over (-5.0, 5.0)
5.18. Plotting Module
1275
SymPy Documentation, Release 0.7.6
Usage
Single plot.
plot parametric(expr x, expr y, range, **kwargs)
If the range is not specied, then a default range of (-10, 10) is used.
Multiple plots with same range.
plot parametric((expr1 x, expr1 y), (expr2 x, expr2 y), range, **kwargs)
If the range is not specied, then a default range of (-10, 10) is used.
Multiple plots with dierent ranges.
plot parametric((expr x, expr y, range), ..., **kwargs)
Range has to be specied for every expression.
Default range may change in the future if a more advanced default range detection algorithm is implemented.
Arguments
expr x : Expression representing the function along x.
expr y : Expression representing the function along y.
range: (u, 0, 5), A 3-tuple denoting the range of the parameter variable.
Keyword Arguments
Arguments for Parametric2DLineSeries class:
adaptive: Boolean. The default value is set to True. Set adaptive to False and specify
nb of points if uniform sampling is required.
depth: int Recursion depth of the adaptive algorithm. A depth of value n samples a
maximum of 2n points.
nb of points: int. Used when the adaptive is set to False. The function is uniformly
sampled at nb of points number of points.
Aesthetics
line color: function which returns a oat.
sympy.plotting.Plot for more details.
Species the color for the plot.
See
If there are multiple plots, then the same Series arguments are applied to all the plots.
If you want to set these options separately, you can index the returned Plot object and
set it.
Arguments for Plot class:
xlabel : str. Label for the x-axis.
ylabel : str. Label for the y-axis.
xscale: {linear, log} Sets the scaling of the x-axis.
yscale: {linear, log} Sets the scaling if the y-axis.
1276
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
axis center: tuple of two oats denoting the coordinates of the center or {center,
auto}
xlim : tuple of two oats, denoting the x-axis limits.
ylim : tuple of two oats, denoting the y-axis limits.
sympy.plotting.plot.plot3d(*args, **kwargs)
Plots a 3D surface plot.
See Also:
Plot (page 1271), SurfaceOver2DRangeSeries (page 1283)
Examples
>>> from sympy import symbols
>>> from sympy.plotting import plot3d
>>> x, y = symbols(x y)
Single plot
>>> plot3d(x*y, (x, -5, 5), (y, -5, 5))
Plot object containing:
[0]: cartesian surface: x*y for x over (-5.0, 5.0) and y over (-5.0, 5.0)
Multiple plots with same range
>>> plot3d(x*y, -x*y, (x, -5, 5), (y, -5, 5))
Plot object containing:
[0]: cartesian surface: x*y for x over (-5.0, 5.0) and y over (-5.0, 5.0)
[1]: cartesian surface: -x*y for x over (-5.0, 5.0) and y over (-5.0, 5.0)
Multiple plots with dierent ranges.
>>> plot3d((x**2 + y**2, (x, -5, 5), (y, -5, 5)),
...
(x*y, (x, -3, 3), (y, -3, 3)))
Plot object containing:
[0]: cartesian surface: x**2 + y**2 for x over (-5.0, 5.0) and y over (-5.0, 5.0)
[1]: cartesian surface: x*y for x over (-3.0, 3.0) and y over (-3.0, 3.0)
Usage
Single plot
plot3d(expr, range x, range y, **kwargs)
If the ranges are not specied, then a default range of (-10, 10) is used.
Multiple plot with the same range.
plot3d(expr1, expr2, range x, range y, **kwargs)
If the ranges are not specied, then a default range of (-10, 10) is used.
Multiple plots with dierent ranges.
plot3d((expr1, range x, range y), (expr2, range x, range y), ..., **kwargs)
Ranges have to be specied for every expression.
5.18. Plotting Module
1277
SymPy Documentation, Release 0.7.6
Default range may change in the future if a more advanced default range detection algorithm is implemented.
Arguments
expr : Expression representing the function along x.
range x: (x, 0, 5), A 3-tuple denoting the range of the x variable.
range y: (y, 0, 5), A 3-tuple denoting the range of the y variable.
Keyword Arguments
Arguments for SurfaceOver2DRangeSeries class:
nb of points x: int. The x range is sampled uniformly at nb of points x of points.
nb of points y: int. The y range is sampled uniformly at nb of points y of points.
Aesthetics:
surface color: Function which returns a oat. Species the color for the surface of the
plot. See sympy.plotting.Plot for more details.
If there are multiple plots, then the same series arguments are applied to all the plots.
If you want to set these options separately, you can index the returned Plot object and
set it.
Arguments for Plot class:
title : str. Title of the plot.
sympy.plotting.plot.plot3d parametric line(*args, **kwargs)
Plots a 3D parametric line plot.
See Also:
Plot (page 1271), Parametric3DLineSeries (page 1283)
Examples
>>> from sympy import symbols, cos, sin
>>> from sympy.plotting import plot3d_parametric_line
>>> u = symbols(u)
Single plot.
>>> plot3d_parametric_line(cos(u), sin(u), u, (u, -5, 5))
Plot object containing:
[0]: 3D parametric cartesian line: (cos(u), sin(u), u) for u over (-5.0, 5.0)
Multiple plots.
>>> plot3d_parametric_line((cos(u), sin(u), u, (u, -5, 5)),
...
(sin(u), u**2, u, (u, -5, 5)))
Plot object containing:
[0]: 3D parametric cartesian line: (cos(u), sin(u), u) for u over (-5.0, 5.0)
[1]: 3D parametric cartesian line: (sin(u), u**2, u) for u over (-5.0, 5.0)
1278
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Usage
Single plot:
plot3d parametric line(expr x, expr y, expr z, range, **kwargs)
If the range is not specied, then a default range of (-10, 10) is used.
Multiple plots.
plot3d parametric line((expr x, expr y, expr z, range), ..., **kwargs)
Ranges have to be specied for every expression.
Default range may change in the future if a more advanced default range detection algorithm is implemented.
Arguments
expr x : Expression representing the function along x.
expr y : Expression representing the function along y.
expr z : Expression representing the function along z.
range: (u, 0, 5), A 3-tuple denoting the range of the parameter variable.
Keyword Arguments
Arguments for Parametric3DLineSeries class.
nb of points: The range is uniformly sampled at nb of points number of points.
Aesthetics:
line color: function which returns a oat.
sympy.plotting.Plot for more details.
Species the color for the plot.
See
If there are multiple plots, then the same series arguments are applied to all the plots.
If you want to set these options separately, you can index the returned Plot object and
set it.
Arguments for Plot class.
title : str. Title of the plot.
sympy.plotting.plot.plot3d parametric surface(*args, **kwargs)
Plots a 3D parametric surface plot.
See Also:
Plot (page 1271), ParametricSurfaceSeries (page 1283)
Examples
>>> from sympy import symbols, cos, sin
>>> from sympy.plotting import plot3d_parametric_surface
>>> u, v = symbols(u v)
Single plot.
5.18. Plotting Module
1279
SymPy Documentation, Release 0.7.6
>>> plot3d_parametric_surface(cos(u + v), sin(u - v), u - v,
...
(u, -5, 5), (v, -5, 5))
Plot object containing:
[0]: parametric cartesian surface: (cos(u + v), sin(u - v), u - v) for u over (-5.0, 5.0) and v
Usage
Single plot.
plot3d parametric surface(expr x, expr y, expr z, range u, range v, **kwargs)
If the ranges is not specied, then a default range of (-10, 10) is used.
Multiple plots.
plot3d parametric surface((expr x, expr y, expr z, range u, range v), ...,
**kwargs)
Ranges have to be specied for every expression.
Default range may change in the future if a more advanced default range detection algorithm is implemented.
Arguments
expr x: Expression representing the function along x.
expr y: Expression representing the function along y.
expr z: Expression representing the function along z.
range u: (u, 0, 5), A 3-tuple denoting the range of the u variable.
range v: (v, 0, 5), A 3-tuple denoting the range of the v variable.
Keyword Arguments
Arguments for ParametricSurfaceSeries class:
nb of points u: int. The u range is sampled uniformly at nb of points v of points
nb of points y: int. The v range is sampled uniformly at nb of points y of points
Aesthetics:
surface color: Function which returns a oat. Species the color for the surface of the
plot. See sympy.plotting.Plot for more details.
If there are multiple plots, then the same series arguments are applied for all the plots.
If you want to set these options separately, you can index the returned Plot object and
set it.
Arguments for Plot class:
title : str. Title of the plot.
sympy.plotting.plot implicit.plot implicit(expr,
x var=None,
**kwargs)
A plot function to plot implicit equations / inequalities.
1280
y var=None,
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Arguments
expr : The equation / inequality that is to be plotted.
x var (optional) : symbol to plot on x-axis or tuple giving symbol and range as (symbol, xmin, xmax)
y var (optional) : symbol to plot on y-axis or tuple giving symbol and range as (symbol, ymin, ymax)
If neither x var nor y var are given then the free symbols in the expression will be assigned in the order they are sorted.
The following keyword arguments can also be used:
adaptive. Boolean. The default value is set to True. It has to be set to False
if you want to use a mesh grid.
depth integer. The depth of recursion for adaptive mesh grid. Default value
is 0. Takes value in the range (0, 4).
points integer. The number of points if adaptive mesh grid is not used. Default value is 200.
title string .The title for the plot.
xlabel string. The label for the x-axis
ylabel string. The label for the y-axis
plot implicit, by default, uses interval arithmetic to plot functions. If the expression cannot be plotted using interval arithmetic, it defaults to a generating a contour using a mesh
grid of xed number of points. By setting adaptive to False, you can force plot implicit to
use the mesh grid. The mesh grid method can be eective when adaptive plotting using
interval arithmetic, fails to plot with small line width.
Examples:
Plot expressions:
>>> from sympy import plot_implicit, cos, sin, symbols, Eq, And
>>> x, y = symbols(x y)
Without any ranges for the symbols in the expression
>>> p1 = plot_implicit(Eq(x**2 + y**2, 5))
With the range for the symbols
>>> p2 = plot_implicit(Eq(x**2 + y**2, 3),
...
(x, -3, 3), (y, -3, 3))
With depth of recursion as argument.
>>> p3 = plot_implicit(Eq(x**2 + y**2, 5),
...
(x, -4, 4), (y, -4, 4), depth = 2)
Using mesh grid and not using adaptive meshing.
>>> p4 = plot_implicit(Eq(x**2 + y**2, 5),
...
(x, -5, 5), (y, -2, 2), adaptive=False)
5.18. Plotting Module
1281
SymPy Documentation, Release 0.7.6
Using mesh grid with number of points as input.
>>> p5 = plot_implicit(Eq(x**2 + y**2, 5),
...
(x, -5, 5), (y, -2, 2),
...
adaptive=False, points=400)
Plotting regions.
>>> p6 = plot_implicit(y > x**2)
Plotting Using boolean conjunctions.
>>> p7 = plot_implicit(And(y > x, y > -x))
When plotting an expression with a single variable (y - 1, for example), specify the x or
the y variable explicitly:
>>> p8 = plot_implicit(y - 1, y_var=y)
>>> p9 = plot_implicit(x - 1, x_var=x)
5.18.4 Series Classes
class sympy.plotting.plot.BaseSeries
Base class for the data objects containing stu to be plotted.
The backend should check if it supports the data series that its given. (eg TextBackend
supports only LineOver1DRange). Its the backend responsibility to know how to use the
class of data series that its given.
Some data series classes are grouped (using a class attribute like is 2Dline) according
to the api they present (based only on convention). The backend is not obliged to use
that api (eg. The LineOver1DRange belongs to the is 2Dline group and presents the
get points method, but the TextBackend does not use the get points method).
class sympy.plotting.plot.Line2DBaseSeries
A base class for 2D lines.
adding the label, steps and only integers options
making is 2Dline true
dening get segments and get color array
class sympy.plotting.plot.LineOver1DRangeSeries(expr, var start end, **kwargs)
Representation for a line consisting of a SymPy expression over a range.
get segments()
Adaptively gets segments for plotting.
The adaptive sampling is done by recursively checking if three points are almost
collinear. If they are not collinear, then more points are added between those points.
References
[1] Adaptive polygonal approximation of parametric curves, Luiz
de Figueiredo.
1282
Henrique
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
class sympy.plotting.plot.Parametric2DLineSeries(expr x, expr y, var start end,
**kwargs)
Representation for a line consisting of two parametric sympy expressions over a range.
get segments()
Adaptively gets segments for plotting.
The adaptive sampling is done by recursively checking if three points are almost
collinear. If they are not collinear, then more points are added between those points.
References
[1] Adaptive polygonal approximation of parametric curves, Luiz
de Figueiredo.
Henrique
class sympy.plotting.plot.Line3DBaseSeries
A base class for 3D lines.
Most of the stu is derived from Line2DBaseSeries.
class sympy.plotting.plot.Parametric3DLineSeries(expr x,
expr y,
expr z,
var start end, **kwargs)
Representation for a 3D line consisting of two parametric sympy expressions and a range.
class sympy.plotting.plot.SurfaceBaseSeries
A base class for 3D surfaces.
class sympy.plotting.plot.SurfaceOver2DRangeSeries(expr,
var start end x,
var start end y, **kwargs)
Representation for a 3D surface consisting of a sympy expression and 2D range.
class sympy.plotting.plot.ParametricSurfaceSeries(expr x,
expr y,
expr z,
var start end u,
var start end v, **kwargs)
Representation for a 3D surface consisting of three parametric sympy expressions and
a range.
class sympy.plotting.plot implicit.ImplicitSeries(expr,
var start end x,
var start end y, has equality,
use interval math,
depth,
nb of points)
Representation for Implicit plot
5.19 Pyglet Plotting Module
This is the documentation for the old plotting module that uses pyglet. This module has some
limitations and is not actively developed anymore. For an alternative you can look at the new
plotting module.
The pyglet plotting module can do nice 2D and 3D plots that can be controlled by console
commands as well as keyboard and mouse, with the only dependency being pyglet.
Here is the simplest usage:
>>> from sympy import var, Plot
>>> var(x y z)
>>> Plot(x*y**3-y*x**3)
5.19. Pyglet Plotting Module
1283
SymPy Documentation, Release 0.7.6
To see lots of plotting examples, see examples/pyglet plotting.py and try running it in
interactive mode (python -i plotting.py):
$ python -i examples/pyglet plotting.py
And type for instance example(7) or example(11).
See also the Plotting Module wiki page for screenshots.
5.19.1 Plot Window Controls
Camera
Sensitivity Modier
Zoom
Rotate View X,Y axis
Rotate View Z axis
Rotate Ordinate Z axis
View XY
View XZ
View YZ
View Perspective
Reset
Axes
Toggle Visible
Toggle Colors
Window
Close
Screenshot
Keys
SHIFT
R and F, Page Up and Down, Numpad + and Arrow Keys, A,S,D,W, Numpad 4,6,8,2
Q and E, Numpad 7 and 9
Z and C, Numpad 1 and 3
F1
F2
F3
F4
X, Numpad 5
Keys
F5
F6
Keys
ESCAPE
F8
The mouse can be used to rotate, zoom, and translate by dragging the left, middle, and right
mouse buttons respectively.
5.19.2 Coordinate Modes
Plot supports several curvilinear coordinate modes, and they are independent for each plotted function. You can specify a coordinate mode explicitly with the mode named argument,
but it can be automatically determined for cartesian or parametric plots, and therefore must
only be specied for polar, cylindrical, and spherical modes.
Specically, Plot(function arguments) and Plot. setitem (i, function arguments)
(accessed using array-index syntax on the Plot instance) will interpret your arguments as a
cartesian plot if you provide one function and a parametric plot if you provide two or three
functions. Similarly, the arguments will be interpreted as a curve is one variable is used, and
a surface if two are used.
Supported mode names by number of variables:
1 (curves): parametric, cartesian, polar
2 (surfaces): parametric, cartesian, cylindrical, spherical
>>> Plot(1, mode=spherical; color=zfade4)
1284
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Note that function parameters are given as option strings of the form key1=value1; key2 =
value2 (spaces are truncated). Keyword arguments given directly to plot apply to the plot
itself.
5.19.3 Specifying Intervals for Variables
The basic format for variable intervals is [var, min, max, steps]. However, the syntax is quite
exible, and arguments not specied are taken from the defaults for the current coordinate
mode:
>>>
>>>
>>>
>>>
>>>
>>>
>>>
Plot(x**2) # implies [x,-5,5,100]
Plot(x**2, [], []) # [x,-1,1,40], [y,-1,1,40]
Plot(x**2-y**2, [100], [100]) # [x,-1,1,100], [y,-1,1,100]
Plot(x**2, [x,-13,13,100])
Plot(x**2, [-13,13]) # [x,-13,13,100]
Plot(x**2, [x,-13,13]) # [x,-13,13,100]
Plot(1*x, [], [x], mode=cylindrical) # [unbound_theta,0,2*Pi,40], [x,-1,1,20]
5.19.4 Using the Interactive Interface
>>> p = Plot(visible=False)
>>> f = x**2
>>> p[1] = f
>>> p[2] = f.diff(x)
>>> p[3] = f.diff(x).diff(x)
>>> p
[1]: x**2, mode=cartesian
[2]: 2*x, mode=cartesian
[3]: 2, mode=cartesian
>>> p.show()
>>> p.clear()
>>> p
<blank plot>
>>> p[1] = x**2+y**2
>>> p[1].style = solid
>>> p[2] = -x**2-y**2
>>> p[2].style = wireframe
>>> p[1].color = z, (0.4,0.4,0.9), (0.9,0.4,0.4)
>>> p[1].style = both
>>> p[2].style = both
>>> p.close()
5.19.5 Using Custom Color Functions
The following code plots a saddle and color it by the magnitude of its gradient:
>>>
>>>
>>>
>>>
fz = x**2-y**2
Fx, Fy, Fz = fz.diff(x), fz.diff(y), 0
p[1] = fz, style=solid
p[1].color = (Fx**2 + Fy**2 + Fz**2)**(0.5)
The coloring algorithm works like this:
1. Evaluate the color function(s) across the curve or surface.
5.19. Pyglet Plotting Module
1285
SymPy Documentation, Release 0.7.6
2. Find the minimum and maximum value of each component.
3. Scale each component to the color gradient.
When not specied explicitly, the default color gradient is f(0.0)=(0.4,0.4,0.4) ->
f(1.0)=(0.9,0.9,0.9). In our case, everything is gray-scale because we have applied the default color gradient uniformly for each color component. When dening a color scheme in
this way, you might want to supply a color gradient as well:
>>> p[1].color = (Fx**2 + Fy**2 + Fz**2)**(0.5), (0.1,0.1,0.9), (0.9,0.1,0.1)
Heres a color gradient with four steps:
>>> gradient = [ 0.0, (0.1,0.1,0.9), 0.3, (0.1,0.9,0.1),
...
0.7, (0.9,0.9,0.1), 1.0, (1.0,0.0,0.0) ]
>>> p[1].color = (Fx**2 + Fy**2 + Fz**2)**(0.5), gradient
The other way to specify a color scheme is to give a separate function for each component r,
g, b. With this syntax, the default color scheme is dened:
>>> p[1].color = z,y,x, (0.4,0.4,0.4), (0.9,0.9,0.9)
This maps z->red, y->green, and x->blue. In some cases, you might prefer to use the following
alternative syntax:
>>> p[1].color = z,(0.4,0.9), y,(0.4,0.9), x,(0.4,0.9)
You can still use multi-step gradients with three-function color schemes.
5.19.6 Plotting Geometric Entities
The plotting module is capable of plotting some 2D geometric entities like line, circle and
ellipse. The following example plots a circle and a tangent line at a random point on the
ellipse.
In [1]: p = Plot(axes=label axes=True)
In [2]: c = Circle(Point(0,0), 1)
In [3]: t = c.tangent line(c.random point())
In [4]: p[0] = c
In [5]: p[1] = t
Plotting polygons (Polygon, RegularPolygon, Triangle) are not supported directly. However a
polygon can be plotted through a loop as follows.
In [6]: p = Plot(axes=label axes=True)
In [7]: t = RegularPolygon(Point(0,0), 1, 5)
In [8]: for i in range(len(t.sides)):
....:
p[i] = t.sides[i]
1286
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
5.20 Assumptions module
5.20.1 Contents
Ask
Module for querying SymPy objects about assumptions.
class sympy.assumptions.ask.Q
Supported ask keys.
sympy.assumptions.ask.ask(proposition,
assumptions=True,
text=AssumptionsContext([]))
Method for inferring properties about objects.
con-
Syntax
ask(proposition)
ask(proposition, assumptions)
where proposition is any boolean expression
Examples
>>> from sympy import ask, Q, pi
>>> from sympy.abc import x, y
>>> ask(Q.rational(pi))
False
>>> ask(Q.even(x*y), Q.even(x) & Q.integer(y))
True
>>> ask(Q.prime(x*y), Q.integer(x) & Q.integer(y))
False
Remarks Relations in assumptions are not implemented (yet), so the following will not
give a meaningful result.
>>> ask(Q.positive(x), Q.is_true(x > 0))
It is however a work in progress.
sympy.assumptions.ask.ask full inference(proposition,
known facts cnf)
Method for inferring properties about objects.
assumptions,
sympy.assumptions.ask.compute known facts(known facts, known facts keys)
Compute the various forms of knowledge compilation used by the assumptions system.
This function is typically applied to the variables known facts and known facts keys
dened at the bottom of this le.
sympy.assumptions.ask.register handler(key, handler)
Register a handler in the ask system. key must be a string and handler a class inheriting
from AskHandler:
>>> from sympy.assumptions import register_handler, ask, Q
>>> from sympy.assumptions.handlers import AskHandler
>>> class MersenneHandler(AskHandler):
5.20. Assumptions module
1287
SymPy Documentation, Release 0.7.6
...
# Mersenne numbers are in the form 2**n + 1, n integer
...
@staticmethod
...
def Integer(expr, assumptions):
...
import math
...
return ask(Q.integer(math.log(expr + 1, 2)))
>>> register_handler(mersenne, MersenneHandler)
>>> ask(Q.mersenne(7))
True
sympy.assumptions.ask.remove handler(key, handler)
Removes a handler from the ask system. Same syntax as register handler
Assume
class sympy.assumptions.assume.AppliedPredicate
The class of expressions resulting from applying a Predicate.
Examples
>>> from sympy import Q, Symbol
>>> x = Symbol(x)
>>> Q.integer(x)
Q.integer(x)
>>> type(Q.integer(x))
<class sympy.assumptions.assume.AppliedPredicate>
arg
Return the expression used by this assumption.
Examples
>>>
>>>
>>>
>>>
x +
from sympy import Q, Symbol
x = Symbol(x)
a = Q.integer(x + 1)
a.arg
1
class sympy.assumptions.assume.AssumptionsContext
Set representing assumptions.
This is used to represent global assumptions, but you can also use this class to create
your own local assumptions contexts. It is basically a thin wrapper to Pythons set, so
see its documentation for advanced usage.
Examples
>>> from sympy import AppliedPredicate, Q
>>> from sympy.assumptions.assume import global_assumptions
>>> global_assumptions
AssumptionsContext()
>>> from sympy.abc import x
>>> global_assumptions.add(Q.real(x))
1288
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> global_assumptions
AssumptionsContext([Q.real(x)])
>>> global_assumptions.remove(Q.real(x))
>>> global_assumptions
AssumptionsContext()
>>> global_assumptions.clear()
add(*assumptions)
Add an assumption.
class sympy.assumptions.assume.Predicate
A predicate is a function that returns a boolean value.
Predicates merely wrap their argument and remain unevaluated:
>>> from sympy import Q, ask, Symbol, S
>>> x = Symbol(x)
>>> Q.prime(7)
Q.prime(7)
To obtain the truth value of an expression containing predicates, use the function ask:
>>> ask(Q.prime(7))
True
The tautological predicate Q.ist rue can be used to wrap other objects:
>>> Q.is_true(x > 1)
Q.is_true(x > 1)
>>> Q.is_true(S(1) < x)
Q.is_true(1 < x)
eval(expr, assumptions=True)
Evaluate self(expr) under the given assumptions.
This uses only direct resolution methods, not logical inference.
sympy.assumptions.assume.assuming(*args, **kwds)
Context manager for assumptions
Examples
>>> from sympy.assumptions import assuming, Q, ask
>>> from sympy.abc import x, y
>>> print(ask(Q.integer(x + y)))
None
>>> with assuming(Q.integer(x), Q.integer(y)):
...
print(ask(Q.integer(x + y)))
True
Rene
sympy.assumptions.refine.refine(expr, assumptions=True)
Simplify an expression using assumptions.
5.20. Assumptions module
1289
SymPy Documentation, Release 0.7.6
Gives the form of expr that would be obtained if symbols in it were replaced by explicit
numerical expressions satisfying the assumptions.
Examples
>>> from sympy import refine, sqrt, Q
>>> from sympy.abc import x
>>> refine(sqrt(x**2), Q.real(x))
Abs(x)
>>> refine(sqrt(x**2), Q.positive(x))
x
sympy.assumptions.refine.refine Pow(expr, assumptions)
Handler for instances of Pow.
>>>
>>>
>>>
>>>
>>>
1
>>>
-1
from sympy import Symbol, Q
from sympy.assumptions.refine import refine_Pow
from sympy.abc import x,y,z
refine_Pow((-1)**x, Q.real(x))
refine_Pow((-1)**x, Q.even(x))
refine_Pow((-1)**x, Q.odd(x))
For powers of -1, even parts of the exponent can be simplied:
>>> refine_Pow((-1)**(x+y), Q.even(x))
(-1)**y
>>> refine_Pow((-1)**(x+y+z), Q.odd(x) & Q.odd(z))
(-1)**y
>>> refine_Pow((-1)**(x+y+2), Q.odd(x))
(-1)**(y + 1)
>>> refine_Pow((-1)**(x+3), True)
(-1)**(x + 1)
sympy.assumptions.refine.refine Relational(expr, assumptions)
Handler for Relational
>>> from sympy.assumptions.refine import refine_Relational
>>> from sympy.assumptions.ask import Q
>>> from sympy.abc import x
>>> refine_Relational(x<0, ~Q.is_true(x<0))
False
sympy.assumptions.refine.refine abs(expr, assumptions)
Handler for the absolute value.
Examples
>>>
>>>
>>>
>>>
>>>
x
1290
from sympy import Symbol, Q, refine, Abs
from sympy.assumptions.refine import refine_abs
from sympy.abc import x
refine_abs(Abs(x), Q.real(x))
refine_abs(Abs(x), Q.positive(x))
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> refine_abs(Abs(x), Q.negative(x))
-x
sympy.assumptions.refine.refine exp(expr, assumptions)
Handler for exponential function.
>>>
>>>
>>>
>>>
>>>
1
from sympy import Symbol, Q, exp, I, pi
from sympy.assumptions.refine import refine_exp
from sympy.abc import x
refine_exp(exp(pi*I*2*x), Q.real(x))
refine_exp(exp(pi*I*2*x), Q.integer(x))
Handlers
Contents
Calculus This module contains query handlers responsible for calculus queries: innitesimal, bounded, etc.
class sympy.assumptions.handlers.calculus.AskBoundedHandler
Handler for key bounded.
Test that an expression is bounded respect to all its variables.
Examples of usage:
>>> from sympy import Symbol, Q
>>> from sympy.assumptions.handlers.calculus import AskBoundedHandler
>>> from sympy.abc import x
>>> a = AskBoundedHandler()
>>> a.Symbol(x, Q.positive(x)) == None
True
>>> a.Symbol(x, Q.bounded(x))
True
static Add(expr, assumptions)
Return True if expr is bounded, False if not and None if unknown.
Truth Table:
B
B
B
+
-
U
x
?
?
?
U
+
U
U
U
?
?
?
?
?
?
+
?
U
U
?
?
B = Bounded
U = Unbounded
? = unknown boundedness
+ = positive sign
- = negative sign
x = sign unknown
5.20. Assumptions module
1291
SymPy Documentation, Release 0.7.6
All Bounded -> True
1 Unbounded and the rest Bounded -> False
>1 Unbounded, all with same known sign -> False
Any Unknown and unknown sign -> None
Else -> None
When the signs are not the same you can have an undened result as in oo oo, hence bounded is also undened.
static Mul(expr, assumptions)
Return True if expr is bounded, False if not and None if unknown.
Truth Table:
B U ?
s
B B U ?
U
U U
?
?
/s
?
B = Bounded
U = Unbounded
? = unknown boundedness
s = signed (hence nonzero)
/s = not signed
static Pow(expr, assumptions)
Unbounded ** NonZero -> Unbounded Bounded ** Bounded -> Bounded Abs()<=1
** Positive -> Bounded Abs()>=1 ** Negative -> Bounded Otherwise unknown
static Symbol(expr, assumptions)
Handles Symbol.
Examples:
>>> from sympy import Symbol, Q
>>> from sympy.assumptions.handlers.calculus import AskBoundedHandler
>>> from sympy.abc import x
>>> a = AskBoundedHandler()
>>> a.Symbol(x, Q.positive(x)) == None
True
>>> a.Symbol(x, Q.bounded(x))
True
class sympy.assumptions.handlers.calculus.AskInfinitesimalHandler
Handler for key innitesimal Test that a given expression is equivalent to an innitesimal number
static Add(expr, assumptions)
Innitesimal*Bounded -> Innitesimal
static Mul(expr, assumptions)
Innitesimal*Bounded -> Innitesimal
1292
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
static Pow(expr, assumptions)
Innitesimal*Bounded -> Innitesimal
nTheory Handlers for keys related to number theory: prime, even, odd, etc.
class sympy.assumptions.handlers.ntheory.AskOddHandler
Handler for key odd Test that an expression represents an odd number
class sympy.assumptions.handlers.ntheory.AskPrimeHandler
Handler for key prime Test that an expression represents a prime number. When the
expression is a number the result, when True, is subject to the limitations of isprime()
which is used to return the result.
static Pow(expr, assumptions)
Integer**Integer -> !Prime
Order AskHandlers related to order relations: positive, negative, etc.
class sympy.assumptions.handlers.order.AskNegativeHandler
This is called by ask() when key=negative
Test that an expression is less (strict) than zero.
Examples:
>>> from sympy import ask, Q, pi
>>> ask(Q.negative(pi+1)) # this calls AskNegativeHandler.Add
False
>>> ask(Q.negative(pi**2)) # this calls AskNegativeHandler.Pow
False
static Add(expr, assumptions)
Positive + Positive -> Positive, Negative + Negative -> Negative
static Pow(expr, assumptions)
Real ** Even -> NonNegative Real ** Odd -> same as base NonNegative ** Positive
-> NonNegative
class sympy.assumptions.handlers.order.AskNonZeroHandler
Handler for key zero Test that an expression is not identically zero
class sympy.assumptions.handlers.order.AskPositiveHandler
Handler for key positive Test that an expression is greater (strict) than zero
Sets
Handlers for predicates related to set membership: integer, rational, etc.
class sympy.assumptions.handlers.sets.AskAlgebraicHandler
Handler for Q.algebraic key.
class sympy.assumptions.handlers.sets.AskAntiHermitianHandler
Handler for Q.antihermitian Test that an expression belongs to the eld of anti-Hermitian
operators, that is, operators in the form x*I, where x is Hermitian
static Add(expr, assumptions)
Antihermitian + Antihermitian -> Antihermitian Antihermitian + !Antihermitian ->
!Antihermitian
static Mul(expr, assumptions)
As long as there is at most only one noncommutative term: Hermitian*Hermitian
5.20. Assumptions module
1293
SymPy Documentation, Release 0.7.6
-> !Antihermitian Hermitian*Antihermitian
tian*Antihermitian -> !Antihermitian
->
Antihermitian
Antihermi-
static Pow(expr, assumptions)
Hermitian**Integer -> !Antihermitian Antihermitian**Even -> !Antihermitian Antihermitian**Odd -> Antihermitian
class sympy.assumptions.handlers.sets.AskComplexHandler
Handler for Q.complex Test that an expression belongs to the eld of complex numbers
class sympy.assumptions.handlers.sets.AskExtendedRealHandler
Handler for Q.extended real Test that an expression belongs to the eld of extended real
numbers, that is real numbers union {Innity, -Innity}
class sympy.assumptions.handlers.sets.AskHermitianHandler
Handler for Q.hermitian Test that an expression belongs to the eld of Hermitian operators
static Add(expr, assumptions)
Hermitian + Hermitian -> Hermitian Hermitian + !Hermitian -> !Hermitian
static Mul(expr, assumptions)
As long as there is at most only one noncommutative term: Hermitian*Hermitian ->
Hermitian Hermitian*Antihermitian -> !Hermitian Antihermitian*Antihermitian ->
Hermitian
static Pow(expr, assumptions)
Hermitian**Integer -> Hermitian
class sympy.assumptions.handlers.sets.AskImaginaryHandler
Handler for Q.imaginary Test that an expression belongs to the eld of imaginary numbers, that is, numbers in the form x*I, where x is real
static Add(expr, assumptions)
Imaginary + Imaginary -> Imaginary Imaginary + Complex -> ? Imaginary + Real
-> !Imaginary
static Mul(expr, assumptions)
Real*Imaginary -> Imaginary Imaginary*Imaginary -> Real
static Pow(expr, assumptions)
Imaginary**Odd -> Imaginary Imaginary**Even -> Real b**Imaginary -> !Imaginary if exponent is an integer multiple of I*pi/log(b) Imaginary**Real -> ? Positive**Real -> Real Negative**Integer -> Real Negative**(Integer/2) -> Imaginary
Negative**Real -> not Imaginary if exponent is not Rational
class sympy.assumptions.handlers.sets.AskIntegerHandler
Handler for Q.integer Test that an expression belongs to the eld of integer numbers
static Add(expr, assumptions)
Integer + Integer -> Integer Integer + !Integer -> !Integer !Integer + !Integer -> ?
static Mul(expr, assumptions)
Integer*Integer -> Integer Integer*Irrational -> !Integer Odd/Even -> !Integer Integer*Rational -> ?
static Pow(expr, assumptions)
Integer + Integer -> Integer Integer + !Integer -> !Integer !Integer + !Integer -> ?
class sympy.assumptions.handlers.sets.AskRationalHandler
Handler for Q.rational Test that an expression belongs to the eld of rational numbers
1294
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
static Add(expr, assumptions)
Rational + Rational -> Rational Rational + !Rational -> !Rational !Rational + !Rational -> ?
static Mul(expr, assumptions)
Rational + Rational -> Rational Rational + !Rational -> !Rational !Rational + !Rational -> ?
static Pow(expr, assumptions)
Rational ** Integer -> Rational Irrational ** Rational -> Irrational Rational ** Irrational -> ?
class sympy.assumptions.handlers.sets.AskRealHandler
Handler for Q.real Test that an expression belongs to the eld of real numbers
static Add(expr, assumptions)
Real + Real -> Real Real + (Complex & !Real) -> !Real
static Mul(expr, assumptions)
Real*Real -> Real Real*Imaginary -> !Real Imaginary*Imaginary -> Real
static Pow(expr, assumptions)
Real**Integer -> Real Positive**Real -> Real Real**(Integer/Even) -> Real if base is
nonnegative Real**(Integer/Odd) -> Real Imaginary**(Integer/Even) -> Real Imaginary**(Integer/Odd) -> not Real Imaginary**Real -> ? since Real could be 0 (giving
real) or 1 (giving imaginary) b**Imaginary -> Real if log(b) is imaginary and b !=
0 and exponent != integer multiple of I*pi/log(b) Real**Real -> ? e.g. sqrt(-1) is
imaginary and sqrt(2) is not
Queries are used to ask information about expressions. Main method for this is ask():
sympy.assumptions.ask.ask(proposition,
assumptions=True,
text=AssumptionsContext([]))
Method for inferring properties about objects.
con-
Syntax
ask(proposition)
ask(proposition, assumptions)
where proposition is any boolean expression
Examples
>>> from sympy import ask, Q, pi
>>> from sympy.abc import x, y
>>> ask(Q.rational(pi))
False
>>> ask(Q.even(x*y), Q.even(x) & Q.integer(y))
True
>>> ask(Q.prime(x*y), Q.integer(x) & Q.integer(y))
False
Remarks Relations in assumptions are not implemented (yet), so the following will not
give a meaningful result.
>>> ask(Q.positive(x), Q.is_true(x > 0))
It is however a work in progress.
5.20. Assumptions module
1295
SymPy Documentation, Release 0.7.6
5.20.2 Querying
asks optional second argument should be a boolean expression involving assumptions about
objects in expr. Valid values include:
Q.integer(x)
Q.positive(x)
Q.integer(x) & Q.positive(x)
etc.
Q is a class in sympy.assumptions holding known predicates.
See documentation for the logic module for a complete list of valid boolean expressions.
You can also dene a context so you dont have to pass that argument each time to function
ask(). This is done by using the assuming context manager from module sympy.assumptions.
>>> from sympy import *
>>> x = Symbol(x)
>>> y = Symbol(y)
>>> facts = Q.positive(x), Q.positive(y)
>>> with assuming(*facts):
...
print(ask(Q.positive(2*x + y)))
True
5.20.3 Supported predicates
bounded
Test that a function is bounded with respect to its variables. For example, sin(x) is a bounded
functions, but exp(x) is not.
Examples:
>>> from sympy import *
>>> x = Symbol(x)
>>> ask(Q.bounded(exp(x)), ~Q.bounded(x))
False
>>> ask(Q.bounded(exp(x)) , Q.bounded(x))
True
>>> ask(Q.bounded(sin(x)), ~Q.bounded(x))
True
commutative
Test that objects are commutative. By default, symbols in SymPy are considered commutative
except otherwise stated.
Examples:
>>> from sympy import *
>>> x, y = symbols(x,y)
>>> ask(Q.commutative(x))
True
>>> ask(Q.commutative(x), ~Q.commutative(x))
False
1296
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> ask(Q.commutative(x*y), ~Q.commutative(x))
False
complex
Test that expression belongs to the eld of complex numbers.
Examples:
>>> from sympy import *
>>> ask(Q.complex(2))
True
>>> ask(Q.complex(I))
True
>>> x, y = symbols(x,y)
>>> ask(Q.complex(x+I*y), Q.real(x) & Q.real(y))
True
even
Test that expression represents an even number, that is, an number that can be written in the
form 2*n, n integer.
Examples:
>>> from sympy import *
>>> ask(Q.even(2))
True
>>> n = Symbol(n)
>>> ask(Q.even(2*n), Q.integer(n))
True
extended real
Test that an expression belongs to the eld of extended real numbers, that is, real numbers
union {Innity, -Innity}.
Examples:
>>> from sympy import *
>>> ask(Q.extended_real(oo))
True
>>> ask(Q.extended_real(2))
True
>>> ask(Q.extended_real(x), Q.real(x))
True
imaginary
Test that an expression belongs to the set of imaginary numbers, that is, it can be
written as x*I, where x is real and I is the imaginary unit.
Examples:
5.20. Assumptions module
1297
SymPy Documentation, Release 0.7.6
>>> from sympy import *
>>> ask(Q.imaginary(2*I))
True
>>> x = Symbol(x)
>>> ask(Q.imaginary(x*I), Q.real(x))
True
innitesimal
Test that an expression is equivalent to an innitesimal number.
Examples:
>>> from sympy import *
>>> ask(Q.infinitesimal(1/oo))
True
>>> x, y = symbols(x,y)
>>> ask(Q.infinitesimal(2*x), Q.infinitesimal(x))
True
>>> ask(Q.infinitesimal(x*y), Q.infinitesimal(x) & Q.bounded(y))
True
integer
Test that an expression belongs to the set of integer numbers.
Examples:
>>> from sympy import *
>>> ask(Q.integer(2))
True
>>> ask(Q.integer(sqrt(2)))
False
>>> x = Symbol(x)
>>> ask(Q.integer(x/2), Q.even(x))
True
irrational
Test that an expression represents an irrational number.
Examples:
>>> from sympy import *
>>> ask(Q.irrational(pi))
True
>>> ask(Q.irrational(sqrt(2)))
True
>>> ask(Q.irrational(x*sqrt(2)), Q.rational(x))
True
rational
Test that an expression represents a rational number.
1298
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples:
>>> from sympy import *
>>> ask(Q.rational(Rational(3, 4)))
True
>>> x, y = symbols(x,y)
>>> ask(Q.rational(x/2), Q.integer(x))
True
>>> ask(Q.rational(x/y), Q.integer(x) & Q.integer(y))
True
negative
Test that an expression is less (strict) than zero.
Examples:
>>> from sympy import *
>>> ask(Q.negative(0.3))
False
>>> x = Symbol(x)
>>> ask(Q.negative(-x), Q.positive(x))
True
Remarks
negative numbers are dened as real numbers that are not zero nor positive, so complex
numbers (with nontrivial imaginary coecients) will return False for this predicate. The
same applies to Q.positive.
positive
Test that a given expression is greater (strict) than zero.
Examples:
>>> from sympy import *
>>> ask(Q.positive(0.3))
True
>>> x = Symbol(x)
>>> ask(Q.positive(-x), Q.negative(x))
True
Remarks
see Remarks for negative
prime
Test that an expression represents a prime number.
Examples:
5.20. Assumptions module
1299
SymPy Documentation, Release 0.7.6
>>> from sympy import *
>>> ask(Q.prime(13))
True
Remarks: Use sympy.ntheory.isprime to test numeric values eciently.
real
Test that an expression belongs to the eld of real numbers.
Examples:
>>> from sympy import *
>>> ask(Q.real(sqrt(2)))
True
>>> x, y = symbols(x,y)
>>> ask(Q.real(x*y), Q.real(x) & Q.real(y))
True
odd
Test that an expression represents an odd number.
Examples:
>>> from sympy import *
>>> ask(Q.odd(3))
True
>>> n = Symbol(n)
>>> ask(Q.odd(2*n + 1), Q.integer(n))
True
nonzero
Test that an expression is not zero.
Examples:
>>> from sympy import *
>>> x = Symbol(x)
>>> ask(Q.nonzero(x), Q.positive(x) | Q.negative(x))
True
5.20.4 Design
Each time ask is called, the appropriate Handler for the current key is called. This is always a
subclass of sympy.assumptions.AskHandler. Its classmethods have the names of the classes
it supports. For example, a (simplied) AskHandler for the ask positive would look like this:
class AskPositiveHandler(CommonHandler):
def Mul(self):
# return True if all arguments in self.expr.args are positive
...
1300
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
def Add(self):
for arg in self.expr.args:
if not ask(arg, positive, self.assumptions):
break
else:
# if all arguments are positive
return True
...
The .Mul() method is called when self.expr is an instance of Mul, the Add method would be
called when self.expr is an instance of Add and so on.
5.20.5 Extensibility
You can dene new queries or support new types by subclassing sympy.assumptions.AskHandler
and registering that handler for a particular key by calling register handler:
sympy.assumptions.ask.register handler(key, handler)
Register a handler in the ask system. key must be a string and handler a class inheriting
from AskHandler:
>>> from sympy.assumptions import register_handler, ask, Q
>>> from sympy.assumptions.handlers import AskHandler
>>> class MersenneHandler(AskHandler):
...
# Mersenne numbers are in the form 2**n + 1, n integer
...
@staticmethod
...
def Integer(expr, assumptions):
...
import math
...
return ask(Q.integer(math.log(expr + 1, 2)))
>>> register_handler(mersenne, MersenneHandler)
>>> ask(Q.mersenne(7))
True
You can undo this operation by calling remove handler.
sympy.assumptions.ask.remove handler(key, handler)
Removes a handler from the ask system. Same syntax as register handler
You can support new types 4 by adding a handler to an existing key. In the following example,
we will create a new type MyType and extend the key prime to accept this type (and return
True)
>>> from sympy.core import Basic
>>> from sympy.assumptions import register_handler
>>> from sympy.assumptions.handlers import AskHandler
>>> class MyType(Basic):
...
pass
>>> class MyAskHandler(AskHandler):
...
@staticmethod
...
def MyType(expr, assumptions):
...
return True
>>> a = MyType()
>>> register_handler(prime, MyAskHandler)
>>> ask(Q.prime(a))
True
4
New type must inherit from Basic, otherwise an exception will be raised. This is a bug and should be xed.
5.20. Assumptions module
1301
SymPy Documentation, Release 0.7.6
5.20.6 Performance improvements
On queries that involve symbolic coecients, logical inference is used. Work on improving
satisable function (sympy.logic.inference.satisable) should result in notable speed improvements.
Logic inference used in one ask could be used to speed up further queries, and current system does not take advantage of this. For example, a truth maintenance system
(http://en.wikipedia.org/wiki/Truth maintenance system) could be implemented.
5.20.7 Misc
You can nd more examples
sympy/assumptions/tests/
in
the
in
the
form
of
test
under
directory
5.21 Term rewriting
Term rewriting is a very general class of functionalities which are used to convert expressions of one type in terms of expressions of dierent kind. For example expanding, combining and converting expressions apply to term rewriting, and also simplication routines
can be included here. Currently SymPy has several functions and basic built-in methods for
performing various types of rewriting.
5.21.1 Expanding
The simplest rewrite rule is expanding expressions into a sparse form. Expanding has several avors and include expanding complex valued expressions, arithmetic expand of products and powers but also expanding functions in terms of more general functions is possible.
Below are listed all currently available expand rules.
Expanding of arithmetic expressions involving products and powers:
>>> from sympy import *
>>> x, y, z = symbols(x,y,z)
>>> ((x + y)*(x - y)).expand(basic=True)
x**2 - y**2
>>> ((x + y + z)**2).expand(basic=True)
x**2 + 2*x*y + 2*x*z + y**2 + 2*y*z + z**2
Arithmetic expand is done by default in expand() so the keyword basic can be omitted. However you can set basic=False to avoid this type of expand if you use rules described below.
This give complete control on what is done with the expression.
Another type of expand rule is expanding complex valued expressions and putting them into
a normal form. For this complex keyword is used. Note that it will always perform arithmetic
expand to obtain the desired normal form:
>>> (x + I*y).expand(complex=True)
re(x) + I*re(y) + I*im(x) - im(y)
>>> sin(x + I*y).expand(complex=True)
sin(re(x) - im(y))*cosh(re(y) + im(x)) + I*cos(re(x) - im(y))*sinh(re(y) + im(x))
1302
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Note also that the same behavior can be obtained by using as real imag() method. However
it will return a tuple containing the real part in the rst place and the imaginary part in the
other. This can be also done in a two step process by using collect function:
>>> (x + I*y).as_real_imag()
(re(x) - im(y), re(y) + im(x))
>>> collect((x + I*y).expand(complex=True), I, evaluate=False)
{1: re(x) - im(y), I: re(y) + im(x)}
There is also possibility for expanding expressions in terms of expressions of dierent kind.
This is very general type of expanding and usually you would use rewrite() to do specic
type of rewrite:
>>> GoldenRatio.expand(func=True)
1/2 + sqrt(5)/2
5.21.2 Common Subexpression Detection and Collection
Before evaluating a large expression, it is often useful to identify common subexpressions,
collect them and evaluate them at once. This is implemented in the cse function. Examples:
>>> from sympy import cse, sqrt, sin, pprint
>>> from sympy.abc import x
>>> pprint(cse(sqrt(sin(x))), use_unicode=True)
________
[], \ sin(x)
>>> pprint(cse(sqrt(sin(x)+5)*sqrt(sin(x)+4)), use_unicode=True)
________
________
[(x, sin(x))], \ x + 4 \ x + 5
>>> pprint(cse(sqrt(sin(x+1) + 5 + cos(y))*sqrt(sin(x+1) + 4 + cos(y))),
...
use_unicode=True)
________
________
[(x, sin(x + 1) + cos(y))], \ x + 4 \ x + 5
>>> pprint(cse((x-y)*(z-y) + sqrt((x-y)*(z-y))), use_unicode=True)
____
[(x, -y), (x1, (x + x)(x + z))], \ x1 + x1
Optimizations to be performed before and after common subexpressions elimination can be
passed in theoptimizations optional argument. A set of predened basic optimizations can
be applied by passing optimizations=basic:
>>> pprint(cse((x-y)*(z-y) + sqrt((x-y)*(z-y)), optimizations=basic),
...
use_unicode=True)
____
[(x, -(x - y)(y - z))], \ x + x
However, these optimizations can be very slow for large expressions. Moreover, if speed is a
concern, one can pass the option order=none. Order of terms will then be dependent on
hashing algorithm implementation, but speed will be greatly improved.
More information:
5.21. Term rewriting
1303
SymPy Documentation, Release 0.7.6
5.22 Series Expansions
The series module implements series expansions as a function and many related functions.
5.22.1 Limits
The main purpose of this module is the computation of limits.
sympy.series.limits.limit(e, z, z0, dir=+)
Compute the limit of e(z) at the point z0.
z0 can be any expression, including oo and -oo.
For dir=+ (default) it calculates the limit from the right (z->z0+) and for dir=- the
limit from the left (z->z0-). For innite z0 (oo or -oo), the dir argument is determined
from the direction of the innity (i.e., dir=- for oo).
Notes
First we try some heuristics for easy and frequent cases like x, 1/x, x**2 and similar, so that its fast. For all other cases, we use the Gruntz algorithm (see the gruntz()
function).
Examples
>>>
>>>
>>>
1
>>>
oo
>>>
-oo
>>>
0
from sympy import limit, sin, Symbol, oo
from sympy.abc import x
limit(sin(x)/x, x, 0)
limit(1/x, x, 0, dir=+)
limit(1/x, x, 0, dir=-)
limit(1/x, x, oo)
class sympy.series.limits.Limit
Represents an unevaluated limit.
Examples
>>> from sympy import Limit, sin, Symbol
>>> from sympy.abc import x
>>> Limit(sin(x)/x, x, 0)
Limit(sin(x)/x, x, 0)
>>> Limit(1/x, x, 0, dir=-)
Limit(1/x, x, 0, dir=-)
doit(**hints)
Evaluates limit
As is explained above, the workhorse for limit computations is the function gruntz() which
implements Gruntz algorithm for computing limits.
1304
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
The Gruntz Algorithm
This section explains the basics of the algorithm used for computing limits. Most of the time
the limit() function should just work. However it is still useful to keep in mind how it is
implemented in case something does not work as expected.
First we dene an ordering on functions. Suppose f (x) and g(x) are two real-valued functions
such that limx f (x) = and similarly limx g(x) = . We shall say that f (x) dominates
(x)a
g(x), written f (x) g(x), if for all a, b R>0 we have limx fg(x)
b = . We also say that f (x) and
g(x) are of the same comparability class if neither f (x) g(x) nor g(x) f (x) and shall denote
it as f (x) g(x).
Note that whenever a, b R>0 then af (x)b f (x), and we shall use this to extend the denition
of to all functions which tend to 0 or as x . Thus we declare that f (x) 1/f (x) and
f (x) f (x).
It is easy to show the following examples:
ex x m
2
ex emx
x
ee ex
xm xn
1
ex+ x ex+log x ex .
From the above denition, it is possible to prove the following property:
Suppose , g1 , g2 , . . . are functions of x, limx = 0 and gi for all i. Let c1 , c2 ,
R with c1 < c2 < . . . .
Then limx i gi ci = limx g1 c1 .
For g1 = g and as above we also have the following easy result:
limx g c = 0 for c > 0
limx g c = for c < 0, where the sign is determined by the (eventual) sign of g
limx g 0 = limx g.
Using these results yields the following strategy for computing limx f (x):
1. Find the set of most rapidly varying subexpressions (MRV set) of f (x). That is, from the
set of all subexpressions of f (x), nd the elements that are maximal under the relation
.
2. Choose a function that is in the same comparability class as the elements in the MRV
set, such that limx = 0.
3. Expand f (x) as a series in in such a way that the antecedents of the above theorem are
satised.
4. Apply the theorem and conclude the computation of limx f (x), possibly by recursively
working on g1 (x).
Notes
This exposition glossed over several details. Many are described in the le gruntz.py, and all
can be found in Gruntz very readable thesis. The most important points that have not been
explained are:
1. Given f(x) and g(x), how do we determine if f (x) g(x), g(x) f (x) or g(x) f (x)?
5.22. Series Expansions
1305
SymPy Documentation, Release 0.7.6
2. How do we nd the MRV set of an expression?
3. How do we compute series expansions?
4. Why does the algorithm terminate?
If you are interested, be sure to take a look at Gruntz Thesis.
Reference
sympy.series.gruntz.gruntz(e, z, z0, dir=+)
Compute the limit of e(z) at the point z0 using the Gruntz algorithm.
z0 can be any expression, including oo and -oo.
For dir=+ (default) it calculates the limit from the right (z->z0+) and for dir=- the
limit from the left (z->z0-). For innite z0 (oo or -oo), the dir argument doesnt matter.
This algorithm is fully described in the module docstring in the gruntz.py le. It relies
heavily on the series expansion. Most frequently, gruntz() is only used if the faster limit()
function (which uses heuristics) fails.
sympy.series.gruntz.compare(a, b, x)
Returns < if a<b, = for a == b, > for a>b
sympy.series.gruntz.rewrite(e, Omega, x, wsym)
e(x) ... the function Omega ... the mrv set wsym ... the symbol which is going to be used
for w
Returns the rewritten e in terms of w and log(w). See test rewrite1() for examples and
correct results.
sympy.series.gruntz.build expression tree(Omega, rewrites)
Helper function for rewrite.
We need to sort Omega (mrv set) so that we replace an expression before we replace any
expression in terms of which it has to be rewritten:
e1 ---> e2 ---> e3
\
-> e4
Here we can do e1, e2, e3, e4 or e1, e2, e4, e3. To do this we assemble the nodes into a
tree, and sort them by height.
This function builds the tree, rewrites then sorts the nodes.
sympy.series.gruntz.mrv leadterm(*args, **kwargs)
Returns (c0, e0) for e.
sympy.series.gruntz.calculate series(e, x, logx=None)
Calculates at least one term of the series of e in x.
This is a place that fails most often, so it is in its own function.
sympy.series.gruntz.limitinf(*args, **kwargs)
Limit e(x) for x-> oo
sympy.series.gruntz.sign(*args, **kwargs)
Returns a sign of an expression e(x) for x->oo.
e > 0 for x sufficiently large ... 1
e == 0 for x sufficiently large ... 0
e < 0 for x sufficiently large ... -1
1306
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
The result of this function is currently undened if e changes sign arbitarily often for
arbitrarily large x (e.g. sin(x)).
Note that this returns zero only if e is constantly zero for x suciently large. [If e is
constant, of course, this is just the same thing as the sign of e.]
sympy.series.gruntz.mrv(e, x)
Returns a SubsSet of most rapidly varying (mrv) subexpressions of e, and e rewritten
in terms of these
sympy.series.gruntz.mrv max1(f, g, exps, x)
Computes the maximum of two sets of expressions f and g, which are in the same comparability class, i.e. mrv max1() compares (two elements of) f and g and returns the set,
which is in the higher comparability class of the union of both, if they have the same
order of variation. Also returns exps, with the appropriate substitutions made.
sympy.series.gruntz.mrv max3(f, expsf, g, expsg, union, expsboth, x)
Computes the maximum of two sets of expressions f and g, which are in the same comparability class, i.e. max() compares (two elements of) f and g and returns either (f, expsf)
[if f is larger], (g, expsg) [if g is larger] or (union, expsboth) [if f, g are of the same class].
class sympy.series.gruntz.SubsSet
Stores (expr, dummy) pairs, and how to rewrite expr-s.
The gruntz algorithm needs to rewrite certain expressions in term of a new variable w.
We cannot use subs, because it is just too smart for us. For example:
> Omega=[exp(exp( p - exp(- p))/(1 - 1/ p)), exp(exp( p))]
> O2=[exp(-exp( p) + exp(-exp(- p))*exp( p)/(1 - 1/ p))/ w, 1/ w]
> e = exp(exp( p - exp(- p))/(1 - 1/ p)) - exp(exp( p))
> e.subs(Omega[0],O2[0]).subs(Omega[1],O2[1])
-1/w + exp(exp(p)*exp(-exp(-p))/(1 - 1/p))
is really not what we want!
So we do it the hard way and keep track of all the things we potentially want to substitute
by dummy variables. Consider the expression:
exp(x - exp(-x)) + exp(x) + x.
The mrv set is {exp(x), exp(-x), exp(x - exp(-x))}. We introduce corresponding dummy
variables d1, d2, d3 and rewrite:
d3 + d1 + x.
This class rst of all keeps track of the mapping expr->variable, i.e. will at this stage be
a dictionary:
{exp(x): d1, exp(-x): d2, exp(x - exp(-x)): d3}.
[It turns out to be more convenient this way round.] But sometimes expressions in the
mrv set have other expressions from the mrv set as subexpressions, and we need to keep
track of that as well. In this case, d3 is really exp(x - d2), so rewrites at this stage is:
{d3: exp(x-d2)}.
The function rewrite uses all this information to correctly rewrite our expression in terms
of w. In this case w can be choosen to be exp(-x), i.e. d2. The correct rewriting then is:
exp(-w)/w + 1/w + x.
5.22. Series Expansions
1307
SymPy Documentation, Release 0.7.6
meets(s2)
Tell whether or not self and s2 have non-empty intersection
union(s2, exps=None)
Compute the union of self and s2, adjusting exps
5.22.2 More Intuitive Series Expansion
This is achieved by creating a wrapper around Basic.series(). This allows for the use of series(x*cos(x),x), which is possibly more intuative than (x*cos(x)).series(x).
Examples
>>>
>>>
>>>
1 -
from sympy import Symbol, cos, series
x = Symbol(x)
series(cos(x),x)
x**2/2 + x**4/24 + O(x**6)
Reference
sympy.series.series.series(expr, x=None, x0=0, n=6, dir=+)
Series expansion of expr around point x = x0.
See the doctring of Expr.series() for complete details of this wrapper.
5.22.3 Order Terms
This module also implements automatic keeping track of the order of your expansion.
Examples
>>> from sympy import Symbol, Order
>>> x = Symbol(x)
>>> Order(x) + x**2
O(x)
>>> Order(x) + 1
1 + O(x)
Reference
class sympy.series.order.Order
Represents the limiting behavior of some function
The order of a function characterizes the function based on the limiting behavior of the
function as it goes to some limit. Only taking the limit point to be a number is currently
supported. This is expressed in big O notation [R345] (page 1912).
The formal denition for the order of a function g(x) about a point a is such that g(x) =
O(f (x)) as x a if and only if for any > 0 there exists a M > 0 such that |g(x)| M |f (x)|
for |x a| < . This is equivalent to limxa sup |g(x)/f (x)| < .
1308
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Lets illustrate it on the following example by taking the expansion of sin(x) about 0:
sin(x) = x x3 /3! + O(x5 )
where in this case O(x5 ) = x5 /5! x7 /7! + . By the denition of O, for any > 0 there is
an M such that:
|x5 /5! x7 /7! + ....| <= M |x5 | for |x| <
or by the alternate denition:
lim |(x5 /5! x7 /7! + ....)/x5 | <
x0
which surely is true, because
lim |(x5 /5! x7 /7! + ....)/x5 | = 1/5!
x0
As it is usually used, the order of a function can be intuitively thought of representing
all terms of powers greater than the one specied. For example, O(x3 ) corresponds to
any terms proportional to x3 , x4 , . . . and any higher power. For a polynomial, this leaves
terms proportional to x2 , x and constants.
Notes
In O(f(x), x) the expression f(x) is assumed to have a leading term. O(f(x), x) is
automatically transformed to O(f(x).as leading term(x),x).
O(expr*f(x), x) is O(f(x), x)
O(expr, x) is O(1)
O(0, x) is 0.
Multivariate O is also supported:
O(f(x,
y),
x,
y)
is
transformed
y).as leading term(x,y).as leading term(y), x, y)
to
O(f(x,
In the multivariate case, it is assumed the limits w.r.t. the various symbols commute.
If no symbols are passed then all symbols in the expression are used and the limit point
is assumed to be zero.
References
[R345] (page 1912)
Examples
5.22. Series Expansions
1309
SymPy Documentation, Release 0.7.6
>>> from sympy import O, oo, cos, pi
>>> from sympy.abc import x, y
>>> O(x
O(x)
>>> O(x
O(x)
>>> O(x
O(x**2,
+ x**2)
+ x**2, (x, 0))
+ x**2, (x, oo))
(x, oo))
>>> O(1 + x*y)
O(1, x, y)
>>> O(1 + x*y, (x, 0), (y, 0))
O(1, x, y)
>>> O(1 + x*y, (x, oo), (y, oo))
O(x*y, (x, oo), (y, oo))
>>> O(1) in
True
>>> O(1, x)
False
>>> O(x) in
True
>>> O(x**2)
True
O(1, x)
in O(1)
O(1, x)
in O(x)
>>> O(x)*x
O(x**2)
>>> O(x) - O(x)
O(x)
>>> O(cos(x))
O(1)
>>> O(cos(x), (x, pi/2))
O(x - pi/2, (x, pi/2))
contains(*args, **kwargs)
Return True if expr belongs to Order(self.expr, *self.variables). Return False if self
belongs to expr. Return None if the inclusion relation cannot be determined (e.g.
when self and expr have dierent symbols).
5.22.4 Series Acceleration
TODO
Reference
sympy.series.acceleration.richardson(A, k, n, N)
Calculate an approximation for lim k->oo A(k) using Richardson extrapolation with the
terms A(n), A(n+1), ..., A(n+N+1). Choosing N = 2*n often gives good results.
A simple example is to calculate exp(1) using the limit denition. This limit converges
slowly; n = 100 only produces two accurate digits:
1310
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> from sympy.abc import n
>>> e = (1 + 1/n)**n
>>> print(round(e.subs(n, 100).evalf(), 10))
2.7048138294
Richardson extrapolation with 11 appropriately chosen terms gives a value that is accurate to the indicated precision:
>>> from sympy import E
>>> from sympy.series.acceleration import richardson
>>> print(round(richardson(e, n, 10, 20).evalf(), 10))
2.7182818285
>>> print(round(E.evalf(), 10))
2.7182818285
Another useful application is to speed up convergence of series. Computing 100 terms
of the zeta(2) series 1/k**2 yields only two accurate digits:
>>> from sympy.abc import k, n
>>> from sympy import Sum
>>> A = Sum(k**-2, (k, 1, n))
>>> print(round(A.subs(n, 100).evalf(), 10))
1.6349839002
Richardson extrapolation performs much better:
>>> from sympy import pi
>>> print(round(richardson(A, n, 10, 20).evalf(), 10))
1.6449340668
>>> print(round(((pi**2)/6).evalf(), 10))
# Exact value
1.6449340668
sympy.series.acceleration.shanks(A, k, n, m=1)
Calculate an approximation for lim k->oo A(k) using the n-term Shanks transformation S(A)(n). With m > 1, calculate the m-fold recursive Shanks transformation
S(S(...S(A)...))(n).
The Shanks transformation is useful for summing Taylor series that converge slowly near
a pole or singularity, e.g. for log(2):
>>> from sympy.abc import k, n
>>> from sympy import Sum, Integer
>>> from sympy.series.acceleration import shanks
>>> A = Sum(Integer(-1)**(k+1) / k, (k, 1, n))
>>> print(round(A.subs(n, 100).doit().evalf(), 10))
0.6881721793
>>> print(round(shanks(A, n, 25).evalf(), 10))
0.6931396564
>>> print(round(shanks(A, n, 25, 5).evalf(), 10))
0.6931471806
The correct value is 0.6931471805599453094172321215.
5.22.5 Residues
TODO
5.22. Series Expansions
1311
SymPy Documentation, Release 0.7.6
Reference
sympy.series.residues.residue(expr, x, x0)
Finds the residue of expr at the point x=x0.
The residue is dened as the coecient of 1/(x-x0) in the power series expansion about
x=x0.
References
1.http://en.wikipedia.org/wiki/Residue theorem
Examples
>>>
>>>
>>>
1
>>>
0
>>>
2
from sympy import Symbol, residue, sin
x = Symbol(x)
residue(1/x, x, 0)
residue(1/x**2, x, 0)
residue(2/sin(x), x, 0)
This function is essential for the Residue Theorem [1].
5.23 Sets
5.23.1 Set
class sympy.sets.sets.Set
The base class for any kind of set.
This is not meant to be used directly as a container of items. It does not behave like the
builtin set; see FiniteSet (page 1319) for that.
Real intervals are represented by the Interval (page 1317) class and unions of sets by
the Union (page 1320) class. The empty set is represented by the EmptySet (page 1323)
class and available as a singleton as S.EmptySet.
Attributes
is
is
is
is
Complement
EmptySet
Intersection
UniversalSet
boundary
The boundary or frontier of a set
A point x is on the boundary of a set S if
1.x is in the closure of S. I.e. Every neighborhood of x contains a point in S.
1312
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
2.x is not in the interior of S. I.e. There does not exist an open set centered on x
contained entirely within S.
There are the points on the outer rim of S. If S is open then these points need not
actually be contained within S.
For example, the boundary of an interval is its start and end points. This is true
regardless of whether or not the interval is open.
Examples
>>>
>>>
{0,
>>>
{0,
from sympy import Interval
Interval(0, 1).boundary
1}
Interval(0, 1, True, False).boundary
1}
complement(universe)
The complement of self w.r.t the given the universe.
Examples
>>> from sympy import Interval, S
>>> Interval(0, 1).complement(S.Reals)
(-oo, 0) U (1, oo)
>>> Interval(0, 1).complement(S.UniversalSet)
UniversalSet() \ [0, 1]
contains(other)
Returns True if other is contained in self as an element.
As a shortcut it is possible to use the in operator:
Examples
>>> from sympy import Interval
>>> Interval(0, 1).contains(0.5)
True
>>> 0.5 in Interval(0, 1)
True
inf
The inmum of self
Examples
>>> from sympy import Interval, Union
>>> Interval(0, 1).inf
0
>>> Union(Interval(0, 1), Interval(2, 3)).inf
0
5.23. Sets
1313
SymPy Documentation, Release 0.7.6
intersect(other)
Returns the intersection of self and other.
>>> from sympy import Interval
>>> Interval(1, 3).intersect(Interval(1, 2))
[1, 2]
intersection(other)
Alias for intersect() (page 1313)
is disjoint(other)
Returns True if self and other are disjoint
References
[R346] (page 1912)
Examples
>>> from sympy import Interval
>>> Interval(0, 2).is_disjoint(Interval(1, 2))
False
>>> Interval(0, 2).is_disjoint(Interval(3, 4))
True
is proper subset(other)
Returns True if self is a proper subset of other.
Examples
>>> from sympy import Interval
>>> Interval(0, 0.5).is_proper_subset(Interval(0, 1))
True
>>> Interval(0, 1).is_proper_subset(Interval(0, 1))
False
is proper superset(other)
Returns True if self is a proper superset of other.
Examples
>>> from sympy import Interval
>>> Interval(0, 1).is_proper_superset(Interval(0, 0.5))
True
>>> Interval(0, 1).is_proper_superset(Interval(0, 1))
False
is subset(other)
Returns True if self is a subset of other.
1314
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import Interval
>>> Interval(0, 0.5).is_subset(Interval(0, 1))
True
>>> Interval(0, 1).is_subset(Interval(0, 1, left_open=True))
False
is superset(other)
Returns True if self is a superset of other.
Examples
>>> from sympy import Interval
>>> Interval(0, 0.5).is_superset(Interval(0, 1))
False
>>> Interval(0, 1).is_superset(Interval(0, 1, left_open=True))
True
isdisjoint(other)
Alias for is disjoint() (page 1314)
issubset(other)
Alias for is subset() (page 1314)
issuperset(other)
Alias for is superset() (page 1315)
measure
The (Lebesgue) measure of self
Examples
>>> from sympy import Interval, Union
>>> Interval(0, 1).measure
1
>>> Union(Interval(0, 1), Interval(2, 3)).measure
2
powerset()
Find the Power set of self.
References
[R347] (page 1912)
Examples
>>> from sympy import FiniteSet, EmptySet
>>> A = EmptySet()
>>> A.powerset()
{EmptySet()}
>>> A = FiniteSet(1, 2)
5.23. Sets
1315
SymPy Documentation, Release 0.7.6
>>> A.powerset() == FiniteSet(FiniteSet(1), FiniteSet(2), FiniteSet(1, 2), EmptySet())
True
subset(*args, **kwargs)
Returns True if other is a subset of self.
sup
The supremum of self
Examples
>>> from sympy import Interval, Union
>>> Interval(0, 1).sup
1
>>> Union(Interval(0, 1), Interval(2, 3)).sup
3
union(other)
Returns the union of self and other.
Examples
As a shortcut it is possible to use the + operator:
>>>
>>>
[0,
>>>
[0,
>>>
(1,
from sympy import Interval, FiniteSet
Interval(0, 1).union(Interval(2, 3))
1] U [2, 3]
Interval(0, 1) + Interval(2, 3)
1] U [2, 3]
Interval(1, 2, True, True) + FiniteSet(2, 3)
2] U {3}
Similarly it is possible to use the - operator for set dierences:
>>>
(1,
>>>
[1,
Interval(0, 2) - Interval(0, 1)
2]
Interval(1, 3) - FiniteSet(2)
2) U (2, 3]
sympy.sets.sets.imageset(*args)
Image of set under transformation f.
If this function cant compute the image, it returns an unevaluated ImageSet object.
f (x)|x self
See Also:
sympy.sets.fancysets.ImageSet (page 1326)
Examples
1316
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> from sympy import Interval, Symbol, imageset, sin, Lambda
>>> x = Symbol(x)
>>> imageset(x, 2*x, Interval(0, 2))
[0, 4]
>>> imageset(lambda x: 2*x, Interval(0, 2))
[0, 4]
>>> imageset(Lambda(x, sin(x)), Interval(-2, 1))
ImageSet(Lambda(x, sin(x)), [-2, 1])
Elementary Sets
5.23.2 Interval
class sympy.sets.sets.Interval
Represents a real interval as a Set.
Usage: Returns an interval with end points start and end.
For left open=True (default left open is False) the interval will be open on the left.
Similarly, for right open=True the interval will be open on the right.
Notes
Only real end points are supported
Interval(a, b) with a > b will return the empty set
Use the evalf() method to turn an Interval into an mpmath mpi interval instance
References
[R348] (page 1912)
Examples
>>>
>>>
[0,
>>>
[0,
from sympy import Symbol, Interval
Interval(0, 1)
1]
Interval(0, 1, False, True)
1)
>>> a = Symbol(a, real=True)
>>> Interval(0, a)
[0, a]
5.23. Sets
1317
SymPy Documentation, Release 0.7.6
Attributes
is
is
is
is
Complement
EmptySet
Intersection
UniversalSet
as relational(symbol)
Rewrite an interval in terms of inequalities and logic operators.
end
The right end point of self.
This property takes the same value as the sup property.
Examples
>>> from sympy import Interval
>>> Interval(0, 1).end
1
is left unbounded
Return True if the left endpoint is negative innity.
is right unbounded
Return True if the right endpoint is positive innity.
left
The left end point of self.
This property takes the same value as the inf property.
Examples
>>> from sympy import Interval
>>> Interval(0, 1).start
0
left open
True if self is left-open.
Examples
>>> from sympy import Interval
>>> Interval(0, 1, left_open=True).left_open
True
>>> Interval(0, 1, left_open=False).left_open
False
right
The right end point of self.
This property takes the same value as the sup property.
1318
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import Interval
>>> Interval(0, 1).end
1
right open
True if self is right-open.
Examples
>>> from sympy import Interval
>>> Interval(0, 1, right_open=True).right_open
True
>>> Interval(0, 1, right_open=False).right_open
False
start
The left end point of self.
This property takes the same value as the inf property.
Examples
>>> from sympy import Interval
>>> Interval(0, 1).start
0
5.23.3 FiniteSet
class sympy.sets.sets.FiniteSet
Represents a nite set of discrete numbers
References
[R349] (page 1912)
Examples
>>> from sympy import FiniteSet
>>> FiniteSet(1, 2, 3, 4)
{1, 2, 3, 4}
>>> 3 in FiniteSet(1, 2, 3, 4)
True
5.23. Sets
1319
SymPy Documentation, Release 0.7.6
Attributes
is
is
is
is
Complement
EmptySet
Intersection
UniversalSet
as relational(symbol)
Rewrite a FiniteSet in terms of equalities and logic operators.
Compound Sets
5.23.4 Union
class sympy.sets.sets.Union
Represents a union of sets as a Set (page 1312).
See Also:
Intersection (page 1321)
References
[R350] (page 1912)
Examples
>>> from sympy import Union, Interval
>>> Union(Interval(1, 2), Interval(3, 4))
[1, 2] U [3, 4]
The Union constructor will always try to merge overlapping intervals, if possible. For
example:
>>> Union(Interval(1, 2), Interval(2, 3))
[1, 3]
Attributes
is
is
is
is
Complement
EmptySet
Intersection
UniversalSet
as relational(symbol)
Rewrite a Union in terms of equalities and logic operators.
static reduce(args)
Simplify a Union (page 1320) using known rules
We rst start with global rules like Merge all FiniteSets
Then we iterate through all pairs and ask the constituent sets if they can simplify
themselves with any other constituent
1320
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
5.23.5 Intersection
class sympy.sets.sets.Intersection
Represents an intersection of sets as a Set (page 1312).
See Also:
Union (page 1320)
References
[R351] (page 1912)
Examples
>>> from sympy import Intersection, Interval
>>> Intersection(Interval(1, 3), Interval(2, 4))
[2, 3]
We often use the .intersect method
>>> Interval(1,3).intersect(Interval(2,4))
[2, 3]
Attributes
is Complement
is EmptySet
is UniversalSet
as relational(symbol)
Rewrite an Intersection in terms of equalities and logic operators
static reduce(args)
Simplify an intersection using known rules
We rst start with global rules like if any empty sets return empty set and distribute
any unions
Then we iterate through all pairs and ask the constituent sets if they can simplify
themselves with any other constituent
5.23.6 ProductSet
class sympy.sets.sets.ProductSet
Represents a Cartesian Product of Sets.
Returns a Cartesian product given several sets as either an iterable or individual arguments.
Can use * operator on any sets for convenient shorthand.
5.23. Sets
1321
SymPy Documentation, Release 0.7.6
Notes
Passes most operations down to the argument sets
Flattens Products of ProductSets
References
[R352] (page 1912)
Examples
>>>
>>>
>>>
[0,
from sympy import Interval, FiniteSet, ProductSet
I = Interval(0, 5); S = FiniteSet(1, 2, 3)
ProductSet(I, S)
5] x {1, 2, 3}
>>> (2, 2) in ProductSet(I, S)
True
>>> Interval(0, 1) * Interval(0, 1) # The unit square
[0, 1] x [0, 1]
>>> coin = FiniteSet(H, T)
>>> set(coin**2)
set([(H, H), (H, T), (T, H), (T, T)])
Attributes
is
is
is
is
Complement
EmptySet
Intersection
UniversalSet
5.23.7 Complement
class sympy.sets.sets.Complement
Represents the set dierence or relative complement of a set with another set.
A B = {x A|x
/ B}
See Also:
Intersection (page 1321), Union (page 1320)
References
http://mathworld.wolfram.com/SetComplement.html
1322
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import Complement, FiniteSet
>>> Complement(FiniteSet(0, 1, 2), FiniteSet(1))
{0, 2}
Attributes
is EmptySet
is Intersection
is UniversalSet
static reduce(A, B)
Simplify a Complement (page 1322).
Singleton Sets
5.23.8 EmptySet
class sympy.sets.sets.EmptySet
Represents the empty set. The empty set is available as a singleton as S.EmptySet.
See Also:
UniversalSet (page 1324)
References
[R353] (page 1912)
Examples
>>> from sympy import S, Interval
>>> S.EmptySet
EmptySet()
>>> Interval(1, 2).intersect(S.EmptySet)
EmptySet()
Attributes
is Complement
is Intersection
is UniversalSet
5.23. Sets
1323
SymPy Documentation, Release 0.7.6
5.23.9 UniversalSet
class sympy.sets.sets.UniversalSet
Represents the set of all things.
S.UniversalSet
The universal set is available as a singleton as
See Also:
EmptySet (page 1323)
References
[R354] (page 1912)
Examples
>>> from sympy import S, Interval
>>> S.UniversalSet
UniversalSet()
>>> Interval(1, 2).intersect(S.UniversalSet)
[1, 2]
Attributes
is Complement
is EmptySet
is Intersection
Special Sets
5.23.10 Naturals
class sympy.sets.fancysets.Naturals
Represents the natural numbers (or counting numbers) which are all positive integers
starting from 1. This set is also available as the Singleton, S.Naturals.
See Also:
Naturals0 (page 1325) non-negative integers (i.e. includes 0, too)
Integers (page 1325) also includes negative integers
Examples
>>> from sympy import S, Interval, pprint
>>> 5 in S.Naturals
True
>>> iterable = iter(S.Naturals)
>>> next(iterable)
1324
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
1
>>>
2
>>>
3
>>>
{1,
next(iterable)
next(iterable)
pprint(S.Naturals.intersect(Interval(0, 10)))
2, ..., 10}
Attributes
is
is
is
is
Complement
EmptySet
Intersection
UniversalSet
5.23.11 Naturals0
class sympy.sets.fancysets.Naturals0
Represents the whole numbers which are all the non-negative integers, inclusive of zero.
See Also:
Naturals (page 1324) positive integers; does not include 0
Integers (page 1325) also includes the negative integers
Attributes
is
is
is
is
Complement
EmptySet
Intersection
UniversalSet
5.23.12 Integers
class sympy.sets.fancysets.Integers
Represents all integers: positive, negative and zero. This set is also available as the
Singleton, S.Integers.
See Also:
Naturals0 (page 1325) non-negative integers
Integers (page 1325) positive and negative integers and zero
Examples
5.23. Sets
1325
SymPy Documentation, Release 0.7.6
>>> from sympy import S, Interval, pprint
>>> 5 in S.Naturals
True
>>> iterable = iter(S.Integers)
>>> next(iterable)
0
>>> next(iterable)
1
>>> next(iterable)
-1
>>> next(iterable)
2
>>> pprint(S.Integers.intersect(Interval(-4, 4)))
{-4, -3, ..., 4}
Attributes
is
is
is
is
Complement
EmptySet
Intersection
UniversalSet
5.23.13 ImageSet
class sympy.sets.fancysets.ImageSet
Image of a set under a mathematical function
Examples
>>> from sympy import Symbol, S, ImageSet, FiniteSet, Lambda
>>> x = Symbol(x)
>>> N = S.Naturals
>>> squares = ImageSet(Lambda(x, x**2), N) # {x**2 for x in N}
>>> 4 in squares
True
>>> 5 in squares
False
>>> FiniteSet(0, 1, 2, 3, 4, 5, 6, 7, 9, 10).intersect(squares)
{1, 4, 9}
>>> square_iterable = iter(squares)
>>> for i in range(4):
...
next(square_iterable)
1
4
9
16
1326
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Attributes
is
is
is
is
Complement
EmptySet
Intersection
UniversalSet
5.24 Simplify
5.24.1 simplify
sympy.simplify.simplify.simplify(expr, ratio=1.7, measure=<function count ops
at 0x9db2b54>, fu=False)
Simplies the given expression.
Simplication is not a well dened term and the exact strategies this function tries can
change in the future versions of SymPy. If your algorithm relies on simplication
(whatever it is), try to determine what you need exactly - is it powsimp()?, radsimp()?,
together()?, logcombine()?, or something else? And use this particular function directly,
because those are well dened and thus your algorithm will be robust.
Nonetheless, especially for interactive use, or when you dont know anything about the
structure of the expression, simplify() tries to apply intelligent heuristics to make the
input expression simpler. For example:
>>> from sympy import simplify, cos, sin
>>> from sympy.abc import x, y
>>> a = (x + x**2)/(x*sin(y)**2 + x*cos(y)**2)
>>> a
(x**2 + x)/(x*sin(y)**2 + x*cos(y)**2)
>>> simplify(a)
x + 1
Note that we could have obtained the same result by using specic simplication functions:
>>> from sympy import trigsimp, cancel
>>> trigsimp(a)
(x**2 + x)/x
>>> cancel(_)
x + 1
In some cases, applying simplify() (page 1327) may actually result in some more complicated expression. The default ratio=1.7 prevents more extreme cases: if (result
length)/(input length) > ratio, then input is returned unmodied. The measure parameter lets you specify the function used to determine how complex an expression is. The
function should take a single argument as an expression and return a number such that
if expression a is more complex than expression b, then measure(a) > measure(b). The
default measure function is count ops(), which returns the total number of operations
in the expression.
For example, if ratio=1, simplify output cant be longer than input.
>>> from sympy import sqrt, simplify, count_ops, oo
>>> root = 1/(sqrt(2)+3)
5.24. Simplify
1327
SymPy Documentation, Release 0.7.6
Since simplify(root) would result in a slightly longer expression, root is returned unchanged instead:
>>> simplify(root, ratio=1) == root
True
If ratio=oo, simplify will be applied anyway:
>>> count_ops(simplify(root, ratio=oo)) > count_ops(root)
True
Note that the shortest expression is not necessary the simplest, so setting ratio to 1 may
not be a good idea. Heuristically, the default value ratio=1.7 seems like a reasonable
choice.
You can easily dene your own measure function based on what you feel should represent
the size or complexity of the input expression. Note that some choices, such as
lambda expr: len(str(expr)) may appear to be good metrics, but have other problems
(in this case, the measure function may slow down simplify too much for very large
expressions). If you dont know what a good metric would be, the default, count ops, is
a good one.
For example:
>>> from sympy import symbols, log
>>> a, b = symbols(a b, positive=True)
>>> g = log(a) + log(b) + log(a)*log(1/b)
>>> h = simplify(g)
>>> h
log(a*b**(-log(a) + 1))
>>> count_ops(g)
8
>>> count_ops(h)
5
So you can see that h is simpler than g using the count ops metric. However, we may not
like how simplify (in this case, using logcombine) has created the b**(log(1/a) + 1)
term. A simple way to reduce this would be to give more weight to powers as operations
in count ops. We can do this by using the visual=True option:
>>> print(count_ops(g, visual=True))
2*ADD + DIV + 4*LOG + MUL
>>> print(count_ops(h, visual=True))
2*LOG + MUL + POW + SUB
>>> from sympy import Symbol, S
>>> def my_measure(expr):
...
POW = Symbol(POW)
...
# Discourage powers by giving POW a weight of 10
...
count = count_ops(expr, visual=True).subs(POW, 10)
...
# Every other operation gets a weight of 1 (the default)
...
count = count.replace(Symbol, type(S.One))
...
return count
>>> my_measure(g)
8
>>> my_measure(h)
14
>>> 15./8 > 1.7 # 1.7 is the default ratio
True
1328
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> simplify(g, measure=my_measure)
-log(a)*log(b) + log(a) + log(b)
Note that because simplify() internally tries many dierent simplication strategies
and then compares them using the measure function, we get a completely dierent result
that is still dierent from the input expression by doing this.
5.24.2 collect
sympy.simplify.simplify.collect(expr, syms, func=None, evaluate=None,
act=False, distribute order term=True)
Collect additive terms of an expression.
ex-
This function collects additive terms of an expression with respect to a list of expression
up to powers with rational exponents. By the term symbol here are meant arbitrary
expressions, which can contain powers, products, sums etc. In other words symbol is a
pattern which will be searched for in the expressions terms.
The input expression is not expanded by collect() (page 1329), so user is expected
to provide an expression is an appropriate form. This makes collect() (page 1329)
more predictable as there is no magic happening behind the scenes. However, it is
important to note, that powers of products are converted to products of powers using
the expand power base() function.
There are two possible types of output. First, if evaluate ag is set, this function will
return an expression with collected terms or else it will return a dictionary with expressions up to rational powers as keys and collected coecients as values.
See Also:
collect const (page 1341), collect sqrt (page 1340), rcollect (page 1331)
Examples
>>> from sympy import S, collect, expand, factor, Wild
>>> from sympy.abc import a, b, c, x, y, z
This function can collect symbolic coecients in polynomials or rational expressions. It
will manage to nd all integer or rational powers of collection variable:
>>> collect(a*x**2 + b*x**2 + a*x - b*x + c, x)
c + x**2*(a + b) + x*(a - b)
The same result can be achieved in dictionary form:
>>>
>>>
a +
>>>
a >>>
c
d = collect(a*x**2 + b*x**2 + a*x - b*x + c, x, evaluate=False)
d[x**2]
b
d[x]
b
d[S.One]
You can also work with multivariate polynomials. However, remember that this function
is greedy so it will care only about a single symbol at time, in specication order:
5.24. Simplify
1329
SymPy Documentation, Release 0.7.6
>>> collect(x**2 + y*x**2 + x*y + y + a*y, [x, y])
x**2*(y + 1) + x*y + y*(a + 1)
Also more complicated expressions can be used as patterns:
>>> from sympy import sin, log
>>> collect(a*sin(2*x) + b*sin(2*x), sin(2*x))
(a + b)*sin(2*x)
>>> collect(a*x*log(x) + b*(x*log(x)), x*log(x))
x*(a + b)*log(x)
You can use wildcards in the pattern:
>>> w = Wild(w1)
>>> collect(a*x**y - b*x**y, w**y)
x**y*(a - b)
It is also possible to work with symbolic powers, although it has more complicated behavior, because in this case powers base and symbolic part of the exponent are treated
as a single symbol:
>>> collect(a*x**c + b*x**c, x)
a*x**c + b*x**c
>>> collect(a*x**c + b*x**c, x**c)
x**c*(a + b)
However if you incorporate rationals to the exponents, then you will get well known
behavior:
>>> collect(a*x**(2*c) + b*x**(2*c), x**c)
x**(2*c)*(a + b)
Note also that all previously stated facts about collect() (page 1329) function apply to
the exponential function, so you can get:
>>> from sympy import exp
>>> collect(a*exp(2*x) + b*exp(2*x), exp(x))
(a + b)*exp(2*x)
If you are interested only in collecting specic powers of some symbols then set exact
ag in arguments:
>>> collect(a*x**7 + b*x**7, x, exact=True)
a*x**7 + b*x**7
>>> collect(a*x**7 + b*x**7, x**7, exact=True)
x**7*(a + b)
You can also apply this function to dierential equations, where derivatives of arbitrary
order can be collected. Note that if you collect with respect to a function or a derivative
of a function, all derivatives of that function will also be collected. Use exact=True to
prevent this from happening:
>>> from sympy import Derivative as D, collect, Function
>>> f = Function(f) (x)
>>> collect(a*D(f,x) + b*D(f,x), D(f,x))
(a + b)*Derivative(f(x), x)
1330
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> collect(a*D(D(f,x),x) + b*D(D(f,x),x), f)
(a + b)*Derivative(f(x), x, x)
>>> collect(a*D(D(f,x),x) + b*D(D(f,x),x), D(f,x), exact=True)
a*Derivative(f(x), x, x) + b*Derivative(f(x), x, x)
>>> collect(a*D(f,x) + b*D(f,x) + a*f + b*f, f)
(a + b)*f(x) + (a + b)*Derivative(f(x), x)
Or you can even match both derivative order and exponent at the same time:
>>> collect(a*D(D(f,x),x)**2 + b*D(D(f,x),x)**2, D(f,x))
(a + b)*Derivative(f(x), x, x)**2
Finally, you can apply a function to each of the collected coecients. For example you
can factorize symbolic coecients of polynomial:
>>> f = expand((x + a + 1)**3)
>>> collect(f, x, factor)
x**3 + 3*x**2*(a + 1) + 3*x*(a + 1)**2 + (a + 1)**3
Note: Arguments are expected to be in expanded form, so you might have to call
expand() prior to calling this function.
sympy.simplify.simplify.rcollect(expr, *vars)
Recursively collect sums in an expression.
See Also:
collect (page 1329), collect const (page 1341), collect sqrt (page 1340)
Examples
>>> from sympy.simplify import rcollect
>>> from sympy.abc import x, y
>>> expr = (x**2*y + x*y + x + y)/(x + y)
>>> rcollect(expr, y)
(x + y*(x**2 + x + 1))/(x + y)
5.24.3 separatevars
sympy.simplify.simplify.separatevars(expr, symbols=[], dict=False, force=False)
Separates variables in an expression, if possible. By default, it separates with respect to
all symbols in an expression and collects constant coecients that are independent of
symbols.
If dict=True then the separated terms will be returned in a dictionary keyed to their
corresponding symbols. By default, all symbols in the expression will appear as keys; if
symbols are provided, then all those symbols will be used as keys, and any terms in the
expression containing other symbols or non-symbols will be returned keyed to the string
5.24. Simplify
1331
SymPy Documentation, Release 0.7.6
coe. (Passing None for symbols will return the expression in a dictionary keyed to
coe.)
If force=True, then bases of powers will be separated regardless of assumptions on the
symbols involved.
Notes
The order of the factors is determined by Mul, so that the separated expressions may
not necessarily be grouped together.
Although factoring is necessary to separate variables in some expressions, it is not necessary in all cases, so one should not count on the returned factors being factored.
Examples
>>> from sympy.abc import x, y, z, alpha
>>> from sympy import separatevars, sin
>>> separatevars((x*y)**y)
(x*y)**y
>>> separatevars((x*y)**y, force=True)
x**y*y**y
>>> e = 2*x**2*z*sin(y)+2*z*x**2
>>> separatevars(e)
2*x**2*z*(sin(y) + 1)
>>> separatevars(e, symbols=(x, y), dict=True)
{coeff: 2*z, x: x**2, y: sin(y) + 1}
>>> separatevars(e, [x, y, alpha], dict=True)
{coeff: 2*z, alpha: 1, x: x**2, y: sin(y) + 1}
If the expression is not really separable, or is only partially separable, separatevars will
do the best it can to separate it by using factoring.
>>> separatevars(x + x*y - 3*x**2)
-x*(3*x - y - 1)
If the expression is not separable then expr is returned unchanged or (if dict=True) then
None is returned.
>>> eq = 2*x + y*sin(x)
>>> separatevars(eq) == eq
True
>>> separatevars(2*x + y*sin(x), symbols=(x, y), dict=True) == None
True
5.24.4 nthroot
sympy.simplify.simplify.nthroot(expr, n, max len=4, prec=15)
compute a real nth-root of a sum of surds
Parameters expr : sum of surds
n : integer
max len : maximum number of surds passed as constants to nsimplify
1332
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.simplify.simplify import nthroot
>>> from sympy import Rational, sqrt
>>> nthroot(90 + 34*sqrt(7), 3)
sqrt(7) + 3
Algorithm
First nsimplify is used to get a candidate root; if it is not a root the minimal polynomial
is computed; the answer is one of its roots.
5.24.5 rad rationalize
sympy.simplify.simplify.rad rationalize(num, den)
Rationalize num/den by removing square roots in the denominator; num and den are
sum of terms whose squares are rationals
Examples
>>> from sympy import sqrt
>>> from sympy.simplify.simplify import rad_rationalize
>>> rad_rationalize(sqrt(3), 1 + sqrt(2)/3)
(-sqrt(3) + sqrt(6)/3, -7/9)
5.24.6 radsimp
sympy.simplify.simplify.radsimp(expr, symbolic=True, max terms=4)
Rationalize the denominator by removing square roots.
Note: the expression returned from radsimp must be used with caution since if the denominator contains symbols, it will be possible to make substitutions that violate the assumptions of the simplication process: that for a denominator matching a + b*sqrt(c),
a != +/-b*sqrt(c). (If there are no symbols, this assumptions is made valid by collecting
terms of sqrt(c) so the match variable a does not contain sqrt(c).) If you do not want
the simplication to occur for symbolic denominators, set symbolic to False.
If there are more than max terms radical terms then the expression is returned unchanged.
Examples
>>>
>>>
>>>
>>>
from
from
from
from
sympy import radsimp, sqrt, Symbol, denom, pprint, I
sympy import factor_terms, fraction, signsimp
sympy.simplify.simplify import collect_sqrt
sympy.abc import a, b, c
5.24. Simplify
1333
SymPy Documentation, Release 0.7.6
>>> radsimp(1/(I + 1))
(1 - I)/2
>>> radsimp(1/(2 + sqrt(2)))
(-sqrt(2) + 2)/2
>>> x,y = map(Symbol, xy)
>>> e = ((2 + 2*sqrt(2))*x + (2 + sqrt(8))*y)/(2 + sqrt(2))
>>> radsimp(e)
sqrt(2)*(x + y)
No simplication beyond removal of the gcd is done. One might want to polish the result
a little, however, by collecting square root terms:
>>> r2 = sqrt(2)
>>> r5 = sqrt(5)
>>> ans = radsimp(1/(y*r2 + x*r2 + a*r5 + b*r5)); pprint(ans)
___
___
___
___
\/ 5 *a + \/ 5 *b - \/ 2 *x - \/ 2 *y
-----------------------------------------2
2
2
2
5*a + 10*a*b + 5*b - 2*x - 4*x*y - 2*y
>>> n, d = fraction(ans)
>>> pprint(factor_terms(signsimp(collect_sqrt(n))/d, radical=True))
___
___
\/ 5 *(a + b) - \/ 2 *(x + y)
-----------------------------------------2
2
2
2
5*a + 10*a*b + 5*b - 2*x - 4*x*y - 2*y
If radicals in the denominator cannot be removed or there is no denominator, the original
expression will be returned.
>>> radsimp(sqrt(2)*x + sqrt(2))
sqrt(2)*x + sqrt(2)
Results with symbols will not always be valid for all substitutions:
>>> eq = 1/(a + b*sqrt(c))
>>> eq.subs(a, b*sqrt(c))
1/(2*b*sqrt(c))
>>> radsimp(eq).subs(a, b*sqrt(c))
nan
If symbolic=False, symbolic denominators will not be transformed (but numeric denominators will still be processed):
>>> radsimp(eq, symbolic=False)
1/(a + b*sqrt(c))
5.24.7 ratsimp
sympy.simplify.simplify.ratsimp(expr)
Put an expression over a common denominator, cancel and reduce.
1334
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import ratsimp
>>> from sympy.abc import x, y
>>> ratsimp(1/x + 1/y)
(x + y)/(x*y)
5.24.8 fraction
sympy.simplify.simplify.fraction(expr, exact=False)
Returns a pair with expressions numerator and denominator. If the given expression is
not a fraction then this function will return the tuple (expr, 1).
This function will not make any attempt to simplify nested fractions or to do any term
rewriting at all.
If only one of the numerator/denominator pair is needed then use numer(expr) or denom(expr) functions respectively.
>>> from sympy import fraction, Rational, Symbol
>>> from sympy.abc import x, y
>>>
(x,
>>>
(x,
fraction(x/y)
y)
fraction(x)
1)
>>> fraction(1/y**2)
(1, y**2)
>>> fraction(x*y/2)
(x*y, 2)
>>> fraction(Rational(1, 2))
(1, 2)
This function will also work ne with assumptions:
>>> k = Symbol(k, negative=True)
>>> fraction(x * y**k)
(x, y**(-k))
If we know nothing about sign of some exponent and exact ag is unset, then structure
this exponents structure will be analyzed and pretty fraction will be returned:
>>> from sympy import exp
>>> fraction(2*x**(-y))
(2, x**y)
>>> fraction(exp(-x))
(1, exp(x))
>>> fraction(exp(-x), exact=True)
(exp(-x), 1)
5.24. Simplify
1335
SymPy Documentation, Release 0.7.6
5.24.9 trigsimp
sympy.simplify.simplify.trigsimp(expr, **opts)
reduces expression by using known trig identities
Notes
method: - Determine the method to use. Valid choices are matching (default), groebner, combined, and fu. If matching, simplify the expression recursively by targeting
common patterns. If groebner, apply an experimental groebner basis algorithm. In this
case further options are forwarded to trigsimp groebner, please refer to its docstring.
If combined, rst run the groebner basis algorithm with small default parameters, then
run the matching algorithm. fu runs the collection of trigonometric transformations
described by Fu, et al. (see the f u docstring).
Examples
>>>
>>>
>>>
>>>
2
from sympy import trigsimp, sin, cos, log
from sympy.abc import x, y
e = 2*sin(x)**2 + 2*cos(x)**2
trigsimp(e)
Simplication occurs wherever trigonometric functions are located.
>>> trigsimp(log(e))
log(2)
Using method = groebner (or combined) might lead to greater simplication.
The old trigsimp routine can be accessed as with method old.
>>> from sympy import coth, tanh
>>> t = 3*tanh(x)**7 - 2/coth(x)**7
>>> trigsimp(t, method=old) == t
True
>>> trigsimp(t)
tanh(x)**7
5.24.10 besselsimp
sympy.simplify.simplify.besselsimp(expr)
Simplify bessel-type functions.
This routine tries to simplify bessel-type functions. Currently it only works on the Bessel J
and I functions, however. It works by looking at all such functions in turn, and eliminating
factors of I and -1 (actually their polar equivalents) in front of the argument. Then,
functions of half-integer order are rewritten using trigonometric functions and functions
of integer order (> 1) are rewritten using functions of low order. Finally, if the expression
was changed, compute factorization of the result with factor().
1336
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> from sympy import besselj, besseli, besselsimp, polar_lift, I, S
>>> from sympy.abc import z, nu
>>> besselsimp(besselj(nu, z*polar_lift(-1)))
exp(I*pi*nu)*besselj(nu, z)
>>> besselsimp(besseli(nu, z*polar_lift(-I)))
exp(-I*pi*nu/2)*besselj(nu, z)
>>> besselsimp(besseli(S(-1)/2, z))
sqrt(2)*cosh(z)/(sqrt(pi)*sqrt(z))
>>> besselsimp(z*besseli(0, z) + z*(besseli(2, z))/2 + besseli(1, z))
3*z*besseli(0, z)/2
5.24.11 powsimp
sympy.simplify.simplify.powsimp(expr, deep=False, combine=all, force=False,
measure=<function count ops at 0x9db2b54>)
reduces expression by combining powers with similar bases and exponents.
Notes
If deep is True then powsimp() will also simplify arguments of functions. By default deep
is set to False.
If force is True then bases will be combined without checking for assumptions, e.g.
sqrt(x)*sqrt(y) -> sqrt(x*y) which is not true if x and y are both negative.
You can make powsimp() only combine bases or only combine exponents by changing
combine=base or combine=exp. By default, combine=all, which does both. combine=base will only combine:
a
a
x * y =>
a
(x*y)
2x
as well as things like 2
=>
x
4
and combine=exp will only combine
a
b
x * x =>
(a + b)
x
combine=exp will strictly only combine exponents in the way that used to be automatic.
Also use deep=True if you need the old behavior.
When combine=all, exp is evaluated rst. Consider the rst example below for when
there could be an ambiguity relating to this. This is done so things like the second
example can be completely combined. If you want base combined rst, do something
like powsimp(powsimp(expr, combine=base), combine=exp).
Examples
>>> from sympy import powsimp, exp, log, symbols
>>> from sympy.abc import x, y, z, n
>>> powsimp(x**y*x**z*y**z, combine=all)
x**(y + z)*y**z
>>> powsimp(x**y*x**z*y**z, combine=exp)
x**(y + z)*y**z
5.24. Simplify
1337
SymPy Documentation, Release 0.7.6
>>> powsimp(x**y*x**z*y**z, combine=base, force=True)
x**y*(x*y)**z
>>> powsimp(x**z*x**y*n**z*n**y, combine=all, force=True)
(n*x)**(y + z)
>>> powsimp(x**z*x**y*n**z*n**y, combine=exp)
n**(y + z)*x**(y + z)
>>> powsimp(x**z*x**y*n**z*n**y, combine=base, force=True)
(n*x)**y*(n*x)**z
>>> x, y = symbols(x y, positive=True)
>>> powsimp(log(exp(x)*exp(y)))
log(exp(x)*exp(y))
>>> powsimp(log(exp(x)*exp(y)), deep=True)
x + y
Radicals with Mul bases will be combined if combine=exp
>>> from sympy import sqrt, Mul
>>> x, y = symbols(x y)
Two radicals are automatically joined through Mul: >>> a=sqrt(x*sqrt(y)) >>> a*a**3
== a**4 True
But if an integer power of that radical has been autoexpanded then Mul does not join the
resulting factors: >>> a**4 # auto expands to a Mul, no longer a Pow x**2*y >>> *a
# so Mul doesnt combine them x**2*y*sqrt(x*sqrt(y)) >>> powsimp( ) # but powsimp
will (x*sqrt(y))**(5/2) >>> powsimp(x*y*a) # but wont when doing so would violate
assumptions x*y*sqrt(x*sqrt(y))
5.24.12 combsimp
sympy.simplify.simplify.combsimp(expr)
Simplify combinatorial expressions.
This function takes as input an expression containing factorials, binomials, Pochhammer
symbol and other combinatorial functions, and tries to minimize the number of those
functions and reduce the size of their arguments. The result is be given in terms of
binomials and factorials.
The algorithm works by rewriting all combinatorial functions as expressions involving
rising factorials (Pochhammer symbols) and applies recurrence relations and other transformations applicable to rising factorials, to reduce their arguments, possibly letting the
resulting rising factorial to cancel. Rising factorials with the second argument being
an integer are expanded into polynomial forms and nally all other rising factorial are
rewritten in terms more familiar functions. If the initial expression contained any combinatorial functions, the result is expressed using binomial coecients and gamma functions. If the initial expression consisted of gamma functions alone, the result is expressed
in terms of gamma functions.
If the result is expressed using gamma functions, the following three additional steps
are performed:
1.Reduce the number of gammas
gamma(x)*gamma(1-x) == pi/sin(pi*x).
1338
by
applying
the
reection
theorem
2.Reduce the number of gammas by applying the multiplication
gamma(x)*gamma(x+1/n)*...*gamma(x+(n-1)/n) == C*gamma(n*x).
theorem
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
3.Reduce the number of prefactors by absorbing them into gammas, where possible.
All transformation rules can be found (or was derived from) here:
1.http://functions.wolfram.com/GammaBetaErf/Pochhammer/17/01/02/
2.http://functions.wolfram.com/GammaBetaErf/Pochhammer/27/01/0005/
Examples
>>> from sympy.simplify import combsimp
>>> from sympy import factorial, binomial
>>> from sympy.abc import n, k
>>> combsimp(factorial(n)/factorial(n - 3))
n*(n - 2)*(n - 1)
>>> combsimp(binomial(n+1, k+1)/binomial(n, k))
(n + 1)/(k + 1)
5.24.13 hypersimp
sympy.simplify.simplify.hypersimp(f, k)
Given combinatorial term f(k) simplify its consecutive term ratio i.e. f(k+1)/f(k). The
input term can be composed of functions and integer sequences which have equivalent
representation in terms of gamma special function.
The algorithm performs three basic steps:
1.Rewrite all functions in terms of gamma, if possible.
2.Rewrite all occurrences of gamma in terms of products of gamma and rising factorial
with integer, absolute constant exponent.
3.Perform simplication of nested fractions, powers and if the resulting expression is
a quotient of polynomials, reduce their total degree.
If f(k) is hypergeometric then as result we arrive with a quotient of polynomials of minimal degree. Otherwise None is returned.
For more information on the implemented algorithm refer to:
1.W. Koepf, Algorithms for m-fold Hypergeometric Summation, Journal of Symbolic
Computation (1995) 20, 399-417
5.24.14 hypersimilar
sympy.simplify.simplify.hypersimilar(f, g, k)
Returns True if f and g are hyper-similar.
Similarity in hypergeometric sense means that a quotient of f(k) and g(k) is a rational
function in k. This procedure is useful in solving recurrence relations.
For more information see hypersimp().
5.24. Simplify
1339
SymPy Documentation, Release 0.7.6
5.24.15 nsimplify
sympy.simplify.simplify.nsimplify(expr,
constants=[],
tolerance=None,
full=False, rational=None)
Find a simple representation for a number or, if there are free symbols or if rational=True, then replace Floats with their Rational equivalents. If no change is made
and rational is not False then Floats will at least be converted to Rationals.
For numerical expressions, a simple formula that numerically matches the given numerical expression is sought (and the input should be possible to evalf to a precision of at
least 30 digits).
Optionally, a list of (rationally independent) constants to include in the formula may be
given.
A lower tolerance may be set to nd less exact matches. If no tolerance is given then the
least precise value will set the tolerance (e.g. Floats default to 15 digits of precision, so
would be tolerance=10**-15).
With full=True, a more extensive search is performed (this is useful to nd simpler numbers when the tolerance is set low).
See Also:
sympy.core.function.nfloat (page 181)
Examples
>>> from sympy import nsimplify, sqrt, GoldenRatio, exp, I, exp, pi
>>> nsimplify(4/(1+sqrt(5)), [GoldenRatio])
-2 + 2*GoldenRatio
>>> nsimplify((1/(exp(3*pi*I/5)+1)))
1/2 - I*sqrt(sqrt(5)/10 + 1/4)
>>> nsimplify(I**I, [pi])
exp(-pi/2)
>>> nsimplify(pi, tolerance=0.01)
22/7
5.24.16 collect sqrt
sympy.simplify.simplify.collect sqrt(expr, evaluate=None)
Return expr with terms having common square roots collected together. If evaluate
is False a count indicating the number of sqrt-containing terms will be returned and, if
non-zero, the terms of the Add will be returned, else the expression itself will be returned
as a single term. If evaluate is True, the expression with any collected terms will be
returned.
Note: since I = sqrt(-1), it is collected, too.
See Also:
collect (page 1329), collect const (page 1341), rcollect (page 1331)
1340
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import sqrt
>>> from sympy.simplify.simplify import collect_sqrt
>>> from sympy.abc import a, b
>>> r2, r3, r5 = [sqrt(i) for i in [2, 3, 5]]
>>> collect_sqrt(a*r2 + b*r2)
sqrt(2)*(a + b)
>>> collect_sqrt(a*r2 + b*r2 + a*r3 + b*r3)
sqrt(2)*(a + b) + sqrt(3)*(a + b)
>>> collect_sqrt(a*r2 + b*r2 + a*r3 + b*r5)
sqrt(3)*a + sqrt(5)*b + sqrt(2)*(a + b)
If evaluate is False then the arguments will be sorted and returned as a list and a count
of the number of sqrt-containing terms will be returned:
>>> collect_sqrt(a*r2 + b*r2 + a*r3 + b*r5, evaluate=False)
((sqrt(3)*a, sqrt(5)*b, sqrt(2)*(a + b)), 3)
>>> collect_sqrt(a*sqrt(2) + b, evaluate=False)
((b, sqrt(2)*a), 1)
>>> collect_sqrt(a + b, evaluate=False)
((a + b,), 0)
5.24.17 collect const
sympy.simplify.simplify.collect const(expr, *vars, **kwargs)
A non-greedy collection of terms with similar number coecients in an Add expr. If
vars is given then only those constants will be targeted. Although any Number can also
be targeted, if this is not desired set Numbers=False and no Float or Rational will be
collected.
See Also:
collect (page 1329), collect sqrt (page 1340), rcollect (page 1331)
Examples
>>> from sympy import sqrt
>>> from sympy.abc import a, s, x, y, z
>>> from sympy.simplify.simplify import collect_const
>>> collect_const(sqrt(3) + sqrt(3)*(1 + sqrt(2)))
sqrt(3)*(sqrt(2) + 2)
>>> collect_const(sqrt(3)*s + sqrt(7)*s + sqrt(3) + sqrt(7))
(sqrt(3) + sqrt(7))*(s + 1)
>>> s = sqrt(2) + 2
>>> collect_const(sqrt(3)*s + sqrt(3) + sqrt(7)*s + sqrt(7))
(sqrt(2) + 3)*(sqrt(3) + sqrt(7))
>>> collect_const(sqrt(3)*s + sqrt(3) + sqrt(7)*s + sqrt(7), sqrt(3))
sqrt(7) + sqrt(3)*(sqrt(2) + 3) + sqrt(7)*(sqrt(2) + 2)
The collection is sign-sensitive, giving higher precedence to the unsigned values:
5.24. Simplify
1341
SymPy Documentation, Release 0.7.6
>>> collect_const(x - y - z)
x - (y + z)
>>> collect_const(-y - z)
-(y + z)
>>> collect_const(2*x - 2*y - 2*z, 2)
2*(x - y - z)
>>> collect_const(2*x - 2*y - 2*z, -2)
2*x - 2*(y + z)
5.24.18 posify
sympy.simplify.simplify.posify(eq)
Return eq (with generic symbols made positive) and a restore dictionary.
Any symbol that has positive=None will be replaced with a positive dummy symbol having the same name. This replacement will allow more symbolic processing of expressions, especially those involving powers and logarithms.
A dictionary that can be sent to subs to restore eq to its original symbols is also returned.
>>>
>>>
>>>
(_x
from sympy import posify, Symbol, log
from sympy.abc import x
posify(x + Symbol(p, positive=True) + Symbol(n, negative=True))
+ n + p, {_x: x})
>> log(1/x).expand() # should be log(1/x) but it comes back as -log(x) log(1/x)
>>> log(posify(1/x)[0]).expand() # take [0] and ignore replacements
-log(_x)
>>> eq, rep = posify(1/x)
>>> log(eq).expand().subs(rep)
-log(x)
>>> posify([x, 1 + x])
([_x, _x + 1], {_x: x})
5.24.19 powdenest
sympy.simplify.simplify.powdenest(eq, force=False, polar=False)
Collect exponents on powers as assumptions allow.
Given (bb**be)**e, this can be simplied as follows:
if bb is positive, or
e is an integer, or
|be| < 1 then this simplies to bb**(be*e)
Given a product of powers raised to a power, (bb1**be1 * bb2**be2...)**e, simplication can be done as follows:
if e is positive, the gcd of all bei can be joined with e;
all non-negative bb can be separated from those that are negative and their gcd can
be joined with e; autosimplication already handles this separation.
integer factors from powers that have integers in the denominator of the exponent
can be removed from any term and the gcd of such integers can be joined with e
1342
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Setting force to True will make symbols that are not explicitly negative behave as though
they are positive, resulting in more denesting.
Setting polar to True will do simplications on the Riemann surface of the logarithm,
also resulting in more denestings.
When there are sums of logs in exp() then a product of powers may be obtained e.g.
exp(3*(log(a) + 2*log(b))) - > a**3*b**6.
Examples
>>> from sympy.abc import a, b, x, y, z
>>> from sympy import Symbol, exp, log, sqrt, symbols, powdenest
>>> powdenest((x**(2*a/3))**(3*x))
(x**(2*a/3))**(3*x)
>>> powdenest(exp(3*x*log(2)))
2**(3*x)
Assumptions may prevent expansion:
>>> powdenest(sqrt(x**2))
sqrt(x**2)
>>> p = symbols(p, positive=True)
>>> powdenest(sqrt(p**2))
p
No other expansion is done.
>>> i, j = symbols(i,j, integer=True)
>>> powdenest((x**x)**(i + j)) # -X-> (x**x)**i*(x**x)**j
x**(x*(i + j))
But exp() will be denested by moving all non-log terms outside of the function; this may
result in the collapsing of the exp to a power with a dierent base:
>>> powdenest(exp(3*y*log(x)))
x**(3*y)
>>> powdenest(exp(y*(log(a) + log(b))))
(a*b)**y
>>> powdenest(exp(3*(log(a) + log(b))))
a**3*b**3
If assumptions allow, symbols can also be moved to the outermost exponent:
>>> i = Symbol(i, integer=True)
>>> powdenest(((x**(2*i))**(3*y))**x)
((x**(2*i))**(3*y))**x
>>> powdenest(((x**(2*i))**(3*y))**x, force=True)
x**(6*i*x*y)
>>> powdenest(((x**(2*a/3))**(3*y/i))**x)
((x**(2*a/3))**(3*y/i))**x
>>> powdenest((x**(2*i)*y**(4*i))**z, force=True)
(x*y**2)**(2*i*z)
5.24. Simplify
1343
SymPy Documentation, Release 0.7.6
>>> n = Symbol(n, negative=True)
>>> powdenest((x**i)**y, force=True)
x**(i*y)
>>> powdenest((n**i)**x, force=True)
(n**i)**x
5.24.20 logcombine
sympy.simplify.simplify.logcombine(expr, force=False)
Takes logarithms and combines them using the following rules:
log(x) + log(y) == log(x*y) if both are not negative
a*log(x) == log(x**a) if x is positive and a is real
If force is True then the assumptions above will be assumed to hold if there is no assumption already in place on a quantity. For example, if a is imaginary or the argument
negative, force will not perform a combination but if a is a symbol with no assumptions
the change will take place.
See Also:
posify (page 1342) replace all symbols with symbols having positive assumptions
Examples
>>> from sympy import Symbol, symbols, log, logcombine, I
>>> from sympy.abc import a, x, y, z
>>> logcombine(a*log(x) + log(y) - log(z))
a*log(x) + log(y) - log(z)
>>> logcombine(a*log(x) + log(y) - log(z), force=True)
log(x**a*y/z)
>>> x,y,z = symbols(x,y,z, positive=True)
>>> a = Symbol(a, real=True)
>>> logcombine(a*log(x) + log(y) - log(z))
log(x**a*y/z)
The transformation is limited to factors and/or terms that contain logs, so the result
depends on the initial state of expansion:
>>> eq = (2 + 3*I)*log(x)
>>> logcombine(eq, force=True) == eq
True
>>> logcombine(eq.expand(), force=True)
log(x**2) + I*log(x**3)
5.24.21 Square Root Denest
sqrtdenest
sympy.simplify.sqrtdenest.sqrtdenest(expr, max iter=3)
Denests sqrts in an expression that contain other square roots if possible, otherwise
returns the expr unchanged. This is based on the algorithms of [1].
1344
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
See Also:
sympy.solvers.solvers.unrad
References
[1] http://researcher.watson.ibm.com/researcher/les/us-fagin/symb85.pdf
[2] D. J. Jerey and A. D. Rich, Symplifying Square Roots of Square Roots by Denesting
(available at http://www.cybertester.com/data/denest.pdf)
Examples
>>> from sympy.simplify.sqrtdenest import sqrtdenest
>>> from sympy import sqrt
>>> sqrtdenest(sqrt(5 + 2 * sqrt(6)))
sqrt(2) + sqrt(3)
5.24.22 Common Subexpresion Elimination
cse
sympy.simplify.cse main.cse(exprs, symbols=None, optimizations=None, postprocess=None, order=canonical)
Perform common subexpression elimination on an expression.
Parameters exprs : list of sympy expressions, or a single sympy expression
The expressions to reduce.
symbols : innite iterator yielding unique Symbols
The symbols used to label the common subexpressions which are
pulled out. The numbered symbols generator is useful. The default
is a stream of symbols of the form x0, x1, etc. This must be an
innite iterator.
optimizations : list of (callable, callable) pairs
The (preprocessor, postprocessor) pairs of external optimization
functions. Optionally basic can be passed for a set of predened
basic optimizations. Such basic optimizations were used by default
in old implementation, however they can be really slow on larger expressions. Now, no pre or post optimizations are made by default.
postprocess : a function which accepts the two return values of cse and
returns the desired form of output from cse, e.g. if you want the
replacements reversed the function might be the following lambda:
lambda r, e: return reversed(r), e
order : string, none or canonical
The order by which Mul and Add arguments are processed. If set to
canonical, arguments will be canonically ordered. If set to none,
ordering will be faster but dependent on expressions hashes, thus
machine dependent and variable. For large expressions where speed
is a concern, use the setting order=none.
5.24. Simplify
1345
SymPy Documentation, Release 0.7.6
Returns replacements : list of (Symbol, expression) pairs
All of the common subexpressions that were replaced. Subexpressions earlier in this list might show up in subexpressions later in this
list.
reduced exprs : list of sympy expressions
The reduced expressions with all of the replacements above.
opt cse
sympy.simplify.cse main.opt cse(exprs, order=canonical)
Find optimization opportunities in Adds, Muls, Pows and negative coecient Muls
Parameters exprs : list of sympy expressions
The expressions to optimize.
order : string, none or canonical
The order by which Mul and Add arguments are processed. For large
expressions where speed is a concern, use the setting order=none.
Returns opt subs : dictionary of expression substitutions
The expression substitutions which can be useful to optimize CSE.
Examples
>>> from sympy.simplify.cse_main import opt_cse
>>> from sympy.abc import x
>>> opt_subs = opt_cse([x**-2])
>>> print(opt_subs)
{x**(-2): 1/(x**2)}
tree cse
sympy.simplify.cse main.tree cse(exprs,
symbols,
opt subs=None,
der=canonical)
Perform raw CSE on expression tree, taking opt subs into account.
or-
Parameters exprs : list of sympy expressions
The expressions to reduce.
symbols : innite iterator yielding unique Symbols
The symbols used to label the common subexpressions which are
pulled out.
opt subs : dictionary of expression substitutions
The expressions to be substituted before any CSE action is performed.
order : string, none or canonical
The order by which Mul and Add arguments are processed. For large
expressions where speed is a concern, use the setting order=none.
1346
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
5.24.23 Hypergeometric Function Expansion
hyperexpand
sympy.simplify.hyperexpand.hyperexpand(f, allow hyper=False, rewrite=default)
Expand hypergeometric functions. If allow hyper is True, allow partial simplication
(that is a result dierent from input, but still containing hypergeometric functions).
Examples
>>> from sympy.simplify.hyperexpand import hyperexpand
>>> from sympy.functions import hyper
>>> from sympy.abc import z
>>> hyperexpand(hyper([], [], z))
exp(z)
Non-hyperegeometric parts of the expression and hypergeometric expressions that are
not recognised are left unchanged:
>>> hyperexpand(1 + hyper([1, 1, 1], [], z))
hyper((1, 1, 1), (), z) + 1
5.24.24 Traversal Tools
use
sympy.simplify.traversaltools.use(expr, func, level=0, args=(), kwargs={})
Use func to transform expr at the given level.
Examples
>>> from sympy import use, expand
>>> from sympy.abc import x, y
>>> f = (x + y)**2*x + 1
>>> use(f, expand, level=2)
x*(x**2 + 2*x*y + y**2) + 1
>>> expand(f)
x**3 + 2*x**2*y + x*y**2 + 1
5.24.25 EPath Tools
EPath class
class sympy.simplify.epathtools.EPath
Manipulate expressions using paths.
EPath grammar in EBNF notation:
5.24. Simplify
1347
SymPy Documentation, Release 0.7.6
literal
number
type
attribute
all
slice
range
query
selector
path
::=
::=
::=
::=
::=
::=
::=
::=
::=
::=
/[A-Za-z ][A-Za-z 0-9]*/
/-?\d+/
literal
literal ?
*
[ number? (: number? (: number?)?)? ]
all | slice
(type | attribute) (| (type | attribute))*
range | query range?
/ selector (/ selector)*
See the docstring of the epath() function.
apply(expr, func, args=None, kwargs=None)
Modify parts of an expression selected by a path.
Examples
>>> from sympy.simplify.epathtools import EPath
>>> from sympy import sin, cos, E
>>> from sympy.abc import x, y, z, t
>>> path = EPath(/*/[0]/Symbol)
>>> expr = [((x, 1), 2), ((3, y), z)]
>>> path.apply(expr, lambda expr: expr**2)
[((x**2, 1), 2), ((3, y**2), z)]
>>> path = EPath(/*/*/Symbol)
>>> expr = t + sin(x + 1) + cos(x + y + E)
>>> path.apply(expr, lambda expr: 2*expr)
t + sin(2*x + 1) + cos(2*x + 2*y + E)
select(expr)
Retrieve parts of an expression selected by a path.
Examples
>>> from sympy.simplify.epathtools import EPath
>>> from sympy import sin, cos, E
>>> from sympy.abc import x, y, z, t
>>> path = EPath(/*/[0]/Symbol)
>>> expr = [((x, 1), 2), ((3, y), z)]
>>> path.select(expr)
[x, y]
>>> path = EPath(/*/*/Symbol)
>>> expr = t + sin(x + 1) + cos(x + y + E)
1348
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> path.select(expr)
[x, x, y]
epath
sympy.simplify.epathtools.epath(path, expr=None,
kwargs=None)
Manipulate parts of an expression selected by a path.
func=None,
args=None,
This function allows to manipulate large nested expressions in single line of code, utilizing techniques to those applied in XML processing standards (e.g. XPath).
If func is None, epath() (page 1349) retrieves elements selected by the path. Otherwise
it applies func to each matching element.
Note that it is more ecient to create an EPath object and use the select and apply
methods of that object, since this will compile the path string only once. This function
should only be used as a convenient shortcut for interactive use.
This is the supported syntax:
select all: /* Equivalent of for arg in args:.
select slice: /[0] or /[1:5] or /[1:5:2] Supports standard Pythons slice syntax.
select by type: /list or /list|tuple Emulates isinstance().
select by attribute: / iter ? Emulates hasattr().
Parameters path : str | EPath
A path as a string or a compiled EPath.
expr : Basic | iterable
An expression or a container of expressions.
func : callable (optional)
A callable that will be applied to matching parts.
args : tuple (optional)
Additional positional arguments to func.
kwargs : dict (optional)
Additional keyword arguments to func.
Examples
>>> from sympy.simplify.epathtools import epath
>>> from sympy import sin, cos, E
>>> from sympy.abc import x, y, z, t
>>> path = /*/[0]/Symbol
>>> expr = [((x, 1), 2), ((3, y), z)]
5.24. Simplify
1349
SymPy Documentation, Release 0.7.6
>>> epath(path, expr)
[x, y]
>>> epath(path, expr, lambda expr: expr**2)
[((x**2, 1), 2), ((3, y**2), z)]
>>> path = /*/*/Symbol
>>> expr = t + sin(x + 1) + cos(x + y + E)
>>>
[x,
>>>
t +
epath(path, expr)
x, y]
epath(path, expr, lambda expr: 2*expr)
sin(2*x + 1) + cos(2*x + 2*y + E)
5.25 Details on the Hypergeometric Function Expansion
Module
This page describes how the function hyperexpand() and related code work. For usage, see
the documentation of the symplify module.
5.25.1 Hypergeometric Function Expansion Algorithm
This section describes the algorithm used to expand hypergeometric functions. Most of it is
based on the papers [Roach1996] (page 1912) and [Roach1997] (page 1912).
Recall that the hypergeometric function is (initially) dened as
)
(
(a1 )n . . . (ap )n z n
a1 , . . . , ap
.
z
=
p Fq
b1 , . . . , bq
(b1 )n . . . (bq )n n!
n=0
It turns out that there are certain dierential operators that can change the ap and pq parameters by integers. If a sequence of such operators is known that converts the set of indices
a0r and b0s into ap and bq , then we shall say the pair ap , bq is reachable from a0r , b0s . Our general
strategy is thus as follows: given a set ap , bq of parameters, try to look up an origin a0r , b0s for
which we know an expression, and then apply the sequence of dierential operators to the
known expression to nd an expression for the Hypergeometric function we are interested
in.
Notation
In the following, the symbol a will always denote a numerator parameter and the symbol b
will always denote a denominator parameter. The subscripts p, q, r, s denote vectors of that
length, so e.g. ap denotes a vector of p numerator parameters. The subscripts i and j denote
running indices, so they should usually be used in conjuction with a for all i. E.g. ai < 4
for all i. Uppercase subscripts I and J denote a chosen, xed index. So for example aI > 0 is
true if the inequality holds for the one index I we are currently interested in.
1350
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Incrementing and decrementing indices
Suppose ai 6= 0. Set A(ai ) =
z d
ai dz
(
+ 1. It is then easy to show that A(ai )p Fq
ap
= p Fq
bq z
d
z
bj 1 dz + 1
ap +ei
bq z
and nd
where ei is the i-th unit vector. Similarly for bj 6= 1 we set B(bj ) =
)
( )
(
p
B(bj )p Fq abqp z = p Fq bqae
z . Thus we can increment upper and decrement lower indices
i
at will, as long as we dont go through zero. The A(ai ) and B(bj ) are called shift operators.
)
( )
(
a1 ...ap
ap
ap +1
d
It is also easy to show that dz
p Fq bq z = b1 ...bq p Fq bq +1 z , where ap +1 is the vector a1 +1, a2 +
1, . . . and similarly for bq + 1. Combining this[ with the shift operators, we arrive at] one(form
) of
q
p
a1 ...ap
ap
d
the Hypergeometric dierential equation: dz j=1 B(bj ) (b1 1)...(bq 1) i=1 A(ai ) p Fq bq z =
0. This holds if all shift operators are dened, i.e. if no ai = 0 and no bj = 1.
Clearing
denominators
multiplying)]through
[
) z we arrive at the following equation:
) and
( by
q ( d
p ( d
ap
d
z dz j=1 z dz + bj 1 z i=1 z dz + ai p Fq bq z = 0. Even though our derivation does not
show it, it can be checked that this equation holds whenever the p Fq is dened.
d
Notice that, under suitable conditions on aI , bJ , each of the operators A(ai ), B(bj ) and z dz
can be expressed in terms of A(aI ) or B(bJ ). Our
) aim is to write the Hypergeometric
( next
ap
dierential equation as follows: [XA(aI ) r]p Fq bq z = 0, for some operator X and some
)
(
( )
ap +eI
constant r to be determined. If r 6= 0, then we can write this as 1
= p Fq abqp z ,
r X p Fq
bq z
and so
1
r X
undoes the shifting of A(aI ), whence it will be called an inverse-shift operator.
d
Now A(aI ) exists if aI 6= 0, and then z dz
= aI A(aI ) a)I . Observe
also
that ))
all the operators
(
(
p ( d
p
d
d
aI A(aI ), so this
A(ai ), B(bj ) and z dz commute. We have i=1 z dz + ai =
i=1,i6=I z dz + ai
gives us (the rst half
half does not have such a nice expression. We nd
) of X. The other
q
q
d
d
z
z dz
+
b
1
=
(a
A(a
)
a
)
j
I
I
I
j=1
j=1 (aI A(aI ) aI + bj 1). Since the rst half had no
dz
q
constant term, we infer r = aI j=1 (bj 1 aI ).
This tells us under which conditions we can un-shift A(aI ), namely when aI 6= 0 and r 6= 0.
Substituting aI 1 for aI then tells us under what conditions we can decrement the index aI .
Doing a similar analysis for B(aJ ), we arrive at the following rules:
An index aI can be decremented if aI 6= 1 and aI 6= bj for all bj .
An index bJ can be incremented if bJ 6= 1 and bJ 6= ai for all ai .
Combined with the conditions (stated above) for the existence of shift operators, we have thus
established the rules of the game!
Reduction of Order
(
Notice that, quite trivially, if aI = bJ , we have p Fq
ap
bq z
(
=
p1 Fq1
with aI omitted, and similarly for bq . We call this reduction of order.
a
p
z
b
q
, where ap means ap
In fact, we can do even better. If aI bJ Z>0 , then it is easy to see that
polynomial in n. It is also easy to see that
nd:
d k n
) z
(z dz
(aI )n
(bJ )n
is actually a
k n
= n z . Combining these two remarks we
If aI bJ Z>0 , then there exists a polynomial
p(n) = p0 + p1 n + . . . )
(of degree aI bJ )
)2
( )
( ) (
(
a
(aI )n
ap
d
d
such that (bJ )n = p(n) and p Fq bq z = p0 + p1 z dz + p2 z dz + . . . p1 Fq1 bp z .
q
5.25. Details on the Hypergeometric Function Expansion Module
1351
SymPy Documentation, Release 0.7.6
Thus any set of parameters ap , bq is reachable from a set of parameters cr , ds where ci dj Z
implies ci < dj . Such a set of parameters cr , ds is called suitable. Our database of known
formulae should only contain suitable origins. The reasons are twofold: rstly, working from
suitable origins is easier, and secondly, a formula for a non-suitable origin can be deduced
from a lower order formula, and we should put this one into the database instead.
Moving Around in the Parameter Space
It remains to investigate the following question: suppose ap , bq and a0p , b0q are both suitable,
and also ai a0i Z, bj b0j Z. When is ap , bq reachable from a0p , b0q ? It is clear that we can
treat all parameters independently that are incongruent mod 1. So assume that ai and bj are
congruent to r mod 1, for all i and j. The same then follows for a0i and b0j .
If r 6= 0, then any such ap , bq is reachable from any a0p , b0q . To see this notice that there exist
constants c, c0 , congruent mod 1, such that ai < c < bj for all i and j, and similarly a0i < c0 < b0j .
If n = c c0 > 0 then we rst inverse-shift all the b0j n times up, and then similarly shift shift
up all the a0i n times. If n < 0 then we rst inverse-shift down the a0i and then shift down the
b0j . This reduces to the case c = c0 . But evidently we can now shift or inverse-shift around the
a0i arbitrarily so long as we keep them less than c, and similarly for the b0j so long as we keep
them bigger than c. Thus ap , bq is reachable from a0p , b0q .
If r = 0 then the problem is slightly more involved. WLOG no parameter is zero. We now
have one additional complication: no parameter can ever move through zero. Hence ap , bq
is reachable from a0p , b0q if and only if the number of ai < 0 equals the number of a0i < 0, and
similarly for the bi and b0i . But in a suitable set of parameters, all bj > 0! This is because the
Hypergeometric function is undened if one of the bj is a non-positive integer and all ai are
smaller than the bj . Hence the number of bj 0 is always zero.
We can thus associate to every suitable set of parameters ap , bq , where no ai = 0, the following
invariants:
For every r [0, 1) the number r of parameters ai r (mod 1), and similarly the number
r of parameters bi r (mod 1).
The number of integers ai with ai < 0.
The above reasoning shows that ap , bq is reachable from a0p , b0q if and only if the invariants
r , r , all agree. Thus in particular being reachable from is a symmetric relation on suitable
parameters without zeros.
Applying the Operators
If all goes well then for a given set of parameters we nd an origin in our database for which
we have a nice formula. We now have to apply (potentially) many dierential operators to it. If
we do this blindly then the result will be very messy. This is because with Hypergeometric type
functions, the derivative is usually expressed as a sum of two contiguous functions. Hence
if we compute N derivatives, then the answer will involve 2N contiguous functions! This is
clearly undesirable. In fact we know from the Hypergeometric dierential equation that we
need at most max(p, q + 1) contiguous functions to express all derivatives.
Hence instead of dierentiating blindly, we will work with a C(z)-module basis: for an origin
a0r , b0s we either store (for particularly pretty answers) or compute a set of N functions (typically N = max(r, s + 1)) with the property that the derivative of any of them is a C(z)-linear
combination of them. In formulae, we store a vector B of N functions, a matrix M and a vector
C (the latter two with entries in C(z)), with the following properties:
1352
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
r Fs
a0r
b0s z
= CB
d
z dz
B = M B.
Then we can compute as many derivatives as we want and we will always end up with C(z)linear combination of at most N special functions.
As hinted above, B, M and C can either all be stored (for particularly pretty answers) or
computed from a single p Fq formula.
Loose Ends
This describes the bulk of the hypergeometric function algorithm. There a few further tricks,
described in the hyperexpand.py source le. The extension to Meijer G-functions is also described there.
5.25.2 Meijer G-Functions of Finite Conuence
Slaters theorem essentially evaluates a G-function as a sum of residues. If all poles are
simple, the resulting series can be recognised as hypergeometric series. Thus a G-function
can be evaluated as a sum of Hypergeometric functions.
If the poles are not simple, the resulting series are not hypergeometric. This is known as
the conuent or logarithmic case (the latter because the resulting series tend to contain
logarithms). The answer depends in a complicated way on the multiplicities of various poles,
and there is no accepted notation for representing it (as far as I know). However if there are
only nitely many multiple poles, we can evaluate the G function as a sum of hypergeometric
functions, plus nitely many extra terms. I could not nd any good reference for this, which
is why I work it out here.
Recall the general setup. We dene
n
m
1
j=1 (1 aj + s)
j=1 (bj s)
G(z) =
z s ds,
2i L qj=m+1 (1 bj + s) pj=n+1 (aj s)
where L is a contour starting and ending at +, enclosing all of the poles of (bj s) for
j = 1, . . . , n once in the negative direction, and no other poles. Also the integral is assumed
absolutely convergent.
In what follows, for any complex numbers a, b, we write a b (mod 1) if and only if there
exists an integer k such that a b = k. Thus there are double poles i ai aj (mod 1) for some
i 6= j n.
We now assume that whenever bj ai (mod 1) for i m, j > n then bj < ai . This means that
no quotient of the relevant gamma functions is a polynomial, and can always be achieved by
reduction of order. Fix a complex number c such that {bi |bi c (mod 1), i m} is not empty.
Enumerate this set as b, b + k1 , . . . , b + ku , with ki non-negative integers. Enumerate similarly
{aj |aj c (mod 1), j > n} as b + l1 , . . . , b + lv . Then li > kj for all i, j. For nite conuence, we
need to assume v u for all such c.
Let c1 , . . . , cw be distinct (mod 1) and exhaust the congruence classes of the bi . I claim
G(z) =
(Fj (z) + Rj (z)),
j=1
5.25. Details on the Hypergeometric Function Expansion Module
1353
SymPy Documentation, Release 0.7.6
where Fj (z) is a hypergeometric function and Rj (z) is a nite sum, both to be specied later.
Indeed corresponding to every cj there is a sequence of poles, at mostly nitely many of them
multiple poles. This is where the j-th term comes from.
Hence x again c, enumerate the relevant bi as b, b + k1 , . . . , b + ku . We will look at the aj
corresponding to a + l1 , . . . , a + lu . The other ai are not treated specially. The corresponding
gamma functions have poles at (potentially) s = b + r for r = 0, 1, . . . . For r lu , pole of the
integrand is simple. We thus set
R(z) =
l
u 1
ress=r+b .
r=0
We nally need to investigate the other poles. Set r = lu + t, t 0. A computation shows
(ki lu t)
1
(1)i
(lu li + 1)t
=
=
,
(li lu t)
(ki lu t)li ki
(lu li + 1)i (lu ki + 1)t
where i = li ki .
Also
(bj lu b t) =
(bj lu b)
,
(1)t (lu + b + 1 bj )t
(1 aj + lu + b + t) = (1 aj + lu + b)(1 aj + lu + b)t
and
ress=b+lu +t (b s) =
Hence
ress=b+lu +t
(1)lu +t
(1)lu (1)t
=
.
(lu + t)!
lu ! (lu + 1)t
m
n
u
(1)i
(1)lu
j=1 (bj lu b)
j=1 (1 aj + lu + b)
p
=z
q
lu ! i=1 (lu ki + 1)i j=n+1 (aj lu b) j=m+1 (1 bj + lu + b)
n
p
u
t
t
(lu li + 1)t j=1 (1 aj + lu + b)t j=n+1 (1) (lu + b + 1 aj )t
t (1)
m
z
,
q
(lu + 1)t i=1 (lu ki + 1)t j=1 (1)t (lu + b + 1 bj )t j=m+1 (1 bj + lu + b)t
b+lu
where the means to omit the terms we treated specially.
We thus arrive at
F (z) = C p+1 Fq
)
1, (1 + lu li ), (1 + lu + b ai )
pmn
(1)
z
,
1 + lu , (1 + lu ki ), (1 + lu + b bi )
where C designates the factor in the residue independent of t. (This result can also be written
in slightly simpler form by converting all the lu etc back to a b , but doing so is going to
require more notation still and is not helpful for computation.)
1354
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
5.25.3 Extending The Hypergeometric Tables
Adding new formulae to the tables is straightforward.
At the top of the le
sympy/simplify/hyperexpand.py, there is a function called add formulae(). Nested in it
are dened two helpers, add(ap, bq, res) and addb(ap, bq, B, C, M), as well as dummys
a, b, c, and z.
The rst step in adding a new formula is by using add(ap, bq, res). This declares hyper(ap,
bq, z) == res. Here ap and bq may use the dummys a, b, and c as free symbols. For example
n
nz
the well-known formula 0 (a)
= (1 z)a is declared by the following line: add((-a, ),
n!
(), (1-z)**a).
From the information provided, the matrices B, C and M will be computed, and the
formula is now available when expanding hypergeometric functions.
Next the test
le sympy/simplify/tests/test hyperexpand.py should be run, in particular the test
test formulae(). This will test the newly added formula numerically. If it fails, there is
(presumably) a typo in what was entered.
Since all newly-added formulae are probably relatively complicated, chances are that the
automatically computed basis is rather suboptimal (there is no good way of testing this, other
than observing very messy output). In this case the matrices B, C and M should be computed
by hand. Then the helper addb can be used to declare a hypergeometric formula with handcomputed basis.
An example
Because this explanation so far might be very theoretical and dicult to understand, we walk
through an explicit example now. We take the Fresnel function C(z) which obeys the following
hypergeometric representation:
( 1
)
2 z4
4
C(z) = z 1 F2 1 5
.
16
2, 4
First we try to add this formula to the lookup table by using the (simpler) function add(ap,
bq, res). The rst two arguments are simply the lists containing the parameter sets of 1 F2 .
The res argument is a little bit more complicated. We only know C(z) in terms of 1 F2 (. . . |f (z))
with f a function of z, in our case
f (z) =
2 z4
.
16
What we need is a formula where the hypergeometric function has only z as argument
We introduce the new complex symbol w and search for a function g(w) such that
1 F2 (. . . |z).
f (g(w)) = w
holds. Then we can replace every z in C(z) by g(w). In the case of our example the function g
could look like
( )
1
i
2
w4 .
g(w) = exp
4
5.25. Details on the Hypergeometric Function Expansion Module
1355
SymPy Documentation, Release 0.7.6
We get these functions mainly by guessing and testing the result. Hence we proceed by
computing f (g(w)) (and simplifying naively)
f (g(w)) =
=
=
2 g(w)4
16
(
( ) 1 )4
2
4
g 2 exp i
4 w
4
2 24
16
( )4 1 4
exp i
w4
4
16
= exp (i) w
=w
and indeed get back w. (In case of branched functions we have to be aware of branch cuts. In
that case we take w to be a positive real number and check the formula. If what we have found
works for positive w, then just replace exp() inside any branched function by exp polar()
and what we get is right for all w.) Hence we can write the formula as
( 1 )
C(g(w)) = g(w) 1 F2 1 4 5 w .
2, 4
and trivially
(
1 F2
)
C
C(g(w))
=
1 5 w =
g(w)
2, 4
1
4
( ) 1)
4
exp i
4 w
( i ) 1
exp 4 w 4
which is exactly what is needed for the third paramenter, res, in add. Finally, the whole
function call to add this rule to the table looks like:
add([S(1)/4],
[S(1)/2, S(5)/4],
fresnelc(exp(pi*I/4)*root(z,4)*2/sqrt(pi)) / (exp(pi*I/4)*root(z,4)*2/sqrt(pi))
)
Using this rule we will nd that it works but the results are not really nice in terms of simplicity
and number of special function instances included. We can obtain much better results by
adding the formula to the lookup table in another way. For this we use the (more complicated)
function addb(ap, bq, B, C, M). The rst two arguments are again the lists containing the
parameter sets of 1 F2 . The remaining three are the matrices mentioned earlier on this page.
We know that the n = max (p, q + 1)-th derivative can be expressed as a linear combination of
lower order derivatives. The matrix B contains the basis {B0 , B1 , . . .} and is of shape n 1. The
best way to get Bi is to take the rst n = max(p, q + 1) derivatives of the expression for p Fq
and take out usefull pieces. In our case we nd that n = max (1, 2 + 1) = 3. For computing the
d
. The rst basis element B0 is set to the expression
derivatives, we have to use the operator z dz
for 1 F2 from above:
(
( ) 1)
) ( 2
exp z 4
exp
C
4
4
B0 =
1
2z 4
d
B0 . For this we can directly use SymPy!
Next we compute z dz
1356
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> from sympy import Symbol, sqrt, exp, I, pi, fresnelc, root, diff, expand
>>> z = Symbol(z)
>>> B0 = sqrt(pi)*exp(-I*pi/4)*fresnelc(2*root(z,4)*exp(I*pi/4)/sqrt(pi))/\
...
(2*root(z,4))
>>> z * diff(B0, z)
z*(cosh(2*sqrt(z))/(4*z) - sqrt(pi)*exp(-I*pi/4)*fresnelc(2*z**(1/4)*exp(I*pi/4)/sqrt(pi))/(8*z**(5/4
>>> expand(_)
cosh(2*sqrt(z))/4 - sqrt(pi)*exp(-I*pi/4)*fresnelc(2*z**(1/4)*exp(I*pi/4)/sqrt(pi))/(8*z**(1/4))
Formatting this result nicely we obtain
B10 =
( ) ( 2
( ) 1)
exp z 4
exp
C
4
4
1
2z
1
4
( )
1
cosh 2 z
4
Computing the second derivative we nd
>>> from sympy import (Symbol, cosh, sqrt, pi, exp, I, fresnelc, root,
...
diff, expand)
>>> z = Symbol(z)
>>> B1prime = cosh(2*sqrt(z))/4 - sqrt(pi)*exp(-I*pi/4)*\
...
fresnelc(2*root(z,4)*exp(I*pi/4)/sqrt(pi))/(8*root(z,4))
>>> z * diff(B1prime, z)
z*(-cosh(2*sqrt(z))/(16*z) + sinh(2*sqrt(z))/(4*sqrt(z)) + sqrt(pi)*exp(-I*pi/4)*fresnelc(2*z**(1/4)*
>>> expand(_)
sqrt(z)*sinh(2*sqrt(z))/4 - cosh(2*sqrt(z))/16 + sqrt(pi)*exp(-I*pi/4)*fresnelc(2*z**(1/4)*exp(I*pi/4
which can be printed as
B20 =
1
16
(
) ( 2
( ) 1)
exp z 4
exp
C
4
4
2z
1
4
( ) 1
( )
1
cosh 2 z + sinh 2 z z
16
4
We see the common pattern and can collect the pieces. Hence it makes sense to choose B1
and B2 as follows
(
1)
2 exp( )z 4
exp(
4 )C
4
B0
1
2z 4
B = B1 =
cosh (2
z)
B2
sinh (2 z) z
(This is in contrast to the basis B = (B0 , B10 , B20 ) that would have been computed automatically
if we used just add(ap, bq, res).)
Because it must hold that p Fq ( |z) = CB the entries of C are obviously
1
C = 0
0
d
Finally we have to compute the entries of the 3 3 matrix M such that z dz
B = M B holds. This
d
is easy. We already computed the rst part z dz B0 above. This gives us the rst row of M . For
the second row we have:
5.25. Details on the Hypergeometric Function Expansion Module
1357
SymPy Documentation, Release 0.7.6
>>> from sympy import Symbol, cosh, sqrt, diff
>>> z = Symbol(z)
>>> B1 = cosh(2*sqrt(z))
>>> z * diff(B1, z)
sqrt(z)*sinh(2*sqrt(z))
and for the third one
>>> from sympy import Symbol, sinh, sqrt, expand, diff
>>> z = Symbol(z)
>>> B2 = sinh(2*sqrt(z))*sqrt(z)
>>> expand(z * diff(B2, z))
sqrt(z)*sinh(2*sqrt(z))/2 + z*cosh(2*sqrt(z))
Now we have computed the entries of this matrix to be
1 1
4 4 0
0 1
M = 0
0
z 21
Note that the entries of C and M should typically be rational functions in z, with rational
coecients. This is all we need to do in order to add a new formula to the lookup table for
hyperexpand.
5.25.4 Implemented Hypergeometric Formulae
A vital part of the algorithm is a relatively large table of hypergeometric function representations. The following automatically generated list contains all the representations implemented in SymPy (of course many more are derived from them). These formulae are mostly
taken from [Luke1969] (page 1912) and [Prudnikov1990] (page 1912). They are all tested
numerically.
( )
z
0 F0 z = e
1 F0
2 F1
(
a, a
2a
2 F1
1
2 z
1358
= 22a1
(
)2a+1
z + 1 + 1
( )
1
1, 1
z = log (z + 1)
2
z
(1
2 F1
( )
a z = (z + 1)a
2 , 1 z
3
2
( )
1
= atanh z
z
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
)
(1
1
2 , 2 z
3
2
2 F1
2 F1
( )
1
= asin z
z
)
(
)2a 1 (
)2a
1 (
a, a + 21
z+1
+
z+1
1
z = 2
2
2
)
(
(
( ))
a, a
= cos 2a asin z
2 F1
1 z
2
(
2 F1
asin ( z)
1, 1
=
3 z
z z + 1
2
(1
2 F1
2 F1
3 F2
1
2 , 2 z
2K (z)
( 1 1 )
2E (z)
2 , 2
z =
1
)
( 1
( ) 2
2 z
1
2 , 1, 1
z
=
atanh z +
log (z + 1)
1
,
2
3
3
3z
2
) (
)
(
)
( 1
4 16
4
1
1
16
2 , 1, 1
z =
z + 1 +
log
z + 1 +
+
3 F2
2, 2
9 9z
3z
2
2
9z
( )
1
z = z b+1 (b 1) ez (b 1, z)
1 F1
b
)
)
(z ) (
1
a
a 12 a+ 12 z2
a+
z
e Ia 12
z =4
2a
2
2
(
1 F1
(
1 F1
)
)
(
)a (
a
a, zei
z = a zei
a+1
5.25. Details on the Hypergeometric Function Expansion Module
1359
SymPy Documentation, Release 0.7.6
1 F1
( 1 )
( )
2
= zi erf zi + ez
1 z
2
(
1 F2
) i (
(
)
(
))
( )
( )
e 4
2 4 z i
2 4 z i
1
4
4
i sinh 2 z S
e
+ cosh 2 z C
e
=
3 5 z
24z
4, 4
(
2 F2
(
)a
)
ai z1
( ) a zei
(
)
z
=
erf
zi
a, zei
3
2a 1
2a 1
2, a + 1
1
2, a
2 F2
( )
1
1, 1
z = ( log (z) + Ei (z))
2, 2
z
z
0 F1
( )
( )
= cosh 2 z
1 z
2
0 F1
)
( )
( ))
1 (
a
= 22a z 2 + 4 I2a1 4 4 z + J2a1 4 4 z (2a)
1 z
(
0 F3
1
2 , a, a
(
0 F3
) (
)2a+1
( i )
( i )
4
4
z = 2 ze i
2
4
I
2
2
ze
J
2 (2a)
2a1
2a1 2 2 ze 4
)
) (
)
)
(
(
( )
( )
( ) 2
1
1
1
a1 a+1
a 12 a+ 12 2
3
1
z
I
z
=
2
4
z
I
a
+
4
z
I
z
a
+
1
a 2
a 2
a 2
a 21 , 2a
2
2
2
(
1 F2
a, a + 12 , 2a
( )
( )
z = z 2b + 12 Ib1 2 z (b)
b
z = Ib+1 ( z) Ib1 ( z) (b + 1)
b, b + 2
sin (b)
(
1 F2
1
2
(
1 F2
1360
)
( )
1
= Shi 2 z
3 3 z
2 z
2, 2
1
2
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
(
1 F2
)
(
)
3 3 i
2 4 z i
4 S
4
e
z
=
e
3 7
3
4z 4
2, 4
(
1 F2
3
4
) i (
)
2 4 z i
e 4
4
C e
=
1 5 z
24z
2, 4
1
4
) ( )2a+1
( )
( )
z
a, a + 12
z =
I2ab
z Ib1
z (b)(2a b + 1)
2a, b, 2a b + 1
2
(
2 F3
)
( )
( ))
1(
1, 1
=
log 2 z + Chi 2 z
3 z
2, 2, 2
z
z
(
2 F3
)
a
a (z)
a
aez
1, 1, a
z
=
((a)
(a,
z))
+
(a
+
1)
(log
(z)
+
E
(z)
+
)
+
1
2
2, 2, a + 1
z (a2 2a + 1)
z (a2 2a + 1) z
(a 1)
(
3 F3
5.25.5 References
5.26 Stats
SymPy statistics module
Introduces a random variable type into the SymPy language.
Random variables may be declared using prebuilt functions such as Normal, Exponential,
Coin, Die, etc... or built with functions like FiniteRV.
Queries on random expressions can be made using the functions
Expression
P(condition)
E(expression)
variance(expression)
density(expression)
sample(expression)
where(condition)
Meaning
Probability
Expected value
Variance
Probability Density Function
Produce a realization
Where the condition is true
5.26.1 Examples
>>>
>>>
>>>
>>>
>>>
from sympy.stats import P, E, variance, Die, Normal
from sympy import Eq, simplify
X, Y = Die(X, 6), Die(Y, 6) # Define two six sided dice
Z = Normal(Z, 0, 1) # Declare a Normal random variable with mean 0, std 1
P(X>3) # Probability X is greater than 3
5.26. Stats
1361
SymPy Documentation, Release 0.7.6
1/2
>>> E(X+Y) # Expectation of the sum of two dice
7
>>> variance(X+Y) # Variance of the sum of two dice
35/6
>>> simplify(P(Z>1)) # Probability of Z being greater than 1
-erf(sqrt(2)/2)/2 + 1/2
5.26.2 Random Variable Types
Finite Types
sympy.stats.DiscreteUniform(name, items)
Create a Finite Random Variable representing a uniform distribution over the input set.
Returns a RandomSymbol.
Examples
>>> from sympy.stats import DiscreteUniform, density
>>> from sympy import symbols
>>> X = DiscreteUniform(X, symbols(a b c)) # equally likely over a, b, c
>>> density(X).dict
{a: 1/3, b: 1/3, c: 1/3}
>>> Y = DiscreteUniform(Y, list(range(5))) # distribution over a range
>>> density(Y).dict
{0: 1/5, 1: 1/5, 2: 1/5, 3: 1/5, 4: 1/5}
sympy.stats.Die(name, sides=6)
Create a Finite Random Variable representing a fair die.
Returns a RandomSymbol.
>>> from sympy.stats import Die, density
>>> D6 = Die(D6, 6) # Six sided Die
>>> density(D6).dict
{1: 1/6, 2: 1/6, 3: 1/6, 4: 1/6, 5: 1/6, 6: 1/6}
>>> D4 = Die(D4, 4) # Four sided Die
>>> density(D4).dict
{1: 1/4, 2: 1/4, 3: 1/4, 4: 1/4}
sympy.stats.Bernoulli(name, p, succ=1, fail=0)
Create a Finite Random Variable representing a Bernoulli process.
Returns a RandomSymbol
>>> from sympy.stats import Bernoulli, density
>>> from sympy import S
1362
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> X = Bernoulli(X, S(3)/4) # 1-0 Bernoulli variable, probability = 3/4
>>> density(X).dict
{0: 1/4, 1: 3/4}
>>> X = Bernoulli(X, S.Half, Heads, Tails) # A fair coin toss
>>> density(X).dict
{Heads: 1/2, Tails: 1/2}
sympy.stats.Coin(name, p=1/2)
Create a Finite Random Variable representing a Coin toss.
Probability p is the chance of gettings Heads. Half by default
Returns a RandomSymbol.
>>> from sympy.stats import Coin, density
>>> from sympy import Rational
>>> C = Coin(C) # A fair coin toss
>>> density(C).dict
{H: 1/2, T: 1/2}
>>> C2 = Coin(C2, Rational(3, 5)) # An unfair coin
>>> density(C2).dict
{H: 3/5, T: 2/5}
sympy.stats.Binomial(name, n, p, succ=1, fail=0)
Create a Finite Random Variable representing a binomial distribution.
Returns a RandomSymbol.
Examples
>>> from sympy.stats import Binomial, density
>>> from sympy import S
>>> X = Binomial(X, 4, S.Half) # Four coin flips
>>> density(X).dict
{0: 1/16, 1: 1/4, 2: 3/8, 3: 1/4, 4: 1/16}
sympy.stats.Hypergeometric(name, N, m, n)
Create a Finite Random Variable representing a hypergeometric distribution.
Returns a RandomSymbol.
Examples
>>> from sympy.stats import Hypergeometric, density
>>> from sympy import S
>>> X = Hypergeometric(X, 10, 5, 3) # 10 marbles, 5 white (success), 3 draws
>>> density(X).dict
{0: 1/12, 1: 5/12, 2: 5/12, 3: 1/12}
sympy.stats.FiniteRV(name, density)
Create a Finite Random Variable given a dict representing the density.
5.26. Stats
1363
SymPy Documentation, Release 0.7.6
Returns a RandomSymbol.
>>> from sympy.stats import FiniteRV, P, E
>>> density = {0: .1, 1: .2, 2: .3, 3: .4}
>>> X = FiniteRV(X, density)
>>> E(X)
2.00000000000000
>>> P(X>=2)
0.700000000000000
Discrete Types
sympy.stats.Geometric(name, p)
Create a discrete random variable with a Geometric distribution.
The density of the Geometric distribution is given by
f (k) := p(1 p)k1
Parameters p: A probability between 0 and 1 :
Returns A RandomSymbol. :
References
[1] http://en.wikipedia.org/wiki/Geometric distribution [2] http://mathworld.wolfram.com/GeometricDis
Examples
>>> from sympy.stats import Geometric, density, E, variance
>>> from sympy import Symbol, S
>>> p = S.One / 5
>>> z = Symbol(z)
>>> X = Geometric(x, p)
>>> density(X)(z)
(4/5)**(z - 1)/5
>>> E(X)
5
>>> variance(X)
20
sympy.stats.Poisson(name, lamda)
Create a discrete random variable with a Poisson distribution.
1364
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
The density of the Poisson distribution is given by
f (k) :=
k e
k!
Parameters lamda: Positive number, a rate :
Returns A RandomSymbol. :
References
[1] http://en.wikipedia.org/wiki/Poisson distribution [2] http://mathworld.wolfram.com/PoissonDistribu
Examples
>>> from sympy.stats import Poisson, density, E, variance
>>> from sympy import Symbol, simplify
>>> rate = Symbol(lambda, positive=True)
>>> z = Symbol(z)
>>> X = Poisson(x, rate)
>>> density(X)(z)
lambda**z*exp(-lambda)/factorial(z)
>>> E(X)
lambda
>>> simplify(variance(X))
lambda
Continuous Types
sympy.stats.Arcsin(name, a=0, b=1)
Create a Continuous Random Variable with an arcsin distribution.
The density of the arcsin distribution is given by
f (x) :=
(x a)(b x)
with x [a, b]. It must hold that < a < b < .
Parameters a : Real number, the left interval boundary
b : Real number, the right interval boundary
Returns A RandomSymbol. :
References
[R382] (page 1912)
5.26. Stats
1365
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.stats import Arcsin, density
>>> from sympy import Symbol, simplify
>>> a = Symbol(a, real=True)
>>> b = Symbol(b, real=True)
>>> z = Symbol(z)
>>> X = Arcsin(x, a, b)
>>> density(X)(z)
1/(pi*sqrt((-a + z)*(b - z)))
sympy.stats.Benini(name, alpha, beta, sigma)
Create a Continuous Random Variable with a Benini distribution.
The density of the Benini distribution is given by
(
)
2 x
x
2 log x
f (x) := e log log [ ]
+
x
x
This is a heavy-tailed distrubtion and is also known as the log-Rayleigh distribution.
Parameters alpha : Real number, > 0, a shape
beta : Real number, > 0, a shape
sigma : Real number, > 0, a scale
Returns A RandomSymbol. :
References
[R383] (page 1912), [R384] (page 1912)
Examples
>>> from sympy.stats import Benini, density
>>> from sympy import Symbol, simplify, pprint
>>>
>>>
>>>
>>>
alpha = Symbol(alpha, positive=True)
beta = Symbol(beta, positive=True)
sigma = Symbol(sigma, positive=True)
z = Symbol(z)
>>> X = Benini(x, alpha, beta, sigma)
>>> D = density(X)(z)
>>> pprint(D, use_unicode=False)
/
/ z \\
/ z \
|
2*beta*log|-----|| - alpha*log|-----| - beta*log
|alpha
\sigma/|
\sigma/
|----- + -----------------|*e
\ z
z
/
1366
2/ z \
|-----|
\sigma/
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
sympy.stats.Beta(name, alpha, beta)
Create a Continuous Random Variable with a Beta distribution.
The density of the Beta distribution is given by
x1 (1 x)1
B(, )
f (x) :=
with x [0, 1].
Parameters alpha : Real number, > 0, a shape
beta : Real number, > 0, a shape
Returns A RandomSymbol. :
References
[R385] (page 1912), [R386] (page 1912)
Examples
>>> from sympy.stats import Beta, density, E, variance
>>> from sympy import Symbol, simplify, pprint, expand_func
>>> alpha = Symbol(alpha, positive=True)
>>> beta = Symbol(beta, positive=True)
>>> z = Symbol(z)
>>> X = Beta(x, alpha, beta)
>>> D = density(X)(z)
>>> pprint(D, use_unicode=False)
alpha - 1
beta - 1
z
*(-z + 1)
--------------------------beta(alpha, beta)
>>> expand_func(simplify(E(X, meijerg=True)))
alpha/(alpha + beta)
>>> simplify(variance(X, meijerg=True))
alpha*beta/((alpha + beta)**2*(alpha + beta + 1))
sympy.stats.BetaPrime(name, alpha, beta)
Create a continuous random variable with a Beta prime distribution.
The density of the Beta prime distribution is given by
f (x) :=
x1 (1 + x)
B(, )
with x > 0.
5.26. Stats
1367
SymPy Documentation, Release 0.7.6
Parameters alpha : Real number, > 0, a shape
beta : Real number, > 0, a shape
Returns A RandomSymbol. :
References
[R387] (page 1913), [R388] (page 1913)
Examples
>>> from sympy.stats import BetaPrime, density
>>> from sympy import Symbol, pprint
>>> alpha = Symbol(alpha, positive=True)
>>> beta = Symbol(beta, positive=True)
>>> z = Symbol(z)
>>> X = BetaPrime(x, alpha, beta)
>>> D = density(X)(z)
>>> pprint(D, use_unicode=False)
alpha - 1
-alpha - beta
z
*(z + 1)
------------------------------beta(alpha, beta)
sympy.stats.Cauchy(name, x0, gamma)
Create a continuous random variable with a Cauchy distribution.
The density of the Cauchy distribution is given by
(
)
1
x x0
1
f (x) := arctan
+
Parameters x0 : Real number, the location
gamma : Real number, > 0, the scale
Returns A RandomSymbol. :
References
[R389] (page 1913), [R390] (page 1913)
Examples
>>> from sympy.stats import Cauchy, density
>>> from sympy import Symbol
1368
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> x0 = Symbol(x0)
>>> gamma = Symbol(gamma, positive=True)
>>> z = Symbol(z)
>>> X = Cauchy(x, x0, gamma)
>>> density(X)(z)
1/(pi*gamma*(1 + (-x0 + z)**2/gamma**2))
sympy.stats.Chi(name, k)
Create a continuous random variable with a Chi distribution.
The density of the Chi distribution is given by
f (x) :=
21k/2 xk1 ex
(k/2)
/2
with x 0.
Parameters k : A positive Integer, k > 0, the number of degrees of freedom
Returns A RandomSymbol. :
References
[R391] (page 1913), [R392] (page 1913)
Examples
>>> from sympy.stats import Chi, density, E, std
>>> from sympy import Symbol, simplify
>>> k = Symbol(k, integer=True)
>>> z = Symbol(z)
>>> X = Chi(x, k)
>>> density(X)(z)
2**(-k/2 + 1)*z**(k - 1)*exp(-z**2/2)/gamma(k/2)
sympy.stats.ChiNoncentral(name, k, l)
Create a continuous random variable with a non-central Chi distribution.
The density of the non-central Chi distribution is given by
e(x + )/2 xk
f (x) :=
Ik/21 (x)
(x)k/2
2
with x 0. Here, I (x) is the modied Bessel function of the rst kind (page 410).
Parameters k : A positive Integer, k > 0, the number of degrees of freedom
l : Shift parameter
Returns A RandomSymbol. :
5.26. Stats
1369
SymPy Documentation, Release 0.7.6
References
[R393] (page 1913)
Examples
>>> from sympy.stats import ChiNoncentral, density, E, std
>>> from sympy import Symbol, simplify
>>> k = Symbol(k, integer=True)
>>> l = Symbol(l)
>>> z = Symbol(z)
>>> X = ChiNoncentral(x, k, l)
>>> density(X)(z)
l*z**k*(l*z)**(-k/2)*exp(-l**2/2 - z**2/2)*besseli(k/2 - 1, l*z)
sympy.stats.ChiSquared(name, k)
Create a continuous random variable with a Chi-squared distribution.
The density of the Chi-squared distribution is given by
f (x) :=
x
k
1
( k ) x 2 1 e 2
2 2
k
2
with x 0.
Parameters k : A positive Integer, k > 0, the number of degrees of freedom
Returns A RandomSymbol. :
References
[R394] (page 1913), [R395] (page 1913)
Examples
>>> from sympy.stats import ChiSquared, density, E, variance
>>> from sympy import Symbol, simplify, combsimp, expand_func
>>> k = Symbol(k, integer=True, positive=True)
>>> z = Symbol(z)
>>> X = ChiSquared(x, k)
>>> density(X)(z)
2**(-k/2)*z**(k/2 - 1)*exp(-z/2)/gamma(k/2)
>>> combsimp(E(X))
k
1370
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> simplify(expand_func(variance(X)))
2*k
sympy.stats.Dagum(name, p, a, b)
Create a continuous random variable with a Dagum distribution.
The density of the Dagum distribution is given by
(
)
( x )ap
ap
b
f (x) :=
x (( x )a + 1)p+1
b
with x > 0.
Parameters p : Real number, p > 0, a shape
a : Real number, a > 0, a shape
b : Real number, b > 0, a scale
Returns A RandomSymbol. :
References
[R396] (page 1913)
Examples
>>> from sympy.stats import Dagum, density
>>> from sympy import Symbol, simplify
>>>
>>>
>>>
>>>
p
b
a
z
=
=
=
=
Symbol(p, positive=True)
Symbol(b, positive=True)
Symbol(a, positive=True)
Symbol(z)
>>> X = Dagum(x, p, a, b)
>>> density(X)(z)
a*p*(z/b)**(a*p)*((z/b)**a + 1)**(-p - 1)/z
sympy.stats.Erlang(name, k, l)
Create a continuous random variable with an Erlang distribution.
The density of the Erlang distribution is given by
f (x) :=
k xk1 ex
(k 1)!
with x [0, ].
Parameters k : Integer
l : Real number, > 0, the rate
Returns A RandomSymbol. :
5.26. Stats
1371
SymPy Documentation, Release 0.7.6
References
[R397] (page 1913), [R398] (page 1913)
Examples
>>> from sympy.stats import Erlang, density, cdf, E, variance
>>> from sympy import Symbol, simplify, pprint
>>> k = Symbol(k, integer=True, positive=True)
>>> l = Symbol(l, positive=True)
>>> z = Symbol(z)
>>> X = Erlang(x, k, l)
>>> D = density(X)(z)
>>> pprint(D, use_unicode=False)
k k - 1 -l*z
l *z
*e
--------------gamma(k)
>>> C = cdf(X, meijerg=True)(z)
>>> pprint(C, use_unicode=False)
/ k*lowergamma(k, 0)
k*lowergamma(k, l*z)
|- ------------------ + -------------------- for z >= 0
<
gamma(k + 1)
gamma(k + 1)
|
\
0
otherwise
>>> simplify(E(X))
k/l
>>> simplify(variance(X))
k/l**2
sympy.stats.Exponential(name, rate)
Create a continuous random variable with an Exponential distribution.
The density of the exponential distribution is given by
f (x) := exp(x)
with x > 0. Note that the expected value is 1/.
Parameters rate : A positive Real number, > 0, the rate (or inverse
scale/inverse mean)
Returns A RandomSymbol. :
References
[R399] (page 1913), [R400] (page 1913)
1372
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.stats import Exponential, density, cdf, E
>>> from sympy.stats import variance, std, skewness
>>> from sympy import Symbol
>>> l = Symbol(lambda, positive=True)
>>> z = Symbol(z)
>>> X = Exponential(x, l)
>>> density(X)(z)
lambda*exp(-lambda*z)
>>> cdf(X)(z)
Piecewise((1 - exp(-lambda*z), z >= 0), (0, True))
>>> E(X)
1/lambda
>>> variance(X)
lambda**(-2)
>>> skewness(X)
2
>>> X = Exponential(x, 10)
>>> density(X)(z)
10*exp(-10*z)
>>> E(X)
1/10
>>> std(X)
1/10
sympy.stats.FDistribution(name, d1, d2)
Create a continuous random variable with a F distribution.
The density of the F distribution is given by
f (x) :=
(d1 x)d1 d2 2
(d1 x+d2 )d1 +d2
xB
( d1
2
, d22
with x > 0.
Parameters d1 : d1 > 0 a parameter
d2 : d2 > 0 a parameter
Returns A RandomSymbol. :
5.26. Stats
1373
SymPy Documentation, Release 0.7.6
References
[R401] (page 1913), [R402] (page 1913)
Examples
>>> from sympy.stats import FDistribution, density
>>> from sympy import Symbol, simplify, pprint
>>> d1 = Symbol(d1, positive=True)
>>> d2 = Symbol(d2, positive=True)
>>> z = Symbol(z)
>>> X = FDistribution(x, d1, d2)
>>> D = density(X)(z)
>>> pprint(D, use_unicode=False)
d2
-______________________________
2
/
d1
-d1 - d2
d2 *\/ (d1*z) *(d1*z + d2)
-------------------------------------/d1 d2\
z*beta|--, --|
\2
2 /
sympy.stats.FisherZ(name, d1, d2)
Create a Continuous Random Variable with an Fishers Z distribution.
The density of the Fishers Z distribution is given by
d /2 d /2
f (x) :=
2d11 d22
ed 1 z
B(d1 /2, d2 /2) (d1 e2z + d2 )(d1 +d2 )/2
Parameters d1 : d1 > 0, degree of freedom
d2 : d2 > 0, degree of freedom
Returns A RandomSymbol. :
References
[R403] (page 1913), [R404] (page 1913)
Examples
>>> from sympy.stats import FisherZ, density
>>> from sympy import Symbol, simplify, pprint
>>> d1 = Symbol(d1, positive=True)
>>> d2 = Symbol(d2, positive=True)
>>> z = Symbol(z)
1374
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> X = FisherZ(x, d1, d2)
>>> D = density(X)(z)
>>> pprint(D, use_unicode=False)
d1
d2
d1
d2
- -- - ---2
2
2
2 /
2*z
\
d1*z
2*d1 *d2 *\d1*e
+ d2/
*e
----------------------------------------/d1 d2\
beta|--, --|
\2
2 /
sympy.stats.Frechet(name, a, s=1, m=0)
Create a continuous random variable with a Frechet distribution.
The density of the Frechet distribution is given by
f (x) :=
xm
s
)1
e(
xm
s )
with x m.
Parameters a : Real number, a (0, ) the shape
s : Real number, s (0, ) the scale
m : Real number, m (, ) the minimum
Returns A RandomSymbol. :
References
[R405] (page 1913)
Examples
>>> from sympy.stats import Frechet, density, E, std
>>> from sympy import Symbol, simplify
>>>
>>>
>>>
>>>
a
s
m
z
=
=
=
=
Symbol(a, positive=True)
Symbol(s, positive=True)
Symbol(m, real=True)
Symbol(z)
>>> X = Frechet(x, a, s, m)
>>> density(X)(z)
a*((-m + z)/s)**(-a - 1)*exp(-((-m + z)/s)**(-a))/s
sympy.stats.Gamma(name, k, theta)
Create a continuous random variable with a Gamma distribution.
5.26. Stats
1375
SymPy Documentation, Release 0.7.6
The density of the Gamma distribution is given by
f (x) :=
x
1
xk1 e
(k)k
with x [0, 1].
Parameters k : Real number, k > 0, a shape
theta : Real number, > 0, a scale
Returns A RandomSymbol. :
References
[R406] (page 1913), [R407] (page 1913)
Examples
>>> from sympy.stats import Gamma, density, cdf, E, variance
>>> from sympy import Symbol, pprint, simplify
>>> k = Symbol(k, positive=True)
>>> theta = Symbol(theta, positive=True)
>>> z = Symbol(z)
>>> X = Gamma(x, k, theta)
>>> D = density(X)(z)
>>> pprint(D, use_unicode=False)
-z
-----k k - 1 theta
theta *z
*e
--------------------gamma(k)
>>> C = cdf(X, meijerg=True)(z)
>>> pprint(C, use_unicode=False)
/
/
z \
|
k*lowergamma|k, -----|
| k*lowergamma(k, 0)
\
theta/
<- ------------------ + ---------------------|
gamma(k + 1)
gamma(k + 1)
|
\
0
for z >= 0
otherwise
>>> E(X)
theta*gamma(k + 1)/gamma(k)
>>> V = simplify(variance(X))
>>> pprint(V, use_unicode=False)
2
k*theta
1376
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
sympy.stats.GammaInverse(name, a, b)
Create a continuous random variable with an inverse Gamma distribution.
The density of the inverse Gamma distribution is given by
(
)
1
f (x) :=
x
exp
()
x
with x > 0.
Parameters a : Real number, a > 0 a shape
b : Real number, b > 0 a scale
Returns A RandomSymbol. :
References
[R408] (page 1913)
Examples
>>> from sympy.stats import GammaInverse, density, cdf, E, variance
>>> from sympy import Symbol, pprint
>>> a = Symbol(a, positive=True)
>>> b = Symbol(b, positive=True)
>>> z = Symbol(z)
>>> X = GammaInverse(x, a, b)
>>> D = density(X)(z)
>>> pprint(D, use_unicode=False)
-b
--a -a - 1
z
b *z
*e
--------------gamma(a)
sympy.stats.Kumaraswamy(name, a, b)
Create a Continuous Random Variable with a Kumaraswamy distribution.
The density of the Kumaraswamy distribution is given by
f (x) := abxa1 (1 xa )b1
with x [0, 1].
Parameters a : Real number, a > 0 a shape
b : Real number, b > 0 a shape
Returns A RandomSymbol. :
5.26. Stats
1377
SymPy Documentation, Release 0.7.6
References
[R409] (page 1913)
Examples
>>> from sympy.stats import Kumaraswamy, density, E, variance
>>> from sympy import Symbol, simplify, pprint
>>> a = Symbol(a, positive=True)
>>> b = Symbol(b, positive=True)
>>> z = Symbol(z)
>>> X = Kumaraswamy(x, a, b)
>>> D = density(X)(z)
>>> pprint(D, use_unicode=False)
b - 1
a - 1 /
a
\
a*b*z
*\- z + 1/
sympy.stats.Laplace(name, mu, b)
Create a continuous random variable with a Laplace distribution.
The density of the Laplace distribution is given by
(
)
1
|x |
f (x) :=
exp
2b
b
Parameters mu : Real number, the location (mean)
b : Real number, b > 0, a scale
Returns A RandomSymbol. :
References
[R410] (page 1913), [R411] (page 1913)
Examples
>>> from sympy.stats import Laplace, density
>>> from sympy import Symbol
>>> mu = Symbol(mu)
>>> b = Symbol(b, positive=True)
>>> z = Symbol(z)
>>> X = Laplace(x, mu, b)
>>> density(X)(z)
exp(-Abs(mu - z)/b)/(2*b)
1378
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
sympy.stats.Logistic(name, mu, s)
Create a continuous random variable with a logistic distribution.
The density of the logistic distribution is given by
f (x) :=
e(x)/s
(
)2
s 1 + e(x)/s
Parameters mu : Real number, the location (mean)
s : Real number, s > 0 a scale
Returns A RandomSymbol. :
References
[R412] (page 1913), [R413] (page 1913)
Examples
>>> from sympy.stats import Logistic, density
>>> from sympy import Symbol
>>> mu = Symbol(mu, real=True)
>>> s = Symbol(s, positive=True)
>>> z = Symbol(z)
>>> X = Logistic(x, mu, s)
>>> density(X)(z)
exp((mu - z)/s)/(s*(exp((mu - z)/s) + 1)**2)
sympy.stats.LogNormal(name, mean, std)
Create a continuous random variable with a log-normal distribution.
The density of the log-normal distribution is given by
(ln x)2
1
f (x) :=
e 22
x 2 2
with x 0.
Parameters mu : Real number, the log-scale
sigma : Real number, 2 > 0 a shape
Returns A RandomSymbol. :
References
[R414] (page 1913), [R415] (page 1913)
5.26. Stats
1379
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.stats import LogNormal, density
>>> from sympy import Symbol, simplify, pprint
>>> mu = Symbol(mu, real=True)
>>> sigma = Symbol(sigma, positive=True)
>>> z = Symbol(z)
>>> X = LogNormal(x, mu, sigma)
>>> D = density(X)(z)
>>> pprint(D, use_unicode=False)
2
-(-mu + log(z))
----------------2
___
2*sigma
\/ 2 *e
-----------------------____
2*\/ pi *sigma*z
>>> X = LogNormal(x, 0, 1) # Mean 0, standard deviation 1
>>> density(X)(z)
sqrt(2)*exp(-log(z)**2/2)/(2*sqrt(pi)*z)
sympy.stats.Maxwell(name, a)
Create a continuous random variable with a Maxwell distribution.
The density of the Maxwell distribution is given by
2
2
2 x2 ex /(2a )
f (x) :=
a3
with x 0.
Parameters a : Real number, a > 0
Returns A RandomSymbol. :
References
[R416] (page 1913), [R417] (page 1913)
Examples
>>> from sympy.stats import Maxwell, density, E, variance
>>> from sympy import Symbol, simplify
>>> a = Symbol(a, positive=True)
>>> z = Symbol(z)
1380
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> X = Maxwell(x, a)
>>> density(X)(z)
sqrt(2)*z**2*exp(-z**2/(2*a**2))/(sqrt(pi)*a**3)
>>> E(X)
2*sqrt(2)*a/sqrt(pi)
>>> simplify(variance(X))
a**2*(-8 + 3*pi)/pi
sympy.stats.Nakagami(name, mu, omega)
Create a continuous random variable with a Nakagami distribution.
The density of the Nakagami distribution is given by
( )
2
21
f (x) :=
x
exp
x2
()
with x > 0.
Parameters mu : Real number,
1
2
a shape
omega : Real number, > 0, the spread
Returns A RandomSymbol. :
References
[R418] (page 1913)
Examples
>>> from sympy.stats import Nakagami, density, E, variance
>>> from sympy import Symbol, simplify, pprint
>>> mu = Symbol(mu, positive=True)
>>> omega = Symbol(omega, positive=True)
>>> z = Symbol(z)
>>> X = Nakagami(x, mu, omega)
>>> D = density(X)(z)
>>> pprint(D, use_unicode=False)
2
-mu*z
------mu
-mu 2*mu - 1 omega
2*mu *omega
*z
*e
---------------------------------gamma(mu)
>>> simplify(E(X, meijerg=True))
sqrt(mu)*sqrt(omega)*gamma(mu + 1/2)/gamma(mu + 1)
5.26. Stats
1381
SymPy Documentation, Release 0.7.6
>>> V = simplify(variance(X, meijerg=True))
>>> pprint(V, use_unicode=False)
2
omega*gamma (mu + 1/2)
omega - ----------------------gamma(mu)*gamma(mu + 1)
sympy.stats.Normal(name, mean, std)
Create a continuous random variable with a Normal distribution.
The density of the Normal distribution is given by
f (x) :=
(x)2
1
e 22
2
Parameters mu : Real number, the mean
sigma : Real number, 2 > 0 the variance
Returns A RandomSymbol. :
References
[R419] (page 1913), [R420] (page 1913)
Examples
>>> from sympy.stats import Normal, density, E, std, cdf, skewness
>>> from sympy import Symbol, simplify, pprint, factor, together, factor_terms
>>> mu = Symbol(mu)
>>> sigma = Symbol(sigma, positive=True)
>>> z = Symbol(z)
>>> X = Normal(x, mu, sigma)
>>> density(X)(z)
sqrt(2)*exp(-(-mu + z)**2/(2*sigma**2))/(2*sqrt(pi)*sigma)
>>> C = simplify(cdf(X))(z) # it needs a little more help...
>>> pprint(C, use_unicode=False)
/ ___
\
|\/ 2 *(-mu + z)|
erf|---------------|
\
2*sigma
/
1
-------------------- + 2
2
>>> simplify(skewness(X))
0
>>> X = Normal(x, 0, 1) # Mean 0, standard deviation 1
>>> density(X)(z)
sqrt(2)*exp(-z**2/2)/(2*sqrt(pi))
1382
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> E(2*X + 1)
1
>>> simplify(std(2*X + 1))
2
sympy.stats.Pareto(name, xm, alpha)
Create a continuous random variable with the Pareto distribution.
The density of the Pareto distribution is given by
f (x) :=
x
m
x+1
with x [xm , ].
Parameters xm : Real number, xm > 0, a scale
alpha : Real number, > 0, a shape
Returns A RandomSymbol. :
References
[R421] (page 1913), [R422] (page 1913)
Examples
>>> from sympy.stats import Pareto, density
>>> from sympy import Symbol
>>> xm = Symbol(xm, positive=True)
>>> beta = Symbol(beta, positive=True)
>>> z = Symbol(z)
>>> X = Pareto(x, xm, beta)
>>> density(X)(z)
beta*xm**beta*z**(-beta - 1)
sympy.stats.QuadraticU(name, a, b)
Create a Continuous Random Variable with a U-quadratic distribution.
The density of the U-quadratic distribution is given by
f (x) := (x )2
with x [a, b].
Parameters a : Real number
b : Real number, a < b
Returns A RandomSymbol. :
5.26. Stats
1383
SymPy Documentation, Release 0.7.6
References
[R423] (page 1913)
Examples
>>> from sympy.stats import QuadraticU, density, E, variance
>>> from sympy import Symbol, simplify, factor, pprint
>>> a = Symbol(a, real=True)
>>> b = Symbol(b, real=True)
>>> z = Symbol(z)
>>> X = QuadraticU(x, a, b)
>>> D = density(X)(z)
>>> pprint(D, use_unicode=False)
/
2
|
/ a
b
\
|12*|- - - - + z|
|
\ 2
2
/
<----------------- for And(a <= z, z <= b)
|
3
|
(-a + b)
|
\
0
otherwise
sympy.stats.RaisedCosine(name, mu, s)
Create a Continuous Random Variable with a raised cosine distribution.
The density of the raised cosine distribution is given by
(
(
))
1
x
f (x) :=
1 + cos
2s
s
with x [ s, + s].
Parameters mu : Real number
s : Real number, s > 0
Returns A RandomSymbol. :
References
[R424] (page 1913)
Examples
>>> from sympy.stats import RaisedCosine, density, E, variance
>>> from sympy import Symbol, simplify, pprint
1384
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> mu = Symbol(mu, real=True)
>>> s = Symbol(s, positive=True)
>>> z = Symbol(z)
>>> X = RaisedCosine(x, mu, s)
>>> D = density(X)(z)
>>> pprint(D, use_unicode=False)
/
/pi*(-mu + z)\
|cos|------------| + 1
|
\
s
/
<--------------------- for And(z <= mu + s, mu - s <= z)
|
2*s
|
\
0
otherwise
sympy.stats.Rayleigh(name, sigma)
Create a continuous random variable with a Rayleigh distribution.
The density of the Rayleigh distribution is given by
f (x) :=
x x2 /22
e
2
with x > 0.
Parameters sigma : Real number, > 0
Returns A RandomSymbol. :
References
[R425] (page 1914), [R426] (page 1914)
Examples
>>> from sympy.stats import Rayleigh, density, E, variance
>>> from sympy import Symbol, simplify
>>> sigma = Symbol(sigma, positive=True)
>>> z = Symbol(z)
>>> X = Rayleigh(x, sigma)
>>> density(X)(z)
z*exp(-z**2/(2*sigma**2))/sigma**2
>>> E(X)
sqrt(2)*sqrt(pi)*sigma/2
>>> variance(X)
-pi*sigma**2/2 + 2*sigma**2
5.26. Stats
1385
SymPy Documentation, Release 0.7.6
sympy.stats.StudentT(name, nu)
Create a continuous random variable with a students t distribution.
The density of the students t distribution is given by
(
) (
) +1
2
+1
x2
2( )
f (x) :=
1+
Parameters nu : Real number, > 0, the degrees of freedom
Returns A RandomSymbol. :
References
[R427] (page 1914), [R428] (page 1914)
Examples
>>> from sympy.stats import StudentT, density, E, variance
>>> from sympy import Symbol, simplify, pprint
>>> nu = Symbol(nu, positive=True)
>>> z = Symbol(z)
>>> X = StudentT(x, nu)
>>> D = density(X)(z)
>>> pprint(D, use_unicode=False)
nu
1
- -- - 2
2
/
2\
|
z |
|1 + --|
\
nu/
-------------------____
/
nu\
\/ nu *beta|1/2, --|
\
2 /
sympy.stats.Triangular(name, a, b, c)
Create a continuous random variable with a triangular distribution.
The density of the triangular distribution is given by
0
for x < a,
2(xa)
(ba)(ca) for a x < c,
2
f (x) := ba
for x = c,
2(bx)
for c < x b,
(ba)(bc)
0
for b < x.
1386
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Parameters a : Real number, a (, )
b : Real number, a < b
c : Real number, a c b
Returns A RandomSymbol. :
References
[R429] (page 1914), [R430] (page 1914)
Examples
>>> from sympy.stats import Triangular, density, E
>>> from sympy import Symbol, pprint
>>>
>>>
>>>
>>>
a
b
c
z
=
=
=
=
Symbol(a)
Symbol(b)
Symbol(c)
Symbol(z)
>>> X = Triangular(x, a,b,c)
>>> pprint(density(X)(z), use_unicode=False)
/
-2*a + 2*z
|----------------- for And(a <= z, z < c)
|(-a + b)*(-a + c)
|
|
2
|
-----for z = c
<
-a + b
|
|
2*b - 2*z
|---------------for And(z <= b, c < z)
|(-a + b)*(b - c)
|
\
0
otherwise
sympy.stats.Uniform(name, left, right)
Create a continuous random variable with a uniform distribution.
The density of the uniform distribution is given by
{
1
for x [a, b]
f (x) := ba
0
otherwise
with x [a, b].
Parameters a : Real number, < a the left boundary
b : Real number, a < b < the right boundary
Returns A RandomSymbol. :
5.26. Stats
1387
SymPy Documentation, Release 0.7.6
References
[R431] (page 1914), [R432] (page 1914)
Examples
>>> from sympy.stats import Uniform, density, cdf, E, variance, skewness
>>> from sympy import Symbol, simplify
>>> a = Symbol(a, negative=True)
>>> b = Symbol(b, positive=True)
>>> z = Symbol(z)
>>> X = Uniform(x, a, b)
>>> density(X)(z)
Piecewise((1/(-a + b), And(a <= z, z <= b)), (0, True))
>>> cdf(X)(z)
-a/(-a + b) + z/(-a + b)
>>> simplify(E(X))
a/2 + b/2
>>> simplify(variance(X))
a**2/12 - a*b/6 + b**2/12
sympy.stats.UniformSum(name, n)
Create a continuous random variable with an Irwin-Hall distribution.
The probability distribution function depends on a single parameter n which is an integer.
The density of the Irwin-Hall distribution is given by
f (x) :=
( )
bxc
1
n
(1)k
(x k)n1
(n 1)!
k
k=0
Parameters n : A positive Integer, n > 0
Returns A RandomSymbol. :
References
[R433] (page 1914), [R434] (page 1914)
Examples
>>> from sympy.stats import UniformSum, density
>>> from sympy import Symbol, pprint
1388
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> n = Symbol(n, integer=True)
>>> z = Symbol(z)
>>> X = UniformSum(x, n)
>>> D = density(X)(z)
>>> pprint(D, use_unicode=False)
floor(z)
___
\
\
k
n - 1 /n\
)
(-1) *(-k + z)
*| |
/
\k/
/__,
k = 0
-------------------------------(n - 1)!
sympy.stats.VonMises(name, mu, k)
Create a Continuous Random Variable with a von Mises distribution.
The density of the von Mises distribution is given by
f (x) :=
e cos(x)
2I0 ()
with x [0, 2].
Parameters mu : Real number, measure of location
k : Real number, measure of concentration
Returns A RandomSymbol. :
References
[R435] (page 1914), [R436] (page 1914)
Examples
>>> from sympy.stats import VonMises, density, E, variance
>>> from sympy import Symbol, simplify, pprint
>>> mu = Symbol(mu)
>>> k = Symbol(k, positive=True)
>>> z = Symbol(z)
>>> X = VonMises(x, mu, k)
>>> D = density(X)(z)
>>> pprint(D, use_unicode=False)
k*cos(mu - z)
e
-----------------2*pi*besseli(0, k)
5.26. Stats
1389
SymPy Documentation, Release 0.7.6
sympy.stats.Weibull(name, alpha, beta)
Create a continuous random variable with a Weibull distribution.
The density of the Weibull distribution is given by
{ ( )k1
k
k x
e(x/)
f (x) :=
0
x0
x<0
Parameters lambda : Real number, > 0 a scale
k : Real number, k > 0 a shape
Returns A RandomSymbol. :
References
[R437] (page 1914), [R438] (page 1914)
Examples
>>> from sympy.stats import Weibull, density, E, variance
>>> from sympy import Symbol, simplify
>>> l = Symbol(lambda, positive=True)
>>> k = Symbol(k, positive=True)
>>> z = Symbol(z)
>>> X = Weibull(x, l, k)
>>> density(X)(z)
k*(z/lambda)**(k - 1)*exp(-(z/lambda)**k)/lambda
>>> simplify(E(X))
lambda*gamma(1 + 1/k)
>>> simplify(variance(X))
lambda**2*(-gamma(1 + 1/k)**2 + gamma(1 + 2/k))
sympy.stats.WignerSemicircle(name, R)
Create a continuous random variable with a Wigner semicircle distribution.
The density of the Wigner semicircle distribution is given by
f (x) :=
2 2
R x2
R2
with x [R, R].
Parameters R : Real number, R > 0, the radius
Returns A RandomSymbol. :
1390
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
References
[R439] (page 1914), [R440] (page 1914)
Examples
>>> from sympy.stats import WignerSemicircle, density, E
>>> from sympy import Symbol, simplify
>>> R = Symbol(R, positive=True)
>>> z = Symbol(z)
>>> X = WignerSemicircle(x, R)
>>> density(X)(z)
2*sqrt(R**2 - z**2)/(pi*R**2)
>>> E(X)
0
sympy.stats.ContinuousRV(symbol, density, set=(-oo, oo))
Create a Continuous Random Variable given the following:
a symbol a probability density function set on which the pdf is valid (defaults to
entire real line)
Returns a RandomSymbol.
Many common continuous random variable types are already implemented. This function
should be necessary only very rarely.
Examples
>>> from sympy import Symbol, sqrt, exp, pi
>>> from sympy.stats import ContinuousRV, P, E
>>> x = Symbol(x)
>>> pdf = sqrt(2)*exp(-x**2/2)/(2*sqrt(pi)) # Normal distribution
>>> X = ContinuousRV(x, pdf)
>>> E(X)
0
>>> P(X>0)
1/2
5.26.3 Interface
sympy.stats.P(condition, given condition=None, numsamples=None, evaluate=True,
**kwargs)
Probability that a condition is true, optionally given a second condition
Parameters expr : Relational containing RandomSymbols
5.26. Stats
1391
SymPy Documentation, Release 0.7.6
The condition of which you want to compute the probability
given condition : Relational containing RandomSymbols
A conditional expression. P(X>1, X>0) is expectation of X>1 given
X>0
numsamples : int
Enables sampling and approximates the probability with this many
samples
evalf : Bool (defaults to True)
If sampling return a number rather than a complex expression
evaluate : Bool (defaults to True)
In case of continuous systems return unevaluated integral
Examples
>>> from sympy.stats import P, Die
>>> from sympy import Eq
>>> X, Y = Die(X, 6), Die(Y, 6)
>>> P(X>3)
1/2
>>> P(Eq(X, 5), X>2) # Probability that X == 5 given that X > 2
1/4
>>> P(X>Y)
5/12
sympy.stats.E(expr, condition=None, numsamples=None, evaluate=True, **kwargs)
Returns the expected value of a random expression
Parameters expr : Expr containing RandomSymbols
The expression of which you want to compute the expectation value
given : Expr containing RandomSymbols
A conditional expression. E(X, X>0) is expectation of X given X > 0
numsamples : int
Enables sampling and approximates the expectation with this many
samples
evalf : Bool (defaults to True)
If sampling return a number rather than a complex expression
evaluate : Bool (defaults to True)
In case of continuous systems return unevaluated integral
Examples
>>> from sympy.stats import E, Die
>>> X = Die(X, 6)
>>> E(X)
7/2
1392
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> E(2*X + 1)
8
>>> E(X, X>3) # Expectation of X given that it is above 3
5
sympy.stats.density(expr, condition=None, evaluate=True, numsamples=None,
**kwargs)
Probability density of a random expression, optionally given a second condition.
This density will take on dierent forms for dierent types of probability spaces. Discrete
variables produce Dicts. Continuous variables produce Lambdas.
Parameters expr : Expr containing RandomSymbols
The expression of which you want to compute the density value
condition : Relational containing RandomSymbols
A conditional expression. density(X>1, X>0) is density of X>1 given
X>0
numsamples : int
Enables sampling and approximates the density with this many samples
Examples
>>> from sympy.stats import density, Die, Normal
>>> from sympy import Symbol
>>> x = Symbol(x)
>>> D = Die(D, 6)
>>> X = Normal(x, 0, 1)
>>> density(D).dict
{1: 1/6, 2: 1/6, 3: 1/6, 4: 1/6, 5: 1/6, 6: 1/6}
>>> density(2*D).dict
{2: 1/6, 4: 1/6, 6: 1/6, 8: 1/6, 10: 1/6, 12: 1/6}
>>> density(X)(x)
sqrt(2)*exp(-x**2/2)/(2*sqrt(pi))
sympy.stats.given(expr, condition=None, **kwargs)
Conditional Random Expression From a random expression and a condition on that expression creates a new probability space from the condition and returns the same expression on that conditional probability space.
Examples
>>>
>>>
>>>
>>>
{4:
from sympy.stats import given, density, Die
X = Die(X, 6)
Y = given(X, X>3)
density(Y).dict
1/3, 5: 1/3, 6: 1/3}
5.26. Stats
1393
SymPy Documentation, Release 0.7.6
Following convention, if the condition is a random symbol then that symbol is considered
xed.
>>> from sympy.stats import Normal
>>> from sympy import pprint
>>> from sympy.abc import z
>>> X = Normal(X, 0, 1)
>>> Y = Normal(Y, 0, 1)
>>> pprint(density(X + Y, Y)(z), use_unicode=False)
2
-(-Y + z)
----------___
2
\/ 2 *e
-----------------____
2*\/ pi
sympy.stats.where(condition, given condition=None, **kwargs)
Returns the domain where a condition is True.
Examples
>>> from sympy.stats import where, Die, Normal
>>> from sympy import symbols, And
>>> D1, D2 = Die(a, 6), Die(b, 6)
>>> a, b = D1.symbol, D2.symbol
>>> X = Normal(x, 0, 1)
>>> where(X**2<1)
Domain: And(-1 < x, x < 1)
>>> where(X**2<1).set
(-1, 1)
>>> where(And(D1<=D2 , D2<3))
Domain: Or(And(a == 1, b == 1), And(a == 1, b == 2), And(a == 2, b == 2))
sympy.stats.variance(X, condition=None, **kwargs)
Variance of a random expression
Expectation of (X-E(X))**2
Examples
>>> from sympy.stats import Die, E, Bernoulli, variance
>>> from sympy import simplify, Symbol
>>> X = Die(X, 6)
>>> p = Symbol(p)
>>> B = Bernoulli(B, p, 1, 0)
1394
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> variance(2*X)
35/3
>>> simplify(variance(B))
p*(-p + 1)
sympy.stats.std(X, condition=None, **kwargs)
Standard Deviation of a random expression
Square root of the Expectation of (X-E(X))**2
Examples
>>> from sympy.stats import Bernoulli, std
>>> from sympy import Symbol, simplify
>>> p = Symbol(p)
>>> B = Bernoulli(B, p, 1, 0)
>>> simplify(std(B))
sqrt(p*(-p + 1))
sympy.stats.sample(expr, condition=None, **kwargs)
A realization of the random expression
Examples
>>> from sympy.stats import Die, sample
>>> X, Y, Z = Die(X, 6), Die(Y, 6), Die(Z, 6)
>>> die_roll = sample(X + Y + Z) # A random realization of three dice
sympy.stats.sample iter(expr, condition=None, numsamples=oo, **kwargs)
Returns an iterator of realizations from the expression given a condition
expr: Random expression to be realized condition: A conditional expression (optional)
numsamples: Length of the iterator (defaults to innity)
See Also:
Sample, sampling P, sampling E, sample iter lambdify, sample iter subs
Examples
>>> from sympy.stats import Normal, sample_iter
>>> X = Normal(X, 0, 1)
>>> expr = X*X + 3
>>> iterator = sample_iter(expr, numsamples=3)
>>> list(iterator)
[12, 4, 7]
5.26. Stats
1395
SymPy Documentation, Release 0.7.6
5.26.4 Mechanics
SymPy Stats employs a relatively complex class hierarchy.
RandomDomains are a mapping of variables to possible values. For example we might say that
the symbol Symbol(x) can take on the values {1, 2, 3, 4, 5, 6}.
class sympy.stats.rv.RandomDomain
A PSpace, or Probability Space, combines a RandomDomain with a density to provide probabilistic information. For example the above domain could be enhanced by a nite density
{1:1/6, 2:1/6, 3:1/6, 4:1/6, 5:1/6, 6:1/6} to fully dene the roll of a fair die named
x.
class sympy.stats.rv.PSpace
A RandomSymbol represents the PSpaces symbol x inside of SymPy expressions.
class sympy.stats.rv.RandomSymbol
The RandomDomain and PSpace classes are almost never directly instantiated. Instead they
are subclassed for a variety of situations.
RandomDomains and PSpaces must be suciently general to represent domains and spaces
of several variables with arbitrarily complex densities. This generality is often unnecessary. Instead we often build SingleDomains and SinglePSpaces to represent single, univariate
events and processes such as a single die or a single normal variable.
class sympy.stats.rv.SinglePSpace
class sympy.stats.rv.SingleDomain
Another common case is to collect together a set of such univariate random variables. A
collection of independent SinglePSpaces or SingleDomains can be brought together to form
a ProductDomain or ProductPSpace. These objects would be useful in representing three
dice rolled together for example.
class sympy.stats.rv.ProductDomain
class sympy.stats.rv.ProductPSpace
The Conditional adjective is added whenever we add a global condition to a RandomDomain
or PSpace. A common example would be three independent dice where we know their sum
to be greater than 12.
class sympy.stats.rv.ConditionalDomain
We specialize further into Finite and Continuous versions of these classes to represent nite
(such as dice) and continuous (such as normals) random variables.
class sympy.stats.frv.FiniteDomain
class sympy.stats.frv.FinitePSpace
class sympy.stats.crv.ContinuousDomain
class sympy.stats.crv.ContinuousPSpace
Additionally there are a few specialized classes that implement certain common random variable types. There is for example a DiePSpace that implements SingleFinitePSpace and a
NormalPSpace that implements SingleContinuousPSpace.
class sympy.stats.frv types.DiePSpace
class sympy.stats.crv types.NormalPSpace
1396
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
RandomVariables can be extracted from these objects using the PSpace.values method.
As previously mentioned SymPy Stats employs a relatively complex class structure. Inheritance is widely used in the implementation of end-level classes. This tactic was chosen to
balance between the need to allow SymPy to represent arbitrarily dened random variables
and optimizing for common cases. This complicates the code but is structured to only be
important to those working on extending SymPy Stats to other random variable types.
Users will not use this class structure. Instead these mechanics are exposed through variable
creation functions Die, Coin, FiniteRV, Normal, Exponential, etc.... These build the appropriate SinglePSpaces and return the corresponding RandomVariable. Conditional and Product
spaces are formed in the natural construction of SymPy expressions and the use of interface
functions E, Given, Density, etc....
sympy.stats.Die()
sympy.stats.Normal()
There are some additional functions that may be useful. They are largely used internally.
sympy.stats.rv.random symbols(expr)
Returns all RandomSymbols within a SymPy Expression.
sympy.stats.rv.pspace(expr)
Returns the underlying Probability Space of a random expression.
For internal use.
Examples
>>> from sympy.stats import pspace, Normal
>>> from sympy.stats.rv import ProductPSpace
>>> X = Normal(X, 0, 1)
>>> pspace(2*X + 1) == X.pspace
True
sympy.stats.rv.rs swap(a, b)
Build a dictionary to swap RandomSymbols based on their underlying symbol.
i.e. if X = (x, pspace1) and Y = (x, pspace2) then X and Y match and the key,
value pair {X:Y} will appear in the result
Inputs: collections a and b of random variables which share common symbols Output:
dict mapping RVs in a to RVs in b
5.27 ODE
5.27.1 User Functions
These are functions that are imported into the global namespace with from sympy import *.
These functions (unlike Hint Functions (page 1405), below) are intended for use by ordinary
users of SymPy.
dsolve()
sympy.solvers.ode.dsolve(eq, func=None, hint=default, simplify=True, ics=None,
xi=None, eta=None, x0=0, n=6, **kwargs)
5.27. ODE
1397
SymPy Documentation, Release 0.7.6
Solves any (supported) kind of ordinary dierential equation and system of ordinary dierential equations.
Examples
>>> from sympy import Function, dsolve, Eq, Derivative, sin, cos, symbols
>>> from sympy.abc import x
>>> f = Function(f)
>>> dsolve(Derivative(f(x), x, x) + 9*f(x), f(x))
f(x) == C1*sin(3*x) + C2*cos(3*x)
>>> eq = sin(x)*cos(f(x)) + cos(x)*sin(f(x))*f(x).diff(x)
>>> dsolve(eq, hint=1st_exact)
[f(x) == -acos(C1/cos(x)) + 2*pi, f(x) == acos(C1/cos(x))]
>>> dsolve(eq, hint=almost_linear)
[f(x) == -acos(-sqrt(C1/cos(x)**2)) + 2*pi, f(x) == -acos(sqrt(C1/cos(x)**2)) + 2*pi,
f(x) == acos(-sqrt(C1/cos(x)**2)), f(x) == acos(sqrt(C1/cos(x)**2))]
>>> t = symbols(t)
>>> x, y = symbols(x, y, function=True)
>>> eq = (Eq(Derivative(x(t),t), 12*t*x(t) + 8*y(t)), Eq(Derivative(y(t),t), 21*x(t) + 7*t*y(t))
>>> dsolve(eq)
[x(t) == C1*x0 + C2*x0*Integral(8*exp(Integral(7*t, t))*exp(Integral(12*t, t))/x0**2, t),
y(t) == C1*y0 + C2(y0*Integral(8*exp(Integral(7*t, t))*exp(Integral(12*t, t))/x0**2, t) +
exp(Integral(7*t, t))*exp(Integral(12*t, t))/x0)]
>>> eq = (Eq(Derivative(x(t),t),x(t)*y(t)*sin(t)), Eq(Derivative(y(t),t),y(t)**2*sin(t)))
>>> dsolve(eq)
set([x(t) == -exp(C1)/(C2*exp(C1) - cos(t)), y(t) == -1/(C1 - cos(t))])
For Single Ordinary Dierential Equation
It is classied under this when number of equation in eq is one. Usage
dsolve(eq, f(x), hint) -> Solve ordinary dierential equation eq for function
f(x), using method hint.
Details
eq can be any supported ordinary dierential equation (see the ode
(page 1457) docstring for supported methods). This can either be an
Equality (page 148), or an expression, which is assumed to be equal to 0.
f(x) is a function of one variable whose derivatives in that variable
make up the ordinary dierential equation eq. In many cases it is not
necessary to provide this; it will be autodetected (and an error raised if it
couldnt be detected).
hint is the solving method that you want dsolve to use. Use
classify ode(eq, f(x)) to get all of the possible hints for an ODE.
The default hint, default, will use whatever hint is returned rst by
classify ode() (page 1400). See Hints below for more options that you
can use for hint.
simplify enables simplication by odesimp() (page 1406). See its docstring for more information. Turn this o, for example, to disable solving
of solutions for func or simplication of arbitrary constants. It will still integrate with this hint. Note that the solution may contain more arbitrary
constants than the order of the ODE with this option enabled.
1398
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
xi and eta are the innitesimal functions of an ordinary dierential
equation. They are the innitesimals of the Lie group of point transformations for which the dierential equation is invariant. The user can
specify values for the innitesimals. If nothing is specied, xi and eta are
calculated using infinitesimals() (page 1404) with the help of various
heuristics.
ics is the set of boundary conditions for the dierential equation. It
should be given in the form of {f(x0): x1, f(x).diff(x).subs(x, x2):
x3} and so on. For now initial conditions are implemented only for power
series solutions of rst-order dierential equations which should be given
in the form of {f(x0): x1} (See issue 4720). If nothing is specied for this
case f(0) is assumed to be C0 and the power series solution is calculated
about 0.
x0 is the point about which the power series solution of a dierential
equation is to be evaluated.
n gives the exponent of the dependent variable up to which the power series
solution of a dierential equation is to be evaluated.
Hints
Aside from the various solving methods, there are also some meta-hints that you
can pass to dsolve() (page 1397):
default: This uses whatever hint is returned rst by classify ode()
(page 1400). This is the default argument to dsolve() (page 1397).
all: To make dsolve() (page 1397) apply all relevant classication hints,
use dsolve(ODE, func, hint=all). This will return a dictionary of
hint:solution terms. If a hint causes dsolve to raise the NotImplementedError, value of that hints key will be the exception object raised. The
dictionary will also include some special keys:
order: The order of the ODE. See also ode order() (page 1478) in deutils.py.
best: The simplest hint; what would be returned by best below.
best hint: The hint that would produce the solution given by best. If
more than one hint produces the best solution, the rst one in the tuple
returned by classify ode() (page 1400) is chosen.
default: The solution that would be returned by default. This is the one
produced by the hint that appears rst in the tuple returned by classify ode() (page 1400).
all Integral: This is the same as all, except if a hint also has a corresponding Integral hint, it only returns the Integral hint. This is useful if all
causes dsolve() (page 1397) to hang because of a dicult or impossible
integral. This meta-hint will also be much faster than all, because integrate() (page 114) is an expensive routine.
best: To have dsolve() (page 1397) try all methods and return the simplest
one. This takes into account whether the solution is solvable in the function,
whether it contains any Integral classes (i.e. unevaluatable integrals), and
which one is the shortest in size.
See also the classify ode() (page 1400) docstring for more info on hints, and
the ode (page 1457) docstring for a list of all supported hints.
Tips
5.27. ODE
1399
SymPy Documentation, Release 0.7.6
You can declare the derivative of an unknown function this way:
>>>
>>>
>>>
>>>
>>>
from sympy import Function,
from sympy.abc import x # x
f = Function(f)(x) # f is
# f_ will be the derivative
f_ = Derivative(f, x)
Derivative
is the independent variable
a function of x
of f with respect to x
See test ode.py for many tests, which serves also as a set of examples for how to
use dsolve() (page 1397).
dsolve() (page 1397) always returns an Equality (page 148) class (except for the
case when the hint is all or all Integral). If possible, it solves the solution explicitly for the function being solved for. Otherwise, it returns an implicit solution.
Arbitrary constants are symbols named C1, C2, and so on.
Because all solutions should be mathematically equivalent, some hints may return
the exact same result for an ODE. Often, though, two dierent hints will return the
same solution formatted dierently. The two should be equivalent. Also note that
sometimes the values of the arbitrary constants in two dierent solutions may not
be the same, because one constant may have absorbed other constants into it.
Do help(ode.ode <hintname>) to get help more information on a specic hint,
where <hintname> is the name of a hint without Integral.
For System Of Ordinary Dierential Equations
Usage
dsolve(eq, func) -> Solve a system of ordinary dierential equations eq for
func being list of functions including x(t), y(t), z(t) where number of functions
in the list depends upon the number of equations provided in eq.
Details
eq can be any supported system of ordinary dierential equations This can
either be an Equality (page 148), or an expression, which is assumed to be
equal to 0.
func holds x(t) and y(t) being functions of one variable which together
with some of their derivatives make up the system of ordinary dierential
equation eq. It is not necessary to provide this; it will be autodetected (and
an error raised if it couldnt be detected).
Hints
The hints are formed by parameters returned by classify sysode, combining
them give hints name used later for forming method name.
classify ode()
sympy.solvers.ode.classify ode(eq, func=None, dict=False, ics=None, **kwargs)
Returns a tuple of possible dsolve() (page 1397) classications for an ODE.
The tuple is ordered so that rst item is the classication that dsolve() (page 1397)
uses to solve the ODE by default. In general, classications at the near the beginning
of the list will produce better solutions faster than those near the end, thought there
are always exceptions. To make dsolve() (page 1397) use a dierent classication, use
1400
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
dsolve(ODE, func, hint=<classification>). See also the dsolve() (page 1397)
docstring for dierent meta-hints you can use.
If dict is true, classify ode() (page 1400) will return a dictionary of hint:match expression terms. This is intended for internal use by dsolve() (page 1397). Note that
because dictionaries are ordered arbitrarily, this will most likely not be in the same order
as the tuple.
You can get help on dierent hints by executing help(ode.ode hintname), where hintname is the name of the hint without Integral.
See allhints (page 1405) or the ode (page 1457) docstring for a list of all supported
hints that can be returned from classify ode() (page 1400).
Notes
These are remarks on hint names.
Integral
If a classication has Integral at the end, it will return the expression with
an unevaluated Integral (page 615) class in it. Note that a hint may do this
anyway if integrate() (page 114) cannot do the integral, though just using an
Integral will do so much faster. Indeed, an Integral hint will always be faster
than its corresponding hint without Integral because integrate() (page 114)
is an expensive routine. If dsolve() (page 1397) hangs, it is probably because
integrate() (page 114) is hanging on a tough or impossible integral. Try using
an Integral hint or all Integral to get it return something.
Note that some hints do not have Integral counterparts. This is because integrate() is not used in solving the ODE for those method. For example, nth order
linear homogeneous ODEs with constant coecients do not require integration
to solve, so there is no nth linear homogeneous constant coeff Integrate
hint. You can easily evaluate any unevaluated Integral (page 615)s in an expression by doing expr.doit().
Ordinals
Some hints contain an ordinal such as 1st linear. This is to help dierentiate
them from other hints, as well as from other methods that may not be implemented yet. If a hint has nth in it, such as the nth linear hints, this means that
the method used to applies to ODEs of any order.
indep and dep
Some hints contain the words indep or dep. These reference the independent
variable and the dependent function, respectively. For example, if an ODE is in
terms of f (x), then indep will refer to x and dep will refer to f .
subs
If a hints has the word subs in it, it means the the ODE is solved by substituting the expression given after the word subs for a single dummy variable.
This is usually in terms of indep and dep as above. The substituted expression
will be written only in characters allowed for names of Python objects, meaning operators will be spelled out. For example, indep/dep will be written as
indep div dep.
coeff
5.27. ODE
1401
SymPy Documentation, Release 0.7.6
The word coeff in a hint refers to the coecients of something in the ODE,
usually of the derivative terms. See the docstring for the individual methods
for more info (help(ode)). This is contrast to coefficients, as in undetermined coefficients, which refers to the common name of a method.
best
Methods that have more than one fundamental way to solve will have a hint for
each sub-method and a best meta-classication. This will evaluate all hints and
return the best, using the same considerations as the normal best meta-hint.
Examples
>>> from sympy import Function, classify_ode, Eq
>>> from sympy.abc import x
>>> f = Function(f)
>>> classify_ode(Eq(f(x).diff(x), 0), f(x))
(separable, 1st_linear, 1st_homogeneous_coeff_best,
1st_homogeneous_coeff_subs_indep_div_dep,
1st_homogeneous_coeff_subs_dep_div_indep,
1st_power_series, lie_group,
nth_linear_constant_coeff_homogeneous,
separable_Integral, 1st_linear_Integral,
1st_homogeneous_coeff_subs_indep_div_dep_Integral,
1st_homogeneous_coeff_subs_dep_div_indep_Integral)
>>> classify_ode(f(x).diff(x, 2) + 3*f(x).diff(x) + 2*f(x) - 4)
(nth_linear_constant_coeff_undetermined_coefficients,
nth_linear_constant_coeff_variation_of_parameters,
nth_linear_constant_coeff_variation_of_parameters_Integral)
checkodesol()
sympy.solvers.ode.checkodesol(ode,
sol,
func=None,
solve for func=True)
Substitutes sol into ode and checks that the result is 0.
order=auto,
This only works when func is one function, like f (x). sol can be a single solution or a list
of solutions. Each solution may be an Equality (page 148) that the solution satises,
e.g. Eq(f(x), C1), Eq(f(x) + C1, 0); or simply an Expr (page 103), e.g. f(x) C1. In most cases it will not be necessary to explicitly identify the function, but if the
function cannot be inferred from the original equation it can be supplied through the
func argument.
If a sequence of solutions is passed, the same sort of container will be used to return the
result for each solution.
It tries the following methods, in order, until it nds zero equivalence:
1.Substitute the solution for f in the original equation. This only works if ode is solved
for f . It will attempt to solve it rst unless solve for func == False.
2.Take n derivatives of the solution, where n is the order of ode, and check to see if
that is equal to the solution. This only works on exact ODEs.
3.Take the 1st, 2nd, ..., nth derivatives of the solution, each time solving for the derivative of f of that order (this will always be possible because f is a linear operator).
Then back substitute each derivative into ode in reverse order.
1402
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
This function returns a tuple. The rst item in the tuple is True if the substitution results
in 0, and False otherwise. The second item in the tuple is what the substitution results
in. It should always be 0 if the rst item is True. Note that sometimes this function
will False, but with an expression that is identically equal to 0, instead of returning
True. This is because simplify() (page 1327) cannot reduce the expression to 0. If an
expression returned by this function vanishes identically, then sol really is a solution to
ode.
If this function seems to hang, it is probably because of a hard simplication.
To use this function to test, test the rst item of the tuple.
Examples
>>> from sympy import Eq, Function, checkodesol, symbols
>>> x, C1 = symbols(x,C1)
>>> f = Function(f)
>>> checkodesol(f(x).diff(x), Eq(f(x), C1))
(True, 0)
>>> assert checkodesol(f(x).diff(x), C1)[0]
>>> assert not checkodesol(f(x).diff(x), x)[0]
>>> checkodesol(f(x).diff(x, 2), x**2)
(False, 2)
homogeneous order()
sympy.solvers.ode.homogeneous order(eq, *symbols)
Returns the order n if g is homogeneous and None if it is not homogeneous.
Determines if a function is homogeneous and if so of what order. A function f (x, y, ) is
homogeneous of order n if f (tx, ty, ) = tn f (x, y, ).
If the function is of two variables, F (x, y), then f being homogeneous of any order is
equivalent to being able to rewrite F (x, y) as G(x/y) or H(y/x). This fact is used to solve
1st order ordinary dierential equations whose coecients are homogeneous of the
same order (see the docstrings of ode 1st homogeneous coeff subs dep div indep()
and ode 1st homogeneous coeff subs indep div dep()).
Symbols can be functions, but every argument of the function must be a symbol, and the
arguments of the function that appear in the expression must match those given in the
list of symbols. If a declared function appears with dierent arguments than given in
the list of symbols, None is returned.
Examples
>>> from sympy import Function, homogeneous_order, sqrt
>>> from sympy.abc import x, y
>>> f = Function(f)
>>> homogeneous_order(f(x), f(x)) is None
True
>>> homogeneous_order(f(x,y), f(y, x), x, y) is None
True
>>> homogeneous_order(f(x), f(x), x)
1
5.27. ODE
1403
SymPy Documentation, Release 0.7.6
>>> homogeneous_order(x**2*f(x)/sqrt(x**2+f(x)**2), x, f(x))
2
>>> homogeneous_order(x**2+f(x), x, f(x)) is None
True
infinitesimals()
sympy.solvers.ode.infinitesimals(eq, func=None, order=None, hint=default,
match=None)
The innitesimal functions of an ordinary dierential equation, (x, y) and (x, y), are the
innitesimals of the Lie group of point transformations for which the dierential equation
is invariant. So, the ODE y 0 = f (x, y) would admit a Lie group x = X(x, y; ) = x + (x, y),
y = Y (x, y; ) = y + (x, y) such that (y )0 = f (x , y ). A change of coordinates, to r(x, y)
and s(x, y), can be performed so this Lie group becomes the translation group, r = r and
s = s + . They are tangents to the coordinate curves of the new system.
Consider the transformation (x, y) (X, Y ) such that the dierential equation remains
invariant. and are the tangents to the transformed coordinates X and Y , at = 0.
(
)
(
)
X(x, y; )
Y (x, y; )
|=0 = ,
|=0 = ,
The innitesimals can be found by solving the following PDE:
>>> from sympy import Function, diff, Eq, pprint
>>> from sympy.abc import x, y
>>> xi, eta, h = map(Function, [xi, eta, h])
>>> h = h(x, y) # dy/dx = h
>>> eta = eta(x, y)
>>> xi = xi(x, y)
>>> genform = Eq(eta.diff(x) + (eta.diff(y) - xi.diff(x))*h
... - (xi.diff(y))*h**2 - xi*(h.diff(x)) - eta*(h.diff(y)), 0)
>>> pprint(genform)
/d
d
\
d
2
d
|--(eta(x, y)) - --(xi(x, y))|*h(x, y) - eta(x, y)*--(h(x, y)) - h (x, y)*--(x
\dy
dx
/
dy
dy
d
d
i(x, y)) - xi(x, y)*--(h(x, y)) + --(eta(x, y)) = 0
dx
dx
Solving the above mentioned PDE is not trivial, and can be solved only by making intelligent assumptions for and (heuristics). Once an innitesimal is found, the attempt
to nd more heuristics stops. This is done to optimise the speed of solving the dierential equation. If a list of all the innitesimals is needed, hint should be agged as
all, which gives the complete list of innitesimals. If the innitesimals for a particular
heuristic needs to be found, it can be passed as a ag to hint.
References
Solving dierential equations by Symmetry Groups, John Starrett, pp. 1 - pp. 14
1404
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import Function, diff
>>> from sympy.solvers.ode import infinitesimals
>>> from sympy.abc import x
>>> f = Function(f)
>>> eq = f(x).diff(x) - x**2*f(x)
>>> infinitesimals(eq)
[{eta(x, f(x)): exp(x**3/3), xi(x, f(x)): 0}]
checkinfsol()
sympy.solvers.ode.checkinfsol(eq, innitesimals, func=None, order=None)
This function is used to check if the given innitesimals are the actual innitesimals of
the given rst order dierential equation. This method is specic to the Lie Group Solver
of ODEs.
As of now, it simply checks, by substituting the innitesimals in the partial dierential
equation.
(
)
h
h
+
h
h2
=0
x
y
x
y
x
y
where , and are the innitesimals and h(x, y) =
dy
dx
The innitesimals should be given in the form of a list of dicts [{xi(x, y): inf,
eta(x, y): inf}], corresponding to the output of the function innitesimals. It returns
a list of values of the form [(True/False, sol)] where sol is the value obtained after
substituting the innitesimals in the PDE. If it is True, then sol would be 0.
5.27.2 Hint Functions
These functions are intended for internal use by dsolve() (page 1397) and others. Unlike
User Functions (page 1397), above, these are not intended for every-day use by ordinary
SymPy users. Instead, functions such as dsolve() (page 1397) should be used. Nonetheless,
these functions contain useful information in their docstrings on the various ODE solving
methods. For this reason, they are documented here.
allhints
sympy.solvers.ode.allhints = (separable, 1st exact, 1st linear, Bernoulli, Riccati special m
This is a list of hints in the order that they should be preferred by classify ode()
(page 1400). In general, hints earlier in the list should produce simpler solutions than
those later in the list (for ODEs that t both). For now, the order of this list is based on
empirical observations by the developers of SymPy.
The hint used by dsolve() (page 1397) for a specic ODE can be overridden (see the
docstring).
In general, Integral hints are grouped at the end of the list, unless there is a method
that returns an unevaluable integral most of the time (which go near the end of the list
anyway). default, all, best, and all Integral meta-hints should not be included in
this list, but best and Integral hints should be included.
5.27. ODE
1405
SymPy Documentation, Release 0.7.6
odesimp
sympy.solvers.ode.odesimp(*args, **kwargs)
Simplies ODEs, including trying to solve for func and running constantsimp()
(page 1407).
It may use knowledge of the type of solution that the hint returns to apply additional
simplications.
It also attempts to integrate any Integral (page 615)s in the expression, if the hint is
not an Integral hint.
This function should have no eect on expressions returned by dsolve() (page 1397),
as dsolve() (page 1397) already calls odesimp() (page 1406), but the individual hint
functions do not call odesimp() (page 1406) (because the dsolve() (page 1397) wrapper
does). Therefore, this function is designed for mainly internal use.
Examples
>>>
>>>
>>>
>>>
from sympy import sin, symbols, dsolve, pprint, Function
from sympy.solvers.ode import odesimp
x , u2, C1= symbols(x,u2,C1)
f = Function(f)
>>>
...
...
>>>
eq = dsolve(x*f(x).diff(x) - f(x) - x*sin(f(x)/x), f(x),
hint=1st_homogeneous_coeff_subs_indep_div_dep_Integral,
simplify=False)
pprint(eq, wrap_line=False)
x
---f(x)
/
|
|
/
1
\
| -|u2 + -------|
|
|
/1 \|
|
|
sin|--||
|
\
\u2//
log(f(x)) = log(C1) +
| ---------------- d(u2)
|
2
|
u2
|
/
>>> pprint(odesimp(eq, f(x), 1, set([C1]),
... hint=1st_homogeneous_coeff_subs_indep_div_dep
... ))
x
--------- = C1
/f(x)\
tan|----|
\2*x /
1406
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
constant renumber
sympy.solvers.ode.constant renumber(expr, symbolname, startnumber, endnumber)
Renumber arbitrary constants in expr to have numbers 1 through N where N is endnumber - startnumber + 1 at most. In the process, this reorders expression terms in a
standard way.
This is a simple function that goes through and renumbers any Symbol (page 122) with
a name in the form symbolname + num where num is in the range from startnumber to
endnumber.
Symbols are renumbered based on .sort key(), so they should be numbered roughly
in the order that they appear in the nal, printed expression. Note that this ordering is
based in part on hashes, so it can produce dierent results on dierent machines.
The structure of this function is very similar to that of constantsimp() (page 1407).
Examples
>>> from sympy import symbols, Eq, pprint
>>> from sympy.solvers.ode import constant_renumber
>>> x, C0, C1, C2, C3, C4 = symbols(x,C:5)
Only constants in the given range (inclusive) are renumbered; the renumbering always
starts from 1:
>>> constant_renumber(C1 + C3 +
C1 + C2 + C4
>>> constant_renumber(C0 + C1 +
C0 + 2*C1 + C2
>>> constant_renumber(C0 + 2*C1
C1 + 3*C2
>>> pprint(C2 + C1*x + C3*x**2)
2
C1*x + C2 + C3*x
>>> pprint(constant_renumber(C2
2
C1 + C2*x + C3*x
C4, C, 1, 3)
C3 + C4, C, 2, 4)
+ C2, C, 0, 1)
+ C1*x + C3*x**2, C, 1, 3))
constantsimp
sympy.solvers.ode.constantsimp(*args, **kwargs)
Simplies an expression with arbitrary constants in it.
This function is written specically to work with dsolve() (page 1397), and is not intended for general use.
Simplication is done by absorbing the arbitrary constants into other arbitrary constants, numbers, and symbols that they are not independent of.
The symbols must all have the same name with numbers after it, for example, C1, C2,
C3. The symbolname here would be C, the startnumber would be 1, and the endnumber
would be 3. If the arbitrary constants are independent of the variable x, then the independent symbol would be x. There is no need to specify the dependent function, such as
f(x), because it already has the independent symbol, x, in it.
5.27. ODE
1407
SymPy Documentation, Release 0.7.6
Because terms are absorbed into arbitrary constants and because constants are
renumbered after simplifying, the arbitrary constants in expr are not necessarily equal
to the ones of the same name in the returned result.
If two or more arbitrary constants are added, multiplied, or raised to the power of each
other, they are rst absorbed together into a single arbitrary constant. Then the new
constant is combined into other terms if necessary.
Absorption of constants is done with limited assistance:
1.terms of Add (page 144)s are collected to try join constants so ex (C1 cos(x)+C2 cos(x))
will simplify to ex C1 cos(x);
2.powers with exponents that are Add (page 144)s are expanded so eC1 +x will be simplied to C1 ex .
Use constant renumber() (page 1407) to renumber constants after simplication or else
arbitrary numbers on constants may appear, e.g. C1 + C3 x.
In rare cases, a single constant can be simplied into two constants. Every dierential
equation solution should have as many arbitrary constants as the order of the dierential
equation. The result here will be technically correct, but it may, for example, have C1
and C2 in an expression, when C1 is actually equal to C2 . Use your discretion in such
situations, and also take advantage of the ability to use hints in dsolve() (page 1397).
Examples
>>> from sympy import symbols
>>> from sympy.solvers.ode import constantsimp
>>> C1, C2, C3, x, y = symbols(C1, C2, C3, x, y)
>>> constantsimp(2*C1*x, set([C1, C2, C3]))
C1*x
>>> constantsimp(C1 + 2 + x, set([C1, C2, C3]))
C1 + x
>>> constantsimp(C1*C2 + 2 + C2 + C3*x, set([C1, C2, C3]))
C1 + C3*x
sol simplicity
sympy.solvers.ode.ode sol simplicity(sol, func, trysolving=True)
Returns an extended integer representing how simple a solution to an ODE is.
The following things are considered, in order from most simple to least:
sol is solved for func.
sol is not solved for func, but can be if passed to solve (e.g., a solution returned by
dsolve(ode, func, simplify=False).
If sol is not solved for func, then base the result on the length of sol, as computed
by len(str(sol)).
If sol has any unevaluated Integral (page 615)s, this will automatically be considered less simple than any of the above.
This function returns an integer such that if solution A is simpler than solution B by above
metric, then ode sol simplicity(sola, func) < ode sol simplicity(solb, func).
Currently, the following are the numbers returned, but if the heuristic is ever improved,
this may change. Only the ordering is guaranteed.
1408
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Simplicity
sol solved for func
sol not solved for func but can be
sol is not solved nor solvable for func
sol contains an Integral (page 615)
Return
-2
-1
len(str(sol))
oo
oo here means the SymPy innity, which should compare greater than any integer.
If you already know solve() (page 1469) cannot solve sol, you can use trysolving=False to skip that step, which is the only potentially slow step. For example,
dsolve() (page 1397) with the simplify=False ag should do this.
If sol is a list of solutions, if the worst solution in the list returns oo it returns that,
otherwise it returns len(str(sol)), that is, the length of the string representation of
the whole list.
Examples
This function is designed to be passed to min as the key argument, such as
min(listofsolutions, key=lambda i: ode sol simplicity(i, f(x))).
>>>
>>>
>>>
>>>
from sympy import symbols, Function, Eq, tan, cos, sqrt, Integral
from sympy.solvers.ode import ode_sol_simplicity
x, C1, C2 = symbols(x, C1, C2)
f = Function(f)
>>> ode_sol_simplicity(Eq(f(x), C1*x**2), f(x))
-2
>>> ode_sol_simplicity(Eq(x**2 + f(x), C1), f(x))
-1
>>> ode_sol_simplicity(Eq(f(x), C1*Integral(2*x, x)), f(x))
oo
>>> eq1 = Eq(f(x)/tan(f(x)/(2*x)), C1)
>>> eq2 = Eq(f(x)/tan(f(x)/(2*x) + f(x)), C2)
>>> [ode_sol_simplicity(eq, f(x)) for eq in [eq1, eq2]]
[26, 33]
>>> min([eq1, eq2], key=lambda i: ode_sol_simplicity(i, f(x)))
f(x)/tan(f(x)/(2*x)) == C1
1st exact
sympy.solvers.ode.ode 1st exact(eq, func, order, match)
Solves 1st order exact ordinary dierential equations.
A 1st order dierential equation is called exact if it is the total dierential of a function.
That is, the dierential equation
P (x, y) x + Q(x, y) y = 0
is exact if there is some function F (x, y) such that P (x, y) = F /x and Q(x, y) = F /y. It
can be shown that a necessary and sucient condition for a rst order ODE to be exact
is that P /y = Q/x. Then, the solution will be as given below:
5.27. ODE
1409
SymPy Documentation, Release 0.7.6
>>>
>>>
>>>
>>>
...
from sympy import Function, Eq, Integral, symbols, pprint
x, y, t, x0, y0, C1= symbols(x,y,t,x0,y0,C1)
P, Q, F= map(Function, [P, Q, F])
pprint(Eq(Eq(F(x, y), Integral(P(t, y), (t, x0, x)) +
Integral(Q(x0, t), (t, y0, y))), C1))
x
y
/
/
|
|
F(x, y) = | P(t, y) dt + | Q(x0, t) dt = C1
|
|
/
/
x0
y0
Where the rst partials of P and Q exist and are continuous in a simply connected region.
A note: SymPy currently has no way to represent inert substitution on an expression, so
the hint 1st exact Integral will return an integral with dy. This is supposed to represent the function that you are solving for.
References
http://en.wikipedia.org/wiki/Exact dierential equation
M. Tenenbaum & H. Pollard, Ordinary Dierential Equations, Dover 1963, pp. 73
# indirect doctest
Examples
>>> from sympy import Function, dsolve, cos, sin
>>> from sympy.abc import x
>>> f = Function(f)
>>> dsolve(cos(f(x)) - (x*sin(f(x)) - f(x)**2)*f(x).diff(x),
... f(x), hint=1st_exact)
x*cos(f(x)) + f(x)**3/3 == C1
1st homogeneous coeff best
sympy.solvers.ode.ode 1st homogeneous coeff best(eq, func, order, match)
Returns
the
best
solution
to
an
ODE
from
the
two
hints
1st homogeneous coeff subs dep div indep and 1st homogeneous coeff subs indep div dep.
This is as determined by ode sol simplicity() (page 1408).
See
the
ode 1st homogeneous coeff subs indep div dep()
(page
ode 1st homogeneous coeff subs dep div indep()
(page
1411)
and
strings for more information on these hints.
Note that there
ode 1st homogeneous coeff best Integral hint.
1412)
docis no
References
http://en.wikipedia.org/wiki/Homogeneous dierential equation
M. Tenenbaum & H. Pollard, Ordinary Dierential Equations, Dover 1963, pp. 59
1410
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
# indirect doctest
Examples
>>>
>>>
>>>
>>>
...
from sympy import Function, dsolve, pprint
from sympy.abc import x
f = Function(f)
pprint(dsolve(2*x*f(x) + (x**2 + f(x)**2)*f(x).diff(x), f(x),
hint=1st_homogeneous_coeff_best, simplify=False))
/
2
\
| 3*x
|
log|----- + 1|
| 2
|
\f (x)
/
log(f(x)) = log(C1) - -------------3
1st homogeneous coeff subs dep div indep
sympy.solvers.ode.ode 1st homogeneous coeff subs dep div indep(eq, func, order,
match)
Solves a 1st order dierential equation with homogeneous coecients using the substi<dependent variable>
tution u1 = <independent
variable> .
This is a dierential equation
P (x, y) + Q(x, y)dy/dx = 0
such that P and Q are homogeneous and of the same order. A function F (x, y) is homogeneous of order n if F (xt, yt) = tn F (x, y). Equivalently, F (x, y) can be rewritten as G(y/x)
or H(x/y). See also the docstring of homogeneous order() (page 1403).
If the coecients P and Q in the dierential equation above are homogeneous functions
of the same order, then it can be shown that the substitution y = u1 x (i.e. u1 = y/x) will
turn the dierential equation into an equation separable in the variables x and u. If h(u1 )
is the function that results from making the substitution u1 = f (x)/x on P (x, f (x)) and g(u2 )
is the function that results from the substitution on Q(x, f (x)) in the dierential equation
P (x, f (x)) + Q(x, f (x))f 0 (x) = 0, then the general solution is:
>>> from sympy import Function, dsolve, pprint
>>> from sympy.abc import x
>>> f, g, h = map(Function, [f, g, h])
>>> genform = g(f(x)/x) + h(f(x)/x)*f(x).diff(x)
>>> pprint(genform)
/f(x)\
/f(x)\ d
g|----| + h|----|*--(f(x))
\ x /
\ x / dx
>>> pprint(dsolve(genform, f(x),
... hint=1st_homogeneous_coeff_subs_dep_div_indep_Integral))
f(x)
---x
/
|
5.27. ODE
1411
SymPy Documentation, Release 0.7.6
|
|
|
|
log(x) = C1 +
-h(u1)
---------------- d(u1)
u1*h(u1) + g(u1)
Where u1 h(u1 ) + g(u1 ) 6= 0 and x 6= 0.
See also the docstrings of ode 1st homogeneous coeff best() (page 1410) and
ode 1st homogeneous coeff subs indep div dep() (page 1412).
References
http://en.wikipedia.org/wiki/Homogeneous dierential equation
M. Tenenbaum & H. Pollard, Ordinary Dierential Equations, Dover 1963, pp. 59
# indirect doctest
Examples
>>>
>>>
>>>
>>>
...
from sympy import Function, dsolve
from sympy.abc import x
f = Function(f)
pprint(dsolve(2*x*f(x) + (x**2 + f(x)**2)*f(x).diff(x), f(x),
hint=1st_homogeneous_coeff_subs_dep_div_indep, simplify=False))
/
3
\
|3*f(x)
f (x)|
log|------ + -----|
| x
3 |
\
x /
log(x) = log(C1) - ------------------3
1st homogeneous coeff subs indep div dep
sympy.solvers.ode.ode 1st homogeneous coeff subs indep div dep(eq, func, order,
match)
Solves a 1st order dierential equation with homogeneous coecients using the substivariable>
tution u2 = <independent
<dependent variable> .
This is a dierential equation
P (x, y) + Q(x, y)dy/dx = 0
such that P and Q are homogeneous and of the same order. A function F (x, y) is homogeneous of order n if F (xt, yt) = tn F (x, y). Equivalently, F (x, y) can be rewritten as G(y/x)
or H(x/y). See also the docstring of homogeneous order() (page 1403).
If the coecients P and Q in the dierential equation above are homogeneous functions
of the same order, then it can be shown that the substitution x = u2 y (i.e. u2 = x/y) will
turn the dierential equation into an equation separable in the variables y and u2 . If
h(u2 ) is the function that results from making the substitution u2 = x/f (x) on P (x, f (x))
1412
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
and g(u2 ) is the function that results from the substitution on Q(x, f (x)) in the dierential
equation P (x, f (x)) + Q(x, f (x))f 0 (x) = 0, then the general solution is:
>>> from sympy import Function, dsolve, pprint
>>> from sympy.abc import x
>>> f, g, h = map(Function, [f, g, h])
>>> genform = g(x/f(x)) + h(x/f(x))*f(x).diff(x)
>>> pprint(genform)
/ x \
/ x \ d
g|----| + h|----|*--(f(x))
\f(x)/
\f(x)/ dx
>>> pprint(dsolve(genform, f(x),
... hint=1st_homogeneous_coeff_subs_indep_div_dep_Integral))
x
---f(x)
/
|
|
-g(u2)
| ---------------- d(u2)
| u2*g(u2) + h(u2)
|
/
f(x) = C1*e
Where u2 g(u2 ) + h(u2 ) 6= 0 and f (x) 6= 0.
See also the docstrings of ode 1st homogeneous coeff best() (page 1410) and
ode 1st homogeneous coeff subs dep div indep() (page 1411).
References
http://en.wikipedia.org/wiki/Homogeneous dierential equation
M. Tenenbaum & H. Pollard, Ordinary Dierential Equations, Dover 1963, pp. 59
# indirect doctest
Examples
>>>
>>>
>>>
>>>
...
...
from sympy import Function, pprint, dsolve
from sympy.abc import x
f = Function(f)
pprint(dsolve(2*x*f(x) + (x**2 + f(x)**2)*f(x).diff(x), f(x),
hint=1st_homogeneous_coeff_subs_indep_div_dep,
simplify=False))
/
2
\
| 3*x
|
log|----- + 1|
| 2
|
\f (x)
/
log(f(x)) = log(C1) - -------------3
5.27. ODE
1413
SymPy Documentation, Release 0.7.6
1st linear
sympy.solvers.ode.ode 1st linear(eq, func, order, match)
Solves 1st order linear dierential equations.
These are dierential equations of the form
dy/dx + P (x)y = Q(x).
These kinds
of dierential equations can be solved in a general way. The integrating
factor e P (x) dx will turn the equation into a separable equation. The general solution is:
>>>
>>>
>>>
>>>
>>>
from sympy import Function, dsolve, Eq, pprint, diff, sin
from sympy.abc import x
f, P, Q = map(Function, [f, P, Q])
genform = Eq(f(x).diff(x) + P(x)*f(x), Q(x))
pprint(genform)
d
P(x)*f(x) + --(f(x)) = Q(x)
dx
>>> pprint(dsolve(genform, f(x), hint=1st_linear_Integral))
/
/
\
|
|
|
|
|
/
|
/
|
|
|
|
|
|
|
| P(x) dx
| - | P(x) dx
|
|
|
|
|
|
|
/
|
/
f(x) = |C1 + | Q(x)*e
dx|*e
|
|
|
\
/
/
References
http://en.wikipedia.org/wiki/Linear dierential equation#First order equation
M. Tenenbaum & H. Pollard, Ordinary Dierential Equations, Dover 1963, pp. 92
# indirect doctest
Examples
>>> f = Function(f)
>>> pprint(dsolve(Eq(x*diff(f(x), x) - f(x), x**2*sin(x)),
... f(x), 1st_linear))
f(x) = x*(C1 - cos(x))
Bernoulli
sympy.solvers.ode.ode Bernoulli(eq, func, order, match)
Solves Bernoulli dierential equations.
1414
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
These are equations of the form
dy/dx + P (x)y = Q(x)y n , n 6= 1.
The substitution w = 1/y 1n will transform an equation of this form into one that is linear
(see the docstring of ode 1st linear() (page 1414)). The general solution is:
>>>
>>>
>>>
>>>
>>>
from sympy import Function, dsolve, Eq, pprint
from sympy.abc import x, n
f, P, Q = map(Function, [f, P, Q])
genform = Eq(f(x).diff(x) + P(x)*f(x), Q(x)*f(x)**n)
pprint(genform)
d
n
P(x)*f(x) + --(f(x)) = Q(x)*f (x)
dx
>>> pprint(dsolve(genform, f(x), hint=Bernoulli_Integral))
1
---1 - n
//
/
\
\
||
|
|
|
||
|
/
|
/
|
||
|
|
|
|
|
||
|
(1 - n)* | P(x) dx
| (-1 + n)* | P(x) dx|
||
|
|
|
|
|
||
|
/
|
/
|
f(x) = ||C1 + (-1 + n)* | -Q(x)*e
dx|*e
|
||
|
|
|
\\
/
/
/
Note that the equation is separable when n = 1 (see the docstring of ode separable()
(page 1421)).
>>> pprint(dsolve(Eq(f(x).diff(x) + P(x)*f(x), Q(x)*f(x)), f(x),
... hint=separable_Integral))
f(x)
/
|
/
| 1
|
| - dy = C1 + | (-P(x) + Q(x)) dx
| y
|
|
/
/
References
http://en.wikipedia.org/wiki/Bernoulli dierential equation
M. Tenenbaum & H. Pollard, Ordinary Dierential Equations, Dover 1963, pp. 95
# indirect doctest
5.27. ODE
1415
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import Function, dsolve, Eq, pprint, log
>>> from sympy.abc import x
>>> f = Function(f)
>>> pprint(dsolve(Eq(x*f(x).diff(x) + f(x), log(x)*f(x)**2),
... f(x), hint=Bernoulli))
1
f(x) = ------------------/
log(x)
1\
x*|C1 + ------ + -|
\
x
x/
Liouville
sympy.solvers.ode.ode Liouville(eq, func, order, match)
Solves 2nd order Liouville dierential equations.
The general form of a Liouville ODE is
( )2
d2 y
dy
dy
+ g(y)
+ h(x) .
2
dx
dx
dx
The general solution is:
>>>
>>>
>>>
>>>
...
>>>
from sympy import Function, dsolve, Eq, pprint, diff
from sympy.abc import x
f, g, h = map(Function, [f, g, h])
genform = Eq(diff(f(x),x,x) + g(f(x))*diff(f(x),x)**2 +
h(x)*diff(f(x),x), 0)
pprint(genform)
2
2
/d
\
d
d
g(f(x))*|--(f(x))| + h(x)*--(f(x)) + ---(f(x)) = 0
\dx
/
dx
2
dx
>>> pprint(dsolve(genform, f(x), hint=Liouville_Integral))
f(x)
/
/
|
|
|
/
|
/
|
|
|
|
| - | h(x) dx
|
| g(y) dy
|
|
|
|
|
/
|
/
C1 + C2* | e
dx +
| e
dy = 0
|
|
/
/
References
Goldstein and Braun, Advanced Methods for the Solution of Dierential Equations,
pp. 98
1416
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
http://www.maplesoft.com/support/help/Maple/view.aspx?path=odeadvisor/Liouville
# indirect doctest
Examples
>>>
>>>
>>>
>>>
...
from sympy import Function, dsolve, Eq, pprint
from sympy.abc import x
f = Function(f)
pprint(dsolve(diff(f(x), x, x) + diff(f(x), x)**2/f(x) +
diff(f(x), x)/x, f(x), hint=Liouville))
________________
________________
[f(x) = -\/ C1 + C2*log(x) , f(x) = \/ C1 + C2*log(x) ]
Riccati special minus2
sympy.solvers.ode.ode Riccati special minus2(eq, func, order, match)
The general Riccati equation has the form
dy/dx = f (x)y 2 + g(x)y + h(x).
While it does not have a general solution [1], the special form, dy/dx = ay 2 bxc , does
have solutions in many cases [2]. This routine returns a solution for a(dy/dx) = by 2 +cy/x+
d/x2 that is obtained by using a suitable change of variables to reduce it to the special
form and is valid when neither a nor b are zero and either c or d is zero.
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
from sympy.abc import x, y, a, b, c, d
from sympy.solvers.ode import dsolve, checkodesol
from sympy import pprint, Function
f = Function(f)
y = f(x)
genform = a*y.diff(x) - (b*y**2 + c*y/x + d/x**2)
sol = dsolve(genform, y)
pprint(sol, wrap_line=False)
/
/
__________________
\\
|
__________________
|
/
2
||
|
/
2
|
\/ 4*b*d - (a + c) *log(x)||
-|a + c - \/ 4*b*d - (a + c) *tan|C1 + ----------------------------||
\
\
2*a
//
f(x) = -----------------------------------------------------------------------2*b*x
>>> checkodesol(genform, sol, order=1)[0]
True
References
1.http://www.maplesoft.com/support/help/Maple/view.aspx?path=odeadvisor/Riccati
2.http://eqworld.ipmnet.ru/en/solutions/ode/ode0106.pdf - http://eqworld.ipmnet.ru/en/solutions/ode
5.27. ODE
1417
SymPy Documentation, Release 0.7.6
nth linear constant coeff homogeneous
sympy.solvers.ode.ode nth linear constant coeff homogeneous(eq,
func,
order, match, returns=sol)
Solves an nth order linear homogeneous dierential equation with constant coecients.
This is an equation of the form
an f (n) (x) + an1 f (n1) (x) + + a1 f 0 (x) + a0 f (x) = 0.
These equations can be solved in a general manner, by taking the roots of the characteristic equation an mn + an1 mn1 + + a1 m + a0 = 0. The solution will then be the sum of
Cn xi erx terms, for each where Cn is an arbitrary constant, r is a root of the characteristic
equation and i is one of each from 0 to the multiplicity of the root - 1 (for example, a
root 3 of multiplicity 2 would create the terms C1 e3x + C2 xe3x ). The exponential is usually
expanded for complex roots using Eulers equation eIx = cos(x) + I sin(x). Complex roots
always come in conjugate pairs in polynomials with real coecients, so the two roots
will be represented (after simplifying the constants) as eax (C1 cos(bx) + C2 sin(bx)).
If SymPy cannot nd exact roots to the characteristic equation, a RootOf (page 1119)
instance will be return instead.
>>> from sympy import Function, dsolve, Eq
>>> from sympy.abc import x
>>> f = Function(f)
>>> dsolve(f(x).diff(x, 5) + 10*f(x).diff(x) - 2*f(x), f(x),
... hint=nth_linear_constant_coeff_homogeneous)
...
f(x) == C1*exp(x*RootOf(_x**5 + 10*_x - 2, 0)) +
C2*exp(x*RootOf(_x**5 + 10*_x - 2, 1)) +
C3*exp(x*RootOf(_x**5 + 10*_x - 2, 2)) +
C4*exp(x*RootOf(_x**5 + 10*_x - 2, 3)) +
C5*exp(x*RootOf(_x**5 + 10*_x - 2, 4))
Note that because this method does not involve
nth linear constant coeff homogeneous Integral hint.
integration,
there
is
no
The following is for internal use:
returns = sol returns the solution to the ODE.
returns = list returns a list of linearly independent solutions, for use with
non homogeneous solution methods like variation of parameters and undetermined coecients. Note that, though the solutions should be linearly independent, this function does not explicitly check that. You can do assert simplify(wronskian(sollist)) != 0 to check for linear independence. Also, assert
len(sollist) == order will need to pass.
returns = both, return a dictionary {sol: <solution to ODE>, list:
<list of linearly independent solutions>}.
References
http://en.wikipedia.org/wiki/Linear dierential equation
neous equation with constant coecients
section:
Nonhomoge-
M. Tenenbaum & H. Pollard, Ordinary Dierential Equations, Dover 1963, pp. 211
1418
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
# indirect doctest
Examples
>>>
>>>
>>>
>>>
...
...
from sympy import Function, dsolve, pprint
from sympy.abc import x
f = Function(f)
pprint(dsolve(f(x).diff(x, 4) + 2*f(x).diff(x, 3) 2*f(x).diff(x, 2) - 6*f(x).diff(x) + 5*f(x), f(x),
hint=nth_linear_constant_coeff_homogeneous))
x
-2*x
f(x) = (C1 + C2*x)*e + (C3*sin(x) + C4*cos(x))*e
nth linear constant coeff undetermined coefficients
sympy.solvers.ode.ode nth linear constant coeff undetermined coefficients(eq,
func,
order,
match)
Solves an nth order linear dierential equation with constant coecients using the
method of undetermined coecients.
This method works on dierential equations of the form
an f (n) (x) + an1 f (n1) (x) + + a1 f 0 (x) + a0 f (x) = P (x),
where P (x) is a function that has a nite number of linearly independent derivatives.
Functions that t this requirement are nite sums functions of the form axi ebx sin(cx + d)
or axi ebx cos(cx + d), where i is a non-negative integer and a, b, c, and d are constants.
For example any polynomial in x, functions like x2 e2x , x sin(x), and ex cos(x) can all be
used. Products of sins and coss have a nite number of derivatives, because they can
be expanded into sin(ax) and cos(bx) terms. However, SymPy currently cannot do that
expansion, so you will need to manually rewrite the expression in terms of the above
to use this method. So, for example, you will need to manually convert sin2 (x) into (1 +
cos(2x))/2 to properly apply the method of undetermined coecients on it.
This method works by creating a trial function from the expression and all of its linear
independent derivatives and substituting them into the original ODE. The coecients for
each term will be a system of linear equations, which are be solved for and substituted,
giving the solution. If any of the trial functions are linearly dependent on the solution
to the homogeneous equation, they are multiplied by sucient x to make them linearly
independent.
References
http://en.wikipedia.org/wiki/Method of undetermined coecients
M. Tenenbaum & H. Pollard, Ordinary Dierential Equations, Dover 1963, pp. 221
# indirect doctest
5.27. ODE
1419
SymPy Documentation, Release 0.7.6
Examples
>>>
>>>
>>>
>>>
...
...
from sympy import Function, dsolve, pprint, exp, cos
from sympy.abc import x
f = Function(f)
pprint(dsolve(f(x).diff(x, 2) + 2*f(x).diff(x) + f(x) 4*exp(-x)*x**2 + cos(2*x), f(x),
hint=nth_linear_constant_coeff_undetermined_coefficients))
/
4\
|
x | -x
4*sin(2*x)
3*cos(2*x)
f(x) = |C1 + C2*x + --|*e
- ---------- + ---------\
3 /
25
25
nth linear constant coeff variation of parameters
sympy.solvers.ode.ode nth linear constant coeff variation of parameters(eq,
func,
order,
match)
Solves an nth order linear dierential equation with constant coecients using the
method of variation of parameters.
This method works on any dierential equations of the form
f (n) (x) + an1 f (n1) (x) + + a1 f 0 (x) + a0 f (x) = P (x).
This method works by assuming that the particular solution takes the form
n
ci (x)yi (x),
x=1
where yi is the ith solution to the homogeneous equation. The solution is then solved
using Wronskians and Cramers Rule. The particular solution is given by
)
n (
Wi (x)
dx yi (x),
W (x)
x=1
where W (x) is the Wronskian of the fundamental system (the system of n linearly independent solutions to the homogeneous equation), and Wi (x) is the Wronskian of the
fundamental system with the ith column replaced with [0, 0, , 0, P (x)].
This method is general enough to solve any nth order inhomogeneous linear dierential equation with constant coecients, but sometimes SymPy cannot simplify the Wronskian well enough to integrate it.
If this method hangs,
try
using
the
nth linear constant coeff variation of parameters Integral
hint
and
simplifying
the
integrals
manually.
Also,
prefer
using
nth linear constant coeff undetermined coefficients when it applies, because
it doesnt use integration, making it faster and more reliable.
Warning, using simplify=False with nth linear constant coe variation of parameters
in dsolve() (page 1397) may cause it to hang, because it will not attempt to simplify the
1420
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Wronskian before integrating. It is recommended that you only use simplify=False with
nth linear constant coe variation of parameters Integral for this method, especially if
the solution to the homogeneous equation has trigonometric functions in it.
References
http://en.wikipedia.org/wiki/Variation of parameters
http://planetmath.org/VariationOfParameters
M. Tenenbaum & H. Pollard, Ordinary Dierential Equations, Dover 1963, pp. 233
# indirect doctest
Examples
>>>
>>>
>>>
>>>
...
...
from sympy import Function, dsolve, pprint, exp, log
from sympy.abc import x
f = Function(f)
pprint(dsolve(f(x).diff(x, 3) - 3*f(x).diff(x, 2) +
3*f(x).diff(x) - f(x) - exp(x)*log(x), f(x),
hint=nth_linear_constant_coeff_variation_of_parameters))
/
3
\
|
2
x *(6*log(x) - 11)| x
f(x) = |C1 + C2*x + C3*x + ------------------|*e
\
36
/
separable
sympy.solvers.ode.ode separable(eq, func, order, match)
Solves separable 1st order dierential equations.
dy
This is any dierential equation that can be written as P (y)
dx = Q(x). The solution can
then just be found by rearranging terms and integrating: P (y) dy = Q(x) dx. This hint
uses sympy.simplify.simplify.separatevars() (page 1331) as its back end, so if a
separable equation is not caught by this solver, it is most likely the fault of that function. separatevars() (page 1331) is smart enough to do most expansion and factoring
necessary to convert a separable equation F (x, y) into the proper form P (x) Q(y). The
general solution is:
>>>
>>>
>>>
>>>
>>>
from sympy import Function, dsolve, Eq, pprint
from sympy.abc import x
a, b, c, d, f = map(Function, [a, b, c, d, f])
genform = Eq(a(x)*b(f(x))*f(x).diff(x), c(x)*d(f(x)))
pprint(genform)
d
a(x)*b(f(x))*--(f(x)) = c(x)*d(f(x))
dx
>>> pprint(dsolve(genform, f(x), hint=separable_Integral))
f(x)
/
/
|
|
| b(y)
| c(x)
| ---- dy = C1 + | ---- dx
5.27. ODE
1421
SymPy Documentation, Release 0.7.6
|
|
d(y)
| a(x)
|
/
References
M. Tenenbaum & H. Pollard, Ordinary Dierential Equations, Dover 1963, pp. 52
# indirect doctest
Examples
>>> from sympy import Function, dsolve, Eq
>>> from sympy.abc import x
>>> f = Function(f)
>>> pprint(dsolve(Eq(f(x)*f(x).diff(x) + x, 3*x*f(x)**2), f(x),
... hint=separable, simplify=False))
/
2
\
2
log\3*f (x) - 1/
x
---------------- = C1 + -6
2
almost linear
sympy.solvers.ode.ode almost linear(eq, func, order, match)
Solves an almost-linear dierential equation.
The general form of an almost linear dierential equation is
f (x)g(y)y + k(x)l(y) + m(x) = 0wherel0 (y) = g(y).
This can be solved by substituting l(y) = u(y). Making the given substitution reduces it
to a linear dierential equation of the form u0 + P (x)u + Q(x) = 0.
The general solution is
>>>
>>>
>>>
>>>
>>>
from sympy import Function, dsolve, Eq, pprint
from sympy.abc import x, y, n
f, g, k, l = map(Function, [f, g, k, l])
genform = Eq(f(x)*(l(y).diff(y)) + k(x)*l(y) + g(x))
pprint(genform)
d
f(x)*--(l(y)) + g(x) + k(x)*l(y) = 0
dy
>>> pprint(dsolve(genform, hint = almost_linear))
/
//
-y*g(x)
\\
|
||
-------for k(x) = 0||
|
||
f(x)
|| -y*k(x)
|
||
|| -------|
||
y*k(x)
||
f(x)
l(y) = |C1 + |<
-----||*e
|
||
f(x)
||
|
||-g(x)*e
||
1422
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
|
|
\
||-------------||
k(x)
\\
otherwise
||
||
//
See Also:
sympy.solvers.ode.ode 1st linear() (page 1414)
References
Joel Moses, Symbolic Integration - The Stormy Decade, Communications of the
ACM, Volume 14, Number 8, August 1971, pp. 558
Examples
>>> from sympy import Function, Derivative, pprint
>>> from sympy.solvers.ode import dsolve, classify_ode
>>> from sympy.abc import x
>>> f = Function(f)
>>> d = f(x).diff(x)
>>> eq = x*d + x*f(x) + 1
>>> dsolve(eq, f(x), hint=almost_linear)
f(x) == (C1 - Ei(x))*exp(-x)
>>> pprint(dsolve(eq, f(x), hint=almost_linear))
-x
f(x) = (C1 - Ei(x))*e
linear coefficients
sympy.solvers.ode.ode linear coefficients(eq, func, order, match)
Solves a dierential equation with linear coecients.
The general form of a dierential equation with linear coecients is
(
)
a1 x + b1 y + c 1
0
y +F
= 0,
a2 x + b2 y + c 2
where a1 , b1 , c1 , a2 , b2 , c2 are constants and a1 b2 a2 b1 6= 0.
This can be solved by substituting:
b2 c1 b1 c2
a2 b1 a1 b2
a1 c2 a2 c1
.
y = y0 +
a2 b1 a1 b2
x = x0 +
This substitution reduces the equation to a homogeneous dierential equation.
See Also:
sympy.solvers.ode.ode 1st homogeneous coeff best()
(page
1410),
sympy.solvers.ode.ode 1st homogeneous coeff subs indep div dep() (page 1412),
sympy.solvers.ode.ode 1st homogeneous coeff subs dep div indep() (page 1411)
5.27. ODE
1423
SymPy Documentation, Release 0.7.6
References
Joel Moses, Symbolic Integration - The Stormy Decade, Communications of the
ACM, Volume 14, Number 8, August 1971, pp. 558
Examples
>>> from sympy import Function, Derivative, pprint
>>> from sympy.solvers.ode import dsolve, classify_ode
>>> from sympy.abc import x
>>> f = Function(f)
>>> df = f(x).diff(x)
>>> eq = (x + f(x) + 1)*df + (f(x) - 6*x + 1)
>>> dsolve(eq, hint=linear_coefficients)
[f(x) == -x - sqrt(C1 + 7*x**2) - 1, f(x) == -x + sqrt(C1 + 7*x**2) - 1]
>>> pprint(dsolve(eq, hint=linear_coefficients))
___________
___________
/
2
/
2
[f(x) = -x - \/ C1 + 7*x
- 1, f(x) = -x + \/ C1 + 7*x
- 1]
separable reduced
sympy.solvers.ode.ode separable reduced(eq, func, order, match)
Solves a dierential equation that can be reduced to the separable form.
The general form of this equation is
y 0 + (y/x)H(xn y) = 0.
This can be solved by substituting u(y) = xn y. The equation then reduces to the separable
u0
form u(powerH(u))
x1 = 0.
The general solution is:
>>>
>>>
>>>
>>>
>>>
from sympy import Function, dsolve, Eq, pprint
from sympy.abc import x, n
f, g = map(Function, [f, g])
genform = f(x).diff(x) + (f(x)/x)*g(x**n*f(x))
pprint(genform)
/ n
\
d
f(x)*g\x *f(x)/
--(f(x)) + --------------dx
x
>>> pprint(dsolve(genform, hint=separable_reduced))
n
x *f(x)
/
|
|
1
|
------------ dy = C1 + log(x)
|
y*(n - g(y))
|
/
1424
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
See Also:
sympy.solvers.ode.ode separable() (page 1421)
References
Joel Moses, Symbolic Integration - The Stormy Decade, Communications of the
ACM, Volume 14, Number 8, August 1971, pp. 558
Examples
>>> from sympy import Function, Derivative, pprint
>>> from sympy.solvers.ode import dsolve, classify_ode
>>> from sympy.abc import x
>>> f = Function(f)
>>> d = f(x).diff(x)
>>> eq = (x - x**2*f(x))*d - f(x)
>>> dsolve(eq, hint=separable_reduced)
[f(x) == (-sqrt(C1*x**2 + 1) + 1)/x, f(x) == (sqrt(C1*x**2 + 1) + 1)/x]
>>> pprint(dsolve(eq, hint=separable_reduced))
___________
___________
/
2
/
2
- \/ C1*x + 1 + 1
\/ C1*x + 1 + 1
[f(x) = --------------------, f(x) = ------------------]
x
x
lie group
sympy.solvers.ode.ode lie group(eq, func, order, match)
This hint implements the Lie group method of solving rst order dierential equations.
The aim is to convert the given dierential equation from the given coordinate given system into another coordinate system where it becomes invariant under the one-parameter
Lie group of translations. The converted ODE is quadrature and can be solved easily.
It makes use of the sympy.solvers.ode.infinitesimals() (page 1404) function which
returns the innitesimals of the transformation.
The coordinates r and s can be found by solving the following Partial Dierential Equations.
r
r
+
=0
x
y
s
s
+
=1
x
y
The dierential equation becomes separable in the new coordinate system
ds
=
dr
5.27. ODE
s
x
r
x
s
+ h(x, y) y
r
+ h(x, y) y
1425
SymPy Documentation, Release 0.7.6
After nding the solution by integration, it is then converted back to the original coordinate system by subsituting r and s in terms of x and y again.
References
Solving dierential equations by Symmetry Groups, John Starrett, pp. 1 - pp. 14
Examples
>>>
>>>
>>>
>>>
...
from sympy import Function, dsolve, Eq, exp, pprint
from sympy.abc import x
f = Function(f)
pprint(dsolve(f(x).diff(x) + 2*x*f(x) - x*exp(-x**2), f(x),
hint=lie_group))
/
2\
2
|
x | -x
f(x) = |C1 + --|*e
\
2 /
1st power series
sympy.solvers.ode.ode 1st power series(eq, func, order, match)
The power series solution is a method which gives the Taylor series expansion to the
solution of a dierential equation.
dy
For a rst order dierential equation dx
= h(x, y), a power series solution exists at a point
x = x0 if h(x, y) is analytic at x0 . The solution is given by
y(x) = y(x0 ) +
Fn (x0 , b)(x x0 )n
,
n!
n=1
where y(x0 ) = b is the value of y at the initial value of x0 . To compute the values of
the Fn (x0 , b) the following algorithm is followed, until the required number of terms are
generated.
1.F1 = h(x0 , b)
2.Fn+1 =
Fn
x
Fn
y F1
References
Travis W. Walker, Analytic power series technique for solving rst-order dierential
equations, p.p 17, 18
Examples
1426
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>>
>>>
>>>
>>>
>>>
>>>
from sympy import Function, Derivative, pprint, exp
from sympy.solvers.ode import dsolve
from sympy.abc import x
f = Function(f)
eq = exp(x)*(f(x).diff(x)) - f(x)
pprint(dsolve(eq, hint=1st_power_series))
3
4
5
C1*x
C1*x
C1*x
/ 6\
f(x) = C1 + C1*x - ----- + ----- + ----- + O\x /
6
24
60
2nd power series ordinary
sympy.solvers.ode.ode 2nd power series ordinary(eq, func, order, match)
Gives a power series solution to a second order homogeneous dierential equation with
polynomial coecients at an ordinary point. A homogenous dierential equation is of
the form
P (x)
d2 y
dy
+ R(x) = 0
+ Q(x)
dx2
dx
For simplicity it is assumed that P (x), Q(x) and R(x) are polynomials, it is sucient that
Q(x)
R(x)
n
n=0 an x ,
P (x) and P (x) exists at x0 . A recurrence relation is obtained by substituting y as
in the dierential equation, and equating the nth term. Using this relation various terms
can be generated.
References
http://tutorial.math.lamar.edu/Classes/DE/SeriesSolutions.aspx
George E. Simmons, Dierential Equations with Applications and Historical Notes,
p.p 176 - 184
Examples
>>>
>>>
>>>
>>>
>>>
from sympy import dsolve, Function, pprint
from sympy.abc import x, y
f = Function(f)
eq = f(x).diff(x, 2) + f(x)
pprint(dsolve(eq, hint=2nd_power_series_ordinary))
/ 4
2
\
/
2
\
|x
x
|
| x
|
/ 6\
f(x) = C2*|-- - -- + 1| + C1*x*|- -- + 1| + O\x /
\24
2
/
\ 6
/
2nd power series regular
sympy.solvers.ode.ode 2nd power series regular(eq, func, order, match)
Gives a power series solution to a second order homogeneous dierential equation with
5.27. ODE
1427
SymPy Documentation, Release 0.7.6
polynomial coecients at a regular point. A second order homogenous dierential equation is of the form
P (x)
d2 y
dy
+ Q(x)
+ R(x) = 0
dx2
dx
2 R(x)
A point is said to regular singular at x0 if x x0 Q(x)
P (x) and (x x0) P (x) are analytic at x0. For
simplicity P (x), Q(x) and R(x) are assumed to be polynomials. The algorithm for nding
the power series solutions is:
1.Try expressing (x x0)P (x) and ((x x0)2 )Q(x) as power series solutions about x0.
Find p0 and q0 which are the constants of the power series expansions.
2.Solve the indicial equation f (m) = m(m 1) + m p0 + q0, to obtain the roots m1 and
m2 of the indicial equation.
3.If m1 m2 is a non integer there exists two series solutions. If m1 = m2, there exists
only one solution. If m1 m2 is an integer, then the existence of one solution is
conrmed. The other solution may or may not exist.
The power series solution is of the form xm n=0 an xn . The coecients are determined
n1
+(m+k)p
nk
by the following recurrence relation. an = k=0 nk
. For the case in which
f (m+n)
m1 m2 is an integer, it can be seen from the recurrence relation that for the lower root
m, when n equals the dierence of both the roots, the denominator becomes zero. So if
the numerator is not equal to zero, a second series solution exists.
References
George E. Simmons, Dierential Equations with Applications and Historical Notes,
p.p 176 - 184
Examples
>>>
>>>
>>>
>>>
>>>
from sympy import dsolve, Function, pprint
from sympy.abc import x, y
f = Function(f)
eq = x*(f(x).diff(x, 2)) + 2*(f(x).diff(x)) + x*f(x)
pprint(dsolve(eq))
/
6
4
2
\
|
x
x
x
|
/ 4
2
\
C1*|- --- + -- - -- + 1|
| x
x
|
\ 720
24
2
/
/ 6\
f(x) = C2*|--- - -- + 1| + ------------------------ + O\x /
\120
6
/
x
5.27.3 Lie heuristics
These functions are intended for internal use of the Lie Group Solver. Nonetheless, they
contain useful information in their docstrings on the algorithms implemented for the various
heuristics.
1428
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
abaco1 simple
sympy.solvers.ode.lie heuristic abaco1 simple(match, comp=False)
The rst heuristic uses the following four sets of assumptions on and
= 0, = f (x)
= 0, = f (y)
= f (x), = 0
= f (y), = 0
The success of this heuristic is determined by algebraic factorisation. For the rst assumption = 0 and to be a function of x, the PDE
h
h
+(
)h
h2
=0
x
y
x
y
x
y
h
reduces to f 0 (x)f h
y = 0 If y is a function of x, then this can usually be integrated easily.
A similar idea is applied to the other 3 assumptions as well.
References
E.S Cheb-Terrab, L.G.S Duarte and L.A,C.P da Mota, Computer Algebra Solving of
First Order ODEs Using Symmetry Methods, pp. 8
abaco1 product
sympy.solvers.ode.lie heuristic abaco1 product(match, comp=False)
The second heuristic uses the following two assumptions on and
= 0, = f (x) g(y)
= f (x) g(y), = 0
2
log(h) is separable in x and y,
The rst assumption of this heuristic holds good if h12 xy
then the separated factors containing x is f (x), and g(y) is obtained by
provided f x
1
f h
1
f x
( f h
) dy
)
is a function of y only.
dy
dy
1
The second assumption holds good if dx
= h(x, y) is rewritten as dx
= h(y,x)
and the same
properties of the rst assumption satisifes. After obtaining f (x) and g(y), the coordinates
are again interchanged, to get as f (x) g(y)
5.27. ODE
1429
SymPy Documentation, Release 0.7.6
References
E.S. Cheb-Terrab, A.D. Roche, Symmetries and First Order ODE Patterns, pp. 7 pp. 8
bivariate
sympy.solvers.ode.lie heuristic bivariate(match, comp=False)
The third heuristic assumes the innitesimals and to be bi-variate polynomials in x
and y. The assumption made here for the logic below is that h is a rational function in x
and y though that may not be necessary for the innitesimals to be bivariate polynomials.
The coecients of the innitesimals are found out by substituting them in the PDE and
grouping similar terms that are polynomials and since they form a linear system, solve
and check for non trivial solutions. The degree of the assumed bivariates are increased
till a certain maximum value.
References
Lie Groups and Dierential Equations pp. 327 - pp. 329
chi
sympy.solvers.ode.lie heuristic chi(match, comp=False)
The aim of the fourth heuristic is to nd the function (x, y) that satisies the PDE
h
h d
dx y = 0.
d
dx
This assumes to be a bivariate polynomial in x and y. By intution, h should be a rational
function in x and y. The method used here is to substitute a general binomial for up to a
certain maximum degree is reached. The coecients of the polynomials, are calculated
by by collecting terms of the same order in x and y.
After nding , the next step is to use = h + , to determine and . This can be done
by dividing by h which would give as the quotient and as the remainder.
References
E.S Cheb-Terrab, L.G.S Duarte and L.A,C.P da Mota, Computer Algebra Solving of
First Order ODEs Using Symmetry Methods, pp. 8
abaco2 similar
sympy.solvers.ode.lie heuristic abaco2 similar(match, comp=False)
This heuristic uses the following two assumptions on and
= g(x), = f (x)
= f (y), = g(y)
For the rst assumption,
1430
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
1.First
h
y
2 h
yy
is calculated. Let us say this value is A
y
2.If this is constant, then h is matched to the form A(x) + B(x)e C then,
f (x) and A(x) f (x) gives g(x)
3.Otherwise
A
X
A
Y
A(x)
dx
C
B(x)
gives
= is calculated. If
a] is a function of x alone
h 0 (x) h
x
h+
b] y
g(x)
= G is a function of x alone. then, e
G dx
gives f (x) and f (x) gives
dy
dy
1
The second assumption holds good if dx
= h(x, y) is rewritten as dx
= h(y,x)
and the same
properties of the rst assumption satisifes. After obtaining f (x) and g(x), the coordinates
are again interchanged, to get as f (x ) and as g(y )
References
E.S. Cheb-Terrab, A.D. Roche, Symmetries and First Order ODE Patterns, pp. 10 pp. 12
function sum
sympy.solvers.ode.lie heuristic function sum(match, comp=False)
This heuristic uses the following two assumptions on and
= 0, = f (x) + g(y)
= f (x) + g(y), = 0
The rst assumption of this heuristic holds good if
2
[(h 2 (h1 ))1 ]
y
x
is separable in x and y,
1.The separated factors containing y is
g
y .
From this g(y) can be determined.
00
2.The separated factors containing x is f (x).
2
1
3.h x
) equals
2 (h
f 00 (x)
f (x)+g(y) .
From this f (x) can be determined.
dy
dy
1
The second assumption holds good if dx
= h(x, y) is rewritten as dx
= h(y,x)
and the same
properties of the rst assumption satisifes. After obtaining f (x) and g(y), the coordinates
are again interchanged, to get as f (x) + g(y).
For both assumptions, the constant factors are separated among g(y) and f 00 (x), such
that f 00 (x) obtained from 3] is the same as that obtained from 2]. If not possible, then
this heuristic fails.
5.27. ODE
1431
SymPy Documentation, Release 0.7.6
References
E.S. Cheb-Terrab, A.D. Roche, Symmetries and First Order ODE Patterns, pp. 7 pp. 8
abaco2 unique unknown
sympy.solvers.ode.lie heuristic abaco2 unique unknown(match, comp=False)
This heuristic assumes the presence of unknown functions or known functions with noninteger powers.
1.A list of all functions and non-integer powers containing x and y
2.Loop over each element f in the list, nd
f
x
f
x
=R
If it is separable in x and y, let X be the factors containing x. Then
a] Check if = X and = X
R satisfy the PDE. If yes, then return and
b] Check if =
R
X
1
and = X
satisfy the PDE. If yes, then return and
If not, then check if
a] = R, = 1
b] = 1, = R1
are solutions.
References
E.S. Cheb-Terrab, A.D. Roche, Symmetries and First Order ODE Patterns, pp. 10 pp. 12
abaco2 unique general
sympy.solvers.ode.lie heuristic abaco2 unique general(match, comp=False)
This heuristic nds if innitesimals of the form = f (x), = g(y) without making any
assumptions on h.
The complete sequence of steps is given in the paper mentioned below.
References
E.S. Cheb-Terrab, A.D. Roche, Symmetries and First Order ODE Patterns, pp. 10 pp. 12
linear
sympy.solvers.ode.lie heuristic linear(match, comp=False)
This heuristic assumes
1. = ax + by + c and
2. = f x + gy + h
1432
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
After substituting the following assumptions in the determining PDE, it reduces to
f + (g a)h bh2 (ax + by + c)
h
h
(f x + gy + c)
x
y
Solving the reduced PDE obtained, using the method of characteristics, becomes impractical. The method followed is grouping similar terms and solving the system of linear
equations obtained. The dierence between the bivariate heuristic is that h need not be
a rational function in this case.
References
E.S. Cheb-Terrab, A.D. Roche, Symmetries and First Order ODE Patterns, pp. 10 pp. 12
5.27.4 System of ODEs
These functions are intended for internal use by dsolve() (page 1397) for system of dierential equations.
system of odes linear 2eq order1 type1
sympy.solvers.ode. linear 2eq order1 type1(x, y, t, r)
It is classied under system of two linear homogeneous rst-order constant-coecient
ordinary dierential equations.
The equations which come under this type are
x0 = ax + by,
y 0 = cx + dy
The characteristics equation is written as
2 + (a + d) + ad bc = 0
and its discriminant is D = (a d)2 + 4bc. There are several cases
1. Case when ad bc 6= 0. The origin of coordinates, x = y = 0, is the only stationary point;
it is - a node if D = 0 - a node if D > 0 and ad bc > 0 - a saddle if D > 0 and ad bc < 0 - a
focus if D < 0 and a + d 6= 0 - a centre if D < 0 and a + d 6= 0.
1.1. If D > 0. The characteristic equation has two distinct real roots 1 and 2 . The
general solution of the system in question is expressed as
x = C1 be1 t + C2 be2 t
5.27. ODE
1433
SymPy Documentation, Release 0.7.6
y = c1 (1 a)e1 t + c2 (2 a)e2 t
where C1 and C2 being arbitary constants
1.2. If D < 0. The characteristics equation has two conjugate roots, 1 = + i and
2 = i. The general solution of the system is given by
x = bet (C1 sin(t) + C2 cos(t))
y = et ([( a)C1 C2 ] sin(t) + [C1 + ( a)C2 cos(t)])
1.3. If D = 0 and a 6= d. The characteristic equation has two equal roots, 1 = 2 . The
general solution of the system is written as
x = 2b(C1 +
a+d
C2
+ C2 t)e 2 t
ad
y = [(d a)C1 + C2 + (d a)C2 t]e
a+d
2 t
1.4. If D = 0 and a = d 6= 0 and b = 0
x = C1 eat , y = (cC1 t + C2 )eat
1.5. If D = 0 and a = d 6= 0 and c = 0
x = (bC1 t + C2 )eat , y = C1 eat
2. Case when ad bc = 0 and a2 + b2 > 0. The whole straight line ax + by = 0 consists of
singular points. The orginal system of dierential equaitons can be rewritten as
x0 = ax + by, y 0 = k(ax + by)
2.1 If a + bk 6= 0, solution will be
x = bC1 + C2 e(a+bk)t , y = aC1 + kC2 e(a+bk)t
2.2 If a + bk = 0, solution will be
x = C1 (bkt 1) + bC2 t, y = k 2 bC1 t + (bk 2 t + 1)C2
1434
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
system of odes linear 2eq order1 type2
sympy.solvers.ode. linear 2eq order1 type2(x, y, t, r)
The equations of this type are
x0 = ax + by + k1, y 0 = cx + dy + k2
The general solution og this system is given by sum of its particular solution and the
general solution of the corresponding homogeneous system is obtained from type1.
1. When ad bc 6= 0. The particular solution will be x = x0 and y = y0 where x0 and y0 are
determined by solving linear system of equations
ax0 + by0 + k1 = 0, cx0 + dy0 + k2 = 0
2.When ad bc = 0 and a2 + b2 > 0. In this case, the system of equation becomes
x0 = ax + by + k1 , y 0 = k(ax + by) + k2
2.1 If = a + bk 6= 0, particular solution is given by
x = b 1 (c1 k c2 )t 2 (ac1 + bc2 )
y = kx + (c2 c1 k)t
2.2 If = a + bk = 0, particular solution is given by
x=
1
b(c2 c1 k)t2 + c1 t
2
y = kx + (c2 c1 k)t
system of odes linear 2eq order1 type3
sympy.solvers.ode. linear 2eq order1 type3(x, y, t, r)
The equations of this type of ode are
x0 = f (t)x + g(t)y
y 0 = g(t)x + f (t)y
5.27. ODE
1435
SymPy Documentation, Release 0.7.6
The solution of such equations is given by
x = eF (C1 eG + C2 eG ), y = eF (C1 eG C2 eG )
where C1 and C2 are arbitary constants, and
F = f (t) dt, G = g(t) dt
system of odes linear 2eq order1 type4
sympy.solvers.ode. linear 2eq order1 type4(x, y, t, r)
The equations of this type of ode are .
x0 = f (t)x + g(t)y
y 0 = g(t)x + f (t)y
The solution is given by
x = F (C1 cos(G) + C2 sin(G)), y = F (C1 sin(G) + C2 cos(G))
where C1 and C2 are arbitary constants, and
F = f (t) dt, G = g(t) dt
system of odes linear 2eq order1 type5
sympy.solvers.ode. linear 2eq order1 type5(x, y, t, r)
The equations of this type of ode are .
x0 = f (t)x + g(t)y
y 0 = ag(t)x + [f (t) + bg(t)]y
The transformation of
x=e
f (t) dt
u, y = e
f (t) dt
v, T =
g(t) dt
leads to a system of constant coecient linear dierential equations
u0 (T ) = v, v 0 (T ) = au + bv
1436
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
system of odes linear 2eq order1 type6
sympy.solvers.ode. linear 2eq order1 type6(x, y, t, r)
The equations of this type of ode are .
x0 = f (t)x + g(t)y
y 0 = a[f (t) + ah(t)]x + a[g(t) h(t)]y
This is solved by rst multiplying the rst equation by a and adding it to the second
equation to obtain
y 0 ax0 = ah(t)(y ax)
Setting U = y ax and integrating the equation we arrive at
y ax = C1 ea
h(t) dt
and on substituing the value of y in rst equation give rise to rst order ODEs. After
solving for x, we can obtain y by substituting the value of x in second equation.
system of odes linear 2eq order1 type7
sympy.solvers.ode. linear 2eq order1 type7(x, y, t, r)
The equations of this type of ode are .
x0 = f (t)x + g(t)y
y 0 = h(t)x + p(t)y
Dierentiating the rst equation and substituting the value of y from second equation
will give a second-order linear equation
gx00 (f g + gp + g 0 )x0 + (f gp g 2 h + f g 0 f 0 g)x = 0
This above equation can be easily integrated if following conditions are satised.
1.f gp g 2 h + f g 0 f 0 g = 0
2.f gp g 2 h + f g 0 f 0 g = ag, f g + gp + g 0 = bg
If rst condition is satised then it is solved by current dsolve solver and in second case it
becomes a constant cocient dierential equation which is also solved by current solver.
Otherwise if the above condition fails then, a particular solution is assumed as x = x0 (t)
and y = y0 (t) Then the general solution is expressed as
g(t)F (t)P (t)
x = C1 x0 (t) + C2 x0 (t)
dt
x20 (t)
5.27. ODE
1437
SymPy Documentation, Release 0.7.6
y = C1 y0 (t) + C2 [
F (t)P (t)
+ y0 (t)
x0 (t)
g(t)F (t)P (t)
dt]
x20 (t)
where C1 and C2 are arbitary constants and
F (t) = e
f (t) dt
, P (t) = e
p(t) dt
system of odes linear 2eq order2 type1
sympy.solvers.ode. linear 2eq order2 type1(x, y, t, r)
System of two constant-coecient second-order linear homogeneous dierential equations
x00 = ax + by
y 00 = cx + dy
The charecteristic equation for above equations
4 (a + d)2 + ad bc = 0
whose discriminant is D = (a d)2 + 4bc 6= 0
1.When ad bc 6= 0
1.1. If D 6= 0. The characteristic equation has four distict roots, 1 , 2 , 3 , 4 . The general
solution of the system is
x = C1 be1 t + C2 be2 t + C3 be3 t + C4 be4 t
y = C1 (21 a)e1 t + C2 (22 a)e2 t + C3 (23 a)e3 t + C4 (24 a)e4 t
where C1 , ..., C4 are arbitary constants.
1.2. If D = 0 and a 6= d:
x = 2C1 (bt +
kt
kt
kt
kt
kt
2bk
2bk
)e 2 + 2C2 (bt +
)e 2 + 2bC3 te 2 + 2bC4 te 2
ad
ad
y = C1 (d a)te 2 + C2 (d a)te
kt
2
kt
+ C3 [(d a)t + 2k]e 2 + C4 [(d a)t 2k]e
where C1 , ..., C4 are arbitary constants and k =
1438
kt
2
2(a + d)
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
1.3. If D = 0 and a = d 6= 0 and b = 0:
x = 2 aC1 e at + 2 aC2 e at
y = cC1 te
at
cC2 te
at
+ C3 e
at
+ C4 e
at
1.4. If D = 0 and a = d 6= 0 and c = 0:
x = bC1 te
at
bC2 te
at
+ C3 e
at
+ C4 e
at
y = 2 aC1 e at + 2 aC2 e at
2.When ad bc = 0 and a2 + b2 > 0. Then the original system becomes
x00 = ax + by
y 00 = k(ax + by)
2.1. If a + bk 6= 0:
x = C1 et
y = C1 ket
a+bk
a+bk
+ C2 et
a+bk
+ C2 ket
+ C3 bt + C4 b
a+bk
C3 at C4 a
2.2. If a + bk = 0:
x = C1 bt3 + C2 bt2 + C3 t + C4
y = kx + 6C1 t + 2C2
system of odes linear 2eq order2 type2
sympy.solvers.ode. linear 2eq order2 type2(x, y, t, r)
The equations in this type are
x00 = a1 x + b1 y + c1
5.27. ODE
1439
SymPy Documentation, Release 0.7.6
y 00 = a2 x + b2 y + c2
The general solution of this system is given by the sum of its particular solution and the
general solution of the homogeneous system. The general solution is given by the linear
system of 2 equation of order 2 and type 1
1. If a1 b2 a2 b1 6= 0. A particular solution will be x = x0 and y = y0 where the constants x0
and y0 are determined by solving the linear algebraic system
a1 x0 + b1 y0 + c1 = 0, a2 x0 + b2 y0 + c2 = 0
2.If a1 b2 a2 b1 = 0 and a21 + b21 > 0. In this case, the system in question becomes
x00 = ax + by + c1 , y 00 = k(ax + by) + c2
2.1. If = a + bk 6= 0, the particular solution will be
x=
1 1
b (c1 k c2 )t2 2 (ac1 + bc2 )
2
1
y = kx + (c2 c1 k)t2
2
2.2. If = a + bk = 0, the particular solution will be
x=
1
1
b(c2 c1 k)t4 + c1 t2
24
2
1
y = kx + (c2 c1 k)t2
2
system of odes linear 2eq order2 type3
sympy.solvers.ode. linear 2eq order2 type3(x, y, t, r)
These type of equation is used for describing the horizontal motion of a pendulum taking
into account the Earth rotation. The solution is given with a2 + 4b > 0:
x = C1 cos(t) + C2 sin(t) + C3 cos(t) + C4 sin(t)
y = C1 sin(t) + C2 cos(t) C3 sin(t) + C4 cos(t)
where C1 , ..., C4 and
=
1440
1 2
1 2
1
1
a+
a + 4b, = a
a + 4b
2
2
2
2
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
system of odes linear 2eq order2 type4
sympy.solvers.ode. linear 2eq order2 type4(x, y, t, r)
These equations are found in the theory of oscillations
x00 + a1 x0 + b1 y 0 + c1 x + d1 y = k1 eit
y 00 + a2 x0 + b2 y 0 + c2 x + d2 y = k2 eit
The general solution of this linear nonhomogeneous system of constant-coecient differential equations is given by the sum of its particular solution and the general solution
of the corresponding homogeneous system (with k1 = k2 = 0)
1.A particular solution is obtained by the method of undetermined coecients:
x = A eit , y = B eit
On substituting these expressions into the original system of dierential equations, one
arrive at a linear nonhomogeneous system of algebraic equations for the coecients A
and B.
2. The general solution of the homogeneous system of dierential equations is determined by a linear combination of linearly independent particular solutions determined
by the method of undetermined coecients in the form of exponentials:
x = Aet , y = Bet
On substituting these expressions into the original system and colleting the coecients
of the unknown A and B, one obtains
(2 + a1 + c1 )A + (b1 + d1 )B = 0
(a2 + c2 )A + (2 + b2 + d2 )B = 0
The determinant of this system must vanish for nontrivial solutions A, B to exist. This
requirement results in the following characteristic equation for
(2 + a1 + c1 )(2 + b2 + d2 ) (b1 + d1 )(a2 + c2 ) = 0
If all roots k1 , ..., k4 of this equation are distict, the general solution of the original system
of the dierential equations has the form
x = C1 (b1 1 + d1 )e1 t C2 (b1 2 + d1 )e2 t C3 (b1 3 + d1 )e3 t C4 (b1 4 + d1 )e4 t
y = C1 (21 + a1 1 + c1 )e1 t + C2 (22 + a1 2 + c1 )e2 t + C3 (23 + a1 3 + c1 )e3 t + C4 (24 + a1 4 + c1 )e4 t
5.27. ODE
1441
SymPy Documentation, Release 0.7.6
system of odes linear 2eq order2 type5
sympy.solvers.ode. linear 2eq order2 type5(x, y, t, r)
The equation which come under this catagory are
x00 = a(ty 0 y)
y 00 = b(tx0 x)
The transformation
u = tx0 x, b = ty 0 y
leads to the rst-order system
u0 = atv, v 0 = btu
The general solution of this system is given by
If ab > 0:
1
u = C1 ae 2
abt2
+ C2 ae 2
abt2
2
2
1
1
v = C1 abe 2 abt C2 abe 2 abt
If ab < 0:
u = C1 a cos(
v = C1
1
1
|ab|t2 ) + C2 a sin(
|ab|t2 )
2
2
1
1
|ab| sin(
|ab|t2 ) + C2 |ab| cos(
|ab|t2 )
2
2
where C1 and C2 are arbitary constants. On substituting the value of u and v in above
equations and integrating the resulting expressions, the general solution will become
u
u
dt, y = C4 t + t
dt
x = C3 t + t
t2
t2
where C3 and C4 are arbitrary constants.
1442
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
system of odes linear 2eq order2 type6
sympy.solvers.ode. linear 2eq order2 type6(x, y, t, r)
The equations are
x00 = f (t)(a1 x + b1 y)
y 00 = f (t)(a2 x + b2 y)
If k1 and k2 are roots of the quadratic equation
k 2 (a1 + b2 )k + a1 b2 a2 b1 = 0
Then by multiplying appropriate constants and adding together original equations we
obtain two independent equations:
z100 = k1 f (t)z1 , z1 = a2 x + (k1 a1 )y
z200 = k2 f (t)z2 , z2 = a2 x + (k2 a1 )y
Solving the equations will give the values of x and y after obtaining the value of z1 and
z2 by solving the dierential equation and substuting the result.
system of odes linear 2eq order2 type7
sympy.solvers.ode. linear 2eq order2 type7(x, y, t, r)
The equations are given as
x00 = f (t)(a1 x0 + b1 y 0 )
y 00 = f (t)(a2 x0 + b2 y 0 )
If k1 and k 2 are roots of the quadratic equation
k 2 (a1 + b2 )k + a1 b2 a2 b1 = 0
Then the system can be reduced by adding together the two equations multiplied by
appropriate constants give following two independent equations:
z100 = k1 f (t)z10 , z1 = a2 x + (k1 a1 )y
5.27. ODE
1443
SymPy Documentation, Release 0.7.6
z200 = k2 f (t)z20 , z2 = a2 x + (k2 a1 )y
Integrating these and returning to the original variables, one arrives at a linear algebraic
system for the unknowns x and y:
a2 x + (k1 a1 )y = C1 ek1 F (t) dt + C2
a2 x + (k2 a1 )y = C3
ek2 F (t) dt + C4
where C1 , ..., C4 are arbitrary constants and F (t) =
f (t) dt
system of odes linear 2eq order2 type8
sympy.solvers.ode. linear 2eq order2 type8(x, y, t, r)
The equation of this catagory are
x00 = af (t)(ty 0 y)
y 00 = bf (t)(tx0 x)
The transformation
u = tx0 x, v = ty 0 y
leads to the system of rst-order equations
u0 = atf (t)v, v 0 = btf (t)u
The general solution of this system has the form
If ab > 0:
u = C1 ae
ab tf (t) dt
+ C2 ae
ab
tf (t) dt
v = C1 abe ab tf (t) dt C2 abe ab tf (t) dt
If ab < 0:
u = C1 a cos(
1444
|ab| tf (t) dt) + C2 a sin( |ab| tf (t) dt)
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
v = C1
|ab| sin( |ab| tf (t) dt) + C2 |ab| cos( |ab| tf (t) dt)
where C1 and C2 are arbitary constants. On substituting the value of u and v in above
equations and integrating the resulting expressions, the general solution will become
u
u
x = C3 t + t
dt,
y
=
C
t
+
t
dt
4
t2
t2
where C3 and C4 are arbitrary constants.
system of odes linear 2eq order2 type9
sympy.solvers.ode. linear 2eq order2 type9(x, y, t, r)
t2 x00 + a1 tx0 + b1 ty 0 + c1 x + d1 y = 0
t2 y 00 + a2 tx0 + b2 ty 0 + c2 x + d2 y = 0
These system of equations are euler type.
The substitution of t = e ( 6= 0) leads to the system of constant coecient linear dierential equations
x00 + (a1 1)x0 + b1 y 0 + c1 x + d1 y = 0
y 00 + a2 x0 + (b2 1)y 0 + c2 x + d2 y = 0
The general solution of the homogeneous system of dierential equations is determined
by a linear combination of linearly independent particular solutions determined by the
method of undetermined coecients in the form of exponentials
x = Aet , y = Bet
On substituting these expressions into the original system and colleting the coecients
of the unknown A and B, one obtains
(2 + (a1 1) + c1 )A + (b1 + d1 )B = 0
(a2 + c2 )A + (2 + (b2 1) + d2 )B = 0
5.27. ODE
1445
SymPy Documentation, Release 0.7.6
The determinant of this system must vanish for nontrivial solutions A, B to exist. This
requirement results in the following characteristic equation for
(2 + (a1 1) + c1 )(2 + (b2 1) + d2 ) (b1 + d1 )(a2 + c2 ) = 0
If all roots k1 , ..., k4 of this equation are distict, the general solution of the original system
of the dierential equations has the form
x = C1 (b1 1 + d1 )e1 t C2 (b1 2 + d1 )e2 t C3 (b1 3 + d1 )e3 t C4 (b1 4 + d1 )e4 t
y = C1 (21 + (a1 1)1 + c1 )e1 t + C2 (22 + (a1 1)2 + c1 )e2 t + C3 (23 + (a1 1)3 + c1 )e3 t + C4 (24 + (a1 1)4 +
system of odes linear 2eq order2 type10
sympy.solvers.ode. linear 2eq order2 type10(x, y, t, r)
The equation of this catagory are
(t2 + t + )2 x00 = ax + by
(t2 + t + )2 y 00 = cx + dy
The transformation
t2
x
1
y
dt, u =
,v =
2
2
+ t +
|t + t + |
|t + t + |
leads to a constant coecient linear system of equations
1
u00 = (a + 2 )u + bv
4
1
v 00 = cu + (d + 2 )v
4
These system of equations obtained can be solved by type1 of System of two constantcoecient second-order linear homogeneous dierential equations.
system of odes linear 2eq order2 type11
sympy.solvers.ode. linear 2eq order2 type11(x, y, t, r)
The equations which comes under this type are
x00 = f (t)(tx0 x) + g(t)(ty 0 y)
1446
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
y 00 = h(t)(tx0 x) + p(t)(ty 0 y)
The transformation
u = tx0 x, v = ty 0 y
leads to the linear system of rst-order equations
u0 = tf (t)u + tg(t)v, v 0 = th(t)u + tp(t)v
On substituting the value of u and v in transformed equation gives value of x and y as
u
v
x = C3 t + t
dt,
y
=
C
t
+
t
dt.
4
2
t
t2
where C3 and C4 are arbitrary constants.
system of odes linear 3eq order1 type1
sympy.solvers.ode. linear 3eq order1 type1(x, y, z, t, r)
x0 = ax
y 0 = bx + cy
z 0 = dx + ky + pz
Solution of such equations are forward substitution. Solving rst equations gives the
value of x, substituting it in second and third equation and solving second equation gives
y and similarly substituting y in third equation give z.
x = C1 eat
y=
z=
bC1 at
e + C2 ect
ac
C1
bk
kC2 ct
(d +
)eat +
e + C3 ept
ap
ac
cp
where C1 , C2 and C3 are arbitrary constants.
5.27. ODE
1447
SymPy Documentation, Release 0.7.6
system of odes linear 3eq order1 type2
sympy.solvers.ode. linear 3eq order1 type2(x, y, z, t, r)
The equations of this type are
x0 = cy bz
y 0 = az cx
z 0 = bx ay
1.First integral:
ax + by + cz = A
(1)
x2 + y 2 + z 2 = B 2
(2)
where A and B are arbitrary constants. It follows from these integrals that the integral
lines are circles formed by the intersection of the planes (1) and sphere (2)
2.Solution:
x = aC0 + kC1 cos(kt) + (cC2 bC3 ) sin(kt)
y = bC0 + kC2 cos(kt) + (aC2 cC3 ) sin(kt)
z = cC0 + kC3 cos(kt) + (bC2 aC3 ) sin(kt)
where k = a2 + b2 + c2 and the four constants of integration, C1 , ..., C4 are constrained by
a single relation,
aC1 + bC2 + cC3 = 0
system of odes linear 3eq order1 type3
sympy.solvers.ode. linear 3eq order1 type3(x, y, z, t, r)
Equations of this system of ODEs
ax0 = bc(y z)
1448
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
by 0 = ac(z x)
cz 0 = ab(x y)
1.First integral:
a2 x + b2 y + c 2 z = A
where A is an arbitary constant. It follows that the integral lines are plane curves.
2.Solution:
x = C0 + kC1 cos(kt) + a1 bc(C2 C3 ) sin(kt)
y = C0 + kC2 cos(kt) + ab1 c(C3 C1 ) sin(kt)
z = C0 + kC3 cos(kt) + abc1 (C1 C2 ) sin(kt)
where k = a2 + b2 + c2 and the four constants of integration, C1 , ..., C4 are constrained by
a single relation
a2 C1 + b2 C2 + c2 C3 = 0
system of odes linear 3eq order1 type4
sympy.solvers.ode. linear 3eq order1 type4(x, y, z, t, r)
Equations:
x0 = (a1 f (t) + g(t))x + a2 f (t)y + a3 f (t)z
y 0 = b1 f (t)x + (b2 f (t) + g(t))y + b3 f (t)z
z 0 = c1 f (t)x + c2 f (t)y + (c3 f (t) + g(t))z
The transformation
x=e
5.27. ODE
g(t) dt
u, y = e
g(t) dt
v, z = e
g(t) dt
w, =
f (t) dt
1449
SymPy Documentation, Release 0.7.6
leads to the system of constant coecient linear dierential equations
u0 = a1 u + a2 v + a3 w
v 0 = b1 u + b2 v + b3 w
w0 = c1 u + c2 v + c3 w
These system of equations are solved by homogeneous linear system of constant coefcients of n equations of rst order. Then substituting the value of u, v and w in transformed equation gives value of x, y and z.
system of odes linear neq order1 type1
sympy.solvers.ode. linear neq order1 type1(match )
System of n rst-order constant-coecient linear nonhomogeneous dierential equation
yk0 = ak1 y1 + ak2 y2 + ... + akn yn ; k = 1, 2, ..., n
or that can be written as y~0 = A.~y where ~y is matrix of yk for k = 1, 2, ...n and A is a n n
matrix.
Since these equations are equivalent to a rst order homogeneous linear dierential
equation. So the general solution will contain n linearly independent parts and solution
will consist some type of exponential functions. Assuming y = ~v ert is a solution of the
system where ~v is a vector of coecients of y1 , ..., yn . Substituting y and y 0 = rvert into the
equation y~0 = A.~y , we get
r~v ert = A~v ert
r~v = A~v
where r comes out to be eigenvalue of A and vector ~v is the eigenvector of A corresponding to r. There are three possiblities of eigenvalues of A
n distinct real eigenvalues
complex conjugate eigenvalues
eigenvalues with multiplicity k
1. When all eigenvalues r1 , .., rn are distinct with n dierent eigenvectors v1 , ...vn then the
solution is given by
~y = C1 er1 t v~1 + C2 er2 t v~2 + ... + Cn ern t v~n
where C1 , C2 , ..., Cn are arbitrary constants.
1450
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
2. When some eigenvalues are complex then in order to make the solution real, we
take a llinear combination: if r = a + bi has an eigenvector ~v = w~1 + iw~2 then to obtain
real-valued solutions to the system, replace the complex-valued solutions erx~v with realvalued solution eax (w~1 cos(bx) w~2 sin(bx)) and for r = a bi replace the solution erx~v with
eax (w~1 sin(bx) + w~2 cos(bx))
3. If some eigenvalues are repeated. Then we get fewer than n linearly independent
eigenvectors, we miss some of the solutions and need to construct the missing ones. We
do this via generalized eigenvectors, vectors which are not eigenvectors but are close
enough that we can use to write down the remaining solutions. For a eigenvalue r with
eigenvector w
~ we obtain w~2 , ..., w~k using
(A rI).w~2 = w
~
(A rI).w~3 = w~2
..
.
(A rI).w~k = wk1
~
~ + tw~2 +
Then the solutions to the system for the eigenspace are ert [w],
~ ert [tw
~ + w~2 ], ert [ t2 w
k1
k2
t
rt t
w~3 ], ..., e [ (k1)! w
~ + (k2)! w~2 + ... + twk1
~ + w~k ]
So, If y~1 , ..., y~n are n solution of obtained from three categories of A, then general solution
to the system y~0 = A.~y
~y = C1 y~1 + C2 y~2 + + Cn y~n
system of odes nonlinear 2eq order1 type1
sympy.solvers.ode. nonlinear 2eq order1 type1(x, y, t, eq)
Equations:
x0 = xn F (x, y)
y 0 = g(y)F (x, y)
Solution:
x = (y),
5.27. ODE
1
dy = t + C2
g(y)F ((y), y)
1451
SymPy Documentation, Release 0.7.6
where
if n 6= 1
= [C1 + (1 n)
1
1
dy] 1n
g(y)
if n = 1
= C1 e
1
g(y)
dy
where C1 and C2 are arbitrary constants.
system of odes nonlinear 2eq order1 type2
sympy.solvers.ode. nonlinear 2eq order1 type2(x, y, t, eq)
Equations:
x0 = ex F (x, y)
y 0 = g(y)F (x, y)
Solution:
x = (y),
1
dy = t + C2
g(y)F ((y), y)
where
if 6= 0
1
= log(C1
if = 0
= C1 +
1
dy)
g(y)
1
dy
g(y)
where C1 and C2 are arbitrary constants.
system of odes nonlinear 2eq order1 type3
sympy.solvers.ode. nonlinear 2eq order1 type3(x, y, t, eq)
Autonomous system of general form
x0 = F (x, y)
1452
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
y 0 = G(x, y)
Assuming y = y(x, C1 ) where C1 is an arbitrary constant is the general solution of the
rst-order equation
F (x, y)yx0 = G(x, y)
Then the general solution of the original system of equations has the form
1
dx = t + C1
F (x, y(x, C1 ))
system of odes nonlinear 2eq order1 type4
sympy.solvers.ode. nonlinear 2eq order1 type4(x, y, t, eq)
Equation:
x0 = f1 (x)g1 (y)(x, y, t)
y 0 = f2 (x)g2 (y)(x, y, t)
First integral:
f2 (x)
dx
f1 (x)
g1 (y)
dy = C
g2 (y)
where C is an arbitrary constant.
On solving the rst integral for x (resp., y ) and on substituting the resulting expression into either equation of the original solution, one arrives at a rs-order equation for
determining y (resp., x ).
system of odes nonlinear 2eq order1 type5
sympy.solvers.ode. nonlinear 2eq order1 type5(func, t, eq)
Clairaut system of ODEs
x = tx0 + F (x0 , y 0 )
y = ty 0 + G(x0 , y 0 )
The following are solutions of the system
5.27. ODE
1453
SymPy Documentation, Release 0.7.6
(i) straight lines:
x = C1 t + F (C1 , C2 ), y = C2 t + G(C1 , C2 )
where C1 and C2 are arbitrary constants;
(ii) envelopes of the above lines;
(iii) continuously dierentiable lines made up from segments of the lines (i) and (ii).
system of odes nonlinear 3eq order1 type1
sympy.solvers.ode. nonlinear 3eq order1 type1(x, y, z, t, eq)
Equations:
ax0 = (b c)yz, by 0 = (c a)zx, cz 0 = (a b)xy
First Integrals:
ax2 + by 2 + cz 2 = C1
a2 x2 + b2 y 2 + c2 z 2 = C2
where C1 and C2 are arbitrary constants. On solving the integrals for y and z and on
substituting the resulting expressions into the rst equation of the system, we arrives at
a separable rst-order equation on x. Similarly doing that for other two equations, we
will arrive at rst order equation on y and z too.
References
-http://eqworld.ipmnet.ru/en/solutions/sysode/sode0401.pdf
system of odes nonlinear 3eq order1 type2
sympy.solvers.ode. nonlinear 3eq order1 type2(x, y, z, t, eq)
Equations:
ax0 = (b c)yzf (x, y, z, t)
by 0 = (c a)zxf (x, y, z, t)
cz 0 = (a b)xyf (x, y, z, t)
1454
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
First Integrals:
ax2 + by 2 + cz 2 = C1
a2 x2 + b2 y 2 + c2 z 2 = C2
where C1 and C2 are arbitrary constants. On solving the integrals for y and z and on
substituting the resulting expressions into the rst equation of the system, we arrives at
a rst-order dierential equations on x. Similarly doing that for other two equations we
will arrive at rst order equation on y and z.
References
-http://eqworld.ipmnet.ru/en/solutions/sysode/sode0402.pdf
system of odes nonlinear 3eq order1 type3
sympy.solvers.ode. nonlinear 3eq order1 type3(x, y, z, t, eq)
Equations:
x0 = cF2 bF3 , y 0 = aF3 cF1 , z 0 = bF1 aF2
where Fn = Fn (x, y, z, t).
1.First Integral:
ax + by + cz = C1 ,
where C is an arbitrary constant.
2. If we assume function Fn to be independent of t,i.e, Fn = Fn (x, y, z) Then, on eliminating
t and z from the rst two equation of the system, one arrives at the rst-order equation
dy
aF3 (x, y, z) cF1 (x, y, z)
=
dx
cF2 (x, y, z) bF3 (x, y, z)
where z = 1c (C1 ax by)
References
-http://eqworld.ipmnet.ru/en/solutions/sysode/sode0404.pdf
5.27. ODE
1455
SymPy Documentation, Release 0.7.6
system of odes nonlinear 3eq order1 type4
sympy.solvers.ode. nonlinear 3eq order1 type4(x, y, z, t, eq)
Equations:
x0 = czF2 byF3 , y 0 = axF3 czF1 , z 0 = byF1 axF2
where Fn = Fn (x, y, z, t)
1.First integral:
ax2 + by 2 + cz 2 = C1
where C is an arbitrary constant.
2. Assuming the function Fn is independent of t: Fn = Fn (x, y, z). Then on eliminating t
and z from the rst two equations of the system, one arrives at the rst-order equation
axF3 (x, y, z) czF1 (x, y, z)
dy
=
dx
czF2 (x, y, z) byF3 (x, y, z)
where z =
1
c (C1
ax2 by 2 )
References
-http://eqworld.ipmnet.ru/en/solutions/sysode/sode0405.pdf
system of odes nonlinear 3eq order1 type5
sympy.solvers.ode. nonlinear 3eq order1 type5(x, y, t, eq)
x0 = x(cF2 bF3 ), y 0 = y(aF3 cF1 ), z 0 = z(bF1 aF2 )
where Fn = Fn (x, y, z, t) and are arbitrary functions.
First Integral:
a
|x| |y| |z| = C1
where C is an arbitrary constant. If the function Fn is independent of t, then, by eliminating t and z from the rst two equations of the system, one arrives at a rst-order
equation.
References
-http://eqworld.ipmnet.ru/en/solutions/sysode/sode0406.pdf
1456
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
5.27.5 Information on the ode module
This module contains dsolve() (page 1397) and dierent helper functions that it uses.
dsolve() (page 1397) solves ordinary dierential equations. See the docstring on the various
functions for their uses. Note that partial dierential equations support is in pde.py. Note
that hint functions have docstrings describing their various methods, but they are intended
for internal use. Use dsolve(ode, func, hint=hint) to solve an ODE using a specic hint.
See also the docstring on dsolve() (page 1397).
Functions in this module
These are the user functions in this module:
dsolve() (page 1397) - Solves ODEs.
classify ode() (page 1400) - Classies ODEs into possible hints for dsolve()
(page 1397).
checkodesol() (page 1402) - Checks if an equation is the solution to an ODE.
homogeneous order() (page 1403) - Returns the homogeneous order of an expression.
infinitesimals() (page 1404) - Returns the innitesimals of the Lie group of
point transformations of an ODE, such that it is invariant.
ode checkinfsol() - Checks if the given innitesimals are the actual innitesimals of a rst order ODE.
These are the non-solver helper functions that are for internal use. The user should
use the various options to dsolve() (page 1397) to obtain the functionality provided
by these functions:
odesimp() (page 1406) - Does all forms of ODE simplication.
ode sol simplicity() (page 1408) - A key function for comparing solutions by
simplicity.
constantsimp() (page 1407) - Simplies arbitrary constants.
constant renumber() (page 1407) - Renumber arbitrary constants.
handle Integral() - Evaluate unevaluated Integrals.
See also the docstrings of these functions.
Currently implemented solver methods
The following methods are implemented for solving ordinary dierential equations. See the
docstrings of the various hint functions for more information on each (run help(ode)):
1st order separable dierential equations.
1st order dierential equations whose coecients or dx and dy are functions homogeneous of the same order.
1st order exact dierential equations.
1st order linear dierential equations.
1st order Bernoulli dierential equations.
Power series solutions for rst order dierential equations.
Lie Group method of solving rst order dierential equations.
2nd order Liouville dierential equations.
5.27. ODE
1457
SymPy Documentation, Release 0.7.6
Power series solutions for second order dierential equations at ordinary and regular
singular points.
nth order linear homogeneous dierential equation with constant coecients.
nth order linear inhomogeneous dierential equation with constant coecients using
the method of undetermined coecients.
nth order linear inhomogeneous dierential equation with constant coecients using
the method of variation of parameters.
Philosophy behind this module
This module is designed to make it easy to add new ODE solving methods without having to
mess with the solving code for other methods. The idea is that there is a classify ode()
(page 1400) function, which takes in an ODE and tells you what hints, if any, will solve the
ODE. It does this without attempting to solve the ODE, so it is fast. Each solving method is a
hint, and it has its own function, named ode <hint>. That function takes in the ODE and any
match expression gathered by classify ode() (page 1400) and returns a solved result. If this
result has any integrals in it, the hint function will return an unevaluated Integral (page 615)
class. dsolve() (page 1397), which is the user wrapper function around all of this, will then
call odesimp() (page 1406) on the result, which, among other things, will attempt to solve the
equation for the dependent variable (the function we are solving for), simplify the arbitrary
constants in the expression, and evaluate any integrals, if the hint allows it.
How to add new solution methods
If you have an ODE that you want dsolve() (page 1397) to be able to solve, try to avoid adding
special case code here. Instead, try nding a general method that will solve your ODE, as well
as others. This way, the ode (page 1457) module will become more robust, and unhindered
by special case hacks. WolphramAlpha and Maples DETools[odeadvisor] function are two
resources you can use to classify a specic ODE. It is also better for a method to work with
an nth order ODE instead of only with specic orders, if possible.
To add a new method, there are a few things that you need to do. First, you need a hint
name for your method. Try to name your hint so that it is unambiguous with all other methods, including ones that may not be implemented yet. If your method uses integrals, also
include a hint Integral hint. If there is more than one way to solve ODEs with your method,
include a hint for each one, as well as a <hint> best hint. Your ode <hint> best() function should choose the best using min with ode sol simplicity as the key argument. See
ode 1st homogeneous coeff best() (page 1410), for example. The function that uses your
method will be called ode <hint>(), so the hint must only use characters that are allowed
in a Python function name (alphanumeric characters and the underscore character). Include a function for every hint, except for Integral hints (dsolve() (page 1397) takes care
of those automatically). Hint names should be all lowercase, unless a word is commonly capitalized (such as Integral or Bernoulli). If you have a hint that you do not want to run with
all Integral that doesnt have an Integral counterpart (such as a best hint that would
defeat the purpose of all Integral), you will need to remove it manually in the dsolve()
(page 1397) code. See also the classify ode() (page 1400) docstring for guidelines on writing a hint name.
Determine in general how the solutions returned by your method compare with other methods
that can potentially solve the same ODEs. Then, put your hints in the allhints (page 1405)
tuple in the order that they should be called. The ordering of this tuple determines which
hints are default. Note that exceptions are ok, because it is easy for the user to choose individual hints with dsolve() (page 1397). In general, Integral variants should go at the end
of the list, and best variants should go before the various hints they apply to. For example,
the undetermined coefficients hint comes before the variation of parameters hint because, even though variation of parameters is more general than undetermined coecients,
1458
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
undetermined coecients generally returns cleaner results for the ODEs that it can solve
than variation of parameters does, and it does not require integration, so it is much faster.
Next, you need to have a match expression or a function that matches the type of the ODE,
which you should put in classify ode() (page 1400) (if the match function is more than just a
few lines, like undetermined coefficients match(), it should go outside of classify ode()
(page 1400)). It should match the ODE without solving for it as much as possible, so that
classify ode() (page 1400) remains fast and is not hindered by bugs in solving code. Be
sure to consider corner cases. For example, if your solution method involves dividing by
something, make sure you exclude the case where that division will be 0.
In most cases, the matching of the ODE will also give you the various parts that you need to
solve it. You should put that in a dictionary (.match() will do this for you), and add that as
matching hints[hint] = matchdict in the relevant part of classify ode() (page 1400).
classify ode() (page 1400) will then send this to dsolve() (page 1397), which will send it to
your function as the match argument. Your function should be named ode <hint>(eq, func,
order, match). If you need to send more information, put it in the match
dictionary. For example, if you had to substitute in a dummy variable in classify ode()
(page 1400) to match the ODE, you will need to pass it to your function using the match dict to
access it. You can access the independent variable using func.args[0], and the dependent
variable (the function you are trying to solve for) as func.func. If, while trying to solve the
ODE, you nd that you cannot, raise NotImplementedError. dsolve() (page 1397) will catch
this error with the all meta-hint, rather than causing the whole routine to fail.
Add a docstring to your function that describes the method employed. Like with anything
else in SymPy, you will need to add a doctest to the docstring, in addition to real tests in
test ode.py. Try to maintain consistency with the other hint functions docstrings. Add
your method to the list at the top of this docstring. Also, add your method to ode.rst in
the docs/src directory, so that the Sphinx docs will pull its docstring into the main SymPy
documentation. Be sure to make the Sphinx documentation by running make html from within
the doc directory to verify that the docstring formats correctly.
If your solution method involves integrating, use C.Integral() instead of integrate()
(page 114). This allows the user to bypass hard/slow integration by using the Integral
variant of your hint. In most cases, calling sympy.core.basic.Basic.doit() (page 93)
will integrate your solution. If this is not the case, you will need to write special code in
handle Integral(). Arbitrary constants should be symbols named C1, C2, and so on. All
solution methods should return an equality instance. If you need an arbitrary number of arbitrary constants, you can use constants = numbered symbols(prefix=C, cls=Symbol,
start=1). If it is possible to solve for the dependent function in a general way, do so. Otherwise, do as best as you can, but do not call solve in your ode <hint>() function. odesimp()
(page 1406) will attempt to solve the solution for you, so you do not need to do that. Lastly,
if your ODE has a common simplication that can be applied to your solutions, you can add
a special case in odesimp() (page 1406) for it. For example, solutions returned from the
1st homogeneous coeff hints often have many log() terms, so odesimp() (page 1406) calls
logcombine() (page 1344) on them (it also helps to write the arbitrary constant as log(C1)
instead of C1 in this case). Also consider common ways that you can rearrange your solution
to have constantsimp() (page 1407) take better advantage of it. It is better to put simplication in odesimp() (page 1406) than in your method, because it can then be turned o with
the simplify ag in dsolve() (page 1397). If you have any extraneous simplication in your
function, be sure to only run it using if match.get(simplify, True):, especially if it can
be slow or if it can reduce the domain of the solution.
Finally, as with every contribution to SymPy, your method will need to be tested. Add a
test for each method in test ode.py. Follow the conventions there, i.e., test the solver using dsolve(eq, f(x), hint=your hint), and also test the solution using checkodesol()
(page 1402) (you can put these in a separate tests and skip/XFAIL if it runs too slow/doesnt
5.27. ODE
1459
SymPy Documentation, Release 0.7.6
work). Be sure to call your hint specically in dsolve() (page 1397), that way the test wont be
broken simply by the introduction of another matching hint. If your method works for higher
order (>1) ODEs, you will need to run sol = constant renumber(sol, C, 1, order)
for each solution, where order is the order of the ODE. This is because constant renumber
renumbers the arbitrary constants by printing order, which is platform dependent. Try to test
every corner case of your solver, including a range of orders if it is a nth order solver, but if
your solver is slow, such as if it involves hard integration, try to keep the test run time down.
Feel free to refactor existing hints to avoid duplicating code or creating inconsistencies. If you
can show that your method exactly duplicates an existing method, including in the simplicity
and speed of obtaining the solutions, then you can remove the old, less general method. The
existing code is tested extensively in test ode.py, so if anything is broken, one of those tests
will surely fail.
5.28 PDE
5.28.1 User Functions
These are functions that are imported into the global namespace with from sympy import *.
They are intended for user use.
pde separate()
sympy.solvers.pde.pde separate(eq, fun, sep, strategy=mul)
Separate variables in partial dierential equation either by additive or multiplicative
separation approach. It tries to rewrite an equation so that one of the specied variables
occurs on a dierent side of the equation than the others.
Parameters
eq Partial dierential equation
fun Original function F(x, y, z)
sep List of separated functions [X(x), u(y, z)]
strategy Separation strategy. You can choose between additive separation (add) and multiplicative separation (mul) which is default.
See Also:
pde separate add, pde separate mul
Examples
>>> from sympy import E, Eq, Function, pde_separate, Derivative as D
>>> from sympy.abc import x, t
>>> u, X, T = map(Function, uXT)
>>> eq = Eq(D(u(x, t), x), E**(u(x, t))*D(u(x, t), t))
>>> pde_separate(eq, u(x, t), [X(x), T(t)], strategy=add)
[exp(-X(x))*Derivative(X(x), x), exp(T(t))*Derivative(T(t), t)]
1460
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> eq = Eq(D(u(x, t), x, 2), D(u(x, t), t, 2))
>>> pde_separate(eq, u(x, t), [X(x), T(t)], strategy=mul)
[Derivative(X(x), x, x)/X(x), Derivative(T(t), t, t)/T(t)]
pde separate add()
sympy.solvers.pde.pde separate add(eq, fun, sep)
Helper function for searching additive separable solutions.
Consider an equation of two independent variables x, y and a dependent variable w, we
look for the product of two functions depending on dierent arguments:
w(x, y, z) = X(x) + y(y, z)
Examples
>>> from sympy import E, Eq, Function, pde_separate_add, Derivative as D
>>> from sympy.abc import x, t
>>> u, X, T = map(Function, uXT)
>>> eq = Eq(D(u(x, t), x), E**(u(x, t))*D(u(x, t), t))
>>> pde_separate_add(eq, u(x, t), [X(x), T(t)])
[exp(-X(x))*Derivative(X(x), x), exp(T(t))*Derivative(T(t), t)]
pde separate mul()
sympy.solvers.pde.pde separate mul(eq, fun, sep)
Helper function for searching multiplicative separable solutions.
Consider an equation of two independent variables x, y and a dependent variable w, we
look for the product of two functions depending on dierent arguments:
w(x, y, z) = X(x) u(y, z)
Examples
>>> from sympy import Function, Eq, pde_separate_mul, Derivative as D
>>> from sympy.abc import x, y
>>> u, X, Y = map(Function, uXY)
>>> eq = Eq(D(u(x, y), x, 2), D(u(x, y), y, 2))
>>> pde_separate_mul(eq, u(x, y), [X(x), Y(y)])
[Derivative(X(x), x, x)/X(x), Derivative(Y(y), y, y)/Y(y)]
pdsolve()
sympy.solvers.pde.pdsolve(eq, func=None, hint=default,
fun=None, **kwargs)
Solves any (supported) kind of partial dierential equation.
dict=False,
solve-
Usage
5.28. PDE
1461
SymPy Documentation, Release 0.7.6
pdsolve(eq, f(x,y), hint) -> Solve partial dierential equation eq for function
f(x,y), using method hint.
Details
eq can be any supported partial dierential equation (see the pde docstring for supported methods). This can either be an Equality, or an expression, which is assumed to be equal to 0.
f(x,y) is a function of two variables whose derivatives in that variable
make up the partial dierential equation. In many cases it is not necessary
to provide this; it will be autodetected (and an error raised if it couldnt be
detected).
hint is the solving method that you want pdsolve to use. Use
classify pde(eq, f(x,y)) to get all of the possible hints for a PDE. The default
hint, default, will use whatever hint is returned rst by classify pde(). See
Hints below for more options that you can use for hint.
solvefun is the convention used for arbitrary functions returned by the
PDE solver. If not set by the user, it is set by default to be F.
Hints
Aside from the various solving methods, there are also some meta-hints that you
can pass to pdsolve():
default: This uses whatever hint is returned rst by classify pde(). This is
the default argument to pdsolve().
all: To make pdsolve apply all relevant classication hints, use pdsolve(PDE,
func, hint=all). This will return a dictionary of hint:solution terms. If a
hint causes pdsolve to raise the NotImplementedError, value of that hints
key will be the exception object raised. The dictionary will also include some
special keys:
order: The order of the PDE. See also ode order() in deutils.py
default: The solution that would be returned by default. This is the one
produced by the hint that appears rst in the tuple returned by classify pde().
all Integral: This is the same as all, except if a hint also has a corresponding Integral hint, it only returns the Integral hint. This is useful if all
causes pdsolve() to hang because of a dicult or impossible integral. This
meta-hint will also be much faster than all, because integrate() is an expensive routine.
See also the classify pde() docstring for more info on hints, and the pde docstring for a list of all supported hints.
Tips
You can declare the derivative of an unknown function this way:
>>>
>>>
>>>
>>>
>>>
>>>
>>>
1462
from sympy import Function, Derivative
from sympy.abc import x, y # x and y are the independent variables
f = Function(f)(x, y) # f is a function of x and y
# fx will be the partial derivative of f with respect to x
fx = Derivative(f, x)
# fy will be the partial derivative of f with respect to y
fy = Derivative(f, y)
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
See test pde.py for many tests, which serves also as a set of examples for how
to use pdsolve().
pdsolve always returns an Equality class (except for the case when the hint is
all or all Integral). Note that it is not possible to get an explicit solution for
f(x, y) as in the case of ODEs
Do help(pde.pde hintname) to get help more information on a specic hint
Examples
>>> from sympy.solvers.pde import pdsolve
>>> from sympy import Function, diff, Eq
>>> from sympy.abc import x, y
>>> f = Function(f)
>>> u = f(x, y)
>>> ux = u.diff(x)
>>> uy = u.diff(y)
>>> eq = Eq(1 + (2*(ux/u)) + (3*(uy/u)))
>>> pdsolve(eq)
f(x, y) == F(3*x - 2*y)*exp(-2*x/13 - 3*y/13)
classify pde()
sympy.solvers.pde.classify pde(eq, func=None, dict=False, **kwargs)
Returns a tuple of possible pdsolve() classications for a PDE.
The tuple is ordered so that rst item is the classication that pdsolve() uses to solve the
PDE by default. In general, classications near the beginning of the list will produce better solutions faster than those near the end, though there are always exceptions. To make
pdsolve use a dierent classication, use pdsolve(PDE, func, hint=<classication>). See
also the pdsolve() docstring for dierent meta-hints you can use.
If dict is true, classify pde() will return a dictionary of hint:match expression terms.
This is intended for internal use by pdsolve(). Note that because dictionaries are ordered
arbitrarily, this will most likely not be in the same order as the tuple.
You can get help on dierent hints by doing help(pde.pde hintname), where hintname is
the name of the hint without Integral.
See sympy.pde.allhints or the sympy.pde docstring for a list of all supported hints that
can be returned from classify pde.
Examples
>>> from sympy.solvers.pde import classify_pde
>>> from sympy import Function, diff, Eq
>>> from sympy.abc import x, y
>>> f = Function(f)
>>> u = f(x, y)
>>> ux = u.diff(x)
>>> uy = u.diff(y)
>>> eq = Eq(1 + (2*(ux/u)) + (3*(uy/u)))
>>> classify_pde(eq)
(1st_linear_constant_coeff_homogeneous,)
5.28. PDE
1463
SymPy Documentation, Release 0.7.6
checkpdesol()
sympy.solvers.pde.checkpdesol(pde, sol, func=None, solve for func=True)
Checks if the given solution satises the partial dierential equation.
pde is the partial dierential equation which can be given in the form of an equation or
an expression. sol is the solution for which the pde is to be checked. This can also be
given in an equation or an expression form. If the function is not provided, the helper
function preprocess from deutils is used to identify the function.
If a sequence of solutions is passed, the same sort of container will be used to return the
result for each solution.
The following methods are currently being implemented to check if the solution satises
the PDE:
1.Directly substitute the solution in the PDE and check. If the solution hasnt been
solved for f, then it will solve for f provided solve for func hasnt been set to False.
If the solution satises the PDE, then a tuple (True, 0) is returned. Otherwise a tuple
(False, expr) where expr is the value obtained after substituting the solution in the PDE.
However if a known solution returns False, it may be due to the inability of doit() to
simplify it to zero.
Examples
>>> from sympy import Function, symbols, diff
>>> from sympy.solvers.pde import checkpdesol, pdsolve
>>> x, y = symbols(x y)
>>> f = Function(f)
>>> eq = 2*f(x,y) + 3*f(x,y).diff(x) + 4*f(x,y).diff(y)
>>> sol = pdsolve(eq)
>>> assert checkpdesol(eq, sol)[0]
>>> eq = x*f(x,y) + f(x,y).diff(x)
>>> checkpdesol(eq, sol)
(False, (x*F(4*x - 3*y) - 6*F(4*x - 3*y)/25 + 4*Subs(Derivative(F(_xi_1), _xi_1), (_xi_1,), (4*x
5.28.2 Hint Methods
These funcions are meant for internal use. However they contain useful information on the
various solving methods.
pde 1st linear constant coeff homogeneous
sympy.solvers.pde.pde 1st linear constant coeff homogeneous(eq, func, order,
match, solvefun)
Solves a rst order linear homogeneous partial dierential equation with constant coefcients.
The general form of this partial dierential equation is
a
1464
df (x, y)
df (x, y)
+b
+ cf (x, y) = 0
dx
dy
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
where a, b and c are constants.
The general solution is of the form:
>>> from sympy.solvers import pdsolve
>>> from sympy.abc import x, y, a, b, c
>>> from sympy import Function, pprint
>>> f = Function(f)
>>> u = f(x,y)
>>> ux = u.diff(x)
>>> uy = u.diff(y)
>>> genform = a*ux + b*uy + c*u
>>> pprint(genform)
d
d
a*--(f(x, y)) + b*--(f(x, y)) + c*f(x, y)
dx
dy
>>> pprint(pdsolve(genform))
-c*(a*x + b*y)
--------------2
2
a + b
f(x, y) = F(-a*y + b*x)*e
References
Viktor Grigoryan, Partial Dierential Equations Math 124A - Fall 2010, pp.7
Examples
>>> from sympy.solvers.pde import (
... pde_1st_linear_constant_coeff_homogeneous)
>>> from sympy import pdsolve
>>> from sympy import Function, diff, pprint
>>> from sympy.abc import x,y
>>> f = Function(f)
>>> pdsolve(f(x,y) + f(x,y).diff(x) + f(x,y).diff(y))
f(x, y) == F(x - y)*exp(-x/2 - y/2)
>>> pprint(pdsolve(f(x,y) + f(x,y).diff(x) + f(x,y).diff(y)))
x
y
- - - 2
2
f(x, y) = F(x - y)*e
pde 1st linear constant coeff
sympy.solvers.pde.pde 1st linear constant coeff(eq, func, order, match, solvefun)
Solves a rst order linear partial dierential equation with constant coecients.
The general form of this partial dierential equation is
a
5.28. PDE
df (x, y)
df (x, y)
+b
+ cf (x, y) = G(x, y)
dx
dy
1465
SymPy Documentation, Release 0.7.6
where a, b and c are constants and G(x, y) can be an arbitrary function in x and y.
The general solution of the PDE is:
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
from sympy.solvers import pdsolve
from sympy.abc import x, y, a, b, c
from sympy import Function, pprint
f = Function(f)
G = Function(G)
u = f(x,y)
ux = u.diff(x)
uy = u.diff(y)
genform = a*u + b*ux + c*uy - G(x,y)
pprint(genform)
d
d
a*f(x, y) + b*--(f(x, y)) + c*--(f(x, y)) - G(x, y)
dx
dy
>>> pprint(pdsolve(genform, hint=1st_linear_constant_coeff_Integral))
//
b*x + c*y
\
||
/
|
||
|
|
||
|
a*xi
|
||
|
------|
||
|
2
2
|
||
|
/b*xi + c*eta -b*eta + c*xi\ b + c
|
||
|
G|------------, -------------|*e
d(xi)|
||
|
|
2
2
2
2
|
|
||
|
\ b + c
b + c
/
|
||
|
|
||
/
|
||
|
f(x, y) = ||F(eta) + -------------------------------------------------------|*
||
2
2
|
\\
b + c
/
\|
||
||
||
||
||
||
||
||
-a*xi ||
-------||
2
2||
b + c ||
e
||
||
/|eta=-b*y + c*x, xi=b*x + c*y
References
Viktor Grigoryan, Partial Dierential Equations Math 124A - Fall 2010, pp.7
1466
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.solvers.pde import pdsolve
>>> from sympy import Function, diff, pprint, exp
>>> from sympy.abc import x,y
>>> f = Function(f)
>>> eq = -2*f(x,y).diff(x) + 4*f(x,y).diff(y) + 5*f(x,y) - exp(x + 3*y)
>>> pdsolve(eq)
f(x, y) == (F(4*x + 2*y) + exp(x/2 + 4*y)/15)*exp(x/2 - y)
pde 1st linear variable coeff
sympy.solvers.pde.pde 1st linear variable coeff(eq, func, order, match, solvefun)
Solves a rst order linear partial dierential equation with variable coecients. The
general form of this partial dierential equation is
a(x, y)
df (x, y)
df (x, y)
+ a(x, y)
+ c(x, y)f (x, y) G(x, y)
dx
dy
where a(x, y), b(x, y), c(x, y) and G(x, y) are arbitrary functions in x and y. This PDE is
converted into an ODE by making the following transformation.
1] as x
2] as the constant in the solution to the dierential equation
dy
dx
= ab
Making the following substitutions reduces it to the linear ODE
a(, )
du
+ c(, )u d(, ) = 0
d
which can be solved using dsolve.
The general form of this PDE is:
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
from sympy.solvers.pde import pdsolve
from sympy.abc import x, y
from sympy import Function, pprint
a, b, c, G, f= [Function(i) for i in [a, b, c, G, f]]
u = f(x,y)
ux = u.diff(x)
uy = u.diff(y)
genform = a(x, y)*u + b(x, y)*ux + c(x, y)*uy - G(x,y)
pprint(genform)
d
d
-G(x, y) + a(x, y)*f(x, y) + b(x, y)*--(f(x, y)) + c(x, y)*--(f(x, y))
dx
dy
References
Viktor Grigoryan, Partial Dierential Equations Math 124A - Fall 2010, pp.7
5.28. PDE
1467
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.solvers.pde import pdsolve
>>> from sympy import Function, diff, pprint, exp
>>> from sympy.abc import x,y
>>> f = Function(f)
>>> eq = x*(u.diff(x)) - y*(u.diff(y)) + y**2*u - y**2
>>> pdsolve(eq)
f(x, y) == F(x*y)*exp(y**2/2) + 1
5.28.3 Information on the pde module
This module contains pdsolve() and dierent helper functions that it uses. It is heavily inspired
by the ode module and hence the basic infrastructure remains the same.
Functions in this module
These are the user functions in this module:
pdsolve() - Solves PDEs
classify pde() - Classies PDEs into possible hints for dsolve().
pde separate() - Separate variables in partial dierential equation either by
additive or multiplicative separation approach.
These are the helper functions in this module:
pde separate add() - Helper function for searching additive separable solutions.
pde separate mul() - Helper function for searching multiplicative
separable solutions.
Currently implemented solver methods
The following methods are implemented for solving partial dierential equations. See the
docstrings of the various pde hint() functions for more information on each (run help(pde)):
1st order linear homogeneous partial dierential equations with constant coecients.
1st order linear general partial dierential equations with constant coecients.
1st order linear partial dierential equations with variable coecients.
5.29 Solvers
The solvers module in SymPy implements methods for solving equations.
5.29.1 Algebraic equations
Use solve() to solve algebraic equations. We suppose all equations are equaled to 0, so
solving x**2 == 1 translates into the following code:
>>> from sympy.solvers import solve
>>> from sympy import Symbol
>>> x = Symbol(x)
>>> solve(x**2 - 1, x)
[-1, 1]
1468
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
The rst argument for solve() is an equation (equaled to zero) and the second argument is
the symbol that we want to solve the equation for.
sympy.solvers.solvers.solve(f, *symbols, **ags)
Algebraically solves equations and systems of equations.
Currently supported are:
univariate polynomial,
transcendental
piecewise combinations of the above
systems of linear and polynomial equations
sytems containing relational expressions.
Input is formed as:
f
a single Expr or Poly that must be zero,
an Equality
a Relational expression or boolean
iterable of one or more of the above
symbols (object(s) to solve for) specied as
none given (other non-numeric objects will be used)
single symbol
denested list of symbols e.g. solve(f, x, y)
ordered iterable of symbols e.g. solve(f, [x, y])
ags
dict=True (default is False) return list (perhaps empty) of solution mappings
set=True (default is False) return list of symbols and set of tuple(s) of solution(s)
exclude=[] (default) dont try to solve for any of the free symbols in exclude;
if expressions are given, the free symbols in them will be extracted automatically.
check=True (default) If False, dont do any testing of solutions. This can be
useful if one wants to include solutions that make any denominator zero.
numerical=True (default) do a fast numerical check if f has only one symbol.
minimal=True (default is False) a very fast, minimal testing.
warn=True (default is False) show a warning if checksol() could not conclude.
simplify=True (default) simplify all but cubic and quartic solutions before
returning them and (if check is not False) use the general simplify function
on the solutions and the expression obtained when they are substituted into
the function which should be zero
force=True (default is False) make positive all symbols without assumptions regarding sign.
5.29. Solvers
1469
SymPy Documentation, Release 0.7.6
rational=True (default) recast Floats as Rational; if this option is not used,
the system containing oats may fail to solve because of issues with polys.
If rational=None, Floats will be recast as rationals but the answer will be
recast as Floats. If the ag is False then nothing will be done to the Floats.
manual=True (default is False) do not use the polys/matrix method to solve
a system of equations, solve them one at a time as you might manually.
implicit=True (default is False) allows solve to return a solution for a pattern in terms of other functions that contain that pattern; this is only needed
if the pattern is inside of some invertible function like cos, exp, ....
particular=True (default is False) instructs solve to try to nd a particular
solution to a linear system with as many zeros as possible; this is very expensive
quick=True (default is False) when using particular=True, use a fast
heuristic instead to nd a solution with many zeros (instead of using the very
slow method guaranteed to nd the largest number of zeros possible)
Notes
assumptions arent checked when solve() input involves relationals or bools.
When the solutions are checked, those that make any denominator zero are automatically
excluded. If you do not want to exclude such solutions then use the check=False option:
>>> from sympy import sin, limit
>>> solve(sin(x)/x) # 0 is excluded
[pi]
If check=False then a solution to the numerator being zero is found: x = 0. In this case,
this is a spurious solution since sin(x)/x has the well known limit (without dicontinuity)
of 1 at x = 0:
>>> solve(sin(x)/x, check=False)
[0, pi]
In the following case, however, the limit exists and is equal to the the value of x = 0 that
is excluded when check=True:
>>>
>>>
[]
>>>
[0]
>>>
0
>>>
0
eq = x**2*(1/x - z**2/x)
solve(eq, x)
solve(eq, x, check=False)
limit(eq, x, 0, -)
limit(eq, x, 0, +)
Examples
The output varies according to the input and can be seen by example:
1470
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> from sympy import solve, Poly, Eq, Function, exp
>>> from sympy.abc import x, y, z, a, b
>>> f = Function(f)
boolean or univariate Relational
>>> solve(x < 3)
And(-oo < re(x), im(x) == 0, re(x) < 3)
to always get a list of solution mappings, use ag dict=True
>>> solve(x - 3, dict=True)
[{x: 3}]
>>> solve([x - 3, y - 1], dict=True)
[{x: 3, y: 1}]
to get a list of symbols and set of solution(s) use ag set=True
>>> solve([x**2 - 3, y - 1], set=True)
([x, y], set([(-sqrt(3), 1), (sqrt(3), 1)]))
single expression and single symbol that is in the expression
>>> solve(x - y, x)
[y]
>>> solve(x - 3, x)
[3]
>>> solve(Eq(x, 3), x)
[3]
>>> solve(Poly(x - 3), x)
[3]
>>> solve(x**2 - y**2, x, set=True)
([x], set([(-y,), (y,)]))
>>> solve(x**4 - 1, x, set=True)
([x], set([(-1,), (1,), (-I,), (I,)]))
single expression with no symbol that is in the expression
>>> solve(3, x)
[]
>>> solve(x - 3, y)
[]
single expression with no symbol given
In this case, all free symbols will be selected as potential symbols to solve
for. If the equation is univariate then a list of solutions is returned; otherwise
as is the case when symbols are given as an iterable of length > 1 a list
of mappings will be returned.
>>> solve(x - 3)
[3]
>>> solve(x**2 - y**2)
[{x: -y}, {x: y}]
>>> solve(z**2*x**2 - z**2*y**2)
[{x: -y}, {x: y}, {z: 0}]
>>> solve(z**2*x - z**2*y**2)
[{x: y**2}, {z: 0}]
5.29. Solvers
1471
SymPy Documentation, Release 0.7.6
when an object other than a Symbol is given as a symbol, it is isolated algebraically
and an implicit solution may be obtained. This is mostly provided as a convenience
to save one from replacing the object with a Symbol and solving for that Symbol. It
will only work if the specied object can be replaced with a Symbol using the subs
method.
>>> solve(f(x) - x, f(x))
[x]
>>> solve(f(x).diff(x) - f(x) - x, f(x).diff(x))
[x + f(x)]
>>> solve(f(x).diff(x) - f(x) - x, f(x))
[-x + Derivative(f(x), x)]
>>> solve(x + exp(x)**2, exp(x), set=True)
([exp(x)], set([(-sqrt(-x),), (sqrt(-x),)]))
>>> from sympy import Indexed, IndexedBase, Tuple, sqrt
>>> A = IndexedBase(A)
>>> eqs = Tuple(A[1] + A[2] - 3, A[1] - A[2] + 1)
>>> solve(eqs, eqs.atoms(Indexed))
{A[1]: 1, A[2]: 2}
To solve for a symbol implicitly, use implicit=True:
>>> solve(x + exp(x), x)
[-LambertW(1)]
>>> solve(x + exp(x), x, implicit=True)
[-exp(x)]
It is possible to solve for anything that can be targeted with subs:
>>> solve(x + 2 + sqrt(3), x + 2)
[-sqrt(3)]
>>> solve((x + 2 + sqrt(3), x + 4 + y), y, x + 2)
{y: -2 + sqrt(3), x + 2: -sqrt(3)}
Nothing heroic is done in this implicit solving so you may end up with a
symbol still in the solution:
>>>
>>>
{y:
>>>
{x:
eqs = (x*y + 3*y + sqrt(3), x + 4 + y)
solve(eqs, y, x + 2)
-sqrt(3)/(x + 3), x + 2: (-2*x - 6 + sqrt(3))/(x + 3)}
solve(eqs, y*x, x)
-y - 4, x*y: -3*y - sqrt(3)}
if you attempt to solve for a number remember that the number you have
obtained does not necessarily mean that the value is equivalent to the
expression obtained:
>>> solve(sqrt(2) - 1, 1)
[sqrt(2)]
>>> solve(x - y + 1, 1) # /!\ -1 is targeted, too
[x/(y - 1)]
>>> [_.subs(z, -1) for _ in solve((x - y + 1).subs(-1, z), 1)]
[-x + y]
To solve for a function within a derivative, use dsolve.
1472
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
single expression and more than 1 symbol
when there is a linear solution
>>> solve(x - y**2, x, y)
[{x: y**2}]
>>> solve(x**2 - y, x, y)
[{y: x**2}]
when undetermined coecients are identied
*that are linear
>>> solve((a + b)*x - b + 2, a, b)
{a: -2, b: 2}
*that are nonlinear
>>> solve((a + b)*x - b**2 + 2, a, b, set=True)
([a, b], set([(-sqrt(2), sqrt(2)), (sqrt(2), -sqrt(2))]))
if there is no linear solution then the rst successful attempt for a nonlinear
solution will be returned
>>> solve(x**2 - y**2, x, y)
[{x: -y}, {x: y}]
>>> solve(x**2 - y**2/exp(x), x, y)
[{x: 2*LambertW(y/2)}]
>>> solve(x**2 - y**2/exp(x), y, x)
[{y: -x*sqrt(exp(x))}, {y: x*sqrt(exp(x))}]
iterable of one or more of the above
involving relationals or bools
>>> solve([x < 3, x - 2])
And(im(x) == 0, re(x) == 2)
>>> solve([x > 3, x - 2])
False
when the system is linear
*with a solution
>>>
{x:
>>>
{x:
>>>
{x:
>>>
{x:
solve([x - 3], x)
3}
solve((x + 5*y - 2,
-3, y: 1}
solve((x + 5*y - 2,
-3, y: 1}
solve((x + 5*y - 2,
-5*y + 2, z: 21*y -
-3*x + 6*y - 15), x, y)
-3*x + 6*y - 15), x, y, z)
-3*x + 6*y - z), z, x, y)
6}
*without a solution
>>> solve([x + 3, x - 3])
[]
when the system is not linear
5.29. Solvers
1473
SymPy Documentation, Release 0.7.6
>>> solve([x**2 + y -2, y**2 - 4], x, y, set=True)
([x, y], set([(-2, -2), (0, 2), (2, -2)]))
if no symbols are given, all free symbols will be selected and a list of mappings
returned
>>> solve([x - 2, x**2 + y])
[{x: 2, y: -4}]
>>> solve([x - 2, x**2 + f(x)], set([f(x), x]))
[{x: 2, f(x): -4}]
if any equation doesnt depend on the symbol(s) given it will be eliminated from
the equation set and an answer may be given implicitly in terms of variables that
were not of interest
>>> solve([x - y, y - 3], x)
{x: y}
sympy.solvers.solvers.solve linear(lhs, rhs=0, symbols=[], exclude=[])
Return a tuple derived from f = lhs - rhs that is either:
(numerator, denominator) of f If this comes back as (0, 1) it means that f is
independent of the symbols in symbols, e.g:
y*cos(x)**2 + y*sin(x)**2 - y = y*(0) = 0
cos(x)**2 + sin(x)**2 = 1
If it comes back as (0, 0) there is no solution to the equation amongst the
symbols given.
If the numerator is not zero then the function is guaranteed to be dependent
on a symbol in symbols.
or
(symbol, solution) where symbol appears linearly in the numerator of f, is in
symbols (if given) and is not in exclude (if given).
No simplication is done to f other than and mul=True expansion, so the solution will correspond strictly to a unique solution.
Examples
>>> from sympy.solvers.solvers import solve_linear
>>> from sympy.abc import x, y, z
These are linear in x and 1/x:
>>>
(x,
>>>
(x,
solve_linear(x + y**2)
-y**2)
solve_linear(1/x - y**2)
y**(-2))
When not linear in x or y then the numerator and denominator are returned.
>>> solve_linear(x**2/y**2 - 3)
(x**2 - 3*y**2, y**2)
1474
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
If the numerator is a symbol then (0, 0) is returned if the solution for that symbol would
have set any denominator to 0:
>>>
(0,
>>>
x
>>>
(x,
solve_linear(1/(1/x - 2))
0)
1/(1/x) # to SymPy, this looks like x ...
solve_linear(1/(1/x)) # so a solution is given
0)
If x is allowed to cancel, then this appears linear, but this sort of cancellation is not done
so the solution will always satisfy the original expression without causing a division by
zero error.
>>> solve_linear(x**2*(1/x - z**2/x))
(x**2*(-z**2 + 1), x)
You can give a list of what you prefer for x candidates:
>>> solve_linear(x + y + z, symbols=[y])
(y, -x - z)
You can also indicate what variables you dont want to consider:
>>> solve_linear(x + y + z, exclude=[x, z])
(y, -x - z)
If only x was excluded then a solution for y or z might be obtained.
sympy.solvers.solvers.solve linear system(system, *symbols, **ags)
Solve system of N linear equations with M variables, which means both under- and
overdetermined systems are supported. The possible number of solutions is zero, one
or innite. Respectively, this procedure will return None or a dictionary with solutions.
In the case of underdetermined systems, all arbitrary parameters are skipped. This may
cause a situation in which an empty dictionary is returned. In that case, all symbols can
be assigned arbitrary values.
Input to this functions is a Nx(M+1) matrix, which means it has to be in augmented
form. If you prefer to enter N equations and M unknowns then use solve(N eqs, M symbols)
instead. Note: a local copy of the matrix is made by this routine so the matrix that is
passed will not be modied.
The algorithm used here is fraction-free Gaussian elimination, which results, after elimination, in an upper-triangular matrix. Then solutions are found using back-substitution.
This approach is more ecient and compact than the Gauss-Jordan method.
>>> from sympy import Matrix, solve_linear_system
>>> from sympy.abc import x, y
Solve the following system:
x + 4 y == 2
-2 x +
y == 14
>>> system = Matrix(( (1, 4, 2), (-2, 1, 14)))
>>> solve_linear_system(system, x, y)
{x: -6, y: 2}
A degenerate system returns an empty dictionary.
5.29. Solvers
1475
SymPy Documentation, Release 0.7.6
>>> system = Matrix(( (0,0,0), (0,0,0) ))
>>> solve_linear_system(system, x, y)
{}
sympy.solvers.solvers.solve linear system LU(matrix, syms)
Solves the augmented matrix system using LUsolve and returns a dictionary in which
solutions are keyed to the symbols of syms as ordered.
The matrix must be invertible.
See Also:
sympy.matrices.LUsolve
Examples
>>> from sympy import Matrix
>>> from sympy.abc import x, y, z
>>> from sympy.solvers.solvers import solve_linear_system_LU
>>>
...
...
...
{x:
solve_linear_system_LU(Matrix([
[1, 2, 0, 1],
[3, 2, 2, 1],
[2, 0, 0, 1]]), [x, y, z])
1/2, y: 1/4, z: -1/2}
sympy.solvers.solvers.solve undetermined coeffs(equ, coes, sym, **ags)
Solve equation of a type p(x; a 1, ..., a k) == q(x) where both p, q are univariate polynomials and f depends on k parameters. The result of this functions is a dictionary with
symbolic values of those parameters with respect to coecients in q.
This functions accepts both Equations class instances and ordinary SymPy expressions.
Specication of parameters and variable is obligatory for eciency and simplicity reason.
>>> from sympy import Eq
>>> from sympy.abc import a, b, c, x
>>> from sympy.solvers import solve_undetermined_coeffs
>>> solve_undetermined_coeffs(Eq(2*a*x + a+b, x), [a, b], x)
{a: 1/2, b: -1/2}
>>> solve_undetermined_coeffs(Eq(a*c*x + a+b, x), [a, b], x)
{a: 1/c, b: -1/c}
sympy.solvers.solvers.nsolve(*args, **kwargs)
Solve a nonlinear equation system numerically:
nsolve(f, [args,] x0, modules=[mpmath], **kwargs)
f is a vector function of symbolic expressions representing the system. args are the
variables. If there is only one variable, this argument can be omitted. x0 is a starting
vector close to a solution.
Use the modules keyword to specify which modules should be used to evaluate the function and the Jacobian matrix. Make sure to use a module that supports matrices. For
more information on the syntax, please see the docstring of lambdify.
1476
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Overdetermined systems are supported.
>>> from sympy import Symbol, nsolve
>>> import sympy
>>> sympy.mpmath.mp.dps = 15
>>> x1 = Symbol(x1)
>>> x2 = Symbol(x2)
>>> f1 = 3 * x1**2 - 2 * x2**2 - 1
>>> f2 = x1**2 - 2 * x1 + x2**2 + 2 * x2 - 8
>>> print(nsolve((f1, f2), (x1, x2), (-1, 1)))
[-1.19287309935246]
[ 1.27844411169911]
For one-dimensional functions the syntax is simplied:
>>> from sympy import sin, nsolve
>>> from sympy.abc import x
>>> nsolve(sin(x), x, 2)
3.14159265358979
>>> nsolve(sin(x), 2)
3.14159265358979
mpmath.ndroot is used, you can nd there more extensive documentation, especially
concerning keyword parameters and available solvers.
sympy.solvers.solvers.check assumptions(expr, **assumptions)
Checks whether expression expr satises all assumptions.
assumptions is a dict of assumptions: {assumption: True|False, ...}.
Examples
>>> from sympy import Symbol, pi, I, exp
>>> from sympy.solvers.solvers import check_assumptions
>>> check_assumptions(-5, integer=True)
True
>>> check_assumptions(pi, real=True, integer=False)
True
>>> check_assumptions(pi, real=True, negative=True)
False
>>> check_assumptions(exp(I*pi/7), real=False)
True
>>> x = Symbol(x, real=True, positive=True)
>>> check_assumptions(2*x + 1, real=True, positive=True)
True
>>> check_assumptions(-2*x - 5, real=True, positive=True)
False
N one is returned if check assumptions() could not conclude.
>>> check_assumptions(2*x - 1, real=True, positive=True)
>>> z = Symbol(z)
>>> check_assumptions(z, real=True)
sympy.solvers.solvers.checksol(f, symbol, sol=None, **ags)
Checks whether sol is a solution of equation f == 0.
5.29. Solvers
1477
SymPy Documentation, Release 0.7.6
Input can be either a single symbol and corresponding value or a dictionary of symbols
and values. When given as a dictionary and ag simplify=True, the values in the dictionary will be simplied. f can be a single equation or an iterable of equations. A solution
must satisfy all equations in f to be considered valid; if a solution does not satisfy any
equation, False is returned; if one or more checks are inconclusive (and none are False)
then None is returned.
Examples
>>> from sympy import symbols
>>> from sympy.solvers import checksol
>>> x, y = symbols(x,y)
>>> checksol(x**4 - 1, x, 1)
True
>>> checksol(x**4 - 1, x, 0)
False
>>> checksol(x**2 + y**2 - 5**2, {x: 3, y: 4})
True
To check if an expression is zero using checksol, pass it as f and send an empty dictionary
for symbol:
>>> checksol(x**2 + x - x*(x + 1), {})
True
None is returned if checksol() could not conclude.
ags:
numerical=True (default) do a fast numerical check if f has only one symbol.
minimal=True (default is False) a very fast, minimal testing.
warn=True (default is False) show a warning if checksol() could not conclude.
simplify=True (default) simplify solution before substituting into function and
simplify the function before trying specic simplications
force=True (default is False) make positive all symbols without assumptions regarding sign.
5.29.2 Ordinary Dierential equations (ODEs)
See ODE (page 1397).
5.29.3 Partial Dierential Equations (PDEs)
See PDE (page 1460).
5.29.4 Deutils (Utilities for solving ODEs and PDEs)
sympy.solvers.deutils.ode order(expr, func)
Returns the order of a given dierential equation with respect to func.
This function is implemented recursively.
1478
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>>
>>>
>>>
>>>
>>>
...
2
>>>
2
>>>
3
from sympy import Function
from sympy.solvers.deutils import ode_order
from sympy.abc import x
f, g = map(Function, [f, g])
ode_order(f(x).diff(x, 2) + f(x).diff(x)**2 +
f(x).diff(x), f(x))
ode_order(f(x).diff(x, 2) + g(x).diff(x, 3), f(x))
ode_order(f(x).diff(x, 2) + g(x).diff(x, 3), g(x))
5.29.5 Recurrence Equations
sympy.solvers.recurr.rsolve(f, y, init=None)
Solve univariate recurrence with rational coecients.
Given k-th order linear recurrence L y = f , or equivalently:
ak (n)y(n + k) + ak1 (n)y(n + k 1) + + a0 (n)y(n) = f (n)
where ai (n), for i = 0, . . . , k, are polynomials or rational functions in n, and f is a hypergeometric function or a sum of a xed number of pairwise dissimilar hypergeometric terms
in n, nds all solutions or returns None, if none were found.
Initial conditions can be given as a dictionary in two forms:
1.{ n 0 : v 0, n 1 : v 1, ..., n m : v m }
2.{ y(n 0) : v 0, y(n 1) : v 1, ..., y(n m) : v m }
or as a list L of values:
L = [ v 0, v 1, ..., v m ]
where L[i] = v i, for i = 0, . . . , m, maps to y(ni ).
See Also:
rsolve poly (page 1480), rsolve ratio (page 1480), rsolve hyper (page 1481)
Examples
Lets consider the following recurrence:
(n 1)y(n + 2) (n2 + 3n 2)y(n + 1) + 2n(n + 1)y(n) = 0
>>> from sympy import Function, rsolve
>>> from sympy.abc import n
>>> y = Function(y)
5.29. Solvers
1479
SymPy Documentation, Release 0.7.6
>>> f = (n - 1)*y(n + 2) - (n**2 + 3*n - 2)*y(n + 1) + 2*n*(n + 1)*y(n)
>>> rsolve(f, y(n))
2**n*C0 + C1*factorial(n)
>>> rsolve(f, y(n), { y(0):0, y(1):3 })
3*2**n - 3*factorial(n)
sympy.solvers.recurr.rsolve poly(coes, f, n, **hints)
Given linear recurrence operator L of order k with polynomial coecients and inhomogeneous equation L y = f , where f is a polynomial, we seek for all polynomial solutions
over eld K of characteristic zero.
The algorithm performs two basic steps:
1.Compute degree N of the general polynomial solution.
2.Find all polynomials of degree N or less of L y = f .
There are two methods for computing the polynomial solutions. If the degree bound
is relatively small, i.e. its smaller than or equal to the order of the recurrence, then
naive method of undetermined coecients is being used. This gives system of algebraic
equations with N + 1 unknowns.
In the other case, the algorithm performs transformation of the initial equation to an
equivalent one, for which the system of algebraic equations has only r indeterminates.
This method is quite sophisticated (in comparison with the naive one) and was invented
together by Abramov, Bronstein and Petkovsek.
It is possible to generalize the algorithm implemented here to the case of linear qdierence and dierential equations.
Lets say that we would like to compute m-th Bernoulli polynomial up to a constant. For
this we can use b(n + 1) b(n) = mnm1 recurrence, which has solution b(n) = Bm + C. For
example:
>>> from sympy import Symbol, rsolve_poly
>>> n = Symbol(n, integer=True)
>>> rsolve_poly([-1, 1], 4*n**3, n)
C0 + n**4 - 2*n**3 + n**2
References
[R376] (page 1914), [R377] (page 1914), [R378] (page 1914)
sympy.solvers.recurr.rsolve ratio(coes, f, n, **hints)
Given linear recurrence operator L of order k with polynomial coecients and inhomogeneous equation L y = f , where f is a polynomial, we seek for all rational solutions over
eld K of characteristic zero.
This procedure accepts only polynomials, however if you are interested in solving recurrence with rational coecients then use rsolve which will pre-process the given
equation and run this procedure with polynomial arguments.
The algorithm performs two basic steps:
1.Compute polynomial v(n) which can be used as universal denominator of any rational
solution of equation L y = f .
1480
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
2.Construct new linear dierence equation by substitution y(n) = u(n)/v(n) and solve
it for u(n) nding all its polynomial solutions. Return None if none were found.
Algorithm implemented here is a revised version of the original Abramovs algorithm,
developed in 1989. The new approach is much simpler to implement and has better
overall eciency. This method can be easily adapted to q-dierence equations case.
Besides nding rational solutions alone, this functions is an important part of Hyper algorithm were it is used to nd particular solution of inhomogeneous part of a recurrence.
See Also:
rsolve hyper (page 1481)
References
[R379] (page 1914)
Examples
>>> from sympy.abc import x
>>> from sympy.solvers.recurr import rsolve_ratio
>>> rsolve_ratio([-2*x**3 + x**2 + 2*x - 1, 2*x**3 + x**2 - 6*x,
... - 2*x**3 - 11*x**2 - 18*x - 9, 2*x**3 + 13*x**2 + 22*x + 8], 0, x)
C2*(2*x - 3)/(2*(x**2 - 1))
sympy.solvers.recurr.rsolve hyper(coes, f, n, **hints)
Given linear recurrence operator L of order k with polynomial coecients and inhomogeneous equation L y = f we seek for all hypergeometric solutions over eld K of
characteristic zero.
The inhomogeneous part can be either hypergeometric or a sum of a xed number of
pairwise dissimilar hypergeometric terms.
The algorithm performs three basic steps:
1.Group together similar hypergeometric terms in the inhomogeneous part of L y = f ,
and nd particular solution using Abramovs algorithm.
2.Compute generating set of L and nd basis in it, so that all solutions are linearly
independent.
3.Form nal solution with the number of arbitrary constants equal to dimension of
basis of L.
Term a(n) is hypergeometric if it is annihilated by rst order linear dierence equations
with polynomial coecients or, in simpler words, if consecutive term ratio is a rational
function.
The output of this procedure is a linear combination of xed number of hypergeometric terms. However the underlying method can generate larger class of solutions DAlembertian terms.
Note also that this method not only computes the kernel of the inhomogeneous equation,
but also reduces in to a basis so that solutions generated by this procedure are linearly
independent
5.29. Solvers
1481
SymPy Documentation, Release 0.7.6
References
[R380] (page 1914), [R381] (page 1914)
Examples
>>> from sympy.solvers import rsolve_hyper
>>> from sympy.abc import x
>>> rsolve_hyper([-1, -1, 1], 0, x)
C0*(1/2 + sqrt(5)/2)**x + C1*(-sqrt(5)/2 + 1/2)**x
>>> rsolve_hyper([-1, 1], 1 + x, x)
C0 + x*(x + 1)/2
5.29.6 Systems of Polynomial Equations
sympy.solvers.polysys.solve poly system(seq, *gens, **args)
Solve a system of polynomial equations.
Examples
>>> from sympy import solve_poly_system
>>> from sympy.abc import x, y
>>> solve_poly_system([x*y - 2*y, 2*y**2 - x**2], x, y)
[(0, 0), (2, -sqrt(2)), (2, sqrt(2))]
sympy.solvers.polysys.solve triangulated(polys, *gens, **args)
Solve a polynomial system using Gianni-Kalkbrenner algorithm.
The algorithm proceeds by computing one Groebner basis in the ground domain and
then by iteratively computing polynomial factorizations in appropriately constructed algebraic extensions of the ground domain.
References
1. Patrizia Gianni, Teo Mora, Algebraic Solution of System of Polynomial Equations
using Groebner Bases, AAECC-5 on Applied Algebra, Algebraic Algorithms and ErrorCorrecting Codes, LNCS 356 247257, 1989
Examples
>>> from sympy.solvers.polysys import solve_triangulated
>>> from sympy.abc import x, y, z
>>> F = [x**2 + y + z - 1, x + y**2 + z - 1, x + y + z**2 - 1]
1482
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> solve_triangulated(F, x, y, z)
[(0, 0, 1), (0, 1, 0), (1, 0, 0)]
5.29.7 Diophantine Equations (DEs)
See Diophantine (page 1483)
5.29.8 Inequalities
See Inequality Solvers (page 1505)
5.30 Diophantine
5.30.1 Diophantine equations
The word Diophantine comes with the name Diophantus, a mathematician lived in the great
city of Alexandria sometime around 250 AD. Often referred to as the father of Algebra, Diophantus in his famous work Arithmetica presented 150 problems that marked the early
beginnings of number theory, the eld of study about integers and their properties. Diophantine equations play a central and an important part in number theory.
We call a Diophantine equation to an equation of the form, f (x1 , x2 , . . . xn ) = 0 where n 2 and
x1 , x2 , . . . xn are integer variables. If we can nd n integers a1 , a2 , . . . an such that x1 = a1 , x2 =
a2 , . . . xn = an satises the above equation, we say that the equation is solvable. You can read
more about Diophantine equations in 5 and 6 .
Currently, following ve types of Diophantine equations can be solved using diophantine()
(page 1487) and other helper functions of the Diophantine module.
Linear Diophantine equations: a1 x1 + a2 x2 + . . . + an xn = b.
General binary quadratic equation: ax2 + bxy + cy 2 + dx + ey + f = 0
Homogeneous ternary quadratic equation: ax2 + by 2 + cz 2 + dxy + eyz + f zx = 0
Extended Pythagorean equation: a1 x21 + a2 x22 + . . . + an x2n = an+1 x2n+1
General sum of squares: x21 + x22 + . . . + x2n = k
5.30.2 Module structure
This module contains diophantine() (page 1487) and helper functions that are needed to
solve certain Diophantine equations. Its structured in the following manner.
diophantine() (page 1487)
diop solve() (page 1488)
* classify diop() (page 1489)
* diop linear() (page 1490)
5
6
Andreescu, Titu. Andrica, Dorin. Cucurezeanu, Ion. An Introduction to Diophantine Equations
Diophantine Equation, Wolfram Mathworld, [online]. Available: http://mathworld.wolfram.com/DiophantineEquation.html
5.30. Diophantine
1483
SymPy Documentation, Release 0.7.6
* diop quadratic() (page 1491)
* diop ternary quadratic() (page 1495)
* diop general pythagorean() (page 1497)
* diop general sum of squares() (page 1497)
merge solution() (page 1500)
When an equation is given to diophantine() (page 1487), it factors the equation(if possible)
and solves the equation given by each factor by calling diop solve() (page 1488) separately.
Then all the results are combined using merge solution() (page 1500).
diop solve() (page 1488) internally uses classify diop() (page 1489) to nd the type of
the equation(and some other details) given to it and then calls the appropriate solver function
based on the type returned. For example, if classify diop() (page 1489) returned linear
as the type of the equation, then diop solve() (page 1488) calls diop linear() (page 1490)
to solve the equation.
Each of the functions, diop linear() (page 1490), diop quadratic() (page 1491),
diop ternary quadratic() (page 1495), diop general pythagorean() (page 1497) and
diop general sum of squares() (page 1497) solves a specic type of equations and the type
can be easily guessed by its name.
Apart from these functions, there are a considerable number of other functions in the Diophantine Module and all of them are listed under User functions and Internal functions.
5.30.3 Tutorial
First, lets import the highest API of the Diophantine module.
>>> from sympy.solvers.diophantine import diophantine
Before we start solving the equations, we need to dene the variables.
>>> from sympy import symbols
>>> x, y, z = symbols(x, y, z, integer=True)
Lets start by solving the easiest type of Diophantine equations, i.e. linear Diophantine equations. Lets solve 2x + 3y = 5. Note that although we write the equation in the above form,
when we input the equation to any of the functions in Diophantine module, it needs to be in
the form eq = 0.
>>> diophantine(2*x + 3*y - 5)
set([(3*t - 5, -2*t + 5)])
Note that stepping one more level below the highest API, we can solve the very same equation
by calling diop solve() (page 1488).
>>> from sympy.solvers.diophantine import diop_solve
>>> diop_solve(2*x + 3*y - 5)
(3*t - 5, -2*t + 5)
Note that it returns a tuple rather than a set. diophantine() (page 1487) always return
a set of tuples. But diop solve() (page 1488) may return a single tuple or a set of tuples
depending on the type of the equation given.
We can also solve this equation by calling diop linear() (page 1490), which is what
diop solve() (page 1488) calls internally.
1484
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> from sympy.solvers.diophantine import diop_linear
>>> diop_linear(2*x + 3*y - 5)
(3*t - 5, -2*t + 5)
If the given equation has no solutions then the outputs will look like below.
>>> diophantine(2*x + 4*y - 3)
set()
>>> diop_solve(2*x + 4*y - 3)
(None, None)
>>> diop_linear(2*x + 4*y - 3)
(None, None)
Note that except for the highest level API, in case of no solutions, a tuple of N one are returned. Size of the tuple is the same as the number of variables. Also, one can specically
set the parameter to be used in the solutions by passing a customized parameter. Consider
the following example.
>>> m = symbols(m, integer=True)
>>> diop_solve(2*x + 3*y - 5, m)
(3*m - 5, -2*m + 5)
Please note that for the moment, user can set the parameter only for linear Diophantine
equations and binary quadratic equations.
Lets try solving a binary quadratic equation which is an equation with two variables and has a
degree of two. Before trying to solve these equations, an idea about various cases associated
with the equation would help a lot. Please refer 7 and 8 for detailed analysis of dierent
cases and the nature of the solutions. Let us dene = b2 4ac w.r.t. the binary quadratic
ax2 + bxy + cy 2 + dx + ey + f = 0.
When < 0, there are either no solutions or only a nite number of solutions.
>>> diophantine(x**2 - 4*x*y + 8*y**2 - 3*x + 7*y - 5)
set([(2, 1), (5, 1)])
In the above equation = (4)2 4 1 8 = 16 and hence only a nite number of solutions
exist.
When = 0 we might have either no solutions or parameterized solutions.
>>> diophantine(3*x**2 - 6*x*y + 3*y**2 - 3*x + 7*y - 5)
set()
>>> diophantine(x**2 - 4*x*y + 4*y**2 - 3*x + 7*y - 5)
set([(-2*t**2 - 7*t + 10, -t**2 - 3*t + 5)])
>>> diophantine(x**2 + 2*x*y + y**2 - 3*x - 3*y)
set([(t, -t), (t, -t + 3)])
The most interesting case is when > 0 and it is not a perfect square. In this case, the
equation has either no solutions or an innte number of solutions. Consider the below cases
where = 8.
>>> diophantine(x**2 - 4*x*y + 2*y**2 - 3*x + 7*y - 5)
set()
>>> from sympy import expand
>>> n = symbols(n, integer=True)
7
Methods to solve Ax2 + Bxy + Cy2 + Dx + Ey + F = 0,[online],
Available:
http://www.alpertron.com.ar/METHODS.HTM
8 Solving the equation ax2+ bxy + cy2 + dx + ey + f= 0, [online], Available: http://www.jpr2718.org/ax2p.pdf
5.30. Diophantine
1485
SymPy Documentation, Release 0.7.6
>>> s = diophantine(x**2 - 2*y**2 - 2*x - 4*y, n)
>>> x_n, y_n = s.pop()
>>> expand(x_n)
-(-2*sqrt(2) + 3)**n/2 + sqrt(2)*(-2*sqrt(2) + 3)**n/2 - sqrt(2)*(2*sqrt(2) + 3)**n/2 - (2*sqrt(2) +
>>> expand(y_n)
-sqrt(2)*(-2*sqrt(2) + 3)**n/4 + (-2*sqrt(2) + 3)**n/2 + sqrt(2)*(2*sqrt(2) + 3)**n/4 + (2*sqrt(2) +
Here n is an integer. Although x n and y n may not look like integers, substituting in specic
values for n (and simplifying) shows that they are. For example consider the following example
where we set n equal to 9.
>>> from sympy import simplify
>>> simplify(x_n.subs({n: 9}))
-9369318
Any binary quadratic of the form ax2 + bxy + cy 2 + dx + ey + f = 0 can be transformed to an
equivalent form X 2 DY 2 = N .
>>> from sympy.solvers.diophantine import find_DN, diop_DN, transformation_to_DN
>>> find_DN(x**2 - 3*x*y + y**2 - 7*x + 5*y - 3)
(5, 920)
So, the above equation is equivalent to the equation X 2 5Y 2 = 920 after a linear transformation. If we want to nd the linear transformation, we can use transformation to DN()
(page 1494)
>>> A, B = transformation_to_DN(x**2 - 3*x*y + y**2 - 7*x + 5*y - 3)
Here A is a 2 X 2 matrix and B is a 2 X 1 matrix such that the transformation
[ ]
[ ]
X
x
=A
+B
Y
y
gives the equation X 2 5Y 2 = 920. Values of A and B are as belows.
>>> A
Matrix([
[1/10, 3/10],
[
0, 1/5]])
>>> B
Matrix([
[ 1/5],
[-11/5]])
We can solve an equation of the form X 2 DY 2 = N by passing D and N to diop DN()
(page 1492)
>>> diop_DN(5, 920)
[]
Unfortunately, our equation does not have solutions.
Now lets turn to homogeneous ternary quadratic equations. These equations are of the form
ax2 +by 2 +cz 2 +dxy +eyz +f zx = 0. These type of equations either have innitely many solutions
or no solutions (except the obvious solution (0, 0, 0))
>>> diophantine(3*x**2 + 4*y**2 - 5*z**2 + 4*x*y + 6*y*z + 7*z*x)
set()
>>> diophantine(3*x**2 + 4*y**2 - 5*z**2 + 4*x*y - 7*y*z + 7*z*x)
set([(-16*p**2 + 28*p*q + 20*q**2, 3*p**2 + 38*p*q - 25*q**2, 4*p**2 - 24*p*q + 68*q**2)])
1486
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
If you are only interested about a base solution rather than the parameterized general solution
(to be more precise, one of the general solutions), you can use diop ternary quadratic()
(page 1495).
>>> from sympy.solvers.diophantine import diop_ternary_quadratic
>>> diop_ternary_quadratic(3*x**2 + 4*y**2 - 5*z**2 + 4*x*y - 7*y*z + 7*z*x)
(-4, 5, 1)
diop ternary quadratic() (page 1495) rst converts the given equation to an equivalent
equation of the form w2 = AX 2 + BY 2 and then it uses descent() (page 1496) to solve the
latter equation. You can refer to the docs of transformation to normal() to nd more on
this. The equation w2 = AX 2 + BY 2 can be solved more easily by using the Aforementioned
descent() (page 1496).
>>> from sympy.solvers.diophantine import descent
>>> descent(3, 1) # solves the equation w**2 = 3*Y**2 + Z**2
(1, 0, 1)
Here the solution tuple is in the order (w, Y, Z)
The extended Pythagorean equation, a1 x21 + a2 x22 + . . . + an x2n = an+1 x2n+1 and the general sum
of squares equation, x21 + x22 + . . . + x2n = k can also be solved using the Diophantine module.
>>> from sympy.abc import a, b, c, d, e, f
>>> diophantine(9*a**2 + 16*b**2 + c**2 + 49*d**2 + 4*e**2 - 25*f**2)
set([(70*t1**2 + 70*t2**2 + 70*t3**2 + 70*t4**2 - 70*t5**2, 105*t1*t5, 420*t2*t5, 60*t3*t5, 210*t4*t5
function diop general pythagorean() (page 1497) can also be called directly to solve the
same equation. This is true about the general sum of squares too. Either you can call
diop general pythagorean() (page 1497) or use the high level API.
>>> diophantine(a**2 + b**2 + c**2 + d**2 + e**2 + f**2 - 112)
set([(8, 4, 4, 4, 0, 0)])
If you want to get a more thorough idea about the the Diophantine module please refer to the
following blog.
http://thilinaatsympy.wordpress.com/
5.30.4 References
5.30.5 User Functions
These are functions that are imported into the global namespace with from sympy import *.
These functions are intended for use by ordinary users of SymPy.
diophantine()
sympy.solvers.diophantine.diophantine(eq, param=t)
Simplify the solution procedure of diophantine equation eq by converting it into a product
of terms which should equal zero.
For example, when solving, x2 y 2 = 0 this is treated as (x + y)(x y) = 0 and x + y = 0
and x y = 0 are solved independently and combined. Each term is solved by calling
diop solve().
5.30. Diophantine
1487
SymPy Documentation, Release 0.7.6
Output of diophantine() is a set of tuples. Each tuple represents a solution of the input
equation. In a tuple, solution for each variable is listed according to the alphabetic order
of input variables. i.e. if we have an equation with two variables a and b, rst element of
the tuple will give the solution for a and the second element will give the solution for b.
See Also:
diop solve
Examples
>>> from sympy.solvers.diophantine import diophantine
>>> from sympy.abc import x, y, z
>>> diophantine(x**2 - y**2)
set([(-t, -t), (t, -t)])
#>>> diophantine(x*(2*x + 3*y - z)) #set([(0, n1, n2), (3*t - z, -2*t + z, z)]) #>>> diophantine(x**2 + 3*x*y + 4*x) #set([(0, n1), (3*t - 4, -t)])
Usage
diophantine(eq, t): Solve the diophantine equation eq. t is the parameter to be used
by diop solve().
Details
eq should be an expression which is assumed to be zero. t is the parameter to be used
in the solution.
diop solve()
sympy.solvers.diophantine.diop solve(eq, param=t)
Solves the diophantine equation eq.
Similar to diophantine() but doesnt try to factor eq as latter does. Uses classify diop() to determine the type of the eqaution and calls the appropriate solver function.
See Also:
diophantine
Examples
>>> from sympy.solvers.diophantine import diop_solve
>>> from sympy.abc import x, y, z, w
>>> diop_solve(2*x + 3*y - 5)
(3*t - 5, -2*t + 5)
>>> diop_solve(4*x + 3*y -4*z + 5)
(3*t + 4*z - 5, -4*t - 4*z + 5, z)
>>> diop_solve(x + 3*y - 4*z + w -6)
(t, -t - 3*y + 4*z + 6, y, z)
1488
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> diop_solve(x**2 + y**2 - 5)
set([(-2, -1), (-2, 1), (2, -1), (2, 1)])
Usage
diop solve(eq, t): Solve diophantine equation, eq using t as a parameter if needed.
Details
eq should be an expression which is assumed to be zero. t is a parameter to be used in
the solution.
classify diop()
sympy.solvers.diophantine.classify diop(eq)
Helper routine used by diop solve() to nd the type of the eq etc.
Returns a tuple containing the type of the diophantine equation along with the variables(free symbols) and their coecients. Variables are returned as a list and coecients are returned as a dict with the key being the respective term and the
constant term is keyed to Integer(1). Type is an element in the set {linear, binary quadratic, general pythagorean, homogeneous ternary quadratic, univariate, general sum of squares}
Examples
>>> from sympy.solvers.diophantine import classify_diop
>>> from sympy.abc import x, y, z, w, t
>>> classify_diop(4*x + 6*y - 4)
([x, y], {1: -4, x: 4, y: 6}, linear)
>>> classify_diop(x + 3*y -4*z + 5)
([x, y, z], {1: 5, x: 1, y: 3, z: -4}, linear)
>>> classify_diop(x**2 + y**2 - x*y + x + 5)
([x, y], {1: 5, x: 1, x**2: 1, y: 0, y**2: 1, x*y: -1}, binary_quadratic)
Usage
classify diop(eq): Return variables, coecients and type of the eq.
Details
eq should be an expression which is assumed to be zero.
5.30. Diophantine
1489
SymPy Documentation, Release 0.7.6
diop linear()
sympy.solvers.diophantine.diop linear(eq, param=t)
Solves linear diophantine equations.
A linear diophantine equation is an equation of the form a1 x1 + a2 x2 + .. + an xn = 0 where
a1 , a2 , ..an are integer constants and x1 , x2 , ..xn are integer variables.
See Also:
diop quadratic,
diop ternary quadratic,
diop general sum of squares
diop general pythagorean,
Examples
>>> from sympy.solvers.diophantine import diop_linear
>>> from sympy.abc import x, y, z, t
>>> from sympy import Integer
>>> diop_linear(2*x - 3*y - 5) #solves equation 2*x - 3*y -5 = 0
(-3*t - 5, -2*t - 5)
Here x = -3*t - 5 and y = -2*t - 5
>>> diop_linear(2*x - 3*y - 4*z -3)
(-3*t - 4*z - 3, -2*t - 4*z - 3, z)
Usage
diop linear(eq): Returns a tuple containing solutions to the diophantine equation eq.
Values in the tuple is arranged in the same order as the sorted variables.
Details
eq is a linear diophantine equation which is assumed to be zero. param is the parameter
to be used in the solution.
base solution linear()
sympy.solvers.diophantine.base solution linear(c, a, b, t=None)
Return the base solution for a linear diophantine equation with two variables.
Used by diop linear() to nd the base solution of a linear Diophantine equation. If t
is given then the parametrized solution is returned.
Examples
>>> from sympy.solvers.diophantine import base_solution_linear
>>> from sympy.abc import t
>>> base_solution_linear(5, 2, 3) # equation 2*x + 3*y = 5
(-5, 5)
>>> base_solution_linear(0, 5, 7) # equation 5*x + 7*y = 0
(0, 0)
1490
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> base_solution_linear(5, 2, 3, t) # equation 2*x + 3*y = 5
(3*t - 5, -2*t + 5)
>>> base_solution_linear(0, 5, 7, t) # equation 5*x + 7*y = 0
(7*t, -5*t)
Usage
base solution linear(c, a, b, t): a, b, c are coecients in ax + by = c and t is the
parameter to be used in the solution.
diop quadratic()
sympy.solvers.diophantine.diop quadratic(eq, param=t)
Solves quadratic diophantine equations.
i.e. equations of the form Ax2 + Bxy + Cy 2 + Dx + Ey + F = 0. Returns a set containing
the tuples (x, y) which contains the solutions. If there are no solutions then (N one, N one)
is returned.
See Also:
diop linear,
diop ternary quadratic,
diop general pythagorean
diop general sum of squares,
References
[R355] (page 1914), [R356] (page 1914)
Examples
>>> from sympy.abc import x, y, t
>>> from sympy.solvers.diophantine import diop_quadratic
>>> diop_quadratic(x**2 + y**2 + 2*x + 2*y + 2, t)
set([(-1, -1)])
Usage
diop quadratic(eq, param): eq is a quadratic binary diophantine equation. param is
used to indicate the parameter to be used in the solution.
Details
eq should be an expression which is assumed to be zero. param is a parameter to be used
in the solution.
5.30. Diophantine
1491
SymPy Documentation, Release 0.7.6
diop DN()
sympy.solvers.diophantine.diop DN(D, N, t=t)
Solves the equation x2 Dy 2 = N .
Mainly concerned in the case D > 0, D is not a perfect square, which is the same as
generalized Pell equation. To solve the generalized Pell equation this function Uses LMM
algorithm. Refer [R357] (page 1914) for more details on the algorithm. Returns one
solution for each class of the solutions. Other solutions of the class can be constructed
according to the values of D and N. Returns a list containing the solution tuples (x, y).
See Also:
find DN, diop bf DN
References
[R357] (page 1914)
Examples
>>> from sympy.solvers.diophantine import diop_DN
>>> diop_DN(13, -4) # Solves equation x**2 - 13*y**2 = -4
[(3, 1), (393, 109), (36, 10)]
The output can be interpreted as follows: There are three fundamental solutions to the
equation x2 13y 2 = 4 given by (3, 1), (393, 109) and (36, 10). Each tuple is in the form
(x, y), i. e solution (3, 1) means that x = 3 and y = 1.
>>> diop_DN(986, 1) # Solves equation x**2 - 986*y**2 = 1
[(49299, 1570)]
Usage
diop DN(D, N, t): D and N are integers as in x2 Dy 2 = N and t is the parameter to be
used in the solutions.
Details
D and N correspond to D and N in the equation. t is the parameter to be used in the
solutions.
cornacchia()
sympy.solvers.diophantine.cornacchia(a, b, m)
Solves ax2 + by 2 = m where gcd(a, b) = 1 = gcd(a, m) and a, b > 0.
Uses the algorithm due to Cornacchia. The method only nds primitive solutions, i.e.
ones with gcd(x, y) = 1. So this method cant be used to nd the solutions of x2 + y 2 = 20
since the only solution to former is (x, y) = (4, 2) and it is not primitive. When a = b =
1, only the solutions with x y are found. For more details, see the References.
1492
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
References
[R358] (page 1914), [R359] (page 1914)
Examples
>>> from sympy.solvers.diophantine import cornacchia
>>> cornacchia(2, 3, 35) # equation 2x**2 + 3y**2 = 35
set([(2, 3), (4, 1)])
>>> cornacchia(1, 1, 25) # equation x**2 + y**2 = 25
set([(4, 3)])
diop bf DN()
sympy.solvers.diophantine.diop bf DN(D, N, t=t)
Uses brute force to solve the equation, x2 Dy 2 = N .
Mainly concerned with the generalized Pell equation which is the case when D > 0, D is
not a perfect square. For more information on the case refer [R360] (page 1914). Let
2
2
(t, u) be the
minimal positive solution of the equation x Dy = 1. Then this method
requires
|N |(t1)
2D
to be small.
See Also:
diop DN
References
[R360] (page 1914)
Examples
>>> from sympy.solvers.diophantine import diop_bf_DN
>>> diop_bf_DN(13, -4)
[(3, 1), (-3, 1), (36, 10)]
>>> diop_bf_DN(986, 1)
[(49299, 1570)]
Usage
diop bf DN(D, N, t): D and N are coecients in x2 Dy 2 = N and t is the parameter to
be used in the solutions.
Details
D and N correspond to D and N in the equation. t is the parameter to be used in the
solutions.
5.30. Diophantine
1493
SymPy Documentation, Release 0.7.6
transformation to DN()
sympy.solvers.diophantine.transformation to DN(eq)
This function transforms general quadratic, ax2 + bxy + cy 2 + dx + ey + f = 0 to more easy
to deal with X 2 DY 2 = N form.
This is used to solve the general quadratic equation by transforming it to the latter form.
Refer [R361] (page 1914) for more detailed information on the transformation. This
function returns a tuple (A, B) where A is a 2 X 2 matrix and B is a 2 X 1 matrix such
that,
Transpose([x y]) = A * Transpose([X Y]) + B
See Also:
find DN
References
[R361] (page 1914)
Examples
>>> from sympy.abc import x, y
>>> from sympy.solvers.diophantine import transformation_to_DN
>>> from sympy.solvers.diophantine import classify_diop
>>> A, B = transformation_to_DN(x**2 - 3*x*y - y**2 - 2*y + 1)
>>> A
Matrix([
[1/26, 3/26],
[
0, 1/13]])
>>> B
Matrix([
[-6/13],
[-4/13]])
A, B returned are such that Transpose((x y)) = A * Transpose((X Y)) + B. Substituting
these values for x and y and a bit of simplifying work will give an equation of the form
x2 Dy 2 = N .
>>> from sympy.abc import X, Y
>>> from sympy import Matrix, simplify, Subs
>>> u = (A*Matrix([X, Y]) + B)[0] # Transformation for x
>>> u
X/26 + 3*Y/26 - 6/13
>>> v = (A*Matrix([X, Y]) + B)[1] # Transformation for y
>>> v
Y/13 - 4/13
Next we will substitute these formulas for x and y and do simplify().
>>> eq = simplify(Subs(x**2 - 3*x*y - y**2 - 2*y + 1, (x, y), (u, v)).doit())
>>> eq
X**2/676 - Y**2/52 + 17/13
By multiplying the denominator appropriately, we can get a Pell equation in the standard
form.
1494
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> eq * 676
X**2 - 13*Y**2 + 884
If only the nal equation is needed, find DN() can be used.
Usage
transformation to DN(eq): where eq is the quadratic to be transformed.
find DN()
sympy.solvers.diophantine.find DN(eq)
This function returns a tuple, (D, N ) of the simplied form, x2 Dy 2 = N , corresponding
to the general quadratic, ax2 + bxy + cy 2 + dx + ey + f = 0.
Solving the general quadratic is then equivalent to solving the equation X 2 DY 2 = N
and transforming the solutions by using the transformation matrices returned by transformation to DN().
See Also:
transformation to DN
References
[R362] (page 1914)
Examples
>>> from sympy.abc import x, y
>>> from sympy.solvers.diophantine import find_DN
>>> find_DN(x**2 - 3*x*y - y**2 - 2*y + 1)
(13, -884)
Interpretation of the output is that we get X 2 13Y 2 = 884 after transforming x2 3xy
y 2 2y + 1 using the transformation returned by transformation to DN().
Usage
find DN(eq): where eq is the quadratic to be transformed.
diop ternary quadratic()
sympy.solvers.diophantine.diop ternary quadratic(eq)
Solves the general quadratic ternary form, ax2 + by 2 + cz 2 + f xy + gyz + hxz = 0.
Returns a tuple (x, y, z) which is a base solution for the above equation. If there are no
solutions, (N one, N one, N one) is returned.
5.30. Diophantine
1495
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.abc import x, y, z
>>> from sympy.solvers.diophantine import diop_ternary_quadratic
>>> diop_ternary_quadratic(x**2 + 3*y**2 - z**2)
(1, 0, 1)
>>> diop_ternary_quadratic(4*x**2 + 5*y**2 - z**2)
(1, 0, 2)
>>> diop_ternary_quadratic(45*x**2 - 7*y**2 - 8*x*y - z**2)
(28, 45, 105)
>>> diop_ternary_quadratic(x**2 - 49*y**2 - z**2 + 13*z*y -8*x*y)
(9, 1, 5)
Usage
diop ternary quadratic(eq): Return a tuple containing a basic solution to eq.
Details
eq should be an homogeneous expression of degree two in three variables and it is assumed to be zero.
square factor()
sympy.solvers.diophantine.square factor(a)
Returns an integer c s.t. a = c2 k, c, k Z. Here k is square free.
Examples
>>>
>>>
2
>>>
6
>>>
1
from sympy.solvers.diophantine import square_factor
square_factor(24)
square_factor(36)
square_factor(1)
descent()
sympy.solvers.diophantine.descent(A, B)
Lagranges descent() with lattice-reduction to nd solutions to x2 = Ay 2 + Bz 2 .
Here A and B should be square free and pairwise prime. Always should be called with
suitable A and B so that the above equation has solutions.
This is more faster than the normal Lagranges descent algorithm because the gaussian
reduction is used.
1496
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
References
[R363] (page 1915)
Examples
>>> from sympy.solvers.diophantine import descent
>>> descent(3, 1) # x**2 = 3*y**2 + z**2
(1, 0, 1)
(x, y, z) = (1, 0, 1) is a solution to the above equation.
>>> descent(41, -113)
(-16, -3, 1)
diop general pythagorean()
sympy.solvers.diophantine.diop general pythagorean(eq, param=m)
Solves the general pythagorean equation, a21 x21 + a22 x22 + ... + a2n x2n a2n+1 x2n+1 = 0.
Returns a tuple which contains a parametrized solution to the equation, sorted in the
same order as the input variables.
Examples
>>> from sympy.solvers.diophantine import diop_general_pythagorean
>>> from sympy.abc import a, b, c, d, e
>>> diop_general_pythagorean(a**2 + b**2 + c**2 - d**2)
(m1**2 + m2**2 - m3**2, 2*m1*m3, 2*m2*m3, m1**2 + m2**2 + m3**2)
>>> diop_general_pythagorean(9*a**2 - 4*b**2 + 16*c**2 + 25*d**2 + e**2)
(10*m1**2 + 10*m2**2 + 10*m3**2 - 10*m4**2, 15*m1**2 + 15*m2**2 + 15*m3**2
+ 15*m4**2, 15*m
Usage
diop general pythagorean(eq, param): where eq is a general pythagorean equation
which is assumed to be zero and param is the base parameter used to construct other
parameters by subscripting.
diop general sum of squares()
sympy.solvers.diophantine.diop general sum of squares(eq, limit=1)
Solves the equation x21 + x22 + ... + x2n k = 0.
Returns at most limit number of solutions. Currently there is no way to set limit using
higher level APIs like diophantine() or diop solve() but that will be xed soon.
5.30. Diophantine
1497
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.solvers.diophantine import diop_general_sum_of_squares
>>> from sympy.abc import a, b, c, d, e, f
>>> diop_general_sum_of_squares(a**2 + b**2 + c**2 + d**2 + e**2 - 2345)
set([(0, 48, 5, 4, 0)])
Usage
general sum of squares(eq, limit) : Here eq is an expression which is assumed to be
zero. Also, eq should be in the form, x21 + x22 + ... + x2n k = 0. At most limit number of
solutions are returned.
Details
When n = 3 if k = 4a (8m + 7) for some a, m Z then there will be no solutions. Refer
[R364] (page 1915) for more details.
Reference
partition()
sympy.solvers.diophantine.partition(n, k=None, zeros=False)
Returns a generator that can be used to generate partitions of an integer n.
A partition of n is a set of positive integers which add upto n. For example, partitions of
3 are 3 , 1 + 2, 1 + 1+ 1. A partition is returned as a tuple. If k equals None, then all
possible partitions are returned irrespective of their size, otherwise only the partitions
of size k are returned. If there are no partions of n with size k then an empty tuple is
returned. If the zero parameter is set to True then a suitable number of zeros are added
at the end of every partition of size less than k.
zero parameter is considered only if k is not None. When the partitions are over, the
last next() call throws the StopIteration exception, so this function should always be
used inside a try - except block.
Examples
>>>
>>>
>>>
(1,
>>>
(1,
>>>
>>>
(3,
>>>
(2,
1498
from sympy.solvers.diophantine import partition
f = partition(5)
next(f)
1, 1, 1, 1)
next(f)
1, 1, 2)
g = partition(5, 3)
next(g)
1, 1)
next(g)
2, 1)
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Details
partition(n, k): Here n is a positive integer and k is the size of the partition which is
also positive integer.
Reference
sum of three squares()
sympy.solvers.diophantine.sum of three squares(n)
Returns a 3-tuple (a, b, c) such that a2 + b2 + c2 = n and a, b, c 0.
Returns (None, None, None) if n = 4a (8m + 7) for some a, m Z. See [R366] (page 1915)
for more details.
References
[R366] (page 1915)
Examples
>>> from sympy.solvers.diophantine import sum_of_three_squares
>>> sum_of_three_squares(44542)
(207, 37, 18)
Usage
sum of three squares(n): Here n is a non-negative integer.
sum of four squares()
sympy.solvers.diophantine.sum of four squares(n)
Returns a 4-tuple (a, b, c, d) such that a2 + b2 + c2 + d2 = n.
Here a, b, c, d 0.
References
[R367] (page 1915)
Examples
>>>
>>>
(8,
>>>
(0,
from sympy.solvers.diophantine import sum_of_four_squares
sum_of_four_squares(3456)
48, 32, 8)
sum_of_four_squares(1294585930293)
1137796, 2161, 1234)
5.30. Diophantine
1499
SymPy Documentation, Release 0.7.6
Usage
sum of four squares(n): Here n is a non-negative integer.
5.30.6 Internal Functions
These functions are intended for the internal use in Diophantine module.
merge solution
sympy.solvers.diophantine.merge solution(var, var t, solution)
This is used to construct the full solution from the solutions of sub equations.
For example when solving the equation (x y)(x2 + y 2 z 2 ) = 0, solutions for each of
the equations x y = 0 and x2 + y 2 z 2 are found independently. Solutions for x y = 0
are (x, y) = (t, t). But we should introduce a value for z when we output the solution
for the original equation. This function converts (t, t) into (t, t, n1 ) where n1 is an integer
parameter.
divisible
sympy.solvers.diophantine.divisible(a, b)
Returns T rue if a is divisible by b and F alse otherwise.
extended euclid
sympy.solvers.diophantine.extended euclid(a, b)
For given a, b returns a tuple containing integers x, y and d such that ax + by = d. Here
d = gcd(a, b).
Examples
>>> from sympy.solvers.diophantine import extended_euclid
>>> extended_euclid(4, 6)
(-1, 1, 2)
>>> extended_euclid(3, 5)
(2, -1, 1)
Usage
extended euclid(a, b): returns x, y and gcd(a, b).
Details
a Any instance of Integer. b Any instance of Integer.
1500
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
PQa
sympy.solvers.diophantine.PQa(P 0, Q 0, D)
Returns useful information needed to solve the Pell equation.
There are six
sequences of integers dened related to the continued fraction represen
P+ D
tation of Q , namely {Pi }, {Qi }, {ai },{Ai }, {Bi }, {Gi }. PQa() Returns these values as
a 6-tuple in the same order as mentioned above. Refer [R368] (page 1915) for more
detailed information.
References
[R368] (page 1915)
Examples
>>> from sympy.solvers.diophantine import PQa
>>> pqa = PQa(13, 4, 5) # (13 + sqrt(5))/4
>>> next(pqa) # (P_0, Q_0, a_0, A_0, B_0, G_0)
(13, 4, 3, 3, 1, -1)
>>> next(pqa) # (P_1, Q_1, a_1, A_1, B_1, G_1)
(-1, 1, 1, 4, 1, 3)
Usage
PQa(P 0, Q 0, D): P 0, Q 0 and D are integers corresponding to P0 , Q0 and D in the
D
. Also its assumed that P02 == Dmod(|Q0 |) and D is square free.
continued fraction P0 +
Q0
equivalent
sympy.solvers.diophantine.equivalent(u, v, r, s, D, N)
Returns True if two solutions (u, v) and (r, s) of x2 Dy 2 = N belongs to the same equivalence class and False otherwise.
Two solutions (u, v) and (r, s) to the above equation fall to the same equivalence class i
both (ur Dvs) and (us vr) are divisible by N . See reference [R369] (page 1915). No
test is performed to test whether (u, v) and (r, s) are actually solutions to the equation.
User should take care of this.
References
[R369] (page 1915)
Examples
5.30. Diophantine
1501
SymPy Documentation, Release 0.7.6
>>> from sympy.solvers.diophantine import equivalent
>>> equivalent(18, 5, -18, -5, 13, -1)
True
>>> equivalent(3, 1, -18, 393, 109, -4)
False
Usage
equivalent(u, v, r, s, D, N): (u, v) and (r, s) are two solutions of the equation
x2 Dy 2 = N and all parameters involved are integers.
simplified
sympy.solvers.diophantine.simplified(x, y, z)
Simplify the solution (x, y, z).
parametrize ternary quadratic
sympy.solvers.diophantine.parametrize ternary quadratic(eq)
Returns the parametrized general solution for the ternary quadratic equation eq which
has the form ax2 + by 2 + cz 2 + f xy + gyz + hxz = 0.
References
[R370] (page 1915)
Examples
>>> from sympy.abc import x, y, z
>>> from sympy.solvers.diophantine import parametrize_ternary_quadratic
>>> parametrize_ternary_quadratic(x**2 + y**2 - z**2)
(2*p*q, p**2 - q**2, p**2 + q**2)
Here p and q are two co-prime integers.
>>> parametrize_ternary_quadratic(3*x**2 + 2*y**2 - z**2 - 2*x*y + 5*y*z - 7*y*z)
(2*p**2 - 2*p*q - q**2, 2*p**2 + 2*p*q - q**2, 2*p**2 - 2*p*q + 3*q**2)
>>> parametrize_ternary_quadratic(124*x**2 - 30*y**2 - 7729*z**2)
(-1410*p**2 - 363263*q**2, 2700*p**2 + 30916*p*q - 695610*q**2, -60*p**2 + 5400*p*q + 15458*q**2
diop ternary quadratic normal
sympy.solvers.diophantine.diop ternary quadratic normal(eq)
Solves the quadratic ternary diophantine equation, ax2 + by 2 + cz 2 = 0.
Here the coecients a, b, and c should be non zero. Otherwise the equation will be a
quadratic binary or univariate equation. If solvable, returns a tuple (x, y, z) that satisifes
the given equation. If the equation does not have integer solutions, (N one, N one, N one) is
returned.
1502
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>>
>>>
>>>
(1,
>>>
(1,
>>>
(4,
from sympy.abc import x, y, z
from sympy.solvers.diophantine import diop_ternary_quadratic_normal
diop_ternary_quadratic_normal(x**2 + 3*y**2 - z**2)
0, 1)
diop_ternary_quadratic_normal(4*x**2 + 5*y**2 - z**2)
0, 2)
diop_ternary_quadratic_normal(34*x**2 - 3*y**2 - 301*z**2)
9, 1)
Usage
diop ternary quadratic normal(eq): where eq is an equation of the form ax2 +by 2 +cz 2 =
0.
ldescent
sympy.solvers.diophantine.ldescent(A, B)
Uses Lagranges method to nd a non trivial solution to w2 = Ax2 + By 2 .
Here, A 6= 0 and B 6= 0 and A and B are square free. Output a tuple (w0 , x0 , y0 ) which is a
solution to the above equation.
References
[R371] (page 1915), [R372] (page 1915)
Examples
>>>
>>>
(1,
>>>
(2,
from sympy.solvers.diophantine import ldescent
ldescent(1, 1) # w^2 = x^2 + y^2
1, 0)
ldescent(4, -7) # w^2 = 4x^2 - 7y^2
-1, 0)
This means that x = 1, y = 0 and w = 2 is a solution to the equation w2 = 4x2 7y 2
>>> ldescent(5, -1) # w^2 = 5x^2 - y^2
(2, 1, -1)
gaussian reduce
sympy.solvers.diophantine.gaussian reduce(w, a, b)
Returns a reduced solution (x, z) to the congruence X 2 aZ 2 0 (mod b) so that x2 + |a|z 2
is minimal.
References
[R373] (page 1915), [R374] (page 1915)
5.30. Diophantine
1503
SymPy Documentation, Release 0.7.6
Details
Here w is a solution of the congruence x2 a (mod b)
holzer
sympy.solvers.diophantine.holzer(x 0, y 0, z 0, a, b, c)
Simplify the solution (x0 , y0 , z0 ) of the equation ax2 + by 2 = cz 2 with a, b, c > 0 and z02 | ab |
to a new reduced solution (x, y, z) such that z 2 | ab |.
prime as sum of two squares
sympy.solvers.diophantine.prime as sum of two squares(p)
Represent a prime p which is congruent to 1 mod 4, as a sum of two squares.
Examples
>>> from sympy.solvers.diophantine import prime_as_sum_of_two_squares
>>> prime_as_sum_of_two_squares(5)
(2, 1)
Reference
pairwise prime
sympy.solvers.diophantine.pairwise prime(a, b, c)
Transform ax2 + by 2 + cz 2 = 0 into an equivalent equation a0 x2 + b0 y 2 + c0 z 2 = 0 where a0 , b0 , c0
are pairwise relatively prime.
Returns a tuple containing a0 , b0 , c0 . gcd(a, b, c) should equal 1 for this to work. The solutions for ax2 + by 2 + cz 2 = 0 can be recovered from the solutions of a0 x2 + b0 y 2 + c0 z 2 = 0.
See Also:
make prime, reocnstruct
Examples
>>> from sympy.solvers.diophantine import pairwise_prime
>>> pairwise_prime(6, 15, 10)
(5, 2, 3)
make prime
sympy.solvers.diophantine.make prime(a, b, c)
Transform the equation ax2 + by 2 + cz 2 = 0 to an equivalent equation a0 x2 + b0 y 2 + c0 z 2 = 0
with gcd(a0 , b0 ) = 1.
Returns a tuple (a0 , b0 , c0 ) which satises above conditions. Note that in the returned tuple
gcd(a0 , c0 ) and gcd(b0 , c0 ) can take any value.
1504
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
See Also:
pairwaise prime, reconstruct
Examples
>>> from sympy.solvers.diophantine import make_prime
>>> make_prime(4, 2, 7)
(2, 1, 14)
reconstruct
sympy.solvers.diophantine.reconstruct(a, b, z)
Reconstruct the z value of an equivalent solution of ax2 + by 2 + cz 2 from the z value of a
solution of a transformed version of the above equation.
5.31 Inequality Solvers
sympy.solvers.inequalities.solve rational inequalities(eqs)
Solve a system of rational inequalities with rational coecients.
See Also:
solve poly inequality (page 1505)
Examples
>>> from sympy.abc import x
>>> from sympy import Poly
>>> from sympy.solvers.inequalities import solve_rational_inequalities
>>> solve_rational_inequalities([[
... ((Poly(-x + 1), Poly(1, x)), >=),
... ((Poly(-x + 1), Poly(1, x)), <=)]])
{1}
>>> solve_rational_inequalities([[
... ((Poly(x), Poly(1, x)), !=),
... ((Poly(-x + 1), Poly(1, x)), >=)]])
(-oo, 0) U (0, 1]
sympy.solvers.inequalities.solve poly inequality(poly, rel)
Solve a polynomial inequality with rational coecients.
See Also:
solve poly inequalities
5.31. Inequality Solvers
1505
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import Poly
>>> from sympy.abc import x
>>> from sympy.solvers.inequalities import solve_poly_inequality
>>> solve_poly_inequality(Poly(x, x, domain=ZZ), ==)
[{0}]
>>> solve_poly_inequality(Poly(x**2 - 1, x, domain=ZZ), !=)
[(-oo, -1), (-1, 1), (1, oo)]
>>> solve_poly_inequality(Poly(x**2 - 1, x, domain=ZZ), ==)
[{-1}, {1}]
sympy.solvers.inequalities.reduce rational inequalities(exprs,
gen,
tional=True)
Reduce a system of rational inequalities with rational coecients.
rela-
Examples
>>> from sympy import Poly, Symbol
>>> from sympy.solvers.inequalities import reduce_rational_inequalities
>>> x = Symbol(x, real=True)
>>> reduce_rational_inequalities([[x**2 <= 0]], x)
x == 0
>>> reduce_rational_inequalities([[x + 2 > 0]], x)
And(-2 < x, x < oo)
>>> reduce_rational_inequalities([[(x + 2, >)]], x)
And(-2 < x, x < oo)
>>> reduce_rational_inequalities([[x + 2]], x)
x == -2
sympy.solvers.inequalities.reduce abs inequality(expr, rel, gen)
Reduce an inequality with nested absolute values.
See Also:
reduce abs inequalities (page 1506)
Examples
>>> from sympy import Q, Abs, Symbol
>>> from sympy.solvers.inequalities import reduce_abs_inequality
>>> x = Symbol(x, real=True)
>>> reduce_abs_inequality(Abs(x - 5) - 3, <, x)
And(2 < x, x < 8)
>>> reduce_abs_inequality(Abs(x + 2)*3 - 13, <, x)
And(-19/3 < x, x < 7/3)
1506
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
sympy.solvers.inequalities.reduce abs inequalities(exprs, gen)
Reduce a system of inequalities with nested absolute values.
See Also:
reduce abs inequality (page 1506)
Examples
>>>
>>>
>>>
>>>
from sympy import Q, Abs, Symbol
from sympy.abc import x
from sympy.solvers.inequalities import reduce_abs_inequalities
x = Symbol(x, real=True)
>>> reduce_abs_inequalities([(Abs(3*x - 5) - 7, <),
... (Abs(x + 25) - 13, >)], x)
And(-2/3 < x, Or(And(-12 < x, x < oo), And(-oo < x, x < -38)), x < 4)
>>> reduce_abs_inequalities([(Abs(x - 4) + Abs(3*x - 5) - 7, <)], x)
And(1/2 < x, x < 4)
sympy.solvers.inequalities.reduce inequalities(inequalities, symbols=[])
Reduce a system of inequalities with rational coecients.
Examples
>>> from sympy import Q, sympify as S, Symbol
>>> from sympy.abc import x, y
>>> from sympy.solvers.inequalities import reduce_inequalities
>>> x = Symbol(x, real=True)
>>> reduce_inequalities(S(0) <= x + 3, [])
And(-3 <= x, x < oo)
>>> x = Symbol(x)
>>> reduce_inequalities(S(0) <= x + y*2 - 1, [x])
-2*y + 1 <= x
sympy.solvers.inequalities.solve univariate inequality(expr,
gen,
tional=True)
Solves a real univariate inequality.
rela-
Examples
>>> from sympy.solvers.inequalities import solve_univariate_inequality
>>> from sympy.core.symbol import Symbol
>>> x = Symbol(x, real=True)
>>> solve_univariate_inequality(x**2 >= 4, x)
Or(And(-oo < x, x <= -2), And(2 <= x, x < oo))
>>> solve_univariate_inequality(x**2 >= 4, x, relational=False)
(-oo, -2] U [2, oo)
5.31. Inequality Solvers
1507
SymPy Documentation, Release 0.7.6
5.32 Tensor Module
A module to manipulate symbolic objects with indices including tensors
5.32.1 Contents
Indexed Objects
Module that denes indexed objects
The classes IndexedBase, Indexed and Idx would represent a matrix element M[i, j] as in the
following graph:
1) The Indexed class represents the entire indexed object.
|
|
M[i, j]
/
\ \
|
|
|
|
|
2) The Idx class represent indices and each Idx can
|
optionally contain information about its range.
|
3) IndexedBase represents the stem of an indexed object, here M.
The stem used by itself is usually taken to represent the entire
array.
There can be any number of indices on an Indexed object. No transformation properties are
implemented in these Base objects, but implicit contraction of repeated indices is supported.
Note that the support for complicated (i.e. non-atomic) integer expressions as indices is
limited. (This should be improved in future releases.)
Examples
To express the above matrix element example you would write:
>>> from sympy.tensor import IndexedBase, Idx
>>> from sympy import symbols
>>> M = IndexedBase(M)
>>> i, j = symbols(i j, cls=Idx)
>>> M[i, j]
M[i, j]
Repeated indices in a product implies a summation, so to express a matrix-vector product in
terms of Indexed objects:
>>> x = IndexedBase(x)
>>> M[i, j]*x[j]
x[j]*M[i, j]
If the indexed objects will be converted to component based arrays, e.g. with the code printers
or the autowrap framework, you also need to provide (symbolic or numerical) dimensions.
This can be done by passing an optional shape parameter to IndexedBase upon construction:
1508
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> dim1, dim2 = symbols(dim1 dim2, integer=True)
>>> A = IndexedBase(A, shape=(dim1, 2*dim1, dim2))
>>> A.shape
(dim1, 2*dim1, dim2)
>>> A[i, j, 3].shape
(dim1, 2*dim1, dim2)
If an IndexedBase object has no shape information, it is assumed that the array is as large as
the ranges of its indices:
>>> n, m = symbols(n m, integer=True)
>>> i = Idx(i, m)
>>> j = Idx(j, n)
>>> M[i, j].shape
(m, n)
>>> M[i, j].ranges
[(0, m - 1), (0, n - 1)]
The above can be compared with the following:
>>> A[i, 2, j].shape
(dim1, 2*dim1, dim2)
>>> A[i, 2, j].ranges
[(0, m - 1), None, (0, n - 1)]
To analyze the structure of indexed expressions, you can use the methods get indices() and
get contraction structure():
>>> from sympy.tensor import get_indices, get_contraction_structure
>>> get_indices(A[i, j, j])
(set([i]), {})
>>> get_contraction_structure(A[i, j, j])
{(j,): set([A[i, j, j]])}
See the appropriate docstrings for a detailed explanation of the output.
class sympy.tensor.indexed.Idx
Represents an integer index as an Integer or integer expression.
There are a number of ways to create an Idx object. The constructor takes two arguments:
label An integer or a symbol that labels the index.
range Optionally you can specify a range as either
Symbol or integer: This is interpreted as a dimension. Lower and upper bounds are
set to 0 and range - 1, respectively.
tuple: The two elements are interpreted as the lower and upper bounds of the range,
respectively.
Note: the Idx constructor is rather pedantic in that it only accepts integer arguments.
The only exception is that you can use oo and -oo to specify an unbounded range. For all
other cases, both label and bounds must be declared as integers, e.g. if n is given as an
argument then n.is integer must return True.
For convenience, if the label is given as a string it is automatically converted to an integer
symbol. (Note: this conversion is not done for range or dimension arguments.)
5.32. Tensor Module
1509
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.tensor import IndexedBase, Idx
>>> from sympy import symbols, oo
>>> n, i, L, U = symbols(n i L U, integer=True)
If a string is given for the label an integer Symbol is created and the bounds are both
None:
>>> idx = Idx(qwerty); idx
qwerty
>>> idx.lower, idx.upper
(None, None)
Both upper and lower bounds can be specied:
>>> idx = Idx(i, (L, U)); idx
i
>>> idx.lower, idx.upper
(L, U)
When only a single bound is given it is interpreted as the dimension and the lower bound
defaults to 0:
>>>
(0,
>>>
(0,
>>>
(0,
idx
n idx
3)
idx
oo)
= Idx(i, n); idx.lower, idx.upper
1)
= Idx(i, 4); idx.lower, idx.upper
= Idx(i, oo); idx.lower, idx.upper
The label can be a literal integer instead of a string/Symbol:
>>> idx = Idx(2, n); idx.lower, idx.upper
(0, n - 1)
>>> idx.label
2
label
Returns the label (Integer or integer expression) of the Idx object.
Examples
>>>
>>>
2
>>>
>>>
j
>>>
j +
from sympy import Idx, Symbol
Idx(2).label
j = Symbol(j, integer=True)
Idx(j).label
Idx(j + 1).label
1
lower
Returns the lower bound of the Index.
1510
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import Idx
>>> Idx(j, 2).lower
0
>>> Idx(j, 5).lower
0
>>> Idx(j).lower is None
True
upper
Returns the upper bound of the Index.
Examples
>>> from sympy import Idx
>>> Idx(j, 2).upper
1
>>> Idx(j, 5).upper
4
>>> Idx(j).upper is None
True
class sympy.tensor.indexed.Indexed
Represents a mathematical object with indices.
>>> from sympy.tensor import Indexed, IndexedBase, Idx
>>> from sympy import symbols
>>> i, j = symbols(i j, cls=Idx)
>>> Indexed(A, i, j)
A[i, j]
It is recommended that Indexed objects are created via IndexedBase:
>>> A = IndexedBase(A)
>>> Indexed(A, i, j) == A[i, j]
True
base
Returns the IndexedBase of the Indexed object.
Examples
>>> from sympy.tensor import Indexed, IndexedBase, Idx
>>> from sympy import symbols
>>> i, j = symbols(i j, cls=Idx)
>>> Indexed(A, i, j).base
A
>>> B = IndexedBase(B)
>>> B == B[i, j].base
True
indices
Returns the indices of the Indexed object.
5.32. Tensor Module
1511
SymPy Documentation, Release 0.7.6
Examples
>>>
>>>
>>>
>>>
(i,
from sympy.tensor import Indexed, Idx
from sympy import symbols
i, j = symbols(i j, cls=Idx)
Indexed(A, i, j).indices
j)
ranges
Returns a list of tuples with lower and upper range of each index.
If an index does not dene the data members upper and lower, the corresponding
slot in the list contains None instead of a tuple.
Examples
>>> from sympy import Indexed,Idx, symbols
>>> Indexed(A, Idx(i, 2), Idx(j, 4), Idx(k, 8)).ranges
[(0, 1), (0, 3), (0, 7)]
>>> Indexed(A, Idx(i, 3), Idx(j, 3), Idx(k, 3)).ranges
[(0, 2), (0, 2), (0, 2)]
>>> x, y, z = symbols(x y z, integer=True)
>>> Indexed(A, x, y, z).ranges
[None, None, None]
rank
Returns the rank of the Indexed object.
Examples
>>> from sympy.tensor import Indexed, Idx
>>> from sympy import symbols
>>> i, j, k, l, m = symbols(i:m, cls=Idx)
>>> Indexed(A, i, j).rank
2
>>> q = Indexed(A, i, j, k, l, m)
>>> q.rank
5
>>> q.rank == len(q.indices)
True
shape
Returns a list with dimensions of each index.
Dimensions is a property of the array, not of the indices. Still, if the IndexedBase
does not dene a shape attribute, it is assumed that the ranges of the indices correspond to the shape of the array.
>>>
>>>
>>>
>>>
>>>
>>>
>>>
1512
from sympy.tensor.indexed import IndexedBase, Idx
from sympy import symbols
n, m = symbols(n m, integer=True)
i = Idx(i, m)
j = Idx(j, m)
A = IndexedBase(A, shape=(n, n))
B = IndexedBase(B)
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>>
(n,
>>>
(m,
A[i, j].shape
n)
B[i, j].shape
m)
class sympy.tensor.indexed.IndexedBase
Represent the base or stem of an indexed object
The IndexedBase class represent an array that contains elements. The main purpose
of this class is to allow the convenient creation of objects of the Indexed class. The
getitem method of IndexedBase returns an instance of Indexed. Alone, without indices, the IndexedBase class can be used as a notation for e.g. matrix equations, resembling what you could do with the Symbol class. But, the IndexedBase class adds
functionality that is not available for Symbol instances:
An IndexedBase object can optionally store shape information. This can be used in
to check array conformance and conditions for numpy broadcasting. (TODO)
An IndexedBase object implements syntactic sugar that allows easy symbolic representation of array operations, using implicit summation of repeated indices.
The IndexedBase object symbolizes a mathematical structure equivalent to arrays,
and is recognized as such for code generation and automatic compilation and wrapping.
>>> from sympy.tensor import IndexedBase, Idx
>>> from sympy import symbols
>>> A = IndexedBase(A); A
A
>>> type(A)
<class sympy.tensor.indexed.IndexedBase>
When an IndexedBase object receives indices, it returns an array with named axes, represented by an Indexed object:
>>> i, j = symbols(i j, integer=True)
>>> A[i, j, 2]
A[i, j, 2]
>>> type(A[i, j, 2])
<class sympy.tensor.indexed.Indexed>
The IndexedBase constructor takes an optional shape argument. If given, it overrides
any shape information in the indices. (But not the index ranges!)
>>>
>>>
>>>
>>>
(m,
>>>
>>>
(o,
m, n, o, p = symbols(m n o p, integer=True)
i = Idx(i, m)
j = Idx(j, n)
A[i, j].shape
n)
B = IndexedBase(B, shape=(o, p))
B[i, j].shape
p)
args
Returns the arguments used to create this IndexedBase object.
5.32. Tensor Module
1513
SymPy Documentation, Release 0.7.6
Examples
>>>
>>>
>>>
(A,
from sympy import IndexedBase
from sympy.abc import x, y
IndexedBase(A, shape=(x, y)).args
(x, y))
label
Returns the label of the IndexedBase object.
Examples
>>> from sympy import IndexedBase
>>> from sympy.abc import x, y
>>> IndexedBase(A, shape=(x, y)).label
A
shape
Returns the shape of the IndexedBase object.
Examples
>>>
>>>
>>>
(x,
from sympy import IndexedBase, Idx, Symbol
from sympy.abc import x, y
IndexedBase(A, shape=(x, y)).shape
y)
Note: If the shape of the IndexedBase is specied, it will override any shape information given by the indices.
>>>
>>>
>>>
>>>
>>>
(x,
>>>
(2,
A = IndexedBase(A, shape=(x, y))
B = IndexedBase(B)
i = Idx(i, 2)
j = Idx(j, 1)
A[i, j].shape
y)
B[i, j].shape
1)
Methods
Module with functions operating on IndexedBase, Indexed and Idx objects
Check shape conformance
Determine indices in resulting expression
etc.
Methods in this module could be implemented by calling methods on Expr objects instead.
When things stabilize this could be a useful refactoring.
sympy.tensor.index methods.get contraction structure(expr)
Determine dummy indices of expr and describe its structure
By dummy we mean indices that are summation indices.
1514
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
The stucture of the expression is determined and described as follows:
1.A conforming summation of Indexed objects is described with a dict where the keys
are summation indices and the corresponding values are sets containing all terms
for which the summation applies. All Add objects in the SymPy expression tree are
described like this.
2.For all nodes in the SymPy expression tree that are not of type Add, the following
applies:
If a node discovers contractions in one of its arguments, the node itself will be stored
as a key in the dict. For that key, the corresponding value is a list of dicts, each of
which is the result of a recursive call to get contraction structure(). The list contains
only dicts for the non-trivial deeper contractions, ommitting dicts with None as the
one and only key.
Note: The presence of expressions among the dictinary keys indicates multiple levels
of index contractions. A nested dict displays nested contractions and may itself contain dicts from a deeper level. In practical calculations the summation in the deepest
nested level must be calculated rst so that the outer expression can access the resulting
indexed object.
Examples
>>> from sympy.tensor.index_methods import get_contraction_structure
>>> from sympy import symbols, default_sort_key
>>> from sympy.tensor import IndexedBase, Idx
>>> x, y, A = map(IndexedBase, [x, y, A])
>>> i, j, k, l = map(Idx, [i, j, k, l])
>>> get_contraction_structure(x[i]*y[i] + A[j, j])
{(i,): set([x[i]*y[i]]), (j,): set([A[j, j]])}
>>> get_contraction_structure(x[i]*y[j])
{None: set([x[i]*y[j]])}
A multiplication of contracted factors results in nested dicts representing the internal
contractions.
>>> d = get_contraction_structure(x[i, i]*y[j, j])
>>> sorted(d.keys(), key=default_sort_key)
[None, x[i, i]*y[j, j]]
In this case, the product has no contractions:
>>> d[None]
set([x[i, i]*y[j, j]])
Factors are contracted rst:
>>> sorted(d[x[i, i]*y[j, j]], key=default_sort_key)
[{(i,): set([x[i, i]])}, {(j,): set([y[j, j]])}]
A parenthesized Add object is also returned as a nested dictionary. The term containing
the parenthesis is a Mul with a contraction among the arguments, so it will be found as
a key in the result. It stores the dictionary resulting from a recursive call on the Add
expression.
5.32. Tensor Module
1515
SymPy Documentation, Release 0.7.6
>>> d = get_contraction_structure(x[i]*(y[i] + A[i, j]*x[j]))
>>> sorted(d.keys(), key=default_sort_key)
[(x[j]*A[i, j] + y[i])*x[i], (i,)]
>>> d[(i,)]
set([(x[j]*A[i, j] + y[i])*x[i]])
>>> d[x[i]*(A[i, j]*x[j] + y[i])]
[{None: set([y[i]]), (j,): set([x[j]*A[i, j]])}]
Powers with contractions in either base or exponent will also be found as keys in the
dictionary, mapping to a list of results from recursive calls:
>>> d = get_contraction_structure(A[j, j]**A[i, i])
>>> d[None]
set([A[j, j]**A[i, i]])
>>> nested_contractions = d[A[j, j]**A[i, i]]
>>> nested_contractions[0]
{(j,): set([A[j, j]])}
>>> nested_contractions[1]
{(i,): set([A[i, i]])}
The description of the contraction structure may appear complicated when represented
with a string in the above examples, but it is easy to iterate over:
>>> from sympy import Expr
>>> for key in d:
...
if isinstance(key, Expr):
...
continue
...
for term in d[key]:
...
if term in d:
...
# treat deepest contraction first
...
pass
...
# treat outermost contactions here
sympy.tensor.index methods.get indices(expr)
Determine the outer indices of expression expr
By outer we mean indices that are not summation indices. Returns a set and a dict. The
set contains outer indices and the dict contains information about index symmetries.
Examples
>>>
>>>
>>>
>>>
>>>
from sympy.tensor.index_methods import get_indices
from sympy import symbols
from sympy.tensor import IndexedBase, Idx
x, y, A = map(IndexedBase, [x, y, A])
i, j, a, z = symbols(i j a z, integer=True)
The indices of the total expression is determined, Repeated indices imply a summation,
for instance the trace of a matrix A:
>>> get_indices(A[i, i])
(set(), {})
In the case of many terms, the terms are required to have identical outer indices. Else
an IndexConformanceException is raised.
1516
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> get_indices(x[i] + A[i, j]*y[j])
(set([i]), {})
Exceptions
An IndexConformanceException means that the terms ar not compatible, e.g.
>>> get_indices(x[i] + y[j])
(...)
IndexConformanceException: Indices are not consistent: x(i) + y(j)
Warning: The concept of outer indices applies recursively, starting on the deepest
level. This implies that dummies inside parenthesis are assumed to be summed rst,
so that the following expression is handled gracefully:
>>> get_indices((x[i] + A[i, j]*y[j])*x[j])
(set([i, j]), {})
This is correct and may appear convenient, but you need to be careful with this as
SymPy will happily .expand() the product, if requested. The resulting expression
would mix the outer j with the dummies inside the parenthesis, which makes it a
dierent expression. To be on the safe side, it is best to avoid such ambiguities by
using unique indices for all contractions that should be held separate.
Tensor
class sympy.tensor.tensor. TensorManager
Class to manage tensor properties.
Notes
Tensors belong to tensor commutation groups; each group has a label comm; there are
predened labels:
0 tensors commuting with any other tensor
1 tensors anticommuting among themselves
2 tensors not commuting, apart with those with comm=0
Other groups can be dened using set comm; tensors in those groups commute with those
with comm=0; by default they do not commute with any other group.
clear()
Clear the TensorManager.
comm i2symbol(i)
Returns the symbol corresponding to the commutation group number.
comm symbols2i(i)
get the commutation group number corresponding to i
i can be a symbol or a number or a string
If i is not already dened its commutation group number is set.
5.32. Tensor Module
1517
SymPy Documentation, Release 0.7.6
get comm(i, j)
Return the commutation parameter for commutation group numbers i, j
see TensorManager.set comm
set comm(i, j, c)
set the commutation parameter c for commutation groups i, j
Parameters i, j : symbols representing commutation groups
c : group commutation number
Notes
i, j can be symbols, strings or numbers, apart from 0, 1 and 2 which are reserved
respectively for commuting, anticommuting tensors and tensors not commuting with
any other group apart with the commuting tensors. For the remaining cases, use
this method to set the commutation rules; by default c=None.
The group commutation number c is assigned in correspondence to the group commutation symbols; it can be
0 commuting
1 anticommuting
None no commutation property
Examples
G and GH do not commute with themselves and commute with each other; A is commuting.
>>> from sympy.tensor.tensor import TensorIndexType, tensor_indices, tensorhead, TensorManag
>>> Lorentz = TensorIndexType(Lorentz)
>>> i0,i1,i2,i3,i4 = tensor_indices(i0:5, Lorentz)
>>> A = tensorhead(A, [Lorentz], [[1]])
>>> G = tensorhead(G, [Lorentz], [[1]], Gcomm)
>>> GH = tensorhead(GH, [Lorentz], [[1]], GHcomm)
>>> TensorManager.set_comm(Gcomm, GHcomm, 0)
>>> (GH(i1)*G(i0)).canon_bp()
G(i0)*GH(i1)
>>> (G(i1)*G(i0)).canon_bp()
G(i1)*G(i0)
>>> (G(i1)*A(i0)).canon_bp()
A(i0)*G(i1)
set comms(*args)
set the commutation group numbers c for symbols i, j
Parameters args : sequence of (i, j, c)
class sympy.tensor.tensor.TensorIndexType
A TensorIndexType is characterized by its name and its metric.
Parameters name : name of the tensor type
metric : metric symmetry or metric object or None
dim : dimension, it can be a symbol or an integer or None
eps dim : dimension of the epsilon tensor
1518
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
dummy fmt : name of the head of dummy indices
Notes
The metric parameter can be: metric = False symmetric metric (in Riemannian geometry)
metric = True antisymmetric metric (for spinor calculus)
metric = None there is no metric
metric can be an object having name and antisym attributes.
If there is a metric the metric is used to raise and lower indices.
In the case of antisymmetric metric, the following raising and lowering conventions will
be adopted:
psi(a) = g(a, b)*psi(-b); chi(-a) = chi(b)*g(-b, -a)
g(-a, b) = delta(-a, b); g(b, -a) = -delta(a, -b)
where delta(-a, b) = delta(b, -a) is the Kronecker delta (see TensorIndex for
the conventions on indices).
If there is no metric it is not possible to raise or lower indices; e.g. the index of the
dening representation of SU(N) is covariant and the conjugate representation is contravariant; for N > 2 they are linearly independent.
eps dim is by default equal to dim, if the latter is an integer; else it can be assigned (for
use in naive dimensional regularization); if eps dim is not an integer epsilon is None.
Examples
>>> from sympy.tensor.tensor import TensorIndexType
>>> Lorentz = TensorIndexType(Lorentz, dummy_fmt=L)
>>> Lorentz.metric
metric(Lorentz,Lorentz)
Examples with metric components data added, this means it is working on a xed basis:
>>> Lorentz.data = [1, -1, -1, -1]
>>> Lorentz
TensorIndexType(Lorentz, 0)
>>> Lorentz.data
[[1 0 0 0]
[0 -1 0 0]
[0 0 -1 0]
[0 0 0 -1]]
5.32. Tensor Module
1519
SymPy Documentation, Release 0.7.6
Attributes
name
metric name
metric antisym
metric
delta
epsilon
dim
dim eps
dummy fmt
data
it is metric or metric.name
the metric tensor
Kronecker delta
the Levi-Civita epsilon tensor
a property to add ndarray values, to work in a specied basis.
class sympy.tensor.tensor.TensorIndex
Represents an abstract tensor index.
Parameters name : name of the index, or True if you want it to be automatically assigned
tensortype : TensorIndexType of the index
is up : ag for contravariant index
Notes
Tensor indices are contracted with the Einstein summation convention.
An index can be in contravariant or in covariant form; in the latter case it is represented
prepending a - to the index name.
Dummy indices have a name with head given by tensortype. dummy fmt
Examples
>>> from sympy.tensor.tensor import TensorIndexType, TensorIndex, TensorSymmetry, TensorType, ge
>>> Lorentz = TensorIndexType(Lorentz, dummy_fmt=L)
>>> i = TensorIndex(i, Lorentz); i
i
>>> sym1 = TensorSymmetry(*get_symmetric_group_sgs(1))
>>> S1 = TensorType([Lorentz], sym1)
>>> A, B = S1(A,B)
>>> A(i)*B(-i)
A(L_0)*B(-L_0)
If you want the index name to be automatically assigned, just put True in the name eld,
it will be generated using the reserved character in front of its name, in order to avoid
conicts with possible existing indices:
>>> i0 = TensorIndex(True, Lorentz)
>>> i0
_i0
>>> i1 = TensorIndex(True, Lorentz)
>>> i1
_i1
>>> A(i0)*B(-i1)
A(_i0)*B(-_i1)
1520
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> A(i0)*B(-i0)
A(L_0)*B(-L_0)
Attributes
name
tensortype
is up
sympy.tensor.tensor.tensor indices(s, typ)
Returns list of tensor indices given their names and their types
Parameters s : string of comma separated names of indices
typ : list of TensorIndexType of the indices
Examples
>>> from sympy.tensor.tensor import TensorIndexType, tensor_indices
>>> Lorentz = TensorIndexType(Lorentz, dummy_fmt=L)
>>> a, b, c, d = tensor_indices(a,b,c,d, Lorentz)
class sympy.tensor.tensor.TensorSymmetry
Monoterm symmetry of a tensor
Parameters bsgs : tuple (base, sgs) BSGS of the symmetry of the tensor
See Also:
sympy.combinatorics.tensor can.get symmetric group sgs (page 275)
Notes
A tensor can have an arbitrary monoterm symmetry provided by its BSGS. Multiterm
symmetries, like the cyclic symmetry of the Riemann tensor, are not covered.
Examples
Dene a symmetric tensor
>>>
>>>
>>>
>>>
>>>
from sympy.tensor.tensor import TensorIndexType, tensor_indices, TensorSymmetry, TensorType,
Lorentz = TensorIndexType(Lorentz, dummy_fmt=L)
sym2 = TensorSymmetry(get_symmetric_group_sgs(2))
S2 = TensorType([Lorentz]*2, sym2)
V = S2(V)
Attributes
base
generators
rank
base of the BSGS
generators of the BSGS
rank of the tensor
5.32. Tensor Module
1521
SymPy Documentation, Release 0.7.6
sympy.tensor.tensor.tensorsymmetry(*args)
Return a TensorSymmetry object.
One can represent a tensor with any monoterm slot symmetry group using a BSGS.
args can be a BSGS args[0] base args[1] sgs
Usually tensors are in (direct products of) representations of the symmetric group; args
can be a list of lists representing the shapes of Young tableaux
Notes
For instance: [[1]] vector [[1]*n] symmetric tensor of rank n [[n]] antisymmetric
tensor of rank n [[2, 2]] monoterm slot symmetry of the Riemann tensor [[1],[1]]
vector*vector [[2],[1],[1] (antisymmetric tensor)*vector*vector
Notice that with the shape [2, 2] we associate only the monoterm symmetries of the
Riemann tensor; this is an abuse of notation, since the shape [2, 2] corresponds usually
to the irreducible representation characterized by the monoterm symmetries and by the
cyclic symmetry.
Examples
Symmetric tensor using a Young tableau
>>>
>>>
>>>
>>>
>>>
from sympy.tensor.tensor import TensorIndexType, TensorType, tensorsymmetry
Lorentz = TensorIndexType(Lorentz, dummy_fmt=L)
sym2 = tensorsymmetry([1, 1])
S2 = TensorType([Lorentz]*2, sym2)
V = S2(V)
Symmetric tensor using a BSGS (base, strong generator set)
>>>
>>>
>>>
>>>
from sympy.tensor.tensor import TensorSymmetry, get_symmetric_group_sgs
sym2 = tensorsymmetry(*get_symmetric_group_sgs(2))
S2 = TensorType([Lorentz]*2, sym2)
V = S2(V)
class sympy.tensor.tensor.TensorType
Class of tensor types.
Parameters index types : list of TensorIndexType of the tensor indices
symmetry : TensorSymmetry of the tensor
Examples
Dene a symmetric tensor
>>>
>>>
>>>
>>>
>>>
1522
from sympy.tensor.tensor import TensorIndexType, tensorsymmetry, TensorType
Lorentz = TensorIndexType(Lorentz, dummy_fmt=L)
sym2 = tensorsymmetry([1, 1])
S2 = TensorType([Lorentz]*2, sym2)
V = S2(V)
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Attributes
index types
symmetry
types
list of TensorIndexType without repetitions
class sympy.tensor.tensor.TensorHead
Tensor head of the tensor
Parameters name : name of the tensor
typ : list of TensorIndexType
comm : commutation group number
Notes
A TensorHead belongs to a commutation group, dened by a symbol on number comm (see
TensorManager.set comm); tensors in a commutation group have the same commutation
properties; by default comm is 0, the group of the commuting tensors.
Examples
>>>
>>>
>>>
>>>
>>>
from sympy.tensor.tensor import TensorIndexType, tensorsymmetry, TensorType
Lorentz = TensorIndexType(Lorentz, dummy_fmt=L)
sym2 = tensorsymmetry([1], [1])
S2 = TensorType([Lorentz]*2, sym2)
A = S2(A)
Examples with ndarray values, the components data assigned to the TensorHead object
are assumed to be in a fully-contravariant representation. In case it is necessary to
assign components data which represents the values of a non-fully covariant tensor, see
the other examples.
>>>
>>>
>>>
>>>
from sympy.tensor.tensor import tensor_indices, tensorhead
Lorentz.data = [1, -1, -1, -1]
i0, i1 = tensor_indices(i0:2, Lorentz)
A.data = [[j+2*i for j in range(4)] for i in range(4)]
in order to retrieve data, it is also necessary to specify abstract indices enclosed by round
brackets, then numerical indices inside square brackets.
>>> A(i0, i1)[0, 0]
0
>>> A(i0, i1)[2, 3] == 3+2*2
True
Notice that square brackets create a valued tensor expression instance:
>>> A(i0, i1)
A(i0, i1)
To view the data, just type:
>>> A.data
[[0 1 2 3]
[2 3 4 5]
5.32. Tensor Module
1523
SymPy Documentation, Release 0.7.6
[4 5 6 7]
[6 7 8 9]]
Turning to a tensor expression, covariant indices get the corresponding components data
corrected by the metric:
>>>
[[0
[2
[4
[6
A(i0,
-1 -2
-3 -4
-5 -6
-7 -8
-i1).data
-3]
-5]
-7]
-9]]
>>> A(-i0, -i1).data
[[0 -1 -2 -3]
[-2 3 4 5]
[-4 5 6 7]
[-6 7 8 9]]
while if all indices are contravariant, the ndarray remains the same
>>> A(i0, i1).data
[[0 1 2 3]
[2 3 4 5]
[4 5 6 7]
[6 7 8 9]]
When all indices are contracted and components data are added to the tensor, accessing
the data will return a scalar, no numpy object. In fact, numpy ndarrays are dropped to
scalars if they contain only one element.
>>> A(i0, -i0)
A(L_0, -L_0)
>>> A(i0, -i0).data
-18
It is also possible to assign components data to an indexed tensor, i.e. a tensor with
specied covariant and contravariant components. In this example, the covariant components data of the Electromagnetic tensor are injected into A:
>>>
>>>
>>>
>>>
...
...
...
...
from sympy import symbols
Ex, Ey, Ez, Bx, By, Bz = symbols(E_x E_y E_z B_x B_y B_z)
c = symbols(c, positive=True)
A(-i0, -i1).data = [
[0, Ex/c, Ey/c, Ez/c],
[-Ex/c, 0, -Bz, By],
[-Ey/c, Bz, 0, -Bx],
[-Ez/c, -By, Bx, 0]]
Now it is possible to retrieve the contravariant form of the Electromagnetic tensor:
>>> A(i0, i1).data
[[0 -E_x/c -E_y/c -E_z/c]
[E_x/c 0 -B_z B_y]
[E_y/c B_z 0 -B_x]
[E_z/c -B_y B_x 0]]
and the mixed contravariant-covariant form:
1524
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> A(i0, -i1).data
[[0 E_x/c E_y/c E_z/c]
[E_x/c 0 B_z -B_y]
[E_y/c -B_z 0 B_x]
[E_z/c B_y -B_x 0]]
To convert the numpys ndarray to a sympy matrix, just cast:
>>> from sympy import Matrix
>>> Matrix(A.data)
Matrix([
[
0, -E_x/c, -E_y/c, -E_z/c],
[E_x/c,
0,
-B_z,
B_y],
[E_y/c,
B_z,
0,
-B_x],
[E_z/c,
-B_y,
B_x,
0]])
Still notice, in this last example, that accessing components data from a tensor without
specifying the indices is equivalent to assume that all indices are contravariant.
It is also possible to store symbolic components data inside a tensor, for example, dene
a four-momentum-like tensor:
>>>
>>>
>>>
>>>
from sympy import symbols
P = tensorhead(P, [Lorentz], [[1]])
E, px, py, pz = symbols(E p_x p_y p_z, positive=True)
P.data = [E, px, py, pz]
The contravariant and covariant components are, respectively:
>>> P(i0).data
[E p_x p_y p_z]
>>> P(-i0).data
[E -p_x -p_y -p_z]
The contraction of a 1-index tensor by itself is usually indicated by a power by two:
>>> P(i0)**2
E**2 - p_x**2 - p_y**2 - p_z**2
As the power by two is clearly identical to P P , it is possible to simply contract the
TensorHead object, without specifying the indices
>>> P**2
E**2 - p_x**2 - p_y**2 - p_z**2
Attributes
name
index types
rank
types
symmetry
comm
equal to typ.types
equal to typ.symmetry
commutation group
commutes with(other)
Returns 0 if self and other commute, 1 if they anticommute.
Returns None if self and other neither commute nor anticommute.
5.32. Tensor Module
1525
SymPy Documentation, Release 0.7.6
class sympy.tensor.tensor.TensExpr
Abstract base class for tensor expressions
Notes
A tensor expression is an expression formed by tensors; currently the sums of tensors
are distributed.
A TensExpr can be a TensAdd or a TensMul.
TensAdd objects are put in canonical form using the Butler-Portugal algorithm for canonicalization under monoterm symmetries.
TensMul objects are formed by products of component tensors, and include a coecient,
which is a SymPy expression.
In the internal representation contracted indices are represented by (ipos1, ipos2,
icomp1, icomp2), where icomp1 is the position of the component tensor with contravariant index, ipos1 is the slot which the index occupies in that component tensor.
Contracted indices are therefore nameless in the internal representation.
get matrix()
Returns ndarray components data as a matrix, if components data are available and
ndarray dimension does not exceed 2.
Examples
>>>
>>>
>>>
>>>
>>>
>>>
from sympy.tensor.tensor import TensorIndexType, tensorsymmetry, TensorType
from sympy import ones
Lorentz = TensorIndexType(Lorentz, dummy_fmt=L)
sym2 = tensorsymmetry([1]*2)
S2 = TensorType([Lorentz]*2, sym2)
A = S2(A)
>>> from sympy.tensor.tensor import tensor_indices, tensorhead
>>> Lorentz.data = [1, -1, -1, -1]
>>> i0, i1 = tensor_indices(i0:2, Lorentz)
>>> A.data = [[j+2*i for j in range(4)] for i in range(4)]
>>> A(i0, i1).get_matrix()
Matrix([
[0, 1, 2, 3],
[2, 3, 4, 5],
[4, 5, 6, 7],
[6, 7, 8, 9]])
It is possible to perform usual operation on matrices, such as the matrix multiplication:
>>> A(i0, i1).get_matrix()*ones(4, 1)
Matrix([
[ 6],
[14],
[22],
[30]])
class sympy.tensor.tensor.TensAdd
Sum of tensors
1526
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Parameters free args : list of the free indices
Notes
Sum of more than one tensor are put automatically in canonical form.
Examples
>>> from sympy.tensor.tensor import TensorIndexType, tensorhead, tensor_indices
>>> Lorentz = TensorIndexType(Lorentz, dummy_fmt=L)
>>> a, b = tensor_indices(a,b, Lorentz)
>>> p, q = tensorhead(p,q, [Lorentz], [[1]])
>>> t = p(a) + q(a); t
p(a) + q(a)
>>> t(b)
p(b) + q(b)
Examples with components data added to the tensor expression:
>>> from sympy import eye
>>> Lorentz.data = [1, -1, -1, -1]
>>> a, b = tensor_indices(a, b, Lorentz)
>>> p.data = [2, 3, -2, 7]
>>> q.data = [2, 3, -2, 7]
>>> t = p(a) + q(a); t
p(a) + q(a)
>>> t(b)
p(b) + q(b)
The following are: 2**2 - 3**2 - 2**2 - 7**2 ==> -58
>>> (p(a)*p(-a)).data
-58
>>> p(a)**2
-58
Attributes
args
rank
free args
tuple of addends
rank of the tensor
list of the free indices in sorted order
canon bp()
canonicalize using the Butler-Portugal algorithm for canonicalization under
monoterm symmetries.
contract metric(g)
Raise or lower indices with the metric g
Parameters g : metric
contract all : if True, eliminate all g which are contracted
5.32. Tensor Module
1527
SymPy Documentation, Release 0.7.6
Notes
see the TensorIndexType docstring for the contraction conventions
static from TIDS list(coe, tids list)
Given a list of coecients and a list of TIDS objects, construct a TensAdd instance,
equivalent to the one that would result from creating single instances of TensMul
and then adding them.
Examples
>>> from sympy.tensor.tensor import TensorIndexType, tensor_indices, tensorhead, TensAdd
>>> Lorentz = TensorIndexType(Lorentz, dummy_fmt=L)
>>> i, j = tensor_indices(i,j, Lorentz)
>>> A, B = tensorhead(A,B, [Lorentz]*2, [[1]*2])
>>> eA = 3*A(i, j)
>>> eB = 2*B(j, i)
>>> t1 = eA._tids
>>> t2 = eB._tids
>>> c1 = eA.coeff
>>> c2 = eB.coeff
>>> TensAdd.from_TIDS_list([c1, c2], [t1, t2])
2*B(i, j) + 3*A(i, j)
If the coecient parameter is a scalar, then it will be applied as a coecient on all
TIDS objects.
>>> TensAdd.from_TIDS_list(4, [t1, t2])
4*A(i, j) + 4*B(i, j)
fun eval(*index tuples)
Return a tensor with free indices substituted according to index tuples
Parameters index types : list of tuples (old index, new index)
Examples
>>> from sympy.tensor.tensor import TensorIndexType, tensor_indices, tensorhead
>>> Lorentz = TensorIndexType(Lorentz, dummy_fmt=L)
>>> i, j, k, l = tensor_indices(i,j,k,l, Lorentz)
>>> A, B = tensorhead(A,B, [Lorentz]*2, [[1]*2])
>>> t = A(i, k)*B(-k, -j) + A(i, -j)
>>> t.fun_eval((i, k),(-j, l))
A(k, L_0)*B(l, -L_0) + A(k, l)
substitute indices(*index tuples)
Return a tensor with free indices substituted according to index tuples
Parameters index types : list of tuples (old index, new index)
Examples
1528
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> from sympy.tensor.tensor import TensorIndexType, tensor_indices, tensorhead
>>> Lorentz = TensorIndexType(Lorentz, dummy_fmt=L)
>>> i, j, k, l = tensor_indices(i,j,k,l, Lorentz)
>>> A, B = tensorhead(A,B, [Lorentz]*2, [[1]*2])
>>> t = A(i, k)*B(-k, -j); t
A(i, L_0)*B(-L_0, -j)
>>> t.substitute_indices((i,j), (j, k))
A(j, L_0)*B(-L_0, -k)
class sympy.tensor.tensor.TensMul
Product of tensors
Parameters coe : SymPy coecient of the tensor
args :
Notes
args[0] list of TensorHead of the component tensors.
args[1] list of (ind, ipos, icomp) where ind is a free index, ipos is the slot position
of ind in the icomp-th component tensor.
args[2] list of tuples representing dummy indices. (ipos1, ipos2, icomp1, icomp2)
indicates that the contravariant dummy index is the ipos1-th slot position in the icomp1th component tensor; the corresponding covariant index is in the ipos2 slot position in
the icomp2-th component tensor.
Attributes
components
types
free
dum
ext rank
rank
coeff
free args
is canon bp
list of TensorHead of the component tensors
list of nonrepeated TensorIndexType
list of (ind, ipos, icomp), see Notes
list of (ipos1, ipos2, icomp1, icomp2), see Notes
rank of the tensor counting the dummy indices
rank of the tensor
SymPy coecient of the tensor
list of the free indices in sorted order
True if the tensor in in canonical form
canon bp()
Canonicalize using the Butler-Portugal algorithm for canonicalization under
monoterm symmetries.
Examples
>>> from sympy.tensor.tensor import TensorIndexType, tensor_indices, tensorhead
>>> Lorentz = TensorIndexType(Lorentz, dummy_fmt=L)
>>> m0, m1, m2 = tensor_indices(m0,m1,m2, Lorentz)
>>> A = tensorhead(A, [Lorentz]*2, [[2]])
>>> t = A(m0,-m1)*A(m1,-m0)
>>> t.canon_bp()
-A(L_0, L_1)*A(-L_0, -L_1)
>>> t = A(m0,-m1)*A(m1,-m2)*A(m2,-m0)
5.32. Tensor Module
1529
SymPy Documentation, Release 0.7.6
>>> t.canon_bp()
0
contract metric(g)
Raise or lower indices with the metric g
Parameters g : metric
Notes
see the TensorIndexType docstring for the contraction conventions
Examples
>>> from sympy.tensor.tensor import TensorIndexType, tensor_indices, tensorhead
>>> Lorentz = TensorIndexType(Lorentz, dummy_fmt=L)
>>> m0, m1, m2 = tensor_indices(m0,m1,m2, Lorentz)
>>> g = Lorentz.metric
>>> p, q = tensorhead(p,q, [Lorentz], [[1]])
>>> t = p(m0)*q(m1)*g(-m0, -m1)
>>> t.canon_bp()
metric(L_0, L_1)*p(-L_0)*q(-L_1)
>>> t.contract_metric(g).canon_bp()
p(L_0)*q(-L_0)
fun eval(*index tuples)
Return a tensor with free indices substituted according to index tuples
index types list of tuples (old index, new index)
Examples
>>> from sympy.tensor.tensor import TensorIndexType, tensor_indices, tensorhead
>>> Lorentz = TensorIndexType(Lorentz, dummy_fmt=L)
>>> i, j, k, l = tensor_indices(i,j,k,l, Lorentz)
>>> A, B = tensorhead(A,B, [Lorentz]*2, [[1]*2])
>>> t = A(i, k)*B(-k, -j); t
A(i, L_0)*B(-L_0, -j)
>>> t.fun_eval((i, k),(-j, l))
A(k, L_0)*B(-L_0, l)
get indices()
Returns the list of indices of the tensor
The indices are listed in the order in which they appear in the component tensors.
The dummy indices are given a name which does not collide with the names of the
free indices.
Examples
1530
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> from sympy.tensor.tensor import TensorIndexType, tensor_indices, tensorhead
>>> Lorentz = TensorIndexType(Lorentz, dummy_fmt=L)
>>> m0, m1, m2 = tensor_indices(m0,m1,m2, Lorentz)
>>> g = Lorentz.metric
>>> p, q = tensorhead(p,q, [Lorentz], [[1]])
>>> t = p(m1)*g(m0,m2)
>>> t.get_indices()
[m1, m0, m2]
perm2tensor(g, canon bp=False)
Returns the tensor corresponding to the permutation g
For further details, see the method in TIDS with the same name.
sorted components()
Returns a tensor with sorted components calling the corresponding method in a
TIDS object.
split()
Returns a list of tensors, whose product is self
Dummy indices contracted among dierent tensor components become free indices
with the same name as the one used to represent the dummy indices.
Examples
>>> from sympy.tensor.tensor import TensorIndexType, tensor_indices, tensorhead
>>> Lorentz = TensorIndexType(Lorentz, dummy_fmt=L)
>>> a, b, c, d = tensor_indices(a,b,c,d, Lorentz)
>>> A, B = tensorhead(A,B, [Lorentz]*2, [[1]*2])
>>> t = A(a,b)*B(-b,c)
>>> t
A(a, L_0)*B(-L_0, c)
>>> t.split()
[A(a, L_0), B(-L_0, c)]
substitute indices(*index tuples)
Return a tensor with free indices substituted according to index tuples
index types list of tuples (old index, new index)
Note: this method will neither raise or lower the indices, it will just replace their
symbol.
Examples
>>> from sympy.tensor.tensor import TensorIndexType, tensor_indices, tensorhead
>>> Lorentz = TensorIndexType(Lorentz, dummy_fmt=L)
>>> i, j, k, l = tensor_indices(i,j,k,l, Lorentz)
>>> A, B = tensorhead(A,B, [Lorentz]*2, [[1]*2])
>>> t = A(i, k)*B(-k, -j); t
A(i, L_0)*B(-L_0, -j)
>>> t.substitute_indices((i,j), (j, k))
A(j, L_0)*B(-L_0, -k)
sympy.tensor.tensor.canon bp(p)
Butler-Portugal canonicalization
5.32. Tensor Module
1531
SymPy Documentation, Release 0.7.6
sympy.tensor.tensor.tensor mul(*a)
product of tensors
sympy.tensor.tensor.riemann cyclic replace(t r)
replace Riemann tensor with an equivalent expression
R(m,n,p,q) -> 2/3*R(m,n,p,q) - 1/3*R(m,q,n,p) + 1/3*R(m,p,n,q)
sympy.tensor.tensor.riemann cyclic(t2)
replace each Riemann tensor with an equivalent expression satisfying the cyclic identity.
This trick is discussed in the reference guide to Cadabra.
Examples
>>>
>>>
>>>
>>>
>>>
>>>
0
from sympy.tensor.tensor import TensorIndexType, tensor_indices, tensorhead, riemann_cyclic
Lorentz = TensorIndexType(Lorentz, dummy_fmt=L)
i, j, k, l = tensor_indices(i,j,k,l, Lorentz)
R = tensorhead(R, [Lorentz]*4, [[2, 2]])
t = R(i,j,k,l)*(R(-i,-j,-k,-l) - 2*R(-i,-k,-j,-l))
riemann_cyclic(t)
5.33 Utilities
This module contains some general purpose utilities that are used across SymPy.
Contents:
5.33.1 Autowrap Module
The autowrap module works very well in tandem with the Indexed classes of the Tensor Module (page 1508). Here is a simple example that shows how to setup a binary routine that
calculates a matrix-vector product.
>>> from sympy.utilities.autowrap import autowrap
>>> from sympy import symbols, IndexedBase, Idx, Eq
>>> A, x, y = map(IndexedBase, [A, x, y])
>>> m, n = symbols(m n, integer=True)
>>> i = Idx(i, m)
>>> j = Idx(j, n)
>>> instruction = Eq(y[i], A[i, j]*x[j]); instruction
y[i] == x[j]*A[i, j]
Because the code printers treat Indexed objects with repeated indices as a summation, the
above equality instance will be translated to low-level code for a matrix vector product. This
is how you tell SymPy to generate the code, compile it and wrap it as a python function:
>>> matvec = autowrap(instruction)
Thats it. Now lets test it with some numpy arrays. The default wrapper backend is f2py. The
wrapper function it provides is set up to accept python lists, which it will silently convert to
numpy arrays. So we can test the matrix vector product like this:
1532
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> M = [[0, 1],
...
[1, 0]]
>>> matvec(M, [2, 3])
[ 3. 2.]
Implementation details
The autowrap module is implemented with a backend consisting of CodeWrapper objects.
The base class CodeWrapper takes care of details about module name, lenames and options.
It also contains the driver routine, which runs through all steps in the correct order, and also
takes care of setting up and removing the temporary working directory.
The actual compilation and wrapping is done by external resources, such as the system installed f2py command. The Cython backend runs a distutils setup script in a subprocess.
Subclasses of CodeWrapper takes care of these backend-dependent details.
API Reference
Module for compiling codegen output, and wrap the binary for use in python.
Note: To use the autowrap module it must rst be imported
>>> from sympy.utilities.autowrap import autowrap
This module provides a common interface for dierent external backends, such as f2py, fwrap,
Cython, SWIG(?) etc. (Currently only f2py and Cython are implemented) The goal is to provide
access to compiled binaries of acceptable performance with a one-button user interface, i.e.
>>> from sympy.abc import x,y
>>> expr = ((x - y)**(25)).expand()
>>> binary_callable = autowrap(expr)
>>> binary_callable(1, 2)
-1.0
The callable returned from autowrap() is a binary python function, not a SymPy object. If
it is desired to use the compiled function in symbolic expressions, it is better to use binary function() which returns a SymPy Function object. The binary callable is attached as
the imp attribute and invoked when a numerical evaluation is requested with evalf(), or
with lambdify().
>>> from sympy.utilities.autowrap import binary_function
>>> f = binary_function(f, expr)
>>> 2*f(x, y) + y
y + 2*f(x, y)
>>> (2*f(x, y) + y).evalf(2, subs={x: 1, y:2})
0.e-110
The idea is that a SymPy user will primarily be interested in working with mathematical
expressions, and should not have to learn details about wrapping tools in order to evaluate
expressions numerically, even if they are computationally expensive.
When is this useful?
5.33. Utilities
1533
SymPy Documentation, Release 0.7.6
1. For computations on large arrays, Python iterations may be too slow, and depending on
the mathematical expression, it may be dicult to exploit the advanced index operations
provided by NumPy.
2. For really long expressions that will be called repeatedly, the compiled binary should be
signicantly faster than SymPys .evalf()
3. If you are generating code with the codegen utility in order to use it in another project,
the automatic python wrappers let you test the binaries immediately from within SymPy.
4. To create customized ufuncs for use with numpy arrays. See ufuncify.
When is this module NOT the best approach?
1. If you are really concerned about speed or memory optimizations, you will probably
get better results by working directly with the wrapper tools and the low level code.
However, the les generated by this utility may provide a useful starting point and
reference code. Temporary les will be left intact if you supply the keyword tempdir=path/to/les/.
2. If the array computation can be handled easily by numpy, and you dont need the binaries
for another project.
class sympy.utilities.autowrap.CodeWrapper(generator, lepath=None, ags=[],
verbose=False)
Base Class for code wrappers
class sympy.utilities.autowrap.CythonCodeWrapper(generator,
lepath=None,
ags=[], verbose=False)
Wrapper that uses Cython
dump pyx(routines, f, prex)
Write a Cython le with python wrappers
This le contains all the denitions of the routines in c code and refers to the header
le.
Arguments
routines List of Routine instances
f File-like object to write the le to
prex The lename prex, used to refer to the proper header le. Only the basename of the prex is used.
class sympy.utilities.autowrap.DummyWrapper(generator, lepath=None, ags=[],
verbose=False)
Class used for testing independent of backends
class sympy.utilities.autowrap.F2PyCodeWrapper(generator,
lepath=None,
ags=[], verbose=False)
Wrapper that uses f2py
class sympy.utilities.autowrap.UfuncifyCodeWrapper(generator, lepath=None,
ags=[], verbose=False)
Wrapper for Ufuncify
dump c(routines, f, prex)
Write a C le with python wrappers
This le contains all the denitions of the routines in c code.
1534
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Arguments
routines List of Routine instances
f File-like object to write the le to
prex The lename prex, used to name the imported module.
sympy.utilities.autowrap.autowrap(*args, **kwargs)
Generates python callable binaries based on the math expression.
Parameters expr :
The SymPy expression that should be wrapped as a binary routine.
language : string, optional
If supplied, (options: C or F95), species the language of the generated code. If None [default], the language is inferred based upon
the specied backend.
backend : string, optional
Backend used to wrap the generated code. Either f2py [default], or
cython.
tempdir : string, optional
Path to directory for temporary les. If this argument is supplied,
the generated code and the wrapper input les are left intact in the
specied path.
args : iterable, optional
An iterable of symbols. Species the argument sequence for the function.
ags : iterable, optional
Additional option ags that will be passed to the backend.
verbose : bool, optional
If True, autowrap will not mute the command line backends. This can
be helpful for debugging.
helpers : iterable, optional
Used to dene auxillary expressions needed for the main expr. If the
main expression needs to call a specialized function it should be put
in the helpers iterable. Autowrap will then make sure that the compiled main expression can link to the helper routine. Items should be
tuples with (<funtion name>, <sympy expression>, <arguments>).
It is mandatory to supply an argument sequence to helper routines.
>>> from sympy.abc import x, y, z :
>>> from sympy.utilities.autowrap import autowrap :
>>> expr = ((x - y + z)**(13)).expand() :
>>> binary func = autowrap(expr) :
>>> binary func(1, 4, 2) :
-1.0 :
5.33. Utilities
1535
SymPy Documentation, Release 0.7.6
sympy.utilities.autowrap.binary function(symfunc, expr, **kwargs)
Returns a sympy function with expr as binary implementation
This is a convenience function that automates the steps needed to autowrap the SymPy
expression and attaching it to a Function object with implemented function().
>>> from sympy.abc import x, y
>>> from sympy.utilities.autowrap import binary_function
>>> expr = ((x - y)**(25)).expand()
>>> f = binary_function(f, expr)
>>> type(f)
<class sympy.core.function.UndefinedFunction>
>>> 2*f(x, y)
2*f(x, y)
>>> f(x, y).evalf(2, subs={x: 1, y: 2})
-1.0
sympy.utilities.autowrap.ufuncify(*args, **kwargs)
Generates a binary function that supports broadcasting on numpy arrays.
Parameters args : iterable
Either a Symbol or an iterable of symbols. Species the argument
sequence for the function.
expr :
A SymPy expression that denes the element wise operation.
language : string, optional
If supplied, (options: C or F95), species the language of the generated code. If None [default], the language is inferred based upon
the specied backend.
backend : string, optional
Backend used to wrap the generated code. Either numpy [default],
cython, or f2py.
tempdir : string, optional
Path to directory for temporary les. If this argument is supplied,
the generated code and the wrapper input les are left intact in the
specied path.
ags : iterable, optional
Additional option ags that will be passed to the backend
verbose : bool, optional
If True, autowrap will not mute the command line backends. This can
be helpful for debugging.
helpers : iterable, optional
Used to dene auxillary expressions needed for the main expr. If the
main expression needs to call a specialized function it should be put
in the helpers iterable. Autowrap will then make sure that the compiled main expression can link to the helper routine. Items should be
tuples with (<funtion name>, <sympy expression>, <arguments>).
It is mandatory to supply an argument sequence to helper routines.
1536
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
References
[1] http://docs.scipy.org/doc/numpy/reference/ufuncs.html
Examples
>>> from sympy.utilities.autowrap import ufuncify
>>> from sympy.abc import x, y
>>> import numpy as np
>>> f = ufuncify((x, y), y + x**2)
>>> type(f)
numpy.ufunc
>>> f([1, 2, 3], 2)
array([ 3., 6., 11.])
>>> f(np.arange(5), 3)
array([ 3., 4., 7., 12., 19.])
For the F2Py and Cython backends, inputs are required to be equal length 1-dimensional
arrays. The F2Py backend will perform type conversion, but the Cython backend will
error if the inputs are not of the expected type.
>>> f fortran = ufuncify((x, y), y + x**2, backend=F2Py)
>>> f fortran(1, 2)
3
>>> f fortran(numpy.array([1, 2, 3]), numpy.array([1.0, 2.0, 3.0]))
array([2., 6., 12.])
>>> f cython = ufuncify((x, y), y + x**2, backend=Cython)
>>> f cython(1, 2)
Traceback (most recent call last):
File <stdin>, line 1, in <module>
TypeError: Argument x has incorrect type (expected numpy.ndarray, got int)
>>> f cython(numpy.array([1.0]), numpy.array([2.0]))
array([ 3.])
Note
The default backend (numpy) will create actual instances of numpy.ufunc. These support ndimensional broadcasting, and implicit type conversion. Use of the other backends
will result in a ufunc-like function, which requires equal length 1-dimensional arrays
for all arguments, and will not perform any type conversions.
5.33.2 Codegen
This module provides functionality to generate directly compilable code from SymPy expressions. The codegen function is the user interface to the code generation functionality in
SymPy. Some details of the implementation is given below for advanced users that may want
to use the framework directly.
Note: The codegen callable is not in the sympy namespace automatically, to use it you must
rst execute
5.33. Utilities
1537
SymPy Documentation, Release 0.7.6
>>> from sympy.utilities.codegen import codegen
Implementation Details
Here we present the most important pieces of the internal structure, as advanced users may
want to use it directly, for instance by subclassing a code generator for a specialized application. It is very likely that you would prefer to use the codegen() function documented
above.
Basic assumptions:
A generic Routine data structure describes the routine that must be translated into
C/Fortran/... code. This data structure covers all features present in one or more of
the supported languages.
Descendants from the CodeGen class transform multiple Routine instances into compilable code. Each derived class translates into a specic language.
In many cases, one wants a simple workow. The friendly functions in the last part are
a simple api on top of the Routine/CodeGen stu. They are easier to use, but are less
powerful.
Routine
The Routine class is a very important piece of the codegen module. Viewing the codegen
utility as a translator of mathematical expressions into a set of statements in a programming
language, the Routine instances are responsible for extracting and storing information about
how the math can be encapsulated in a function call. Thus, it is the Routine constructor that
decides what arguments the routine will need and if there should be a return value.
API Reference
module for generating C, C++, Fortran77, Fortran90 and Octave/Matlab routines that evaluate sympy expressions. This module is work in progress. Only the milestones with a +
character in the list below have been completed.
How is sympy.utilities.codegen dierent from sympy.printing.ccode?
We considered the idea to extend the printing routines for sympy functions in such a way that
it prints complete compilable code, but this leads to a few unsurmountable issues that can
only be tackled with dedicated code generator:
For C, one needs both a code and a header le, while the printing routines generate just
one string. This code generator can be extended to support .pyf les for f2py.
SymPy functions are not concerned with programming-technical issues, such as input,
output and input-output arguments. Other examples are contiguous or non-contiguous
arrays, including headers of other libraries such as gsl or others.
It is highly interesting to evaluate several sympy functions in one C routine, eventually
sharing common intermediate results with the help of the cse routine. This is more than
just printing.
From the programming perspective, expressions with constants should be evaluated in
the code generator as much as possible. This is dierent for printing.
1538
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Basic assumptions
A generic Routine data structure describes the routine that must be translated into
C/Fortran/... code. This data structure covers all features present in one or more of
the supported languages.
Descendants from the CodeGen class transform multiple Routine instances into compilable code. Each derived class translates into a specic language.
In many cases, one wants a simple workow. The friendly functions in the last part are
a simple api on top of the Routine/CodeGen stu. They are easier to use, but are less
powerful.
Milestones
First working version with scalar input arguments, generating C code, tests
Friendly functions that are easier to use than the rigorous Routine/CodeGen workow.
Integer and Real numbers as input and output
Output arguments
InputOutput arguments
Sort input/output arguments properly
Contiguous array arguments (numpy matrices)
Also generate .pyf code for f2py (in autowrap module)
Isolate constants and evaluate them beforehand in double precision
Fortran 90
Octave/Matlab
Common Subexpression Elimination
User dened comments in the generated code
Optional extra include lines for libraries/objects that can eval special functions
Test other C compilers and libraries: gcc, tcc, libtcc, gcc+gsl, ...
Contiguous array arguments (sympy matrices)
Non-contiguous array arguments (sympy matrices)
ccode must raise an error when it encounters something that can not be translated into
c. ccode(integrate(sin(x)/x, x)) does not make sense.
Complex numbers as input and output
A default complex datatype
Include extra information in the header: date, user, hostname, sha1 hash, ...
Fortran 77
C++
Python
...
class sympy.utilities.codegen.Routine(name, arguments, results, local vars)
Generic description of evaluation routine for set of expressions.
A CodeGen class can translate instances of this class into code in a particular language.
The routine specication covers all the features present in these languages. The CodeGen part must raise an exception when certain features are not present in the target
5.33. Utilities
1539
SymPy Documentation, Release 0.7.6
language. For example, multiple return values are possible in Python, but not in C or
Fortran. Another example: Fortran and Python support complex numbers, while C does
not.
result variables
Returns a list of OutputArgument, InOutArgument and Result.
If return values are present, they are at the end ot the list.
variables
Returns a set of all variables possibly used in the routine.
For routines with unnamed return values, the dummies that may or may not be used
will be included in the set.
class sympy.utilities.codegen.DataType(cname, fname, pyname, octname)
Holds strings for a certain datatype in dierent languages.
sympy.utilities.codegen.get default datatype(expr)
Derives an appropriate datatype based on the expression.
class sympy.utilities.codegen.Argument(name, datatype=None, dimensions=None,
precision=None)
An abstract Argument data structure: a name and a data type.
This structure is rened in the descendants below.
class sympy.utilities.codegen.Result(expr,
name=None,
result var=None,
datatype=None, dimensions=None, precision=None)
An expression for a return value.
The name result is used to avoid conicts with the reserved word return in the python
language. It is also shorter than ReturnValue.
These may or may not need a name in the destination (e.g., return(x*y) might return
a value without ever naming it).
class sympy.utilities.codegen.CodeGen(project=project)
Abstract class for the code generators.
dump code(routines, f, prex, header=True, empty=True)
Write the code by calling language specic methods.
The generated le contains all the denitions of the routines in low-level code and
refers to the header le if appropriate.
Parameters routines : list
A list of Routine instances.
f : le-like
Where to write the le.
prex : string
The lename prex, used to refer to the proper header le. Only the
basename of the prex is used.
header : bool, optional
When True, a header comment is included on top of each source le.
[default : True]
empty : bool, optional
1540
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
When True, empty lines are included to structure the source les.
[default : True]
routine(name, expr, argument sequence)
Creates an Routine object that is appropriate for this language.
This implementation is appropriate for at least C/Fortran. Subclasses can override
this if necessary.
Here, we assume at most one return value (the l-value) which must be scalar. Additional outputs are OutputArguments (e.g., pointers on right-hand-side or passby-reference). Matrices are always returned via OutputArguments. If argument sequence is None, arguments will be ordered alphabetically, but with all InputArguments rst, and then OutputArgument and InOutArguments.
write(routines, prex, to les=False, header=True, empty=True)
Writes all the source code les for the given routines.
The generated source is returned as a list of (lename, contents) tuples, or is written
to les (see below). Each lename consists of the given prex, appended with an
appropriate extension.
Parameters routines : list
A list of Routine instances to be written
prex : string
The prex for the output les
to les : bool, optional
When True, the output is written to les. Otherwise, a list of (lename, contents) tuples is returned. [default: False]
header : bool, optional
When True, a header comment is included on top of each source le.
[default: True]
empty : bool, optional
When True, empty lines are included to structure the source les.
[default: True]
class sympy.utilities.codegen.CCodeGen(project=project)
Generator for C code.
The .write() method inherited from CodeGen will output a code le and an interface le,
<prex>.c and <prex>.h respectively.
dump c(routines, f, prex, header=True, empty=True)
Write the code by calling language specic methods.
The generated le contains all the denitions of the routines in low-level code and
refers to the header le if appropriate.
Parameters routines : list
A list of Routine instances.
f : le-like
Where to write the le.
prex : string
5.33. Utilities
1541
SymPy Documentation, Release 0.7.6
The lename prex, used to refer to the proper header le. Only the
basename of the prex is used.
header : bool, optional
When True, a header comment is included on top of each source le.
[default : True]
empty : bool, optional
When True, empty lines are included to structure the source les.
[default : True]
dump h(routines, f, prex, header=True, empty=True)
Writes the C header le.
This le contains all the function declarations.
Parameters routines : list
A list of Routine instances.
f : le-like
Where to write the le.
prex : string
The lename prex, used to construct the include guards. Only the
basename of the prex is used.
header : bool, optional
When True, a header comment is included on top of each source le.
[default : True]
empty : bool, optional
When True, empty lines are included to structure the source les.
[default : True]
get prototype(routine)
Returns a string for the function prototype of the routine.
If the routine has multiple result objects, an CodeGenError is raised.
See: http://en.wikipedia.org/wiki/Function prototype
class sympy.utilities.codegen.FCodeGen(project=project)
Generator for Fortran 95 code
The .write() method inherited from CodeGen will output a code le and an interface le,
<prex>.f90 and <prex>.h respectively.
dump f95(routines, f, prex, header=True, empty=True)
Write the code by calling language specic methods.
The generated le contains all the denitions of the routines in low-level code and
refers to the header le if appropriate.
Parameters routines : list
A list of Routine instances.
f : le-like
Where to write the le.
prex : string
1542
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
The lename prex, used to refer to the proper header le. Only the
basename of the prex is used.
header : bool, optional
When True, a header comment is included on top of each source le.
[default : True]
empty : bool, optional
When True, empty lines are included to structure the source les.
[default : True]
dump h(routines, f, prex, header=True, empty=True)
Writes the interface to a header le.
This le contains all the function declarations.
Parameters routines : list
A list of Routine instances.
f : le-like
Where to write the le.
prex : string
The lename prex.
header : bool, optional
When True, a header comment is included on top of each source le.
[default : True]
empty : bool, optional
When True, empty lines are included to structure the source les.
[default : True]
get interface(routine)
Returns a string for the function interface.
The routine should have a single result object, which can be None. If the routine
has multiple result objects, a CodeGenError is raised.
See: http://en.wikipedia.org/wiki/Function prototype
class sympy.utilities.codegen.OctaveCodeGen(project=project)
Generator for Octave code.
The .write() method inherited from CodeGen will output a code le <prex>.m.
Octave .m les usually contain one function. That function name should match the lename (prefix). If you pass multiple name expr pairs, the latter ones are presumed to be
private functions accessed by the primary function.
You should only pass inputs to argument sequence: outputs are ordered according to
their order in name expr.
dump m(routines, f, prex, header=True, empty=True, inline=True)
Write the code by calling language specic methods.
The generated le contains all the denitions of the routines in low-level code and
refers to the header le if appropriate.
Parameters routines : list
A list of Routine instances.
5.33. Utilities
1543
SymPy Documentation, Release 0.7.6
f : le-like
Where to write the le.
prex : string
The lename prex, used to refer to the proper header le. Only the
basename of the prex is used.
header : bool, optional
When True, a header comment is included on top of each source le.
[default : True]
empty : bool, optional
When True, empty lines are included to structure the source les.
[default : True]
routine(name, expr, argument sequence)
Specialized Routine creation for Octave.
sympy.utilities.codegen.codegen(name expr,
language,
prex=None,
project=project, to les=False, header=True,
empty=True, argument sequence=None)
Generate source code for expressions in a given language.
Parameters name expr : tuple, or list of tuples
A single (name, expression) tuple or a list of (name, expression) tuples. Each tuple corresponds to a routine. If the expression is an
equality (an instance of class Equality) the left hand side is considered
an output argument. If expression is an iterable, then the routine will
have multiple outputs.
language : string
A string that indicates the source code language. This is case insensitive. Currently, C, F95 and Octave are supported. Octave
generates code compatible with both Octave and Matlab.
prex : string, optional
A prex for the names of the les that contain the source code.
Language-dependent suxes will be appended. If omitted, the name
of the rst name expr tuple is used.
project : string, optional
A project name, used for making unique preprocessor instructions.
[default: project]
to les : bool, optional
When True, the code will be written to one or more les with the
given prex, otherwise strings with the names and contents of these
les are returned. [default: False]
header : bool, optional
When True, a header is written on top of each source le. [default:
True]
empty : bool, optional
When True, empty lines are used to structure the code. [default:
True]
1544
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
argument sequence : iterable, optional
Sequence of arguments for the routine in a preferred order. A CodeGenError is raised if required arguments are missing. Redundant
arguments are used without warning. If omitted, arguments will be
ordered alphabetically, but with all input aguments rst, and then
output or in-out arguments.
Examples
>>> from sympy.utilities.codegen import codegen
>>> from sympy.abc import x, y, z
>>> [(c_name, c_code), (h_name, c_header)] = codegen(
...
(f, x+y*z), C, test, header=False, empty=False)
>>> print(c_name)
test.c
>>> print(c_code)
#include test.h
#include <math.h>
double f(double x, double y, double z) {
double f_result;
f_result = x + y*z;
return f_result;
}
>>> print(h_name)
test.h
>>> print(c_header)
#ifndef PROJECT__TEST__H
#define PROJECT__TEST__H
double f(double x, double y, double z);
#endif
Another example using Equality objects to give named outputs. Here the lename (prex) is taken from the rst (name, expr) pair.
>>> from sympy.abc import f, g
>>> from sympy import Eq
>>> [(c_name, c_code), (h_name, c_header)] = codegen(
...
[(myfcn, x + y), (fcn2, [Eq(f, 2*x), Eq(g, y)])],
...
C, header=False, empty=False)
>>> print(c_name)
myfcn.c
>>> print(c_code)
#include myfcn.h
#include <math.h>
double myfcn(double x, double y) {
double myfcn_result;
myfcn_result = x + y;
return myfcn_result;
}
void fcn2(double x, double y, double *f, double *g) {
(*f) = 2*x;
(*g) = y;
}
sympy.utilities.codegen.make routine(name, expr, argument sequence=None, language=F95)
A factory that makes an appropriate Routine from an expression.
5.33. Utilities
1545
SymPy Documentation, Release 0.7.6
Parameters name : string
The name of this routine in the generated code.
expr : expression or list/tuple of expressions
A SymPy expression that the Routine instance will represent. If given
a list or tuple of expressions, the routine will be considered to have
multiple return values and/or output arguments.
argument sequence : list or tuple, optional
List arguments for the routine in a preferred order. If omitted, the
results are language dependent, for example, alphabetical order or
in the same order as the given expressions.
language : string, optional
Specify a target language. The Routine itself should be languageagnostic but the precise way one is created, error checking, etc depend on the language. [default: F95].
A decision about whether to use output arguments or return values
is made :
depending on both the language and the particular mathematical
expressions. :
For an expression of type Equality, the left hand side is typically made
:
into an OutputArgument (or perhaps an InOutArgument if appropriate). :
Otherwise, typically, the calculated expression is made a return values of :
the routine. :
Examples
>>> from sympy.utilities.codegen import make_routine
>>> from sympy.abc import x, y, f, g
>>> from sympy import Eq
>>> r = make_routine(test, [Eq(f, 2*x), Eq(g, x + y)])
>>> [arg.result_var for arg in r.results]
[]
>>> [arg.name for arg in r.arguments]
[x, y, f, g]
>>> [arg.name for arg in r.result_variables]
[f, g]
>>> r.local_vars
set()
Another more complicated example with a mixture of specied and automaticallyassigned names. Also has Matrix output.
>>> from sympy import Matrix
>>> r = make_routine(fcn, [x*y, Eq(f, 1), Eq(g, x + g), Matrix([[x, 2]])])
>>> [arg.result_var for arg in r.results]
[result_5397460570204848505]
>>> [arg.expr for arg in r.results]
1546
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
[x*y]
>>> [arg.name for arg in r.arguments]
[x, y, f, g, out_8598435338387848786]
We can examine the various arguments more closely:
>>> from sympy.utilities.codegen import (InputArgument, OutputArgument,
...
InOutArgument)
>>> [a.name for a in r.arguments if isinstance(a, InputArgument)]
[x, y]
>>>
[f,
>>>
[1,
[a.name for a in r.arguments if isinstance(a, OutputArgument)]
out_8598435338387848786]
[a.expr for a in r.arguments if isinstance(a, OutputArgument)]
Matrix([[x, 2]])]
>>> [a.name for a in r.arguments if isinstance(a, InOutArgument)]
[g]
>>> [a.expr for a in r.arguments if isinstance(a, InOutArgument)]
[g + x]
5.33.3 Decorator
Useful utility decorators.
sympy.utilities.decorator.conserve mpmath dps(func)
After the function nishes, resets the value of mpmath.mp.dps to the value it had before
the function was run.
sympy.utilities.decorator.doctest depends on(exe=None, modules=None, disable viewers=None)
Adds metadata about the depenencies which need to be met for doctesting the docstrings
of the decorated objects.
sympy.utilities.decorator.no attrs in subclass
Dont inherit certain attributes from a base class
>>> from sympy.utilities.decorator import no_attrs_in_subclass
>>> class A(object):
...
x = test
>>> A.x = no_attrs_in_subclass(A, A.x)
>>> class B(A):
...
pass
>>> hasattr(A, x)
True
>>> hasattr(B, x)
False
sympy.utilities.decorator.public(obj)
Append objs name to global all variable (call site).
By using this decorator on functions or classes you achieve the same goal as by lling
all variables manually, you just dont have to repeat your self (objects name). You
5.33. Utilities
1547
SymPy Documentation, Release 0.7.6
also know if object is public at denition site, not at some random location (where all
was set).
Note that in multiple decorator setup, in almost all cases, @public decorator must be
applied before any other decorators, because it relies on the pointer to objects global
namespace. If you apply other decorators rst, @public may end up modifying wrong
namespace.
Example:
>>> from sympy.utilities.decorator import public
>>> __all__
Traceback (most
...
NameError: name
Traceback (most
...
NameError: name
recent call last):
__all__ is not defined
recent call last):
__all__ is not defined
>>> @public
... def some_function():
...
pass
>>> __all__
[some_function]
sympy.utilities.decorator.threaded(func)
Apply func to subelements of an object, including Add.
This decorator is intended to make it uniformly possible to apply a function to all elements
of composite objects, e.g. matrices, lists, tuples and other iterable containers, or just
expressions.
This version of threaded() (page 1548) decorator allows threading over elements of Add
class. If this behavior is not desirable use xthreaded() (page 1548) decorator.
Functions using this decorator must have the following signature:
@threaded
def function(expr, *args, **kwargs):
sympy.utilities.decorator.threaded factory(func, use add)
A factory for threaded decorators.
sympy.utilities.decorator.xthreaded(func)
Apply func to subelements of an object, excluding Add.
This decorator is intended to make it uniformly possible to apply a function to all elements
of composite objects, e.g. matrices, lists, tuples and other iterable containers, or just
expressions.
This version of threaded() (page 1548) decorator disallows threading over elements of
Add class. If this behavior is not desirable use threaded() (page 1548) decorator.
Functions using this decorator must have the following signature:
@xthreaded
def function(expr, *args, **kwargs):
1548
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
5.33.4 Enumerative
This module includes functions and classes for enumerating and counting multiset partitions.
sympy.utilities.enumerative.multiset partitions taocp(multiplicities)
Enumerates partions of a multiset.
Parameters multiplicities :
list of integer multiplicities of the components of the multiset.
See Also:
sympy.utilities.iterables.multiset partititions Takes a multiset as input and directly yields multiset partitions. It dispatches to a number of functions, including
this one, for implementation. Most users will nd it more convenient to use than
multiset partitions taocp.
Examples
>>> from sympy.utilities.enumerative import list_visitor
>>> from sympy.utilities.enumerative import multiset_partitions_taocp
>>> # variables components and multiplicities represent the multiset abb
>>> components = ab
>>> multiplicities = [1, 2]
>>> states = multiset_partitions_taocp(multiplicities)
>>> list(list_visitor(state, components) for state in states)
[[[a, b, b]],
[[a, b], [b]],
[[a], [b, b]],
[[a], [b], [b]]]
Yields
state Internal data structure which encodes a particular partition. This output is then
usually processed by a vistor function which combines the information from this data
structure with the components themselves to produce an actual partition.
Unless they wish to create their own visitor function, users will have little need to
look inside this data structure. But, for reference, it is a 3-element list with components:
f is a frame array, which is used to divide pstack into parts.
lpart points to the base of the topmost part.
pstack is an array of PartComponent objects.
The state output oers a peek into the internal data structures of the enumeration
function. The client should treat this as read-only; any modication of the data
structure will cause unpredictable (and almost certainly incorrect) results. Also, the
components of state are modied in place at each iteration. Hence, the visitor must
be called at each loop iteration. Accumulating the state instances and processing
them later will not work.
5.33. Utilities
1549
SymPy Documentation, Release 0.7.6
sympy.utilities.enumerative.factoring visitor(state, primes)
Use with multiset partitions taocp to enumerate the ways a number can be expressed as
a product of factors. For this usage, the exponents of the prime factors of a number are
arguments to the partition enumerator, while the corresponding prime factors are input
here.
Examples
To enumerate the factorings of a number we can think of the elements of the partition
as being the prime factors and the multiplicities as being their exponents.
>>> from sympy.utilities.enumerative import factoring_visitor
>>> from sympy.utilities.enumerative import multiset_partitions_taocp
>>> from sympy import factorint
>>> primes, multiplicities = zip(*factorint(24).items())
>>> primes
(2, 3)
>>> multiplicities
(3, 1)
>>> states = multiset_partitions_taocp(multiplicities)
>>> list(factoring_visitor(state, primes) for state in states)
[[24], [8, 3], [12, 2], [4, 6], [4, 2, 3], [6, 2, 2], [2, 2, 2, 3]]
sympy.utilities.enumerative.list visitor(state, components)
Return a list of lists to represent the partition.
Examples
>>> from sympy.utilities.enumerative import list_visitor
>>> from sympy.utilities.enumerative import multiset_partitions_taocp
>>> states = multiset_partitions_taocp([1, 2, 1])
>>> s = next(states)
>>> list_visitor(s, abc) # for multiset a b b c
[[a, b, b, c]]
>>> s = next(states)
>>> list_visitor(s, [1, 2, 3]) # for multiset 1 2 2 3
[[1, 2, 2], [3]]
The approach of the function multiset partitions taocp is extended and generalized by the
class MultisetPartitionTraverser.
class sympy.utilities.enumerative.MultisetPartitionTraverser
Has methods to enumerate and count the partitions of a multiset.
This implements a refactored and extended version of Knuths algorithm 7.1.2.5M
[AOCP] (page 1915).
The enumeration methods of this class are generators and return data structures
which can be interpreted by the same visitor functions used for the output of multiset partitions taocp.
See Also:
multiset partitions taocp (page 1549), sympy.utilities.iterables.multiset partititions
1550
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
References
[AOCP] (page 1915), [Factorisatio] (page 1915), [Yorgey] (page 1915)
Examples
>>> from sympy.utilities.enumerative import MultisetPartitionTraverser
>>> m = MultisetPartitionTraverser()
>>> m.count_partitions([4,4,4,2])
127750
>>> m.count_partitions([3,3,3])
686
count partitions(multiplicities)
Returns the number of partitions of a multiset whose components have the multiplicities given in multiplicities.
For larger counts, this method is much faster than calling one of the enumerators
and counting the result. Uses dynamic programming to cut down on the number
of nodes actually explored. The dictionary used in order to accelerate the counting
process is stored in the MultisetPartitionTraverser object and persists across
calls. If the the user does not expect to call count partitions for any additional
multisets, the object should be cleared to save memory. On the other hand, the
cache built up from one count run can signicantly speed up subsequent calls to
count partitions, so it may be advantageous not to clear the object.
Notes
If one looks at the workings of Knuths algorithm M [AOCP] (page 1915), it can be
viewed as a traversal of a binary tree of parts. A part has (up to) two children, the left
child resulting from the spread operation, and the right child from the decrement
operation. The ordinary enumeration of multiset partitions is an in-order traversal of
this tree, and with the partitions corresponding to paths from the root to the leaves.
The mapping from paths to partitions is a little complicated, since the partition would
contain only those parts which are leaves or the parents of a spread link, not those
which are parents of a decrement link.
For counting purposes, it is sucient to count leaves, and this can be done with a
recursive in-order traversal. The number of leaves of a subtree rooted at a particular
part is a function only of that part itself, so memoizing has the potential to speed up
the counting dramatically.
This method follows a computational approach which is similar to the hypothetical
memoized recursive function, but with two dierences:
1.This method is iterative, borrowing its structure from the other enumerations
and maintaining an explicit stack of parts which are in the process of being
counted. (There may be multisets which can be counted reasonably quickly by
this implementation, but which would overow the default Python recursion limit
with a recursive implementation.)
2.Instead of using the part data structure directly, a more compact key is constructed. This saves space, but more importantly coalesces some parts which
would remain separate with physical keys.
5.33. Utilities
1551
SymPy Documentation, Release 0.7.6
Unlike the enumeration functions, there is currently no range version of
count partitions. If someone wants to stretch their brain, it should be possible to
construct one by memoizing with a histogram of counts rather than a single count,
and combining the histograms.
Examples
>>> from sympy.utilities.enumerative import MultisetPartitionTraverser
>>> m = MultisetPartitionTraverser()
>>> m.count_partitions([9,8,2])
288716
>>> m.count_partitions([2,2])
9
>>> del m
enum all(multiplicities)
Enumerate the partitions of a multiset.
See Also:
multiset partitions taocp (page 1549) which provides the same result as this
method, but is about twice as fast. Hence, enum all is primarily useful for testing. Also see the function for a discussion of states and visitors.
Examples
>>> from sympy.utilities.enumerative import list_visitor
>>> from sympy.utilities.enumerative import MultisetPartitionTraverser
>>> m = MultisetPartitionTraverser()
>>> states = m.enum_all([2,2])
>>> list(list_visitor(state, ab) for state in states)
[[[a, a, b, b]],
[[a, a, b], [b]],
[[a, a], [b, b]],
[[a, a], [b], [b]],
[[a, b, b], [a]],
[[a, b], [a, b]],
[[a, b], [a], [b]],
[[a], [a], [b, b]],
[[a], [a], [b], [b]]]
enum large(multiplicities, lb)
Enumerate the partitions of a multiset with lb < num(parts)
Equivalent to enum range(multiplicities, lb, sum(multiplicities))
Parameters multiplicities :
list of multiplicities of the components of the multiset.
lb :
Number of parts in the partition must be greater than this lower
bound.
See Also:
enum all (page 1552), enum small (page 1553), enum range (page 1553)
1552
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.utilities.enumerative import list_visitor
>>> from sympy.utilities.enumerative import MultisetPartitionTraverser
>>> m = MultisetPartitionTraverser()
>>> states = m.enum_large([2,2], 2)
>>> list(list_visitor(state, ab) for state in states)
[[[a, a], [b], [b]],
[[a, b], [a], [b]],
[[a], [a], [b, b]],
[[a], [a], [b], [b]]]
enum range(multiplicities, lb, ub)
Enumerate the partitions of a multiset with lb < num(parts) <= ub.
In particular, if partitions with exactly k parts are desired, call with (multiplicities, k - 1, k). This method generalizes enum all, enum small, and enum large.
Examples
>>> from sympy.utilities.enumerative import list_visitor
>>> from sympy.utilities.enumerative import MultisetPartitionTraverser
>>> m = MultisetPartitionTraverser()
>>> states = m.enum_range([2,2], 1, 2)
>>> list(list_visitor(state, ab) for state in states)
[[[a, a, b], [b]],
[[a, a], [b, b]],
[[a, b, b], [a]],
[[a, b], [a, b]]]
enum small(multiplicities, ub)
Enumerate multiset partitions with no more than ub parts.
Equivalent to enum range(multiplicities, 0, ub)
Parameters multiplicities :
list of multiplicities of the components of the multiset.
ub :
Maximum number of parts
See Also:
enum all (page 1552), enum large (page 1552), enum range (page 1553)
Examples
>>> from sympy.utilities.enumerative import list_visitor
>>> from sympy.utilities.enumerative import MultisetPartitionTraverser
>>> m = MultisetPartitionTraverser()
>>> states = m.enum_small([2,2], 2)
>>> list(list_visitor(state, ab) for state in states)
[[[a, a, b, b]],
[[a, a, b], [b]],
[[a, a], [b, b]],
5.33. Utilities
1553
SymPy Documentation, Release 0.7.6
[[a, b, b], [a]],
[[a, b], [a, b]]]
The implementation is based, in part, on the answer given to exercise 69, in Knuth
[AOCP] (page 1915).
5.33.5 Iterables
cartes
Returns the cartesian product of sequences as a generator.
Examples::
>>> from sympy.utilities.iterables import cartes
>>> list(cartes([1,2,3], ab))
[(1, a), (1, b), (2, a), (2, b), (3, a), (3, b)]
variations
variations(seq, n) Returns all the variations of the list of size n.
Has an optional third argument. Must be a boolean value and makes the method return the
variations with repetition if set to True, or the variations without repetition if set to False.
Examples::
>>> from sympy.utilities.iterables import variations
>>> list(variations([1,2,3], 2))
[(1, 2), (1, 3), (2, 1), (2, 3), (3, 1), (3, 2)]
>>> list(variations([1,2,3], 2, True))
[(1, 1), (1, 2), (1, 3), (2, 1), (2, 2), (2, 3), (3, 1), (3, 2), (3, 3)]
partitions
Although the combinatorics module contains Partition and IntegerPartition classes for investigation and manipulation of partitions, there are a few functions to generate partitions that
can be used as low-level tools for routines: partitions and multiset partitions. The former gives integer partitions, and the latter gives enumerated partitions of elements. There
is also a routine kbins that will give a variety of permutations of partions.
partitions:
>>> from sympy.utilities.iterables import partitions
>>> [p.copy() for s, p in partitions(7, m=2, size=True) if s == 2]
[{1: 1, 6: 1}, {2: 1, 5: 1}, {3: 1, 4: 1}]
multiset partitions:
>>> from sympy.utilities.iterables import multiset_partitions
>>> [p for p in multiset_partitions(3, 2)]
[[[0, 1], [2]], [[0, 2], [1]], [[0], [1, 2]]]
>>> [p for p in multiset_partitions([1, 1, 1, 2], 2)]
[[[1, 1, 1], [2]], [[1, 1, 2], [1]], [[1, 1], [1, 2]]]
1554
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
kbins:
>>> from sympy.utilities.iterables import kbins
>>> def show(k):
...
rv = []
...
for p in k:
...
rv.append(,.join([.join(j) for j in p]))
...
return sorted(rv)
...
>>> show(kbins(ABCD, 2))
[A,BCD, AB,CD, ABC,D]
>>> show(kbins(ABC, 2))
[A,BC, AB,C]
>>> show(kbins(ABC, 2, ordered=0)) # same as multiset_partitions
[A,BC, AB,C, AC,B]
>>> show(kbins(ABC, 2, ordered=1))
[A,BC, A,CB,
B,AC, B,CA,
C,AB, C,BA]
>>> show(kbins(ABC, 2, ordered=10))
[A,BC, AB,C, AC,B,
B,AC, BC,A,
C,AB]
>>> show(kbins(ABC, 2, ordered=11))
[A,BC, A,CB, AB,C, AC,B,
B,AC, B,CA, BA,C, BC,A,
C,AB, C,BA, CA,B, CB,A]
Docstring
sympy.utilities.iterables.binary partitions(n)
Generates the binary partition of n.
A binary partition consists only of numbers that are powers of two. Each step reduces a
2**(k+1) to 2**k and 2**k. Thus 16 is converted to 8 and 8.
Reference: TAOCP 4, section 7.2.1.5, problem 64
Examples
>>>
>>>
...
...
[4,
[2,
[2,
[1,
from sympy.utilities.iterables import binary_partitions
for i in binary_partitions(5):
print(i)
1]
2, 1]
1, 1, 1]
1, 1, 1, 1]
sympy.utilities.iterables.bracelets(n, k)
Wrapper to necklaces to return a free (unrestricted) necklace.
sympy.utilities.iterables.capture(func)
Return the printed output of func().
f unc should be a function without arguments that produces output with print statements.
5.33. Utilities
1555
SymPy Documentation, Release 0.7.6
>>> from sympy.utilities.iterables import capture
>>> from sympy import pprint
>>> from sympy.abc import x
>>> def foo():
...
print(hello world!)
...
>>> hello in capture(foo) # foo, not foo()
True
>>> capture(lambda: pprint(2/x))
2\n-\nx\n
sympy.utilities.iterables.common prefix(*seqs)
Return the subsequence that is a common start of sequences in seqs.
>>>
>>>
[0,
>>>
[0,
>>>
[1,
>>>
[1]
from sympy.utilities.iterables import common_prefix
common_prefix(list(range(3)))
1, 2]
common_prefix(list(range(3)), list(range(4)))
1, 2]
common_prefix([1, 2, 3], [1, 2, 5])
2]
common_prefix([1, 2, 3], [1, 3, 5])
sympy.utilities.iterables.common suffix(*seqs)
Return the subsequence that is a common ending of sequences in seqs.
>>>
>>>
[0,
>>>
[]
>>>
[2,
>>>
[3]
from sympy.utilities.iterables import common_suffix
common_suffix(list(range(3)))
1, 2]
common_suffix(list(range(3)), list(range(4)))
common_suffix([1, 2, 3], [9, 2, 3])
3]
common_suffix([1, 2, 3], [9, 7, 3])
sympy.utilities.iterables.dict merge(*dicts)
Merge dictionaries into a single dictionary.
sympy.utilities.iterables.filter symbols(iterator, exclude)
Only yield elements from iterator that do not occur in exclude.
Parameters iterator : iterable
iterator to take elements from :
exclude : iterable
elements to exclude :
Returns iterator : iterator
ltered iterator :
sympy.utilities.iterables.flatten(iterable, levels=None, cls=None)
Recursively denest iterable containers.
>>> from sympy.utilities.iterables import flatten
>>> flatten([1, 2, 3])
[1, 2, 3]
1556
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> flatten([1, 2, [3]])
[1, 2, 3]
>>> flatten([1, [2, 3], [4, 5]])
[1, 2, 3, 4, 5]
>>> flatten([1.0, 2, (1, None)])
[1.0, 2, 1, None]
If you want to denest only a specied number of levels of nested containers, then set
levels ag to the desired number of levels:
>>> ls = [[(-2, -1), (1, 2)], [(0, 0)]]
>>> flatten(ls, levels=1)
[(-2, -1), (1, 2), (0, 0)]
If cls argument is specied, it will only atten instances of that class, for example:
>>>
>>>
...
...
>>>
[1,
from sympy.core import Basic
class MyOp(Basic):
pass
flatten([MyOp(1, MyOp(2, 3))], cls=MyOp)
2, 3]
adapted from http://kogs-www.informatik.uni-hamburg.de/meine/python tricks
sympy.utilities.iterables.generate bell(n)
Return permutations of [0, 1, ..., n - 1] such that each permutation diers from the last
by the exchange of a single pair of neighbors. The n! permutations are returned as an
iterator. In order to obtain the next permutation from a random starting permutation,
use the next trotterjohnson method of the Permutation class (which generates the
same sequence in a dierent manner).
See Also:
sympy.combinatorics.Permutation.next trotterjohnson
References
http://en.wikipedia.org/wiki/Method ringing
http://stackoverow.com/questions/4856615/recursive-permutation/4857018
http://programminggeeks.com/bell-algorithm-for-permutation/
http://en.wikipedia.org/wiki/Steinhaus%E2%80%93Johnson%E2%80%93Trotter algorithm
Generating involutions, derangements, and relatives by ECO Vincent Vajnovszki,
DMTCS vol 1 issue 12, 2010
Examples
>>> from itertools import permutations
>>> from sympy.utilities.iterables import generate_bell
>>> from sympy import zeros, Matrix
This is the sort of permutation used in the ringing of physical bells, and does not produce
permutations in lexicographical order. Rather, the permutations dier from each other
5.33. Utilities
1557
SymPy Documentation, Release 0.7.6
by exactly one inversion, and the position at which the swapping occurs varies periodically in a simple fashion. Consider the rst few permutations of 4 elements generated
by permutations and generate bell:
>>> list(permutations(range(4)))[:5]
[(0, 1, 2, 3), (0, 1, 3, 2), (0, 2, 1, 3), (0, 2, 3, 1), (0, 3, 1, 2)]
>>> list(generate_bell(4))[:5]
[(0, 1, 2, 3), (0, 1, 3, 2), (0, 3, 1, 2), (3, 0, 1, 2), (3, 0, 2, 1)]
Notice how the 2nd and 3rd lexicographical permutations have 3 elements out of place
whereas each bell permutation always has only two elements out of place relative to
the previous permutation (and so the signature (+/-1) of a permutation is opposite of the
signature of the previous permutation).
How the position of inversion varies across the elements can be seen by tracing out
where the largest number appears in the permutations:
>>> m = zeros(4, 24)
>>> for i, p in enumerate(generate_bell(4)):
...
m[:, i] = Matrix([j - 3 for j in list(p)])
>>> m.print_nonzero(X)
[XXX XXXXXX XXXXXX XXX]
[XX XX XXXX XX XXXX XX XX]
[X XXXX XX XXXX XX XXXX X]
[ XXXXXX XXXXXX XXXXXX ]
# make largest zero
sympy.utilities.iterables.generate derangements(perm)
Routine to generate unique derangements.
TODO: This will be rewritten to use the ECO operator approach once the permutations
branch is in master.
See Also:
sympy.functions.combinatorial.factorials.subfactorial
Examples
>>> from sympy.utilities.iterables import generate_derangements
>>> list(generate_derangements([0, 1, 2]))
[[1, 2, 0], [2, 0, 1]]
>>> list(generate_derangements([0, 1, 2, 3]))
[[1, 0, 3, 2], [1, 2, 3, 0], [1, 3, 0, 2], [2, 0, 3, 1],
[2, 3, 0, 1], [2, 3, 1, 0], [3, 0,
>>> list(generate_derangements([0, 1, 1]))
[]
sympy.utilities.iterables.generate involutions(n)
Generates involutions.
An involution is a permutation that when multiplied by itself equals the identity permutation. In this implementation the involutions are generated using Fixed Points.
Alternatively, an involution can be considered as a permutation that does not contain any
cycles with a length that is greater than two.
Reference: http://mathworld.wolfram.com/PermutationInvolution.html
1558
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.utilities.iterables import generate_involutions
>>> list(generate_involutions(3))
[(0, 1, 2), (0, 2, 1), (1, 0, 2), (2, 1, 0)]
>>> len(list(generate_involutions(4)))
10
sympy.utilities.iterables.generate oriented forest(n)
This algorithm generates oriented forests.
An oriented graph is a directed graph having no symmetric pair of directed edges. A
forest is an acyclic graph, i.e., it has no cycles. A forest can also be described as a
disjoint union of trees, which are graphs in which any two vertices are connected by
exactly one simple path.
Reference:
[1] T. Beyer and S.M. Hedetniemi:
constant time generation of
rooted trees, SIAM J. Computing Vol.
9, No.
4, November 1980 [2]
http://stackoverow.com/questions/1633833/oriented-forest-taocp-algorithm-in-python
Examples
>>> from sympy.utilities.iterables import generate_oriented_forest
>>> list(generate_oriented_forest(4))
[[0, 1, 2, 3], [0, 1, 2, 2], [0, 1, 2, 1], [0, 1, 2, 0],
[0, 1, 1, 1], [0, 1, 1, 0], [0, 1,
sympy.utilities.iterables.group(seq, multiple=True)
Splits a sequence into a list of lists of equal, adjacent elements.
See Also:
multiset (page 1562)
Examples
>>> from sympy.utilities.iterables import group
>>> group([1, 1, 1, 2, 2, 3])
[[1, 1, 1], [2, 2], [3]]
>>> group([1, 1, 1, 2, 2, 3], multiple=False)
[(1, 3), (2, 2), (3, 1)]
>>> group([1, 1, 3, 2, 2, 1], multiple=False)
[(1, 2), (3, 1), (2, 2), (1, 1)]
sympy.utilities.iterables.has dups(seq)
Return True if there are any duplicate elements in seq.
Examples
>>> from sympy.utilities.iterables import has_dups
>>> from sympy import Dict, Set
5.33. Utilities
1559
SymPy Documentation, Release 0.7.6
>>> has_dups((1, 2, 1))
True
>>> has_dups(range(3))
False
>>> all(has_dups(c) is False for c in (set(), Set(), dict(), Dict()))
True
sympy.utilities.iterables.has variety(seq)
Return True if there are any dierent elements in seq.
Examples
>>> from sympy.utilities.iterables import has_variety
>>> has_variety((1, 2, 1))
True
>>> has_variety((1, 1, 1))
False
sympy.utilities.iterables.ibin(n, bits=0, str=False)
Return a list of length bits corresponding to the binary value of n with small bits to the
right (last). If bits is omitted, the length will be the number required to represent n. If
the bits are desired in reversed order, use the [::-1] slice of the returned list.
If a sequence of all bits-length lists starting from [0, 0,..., 0] through [1, 1, ..., 1] are
desired, pass a non-integer for bits, e.g. all.
If the bit string is desired pass str=True.
Examples
>>>
>>>
[1,
>>>
[0,
>>>
[0,
from sympy.utilities.iterables import ibin
ibin(2)
0]
ibin(2, 4)
0, 1, 0]
ibin(2, 4)[::-1]
1, 0, 0]
If all lists corresponding to 0 to 2**n - 1, pass a non-integer for bits:
>>>
>>>
...
(0,
(0,
(1,
(1,
bits = 2
for i in ibin(2, all):
print(i)
0)
1)
0)
1)
If a bit string is desired of a given length, use str=True:
>>> n = 123
>>> bits = 10
>>> ibin(n, bits, str=True)
0001111011
>>> ibin(n, bits, str=True)[::-1]
1560
# small bits left
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
1101111000
>>> list(ibin(3, all, str=True))
[000, 001, 010, 011, 100, 101, 110, 111]
sympy.utilities.iterables.interactive traversal(expr)
Traverse a tree asking a user which branch to choose.
sympy.utilities.iterables.kbins(l, k, ordered=None)
Return sequence l partitioned into k bins.
See Also:
partitions (page 1566), multiset partitions (page 1563)
Examples
>>> from sympy.utilities.iterables import kbins
The default is to give the items in the same order, but grouped into k partitions without
any reordering:
>>> from __future__ import print_function
>>> for p in kbins(list(range(5)), 2):
...
print(p)
...
[[0], [1, 2, 3, 4]]
[[0, 1], [2, 3, 4]]
[[0, 1, 2], [3, 4]]
[[0, 1, 2, 3], [4]]
The ordered ag which is either None (to give the simple partition of the the elements)
or is a 2 digit integer indicating whether the order of the bins and the order of the items
in the bins matters. Given:
A
B
C
D
=
=
=
=
[[0], [1, 2]]
[[1, 2], [0]]
[[2, 1], [0]]
[[0], [2, 1]]
the following values for ordered have the shown meanings:
00
01
10
11
means
means
means
means
A
A
A
A
==
==
==
==
B == C == D
B
D
A
>>> for ordered in [None, 0, 1, 10, 11]:
...
print(ordered = %s % ordered)
...
for p in kbins(list(range(3)), 2, ordered=ordered):
...
print(
%s % p)
...
ordered = None
[[0], [1, 2]]
[[0, 1], [2]]
ordered = 0
[[0, 1], [2]]
[[0, 2], [1]]
[[0], [1, 2]]
5.33. Utilities
1561
SymPy Documentation, Release 0.7.6
ordered = 1
[[0], [1, 2]]
[[0], [2, 1]]
[[1], [0, 2]]
[[1], [2, 0]]
[[2], [0, 1]]
[[2], [1, 0]]
ordered = 10
[[0, 1], [2]]
[[2], [0, 1]]
[[0, 2], [1]]
[[1], [0, 2]]
[[0], [1, 2]]
[[1, 2], [0]]
ordered = 11
[[0], [1, 2]]
[[0, 1], [2]]
[[0], [2, 1]]
[[0, 2], [1]]
[[1], [0, 2]]
[[1, 0], [2]]
[[1], [2, 0]]
[[1, 2], [0]]
[[2], [0, 1]]
[[2, 0], [1]]
[[2], [1, 0]]
[[2, 1], [0]]
sympy.utilities.iterables.minlex(seq, directed=True, is set=False, small=None)
Return a tuple where the smallest element appears rst; if directed is True (default)
then the order is preserved, otherwise the sequence will be reversed if that gives a
smaller ordering.
If every element appears only once then is set can be set to True for more ecient
processing.
If the smallest element is known at the time of calling, it can be passed and the calculation
of the smallest element will be omitted.
Examples
>>>
>>>
(0,
>>>
(0,
>>>
(0,
from sympy.combinatorics.polyhedron import minlex
minlex((1, 2, 0))
1, 2)
minlex((1, 0, 2))
2, 1)
minlex((1, 0, 2), directed=False)
1, 2)
>>> minlex(11010011000, directed=True)
00011010011
>>> minlex(11010011000, directed=False)
00011001011
sympy.utilities.iterables.multiset(seq)
Return the hashable sequence in multiset form with values being the multiplicity of the
item in the sequence.
1562
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
See Also:
group (page 1559)
Examples
>>> from sympy.utilities.iterables import multiset
>>> multiset(mississippi)
{i: 4, m: 1, p: 2, s: 4}
sympy.utilities.iterables.multiset combinations(m, n, g=None)
Return the unique combinations of size n from multiset m.
Examples
>>> from sympy.utilities.iterables import multiset_combinations
>>> from itertools import combinations
>>> [.join(i) for i in multiset_combinations(baby, 3)]
[abb, aby, bby]
>>> def count(f, s): return len(list(f(s, 3)))
The number of combinations depends on the number of letters; the number of unique
combinations depends on how the letters are repeated.
>>> s1 = abracadabra
>>> s2 = banana tree
>>> count(combinations, s1), count(multiset_combinations, s1)
(165, 23)
>>> count(combinations, s2), count(multiset_combinations, s2)
(165, 54)
sympy.utilities.iterables.multiset partitions(multiset, m=None)
Return unique partitions of the given multiset (in list form). If m is None, all multisets
will be returned, otherwise only partitions with m parts will be returned.
If multiset is an integer, a range [0, 1, ..., multiset - 1] will be supplied.
See Also:
partitions
(page
1566),
sympy.combinatorics.partitions.Partition
(page 187), sympy.combinatorics.partitions.IntegerPartition (page 189),
sympy.functions.combinatorial.numbers.nT
Notes
When all the elements are the same in the multiset, the order of the returned partitions
is determined by the partitions routine. If one is counting partitions then it is better
to use the nT function.
Examples
5.33. Utilities
1563
SymPy Documentation, Release 0.7.6
>>> from sympy.utilities.iterables import multiset_partitions
>>> list(multiset_partitions([1, 2, 3, 4], 2))
[[[1, 2, 3], [4]], [[1, 2, 4], [3]], [[1, 2], [3, 4]],
[[1, 3, 4], [2]], [[1, 3], [2, 4]], [[1, 4], [2, 3]],
[[1], [2, 3, 4]]]
>>> list(multiset_partitions([1, 2, 3, 4], 1))
[[[1, 2, 3, 4]]]
Only unique partitions are returned and these will be returned in a canonical order regardless of the order of the input:
>>> a = [1, 2, 2, 1]
>>> ans = list(multiset_partitions(a, 2))
>>> a.sort()
>>> list(multiset_partitions(a, 2)) == ans
True
>>> a = range(3, 1, -1)
>>> (list(multiset_partitions(a)) ==
... list(multiset_partitions(sorted(a))))
True
If m is omitted then all partitions will be returned:
>>> list(multiset_partitions([1, 1, 2]))
[[[1, 1, 2]], [[1, 1], [2]], [[1, 2], [1]], [[1], [1], [2]]]
>>> list(multiset_partitions([1]*3))
[[[1, 1, 1]], [[1], [1, 1]], [[1], [1], [1]]]
Counting
The number of partitions of a set is given by the bell number:
>>> from sympy import bell
>>> len(list(multiset_partitions(5))) == bell(5) == 52
True
The number of partitions of length k from a set of size n is given by the Stirling Number
of the 2nd kind:
>>> def S2(n, k):
...
from sympy import Dummy, binomial, factorial, Sum
...
if k > n:
...
return 0
...
j = Dummy()
...
arg = (-1)**(k-j)*j**n*binomial(k,j)
...
return 1/factorial(k)*Sum(arg,(j,0,k)).doit()
...
>>> S2(5, 2) == len(list(multiset_partitions(5, 2))) == 15
True
These comments on counting apply to sets, not multisets.
sympy.utilities.iterables.multiset permutations(m, size=None, g=None)
Return the unique permutations of multiset m.
1564
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.utilities.iterables import multiset_permutations
>>> from sympy import factorial
>>> [.join(i) for i in multiset_permutations(aab)]
[aab, aba, baa]
>>> factorial(len(banana))
720
>>> len(list(multiset_permutations(banana)))
60
sympy.utilities.iterables.necklaces(n, k, free=False)
A routine to generate necklaces that may (free=True) or may not (free=False) be turned
over to be viewed. The necklaces returned are comprised of n integers (beads) with k
dierent values (colors). Only unique necklaces are returned.
References
http://mathworld.wolfram.com/Necklace.html
Examples
>>> from sympy.utilities.iterables import necklaces, bracelets
>>> def show(s, i):
...
return .join(s[j] for j in i)
The unrestricted necklace is sometimes also referred to as a bracelet (an object that
can be turned over, a sequence that can be reversed) and the term necklace is used to
imply a sequence that cannot be reversed. So ACB == ABC for a bracelet (rotate and
reverse) while the two are dierent for a necklace since rotation alone cannot make the
two sequences the same.
(mnemonic: Bracelets can be viewed Backwards, but Not Necklaces.)
>>> B = [show(ABC, i) for i in bracelets(3, 3)]
>>> N = [show(ABC, i) for i in necklaces(3, 3)]
>>> set(N) - set(B)
set([ACB])
>>> list(necklaces(4, 2))
[(0, 0, 0, 0), (0, 0, 0, 1), (0, 0, 1, 1),
(0, 1, 0, 1), (0, 1, 1, 1), (1, 1, 1, 1)]
>>> [show(.o, i) for i in bracelets(4, 2)]
[...., ...o, ..oo, .o.o, .ooo, oooo]
sympy.utilities.iterables.numbered symbols(prex=x, cls=None, start=0, exclude=[], *args, **assumptions)
Generate an innite stream of Symbols consisting of a prex and increasing subscripts
provided that they do not occur in exclude.
Parameters prex : str, optional
The prex to use. By default, this function will generate symbols of
the form x0, x1, etc.
5.33. Utilities
1565
SymPy Documentation, Release 0.7.6
cls : class, optional
The class to use. By default, it uses Symbol, but you can also use Wild
or Dummy.
start : int, optional
The start number. By default, it is 0.
Returns sym : Symbol
The subscripted symbols.
sympy.utilities.iterables.partitions(n, m=None, k=None, size=False)
Generate all partitions of integer n (>= 0).
Parameters m : integer (default gives partitions of all sizes)
limits number of parts in parition (mnemonic: m, maximum parts)
k : integer (default gives partitions number from 1 through n)
limits the numbers that are kept in the partition (mnemonic: k, keys)
size : bool (default False, only partition is returned)
when True then (M, P) is returned where M is the sum of the multiplicities and P is the generated partition.
Each partition is represented as a dictionary, mapping an integer :
to the number of copies of that integer in the partition. For example,
:
the rst partition of 4 returned is {4: 1}, 4: one of them. :
See Also:
sympy.combinatorics.partitions.Partition
(page
sympy.combinatorics.partitions.IntegerPartition (page 189)
187),
Examples
>>> from sympy.utilities.iterables import partitions
The numbers appearing in the partition (the key of the returned dict) are limited with k:
>>>
...
{2:
{1:
{1:
{1:
for p in partitions(6, k=2):
print(p)
3}
2, 2: 2}
4, 2: 1}
6}
The maximum number of parts in the partion (the sum of the values in the returned dict)
are limited with m:
>>>
...
...
{6:
{1:
{2:
{3:
1566
for p in partitions(6, m=2):
print(p)
1}
1, 5: 1}
1, 4: 1}
2}
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Note that the same dictionary object is returned each time. This is for speed: generating each partition goes quickly, taking constant time, independent of n.
>>> [p for p in partitions(6, k=2)]
[{1: 6}, {1: 6}, {1: 6}, {1: 6}]
If you want to build a list of the returned dictionaries then make a copy of them:
>>> [p.copy() for p in partitions(6, k=2)]
[{2: 3}, {1: 2, 2: 2}, {1: 4, 2: 1}, {1: 6}]
>>> [(M, p.copy()) for M, p in partitions(6, k=2, size=True)]
[(3, {2: 3}), (4, {1: 2, 2: 2}), (5, {1: 4, 2: 1}), (6, {1: 6})]
Reference: modied from Tim Peters version to allow for k and m values:
code.activestate.com/recipes/218332-generator-for-integer-partitions/
sympy.utilities.iterables.postfixes(seq)
Generate all postxes of a sequence.
Examples
>>> from sympy.utilities.iterables import postfixes
>>> list(postfixes([1,2,3,4]))
[[4], [3, 4], [2, 3, 4], [1, 2, 3, 4]]
sympy.utilities.iterables.postorder traversal(node, keys=None)
Do a postorder traversal of a tree.
This generator recursively yields nodes that it has visited in a postorder fashion. That
is, it descends through the tree depth-rst to yield all of a nodes childrens postorder
traversal before yielding the node itself.
Parameters node : sympy expression
The expression to traverse.
keys : (default None) sort key(s)
The key(s) used to sort args of Basic objects. When None, args of
Basic objects are processed in arbitrary order. If key is dened, it
will be passed along to ordered() as the only key(s) to use to sort the
arguments; if key is simply True then the default keys of ordered will
be used (node count and default sort key).
Examples
>>> from sympy.utilities.iterables import postorder_traversal
>>> from sympy.abc import w, x, y, z
The nodes are returned in the order that they are encountered unless key is given; simply
passing key=True will guarantee that the traversal is unique.
>>>
[z,
>>>
[w,
list(postorder_traversal(w
y, x, x + y, z*(x + y), w,
list(postorder_traversal(w
z, x, y, x + y, z*(x + y),
5.33. Utilities
+
w
+
w
(x + y)*z))
+ z*(x + y)]
(x + y)*z, keys=True))
+ z*(x + y)]
1567
SymPy Documentation, Release 0.7.6
Yields
subtree [sympy expression] All of the subtrees in the tree.
sympy.utilities.iterables.prefixes(seq)
Generate all prexes of a sequence.
Examples
>>> from sympy.utilities.iterables import prefixes
>>> list(prefixes([1,2,3,4]))
[[1], [1, 2], [1, 2, 3], [1, 2, 3, 4]]
sympy.utilities.iterables.reshape(seq, how)
Reshape the sequence according to the template in how.
Examples
>>> from sympy.utilities import reshape
>>> seq = list(range(1, 9))
>>> reshape(seq, [4]) # lists of 4
[[1, 2, 3, 4], [5, 6, 7, 8]]
>>> reshape(seq, (4,)) # tuples of 4
[(1, 2, 3, 4), (5, 6, 7, 8)]
>>> reshape(seq, (2, 2)) # tuples of 4
[(1, 2, 3, 4), (5, 6, 7, 8)]
>>> reshape(seq, (2, [2])) # (i, i, [i, i])
[(1, 2, [3, 4]), (5, 6, [7, 8])]
>>> reshape(seq, ((2,), [2])) # etc....
[((1, 2), [3, 4]), ((5, 6), [7, 8])]
>>> reshape(seq, (1, [2], 1))
[(1, [2, 3], 4), (5, [6, 7], 8)]
>>> reshape(tuple(seq), ([[1], 1, (2,)],))
(([[1], 2, (3, 4)],), ([[5], 6, (7, 8)],))
>>> reshape(tuple(seq), ([1], 1, (2,)))
(([1], 2, (3, 4)), ([5], 6, (7, 8)))
>>> reshape(list(range(12)), [2, [3], set([2]), (1, (3,), 1)])
[[0, 1, [2, 3, 4], set([5, 6]), (7, (8, 9, 10), 11)]]
sympy.utilities.iterables.rotate left(x, y)
Left rotates a list x by the number of steps specied in y.
1568
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>>
>>>
>>>
[1,
from sympy.utilities.iterables import rotate_left
a = [0, 1, 2]
rotate_left(a, 1)
2, 0]
sympy.utilities.iterables.rotate right(x, y)
Right rotates a list x by the number of steps specied in y.
Examples
>>>
>>>
>>>
[2,
from sympy.utilities.iterables import rotate_right
a = [0, 1, 2]
rotate_right(a, 1)
0, 1]
sympy.utilities.iterables.runs(seq, op=<built-in function gt>)
Group the sequence into lists in which successive elements all compare the same with
the comparison operator, op: op(seq[i + 1], seq[i]) is True from all elements in a run.
Examples
>>> from sympy.utilities.iterables import runs
>>> from operator import ge
>>> runs([0, 1, 2, 2, 1, 4, 3, 2, 2])
[[0, 1, 2], [2], [1, 4], [3], [2], [2]]
>>> runs([0, 1, 2, 2, 1, 4, 3, 2, 2], op=ge)
[[0, 1, 2, 2], [1, 4], [3], [2, 2]]
sympy.utilities.iterables.sift(seq, keyfunc)
Sift the sequence, seq into a dictionary according to keyfunc.
OUTPUT: each element in expr is stored in a list keyed to the value of keyfunc for the
element.
See Also:
ordered
Examples
>>> from sympy.utilities import sift
>>> from sympy.abc import x, y
>>> from sympy import sqrt, exp
>>> sift(range(5), lambda x: x % 2)
{0: [0, 2, 4], 1: [1, 3]}
sift() returns a defaultdict() object, so any key that has no matches will give [].
>>> sift([x], lambda x: x.is_commutative)
{True: [x]}
5.33. Utilities
1569
SymPy Documentation, Release 0.7.6
>>> _[False]
[]
Sometimes you wont know how many keys you will get:
>>> sift([sqrt(x), exp(x), (y**x)**2],
...
lambda x: x.as_base_exp()[0])
{E: [exp(x)], x: [sqrt(x)], y: [y**(2*x)]}
If you need to sort the sifted items it might be better to use ordered which can economically apply multiple sort keys to a squence while sorting.
sympy.utilities.iterables.subsets(seq, k=None, repetition=False)
Generates all k-subsets (combinations) from an n-element set, seq.
A k-subset of an n-element set is any subset of length exactly k. The number of k-subsets
of an n-element set is given by binomial(n, k), whereas there are 2**n subsets all together.
If k is None then all 2**n subsets will be returned from shortest to longest.
Examples
>>> from sympy.utilities.iterables import subsets
subsets(seq, k) will return the n!/k!/(n - k)! k-subsets (combinations) without repetition,
i.e. once an item has been removed, it can no longer be taken:
>>> list(subsets([1,
[(1, 2)]
>>> list(subsets([1,
[(), (1,), (2,), (1,
>>> list(subsets([1,
[(1, 2), (1, 3), (2,
2], 2))
2]))
2)]
2, 3], 2))
3)]
subsets(seq, k, repetition=True) will return the (n - 1 + k)!/k!/(n - 1)! combinations with
repetition:
>>> list(subsets([1, 2], 2, repetition=True))
[(1, 1), (1, 2), (2, 2)]
If you ask for more items than are in the set you get the empty set unless you allow
repetitions:
>>> list(subsets([0, 1], 3, repetition=False))
[]
>>> list(subsets([0, 1], 3, repetition=True))
[(0, 0, 0), (0, 0, 1), (0, 1, 1), (1, 1, 1)]
sympy.utilities.iterables.take(iter, n)
Return n items from iter iterator.
sympy.utilities.iterables.topological sort(graph, key=None)
Topological sort of graphs vertices.
Parameters graph : tuple[list, list[tuple[T, T]]
A tuple consisting of a list of vertices and a list of edges of a graph to
be sorted topologically.
key : callable[T] (optional)
1570
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Ordering key for vertices on the same level. By default the natural
(e.g. lexicographic) ordering is used (in this case the base type must
implement ordering relations).
Examples
Consider a graph:
+---+
+---+
+---+
| 7 |\
| 5 |
| 3 |
+---+ \
+---+
+---+
|
\ /
/ |
| / \ /
\ /
|
V V
V V
|
+----+
+---+ |
| 11 |
| 8 | |
+----+
+---+ |
/
|
| | \
| \
\ /
/ \ |
V \
V V
/ V V
+---+ \
+---+ | +----+
| 2 | | | 9 | | | 10 |
+---+ | +---+ | +----+
\
/
where vertices are integers. This graph can be encoded using elementary Pythons data
structures as follows:
>>> V = [2, 3, 5, 7, 8, 9, 10, 11]
>>> E = [(7, 11), (7, 8), (5, 11), (3, 8), (3, 10),
...
(11, 2), (11, 9), (11, 10), (8, 9)]
To compute a topological sort for graph (V, E) issue:
>>> from sympy.utilities.iterables import topological_sort
>>> topological_sort((V, E))
[3, 5, 7, 8, 11, 2, 9, 10]
If specic tie breaking approach is needed, use key parameter:
>>> topological_sort((V, E), key=lambda v: -v)
[7, 5, 11, 3, 10, 8, 9, 2]
Only acyclic graphs can be sorted. If the input graph has a cycle, then ValueError will
be raised:
>>> topological_sort((V, E + [(10, 7)]))
Traceback (most recent call last):
...
ValueError: cycle detected
Traceback (most recent call last):
...
ValueError: cycle detected
See Also:
http://en.wikipedia.org/wiki/Topological sorting
5.33. Utilities
1571
SymPy Documentation, Release 0.7.6
sympy.utilities.iterables.unflatten(iter, n=2)
Group iter into tuples of length n. Raise an error if the length of iter is not a multiple
of n.
sympy.utilities.iterables.uniq(seq, result=None)
Yield unique elements from seq as an iterator. The second parameter result is used
internally; it is not necessary to pass anything for this.
Examples
>>> from sympy.utilities.iterables import uniq
>>> dat = [1, 4, 1, 5, 4, 2, 1, 2]
>>> type(uniq(dat)) in (list, tuple)
False
>>> list(uniq(dat))
[1, 4, 5, 2]
>>> list(uniq(x for x in dat))
[1, 4, 5, 2]
>>> list(uniq([[1], [2, 1], [1]]))
[[1], [2, 1]]
sympy.utilities.iterables.unrestricted necklace(*args, **kwargs)
Wrapper to necklaces to return a free (unrestricted) necklace.
sympy.utilities.iterables.variations(seq, n, repetition=False)
Returns a generator of the n-sized variations of seq (size N). repetition controls
whether items in seq can appear more than once;
See Also:
sympy.core.compatibility.permutations, sympy.core.compatibility.product
Examples
variations(seq, n) will return N! / (N - n)! permutations without repetition of seqs elements:
>>> from sympy.utilities.iterables import variations
>>> list(variations([1, 2], 2))
[(1, 2), (2, 1)]
variations(seq, n, True) will return the N**n permutations obtained by allowing repetition of elements:
>>> list(variations([1, 2], 2, repetition=True))
[(1, 1), (1, 2), (2, 1), (2, 2)]
If you ask for more items than are in the set you get the empty set unless you allow
repetitions:
>>> list(variations([0, 1], 3, repetition=False))
[]
>>> list(variations([0, 1], 3, repetition=True))[:4]
[(0, 0, 0), (0, 0, 1), (0, 1, 0), (0, 1, 1)]
1572
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
5.33.6 Lambdify
This module provides convenient functions to transform sympy expressions to lambda functions which can be used to calculate numerical values very fast.
sympy.utilities.lambdify.implemented function(symfunc, implementation)
Add numerical implementation to function symfunc.
symfunc can be an UndefinedFunction instance, or a name string. In the latter case we
create an UndefinedFunction instance with that name.
Be aware that this is a quick workaround, not a general method to create special symbolic
functions. If you want to create a symbolic function to be used by all the machinery of
SymPy you should subclass the Function class.
Parameters symfunc : str or UndefinedFunction instance
If str, then create new UndefinedFunction with this as name. If
symf unc is a sympy function, attach implementation to it.
implementation : callable
numerical implementation to be called by evalf() or lambdify
Returns afunc : sympy.FunctionClass instance
function with attached implementation
Examples
>>>
>>>
>>>
>>>
>>>
>>>
5
from sympy.abc import x
from sympy.utilities.lambdify import lambdify, implemented_function
from sympy import Function
f = implemented_function(Function(f), lambda x: x+1)
lam_f = lambdify(x, f(x))
lam_f(4)
sympy.utilities.lambdify.lambdastr(args, expr, printer=None, dummify=False)
Returns a string that can be evaluated to a lambda function.
Examples
>>> from sympy.abc import x, y, z
>>> from sympy.utilities.lambdify import lambdastr
>>> lambdastr(x, x**2)
lambda x: (x**2)
>>> lambdastr((x,y,z), [z,y,x])
lambda x,y,z: ([z, y, x])
Although tuples may not appear as arguments to lambda in Python 3, lambdastr will create a lambda function that will unpack the original arguments so that nested arguments
can be handled:
>>> lambdastr((x, (y, z)), x + y)
lambda _0,_1: (lambda x,y,z: (x + y))(*list(__flatten_args__([_0,_1])))
5.33. Utilities
1573
SymPy Documentation, Release 0.7.6
sympy.utilities.lambdify.lambdify(args, expr, modules=None, printer=None,
use imps=True, dummify=True)
Returns a lambda function for fast calculation of numerical values.
If not specied dierently by the user, SymPy functions are replaced as far as possible
by either python-math, numpy (if available) or mpmath functions - exactly in this order.
To change this behavior, the modules argument can be used. It accepts:
the strings math, mpmath, numpy, numexpr, sympy
any modules (e.g. math)
dictionaries that map names of sympy functions to arbitrary functions
lists that contain a mix of the arguments above, with higher priority given to entries
appearing rst.
The default behavior is to substitute all arguments in the provided expression with
dummy symbols. This allows for applied functions (e.g. f(t)) to be supplied as arguments. Call the function with dummify=False if dummy substitution is unwanted (and
args is not a string). If you want to view the lambdied function or provide sympy as
the module, you should probably set dummify=False.
For functions involving large array calculations, numexpr can provide a signicant speedup over numpy.
Please note that the available functions for numexpr are more limited than numpy but can be expanded with implemented function
and user dened subclasses of Function.
If specied, numexpr may be the
only option in modules. The ocial list of numexpr functions can be found at:
https://github.com/pydata/numexpr#supported-functions
Examples
>>>
>>>
>>>
>>>
from
from
from
from
sympy.utilities.lambdify import implemented_function
sympy import sqrt, sin, Matrix
sympy import Function
sympy.abc import w, x, y, z
>>> f = lambdify(x, x**2)
>>> f(2)
4
>>> f = lambdify((x, y, z), [z, y, x])
>>> f(1,2,3)
[3, 2, 1]
>>> f = lambdify(x, sqrt(x))
>>> f(4)
2.0
>>> f = lambdify((x, y), sin(x*y)**2)
>>> f(0, 5)
0.0
>>> row = lambdify((x, y), Matrix((x, x + y)).T, modules=sympy)
>>> row(1, 2)
Matrix([[1, 3]])
Tuple arguments are handled and the lambdied function should be called with the same
type of arguments as were used to create the function.:
>>> f = lambdify((x, (y, z)), x + y)
>>> f(1, (2, 4))
3
1574
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
A more robust way of handling this is to always work with attened arguments:
>>>
>>>
>>>
>>>
>>>
10
from sympy.utilities.iterables import flatten
args = w, (x, (y, z))
vals = 1, (2, (3, 4))
f = lambdify(flatten(args), w + x + y + z)
f(*flatten(vals))
Functions present in expr can also carry their own numerical implementations, in a
callable attached to the imp attribute. Usually you attach this using the implemented function factory:
>>> f = implemented_function(Function(f), lambda x: x+1)
>>> func = lambdify(x, f(x))
>>> func(4)
5
lambdify always prefers imp implementations to implementations in other namespaces, unless the use imps input parameter is False.
Deprecation Warnings
In previous releases lambdify replaced Matrix with numpy.matrix by default. As of
release 0.7.6 numpy.array is being transitioned to the default. In release 0.7.7 this
transition will be complete. For now, to use the new default behavior you must pass in
[{ImmutableMatrix: numpy.array}, numpy] to the modules kwarg.
>>> from sympy import lambdify, Matrix
>>> from sympy.abc import x, y
>>> import numpy
>>> mat2array = [{ImmutableMatrix: numpy.array}, numpy]
>>> f = lambdify((x, y), Matrix([x, y]), modules=mat2array)
>>> f(1, 2)
array([[1],
[2]])
Usage
1.Use one of the provided modules:
>>>
>>>
>>>
>>>
from sympy import sin, tan, gamma
from sympy.utilities.lambdify import lambdastr
from sympy.abc import x, y
f = lambdify(x, sin(x), math)
Attention: Functions that are not in the math module will throw a name
error when the lambda function is evaluated! So this would be better:
>>> f = lambdify(x, sin(x)*gamma(x), (math, mpmath, sympy))
2.Use some other module:
>>> import numpy
>>> f = lambdify((x,y), tan(x*y), numpy)
5.33. Utilities
1575
SymPy Documentation, Release 0.7.6
Attention: There are naming dierences between numpy and sympy. So if
you simply take the numpy module, e.g. sympy.atan will not be translated to
numpy.arctan. Use the modied module instead by passing the string numpy:
>>> f = lambdify((x,y), tan(x*y), numpy)
>>> f(1, 2)
-2.18503986326
>>> from numpy import array
>>> f(array([1, 2, 3]), array([2, 3, 5]))
[-2.18503986 -0.29100619 -0.8559934 ]
3.Use a dictionary dening custom functions:
>>> def my_cool_function(x): return sin(%s) is cool % x
>>> myfuncs = {sin : my_cool_function}
>>> f = lambdify(x, sin(x), myfuncs); f(1)
sin(1) is cool
5.33.7 Memoization
sympy.utilities.memoization.assoc recurrence memo(base seq)
Memo decorator for associated sequences dened by recurrence starting from base
base seq(n) callable to get base sequence elements
XXX works only for Pn0 = base seq(0) cases XXX works only for m <= n cases
sympy.utilities.memoization.recurrence memo(initial)
Memo decorator for sequences dened by recurrence
See usage examples e.g. in the specfun/combinatorial module
5.33.8 Miscellaneous
Miscellaneous stu that doesnt really t anywhere else.
sympy.utilities.misc.debug(*args)
Print *args if SYMPY DEBUG is True, else do nothing.
sympy.utilities.misc.debug decorator(func)
If SYMPY DEBUG is True, it will print a nice execution tree with arguments and results
of all decorated functions, else do nothing.
sympy.utilities.misc.find executable(executable, path=None)
Try to nd executable in the directories listed in path (a string listing directories separated by os.pathsep; defaults to os.environ[PATH]). Returns the complete lename
or None if not found
sympy.utilities.misc.rawlines(s)
Return a cut-and-pastable string that, when printed, is equivalent to the input. The
string returned is formatted so it can be indented nicely within tests; in some cases it is
wrapped in the dedent function which has to be imported from textwrap.
1576
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
Note: because there are characters in the examples below that need to be escaped because they are themselves within a triple quoted docstring, expressions below look more
complicated than they would be if they were printed in an interpreter window.
>>>
>>>
>>>
>>>
(
from sympy.utilities.misc import rawlines
from sympy import TableForm
s = str(TableForm([[1, 10]], headings=(None, [a, bee])))
print(rawlines(s)) # the \ appears as \ when printed
a bee\n
-----\n
1 10
)
>>> print(rawlines(this
... that))
dedent(\
this
that)
>>> print(rawlines(this
... that
... ))
dedent(\
this
that
)
>>> s = this
... is a triple
...
>>> print(rawlines(s))
dedent(\
this
is a triple
)
>>> print(rawlines(this
... that
...
))
(
this\n
that\n
5.33.9 PKGDATA
pkgdata is a simple, extensible way for a package to acquire data le resources.
The getResource function is equivalent to the standard idioms, such as the following minimal
implementation:
import sys, os
5.33. Utilities
1577
SymPy Documentation, Release 0.7.6
def getResource(identifier, pkgname=__name__):
pkgpath = os.path.dirname(sys.modules[pkgname].__file__)
path = os.path.join(pkgpath, identifier)
return open(os.path.normpath(path), mode=rb)
When a loader is present on the module given by name , it will defer getResource to its
get data implementation and return it as a le-like object (such as StringIO).
sympy.utilities.pkgdata.get resource(identier, pkgname=sympy.utilities.pkgdata)
Acquire a readable object for a given package name and identier. An IOError will be
raised if the resource can not be found.
For example:
mydata = get_resource(mypkgdata.jpg).read()
Note that the package name must be fully qualied, if given, such that it would be found
in sys.modules.
In some cases, getResource will return a real le object. In that case, it may be useful to
use its name attribute to get the path rather than use it as a le-like object. For example,
you may be handing data o to a C API.
5.33.10 pytest
py.test hacks to support XFAIL/XPASS
sympy.utilities.pytest.SKIP(reason)
Similar to skip(), but this is a decorator.
sympy.utilities.pytest.raises(expectedException, code=None)
Tests that code raises the exception expectedException.
code may be a callable, such as a lambda expression or function name.
If code is not given or None, raises will return a context manager for use in with statements; the code to execute then comes from the scope of the with.
raises() does nothing if the callable raises the expected exception, otherwise it raises
an AssertionError.
Examples
>>> from sympy.utilities.pytest import raises
>>> raises(ZeroDivisionError, lambda: 1/0)
>>> raises(ZeroDivisionError, lambda: 1/2)
Traceback (most recent call last):
...
AssertionError: DID NOT RAISE
Traceback (most recent call last):
...
AssertionError: DID NOT RAISE
>>> with raises(ZeroDivisionError):
...
n = 1/0
>>> with raises(ZeroDivisionError):
1578
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
...
n = 1/2
Traceback (most
...
AssertionError:
Traceback (most
...
AssertionError:
recent call last):
DID NOT RAISE
recent call last):
DID NOT RAISE
Note that you cannot test multiple statements via with raises:
>>> with raises(ZeroDivisionError):
...
n = 1/0
# will execute and raise, aborting the with
...
n = 9999/0 # never executed
This is just what with is supposed to do: abort the contained statement sequence at the
rst exception and let the context manager deal with the exception.
To test multiple statements, youll need a separate with for each:
>>> with raises(ZeroDivisionError):
...
n = 1/0
# will execute and raise
>>> with raises(ZeroDivisionError):
...
n = 9999/0 # will also execute and raise
5.33.11 Randomised Testing
Helpers for randomized testing
sympy.utilities.randtest.comp(z1, z2, tol)
Return a bool indicating whether the error between z1 and z2 is <= tol.
If z2 is non-zero and |z1| > 1 the error is normalized by |z1|, so if you want the absolute
error, call this as comp(z1 - z2, 0, tol).
sympy.utilities.randtest.random complex number(a=2, b=-1, c=3, d=1, rational=False)
Return a random complex number.
To reduce chance of hitting branch cuts or anything, we guarantee b <= Im z <= d, a
<= Re z <= c
sympy.utilities.randtest.test derivative numerically(f, z, tol=1e-06, a=2, b=-1,
c=3, d=1)
Test numerically that the symbolically computed derivative of f with respect to z is correct.
This routine does not test whether there are Floats present with precision higher than
15 digits so if there are, your results may not be what you expect due to round-o errors.
Examples
>>> from sympy import sin
>>> from sympy.abc import x
>>> from sympy.utilities.randtest import test_derivative_numerically as td
>>> td(sin(x), x)
True
5.33. Utilities
1579
SymPy Documentation, Release 0.7.6
sympy.utilities.randtest.verify numerically(f, g, z=None, tol=1e-06, a=2, b=-1,
c=3, d=1)
Test numerically that f and g agree when evaluated in the argument z.
If z is None, all symbols will be tested. This routine does not test whether there are
Floats present with precision higher than 15 digits so if there are, your results may not
be what you expect due to round- o errors.
Examples
>>> from sympy import sin, cos
>>> from sympy.abc import x
>>> from sympy.utilities.randtest import verify_numerically as tn
>>> tn(sin(x)**2 + cos(x)**2, 1, x)
True
5.33.12 Run Tests
This is our testing framework.
Goals:
it should be compatible with py.test and operate very similarly (or identically)
doesnt require any external dependencies
preferably all the functionality should be in this le only
no magic, just import the test le and execute the test functions, thats it
portable
class sympy.utilities.runtests.PyTestReporter(verbose=False, tb=short, colors=True,
force colors=False,
split=None)
Py.test like reporter. Should produce output identical to py.test.
write(text, color=, align=left, width=None, force colors=False)
Prints a text on the screen.
It uses sys.stdout.write(), so no readline library is necessary.
Parameters color : choose from the colors below, means default color
align : left/right, left is a normal print, right is aligned on
the right-hand side of the screen, lled with spaces if necessary
width : the screen width
class sympy.utilities.runtests.Reporter
Parent class for all reporters.
class sympy.utilities.runtests.SymPyDocTestFinder(verbose=False,
parser=<doctest.DocTestParser
instance
at
0x9ccba2c>,
recurse=True,
exclude empty=True)
A class used to extract the DocTests that are relevant to a given object, from its docstring and the docstrings of its contained objects. Doctests can currently be extracted
1580
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
from the following object types: modules, functions, classes, methods, staticmethods,
classmethods, and properties.
Modied from doctests version by looking harder for code in the case that it looks like
the the code comes from a dierent module. In the case of decorated functions (e.g.
@vectorize) they appear to come from a dierent module (e.g. multidemensional) even
though their code is not there.
class sympy.utilities.runtests.SymPyDocTestRunner(checker=None,
verbose=None, optionags=0)
A class used to run DocTest test cases, and accumulate statistics. The run method is
used to process a single DocTest case. It returns a tuple (f, t), where t is the number
of test cases tried, and f is the number of test cases that failed.
Modied from the doctest version to not reset the sys.displayhook (see issue 5140).
See the docstring of the original DocTestRunner for more information.
run(test, compileags=None, out=None, clear globs=True)
Run the examples in test, and display the results using the writer function out.
The examples are run in the namespace test.globs. If clear globs is true (the
default), then this namespace will be cleared after the test runs, to help with garbage
collection. If you would like to examine the namespace after the test completes, then
use clear globs=False.
compileflags gives the set of ags that should be used by the Python compiler when
running the examples. If not specied, then it will default to the set of future-import
ags that apply to globs.
The output of each example is checked using SymPyDocTestRunner.check output,
and the results are formatted by the SymPyDocTestRunner.report * methods.
class sympy.utilities.runtests.SymPyOutputChecker
Compared to the OutputChecker from the stdlib our OutputChecker class supports numerical comparison of oats occuring in the output of the doctest examples
check output(want, got, optionags)
Return True i the actual output from an example (got) matches the expected output (want). These strings are always considered to match if they are identical; but
depending on what option ags the test runner is using, several non-exact match
types are also possible. See the documentation for T estRunner for more information
about option ags.
sympy.utilities.runtests.SymPyTestResults
alias of TestResults
sympy.utilities.runtests.convert to native paths(lst)
Converts a list of / separated paths into a list of native (os.sep separated) paths and
converts to lowercase if the system is case insensitive.
sympy.utilities.runtests.doctest(*paths, **kwargs)
Runs doctests in all *.py les in the sympy directory which match any of the given strings
in paths or all tests if paths=[].
Notes:
Paths can be entered in native system format or in unix, forward-slash format.
Files that are on the blacklist can be tested by providing their path; they are only
excluded if no paths are given.
5.33. Utilities
1581
SymPy Documentation, Release 0.7.6
Examples
>>> import sympy
Run all tests:
>>> sympy.doctest()
Run one le:
>>> sympy.doctest(sympy/core/basic.py)
>>> sympy.doctest(polynomial.rst)
Run all tests in sympy/functions/ and some particular le:
>>> sympy.doctest(/functions, basic.py)
Run any le having polynomial in its name, doc/src/modules/polynomial.rst,
sympy/functions/special/polynomials.py, and sympy/polys/polynomial.py:
>>> sympy.doctest(polynomial)
The split option can be passed to split the test run into parts. The split currently only
splits the test les, though this may change in the future. split should be a string of
the form a/b, which will run part a of b. Note that the regular doctests and the Sphinx
doctests are split independently. For instance, to run the rst half of the test suite:
>>> sympy.doctest(split=1/2)
The subprocess and verbose options are the same as with the function test(). See the
docstring of that function for more information.
sympy.utilities.runtests.get sympy dir()
Returns the root sympy directory and set the global value indicating whether the system
is case sensitive or not.
sympy.utilities.runtests.run all tests(test args=(),
test kwargs={},
doctest args=(),
doctest kwargs={},
examples args=(),
examples kwargs={quiet: True})
Run all tests.
Right now, this runs the regular tests (bin/test), the doctests (bin/doctest), the examples
(examples/all.py), and the sage tests (see sympy/external/tests/test sage.py).
This is what setup.py test uses.
You can pass arguments and keyword arguments to the test functions that support them
(for now, test, doctest, and the examples). See the docstrings of those functions for a
description of the available options.
For example, to run the solvers tests with colors turned o:
>>> from sympy.utilities.runtests import run_all_tests
>>> run_all_tests(test_args=(solvers,),
... test_kwargs={colors:False})
1582
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
sympy.utilities.runtests.run in subprocess with hash randomization(function,
function args=(),
function kwargs={},
command=/home/vagrant/doc
virtualenv/bin/python,
module=sympy.utilities.runtests,
force=False)
Run a function in a Python subprocess with hash randomization enabled.
If hash randomization is not supported by the version of Python given, it returns False.
Otherwise, it returns the exit value of the command. The function is passed to sys.exit(),
so the return value of the function will be the return value.
The environment variable PYTHONHASHSEED is used to seed Pythons hash randomization. If it is set, this function will return False, because starting a new subprocess
is unnecessary in that case. If it is not set, one is set at random, and the tests are run.
Note that if this environment variable is set when Python starts, hash randomization is
automatically enabled. To force a subprocess to be created even if PYTHONHASHSEED
is set, pass force=True. This ag will not force a subprocess in Python versions that do
not support hash randomization (see below), because those versions of Python do not
support the -R ag.
function should be a string name of a function that is importable from the module module, like test. The default for module is sympy.utilities.runtests. function args and
function kwargs should be a repr-able tuple and dict, respectively. The default Python
command is sys.executable, which is the currently running Python command.
This function is necessary because the seed for hash randomization must be set by the
environment variable before Python starts. Hence, in order to use a predetermined seed
for tests, we must start Python in a separate subprocess.
Hash randomization was added in the minor Python versions 2.6.8, 2.7.3, 3.1.5, and
3.2.3, and is enabled by default in all Python versions after and including 3.3.0.
Examples
>>> from sympy.utilities.runtests import (
... run_in_subprocess_with_hash_randomization)
>>> # run the core tests in verbose mode
>>> run_in_subprocess_with_hash_randomization(_test,
... function_args=(core,),
... function_kwargs={verbose: True})
# Will return 0 if sys.executable supports hash randomization and tests
# pass, 1 if they fail, and False if it does not support hash
# randomization.
sympy.utilities.runtests.split list(l, split)
Splits a list into part a of b
split should be a string of the form a/b. For instance, 1/3 would give the split one of
three.
If the length of the list is not divisible by the number of splits, the last split will have
more items.
5.33. Utilities
1583
SymPy Documentation, Release 0.7.6
>>>
>>>
>>>
[0,
>>>
[3,
>>>
[6,
from sympy.utilities.runtests import split_list
a = list(range(10))
split_list(a, 1/3)
1, 2]
split_list(a, 2/3)
4, 5]
split_list(a, 3/3)
7, 8, 9]
sympy.utilities.runtests.sympytestfile(lename,
module relative=True,
name=None,
package=None,
globs=None,
verbose=None,
report=True,
optionags=0,
extraglobs=None,
raise on error=False,
parser=<doctest.DocTestParser instance
at 0x9c7360c>, encoding=None)
Test examples in the given le. Return (#failures, #tests).
Optional keyword arg module relative species how lenames should be interpreted:
If module relative is True (the default), then filename species a module-relative
path. By default, this path is relative to the calling modules directory; but if the
package argument is specied, then it is relative to that package. To ensure osindependence, filename should use / characters to separate path segments, and
should not be an absolute path (i.e., it may not begin with /).
If module relative is False, then filename species an os-specic path. The path
may be absolute or relative (to the current working directory).
Optional keyword arg name gives the name of the test; by default use the les basename.
Optional keyword argument package is a Python package or the name of a Python package whose directory should be used as the base directory for a module relative lename.
If no package is specied, then the calling modules directory is used as the base directory for module relative lenames. It is an error to specify package if module relative
is False.
Optional keyword arg globs gives a dict to be used as the globals when executing examples; by default, use {}. A copy of this dict is actually used for each docstring, so that
each docstrings examples start with a clean slate.
Optional keyword arg extraglobs gives a dictionary that should be merged into the
globals that are used to execute examples. By default, no extra globals are used.
Optional keyword arg verbose prints lots of stu if true, prints only failures if false; by
default, its true i -v is in sys.argv.
Optional keyword arg report prints a summary at the end when true, else prints nothing
at the end. In verbose mode, the summary is detailed, else very brief (in fact, empty if
all tests passed).
Optional keyword arg optionflags ors together module constants, and defaults to 0.
Possible values (see the docs for details):
DONT ACCEPT TRUE FOR 1
DONT ACCEPT BLANKLINE
NORMALIZE WHITESPACE
ELLIPSIS
SKIP
1584
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
IGNORE EXCEPTION DETAIL
REPORT UDIFF
REPORT CDIFF
REPORT NDIFF
REPORT ONLY FIRST FAILURE
Optional keyword arg raise on error raises an exception on the rst unexpected exception or failure. This allows failures to be post-mortem debugged.
Optional keyword arg parser species a DocTestParser (or subclass) that should be used
to extract tests from the les.
Optional keyword arg encoding species an encoding that should be used to convert the
le to unicode.
Advanced tomfoolery: testmod runs methods of a local instance of class doctest.Tester,
then merges the results into (or creates) global Tester instance doctest.master. Methods
of doctest.master can be called directly too, if you want to do something unusual. Passing
report=0 to testmod is especially useful then, to delay displaying a summary. Invoke
doctest.master.summarize(verbose) when youre done ddling.
sympy.utilities.runtests.test(*paths, **kwargs)
Run tests in the specied test *.py les.
Tests in a particular test *.py le are run if any of the given strings in paths matches a
part of the test les path. If paths=[], tests in all test *.py les are run.
Notes:
If sort=False, tests are run in random order (not default).
Paths can be entered in native system format or in unix, forward-slash format.
Files that are on the blacklist can be tested by providing their path; they are only
excluded if no paths are given.
Explanation of test results
Output
.
F
X
f
s
w
T
K
Meaning
passed
failed
XPassed (expected to fail but passed)
XFAILed (expected to fail and indeed failed)
skipped
slow
timeout (e.g., when --timeout is used)
KeyboardInterrupt (when running the slow tests with --slow, you can
interrupt one of them without killing the test runner)
Colors have no additional meaning and are used just to facilitate interpreting the output.
Examples
>>> import sympy
Run all tests:
5.33. Utilities
1585
SymPy Documentation, Release 0.7.6
>>> sympy.test()
Run one le:
>>> sympy.test(sympy/core/tests/test_basic.py)
>>> sympy.test(_basic)
Run all tests in sympy/functions/ and some particular le:
>>> sympy.test(sympy/core/tests/test_basic.py,
...
sympy/functions)
Run all tests in sympy/core and sympy/utilities:
>>> sympy.test(/core, /util)
Run specic test from a le:
>>> sympy.test(sympy/core/tests/test_basic.py,
...
kw=test_equality)
Run specic test from any le:
>>> sympy.test(kw=subs)
Run the tests with verbose mode on:
>>> sympy.test(verbose=True)
Dont sort the test output:
>>> sympy.test(sort=False)
Turn on post-mortem pdb:
>>> sympy.test(pdb=True)
Turn o colors:
>>> sympy.test(colors=False)
Force colors, even when the output is not to a terminal (this is useful, e.g., if you are
piping to less -r and you still want colors)
>>> sympy.test(force_colors=False)
The traceback verboseness can be set to short or no (default is short)
>>> sympy.test(tb=no)
The split option can be passed to split the test run into parts. The split currently only
splits the test les, though this may change in the future. split should be a string of the
form a/b, which will run part a of b. For instance, to run the rst half of the test suite:
>>> sympy.test(split=1/2)
You can disable running the tests in a separate subprocess using subprocess=False.
This is done to support seeding hash randomization, which is enabled by default in the
Python versions where it is supported. If subprocess=False, hash randomization is enabled/disabled according to whether it has been enabled or not in the calling Python
1586
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
process. However, even if it is enabled, the seed cannot be printed unless it is called
from a new Python process.
Hash randomization was added in the minor Python versions 2.6.8, 2.7.3, 3.1.5, and
3.2.3, and is enabled by default in all Python versions after and including 3.3.0.
If hash randomization is not supported subprocess=False is used automatically.
>>> sympy.test(subprocess=False)
To set the hash randomization seed, set the environment variable PYTHONHASHSEED before
running the tests. This can be done from within Python using
>>> import os
>>> os.environ[PYTHONHASHSEED] = 42
Or from the command line using
$ PYTHONHASHSEED=42 ./bin/test
If the seed is not set, a random seed will be chosen.
Note that to reproduce the same hash values, you must use both the same seed as well
as the same architecture (32-bit vs. 64-bit).
5.33.13 Source Code Inspection
This module adds several functions for interactive source code inspection.
sympy.utilities.source.get class(lookup view)
Convert a string version of a class name to the object.
For example, get class(sympy.core.Basic) will return class Basic located in module
sympy.core
sympy.utilities.source.get mod func(callback)
splits the string path to a class into a string path to the module and the name of the class.
For example:
>>> from sympy.utilities.source import get_mod_func
>>> get_mod_func(sympy.core.basic.Basic)
(sympy.core.basic, Basic)
sympy.utilities.source.source(object)
Prints the source code of a given object.
5.33.14 Timing Utilities
Simple tools for timing functions execution, when IPython is not available.
sympy.utilities.timeutils.timed(func, setup=pass, limit=None)
Adaptively measure execution time of a function.
5.33. Utilities
1587
SymPy Documentation, Release 0.7.6
5.34 Parsing input
5.34.1 Parsing Functions Reference
sympy.parsing.sympy parser.parse expr(s,
local dict=None,
transformations=(<function
lambda notation
at
0x10ad264c>,
<function
auto symbol
at 0x10ad2614>, <function auto number at
0x10ad26f4>, <function factorial notation
at 0x10ad2684>), global dict=None, evaluate=True)
Converts the string s to a SymPy expression, in local dict
Parameters s : str
The string to parse.
local dict : dict, optional
A dictionary of local variables to use when parsing.
global dict : dict, optional
A dictionary of global variables. By default, this is initialized with
from sympy import *; provide this parameter to override this behavior (for instance, to parse Q & S).
transformations : tuple, optional
A tuple of transformation functions used to modify the tokens of the
parsed expression before evaluation. The default transformations
convert numeric literals into their SymPy equivalents, convert undened variables into SymPy symbols, and allow the use of standard
mathematical factorial notation (e.g. x!).
evaluate : bool, optional
When False, the order of the arguments will remain as they were in
the string and automatic simplication that would normally occur is
suppressed. (see examples)
See Also:
stringify expr,
eval expr,
plicit multiplication application
standard transformations,
im-
Examples
>>> from sympy.parsing.sympy_parser import parse_expr
>>> parse_expr(1/2)
1/2
>>> type(_)
<class sympy.core.numbers.Half>
>>> from sympy.parsing.sympy_parser import standard_transformations,\
... implicit_multiplication_application
>>> transformations = (standard_transformations +
...
(implicit_multiplication_application,))
>>> parse_expr(2x, transformations=transformations)
2*x
1588
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
When evaluate=False, some automatic simplications will not occur:
>>> parse_expr(2**3), parse_expr(2**3, evaluate=False)
(8, 2**3)
In addition the order of the arguments will not be made canonical. This feature allows
one to tell exactly how the expression was entered:
>>> a = parse_expr(1 + x, evaluate=False)
>>> b = parse_expr(x + 1, evaluate=0)
>>> a == b
False
>>> a.args
(1, x)
>>> b.args
(x, 1)
sympy.parsing.sympy parser.stringify expr(s, local dict, global dict, transformations)
Converts the string s to Python code, in local dict
Generally, parse expr should be used.
sympy.parsing.sympy parser.eval expr(code, local dict, global dict)
Evaluate Python code generated by stringify expr.
Generally, parse expr should be used.
sympy.parsing.sympy tokenize.printtoken(type, token, srow scol, erow ecol, line)
sympy.parsing.sympy tokenize.tokenize(readline, tokeneater=<function printtoken
at 0x11000c6c>)
The tokenize() function accepts two parameters: one representing the input stream, and
one providing an output mechanism for tokenize().
The rst parameter, readline, must be a callable object which provides the same interface
as the readline() method of built-in le objects. Each call to the function should return
one line of input as a string.
The second parameter, tokeneater, must also be a callable object. It is called once
for each token, with ve arguments, corresponding to the tuples generated by generate tokens().
sympy.parsing.sympy tokenize.untokenize(iterable)
Transform tokens back into Python source code.
Each element returned by the iterable must be a token sequence with at least two elements, a token number and token value. If only two tokens are passed, the resulting
output is poor.
Round-trip invariant for full input: Untokenized source will match input source exactly
Round-trip invariant for limited intput:
# Output text will tokenize the back to the input
t1 = [tok[:2] for tok in generate_tokens(f.readline)]
newcode = untokenize(t1)
readline = iter(newcode.splitlines(1)).next
t2 = [tok[:2] for tok in generate_tokens(readline)]
if t1 != t2:
raise ValueError(t1 should be equal to t2)
5.34. Parsing input
1589
SymPy Documentation, Release 0.7.6
sympy.parsing.sympy tokenize.generate tokens(readline)
The generate tokens() generator requires one argment, readline, which must be a
callable object which provides the same interface as the readline() method of built-in
le objects. Each call to the function should return one line of input as a string. Alternately, readline can be a callable function terminating with StopIteration:
readline = open(myfile).next
# Example of alternate readline
The generator produces 5-tuples with these members: the token type; the token string;
a 2-tuple (srow, scol) of ints specifying the row and column where the token begins in
the source; a 2-tuple (erow, ecol) of ints specifying the row and column where the token
ends in the source; and the line on which the token was found. The line passed is the
logical line; continuation lines are included.
sympy.parsing.sympy tokenize.group(*choices)
sympy.parsing.sympy tokenize.any(*choices)
sympy.parsing.sympy tokenize.maybe(*choices)
sympy.parsing.maxima.parse maxima(str, globals=None, name dict={})
sympy.parsing.mathematica.mathematica(s)
5.34.2 Parsing Exceptions Reference
class sympy.parsing.sympy tokenize.TokenError
class sympy.parsing.sympy tokenize.StopTokenizing
5.34.3 Parsing Transformations Reference
A transformation is a function that accepts the arguments tokens, local dict, global dict
and returns a list of transformed tokens. They can be used by passing a list of functions to
parse expr() and are applied in the order given.
sympy.parsing.sympy parser.standard transformations = (<function lambda notation at 0x10ad26
Standard transformations for parse expr(). Inserts calls to Symbol, Integer, and other
SymPy datatypes and allows the use of standard factorial notation (e.g. x!).
sympy.parsing.sympy parser.split symbols(tokens, local dict, global dict)
Splits symbol names for implicit multiplication.
Intended to let expressions like xyz be parsed as x*y*z. Does not split Greek character names, so theta will not become t*h*e*t*a. Generally this should be used with
implicit multiplication.
sympy.parsing.sympy parser.split symbols custom(predicate)
Creates a transformation that splits symbol names.
predicate should return True if the symbol name is to be split.
For instance, to retain the default behavior but avoid splitting certain symbol names, a
predicate like this would work:
>>>
...
...
>>>
...
1590
from sympy.parsing.sympy_parser import (parse_expr, _token_splittable,
standard_transformations, implicit_multiplication,
split_symbols_custom)
def can_split(symbol):
if symbol not in (list, of, unsplittable, names):
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
...
return _token_splittable(symbol)
...
return False
...
>>> transformation = split_symbols_custom(can_split)
>>> parse_expr(unsplittable, transformations=standard_transformations +
... (transformation, implicit_multiplication))
unsplittable
sympy.parsing.sympy parser.implicit multiplication(result,
global dict)
Makes the multiplication operator optional in most cases.
local dict,
Use this before implicit application(), otherwise expressions like sin 2x will be
parsed as x * sin(2) rather than sin(2*x).
Example:
>>> from sympy.parsing.sympy_parser import (parse_expr,
... standard_transformations, implicit_multiplication)
>>> transformations = standard_transformations + (implicit_multiplication,)
>>> parse_expr(3 x y, transformations=transformations)
3*x*y
sympy.parsing.sympy parser.implicit application(result, local dict, global dict)
Makes parentheses optional in some cases for function calls.
Use this after implicit multiplication(), otherwise expressions like sin 2x will be
parsed as x * sin(2) rather than sin(2*x).
Example:
>>> from sympy.parsing.sympy_parser import (parse_expr,
... standard_transformations, implicit_application)
>>> transformations = standard_transformations + (implicit_application,)
>>> parse_expr(cot z + csc z, transformations=transformations)
cot(z) + csc(z)
sympy.parsing.sympy parser.function exponentiation(tokens,
global dict)
Allows functions to be exponentiated, e.g. cos**2(x).
local dict,
Example:
>>> from sympy.parsing.sympy_parser import (parse_expr,
... standard_transformations, function_exponentiation)
>>> transformations = standard_transformations + (function_exponentiation,)
>>> parse_expr(sin**4(x), transformations=transformations)
sin(x)**4
sympy.parsing.sympy parser.implicit multiplication application(result,
local dict,
global dict)
Allows a slightly relaxed syntax.
Parentheses for single-argument method calls are optional.
Multiplication is implicit.
Symbol names can be split (i.e. spaces are not needed between symbols).
Functions can be exponentiated.
Example:
5.34. Parsing input
1591
SymPy Documentation, Release 0.7.6
>>> from sympy.parsing.sympy_parser import (parse_expr,
... standard_transformations, implicit_multiplication_application)
>>> parse_expr(10sin**2 x**2 + 3xyz + tan theta,
... transformations=(standard_transformations +
... (implicit_multiplication_application,)))
3*x*y*z + 10*sin(x**2)**2 + tan(theta)
sympy.parsing.sympy parser.rationalize(tokens, local dict, global dict)
Converts oats into Rational. Run AFTER auto number.
sympy.parsing.sympy parser.convert xor(tokens, local dict, global dict)
Treats XOR, , as exponentiation, **.
These are included in :data:sympy.parsing.sympy parser.standard transformations and
generally dont need to be manually added by the user.
sympy.parsing.sympy parser.factorial notation(tokens, local dict, global dict)
Allows standard notation for factorial.
sympy.parsing.sympy parser.auto symbol(tokens, local dict, global dict)
Inserts calls to Symbol for undened variables.
sympy.parsing.sympy parser.auto number(tokens, local dict, global dict)
Converts numeric literals to use SymPy equivalents.
Complex numbers use I; integer literals use Integer, oat literals use Float, and repeating decimals use Rational.
5.35 Calculus
Some calculus-related methods waiting to nd a better place in the SymPy modules tree.
sympy.calculus.euler.euler equations(L, funcs=(), vars=())
Find the Euler-Lagrange equations [R1] (page 1915) for a given Lagrangian.
Parameters L : Expr
The Lagrangian that should be a function of the functions listed in
the second argument and their derivatives.
For example, in the case of two functions f (x, y), g(x, y) and two independent variables x, y the Lagrangian would have the form:
(
)
f (x, y) f (x, y) g(x, y) g(x, y)
L f (x, y), g(x, y),
,
,
,
, x, y
x
y
x
y
In many cases it is not necessary to provide anything, except the Lagrangian, it will be autodetected (and an error raised if this couldnt
be done).
funcs : Function or an iterable of Functions
The functions that the Lagrangian depends on. The Euler equations
are dierential equations for each of these functions.
vars : Symbol or an iterable of Symbols
The Symbols that are the independent variables of the functions.
1592
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Returns eqns : list of Eq
The list of dierential equations, one for each function.
References
[R1] (page 1915)
Examples
>>> from sympy import Symbol, Function
>>> from sympy.calculus.euler import euler_equations
>>> x = Function(x)
>>> t = Symbol(t)
>>> L = (x(t).diff(t))**2/2 - x(t)**2/2
>>> euler_equations(L, x(t), t)
[-x(t) - Derivative(x(t), t, t) == 0]
>>> u = Function(u)
>>> x = Symbol(x)
>>> L = (u(t, x).diff(t))**2/2 - (u(t, x).diff(x))**2/2
>>> euler_equations(L, u(t, x), [t, x])
[-Derivative(u(t, x), t, t) + Derivative(u(t, x), x, x) == 0]
sympy.calculus.singularities.singularities(expr, sym)
Finds singularities for a function. Currently supported functions are: - univariate real
rational functions
References
[R2] (page 1915)
Examples
>>> from sympy.calculus.singularities import singularities
>>> from sympy import Symbol
>>> x = Symbol(x, real=True)
>>> singularities(x**2 + x + 1, x)
()
>>> singularities(1/(x + 1), x)
(-1,)
5.35.1 Finite dierence weights
This module implements an algorithm for ecient generation of nite dierence weights for
ordinary dierentials of functions for derivatives from 0 (interpolation) up to arbitrary order.
The core algorithm is provided in the nite dierence weight generating function (nite di weights), and two convenience functions are provided for:
estimating a derivative (or interpolate) directly from a series of points is
provided (apply finite diff).
5.35. Calculus
also
1593
SymPy Documentation, Release 0.7.6
making a nite dierence approximation of a Derivative instance
(as finite diff).
sympy.calculus.finite diff.apply finite diff(order, x list, y list, x0)
Calculates the nite dierence approximation of the derivative of requested order at x0
from points provided in x list and y list.
Parameters order: int :
order of derivative to approximate. 0 corresponds to interpolation.
x list: sequence :
Strictly monotonically increasing sequence of values for the independent variable.
y list: sequence :
The function value at corresponding values for the independent variable in x list.
x0: Number or Symbol :
At what value of the independent variable the derivative should be
evaluated.
Returns sympy.core.add.Add or sympy.core.numbers.Number :
The nite dierence expression approximating the requested derivative order at x0.
See Also:
sympy.calculus.finite diff.finite diff weights (page 1596)
Notes
Order = 0 corresponds to interpolation. Only supply so many points you think makes
sense to around x0 when extracting the derivative (the function need to be well behaved
within that region). Also beware of Runges phenomenon.
References
Fortran 90 implementation with Python interface for numerics: nitedi
Examples
>>> from sympy.calculus import apply_finite_diff
>>> cube = lambda arg: (1.0*arg)**3
>>> xlist = range(-3,3+1)
>>> apply_finite_diff(2, xlist, map(cube, xlist), 2) - 12
-3.55271367880050e-15
we see that the example above only contain rounding errors. apply nite di can also
be used on more abstract objects:
1594
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>>
>>>
>>>
>>>
>>>
>>>
(-1
from sympy import IndexedBase, Idx
from sympy.calculus import apply_finite_diff
x, y = map(IndexedBase, xy)
i = Idx(i)
x_list, y_list = zip(*[(x[i+j], y[i+j]) for j in range(-1,2)])
apply_finite_diff(1, x_list, y_list, x[i])
+ (x[i + 1] - x[i])/(-x[i - 1] + x[i]))*y[i]/(x[i + 1] - x[i]) + (-x[i - 1] + x[i])*y[i + 1]
sympy.calculus.finite diff.as finite diff(derivative,
points=1,
x0=None,
wrt=None)
Returns an approximation of a derivative of a function in the form of a nite dierence
formula. The expression is a weighted sum of the function at a number of discrete values
of (one of) the independent variable(s).
Parameters derivative: a Derivative instance (needs to have an variables :
and expr attribute).
points: sequence or coecient, optional :
If sequence: discrete values (length >= order+1) of the independent
variable used for generating the nite dierence weights. If it is a coecient, it will be used as the step-size for generating an equidistant
sequence of length order+1 centered around x0. default: 1 (step-size
1)
x0: number or Symbol, optional :
the value of the independent variable (wrt) at which the derivative is
to be approximated. default: same as wrt
wrt: Symbol, optional :
with respect to the variable for which the (partial) derivative is to
be approximated for. If not provided it is required that the Derivative
is ordinary. default: None
See Also:
sympy.calculus.finite diff.apply finite diff
(page
sympy.calculus.finite diff.finite diff weights (page 1596)
1594),
Examples
>>> from sympy import symbols, Function, exp, sqrt, Symbol, as_finite_diff
>>> x, h = symbols(x h)
>>> f = Function(f)
>>> as_finite_diff(f(x).diff(x))
-f(x - 1/2) + f(x + 1/2)
The default step size and number of points are 1 and order + 1 respectively. We can
change the step size by passing a symbol as a parameter:
>>> as_finite_diff(f(x).diff(x), h)
-f(-h/2 + x)/h + f(h/2 + x)/h
We can also specify the discretized values to be used in a sequence:
5.35. Calculus
1595
SymPy Documentation, Release 0.7.6
>>> as_finite_diff(f(x).diff(x), [x, x+h, x+2*h])
-3*f(x)/(2*h) + 2*f(h + x)/h - f(2*h + x)/(2*h)
The algorithm is not restricted to use equidistant spacing, nor do we need to make the
approximation around x0, but we can get an expression estimating the derivative at an
oset:
>>> e, sq2 = exp(1), sqrt(2)
>>> xl = [x-h, x+h, x+e*h]
>>> as_finite_diff(f(x).diff(x, 1), xl, x+h*sq2)
2*h*((h + sqrt(2)*h)/(2*h) - (-sqrt(2)*h + h)/(2*h))*f(E*h + x)/((-h + E*h)*(h + E*h)) + (-(-sqr
Partial derivatives are also supported:
>>> y = Symbol(y)
>>> d2fdxdy=f(x,y).diff(x,y)
>>> as_finite_diff(d2fdxdy, wrt=x)
-f(x - 1/2, y) + f(x + 1/2, y)
sympy.calculus.finite diff.finite diff weights(order, x list, x0)
Calculates the nite dierence weights for an arbitrarily spaced one-dimensional grid
(x list) for derivatives at x0 of order 0, 1, ..., up to order using a recursive formula.
Parameters order : int
Up to what derivative order weights should be calculated. 0 corresponds to interpolation.
x list: sequence :
Strictly monotonically increasing sequence of values for the independent variable.
x0: Number or Symbol :
At what value of the independent variable the nite dierence
weights should be generated.
Returns list :
A list of sublists, each corresponding to coecients for increasing
derivative order, and each containing lists of coecients for increasing accuracy.
See Also:
sympy.calculus.finite diff.apply finite diff (page 1594)
Notes
If weights for a nite dierence approximation of the 3rd order derivative is wanted,
weights for 0th, 1st and 2nd order are calculated for free, so are formulae using fewer
and fewer of the parameters. This is something one can take advantage of to save computational cost.
References
[R3] (page 1915)
1596
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import S
>>> from sympy.calculus import finite_diff_weights
>>> finite_diff_weights(1, [-S(1)/2, S(1)/2, S(3)/2, S(5)/2], 0)
[[[1, 0, 0, 0],
[1/2, 1/2, 0, 0],
[3/8, 3/4, -1/8, 0],
[5/16, 15/16, -5/16, 1/16]],
[[0, 0, 0, 0], [-1, 1, 0, 0], [-1, 1, 0, 0], [-23/24, 7/8, 1/8, -1/24]]]
the result is two subslists, the rst is for the 0:th derivative (interpolation) and the second
for the rst derivative (we gave 1 as the parameter of order so this is why we get no list
for a higher order derivative). Each sublist contains the most accurate formula in the
end (all points used).
Beware of the oset in the lower accuracy formulae when looking at a centered dierence:
>>> from sympy import S
>>> from sympy.calculus import finite_diff_weights
>>> finite_diff_weights(1, [-S(5)/2, -S(3)/2, -S(1)/2, S(1)/2,
...
S(3)/2, S(5)/2], 0)
[[[1, 0, 0, 0, 0, 0],
[-3/2, 5/2, 0, 0, 0, 0],
[3/8, -5/4, 15/8, 0, 0, 0],
[1/16, -5/16, 15/16, 5/16, 0, 0],
[3/128, -5/32, 45/64, 15/32, -5/128, 0],
[3/256, -25/256, 75/128, 75/128, -25/256, 3/256]],
[[0, 0, 0, 0, 0, 0],
[-1, 1, 0, 0, 0, 0],
[1, -3, 2, 0, 0, 0],
[1/24, -1/8, -7/8, 23/24, 0, 0],
[0, 1/24, -9/8, 9/8, -1/24, 0],
[-3/640, 25/384, -75/64, 75/64, -25/384, 3/640]]]
The capability to generate weights at arbitrary points can be used e.g. to minimize
Runges phenomenon by using Chebyshev nodes:
>>> from sympy import cos, symbols, pi, simplify
>>> from sympy.calculus import finite_diff_weights
>>> N, (h, x) = 4, symbols(h x)
>>> x_list = [x+h*cos(i*pi/(N)) for i in range(N,-1,-1)] # chebyshev nodes
>>> print(x_list)
[-h + x, -sqrt(2)*h/2 + x, x, sqrt(2)*h/2 + x, h + x]
>>> mycoeffs = finite_diff_weights(1, x_list, 0)[1][4]
>>> [simplify(c) for c in mycoeffs]
[(h**3/2 + h**2*x - 3*h*x**2 - 4*x**3)/h**4,
(-sqrt(2)*h**3 - 4*h**2*x + 3*sqrt(2)*h*x**2 + 8*x**3)/h**4,
6*x/h**2 - 8*x**3/h**4,
(sqrt(2)*h**3 - 4*h**2*x - 3*sqrt(2)*h*x**2 + 8*x**3)/h**4,
(-h**3/2 + h**2*x + 3*h*x**2 - 4*x**3)/h**4]
5.36 Physics Module
A module that helps solving problems in physics
5.36. Physics Module
1597
SymPy Documentation, Release 0.7.6
5.36.1 Contents
Hydrogen Wavefunctions
sympy.physics.hydrogen.E nl(n, Z=1)
Returns the energy of the state (n, l) in Hartree atomic units.
The energy doesnt depend on l.
Examples
>>> from sympy import var
>>> from sympy.physics.hydrogen import E_nl
>>> var(n Z)
(n, Z)
>>> E_nl(n, Z)
-Z**2/(2*n**2)
>>> E_nl(1)
-1/2
>>> E_nl(2)
-1/8
>>> E_nl(3)
-1/18
>>> E_nl(3, 47)
-2209/18
sympy.physics.hydrogen.E nl dirac(n, l, spin up=True, Z=1, c=137.035999037000)
Returns the relativistic energy of the state (n, l, spin) in Hartree atomic units.
The energy is calculated from the Dirac equation. The rest mass energy is not included.
n, l quantum numbers n and l
spin up True if the electron spin is up (default), otherwise down
Z atomic number (1 for Hydrogen, 2 for Helium, ...)
c speed of light in atomic units.
http://arxiv.org/abs/1012.3627
Default value is 137.035999037, taken from:
Examples
>>> from sympy.physics.hydrogen import E_nl_dirac
>>> E_nl_dirac(1, 0)
-0.500006656595360
>>> E_nl_dirac(2, 0)
-0.125002080189006
>>> E_nl_dirac(2, 1)
-0.125000416028342
>>> E_nl_dirac(2, 1, False)
-0.125002080189006
>>> E_nl_dirac(3, 0)
-0.0555562951740285
>>> E_nl_dirac(3, 1)
-0.0555558020932949
1598
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> E_nl_dirac(3, 1, False)
-0.0555562951740285
>>> E_nl_dirac(3, 2)
-0.0555556377366884
>>> E_nl_dirac(3, 2, False)
-0.0555558020932949
sympy.physics.hydrogen.R nl(n, l, r, Z=1)
Returns the Hydrogen radial wavefunction R {nl}.
n, l quantum numbers n and l
r radial coordinate
Z atomic number (1 for Hydrogen, 2 for Helium, ...)
Everything is in Hartree atomic units.
Examples
>>> from sympy.physics.hydrogen import R_nl
>>> from sympy import var
>>> var(r Z)
(r, Z)
>>> R_nl(1, 0, r, Z)
2*sqrt(Z**3)*exp(-Z*r)
>>> R_nl(2, 0, r, Z)
sqrt(2)*(-Z*r + 2)*sqrt(Z**3)*exp(-Z*r/2)/4
>>> R_nl(2, 1, r, Z)
sqrt(6)*Z*r*sqrt(Z**3)*exp(-Z*r/2)/12
For Hydrogen atom, you can just use the default value of Z=1:
>>> R_nl(1, 0, r)
2*exp(-r)
>>> R_nl(2, 0, r)
sqrt(2)*(-r + 2)*exp(-r/2)/4
>>> R_nl(3, 0, r)
2*sqrt(3)*(2*r**2/9 - 2*r + 3)*exp(-r/3)/27
For Silver atom, you would use Z=47:
>>> R_nl(1, 0, r, Z=47)
94*sqrt(47)*exp(-47*r)
>>> R_nl(2, 0, r, Z=47)
47*sqrt(94)*(-47*r + 2)*exp(-47*r/2)/4
>>> R_nl(3, 0, r, Z=47)
94*sqrt(141)*(4418*r**2/9 - 94*r + 3)*exp(-47*r/3)/27
The normalization of the radial wavefunction is:
>>>
>>>
1
>>>
1
>>>
1
from sympy import integrate, oo
integrate(R_nl(1, 0, r)**2 * r**2, (r, 0, oo))
integrate(R_nl(2, 0, r)**2 * r**2, (r, 0, oo))
integrate(R_nl(2, 1, r)**2 * r**2, (r, 0, oo))
5.36. Physics Module
1599
SymPy Documentation, Release 0.7.6
It holds for any atomic number:
>>> integrate(R_nl(1, 0, r, Z=2)**2 * r**2, (r, 0, oo))
1
>>> integrate(R_nl(2, 0, r, Z=3)**2 * r**2, (r, 0, oo))
1
>>> integrate(R_nl(2, 1, r, Z=4)**2 * r**2, (r, 0, oo))
1
Matrices
Known matrices related to physics
sympy.physics.matrices.mdft(n)
Returns an expression of a discrete Fourier transform as a matrix multiplication. It is an
n X n matrix.
References
[R320] (page 1915)
Examples
>>> from sympy.physics.matrices import mdft
>>> mdft(3)
Matrix([
[sqrt(3)/3,
sqrt(3)/3,
sqrt(3)/3],
[sqrt(3)/3, sqrt(3)*exp(-2*I*pi/3)/3, sqrt(3)*exp(-4*I*pi/3)/3],
[sqrt(3)/3, sqrt(3)*exp(-4*I*pi/3)/3, sqrt(3)*exp(-8*I*pi/3)/3]])
sympy.physics.matrices.mgamma(mu, lower=False)
Returns a Dirac gamma matrix in the standard (Dirac) representation.
If you want , use gamma(mu, True).
We use a convention:
5 = i 0 1 2 3
5 = i 0 1 2 3 = 5
References
[R321] (page 1915)
Examples
>>> from sympy.physics.matrices import mgamma
>>> mgamma(1)
Matrix([
[ 0, 0, 0, 1],
[ 0, 0, 1, 0],
1600
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
[ 0, -1, 0, 0],
[-1, 0, 0, 0]])
sympy.physics.matrices.msigma(i)
Returns a Pauli matrix i with i = 1, 2, 3
References
[R322] (page 1915)
Examples
>>> from sympy.physics.matrices import msigma
>>> msigma(1)
Matrix([
[0, 1],
[1, 0]])
sympy.physics.matrices.pat matrix(m, dx, dy, dz)
Returns the Parallel Axis Theorem matrix to translate the inertia matrix a distance of
(dx, dy, dz) for a body of mass m.
Examples
To translate a body having a mass of 2 units a distance of 1 unit along the x-axis we get:
>>> from sympy.physics.matrices import pat_matrix
>>> pat_matrix(2, 1, 0, 0)
Matrix([
[0, 0, 0],
[0, 2, 0],
[0, 0, 2]])
Pauli Algebra
This module implements Pauli algebra by subclassing Symbol. Only algebraic properties of
Pauli matrices are used (we dont use the Matrix class).
See the documentation to the class Pauli for examples.
References
sympy.physics.paulialgebra.evaluate pauli product(arg)
Help function to evaluate Pauli matrices product with symbolic objects
Parameters arg: symbolic expression that contains Paulimatrices :
5.36. Physics Module
1601
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.physics.paulialgebra import Pauli, evaluate_pauli_product
>>> from sympy import I
>>> evaluate_pauli_product(I*Pauli(1)*Pauli(2))
-sigma3
>>> from sympy.abc import x,y
>>> evaluate_pauli_product(x**2*Pauli(2)*Pauli(1))
-I*x**2*sigma3
Quantum Harmonic Oscillator in 1-D
sympy.physics.qho 1d.E n(n, omega)
Returns the Energy of the One-dimensional harmonic oscillator
n the nodal quantum number
omega the harmonic oscillator angular frequency
The unit of the returned value matches the unit of hw, since the energy is calculated as:
E n = hbar * omega*(n + 1/2)
Examples
>>> from sympy.physics.qho_1d import E_n
>>> from sympy import var
>>> var(x omega)
(x, omega)
>>> E_n(x, omega)
hbar*omega*(x + 1/2)
sympy.physics.qho 1d.psi n(n, x, m, omega)
Returns the wavefunction psi {n} for the One-dimensional harmonic oscillator.
n the nodal quantum number. Corresponds to the number of nodes in the wavefunction. n >= 0
x x coordinate
m mass of the particle
omega angular frequency of the oscillator
Examples
>>> from sympy.physics.qho_1d import psi_n
>>> from sympy import var
>>> var(x m omega)
(x, m, omega)
>>> psi_n(0, x, m, omega)
(m*omega)**(1/4)*exp(-m*omega*x**2/(2*hbar))/(hbar**(1/4)*pi**(1/4))
1602
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Quantum Harmonic Oscillator in 3-D
sympy.physics.sho.E nl(n, l, hw)
Returns the Energy of an isotropic harmonic oscillator
n the nodal quantum number
l the orbital angular momentum
hw the harmonic oscillator parameter.
The unit of the returned value matches the unit of hw, since the energy is calculated as:
E nl = (2*n + l + 3/2)*hw
Examples
>>> from sympy.physics.sho import E_nl
>>> from sympy import symbols
>>> x, y, z = symbols(x, y, z)
>>> E_nl(x, y, z)
z*(2*x + y + 3/2)
sympy.physics.sho.R nl(n, l, nu, r)
Returns the radial wavefunction R {nl} for a 3d isotropic harmonic oscillator.
n the nodal quantum number. Corresponds to the number of nodes in the wavefunction. n >= 0
l the quantum number for orbital angular momentum
nu mass-scaled frequency: nu = m*omega/(2*hbar) where m0 isthemassandomega the frequency of the oscillator. (in atomic units nu == omega/2)
r Radial coordinate
Examples
>>> from sympy.physics.sho import R_nl
>>> from sympy import var
>>> var(r nu l)
(r, nu, l)
>>> R_nl(0, 0, 1, r)
2*2**(3/4)*exp(-r**2)/pi**(1/4)
>>> R_nl(1, 0, 1, r)
4*2**(1/4)*sqrt(3)*(-2*r**2 + 3/2)*exp(-r**2)/(3*pi**(1/4))
l, nu and r may be symbolic:
>>> R_nl(0, 0, nu, r)
2*2**(3/4)*sqrt(nu**(3/2))*exp(-nu*r**2)/pi**(1/4)
>>> R_nl(0, l, 1, r)
r**l*sqrt(2**(l + 3/2)*2**(l + 2)/factorial2(2*l + 1))*exp(-r**2)/pi**(1/4)
The normalization of the radial wavefunction is:
>>> from sympy import Integral, oo
>>> Integral(R_nl(0, 0, 1, r)**2 * r**2, (r, 0, oo)).n()
1.00000000000000
5.36. Physics Module
1603
SymPy Documentation, Release 0.7.6
>>> Integral(R_nl(1, 0, 1, r)**2 * r**2, (r, 0, oo)).n()
1.00000000000000
>>> Integral(R_nl(1, 1, 1, r)**2 * r**2, (r, 0, oo)).n()
1.00000000000000
Second Quantization
Second quantization operators and states for bosons.
This follow the formulation of Fetter and Welecka, Quantum Theory of Many-Particle Systems.
class sympy.physics.secondquant.Dagger
Hermitian conjugate of creation/annihilation operators.
Examples
>>> from sympy import I
>>> from sympy.physics.secondquant import Dagger, B, Bd
>>> Dagger(2*I)
-2*I
>>> Dagger(B(0))
CreateBoson(0)
>>> Dagger(Bd(0))
AnnihilateBoson(0)
classmethod eval(arg)
Evaluates the Dagger instance.
Examples
>>> from sympy import I
>>> from sympy.physics.secondquant import Dagger, B, Bd
>>> Dagger(2*I)
-2*I
>>> Dagger(B(0))
CreateBoson(0)
>>> Dagger(Bd(0))
AnnihilateBoson(0)
The eval() method is called automatically.
class sympy.physics.secondquant.KroneckerDelta
The discrete, or Kronecker, delta function.
A function that takes in two integers i and j. It returns 0 if i and j are not equal or it
returns 1 if i and j are equal.
Parameters i : Number, Symbol
The rst index of the delta function.
j : Number, Symbol
The second index of the delta function.
1604
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
See Also:
eval
(page
(page 377)
1605),
sympy.functions.special.delta functions.DiracDelta
References
[R344] (page 1915)
Examples
A simple example with integer indices:
>>> from sympy.functions.special.tensor_functions import KroneckerDelta
>>> KroneckerDelta(1, 2)
0
>>> KroneckerDelta(3, 3)
1
Symbolic indices:
>>> from sympy.abc import
>>> KroneckerDelta(i, j)
KroneckerDelta(i, j)
>>> KroneckerDelta(i, i)
1
>>> KroneckerDelta(i, i +
0
>>> KroneckerDelta(i, i +
KroneckerDelta(i, i + k +
i, j, k
1)
1 + k)
1)
classmethod eval(i, j)
Evaluates the discrete delta function.
Examples
>>> from sympy.functions.special.tensor_functions import KroneckerDelta
>>> from sympy.abc import i, j, k
>>> KroneckerDelta(i,
KroneckerDelta(i, j)
>>> KroneckerDelta(i,
1
>>> KroneckerDelta(i,
0
>>> KroneckerDelta(i,
KroneckerDelta(i, i +
j)
i)
i + 1)
i + 1 + k)
k + 1)
# indirect doctest
indices contain equal information
Returns True if indices are either both above or below fermi.
5.36. Physics Module
1605
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.functions.special.tensor_functions import KroneckerDelta
>>> from sympy import Symbol
>>> a = Symbol(a, above_fermi=True)
>>> i = Symbol(i, below_fermi=True)
>>> p = Symbol(p)
>>> q = Symbol(q)
>>> KroneckerDelta(p, q).indices_contain_equal_information
True
>>> KroneckerDelta(p, q+1).indices_contain_equal_information
True
>>> KroneckerDelta(i, p).indices_contain_equal_information
False
is above fermi
True if Delta can be non-zero above fermi
See Also:
is below fermi
(page
1606),
is only above fermi (page 1607)
is only below fermi
(page
1607),
Examples
>>> from sympy.functions.special.tensor_functions import KroneckerDelta
>>> from sympy import Symbol
>>> a = Symbol(a, above_fermi=True)
>>> i = Symbol(i, below_fermi=True)
>>> p = Symbol(p)
>>> q = Symbol(q)
>>> KroneckerDelta(p, a).is_above_fermi
True
>>> KroneckerDelta(p, i).is_above_fermi
False
>>> KroneckerDelta(p, q).is_above_fermi
True
is below fermi
True if Delta can be non-zero below fermi
See Also:
is above fermi
(page
1606),
is only below fermi (page 1607)
is only above fermi
(page
1607),
Examples
>>> from sympy.functions.special.tensor_functions import KroneckerDelta
>>> from sympy import Symbol
>>> a = Symbol(a, above_fermi=True)
>>> i = Symbol(i, below_fermi=True)
>>> p = Symbol(p)
>>> q = Symbol(q)
>>> KroneckerDelta(p, a).is_below_fermi
False
1606
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> KroneckerDelta(p, i).is_below_fermi
True
>>> KroneckerDelta(p, q).is_below_fermi
True
is only above fermi
True if Delta is restricted to above fermi
See Also:
is above fermi (page 1606), is below fermi (page 1606), is only below fermi
(page 1607)
Examples
>>> from sympy.functions.special.tensor_functions import KroneckerDelta
>>> from sympy import Symbol
>>> a = Symbol(a, above_fermi=True)
>>> i = Symbol(i, below_fermi=True)
>>> p = Symbol(p)
>>> q = Symbol(q)
>>> KroneckerDelta(p, a).is_only_above_fermi
True
>>> KroneckerDelta(p, q).is_only_above_fermi
False
>>> KroneckerDelta(p, i).is_only_above_fermi
False
is only below fermi
True if Delta is restricted to below fermi
See Also:
is above fermi (page 1606), is below fermi (page 1606), is only above fermi
(page 1607)
Examples
>>> from sympy.functions.special.tensor_functions import KroneckerDelta
>>> from sympy import Symbol
>>> a = Symbol(a, above_fermi=True)
>>> i = Symbol(i, below_fermi=True)
>>> p = Symbol(p)
>>> q = Symbol(q)
>>> KroneckerDelta(p, i).is_only_below_fermi
True
>>> KroneckerDelta(p, q).is_only_below_fermi
False
>>> KroneckerDelta(p, a).is_only_below_fermi
False
killable index
Returns the index which is preferred to substitute in the nal expression.
The index to substitute is the index with less information regarding fermi level. If
indices contain same information, a is preferred before b.
See Also:
5.36. Physics Module
1607
SymPy Documentation, Release 0.7.6
preferred index (page 1608)
Examples
>>>
>>>
>>>
>>>
>>>
>>>
>>>
p
>>>
p
>>>
j
from sympy.functions.special.tensor_functions import KroneckerDelta
from sympy import Symbol
a = Symbol(a, above_fermi=True)
i = Symbol(i, below_fermi=True)
j = Symbol(j, below_fermi=True)
p = Symbol(p)
KroneckerDelta(p, i).killable_index
KroneckerDelta(p, a).killable_index
KroneckerDelta(i, j).killable_index
preferred index
Returns the index which is preferred to keep in the nal expression.
The preferred index is the index with more information regarding fermi level. If
indices contain same information, a is preferred before b.
See Also:
killable index (page 1607)
Examples
>>>
>>>
>>>
>>>
>>>
>>>
>>>
i
>>>
a
>>>
i
from sympy.functions.special.tensor_functions import KroneckerDelta
from sympy import Symbol
a = Symbol(a, above_fermi=True)
i = Symbol(i, below_fermi=True)
j = Symbol(j, below_fermi=True)
p = Symbol(p)
KroneckerDelta(p, i).preferred_index
KroneckerDelta(p, a).preferred_index
KroneckerDelta(i, j).preferred_index
class sympy.physics.secondquant.AnnihilateBoson
Bosonic annihilation operator.
Examples
>>> from sympy.physics.secondquant import B
>>> from sympy.abc import x
>>> B(x)
AnnihilateBoson(x)
apply operator(state)
Apply state to self if self is not symbolic and state is a FockStateKet, else multiply
self by state.
1608
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.physics.secondquant import B, BKet
>>> from sympy.abc import x, y, n
>>> B(x).apply_operator(y)
y*AnnihilateBoson(x)
>>> B(0).apply_operator(BKet((n,)))
sqrt(n)*FockStateBosonKet((n - 1,))
class sympy.physics.secondquant.CreateBoson
Bosonic creation operator.
apply operator(state)
Apply state to self if self is not symbolic and state is a FockStateKet, else multiply
self by state.
Examples
>>> from sympy.physics.secondquant import B, Dagger, BKet
>>> from sympy.abc import x, y, n
>>> Dagger(B(x)).apply_operator(y)
y*CreateBoson(x)
>>> B(0).apply_operator(BKet((n,)))
sqrt(n)*FockStateBosonKet((n - 1,))
class sympy.physics.secondquant.AnnihilateFermion
Fermionic annihilation operator.
apply operator(state)
Apply state to self if self is not symbolic and state is a FockStateKet, else multiply
self by state.
Examples
>>> from sympy.physics.secondquant import B, Dagger, BKet
>>> from sympy.abc import x, y, n
>>> Dagger(B(x)).apply_operator(y)
y*CreateBoson(x)
>>> B(0).apply_operator(BKet((n,)))
sqrt(n)*FockStateBosonKet((n - 1,))
is only q annihilator
Always destroy a quasi-particle? (annihilate hole or annihilate particle)
>>>
>>>
>>>
>>>
>>>
from sympy import Symbol
from sympy.physics.secondquant import F
a = Symbol(a, above_fermi=True)
i = Symbol(i, below_fermi=True)
p = Symbol(p)
>>> F(a).is_only_q_annihilator
True
>>> F(i).is_only_q_annihilator
False
5.36. Physics Module
1609
SymPy Documentation, Release 0.7.6
>>> F(p).is_only_q_annihilator
False
is only q creator
Always create a quasi-particle? (create hole or create particle)
>>>
>>>
>>>
>>>
>>>
from sympy import Symbol
from sympy.physics.secondquant import F
a = Symbol(a, above_fermi=True)
i = Symbol(i, below_fermi=True)
p = Symbol(p)
>>> F(a).is_only_q_creator
False
>>> F(i).is_only_q_creator
True
>>> F(p).is_only_q_creator
False
is q annihilator
Can we destroy a quasi-particle? (annihilate hole or annihilate particle) If so, would
that be above or below the fermi surface?
>>>
>>>
>>>
>>>
>>>
from sympy import Symbol
from sympy.physics.secondquant import F
a = Symbol(a, above_fermi=1)
i = Symbol(i, below_fermi=1)
p = Symbol(p)
>>> F(a).is_q_annihilator
1
>>> F(i).is_q_annihilator
0
>>> F(p).is_q_annihilator
1
is q creator
Can we create a quasi-particle? (create hole or create particle) If so, would that be
above or below the fermi surface?
>>>
>>>
>>>
>>>
>>>
from sympy import Symbol
from sympy.physics.secondquant import F
a = Symbol(a, above_fermi=True)
i = Symbol(i, below_fermi=True)
p = Symbol(p)
>>> F(a).is_q_creator
0
>>> F(i).is_q_creator
-1
>>> F(p).is_q_creator
-1
class sympy.physics.secondquant.CreateFermion
Fermionic creation operator.
apply operator(state)
Apply state to self if self is not symbolic and state is a FockStateKet, else multiply
self by state.
1610
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.physics.secondquant import B, Dagger, BKet
>>> from sympy.abc import x, y, n
>>> Dagger(B(x)).apply_operator(y)
y*CreateBoson(x)
>>> B(0).apply_operator(BKet((n,)))
sqrt(n)*FockStateBosonKet((n - 1,))
is only q annihilator
Always destroy a quasi-particle? (annihilate hole or annihilate particle)
>>>
>>>
>>>
>>>
>>>
from sympy import Symbol
from sympy.physics.secondquant import Fd
a = Symbol(a, above_fermi=True)
i = Symbol(i, below_fermi=True)
p = Symbol(p)
>>> Fd(a).is_only_q_annihilator
False
>>> Fd(i).is_only_q_annihilator
True
>>> Fd(p).is_only_q_annihilator
False
is only q creator
Always create a quasi-particle? (create hole or create particle)
>>>
>>>
>>>
>>>
>>>
from sympy import Symbol
from sympy.physics.secondquant import Fd
a = Symbol(a, above_fermi=True)
i = Symbol(i, below_fermi=True)
p = Symbol(p)
>>> Fd(a).is_only_q_creator
True
>>> Fd(i).is_only_q_creator
False
>>> Fd(p).is_only_q_creator
False
is q annihilator
Can we destroy a quasi-particle? (annihilate hole or annihilate particle) If so, would
that be above or below the fermi surface?
>>>
>>>
>>>
>>>
>>>
from sympy import Symbol
from sympy.physics.secondquant import Fd
a = Symbol(a, above_fermi=1)
i = Symbol(i, below_fermi=1)
p = Symbol(p)
>>> Fd(a).is_q_annihilator
0
>>> Fd(i).is_q_annihilator
-1
>>> Fd(p).is_q_annihilator
-1
5.36. Physics Module
1611
SymPy Documentation, Release 0.7.6
is q creator
Can we create a quasi-particle? (create hole or create particle) If so, would that be
above or below the fermi surface?
>>>
>>>
>>>
>>>
>>>
from sympy import Symbol
from sympy.physics.secondquant import Fd
a = Symbol(a, above_fermi=True)
i = Symbol(i, below_fermi=True)
p = Symbol(p)
>>> Fd(a).is_q_creator
1
>>> Fd(i).is_q_creator
0
>>> Fd(p).is_q_creator
1
class sympy.physics.secondquant.FockState
Many particle Fock state with a sequence of occupation numbers.
Anywhere you can have a FockState, you can also have S.Zero. All code must check for
this!
Base class to represent FockStates.
class sympy.physics.secondquant.FockStateBra
Representation of a bra.
class sympy.physics.secondquant.FockStateKet
Representation of a ket.
class sympy.physics.secondquant.FockStateBosonKet
Many particle Fock state with a sequence of occupation numbers.
Occupation numbers can be any integer >= 0.
Examples
>>> from sympy.physics.secondquant import BKet
>>> BKet([1, 2])
FockStateBosonKet((1, 2))
class sympy.physics.secondquant.FockStateBosonBra
Describes a collection of BosonBra particles.
Examples
>>> from sympy.physics.secondquant import BBra
>>> BBra([1, 2])
FockStateBosonBra((1, 2))
sympy.physics.secondquant.BBra
alias of FockStateBosonBra (page 1612)
sympy.physics.secondquant.BKet
alias of FockStateBosonKet (page 1612)
sympy.physics.secondquant.FBra
alias of FockStateFermionBra
1612
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
sympy.physics.secondquant.FKet
alias of FockStateFermionKet
sympy.physics.secondquant.F
alias of AnnihilateFermion (page 1609)
sympy.physics.secondquant.Fd
alias of CreateFermion (page 1610)
sympy.physics.secondquant.B
alias of AnnihilateBoson (page 1608)
sympy.physics.secondquant.Bd
alias of CreateBoson (page 1609)
sympy.physics.secondquant.apply operators(e)
Take a sympy expression with operators and states and apply the operators.
Examples
>>> from sympy.physics.secondquant import apply_operators
>>> from sympy import sympify
>>> apply_operators(sympify(3)+4)
7
class sympy.physics.secondquant.InnerProduct
An unevaluated inner product between a bra and ket.
Currently this class just reduces things to a product of Kronecker Deltas. In the future,
we could introduce abstract states like |a> and |b>, and leave the inner product unevaluated as <a|b>.
bra
Returns the bra part of the state
ket
Returns the ket part of the state
class sympy.physics.secondquant.BosonicBasis
Base class for a basis set of bosonic Fock states.
class sympy.physics.secondquant.VarBosonicBasis(n max)
A single state, variable particle number basis set.
Examples
>>> from sympy.physics.secondquant import VarBosonicBasis
>>> b = VarBosonicBasis(5)
>>> b
[FockState((0,)), FockState((1,)), FockState((2,)),
FockState((3,)), FockState((4,))]
index(state)
Returns the index of state in basis.
5.36. Physics Module
1613
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.physics.secondquant import VarBosonicBasis
>>> b = VarBosonicBasis(3)
>>> state = b.state(1)
>>> b
[FockState((0,)), FockState((1,)), FockState((2,))]
>>> state
FockStateBosonKet((1,))
>>> b.index(state)
1
state(i)
The state of a single basis.
Examples
>>> from sympy.physics.secondquant import VarBosonicBasis
>>> b = VarBosonicBasis(5)
>>> b.state(3)
FockStateBosonKet((3,))
class sympy.physics.secondquant.FixedBosonicBasis(n particles, n levels)
Fixed particle number basis set.
Examples
>>> from sympy.physics.secondquant import FixedBosonicBasis
>>> b = FixedBosonicBasis(2, 2)
>>> state = b.state(1)
>>> b
[FockState((2, 0)), FockState((1, 1)), FockState((0, 2))]
>>> state
FockStateBosonKet((1, 1))
>>> b.index(state)
1
index(state)
Returns the index of state in basis.
Examples
>>> from sympy.physics.secondquant import FixedBosonicBasis
>>> b = FixedBosonicBasis(2, 3)
>>> b.index(b.state(3))
3
state(i)
Returns the state that lies at index i of the basis
1614
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.physics.secondquant import FixedBosonicBasis
>>> b = FixedBosonicBasis(2, 3)
>>> b.state(3)
FockStateBosonKet((1, 0, 1))
class sympy.physics.secondquant.Commutator
The Commutator: [A, B] = A*B - B*A
The arguments are ordered according to . cmp ()
>>> from sympy import symbols
>>> from sympy.physics.secondquant import Commutator
>>> A, B = symbols(A,B, commutative=False)
>>> Commutator(B, A)
-Commutator(A, B)
Evaluate the commutator with .doit()
>>> comm = Commutator(A,B); comm
Commutator(A, B)
>>> comm.doit()
A*B - B*A
For two second quantization operators the commutator is evaluated immediately:
>>>
>>>
>>>
>>>
from sympy.physics.secondquant import Fd, F
a = symbols(a, above_fermi=True)
i = symbols(i, below_fermi=True)
p,q = symbols(p,q)
>>> Commutator(Fd(a),Fd(i))
2*NO(CreateFermion(a)*CreateFermion(i))
But for more complicated expressions, the evaluation is triggered by a call to .doit()
>>> comm = Commutator(Fd(p)*Fd(q),F(i)); comm
Commutator(CreateFermion(p)*CreateFermion(q), AnnihilateFermion(i))
>>> comm.doit(wicks=True)
-KroneckerDelta(i, p)*CreateFermion(q) +
KroneckerDelta(i, q)*CreateFermion(p)
doit(**hints)
Enables the computation of complex expressions.
Examples
>>>
>>>
>>>
>>>
>>>
>>>
0
from sympy.physics.secondquant import Commutator, F, Fd
from sympy import symbols
i, j = symbols(i,j, below_fermi=True)
a, b = symbols(a,b, above_fermi=True)
c = Commutator(Fd(a)*F(i),Fd(b)*F(j))
c.doit(wicks=True)
classmethod eval(a, b)
The Commutator [A,B] is on canonical form if A < B.
5.36. Physics Module
1615
SymPy Documentation, Release 0.7.6
Examples
>>>
>>>
>>>
>>>
>>>
0
from sympy.physics.secondquant import Commutator, F, Fd
from sympy.abc import x
c1 = Commutator(F(x), Fd(x))
c2 = Commutator(Fd(x), F(x))
Commutator.eval(c1, c2)
sympy.physics.secondquant.matrix rep(op, basis)
Find the representation of an operator in a basis.
Examples
>>> from sympy.physics.secondquant import VarBosonicBasis, B, matrix_rep
>>> b = VarBosonicBasis(5)
>>> o = B(0)
>>> matrix_rep(o, b)
Matrix([
[0, 1,
0,
0, 0],
[0, 0, sqrt(2),
0, 0],
[0, 0,
0, sqrt(3), 0],
[0, 0,
0,
0, 2],
[0, 0,
0,
0, 0]])
sympy.physics.secondquant.contraction(a, b)
Calculates contraction of Fermionic operators a and b.
Examples
>>>
>>>
>>>
>>>
>>>
from
from
p, q
a, b
i, j
sympy import symbols
sympy.physics.secondquant import F, Fd, contraction
= symbols(p,q)
= symbols(a,b, above_fermi=True)
= symbols(i,j, below_fermi=True)
A contraction is non-zero only if a quasi-creator is to the right of a quasi-annihilator:
>>> contraction(F(a),Fd(b))
KroneckerDelta(a, b)
>>> contraction(Fd(i),F(j))
KroneckerDelta(i, j)
For general indices a non-zero result restricts the indices to below/above the fermi surface:
>>> contraction(Fd(p),F(q))
KroneckerDelta(_i, q)*KroneckerDelta(p, q)
>>> contraction(F(p),Fd(q))
KroneckerDelta(_a, q)*KroneckerDelta(p, q)
Two creators or two annihilators always vanishes:
>>> contraction(F(p),F(q))
0
1616
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> contraction(Fd(p),Fd(q))
0
sympy.physics.secondquant.wicks(e, **kw args)
Returns the normal ordered equivalent of an expression using Wicks Theorem.
Examples
>>> from sympy import symbols, Function, Dummy
>>> from sympy.physics.secondquant import wicks, F, Fd, NO
>>> p,q,r = symbols(p,q,r)
>>> wicks(Fd(p)*F(q))
d(p, q)*d(q, _i) + NO(CreateFermion(p)*AnnihilateFermion(q))
By default, the expression is expanded:
>>> wicks(F(p)*(F(q)+F(r)))
NO(AnnihilateFermion(p)*AnnihilateFermion(q)) + NO(
AnnihilateFermion(p)*AnnihilateFermion(r))
With the keyword keep only fully contracted=True, only fully contracted terms are returned.
By request, the result can be simplied in the following order: KroneckerDelta
functions are evaluated Dummy variables are substituted consistently across terms
>>> p, q, r = symbols(p q r, cls=Dummy)
>>> wicks(Fd(p)*(F(q)+F(r)), keep_only_fully_contracted=True)
KroneckerDelta(_i, _q)*KroneckerDelta(
_p, _q) + KroneckerDelta(_i, _r)*KroneckerDelta(_p, _r)
class sympy.physics.secondquant.NO
This Object is used to represent normal ordering brackets.
i.e. {abcd} sometimes written :abcd:
Applying the function NO(arg) to an argument means that all operators in the argument
will be assumed to anticommute, and have vanishing contractions. This allows an immediate reordering to canonical form upon object creation.
>>> from sympy import symbols
>>> from sympy.physics.secondquant import NO, F, Fd
>>> p,q = symbols(p,q)
>>> NO(Fd(p)*F(q))
NO(CreateFermion(p)*AnnihilateFermion(q))
>>> NO(F(q)*Fd(p))
-NO(CreateFermion(p)*AnnihilateFermion(q))
Note: If you want to generate a normal ordered equivalent of an expression, you should
use the function wicks(). This class only indicates that all operators inside the brackets
anticommute, and have vanishing contractions. Nothing more, nothing less.
doit(**kw args)
Either removes the brackets or enables complex computations in its arguments.
5.36. Physics Module
1617
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.physics.secondquant import NO, Fd, F
>>> from textwrap import fill
>>> from sympy import symbols, Dummy
>>> p,q = symbols(p,q, cls=Dummy)
>>> print(fill(str(NO(Fd(p)*F(q)).doit())))
KroneckerDelta(_a, _p)*KroneckerDelta(_a,
_q)*CreateFermion(_a)*AnnihilateFermion(_a) + KroneckerDelta(_a,
_p)*KroneckerDelta(_i, _q)*CreateFermion(_a)*AnnihilateFermion(_i) KroneckerDelta(_a, _q)*KroneckerDelta(_i,
_p)*AnnihilateFermion(_a)*CreateFermion(_i) - KroneckerDelta(_i,
_p)*KroneckerDelta(_i, _q)*AnnihilateFermion(_i)*CreateFermion(_i)
get subNO(i)
Returns a NO() without FermionicOperator at index i.
Examples
>>> from sympy import symbols
>>> from sympy.physics.secondquant import F, NO
>>> p,q,r = symbols(p,q,r)
>>> NO(F(p)*F(q)*F(r)).get_subNO(1)
NO(AnnihilateFermion(p)*AnnihilateFermion(r))
has q annihilators
Return 0 if the rightmost argument of the rst argument is a not a q annihilator, else
1 if it is above fermi or -1 if it is below fermi.
Examples
>>> from sympy import symbols
>>> from sympy.physics.secondquant import NO, F, Fd
>>>
>>>
>>>
-1
>>>
1
>>>
0
a = symbols(a, above_fermi=True)
i = symbols(i, below_fermi=True)
NO(Fd(a)*Fd(i)).has_q_annihilators
NO(F(i)*F(a)).has_q_annihilators
NO(Fd(a)*F(i)).has_q_annihilators
has q creators
Return 0 if the leftmost argument of the rst argument is a not a q creator, else 1 if
it is above fermi or -1 if it is below fermi.
Examples
>>> from sympy import symbols
>>> from sympy.physics.secondquant import NO, F, Fd
1618
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>>
>>>
>>>
1
>>>
-1
>>>
0
a = symbols(a, above_fermi=True)
i = symbols(i, below_fermi=True)
NO(Fd(a)*Fd(i)).has_q_creators
NO(F(i)*F(a)).has_q_creators
NO(Fd(i)*F(a)).has_q_creators
iter q annihilators()
Iterates over the annihilation operators.
Examples
>>>
>>>
>>>
>>>
>>>
from
i, j
a, b
from
no =
sympy import symbols
= symbols(i j, below_fermi=True)
= symbols(a b, above_fermi=True)
sympy.physics.secondquant import NO, F, Fd
NO(Fd(a)*F(i)*F(b)*Fd(j))
>>> no.iter_q_creators()
<generator object... at 0x...>
>>> list(no.iter_q_creators())
[0, 1]
>>> list(no.iter_q_annihilators())
[3, 2]
iter q creators()
Iterates over the creation operators.
Examples
>>>
>>>
>>>
>>>
>>>
from
i, j
a, b
from
no =
sympy import symbols
= symbols(i j, below_fermi=True)
= symbols(a b, above_fermi=True)
sympy.physics.secondquant import NO, F, Fd
NO(Fd(a)*F(i)*F(b)*Fd(j))
>>> no.iter_q_creators()
<generator object... at 0x...>
>>> list(no.iter_q_creators())
[0, 1]
>>> list(no.iter_q_annihilators())
[3, 2]
sympy.physics.secondquant.evaluate deltas(e)
We evaluate KroneckerDelta symbols in the expression assuming Einstein summation.
If one index is repeated it is summed over and in eect substituted with the other one.
If both indices are repeated we substitute according to what is the preferred index. this
is determined by KroneckerDelta.preferred index and KroneckerDelta.killable index.
In case there are no possible substitutions or if a substitution would imply a loss of
information, nothing is done.
5.36. Physics Module
1619
SymPy Documentation, Release 0.7.6
In case an index appears in more than one KroneckerDelta, the resulting substitution
depends on the order of the factors. Since the ordering is platform dependent, the literal
expression resulting from this function may be hard to predict.
Examples
We assume the following:
>>>
>>>
>>>
>>>
>>>
>>>
>>>
from sympy import symbols, Function, Dummy, KroneckerDelta
from sympy.physics.secondquant import evaluate_deltas
i,j = symbols(i j, below_fermi=True, cls=Dummy)
a,b = symbols(a b, above_fermi=True, cls=Dummy)
p,q = symbols(p q, cls=Dummy)
f = Function(f)
t = Function(t)
The order of preference for these indices according to KroneckerDelta is (a, b, i, j, p, q).
Trivial cases:
>>> evaluate_deltas(KroneckerDelta(i,j)*f(i))
f(_j)
>>> evaluate_deltas(KroneckerDelta(i,j)*f(j))
f(_i)
>>> evaluate_deltas(KroneckerDelta(i,p)*f(p))
f(_i)
>>> evaluate_deltas(KroneckerDelta(q,p)*f(p))
f(_q)
>>> evaluate_deltas(KroneckerDelta(q,p)*f(q))
f(_p)
# d_ij f(i) -> f(j)
# d_ij f(j) -> f(i)
# d_ip f(p) -> f(i)
# d_qp f(p) -> f(q)
# d_qp f(q) -> f(p)
More interesting cases:
>>> evaluate_deltas(KroneckerDelta(i,p)*t(a,i)*f(p,q))
f(_i, _q)*t(_a, _i)
>>> evaluate_deltas(KroneckerDelta(a,p)*t(a,i)*f(p,q))
f(_a, _q)*t(_a, _i)
>>> evaluate_deltas(KroneckerDelta(p,q)*f(p,q))
f(_p, _p)
Finally, here are some cases where nothing is done, because that would imply a loss of
information:
>>> evaluate_deltas(KroneckerDelta(i,p)*f(q))
f(_q)*KroneckerDelta(_i, _p)
>>> evaluate_deltas(KroneckerDelta(i,p)*f(i))
f(_i)*KroneckerDelta(_i, _p)
class sympy.physics.secondquant.AntiSymmetricTensor
Stores upper and lower indices in separate Tuples.
Each group of indices is assumed to be antisymmetric.
Examples
1620
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> from sympy import symbols
>>> from sympy.physics.secondquant import AntiSymmetricTensor
>>> i, j = symbols(i j, below_fermi=True)
>>> a, b = symbols(a b, above_fermi=True)
>>> AntiSymmetricTensor(v, (a, i), (b, j))
AntiSymmetricTensor(v, (a, i), (b, j))
>>> AntiSymmetricTensor(v, (i, a), (b, j))
-AntiSymmetricTensor(v, (a, i), (b, j))
As you can see, the indices are automatically sorted to a canonical form.
doit(**kw args)
Returns self.
Examples
>>> from sympy import symbols
>>> from sympy.physics.secondquant import AntiSymmetricTensor
>>> i, j = symbols(i,j, below_fermi=True)
>>> a, b = symbols(a,b, above_fermi=True)
>>> AntiSymmetricTensor(v, (a, i), (b, j)).doit()
AntiSymmetricTensor(v, (a, i), (b, j))
lower
Returns the lower indices.
Examples
>>> from sympy import symbols
>>> from sympy.physics.secondquant import AntiSymmetricTensor
>>> i, j = symbols(i,j, below_fermi=True)
>>> a, b = symbols(a,b, above_fermi=True)
>>> AntiSymmetricTensor(v, (a, i), (b, j))
AntiSymmetricTensor(v, (a, i), (b, j))
>>> AntiSymmetricTensor(v, (a, i), (b, j)).lower
(b, j)
symbol
Returns the symbol of the tensor.
Examples
>>> from sympy import symbols
>>> from sympy.physics.secondquant import AntiSymmetricTensor
>>> i, j = symbols(i,j, below_fermi=True)
>>> a, b = symbols(a,b, above_fermi=True)
>>> AntiSymmetricTensor(v, (a, i), (b, j))
AntiSymmetricTensor(v, (a, i), (b, j))
>>> AntiSymmetricTensor(v, (a, i), (b, j)).symbol
v
upper
Returns the upper indices.
5.36. Physics Module
1621
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import symbols
>>> from sympy.physics.secondquant import AntiSymmetricTensor
>>> i, j = symbols(i,j, below_fermi=True)
>>> a, b = symbols(a,b, above_fermi=True)
>>> AntiSymmetricTensor(v, (a, i), (b, j))
AntiSymmetricTensor(v, (a, i), (b, j))
>>> AntiSymmetricTensor(v, (a, i), (b, j)).upper
(a, i)
sympy.physics.secondquant.substitute dummies(expr,
new indices=False,
pretty indices={})
Collect terms by substitution of dummy variables.
This routine allows simplication of Add expressions containing terms which dier only
due to dummy variables.
The idea is to substitute all dummy variables consistently depending on the structure of
the term. For each term, we obtain a sequence of all dummy variables, where the order is
determined by the index range, what factors the index belongs to and its position in each
factor. See get ordered dummies() for more inforation about the sorting of dummies.
The index sequence is then substituted consistently in each term.
Examples
>>>
>>>
>>>
>>>
>>>
from sympy import symbols, Function, Dummy
from sympy.physics.secondquant import substitute_dummies
a,b,c,d = symbols(a b c d, above_fermi=True, cls=Dummy)
i,j = symbols(i j, below_fermi=True, cls=Dummy)
f = Function(f)
>>> expr = f(a,b) + f(c,d); expr
f(_a, _b) + f(_c, _d)
Since a, b, c and d are equivalent summation indices, the expression can be simplied
to a single term (for which the dummy indices are still summed over)
>>> substitute_dummies(expr)
2*f(_a, _b)
Controlling output:
By default the dummy symbols that are already present in the expression will be reused in
a dierent permuation. However, if new indices=True, new dummies will be generated
and inserted. The keyword pretty indices can be used to control this generation of new
symbols.
By default the new dummies will be generated on the form i 1, i 2, a 1, etc. If you supply
a dictionary with key:value pairs in the form:
{ index group: string of letters }
The letters will be used as labels for the new dummy symbols. The index groups must
be one of above, below or general.
>>> expr = f(a,b,i,j)
>>> my_dummies = { above:st, below:uv }
1622
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> substitute_dummies(expr, new_indices=True, pretty_indices=my_dummies)
f(_s, _t, _u, _v)
If we run out of letters, or if there is no keyword for some index group the default dummy
generator will be used as a fallback:
>>> p,q = symbols(p q, cls=Dummy) # general indices
>>> expr = f(p,q)
>>> substitute_dummies(expr, new_indices=True, pretty_indices=my_dummies)
f(_p_0, _p_1)
class sympy.physics.secondquant.PermutationOperator
Represents the index permutation operator P(ij).
P(ij)*f(i)*g(j) = f(i)*g(j) - f(j)*g(i)
get permuted(expr)
Returns -expr with permuted indices.
>>> from sympy import symbols, Function
>>> from sympy.physics.secondquant import PermutationOperator
>>> p,q = symbols(p,q)
>>> f = Function(f)
>>> PermutationOperator(p,q).get_permuted(f(p,q))
-f(q, p)
sympy.physics.secondquant.simplify index permutations(expr,
permutation operators)
Performs simplication by introducing PermutationOperators where appropriate.
Schematically: [abij] - [abji] - [baij] + [baji] -> P(ab)*P(ij)*[abij]
permutation operators is a list of PermutationOperators to consider.
If permutation operators=[P(ab),P(ij)] we will try to introduce the permutation operators
P(ij) and P(ab) in the expression. If there are other possible simplications, we ignore
them.
>>>
>>>
>>>
>>>
>>>
>>>
from sympy import symbols, Function
from sympy.physics.secondquant import simplify_index_permutations
from sympy.physics.secondquant import PermutationOperator
p,q,r,s = symbols(p,q,r,s)
f = Function(f)
g = Function(g)
>>> expr = f(p)*g(q) - f(q)*g(p); expr
f(p)*g(q) - f(q)*g(p)
>>> simplify_index_permutations(expr,[PermutationOperator(p,q)])
f(p)*g(q)*PermutationOperator(p, q)
>>> PermutList = [PermutationOperator(p,q),PermutationOperator(r,s)]
>>> expr = f(p,r)*g(q,s) - f(q,r)*g(p,s) + f(q,s)*g(p,r) - f(p,s)*g(q,r)
>>> simplify_index_permutations(expr,PermutList)
f(p, r)*g(q, s)*PermutationOperator(p, q)*PermutationOperator(r, s)
Wigner Symbols
Wigner, Clebsch-Gordan, Racah, and Gaunt coecients
5.36. Physics Module
1623
SymPy Documentation, Release 0.7.6
Collection of functions for calculating Wigner 3j, 6j, 9j, Clebsch-Gordan, Racah as well as
Gaunt coecients exactly, all evaluating to a rational number times the square root of a
rational number [Rasch03] (page 1915).
Please see the description of the individual functions for further details and examples.
References
Credits and Copyright
This code was taken from Sage with the permission of all authors:
https://groups.google.com/forum/#!topic/sage-devel/M4NZdu-7O38
AUTHORS:
Jens Rasch (2009-03-24): initial version for Sage
Jens Rasch (2009-05-31): updated to sage-4.0
Copyright (C) 2008 Jens Rasch <[email protected]>
sympy.physics.wigner.clebsch gordan(j 1, j 2, j 3, m 1, m 2, m 3)
Calculates the Clebsch-Gordan coecient hj1 m1 j2 m2 |j3 m3 i.
The reference for this function is [Edmonds74] (page 1916).
INPUT:
j 1, j 2, j 3, m 1, m 2, m 3 - integer or half integer
OUTPUT:
Rational number times the square root of a rational number.
EXAMPLES:
>>> from sympy import S
>>> from sympy.physics.wigner import clebsch_gordan
>>> clebsch_gordan(S(3)/2, S(1)/2, 2, S(3)/2, S(1)/2, 2)
1
>>> clebsch_gordan(S(3)/2, S(1)/2, 1, S(3)/2, -S(1)/2, 1)
sqrt(3)/2
>>> clebsch_gordan(S(3)/2, S(1)/2, 1, -S(1)/2, S(1)/2, 0)
-sqrt(2)/2
NOTES:
The Clebsch-Gordan coecient will be evaluated via its relation to Wigner 3j symbols:
hj1 m1 j2 m2 |j3 m3 i = (1)j1 j2 +m3 2j3 + 1 W igner3j(j1 , j2 , j3 , m1 , m2 , m3 )
See also the documentation on Wigner 3j symbols which exhibit much higher symmetry
relations than the Clebsch-Gordan coecient.
AUTHORS:
Jens Rasch (2009-03-24): initial version
sympy.physics.wigner.gaunt(l 1, l 2, l 3, m 1, m 2, m 3, prec=None)
Calculate the Gaunt coecient.
The Gaunt coecient is dened as the integral over three spherical harmonics:
Y (j1 , j2 , j3 , m1 , m2 , m3 ) = Yl1 ,m1 ()Yl2 ,m2 ()Yl3 ,m3 ()d = (2l1 + 1)(2l2 + 1)(2l3 + 1)/(4) Y (j1 , j2 , j3 , 0, 0, 0) Y (
1624
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
INPUT:
l 1, l 2, l 3, m 1, m 2, m 3 - integer
prec - precision, default: None. Providing a precision can drastically speed up the
calculation.
OUTPUT:
Rational number times the square root of a rational number (if prec=None), or real number if a precision is given.
Examples
>>> from sympy.physics.wigner import gaunt
>>> gaunt(1,0,1,1,0,-1)
-1/(2*sqrt(pi))
>>> gaunt(1000,1000,1200,9,3,-12).n(64)
0.00689500421922113448...
It is an error to use non-integer values for l and m:
sage: gaunt(1.2,0,1.2,0,0,0)
Traceback (most recent call last):
...
ValueError: l values must be integer
sage: gaunt(1,0,1,1.1,0,-1.1)
Traceback (most recent call last):
...
ValueError: m values must be integer
NOTES:
The Gaunt coecient obeys the following symmetry rules:
invariant under any permutation of the columns
Y (j1 , j2 , j3 , m1 , m2 , m3 ) = Y (j3 , j1 , j2 , m3 , m1 , m2 ) = Y (j2 , j3 , j1 , m2 , m3 , m1 ) = Y (j3 , j2 , j1 , m3 , m2 , m1 ) = Y (j1 , j
invariant under space inection, i.e.
Y (j1 , j2 , j3 , m1 , m2 , m3 ) = Y (j1 , j2 , j3 , m1 , m2 , m3 )
symmetric with respect to the 72 Regge symmetries as inherited for the 3j symbols
[Regge58] (page 1916)
zero for l1 , l2 , l3 not fullling triangle relation
zero for violating any one of the conditions: l1 |m1 |, l2 |m2 |, l3 |m3 |
non-zero only for an even sum of the li , i.e. J = l1 + l2 + l3 = 2n for n in N
ALGORITHM:
This function uses the algorithm of [Liberatodebrito82] (page 1916) to calculate the value
of the Gaunt coecient exactly. Note that the formula contains alternating sums over
large factorials and is therefore unsuitable for nite precision arithmetic and only useful
for a computer algebra system [Rasch03] (page 1915).
5.36. Physics Module
1625
SymPy Documentation, Release 0.7.6
REFERENCES:
AUTHORS:
Jens Rasch (2009-03-24): initial version for Sage
sympy.physics.wigner.racah(aa, bb, cc, dd, ee, , prec=None)
Calculate the Racah symbol W (a, b, c, d; e, f ).
INPUT:
a, ..., f - integer or half integer
prec - precision, default: None. Providing a precision can drastically speed up the
calculation.
OUTPUT:
Rational number times the square root of a rational number (if prec=None), or real number if a precision is given.
Examples
>>> from sympy.physics.wigner import racah
>>> racah(3,3,3,3,3,3)
-1/14
NOTES:
The Racah symbol is related to the Wigner 6j symbol:
W igner6j(j1 , j2 , j3 , j4 , j5 , j6 ) = (1)j1 +j2 +j4 +j5 W (j1 , j2 , j5 , j4 , j3 , j6 )
Please see the 6j symbol for its much richer symmetries and for additional properties.
ALGORITHM:
This function uses the algorithm of [Edmonds74] (page 1916) to calculate the value of the
6j symbol exactly. Note that the formula contains alternating sums over large factorials
and is therefore unsuitable for nite precision arithmetic and only useful for a computer
algebra system [Rasch03] (page 1915).
AUTHORS:
Jens Rasch (2009-03-24): initial version
sympy.physics.wigner.wigner 3j(j 1, j 2, j 3, m 1, m 2, m 3)
Calculate the Wigner 3j symbol W igner3j(j1 , j2 , j3 , m1 , m2 , m3 ).
INPUT:
j 1, j 2, j 3, m 1, m 2, m 3 - integer or half integer
OUTPUT:
Rational number times the square root of a rational number.
Examples
>>> from sympy.physics.wigner import wigner_3j
>>> wigner_3j(2, 6, 4, 0, 0, 0)
sqrt(715)/143
>>> wigner_3j(2, 6, 4, 0, 0, 1)
0
1626
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
It is an error to have arguments that are not integer or half integer values:
sage: wigner 3j(2.1, 6, 4, 0, 0, 0)
Traceback (most recent call last):
...
ValueError: j values must be integer or half integer
sage: wigner 3j(2, 6, 4, 1, 0, -1.1)
Traceback (most recent call last):
...
ValueError: m values must be integer or half integer
NOTES:
The Wigner 3j symbol obeys the following symmetry rules:
invariant under any permutation of the columns (with the exception of a sign change
where J := j1 + j2 + j3 ):
W igner3j(j1 , j2 , j3 , m1 , m2 , m3 ) = W igner3j(j3 , j1 , j2 , m3 , m1 , m2 ) = W igner3j(j2 , j3 , j1 , m2 , m3 , m1 ) = (1)J W
invariant under space inection, i.e.
W igner3j(j1 , j2 , j3 , m1 , m2 , m3 ) = (1)J W igner3j(j1 , j2 , j3 , m1 , m2 , m3 )
symmetric with respect to the 72 additional symmetries based on the work by
[Regge58] (page 1916)
zero for j1 , j2 , j3 not fullling triangle relation
zero for m1 + m2 + m3 6= 0
zero for violating any one of the conditions j1 |m1 |, j2 |m2 |, j3 |m3 |
ALGORITHM:
This function uses the algorithm of [Edmonds74] (page 1916) to calculate the value of the
3j symbol exactly. Note that the formula contains alternating sums over large factorials
and is therefore unsuitable for nite precision arithmetic and only useful for a computer
algebra system [Rasch03] (page 1915).
REFERENCES:
AUTHORS:
Jens Rasch (2009-03-24): initial version
sympy.physics.wigner.wigner 6j(j 1, j 2, j 3, j 4, j 5, j 6, prec=None)
Calculate the Wigner 6j symbol W igner6j(j1 , j2 , j3 , j4 , j5 , j6 ).
INPUT:
j 1, ..., j 6 - integer or half integer
prec - precision, default: None. Providing a precision can drastically speed up the
calculation.
OUTPUT:
Rational number times the square root of a rational number (if prec=None), or real number if a precision is given.
5.36. Physics Module
1627
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.physics.wigner import wigner_6j
>>> wigner_6j(3,3,3,3,3,3)
-1/14
>>> wigner_6j(5,5,5,5,5,5)
1/52
It is an error to have arguments that are not integer or half integer values or do not fulll
the triangle relation:
sage: wigner 6j(2.5,2.5,2.5,2.5,2.5,2.5)
Traceback (most recent call last):
...
ValueError: j values must be integer or half integer and fulfill the triangle relation
sage: wigner 6j(0.5,0.5,1.1,0.5,0.5,1.1)
Traceback (most recent call last):
...
ValueError: j values must be integer or half integer and fulfill the triangle relation
NOTES:
The Wigner 6j symbol is related to the Racah symbol but exhibits more symmetries as
detailed below.
W igner6j(j1 , j2 , j3 , j4 , j5 , j6 ) = (1)j1 +j2 +j4 +j5 W (j1 , j2 , j5 , j4 , j3 , j6 )
The Wigner 6j symbol obeys the following symmetry rules:
Wigner 6j symbols are left invariant under any permutation of the columns:
W igner6j(j1 , j2 , j3 , j4 , j5 , j6 ) = W igner6j(j3 , j1 , j2 , j6 , j4 , j5 ) = W igner6j(j2 , j3 , j1 , j5 , j6 , j4 ) = W igner6j(j3 , j2 , j
They are invariant under the exchange of the upper and lower arguments in each of
any two columns, i.e.
W igner6j(j1 , j2 , j3 , j4 , j5 , j6 ) = W igner6j(j1 , j5 , j6 , j4 , j2 , j3 ) = W igner6j(j4 , j2 , j6 , j1 , j5 , j3 ) = W igner6j(j4 , j5 , j
additional 6 symmetries [Regge59] (page 1916) giving rise to 144 symmetries in
total
only non-zero if any triple of js fulll a triangle relation
ALGORITHM:
This function uses the algorithm of [Edmonds74] (page 1916) to calculate the value of the
6j symbol exactly. Note that the formula contains alternating sums over large factorials
and is therefore unsuitable for nite precision arithmetic and only useful for a computer
algebra system [Rasch03] (page 1915).
REFERENCES:
sympy.physics.wigner.wigner 9j(j 1, j 2, j 3, j 4, j 5, j 6, j 7, j 8, j 9, prec=None)
Calculate the Wigner 9j symbol W igner9j(j1 , j2 , j3 , j4 , j5 , j6 , j7 , j8 , j9 ).
INPUT:
j 1, ..., j 9 - integer or half integer
prec - precision, default: None. Providing a precision can drastically speed up the
calculation.
1628
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
OUTPUT:
Rational number times the square root of a rational number (if prec=None), or real number if a precision is given.
Examples
>>> from sympy.physics.wigner import wigner_9j
>>> wigner_9j(1,1,1, 1,1,1, 1,1,0 ,prec=64) # ==1/18
0.05555555...
It is an error to have arguments that are not integer or half integer values or do not fulll
the triangle relation:
sage: wigner 9j(0.5,0.5,0.5, 0.5,0.5,0.5, 0.5,0.5,0.5,prec=64)
Traceback (most recent call last):
...
ValueError: j values must be integer or half integer and fulfill the triangle relation
sage: wigner 9j(1,1,1, 0.5,1,1.5, 0.5,1,2.5,prec=64)
Traceback (most recent call last):
...
ValueError: j values must be integer or half integer and fulfill the triangle relation
ALGORITHM:
This function uses the algorithm of [Edmonds74] (page 1916) to calculate the value of the
3j symbol exactly. Note that the formula contains alternating sums over large factorials
and is therefore unsuitable for nite precision arithmetic and only useful for a computer
algebra system [Rasch03] (page 1915).
Units
Introduction
This module provides around 200 predened units that are commonly used in the sciences.
Additionally, it provides the Unit class which allows you to dene your own units.
Examples
All examples in this tutorial are computable, so one can just copy and paste them into a Python
shell and do something useful with them. All computations were done using the following
setup:
>>> from sympy.physics.units import *
Dimensionless quantities
less) quantities.
Provides variables for commonly written dimensionless (unit-
>>> 2*ten
20
>>> 20*percent
1/5
>>> 300*kilo*20*percent
5.36. Physics Module
1629
SymPy Documentation, Release 0.7.6
60000
>>> nano*deg
pi/180000000000
Base units The SI base units are dened variable name that are commonly used in written
and verbal communication. The singular abbreviated versions are what is used for display
purposes, but the plural non-abbreviated versions are dened in case it helps readability.
>>> 5*meters
5*m
>>> milli*kilogram
kg/1000
>>> gram
kg/1000
Note that British Imperial and U.S. customary units are not included. We strongly urge the
use of SI units; only Myanmar (Burma), Liberia, and the United States have not ocially
accepted the SI system.
Derived units
Common SI derived units.
>>> joule
kg*m**2/s**2
Docstring
Physical units and dimensions.
The base class is Unit, where all here dened units (200) inherit from.
The nd unit function can help you nd units for a given quantity:
>>> import sympy.physics.units as u
>>> u.find_unit(coul)
[coulomb, coulombs]
>>> u.find_unit(u.charge)
[C, charge, coulomb, coulombs]
>>> u.coulomb
A*s
Units are always given in terms of base units that have a name and an abbreviation:
>>> u.A.name
ampere
>>> u.ampere.abbrev
A
The generic name for a unit (like length, mass, etc...) can help you nd units:
>>> u.find_unit(magnet)
[magnetic_flux, magnetic_constant, magnetic_flux_density]
>>> u.find_unit(u.magnetic_flux)
[Wb, wb, weber, webers, magnetic_flux]
If, for a given session, you wish to add a unit you may do so:
1630
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> u.find_unit(gal)
[]
>>> u.gal = 4*u.quart
>>> u.gal/u.inch**3
231
To see a given quantity in terms of some other unit, divide by the desired unit:
>>> mph = u.miles/u.hours
>>> (u.m/u.s/mph).n(2)
2.2
The units are dened in terms of base units, so when you divide similar units you will obtain
a pure number. This means, for example, that if you divide a real-world mass (like grams) by
the atomic mass unit (amu) you will obtain Avogadros number. To obtain the answer in moles
you should divide by the unit avogadro:
>>> u.grams/u.amu
602214179000000000000000
>>> _/u.avogadro
mol
For chemical calculations the unit mmu (molar mass unit) has been dened so this conversion is
handled automatically. For example, the number of moles in 1 kg of water might be calculated
as:
>>> u.kg/(18*u.mmu).n(3)
55.5*mol
If you need the number of atoms in a mol as a pure number you can use avogadro number but
if you need it as a dimensional quantity you should use avogadro constant. (avogadro is a
shorthand for the dimensional quantity.)
>>> u.avogadro_number
602214179000000000000000
>>> u.avogadro_constant
602214179000000000000000/mol
class sympy.physics.units.Unit
Base class for base unit of physical units.
>>> from sympy.physics.units import Unit
>>> Unit(meter, m)
m
Other units are derived from base units:
>>> import sympy.physics.units as u
>>> cm = u.m/100
>>> 100*u.cm
m
sympy.physics.units.find unit(quantity)
Return a list of matching units names. if quantity is a string units containing the string
quantity if quantity is a unit units having matching base units
5.36. Physics Module
1631
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.physics import units as u
>>> u.find_unit(charge)
[charge]
>>> u.find_unit(u.charge)
[C, charge, coulomb, coulombs]
>>> u.find_unit(volt)
[volt, volts, voltage]
>>> u.find_unit(u.inch**3)[:5]
[l, cl, dl, ml, liter]
High energy physics
Abstract
Contains docstrings for methods in high energy physics.
Gamma matrices
High energy physics
class sympy.physics.hep.gamma matrices. LorentzContainer
Helper to collect LorentzIndex indices in various dimensions.
It collects LorentzIndex TensorIndexType that have been implemented in the code, and
stores them in a dict()
class sympy.physics.hep.gamma matrices.GammaMatrixHead
Class to wrap a TensorHead for gamma matrices.
dim dimension of the gamma matrix. eps dim correction for dimensional regularization,
use None if not needed.
Examples
>>> from sympy.physics.hep.gamma_matrices import GammaMatrixHead
>>> from sympy.tensor.tensor import tensor_indices
>>> G = GammaMatrixHead()
>>> i = tensor_indices(i, G.LorentzIndex)
>>> G(i)
gamma(i, auto_left, -auto_right)
Note that there is already an instance of GammaMatrixHead in four dimensions: GammaMatrix, which is simply declare as
GammaMatrix = GammaMatrixHead()
>>> from sympy.physics.hep.gamma_matrices import GammaMatrix
>>> from sympy.tensor.tensor import tensor_indices
>>> i = tensor_indices(i, GammaMatrix.LorentzIndex)
>>> GammaMatrix(i)
gamma(i, auto_left, -auto_right)
1632
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
To access the metric tensor
>>> GammaMatrix.LorentzIndex.metric
metric(LorentzIndex,LorentzIndex)
static extract type tens(expression)
Extract from a TensExpr all elements of this type.
Returns two tensor expressions:
the rst contains all TensorHead of this type.
the second contains all remaining.
gamma trace(t)
trace of a single line of gamma matrices
Examples
>>> from sympy.physics.hep.gamma_matrices import GammaMatrix as G
>>> from sympy.tensor.tensor import tensor_indices, tensorhead
>>> p, q = tensorhead(p, q, [G.LorentzIndex], [[1]])
>>> i0,i1,i2,i3,i4,i5 = tensor_indices(i0:6, G.LorentzIndex)
>>> ps = p(i0)*G(-i0)
>>> qs = q(i0)*G(-i0)
>>> G.gamma_trace(G(i0)*G(i1))
4*metric(i0, i1)
>>> G.gamma_trace(ps*ps) - 4*p(i0)*p(-i0)
0
>>> G.gamma_trace(ps*qs + ps*ps) - 4*p(i0)*p(-i0) - 4*p(i0)*q(-i0)
0
static simplify gpgp(ex, sort=True)
simplify products G(i)*p(-i)*G(j)*p(-j) -> p(i)*p(-i)
Examples
>>> from sympy.physics.hep.gamma_matrices import GammaMatrix as G
>>> from sympy.tensor.tensor import tensor_indices, tensorhead
>>> p, q = tensorhead(p, q, [G.LorentzIndex], [[1]])
>>> i0,i1,i2,i3,i4,i5 = tensor_indices(i0:6, G.LorentzIndex)
>>> ps = p(i0)*G(-i0)
>>> qs = q(i0)*G(-i0)
>>> G.simplify_gpgp(ps*qs*qs)
gamma(-L_0, auto_left, -auto_right)*p(L_0)*q(L_1)*q(-L_1)
static simplify lines(ex)
simplify a product of gamma matrices
Examples
>>>
>>>
>>>
>>>
from sympy.physics.hep.gamma_matrices import GammaMatrix, DiracSpinorIndex
from sympy.tensor.tensor import tensor_indices
i0,i1,i2,i3,i4,i5 = tensor_indices(i0:6, GammaMatrix.LorentzIndex)
s0,s1,s2,s3,s4,s5,s6,s7 = tensor_indices(s0:8, DiracSpinorIndex)
5.36. Physics Module
1633
SymPy Documentation, Release 0.7.6
>>> G = GammaMatrix
>>> t = G(i1,s1,-s2)*G(i4,s7,-s6)*G(i2,s2,-s3)*G(i3,s4,-s5)*G(i5,s6,-s7)
>>> G.simplify_lines(t)
4*gamma(i3, s4, -s5)*gamma(i1, s1, -S_0)*gamma(i2, S_0, -s3)*metric(i4, i5)
The Physics Vector Module
Abstract
In this documentation the components of the sympy.physics.vector module have been discussed. vector has been written to facilitate the operations pertaining to 3-dimensional
vectors, as functions of time or otherwise, in sympy.physics.
References for Physics/Vector
Guide to Vector
Vector & ReferenceFrame In vector, vectors and reference frames are the building
blocks of dynamic systems. This document will describe these mathematically and describe
how to use them with this modules code.
Vector A vector is a geometric object that has a magnitude (or length) and a direction.
Vectors in 3-space are often represented on paper as:
Vector on page
Vector Algebra
Vector out
of page
Vector
into page
Vector algebra is the rst topic to be discussed.
Two vectors are said to be equal if and only if (i) they have the same same magnitude and
orientation.
Vector Operations Multiple algebraic operations can be done with vectors: addition between vectors, scalar multiplication, and vector multiplication.
Vector addition as based on the parallelogram law.
a+b
1634
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Vector addition is also commutative:
a+b=b+a
(a + b) + c = a + (b + c)
Scalar multiplication is the product of a vector and a scalar; the result is a vector with the
same orientation but whose magnitude is scaled by the scalar. Note that multiplication by
-1 is equivalent to rotating the vector by 180 degrees about an arbitrary axis in the plane
perpendicular to the vector.
2a
-a
A unit vector is simply a vector whose magnitude is equal to 1. Given any vector v we can
dene a unit vector as:
v
^
nv =
kvk
Note that every vector can be written as the product of a scalar and unit vector.
Three vector products are implemented in vector: the dot product, the cross product, and
the outer product.
The dot product operation maps two vectors to a scalar. It is dened as:
a b = kakkbk cos()
where is the angle between a and b.
The dot product of two unit vectors represent the magnitude of the common direction; for
other vectors, it is the product of the magnitude of the common direction and the two vectors
magnitudes. The dot product of two perpendicular is zero. The gure below shows some
examples:
a
a
ab=0
a
aa=1
c
45
a c = 1/sqrt(2)
The dot product is commutative:
ab=ba
5.36. Physics Module
1635
SymPy Documentation, Release 0.7.6
The cross product vector multiplication operation of two vectors returns a vector:
ab=c
The vector c has the following properties: its orientation is perpendicular to both a and b,
its magnitude is dened as kck = kakkbk sin() (where is the angle between a and b), and
has a sense dened by using the right hand rule between kakkbk. The gure below shows
this:
c / sqrt(2)
45
axa=0
The cross product has the following properties:
It is not commutative:
a b 6= b a
a b = b a
and not associative:
(a b) c 6= a (b c)
Two parallel vectors will have a zero cross product.
The outer product between two vectors will not be not be discussed here, but instead in the
inertia section (that is where it is used). Other useful vector properties and relationships are:
(a + b) = a + b
a (b + c) = a b + a c
a (b + c) = a b + a b
(a b) c gives the scalar triple product.
a (b c) does not work, as you cannot cross a vector and a scalar.
(a b) c = a (b c)
(a b) c = (b c) a = (c a) b
(a b) c = b(a c) a(b c)
a (b c) = b(a c) c(a b)
1636
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Alternative Representation If we have three non-coplanar unit vectors ^
nx , ^
ny , ^
nz , we can
represent any vector a as a = ax^
n x + ay ^
ny + az^
nz . In this situation ^
nx , ^
ny , ^
nz are referred to
as a basis. ax , ay , az are called the measure numbers. Usually the unit vectors are mutually
perpendicular, in which case we can refer to them as an orthonormal basis, and they are
usually right-handed.
To test equality between two vectors, now we can do the following. With vectors:
a = ax^
nx + ay^
ny + az^
nz
b = bx^
nx + by^
ny + bz^
nz
We can claim equality if: ax = bx , ay = by , az = bz .
Vector addition is then represented, for the same two vectors, as:
a + b = (ax + bx )^
nx + (ay + by )^
ny + (az + bz )^
nz
Multiplication operations are now dened as:
b = bx^
nx + by^
ny + bz^
nz
a b = ax bx + ay by + az bz
^
nx ^
ny ^
nz
a b = det ax ay az
bx by bz
ax ay az
(a b) c = det bx by bz
cx cy cz
To write a vector in a given basis, we can do the follow:
a = (a ^
nx )^
nx + (a ^
ny )^
ny + (a ^
nz )^
nz
Examples
Some numeric examples of these operations follow:
a=^
nx + 5^
ny
b=^
ny + ^
nz
a+b=^
nx + 6^
ny + ^
nz
ab=5
a ^
ny = 5
a ^
nz = 0
a b = 5^
nx ^
ny + ^
nz
b a = 5^
nx + ^
ny ^
nz
Vector Calculus To deal with the calculus of vectors with moving object, we have to introduce the concept of a reference frame. A classic example is a train moving along its tracks,
5.36. Physics Module
1637
SymPy Documentation, Release 0.7.6
with you and a friend inside. If both you and your friend are sitting, the relative velocity between the two of you is zero. From an observer outside the train, you will both have velocity
though.
We will now apply more rigor to this denition. A reference frame is a virtual platform
which we choose to observe vector quantities from. If we have a reference frame N, vector
a is said to be xed in the frame N if none of its properties ever change when observed from
N. We will typically assign a xed orthonormal basis vector set with each reference frame; N
will have ^
nx , ^
ny , ^
nz as its basis vectors.
Derivatives of Vectors A vector which is not xed in a reference frame therefore has changing properties when observed from that frame. Calculus is the study of change, and in order
to deal with the peculiarities of vectors xed and not xed in dierent reference frames, we
need to be more explicit in our denitions.
Fixed in:
d
c
A B
cx
dx
x
e
x
f
In the above gure, we have vectors c, d, e, f. If one were to take the derivative of e with
respect to :
de
d
it is not clear what the derivative is. If you are observing from frame A, it is clearly non-zero.
If you are observing from frame B, the derivative is zero. We will therefore introduce the
frame as part of the derivative notation:
A
de
d
B
de
d
A
dc
d
B
dc
d
1638
6= 0, the derivative of e with respect to in the reference frame A
= 0, the derivative of e with respect to in the reference frame B
= 0, the derivative of c with respect to in the reference frame A
6= 0, the derivative of c with respect to in the reference frame B
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Here are some additional properties of derivatives of vectors in specic frames:
A
d
(a + b) =
dt
A
d
a =
dt
A
d
(a b) =
dt
da A db
+
dt
dt
A
d
da
a+
dt
dt
A
A
da
db
b+a
dt
dt
Relating Sets of Basis Vectors We need to now dene the relationship between two different reference frames; or how to relate the basis vectors of one frame to another. We can
do this using a direction cosine matrix (DCM). The direction cosine matrix relates the basis
vectors of one frame to another, in the following fashion:
bx
^
ax
[
] ^
^
ay = A CB ^
by
^
^
az
bz
When two frames (say, A & B) are initially aligned, then one frame has all of its basis vectors
rotated around an axis which is aligned with a basis vector, we say the frames are related by
a simple rotation. The gure below shows this:
az
bz
A
ax
by
ay
bx
The above rotation is a simple rotation about the Z axis by an angle . Note that after the
rotation, the basis vectors ^
az and ^
bz are still aligned.
This rotation can be characterized by the following direction cosine matrix:
cos() sin() 0
A B
C = sin() cos() 0
0
0
1
Simple rotations about the X and Y axes are dened by:
1
0
0
DCM for x-axis rotation: 0 cos() sin()
0 sin() cos()
cos() 0 sin()
0
1
0
DCM for y-axis rotation:
sin() 0 cos()
5.36. Physics Module
1639
SymPy Documentation, Release 0.7.6
Rotation in the positive direction here will be dened by using the right-hand rule.
The direction cosine matrix is also involved with the denition of the dot product between
sets of basis vectors. If we have two reference frames with associated basis vectors, their
direction cosine matrix can be dened as:
^
ax ^
bx ^
ax ^
by ^
ax ^
bz
Cxx Cxy Cxz
Cyx Cyy Cyz = ^
ay ^
bx ^
ay ^
by ^
ay ^
bz
^
^
^
Czx Czy Czz
^
az bx ^
az by ^
az b z
Additionally, the direction cosine matrix is orthogonal, in that:
A
CB = (B CA )1
= (B CA )T
If we have reference frames A and B, which in this example have undergone a simple z-axis
rotation by an amount , we will have two sets of basis vectors. We can then dene two
vectors: a = ^
ax + ^
ay + ^
az and b = ^
bx + ^
by + ^
bz . If we wish to express b in the A frame, we do
the following:
b =^
bx + ^
by + ^
bz
[
]
[
]
[
]
b= ^
ax + ^
ay + ^
az
ax (^
bx + ^
by + ^
bz ) ^
ay (^
bx + ^
by + ^
bz ) ^
az (^
bx + ^
by + ^
bz ) ^
b = (cos() sin())^
ax + (sin() + cos())^
ay + ^
az
And if we wish to express a in the B, we do:
a =^
ax + ^
ay + ^
az
]
]
[
]
[
[
bz
by + ^
bx + ^
a= ^
bz (^
ax + ^
ay + ^
az ) ^
by (^
ax + ^
ay + ^
az ) ^
bx (^
ax + ^
ay + ^
az ) ^
a = (cos() + sin())^
bx + ( sin() + cos())^
by + ^
bz
Derivatives with Multiple Frames If we have reference frames A and B we will have
two sets of basis vectors. We can then dene two vectors: a = ax^
ax + ay^
ay + az^
az and b =
^
^
^
bx bx + by by + bz bz . If we want to take the derivative of b in the reference frame A, we must
rst express it in A, and the take the derivatives of the measure numbers:
d(b ^
ay )
db
d(b ^
ax )
d(b ^
az )
^
^
^
=
ax +
ay +
az +
dx
dx
dx
dx
Examples
1640
An example of vector calculus:
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
l
by
ay
ax
bx
In this example we have two bodies, each with an attached reference frame. We will say that
and x are functions of time. We wish to know the time derivative of vector c in both the A
and B frames.
First, we need to dene c; c = x^
bx + l^
by . This provides a denition in the B frame. We can
now do the following:
B
dc
dl
dx ^
bx + ^
by
=
dt
dt
dt
= x ^
bx
To take the derivative in the A frame, we have to
cos()
A B
0
C =
sin()
rst relate the two frames:
0 sin()
1
0
0 cos()
Now we can do the following:
d(c ^
ay )
dc
d(c ^
ax )
d(c ^
az )
^
^
^
=
ax +
ay +
az
dt
dt
dt
dt
d(cos()x)
d(l)
d( sin()x)
^
^
^
=
ax +
ay +
az
dt
dt
dt
= ( sin()x + cos()x)^
ax + ( cos()x + sin()x)^
az
Note that this is the time derivative of c in A, and is expressed in the A frame. We can express
5.36. Physics Module
1641
SymPy Documentation, Release 0.7.6
it in the B frame however, and the expression will still be valid:
A
dc
= ( sin()x + cos()x)^
ax + ( cos()x + sin()x)^
az
dt
= x ^
bx x^
bz
Note the dierence in expression complexity between the two forms. They are equivalent, but
one is much simpler. This is an extremely important concept, as dening vectors in the more
complex forms can vastly slow down formulation of the equations of motion and increase their
length, sometimes to a point where they cannot be shown on screen.
Using Vectors and Reference Frames We have waited until after all of the relevant mathematical relationships have been dened for vectors and reference frames to introduce code.
This is due to how vectors are formed. When starting any problem in vector, one of the rst
steps is dening a reference frame (remember to import sympy.physics.vector rst):
>>> from sympy.physics.vector import *
>>> N = ReferenceFrame(N)
Now we have created a reference frame, N. To have access to any basis vectors, rst a
reference frame needs to be created. Now that we have made and object representing N, we
can access its basis vectors:
>>> N.x
N.x
>>> N.y
N.y
>>> N.z
N.z
Vector Algebra, in physics.vector
vectors.:
We can now do basic algebraic operations on these
>>> N.x == N.x
True
>>> N.x == N.y
False
>>> N.x + N.y
N.x + N.y
>>> 2 * N.x + N.y
2*N.x + N.y
Remember, dont add a scalar quantity to a vector (N.x + 5); this will raise an error. At this
point, well use SymPys Symbol in our vectors. Remember to refer to SymPys Gotchas and
Pitfalls when dealing with symbols.:
>>> from sympy import Symbol, symbols
>>> x = Symbol(x)
>>> x * N.x
x*N.x
>>> x*(N.x + N.y)
x*N.x + x*N.y
In vector multiple interfaces to vector multiplication have been implemented, at the operator
level, method level, and function level. The vector dot product can work as follows:
1642
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>>
1
>>>
0
>>>
1
>>>
0
>>>
1
>>>
0
N.x & N.x
N.x & N.y
N.x.dot(N.x)
N.x.dot(N.y)
dot(N.x, N.x)
dot(N.x, N.y)
The ocial interface is the function interface; this is what will be used in all examples. This
is to avoid confusion with the attribute and methods being next to each other, and in the case
of the operator operation priority. The operators used in vector for vector multiplication do
not posses the correct order of operations; this can lead to errors. Care with parentheses is
needed when using operators to represent vector multiplication.
The cross product is the other vector multiplication which will be discussed here. It oers
similar interfaces to the dot product, and comes with the same warnings.
>>> N.x ^ N.x
0
>>> N.x ^ N.y
N.z
>>> N.x.cross(N.x)
0
>>> N.x.cross(N.z)
- N.y
>>> cross(N.x, N.y)
N.z
>>> N.x ^ (N.y + N.z)
- N.y + N.z
Two additional operations can be done with vectors: normalizing the vector to length 1, and
getting its magnitude. These are done as follows:
>>> (N.x + N.y).normalize()
sqrt(2)/2*N.x + sqrt(2)/2*N.y
>>> (N.x + N.y).magnitude()
sqrt(2)
Vectors are often expressed in a matrix form, especially for numerical purposes. Since the
matrix form does not contain any information about the reference frame the vector is dened
in, you must provide a reference frame to extract the measure numbers from the vector. There
is a convenience function to do this:
>>> (x * N.x + 2 * x * N.y + 3 * x * N.z).to_matrix(N)
Matrix([
[ x],
[2*x],
[3*x]])
Vector Calculus, in physics.vector We have already introduced our rst reference frame.
We can take the derivative in that frame right now, if we desire:
5.36. Physics Module
1643
SymPy Documentation, Release 0.7.6
>>> (x * N.x + N.y).diff(x, N)
N.x
SymPy has a diff function, but it does not currently work with vector Vectors, so please use
Vectors diff method. The reason for this is that when dierentiating a Vector, the frame
of reference must be specied in addition to what you are taking the derivative with respect
to; SymPys diff function doesnt t this mold.
The more interesting case arise with multiple reference frames. If we introduce a second
reference frame, A, we now have two frames. Note that at this point we can add components
of N and A together, but cannot perform vector multiplication, as no relationship between the
two frames has been dened.
>>> A = ReferenceFrame(A)
>>> A.x + N.x
A.x + N.x
If we want to do vector multiplication, rst we have to dene and orientation. The orient
method of ReferenceFrame provides that functionality.
>>> A.orient(N, Axis, [x, N.y])
If we desire, we can view the DCM between these two frames at any time. This can be
calculated with the dcm method. This code: N.dcm(A) gives the dcm N CA .
This orients the A frame relative to the N frame by a simple rotation around the Y axis, by an
amount x. Other, more complicated rotation types include Body rotations, Space rotations,
quaternions, and arbitrary axis rotations. Body and space rotations are equivalent to doing 3
simple rotations in a row, each about a basis vector in the new frame. An example follows:
>>> N = ReferenceFrame(N)
>>> Bp = ReferenceFrame(Bp)
>>> Bpp = ReferenceFrame(Bpp)
>>> B = ReferenceFrame(B)
>>> q1,q2,q3 = symbols(q1 q2 q3)
>>> Bpp.orient(N,Axis, [q1, N.x])
>>> Bp.orient(Bpp,Axis, [q2, Bpp.y])
>>> B.orient(Bp,Axis, [q3, Bp.z])
>>> N.dcm(B)
Matrix([
[
cos(q2)*cos(q3),
-sin(q3)*cos(q2),
sin(
[sin(q1)*sin(q2)*cos(q3) + sin(q3)*cos(q1), -sin(q1)*sin(q2)*sin(q3) + cos(q1)*cos(q3), -sin(q1)*cos(
[sin(q1)*sin(q3) - sin(q2)*cos(q1)*cos(q3), sin(q1)*cos(q3) + sin(q2)*sin(q3)*cos(q1), cos(q1)*cos(
>>> B.orient(N,Body,[q1,q2,q3],XYZ)
>>> N.dcm(B)
Matrix([
[
cos(q2)*cos(q3),
-sin(q3)*cos(q2),
sin(
[sin(q1)*sin(q2)*cos(q3) + sin(q3)*cos(q1), -sin(q1)*sin(q2)*sin(q3) + cos(q1)*cos(q3), -sin(q1)*cos(
[sin(q1)*sin(q3) - sin(q2)*cos(q1)*cos(q3), sin(q1)*cos(q3) + sin(q2)*sin(q3)*cos(q1), cos(q1)*cos(
Space orientations are similar to body orientation, but applied from the frame to body. Body
and space rotations can involve either two or three axes: XYZ works, as does YZX, ZXZ,
YXY, etc. What is key is that each simple rotation is about a dierent axis than the previous
one; ZZX does not completely orient a set of basis vectors in 3 space.
Sometimes it will be more convenient to create a new reference frame and orient relative to an
existing one in one step. The orientnew method allows for this functionality, and essentially
wraps the orient method. All of the things you can do in orient, you can do in orientnew.
1644
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> C = N.orientnew(C, Axis, [q1, N.x])
Quaternions (or Euler Parameters) use 4 value to characterize the orientation of the frame.
This and arbitrary axis rotations are described in the orient and orientnew method help, or
in the references [Kane1983] (page 1916).
Finally, before starting multiframe calculus operations, we will introduce another vector tool:
dynamicsymbols. dynamicsymbols is a shortcut function to create undened functions of time
within SymPy. The derivative of such a dynamicsymbol is shown below.
>>> from sympy import diff
>>> q1, q2, q3 = dynamicsymbols(q1 q2 q3)
>>> diff(q1, Symbol(t))
Derivative(q1(t), t)
The dynamicsymbol printing is not very clear above; we will also introduce a few other tools
here. We can use vprint instead of print for non-interactive sessions.
>>> q1
q1(t)
>>> q1d = diff(q1, Symbol(t))
>>> vprint(q1)
q1
>>> vprint(q1d)
q1
For interactive sessions use init vprinting. There also exist analogs for SymPys vprint,
vpprint, and latex, vlatex.
>>>
>>>
>>>
q1
>>>
q1
from sympy.physics.vector import init_vprinting
init_vprinting(pretty_print=False)
q1
q1d
A dynamicsymbol should be used to represent any time varying quantity in vector, whether
it is a coordinate, varying position, or force. The primary use of a dynamicsymbol is for
speeds and coordinates (of which there will be more discussion in the Kinematics Section of
the documentation).
Now we will dene the orientation of our new frames with a dynamicsymbol, and can take
derivatives and time derivatives with ease. Some examples follow.
>>> N = ReferenceFrame(N)
>>> B = N.orientnew(B, Axis, [q1, N.x])
>>> (B.y*q2 + B.z).diff(q2, N)
B.y
>>> (B.y*q2 + B.z).dt(N)
(-q1 + q2)*B.y + q2*q1*B.z
Note that the output vectors are kept in the same frames that they were provided in. This
remains true for vectors with components made of basis vectors from multiple frames:
>>> (B.y*q2 + B.z + q2*N.x).diff(q2, N)
B.y + N.x
5.36. Physics Module
1645
SymPy Documentation, Release 0.7.6
How Vectors are Coded What follows is a short description of how vectors are dened by
the code in vector. It is provided for those who want to learn more about how this part of
sympy.physics.vector works, and does not need to be read to use this module; dont read it
unless you want to learn how this module was implemented.
Every Vectors main information is stored in the args attribute, which stores the three measure numbers for each basis vector in a frame, for every relevant frame. A vector does not
exist in code until a ReferenceFrame is created. At this point, the x, y, and z attributes of the
reference frame are immutable Vectors which have measure numbers of [1,0,0], [0,1,0], and
[0,0,1] associated with that ReferenceFrame. Once these vectors are accessible, new vectors
can be created by doing algebraic operations with the basis vectors. A vector can have components from multiple frames though. That is why args is a list; it has as many elements in
the list as there are unique ReferenceFrames in its components, i.e. if there are A and B frame
basis vectors in our new vector, args is of length 2; if it has A, B, and C frame basis vector,
args is of length three.
Each element in the args list is a 2-tuple; the rst element is a SymPy Matrix (this is where
the measure numbers for each set of basis vectors are stored) and the second element is a
ReferenceFrame to associate those measure numbers with.
ReferenceFrame stores a few things. First, it stores the name you supply it on creation (name
attribute). It also stores the direction cosine matrices, dened upon creation with the orientnew method, or calling the orient method after creation. The direction cosine matrices
are represented by SymPys Matrix, and are part of a dictionary where the keys are the ReferenceFrame and the value the Matrix; these are set bi-directionally; in that when you orient
A to N you are setting As orientation dictionary to include N and its Matrix, but also you
also are setting Ns orientation dictionary to include A and its Matrix (that DCM being the
transpose of the other).
Vector: Kinematics This document will give some mathematical background to describing
a systems kinematics as well as how to represent the kinematics in physics.vector.
Introduction to Kinematics The rst topic is rigid motion kinematics. A rigid body is an
idealized representation of a physical object which has mass and rotational inertia. Rigid
bodies are obviously not exible. We can break down rigid body motion into translational motion, and rotational motion (when dealing with particles, we only have translational motion).
Rotational motion can further be broken down into simple rotations and general rotations.
Translation of a rigid body is dened as a motion where the orientation of the body does not
change during the motion; or during the motion any line segment would be parallel to itself
at the start of the motion.
Simple rotations are rotations in which the orientation of the body may change, but there is
always one line which remains parallel to itself at the start of the motion.
General rotations are rotations which there is not always one line parallel to itself at the start
of the motion.
Angular Velocity The angular velocity of a rigid body refers to the rate of change of its
orientation. The angular velocity of a body is written down as: N B , or the angular velocity of
B in N, which is a vector. Note that here, the term rigid body was used, but reference frames
can also have angular velocities. Further discussion of the distinction between a rigid body
and a reference frame will occur later when describing the code representation.
Angular velocity is dened as being positive in the direction which causes the orientation
angles to increase (for simple rotations, or series of simple rotations).
1646
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
nx
nx
nz
ny
wB=q nx
The angular velocity vector represents the time derivative of the orientation. As a time derivative vector quantity, like those covered in the Vector & ReferenceFrame documentation, this
quantity (angular velocity) needs to be dened in a reference frame. That is what the N is in
the above denition of angular velocity; the frame in which the angular velocity is dened in.
The angular velocity of B in N can also be dened by:
N
B = (
N ^
N ^
d^
by ^ ^
dbz ^ ^
dbx ^ ^
bz )bx + (
bx )by + (
by )bz
dt
dt
dt
It is also common for a bodys angular velocity to be written as:
N
B = wx^
bx + wy^
by + wz^
bz
There are a few additional important points relating to angular velocity. The rst is the addition theorem for angular velocities, a way of relating the angular velocities of multiple bodies
and frames. The theorem follows:
N
D =N A +A B +B C +C D
This is also shown in the following example:
5.36. Physics Module
1647
SymPy Documentation, Release 0.7.6
N
nx
nz
ax
ny
q1
q2
cy
bz
q3
A = 0
B = q1^
ax
B C
= q2^
bz
C
D = q3^
cy
D = q1^
ax q2^
bz + q3^
cy
Note the signs used in the angular velocity denitions, which are related to how the displacement angle is dened in this case.
This theorem makes dening angular velocities of multibody systems much easier, as the
angular velocity of a body in a chain needs to only be dened to the previous body in order to
be fully dened (and the rst body needs to be dened in the desired reference frame). The
following gure shows an example of when using this theorem can make things easier.
p2
w1
w2
w3
p3
p1
Here we can easily write the angular velocity of the body D in the reference frame of the rst
body A:
A
^1 + w2 p^2 + w3 p
^3
D = w1 p
It is very important to remember to only use this with angular velocities; you cannot use this
theorem with the velocities of points.
1648
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
There is another theorem commonly used: the derivative theorem. It provides an alternative
method (which can be easier) to calculate the time derivative of a vector in a reference frame:
N
B
dv
dv N B
=
+ v
dt
dt
The vector v can be any vector quantity: a position vector, a velocity vector, angular velocity
vector, etc. Instead of taking the time derivative of the vector in N, we take it in B, where B
can be any reference frame or body, usually one in which it is easy to take the derivative on v
in (v is usually composed only of the basis vector set belonging to B). Then we add the cross
product of the angular velocity of our newer frame, N B and our vector quantity v. Again, you
can choose any alternative frame for this. Examples follow:
Angular Acceleration Angular acceleration refers to the time rate of change of the angular
velocity vector. Just as the angular velocity vector is for a body and is specied in a frame,
the angular acceleration vector is for a body and is specied in a frame: N B , or the angular
acceleration of B in N, which is a vector.
Calculating the angular acceleration is relatively straight forward:
N B
N N
d B
dt
Note that this can be calculated with the derivative theorem, and when the angular velocity
is dened in a body xed frame, becomes quite simple:
N B
N N
d B
dt
d B N B N B
+
dt
N
if B = wx^
bx + wy^
by + wz^
bz
N B
B N
then
N B
B N
d B
+
dt
B N B
{z
}
this is 0 by denition
dwx ^
dwy ^
dwz ^
N B
=
bx +
by +
bz
dt
dt
dt
N B
= wx^
bx + wy^
by + wz^
bz
Again, this is only for the case in which the angular velocity of the body is dened in body
xed components.
Point Velocity & Acceleration Consider a point, P : we can dene some characteristics of
the point. First, we can dene a position vector from some other point to P . Second, we can
dene the velocity vector of P in a reference frame of our choice. Third, we can dene the
acceleration vector of P in a reference frame of our choice.
These three quantities are read as:
rOP , the position vector from O to P
N P
v , the velocity of P in the reference frame N
N P
a , the acceleration of P in the reference frame N
5.36. Physics Module
1649
SymPy Documentation, Release 0.7.6
Note that the position vector does not have a frame associated with it; this is because there
is no time derivative involved, unlike the velocity and acceleration vectors.
We can nd these quantities for a simple example easily:
rOP
Lets dene: rOP = qx^
nx + qy^
ny
N
drOP
dt
N P
then we can calculate: v = qx^
nx + qy^
ny
N P
v =
N N P
d v
dt
N P
a = qx^
nx + qy^
ny
N
and : aP =
It is critical to understand in the above example that the point O is xed in the reference frame
N. There is no addition theorem for translational velocities; alternatives will be discussed
later though. Also note that the position of every point might not always need to be dened
to form the dynamic equations of motion. When you dont want to dene the position vector
of a point, you can start by just dening the velocity vector. For the above example:
Let us instead dene the velocity vector as:
N P
then acceleration can be written as:
N P
v = ux^
nx + uy^
ny
a = u x^
nx + u y^
ny
There will often be cases when the velocity of a point is desired and a related points velocity
is known. For the cases in which we have two points xed on a rigid body, we use the 2-Point
Theorem:
1650
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
xed in B
xed in B
xed in N
Lets say we know the velocity of the point S and the angular velocity of the body B, both
dened in the reference frame N. We can calculate the velocity and acceleration of the point
P in N as follows:
v =N vS +N B rSP
N P
a =N aS +N B rSP +N B (N B rSP )
N P
When only one of the two points is xed on a body, the 1 point theorem is used instead.
not xed in B
xed in B
xed in N
Here, the velocity of point S is known in the frame N, the angular velocity of B is known in
N, and the velocity of the point P is known in the frame associated with body B. We can then
5.36. Physics Module
1651
SymPy Documentation, Release 0.7.6
write the velocity and acceleration of P in N as:
v =B vP +N vS +N B rSP
N P
a =B aS +N aO +N B rSP +N B (N B rSP ) + 2N B B vP
N P
Examples of applications of the 1 point and 2 point theorem follow.
bz
B
nz
N
nx
by
bx
P
rOP
ny
This example has a disc translating and rotating in a plane. We can easily dene the angular
velocity of the body B and velocity of the point O:
N
B = u3^
nz = u3^
bz
N O
v = u1^
nx + u2^
ny
and accelerations can be written as:
= u3^
nz = u3^
bz
N B
N O
a = u1^
nx + u2^
ny
We can use the 2 point theorem to calculate the velocity and acceleration of point P now.
rOP = R^
bx
v =N vO +N B rOP
N P
v = u1^
nx + u2^
ny + u3^
bz R^
bx = u1^
nx + u2^
ny + u3 R^
by
N P
a =N aO +N B rOP +N B (N B rOP )
N P
a = u1^
nx + u2^
ny + u3^
bz R^
bx + u3^
bz (u3^
bz R^
bx )
N P
a = u1^
nx + u2^
ny + Ru3^
by Ru23^
bx
N P
N
O
nx
1652
ny
by
cy
bx
cx
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
In this example we have a double pendulum. We can use the two point theorem twice here in
order to nd the velocity of points Q and P ; point Os velocity is zero in N.
rOQ = l^
bx
rQP = l^
cx
N B
= u1^
bz
N
C = u2^
cz
v =N vO +N B rOQ
N Q
v = u1 l^
by
N Q
v =N vQ +N C rQP
N Q
v = u1 l^
by + u2^
cz l^
cx
N Q
v = u1 l^
by + u2 l^
cy
N P
nz
N
O
nx
ny
q1
Q
q2
P
cz
cy
cx
In this example we have a particle moving on a ring; the ring is supported by a rod which can
rotate about the ^
nx axis. First we use the two point theorem to nd the velocity of the center
point of the ring, Q, then use the 1 point theorem to nd the velocity of the particle on the
5.36. Physics Module
1653
SymPy Documentation, Release 0.7.6
ring.
N
C = u1^
nx
rOQ = l^
cz
N Q
v = u1 l^
cy
QP
= R(cos(q2 )^
cx + sin(q2 )^
cy )
C P
v = Ru2 (sin(q2 )^
cx + cos(q2 )^
cy )
v =C vP +N vQ +N C rQP
N P
v = Ru2 (sin(q2 )^
cx + cos(q2 )^
cy ) + u1 l^
cy + u1^
cx R(cos(q2 )^
cx + sin(q2 )^
cy
N P
v = Ru2 sin(q2 )^
cx + (Ru2 cos(q2 ) + u1 l)^
cy + Ru1 sin(q2 )^
cz
N P
A nal topic in the description of velocities of points is that of rolling, or rather, rolling without
slip. Two bodies are said to be rolling without slip if and only if the point of contact on each
body has the same velocity in another frame. See the following gure:
N
P
Pa is the contact point on body A
Pb is the contact point on body B
Rolling without slip i:
N Pa
= NvPb
This is commonly used to form the velocity of a point on one object rolling on another xed
object, such as in the following example:
Kinematics in physics.vector It should be clear by now that the topic of kinematics here
has been mostly describing the correct way to manipulate vectors into representing the velocities of points. Within vector there are convenient methods for storing these velocities
associated with frames and points. Well now revisit the above examples and show how to
represent them in sympy (page 704).
The topic of reference frame creation has already been covered. When a ReferenceFrame is
created though, it automatically calculates the angular velocity of the frame using the time
derivative of the DCM and the angular velocity denition.
>>> from sympy import Symbol, sin, cos
>>> from sympy.physics.vector import *
>>> N = ReferenceFrame(N)
>>> q1 = dynamicsymbols(q1)
>>> A = N.orientnew(A, Axis, [q1, N.x])
>>> A.ang_vel_in(N)
q1*N.x
Note that the angular velocity can be dened in an alternate way:
1654
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> B = ReferenceFrame(B)
>>> u1 = dynamicsymbols(u1)
>>> B.set_ang_vel(N, u1 * B.y)
>>> B.ang_vel_in(N)
u1*B.y
>>> N.ang_vel_in(B)
- u1*B.y
Both upon frame creation during orientnew and when calling set ang vel, the angular velocity is set in both frames involved, as seen above.
nz
nx
ax
ny
q1
q2
cy
bz
q3
Here we have multiple bodies with angular velocities dened relative to each other. This is
coded as:
>>> N = ReferenceFrame(N)
>>> A = ReferenceFrame(A)
>>> B = ReferenceFrame(B)
>>> C = ReferenceFrame(C)
>>> D = ReferenceFrame(D)
>>> u1, u2, u3 = dynamicsymbols(u1 u2 u3)
>>> A.set_ang_vel(N, 0)
>>> B.set_ang_vel(A, u1 * A.x)
>>> C.set_ang_vel(B, -u2 * B.z)
>>> D.set_ang_vel(C, u3 * C.y)
>>> D.ang_vel_in(N)
u3*C.y - u2*B.z + u1*A.x
In vector the shortest path between two frames is used when nding the angular velocity.
That would mean if we went back and set:
>>> D.set_ang_vel(N, 0)
>>> D.ang_vel_in(N)
0
The path that was just dened is what is used. This can cause problems though, as now the
angular velocity denitions are inconsistent. It is recommended that you avoid doing this.
Points are a translational analog to the rotational ReferenceFrame. Creating a Point can be
done in two ways, like ReferenceFrame:
5.36. Physics Module
1655
SymPy Documentation, Release 0.7.6
>>> O = Point(O)
>>> P = O.locatenew(P, 3 * N.x + N.y)
>>> P.pos_from(O)
3*N.x + N.y
>>> Q = Point(Q)
>>> Q.set_pos(P, N.z)
>>> Q.pos_from(P)
N.z
>>> Q.pos_from(O)
3*N.x + N.y + N.z
Similar to ReferenceFrame, the position vector between two points is found by the shortest
path (number of intermediate points) between them. Unlike rotational motion, there is no
addition theorem for the velocity of points. In order to have the velocity of a Point in a
ReferenceFrame, you have to set the value.
>>> O = Point(O)
>>> O.set_vel(N, u1*N.x)
>>> O.vel(N)
u1*N.x
For both translational and rotational accelerations, the value is computed by taking the time
derivative of the appropriate velocity, unless the user sets it otherwise.
>>> O.acc(N)
u1*N.x
>>> O.set_acc(N, u2*u1*N.y)
>>> O.acc(N)
u1*u2*N.y
Next is a description of the 2 point and 1 point theorems, as used in sympy.
bz
B
nz
N
nx
by
bx
ny
P
rOP
First is the translating, rotating disc.
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
N = ReferenceFrame(N)
u1, u2, u3 = dynamicsymbols(u1 u2 u3)
R = Symbol(R)
B = ReferenceFrame(B)
O = Point(O)
O.set_vel(N, u1 * N.x + u2 * N.y)
P = O.locatenew(P, R * B.x)
B.set_ang_vel(N, u3 * B.z)
P.v2pt_theory(O, N, B)
1656
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
u1*N.x + u2*N.y + R*u3*B.y
>>> P.a2pt_theory(O, N, B)
u1*N.x + u2*N.y - R*u3**2*B.x + R*u3*B.y
We will also cover implementation of the 1 point theorem.
nz
N
O
nx
ny
q1
Q
q2
P
cz
cy
cx
This is the particle moving on a ring, again.
>>> N = ReferenceFrame(N)
>>> u1, u2 = dynamicsymbols(u1 u2)
>>> q1, q2 = dynamicsymbols(q1 q2)
>>> l = Symbol(l)
>>> R = Symbol(R)
>>> C = N.orientnew(C, Axis, [q1, N.x])
>>> C.set_ang_vel(N, u1 * N.x)
>>> O = Point(O)
>>> O.set_vel(N, 0)
>>> Q = O.locatenew(Q, -l * C.z)
>>> P = Q.locatenew(P, R * (cos(q2) * C.x + sin(q2) * C.y))
>>> P.set_vel(C, R * u2 * (-sin(q2) * C.x + cos(q2) * C.y))
>>> Q.v2pt_theory(O, N, C)
l*u1*C.y
>>> P.v1pt_theory(Q, N, C)
- R*u2*sin(q2)*C.x + (R*u2*cos(q2) + l*u1)*C.y + R*u1*sin(q2)*C.z
Potential Issues/Advanced Topics/Future Features in Physics/Vector Module This
document will describe some of the more advanced functionality that this module oers but
which is not part of the ocial interface. Here, some of the features that will be implemented in the future will also be covered, along with unanswered questions about proper
functionality. Also, common problems will be discussed, along with some solutions.
5.36. Physics Module
1657
SymPy Documentation, Release 0.7.6
Inertia (Dyadics) A dyadic tensor is a second order tensor formed by the juxtaposition of a
pair of vectors. There are various operations dened with respect to dyadics, which have been
implemented in vector in the form of class Dyadic. To know more, refer to the Dyadic and
Vector class APIs. Dyadics are used to dene the inertia of bodies within mechanics. Inertia
dyadics can be dened explicitly but the inertia function is typically much more convenient
for the user:
>>> from sympy.physics.mechanics import ReferenceFrame, inertia
>>> N = ReferenceFrame(N)
Supply a reference frame and the moments of inertia if the object
is symmetrical:
>>> inertia(N, 1, 2, 3)
(N.x|N.x) + 2*(N.y|N.y) + 3*(N.z|N.z)
Supply a reference frame along with the products and moments of inertia
for a general object:
>>> inertia(N, 1, 2, 3, 4, 5, 6)
(N.x|N.x) + 4*(N.x|N.y) + 6*(N.x|N.z) + 4*(N.y|N.x) + 2*(N.y|N.y) + 5*(N.y|N.z) + 6*(N.z|N.x) + 5*(N.z|N.
Notice that the inertia function returns a dyadic with each component represented as two
unit vectors separated by a |. Refer to the Dyadic (page 1695) section for more information
about dyadics.
Inertia is often expressed in a matrix, or tensor, form, especially for numerical purposes.
Since the matrix form does not contain any information about the reference frame(s) the
inertia dyadic is dened in, you must provide one or two reference frames to extract the
measure numbers from the dyadic. There is a convenience function to do this:
>>> inertia(N, 1, 2, 3, 4, 5, 6).to_matrix(N)
Matrix([
[1, 4, 6],
[4, 2, 5],
[6, 5, 3]])
Common Issues Here issues with numerically integrating code, choice of dynamicsymbols
for coordinate and speed representation, printing, dierentiating, and substitution will occur.
Printing The default printing options are to use sorting for Vector and Dyadic measure
numbers, and have unsorted output from the vprint, vpprint, and vlatex functions. If you
are printing something large, please use one of those functions, as the sorting can increase
printing time from seconds to minutes.
Substitution Substitution into large expressions can be slow, and take a few minutes.
Acceleration of Points At a minimum, points need to have their velocities dened, as the
acceleration can be calculated by taking the time derivative of the velocity in the same frame.
If the 1 point or 2 point theorems were used to compute the velocity, the time derivative of the
velocity expression will most likely be more complex than if you were to use the acceleration
level 1 point and 2 point theorems. Using the acceleration level methods can result in shorted
expressions at this point, which will result in shorter expressions later (such as when forming
Kanes equations).
1658
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Advanced Interfaces Here we will cover advanced options in: ReferenceFrame, dynamicsymbols, and some associated functionality.
ReferenceFrame ReferenceFrame is shown as having a .name attribute and .x, .y, and .z
attributes for accessing the basis vectors, as well as a fairly rigidly dened print output. If
you wish to have a dierent set of indices dened, there is an option for this. This will also
require a dierent interface for accessing the basis vectors.
>>> from sympy.physics.vector import ReferenceFrame, vprint, vpprint, vlatex
>>> N = ReferenceFrame(N, indices=[i, j, k])
>>> N[i]
N[i]
>>> N.x
N[i]
>>> vlatex(N.x)
\\mathbf{\\hat{n}_{i}}
Also, the latex output can have custom strings; rather than just indices though, the entirety
of each basis vector can be specied. The custom latex strings can occur without custom
indices, and also overwrites the latex string that would be used if there were custom indices.
>>> from sympy.physics.vector import ReferenceFrame, vlatex
>>> N = ReferenceFrame(N, latexs=[n1,\mathbf{n}_2,cat])
>>> vlatex(N.x)
n1
>>> vlatex(N.y)
\\mathbf{n}_2
>>> vlatex(N.z)
cat
dynamicsymbols The dynamicsymbols function also has hidden functionality; the variable
which is associated with time can be changed, as well as the notation for printing derivatives.
>>> from sympy import symbols
>>> from sympy.physics.vector import dynamicsymbols, vprint
>>> q1 = dynamicsymbols(q1)
>>> q1
q1(t)
>>> dynamicsymbols._t = symbols(T)
>>> q2 = dynamicsymbols(q2)
>>> q2
q2(T)
>>> q1
q1(t)
>>> q1d = dynamicsymbols(q1, 1)
>>> vprint(q1d)
q1
>>> dynamicsymbols._str = d
>>> vprint(q1d)
q1d
>>> dynamicsymbols._str = \
>>> dynamicsymbols._t = symbols(t)
Note that only dynamic symbols created after the change are dierent. The same is not true
for the .s tr attribute; this aects the printing output only, so dynamic symbols created before
or after will print the same way.
5.36. Physics Module
1659
SymPy Documentation, Release 0.7.6
Also note that Vectors .dt method uses the . t attribute of dynamicsymbols, along with a
number of other important functions and methods. Dont mix and match symbols representing
time.
Scalar and Vector Field Functionality
Introduction
Vectors and Scalars In physics, we deal with two kinds of quantities scalars and vectors.
A scalar is an entity which only has a magnitude no direction. Examples of scalar quantities
include mass, electric charge, temperature, distance, etc.
A vector, on the other hand, is an entity that is characterized by a magnitude and a direction.
Examples of vector quantities are displacement, velocity, magnetic eld, etc.
A scalar can be depicted just by a number, for e.g. a temperature of 300 K. On the other
hand, vectorial quantities like acceleration are usually denoted by a vector. Given a vector
V, the magnitude of the corresponding quantity can be calculated as the magnitude of the
vector itself kVk, while the direction would be specied by a unit vector in the direction of the
V
original vector, ^
V = kVk
.
For example, consider a displacement of (3^
i + 4^
j + 5^
k) m, where , as per standard convention,
^
i, ^
j and ^
k represent unit vectors in the X, Y and Z directions
respectively. Therefore, it can
be concluded that the distance traveled is k3^
i + 4^
j + 5^
kk m = 5 2 m. The direction of travel is
3 ^
4 ^
5 ^
given by the unit vector 5
i + 5
j + 5
k.
2
2
2
Fields In general, a f ield is a vector or scalar quantity that can be specied everywhere in
space as a function of position (Note that in general a eld may also be dependent on time
and other custom variables). In this module, we deal with 3-dimensional spaces only. Hence,
a eld is dened as a function of the x, y and z coordinates corresponding to a location in 3D
space.
For example, temperate in 3 dimensional space (a temperature eld) can be written as T (x, y, z)
a scalar function of the position. An example of a scalar eld in electromagnetism is the
electric potential.
In a similar manner, a vector eld can be dened as a vectorial function of the location (x, y, z)
of any point in space.
For instance, every point on the earth may be considered to be in the gravitational force eld
of the earth. We may specify the eld by the magnitude and the direction of acceleration due
to gravity (i.e. force per unit mass ) g(x, y, z) at every point in space.
To give an example from electromagnetism, consider an electric potential of form 2x2 y, a
scalar eld in 3D space. The corresponding conservative electric eld can be computed as
2^
the gradient of the electric potential function, and expressed as 4xy^
i + 2x
j. The magnitude of
this electric eld can in turn be expressed as a scalar eld of the form 4x4 + 16x2 y 2 .
Implementation of elds in sympy.physics.vector In sympy.physics.vector, every ReferenceFrame instance is assigned basis vectors corresponding to the X, Y and Z directions.
These can be accessed using the attributes named x, y and z respectively. Hence, to dene
a vector v of the form 3^
i + 4^
j + 5^
k with respect to a given frame R, you would do
1660
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> from sympy.physics.vector import ReferenceFrame
>>> R = ReferenceFrame(R)
>>> v = 3*R.x + 4*R.y + 5*R.z
Vector math and basic calculus operations with respect to vectors have already been elaborated upon in other sections of this modules documentation.
On the other hand, base scalars (or coordinate variables) are implemented as special SymPy
Symbol s assigned to every frame, one for each direction from X, Y and Z. For a frame R, the
X, Y and Z base scalar Symbol s can be accessed using the R[0], R[1] and R[2] expressions
respectively.
Therefore, to generate the expression for the aforementioned electric potential eld 2x2 y, you
would have to do
>>> from sympy.physics.vector import ReferenceFrame
>>> R = ReferenceFrame(R)
>>> electric_potential = 2*R[0]**2*R[1]
>>> electric_potential
2*R_x**2*R_y
In string representation, R x denotes the X base scalar assigned to ReferenceFrame R. Essentially, R x is the string representation of R[0].
Scalar elds can be treated just as any other SymPy expression, for any math/calculus functionality. Hence, to dierentiate the above electric potential with respect to x (i.e. R[0]), you
would have to use the diff method.
>>> from sympy.physics.vector import ReferenceFrame
>>> R = ReferenceFrame(R)
>>> electric_potential = 2*R[0]**2*R[1]
>>> from sympy import diff
>>> diff(electric_potential, R[0])
4*R_x*R_y
Like vectors (and vector elds), scalar elds can also be re-expressed in other frames of
reference, apart from the one they were dened in assuming that an orientation relationship
exists between the concerned frames. This can be done using the express method, in a way
similar to vectors - but with the variables parameter set to True.
>>> from sympy.physics.vector import ReferenceFrame
>>> R = ReferenceFrame(R)
>>> electric_potential = 2*R[0]**2*R[1]
>>> from sympy.physics.vector import dynamicsymbols, express
>>> q = dynamicsymbols(q)
>>> R1 = R.orientnew(R1, rot_type = Axis, amounts = [q, R.z])
>>> express(electric_potential, R1, variables=True)
2*(R1_x*sin(q(t)) + R1_y*cos(q(t)))*(R1_x*cos(q(t)) - R1_y*sin(q(t)))**2
Moreover, considering scalars can also be functions of time just as vectors, dierentiation
with respect to time is also possible. Depending on the Symbol s present in the expression
and the frame with respect to which the time dierentiation is being done, the output will
change/remain the same.
>>>
>>>
>>>
>>>
>>>
from sympy.physics.vector import ReferenceFrame
R = ReferenceFrame(R)
electric_potential = 2*R[0]**2*R[1]
q = dynamicsymbols(q)
R1 = R.orientnew(R1, rot_type = Axis, amounts = [q, R.z])
5.36. Physics Module
1661
SymPy Documentation, Release 0.7.6
>>> from sympy.physics.vector import time_derivative
>>> time_derivative(electric_potential, R)
0
>>> time_derivative(electric_potential, R1).simplify()
(R1_x*cos(q(t)) - R1_y*sin(q(t)))*(3*R1_x**2*cos(2*q(t)) - R1_x**2 6*R1_x*R1_y*sin(2*q(t)) - 3*R1_y**2*cos(2*q(t)) - R1_y**2)*Derivative(q(t), t)
Field operators and other related functions Here we describe some basic eld-related
functionality implemented in sympy.physics.vector
Curl A curl is a mathematical operator that describes an innitesimal rotation of a vector
in 3D space. The direction is determined by the right-hand rule (along the axis of rotation),
and the magnitude is given by the magnitude of rotation.
In the 3D Cartesian system, the curl of a 3D vector F , denoted by F is given by (
)
(
)
)
Fy ^ ( Fx
Fy
Fz ^
Fx ^
z
F = F
i
+
j
+
k
y
z
z
x
x
y
where Fx denotes the X component of vector F.
To compute the curl of a vector eld in physics.vector, you would do
>>> from sympy.physics.vector import ReferenceFrame
>>> R = ReferenceFrame(R)
>>> from sympy.physics.vector import curl
>>> field = R[0]*R[1]*R[2]*R.x
>>> curl(field, R)
R_x*R_y*R.y - R_x*R_z*R.z
Divergence Divergence is a vector operator that measures the magnitude of a vector elds
source or sink at a given point, in terms of a signed scalar.
The divergence operator always returns a scalar after operating on a vector.
In the 3D Cartesian system, the divergence of a 3D vector F, denoted by F is given by F=
U
x
V
y
W
z
where U , V and W denote the X, Y and Z components of F respectively.
To compute the divergence of a vector eld in physics.vector, you would do
>>> from sympy.physics.vector import ReferenceFrame
>>> R = ReferenceFrame(R)
>>> from sympy.physics.vector import divergence
>>> field = R[0]*R[1]*R[2] * (R.x+R.y+R.z)
>>> divergence(field, R)
R_x*R_y + R_x*R_z + R_y*R_z
Gradient Consider a scalar eld f (x, y, z) in 3D space. The gradient of this eld is dened
as the vector of the 3 partial derivatives of f with respect to x, y and z in the X, Y and Z
directions respectively.
In the 3D Cartesian system, the divergence of a scalar eld f , denoted by f is given by f = f^
i + f^
j + f ^
k
x
1662
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
To compute the divergence of a vector eld in physics.vector, you would do
>>> from sympy.physics.vector import ReferenceFrame
>>> R = ReferenceFrame(R)
>>> from sympy.physics.vector import gradient
>>> scalar_field = R[0]*R[1]*R[2]
>>> gradient(scalar_field, R)
R_y*R_z*R.x + R_x*R_z*R.y + R_x*R_y*R.z
Conservative and Solenoidal elds In vector calculus, a conservative eld is a eld that is
the gradient of some scalar eld. Conservative elds have the property that their line integral
over any path depends only on the end-points, and is independent of the path between them.
A conservative vector eld is also said to be irrotational, since the curl of a conservative
eld is always zero.
In physics, conservative elds represent forces in physical systems where energy is conserved.
To check if a vector eld is conservative in physics.vector, use the is conservative function.
>>> from sympy.physics.vector import ReferenceFrame, is_conservative
>>> R = ReferenceFrame(R)
>>> field = R[1]*R[2]*R.x + R[0]*R[2]*R.y + R[0]*R[1]*R.z
>>> is_conservative(field)
True
>>> curl(field, R)
0
A solenoidal eld, on the other hand, is a vector eld whose divergence is zero at all points
in space.
To check if a vector eld is solenoidal in physics.vector, use the is solenoidal function.
>>> from sympy.physics.vector import ReferenceFrame, is_solenoidal
>>> R = ReferenceFrame(R)
>>> field = R[1]*R[2]*R.x + R[0]*R[2]*R.y + R[0]*R[1]*R.z
>>> is_solenoidal(field)
True
>>> divergence(field, R)
0
Scalar potential functions We have previously mentioned that every conservative eld
can be dened as the gradient of some scalar eld. This scalar eld is also called the scalar
potential eld corresponding to the aforementioned conservative eld.
The scalar potential function in physics.vector calculates the scalar potential eld corresponding to a given conservative vector eld in 3D space - minus the extra constant of
integration, of course.
Example of usage >>> from sympy.physics.vector import ReferenceFrame, scalar_potential
>>> R = ReferenceFrame(R)
>>> conservative_field = 4*R[0]*R[1]*R[2]*R.x + 2*R[0]**2*R[2]*R.y + 2*R[0]**2*R[1]*R.z
>>> scalar_potential(conservative_field, R)
2*R_x**2*R_y*R_z
5.36. Physics Module
1663
SymPy Documentation, Release 0.7.6
Providing a non-conservative vector eld as an argument to scalar potential raises a ValueError.
The scalar potential dierence, or simply potential dierence, corresponding to a conservative vector eld can be dened as the dierence between the values of its scalar potential
function at two points in space. This is useful in calculating a line integral with respect to a
conservative function, since it depends only on the endpoints of the path.
This computation is performed as follows in physics.vector.
>>>
>>>
>>>
>>>
>>>
>>>
>>>
4
from sympy.physics.vector import ReferenceFrame, Point
from sympy.physics.vector import scalar_potential_difference
R = ReferenceFrame(R)
O = Point(O)
P = O.locatenew(P, 1*R.x + 2*R.y + 3*R.z)
vectfield = 4*R[0]*R[1]*R.x + 2*R[0]**2*R.y
scalar_potential_difference(vectfield, R, O, P, O)
If provided with a scalar expression instead of a vector eld, scalar potential difference
returns the dierence between the values of that scalar eld at the two given points in space.
Vector API
Essential Classes
CoordinateSym
class sympy.physics.vector.frame.CoordinateSym
A coordinate symbol/base scalar associated wrt a Reference Frame.
Ideally, users should not instantiate this class. Instances of this class must only be accessed through the corresponding frame as frame[index].
CoordinateSyms having the same frame and index parameters are equal (even though
they may be instantiated separately).
Parameters name : string
The display name of the CoordinateSym
frame : ReferenceFrame
The reference frame this base scalar belongs to
index : 0, 1 or 2
The index of the dimension denoted by this coordinate variable
Examples
>>> from sympy.physics.vector import ReferenceFrame, CoordinateSym
>>> A = ReferenceFrame(A)
>>> A[1]
A_y
>>> type(A[0])
<class sympy.physics.vector.frame.CoordinateSym>
>>> a_y = CoordinateSym(a_y, A, 1)
>>> a_y == A[1]
True
1664
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
ReferenceFrame
class sympy.physics.vector.frame.ReferenceFrame(name,
indices=None,
latexs=None, variables=None)
A reference frame in classical mechanics.
ReferenceFrame is a class used to represent a reference frame in classical mechanics.
It has a standard basis of three unit vectors in the frames x, y, and z directions.
It also can have a rotation relative to a parent frame; this rotation is dened by a direction
cosine matrix relating this frames basis vectors to the parent frames basis vectors. It
can also have an angular velocity vector, dened in another frame.
ang acc in(otherframe)
Returns the angular acceleration Vector of the ReferenceFrame.
Eectively returns the Vector: N alpha B which represent the angular acceleration
of B in N, where B is self, and N is otherframe.
Parameters otherframe : ReferenceFrame
The ReferenceFrame which the angular acceleration is returned in.
Examples
>>> from sympy.physics.vector import ReferenceFrame, Vector
>>> N = ReferenceFrame(N)
>>> A = ReferenceFrame(A)
>>> V = 10 * N.x
>>> A.set_ang_acc(N, V)
>>> A.ang_acc_in(N)
10*N.x
ang vel in(otherframe)
Returns the angular velocity Vector of the ReferenceFrame.
Eectively returns the Vector: N omega B which represent the angular velocity of
B in N, where B is self, and N is otherframe.
Parameters otherframe : ReferenceFrame
The ReferenceFrame which the angular velocity is returned in.
Examples
>>> from sympy.physics.vector import ReferenceFrame, Vector
>>> N = ReferenceFrame(N)
>>> A = ReferenceFrame(A)
>>> V = 10 * N.x
>>> A.set_ang_vel(N, V)
>>> A.ang_vel_in(N)
10*N.x
dcm(otherframe)
The direction cosine matrix between frames.
This gives the DCM between this frame and the otherframe. The format is N.xyz =
N.dcm(B) * B.xyz A SymPy Matrix is returned.
Parameters otherframe : ReferenceFrame
5.36. Physics Module
1665
SymPy Documentation, Release 0.7.6
The otherframe which the DCM is generated to.
Examples
>>> from sympy.physics.vector import ReferenceFrame, Vector
>>> from sympy import symbols
>>> q1 = symbols(q1)
>>> N = ReferenceFrame(N)
>>> A = N.orientnew(A, Axis, [q1, N.x])
>>> N.dcm(A)
Matrix([
[1,
0,
0],
[0, cos(q1), -sin(q1)],
[0, sin(q1), cos(q1)]])
orient(parent, rot type, amounts, rot order=)
Denes the orientation of this frame relative to a parent frame.
Parameters parent : ReferenceFrame
The frame that this ReferenceFrame will have its orientation matrix
dened in relation to.
rot type : str
The type of orientation matrix that is being created. Supported types
are Body, Space, Quaternion, and Axis. See examples for correct
usage.
amounts : list OR value
The quantities that the orientation matrix will be dened by.
rot order : str
If applicable, the order of a series of rotations.
Examples
>>>
>>>
>>>
>>>
>>>
from sympy.physics.vector import ReferenceFrame, Vector
from sympy import symbols
q0, q1, q2, q3, q4 = symbols(q0 q1 q2 q3 q4)
N = ReferenceFrame(N)
B = ReferenceFrame(B)
Now we have a choice of how to implement the orientation. First is Body. Body
orientation takes this reference frame through three successive simple rotations.
Acceptable rotation orders are of length 3, expressed in XYZ or 123, and cannot
have a rotation about about an axis twice in a row.
>>> B.orient(N, Body, [q1, q2, q3], 123)
>>> B.orient(N, Body, [q1, q2, 0], ZXZ)
>>> B.orient(N, Body, [0, 0, 0], XYX)
Next is Space. Space is like Body, but the rotations are applied in the opposite order.
>>> B.orient(N, Space, [q1, q2, q3], 312)
1666
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Next is Quaternion. This orients the new ReferenceFrame with Quaternions, dened
as a nite rotation about lambda, a unit vector, by some amount theta. This orientation is described by four parameters: q0 = cos(theta/2) q1 = lambda x sin(theta/2)
q2 = lambda y sin(theta/2) q3 = lambda z sin(theta/2) Quaternion does not take in
a rotation order.
>>> B.orient(N, Quaternion, [q0, q1, q2, q3])
Last is Axis. This is a rotation about an arbitrary, non-time-varying axis by some
angle. The axis is supplied as a Vector. This is how simple rotations are dened.
>>> B.orient(N, Axis, [q1, N.x + 2 * N.y])
orientnew(newname, rot type, amounts, rot order=, variables=None,
dices=None, latexs=None)
Creates a new ReferenceFrame oriented with respect to this Frame.
in-
See ReferenceFrame.orient() for acceptable rotation types, amounts, and orders.
Parent is going to be self.
Parameters parent : ReferenceFrame
The frame that this ReferenceFrame will have its orientation matrix
dened in relation to.
rot type : str
The type of orientation matrix that is being created. Supported types
are Body, Space, Quaternion, and Axis. See examples for correct
usage.
amounts : list OR value
The quantities that the orientation matrix will be dened by.
rot order : str
If applicable, the order of a series of rotations.
Examples
>>>
>>>
>>>
>>>
>>>
from sympy.physics.vector import ReferenceFrame, Vector
from sympy import symbols
q0, q1, q2, q3, q4 = symbols(q0 q1 q2 q3 q4)
N = ReferenceFrame(N)
B = ReferenceFrame(B)
Now we have a choice of how to implement the orientation. First is Body. Body
orientation takes this reference frame through three successive simple rotations.
Acceptable rotation orders are of length 3, expressed in XYZ or 123, and cannot
have a rotation about about an axis twice in a row.
>>> B.orient(N, Body, [q1, q2, q3], 123)
>>> B.orient(N, Body, [q1, q2, 0], ZXZ)
>>> B.orient(N, Body, [0, 0, 0], XYX)
Next is Space. Space is like Body, but the rotations are applied in the opposite order.
>>> B.orient(N, Space, [q1, q2, q3], 312)
5.36. Physics Module
1667
SymPy Documentation, Release 0.7.6
Next is Quaternion. This orients the new ReferenceFrame with Quaternions, dened
as a nite rotation about lambda, a unit vector, by some amount theta. This orientation is described by four parameters: q0 = cos(theta/2) q1 = lambda x sin(theta/2)
q2 = lambda y sin(theta/2) q3 = lambda z sin(theta/2) Quaternion does not take in
a rotation order.
>>> B.orient(N, Quaternion, [q0, q1, q2, q3])
Last is Axis. This is a rotation about an arbitrary, non-time-varying axis by some
angle. The axis is supplied as a Vector. This is how simple rotations are dened.
>>> B.orient(N, Axis, [q1, N.x + 2 * N.y])
set ang acc(otherframe, value)
Dene the angular acceleration Vector in a ReferenceFrame.
Denes the angular acceleration of this ReferenceFrame, in another. Angular acceleration can be dened with respect to multiple dierent ReferenceFrames. Care
must be taken to not create loops which are inconsistent.
Parameters otherframe : ReferenceFrame
A ReferenceFrame to dene the angular acceleration in
value : Vector
The Vector representing angular acceleration
Examples
>>> from sympy.physics.vector import ReferenceFrame, Vector
>>> N = ReferenceFrame(N)
>>> A = ReferenceFrame(A)
>>> V = 10 * N.x
>>> A.set_ang_acc(N, V)
>>> A.ang_acc_in(N)
10*N.x
set ang vel(otherframe, value)
Dene the angular velocity vector in a ReferenceFrame.
Denes the angular velocity of this ReferenceFrame, in another. Angular velocity
can be dened with respect to multiple dierent ReferenceFrames. Care must be
taken to not create loops which are inconsistent.
Parameters otherframe : ReferenceFrame
A ReferenceFrame to dene the angular velocity in
value : Vector
The Vector representing angular velocity
Examples
>>>
>>>
>>>
>>>
1668
from sympy.physics.vector import ReferenceFrame, Vector
N = ReferenceFrame(N)
A = ReferenceFrame(A)
V = 10 * N.x
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> A.set_ang_vel(N, V)
>>> A.ang_vel_in(N)
10*N.x
variable map(otherframe)
Returns a dictionary which expresses the coordinate variables of this frame in terms
of the variables of otherframe.
If Vector.simp is True, returns a simplied version of the mapped values. Else, returns them without simplication.
Simplication of the expressions may take time.
Parameters otherframe : ReferenceFrame
The other frame to map the variables to
Examples
>>> from sympy.physics.vector import ReferenceFrame, dynamicsymbols
>>> A = ReferenceFrame(A)
>>> q = dynamicsymbols(q)
>>> B = A.orientnew(B, Axis, [q, A.z])
>>> A.variable_map(B)
{A_x: B_x*cos(q(t)) - B_y*sin(q(t)), A_y: B_x*sin(q(t)) + B_y*cos(q(t)), A_z: B_z}
x
The basis Vector for the ReferenceFrame, in the x direction.
y
The basis Vector for the ReferenceFrame, in the y direction.
z
The basis Vector for the ReferenceFrame, in the z direction.
Vector
class sympy.physics.vector.vector.Vector(inlist)
The class used to dene vectors.
It along with ReferenceFrame are the building blocks of describing a classical mechanics
system in PyDy and sympy.physics.vector.
Attributes
simp
Boolean
Let certain methods use trigsimp on their outputs
cross(other)
The cross product operator for two Vectors.
Returns a Vector, expressed in the same ReferenceFrames as self.
Parameters other : Vector
The Vector which we are crossing with
5.36. Physics Module
1669
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.physics.vector import ReferenceFrame, Vector
>>> from sympy import symbols
>>> q1 = symbols(q1)
>>> N = ReferenceFrame(N)
>>> N.x ^ N.y
N.z
>>> A = N.orientnew(A, Axis, [q1, N.x])
>>> A.x ^ N.y
N.z
>>> N.y ^ A.x
- sin(q1)*A.y - cos(q1)*A.z
diff(wrt, otherframe)
Takes the partial derivative, with respect to a value, in a frame.
Returns a Vector.
Parameters wrt : Symbol
What the partial derivative is taken with respect to.
otherframe : ReferenceFrame
The ReferenceFrame that the partial derivative is taken in.
Examples
>>> from sympy.physics.vector import ReferenceFrame, Vector, dynamicsymbols
>>> from sympy import Symbol
>>> Vector.simp = True
>>> t = Symbol(t)
>>> q1 = dynamicsymbols(q1)
>>> N = ReferenceFrame(N)
>>> A = N.orientnew(A, Axis, [q1, N.y])
>>> A.x.diff(t, N)
- q1*A.z
doit(**hints)
Calls .doit() on each term in the Vector
dot(other)
Dot product of two vectors.
Returns a scalar, the dot product of the two Vectors
Parameters other : Vector
The Vector which we are dotting with
Examples
>>>
>>>
>>>
>>>
>>>
1670
from sympy.physics.vector import ReferenceFrame, dot
from sympy import symbols
q1 = symbols(q1)
N = ReferenceFrame(N)
dot(N.x, N.x)
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
1
>>> dot(N.x, N.y)
0
>>> A = N.orientnew(A, Axis, [q1, N.x])
>>> dot(N.y, A.y)
cos(q1)
dt(otherframe)
Returns a Vector which is the time derivative of the self Vector, taken in frame otherframe.
Calls the global time derivative method
Parameters otherframe : ReferenceFrame
The frame to calculate the time derivative in
express(otherframe, variables=False)
Returns a Vector equivalent to this one, expressed in otherframe. Uses the global
express method.
Parameters otherframe : ReferenceFrame
The frame for this Vector to be described in
variables : boolean
If True, the coordinate symbols(if present) in this Vector are reexpressed in terms otherframe
Examples
>>> from sympy.physics.vector import ReferenceFrame, Vector, dynamicsymbols
>>> q1 = dynamicsymbols(q1)
>>> N = ReferenceFrame(N)
>>> A = N.orientnew(A, Axis, [q1, N.y])
>>> A.x.express(N)
cos(q1)*N.x - sin(q1)*N.z
magnitude()
Returns the magnitude (Euclidean norm) of self.
normalize()
Returns a Vector of magnitude 1, codirectional with self.
outer(other)
Outer product between two Vectors.
A rank increasing operation, which returns a Dyadic from two Vectors
Parameters other : Vector
The Vector to take the outer product with
Examples
>>> from sympy.physics.vector import ReferenceFrame, outer
>>> N = ReferenceFrame(N)
>>> outer(N.x, N.x)
(N.x|N.x)
5.36. Physics Module
1671
SymPy Documentation, Release 0.7.6
separate()
The constituents of this vector in dierent reference frames, as per its denition.
Returns a dict mapping each ReferenceFrame to the corresponding constituent Vector.
Examples
>>> from sympy.physics.vector import ReferenceFrame
>>> R1 = ReferenceFrame(R1)
>>> R2 = ReferenceFrame(R2)
>>> v = R1.x + R2.x
>>> v.separate() == {R1: R1.x, R2: R2.x}
True
simplify()
Returns a simplied Vector.
subs(*args, **kwargs)
Substituion on the Vector.
Examples
>>> from sympy.physics.vector import ReferenceFrame
>>> from sympy import Symbol
>>> N = ReferenceFrame(N)
>>> s = Symbol(s)
>>> a = N.x * s
>>> a.subs({s: 2})
2*N.x
to matrix(reference frame)
Returns the matrix form of the vector with respect to the given frame.
Parameters reference frame : ReferenceFrame
The reference frame that the rows of the matrix correspond to.
Returns matrix : ImmutableMatrix, shape(3,1)
The matrix that gives the 1D vector.
Examples
>>> from sympy import symbols
>>> from sympy.physics.vector import ReferenceFrame
>>> from sympy.physics.mechanics.functions import inertia
>>> a, b, c = symbols(a, b, c)
>>> N = ReferenceFrame(N)
>>> vector = a * N.x + b * N.y + c * N.z
>>> vector.to_matrix(N)
Matrix([
[a],
[b],
[c]])
>>> beta = symbols(beta)
1672
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> A = N.orientnew(A, Axis, (beta, N.x))
>>> vector.to_matrix(A)
Matrix([
[
a],
[ b*cos(beta) + c*sin(beta)],
[-b*sin(beta) + c*cos(beta)]])
Dyadic
class sympy.physics.vector.dyadic.Dyadic(inlist)
A Dyadic object.
See: http://en.wikipedia.org/wiki/Dyadic tensor Kane, T., Levinson, D. Dynamics Theory
and Applications. 1985 McGraw-Hill
A more powerful way to represent a rigid bodys inertia. While it is more complex, by
choosing Dyadic components to be in body xed basis vectors, the resulting matrix is
equivalent to the inertia tensor.
cross(other)
For a cross product in the form: Dyadic x Vector.
Parameters other : Vector
The Vector that we are crossing this Dyadic with
Examples
>>> from sympy.physics.vector import ReferenceFrame, outer, cross
>>> N = ReferenceFrame(N)
>>> d = outer(N.x, N.x)
>>> cross(d, N.y)
(N.x|N.z)
doit(**hints)
Calls .doit() on each term in the Dyadic
dot(other)
The inner product operator for a Dyadic and a Dyadic or Vector.
Parameters other : Dyadic or Vector
The other Dyadic or Vector to take the inner product with
Examples
>>> from sympy.physics.vector import ReferenceFrame, outer
>>> N = ReferenceFrame(N)
>>> D1 = outer(N.x, N.y)
>>> D2 = outer(N.y, N.y)
>>> D1.dot(D2)
(N.x|N.y)
>>> D1.dot(N.y)
N.x
dt(frame)
Take the time derivative of this Dyadic in a frame.
5.36. Physics Module
1673
SymPy Documentation, Release 0.7.6
This function calls the global time derivative method
Parameters frame : ReferenceFrame
The frame to take the time derivative in
Examples
>>> from sympy.physics.vector import ReferenceFrame, outer, dynamicsymbols
>>> N = ReferenceFrame(N)
>>> q = dynamicsymbols(q)
>>> B = N.orientnew(B, Axis, [q, N.z])
>>> d = outer(N.x, N.x)
>>> d.dt(B)
- q*(N.y|N.x) - q*(N.x|N.y)
express(frame1, frame2=None)
Expresses this Dyadic in alternate frame(s)
The rst frame is the list side expression, the second frame is the right side; if Dyadic
is in form A.x|B.y, you can express it in two dierent frames. If no second frame is
given, the Dyadic is expressed in only one frame.
Calls the global express function
Parameters frame1 : ReferenceFrame
The frame to express the left side of the Dyadic in
frame2 : ReferenceFrame
If provided, the frame to express the right side of the Dyadic in
Examples
>>> from sympy.physics.vector import ReferenceFrame, outer, dynamicsymbols
>>> N = ReferenceFrame(N)
>>> q = dynamicsymbols(q)
>>> B = N.orientnew(B, Axis, [q, N.z])
>>> d = outer(N.x, N.x)
>>> d.express(B, N)
cos(q)*(B.x|N.x) - sin(q)*(B.y|N.x)
simplify()
Returns a simplied Dyadic.
subs(*args, **kwargs)
Substituion on the Dyadic.
Examples
>>>
>>>
>>>
>>>
>>>
1674
from sympy.physics.vector import ReferenceFrame
from sympy import Symbol
N = ReferenceFrame(N)
s = Symbol(s)
a = s * (N.x|N.x)
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> a.subs({s: 2})
2*(N.x|N.x)
to matrix(reference frame, second reference frame=None)
Returns the matrix form of the dyadic with respect to one or two reference frames.
Parameters reference frame : ReferenceFrame
The reference frame that the rows and columns of the matrix correspond to. If a second reference frame is provided, this only corresponds to the rows of the matrix.
second reference frame : ReferenceFrame, optional, default=None
The reference frame that the columns of the matrix correspond to.
Returns matrix : ImmutableMatrix, shape(3,3)
The matrix that gives the 2D tensor form.
Examples
>>> from sympy import symbols
>>> from sympy.physics.vector import ReferenceFrame, Vector
>>> Vector.simp = True
>>> from sympy.physics.mechanics import inertia
>>> Ixx, Iyy, Izz, Ixy, Iyz, Ixz = symbols(Ixx, Iyy, Izz, Ixy, Iyz, Ixz)
>>> N = ReferenceFrame(N)
>>> inertia_dyadic = inertia(N, Ixx, Iyy, Izz, Ixy, Iyz, Ixz)
>>> inertia_dyadic.to_matrix(N)
Matrix([
[Ixx, Ixy, Ixz],
[Ixy, Iyy, Iyz],
[Ixz, Iyz, Izz]])
>>> beta = symbols(beta)
>>> A = N.orientnew(A, Axis, (beta, N.x))
>>> inertia_dyadic.to_matrix(A)
Matrix([
[
Ixx,
Ixy*cos(beta) + I
[ Ixy*cos(beta) + Ixz*sin(beta), Iyy*cos(2*beta)/2 + Iyy/2 + Iyz*sin(2*beta) - Izz*cos(2*bet
[-Ixy*sin(beta) + Ixz*cos(beta),
-Iyy*sin(2*beta)/2 + Iyz*cos(2*beta) + Izz*s
Kinematics (Docstrings)
Point
class sympy.physics.vector.point.Point(name)
This object represents a point in a dynamic system.
It stores the: position, velocity, and acceleration of a point. The position is a vector
dened as the vector distance from a parent point to this point.
a1pt theory(otherpoint, outframe, interframe)
Sets the acceleration of this point with the 1-point theory.
The 1-point theory for point acceleration looks like this:
N aP = B aP + N aO + N alphaB x rOP + N omegaB x (N omegaB x
rOP) + 2 N omegaB x B vP
5.36. Physics Module
1675
SymPy Documentation, Release 0.7.6
where O is a point xed in B, P is a point moving in B, and B is rotating in frame N.
Parameters otherpoint : Point
The rst point of the 1-point theory (O)
outframe : ReferenceFrame
The frame we want this points acceleration dened in (N)
xedframe : ReferenceFrame
The intermediate frame in this calculation (B)
Examples
>>> from sympy.physics.vector import Point, ReferenceFrame
>>> from sympy.physics.vector import Vector, dynamicsymbols
>>> q = dynamicsymbols(q)
>>> q2 = dynamicsymbols(q2)
>>> qd = dynamicsymbols(q, 1)
>>> q2d = dynamicsymbols(q2, 1)
>>> N = ReferenceFrame(N)
>>> B = ReferenceFrame(B)
>>> B.set_ang_vel(N, 5 * B.y)
>>> O = Point(O)
>>> P = O.locatenew(P, q * B.x)
>>> P.set_vel(B, qd * B.x + q2d * B.y)
>>> O.set_vel(N, 0)
>>> P.a1pt_theory(O, N, B)
(-25*q + q)*B.x + q2*B.y - 10*q*B.z
a2pt theory(otherpoint, outframe, xedframe)
Sets the acceleration of this point with the 2-point theory.
The 2-point theory for point acceleration looks like this:
N aP = N aO + N alphaB x rOP + N omegaB x (N omegaB x rOP)
where O and P are both points xed in frame B, which is rotating in frame N.
Parameters otherpoint : Point
The rst point of the 2-point theory (O)
outframe : ReferenceFrame
The frame we want this points acceleration dened in (N)
xedframe : ReferenceFrame
The frame in which both points are xed (B)
Examples
>>>
>>>
>>>
>>>
>>>
>>>
>>>
1676
from sympy.physics.vector import Point, ReferenceFrame, dynamicsymbols
q = dynamicsymbols(q)
qd = dynamicsymbols(q, 1)
N = ReferenceFrame(N)
B = N.orientnew(B, Axis, [q, N.z])
O = Point(O)
P = O.locatenew(P, 10 * B.x)
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> O.set_vel(N, 5 * N.x)
>>> P.a2pt_theory(O, N, B)
- 10*q**2*B.x + 10*q*B.y
acc(frame)
The acceleration Vector of this Point in a ReferenceFrame.
Parameters frame : ReferenceFrame
The frame in which the returned acceleration vector will be dened
in
Examples
>>> from sympy.physics.vector import Point, ReferenceFrame
>>> N = ReferenceFrame(N)
>>> p1 = Point(p1)
>>> p1.set_acc(N, 10 * N.x)
>>> p1.acc(N)
10*N.x
locatenew(name, value)
Creates a new point with a position dened from this point.
Parameters name : str
The name for the new point
value : Vector
The position of the new point relative to this point
Examples
>>>
>>>
>>>
>>>
from sympy.physics.vector import ReferenceFrame, Point
N = ReferenceFrame(N)
P1 = Point(P1)
P2 = P1.locatenew(P2, 10 * N.x)
pos from(otherpoint)
Returns a Vector distance between this Point and the other Point.
Parameters otherpoint : Point
The otherpoint we are locating this one relative to
Examples
>>> from sympy.physics.vector import Point, ReferenceFrame
>>> N = ReferenceFrame(N)
>>> p1 = Point(p1)
>>> p2 = Point(p2)
>>> p1.set_pos(p2, 10 * N.x)
>>> p1.pos_from(p2)
10*N.x
5.36. Physics Module
1677
SymPy Documentation, Release 0.7.6
set acc(frame, value)
Used to set the acceleration of this Point in a ReferenceFrame.
Parameters value : Vector
The vector value of this points acceleration in the frame
frame : ReferenceFrame
The frame in which this points acceleration is dened
Examples
>>> from sympy.physics.vector import Point, ReferenceFrame
>>> N = ReferenceFrame(N)
>>> p1 = Point(p1)
>>> p1.set_acc(N, 10 * N.x)
>>> p1.acc(N)
10*N.x
set pos(otherpoint, value)
Used to set the position of this point w.r.t. another point.
Parameters value : Vector
The vector which denes the location of this point
point : Point
The other point which this points location is dened relative to
Examples
>>> from sympy.physics.vector import Point, ReferenceFrame
>>> N = ReferenceFrame(N)
>>> p1 = Point(p1)
>>> p2 = Point(p2)
>>> p1.set_pos(p2, 10 * N.x)
>>> p1.pos_from(p2)
10*N.x
set vel(frame, value)
Sets the velocity Vector of this Point in a ReferenceFrame.
Parameters value : Vector
The vector value of this points velocity in the frame
frame : ReferenceFrame
The frame in which this points velocity is dened
Examples
>>>
>>>
>>>
>>>
1678
from sympy.physics.vector import Point, ReferenceFrame
N = ReferenceFrame(N)
p1 = Point(p1)
p1.set_vel(N, 10 * N.x)
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> p1.vel(N)
10*N.x
v1pt theory(otherpoint, outframe, interframe)
Sets the velocity of this point with the 1-point theory.
The 1-point theory for point velocity looks like this:
N vP = B vP + N vO + N omegaB x rOP
where O is a point xed in B, P is a point moving in B, and B is rotating in frame N.
Parameters otherpoint : Point
The rst point of the 2-point theory (O)
outframe : ReferenceFrame
The frame we want this points velocity dened in (N)
interframe : ReferenceFrame
The intermediate frame in this calculation (B)
Examples
>>> from sympy.physics.vector import Point, ReferenceFrame
>>> from sympy.physics.vector import Vector, dynamicsymbols
>>> q = dynamicsymbols(q)
>>> q2 = dynamicsymbols(q2)
>>> qd = dynamicsymbols(q, 1)
>>> q2d = dynamicsymbols(q2, 1)
>>> N = ReferenceFrame(N)
>>> B = ReferenceFrame(B)
>>> B.set_ang_vel(N, 5 * B.y)
>>> O = Point(O)
>>> P = O.locatenew(P, q * B.x)
>>> P.set_vel(B, qd * B.x + q2d * B.y)
>>> O.set_vel(N, 0)
>>> P.v1pt_theory(O, N, B)
q*B.x + q2*B.y - 5*q*B.z
v2pt theory(otherpoint, outframe, xedframe)
Sets the velocity of this point with the 2-point theory.
The 2-point theory for point velocity looks like this:
N vP = N vO + N omegaB x rOP
where O and P are both points xed in frame B, which is rotating in frame N.
Parameters otherpoint : Point
The rst point of the 2-point theory (O)
outframe : ReferenceFrame
The frame we want this points velocity dened in (N)
xedframe : ReferenceFrame
The frame in which both points are xed (B)
5.36. Physics Module
1679
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.physics.vector import Point, ReferenceFrame, dynamicsymbols
>>> q = dynamicsymbols(q)
>>> qd = dynamicsymbols(q, 1)
>>> N = ReferenceFrame(N)
>>> B = N.orientnew(B, Axis, [q, N.z])
>>> O = Point(O)
>>> P = O.locatenew(P, 10 * B.x)
>>> O.set_vel(N, 5 * N.x)
>>> P.v2pt_theory(O, N, B)
5*N.x + 10*q*B.y
vel(frame)
The velocity Vector of this Point in the ReferenceFrame.
Parameters frame : ReferenceFrame
The frame in which the returned velocity vector will be dened in
Examples
>>> from sympy.physics.vector import Point, ReferenceFrame
>>> N = ReferenceFrame(N)
>>> p1 = Point(p1)
>>> p1.set_vel(N, 10 * N.x)
>>> p1.vel(N)
10*N.x
kinematic equations
sympy.physics.vector.functions.kinematic equations(speeds, coords,
rot order=)
Gives equations relating the qdots to us for a rotation type.
rot type,
Supply rotation type and order as in orient. Speeds are assumed to be body-xed; if we
are dening the orientation of B in A using by rot type, the angular velocity of B in A is
assumed to be in the form: speed[0]*B.x + speed[1]*B.y + speed[2]*B.z
Parameters speeds : list of length 3
The body xed angular velocity measure numbers.
coords : list of length 3 or 4
The coordinates used to dene the orientation of the two frames.
rot type : str
The type of rotation used to create the equations. Body, Space, or
Quaternion only
rot order : str
If applicable, the order of a series of rotations.
1680
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.physics.vector import dynamicsymbols
>>> from sympy.physics.vector import kinematic_equations, vprint
>>> u1, u2, u3 = dynamicsymbols(u1 u2 u3)
>>> q1, q2, q3 = dynamicsymbols(q1 q2 q3)
>>> vprint(kinematic_equations([u1,u2,u3], [q1,q2,q3], body, 313),
...
order=None)
[-(u1*sin(q3) + u2*cos(q3))/sin(q2) + q1, -u1*cos(q3) + u2*sin(q3) + q2, (u1*sin(q3) + u2*cos(
sympy.physics.vector.functions.partial velocity(vel list, u list, frame)
Returns a list of partial velocities.
For a list of velocity or angular velocity vectors the partial derivatives with respect to
the supplied generalized speeds are computed, in the specied ReferenceFrame.
The output is a list of lists. The outer list has a number of elements equal to the number
of supplied velocity vectors. The inner lists are, for each velocity vector, the partial
derivatives of that velocity vector with respect to the generalized speeds supplied.
Parameters vel list : list
List of velocities of Points and angular velocities of ReferenceFrames
u list : list
List of independent generalized speeds.
frame : ReferenceFrame
The ReferenceFrame the partial derivatives are going to be taken in.
Examples
>>> from sympy.physics.vector import Point, ReferenceFrame
>>> from sympy.physics.vector import dynamicsymbols
>>> from sympy.physics.vector import partial_velocity
>>> u = dynamicsymbols(u)
>>> N = ReferenceFrame(N)
>>> P = Point(P)
>>> P.set_vel(N, u * N.x)
>>> vel_list = [P.vel(N)]
>>> u_list = [u]
>>> partial_velocity(vel_list, u_list, N)
[[N.x]]
sympy.physics.vector.functions.get motion params(frame, **kwargs)
Returns the three motion parameters - (acceleration, velocity, and position) as vectorial
functions of time in the given frame.
If a higher order dierential function is provided, the lower order functions are used
as boundary conditions. For example, given the acceleration, the velocity and position
parameters are taken as boundary conditions.
The values of time at which the boundary conditions are specied are taken from
timevalue1(for position boundary condition) and timevalue2(for velocity boundary condition).
If any of the boundary conditions are not provided, they are taken to be zero by default
(zero vectors, in case of vectorial inputs). If the boundary conditions are also functions
5.36. Physics Module
1681
SymPy Documentation, Release 0.7.6
of time, they are converted to constants by substituting the time values in the dynamicsymbols. t time Symbol.
This function can also be used for calculating rotational motion parameters. Have a look
at the Parameters and Examples for more clarity.
Parameters frame : ReferenceFrame
The frame to express the motion parameters in
acceleration : Vector
Acceleration of the object/frame as a function of time
velocity : Vector
Velocity as function of time or as boundary condition of velocity at
time = timevalue1
position : Vector
Velocity as function of time or as boundary condition of velocity at
time = timevalue1
timevalue1 : sympyable
Value of time for position boundary condition
timevalue2 : sympyable
Value of time for velocity boundary condition
Examples
>>> from sympy.physics.vector import ReferenceFrame, get_motion_params, dynamicsymbols
>>> from sympy import symbols
>>> R = ReferenceFrame(R)
>>> v1, v2, v3 = dynamicsymbols(v1 v2 v3)
>>> v = v1*R.x + v2*R.y + v3*R.z
>>> get_motion_params(R, position = v)
(v1*R.x + v2*R.y + v3*R.z, v1*R.x + v2*R.y + v3*R.z, v1*R.x + v2*R.y + v3*R.z)
>>> a, b, c = symbols(a b c)
>>> v = a*R.x + b*R.y + c*R.z
>>> get_motion_params(R, velocity = v)
(0, a*R.x + b*R.y + c*R.z, a*t*R.x + b*t*R.y + c*t*R.z)
>>> parameters = get_motion_params(R, acceleration = v)
>>> parameters[1]
a*t*R.x + b*t*R.y + c*t*R.z
>>> parameters[2]
a*t**2/2*R.x + b*t**2/2*R.y + c*t**2/2*R.z
Printing (Docstrings)
init printing
1682
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
sympy.physics.vector.printing.init printing(pretty print=True,
order=None,
use unicode=None,
use latex=None, wrap line=None,
num columns=None,
no global=False,
ip=None,
euler=False,
forecolor=Black,
backcolor=Transparent,
fontsize=10pt,
latex mode=equation*,
print builtin=True,
str printer=None,
pretty printer=None,
latex printer=None)
Initializes pretty-printer depending on the environment.
Parameters pretty print: boolean :
If True, use pretty print to stringify or the provided pretty printer; if
False, use sstrrepr to stringify or the provided string printer.
order: string or None :
There are a few dierent settings for this parameter: lex (default),
which is lexographic order; grlex, which is graded lexographic order;
grevlex, which is reversed graded lexographic order; old, which is
used for compatibility reasons and for long expressions; None, which
sets it to lex.
use unicode: boolean or None :
If True, use unicode characters; if False, do not use unicode characters.
use latex: string, boolean, or None :
If True, use default latex rendering in GUI interfaces (png and mathjax); if False, do not use latex rendering; if png, enable latex rendering with an external latex compiler, falling back to matplotlib if
external compilation fails; if matplotlib, enable latex rendering with
matplotlib; if mathjax, enable latex text generation, for example
MathJax rendering in IPython notebook or text rendering in LaTeX
documents
wrap line: boolean :
If True, lines will wrap at the end; if False, they will not wrap but
continue as one line. This is only relevant if prettyp rint is True.
num columns: int or None :
If int, number of columns before wrapping is set to num columns; if
None, number of columns before wrapping is set to terminal width.
This is only relevant if prettyp rint is True.
no global: boolean :
If True, the settings become system wide; if False, use just for this
console/session.
ip: An interactive console :
This can either be an instance of IPython, or a class that derives from
code.InteractiveConsole.
5.36. Physics Module
1683
SymPy Documentation, Release 0.7.6
euler: boolean, optional, default=False :
Loads the euler package in the LaTeX preamble for handwritten style
fonts (http://www.ctan.org/pkg/euler).
forecolor: string, optional, default=Black :
DVI setting for foreground color.
backcolor: string, optional, default=Transparent :
DVI setting for background color.
fontsize: string, optional, default=10pt :
A font size to pass to the LaTeX documentclass function in the preamble.
latex mode: string, optional, default=equation* :
The mode used in the LaTeX printer.
{inline|plain|equation|equation*}.
Can
be
one
of:
print builtin: boolean, optional, default=True :
If true then oats and integers will be printed. If false the printer will
only print SymPy types.
str printer: function, optional, default=None :
A custom string printer
sympy.printing.sstrrepr().
function.
This
should
mimic
pretty printer: function, optional, default=None :
A custom pretty printer. This should mimic sympy.printing.pretty().
latex printer: function, optional, default=None :
A custom LaTeX printer. This should mimic sympy.printing.latex()
This should mimic sympy.printing.latex().
Examples
>>> from sympy.interactive import init_printing
>>> from sympy import Symbol, sqrt
>>> from sympy.abc import x, y
>>> sqrt(5)
sqrt(5)
>>> init_printing(pretty_print=True)
>>> sqrt(5)
___
\/ 5
>>> theta = Symbol(theta)
>>> init_printing(use_unicode=True)
>>> theta
\u03b8
>>> init_printing(use_unicode=False)
>>> theta
theta
>>> init_printing(order=lex)
>>> str(y + x + y**2 + x**2)
x**2 + x + y**2 + y
>>> init_printing(order=grlex)
1684
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> str(y + x + y**2 + x**2)
x**2 + x + y**2 + y
>>> init_printing(order=grevlex)
>>> str(y * x**2 + x * y**2)
x**2*y + x*y**2
>>> init_printing(order=old)
>>> str(x**2 + y**2 + x + y)
x**2 + x + y**2 + y
>>> init_printing(num_columns=10)
>>> x**2 + x + y**2 + y
x + y +
x**2 + y**2
vprint
sympy.physics.vector.printing.vprint(expr, **settings)
Function for printing of expressions generated in the sympy.physics vector package.
Extends SymPys StrPrinter, takes the same setting accepted by SymPys sstr(), and is
equivalent to print(sstr(f oo)).
Parameters expr : valid SymPy object
SymPy expression to print.
settings : args
Same as the settings accepted by SymPys sstr().
Examples
>>> from sympy.physics.vector import vprint, dynamicsymbols
>>> u1 = dynamicsymbols(u1)
>>> print(u1)
u1(t)
>>> vprint(u1)
u1
vpprint
sympy.physics.vector.printing.vpprint(expr, **settings)
Function for pretty printing of expressions generated in the sympy.physics vector package.
Mainly used for expressions not inside a vector; the output of running scripts and generating equations of motion. Takes the same options as SymPys pretty print(); see that
function for more information.
Parameters expr : valid SymPy object
SymPy expression to pretty print
settings : args
Same as those accepted by SymPys pretty print.
5.36. Physics Module
1685
SymPy Documentation, Release 0.7.6
vlatex
sympy.physics.vector.printing.vlatex(expr, **settings)
Function for printing latex representation of sympy.physics.vector objects.
For latex representation of Vectors, Dyadics, and dynamicsymbols. Takes the same options as SymPys latex(); see that function for more information;
Parameters expr : valid SymPy object
SymPy expression to represent in LaTeX form
settings : args
Same as latex()
Examples
>>> from sympy.physics.vector import vlatex, ReferenceFrame, dynamicsymbols
>>> N = ReferenceFrame(N)
>>> q1, q2 = dynamicsymbols(q1 q2)
>>> q1d, q2d = dynamicsymbols(q1 q2, 1)
>>> q1dd, q2dd = dynamicsymbols(q1 q2, 2)
>>> vlatex(N.x + N.y)
\\mathbf{\\hat{n}_x} + \\mathbf{\\hat{n}_y}
>>> vlatex(q1 + q2)
q_{1} + q_{2}
>>> vlatex(q1d)
\\dot{q}_{1}
>>> vlatex(q1 * q2d)
q_{1} \\dot{q}_{2}
>>> vlatex(q1dd * q1 / q1d)
\\frac{q_{1} \\ddot{q}_{1}}{\\dot{q}_{1}}
Essential Functions (Docstrings)
dynamicsymbols
sympy.physics.vector.dynamicsymbols(names, level=0)
Uses symbols and Function for functions of time.
Creates a SymPy UndenedFunction, which is then initialized as a function of a variable,
the default being Symbol(t).
Parameters names : str
Names of the dynamic symbols you want to create; works the same
way as inputs to symbols
level : int
Level of dierentiation of the returned function; d/dt once of t, twice
of t, etc.
Examples
1686
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> from sympy.physics.vector import dynamicsymbols
>>> from sympy import diff, Symbol
>>> q1 = dynamicsymbols(q1)
>>> q1
q1(t)
>>> diff(q1, Symbol(t))
Derivative(q1(t), t)
dot
sympy.physics.vector.functions.dot(vec1, vec2)
Dot product convenience wrapper for Vector.dot(): Dot product of two vectors.
Returns a scalar, the dot product of the two Vectors
Parameters other : Vector
The Vector which we are dotting with
Examples
>>> from sympy.physics.vector import ReferenceFrame, dot
>>> from sympy import symbols
>>> q1 = symbols(q1)
>>> N = ReferenceFrame(N)
>>> dot(N.x, N.x)
1
>>> dot(N.x, N.y)
0
>>> A = N.orientnew(A, Axis, [q1, N.x])
>>> dot(N.y, A.y)
cos(q1)
cross
sympy.physics.vector.functions.cross(vec1, vec2)
Cross product convenience wrapper for Vector.cross(): The cross product operator for
two Vectors.
Returns a Vector, expressed in the same ReferenceFrames as self.
Parameters other : Vector
The Vector which we are crossing with
Examples
>>>
>>>
>>>
>>>
>>>
N.z
>>>
>>>
N.z
from sympy.physics.vector import ReferenceFrame, Vector
from sympy import symbols
q1 = symbols(q1)
N = ReferenceFrame(N)
N.x ^ N.y
A = N.orientnew(A, Axis, [q1, N.x])
A.x ^ N.y
5.36. Physics Module
1687
SymPy Documentation, Release 0.7.6
>>> N.y ^ A.x
- sin(q1)*A.y - cos(q1)*A.z
outer
sympy.physics.vector.functions.outer(vec1, vec2)
Outer product convenience wrapper for Vector.outer(): Outer product between two Vectors.
A rank increasing operation, which returns a Dyadic from two Vectors
Parameters other : Vector
The Vector to take the outer product with
Examples
>>> from sympy.physics.vector import ReferenceFrame, outer
>>> N = ReferenceFrame(N)
>>> outer(N.x, N.x)
(N.x|N.x)
express
sympy.physics.vector.functions.express(expr,
frame,
ables=False)
Global function for express functionality.
frame2=None,
vari-
Re-expresses a Vector, scalar(sympyable) or Dyadic in given frame.
Refer to the local methods of Vector and Dyadic for details. If variables is True, then
the coordinate variables (CoordinateSym instances) of other frames present in the vector/scalar eld or dyadic expression are also substituted in terms of the base scalars of
this frame.
Parameters expr : Vector/Dyadic/scalar(sympyable)
The expression to re-express in ReferenceFrame frame
frame: ReferenceFrame :
The reference frame to express expr in
frame2 : ReferenceFrame
The other frame required for re-expression(only for Dyadic expr)
variables : boolean
Species whether to substitute the coordinate variables present in
expr, in terms of those of frame
Examples
>>>
>>>
>>>
>>>
>>>
1688
from sympy.physics.vector import ReferenceFrame, outer, dynamicsymbols
N = ReferenceFrame(N)
q = dynamicsymbols(q)
B = N.orientnew(B, Axis, [q, N.z])
d = outer(N.x, N.x)
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> from sympy.physics.vector import express
>>> express(d, B, N)
cos(q)*(B.x|N.x) - sin(q)*(B.y|N.x)
>>> express(B.x, N)
cos(q)*N.x + sin(q)*N.y
>>> express(N[0], B, variables=True)
B_x*cos(q(t)) - B_y*sin(q(t))
time derivative
sympy.physics.vector.functions.time derivative(expr, frame, order=1)
Calculate the time derivative of a vector/scalar eld function or dyadic expression in
given frame.
Parameters expr : Vector/Dyadic/sympifyable
The expression whose time derivative is to be calculated
frame : ReferenceFrame
The reference frame to calculate the time derivative in
order : integer
The order of the derivative to be calculated
References
http://en.wikipedia.org/wiki/Rotating reference frame#Time derivatives in the two frames
Examples
>>> from sympy.physics.vector import ReferenceFrame, dynamicsymbols
>>> from sympy import Symbol
>>> q1 = Symbol(q1)
>>> u1 = dynamicsymbols(u1)
>>> N = ReferenceFrame(N)
>>> A = N.orientnew(A, Axis, [q1, N.x])
>>> v = u1 * N.x
>>> A.set_ang_vel(N, 10*A.x)
>>> from sympy.physics.vector import time_derivative
>>> time_derivative(v, N)
u1*N.x
>>> time_derivative(u1*A[0], N)
N_x*Derivative(u1(t), t)
>>> B = N.orientnew(B, Axis, [u1, N.z])
>>> from sympy.physics.vector import outer
>>> d = outer(N.x, N.x)
>>> time_derivative(d, B)
- u1*(N.y|N.x) - u1*(N.x|N.y)
Docstrings for basic eld functions
Field operation functions
elds in general.
5.36. Physics Module
These functions implement some basic operations pertaining to
1689
SymPy Documentation, Release 0.7.6
curl
sympy.physics.vector.fieldfunctions.curl(vect, frame)
Returns the curl of a vector eld computed wrt the coordinate symbols of the given
frame.
Parameters vect : Vector
The vector operand
frame : ReferenceFrame
The reference frame to calculate the curl in
Examples
>>> from sympy.physics.vector import ReferenceFrame
>>> from sympy.physics.vector import curl
>>> R = ReferenceFrame(R)
>>> v1 = R[1]*R[2]*R.x + R[0]*R[2]*R.y + R[0]*R[1]*R.z
>>> curl(v1, R)
0
>>> v2 = R[0]*R[1]*R[2]*R.x
>>> curl(v2, R)
R_x*R_y*R.y - R_x*R_z*R.z
divergence
sympy.physics.vector.fieldfunctions.divergence(vect, frame)
Returns the divergence of a vector eld computed wrt the coordinate symbols of the
given frame.
Parameters vect : Vector
The vector operand
frame : ReferenceFrame
The reference frame to calculate the divergence in
Examples
>>> from sympy.physics.vector import ReferenceFrame
>>> from sympy.physics.vector import divergence
>>> R = ReferenceFrame(R)
>>> v1 = R[0]*R[1]*R[2] * (R.x+R.y+R.z)
>>> divergence(v1, R)
R_x*R_y + R_x*R_z + R_y*R_z
>>> v2 = 2*R[1]*R[2]*R.y
>>> divergence(v2, R)
2*R_z
gradient
sympy.physics.vector.fieldfunctions.gradient(scalar, frame)
Returns the vector gradient of a scalar eld computed wrt the coordinate symbols of the
given frame.
Parameters scalar : sympiable
1690
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
The scalar eld to take the gradient of
frame : ReferenceFrame
The frame to calculate the gradient in
Examples
>>> from sympy.physics.vector import ReferenceFrame
>>> from sympy.physics.vector import gradient
>>> R = ReferenceFrame(R)
>>> s1 = R[0]*R[1]*R[2]
>>> gradient(s1, R)
R_y*R_z*R.x + R_x*R_z*R.y + R_x*R_y*R.z
>>> s2 = 5*R[0]**2*R[2]
>>> gradient(s2, R)
10*R_x*R_z*R.x + 5*R_x**2*R.z
scalar potential
sympy.physics.vector.fieldfunctions.scalar potential(eld, frame)
Returns the scalar potential function of a eld in a given frame (without the added integration constant).
Parameters eld : Vector
The vector eld whose scalar potential function is to be calculated
frame : ReferenceFrame
The frame to do the calculation in
Examples
>>> from sympy.physics.vector import ReferenceFrame
>>> from sympy.physics.vector import scalar_potential, gradient
>>> R = ReferenceFrame(R)
>>> scalar_potential(R.z, R) == R[2]
True
>>> scalar_field = 2*R[0]**2*R[1]*R[2]
>>> grad_field = gradient(scalar_field, R)
>>> scalar_potential(grad_field, R)
2*R_x**2*R_y*R_z
scalar potential dierence
sympy.physics.vector.fieldfunctions.scalar potential difference(eld, frame,
point1,
point2, origin)
Returns the scalar potential dierence between two points in a certain frame, wrt a given
eld.
If a scalar eld is provided, its values at the two points are considered. If a conservative
vector eld is provided, the values of its scalar potential function at the two points are
used.
5.36. Physics Module
1691
SymPy Documentation, Release 0.7.6
Returns (potential at position 2) - (potential at position 1)
Parameters eld : Vector/sympyable
The eld to calculate wrt
frame : ReferenceFrame
The frame to do the calculations in
point1 : Point
The initial Point in given frame
position2 : Point
The second Point in the given frame
origin : Point
The Point to use as reference point for position vector calculation
Examples
>>> from sympy.physics.vector import ReferenceFrame, Point
>>> from sympy.physics.vector import scalar_potential_difference
>>> R = ReferenceFrame(R)
>>> O = Point(O)
>>> P = O.locatenew(P, R[0]*R.x + R[1]*R.y + R[2]*R.z)
>>> vectfield = 4*R[0]*R[1]*R.x + 2*R[0]**2*R.y
>>> scalar_potential_difference(vectfield, R, O, P, O)
2*R_x**2*R_y
>>> Q = O.locatenew(O, 3*R.x + R.y + 2*R.z)
>>> scalar_potential_difference(vectfield, R, P, Q, O)
-2*R_x**2*R_y + 18
Checking the type of vector eld
is conservative
sympy.physics.vector.fieldfunctions.is conservative(eld)
Checks if a eld is conservative.
Examples
>>> from sympy.physics.vector import ReferenceFrame
>>> from sympy.physics.vector import is_conservative
>>> R = ReferenceFrame(R)
>>> is_conservative(R[1]*R[2]*R.x + R[0]*R[2]*R.y + R[0]*R[1]*R.z)
True
>>> is_conservative(R[2] * R.y)
False
Paramaters
eld [Vector] The eld to check for conservative property
1692
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
is solenoidal
sympy.physics.vector.fieldfunctions.is solenoidal(eld)
Checks if a eld is solenoidal.
Examples
>>> from sympy.physics.vector import ReferenceFrame
>>> from sympy.physics.vector import is_solenoidal
>>> R = ReferenceFrame(R)
>>> is_solenoidal(R[1]*R[2]*R.x + R[0]*R[2]*R.y + R[0]*R[1]*R.z)
True
>>> is_solenoidal(R[1] * R.y)
False
Paramaters
eld [Vector] The eld to check for solenoidal property
Classical Mechanics
Abstract
In this documentation many components of the physics/mechanics module will be discussed. mechanics has been written to allow for creation of symbolic equations of motion
for complicated multibody systems.
Vector
This module derives the vector-related abilities and related functionalities from
physics.vector. Please have a look at the documentation of physics.vector and its
necessary API to understand the vector capabilities of mechanics.
Mechanics
In physics, mechanics describes conditions of rest (statics) or motion (dynamics). There are a
few common steps to all mechanics problems. First, an idealized representation of a system is
described. Next, we use physical laws to generate equations that dene the systems behavior.
Then, we solve these equations, sometimes analytically but usually numerically. Finally, we
extract information from these equations and solutions. The current scope of the module is
multi-body dynamics: the motion of systems of multiple particles and/or rigid bodies. For
example, this module could be used to understand the motion of a double pendulum, planets,
robotic manipulators, bicycles, and any other system of rigid bodies that may fascinate us.
Often, the objective in multi-body dynamics is to obtain the trajectory of a system of rigid
bodies through time. The challenge for this task is to rst formulate the equations of motion
of the system. Once they are formulated, they must be solved, that is, integrated forward in
time. When digital computers came around, solving became the easy part of the problem.
5.36. Physics Module
1693
SymPy Documentation, Release 0.7.6
Now, we can tackle more complicated problems, which leaves the challenge of formulating
the equations.
The term equations of motion is used to describe the application of Newtons second law to
multi-body systems. The form of the equations of motion depends on the method used to generate them. This package implements two of these methods: Kanes method and Lagranges
method. This module facilitates the formulation of equations of motion, which can then be
solved (integrated) using generic ordinary dierential equation (ODE) solvers.
The approach to a particular class of dynamics problems, that of forward dynamics, has the
following steps:
1. describing the systems geometry and conguration,
2. specifying the way the system can move, including constraints on its motion
3. describing the external forces and moments on the system,
4. combining the above information according to Newtons second law (F = ma), and
5. organizing the resulting equations so that they can be integrated to obtain the systems
trajectory through time.
Together with the rest of SymPy, this module performs steps 4 and 5, provided that the user
can perform 1 through 3 for the module. That is to say, the user must provide a complete
representation of the free body diagrams that themselves represent the system, with which
this code can provide equations of motion in a form amenable to numerical integration. Step
5 above amounts to arduous algebra for even fairly simple multi-body systems. Thus, it is
desirable to use a symbolic math package, such as Sympy, to perform this step. It is for
this reason that this module is a part of Sympy. Step 4 amounts to this specic module,
sympy.physics.mechanics.
Guide to Mechanics
Masses, Inertias, Particles and Rigid Bodies in Physics/Mechanics This document will
describe how to represent masses and inertias in mechanics and use of the RigidBody and
Particle classes.
It is assumed that the reader is familiar with the basics of these topics, such as nding the
center of mass for a system of particles, how to manipulate an inertia tensor, and the denition
of a particle and rigid body. Any advanced dynamics text can provide a reference for these
details.
Mass The only requirement for a mass is that it needs to be a sympify-able expression. Keep
in mind that masses can be time varying.
Particle Particles are created with the class Particle in mechanics. A Particle object has
an associated point and an associated mass which are the only two attributes of the object.:
>>>
>>>
>>>
>>>
>>>
>>>
from sympy.physics.mechanics import Particle, Point
from sympy import Symbol
m = Symbol(m)
po = Point(po)
# create a particle container
pa = Particle(pa, po, m)
1694
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
The associated point contains the position, velocity and acceleration of the particle. mechanics allows one to perform kinematic analysis of points separate from their association with
masses.
Inertia See the Inertia (Dyadics) section in Advanced Topics part of physics/vector docs.
Rigid Body Rigid bodies are created in a similar fashion as particles. The RigidBody class
generates objects with four attributes: mass, center of mass, a reference frame, and an inertia
tuple:
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
from sympy import Symbol
from sympy.physics.mechanics import ReferenceFrame, Point, RigidBody
from sympy.physics.mechanics import outer
m = Symbol(m)
A = ReferenceFrame(A)
P = Point(P)
I = outer(A.x, A.x)
# create a rigid body
B = RigidBody(B, P, A, m, (I, P))
The mass is specied exactly as is in a particle. Similar to the Particles .point, the RigidBodys center of mass, .masscenter must be specied. The reference frame is stored in an
analogous fashion and holds information about the bodys orientation and angular velocity.
Finally, the inertia for a rigid body needs to be specied about a point. In mechanics, you are
allowed to specify any point for this. The most common is the center of mass, as shown in the
above code. If a point is selected which is not the center of mass, ensure that the position
between the point and the center of mass has been dened. The inertia is specied as a tuple
of length two with the rst entry being a Dyadic and the second entry being a Point of which
the inertia dyadic is dened about.
Dyadic In mechanics, dyadics are used to represent inertia ([Kane1985] (page 1916),
[WikiDyadics] (page 1916), [WikiDyadicProducts] (page 1916)). A dyadic is a linear polynomial of component unit dyadics, similar to a vector being a linear polynomial of component
unit vectors. A dyadic is the outer product between two vectors which returns a new quantity
representing the juxtaposition of these two vectors. For example:
^
ax ^
ax = ^
ax^
ax
^
ax ^
ay = ^
ax^
ay
Where ^
ax^
ax and ^
ax^
ay are the outer products obtained by multiplying the left side as a column
vector by the right side as a row vector. Note that the order is signicant.
Some additional properties of a dyadic are:
(xv) w = v (xw) = x(v w)
v (w + u) = v w + v u
(v + w) u = v u + w u
a
A vector in a reference frame can be represented as b or a^
i + b^
j + c^
k. Similarly, a dyadic can
c
5.36. Physics Module
1695
SymPy Documentation, Release 0.7.6
be represented in tensor form:
a11
a21
a31
a12
a22
a32
a13
a23
a33
or in dyadic form:
a11^
ax^
ax + a12^
ax^
ay + a13^
ax^
az + a21^
ay^
ax + a22^
ay^
ay + a23^
ay^
az + a31^
az^
ax + a32^
az^
ay + a33^
az^
az
Just as with vectors, the later representation makes it possible to keep track of which frames
the dyadic is dened with respect to. Also, the two components of each term in the dyadic
need not be in the same frame. The following is valid:
^
ax ^
by = ^
ax^
by
Dyadics can also be crossed and dotted with vectors; again, order matters:
^
ax^
ax ^
ax = ^
ax
^
ay^
ax ^
ax = ^
ay
^
ax^
ay ^
ax = 0
^
ax ^
ax^
ax = ^
ax
^
ax ^
ax^
ay = ^
ay
^
ax ^
ay^
ax = 0
^
ax ^
ay^
ax = ^
az^
ax
^
ax ^
ax^
ax = 0
^
ay^
ax ^
az = ^
ay^
ay
One can also take the time derivative of dyadics or express them in dierent frames, just like
with vectors.
Linear Momentum
The linear momentum of a particle P is dened as:
LP = mv
where m is the mass of the particle P and v is the velocity of the particle in the inertial
frame.[Likins1973] .
Similarly the linear momentum of a rigid body is dened as:
LB = mv
where m is the mass of the rigid body, B, and v is the velocity of the mass center of B in the
inertial frame.
1696
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Angular Momentum The angular momentum of a particle P about an arbitrary point O in
an inertial frame N is dened as:
N
HP /O = r mv
where r is a position vector from point O to the particle of mass m and v is the velocity of the
particle in the inertial frame.
Similarly the angular momentum of a rigid body B about a point O in an inertial frame N is
dened as:
N
HB/O =N HB/B +N HB
/O
where the angular momentum of the body about its mass center is:
N
HB/B = I
and the angular momentum of the mass center about O is:
N
HB
/O
= r mv
where I is the central inertia dyadic of rigid body B, is the inertial angular velocity of B,
r is a position vector from point O to the mass center of B, m is the mass of B and v is the
velocity of the mass center in the inertial frame.
Using momenta functions in Mechanics
momenta functions in mechanics.
The following example shows how to use the
One begins by creating the requisite symbols to describe the system. Then the reference
frame is created and the kinematics are done.
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
from sympy import symbols
from sympy.physics.mechanics import dynamicsymbols, ReferenceFrame
from sympy.physics.mechanics import RigidBody, Particle, Point, outer
from symp.physics.mechanics import linear momentum, angular momentum
m, M, l1 = symbols(m M l1)
q1d = dynamicsymbols(q1d)
N = ReferenceFrame(N)
O = Point(O)
O.set vel(N, 0 * N.x)
Ac = O.locatenew(Ac, l1 * N.x)
P = Ac.locatenew(P, l1 * N.x)
a = ReferenceFrame(a)
a.set ang vel(N, q1d * N.z)
Ac.v2pt theory(O, N, a)
P.v2pt theory(O, N, a)
Finally, the bodies that make up the system are created. In this case the system consists of a
particle Pa and a RigidBody A.
>> Pa = Particle(Pa, P, m)
>> I = outer(N.z, N.z)
>> A = RigidBody(A, Ac, a, M, (I, Ac))
5.36. Physics Module
1697
SymPy Documentation, Release 0.7.6
Then one can either choose to evaluate the the momenta of individual components of the
system or of the entire system itself.
>> linear momentum(N,A)
M*l1*q1d*N.y
>> angular momentum(O, N, Pa)
4*l1**2*m*q1d*N.z
>> linear momentum(N, A, Pa)
(M*l1*q1d + 2*l1*m*q1d)*N.y
>> angular momentum(O, N, A, Pa)
(4*l1**2*m*q1d + q1d)*N.z
It should be noted that the user can determine either momenta in any frame in mechanics as
the user is allowed to specify the reference frame when calling the function. In other words
the user is not limited to determining just inertial linear and angular momenta. Please refer
to the docstrings on each function to learn more about how each function works precisely.
Kinetic Energy
The kinetic energy of a particle P is dened as
TP =
1
mv2
2
where m is the mass of the particle P and v is the velocity of the particle in the inertial frame.
Similarly the kinetic energy of a rigid body B is dened as
TB = Tt + Tr
where the translational kinetic energy is given by:
Tt =
1
mv v
2
and the rotational kinetic energy is given by:
Tr =
1
I
2
where m is the mass of the rigid body, v is the velocity of the mass center in the inertial
frame, is the inertial angular velocity of the body and I is the central inertia dyadic.
Potential Energy Potential energy is dened as the energy possessed by a body or system
by virtue of its position or arrangement.
Since there are a variety of denitions for potential energy, this is not discussed further here.
One can learn more about this in any elementary text book on dynamics.
Lagrangian
The Lagrangian of a body or a system of bodies is dened as:
L=T V
where T and V are the kinetic and potential energies respectively.
1698
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Using energy functions in Mechanics The following example shows how to use the energy functions in mechanics.
As was discussed above in the momenta functions, one rst creates the system by going
through an identical procedure.
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
from sympy import symbols
from sympy.physics.mechanics import dynamicsymbols, ReferenceFrame, outer
from sympy.physics.mechanics import RigidBody, Particle, mechanics printing
from symp.physics.mechanics import kinetic energy, potential energy, Point
mechanics printing()
m, M, l1, g, h, H = symbols(m M l1 g h H)
omega = dynamicsymbols(omega)
N = ReferenceFrame(N)
O = Point(O)
O.set vel(N, 0 * N.x)
Ac = O.locatenew(Ac, l1 * N.x)
P = Ac.locatenew(P, l1 * N.x)
a = ReferenceFrame(a)
a.set ang vel(N, omega * N.z)
Ac.v2pt theory(O, N, a)
P.v2pt theory(O, N, a)
Pa = Particle(Pa, P, m)
I = outer(N.z, N.z)
A = RigidBody(A, Ac, a, M, (I, Ac))
The user can then determine the kinetic energy of any number of entities of the system:
>> kinetic energy(N, Pa)
2*l1**2*m*q1d**2
>> kinetic energy(N, Pa, A)
M*l1**2*q1d**2/2 + 2*l1**2*m*q1d**2 + q1d**2/2
It should be noted that the user can determine either kinetic energy relative to any frame in
mechanics as the user is allowed to specify the reference frame when calling the function. In
other words the user is not limited to determining just inertial kinetic energy.
For potential energies, the user must rst specify the potential energy of every entity of the
system using the set potential energy method. The potential energy of any number of entities comprising the system can then be determined:
>> Pa.set potential energy(m * g * h)
>> A.set potential energy(M * g * H)
>> potential energy(A, Pa)
H*M*g + g*h*m
One can also determine the Lagrangian for this system:
>> Lagrangian(Pa, A)
-H*M*g + M*l1**2*q1d**2/2 - g*h*m + 2*l1**2*m*q1d**2 + q1d**2/2
Please refer to the docstrings to learn more about each function.
Kanes Method in Physics/Mechanics mechanics provides functionality for deriving
equations of motion using Kanes method [Kane1985] (page 1916). This document will describe Kanes method as used in this module, but not how the equations are actually derived.
5.36. Physics Module
1699
SymPy Documentation, Release 0.7.6
Structure of Equations In mechanics we are assuming there are 5 basic sets of equations
needed to describe a system. They are: holonomic constraints, non-holonomic constraints,
kinematic dierential equations, dynamic equations, and dierentiated non-holonomic equations.
fh (q, t) = 0
knh (q, t)u + fnh (q, t) = 0
kkq_ (q, t)q + kku (q, t)u + fk (q, t) = 0
kd (q, t)u + fd (q, q,
u, t) = 0
kdnh (q, t)u + fdnh (q, q,
u, t) = 0
In mechanics holonomic constraints are only used for the linearization process; it is assumed
that they will be too complicated to solve for the dependent coordinate(s). If you are able to
easily solve a holonomic constraint, you should consider redening your problem in terms of
a smaller set of coordinates. Alternatively, the time-dierentiated holonomic constraints can
be supplied.
Kanes method forms two expressions, Fr and Fr , whose sum is zero. In this module, these
expressions are rearranged into the following form:
M(q, t)u = f(q, q,
u, t)
For a non-holonomic system with o total speeds and m motion constraints, we will get o - m
equations. The mass-matrix/forcing equations are then augmented in the following fashion:
[
]
kd (q, t)
M(q, t) =
kdnh (q, t)
[
]
fd (q, q,
u, t)
u, t) =
(forcing) (q, q,
fdnh (q, q,
u, t)
Kanes Method in Physics/Mechanics The formulation of the equations of motion in mechanics starts with creation of a KanesMethod object. Upon initialization of the KanesMethod
object, an inertial reference frame needs to be supplied. along with some basic system information, suchs as coordinates and speeds
>>>
>>>
>>>
>>>
>>>
from sympy.physics.mechanics import *
N = ReferenceFrame(N)
q1, q2, u1, u2 = dynamicsymbols(q1 q2 u1 u2)
q1d, q2d, u1d, u2d = dynamicsymbols(q1 q2 u1 u2, 1)
KM = KanesMethod(N, [q1, q2], [u1, u2])
It is also important to supply the order of coordinates and speeds properly if there are dependent coordinates and speeds. They must be supplied after independent coordinates and
speeds or as a keyword argument; this is shown later.
>>>
>>>
>>>
>>>
>>>
q1, q2, q3, q4 = dynamicsymbols(q1 q2 q3 q4)
u1, u2, u3, u4 = dynamicsymbols(u1 u2 u3 u4)
# Here we will assume q2 is dependent, and u2 and u3 are dependent
# We need the constraint equations to enter them though
KM = KanesMethod(N, [q1, q3, q4], [u1, u4])
Additionally, if there are auxiliary speeds, they need to be identied here. See the examples
for more information on this. In this example u4 is the auxiliary speed.
1700
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> KM = KanesMethod(N, [q1, q3, q4], [u1, u2, u3], u_auxiliary=[u4])
Kinematic dierential equations must also be supplied; there are to be provided as a list of
expressions which are each equal to zero. A trivial example follows:
>>> kd = [q1d - u1, q2d - u2]
Turning on mechanics printing() makes the expressions signicantly shorter and is recommended. Alternatively, the mprint and mpprint commands can be used.
If there are non-holonomic constraints, dependent speeds need to be specied (and so do
dependent coordinates, but they only come into play when linearizing the system). The constraints need to be supplied in a list of expressions which are equal to zero, trivial motion and
conguration constraints are shown below:
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
...
...
...
...
N = ReferenceFrame(N)
q1, q2, q3, q4 = dynamicsymbols(q1 q2 q3 q4)
q1d, q2d, q3d, q4d = dynamicsymbols(q1 q2 q3 q4, 1)
u1, u2, u3, u4 = dynamicsymbols(u1 u2 u3 u4)
#Here we will assume q2 is dependent, and u2 and u3 are dependent
speed_cons = [u2 - u1, u3 - u1 - u4]
coord_cons = [q2 - q1]
q_ind = [q1, q3, q4]
q_dep = [q2]
u_ind = [u1, u4]
u_dep = [u2, u3]
kd = [q1d - u1, q2d - u2, q3d - u3, q4d - u4]
KM = KanesMethod(N, q_ind, u_ind, kd,
q_dependent=q_dep,
configuration_constraints=coord_cons,
u_dependent=u_dep,
velocity_constraints=speed_cons)
A dictionary returning the solved qs
can also be solved for:
>>> mechanics_printing(pretty_print=False)
>>> KM.kindiffdict()
{q1: u1, q2: u2, q3: u3, q4: u4}
The nal step in forming the equations of motion is supplying a list of bodies and particles, and
a list of 2-tuples of the form (Point, Vector) or (ReferenceFrame, Vector) to represent
applied forces and torques.
>>> N = ReferenceFrame(N)
>>> q, u = dynamicsymbols(q u)
>>> qd, ud = dynamicsymbols(q u, 1)
>>> P = Point(P)
>>> P.set_vel(N, u * N.x)
>>> Pa = Particle(Pa, P, 5)
>>> BL = [Pa]
>>> FL = [(P, 7 * N.x)]
>>> KM = KanesMethod(N, [q], [u], [qd - u])
>>> (fr, frstar) = KM.kanes_equations(FL, BL)
>>> KM.mass_matrix
Matrix([[5]])
>>> KM.forcing
Matrix([[7]])
When there are motion constraints, the mass matrix is augmented by the kdnh (q, t) matrix, and
5.36. Physics Module
1701
SymPy Documentation, Release 0.7.6
the forcing vector by the fdnh (q, q,
u, t) vector.
There are also the full mass matrix and full forcing vector terms, these include the kinematic dierential equations; the mass matrix is of size (n + o) x (n + o), or square and the
size of all coordinates and speeds.
>>> KM.mass_matrix_full
Matrix([
[1, 0],
[0, 5]])
>>> KM.forcing_full
Matrix([
[u],
[7]])
Exploration of the provided examples is encouraged in order to gain more understanding of
the KanesMethod object.
Lagranges Method in Physics/Mechanics mechanics provides functionality for deriving equations of motion using Lagranges method. This document will describe Lagranges
method as used in this module, but not how the equations are actually derived.
Structure of Equations In mechanics we are assuming there are 3 basic sets of equations needed to describe a system; the constraint equations, the time dierentiated constraint
equations and the dynamic equations.
mc (q, t)q + fc (q, t) = 0
mdc (q,
q, t)
q + fdc (q,
q, t) = 0
md (q,
q, t)
q + c (q, t) + fd (q,
q, t) = 0
In this module, the expressions formed by using Lagranges equations of the second kind are
rearranged into the following form:
M(q, t)x = f(q, q,
t)
where in the case of a system without constraints:
x = q
For a constrained system with n generalized speeds and m constraints, we will get n - m
equations. The mass-matrix/forcing equations are then augmented in the following fashion:
[ ]
q
x=
[
]
M(q, t) = md (q, t)
c (q, t)
[
]
t)
F(q,
q, t) = fd (q, q,
Lagranges Method in Physics/Mechanics The formulation of the equations of motion in
mechanics using Lagranges Method starts with the creation of generalized coordinates and
a Lagrangian. The Lagrangian can either be created with the Lagrangian function or can be
a user supplied function. In this case we will supply the Lagrangian.
1702
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>>
>>>
>>>
>>>
from sympy.physics.mechanics import *
q1, q2 = dynamicsymbols(q1 q2)
q1d, q2d = dynamicsymbols(q1 q2, 1)
L = q1d**2 + q2d**2
To formulate the equations of motion we create a LagrangesMethod object. The Lagrangian
and generalized coordinates need to be supplied upon initialization.
>>> LM = LagrangesMethod(L, [q1, q2])
With that the equations of motion can be formed.
>>> mechanics_printing(pretty_print=False)
>>> LM.form_lagranges_equations()
Matrix([
[2*q1],
[2*q2]])
It is possible to obtain the mass matrix and the forcing vector.
>>> LM.mass_matrix
Matrix([
[2, 0],
[0, 2]])
>>> LM.forcing
Matrix([
[0],
[0]])
If there are any holonomic or non-holonomic constraints, they must be supplied as keyword
arguments (hol coneqs and nonhol coneqs respectively) in a list of expressions which are
equal to zero. Modifying the example above, the equations of motion can then be generated:
>>> LM = LagrangesMethod(L, [q1, q2], hol_coneqs=[q1 - q2])
When the equations of motion are generated in this case, the Lagrange multipliers are introduced; they are represented by lam1 in this case. In general, there will be as many multipliers
as there are constraint equations.
>>> LM.form_lagranges_equations()
Matrix([
[ lam1 + 2*q1],
[-lam1 + 2*q2]])
Also in the case of systems with constraints, the full mass matrix is augmented by the kdc (q, t)
matrix, and the forcing vector by the fdc (q, q,
t) vector. The full mass matrix is of size (2n +
o) x (2n + o), i.e. its a square matrix.
>>> LM.mass_matrix_full
Matrix([
[1, 0, 0, 0, 0],
[0, 1, 0, 0, 0],
[0, 0, 2, 0, -1],
[0, 0, 0, 2, 1],
[0, 0, 1, -1, 0]])
>>> LM.forcing_full
Matrix([
[q1],
5.36. Physics Module
1703
SymPy Documentation, Release 0.7.6
[q2],
[ 0],
[ 0],
[ 0]])
If there are any non-conservative forces or moments acting on the system, they must also be
supplied as keyword arguments in a list of 2-tuples of the form (Point, Vector) or (ReferenceFrame, Vector) where the Vector represents the non-conservative forces and torques.
Along with this 2-tuple, the inertial frame must also be specied as a keyword argument. This
is shown below by modifying the example above:
>>> N = ReferenceFrame(N)
>>> P = Point(P)
>>> P.set_vel(N, q1d * N.x)
>>> FL = [(P, 7 * N.x)]
>>> LM = LagrangesMethod(L, [q1, q2], forcelist=FL, frame=N)
>>> LM.form_lagranges_equations()
Matrix([
[2*q1 - 7],
[
2*q2]])
Exploration of the provided examples is encouraged in order to gain more understanding of
the LagrangesMethod object.
Linearization in Physics/Mechanics mechanics includes methods for linearizing the generated equations of motion (EOM) about an operating point (also known as the trim condition). Note that this operating point doesnt have to be an equilibrium position, it just needs
to satisfy the equations of motion.
Linearization is accomplished by taking the rst order Taylor expansion of the EOM about
the operating point. When there are no dependent coordinates or speeds this is simply the
jacobian of the right hand side about q and u. However, in the presence of constraints more
care needs to be taken. The linearization methods provided here handle these constraints
correctly.
Background
eral form:
In mechanics we assume all systems can be represented in the following genfc (q, t) = 0l1
fv (q, u, t) = 0m1
fa (q, q,
u, u,
t) = 0m1
f0 (q, q,
t) + f1 (q, u, t) = 0n1
f2 (q, u, u,
t) + f3 (q, q,
u, r, t) + f4 (q, , t) = 0(om+k)1
where
q, q Rn
u, u Ro
r Rs
Rk
In this form,
1704
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
fc represents the conguration constraint equations
fv represents the velocity constraint equations
fa represents the acceleration constraint equations
f0 and f1 form the kinematic dierential equations
f2 , f3 , and f4 form the dynamic dierential equations
q and q are the generalized coordinates and their derivatives
u and u are the generalized speeds and their derivatives
r is the system inputs
is the Lagrange multipliers
This generalized form is held inside the Linearizer class, which performs the actual linearization. Both KanesMethod and LagrangesMethod objects have methods for forming the
linearizer using the to linearizer class method.
A Note on Dependent Coordinates and Speeds
If the system being linearized contains constraint equations, this results in not all generalized coordinates being independent (i.e. q1 may depend on q2 ). With l conguration
constraints, and m velocity constraints, there are l dependent coordinates and m dependent speeds.
In general, you may pick any of the coordinates and speeds to be dependent, but in practice some choices may result in undesirable singularites. Methods for deciding which
coordinates/speeds to make dependent is behind the scope of this guide. For more information, please see [Blajer1994] (page 1916).
Once the system is coerced into the generalized form, the linearized EOM can be solved for.
The methods provided in mechanics allow for two dierent forms of the linearized EOM:
M , A, and B In this form, the forcing matrix is linearized into two separate matrices A and
B. This is the default form of the linearized EOM. The resulting equations are:
[ ]
q
[ ]
qi
M u = A
+ B r
ui
where
M R(n+o+k)(n+o+k)
A R(n+o+k)(nl+om)
B R(n+o+k)s
Note that qi and ui are just the independent coordinates and speeds, while q and u contains both the independent and dependent coordinates and speeds.
A and B In this form, the linearized EOM are brought into explicit rst order form, in terms of
just the independent coordinates and speeds. This form is often used in stability analysis
or control theory. The resulting equations are:
[ ]
[ ]
[ ]
qi
qi
=A
+ B r
ui
ui
5.36. Physics Module
1705
SymPy Documentation, Release 0.7.6
where
A R(nl+om)(nl+om)
B R(nl+om)s
To use this form set A and B=True in the linearize class method.
Linearizing Kanes Equations After initializing the KanesMethod object and forming Fr
and Fr using the kanes equations class method, linearization can be accomplished in a couple ways. The dierent methods will be demonstrated with a simple pendulum system:
>>>
>>>
>>>
>>>
>>>
>>>
from sympy import symbols, Matrix
from sympy.physics.mechanics import *
q1 = dynamicsymbols(q1)
u1 = dynamicsymbols(u1)
q1d = dynamicsymbols(q1, 1)
L, m, t, g = symbols(L, m, t, g)
>>>
>>>
>>>
>>>
# Compose world frame
N = ReferenceFrame(N)
pN = Point(N*)
pN.set_vel(N, 0)
# Angle of pendulum
# Angular velocity
>>> # A.x is along the pendulum
>>> A = N.orientnew(A, axis, [q1, N.z])
>>> A.set_ang_vel(N, u1*N.z)
>>>
>>>
>>>
>>>
# Locate point P relative to the origin N*
P = pN.locatenew(P, L*A.x)
vel_P = P.v2pt_theory(pN, N, A)
pP = Particle(pP, P, m)
>>> # Create Kinematic Differential Equations
>>> kde = Matrix([q1d - u1])
>>> # Input the force resultant at P
>>> R = m*g*N.x
>>> # Solve for eom with kanes method
>>> KM = KanesMethod(N, q_ind=[q1], u_ind=[u1], kd_eqs=kde)
>>> fr, frstar = KM.kanes_equations([(P, R)], [pP])
1. Using the Linearizer class directly: A linearizer object can be created using the
to linearizer class method. This coerces the representation found in the KanesMethod object into the generalized form described above. As the independent and dependent coordinates and speeds are specied upon creation of the KanesMethod object, there is no need to
speciy them here.
>>> linearizer = KM.to_linearizer()
The linearized EOM can then be formed with the linearize method of the Linearizer object:
>>> M, A, B = linearizer.linearize()
>>> M
1706
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Matrix([
[1,
0],
[0, -L**2*m]])
>>> A
Matrix([
[
0, 1],
[L*g*m*cos(q1(t)), 0]])
>>> B
Matrix(0, 0, [])
Alternatively, the A and B form can be generated instead by specifying A and B=True:
>>> A, B = linearizer.linearize(A_and_B=True)
>>> A
Matrix([
[
0, 1],
[-g*cos(q1(t))/L, 0]])
>>> B
Matrix(0, 0, [])
An operating point can also be specied as a dictionary or an iterable of dictionaries. This
will evaluate the linearized form at the specied point before returning the matrices:
>>> op_point = {q1: 0, u1: 0}
>>> A_op, B_op = linearizer.linearize(A_and_B=True, op_point=op_point)
>>> A_op
Matrix([
[
0, 1],
[-g/L, 0]])
Note that the same eect can be had by applying msubs to the matrices generated without
the op point kwarg:
>>> assert msubs(A, op_point) == A_op
Sometimes the returned matrices may not be in the most simplied form. Simplication can
be performed after the fact, or the Linearizer object can be made to perform simplication
internally by setting the simplify kwarg to True.
2. Using the linearize class method: The linearize method of the KanesMethod class
is provided as a nice wrapper that calls to linearizer internally, performs the linearization,
and returns the result. Note that all the kwargs available in the linearize method described
above are also available here:
>>> A, B, inp_vec = KM.linearize(A_and_B=True, op_point=op_point, new_method=True)
>>> A
Matrix([
[
0, 1],
[-g/L, 0]])
The additional output inp vec is a vector containing all found dynamicsymbols not included
in the generalized coordinate or speed vectors. These are assumed to be inputs to the system,
forming the r vector described in the background above. In this example there are no inputs,
so the vector is empty:
>>> inp_vec
Matrix(0, 0, [])
5.36. Physics Module
1707
SymPy Documentation, Release 0.7.6
Whats with the new method kwarg?
Previous releases of SymPy contained a linearization method for KanesM ethod objects.
This method is deprecated, and will be removed from future releases. Until then, you
must set new method=True in all calls to KanesMethod.linearize. After the old method
is removed, this kwarg will no longer be needed.
Linearizing Lagranges Equations Linearization of Lagranges equations proceeds much
the same as that of Kanes equations. As before, the process will be demonstrated with a
simple pendulum system:
>>>
>>>
>>>
>>>
>>>
>>>
# Redefine A and P in terms of q1d, not u1
A = N.orientnew(A, axis, [q1, N.z])
A.set_ang_vel(N, q1d*N.z)
P = pN.locatenew(P, L*A.x)
vel_P = P.v2pt_theory(pN, N, A)
pP = Particle(pP, P, m)
>>>
>>>
>>>
>>>
# Solve for eom with Lagranges method
Lag = Lagrangian(N, pP)
LM = LagrangesMethod(Lag, [q1], forcelist=[(P, R)], frame=N)
lag_eqs = LM.form_lagranges_equations()
1. Using the Linearizer class directly: A Linearizer object can be formed from a LagrangesMethod object using the to linearizer class method. The only dierence between
this process and that of the KanesMethod class is that the LagrangesMethod object doesnt already have its independent and dependent coordinates and speeds specied internally. These
must be specied in the call to to linearizer. In this example there are no dependent coordinates and speeds, but if there were they would be included in the q dep and qd dep kwargs:
>>> linearizer = LM.to_linearizer(q_ind=[q1], qd_ind=[q1d])
Once in this form, everything is the same as it was before with the KanesMethod example:
>>> A, B = linearizer.linearize(A_and_B=True, op_point=op_point)
>>> A
Matrix([
[
0, 1],
[-g/L, 0]])
2. Using the linearize class method: Similar to KanesMethod, the LagrangesMethod
class also provides a linearize method as a nice wrapper that calls to linearizer internally,
performs the linearization, and returns the result. As before, the only dierence is that the
independent and dependent coordinates and speeds must be specied in the call as well:
>>> A, B, inp_vec = LM.linearize(q_ind=[q1], qd_ind=[q1d], A_and_B=True, op_point=op_point)
>>> A
Matrix([
[
0, 1],
[-g/L, 0]])
1708
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Potential Issues While the Linearizer class should be able to linearize all systems, there
are some potential issues that could occur. These are discussed below, along with some
troubleshooting tips for solving them.
1. Symbolic linearization with A and B=True is slow This could be due to a number of
things, but the most likely one is that solving a large linear system symbolically is an expensive
operation. Specifying an operating point will reduce the expression size and speed this up.
If a purely symbolic solution is desired though (for application of many operating points at
a later period, for example) a way to get around this is to evaluate with A and B=False, and
then solve manually after applying the operating point:
>>> M, A, B = linearizer.linearize()
>>> M_op = msubs(M, op_point)
>>> A_op = msubs(A, op_point)
>>> perm_mat = linearizer.perm_mat
>>> A_lin = perm_mat.T * M_op.LUsolve(A_op)
>>> A_lin
Matrix([
[
0, 1],
[-g/L, 0]])
The fewer symbols in A and M before solving, the faster this solution will be. Thus, for large
expressions, it may be to your benet to delay conversion to the A and B form until most
symbols are subbed in for their numeric values.
2. The linearized form has nan, zoo, or oo as matrix elements There are two potential
causes for this. The rst (and the one you should check rst) is that some choices of dependent
coordinates will result in singularities at certain operating points. Coordinate partitioning in a
systemic manner to avoid this is beyond the scope of this guide; see [Blajer1994] (page 1916)
for more information.
The other potential cause for this is that the matrices may not have been in the most reduced
form before the operating point was substituted in. A simple example of this behavior is:
>>>
>>>
>>>
>>>
nan
from sympy import sin, tan
expr = sin(q1)/tan(q1)
op_point = {q1: 0}
expr.subs(op_point)
Note that if this expression was simplied before substitution, the correct value results:
>>> expr.simplify().subs(op_point)
1
A good way of avoiding this hasnt been found yet. For expressions of reasonable size, using
msubs with smart=True will apply an algorithm that tries to avoid these conditions. For large
expressions though this is extremely time consuming.
>>> msubs(expr, op_point, smart=True)
1
Further Examples The pendulum example used above was simple, but didnt include any
dependent coordinates or speeds. For a more thorough example, the same pendulum was
linearized with dependent coordinates using both Kanes and Lagranges methods:
5.36. Physics Module
1709
SymPy Documentation, Release 0.7.6
Nonminimal Coordinates Pendulum In this example we demonstrate the use of the functionality provided in mechanics for deriving the equations of motion (EOM) for a pendulum
with a nonminimal set of coordinates. As the pendulum is a one degree of freedom system, it
can be described using one coordinate and one speed (the pendulum angle, and the angular
velocity respectively). Choosing instead to describe the system using the x and y coordinates
of the mass results in a need for constraints. The system is shown below:
Ay
Ny
Ax
Nx
q1
L
q2
u1
u2
m
The system will be modeled using both Kanes and Lagranges methods, and the resulting
EOM linearized. While this is a simple problem, it should illustrate the use of the linearization
methods in the presence of constraints.
Kanes Method First we need to create the dynamicsymbols needed to describe the system
as shown in the above diagram. In this case, the generalized coordinates q1 and q2 represent
the mass x and y coordinates in the inertial N frame. Likewise, the generalized speeds u1 and
u2 represent the velocities in these directions. We also create some symbols to represent the
length and mass of the pendulum, as well as gravity and time.
>>>
>>>
>>>
>>>
>>>
>>>
>>>
from sympy.physics.mechanics import *
from sympy import symbols, atan, Matrix, solve
# Create generalized coordinates and speeds for this non-minimal realization
# q1, q2 = N.x and N.y coordinates of pendulum
# u1, u2 = N.x and N.y velocities of pendulum
q1, q2 = dynamicsymbols(q1:3)
q1d, q2d = dynamicsymbols(q1:3, level=1)
1710
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> u1, u2 = dynamicsymbols(u1:3)
>>> u1d, u2d = dynamicsymbols(u1:3, level=1)
>>> L, m, g, t = symbols(L, m, g, t)
Next, we create a world coordinate frame N , and its origin point N . The velocity of the origin
is set to 0. A second coordinate frame A is oriented such that its x-axis is along the pendulum
(as shown in the diagram above).
>>>
>>>
>>>
>>>
# Compose world frame
N = ReferenceFrame(N)
pN = Point(N*)
pN.set_vel(N, 0)
>>> # A.x is along the pendulum
>>> theta1 = atan(q2/q1)
>>> A = N.orientnew(A, axis, [theta1, N.z])
Locating the pendulum mass is then as easy as specifying its location with in terms of its x
and y coordinates in the world frame. A Particle object is then created to represent the
mass at this location.
>>> # Locate the pendulum mass
>>> P = pN.locatenew(P1, q1*N.x + q2*N.y)
>>> pP = Particle(pP, P, m)
The kinematic dierential equations (KDEs) relate the derivatives of the generalized coordinates to the generalized speeds. In this case the speeds are the derivatives, so these are
simple. A dictionary is also created to map q to u:
>>> # Calculate the kinematic differential equations
>>> kde = Matrix([q1d - u1,
...
q2d - u2])
>>> dq_dict = solve(kde, [q1d, q2d])
The velocity of the mass is then the time derivative of the position from the origin N :
>>> # Set velocity of point P
>>> P.set_vel(N, P.pos_from(pN).dt(N).subs(dq_dict))
As this system has more coordinates than degrees of freedom, constraints are needed. The
conguration constraints relate the coordinates to each other. In this case the constraint is
that the distance from the origin to the mass is always the length L (the pendulum doesnt
get longer). Likewise, the velocity constraint is that the mass velocity in the A.x direction is
always 0 (no radial velocity).
>>> f_c = Matrix([P.pos_from(pN).magnitude() - L])
>>> f_v = Matrix([P.vel(N).express(A).dot(A.x)])
>>> f_v.simplify()
The force on the system is just gravity, at point P.
>>> # Input the force resultant at P
>>> R = m*g*N.x
With the problem setup, the equations of motion can be generated using the KanesMethod
class. As there are constraints, dependent and independent coordinates need to be provided
to the class. In this case well use q2 and u2 as the independent coordinates and speeds:
5.36. Physics Module
1711
SymPy Documentation, Release 0.7.6
>>> # Derive the equations of motion using the KanesMethod class.
>>> KM = KanesMethod(N, q_ind=[q2], u_ind=[u2], q_dependent=[q1],
...
u_dependent=[u1], configuration_constraints=f_c,
...
velocity_constraints=f_v, kd_eqs=kde)
>>> (fr, frstar) = KM.kanes_equations([(P, R)], [pP])
For linearization, operating points can be specied on the call, or be substituted in afterwards. In this case well provide them in the call, supplied in a list. The A and B=True kwarg
indicates to solve invert the M matrix and solve for just the explicit linearized A and B matrices. The simplify=True kwarg indicates to simplify inside the linearize call, and return
the presimplied matrices. The cost of doing this is small for simple systems, but for larger
systems this can be a costly operation, and should be avoided.
>>> # Set the operating point to be straight down, and non-moving
>>> q_op = {q1: L, q2: 0}
>>> u_op = {u1: 0, u2: 0}
>>> ud_op = {u1d: 0, u2d: 0}
>>> # Perform the linearization
>>> A, B, inp_vec = KM.linearize(op_point=[q_op, u_op, ud_op], A_and_B=True,
...
new_method=True, simplify=True)
>>> A
Matrix([
[
0, 1],
[-g/L, 0]])
>>> B
Matrix(0, 0, [])
The resulting A matrix has dimensions 2 x 2, while the number of total states is len(q) +
len(u) = 2 + 2 = 4. This is because for constrained systems the resulting A and B form has
a partitioned state vector only containing the independent coordinates and speeds. Written
out mathematically, the system linearized about this point would be written as:
[ ] [
][ ]
0 1 q2
q2
= g
u2
0 u2
L
Lagranges Method The derivation using Lagranges method is very similar to the approach using Kanes method described above. As before, we rst create the dynamicsymbols
needed to describe the system. In this case, the generalized coordinates q1 and q2 represent
the mass x and y coordinates in the inertial N frame. This results in the time derivatives q1 and
q2 representing the velocities in these directions. We also create some symbols to represent
the length and mass of the pendulum, as well as gravity and time.
>>>
>>>
>>>
>>>
>>>
from sympy.physics.mechanics import *
from sympy import symbols, atan, Matrix
q1, q2 = dynamicsymbols(q1:3)
q1d, q2d = dynamicsymbols(q1:3, level=1)
L, m, g, t = symbols(L, m, g, t)
Next, we create a world coordinate frame N , and its origin point N . The velocity of the origin
is set to 0. A second coordinate frame A is oriented such that its x-axis is along the pendulum
(as shown in the diagram above).
>>> # Compose World Frame
>>> N = ReferenceFrame(N)
>>> pN = Point(N*)
1712
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>>
>>>
>>>
>>>
pN.set_vel(N, 0)
# A.x is along the pendulum
theta1 = atan(q2/q1)
A = N.orientnew(A, axis, [theta1, N.z])
Locating the pendulum mass is then as easy as specifying its location with in terms of its x
and y coordinates in the world frame. A Particle object is then created to represent the
mass at this location.
>>>
>>>
>>>
>>>
# Create point P, the pendulum mass
P = pN.locatenew(P1, q1*N.x + q2*N.y)
P.set_vel(N, P.pos_from(pN).dt(N))
pP = Particle(pP, P, m)
As this system has more coordinates than degrees of freedom, constraints are needed. In this
case only a single holonomic constraints is needed: the distance from the origin to the mass
is always the length L (the pendulum doesnt get longer).
>>> # Holonomic Constraint Equations
>>> f_c = Matrix([q1**2 + q2**2 - L**2])
The force on the system is just gravity, at point P.
>>> # Input the force resultant at P
>>> R = m*g*N.x
With the problem setup, the Lagrangian can be calculated, and the equations of motion
formed. Note that the call to LagrangesMethod includes the Lagrangian, the generalized
coordinates, the constraints (specied by hol coneqs or nonhol coneqs), the list of (body,
force) pairs, and the inertial frame. In contrast to the KanesMethod initializer, independent
and dependent coordinates are not partitioned inside the LagrangesMethod object. Such a
partition is supplied later.
>>>
>>>
>>>
>>>
# Calculate the lagrangian, and form the equations of motion
Lag = Lagrangian(N, pP)
LM = LagrangesMethod(Lag, [q1, q2], hol_coneqs=f_c, forcelist=[(P, R)], frame=N)
lag_eqs = LM.form_lagranges_equations()
Next, we compose the operating point dictionary, set in the hanging at rest position:
>>> # Compose operating point
>>> op_point = {q1: L, q2: 0, q1d: 0, q2d: 0, q1d.diff(t): 0, q2d.diff(t): 0}
As there are constraints in the formulation, there will be corresponding Lagrange Multipliers.
These may appear inside the linearized form as well, and thus should also be included inside
the operating point dictionary. Fortunately, the LagrangesMethod class provides an easy way
of solving for the multipliers at a given operating point using the solve multipliers method.
>>> # Solve for multiplier operating point
>>> lam_op = LM.solve_multipliers(op_point=op_point)
With this solution, linearization can be completed. Note that in contrast to the KanesMethod
approach, the LagrangesMethod.linearize method also requires the partitioning of the generalized coordinates and their time derivatives into independent and dependent vectors. This
is the same as what was passed into the KanesMethod constructor above:
>>> op_point.update(lam_op)
>>> # Perform the Linearization
5.36. Physics Module
1713
SymPy Documentation, Release 0.7.6
>>> A, B, inp_vec = LM.linearize([q2], [q2d], [q1], [q1d],
...
op_point=op_point, A_and_B=True)
>>> A
Matrix([
[
0, 1],
[-g/L, 0]])
>>> B
Matrix(0, 0, [])
The resulting A matrix has dimensions 2 x 2, while the number of total states is 2*len(q)
= 4. This is because for constrained systems the resulting A and B form has a partitioned
state vector only containing the independent coordinates and their derivatives. Written out
mathematically, the system linearized about this point would be written as:
[ ] [
][ ]
0 1 q2
q2
= g
q2
0 q2
L
Examples for Physics/Mechanics Here are some examples that illustrate how one typically uses this module. We have ordered the examples roughly according to increasing diculty. If you have used this module to do something others might nd useful or interesting,
consider adding it here!
A rolling disc The disc is assumed to be innitely thin, in contact with the ground at only
1 point, and it is rolling without slip on the ground. See the image below.
nz
ny
rz
R
ry
rx
nx
1714
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
We model the rolling disc in three dierent ways, to show more of the functionality of this
module.
A rolling disc, with Kanes method Here the denition of the rolling discs kinematics is
formed from the contact point up, removing the need to introduce generalized speeds. Only
3 conguration and three speed variables are need to describe this system, along with the
discs mass and radius, and the local gravity (note that mass will drop out).
>>>
>>>
>>>
>>>
>>>
>>>
from sympy import symbols, sin, cos, tan
from sympy.physics.mechanics import *
q1, q2, q3, u1, u2, u3 = dynamicsymbols(q1 q2 q3 u1 u2 u3)
q1d, q2d, q3d, u1d, u2d, u3d = dynamicsymbols(q1 q2 q3 u1 u2 u3, 1)
r, m, g = symbols(r m g)
mechanics_printing(pretty_print=False)
The kinematics are formed by a series of simple rotations. Each simple rotation creates a new
frame, and the next rotation is dened by the new frames basis vectors. This example uses a
3-1-2 series of rotations, or Z, X, Y series of rotations. Angular velocity for this is dened using
the second frames basis (the lean frame); it is for this reason that we dened intermediate
frames, rather than using a body-three orientation.
>>>
>>>
>>>
>>>
>>>
>>>
N = ReferenceFrame(N)
Y = N.orientnew(Y, Axis, [q1, N.z])
L = Y.orientnew(L, Axis, [q2, Y.x])
R = L.orientnew(R, Axis, [q3, L.y])
w_R_N_qd = R.ang_vel_in(N)
R.set_ang_vel(N, u1 * L.x + u2 * L.y + u3 * L.z)
This is the translational kinematics. We create a point with no velocity in N; this is the contact
point between the disc and ground. Next we form the position vector from the contact point
to the discs center of mass. Finally we form the velocity and acceleration of the disc.
>>> C = Point(C)
>>> C.set_vel(N, 0)
>>> Dmc = C.locatenew(Dmc, r * L.z)
>>> Dmc.v2pt_theory(C, N, R)
r*u2*L.x - r*u1*L.y
This is a simple way to form the inertia dyadic. The inertia of the disc does not change within
the lean frame as the disc rolls; this will make for simpler equations in the end.
>>> I = inertia(L, m / 4 * r**2, m / 2 * r**2, m / 4 * r**2)
>>> mprint(I)
m*r**2/4*(L.x|L.x) + m*r**2/2*(L.y|L.y) + m*r**2/4*(L.z|L.z)
Kinematic dierential equations; how the generalized coordinate time derivatives relate to
generalized speeds.
>>> kd = [dot(R.ang_vel_in(N) - w_R_N_qd, uv) for uv in L]
Creation of the force list; it is the gravitational force at the center of mass of the disc. Then
we create the disc by assigning a Point to the center of mass attribute, a ReferenceFrame to
the frame attribute, and mass and inertia. Then we form the body list.
>>> ForceList = [(Dmc, - m * g * Y.z)]
>>> BodyD = RigidBody(BodyD, Dmc, R, m, (I, Dmc))
>>> BodyList = [BodyD]
5.36. Physics Module
1715
SymPy Documentation, Release 0.7.6
Finally we form the equations of motion, using the same steps we did before. Specify inertial
frame, supply generalized coordinates and speeds, supply kinematic dierential equation dictionary, compute Fr from the force list and Fr* from the body list, compute the mass matrix
and forcing terms, then solve for the u dots (time derivatives of the generalized speeds).
>>> KM = KanesMethod(N, q_ind=[q1, q2, q3], u_ind=[u1, u2, u3], kd_eqs=kd)
>>> (fr, frstar) = KM.kanes_equations(ForceList, BodyList)
>>> MM = KM.mass_matrix
>>> forcing = KM.forcing
>>> rhs = MM.inv() * forcing
>>> kdd = KM.kindiffdict()
>>> rhs = rhs.subs(kdd)
>>> rhs.simplify()
>>> mprint(rhs)
Matrix([
[(4*g*sin(q2) + 6*r*u2*u3 - r*u3**2*tan(q2))/(5*r)],
[
-2*u1*u3/3],
[
(-2*u2 + u3*tan(q2))*u1]])
A rolling disc, with Kanes method and constraint forces We will now revisit the rolling
disc example, except this time we are bringing the non-contributing (constraint) forces into
evidence. See [Kane1985] (page 1916) for a more thorough explanation of this. Here, we will
turn on the automatic simplifcation done when doing vector operations. It makes the outputs
nicer for small problems, but can cause larger vector operations to hang.
>>>
>>>
>>>
>>>
>>>
>>>
from sympy import symbols, sin, cos, tan
from sympy.physics.mechanics import *
mechanics_printing(pretty_print=False)
q1, q2, q3, u1, u2, u3 = dynamicsymbols(q1 q2 q3 u1 u2 u3)
q1d, q2d, q3d, u1d, u2d, u3d = dynamicsymbols(q1 q2 q3 u1 u2 u3, 1)
r, m, g = symbols(r m g)
These two lines introduce the extra quantities needed to nd the constraint forces.
>>> u4, u5, u6, f1, f2, f3 = dynamicsymbols(u4 u5 u6 f1 f2 f3)
Most of the main code is the same as before.
>>>
>>>
>>>
>>>
>>>
>>>
N = ReferenceFrame(N)
Y = N.orientnew(Y, Axis, [q1, N.z])
L = Y.orientnew(L, Axis, [q2, Y.x])
R = L.orientnew(R, Axis, [q3, L.y])
w_R_N_qd = R.ang_vel_in(N)
R.set_ang_vel(N, u1 * L.x + u2 * L.y + u3 * L.z)
The denition of rolling without slip necessitates that the velocity of the contact point is zero;
as part of bringing the constraint forces into evidence, we have to introduce speeds at this
point, which will by denition always be zero. They are normal to the ground, along the path
which the disc is rolling, and along the ground in an perpendicular direction.
>>>
>>>
>>>
>>>
>>>
>>>
C = Point(C)
C.set_vel(N, u4 * L.x + u5 * (Y.z ^ L.x) + u6 * Y.z)
Dmc = C.locatenew(Dmc, r * L.z)
vel = Dmc.v2pt_theory(C, N, R)
I = inertia(L, m / 4 * r**2, m / 2 * r**2, m / 4 * r**2)
kd = [dot(R.ang_vel_in(N) - w_R_N_qd, uv) for uv in L]
1716
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Just as we previously introduced three speeds as part of this process, we also introduce three
forces; they are in the same direction as the speeds, and represent the constraint forces in
those directions.
>>> ForceList = [(Dmc, - m * g * Y.z), (C, f1 * L.x + f2 * (Y.z ^ L.x) + f3 * Y.z)]
>>> BodyD = RigidBody(BodyD, Dmc, R, m, (I, Dmc))
>>> BodyList = [BodyD]
>>> KM = KanesMethod(N, q_ind=[q1, q2, q3], u_ind=[u1, u2, u3], kd_eqs=kd,
...
u_auxiliary=[u4, u5, u6])
>>> (fr, frstar) = KM.kanes_equations(ForceList, BodyList)
>>> MM = KM.mass_matrix
>>> forcing = KM.forcing
>>> rhs = MM.inv() * forcing
>>> kdd = KM.kindiffdict()
>>> rhs = rhs.subs(kdd)
>>> rhs.simplify()
>>> mprint(rhs)
Matrix([
[(4*g*sin(q2) + 6*r*u2*u3 - r*u3**2*tan(q2))/(5*r)],
[
-2*u1*u3/3],
[
(-2*u2 + u3*tan(q2))*u1]])
>>> from sympy import trigsimp, signsimp, collect, factor_terms
>>> def simplify_auxiliary_eqs(w):
...
return signsimp(trigsimp(collect(collect(factor_terms(w), f2), m*r)))
>>> mprint(KM.auxiliary_eqs.applyfunc(simplify_auxiliary_eqs))
Matrix([
[
-m*r*(u1*u3 + u2) + f1],
[-m*r*u1**2*sin(q2) - m*r*u2*u3/cos(q2) + m*r*cos(q2)*u1 + f2],
[
-g*m + m*r*(u1**2*cos(q2) + sin(q2)*u1) + f3]])
A rolling disc using Lagranges Method Here the rolling disc is formed from the contact
point up, removing the need to introduce generalized speeds. Only 3 conguration and 3
speed variables are needed to describe this system, along with the discs mass and radius,
and the local gravity.
>>>
>>>
>>>
>>>
>>>
>>>
from sympy import symbols, cos, sin
from sympy.physics.mechanics import *
mechanics_printing(pretty_print=False)
q1, q2, q3 = dynamicsymbols(q1 q2 q3)
q1d, q2d, q3d = dynamicsymbols(q1 q2 q3, 1)
r, m, g = symbols(r m g)
The kinematics are formed by a series of simple rotations. Each simple rotation creates a new
frame, and the next rotation is dened by the new frames basis vectors. This example uses
a 3-1-2 series of rotations, or Z, X, Y series of rotations. Angular velocity for this is dened
using the second frames basis (the lean frame).
>>>
>>>
>>>
>>>
N
Y
L
R
=
=
=
=
ReferenceFrame(N)
N.orientnew(Y, Axis, [q1, N.z])
Y.orientnew(L, Axis, [q2, Y.x])
L.orientnew(R, Axis, [q3, L.y])
This is the translational kinematics. We create a point with no velocity in N; this is the contact
point between the disc and ground. Next we form the position vector from the contact point
to the discs center of mass. Finally we form the velocity and acceleration of the disc.
5.36. Physics Module
1717
SymPy Documentation, Release 0.7.6
>>> C = Point(C)
>>> C.set_vel(N, 0)
>>> Dmc = C.locatenew(Dmc, r * L.z)
>>> Dmc.v2pt_theory(C, N, R)
r*(sin(q2)*q1 + q3)*L.x - r*q2*L.y
Forming the inertia dyadic.
>>> I = inertia(L, m / 4 * r**2, m / 2 * r**2, m / 4 * r**2)
>>> mprint(I)
m*r**2/4*(L.x|L.x) + m*r**2/2*(L.y|L.y) + m*r**2/4*(L.z|L.z)
>>> BodyD = RigidBody(BodyD, Dmc, R, m, (I, Dmc))
We then set the potential energy and determine the Lagrangian of the rolling disc.
>>> BodyD.set_potential_energy(- m * g * r * cos(q2))
>>> Lag = Lagrangian(N, BodyD)
Then the equations of motion are generated by initializing the LagrangesMethod object. Finally we solve for the generalized accelerations(q double dots) with the rhs method.
>>> q = [q1, q2, q3]
>>> l = LagrangesMethod(Lag, q)
>>> le = l.form_lagranges_equations()
>>> le.simplify(); le
Matrix([
[m*r**2*(12*sin(q2)*q3 + 10*sin(2*q2)*q1*q2 + 12*cos(q2)*q2*q3 - 5*cos(2*q2)*q1 + 7*q1)/8],
[
m*r*(8*g*sin(q2) - 5*r*sin(2*q2)*q1**2 - 12*r*cos(q2)*q1*q3 + 10*r*q2)/8],
[
3*m*r**2*(sin(q2)*q1 + cos(q2)*q1*q2 + q3)/2]]
>>> lrhs = l.rhs(); lrhs.simplify(); lrhs
Matrix([
[
q1],
[
q2],
[
q3],
[
-2*(2*tan(q2)*q1 + 3*q3/cos(q2))*q2],
[-4*g*sin(q2)/(5*r) + sin(2*q2)*q1**2/2 + 6*cos(q2)*q1*q3/5],
[
(-5*cos(q2)*q1 + 6*tan(q2)*q3 + 4*q1/cos(q2))*q2]])
A bicycle The bicycle is an interesting system in that it has multiple rigid bodies, nonholonomic constraints, and a holonomic constraint. The linearized equations of motion are
presented in [Meijaard2007] (page 1916). This example will go through construction of the
equations of motion in mechanics.
>>> from sympy import *
>>> from sympy.physics.mechanics import *
>>> print(Calculation of Linearized Bicycle \A\ Matrix,
...
with States: Roll, Steer, Roll Rate, Steer Rate)
Calculation of Linearized Bicycle A Matrix, with States: Roll, Steer, Roll Rate, Steer Rate
Note that this code has been crudely ported from Autolev, which is the reason for some of the
unusual naming conventions. It was purposefully as similar as possible in order to aid initial
porting & debugging. We set Vector.simp to False (in case it has been set True elsewhere),
since it slows down the computations:
>>> Vector.simp = False
>>> mechanics_printing(pretty_print=False)
1718
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Declaration of Coordinates & Speeds: A simple denition for qdots, qd = u,is used in this
code. Speeds are: yaw frame ang. rate, roll frame ang. rate, rear wheel frame ang. rate
(spinning motion), frame ang. rate (pitching motion), steering frame ang. rate, and front
wheel ang. rate (spinning motion). Wheel positions are ignorable coordinates, so they are
not introduced.
>>>
>>>
>>>
>>>
q1, q2, q3, q4, q5 = dynamicsymbols(q1 q2 q3 q4 q5)
q1d, q2d, q4d, q5d = dynamicsymbols(q1 q2 q4 q5, 1)
u1, u2, u3, u4, u5, u6 = dynamicsymbols(u1 u2 u3 u4 u5 u6)
u1d, u2d, u3d, u4d, u5d, u6d = dynamicsymbols(u1 u2 u3 u4 u5 u6, 1)
Declaration of Systems Parameters: The below symbols should be fairly self-explanatory.
>>>
>>>
>>>
>>>
>>>
...
>>>
>>>
WFrad, WRrad, htangle, forkoffset = symbols(WFrad WRrad htangle forkoffset)
forklength, framelength, forkcg1 = symbols(forklength framelength forkcg1)
forkcg3, framecg1, framecg3, Iwr11 = symbols(forkcg3 framecg1 framecg3 Iwr11)
Iwr22, Iwf11, Iwf22, Iframe11 = symbols(Iwr22 Iwf11 Iwf22 Iframe11)
Iframe22, Iframe33, Iframe31, Ifork11 = \
symbols(Iframe22 Iframe33 Iframe31 Ifork11)
Ifork22, Ifork33, Ifork31, g = symbols(Ifork22 Ifork33 Ifork31 g)
mframe, mfork, mwf, mwr = symbols(mframe mfork mwf mwr)
Set up reference frames for the system: N - inertial Y - yaw R - roll WR - rear wheel, rotation
angle is ignorable coordinate so not oriented Frame - bicycle frame TempFrame - statically
rotated frame for easier reference inertia denition Fork - bicycle fork TempFork - statically
rotated frame for easier reference inertia denition WF - front wheel, again posses a ignorable
coordinate
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
N = ReferenceFrame(N)
Y = N.orientnew(Y, Axis, [q1, N.z])
R = Y.orientnew(R, Axis, [q2, Y.x])
Frame = R.orientnew(Frame, Axis, [q4 + htangle, R.y])
WR = ReferenceFrame(WR)
TempFrame = Frame.orientnew(TempFrame, Axis, [-htangle, Frame.y])
Fork = Frame.orientnew(Fork, Axis, [q5, Frame.x])
TempFork = Fork.orientnew(TempFork, Axis, [-htangle, Fork.y])
WF = ReferenceFrame(WF)
Kinematics of the Bicycle: First block of code is forming the positions of the relevant points
rear wheel contact -> rear wheels center of mass -> frames center of mass + frame/fork
connection -> forks center of mass + front wheels center of mass -> front wheel contact
point.
>>>
>>>
>>>
>>>
>>>
>>>
>>>
...
WR_cont = Point(WR_cont)
WR_mc = WR_cont.locatenew(WR_mc, WRrad * R.z)
Steer = WR_mc.locatenew(Steer, framelength * Frame.z)
Frame_mc = WR_mc.locatenew(Frame_mc, -framecg1 * Frame.x + framecg3 * Frame.z)
Fork_mc = Steer.locatenew(Fork_mc, -forkcg1 * Fork.x + forkcg3 * Fork.z)
WF_mc = Steer.locatenew(WF_mc, forklength * Fork.x + forkoffset * Fork.z)
WF_cont = WF_mc.locatenew(WF_cont, WFrad*(dot(Fork.y, Y.z)*Fork.y - \
Y.z).normalize())
Set the angular velocity of each frame: Angular accelerations end up being calculated automatically by dierentiating the angular velocities when rst needed. :: u1 is yaw rate u2 is
roll rate u3 is rear wheel rate u4 is frame pitch rate u5 is fork steer rate u6 is front wheel
rate
5.36. Physics Module
1719
SymPy Documentation, Release 0.7.6
>>>
>>>
>>>
>>>
>>>
>>>
Y.set_ang_vel(N, u1 * Y.z)
R.set_ang_vel(Y, u2 * R.x)
WR.set_ang_vel(Frame, u3 * Frame.y)
Frame.set_ang_vel(R, u4 * Frame.y)
Fork.set_ang_vel(Frame, u5 * Fork.x)
WF.set_ang_vel(Fork, u6 * Fork.y)
Form the velocities of the points, using the 2-point theorem. Accelerations again are calculated automatically when rst needed.
>>> WR_cont.set_vel(N, 0)
>>> WR_mc.v2pt_theory(WR_cont, N, WR)
WRrad*(u1*sin(q2) + u3 + u4)*R.x - WRrad*u2*R.y
>>> Steer.v2pt_theory(WR_mc, N, Frame)
WRrad*(u1*sin(q2) + u3 + u4)*R.x - WRrad*u2*R.y
>>> Frame_mc.v2pt_theory(WR_mc, N, Frame)
WRrad*(u1*sin(q2) + u3 + u4)*R.x - WRrad*u2*R.y
>>> Fork_mc.v2pt_theory(Steer, N, Fork)
WRrad*(u1*sin(q2) + u3 + u4)*R.x - WRrad*u2*R.y
>>> WF_mc.v2pt_theory(Steer, N, Fork)
WRrad*(u1*sin(q2) + u3 + u4)*R.x - WRrad*u2*R.y
>>> WF_cont.v2pt_theory(WF_mc, N, WF)
WRrad*(u1*sin(q2) + u3 + u4)*R.x - WRrad*u2*R.y
+ framelength*(u1*sin(q2) + u4)*Frame.x - framelength
+ framecg3*(u1*sin(q2) + u4)*Frame.x + (-framecg1*(u1
+ framelength*(u1*sin(q2) + u4)*Frame.x - framelength
+ framelength*(u1*sin(q2) + u4)*Frame.x - framelength
+ framelength*(u1*sin(q2) + u4)*Frame.x - framelength
Sets the inertias of each body. Uses the inertia frame to construct the inertia dyadics. Wheel
inertias are only dened by principal moments of inertia, and are in fact constant in the frame
and fork reference frames; it is for this reason that the orientations of the wheels does not
need to be dened. The frame and fork inertias are dened in the Temp frames which are
xed to the appropriate body frames; this is to allow easier input of the reference values of
the benchmark paper. Note that due to slightly dierent orientations, the products of inertia
need to have their signs ipped; this is done later when entering the numerical value.
>>>
...
>>>
>>>
>>>
Frame_I = (inertia(TempFrame, Iframe11, Iframe22, Iframe33, 0, 0,
Iframe31), Frame_mc)
Fork_I = (inertia(TempFork, Ifork11, Ifork22, Ifork33, 0, 0, Ifork31), Fork_mc)
WR_I = (inertia(Frame, Iwr11, Iwr22, Iwr11), WR_mc)
WF_I = (inertia(Fork, Iwf11, Iwf22, Iwf11), WF_mc)
Declaration of the RigidBody containers.
>>>
>>>
>>>
>>>
BodyFrame = RigidBody(BodyFrame, Frame_mc, Frame, mframe, Frame_I)
BodyFork = RigidBody(BodyFork, Fork_mc, Fork, mfork, Fork_I)
BodyWR = RigidBody(BodyWR, WR_mc, WR, mwr, WR_I)
BodyWF = RigidBody(BodyWF, WF_mc, WF, mwf, WF_I)
>>> print(Before Forming the List of Nonholonomic Constraints.)
Before Forming the List of Nonholonomic Constraints.
The kinematic dierential equations; they are dened quite simply. Each entry in this list is
equal to zero.
>>> kd = [q1d - u1, q2d - u2, q4d - u4, q5d - u5]
The nonholonomic constraints are the velocity of the front wheel contact point dotted into the
X, Y, and Z directions; the yaw frame is used as it is closer to the front wheel (1 less DCM
connecting them). These constraints force the velocity of the front wheel contact point to be
0 in the inertial frame; the X and Y direction constraints enforce a no-slip condition, and the
Z direction constraint forces the front wheel contact point to not move away from the ground
1720
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
frame, essentially replicating the holonomic constraint which does not allow the frame pitch
to change in an invalid fashion.
>>> conlist_speed = [WF_cont.vel(N) & Y.x,
...
WF_cont.vel(N) & Y.y,
...
WF_cont.vel(N) & Y.z]
The holonomic constraint is that the position from the rear wheel contact point to the front
wheel contact point when dotted into the normal-to-ground plane direction must be zero;
eectively that the front and rear wheel contact points are always touching the ground plane.
This is actually not part of the dynamic equations, but instead is necessary for the linearization
process.
>>> conlist_coord = [WF_cont.pos_from(WR_cont) & Y.z]
The force list; each body has the appropriate gravitational force applied at its center of mass.
>>> FL = [(Frame_mc, -mframe * g * Y.z), (Fork_mc, -mfork * g * Y.z),
...
(WF_mc, -mwf * g * Y.z), (WR_mc, -mwr * g * Y.z)]
>>> BL = [BodyFrame, BodyFork, BodyWR, BodyWF]
The N frame is the inertial frame, coordinates are supplied in the order of independent, dependent coordinates. The kinematic dierential equations are also entered here. Here the
independent speeds are specied, followed by the dependent speeds, along with the nonholonomic constraints. The dependent coordinate is also provided, with the holonomic constraint. Again, this is only comes into play in the linearization process, but is necessary for
the linearization to correctly work.
>>> KM = KanesMethod(N, q_ind=[q1, q2, q5],
...
q_dependent=[q4], configuration_constraints=conlist_coord,
...
u_ind=[u2, u3, u5],
...
u_dependent=[u1, u4, u6], velocity_constraints=conlist_speed,
...
kd_eqs=kd)
>>> print(Before Forming Generalized Active and Inertia Forces, Fr and Fr*)
Before Forming Generalized Active and Inertia Forces, Fr and Fr*
>>> (fr, frstar) = KM.kanes_equations(FL, BL)
>>> print(Base Equations of Motion Computed)
Base Equations of Motion Computed
This is the start of entering in the numerical values from the benchmark paper to validate the
eigenvalues of the linearized equations from this model to the reference eigenvalues. Look
at the aforementioned paper for more information. Some of these are intermediate values,
used to transform values from the paper into the coordinate systems used in this model.
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
...
>>>
...
>>>
PaperRadRear
PaperRadFront
HTA
TrailPaper
rake
PaperWb
PaperFrameCgX
PaperFrameCgZ
PaperForkCgX
PaperForkCgZ
FrameLength
=
=
=
=
=
=
=
=
=
=
=
FrameCGNorm
FrameCGPar
0.3
0.35
evalf.N(pi/2-pi/10)
0.08
evalf.N(-(TrailPaper*sin(HTA)-(PaperRadFront*cos(HTA))))
1.02
0.3
0.9
0.9
0.7
evalf.N(PaperWb*sin(HTA) - (rake - \
(PaperRadFront - PaperRadRear)*cos(HTA)))
evalf.N((PaperFrameCgZ - PaperRadRear - \
(PaperFrameCgX/sin(HTA))*cos(HTA))*sin(HTA))
evalf.N((PaperFrameCgX / sin(HTA) + \
5.36. Physics Module
1721
SymPy Documentation, Release 0.7.6
...
...
>>>
>>>
>>>
>>>
...
>>>
...
>>>
...
tempa
tempb
tempc
PaperForkL
=
=
=
=
ForkCGNorm
ForkCGPar
(PaperFrameCgZ - PaperRadRear - \
PaperFrameCgX / sin(HTA) * cos(HTA)) * cos(HTA)))
evalf.N((PaperForkCgZ - PaperRadFront))
evalf.N((PaperWb-PaperForkCgX))
evalf.N(sqrt(tempa**2 + tempb**2))
evalf.N((PaperWb*cos(HTA) - \
(PaperRadFront - PaperRadRear)*sin(HTA)))
evalf.N(rake + (tempc * sin(pi/2 - \
HTA - acos(tempa/tempc))))
evalf.N(tempc * cos((pi/2 - HTA) - \
acos(tempa/tempc)) - PaperForkL)
Here is the nal assembly of the numerical values. The symbol v is the forward speed of the
bicycle (a concept which only makes sense in the upright, static equilibrium case?). These
are in a dictionary which will later be substituted in. Again the sign on the product of inertia
values is ipped here, due to dierent orientations of coordinate systems.
>>>
>>>
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
>>>
>>>
v = Symbol(v)
val_dict = {
WFrad: PaperRadFront,
WRrad: PaperRadRear,
htangle: HTA,
forkoffset: rake,
forklength: PaperForkL,
framelength: FrameLength,
forkcg1: ForkCGPar,
forkcg3: ForkCGNorm,
framecg1: FrameCGNorm,
framecg3: FrameCGPar,
Iwr11: 0.0603,
Iwr22: 0.12,
Iwf11: 0.1405,
Iwf22: 0.28,
Ifork11: 0.05892,
Ifork22: 0.06,
Ifork33: 0.00708,
Ifork31: 0.00756,
Iframe11: 9.2,
Iframe22: 11,
Iframe33: 2.8,
Iframe31: -2.4,
mfork: 4,
mframe: 85,
mwf: 3,
mwr: 2,
g: 9.81,
q1: 0,
q2: 0,
q4: 0,
q5: 0,
u1: 0,
u2: 0,
u3: v/PaperRadRear,
u4: 0,
u5: 0,
u6: v/PaperRadFront}
kdd = KM.kindiffdict()
print(Before Linearization of the \Forcing\ Term)
1722
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Before Linearization of the Forcing Term
Linearizes the forcing vector; the equations are set up as MM udot = forcing, where MM is the
mass matrix, udot is the vector representing the time derivatives of the generalized speeds,
and forcing is a vector which contains both external forcing terms and internal forcing terms,
such as centripetal or Coriolis forces. This actually returns a matrix with as many rows as total
coordinates and speeds, but only as many columns as independent coordinates and speeds.
(Note that below this is commented out, as it takes a few minutes to run, which is not good
when performing the doctests)
>>> # forcing_lin = KM.linearize()[0].subs(sub_dict)
As mentioned above, the size of the linearized forcing terms is expanded to include both qs
and us, so the mass matrix must have this done as well. This will likely be changed to be part
of the linearized process, for future reference.
>>> MM_full = (KM._k_kqdot).row_join(zeros(4, 6)).col_join(
...
(zeros(6, 4)).row_join(KM.mass_matrix))
>>> print(Before Substitution of Numerical Values)
Before Substitution of Numerical Values
I think this is pretty self explanatory. It takes a really long time though. Ive experimented with
using evalf with substitution, this failed due to maximum recursion depth being exceeded; I
also tried lambdifying this, and it is also not successful. (again commented out due to speed)
>>> # MM_full = MM_full.subs(val_dict)
>>> # forcing_lin = forcing_lin.subs(val_dict)
>>> # print(Before .evalf() call)
>>> # MM_full = MM_full.evalf()
>>> # forcing_lin = forcing_lin.evalf()
Finally, we construct an A matrix for the form xdot = A x (x being the state vector, although
in this case, the sizes are a little o). The following line extracts only the minimum entries
required for eigenvalue analysis, which correspond to rows and columns for lean, steer, lean
rate, and steer rate. (this is all commented out due to being dependent on the above code,
which is also commented out):
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
#
#
#
#
#
#
#
#
#
#
#
#
#
Amat = MM_full.inv() * forcing_lin
A = Amat.extract([1,2,4,6],[1,2,3,5])
print(A)
print(v = 1)
print(A.subs(v, 1).eigenvals())
print(v = 2)
print(A.subs(v, 2).eigenvals())
print(v = 3)
print(A.subs(v, 3).eigenvals())
print(v = 4)
print(A.subs(v, 4).eigenvals())
print(v = 5)
print(A.subs(v, 5).eigenvals())
Upon running the above code yourself, enabling the commented out lines, compare the computed eigenvalues to those is the referenced paper. This concludes the bicycle example.
Nonminimal Coordinates Pendulum In this example we demonstrate the use of the functionality provided in mechanics for deriving the equations of motion (EOM) for a pendulum
5.36. Physics Module
1723
SymPy Documentation, Release 0.7.6
with a nonminimal set of coordinates. As the pendulum is a one degree of freedom system, it
can be described using one coordinate and one speed (the pendulum angle, and the angular
velocity respectively). Choosing instead to describe the system using the x and y coordinates
of the mass results in a need for constraints. The system is shown below:
Ay
Ny
Ax
Nx
q1
L
q2
u1
u2
m
The system will be modeled using both Kanes and Lagranges methods, and the resulting
EOM linearized. While this is a simple problem, it should illustrate the use of the linearization
methods in the presence of constraints.
Kanes Method First we need to create the dynamicsymbols needed to describe the system
as shown in the above diagram. In this case, the generalized coordinates q1 and q2 represent
the mass x and y coordinates in the inertial N frame. Likewise, the generalized speeds u1 and
u2 represent the velocities in these directions. We also create some symbols to represent the
length and mass of the pendulum, as well as gravity and time.
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
from sympy.physics.mechanics import *
from sympy import symbols, atan, Matrix, solve
# Create generalized coordinates and speeds for this non-minimal realization
# q1, q2 = N.x and N.y coordinates of pendulum
# u1, u2 = N.x and N.y velocities of pendulum
q1, q2 = dynamicsymbols(q1:3)
q1d, q2d = dynamicsymbols(q1:3, level=1)
u1, u2 = dynamicsymbols(u1:3)
u1d, u2d = dynamicsymbols(u1:3, level=1)
1724
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> L, m, g, t = symbols(L, m, g, t)
Next, we create a world coordinate frame N , and its origin point N . The velocity of the origin
is set to 0. A second coordinate frame A is oriented such that its x-axis is along the pendulum
(as shown in the diagram above).
>>>
>>>
>>>
>>>
# Compose world frame
N = ReferenceFrame(N)
pN = Point(N*)
pN.set_vel(N, 0)
>>> # A.x is along the pendulum
>>> theta1 = atan(q2/q1)
>>> A = N.orientnew(A, axis, [theta1, N.z])
Locating the pendulum mass is then as easy as specifying its location with in terms of its x
and y coordinates in the world frame. A Particle object is then created to represent the
mass at this location.
>>> # Locate the pendulum mass
>>> P = pN.locatenew(P1, q1*N.x + q2*N.y)
>>> pP = Particle(pP, P, m)
The kinematic dierential equations (KDEs) relate the derivatives of the generalized coordinates to the generalized speeds. In this case the speeds are the derivatives, so these are
simple. A dictionary is also created to map q to u:
>>> # Calculate the kinematic differential equations
>>> kde = Matrix([q1d - u1,
...
q2d - u2])
>>> dq_dict = solve(kde, [q1d, q2d])
The velocity of the mass is then the time derivative of the position from the origin N :
>>> # Set velocity of point P
>>> P.set_vel(N, P.pos_from(pN).dt(N).subs(dq_dict))
As this system has more coordinates than degrees of freedom, constraints are needed. The
conguration constraints relate the coordinates to each other. In this case the constraint is
that the distance from the origin to the mass is always the length L (the pendulum doesnt
get longer). Likewise, the velocity constraint is that the mass velocity in the A.x direction is
always 0 (no radial velocity).
>>> f_c = Matrix([P.pos_from(pN).magnitude() - L])
>>> f_v = Matrix([P.vel(N).express(A).dot(A.x)])
>>> f_v.simplify()
The force on the system is just gravity, at point P.
>>> # Input the force resultant at P
>>> R = m*g*N.x
With the problem setup, the equations of motion can be generated using the KanesMethod
class. As there are constraints, dependent and independent coordinates need to be provided
to the class. In this case well use q2 and u2 as the independent coordinates and speeds:
>>> # Derive the equations of motion using the KanesMethod class.
>>> KM = KanesMethod(N, q_ind=[q2], u_ind=[u2], q_dependent=[q1],
...
u_dependent=[u1], configuration_constraints=f_c,
5.36. Physics Module
1725
SymPy Documentation, Release 0.7.6
...
velocity_constraints=f_v, kd_eqs=kde)
>>> (fr, frstar) = KM.kanes_equations([(P, R)], [pP])
For linearization, operating points can be specied on the call, or be substituted in afterwards. In this case well provide them in the call, supplied in a list. The A and B=True kwarg
indicates to solve invert the M matrix and solve for just the explicit linearized A and B matrices. The simplify=True kwarg indicates to simplify inside the linearize call, and return
the presimplied matrices. The cost of doing this is small for simple systems, but for larger
systems this can be a costly operation, and should be avoided.
>>> # Set the operating point to be straight down, and non-moving
>>> q_op = {q1: L, q2: 0}
>>> u_op = {u1: 0, u2: 0}
>>> ud_op = {u1d: 0, u2d: 0}
>>> # Perform the linearization
>>> A, B, inp_vec = KM.linearize(op_point=[q_op, u_op, ud_op], A_and_B=True,
...
new_method=True, simplify=True)
>>> A
Matrix([
[
0, 1],
[-g/L, 0]])
>>> B
Matrix(0, 0, [])
The resulting A matrix has dimensions 2 x 2, while the number of total states is len(q) +
len(u) = 2 + 2 = 4. This is because for constrained systems the resulting A and B form has
a partitioned state vector only containing the independent coordinates and speeds. Written
out mathematically, the system linearized about this point would be written as:
[ ] [
][ ]
0 1 q2
q2
= g
u2
0 u2
L
Lagranges Method The derivation using Lagranges method is very similar to the approach using Kanes method described above. As before, we rst create the dynamicsymbols
needed to describe the system. In this case, the generalized coordinates q1 and q2 represent
the mass x and y coordinates in the inertial N frame. This results in the time derivatives q1 and
q2 representing the velocities in these directions. We also create some symbols to represent
the length and mass of the pendulum, as well as gravity and time.
>>>
>>>
>>>
>>>
>>>
from sympy.physics.mechanics import *
from sympy import symbols, atan, Matrix
q1, q2 = dynamicsymbols(q1:3)
q1d, q2d = dynamicsymbols(q1:3, level=1)
L, m, g, t = symbols(L, m, g, t)
Next, we create a world coordinate frame N , and its origin point N . The velocity of the origin
is set to 0. A second coordinate frame A is oriented such that its x-axis is along the pendulum
(as shown in the diagram above).
>>>
>>>
>>>
>>>
>>>
# Compose World Frame
N = ReferenceFrame(N)
pN = Point(N*)
pN.set_vel(N, 0)
# A.x is along the pendulum
1726
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> theta1 = atan(q2/q1)
>>> A = N.orientnew(A, axis, [theta1, N.z])
Locating the pendulum mass is then as easy as specifying its location with in terms of its x
and y coordinates in the world frame. A Particle object is then created to represent the
mass at this location.
>>>
>>>
>>>
>>>
# Create point P, the pendulum mass
P = pN.locatenew(P1, q1*N.x + q2*N.y)
P.set_vel(N, P.pos_from(pN).dt(N))
pP = Particle(pP, P, m)
As this system has more coordinates than degrees of freedom, constraints are needed. In this
case only a single holonomic constraints is needed: the distance from the origin to the mass
is always the length L (the pendulum doesnt get longer).
>>> # Holonomic Constraint Equations
>>> f_c = Matrix([q1**2 + q2**2 - L**2])
The force on the system is just gravity, at point P.
>>> # Input the force resultant at P
>>> R = m*g*N.x
With the problem setup, the Lagrangian can be calculated, and the equations of motion
formed. Note that the call to LagrangesMethod includes the Lagrangian, the generalized
coordinates, the constraints (specied by hol coneqs or nonhol coneqs), the list of (body,
force) pairs, and the inertial frame. In contrast to the KanesMethod initializer, independent
and dependent coordinates are not partitioned inside the LagrangesMethod object. Such a
partition is supplied later.
>>>
>>>
>>>
>>>
# Calculate the lagrangian, and form the equations of motion
Lag = Lagrangian(N, pP)
LM = LagrangesMethod(Lag, [q1, q2], hol_coneqs=f_c, forcelist=[(P, R)], frame=N)
lag_eqs = LM.form_lagranges_equations()
Next, we compose the operating point dictionary, set in the hanging at rest position:
>>> # Compose operating point
>>> op_point = {q1: L, q2: 0, q1d: 0, q2d: 0, q1d.diff(t): 0, q2d.diff(t): 0}
As there are constraints in the formulation, there will be corresponding Lagrange Multipliers.
These may appear inside the linearized form as well, and thus should also be included inside
the operating point dictionary. Fortunately, the LagrangesMethod class provides an easy way
of solving for the multipliers at a given operating point using the solve multipliers method.
>>> # Solve for multiplier operating point
>>> lam_op = LM.solve_multipliers(op_point=op_point)
With this solution, linearization can be completed. Note that in contrast to the KanesMethod
approach, the LagrangesMethod.linearize method also requires the partitioning of the generalized coordinates and their time derivatives into independent and dependent vectors. This
is the same as what was passed into the KanesMethod constructor above:
>>> op_point.update(lam_op)
>>> # Perform the Linearization
>>> A, B, inp_vec = LM.linearize([q2], [q2d], [q1], [q1d],
...
op_point=op_point, A_and_B=True)
5.36. Physics Module
1727
SymPy Documentation, Release 0.7.6
>>> A
Matrix([
[
0, 1],
[-g/L, 0]])
>>> B
Matrix(0, 0, [])
The resulting A matrix has dimensions 2 x 2, while the number of total states is 2*len(q)
= 4. This is because for constrained systems the resulting A and B form has a partitioned
state vector only containing the independent coordinates and their derivatives. Written out
mathematically, the system linearized about this point would be written as:
][ ]
[ ] [
0 1 q2
q2
= g
q2
0 q2
L
Potential Issues/Advanced Topics/Future Features in Physics/Mechanics This document will describe some of the more advanced functionality that this module oers but which
is not part of the ocial interface. Here, some of the features that will be implemented in
the future will also be covered, along with unanswered questions about proper functionality.
Also, common problems will be discussed, along with some solutions.
Common Issues Here issues with numerically integrating code, choice of dynamicsymbols
for coordinate and speed representation, printing, dierentiating, and substitution will occur.
Numerically Integrating Code See Future Features: Code Output
Dierentiating Dierentiation of very large expressions can take some time in SymPy; it
is possible for large expressions to take minutes for the derivative to be evaluated. This will
most commonly come up in linearization.
Choice of Coordinates and Speeds The Kane object is set up with the assumption that
the generalized speeds are not the same symbol as the time derivatives of the generalized
coordinates. This isnt to say that they cant be the same, just that they have to have a dierent
symbol. If you did this:
>> KM.coords([q1, q2, q3])
>> KM.speeds([q1d, q2d, q3d])
Your code would not work. Currently, kinematic dierential equations are required to be
provided. It is at this point that we hope the user will discover they should not attempt the
behavior shown in the code above.
This behavior might not be true for other methods of forming the equations of motion though.
Printing The default printing options are to use sorting for Vector and Dyad measure numbers, and have unsorted output from the mprint, mpprint, and mlatex functions. If you are
printing something large, please use one of those functions, as the sorting can increase printing time from seconds to minutes.
1728
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Dierentiating Dierentiation of very large expressions can take some time in SymPy; it
is possible for large expressions to take minutes for the derivative to be evaluated. This will
most commonly come up in linearization.
Substitution There are two common issues with substitution in mechanics:
When subbing in expressions for dynamicsymbols, sympys normal subs will substitute
in for derivatives of the dynamic symbol as well:
>>> from sympy.physics.mechanics import dynamicsymbols
>>> x = dynamicsymbols(x)
>>> expr = x.diff() + x
>>> sub_dict = {x: 1}
>>> expr.subs(sub_dict)
Derivative(1, t) + 1
In this case, x was replaced with 1 inside the Derivative as well, which is undesired.
Substitution into large expressions can be slow.
If your substitution is simple (direct replacement of expressions with other expressions, such
as when evaluating at an operating point) it is recommended to use the provided msubs function, as it is signicantly faster, and handles the derivative issue appropriately:
>>> from sympy.physics.mechanics import msubs
>>> msubs(expr, sub_dict)
Derivative(x(t), t) + 1
Linearization Currently, the linearization methods dont support cases where there are
non-coordinate, non-speed dynamic symbols outside of the dynamic equations. It also does
not support cases where time derivatives of these types of dynamic symbols show up. This
means if you have kinematic dierential equations which have a non-coordinate, non-speed
dynamic symbol, it will not work. It also means if you have dened a system parameter (say
a length or distance or mass) as a dynamic symbol, its time derivative is likely to show up in
the dynamic equations, and this will prevent linearization.
Acceleration of Points At a minimum, points need to have their velocities dened, as the
acceleration can be calculated by taking the time derivative of the velocity in the same frame.
If the 1 point or 2 point theorems were used to compute the velocity, the time derivative of the
velocity expression will most likely be more complex than if you were to use the acceleration
level 1 point and 2 point theorems. Using the acceleration level methods can result in shorted
expressions at this point, which will result in shorter expressions later (such as when forming
Kanes equations).
Advanced Interfaces
Advanced Functionality Remember that the Kane object supports bodies which have timevarying masses and inertias, although this functionality isnt completely compatible with the
linearization method.
Operators were discussed earlier as a potential way to do mathematical operations on Vector
and Dyad objects. The majority of the code in this module is actually coded with them, as it
can (subjectively) result in cleaner, shorter, more readable code. If using this interface in
5.36. Physics Module
1729
SymPy Documentation, Release 0.7.6
your code, remember to take care and use parentheses; the default order of operations in
Python results in addition occurring before some of the vector products, so use parentheses
liberally.
Future Features This will cover the planned features to be added to this submodule.
Code Output A function for generating code output for numerical integration is the highest
priority feature to implement next. There are a number of considerations here.
Code output for C (using the GSL libraries), Fortran 90 (using LSODA), MATLAB, and SciPy
is the goal. Things to be considered include: use of cse on large expressions for MATLAB and
SciPy, which are interpretive. It is currently unclear whether compiled languages will benet
from common subexpression elimination, especially considering that it is a common part of
compiler optimization, and there can be a signicant time penalty when calling cse.
Care needs to be taken when constructing the strings for these expressions, as well as handling of input parameters, and other dynamic symbols. How to deal with output quantities
when integrating also needs to be decided, with the potential for multiple options being considered.
References for Physics/Mechanics
Mechanics API
Masses, Inertias & Particles, RigidBodys (Docstrings)
Particle
class sympy.physics.mechanics.particle.Particle(name, point, mass)
A particle.
Particles have a non-zero mass and lack spatial extension; they take up no space.
Values need to be supplied on initialization, but can be changed later.
Parameters name : str
Name of particle
point : Point
A physics/mechanics Point which represents the position, velocity,
and acceleration of this Particle
mass : sympifyable
A SymPy expression representing the Particles mass
Examples
>>>
>>>
>>>
>>>
>>>
1730
from sympy.physics.mechanics import Particle, Point
from sympy import Symbol
po = Point(po)
m = Symbol(m)
pa = Particle(pa, po, m)
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> # Or you could change these later
>>> pa.mass = m
>>> pa.point = po
angular momentum(point, frame)
Angular momentum of the particle about the point.
The angular momentum H, about some point O of a particle, P, is given by:
H=rxm*v
where r is the position vector from point O to the particle P, m is the mass of the
particle, and v is the velocity of the particle in the inertial frame, N.
Parameters point : Point
The point about which angular momentum of the particle is desired.
frame : ReferenceFrame
The frame in which angular momentum is desired.
Examples
>>> from sympy.physics.mechanics import Particle, Point, ReferenceFrame
>>> from sympy.physics.mechanics import dynamicsymbols
>>> m, v, r = dynamicsymbols(m v r)
>>> N = ReferenceFrame(N)
>>> O = Point(O)
>>> A = O.locatenew(A, r * N.x)
>>> P = Particle(P, A, m)
>>> P.point.set_vel(N, v * N.y)
>>> P.angular_momentum(O, N)
m*r*v*N.z
get mass()
Mass of the particle.
get point()
Point of the particle.
kinetic energy(frame)
Kinetic energy of the particle
The kinetic energy, T, of a particle, P, is given by
T = 1/2 m v2
where m is the mass of particle P, and v is the velocity of the particle in the supplied
ReferenceFrame.
Parameters frame : ReferenceFrame
The Particles velocity is typically dened with respect to an inertial
frame but any relevant frame in which the velocity is known can be
supplied.
5.36. Physics Module
1731
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.physics.mechanics import Particle, Point, ReferenceFrame
>>> from sympy import symbols
>>> m, v, r = symbols(m v r)
>>> N = ReferenceFrame(N)
>>> O = Point(O)
>>> P = Particle(P, O, m)
>>> P.point.set_vel(N, v * N.y)
>>> P.kinetic_energy(N)
m*v**2/2
linear momentum(frame)
Linear momentum of the particle.
The linear momentum L, of a particle P, with respect to frame N is given by
L=m*v
where m is the mass of the particle, and v is the velocity of the particle in the frame
N.
Parameters frame : ReferenceFrame
The frame in which linear momentum is desired.
Examples
>>> from sympy.physics.mechanics import Particle, Point, ReferenceFrame
>>> from sympy.physics.mechanics import dynamicsymbols
>>> m, v = dynamicsymbols(m v)
>>> N = ReferenceFrame(N)
>>> P = Point(P)
>>> A = Particle(A, P, m)
>>> P.set_vel(N, v * N.x)
>>> A.linear_momentum(N)
m*v*N.x
mass
Mass of the particle.
point
Point of the particle.
potential energy
The potential energy of the Particle.
Examples
>>> from sympy.physics.mechanics import Particle, Point
>>> from sympy import symbols
>>> m, g, h = symbols(m g h)
>>> O = Point(O)
>>> P = Particle(P, O, m)
>>> P.set_potential_energy(m * g * h)
>>> P.potential_energy
g*h*m
1732
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
set potential energy(scalar)
Used to set the potential energy of the Particle.
Parameters scalar : Sympifyable
The potential energy (a scalar) of the Particle.
Examples
>>>
>>>
>>>
>>>
>>>
>>>
from sympy.physics.mechanics import Particle, Point
from sympy import symbols
m, g, h = symbols(m g h)
O = Point(O)
P = Particle(P, O, m)
P.set_potential_energy(m * g * h)
RigidBody
class sympy.physics.mechanics.rigidbody.RigidBody(name, masscenter,
mass, inertia)
An idealized rigid body.
frame,
This is essentially a container which holds the various components which describe a rigid
body: a name, mass, center of mass, reference frame, and inertia.
All of these need to be supplied on creation, but can be changed afterwards.
Examples
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
from sympy import Symbol
from sympy.physics.mechanics import ReferenceFrame, Point, RigidBody
from sympy.physics.mechanics import outer
m = Symbol(m)
A = ReferenceFrame(A)
P = Point(P)
I = outer (A.x, A.x)
inertia_tuple = (I, P)
B = RigidBody(B, P, A, m, inertia_tuple)
# Or you could change them afterwards
m2 = Symbol(m2)
B.mass = m2
Attributes
name
masscenter
frame
mass
inertia
string
Point
ReferenceFrame
Sympifyable
(Dyadic,
Point)
5.36. Physics Module
The bodys name.
The point which represents the center of mass of the
rigid body.
The ReferenceFrame which the rigid body is xed in.
The bodys mass.
The bodys inertia about a point; stored in a tuple as
shown above.
1733
SymPy Documentation, Release 0.7.6
angular momentum(point, frame)
Angular momentum of the rigid body.
The angular momentum H, about some point O, of a rigid body B, in a frame N is
given by
H = I* . omega + r* x (M * v)
where I* is the central inertia dyadic of B, omega is the angular velocity of body B
in the frame, N, r* is the position vector from point O to the mass center of B, and v
is the velocity of point O in the frame, N.
Parameters point : Point
The point about which angular momentum is desired.
frame : ReferenceFrame
The frame in which angular momentum is desired.
Examples
>>> from sympy.physics.mechanics import Point, ReferenceFrame, outer
>>> from sympy.physics.mechanics import RigidBody, dynamicsymbols
>>> M, v, r, omega = dynamicsymbols(M v r omega)
>>> N = ReferenceFrame(N)
>>> b = ReferenceFrame(b)
>>> b.set_ang_vel(N, omega * b.x)
>>> P = Point(P)
>>> P.set_vel(N, 1 * N.x)
>>> I = outer (b.x, b.x)
>>> Inertia_tuple = (I, P)
>>> B = RigidBody(B, P, b, M, Inertia_tuple)
>>> B.angular_momentum(P, N)
omega*b.x
central inertia
The bodys central inertia dyadic.
kinetic energy(frame)
Kinetic energy of the rigid body
The kinetic energy, T, of a rigid body, B, is given by
T = 1/2 (I omega2 + m v2)
where I and m are the central inertia dyadic and mass of rigid body B, respectively,
omega is the bodys angular velocity and v is the velocity of the bodys mass center
in the supplied ReferenceFrame.
Parameters frame : ReferenceFrame
The RigidBodys angular velocity and the velocity of its mass center
are typically dened with respect to an inertial frame but any relevant
frame in which the velocities are known can be supplied.
Examples
1734
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> from sympy.physics.mechanics import Point, ReferenceFrame, outer
>>> from sympy.physics.mechanics import RigidBody
>>> from sympy import symbols
>>> M, v, r, omega = symbols(M v r omega)
>>> N = ReferenceFrame(N)
>>> b = ReferenceFrame(b)
>>> b.set_ang_vel(N, omega * b.x)
>>> P = Point(P)
>>> P.set_vel(N, v * N.x)
>>> I = outer (b.x, b.x)
>>> inertia_tuple = (I, P)
>>> B = RigidBody(B, P, b, M, inertia_tuple)
>>> B.kinetic_energy(N)
M*v**2/2 + omega**2/2
linear momentum(frame)
Linear momentum of the rigid body.
The linear momentum L, of a rigid body B, with respect to frame N is given by
L = M * v*
where M is the mass of the rigid body and v* is the velocity of the mass center of B
in the frame, N.
Parameters frame : ReferenceFrame
The frame in which linear momentum is desired.
Examples
>>> from sympy.physics.mechanics import Point, ReferenceFrame, outer
>>> from sympy.physics.mechanics import RigidBody, dynamicsymbols
>>> M, v = dynamicsymbols(M v)
>>> N = ReferenceFrame(N)
>>> P = Point(P)
>>> P.set_vel(N, v * N.x)
>>> I = outer (N.x, N.x)
>>> Inertia_tuple = (I, P)
>>> B = RigidBody(B, P, N, M, Inertia_tuple)
>>> B.linear_momentum(N)
M*v*N.x
potential energy
The potential energy of the RigidBody.
Examples
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
from sympy.physics.mechanics import RigidBody, Point, outer, ReferenceFrame
from sympy import symbols
M, g, h = symbols(M g h)
b = ReferenceFrame(b)
P = Point(P)
I = outer (b.x, b.x)
Inertia_tuple = (I, P)
B = RigidBody(B, P, b, M, Inertia_tuple)
B.set_potential_energy(M * g * h)
5.36. Physics Module
1735
SymPy Documentation, Release 0.7.6
>>> B.potential_energy
M*g*h
set potential energy(scalar)
Used to set the potential energy of this RigidBody.
Parameters scalar: Sympifyable :
The potential energy (a scalar) of the RigidBody.
Examples
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
from sympy.physics.mechanics import Particle, Point, outer
from sympy.physics.mechanics import RigidBody, ReferenceFrame
from sympy import symbols
b = ReferenceFrame(b)
M, g, h = symbols(M g h)
P = Point(P)
I = outer (b.x, b.x)
Inertia_tuple = (I, P)
B = RigidBody(B, P, b, M, Inertia_tuple)
B.set_potential_energy(M * g * h)
inertia
sympy.physics.mechanics.functions.inertia(frame, ixx, iyy, izz, ixy=0, iyz=0,
izx=0)
Simple way to create inertia Dyadic object.
If you dont know what a Dyadic is, just treat this like the inertia tensor. Then, do the
easy thing and dene it in a body-xed frame.
Parameters frame : ReferenceFrame
The frame the inertia is dened in
ixx : Sympifyable
the xx element in the inertia dyadic
iyy : Sympifyable
the yy element in the inertia dyadic
izz : Sympifyable
the zz element in the inertia dyadic
ixy : Sympifyable
the xy element in the inertia dyadic
iyz : Sympifyable
the yz element in the inertia dyadic
izx : Sympifyable
the zx element in the inertia dyadic
1736
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.physics.mechanics import ReferenceFrame, inertia
>>> N = ReferenceFrame(N)
>>> inertia(N, 1, 2, 3)
(N.x|N.x) + 2*(N.y|N.y) + 3*(N.z|N.z)
inertia of point mass
sympy.physics.mechanics.functions.inertia of point mass(mass, pos vec, frame)
Inertia dyadic of a point mass relative to point O.
Parameters mass : Sympifyable
Mass of the point mass
pos vec : Vector
Position from point O to point mass
frame : ReferenceFrame
Reference frame to express the dyadic in
Examples
>>> from sympy import symbols
>>> from sympy.physics.mechanics import ReferenceFrame, inertia_of_point_mass
>>> N = ReferenceFrame(N)
>>> r, m = symbols(r m)
>>> px = r * N.x
>>> inertia_of_point_mass(m, px, N)
m*r**2*(N.y|N.y) + m*r**2*(N.z|N.z)
linear momentum
sympy.physics.mechanics.functions.linear momentum(frame, *body)
Linear momentum of the system.
This function returns the linear momentum of a system of Particles and/or RigidBodys.
The linear momentum of a system is equal to the vector sum of the linear momentum of
its constituents. Consider a system, S, comprised of a rigid body, A, and a particle, P. The
linear momentum of the system, L, is equal to the vector sum of the linear momentum
of the particle, L1, and the linear momentum of the rigid body, L2, i.e.
L = L1 + L2
Parameters frame : ReferenceFrame
The frame in which linear momentum is desired.
body1, body2, body3... : Particle and/or RigidBody
The body (or bodies) whose linear momentum is required.
5.36. Physics Module
1737
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.physics.mechanics import Point, Particle, ReferenceFrame
>>> from sympy.physics.mechanics import RigidBody, outer, linear_momentum
>>> N = ReferenceFrame(N)
>>> P = Point(P)
>>> P.set_vel(N, 10 * N.x)
>>> Pa = Particle(Pa, P, 1)
>>> Ac = Point(Ac)
>>> Ac.set_vel(N, 25 * N.y)
>>> I = outer(N.x, N.x)
>>> A = RigidBody(A, Ac, N, 20, (I, Ac))
>>> linear_momentum(N, A, Pa)
10*N.x + 500*N.y
angular momentum
sympy.physics.mechanics.functions.angular momentum(point, frame, *body)
Angular momentum of a system
This function returns the angular momentum of a system of Particles and/or RigidBodys.
The angular momentum of such a system is equal to the vector sum of the angular momentum of its constituents. Consider a system, S, comprised of a rigid body, A, and a
particle, P. The angular momentum of the system, H, is equal to the vector sum of the
angular momentum of the particle, H1, and the angular momentum of the rigid body,
H2, i.e.
H = H1 + H2
Parameters point : Point
The point about which angular momentum of the system is desired.
frame : ReferenceFrame
The frame in which angular momentum is desired.
body1, body2, body3... : Particle and/or RigidBody
The body (or bodies) whose angular momentum is required.
Examples
>>> from sympy.physics.mechanics import Point, Particle, ReferenceFrame
>>> from sympy.physics.mechanics import RigidBody, outer, angular_momentum
>>> N = ReferenceFrame(N)
>>> O = Point(O)
>>> O.set_vel(N, 0 * N.x)
>>> P = O.locatenew(P, 1 * N.x)
>>> P.set_vel(N, 10 * N.x)
>>> Pa = Particle(Pa, P, 1)
>>> Ac = O.locatenew(Ac, 2 * N.y)
>>> Ac.set_vel(N, 5 * N.y)
>>> a = ReferenceFrame(a)
>>> a.set_ang_vel(N, 10 * N.z)
>>> I = outer(N.z, N.z)
>>> A = RigidBody(A, Ac, a, 20, (I, Ac))
>>> angular_momentum(O, N, Pa, A)
10*N.z
1738
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
kinetic energy
sympy.physics.mechanics.functions.kinetic energy(frame, *body)
Kinetic energy of a multibody system.
This function returns the kinetic energy of a system of Particles and/or RigidBodys.
The kinetic energy of such a system is equal to the sum of the kinetic energies of its
constituents. Consider a system, S, comprising a rigid body, A, and a particle, P. The
kinetic energy of the system, T, is equal to the vector sum of the kinetic energy of the
particle, T1, and the kinetic energy of the rigid body, T2, i.e.
T = T1 + T2
Kinetic energy is a scalar.
Parameters frame : ReferenceFrame
The frame in which the velocity or angular velocity of the body is
dened.
body1, body2, body3... : Particle and/or RigidBody
The body (or bodies) whose kinetic energy is required.
Examples
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
350
from sympy.physics.mechanics import Point, Particle, ReferenceFrame
from sympy.physics.mechanics import RigidBody, outer, kinetic_energy
N = ReferenceFrame(N)
O = Point(O)
O.set_vel(N, 0 * N.x)
P = O.locatenew(P, 1 * N.x)
P.set_vel(N, 10 * N.x)
Pa = Particle(Pa, P, 1)
Ac = O.locatenew(Ac, 2 * N.y)
Ac.set_vel(N, 5 * N.y)
a = ReferenceFrame(a)
a.set_ang_vel(N, 10 * N.z)
I = outer(N.z, N.z)
A = RigidBody(A, Ac, a, 20, (I, Ac))
kinetic_energy(N, Pa, A)
potential energy
sympy.physics.mechanics.functions.potential energy(*body)
Potential energy of a multibody system.
This function returns the potential energy of a system of Particles and/or RigidBodys.
The potential energy of such a system is equal to the sum of the potential energy of its
constituents. Consider a system, S, comprising a rigid body, A, and a particle, P. The
potential energy of the system, V, is equal to the vector sum of the potential energy of
the particle, V1, and the potential energy of the rigid body, V2, i.e.
V = V1 + V2
Potential energy is a scalar.
Parameters body1, body2, body3... : Particle and/or RigidBody
The body (or bodies) whose potential energy is required.
5.36. Physics Module
1739
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.physics.mechanics import Point, Particle, ReferenceFrame
>>> from sympy.physics.mechanics import RigidBody, outer, potential_energy
>>> from sympy import symbols
>>> M, m, g, h = symbols(M m g h)
>>> N = ReferenceFrame(N)
>>> O = Point(O)
>>> O.set_vel(N, 0 * N.x)
>>> P = O.locatenew(P, 1 * N.x)
>>> Pa = Particle(Pa, P, m)
>>> Ac = O.locatenew(Ac, 2 * N.y)
>>> a = ReferenceFrame(a)
>>> I = outer(N.z, N.z)
>>> A = RigidBody(A, Ac, a, M, (I, Ac))
>>> Pa.set_potential_energy(m * g * h)
>>> A.set_potential_energy(M * g * h)
>>> potential_energy(Pa, A)
M*g*h + g*h*m
Lagrangian
sympy.physics.mechanics.functions.Lagrangian(frame, *body)
Lagrangian of a multibody system.
This function returns the Lagrangian of a system of Particles and/or RigidBodys. The
Lagrangian of such a system is equal to the dierence between the kinetic energies and
potential energies of its constituents. If T and V are the kinetic and potential energies
of a system then its Lagrangian, L, is dened as
L=T-V
The Lagrangian is a scalar.
Parameters frame : ReferenceFrame
The frame in which the velocity or angular velocity of the body is
dened to determine the kinetic energy.
body1, body2, body3... : Particle and/or RigidBody
The body (or bodies) whose Lagrangian is required.
Examples
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
1740
from sympy.physics.mechanics import Point, Particle, ReferenceFrame
from sympy.physics.mechanics import RigidBody, outer, Lagrangian
from sympy import symbols
M, m, g, h = symbols(M m g h)
N = ReferenceFrame(N)
O = Point(O)
O.set_vel(N, 0 * N.x)
P = O.locatenew(P, 1 * N.x)
P.set_vel(N, 10 * N.x)
Pa = Particle(Pa, P, 1)
Ac = O.locatenew(Ac, 2 * N.y)
Ac.set_vel(N, 5 * N.y)
a = ReferenceFrame(a)
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> a.set_ang_vel(N, 10 * N.z)
>>> I = outer(N.z, N.z)
>>> A = RigidBody(A, Ac, a, 20, (I, Ac))
>>> Pa.set_potential_energy(m * g * h)
>>> A.set_potential_energy(M * g * h)
>>> Lagrangian(N, Pa, A)
-M*g*h - g*h*m + 350
Kanes Method & Lagranges Method (Docstrings)
KaneMethod
class sympy.physics.mechanics.kane.KanesMethod(frame,
q ind,
u ind,
kd eqs=None,
q dependent=None,
conguration constraints=None,
u dependent=None,
velocity constraints=None,
acceleration constraints=None,
u auxiliary=None)
Kanes method object.
This object is used to do the book-keeping as you go through and form equations of
motion in the way Kane presents in: Kane, T., Levinson, D. Dynamics Theory and Applications. 1985 McGraw-Hill
The attributes are for equations in the form [M] udot = forcing.
Examples
This is a simple example for a one degree of freedom translational spring-mass-damper.
In this example, we rst need to do the kinematics. This involves creating generalized
speeds and coordinates and their derivatives. Then we create a point and set its velocity
in a frame.
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
from sympy import symbols
from sympy.physics.mechanics import dynamicsymbols, ReferenceFrame
from sympy.physics.mechanics import Point, Particle, KanesMethod
q, u = dynamicsymbols(q u)
qd, ud = dynamicsymbols(q u, 1)
m, c, k = symbols(m c k)
N = ReferenceFrame(N)
P = Point(P)
P.set_vel(N, u * N.x)
Next we need to arrange/store information in the way that KanesMethod requires. The
kinematic dierential equations need to be stored in a dict. A list of forces/torques must
be constructed, where each entry in the list is a (Point, Vector) or (ReferenceFrame,
Vector) tuple, where the Vectors represent the Force or Torque. Next a particle needs
to be created, and it needs to have a point and mass assigned to it. Finally, a list of all
bodies and particles needs to be created.
>>> kd = [qd - u]
>>> FL = [(P, (-k * q - c * u) * N.x)]
5.36. Physics Module
1741
SymPy Documentation, Release 0.7.6
>>> pa = Particle(pa, P, m)
>>> BL = [pa]
Finally we can generate the equations of motion. First we create the KanesMethod object
and supply an inertial frame, coordinates, generalized speeds, and the kinematic dierential equations. Additional quantities such as conguration and motion constraints,
dependent coordinates and speeds, and auxiliary speeds are also supplied here (see the
online documentation). Next we form FR* and FR to complete: Fr + Fr* = 0. We have
the equations of motion at this point. It makes sense to rearrnge them though, so we
calculate the mass matrix and the forcing terms, for E.o.M. in the form: [MM] udot =
forcing, where MM is the mass matrix, udot is a vector of the time derivatives of the
generalized speeds, and forcing is a vector representing forcing terms.
>>> KM = KanesMethod(N, q_ind=[q], u_ind=[u], kd_eqs=kd)
>>> (fr, frstar) = KM.kanes_equations(FL, BL)
>>> MM = KM.mass_matrix
>>> forcing = KM.forcing
>>> rhs = MM.inv() * forcing
>>> rhs
Matrix([[(-c*u(t) - k*q(t))/m]])
>>> KM.linearize(A_and_B=True, new_method=True)[0]
Matrix([
[
0,
1],
[-k/m, -c/m]])
Please look at the documentation pages for more information on how to perform linearization and how to deal with dependent coordinates & speeds, and how do deal with
bringing non-contributing forces into evidence.
Attributes
q, u
Matrix
bodylist
iterable
forcelist
iterable
auxiliary
Matrix
mass matrix Matrix
forcing
Matrix
mass matrix full
Matrix
forcMaing full
trix
Matrices of the generalized coordinates and speeds
Iterable of Point and RigidBody objects in the system.
Iterable of (Point, vector) or (ReferenceFrame, vector) tuples
describing the forces on the system.
If applicable, the set of auxiliary Kanes equations used to solve
for non-contributing forces.
The systems mass matrix
The systems forcing vector
The mass matrix for the us and qs
The forcing vector for the us and qs
auxiliary eqs
A matrix containing the auxiliary equations.
forcing
The forcing vector of the system.
forcing full
The forcing vector of the system, augmented by the kinematic dierential equations.
1742
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
kanes equations(FL, BL)
Method to form Kanes equations, Fr + Fr* = 0.
Returns (Fr, Fr*). In the case where auxiliary generalized speeds are present (say, s
auxiliary speeds, o generalized speeds, and m motion constraints) the length of the
returned vectors will be o - m + s in length. The rst o - m equations will be the
constrained Kanes equations, then the s auxiliary Kanes equations. These auxiliary
equations can be accessed with the auxiliary eqs().
Parameters FL : list
Takes in a list of (Point, Vector) or (ReferenceFrame, Vector) tuples
which represent the force at a point or torque on a frame.
BL : list
A list of all RigidBodys and Particles in the system.
kindiffdict()
Returns a dictionary mapping q to u.
linearize(**kwargs)
Linearize the equations of motion about a symbolic operating point.
If kwarg A and B is False (default), returns M, A, B, r for the linearized form, M*[q,
u]T = A*[q ind, u ind]T + B*r.
If kwarg A and B is True, returns A, B, r for the linearized form dx = A*x + B*r, where
x = [q ind, u ind]T. Note that this is computationally intensive if there are many
symbolic parameters. For this reason, it may be more desirable to use the default
A and B=False, returning M, A, and B. Values may then be substituted in to these
matrices, and the state space form found as A = P.T*M.inv()*A, B = P.T*M.inv()*B,
where P = Linearizer.perm mat.
In both cases, r is found as all dynamicsymbols in the equations of motion that are
not part of q, u, q, or u. They are sorted in canonical form.
The operating points may be also entered using the op point kwarg. This takes a
dictionary of {symbol: value}, or a an iterable of such dictionaries. The values may
be numberic or symbolic. The more values you can specify beforehand, the faster
this computation will run.
As part of the deprecation cycle, the new method will not be used unless the kwarg
new method is set to True. If the kwarg is missing, or set to false, the old linearization
method will be used. After next release the need for this kwarg will be removed.
For more documentation, please see the Linearizer class.
mass matrix
The mass matrix of the system.
mass matrix full
The mass matrix of the system, augmented by the kinematic dierential equations.
rhs(inv method=None)
Returns the systems equations of motion in rst order form.
The output of this will be the right hand side of:
[qdot, udot].T = f(q, u, t)
Or, the equations of motion in rst order form. The right hand side is what is needed
by most numerical ODE integrators.
Parameters inv method : str
5.36. Physics Module
1743
SymPy Documentation, Release 0.7.6
The specic sympy inverse matrix calculation method to use. For a
list of valid methods, see inv()
to linearizer()
Returns an instance of the Linearizer class, initiated from the data in the KanesMethod class. This may be more desirable than using the linearize class method,
as the Linearizer object will allow more ecient recalculation (i.e. about varying
operating points).
LagrangesMethod
class sympy.physics.mechanics.lagrange.LagrangesMethod(Lagrangian,
qs,
coneqs=None,
forcelist=None,
frame=None,
hol coneqs=None,
nonhol coneqs=None)
Lagranges method object.
This object generates the equations of motion in a two step procedure. The rst step involves the initialization of LagrangesMethod by supplying the Lagrangian and the generalized coordinates, at the bare minimum. If there are any constraint equations, they can
be supplied as keyword arguments. The Lagrange multipliers are automatically generated and are equal in number to the constraint equations. Similarly any non-conservative
forces can be supplied in an iterable (as described below and also shown in the example)
along with a ReferenceFrame. This is also discussed further in the init method.
Examples
This is a simple example for a one degree of freedom translational spring-mass-damper.
In this example, we rst need to do the kinematics. This involves creating generalized
coordinates and their derivatives. Then we create a point and set its velocity in a frame.
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
from sympy.physics.mechanics import LagrangesMethod, Lagrangian
from sympy.physics.mechanics import ReferenceFrame, Particle, Point
from sympy.physics.mechanics import dynamicsymbols, kinetic_energy
from sympy import symbols
q = dynamicsymbols(q)
qd = dynamicsymbols(q, 1)
m, k, b = symbols(m k b)
N = ReferenceFrame(N)
P = Point(P)
P.set_vel(N, qd * N.x)
We need to then prepare the information as required by LagrangesMethod to generate equations of motion. First we create the Particle, which has a point attached to it.
Following this the lagrangian is created from the kinetic and potential energies. Then,
an iterable of nonconservative forces/torques must be constructed, where each item is
a (Point, Vector) or (ReferenceFrame, Vector) tuple, with the Vectors representing the
nonconservative forces or torques.
>>>
>>>
>>>
>>>
1744
Pa = Particle(Pa, P, m)
Pa.set_potential_energy(k * q**2 / 2.0)
L = Lagrangian(N, Pa)
fl = [(P, -b * qd * N.x)]
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Finally we can generate the equations of motion. First we create the LagrangesMethod
object. To do this one must supply the Lagrangian, and the generalized coordinates.
The constraint equations, the forcelist, and the inertial frame may also be provided,
if relevant. Next we generate Lagranges equations of motion, such that: Lagranges
equations of motion = 0. We have the equations of motion at this point.
>>> l = LagrangesMethod(L, [q], forcelist = fl, frame = N)
>>> print(l.form_lagranges_equations())
Matrix([[b*Derivative(q(t), t) + 1.0*k*q(t) + m*Derivative(q(t), t, t)]])
We can also solve for the states using the rhs method.
>>> print(l.rhs())
Matrix([[Derivative(q(t), t)], [(-b*Derivative(q(t), t) - 1.0*k*q(t))/m]])
Please refer to the docstrings on each method for more details.
Attributes
q, u
Matrix
forcelist
iterable
mass matrix Matrix
forcing
Matrix
mass matrix full
Matrix
forcMaing full
trix
Matrices of the generalized coordinates and speeds
Iterable of (Point, vector) or (ReferenceFrame, vector) tuples
describing the forces on the system.
The systems mass matrix
The systems forcing vector
The mass matrix for the qdots, qdoubledots, and the
lagrange multipliers (lam)
The forcing vector for the qdots, qdoubledots and lagrange
multipliers (lam)
forcing
Returns the forcing vector from lagranges equations method.
forcing full
Augments qdots to the forcing vector above.
form lagranges equations()
Method to form Lagranges equations of motion.
Returns a vector of equations of motion using Lagranges equations of the second
kind.
linearize(q ind=None, qd ind=None, q dep=None, qd dep=None, **kwargs)
Linearize the equations of motion about a symbolic operating point.
If kwarg A and B is False (default), returns M, A, B, r for the linearized form, M*[q,
u]T = A*[q ind, u ind]T + B*r.
If kwarg A and B is True, returns A, B, r for the linearized form dx = A*x + B*r, where
x = [q ind, u ind]T. Note that this is computationally intensive if there are many
symbolic parameters. For this reason, it may be more desirable to use the default
A and B=False, returning M, A, and B. Values may then be substituted in to these
matrices, and the state space form found as A = P.T*M.inv()*A, B = P.T*M.inv()*B,
where P = Linearizer.perm mat.
In both cases, r is found as all dynamicsymbols in the equations of motion that are
not part of q, u, q, or u. They are sorted in canonical form.
5.36. Physics Module
1745
SymPy Documentation, Release 0.7.6
The operating points may be also entered using the op point kwarg. This takes a
dictionary of {symbol: value}, or a an iterable of such dictionaries. The values may
be numberic or symbolic. The more values you can specify beforehand, the faster
this computation will run.
For more documentation, please see the Linearizer class.
mass matrix
Returns the mass matrix, which is augmented by the Lagrange multipliers, if necessary.
If the system is described by n generalized coordinates and there are no constraint
equations then an n X n matrix is returned.
If there are n generalized coordinates and m constraint equations have been supplied during initialization then an n X (n+m) matrix is returned. The (n + m - 1)th
and (n + m)th columns contain the coecients of the Lagrange multipliers.
mass matrix full
Augments the coecients of qdots to the mass matrix.
rhs(inv method=None, **kwargs)
Returns equations that can be solved numerically
Parameters inv method : str
The specic sympy inverse matrix calculation method to use. For a
list of valid methods, see inv()
solve multipliers(op point=None, sol type=dict)
Solves for the values of the lagrange multipliers symbolically at the specied operating point
Parameters op point : dict or iterable of dicts, optional
Point at which to solve at. The operating point is specied as a dictionary or iterable of dictionaries of {symbol: value}. The value may
be numeric or symbolic itself.
sol type : str, optional
Solution return type. Valid options are: - dict: A dict of {symbol :
value} (default) - Matrix: An ordered column matrix of the solution
to linearizer(q ind=None, qd ind=None, q dep=None, qd dep=None)
Returns an instance of the Linearizer class, initiated from the data in the LagrangesMethod class. This may be more desirable than using the linearize class method,
as the Linearizer object will allow more ecient recalculation (i.e. about varying
operating points).
Parameters q ind, qd ind : array like, optional
The independent generalized coordinates and speeds.
q dep, qd dep : array like, optional
The dependent generalized coordinates and speeds.
Linearization (Docstrings)
Linearizer
1746
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
class sympy.physics.mechanics.linearize.Linearizer(f 0, f 1, f 2, f 3, f 4, f c,
f v, f a, q, u, q i=None,
q d=None,
u i=None,
u d=None,
r=None,
lams=None)
This object holds the general model form for a dynamic system. This model is used for
computing the linearized form of the system, while properly dealing with constraints
leading to dependent coordinates and speeds.
Attributes
f 0, f 1, f 2, f 3, f 4,
f c, f v, f a
q, u, r
q i, u i
q d, u d
perm mat
Matrix
Matrix
Matrix
Matrix
Matrix
Matrices holding the general system form.
Matrices holding the generalized coordinates,
speeds, and input vectors.
Matrices of the independent generalized coordinates
and speeds.
Matrices of the dependent generalized coordinates
and speeds.
Permutation matrix such that [q ind, u ind]T =
perm mat*[q, u]T
linearize(op point=None, A and B=False, simplify=False)
Linearize the system about the operating point. Note that q op, u op, qd op, ud op
must satisfy the equations of motion. These may be either symbolic or numeric.
Parameters op point : dict or iterable of dicts, optional
Dictionary or iterable of dictionaries containing the operating point
conditions. These will be substituted in to the linearized system before the linearization is complete. Leave blank if you want a completely symbolic form. Note that any reduction in symbols (whether
substituted for numbers or expressions with a common parameter)
will result in faster runtime.
A and B : bool, optional
If A and B=False (default), (M, A, B) is returned for forming [M]*[q,
u]T = [A]*[q ind, u ind]T + [B]r. If A and B=True, (A, B) is returned
for forming dx = [A]x + [B]r, where x = [q ind, u ind]T.
simplify : bool, optional
Determines if returned values are simplied before return. For large
expressions this may be time consuming. Default is False.
Potential Issues
Note that the process of solving with A and B=True is computationally intensive if
there are many symbolic parameters. For this reason, it may be more desirable to
use the default A and B=False, returning M, A, and B. More values may then be
substituted in to these matrices later on. The state space form can then be found as
A = P.T*M.LUsolve(A), B = P.T*M.LUsolve(B), where P = Linearizer.perm mat.
Expression Manipulation (Docstrings)
5.36. Physics Module
1747
SymPy Documentation, Release 0.7.6
msubs
sympy.physics.mechanics.msubs(expr, *sub dicts, **kwargs)
A custom subs for use on expressions derived in physics.mechanics.
Traverses the expression tree once, performing the subs found in sub dicts. Terms inside
Derivative expressions are ignored:
>>> from sympy.physics.mechanics import dynamicsymbols, msubs
>>> x = dynamicsymbols(x)
>>> msubs(x.diff() + x, {x: 1})
Derivative(x(t), t) + 1
Note that sub dicts can be a single dictionary, or several dictionaries:
>>>
>>>
>>>
>>>
10
x, y, z = dynamicsymbols(x, y, z)
sub1 = {x: 1, y: 2}
sub2 = {z: 3, x.diff(): 4}
msubs(x.diff() + x + y + z, sub1, sub2)
If smart=True (default False), also checks for conditions that may result in nan, but if
simplied would yield a valid expression. For example:
>>> from sympy import sin, tan
>>> (sin(x)/tan(x)).subs(x, 0)
nan
>>> msubs(sin(x)/tan(x), {x: 0}, smart=True)
1
It does this by rst replacing all tan with sin/cos. Then each node is traversed. If
the node is a fraction, subs is rst evaluated on the denominator. If this results in 0,
simplication of the entire fraction is attempted. Using this selective simplication, only
subexpressions that result in 1/0 are targeted, resulting in faster performance.
nd dynamicsymbols
sympy.physics.mechanics.find dynamicsymbols(expression, exclude=None)
Find all dynamicsymbols in expression.
>>> from sympy.physics.mechanics import dynamicsymbols, find_dynamicsymbols
>>> x, y = dynamicsymbols(x, y)
>>> expr = x + x.diff()*y
>>> find_dynamicsymbols(expr)
set([x(t), y(t), Derivative(x(t), t)])
If the optional exclude kwarg is used, only dynamicsymbols not in the iterable exclude
are returned.
>>> find_dynamicsymbols(expr, [x, y])
set([Derivative(x(t), t)])
Printing (Docstrings)
mechanics printing This
time derivative printing.
1748
function
is
the
same
as
physics.vectors
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
mprint This function is the same as physics.vectors vprint.
This function is the same as physics.vectors vpprint.
mpprint
mlatex This function is the same as physics.vectors vlatex.
Quantum Mechanics
Abstract
Contains Docstrings of Physics-Quantum module
Quantum Functions
Anticommutator The anti-commutator: {A,B} = A*B + B*A.
class sympy.physics.quantum.anticommutator.AntiCommutator
The standard anticommutator, in an unevaluated state.
Evaluating an anticommutator is dened [R327] (page 1916) as: {A, B} = A*B + B*A.
This class returns the anticommutator in an unevaluated form. To evaluate the anticommutator, use the .doit() method.
Cannonical ordering of an anticommutator is {A, B} for A < B. The arguments of the
anticommutator are put into canonical order using cmp . If B < A, then {A, B} is
returned as {B, A}.
Parameters A : Expr
The rst argument of the anticommutator {A,B}.
B : Expr
The second argument of the anticommutator {A,B}.
References
[R327] (page 1916)
Examples
>>>
>>>
>>>
>>>
>>>
>>>
from sympy import symbols
from sympy.physics.quantum import AntiCommutator
from sympy.physics.quantum import Operator, Dagger
x, y = symbols(x,y)
A = Operator(A)
B = Operator(B)
Create an anticommutator and use doit() to multiply them out.
5.36. Physics Module
1749
SymPy Documentation, Release 0.7.6
>>> ac = AntiCommutator(A,B); ac
{A,B}
>>> ac.doit()
A*B + B*A
The commutator orders it arguments in canonical order:
>>> ac = AntiCommutator(B,A); ac
{A,B}
Commutative constants are factored out:
>>> AntiCommutator(3*x*A,x*y*B)
3*x**2*y*{A,B}
Adjoint operations applied to the anticommutator are properly applied to the arguments:
>>> Dagger(AntiCommutator(A,B))
{Dagger(A),Dagger(B)}
doit(**hints)
Evaluate anticommutator
Clebsch-Gordan Coecients Clebsch-Gordon Coecients.
class sympy.physics.quantum.cg.CG
Class for Clebsch-Gordan coecient
Clebsch-Gordan coecients describe the angular momentum coupling between two systems. The coecients give the expansion of a coupled total angular momentum state
and an uncoupled tensor product state. The Clebsch-Gordan coecients are dened as
[R328] (page 1916):
,m1
Cjj21,m
= hj1 , m1 ; j2 , m2 |j3 , m3 i
2 ,j3 ,m3
Parameters j1, m1, j2, m2, j3, m3 : Number, Symbol
Terms determining the angular momentum of coupled angular momentum systems.
See Also:
Wigner3j (page 1751) Wigner-3j symbols
References
[R328] (page 1916)
Examples
Dene a Clebsch-Gordan coecient and evaluate its value
1750
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> from sympy.physics.quantum.cg import CG
>>> from sympy import S
>>> cg = CG(S(3)/2, S(3)/2, S(1)/2, -S(1)/2, 1, 1)
>>> cg
CG(3/2, 3/2, 1/2, -1/2, 1, 1)
>>> cg.doit()
sqrt(3)/2
class sympy.physics.quantum.cg.Wigner3j
Class for the Wigner-3j symbols
Wigner 3j-symbols are coecients determined by the coupling of two angular momenta.
When created, they are expressed as symbolic quantities that, for numerical parameters,
can be evaluated using the .doit() method [R329] (page 1916).
Parameters j1, m1, j2, m2, j3, m3 : Number, Symbol
Terms determining the angular momentum of coupled angular momentum systems.
See Also:
CG (page 1750) Clebsch-Gordan coecients
References
[R329] (page 1916)
Examples
Declare a Wigner-3j coecient and calcualte its value
>>> from sympy.physics.quantum.cg import Wigner3j
>>> w3j = Wigner3j(6,0,4,0,2,0)
>>> w3j
Wigner3j(6, 0, 4, 0, 2, 0)
>>> w3j.doit()
sqrt(715)/143
class sympy.physics.quantum.cg.Wigner6j
Class for the Wigner-6j symbols
See Also:
Wigner3j (page 1751) Wigner-3j symbols
class sympy.physics.quantum.cg.Wigner9j
Class for the Wigner-9j symbols
See Also:
Wigner3j (page 1751) Wigner-3j symbols
sympy.physics.quantum.cg.cg simp(e)
Simplify and combine CG coecients
This function uses various symmetry and properties of sums and products of ClebschGordan coecients to simplify statements involving these terms [R330] (page 1916).
5.36. Physics Module
1751
SymPy Documentation, Release 0.7.6
See Also:
CG (page 1750) Clebsh-Gordan coecients
References
[R330] (page 1916)
Examples
Simplify the sum over CG(a,alpha,0,0,a,alpha) for all alpha to 2*a+1
>>>
>>>
>>>
>>>
>>>
3
from sympy.physics.quantum.cg import CG, cg_simp
a = CG(1,1,0,0,1,1)
b = CG(1,0,0,0,1,0)
c = CG(1,-1,0,0,1,-1)
cg_simp(a+b+c)
Commutator The commutator: [A,B] = A*B - B*A.
class sympy.physics.quantum.commutator.Commutator
The standard commutator, in an unevaluated state.
Evaluating a commutator is dened [R331] (page 1916) as: [A, B] = A*B - B*A. This
class returns the commutator in an unevaluated form. To evaluate the commutator, use
the .doit() method.
Cannonical ordering of a commutator is [A, B] for A < B. The arguments of the commutator are put into canonical order using cmp . If B < A, then [B, A] is returned as
-[A, B].
Parameters A : Expr
The rst argument of the commutator [A,B].
B : Expr
The second argument of the commutator [A,B].
References
[R331] (page 1916)
Examples
>>>
>>>
>>>
>>>
>>>
from sympy.physics.quantum import Commutator, Dagger, Operator
from sympy.abc import x, y
A = Operator(A)
B = Operator(B)
C = Operator(C)
Create a commutator and use .doit() to evaluate it:
1752
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> comm = Commutator(A, B)
>>> comm
[A,B]
>>> comm.doit()
A*B - B*A
The commutator orders it arguments in canonical order:
>>> comm = Commutator(B, A); comm
-[A,B]
Commutative constants are factored out:
>>> Commutator(3*x*A, x*y*B)
3*x**2*y*[A,B]
Using .expand(commutator=True), the standard commutator expansion rules can be
applied:
>>> Commutator(A+B, C).expand(commutator=True)
[A,C] + [B,C]
>>> Commutator(A, B+C).expand(commutator=True)
[A,B] + [A,C]
>>> Commutator(A*B, C).expand(commutator=True)
[A,C]*B + A*[B,C]
>>> Commutator(A, B*C).expand(commutator=True)
[A,B]*C + B*[A,C]
Adjoint operations applied to the commutator are properly applied to the arguments:
>>> Dagger(Commutator(A, B))
-[Dagger(A),Dagger(B)]
doit(**hints)
Evaluate commutator
Constants Constants (like hbar) related to quantum mechanics.
Dagger
Hermitian conjugation.
class sympy.physics.quantum.dagger.Dagger
General Hermitian conjugate operation.
Take the Hermetian conjugate of an argument [R332] (page 1916). For matrices this
operation is equivalent to transpose and complex conjugate [R333] (page 1916).
Parameters arg : Expr
The sympy expression that we want to take the dagger of.
References
[R332] (page 1916), [R333] (page 1916)
5.36. Physics Module
1753
SymPy Documentation, Release 0.7.6
Examples
Daggering various quantum objects:
>>> from sympy.physics.quantum.dagger import Dagger
>>> from sympy.physics.quantum.state import Ket, Bra
>>> from sympy.physics.quantum.operator import Operator
>>> Dagger(Ket(psi))
<psi|
>>> Dagger(Bra(phi))
|phi>
>>> Dagger(Operator(A))
Dagger(A)
Inner and outer products:
>>> from sympy.physics.quantum import InnerProduct, OuterProduct
>>> Dagger(InnerProduct(Bra(a), Ket(b)))
<b|a>
>>> Dagger(OuterProduct(Ket(a), Bra(b)))
|b><a|
Powers, sums and products:
>>> A = Operator(A)
>>> B = Operator(B)
>>> Dagger(A*B)
Dagger(B)*Dagger(A)
>>> Dagger(A+B)
Dagger(A) + Dagger(B)
>>> Dagger(A**2)
Dagger(A)**2
Dagger also seamlessly handles complex numbers and matrices:
>>> from sympy import Matrix, I
>>> m = Matrix([[1,I],[2,I]])
>>> m
Matrix([
[1, I],
[2, I]])
>>> Dagger(m)
Matrix([
[ 1, 2],
[-I, -I]])
Inner Product Symbolic inner product.
class sympy.physics.quantum.innerproduct.InnerProduct
An unevaluated inner product between a Bra and a Ket [1].
Parameters bra : BraBase or subclass
The bra on the left side of the inner product.
ket : KetBase or subclass
The ket on the right side of the inner product.
1754
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
References
[R336] (page 1916)
Examples
Create an InnerProduct and check its properties:
>>> from sympy.physics.quantum import Bra, Ket, InnerProduct
>>> b = Bra(b)
>>> k = Ket(k)
>>> ip = b*k
>>> ip
<b|k>
>>> ip.bra
<b|
>>> ip.ket
|k>
In simple products of kets and bras inner products will be automatically identied and
created:
>>> b*k
<b|k>
But in more complex expressions, there is ambiguity in whether inner or outer products
should be created:
>>> k*b*k*b
|k><b|*|k>*<b|
A user can force the creation of a inner products in a complex expression by using parentheses to group the bra and ket:
>>> k*(b*k)*b
<b|k>*|k>*<b|
Notice how the inner product <b|k> moved to the left of the expression because inner
products are commutative complex numbers.
Tensor Product
Abstract tensor product.
class sympy.physics.quantum.tensorproduct.TensorProduct
The tensor product of two or more arguments.
For matrices, this uses matrix tensor product to compute the Kronecker or tensor product matrix. For other objects a symbolic TensorProduct instance is returned. The tensor
product is a non-commutative multiplication that is used primarily with operators and
states in quantum mechanics.
Currently, the tensor product distinguishes between commutative and non- commutative arguments. Commutative arguments are assumed to be scalars and are pulled out
in front of the TensorProduct. Non-commutative arguments remain in the resulting
TensorProduct.
Parameters args : tuple
A sequence of the objects to take the tensor product of.
5.36. Physics Module
1755
SymPy Documentation, Release 0.7.6
Examples
Start with a simple tensor product of sympy matrices:
>>> from sympy import I, Matrix, symbols
>>> from sympy.physics.quantum import TensorProduct
>>> m1 = Matrix([[1,2],[3,4]])
>>> m2 = Matrix([[1,0],[0,1]])
>>> TensorProduct(m1, m2)
Matrix([
[1, 0, 2, 0],
[0, 1, 0, 2],
[3, 0, 4, 0],
[0, 3, 0, 4]])
>>> TensorProduct(m2, m1)
Matrix([
[1, 2, 0, 0],
[3, 4, 0, 0],
[0, 0, 1, 2],
[0, 0, 3, 4]])
We can also construct tensor products of non-commutative symbols:
>>>
>>>
>>>
>>>
>>>
AxB
from sympy import Symbol
A = Symbol(A,commutative=False)
B = Symbol(B,commutative=False)
tp = TensorProduct(A, B)
tp
We can take the dagger of a tensor product (note the order does NOT reverse like the
dagger of a normal product):
>>> from sympy.physics.quantum import Dagger
>>> Dagger(tp)
Dagger(A)xDagger(B)
Expand can be used to distribute a tensor product across addition:
>>> C = Symbol(C,commutative=False)
>>> tp = TensorProduct(A+B,C)
>>> tp
(A + B)xC
>>> tp.expand(tensorproduct=True)
AxC + BxC
sympy.physics.quantum.tensorproduct.tensor product simp(e, **hints)
Try to simplify and combine TensorProducts.
In general this will try to pull expressions inside of TensorProducts. It currently only
works for relatively simple cases where the products have only scalars, raw TensorProducts, not Add, Pow, Commutators of TensorProducts. It is best to see what it does
by showing examples.
1756
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>>
>>>
>>>
>>>
>>>
>>>
>>>
from sympy.physics.quantum import tensor_product_simp
from sympy.physics.quantum import TensorProduct
from sympy import Symbol
A = Symbol(A,commutative=False)
B = Symbol(B,commutative=False)
C = Symbol(C,commutative=False)
D = Symbol(D,commutative=False)
First see what happens to products of tensor products:
>>> e = TensorProduct(A,B)*TensorProduct(C,D)
>>> e
AxB*CxD
>>> tensor_product_simp(e)
(A*C)x(B*D)
This is the core logic of this function, and it works inside, powers, sums, commutators
and anticommutators as well:
>>> tensor_product_simp(e**2)
(A*C)x(B*D)**2
States and Operators
Cartesian Operators and States
mentum.
Operators and states for 1D cartesian position and mo-
TODO:
Add 3D classes to mappings in operatorset.py
class sympy.physics.quantum.cartesian.XOp
1D cartesian position operator.
class sympy.physics.quantum.cartesian.YOp
Y cartesian coordinate operator (for 2D or 3D systems)
class sympy.physics.quantum.cartesian.ZOp
Z cartesian coordinate operator (for 3D systems)
class sympy.physics.quantum.cartesian.PxOp
1D cartesian momentum operator.
class sympy.physics.quantum.cartesian.XKet
1D cartesian position eigenket.
position
The position of the state.
class sympy.physics.quantum.cartesian.XBra
1D cartesian position eigenbra.
position
The position of the state.
class sympy.physics.quantum.cartesian.PxKet
1D cartesian momentum eigenket.
5.36. Physics Module
1757
SymPy Documentation, Release 0.7.6
momentum
The momentum of the state.
class sympy.physics.quantum.cartesian.PxBra
1D cartesian momentum eigenbra.
momentum
The momentum of the state.
class sympy.physics.quantum.cartesian.PositionState3D
Base class for 3D cartesian position eigenstates
position x
The x coordinate of the state
position y
The y coordinate of the state
position z
The z coordinate of the state
class sympy.physics.quantum.cartesian.PositionKet3D
3D cartesian position eigenket
class sympy.physics.quantum.cartesian.PositionBra3D
3D cartesian position eigenbra
Hilbert Space
Hilbert spaces for quantum mechanics.
Authors: * Brian Granger * Matt Curry
class sympy.physics.quantum.hilbert.HilbertSpace
An abstract Hilbert space for quantum mechanics.
In short, a Hilbert space is an abstract vector space that is complete with inner products
dened [R334] (page 1916).
References
[R334] (page 1916)
Examples
>>> from sympy.physics.quantum.hilbert import HilbertSpace
>>> hs = HilbertSpace()
>>> hs
H
dimension
Return the Hilbert dimension of the space.
class sympy.physics.quantum.hilbert.ComplexSpace
Finite dimensional Hilbert space of complex vectors.
The elements of this Hilbert space are n-dimensional complex valued vectors with the
usual inner product that takes the complex conjugate of the vector on the right.
A classic example of this type of Hilbert space is spin-1/2, which is ComplexSpace(2).
Generalizing to spin-s, the space is ComplexSpace(2*s+1). Quantum computing with N
qubits is done with the direct product space ComplexSpace(2)**N.
1758
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import symbols
>>> from sympy.physics.quantum.hilbert import ComplexSpace
>>> c1 = ComplexSpace(2)
>>> c1
C(2)
>>> c1.dimension
2
>>> n = symbols(n)
>>> c2 = ComplexSpace(n)
>>> c2
C(n)
>>> c2.dimension
n
class sympy.physics.quantum.hilbert.L2
The Hilbert space of square integrable functions on an interval.
An L2 object takes in a single sympy Interval argument which represents the interval its
functions (vectors) are dened on.
Examples
>>> from sympy import Interval, oo
>>> from sympy.physics.quantum.hilbert import L2
>>> hs = L2(Interval(0,oo))
>>> hs
L2([0, oo))
>>> hs.dimension
oo
>>> hs.interval
[0, oo)
class sympy.physics.quantum.hilbert.FockSpace
The Hilbert space for second quantization.
Technically, this Hilbert space is a innite direct sum of direct products of single particle
Hilbert spaces [R335] (page 1916). This is a mess, so we have a class to represent it
directly.
References
[R335] (page 1916)
Examples
>>>
>>>
>>>
F
>>>
oo
from sympy.physics.quantum.hilbert import FockSpace
hs = FockSpace()
hs
hs.dimension
5.36. Physics Module
1759
SymPy Documentation, Release 0.7.6
Operator
Quantum mechanical operators.
TODO:
Fix early 0 in apply operators.
Debug and test apply operators.
Get cse working with classes in this le.
Doctests and documentation of special methods for InnerProduct, Commutator, AntiCommutator, represent, apply operators.
class sympy.physics.quantum.operator.Operator
Base class for non-commuting quantum operators.
An operator maps between quantum states [R337] (page 1916). In quantum mechanics,
observables (including, but not limited to, measured physical values) are represented as
Hermitian operators [R338] (page 1916).
Parameters args : tuple
The list of numbers or parameters that uniquely specify the operator.
For time-dependent operators, this will include the time.
References
[R337] (page 1916), [R338] (page 1916)
Examples
Create an operator and examine its attributes:
>>> from sympy.physics.quantum import Operator
>>> from sympy import symbols, I
>>> A = Operator(A)
>>> A
A
>>> A.hilbert_space
H
>>> A.label
(A,)
>>> A.is_commutative
False
Create another operator and do some arithmetic operations:
>>> B = Operator(B)
>>> C = 2*A*A + I*B
>>> C
2*A**2 + I*B
Operators dont commute:
>>> A.is_commutative
False
>>> B.is_commutative
False
>>> A*B == B*A
False
1760
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Polymonials of operators respect the commutation properties:
>>> e = (A+B)**3
>>> e.expand()
A*B*A + A*B**2 + A**2*B + A**3 + B*A*B + B*A**2 + B**2*A + B**3
Operator inverses are handle symbolically:
>>> A.inv()
A**(-1)
>>> A*A.inv()
1
class sympy.physics.quantum.operator.HermitianOperator
A Hermitian operator that satises H == Dagger(H).
Parameters args : tuple
The list of numbers or parameters that uniquely specify the operator.
For time-dependent operators, this will include the time.
Examples
>>> from sympy.physics.quantum import Dagger, HermitianOperator
>>> H = HermitianOperator(H)
>>> Dagger(H)
H
class sympy.physics.quantum.operator.UnitaryOperator
A unitary operator that satises U*Dagger(U) == 1.
Parameters args : tuple
The list of numbers or parameters that uniquely specify the operator.
For time-dependent operators, this will include the time.
Examples
>>> from sympy.physics.quantum import Dagger, UnitaryOperator
>>> U = UnitaryOperator(U)
>>> U*Dagger(U)
1
class sympy.physics.quantum.operator.IdentityOperator(*args, **hints)
An identity operator I that satises op * I == I * op == op for any operator op.
Parameters N : Integer
Optional parameter that species the dimension of the Hilbert space
of operator. This is used when generating a matrix representation.
Examples
>>> from sympy.physics.quantum import IdentityOperator
>>> IdentityOperator()
I
5.36. Physics Module
1761
SymPy Documentation, Release 0.7.6
class sympy.physics.quantum.operator.OuterProduct
An unevaluated outer product between a ket and bra.
This constructs an outer product between any subclass of KetBase and BraBase as
|a><b|. An OuterProduct inherits from Operator as they act as operators in quantum
expressions. For reference see [R339] (page 1916).
Parameters ket : KetBase
The ket on the left side of the outer product.
bar : BraBase
The bra on the right side of the outer product.
References
[R339] (page 1916)
Examples
Create a simple outer product by hand and take its dagger:
>>> from sympy.physics.quantum import Ket, Bra, OuterProduct, Dagger
>>> from sympy.physics.quantum import Operator
>>> k = Ket(k)
>>> b = Bra(b)
>>> op = OuterProduct(k, b)
>>> op
|k><b|
>>> op.hilbert_space
H
>>> op.ket
|k>
>>> op.bra
<b|
>>> Dagger(op)
|b><k|
In simple products of kets and bras outer products will be automatically identied and
created:
>>> k*b
|k><b|
But in more complex expressions, outer products are not automatically created:
>>> A = Operator(A)
>>> A*k*b
A*|k>*<b|
A user can force the creation of an outer product in a complex expression by using parentheses to group the ket and bra:
>>> A*(k*b)
A*|k><b|
1762
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
bra
Return the bra on the right side of the outer product.
ket
Return the ket on the left side of the outer product.
class sympy.physics.quantum.operator.DifferentialOperator
An operator for representing the dierential operator, i.e. d/dx
It is initialized by passing two arguments. The rst is an arbitrary expression that involves a function, such as Derivative(f(x), x). The second is the function (e.g. f(x))
which we are to replace with the Wavefunction that this DifferentialOperator is applied to.
Parameters expr : Expr
The arbitrary expression which the appropriate Wavefunction is to be
substituted into
func : Expr
A function (e.g. f(x)) which is to be replaced with the appropriate
Wavefunction when this DierentialOperator is applied
Examples
You can dene a completely arbitrary expression and specify where the Wavefunction is
to be substituted
>>> from sympy import Derivative, Function, Symbol
>>> from sympy.physics.quantum.operator import DifferentialOperator
>>> from sympy.physics.quantum.state import Wavefunction
>>> from sympy.physics.quantum.qapply import qapply
>>> f = Function(f)
>>> x = Symbol(x)
>>> d = DifferentialOperator(1/x*Derivative(f(x), x), f(x))
>>> w = Wavefunction(x**2, x)
>>> d.function
f(x)
>>> d.variables
(x,)
>>> qapply(d*w)
Wavefunction(2, x)
expr
Returns the arbitary expression which is to have the Wavefunction substituted into
it
Examples
>>> from sympy.physics.quantum.operator import DifferentialOperator
>>> from sympy import Function, Symbol, Derivative
>>> x = Symbol(x)
>>> f = Function(f)
>>> d = DifferentialOperator(Derivative(f(x), x), f(x))
>>> d.expr
Derivative(f(x), x)
>>> y = Symbol(y)
5.36. Physics Module
1763
SymPy Documentation, Release 0.7.6
>>> d = DifferentialOperator(Derivative(f(x, y), x) +
...
Derivative(f(x, y), y), f(x, y))
>>> d.expr
Derivative(f(x, y), x) + Derivative(f(x, y), y)
free symbols
Return the free symbols of the expression.
function
Returns the function which is to be replaced with the Wavefunction
Examples
>>> from sympy.physics.quantum.operator import DifferentialOperator
>>> from sympy import Function, Symbol, Derivative
>>> x = Symbol(x)
>>> f = Function(f)
>>> d = DifferentialOperator(Derivative(f(x), x), f(x))
>>> d.function
f(x)
>>> y = Symbol(y)
>>> d = DifferentialOperator(Derivative(f(x, y), x) +
...
Derivative(f(x, y), y), f(x, y))
>>> d.function
f(x, y)
variables
Returns the variables with which the function in the specied arbitrary expression
is evaluated
Examples
>>> from sympy.physics.quantum.operator import DifferentialOperator
>>> from sympy import Symbol, Function, Derivative
>>> x = Symbol(x)
>>> f = Function(f)
>>> d = DifferentialOperator(1/x*Derivative(f(x), x), f(x))
>>> d.variables
(x,)
>>> y = Symbol(y)
>>> d = DifferentialOperator(Derivative(f(x, y), x) +
...
Derivative(f(x, y), y), f(x, y))
>>> d.variables
(x, y)
Operator/State Helper Functions A module for mapping operators to their corresponding
eigenstates and vice versa
It contains a global dictionary with eigenstate-operator pairings. If a new state-operator pair
is created, this dictionary should be updated as well.
It also contains functions operators to state and state to operators for mapping between the
two. These can handle both classes and instances of operators and states. See the individual
function descriptions for details.
1764
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
TODO List: - Update the dictionary with a complete list of state-operator pairs
sympy.physics.quantum.operatorset.operators to state(operators, **options)
Returns the eigenstate of the given operator or set of operators
A global function for mapping operator classes to their associated states. It takes either
an Operator or a set of operators and returns the state associated with these.
This function can handle both instances of a given operator or just the class itself (i.e.
both XOp() and XOp)
There are multiple use cases to consider:
1) A class or set of classes is passed: First, we try to instantiate default instances for these
operators. If this fails, then the class is simply returned. If we succeed in instantiating
default instances, then we try to call state. operators to state on the operator instances.
If this fails, the class is returned. Otherwise, the instance returned by operators to state
is returned.
2) An instance or set of instances is passed: In this case, state. operators to state is
called on the instances passed. If this fails, a state class is returned. If the method
returns an instance, that instance is returned.
In both cases, if the operator class or set does not exist in the state mapping dictionary,
None is returned.
Parameters arg: Operator or set :
The class or instance of the operator or set of operators to be mapped
to a state
Examples
>>> from sympy.physics.quantum.cartesian import XOp, PxOp
>>> from sympy.physics.quantum.operatorset import operators_to_state
>>> from sympy.physics.quantum.operator import Operator
>>> operators_to_state(XOp)
|x>
>>> operators_to_state(XOp())
|x>
>>> operators_to_state(PxOp)
|px>
>>> operators_to_state(PxOp())
|px>
>>> operators_to_state(Operator)
|psi>
>>> operators_to_state(Operator())
|psi>
sympy.physics.quantum.operatorset.state to operators(state, **options)
Returns the operator or set of operators corresponding to the given eigenstate
A global function for mapping state classes to their associated operators or sets of operators. It takes either a state class or instance.
This function can handle both instances of a given state or just the class itself (i.e. both
XKet() and XKet)
There are multiple use cases to consider:
1) A state class is passed: In this case, we rst try instantiating a default instance of
the class. If this succeeds, then we try to call state. state to operators on that instance.
5.36. Physics Module
1765
SymPy Documentation, Release 0.7.6
If the creation of the default instance or if the calling of state to operators fails, then
either an operator class or set of operator classes is returned. Otherwise, the appropriate
operator instances are returned.
2) A state instance is returned: Here, state. state to operators is called for the instance.
If this fails, then a class or set of operator classes is returned. Otherwise, the instances
are returned.
In either case, if the states class does not exist in state mapping, None is returned.
Parameters arg: StateBase class or instance (or subclasses) :
The class or instance of the state to be mapped to an operator or set
of operators
Examples
>>>
>>>
>>>
>>>
X
>>>
X
>>>
Px
>>>
Px
>>>
Px
>>>
X
>>>
O
>>>
O
Qapply
from sympy.physics.quantum.cartesian import XKet, PxKet, XBra, PxBra
from sympy.physics.quantum.operatorset import state_to_operators
from sympy.physics.quantum.state import Ket, Bra
state_to_operators(XKet)
state_to_operators(XKet())
state_to_operators(PxKet)
state_to_operators(PxKet())
state_to_operators(PxBra)
state_to_operators(XBra)
state_to_operators(Ket)
state_to_operators(Bra)
Logic for applying operators to states.
Todo: * Sometimes the nal result needs to be expanded, we should do this by hand.
sympy.physics.quantum.qapply.qapply(e, **options)
Apply operators to states in a quantum expression.
Parameters e : Expr
The expression containing operators and states. This expression tree
will be walked to nd operators acting on states symbolically.
options : dict
A dict of key/value pairs that determine how the operator actions are
carried out.
The following options are valid:
dagger: try to apply Dagger operators to the left (default: False).
ip doit: call .doit() in inner products when they are encountered
(default: True).
1766
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Returns e : Expr
The original expression, but with the operators applied to states.
Represent
Logic for representing operators in state in various bases.
TODO:
Get represent working with continuous hilbert spaces.
Document default basis functionality.
sympy.physics.quantum.represent.represent(expr, **options)
Represent the quantum expression in the given basis.
In quantum mechanics abstract states and operators can be represented in various basis
sets. Under this operation the follow transforms happen:
Ket -> column vector or function
Bra -> row vector of function
Operator -> matrix or dierential operator
This function is the top-level interface for this action.
This function walks the sympy expression tree looking for QExpr instances that have a
represent method. This method is then called and the object is replaced by the representation returned by this method. By default, the represent method will dispatch
to other methods that handle the representation logic for a particular basis set. The
naming convention for these methods is the following:
def represent FooBasis(self, e, basis, **options)
This function will have the logic for representing instances of its class in the basis set
having a class named FooBasis.
Parameters expr : Expr
The expression to represent.
basis : Operator, basis set
An object that contains the information about the basis set. If an operator is used, the basis is assumed to be the orthonormal eigenvectors
of that operator. In general though, the basis argument can be any
object that contains the basis set information.
options : dict
Key/value pairs of options that are passed to the underlying method
that nds the representation. These options can be used to control
how the representation is done. For example, this is where the size
of the basis set would be set.
Returns e : Expr
The SymPy expression of the represented quantum expression.
Examples
Here we subclass Operator and Ket to create the z-spin operator and its spin 1/2 up
eigenstate. By denining the represent SzOp method, the ket can be represented in
the z-spin basis.
5.36. Physics Module
1767
SymPy Documentation, Release 0.7.6
>>> from sympy.physics.quantum import Operator, represent, Ket
>>> from sympy import Matrix
>>> class SzUpKet(Ket):
...
def _represent_SzOp(self, basis, **options):
...
return Matrix([1,0])
...
>>> class SzOp(Operator):
...
pass
...
>>> sz = SzOp(Sz)
>>> up = SzUpKet(up)
>>> represent(up, basis=sz)
Matrix([
[1],
[0]])
Here we see an example of representations in a continuous basis. We see that the result
of representing various combinations of cartesian position operators and kets give us
continuous expressions involving DiracDelta functions.
>>> from sympy.physics.quantum.cartesian import XOp, XKet, XBra
>>> X = XOp()
>>> x = XKet()
>>> y = XBra(y)
>>> represent(X*x)
x*DiracDelta(x - x_2)
>>> represent(X*x*y)
x*DiracDelta(x - x_3)*DiracDelta(x_1 - y)
sympy.physics.quantum.represent.rep innerproduct(expr, **options)
Returns an innerproduct like representation (e.g. <x|x>) for the given state.
Attempts to calculate inner product with a bra from the specied basis. Should only be
passed an instance of KetBase or BraBase
Parameters expr : KetBase or BraBase
The expression to be represented
Examples
>>> from sympy.physics.quantum.represent import rep_innerproduct
>>> from sympy.physics.quantum.cartesian import XOp, XKet, PxOp, PxKet
>>> rep_innerproduct(XKet())
DiracDelta(x - x_1)
>>> rep_innerproduct(XKet(), basis=PxOp())
sqrt(2)*exp(-I*px_1*x/hbar)/(2*sqrt(hbar)*sqrt(pi))
>>> rep_innerproduct(PxKet(), basis=XOp())
sqrt(2)*exp(I*px*x_1/hbar)/(2*sqrt(hbar)*sqrt(pi))
sympy.physics.quantum.represent.rep expectation(expr, **options)
Returns an <x|A|x> type representation for the given operator.
Parameters expr : Operator
Operator to be represented in the specied basis
1768
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.physics.quantum.cartesian import XOp, XKet, PxOp, PxKet
>>> from sympy.physics.quantum.represent import rep_expectation
>>> rep_expectation(XOp())
x_1*DiracDelta(x_1 - x_2)
>>> rep_expectation(XOp(), basis=PxOp())
<px_2|*X*|px_1>
>>> rep_expectation(XOp(), basis=PxKet())
<px_2|*X*|px_1>
sympy.physics.quantum.represent.integrate result(orig expr, result, **options)
Returns the result of integrating over any unities (|x><x|) in the given expression. Intended for integrating over the result of representations in continuous bases.
This function integrates over any unities that may have been inserted into the quantum
expression and returns the result. It uses the interval of the Hilbert space of the basis
state passed to it in order to gure out the limits of integration. The unities option must
be specied for this to work.
Note: This is mostly used internally by represent(). Examples are given merely to show
the use cases.
Parameters orig expr : quantum expression
The original expression which was to be represented
result: Expr :
The resulting representation that we wish to integrate over
Examples
>>> from sympy import symbols, DiracDelta
>>> from sympy.physics.quantum.represent import integrate_result
>>> from sympy.physics.quantum.cartesian import XOp, XKet
>>> x_ket = XKet()
>>> X_op = XOp()
>>> x, x_1, x_2 = symbols(x, x_1, x_2)
>>> integrate_result(X_op*x_ket, x*DiracDelta(x-x_1)*DiracDelta(x_1-x_2))
x*DiracDelta(x - x_1)*DiracDelta(x_1 - x_2)
>>> integrate_result(X_op*x_ket, x*DiracDelta(x-x_1)*DiracDelta(x_1-x_2),
...
unities=[1])
x*DiracDelta(x - x_2)
sympy.physics.quantum.represent.get basis(expr, **options)
Returns a basis state instance corresponding to the basis specied in options=s. If no
basis is specied, the function tries to form a default basis state of the given expression.
There are three behaviors:
1.The basis specied in options is already an instance of StateBase. If this is the case,
it is simply returned. If the class is specied but not an instance, a default instance
is returned.
2.The basis specied is an operator or set of operators. If this is the case, the operator to state mapping method is used.
5.36. Physics Module
1769
SymPy Documentation, Release 0.7.6
3.No basis is specied. If expr is a state, then a default instance of its class is returned.
If expr is an operator, then it is mapped to the corresponding state. If it is neither,
then we cannot obtain the basis state.
If the basis cannot be mapped, then it is not changed.
This will be called from within represent, and represent will only pass QExprs.
TODO (?): Support for Muls and other types of expressions?
Parameters expr : Operator or StateBase
Expression whose basis is sought
Examples
>>>
>>>
>>>
>>>
>>>
|x>
>>>
|x>
>>>
|px>
>>>
|px>
from sympy.physics.quantum.represent import get_basis
from sympy.physics.quantum.cartesian import XOp, XKet, PxOp, PxKet
x = XKet()
X = XOp()
get_basis(x)
get_basis(X)
get_basis(x, basis=PxOp())
get_basis(x, basis=PxKet)
sympy.physics.quantum.represent.enumerate states(*args, **options)
Returns instances of the given state with dummy indices appended
Operates in two dierent modes:
1.Two arguments are passed to it. The rst is the base state which is to be indexed,
and the second argument is a list of indices to append.
2.Three arguments are passed. The rst is again the base state to be indexed. The
second is the start index for counting. The nal argument is the number of kets you
wish to receive.
Tries to call state. enumerate state. If this fails, returns an empty list
Parameters args : list
See list of operation modes above for explanation
Examples
>>> from sympy.physics.quantum.cartesian import XBra, XKet
>>> from sympy.physics.quantum.represent import enumerate_states
>>> test = XKet(foo)
>>> enumerate_states(test, 1, 3)
[|foo_1>, |foo_2>, |foo_3>]
>>> test2 = XBra(bar)
>>> enumerate_states(test2, [4, 5, 10])
[<bar_4|, <bar_5|, <bar_10|]
1770
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Spin
Quantum mechanical angular momemtum.
class sympy.physics.quantum.spin.Rotation
Wigner D operator in terms of Euler angles.
Denes the rotation operator in terms of the Euler angles dened by the z-y-z convention
for a passive transformation. That is the coordinate axes are rotated rst about the zaxis, giving the new x-y-z axes. Then this new coordinate system is rotated about the
new y-axis, giving new x-y-z axes. Then this new coordinate system is rotated about
the z-axis. Conventions follow those laid out in [R340] (page 1916).
Parameters alpha : Number, Symbol
First Euler Angle
beta : Number, Symbol
Second Euler angle
gamma : Number, Symbol
Third Euler angle
See Also:
WignerD (page 1773) Symbolic Wigner-D function
D (page 1771) Wigner-D function
d (page 1772) Wigner small-d function
References
[R340] (page 1916)
Examples
A simple example rotation operator:
>>> from sympy import pi
>>> from sympy.physics.quantum.spin import Rotation
>>> Rotation(pi, 0, pi/2)
R(pi,0,pi/2)
With symbolic Euler angles and calculating the inverse rotation operator:
>>> from sympy import symbols
>>> a, b, c = symbols(a b c)
>>> Rotation(a, b, c)
R(a,b,c)
>>> Rotation(a, b, c).inverse()
R(-c,-b,-a)
classmethod D(j, m, mp, alpha, beta, gamma)
Wigner D-function.
Returns an instance of the WignerD class corresponding to the Wigner-D function
specied by the parameters.
Parameters j : Number
Total angular momentum
5.36. Physics Module
1771
SymPy Documentation, Release 0.7.6
m : Number
Eigenvalue of angular momentum along axis after rotation
mp : Number
Eigenvalue of angular momentum along rotated axis
alpha : Number, Symbol
First Euler angle of rotation
beta : Number, Symbol
Second Euler angle of rotation
gamma : Number, Symbol
Third Euler angle of rotation
See Also:
WignerD (page 1773) Symbolic Wigner-D function
Examples
Return the Wigner-D matrix element for a dened rotation, both numerical and symbolic:
>>> from sympy.physics.quantum.spin import Rotation
>>> from sympy import pi, symbols
>>> alpha, beta, gamma = symbols(alpha beta gamma)
>>> Rotation.D(1, 1, 0,pi, pi/2,-pi)
WignerD(1, 1, 0, pi, pi/2, -pi)
classmethod d(j, m, mp, beta)
Wigner small-d function.
Returns an instance of the WignerD class corresponding to the Wigner-D function
specied by the parameters with the alpha and gamma angles given as 0.
Parameters j : Number
Total angular momentum
m : Number
Eigenvalue of angular momentum along axis after rotation
mp : Number
Eigenvalue of angular momentum along rotated axis
beta : Number, Symbol
Second Euler angle of rotation
See Also:
WignerD (page 1773) Symbolic Wigner-D function
1772
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
Return the Wigner-D matrix element for a dened rotation, both numerical and symbolic:
>>> from sympy.physics.quantum.spin import Rotation
>>> from sympy import pi, symbols
>>> beta = symbols(beta)
>>> Rotation.d(1, 1, 0, pi/2)
WignerD(1, 1, 0, 0, pi/2, 0)
class sympy.physics.quantum.spin.WignerD
Wigner-D function
The Wigner D-function gives the matrix elements of the rotation operator in the jmrepresentation. For the Euler angles , , , the D-function is dened such that:
< j, m|R(, , )|j 0 , m0 >= jj 0 D(j, m, m0 , , , )
Where the rotation operator is as dened by the Rotation class [R341] (page 1916).
The Wigner D-function dened in this way gives:
0
D(j, m, m0 , , , ) = eim d(j, m, m0 , )eim
Where d is the Wigner small-d function, which is given by Rotation.d.
The Wigner small-d function gives the component of the Wigner D-function that is determined by the second Euler angle. That is the Wigner D-function is:
0
D(j, m, m0 , , , ) = eim d(j, m, m0 , )eim
Where d is the small-d function. The Wigner D-function is given by Rotation.D.
Note that to evaluate the D-function, the j, m and mp parameters must be integer or half
integer numbers.
Parameters j : Number
Total angular momentum
m : Number
Eigenvalue of angular momentum along axis after rotation
mp : Number
Eigenvalue of angular momentum along rotated axis
alpha : Number, Symbol
First Euler angle of rotation
beta : Number, Symbol
Second Euler angle of rotation
gamma : Number, Symbol
Third Euler angle of rotation
See Also:
Rotation (page 1771) Rotation operator
5.36. Physics Module
1773
SymPy Documentation, Release 0.7.6
References
[R341] (page 1916)
Examples
Evaluate the Wigner-D matrix elements of a simple rotation:
>>> from sympy.physics.quantum.spin import Rotation
>>> from sympy import pi
>>> rot = Rotation.D(1, 1, 0, pi, pi/2, 0)
>>> rot
WignerD(1, 1, 0, pi, pi/2, 0)
>>> rot.doit()
sqrt(2)/2
Evaluate the Wigner-d matrix elements of a simple rotation
>>> rot = Rotation.d(1, 1, 0, pi/2)
>>> rot
WignerD(1, 1, 0, 0, pi/2, 0)
>>> rot.doit()
-sqrt(2)/2
class sympy.physics.quantum.spin.JxKet
Eigenket of Jx.
See JzKet for the usage of spin eigenstates.
See Also:
JzKet (page 1774) Usage of spin states
class sympy.physics.quantum.spin.JxBra
Eigenbra of Jx.
See JzKet for the usage of spin eigenstates.
See Also:
JzKet (page 1774) Usage of spin states
class sympy.physics.quantum.spin.JyKet
Eigenket of Jy.
See JzKet for the usage of spin eigenstates.
See Also:
JzKet (page 1774) Usage of spin states
class sympy.physics.quantum.spin.JyBra
Eigenbra of Jy.
See JzKet for the usage of spin eigenstates.
See Also:
JzKet (page 1774) Usage of spin states
1774
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
class sympy.physics.quantum.spin.JzKet
Eigenket of Jz.
Spin state which is an eigenstate of the Jz operator. Uncoupled states, that is states
representing the interaction of multiple separate spin states, are dened as a tensor
product of states.
Parameters j : Number, Symbol
Total spin angular momentum
m : Number, Symbol
Eigenvalue of the Jz spin operator
See Also:
JzKetCoupled (page 1777) Coupled eigenstates
TensorProduct Used to specify uncoupled states
uncouple (page 1780) Uncouples states given coupling parameters
couple (page 1779) Couples uncoupled states
Examples
Normal States:
Dening simple spin states, both numerical and symbolic:
>>> from sympy.physics.quantum.spin import JzKet, JxKet
>>> from sympy import symbols
>>> JzKet(1, 0)
|1,0>
>>> j, m = symbols(j m)
>>> JzKet(j, m)
|j,m>
Rewriting the JzKet in terms of eigenkets of the Jx operator: Note: that the resulting
eigenstates are JxKets
>>> JzKet(1,1).rewrite(Jx)
|1,-1>/2 - sqrt(2)*|1,0>/2 + |1,1>/2
Get the vector representation of a state in terms of the basis elements of the Jx operator:
>>> from sympy.physics.quantum.represent import represent
>>> from sympy.physics.quantum.spin import Jx, Jz
>>> represent(JzKet(1,-1), basis=Jx)
Matrix([
[
1/2],
[sqrt(2)/2],
[
1/2]])
Apply innerproducts between states:
>>> from sympy.physics.quantum.innerproduct import InnerProduct
>>> from sympy.physics.quantum.spin import JxBra
>>> i = InnerProduct(JxBra(1,1), JzKet(1,1))
>>> i
<1,1|1,1>
5.36. Physics Module
1775
SymPy Documentation, Release 0.7.6
>>> i.doit()
1/2
Uncoupled States:
Dene an uncoupled state as a TensorProduct between two Jz eigenkets:
>>> from sympy.physics.quantum.tensorproduct import TensorProduct
>>> j1,m1,j2,m2 = symbols(j1 m1 j2 m2)
>>> TensorProduct(JzKet(1,0), JzKet(1,1))
|1,0>x|1,1>
>>> TensorProduct(JzKet(j1,m1), JzKet(j2,m2))
|j1,m1>x|j2,m2>
A TensorProduct can be rewritten, in which case the eigenstates that make up the tensor
product is rewritten to the new basis:
>>> TensorProduct(JzKet(1,1),JxKet(1,1)).rewrite(Jz)
|1,1>x|1,-1>/2 + sqrt(2)*|1,1>x|1,0>/2 + |1,1>x|1,1>/2
The represent method for TensorProducts gives the vector representation of the state.
Note that the state in the product basis is the equivalent of the tensor product of the
vector representation of the component eigenstates:
>>> represent(TensorProduct(JzKet(1,0),JzKet(1,1)))
Matrix([
[0],
[0],
[0],
[1],
[0],
[0],
[0],
[0],
[0]])
>>> represent(TensorProduct(JzKet(1,1),JxKet(1,1)), basis=Jz)
Matrix([
[
1/2],
[sqrt(2)/2],
[
1/2],
[
0],
[
0],
[
0],
[
0],
[
0],
[
0]])
class sympy.physics.quantum.spin.JzBra
Eigenbra of Jz.
See the JzKet for the usage of spin eigenstates.
See Also:
JzKet (page 1774) Usage of spin states
class sympy.physics.quantum.spin.JxKetCoupled
Coupled eigenket of Jx.
See JzKetCoupled for the usage of coupled spin eigenstates.
1776
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
See Also:
JzKetCoupled (page 1777) Usage of coupled spin states
class sympy.physics.quantum.spin.JxBraCoupled
Coupled eigenbra of Jx.
See JzKetCoupled for the usage of coupled spin eigenstates.
See Also:
JzKetCoupled (page 1777) Usage of coupled spin states
class sympy.physics.quantum.spin.JyKetCoupled
Coupled eigenket of Jy.
See JzKetCoupled for the usage of coupled spin eigenstates.
See Also:
JzKetCoupled (page 1777) Usage of coupled spin states
class sympy.physics.quantum.spin.JyBraCoupled
Coupled eigenbra of Jy.
See JzKetCoupled for the usage of coupled spin eigenstates.
See Also:
JzKetCoupled (page 1777) Usage of coupled spin states
class sympy.physics.quantum.spin.JzKetCoupled
Coupled eigenket of Jz
Spin state that is an eigenket of Jz which represents the coupling of separate spin spaces.
The arguments for creating instances of JzKetCoupled are j, m, jn and an optional jcoupling argument. The j and m options are the total angular momentum quantum numbers, as used for normal states (e.g. JzKet).
The other required parameter in jn, which is a tuple dening the jn angular momentum
quantum numbers of the product spaces. So for example, if a state represented the
coupling of the product basis state |j1 , m1 i|j2 , m2 i, the jn for this state would be (j1,j2).
The nal option is jcoupling, which is used to dene how the spaces specied by jn
are coupled, which includes both the order these spaces are coupled together and the
quantum numbers that arise from these couplings. The jcoupling parameter itself is
a list of lists, such that each of the sublists denes a single coupling between the spin
spaces. If there are N coupled angular momentum spaces, that is jn has N elements, then
there must be N-1 sublists. Each of these sublists making up the jcoupling parameter
have length 3. The rst two elements are the indicies of the product spaces that are
considered to be coupled together. For example, if we want to couple j1 and j4 , the
indicies would be 1 and 4. If a state has already been coupled, it is referenced by the
smallest index that is coupled, so if j2 and j4 has already been coupled to some j24 , then
this value can be coupled by referencing it with index 2. The nal element of the sublist
is the quantum number of the coupled state. So putting everything together, into a
valid sublist for jcoupling, if j1 and j2 are coupled to an angular momentum space with
quantum number j12 with the value j12, the sublist would be (1,2,j12), N-1 of these
sublists are used in the list for jcoupling.
Note the jcoupling parameter is optional, if it is not specied, the default coupling is
taken. This default value is to coupled the spaces in order and take the quantum number
5.36. Physics Module
1777
SymPy Documentation, Release 0.7.6
of the coupling to be the maximum value. For example, if the spin spaces are j1 , j2 , j3 ,
j4 , then the default coupling couples j1 and j2 to j12 = j1 + j2 , then, j12 and j3 are coupled
to j123 = j12 + j3 , and nally j123 and j4 to j = j123 + j4 . The jcoupling value that would
correspond to this is:
((1,2,j1+j2),(1,3,j1+j2+j3))
Parameters args : tuple
The arguments that must be passed are j, m, jn, and jcoupling. The j
value is the total angular momentum. The m value is the eigenvalue of
the Jz spin operator. The jn list are the j values of argular momentum
spaces coupled together. The jcoupling parameter is an optional
parameter dening how the spaces are coupled together. See the
above description for how these coupling parameters are dened.
See Also:
JzKet (page 1774) Normal spin eigenstates
uncouple (page 1780) Uncoupling of coupling spin states
couple (page 1779) Coupling of uncoupled spin states
Examples
Dening simple spin states, both numerical and symbolic:
>>> from sympy.physics.quantum.spin import JzKetCoupled
>>> from sympy import symbols
>>> JzKetCoupled(1, 0, (1, 1))
|1,0,j1=1,j2=1>
>>> j, m, j1, j2 = symbols(j m j1 j2)
>>> JzKetCoupled(j, m, (j1, j2))
|j,m,j1=j1,j2=j2>
Dening coupled spin states for more than 2 coupled spaces with various coupling parameters:
>>> JzKetCoupled(2, 1, (1, 1, 1))
|2,1,j1=1,j2=1,j3=1,j(1,2)=2>
>>> JzKetCoupled(2, 1, (1, 1, 1), ((1,2,2),(1,3,2)) )
|2,1,j1=1,j2=1,j3=1,j(1,2)=2>
>>> JzKetCoupled(2, 1, (1, 1, 1), ((2,3,1),(1,2,2)) )
|2,1,j1=1,j2=1,j3=1,j(2,3)=1>
Rewriting the JzKetCoupled in terms of eigenkets of the Jx operator: Note: that the
resulting eigenstates are JxKetCoupled
>>> JzKetCoupled(1,1,(1,1)).rewrite(Jx)
|1,-1,j1=1,j2=1>/2 - sqrt(2)*|1,0,j1=1,j2=1>/2 + |1,1,j1=1,j2=1>/2
The rewrite method can be used to convert a coupled state to an uncoupled state. This
is done by passing coupled=False to the rewrite function:
>>> JzKetCoupled(1, 0, (1, 1)).rewrite(Jz, coupled=False)
-sqrt(2)*|1,-1>x|1,1>/2 + sqrt(2)*|1,1>x|1,-1>/2
Get the vector representation of a state in terms of the basis elements of the Jx operator:
1778
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> from sympy.physics.quantum.represent import represent
>>> from sympy.physics.quantum.spin import Jx
>>> from sympy import S
>>> represent(JzKetCoupled(1,-1,(S(1)/2,S(1)/2)), basis=Jx)
Matrix([
[
0],
[
1/2],
[sqrt(2)/2],
[
1/2]])
class sympy.physics.quantum.spin.JzBraCoupled
Coupled eigenbra of Jz.
See the JzKetCoupled for the usage of coupled spin eigenstates.
See Also:
JzKetCoupled (page 1777) Usage of coupled spin states
sympy.physics.quantum.spin.couple(expr, jcoupling list=None)
Couple a tensor product of spin states
This function can be used to couple an uncoupled tensor product of spin states. All of the
eigenstates to be coupled must be of the same class. It will return a linear combination
of eigenstates that are subclasses of CoupledSpinState determined by Clebsch-Gordan
angular momentum coupling coecients.
Parameters expr : Expr
An expression involving TensorProducts of spin states to be coupled.
Each state must be a subclass of SpinState and they all must be the
same class.
jcoupling list : list or tuple
Elements of this list are sub-lists of length 2 specifying the order of
the coupling of the spin spaces. The length of this must be N-1, where
N is the number of states in the tensor product to be coupled. The
elements of this sublist are the same as the rst two elements of each
sublist in the jcoupling parameter dened for JzKetCoupled. If this
parameter is not specied, the default value is taken, which couples
the rst and second product basis spaces, then couples this new coupled space to the third product space, etc
Examples
Couple a tensor product of numerical states for two spaces:
>>> from sympy.physics.quantum.spin import JzKet, couple
>>> from sympy.physics.quantum.tensorproduct import TensorProduct
>>> couple(TensorProduct(JzKet(1,0), JzKet(1,1)))
-sqrt(2)*|1,1,j1=1,j2=1>/2 + sqrt(2)*|2,1,j1=1,j2=1>/2
Numerical coupling of three spaces using the default coupling method, i.e. rst and
second spaces couple, then this couples to the third space:
>>> couple(TensorProduct(JzKet(1,1), JzKet(1,1), JzKet(1,0)))
sqrt(6)*|2,2,j1=1,j2=1,j3=1,j(1,2)=2>/3 + sqrt(3)*|3,2,j1=1,j2=1,j3=1,j(1,2)=2>/3
5.36. Physics Module
1779
SymPy Documentation, Release 0.7.6
Perform this same coupling, but we dene the coupling to rst couple the rst and third
spaces:
>>> couple(TensorProduct(JzKet(1,1), JzKet(1,1), JzKet(1,0)), ((1,3),(1,2)) )
sqrt(2)*|2,2,j1=1,j2=1,j3=1,j(1,3)=1>/2 - sqrt(6)*|2,2,j1=1,j2=1,j3=1,j(1,3)=2>/6 + sqrt(3)*|3,2,
Couple a tensor product of symbolic states:
>>> from sympy import symbols
>>> j1,m1,j2,m2 = symbols(j1 m1 j2 m2)
>>> couple(TensorProduct(JzKet(j1,m1), JzKet(j2,m2)))
Sum(CG(j1, m1, j2, m2, j, m1 + m2)*|j,m1 + m2,j1=j1,j2=j2>, (j, m1 + m2, j1 + j2))
sympy.physics.quantum.spin.uncouple(expr, jn=None, jcoupling list=None)
Uncouple a coupled spin state
Gives the uncoupled representation of a coupled spin state. Arguments must be either
a spin state that is a subclass of CoupledSpinState or a spin state that is a subclass of
SpinState and an array giving the j values of the spaces that are to be coupled
Parameters expr : Expr
The expression containing states that are to be coupled. If the states
are a subclass of SpinState, the jn and jcoupling parameters must
be dened. If the states are a subclass of CoupledSpinState, jn and
jcoupling will be taken from the state.
jn : list or tuple
The list of the j-values that are coupled. If state is a CoupledSpinState, this parameter is ignored. This must be dened if state is not
a subclass of CoupledSpinState. The syntax of this parameter is the
same as the jn parameter of JzKetCoupled.
jcoupling list : list or tuple
The list dening how the j-values are coupled together. If state is a
CoupledSpinState, this parameter is ignored. This must be dened
if state is not a subclass of CoupledSpinState. The syntax of this parameter is the same as the jcoupling parameter of JzKetCoupled.
Examples
Uncouple a numerical state using a CoupledSpinState state:
>>> from sympy.physics.quantum.spin import JzKetCoupled, uncouple
>>> from sympy import S
>>> uncouple(JzKetCoupled(1, 0, (S(1)/2, S(1)/2)))
sqrt(2)*|1/2,-1/2>x|1/2,1/2>/2 + sqrt(2)*|1/2,1/2>x|1/2,-1/2>/2
Perform the same calculation using a SpinState state:
>>> from sympy.physics.quantum.spin import JzKet
>>> uncouple(JzKet(1, 0), (S(1)/2, S(1)/2))
sqrt(2)*|1/2,-1/2>x|1/2,1/2>/2 + sqrt(2)*|1/2,1/2>x|1/2,-1/2>/2
Uncouple a numerical state of three coupled spaces using a CoupledSpinState state:
>>> uncouple(JzKetCoupled(1, 1, (1, 1, 1), ((1,3,1),(1,2,1)) ))
|1,-1>x|1,1>x|1,1>/2 - |1,0>x|1,0>x|1,1>/2 + |1,1>x|1,0>x|1,0>/2 - |1,1>x|1,1>x|1,-1>/2
1780
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Perform the same calculation using a SpinState state:
>>> uncouple(JzKet(1, 1), (1, 1, 1), ((1,3,1),(1,2,1)) )
|1,-1>x|1,1>x|1,1>/2 - |1,0>x|1,0>x|1,1>/2 + |1,1>x|1,0>x|1,0>/2 - |1,1>x|1,1>x|1,-1>/2
Uncouple a symbolic state using a CoupledSpinState state:
>>> from sympy import symbols
>>> j,m,j1,j2 = symbols(j m j1 j2)
>>> uncouple(JzKetCoupled(j, m, (j1, j2)))
Sum(CG(j1, m1, j2, m2, j, m)*|j1,m1>x|j2,m2>, (m1, -j1, j1), (m2, -j2, j2))
Perform the same calculation using a SpinState state
>>> uncouple(JzKet(j, m), (j1, j2))
Sum(CG(j1, m1, j2, m2, j, m)*|j1,m1>x|j2,m2>, (m1, -j1, j1), (m2, -j2, j2))
State Dirac notation for states.
class sympy.physics.quantum.state.KetBase
Base class for Kets.
This class denes the dual property and the brackets for printing. This is an abstract
base class and you should not instantiate it directly, instead use Ket.
class sympy.physics.quantum.state.BraBase
Base class for Bras.
This class denes the dual property and the brackets for printing. This is an abstract
base class and you should not instantiate it directly, instead use Bra.
class sympy.physics.quantum.state.StateBase
Abstract base class for general abstract states in quantum mechanics.
All other state classes dened will need to inherit from this class. It carries the basic
structure for all other states such as dual, eval adjoint and label.
This is an abstract base class and you should not instantiate it directly, instead use State.
dual
Return the dual state of this one.
classmethod dual class()
Return the class used to construt the dual.
operators
Return the operator(s) that this state is an eigenstate of
class sympy.physics.quantum.state.State
General abstract quantum state used as a base class for Ket and Bra.
class sympy.physics.quantum.state.Ket
A general time-independent Ket in quantum mechanics.
Inherits from State and KetBase. This class should be used as the base class for all
physical, time-independent Kets in a system. This class and its subclasses will be the
main classes that users will use for expressing Kets in Dirac notation [R342] (page 1916).
Parameters args : tuple
5.36. Physics Module
1781
SymPy Documentation, Release 0.7.6
The list of numbers or parameters that uniquely specify the ket.
This will usually be its symbol or its quantum numbers. For timedependent state, this will include the time.
References
[R342] (page 1916)
Examples
Create a simple Ket and looking at its properties:
>>> from sympy.physics.quantum import Ket, Bra
>>> from sympy import symbols, I
>>> k = Ket(psi)
>>> k
|psi>
>>> k.hilbert_space
H
>>> k.is_commutative
False
>>> k.label
(psi,)
Kets know about their associated bra:
>>> k.dual
<psi|
>>> k.dual_class()
<class sympy.physics.quantum.state.Bra>
Take a linear combination of two kets:
>>> k0 = Ket(0)
>>> k1 = Ket(1)
>>> 2*I*k0 - 4*k1
2*I*|0> - 4*|1>
Compound labels are passed as tuples:
>>> n, m = symbols(n,m)
>>> k = Ket(n,m)
>>> k
|nm>
class sympy.physics.quantum.state.Bra
A general time-independent Bra in quantum mechanics.
Inherits from State and BraBase. A Bra is the dual of a Ket [R343] (page 1916). This
class and its subclasses will be the main classes that users will use for expressing Bras
in Dirac notation.
Parameters args : tuple
The list of numbers or parameters that uniquely specify the ket.
This will usually be its symbol or its quantum numbers. For timedependent state, this will include the time.
1782
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
References
[R343] (page 1916)
Examples
Create a simple Bra and look at its properties:
>>> from sympy.physics.quantum import Ket, Bra
>>> from sympy import symbols, I
>>> b = Bra(psi)
>>> b
<psi|
>>> b.hilbert_space
H
>>> b.is_commutative
False
Bras know about their dual Kets:
>>> b.dual
|psi>
>>> b.dual_class()
<class sympy.physics.quantum.state.Ket>
Like Kets, Bras can have compound labels and be manipulated in a similar manner:
>>> n, m = symbols(n,m)
>>> b = Bra(n,m) - I*Bra(m,n)
>>> b
-I*<mn| + <nm|
Symbols in a Bra can be substituted using .subs:
>>> b.subs(n,m)
<mm| - I*<mm|
class sympy.physics.quantum.state.TimeDepState
Base class for a general time-dependent quantum state.
This class is used as a base class for any time-dependent state. The main dierence
between this class and the time-independent state is that this class takes a second argument that is the time in addition to the usual label argument.
Parameters args : tuple
The list of numbers or parameters that uniquely specify the ket.
This will usually be its symbol or its quantum numbers. For timedependent state, this will include the time as the nal argument.
label
The label of the state.
time
The time of the state.
class sympy.physics.quantum.state.TimeDepBra
General time-dependent Bra in quantum mechanics.
This inherits from TimeDepState and BraBase and is the main class that should be used
for Bras that vary with time. Its dual is a TimeDepBra.
5.36. Physics Module
1783
SymPy Documentation, Release 0.7.6
Parameters args : tuple
The list of numbers or parameters that uniquely specify the ket.
This will usually be its symbol or its quantum numbers. For timedependent state, this will include the time as the nal argument.
Examples
>>> from sympy.physics.quantum import TimeDepBra
>>> from sympy import symbols, I
>>> b = TimeDepBra(psi, t)
>>> b
<psi;t|
>>> b.time
t
>>> b.label
(psi,)
>>> b.hilbert_space
H
>>> b.dual
|psi;t>
class sympy.physics.quantum.state.TimeDepKet
General time-dependent Ket in quantum mechanics.
This inherits from TimeDepState and KetBase and is the main class that should be used
for Kets that vary with time. Its dual is a TimeDepBra.
Parameters args : tuple
The list of numbers or parameters that uniquely specify the ket.
This will usually be its symbol or its quantum numbers. For timedependent state, this will include the time as the nal argument.
Examples
Create a TimeDepKet and look at its attributes:
>>> from sympy.physics.quantum import TimeDepKet
>>> k = TimeDepKet(psi, t)
>>> k
|psi;t>
>>> k.time
t
>>> k.label
(psi,)
>>> k.hilbert_space
H
TimeDepKets know about their dual bra:
>>> k.dual
<psi;t|
>>> k.dual_class()
<class sympy.physics.quantum.state.TimeDepBra>
class sympy.physics.quantum.state.Wavefunction
Class for representations in continuous bases
1784
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
This class takes an expression and coordinates in its constructor. It can be used to easily
calculate normalizations and probabilities.
Parameters expr : Expr
The expression representing the functional form of the w.f.
coords : Symbol or tuple
The coordinates to be integrated over, and their bounds
Examples
Particle in a box, specifying bounds in the more primitive way of using Piecewise:
>>> from sympy import Symbol, Piecewise, pi, N
>>> from sympy.functions import sqrt, sin
>>> from sympy.physics.quantum.state import Wavefunction
>>> x = Symbol(x, real=True)
>>> n = 1
>>> L = 1
>>> g = Piecewise((0, x < 0), (0, x > L), (sqrt(2//L)*sin(n*pi*x/L), True))
>>> f = Wavefunction(g, x)
>>> f.norm
1
>>> f.is_normalized
True
>>> p = f.prob()
>>> p(0)
0
>>> p(L)
0
>>> p(0.5)
2
>>> p(0.85*L)
2*sin(0.85*pi)**2
>>> N(p(0.85*L))
0.412214747707527
Additionally, you can specify the bounds of the function and the indices in a more compact
way:
>>> from sympy import symbols, pi, diff
>>> from sympy.functions import sqrt, sin
>>> from sympy.physics.quantum.state import Wavefunction
>>> x, L = symbols(x,L, positive=True)
>>> n = symbols(n, integer=True, positive=True)
>>> g = sqrt(2/L)*sin(n*pi*x/L)
>>> f = Wavefunction(g, (x, 0, L))
>>> f.norm
1
>>> f(L+1)
0
>>> f(L-1)
sqrt(2)*sin(pi*n*(L - 1)/L)/sqrt(L)
>>> f(-1)
0
>>> f(0.85)
sqrt(2)*sin(0.85*pi*n/L)/sqrt(L)
5.36. Physics Module
1785
SymPy Documentation, Release 0.7.6
>>> f(0.85, n=1, L=1)
sqrt(2)*sin(0.85*pi)
>>> f.is_commutative
False
All arguments are automatically sympied, so you can dene the variables as strings
rather than symbols:
>>> expr = x**2
>>> f = Wavefunction(expr, x)
>>> type(f.variables[0])
<class sympy.core.symbol.Symbol>
Derivatives of Wavefunctions will return Wavefunctions:
>>> diff(f, x)
Wavefunction(2*x, x)
expr
Return the expression which is the functional form of the Wavefunction
Examples
>>> from sympy.physics.quantum.state import Wavefunction
>>> from sympy import symbols
>>> x, y = symbols(x, y)
>>> f = Wavefunction(x**2, x)
>>> f.expr
x**2
is commutative
Override Functions is commutative so that order is preserved in represented expressions
is normalized
Returns true if the Wavefunction is properly normalized
Examples
>>> from sympy import symbols, pi
>>> from sympy.functions import sqrt, sin
>>> from sympy.physics.quantum.state import Wavefunction
>>> x, L = symbols(x,L, positive=True)
>>> n = symbols(n, integer=True, positive=True)
>>> g = sqrt(2/L)*sin(n*pi*x/L)
>>> f = Wavefunction(g, (x, 0, L))
>>> f.is_normalized
True
limits
Return the limits of the coordinates which the w.f. depends on If no limits are specied, defaults to (-oo, oo).
1786
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>>
>>>
>>>
>>>
>>>
{x:
>>>
>>>
{x:
>>>
>>>
{x:
from sympy.physics.quantum.state import Wavefunction
from sympy import symbols
x, y = symbols(x, y)
f = Wavefunction(x**2, (x, 0, 1))
f.limits
(0, 1)}
f = Wavefunction(x**2, x)
f.limits
(-oo, oo)}
f = Wavefunction(x**2 + y**2, x, (y, -1, 2))
f.limits
(-oo, oo), y: (-1, 2)}
norm
Return the normalization of the specied functional form.
This function integrates over the coordinates of the Wavefunction, with the bounds
specied.
Examples
>>> from sympy import symbols, pi
>>> from sympy.functions import sqrt, sin
>>> from sympy.physics.quantum.state import Wavefunction
>>> x, L = symbols(x,L, positive=True)
>>> n = symbols(n, integer=True, positive=True)
>>> g = sqrt(2/L)*sin(n*pi*x/L)
>>> f = Wavefunction(g, (x, 0, L))
>>> f.norm
1
>>> g = sin(n*pi*x/L)
>>> f = Wavefunction(g, (x, 0, L))
>>> f.norm
sqrt(2)*sqrt(L)/2
normalize()
Return a normalized version of the Wavefunction
Examples
>>> from sympy import symbols, pi
>>> from sympy.functions import sqrt, sin
>>> from sympy.physics.quantum.state import Wavefunction
>>> x = symbols(x, real=True)
>>> L = symbols(L, positive=True)
>>> n = symbols(n, integer=True, positive=True)
>>> g = sin(n*pi*x/L)
>>> f = Wavefunction(g, (x, 0, L))
>>> f.normalize()
Wavefunction(sqrt(2)*sin(pi*n*x/L)/sqrt(L), (x, 0, L))
prob()
Return the absolute magnitude of the w.f., |(x)|2
5.36. Physics Module
1787
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import symbols, pi
>>> from sympy.functions import sqrt, sin
>>> from sympy.physics.quantum.state import Wavefunction
>>> x, L = symbols(x,L, real=True)
>>> n = symbols(n, integer=True)
>>> g = sin(n*pi*x/L)
>>> f = Wavefunction(g, (x, 0, L))
>>> f.prob()
Wavefunction(sin(pi*n*x/L)**2, x)
variables
Return the coordinates which the wavefunction depends on
Examples
>>> from sympy.physics.quantum.state import Wavefunction
>>> from sympy import symbols
>>> x,y = symbols(x,y)
>>> f = Wavefunction(x*y, x, y)
>>> f.variables
(x, y)
>>> g = Wavefunction(x*y, x)
>>> g.variables
(x,)
Quantum Computation
Circuit Plot Matplotlib based plotting of quantum circuits.
Todo:
Optimize printing of large circuits.
Get this to work with single gates.
Do a better job checking the form of circuits to make sure it is a Mul of Gates.
Get multi-target gates plotting.
Get initial and nal states to plot.
Get measurements to plot. Might need to rethink measurement as a gate issue.
Get scale and gsize to be handled in a better way.
Write some tests/examples!
sympy.physics.quantum.circuitplot.labeller(n, symbol=q)
Autogenerate labels for wires of quantum circuits.
Parameters n : int
number of qubits in the circuit
symbol : string
A character string to precede all gate labels. E.g. q 0, q 1, etc.
1788
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> from sympy.physics.quantum.circuitplot import labeller :
>>> labeller(2) :
[q 1, q 0] :
>>> labeller(3,j) :
[j 2, j 1, j 0] :
class sympy.physics.quantum.circuitplot.Mz
Mock-up of a z measurement gate.
This is in circuitplot rather than gate.py because its not a real gate, it just draws one.
class sympy.physics.quantum.circuitplot.Mx
Mock-up of an x measurement gate.
This is in circuitplot rather than gate.py because its not a real gate, it just draws one.
sympy.physics.quantum.circuitplot.CreateCGate(name, latexname=None)
Use a lexical closure to make a controlled gate.
Gates
An implementation of gates that act on qubits.
Gates are unitary operators that act on the space of qubits.
Medium Term Todo:
Optimize Gate. apply operators Qubit to remove the creation of many intermediate Qubit
objects.
Add commutation relationships to all operators and use this in gate sort.
Fix gate sort and gate simp.
Get multi-target UGates plotting properly.
Get UGate to work with either sympy/numpy matrices and output either format. This
should also use the matrix slots.
class sympy.physics.quantum.gate.Gate
Non-controlled unitary gate operator that acts on qubits.
This is a general abstract gate that needs to be subclassed to do anything useful.
Parameters label : tuple, int
A list of the target qubits (as ints) that the gate will apply to.
get target matrix(format=sympy)
The matrix rep. of the target part of the gate.
Parameters format : str
The format string (sympy,numpy, etc.)
min qubits
The minimum number of qubits this gate needs to act on.
nqubits
The total number of qubits this gate acts on.
For controlled gate subclasses this includes both target and control qubits, so that,
for examples the CNOT gate acts on 2 qubits.
targets
A tuple of target qubits.
5.36. Physics Module
1789
SymPy Documentation, Release 0.7.6
class sympy.physics.quantum.gate.CGate
A general unitary gate with control qubits.
A general control gate applies a target gate to a set of targets if all of the control qubits
have a particular values (set by CGate.control value).
Parameters label : tuple
The label in this case has the form (controls, gate), where controls is
a tuple/list of control qubits (as ints) and gate is a Gate instance that
is the target operator.
controls
A tuple of control qubits.
decompose(**options)
Decompose the controlled gate into CNOT and single qubits gates.
eval controls(qubit)
Return True/False to indicate if the controls are satised.
gate
The non-controlled gate that will be applied to the targets.
min qubits
The minimum number of qubits this gate needs to act on.
nqubits
The total number of qubits this gate acts on.
For controlled gate subclasses this includes both target and control qubits, so that,
for examples the CNOT gate acts on 2 qubits.
plot gate(circ plot, gate idx)
Plot the controlled gate. If simplify cgate is true, simplify C-X and C-Z gates into
their more familiar forms.
targets
A tuple of target qubits.
class sympy.physics.quantum.gate.UGate
General gate specied by a set of targets and a target matrix.
Parameters label : tuple
A tuple of the form (targets, U), where targets is a tuple of the target
qubits and U is a unitary matrix with dimension of len(targets).
get target matrix(format=sympy)
The matrix rep. of the target part of the gate.
Parameters format : str
The format string (sympy,numpy, etc.)
targets
A tuple of target qubits.
class sympy.physics.quantum.gate.OneQubitGate
A single qubit unitary gate base class.
class sympy.physics.quantum.gate.TwoQubitGate
A two qubit unitary gate base class.
class sympy.physics.quantum.gate.IdentityGate
The single qubit identity gate.
1790
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Parameters target : int
The target qubit this gate will apply to.
class sympy.physics.quantum.gate.HadamardGate
The single qubit Hadamard gate.
Parameters target : int
The target qubit this gate will apply to.
Examples
>>> from sympy import sqrt
>>> from sympy.physics.quantum.qubit import Qubit
>>> from sympy.physics.quantum.gate import HadamardGate
>>> from sympy.physics.quantum.qapply import qapply
>>> qapply(HadamardGate(0)*Qubit(1))
sqrt(2)*|0>/2 - sqrt(2)*|1>/2
>>> # Hadamard on bell state, applied on 2 qubits.
>>> psi = 1/sqrt(2)*(Qubit(00)+Qubit(11))
>>> qapply(HadamardGate(0)*HadamardGate(1)*psi)
sqrt(2)*|00>/2 + sqrt(2)*|11>/2
class sympy.physics.quantum.gate.XGate
The single qubit X, or NOT, gate.
Parameters target : int
The target qubit this gate will apply to.
class sympy.physics.quantum.gate.YGate
The single qubit Y gate.
Parameters target : int
The target qubit this gate will apply to.
class sympy.physics.quantum.gate.ZGate
The single qubit Z gate.
Parameters target : int
The target qubit this gate will apply to.
class sympy.physics.quantum.gate.TGate
The single qubit pi/8 gate.
This gate rotates the phase of the state by pi/4 if the state is |1> and does nothing if the
state is |0>.
Parameters target : int
The target qubit this gate will apply to.
class sympy.physics.quantum.gate.PhaseGate
The single qubit phase, or S, gate.
This gate rotates the phase of the state by pi/2 if the state is |1> and does nothing if the
state is |0>.
Parameters target : int
The target qubit this gate will apply to.
5.36. Physics Module
1791
SymPy Documentation, Release 0.7.6
class sympy.physics.quantum.gate.SwapGate
Two qubit SWAP gate.
This gate swap the values of the two qubits.
Parameters label : tuple
A tuple of the form (target1, target2).
decompose(**options)
Decompose the SWAP gate into CNOT gates.
class sympy.physics.quantum.gate.CNotGate
Two qubit controlled-NOT.
This gate performs the NOT or X gate on the target qubit if the control qubits all have
the value 1.
Parameters label : tuple
A tuple of the form (control, target).
Examples
>>>
>>>
>>>
>>>
>>>
|11>
from sympy.physics.quantum.gate import CNOT
from sympy.physics.quantum.qapply import qapply
from sympy.physics.quantum.qubit import Qubit
c = CNOT(1,0)
qapply(c*Qubit(10)) # note that qubits are indexed from right to left
controls
A tuple of control qubits.
gate
The non-controlled gate that will be applied to the targets.
min qubits
The minimum number of qubits this gate needs to act on.
targets
A tuple of target qubits.
sympy.physics.quantum.gate.CNOT
alias of CNotGate (page 1792)
sympy.physics.quantum.gate.SWAP
alias of SwapGate (page 1791)
sympy.physics.quantum.gate.H
alias of HadamardGate (page 1791)
sympy.physics.quantum.gate.X
alias of XGate (page 1791)
sympy.physics.quantum.gate.Y
alias of YGate (page 1791)
sympy.physics.quantum.gate.Z
alias of ZGate (page 1791)
sympy.physics.quantum.gate.T
alias of TGate (page 1791)
1792
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
sympy.physics.quantum.gate.S
alias of PhaseGate (page 1791)
sympy.physics.quantum.gate.Phase
alias of PhaseGate (page 1791)
sympy.physics.quantum.gate.normalized(normalize)
Set ag controlling normalization of Hadamard gates by 1/sqrt(2).
This is a global setting that can be used to simplify the look of various expressions, by
leaving o the leading 1/sqrt(2) of the Hadamard gate.
Parameters normalize : bool
Should the Hadamard gate include the 1/sqrt(2) normalization factor?
When True, the Hadamard gate will have the 1/sqrt(2). When False,
the Hadamard gate will not have this factor.
sympy.physics.quantum.gate.gate sort(circuit)
Sorts the gates while keeping track of commutation relations
This function uses a bubble sort to rearrange the order of gate application. Keeps track
of Quantum computations special commutation relations (e.g. things that apply to the
same Qubit do not commute with each other)
circuit is the Mul of gates that are to be sorted.
sympy.physics.quantum.gate.gate simp(circuit)
Simplies gates symbolically
It rst sorts gates using gate sort. It then applies basic simplication rules to the circuit,
e.g., XGate**2 = Identity
sympy.physics.quantum.gate.random circuit(ngates, nqubits, gate space=(<class
sympy.physics.quantum.gate.XGate>,
<class sympy.physics.quantum.gate.YGate>,
<class sympy.physics.quantum.gate.ZGate>,
<class sympy.physics.quantum.gate.PhaseGate>,
<class sympy.physics.quantum.gate.TGate>,
<class sympy.physics.quantum.gate.HadamardGate>,
<class sympy.physics.quantum.gate.CNotGate>,
<class sympy.physics.quantum.gate.SwapGate>))
Return a random circuit of ngates and nqubits.
This uses an equally weighted sample of (X, Y, Z, S, T, H, CNOT, SWAP) gates.
Parameters ngates : int
The number of gates in the circuit.
nqubits : int
The number of qubits in the circuit.
gate space : tuple
A tuple of the gate classes that will be used in the circuit. Repeating
gate classes multiple times in this tuple will increase the frequency
they appear in the random circuit.
class sympy.physics.quantum.gate.CGateS
Version of CGate that allows gate simplications. I.e. cnot looks like an oplus, cphase
has dots, etc.
5.36. Physics Module
1793
SymPy Documentation, Release 0.7.6
Grovers Algorithm
Grovers algorithm and helper functions.
Todo:
W gate construction (or perhaps -W gate based on Mermins book)
Generalize the algorithm for an unknown function that returns 1 on multiple qubit states,
not just one.
Implement represent ZGate in OracleGate
class sympy.physics.quantum.grover.OracleGate
A black box gate.
The gate marks the desired qubits of an unknown function by ipping the sign of the
qubits. The unknown function returns true when it nds its desired qubits and false
otherwise.
Parameters qubits : int
Number of qubits.
oracle : callable
A callable function that returns a boolean on a computational basis.
Examples
Apply an Oracle gate that ips the sign of |2> on dierent qubits:
>>>
>>>
>>>
>>>
>>>
>>>
-|2>
>>>
|3>
from sympy.physics.quantum.qubit import IntQubit
from sympy.physics.quantum.qapply import qapply
from sympy.physics.quantum.grover import OracleGate
f = lambda qubits: qubits == IntQubit(2)
v = OracleGate(2, f)
qapply(v*IntQubit(2))
qapply(v*IntQubit(3))
search function
The unknown function that helps nd the sought after qubits.
targets
A tuple of target qubits.
class sympy.physics.quantum.grover.WGate
General n qubit W Gate in Grovers algorithm.
The gate performs the operation 2|phi><phi| - 1 on some qubits. |phi> = (tensor
product of n Hadamards)*(|0> with n qubits)
Parameters nqubits : int
The number of qubits to operate on
sympy.physics.quantum.grover.superposition basis(nqubits)
Creates an equal superposition of the computational basis.
Parameters nqubits : int
The number of qubits.
Returns state : Qubit
An equal superposition of the computational basis with nqubits.
1794
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
Create an equal superposition of 2 qubits:
>>> from sympy.physics.quantum.grover import superposition_basis
>>> superposition_basis(2)
|0>/2 + |1>/2 + |2>/2 + |3>/2
sympy.physics.quantum.grover.grover iteration(qstate, oracle)
Applies one application of the Oracle and W Gate, WV.
Parameters qstate : Qubit
A superposition of qubits.
oracle : OracleGate
The black box operator that ips the sign of the desired basis qubits.
Returns Qubit : The qubits after applying the Oracle and W gate.
Examples
Perform one iteration of grovers algorithm to see a phase change:
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
|2>
from sympy.physics.quantum.qapply import qapply
from sympy.physics.quantum.qubit import IntQubit
from sympy.physics.quantum.grover import OracleGate
from sympy.physics.quantum.grover import superposition_basis
from sympy.physics.quantum.grover import grover_iteration
numqubits = 2
basis_states = superposition_basis(numqubits)
f = lambda qubits: qubits == IntQubit(2)
v = OracleGate(numqubits, f)
qapply(grover_iteration(basis_states, v))
sympy.physics.quantum.grover.apply grover(oracle, nqubits, iterations=None)
Applies grovers algorithm.
Parameters oracle : callable
The unknown callable function that returns true when applied to the
desired qubits and false otherwise.
Returns state : Expr
The resulting state after Grovers algorithm has been iterated.
Examples
Apply grovers algorithm to an even superposition of 2 qubits:
>>>
>>>
>>>
>>>
>>>
|2>
from sympy.physics.quantum.qapply import qapply
from sympy.physics.quantum.qubit import IntQubit
from sympy.physics.quantum.grover import apply_grover
f = lambda qubits: qubits == IntQubit(2)
qapply(apply_grover(f, 2))
5.36. Physics Module
1795
SymPy Documentation, Release 0.7.6
QFT
An implementation of qubits and gates acting on them.
Todo:
Update docstrings.
Update tests.
Implement apply using decompose.
Implement represent using decompose or something smarter. For this to work we rst
have to implement represent for SWAP.
Decide if we want upper index to be inclusive in the constructor.
Fix the printing of Rk gates in plotting.
class sympy.physics.quantum.qft.QFT
The forward quantum Fourier transform.
decompose()
Decomposes QFT into elementary gates.
class sympy.physics.quantum.qft.IQFT
The inverse quantum Fourier transform.
decompose()
Decomposes IQFT into elementary gates.
class sympy.physics.quantum.qft.RkGate
This is the R k gate of the QTF.
sympy.physics.quantum.qft.Rk
alias of RkGate (page 1796)
Qubit
Qubits for quantum computing.
Todo: * Finish implementing measurement logic. This should include POVM. * Update docstrings. * Update tests.
class sympy.physics.quantum.qubit.Qubit
A multi-qubit ket in the computational (z) basis.
We use the normal convention that the least signicant qubit is on the right, so |00001>
has a 1 in the least signicant qubit.
Parameters values : list, str
The qubit values as a list of ints ([0,0,0,1,1,]) or a string (011).
Examples
Create a qubit in a couple of dierent ways and look at their attributes:
>>> from sympy.physics.quantum.qubit import Qubit
>>> Qubit(0,0,0)
|000>
>>> q = Qubit(0101)
>>> q
|0101>
1796
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>>
4
>>>
4
>>>
4
>>>
(0,
q.nqubits
len(q)
q.dimension
q.qubit_values
1, 0, 1)
We can ip the value of an individual qubit:
>>> q.flip(1)
|0111>
We can take the dagger of a Qubit to get a bra:
>>> from sympy.physics.quantum.dagger import Dagger
>>> Dagger(q)
<0101|
>>> type(Dagger(q))
<class sympy.physics.quantum.qubit.QubitBra>
Inner products work as expected:
>>> ip = Dagger(q)*q
>>> ip
<0101|0101>
>>> ip.doit()
1
class sympy.physics.quantum.qubit.QubitBra
A multi-qubit bra in the computational (z) basis.
We use the normal convention that the least signicant qubit is on the right, so |00001>
has a 1 in the least signicant qubit.
Parameters values : list, str
The qubit values as a list of ints ([0,0,0,1,1,]) or a string (011).
See Also:
Qubit (page 1796) Examples using qubits
class sympy.physics.quantum.qubit.IntQubit
A qubit ket that store integers as binary numbers in qubit values.
The dierences between this class and Qubit are:
The form of the constructor.
The qubit values are printed as their corresponding integer, rather than the raw
qubit values. The internal storage format of the qubit values in the same as Qubit.
Parameters values : int, tuple
If a single argument, the integer we want to represent in the qubit
values. This integer will be represented using the fewest possible
number of qubits. If a pair of integers, the rst integer gives the
integer to represent in binary form and the second integer gives the
number of qubits to use.
5.36. Physics Module
1797
SymPy Documentation, Release 0.7.6
Examples
Create a qubit for the integer 5:
>>>
>>>
>>>
>>>
|5>
from sympy.physics.quantum.qubit import IntQubit
from sympy.physics.quantum.qubit import Qubit
q = IntQubit(5)
q
We can also create an IntQubit by passing a Qubit instance.
>>>
>>>
|5>
>>>
5
>>>
3
>>>
(1,
q = IntQubit(Qubit(101))
q
q.as_int()
q.nqubits
q.qubit_values
0, 1)
We can go back to the regular qubit form.
>>> Qubit(q)
|101>
class sympy.physics.quantum.qubit.IntQubitBra
A qubit bra that store integers as binary numbers in qubit values.
sympy.physics.quantum.qubit.qubit to matrix(qubit, format=sympy)
Coverts an Add/Mul of Qubit objects into its matrix representation
This function is the inverse of matrix to qubit and is a shorthand for represent(qubit).
sympy.physics.quantum.qubit.matrix to qubit(matrix)
Convert from the matrix repr. to a sum of Qubit objects.
Parameters matrix : Matrix, numpy.matrix, scipy.sparse
The matrix to build the Qubit representation of. This works with
sympy matrices, numpy matrices and scipy.sparse sparse matrices.
Examples
Represent a state and then go back to its qubit form:
>>>
>>>
>>>
>>>
>>>
|01>
from sympy.physics.quantum.qubit import matrix_to_qubit, Qubit
from sympy.physics.quantum.gate import Z
from sympy.physics.quantum.represent import represent
q = Qubit(01)
matrix_to_qubit(represent(q))
sympy.physics.quantum.qubit.matrix to density(mat)
Works by nding the eigenvectors and eigenvalues of the matrix. We know we can decompose rho by doing: sum(EigenVal*|Eigenvect><Eigenvect|)
1798
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
sympy.physics.quantum.qubit.measure all(qubit,
format=sympy,
ize=True)
Perform an ensemble measurement of all qubits.
normal-
Parameters qubit : Qubit, Add
The qubit to measure. This can be any Qubit or a linear combination
of them.
format : str
The format of the intermediate matrices to use. Possible values
are (sympy,numpy,scipy.sparse). Currently only sympy is implemented.
Returns result : list
A list that consists of primitive states and their probabilities.
Examples
>>> from sympy.physics.quantum.qubit import Qubit, measure_all
>>> from sympy.physics.quantum.gate import H, X, Y, Z
>>> from sympy.physics.quantum.qapply import qapply
>>> c = H(0)*H(1)*Qubit(00)
>>> c
H(0)*H(1)*|00>
>>> q = qapply(c)
>>> measure_all(q)
[(|00>, 1/4), (|01>, 1/4), (|10>, 1/4), (|11>, 1/4)]
sympy.physics.quantum.qubit.measure partial(qubit, bits, format=sympy, normalize=True)
Perform a partial ensemble measure on the specifed qubits.
Parameters qubits : Qubit
The qubit to measure. This can be any Qubit or a linear combination
of them.
bits : tuple
The qubits to measure.
format : str
The format of the intermediate matrices to use. Possible values
are (sympy,numpy,scipy.sparse). Currently only sympy is implemented.
Returns result : list
A list that consists of primitive states and their probabilities.
Examples
>>> from sympy.physics.quantum.qubit import Qubit, measure_partial
>>> from sympy.physics.quantum.gate import H, X, Y, Z
>>> from sympy.physics.quantum.qapply import qapply
5.36. Physics Module
1799
SymPy Documentation, Release 0.7.6
>>> c = H(0)*H(1)*Qubit(00)
>>> c
H(0)*H(1)*|00>
>>> q = qapply(c)
>>> measure_partial(q, (0,))
[(sqrt(2)*|00>/2 + sqrt(2)*|10>/2, 1/2), (sqrt(2)*|01>/2 + sqrt(2)*|11>/2, 1/2)]
sympy.physics.quantum.qubit.measure partial oneshot(qubit,
bits,
mat=sympy)
Perform a partial oneshot measurement on the specied qubits.
for-
A oneshot measurement is equivalent to performing a measurement on a quantum system. This type of measurement does not return the probabilities like an ensemble measurement does, but rather returns one of the possible resulting states. The exact state
that is returned is determined by picking a state randomly according to the ensemble
probabilities.
Parameters qubits : Qubit
The qubit to measure. This can be any Qubit or a linear combination
of them.
bits : tuple
The qubits to measure.
format : str
The format of the intermediate matrices to use. Possible values
are (sympy,numpy,scipy.sparse). Currently only sympy is implemented.
Returns result : Qubit
The qubit that the system collapsed to upon measurement.
sympy.physics.quantum.qubit.measure all oneshot(qubit, format=sympy)
Perform a oneshot ensemble measurement on all qubits.
A oneshot measurement is equivalent to performing a measurement on a quantum system. This type of measurement does not return the probabilities like an ensemble measurement does, but rather returns one of the possible resulting states. The exact state
that is returned is determined by picking a state randomly according to the ensemble
probabilities.
Parameters qubits : Qubit
The qubit to measure. This can be any Qubit or a linear combination
of them.
format : str
The format of the intermediate matrices to use. Possible values
are (sympy,numpy,scipy.sparse). Currently only sympy is implemented.
Returns result : Qubit
The qubit that the system collapsed to upon measurement.
Shors Algorithm
Shors algorithm and helper functions.
Todo:
1800
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Get the CMod gate working again using the new Gate API.
Fix everything.
Update docstrings and reformat.
Remove print statements. We may want to think about a better API for this.
class sympy.physics.quantum.shor.CMod
A controlled mod gate.
This is black box controlled Mod function for use by shors algorithm. TODO implement
a decompose property that returns how to do this in terms of elementary gates
N
N is the type of modular arithmetic we are doing.
a
Base of the controlled mod function.
t
Size of 1/2 input register. First 1/2 holds output.
sympy.physics.quantum.shor.period find(a, N)
Finds the period of a in modulo N arithmetic
This is quantum part of Shors algorithm.It takes two registers, puts rst in superposition
of states with Hadamards so: |k>|0> with k being all possible choices. It then does a
controlled mod and a QFT to determine the order of a.
sympy.physics.quantum.shor.shor(N)
This function implements Shors factoring algorithm on the Integer N
The algorithm starts by picking a random number (a) and seeing if it is coprime with N.
If it isnt, then the gcd of the two numbers is a factor and we are done. Otherwise, it
begins the period nding subroutine which nds the period of a in modulo N arithmetic.
This period, if even, can be used to calculate factors by taking a**(r/2)-1 and a**(r/2)+1.
These values are returned.
Analytic Solutions
Particle in a Box
1D quantum particle in a box.
class sympy.physics.quantum.piab.PIABHamiltonian
Particle in a box Hamiltonian operator.
class sympy.physics.quantum.piab.PIABKet
Particle in a box eigenket.
class sympy.physics.quantum.piab.PIABBra
Particle in a box eigenbra.
Optics Module
Abstract
Contains docstrings of Physics-Optics module
5.36. Physics Module
1801
SymPy Documentation, Release 0.7.6
Gaussian Optics
Gaussian optics.
The module implements:
Ray transfer matrices for geometrical and gaussian optics.
See RayTransferMatrix, GeometricRay and BeamParameter
Conjugation relations for geometrical and gaussian optics.
See geometric conj*, gauss conj and conjugate gauss beams
The conventions for the distances are as follows:
focal distance positive for convergent lenses
object distance positive for real objects
image distance positive for real images
class sympy.physics.optics.gaussopt.RayTransferMatrix
Base class for a Ray Transfer Matrix.
It should be used if there isnt already a more specic subclass mentioned in See Also.
Parameters parameters : A, B, C and D or 2x2 matrix (Matrix(2, 2, [A, B, C,
D]))
See Also:
GeometricRay (page 1805), BeamParameter (page 1807), FreeSpace (page 1803), FlatRefraction (page 1804), CurvedRefraction (page 1804), FlatMirror (page 1804),
CurvedMirror (page 1805), ThinLens (page 1805)
References
[R323] (page 1917)
Examples
>>> from sympy.physics.optics import RayTransferMatrix, ThinLens
>>> from sympy import Symbol, Matrix
>>> mat = RayTransferMatrix(1, 2, 3, 4)
>>> mat
Matrix([
[1, 2],
[3, 4]])
>>> RayTransferMatrix(Matrix([[1, 2], [3, 4]]))
Matrix([
[1, 2],
[3, 4]])
>>> mat.A
1
1802
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> f = Symbol(f)
>>> lens = ThinLens(f)
>>> lens
Matrix([
[
1, 0],
[-1/f, 1]])
>>> lens.C
-1/f
A
The A parameter of the Matrix.
Examples
>>> from sympy.physics.optics import RayTransferMatrix
>>> mat = RayTransferMatrix(1, 2, 3, 4)
>>> mat.A
1
B
The B parameter of the Matrix.
Examples
>>> from sympy.physics.optics import RayTransferMatrix
>>> mat = RayTransferMatrix(1, 2, 3, 4)
>>> mat.B
2
C
The C parameter of the Matrix.
Examples
>>> from sympy.physics.optics import RayTransferMatrix
>>> mat = RayTransferMatrix(1, 2, 3, 4)
>>> mat.C
3
D
The D parameter of the Matrix.
Examples
>>> from sympy.physics.optics import RayTransferMatrix
>>> mat = RayTransferMatrix(1, 2, 3, 4)
>>> mat.D
4
class sympy.physics.optics.gaussopt.FreeSpace
Ray Transfer Matrix for free space.
5.36. Physics Module
1803
SymPy Documentation, Release 0.7.6
Parameters distance :
See Also:
RayTransferMatrix (page 1802)
Examples
>>> from sympy.physics.optics import FreeSpace
>>> from sympy import symbols
>>> d = symbols(d)
>>> FreeSpace(d)
Matrix([
[1, d],
[0, 1]])
class sympy.physics.optics.gaussopt.FlatRefraction
Ray Transfer Matrix for refraction.
Parameters n1 : refractive index of one medium
n2 : refractive index of other medium
See Also:
RayTransferMatrix (page 1802)
Examples
>>> from sympy.physics.optics import FlatRefraction
>>> from sympy import symbols
>>> n1, n2 = symbols(n1 n2)
>>> FlatRefraction(n1, n2)
Matrix([
[1,
0],
[0, n1/n2]])
class sympy.physics.optics.gaussopt.CurvedRefraction
Ray Transfer Matrix for refraction on curved interface.
Parameters R : radius of curvature (positive for concave)
n1 : refractive index of one medium
n2 : refractive index of other medium
See Also:
RayTransferMatrix (page 1802)
Examples
>>> from sympy.physics.optics import CurvedRefraction
>>> from sympy import symbols
>>> R, n1, n2 = symbols(R n1 n2)
>>> CurvedRefraction(R, n1, n2)
Matrix([
[
1,
0],
[(n1 - n2)/(R*n2), n1/n2]])
1804
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
class sympy.physics.optics.gaussopt.FlatMirror
Ray Transfer Matrix for reection.
See Also:
RayTransferMatrix (page 1802)
Examples
>>> from sympy.physics.optics import FlatMirror
>>> FlatMirror()
Matrix([
[1, 0],
[0, 1]])
class sympy.physics.optics.gaussopt.CurvedMirror
Ray Transfer Matrix for reection from curved surface.
Parameters R : radius of curvature (positive for concave)
See Also:
RayTransferMatrix (page 1802)
Examples
>>> from sympy.physics.optics import CurvedMirror
>>> from sympy import symbols
>>> R = symbols(R)
>>> CurvedMirror(R)
Matrix([
[
1, 0],
[-2/R, 1]])
class sympy.physics.optics.gaussopt.ThinLens
Ray Transfer Matrix for a thin lens.
Parameters f : the focal distance
See Also:
RayTransferMatrix (page 1802)
Examples
>>> from sympy.physics.optics import ThinLens
>>> from sympy import symbols
>>> f = symbols(f)
>>> ThinLens(f)
Matrix([
[
1, 0],
[-1/f, 1]])
class sympy.physics.optics.gaussopt.GeometricRay
Representation for a geometric ray in the Ray Transfer Matrix formalism.
5.36. Physics Module
1805
SymPy Documentation, Release 0.7.6
Parameters h : height, and
angle : angle, or
matrix : a 2x1 matrix (Matrix(2, 1, [height, angle]))
Examples :
======= :
>>> from sympy.physics.optics import GeometricRay, FreeSpace :
>>> from sympy import symbols, Matrix :
>>> d, h, angle = symbols(d, h, angle) :
>>> GeometricRay(h, angle) :
Matrix([ :
[ h], :
[angle]]) :
>>> FreeSpace(d)*GeometricRay(h, angle) :
Matrix([ :
[angle*d + h], :
[ angle]]) :
>>> GeometricRay( Matrix( ((h,), (angle,)) ) ) :
Matrix([ :
[ h], :
[angle]]) :
See Also:
RayTransferMatrix (page 1802)
angle
The angle with the optical axis.
Examples
>>> from sympy.physics.optics import GeometricRay
>>> from sympy import symbols
>>> h, angle = symbols(h, angle)
>>> gRay = GeometricRay(h, angle)
>>> gRay.angle
angle
height
The distance from the optical axis.
Examples
>>>
>>>
>>>
>>>
1806
from sympy.physics.optics import GeometricRay
from sympy import symbols
h, angle = symbols(h, angle)
gRay = GeometricRay(h, angle)
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> gRay.height
h
class sympy.physics.optics.gaussopt.BeamParameter
Representation for a gaussian ray in the Ray Transfer Matrix formalism.
Parameters wavelen : the wavelength,
z : the distance to waist, and
w : the waist, or
z r : the rayleigh range
See Also:
RayTransferMatrix (page 1802)
References
[R324] (page 1917)
Examples
>>>
>>>
>>>
1 +
from sympy.physics.optics import BeamParameter
p = BeamParameter(530e-9, 1, w=1e-3)
p.q
1.88679245283019*I*pi
>>> p.q.n()
1.0 + 5.92753330865999*I
>>> p.w_0.n()
0.00100000000000000
>>> p.z_r.n()
5.92753330865999
>>> from sympy.physics.optics import FreeSpace
>>> fs = FreeSpace(10)
>>> p1 = fs*p
>>> p.w.n()
0.00101413072159615
>>> p1.w.n()
0.00210803120913829
divergence
Half of the total angular spread.
Examples
>>> from sympy.physics.optics import BeamParameter
>>> p = BeamParameter(530e-9, 1, w=1e-3)
>>> p.divergence
0.00053/pi
gouy
The Gouy phase.
5.36. Physics Module
1807
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.physics.optics import BeamParameter
>>> p = BeamParameter(530e-9, 1, w=1e-3)
>>> p.gouy
atan(0.53/pi)
q
The complex parameter representing the beam.
Examples
>>>
>>>
>>>
1 +
from sympy.physics.optics import BeamParameter
p = BeamParameter(530e-9, 1, w=1e-3)
p.q
1.88679245283019*I*pi
radius
The radius of curvature of the phase front.
Examples
>>> from sympy.physics.optics import BeamParameter
>>> p = BeamParameter(530e-9, 1, w=1e-3)
>>> p.radius
0.2809/pi**2 + 1
w
The beam radius at 1/e2 intensity.
See Also:
w 0 (page 1808) the minimal radius of beam
Examples
>>> from sympy.physics.optics import BeamParameter
>>> p = BeamParameter(530e-9, 1, w=1e-3)
>>> p.w
0.001*sqrt(0.2809/pi**2 + 1)
w0
The beam waist (minimal radius).
See Also:
w (page 1808) the beam radius at 1/e2 intensity
Examples
1808
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> from sympy.physics.optics import BeamParameter
>>> p = BeamParameter(530e-9, 1, w=1e-3)
>>> p.w_0
0.00100000000000000
waist approximation limit
The minimal waist for which the gauss beam approximation is valid.
The gauss beam is a solution to the paraxial equation. For curvatures that are too
great it is not a valid approximation.
Examples
>>> from sympy.physics.optics import BeamParameter
>>> p = BeamParameter(530e-9, 1, w=1e-3)
>>> p.waist_approximation_limit
1.06e-6/pi
sympy.physics.optics.gaussopt.waist2rayleigh(w, wavelen)
Calculate the rayleigh range from the waist of a gaussian beam.
See Also:
rayleigh2waist (page 1809), BeamParameter (page 1807)
Examples
>>> from sympy.physics.optics import waist2rayleigh
>>> from sympy import symbols
>>> w, wavelen = symbols(w wavelen)
>>> waist2rayleigh(w, wavelen)
pi*w**2/wavelen
sympy.physics.optics.gaussopt.rayleigh2waist(z r, wavelen)
Calculate the waist from the rayleigh range of a gaussian beam.
See Also:
waist2rayleigh (page 1809), BeamParameter (page 1807)
Examples
>>> from sympy.physics.optics import rayleigh2waist
>>> from sympy import symbols
>>> z_r, wavelen = symbols(z_r wavelen)
>>> rayleigh2waist(z_r, wavelen)
sqrt(wavelen*z_r)/sqrt(pi)
sympy.physics.optics.gaussopt.geometric conj ab(a, b)
Conjugation relation for geometrical beams under paraxial conditions.
Takes the distances to the optical element and returns the needed focal distance.
See Also:
geometric conj af (page 1810), geometric conj bf (page 1810)
5.36. Physics Module
1809
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.physics.optics import geometric_conj_ab
>>> from sympy import symbols
>>> a, b = symbols(a b)
>>> geometric_conj_ab(a, b)
a*b/(a + b)
sympy.physics.optics.gaussopt.geometric conj af(a, f)
Conjugation relation for geometrical beams under paraxial conditions.
Takes the object distance (for geometric conj af) or the image distance (for geometric conj bf) to the optical element and the focal distance. Then it returns the other
distance needed for conjugation.
See Also:
geometric conj ab (page 1809)
Examples
>>> from sympy.physics.optics.gaussopt import geometric_conj_af, geometric_conj_bf
>>> from sympy import symbols
>>> a, b, f = symbols(a b f)
>>> geometric_conj_af(a, f)
a*f/(a - f)
>>> geometric_conj_bf(b, f)
b*f/(b - f)
sympy.physics.optics.gaussopt.geometric conj bf(a, f)
Conjugation relation for geometrical beams under paraxial conditions.
Takes the object distance (for geometric conj af) or the image distance (for geometric conj bf) to the optical element and the focal distance. Then it returns the other
distance needed for conjugation.
See Also:
geometric conj ab (page 1809)
Examples
>>> from sympy.physics.optics.gaussopt import geometric_conj_af, geometric_conj_bf
>>> from sympy import symbols
>>> a, b, f = symbols(a b f)
>>> geometric_conj_af(a, f)
a*f/(a - f)
>>> geometric_conj_bf(b, f)
b*f/(b - f)
sympy.physics.optics.gaussopt.gaussian conj(s in, z r in, f)
Conjugation relation for gaussian beams.
Parameters s in : the distance to optical element from the waist
z r in : the rayleigh range of the incident beam
f : the focal length of the optical element
1810
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Returns a tuple containing (s out, z r out, m) :
s out : the distance between the new waist and the optical element
z r out : the rayleigh range of the emergent beam
m : the ration between the new and the old waists
Examples
>>> from sympy.physics.optics import gaussian_conj
>>> from sympy import symbols
>>> s_in, z_r_in, f = symbols(s_in z_r_in f)
>>> gaussian_conj(s_in, z_r_in, f)[0]
1/(-1/(s_in + z_r_in**2/(-f + s_in)) + 1/f)
>>> gaussian_conj(s_in, z_r_in, f)[1]
z_r_in/(1 - s_in**2/f**2 + z_r_in**2/f**2)
>>> gaussian_conj(s_in, z_r_in, f)[2]
1/sqrt(1 - s_in**2/f**2 + z_r_in**2/f**2)
sympy.physics.optics.gaussopt.conjugate gauss beams(wavelen,
waist in,
waist out, **kwargs)
Find the optical setup conjugating the object/image waists.
Parameters wavelen : the wavelength of the beam
waist in and waist out : the waists to be conjugated
f : the focal distance of the element used in the conjugation
Returns a tuple containing (s in, s out, f) :
s in : the distance before the optical element
s out : the distance after the optical element
f : the focal distance of the optical element
Examples
>>> from sympy.physics.optics import conjugate_gauss_beams
>>> from sympy import symbols, factor
>>> l, w_i, w_o, f = symbols(l w_i w_o f)
>>> conjugate_gauss_beams(l, w_i, w_o, f=f)[0]
f*(-sqrt(w_i**2/w_o**2 - pi**2*w_i**4/(f**2*l**2)) + 1)
>>> factor(conjugate_gauss_beams(l, w_i, w_o, f=f)[1])
f*w_o**2*(w_i**2/w_o**2 - sqrt(w_i**2/w_o**2 pi**2*w_i**4/(f**2*l**2)))/w_i**2
>>> conjugate_gauss_beams(l, w_i, w_o, f=f)[2]
f
5.36. Physics Module
1811
SymPy Documentation, Release 0.7.6
Medium
Contains
Medium
class sympy.physics.optics.medium.Medium
This class represents an optical medium. The prime reason to implement this is to facilitate refraction, Fermats priciple, etc.
An optical medium is a material through which electromagnetic waves propagate. The
permittivity and permeability of the medium dene how electromagnetic waves propagate in it.
Parameters name: string :
The display name of the Medium.
permittivity: Sympifyable :
Electric permittivity of the space.
permeability: Sympifyable :
Magnetic permeability of the space.
n: Sympifyable :
Index of refraction of the medium.
References
[R325] (page 1917)
Examples
>>> from sympy.abc import epsilon, mu
>>> from sympy.physics.optics import Medium
>>> m1 = Medium(m1)
>>> m2 = Medium(m2, epsilon, mu)
>>> m1.intrinsic_impedance
149896229*pi*kg*m**2/(1250000*A**2*s**3)
>>> m2.refractive_index
299792458*m*sqrt(epsilon*mu)/s
intrinsic impedance
Returns intrinsic impedance of the medium.
The intrinsic impedance of a medium is the ratio of the transverse components of the
electric and magnetic elds of the electromagnetic wave travelling in the medium.
In a region with no electrical conductivity it simplies to the square root of ratio of
magnetic permeability to electric permittivity.
Examples
1812
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> from sympy.physics.optics import Medium
>>> m = Medium(m)
>>> m.intrinsic_impedance
149896229*pi*kg*m**2/(1250000*A**2*s**3)
permeability
Returns magnetic permeability of the medium.
Examples
>>> from sympy.physics.optics import Medium
>>> m = Medium(m)
>>> m.permeability
pi*kg*m/(2500000*A**2*s**2)
permittivity
Returns electric permittivity of the medium.
Examples
>>> from sympy.physics.optics import Medium
>>> m = Medium(m)
>>> m.permittivity
625000*A**2*s**4/(22468879468420441*pi*kg*m**3)
refractive index
Returns refractive index of the medium.
Examples
>>> from sympy.physics.optics import Medium
>>> m = Medium(m)
>>> m.refractive_index
1
speed
Returns speed of the electromagnetic wave travelling in the medium.
Examples
>>> from sympy.physics.optics import Medium
>>> m = Medium(m)
>>> m.speed
299792458*m/s
Utilities
Contains
refraction angle
5.36. Physics Module
1813
SymPy Documentation, Release 0.7.6
deviation
lens makers formula
mirror formula
lens formula
sympy.physics.optics.utils.refraction angle(incident, medium1, medium2, normal=None, plane=None)
This function calculates transmitted vector after refraction at planar surface. medium1
and medium2 can be M edium or any sympiable object.
If incident is an object of Ray3D, normal also has to be an instance of Ray3D in order to
get the output as a Ray3D. Please note that if plane of separation is not provided and
normal is an instance of Ray3D, normal will be assumed to be intersecting incident ray
at the plane of separation. This will not be the case when normal is a M atrix or any other
sequence. If incident is an instance of Ray3D and plane has not been provided and normal
is not Ray3D, output will be a M atrix.
Parameters incident : Matrix, Ray3D, tuple or list
Incident vector
medium1 : sympy.physics.optics.medium.Medium or sympiable
Medium 1 or its refractive index
medium2 : sympy.physics.optics.medium.Medium or sympiable
Medium 2 or its refractive index
normal : Matrix, Ray3D, tuple or list
Normal vector
plane : Plane
Plane of separation of the two media.
Examples
>>> from sympy.physics.optics import refraction_angle
>>> from sympy.geometry import Point3D, Ray3D, Plane
>>> from sympy.matrices import Matrix
>>> from sympy import symbols
>>> n = Matrix([0, 0, 1])
>>> P = Plane(Point3D(0, 0, 0), normal_vector=[0, 0, 1])
>>> r1 = Ray3D(Point3D(-1, -1, 1), Point3D(0, 0, 0))
>>> refraction_angle(r1, 1, 1, n)
Matrix([
[ 1],
[ 1],
[-1]])
>>> refraction_angle(r1, 1, 1, plane=P)
Ray3D(Point3D(0, 0, 0), Point3D(1, 1, -1))
With dierent index of refraction of the two media
>>> n1, n2 = symbols(n1, n2)
>>> refraction_angle(r1, n1, n2, n)
Matrix([
[
n1/n2],
1814
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
[
n1/n2],
[-sqrt(3)*sqrt(-2*n1**2/(3*n2**2) + 1)]])
>>> refraction_angle(r1, n1, n2, plane=P)
Ray3D(Point3D(0, 0, 0), Point3D(n1/n2, n1/n2, -sqrt(3)*sqrt(-2*n1**2/(3*n2**2) + 1)))
sympy.physics.optics.utils.deviation(incident,
medium1,
medium2,
normal=None, plane=None)
This function calculates the angle of deviation of a ray due to refraction at planar surface.
Parameters incident : Matrix, Ray3D, tuple or list
Incident vector
medium1 : sympy.physics.optics.medium.Medium or sympiable
Medium 1 or its refractive index
medium2 : sympy.physics.optics.medium.Medium or sympiable
Medium 2 or its refractive index
normal : Matrix, Ray3D, tuple or list
Normal vector
plane : Plane
Plane of separation of the two media.
Examples
>>> from sympy.physics.optics import deviation
>>> from sympy.geometry import Point3D, Ray3D, Plane
>>> from sympy.matrices import Matrix
>>> from sympy import symbols
>>> n1, n2 = symbols(n1, n2)
>>> n = Matrix([0, 0, 1])
>>> P = Plane(Point3D(0, 0, 0), normal_vector=[0, 0, 1])
>>> r1 = Ray3D(Point3D(-1, -1, 1), Point3D(0, 0, 0))
>>> deviation(r1, 1, 1, n)
0
>>> deviation(r1, n1, n2, plane=P)
-acos(-sqrt(-2*n1**2/(3*n2**2) + 1)) + acos(-sqrt(3)/3)
sympy.physics.optics.utils.lens makers formula(n lens, n surr, r1, r2)
This function calculates focal length of a thin lens. It follows cartesian sign convention.
Parameters n lens : Medium or sympiable
Index of refraction of lens.
n surr : Medium or sympiable
Index of reection of surrounding.
r1 : sympiable
Radius of curvature of rst surface.
r2 : sympiable
Radius of curvature of second surface.
5.36. Physics Module
1815
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.physics.optics import lens_makers_formula
>>> lens_makers_formula(1.33, 1, 10, -10)
15.1515151515151
sympy.physics.optics.utils.mirror formula(focal length=None,
u=None,
v=None)
This function provides one of the three parameters when two of them are supplied. This
is valid only for paraxial rays.
Parameters focal length : sympiable
Focal length of the mirror.
u : sympiable
Distance of object from the pole on the principal axis.
v : sympiable
Distance of the image from the pole on the principal axis.
Examples
>>> from sympy.physics.optics import mirror_formula
>>> from sympy.abc import f, u, v
>>> mirror_formula(focal_length=f, u=u)
f*u/(-f + u)
>>> mirror_formula(focal_length=f, v=v)
f*v/(-f + v)
>>> mirror_formula(u=u, v=v)
u*v/(u + v)
sympy.physics.optics.utils.lens formula(focal length=None, u=None, v=None)
This function provides one of the three parameters when two of them are supplied. This
is valid only for paraxial rays.
Parameters focal length : sympiable
Focal length of the mirror.
u : sympiable
Distance of object from the optical center on the principal axis.
v : sympiable
Distance of the image from the optical center on the principal axis.
Examples
>>> from sympy.physics.optics import lens_formula
>>> from sympy.abc import f, u, v
>>> lens_formula(focal_length=f, u=u)
f*u/(f + u)
>>> lens_formula(focal_length=f, v=v)
f*v/(f - v)
>>> lens_formula(u=u, v=v)
u*v/(u - v)
1816
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Waves
This module has all the classes and functions related to waves in optics.
Contains
TWave
class sympy.physics.optics.waves.TWave(amplitude, frequency=None, phase=0,
time period=None, n=n)
This is a simple transverse sine wave travelling in a one dimensional space. Basic properties are required at the time of creation of the object but they can be changed later
with respective methods provided.
It has been represented as A cos(k x t + ) where A is amplitude, is angular
velocity, kiswavenumber, : math : x is a spatial variable to represent the position on the
dimension on which the wave propagates and is phase angle of the wave.
Raises ValueError : When neither frequency nor time period is provided
or they are not consistent.
TypeError : When anyting other than TWave objects is added.
Examples
>>> from sympy import symbols
>>> from sympy.physics.optics import TWave
>>> A1, phi1, A2, phi2, f = symbols(A1, phi1, A2, phi2, f)
>>> w1 = TWave(A1, f, phi1)
>>> w2 = TWave(A2, f, phi2)
>>> w3 = w1 + w2 # Superposition of two waves
>>> w3
TWave(sqrt(A1**2 + 2*A1*A2*cos(phi1 - phi2) + A2**2), f,
atan2(A1*cos(phi1) + A2*cos(phi2), A1*sin(phi1) + A2*sin(phi2)))
>>> w3.amplitude
sqrt(A1**2 + 2*A1*A2*cos(phi1 - phi2) + A2**2)
>>> w3.phase
atan2(A1*cos(phi1) + A2*cos(phi2), A1*sin(phi1) + A2*sin(phi2))
>>> w3.speed
299792458*m/(n*s)
>>> w3.angular_velocity
2*pi*f
Arguments
amplitude [Sympifyable] Amplitude of the wave.
frequency [Sympifyable] Frequency of the wave.
phase [Sympifyable] Phase angle of the wave.
time period [Sympifyable] Time period of the wave.
n [Sympifyable] Refractive index of the medium.
amplitude
Returns the amplitude of the wave.
5.36. Physics Module
1817
SymPy Documentation, Release 0.7.6
Examples
>>>
>>>
>>>
>>>
>>>
A
from sympy import symbols
from sympy.physics.optics import TWave
A, phi, f = symbols(A, phi, f)
w = TWave(A, f, phi)
w.amplitude
angular velocity
Returns angular velocity of the wave.
Examples
>>> from sympy import symbols
>>> from sympy.physics.optics import TWave
>>> A, phi, f = symbols(A, phi, f)
>>> w = TWave(A, f, phi)
>>> w.angular_velocity
2*pi*f
frequency
Returns the frequency of the wave.
Examples
>>>
>>>
>>>
>>>
>>>
f
from sympy import symbols
from sympy.physics.optics import TWave
A, phi, f = symbols(A, phi, f)
w = TWave(A, f, phi)
w.frequency
phase
Returns the phase angle of the wave.
Examples
>>>
>>>
>>>
>>>
>>>
phi
from sympy import symbols
from sympy.physics.optics import TWave
A, phi, f = symbols(A, phi, f)
w = TWave(A, f, phi)
w.phase
speed
Returns the speed of travelling wave. It is medium dependent.
1818
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import symbols
>>> from sympy.physics.optics import TWave
>>> A, phi, f = symbols(A, phi, f)
>>> w = TWave(A, f, phi)
>>> w.speed
299792458*m/(n*s)
time period
Returns the time period of the wave.
Examples
>>>
>>>
>>>
>>>
>>>
1/f
from sympy import symbols
from sympy.physics.optics import TWave
A, phi, f = symbols(A, phi, f)
w = TWave(A, f, phi)
w.time_period
wavelength
Returns wavelength of the wave. It depends on the medium of the wave.
Examples
>>> from sympy import symbols
>>> from sympy.physics.optics import TWave
>>> A, phi, f = symbols(A, phi, f)
>>> w = TWave(A, f, phi)
>>> w.wavelength
299792458*m/(f*n*s)
wavenumber
Returns wavenumber of the wave.
Examples
>>> from sympy import symbols
>>> from sympy.physics.optics import TWave
>>> A, phi, f = symbols(A, phi, f)
>>> w = TWave(A, f, phi)
>>> w.wavenumber
pi*f*n*s/(149896229*m)
Unit systems
This module integrates unit systems into SymPy, allowing a user choose which system to use
when doing their computations and providing utilities to display and convert units.
5.36. Physics Module
1819
SymPy Documentation, Release 0.7.6
Unit systems are composed of units and constants, which are themselves described from
dimensions and numbers, and possibly a prex. Quantities are dened by their unit and their
numerical value, with respect to the current system.
The main advantage of this implementation over the old unit module is that it divides the
units in unit systems, so that the user can decide which units to use, instead of having all in
the name space (for example astrophysicists can only use units with ua, Earth or Sun masses,
the theoricists will use natural system, etc.). Moreover it allows a better control over the
dimensions and conversions.
Ideas about future developments can be found on the Github wiki.
Philosophy behind unit systems
Dimensions
Introduction At the root of unit systems are dimension systems, whose structure mainly
determines the one of unit systems. Our denition could seem rough but they are largely
sucient for our purposes.
A dimension will be dened as a property which is measurable and assigned to a specic
phenomenom. In this sense dimensions are dierent from pure numbers because they carry
some extra-sense, and for this reason two dierent dimensions can not be added. For example
time or length are dimensions, but also any other things which has some sense for us, like
angle, number of particles (moles...) or information (bits...).
From this point of view the only truly dimensionless quantity are pure numbers. The idea
of being dimensionless is very system-dependent, as can be seen from the (c, ~, G), in which
all units appears to be dimensionless in the usual common sense. This is unavoidable for
computability of generic unit systems (but at the end we can tell the program what is dimensionless).
Dimensions can be composed together by taking their product or their ratio (to be dened
below). For example the velocity is dened as length divided by time, or we can see the length
as velocity multiplied by time, depending of what we see as the more fundamental: in general
we can select a set of base dimensions from which we can describe all the others.
Group structure After this short introduction whose aim was to introduce the dimensions
from an intuitive perspective, we describe the mathematical structure. A dimension system
with n independent dimensions {di }i=1,...,n is described by a multiplicative group G:
there an identity element 1 corresponding to pure numbers;
the product D3 = D1 D2 of two elements D1 , D2 G is also in G;
any element D G has an inverse D1 G.
We denote
Dn = D D,
|
{z
}
n times
and by denition D0 = 1. The {di }i=1,...,n are called generators of the group since any element
D G can be expressed as the product of powers of the generators:
D=
dai i ,
ai Z.
i=1
1820
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
The identity is given for ai = 0, i, while we recover the generator di for ai = 1, aj = 0, j 6= i.
This group has the following properties:
1. abelian, since the generator commutes, [di , dj ] = 0;
2. countable (innite but discrete) since the elements are indexed by the powers of the
generators 9 .
One can change the dimension basis {d0i }i=1,...,n by taking some combination of the old generators:
d0i
dj ij .
j=1
Linear space representation It is possible to use the linear space Zn as a representation
of the group since the power coecients ai carry all the information one needs (we do not
distinguish between the element of the group and its representation):
a1
..
(di )j = ij ,
D = . .
an
The change of basis to d0i follows the usual rule of change of basis for linear space, the matrix
being given by the coecients Pij , which are simply the coecients of the new vectors in
term of the old basis:
d0i = Pij dj .
We will use this last solution in our algorithm.
An example In order to illustrate all this formalism, we end this section with a specic
example, the MKS system (m, kg, s) with dimensions (L: length, M: mass, T: time). They are
represented as (we will always sort the vectors in alphabetic order)
1
0
0
L = 0 ,
M = 1 ,
T = 0 .
0
0
1
Other dimensions can be derived, for example velocity V or action A
V = LT 1 ,
1
V = 0 ,
1
A = M L2 T 2 ,
2
A = 1 .
2
We can change the basis to go to the natural system (m, c, ~) with dimension (L: length, V:
velocity, A: action) 10 . In this basis the generators are
1
0
0
A = 0 ,
L = 1 ,
V = 0 ,
0
0
1
9 In general we will consider only dimensions with a maximum coecient, so we can only a truncation of the
group; but this is not useful for the algorithm.
10 We anticipate a little by considering c and ~ as units and not as physical constants.
5.36. Physics Module
1821
SymPy Documentation, Release 0.7.6
whereas the mass and time are given by
T = LV 1 ,
0
T = 1 ,
1
M = AV 2 ,
1
M = 0 .
2
Finally the inverse change of basis matrix P 1 is obtained by gluing the vectors expressed in
the old basis:
2 1 1
P 1 = 1 0 0 .
2 0 1
To nd the change of basis matrix we just have to
0 1
P = 1 0
0 2
take the inverse
0
1 .
1
Units Units are constructed upon dimensions: they give them an ordering, or equivalently
a magnitude. First one has to choose some origin, usually based on some physical measurement. Then the other units of the same dimensions follow by simple rescaling (we omit for
now the case of unit with oset, such as the temperature).
For example consider the length dimension. We start by choosing the origin to be the meter
m, and we dene the other units from it: the kilometer is 103 m, the foot is 0.3048 m, etc.
After dening several units of dierent dimensions we can form a unit system, which is basically a dimension system with a notion of scale.
Constants Physical constants are really just units. They indicate that we did not understand
that two dimensions are in fact the same. For example we see a velocity for the light dierent
from 1 because we do not think that time is the same as space (which is normal because of
our sense; but it is dierent at the fundamental level). For example, once there was the heat
constant which allowed to convert between joules and calories since people did not know
that heat was energy. As soon as they understood it they xed this constant to 1 (this is a very
schematic story).
We can interpret the fact that now we x the value of fundamental constants in the SI as
showing that they are units (and we use them to dene the other usual units).
Quantities A quantity is dened from a factor and a unit. According to this construction,
they are very similar to unit since at the very end a quantity is just a scale (made from the
quantity and unit factors) and a dimension. What distinguishes units is that they play the role
of some xed references, and we always have a denite number of them.
The need for a reference It is not possible to dene from scratch units and unit systems:
one needs to dene some references, and then build the rest over them. Said in another way,
we need an origin for the scales of our units (i.e. a unit with factor 1), and to be sure that all
units of a given dimension are dened consistently we need to use the same origin for all of
them. This can happend if we want to use a derived unit as a base units in another system: we
should not dene it as having a scale 1, because, even if it is inconsistent inside the system,
1822
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
we could not convert to the rst system since we have two dierent units (from our point of
view) of same scale (which means they are equal for the computer).
We will say that the dimensions and scales dened outside systems are canonical, because
we use them for all computations. On the other side the dimensions and scales obtained with
reference to a system are called physical, because they ultimately carry a sense.
Lets use a concrete (and important) example: the case of the mass units. We would like to
dene the gram as the origin. We would like to dene the gram as the canonical origin for
the mass, so we assign it a scale 1. Then we can dene a system (e.g. in chemistry) that take
it as a base unit. The MKS system prefers to use the kilogram; a naive choice would be to
attribute it a scale if 1 since it is a base, but we see that we could not convert to the chemistry
system because g and kg have both been given the same factor. So we need to dene kg as
1000 g, and only then use it as a base in MKS. But as soon as we ask the question what is
the factor of kg in MKS?, we get the answer 1, since it is a base unit.
Thus we will dene all computations without referring to a system, and it is only at the end
that we can plug the result into a system to give the context we are interested in.
Literature
Examples
In the following sections we give few examples of what can be do with this module.
Dimensional analysis We will start from the second Newtons law
ma = F
where m, a and F are the mass, the acceleration and the force respectively. Knowing the
dimensions of m (M ) and a (LT 2 ), we will determine the dimension of F ; obviously we will
nd that it is a force: M LT 2 .
From there we will use the expression of the gravitational force between the particle of mass
m and the body of mass M , at a distance r
F =
GmM
r2
to determine the dimension of the Newtons constant G. The result should be L3 M 1 T 2 .
>>> from __future__ import division
>>> from sympy import solve, Symbol, symbols
>>> from sympy.physics.unitsystems.systems import mks_dim as mks
>>> from sympy.physics.unitsystems.simplifiers import dim_simplify
>>> length, mass, time = mks[length], mks[mass], mks[time]
>>> acceleration = mks[acceleration]
>>> m, a, F = symbols(m a F)
>>> newton = m*a - F
>>> sol = solve(newton, F)[0]
>>> force = dim_simplify(sol.subs({m: mass, a: acceleration}))
>>> force
{length: 1, mass: 1, time: -2}
5.36. Physics Module
1823
SymPy Documentation, Release 0.7.6
>>> force == mks[force]
True
>>> M, r, G = symbols(M r G)
>>> grav_force = F - G * m * M / r**2
>>> sol = solve(grav_force, G)[0]
>>> const = dim_simplify(sol.subs({m: mass, M: mass, r: length, F: force}))
>>> const
{length: 3, mass: -1, time: -2}
Note that one should rst solve the equation, and then substitute with the dimensions.
Equation with quantities
Using Keplers third law
T2
4 2
=
a3
GM
we can nd the Venus orbital period using the known values for the other variables (taken
from Wikipedia). The result should be 224.701 days.
>>> from __future__ import division
>>> from sympy import solve, Symbol, symbols, pi
>>> from sympy.physics.unitsystems import Unit, Quantity as Q
>>> from sympy.physics.unitsystems.simplifiers import qsimplify
>>> from sympy.physics.unitsystems.systems import mks
>>> m, kg, s = mks[m], mks[kg], mks[s]
>>> T, a, M, G = symbols(T a M G)
>>> venus_a = Q(108208000e3, m)
>>> solar_mass = Q(1.9891e30, kg)
>>> venus_subs = {a: venus_a, M: solar_mass, G: mks[G]}
>>> Tsol = solve(T**2 / a**3 - 4*pi**2 / G / M, T)[1]
>>> q = qsimplify(Tsol.subs(venus_subs))
>>> day = Unit(s.dim, abbrev=day, factor=86400)
>>> print(q.convert_to(qsimplify(day)))
224.667 day
We could also have the solar mass and the day as units coming from the astrophysical system,
but I wanted to show how to create a unit that one needs.
We can see in this example that intermediate dimensions can be ill-dened, such as sqrt(G),
but one should check that the nal result - when all dimensions are combined - is well dened.
Dimensions and dimension systems
Denition of physical dimensions.
Unit systems will be constructed on top of these dimensions.
Most of the examples in the doc use MKS system and are presented from the computer point
of view: from a human point, adding length to time is not legal in MKS but it is in natural
system; for a computer in natural system there is no time dimension (but a velocity dimension
instead) - in the basis - so the question of adding time to length has no meaning.
class sympy.physics.unitsystems.dimensions.Dimension
This class represent the dimension of a physical quantities.
The dimensions may have a name and a symbol. All other arguments are dimensional
powers. They represent a characteristic of a quantity, giving an interpretation to it: for
1824
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
example (in classical mechanics) we know that time is dierent from temperature, and
dimensions make this dierence (but they do not provide any measure of these quantites).
>>> from sympy.physics.unitsystems.dimensions import Dimension
>>> length = Dimension(length=1)
>>> length
{length: 1}
>>> time = Dimension(time=1)
Dimensions behave like a dictionary where the key is the name and the value corresponds
to the exponent.
Dimensions can be composed using multiplication, division and exponentiation (by a
number) to give new dimensions. Addition and subtraction is dened only when the two
objects are the same dimension.
>>> velocity = length.div(time)
>>> velocity
{length: 1, time: -1}
>>> length.add(length)
{length: 1}
>>> length.pow(2)
{length: 2}
Dening addition-like operations will help when doing dimensional analysis.
Note that two dimensions are equal if they have the same powers, even if their names
and/or symbols dier.
>>> Dimension(length=1) == Dimension(length=1, name=length)
True
>>> Dimension(length=1) == Dimension(length=1, symbol=L)
True
>>> Dimension(length=1) == Dimension(length=1, name=length,
...
symbol=L)
True
add(other)
Dene the addition for Dimension.
Addition of dimension has a sense only if the second object is the same dimension
(we dont add length to time).
get(k[, d ]) D[k] if k in D, else d. d defaults to None.
has integer powers
Check if the dimension object has only integer powers.
All the dimension powers should be integers, but rational powers may appear in
intermediate steps. This method may be used to check that the nal result is welldened.
is dimensionless
Check if the dimension object really has a dimension.
A dimension should have at least one component with non-zero power.
items() list of Ds (key, value) pairs, as 2-tuples
keys() list of Ds keys
values() list of Ds values
5.36. Physics Module
1825
SymPy Documentation, Release 0.7.6
class sympy.physics.unitsystems.dimensions.DimensionSystem(base,
dims=(),
name=,
descr=)
DimensionSystem represents a coherent set of dimensions.
In a system dimensions are of three types:
base dimensions;
derived dimensions: these are dened in terms of the base dimensions (for example
velocity is dened from the division of length by time);
canonical dimensions: these are used to dene systems because one has to start
somewhere: we can not build ex nihilo a system (see the discussion in the documentation for more details).
All intermediate computations will use the canonical basis, but at the end one can choose
to print result in some other basis.
In a system dimensions can be represented as a vector, where the components represent
the powers associated to each base dimension.
can transf matrix
Compute the canonical transformation matrix from the canonical to the base dimension basis.
It is the inverse of the matrix computed with inv can transf matrix().
dim
Give the dimension of the system.
That is return the number of dimensions forming the basis.
dim can vector(dim)
Vector representation in terms of the canonical base dimensions.
dim vector(dim)
Vector representation in terms of the base dimensions.
extend(base, dims=(), name=, description=)
Extend the current system into a new one.
Take the base and normal units of the current system to merge them to the base and
normal units given in argument. If not provided, name and description are overriden
by empty strings.
get dim(dim)
Find a specic dimension which is part of the system.
dim can be a string or a dimension object. If no dimension is found, then return
None.
inv can transf matrix
Compute the inverse transformation matrix from the base to the canonical dimension
basis.
It corresponds to the matrix where columns are the vector of base dimensions in
canonical basis.
This matrix will almost never be used because dimensions are always dene with
respect to the canonical basis, so no work has to be done to get them in this basis.
Nonetheless if this matrix is not square (or not invertible) it means that we have
chosen a bad basis.
is consistent
Check if the system is well dened.
1826
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
list can dims
List all canonical dimension names.
print dim base(dim)
Give the string expression of a dimension in term of the basis.
Dimensions are displayed by decreasing power.
static sort dims(dims)
Sort dimensions given in argument using their str function.
This function will ensure that we get always the same tuple for a given set of dimensions.
Unit prexes
Module dening unit prexe class and some constants.
Constant dict for SI and binary prexes are dened as PREFIXES and BIN PREFIXES.
class sympy.physics.unitsystems.prefixes.Prefix(name,
abbrev,
base=10)
This class represent prexes, with their name, symbol and factor.
exponent,
Prexes are used to create derived units from a given unit. They should always be encapsulated into units.
The factor is constructed from a base (default is 10) to some power, and it gives the total
multiple or fraction. For example the kilometer km is constructed from the meter (factor
1) and the kilo (10 to the power 3, i.e. 1000). The base can be changed to allow e.g.
binary prexes.
A prex multiplied by something will always return the product of this other object times
the factor, except if the other object:
is a prex and they can be combined into a new prex;
denes multiplication with prexes (which is the case for the Unit class).
Units and unit systems
Unit system for physical quantities; include denition of constants.
class sympy.physics.unitsystems.units.Unit
Class for the units.
A unit is dened by two things:
a dimension;
a factor.
The factor represents the position of the unit with respect to the canonical unit of this
dimension. For example if we choose the gram to be the canonical dimension for the
mass, then by denition its factor is 1; on the other hand the factor dened here for
kilogram is 1000, even when it is a base unit. The explanation is that here we do not
have dened any system that we could use as a reference: here the canonical unit is the
only scale, and thus the only available origin.
Additionnaly one can add a prex and an abbreviation. The only utility of the former is to
provide a shorthand for some units, but it is never used among computations; it appears
only when dening and printing units. The same remark applies to the abbreviation.
5.36. Physics Module
1827
SymPy Documentation, Release 0.7.6
All operations (pow, mul, etc.) are dened as the corresponding ones acting on the factor
(a number) and the dimension.
abbrev
Symbol representing the unit name.
Prepend the abbreviation with the prex symbol if it is denes.
abbrev dim
Abbreviation which use only intrinsinc properties of the unit.
as quantity
Convert the unit to a quantity.
The quantity unit is given by the unit of factor 1 and with identical dimension.
>>> from sympy.physics.unitsystems.dimensions import Dimension
>>> from sympy.physics.unitsystems.units import Unit
>>> length = Dimension(length=1)
>>> u = Unit(length, factor=10)
>>> q = u.as_quantity
>>> q.factor
10
>>> q.unit == Unit(length)
True
factor
Overall magnitude of the unit.
is compatible(other)
Test if argument is a unit and has the same dimension as self.
This function is used to verify that some operations can be done.
class sympy.physics.unitsystems.units.Constant
Physical constant.
In our framework a constant is considered as a unit, to which humans givesa special
sense, because we believe that they give us a special information on nature; but it is just
a demonstration of our ignorance.
class sympy.physics.unitsystems.units.UnitSystem(base, units=(), name=, descr=)
UnitSystem represents a coherent set of units.
A unit system is basically a dimension system with notions of scales. Many of the methods
are dened in the same way.
It is much better if all base units have a symbol.
dim
Give the dimension of the system.
That is return the number of units forming the basis.
extend(base, units=(), name=, description=)
Extend the current system into a new one.
Take the base and normal units of the current system to merge them to the base and
normal units given in argument. If not provided, name and description are overriden
by empty strings.
get unit(unit)
Find a specic unit which is part of the system.
unit can be a string or a dimension object. If no unit is found, then return None.
1828
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
is consistent
Check if the underlying dimension system is consistent.
print unit base(unit)
Give the string expression of a unit in term of the basis.
Units are displayed by decreasing power.
Physical quantities
Physical quantities.
class sympy.physics.unitsystems.quantities.Quantity
Physical quantity.
A quantity is dened from a factor and a unit.
add(other)
Add two quantities.
If the other object is not a quantity, raise an error. Two quantities can be added only
if they have the same unit: so we convert rst the other quantity to the same unit
and, if it succedded, then we add the factors.
as unit
Convert the quantity to a unit.
convert to(unit)
Convert the quantity to another (compatible) unit.
5.37 Category Theory Module
5.37.1 Introduction
The category theory module for SymPy will allow manipulating diagrams within a single category, including drawing them in TikZ and deciding whether they are commutative or not.
The general reference work this module tries to follow is
[JoyOfCats] J. Adamek, H. Herrlich. G. E. Strecker: Abstract and Concrete
Categories. The Joy of Cats.
The latest version of this book should be available for free download from
katmat.math.uni-bremen.de/acc/acc.pdf
The module is still in its pre-embryonic stage.
5.37.2 Base Class Reference
This section lists the classes which implement some of the basic notions in category theory:
objects, morphisms, categories, and diagrams.
class sympy.categories.Object
The base class for any kind of object in an abstract category.
While technically any instance of Basic will do, this class is the recommended way to
create abstract objects in abstract categories.
5.37. Category Theory Module
1829
SymPy Documentation, Release 0.7.6
class sympy.categories.Morphism
The base class for any morphism in an abstract category.
In abstract categories, a morphism is an arrow between two category objects. The object
where the arrow starts is called the domain, while the object where the arrow ends is
called the codomain.
Two morphisms between the same pair of objects are considered to be the same morphisms. To distinguish between morphisms between the same objects use NamedMorphism (page 1831).
It is prohibited to instantiate this class. Use one of the derived classes instead.
See Also:
IdentityMorphism (page 1833), NamedMorphism (page 1831), CompositeMorphism
(page 1831)
codomain
Returns the codomain of the morphism.
Examples
>>> from sympy.categories import Object, NamedMorphism
>>> A = Object(A)
>>> B = Object(B)
>>> f = NamedMorphism(A, B, f)
>>> f.codomain
Object(B)
compose(other)
Composes self with the supplied morphism.
The order of elements in the composition is the usual order, i.e., to construct g f
use g.compose(f).
Examples
>>> from sympy.categories import Object, NamedMorphism
>>> A = Object(A)
>>> B = Object(B)
>>> C = Object(C)
>>> f = NamedMorphism(A, B, f)
>>> g = NamedMorphism(B, C, g)
>>> g * f
CompositeMorphism((NamedMorphism(Object(A), Object(B), f),
NamedMorphism(Object(B), Object(C), g)))
>>> (g * f).domain
Object(A)
>>> (g * f).codomain
Object(C)
domain
Returns the domain of the morphism.
1830
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.categories import Object, NamedMorphism
>>> A = Object(A)
>>> B = Object(B)
>>> f = NamedMorphism(A, B, f)
>>> f.domain
Object(A)
class sympy.categories.NamedMorphism
Represents a morphism which has a name.
Names are used to distinguish between morphisms which have the same domain and
codomain: two named morphisms are equal if they have the same domains, codomains,
and names.
See Also:
Morphism (page 1829)
Examples
>>> from sympy.categories import Object, NamedMorphism
>>> A = Object(A)
>>> B = Object(B)
>>> f = NamedMorphism(A, B, f)
>>> f
NamedMorphism(Object(A), Object(B), f)
>>> f.name
f
name
Returns the name of the morphism.
Examples
>>>
>>>
>>>
>>>
>>>
f
from sympy.categories import Object, NamedMorphism
A = Object(A)
B = Object(B)
f = NamedMorphism(A, B, f)
f.name
class sympy.categories.CompositeMorphism
Represents a morphism which is a composition of other morphisms.
Two composite morphisms are equal if the morphisms they were obtained from (components) are the same and were listed in the same order.
The arguments to the constructor for this class should be listed in diagram order: to
obtain the composition g f from the instances of Morphism (page 1829) g and f use
CompositeMorphism(f, g).
5.37. Category Theory Module
1831
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.categories import Object, NamedMorphism, CompositeMorphism
>>> A = Object(A)
>>> B = Object(B)
>>> C = Object(C)
>>> f = NamedMorphism(A, B, f)
>>> g = NamedMorphism(B, C, g)
>>> g * f
CompositeMorphism((NamedMorphism(Object(A), Object(B), f),
NamedMorphism(Object(B), Object(C), g)))
>>> CompositeMorphism(f, g) == g * f
True
codomain
Returns the codomain of this composite morphism.
The codomain of the composite morphism is the codomain of its last component.
Examples
>>> from sympy.categories import Object, NamedMorphism
>>> A = Object(A)
>>> B = Object(B)
>>> C = Object(C)
>>> f = NamedMorphism(A, B, f)
>>> g = NamedMorphism(B, C, g)
>>> (g * f).codomain
Object(C)
components
Returns the components of this composite morphism.
Examples
>>> from sympy.categories import Object, NamedMorphism
>>> A = Object(A)
>>> B = Object(B)
>>> C = Object(C)
>>> f = NamedMorphism(A, B, f)
>>> g = NamedMorphism(B, C, g)
>>> (g * f).components
(NamedMorphism(Object(A), Object(B), f),
NamedMorphism(Object(B), Object(C), g))
domain
Returns the domain of this composite morphism.
The domain of the composite morphism is the domain of its rst component.
Examples
1832
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> from sympy.categories import Object, NamedMorphism
>>> A = Object(A)
>>> B = Object(B)
>>> C = Object(C)
>>> f = NamedMorphism(A, B, f)
>>> g = NamedMorphism(B, C, g)
>>> (g * f).domain
Object(A)
flatten(new name)
Forgets the composite structure of this morphism.
If new name is not empty, returns a NamedMorphism (page 1831) with the supplied
name, otherwise returns a Morphism (page 1829). In both cases the domain of the
new morphism is the domain of this composite morphism and the codomain of the
new morphism is the codomain of this composite morphism.
Examples
>>> from sympy.categories import Object, NamedMorphism
>>> A = Object(A)
>>> B = Object(B)
>>> C = Object(C)
>>> f = NamedMorphism(A, B, f)
>>> g = NamedMorphism(B, C, g)
>>> (g * f).flatten(h)
NamedMorphism(Object(A), Object(C), h)
class sympy.categories.IdentityMorphism
Represents an identity morphism.
An identity morphism is a morphism with equal domain and codomain, which acts as an
identity with respect to composition.
See Also:
Morphism (page 1829)
Examples
>>> from sympy.categories import Object, NamedMorphism, IdentityMorphism
>>> A = Object(A)
>>> B = Object(B)
>>> f = NamedMorphism(A, B, f)
>>> id_A = IdentityMorphism(A)
>>> id_B = IdentityMorphism(B)
>>> f * id_A == f
True
>>> id_B * f == f
True
class sympy.categories.Category
An (abstract) category.
A category [JoyOfCats] is a quadruple K = (O, hom, id, ) consisting of
a (set-theoretical) class O, whose members are called K-objects,
5.37. Category Theory Module
1833
SymPy Documentation, Release 0.7.6
for each pair (A, B) of K-objects, a set hom(A, B) whose members are called Kmorphisms from A to B,
for a each K-object A, a morphism id : A A, called the K-identity of A,
a composition law associating with every K-morphisms f : A B and g : B C a
K-morphism g f : A C, called the composite of f and g.
Composition is associative, K-identities are identities with respect to composition, and
the sets hom(A, B) are pairwise disjoint.
This class knows nothing about its objects and morphisms. Concrete cases of (abstract)
categories should be implemented as classes derived from this one.
Certain instances of Diagram (page 1835) can be asserted to be commutative in a Category (page 1833) by supplying the argument commutative diagrams in the constructor.
See Also:
Diagram (page 1835)
Examples
>>> from sympy.categories import Object, NamedMorphism, Diagram, Category
>>> from sympy import FiniteSet
>>> A = Object(A)
>>> B = Object(B)
>>> C = Object(C)
>>> f = NamedMorphism(A, B, f)
>>> g = NamedMorphism(B, C, g)
>>> d = Diagram([f, g])
>>> K = Category(K, commutative_diagrams=[d])
>>> K.commutative_diagrams == FiniteSet(d)
True
commutative diagrams
Returns the FiniteSet of diagrams which are known to be commutative in this category.
>>> from sympy.categories import Object, NamedMorphism, Diagram, Category
>>> from sympy import FiniteSet
>>> A = Object(A)
>>> B = Object(B)
>>> C = Object(C)
>>> f = NamedMorphism(A, B, f)
>>> g = NamedMorphism(B, C, g)
>>> d = Diagram([f, g])
>>> K = Category(K, commutative_diagrams=[d])
>>> K.commutative_diagrams == FiniteSet(d)
True
name
Returns the name of this category.
Examples
1834
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> from sympy.categories import Category
>>> K = Category(K)
>>> K.name
K
objects
Returns the class of objects of this category.
Examples
>>> from sympy.categories import Object, Category
>>> from sympy import FiniteSet
>>> A = Object(A)
>>> B = Object(B)
>>> K = Category(K, FiniteSet(A, B))
>>> K.objects
Class({Object(A), Object(B)})
class sympy.categories.Diagram
Represents a diagram in a certain category.
Informally, a diagram is a collection of objects of a category and certain morphisms
between them. A diagram is still a monoid with respect to morphism composition; i.e.,
identity morphisms, as well as all composites of morphisms included in the diagram
belong to the diagram. For a more formal approach to this notion see [Pare1970].
The components of composite morphisms are also added to the diagram. No properties
are assigned to such morphisms by default.
A commutative diagram is often accompanied by a statement of the following kind: if
such morphisms with such properties exist, then such morphisms which such properties exist and the diagram is commutative. To represent this, an instance of Diagram
(page 1835) includes a collection of morphisms which are the premises and another collection of conclusions. premises and conclusions associate morphisms belonging to
the corresponding categories with the FiniteSets of their properties.
The set of properties of a composite morphism is the intersection of the sets of properties
of its components. The domain and codomain of a conclusion morphism should be among
the domains and codomains of the morphisms listed as the premises of a diagram.
No checks are carried out of whether the supplied object and morphisms do belong to
one and the same category.
References
[Pare1970] B. Pareigis: Categories and functors. Academic Press, 1970.
Examples
>>>
>>>
>>>
>>>
>>>
from sympy.categories import Object, NamedMorphism, Diagram
from sympy import FiniteSet, pprint, default_sort_key
A = Object(A)
B = Object(B)
C = Object(C)
5.37. Category Theory Module
1835
SymPy Documentation, Release 0.7.6
>>> f = NamedMorphism(A, B, f)
>>> g = NamedMorphism(B, C, g)
>>> d = Diagram([f, g])
>>> premises_keys = sorted(d.premises.keys(), key=default_sort_key)
>>> pprint(premises_keys, use_unicode=False)
[g*f:A-->C, id:A-->A, id:B-->B, id:C-->C, f:A-->B, g:B-->C]
>>> pprint(d.premises, use_unicode=False)
{g*f:A-->C: EmptySet(), id:A-->A: EmptySet(), id:B-->B: EmptySet(), id:C-->C:
EmptySet(), f:A-->B: EmptySet(), g:B-->C: EmptySet()}
>>> d = Diagram([f, g], {g * f: unique})
>>> pprint(d.conclusions)
{g*f:A-->C: {unique}}
conclusions
Returns the conclusions of this diagram.
Examples
>>> from sympy.categories import Object, NamedMorphism
>>> from sympy.categories import IdentityMorphism, Diagram
>>> from sympy import FiniteSet
>>> A = Object(A)
>>> B = Object(B)
>>> C = Object(C)
>>> f = NamedMorphism(A, B, f)
>>> g = NamedMorphism(B, C, g)
>>> d = Diagram([f, g])
>>> IdentityMorphism(A) in d.premises.keys()
True
>>> g * f in d.premises.keys()
True
>>> d = Diagram([f, g], {g * f: unique})
>>> d.conclusions[g * f] == FiniteSet(unique)
True
hom(A, B)
Returns a 2-tuple of sets of morphisms between objects A and B: one set of morphisms listed as premises, and the other set of morphisms listed as conclusions.
See Also:
Object (page 1829), Morphism (page 1829)
Examples
>>> from sympy.categories import Object, NamedMorphism, Diagram
>>> from sympy import pretty
>>> A = Object(A)
>>> B = Object(B)
>>> C = Object(C)
>>> f = NamedMorphism(A, B, f)
>>> g = NamedMorphism(B, C, g)
>>> d = Diagram([f, g], {g * f: unique})
>>> print(pretty(d.hom(A, C), use_unicode=False))
({g*f:A-->C}, {g*f:A-->C})
1836
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
is subdiagram(diagram)
Checks whether diagram is a subdiagram of self. Diagram D0 is a subdiagram of D
if all premises (conclusions) of D0 are contained in the premises (conclusions) of D.
The morphisms contained both in D0 and D should have the same properties for D0
to be a subdiagram of D.
Examples
>>> from sympy.categories import Object, NamedMorphism, Diagram
>>> A = Object(A)
>>> B = Object(B)
>>> C = Object(C)
>>> f = NamedMorphism(A, B, f)
>>> g = NamedMorphism(B, C, g)
>>> d = Diagram([f, g], {g * f: unique})
>>> d1 = Diagram([f])
>>> d.is_subdiagram(d1)
True
>>> d1.is_subdiagram(d)
False
objects
Returns the FiniteSet of objects that appear in this diagram.
Examples
>>> from sympy.categories import Object, NamedMorphism, Diagram
>>> A = Object(A)
>>> B = Object(B)
>>> C = Object(C)
>>> f = NamedMorphism(A, B, f)
>>> g = NamedMorphism(B, C, g)
>>> d = Diagram([f, g])
>>> d.objects
{Object(A), Object(B), Object(C)}
premises
Returns the premises of this diagram.
Examples
>>> from sympy.categories import Object, NamedMorphism
>>> from sympy.categories import IdentityMorphism, Diagram
>>> from sympy import pretty
>>> A = Object(A)
>>> B = Object(B)
>>> f = NamedMorphism(A, B, f)
>>> id_A = IdentityMorphism(A)
>>> id_B = IdentityMorphism(B)
>>> d = Diagram([f])
>>> print(pretty(d.premises, use_unicode=False))
{id:A-->A: EmptySet(), id:B-->B: EmptySet(), f:A-->B: EmptySet()}
5.37. Category Theory Module
1837
SymPy Documentation, Release 0.7.6
subdiagram from objects(objects)
If objects is a subset of the objects of self, returns a diagram which has as premises
all those premises of self which have a domains and codomains in objects, likewise
for conclusions. Properties are preserved.
Examples
>>> from sympy.categories import Object, NamedMorphism, Diagram
>>> from sympy import FiniteSet
>>> A = Object(A)
>>> B = Object(B)
>>> C = Object(C)
>>> f = NamedMorphism(A, B, f)
>>> g = NamedMorphism(B, C, g)
>>> d = Diagram([f, g], {f: unique, g*f: veryunique})
>>> d1 = d.subdiagram_from_objects(FiniteSet(A, B))
>>> d1 == Diagram([f], {f: unique})
True
5.37.3 Diagram Drawing
This section lists the classes which allow automatic drawing of diagrams.
class sympy.categories.diagram drawing.DiagramGrid(diagram,
**hints)
Constructs and holds the tting of the diagram into a grid.
groups=None,
The mission of this class is to analyse the structure of the supplied diagram and to place
its objects on a grid such that, when the objects and the morphisms are actually drawn,
the diagram would be readable, in the sense that there will not be many intersections of
moprhisms. This class does not perform any actual drawing. It does strive nevertheless
to oer sucient metadata to draw a diagram.
Consider the following simple diagram.
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
from sympy.categories import Object, NamedMorphism
from sympy.categories import Diagram, DiagramGrid
from sympy import pprint
A = Object(A)
B = Object(B)
C = Object(C)
f = NamedMorphism(A, B, f)
g = NamedMorphism(B, C, g)
diagram = Diagram([f, g])
The simplest way to have a diagram laid out is the following:
>>> grid = DiagramGrid(diagram)
>>> (grid.width, grid.height)
(2, 2)
>>> pprint(grid)
A B
C
Sometimes one sees the diagram as consisting of logical groups. One can advise DiagramGrid as to such groups by employing the groups keyword argument.
1838
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Consider the following diagram:
>>>
>>>
>>>
>>>
>>>
>>>
D = Object(D)
f = NamedMorphism(A, B, f)
g = NamedMorphism(B, C, g)
h = NamedMorphism(D, A, h)
k = NamedMorphism(D, B, k)
diagram = Diagram([f, g, h, k])
Lay it out with generic layout:
>>> grid = DiagramGrid(diagram)
>>> pprint(grid)
A B D
C
Now, we can group the objects A and D to have them near one another:
>>> grid = DiagramGrid(diagram, groups=[[A, D], B, C])
>>> pprint(grid)
B
C
A
Note how the positioning of the other objects changes.
Further indications can be supplied to the constructor of DiagramGrid (page 1838) using keyword arguments. The currently supported hints are explained in the following
paragraphs.
DiagramGrid (page 1838) does not automatically guess which layout would suit the supplied diagram better. Consider, for example, the following linear diagram:
>>>
>>>
>>>
>>>
>>>
>>>
E = Object(E)
f = NamedMorphism(A, B, f)
g = NamedMorphism(B, C, g)
h = NamedMorphism(C, D, h)
i = NamedMorphism(D, E, i)
diagram = Diagram([f, g, h, i])
When laid out with the generic layout, it does not get to look linear:
>>> grid = DiagramGrid(diagram)
>>> pprint(grid)
A B
C
D
E
To get it laid out in a line, use layout=sequential:
>>> grid = DiagramGrid(diagram, layout=sequential)
>>> pprint(grid)
A B C D E
One may sometimes need to transpose the resulting layout. While this can always be
done by hand, DiagramGrid (page 1838) provides a hint for that purpose:
5.37. Category Theory Module
1839
SymPy Documentation, Release 0.7.6
>>> grid = DiagramGrid(diagram, layout=sequential, transpose=True)
>>> pprint(grid)
A
B
C
D
E
Separate hints can also be provided for each group.
For an example, refer
to tests/test drawing.py, and see the dierent ways in which the ve lemma
[FiveLemma] can be laid out.
See Also:
Diagram
References
[FiveLemma] http://en.wikipedia.org/wiki/Five lemma
height
Returns the number of rows in this diagram layout.
Examples
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
2
from sympy.categories import Object, NamedMorphism
from sympy.categories import Diagram, DiagramGrid
A = Object(A)
B = Object(B)
C = Object(C)
f = NamedMorphism(A, B, f)
g = NamedMorphism(B, C, g)
diagram = Diagram([f, g])
grid = DiagramGrid(diagram)
grid.height
morphisms
Returns those morphisms (and their properties) which are suciently meaningful
to be drawn.
Examples
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
1840
from sympy.categories import Object, NamedMorphism
from sympy.categories import Diagram, DiagramGrid
A = Object(A)
B = Object(B)
C = Object(C)
f = NamedMorphism(A, B, f)
g = NamedMorphism(B, C, g)
diagram = Diagram([f, g])
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> grid = DiagramGrid(diagram)
>>> grid.morphisms
{NamedMorphism(Object(A), Object(B), f): EmptySet(),
NamedMorphism(Object(B), Object(C), g): EmptySet()}
width
Returns the number of columns in this diagram layout.
Examples
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
2
from sympy.categories import Object, NamedMorphism
from sympy.categories import Diagram, DiagramGrid
A = Object(A)
B = Object(B)
C = Object(C)
f = NamedMorphism(A, B, f)
g = NamedMorphism(B, C, g)
diagram = Diagram([f, g])
grid = DiagramGrid(diagram)
grid.width
class sympy.categories.diagram drawing.ArrowStringDescription(unit,
curving,
curving amount,
looping start,
looping end,
horizontal direction,
vertical direction,
label position,
label)
Stores the information necessary for producing an Xy-pic description of an arrow.
The principal goal of this class is to abstract away the string representation of an arrow
and to also provide the functionality to produce the actual Xy-pic string.
unit sets the unit which will be used to specify the amount of curving and other distances. horizontal direction should be a string of r or l specifying the horizontal
oset of the target cell of the arrow relatively to the current one. vertical direction
should specify the vertical oset using a series of either d or u. label position
should be either , , or | to specify that the label should be positioned above the
arrow, below the arrow or just over the arrow, in a break. Note that the notions above
and below are relative to arrow direction. label stores the morphism label.
This works as follows (disregard the yet unexplained arguments):
>>> from sympy.categories.diagram_drawing import ArrowStringDescription
>>> astr = ArrowStringDescription(
... unit=mm, curving=None, curving_amount=None,
... looping_start=None, looping_end=None, horizontal_direction=d,
... vertical_direction=r, label_position=_, label=f)
>>> print(str(astr))
\ar[dr]_{f}
5.37. Category Theory Module
1841
SymPy Documentation, Release 0.7.6
curving should be one of , to specify in which direction the arrow is going to
curve. curving amount is a number describing how many units the morphism is going
to curve:
>>> astr = ArrowStringDescription(
... unit=mm, curving=^, curving_amount=12,
... looping_start=None, looping_end=None, horizontal_direction=d,
... vertical_direction=r, label_position=_, label=f)
>>> print(str(astr))
\ar@/^12mm/[dr]_{f}
looping start and looping end are currently only used for loop morphisms, those which
have the same domain and codomain. These two attributes should store a valid Xy-pic
direction and specify, correspondingly, the direction the arrow gets out into and the
direction the arrow gets back from:
>>> astr = ArrowStringDescription(
... unit=mm, curving=None, curving_amount=None,
... looping_start=u, looping_end=l, horizontal_direction=,
... vertical_direction=, label_position=_, label=f)
>>> print(str(astr))
\ar@(u,l)[]_{f}
label displacement controls how far the arrow label is from the ends of the arrow. For
example, to position the arrow label near the arrow head, use >:
>>> astr = ArrowStringDescription(
... unit=mm, curving=^, curving_amount=12,
... looping_start=None, looping_end=None, horizontal_direction=d,
... vertical_direction=r, label_position=_, label=f)
>>> astr.label_displacement = >
>>> print(str(astr))
\ar@/^12mm/[dr]_>{f}
Finally, arrow style is used to specify the arrow style. To get a dashed arrow, for example, use {>} as arrow style:
>>> astr = ArrowStringDescription(
... unit=mm, curving=^, curving_amount=12,
... looping_start=None, looping_end=None, horizontal_direction=d,
... vertical_direction=r, label_position=_, label=f)
>>> astr.arrow_style = {-->}
>>> print(str(astr))
\ar@/^12mm/@{-->}[dr]_{f}
See Also:
XypicDiagramDrawer (page 1843)
Notes
Instances of ArrowStringDescription (page 1841) will be constructed by XypicDiagramDrawer (page 1843) and provided for further use in formatters. The user is not
expected to construct instances of ArrowStringDescription (page 1841) themselves.
To be able to properly utilise this class, the reader is encouraged to checkout the Xy-pic
user guide, available at [Xypic].
1842
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
References
[Xypic] http://www.tug.org/applications/Xy-pic/
class sympy.categories.diagram drawing.XypicDiagramDrawer
Given a Diagram and the corresponding DiagramGrid (page 1838), produces the Xy-pic
representation of the diagram.
The most important method in this class is draw. Consider the following triangle diagram:
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
from sympy.categories import Object, NamedMorphism, Diagram
from sympy.categories import DiagramGrid, XypicDiagramDrawer
A = Object(A)
B = Object(B)
C = Object(C)
f = NamedMorphism(A, B, f)
g = NamedMorphism(B, C, g)
diagram = Diagram([f, g], {g * f: unique})
To draw this diagram, its objects need to be laid out with a DiagramGrid (page 1838):
>>> grid = DiagramGrid(diagram)
Finally, the drawing:
>>> drawer = XypicDiagramDrawer()
>>> print(drawer.draw(diagram, grid))
\xymatrix{
A \ar[d]_{g\circ f} \ar[r]^{f} & B \ar[ld]^{g} \\
C &
}
For further details see the docstring of this method.
To control the appearance of the arrows, formatters are used. The dictionary arrow formatters maps morphisms to formatter functions. A formatter is accepts an ArrowStringDescription (page 1841) and is allowed to modify any of the arrow properties
exposed thereby. For example, to have all morphisms with the property unique appear
as dashed arrows, and to have their names prepended with !, the following should be
done:
>>> def formatter(astr):
...
astr.label = \exists ! + astr.label
...
astr.arrow_style = {-->}
>>> drawer.arrow_formatters[unique] = formatter
>>> print(drawer.draw(diagram, grid))
\xymatrix{
A \ar@{-->}[d]_{\exists !g\circ f} \ar[r]^{f} & B \ar[ld]^{g} \\
C &
}
To modify the appearance of all arrows in the diagram, set default arrow formatter.
For example, to place all morphism labels a little bit farther from the arrow head so that
they look more centred, do as follows:
>>> def default_formatter(astr):
...
astr.label_displacement = (0.45)
>>> drawer.default_arrow_formatter = default_formatter
>>> print(drawer.draw(diagram, grid))
5.37. Category Theory Module
1843
SymPy Documentation, Release 0.7.6
\xymatrix{
A \ar@{-->}[d]_(0.45){\exists !g\circ f} \ar[r]^(0.45){f} & B \ar[ld]^(0.45){g} \\
C &
}
In some diagrams some morphisms are drawn as curved arrows. Consider the following
diagram:
>>> D = Object(D)
>>> E = Object(E)
>>> h = NamedMorphism(D, A, h)
>>> k = NamedMorphism(D, B, k)
>>> diagram = Diagram([f, g, h, k])
>>> grid = DiagramGrid(diagram)
>>> drawer = XypicDiagramDrawer()
>>> print(drawer.draw(diagram, grid))
\xymatrix{
A \ar[r]_{f} & B \ar[d]^{g} & D \ar[l]^{k} \ar@/_3mm/[ll]_{h} \\
& C &
}
To control how far the morphisms are curved by default, one can use the unit and default curving amount attributes:
>>> drawer.unit = cm
>>> drawer.default_curving_amount = 1
>>> print(drawer.draw(diagram, grid))
\xymatrix{
A \ar[r]_{f} & B \ar[d]^{g} & D \ar[l]^{k} \ar@/_1cm/[ll]_{h} \\
& C &
}
In some diagrams, there are multiple curved morphisms between the same two objects.
To control by how much the curving changes between two such successive morphisms,
use default curving step:
>>> drawer.default_curving_step = 1
>>> h1 = NamedMorphism(A, D, h1)
>>> diagram = Diagram([f, g, h, k, h1])
>>> grid = DiagramGrid(diagram)
>>> print(drawer.draw(diagram, grid))
\xymatrix{
A \ar[r]_{f} \ar@/^1cm/[rr]^{h_{1}} & B \ar[d]^{g} & D \ar[l]^{k} \ar@/_2cm/[ll]_{h} \\
& C &
}
The default value of default curving step is 4 units.
See Also:
draw (page 1844), ArrowStringDescription (page 1841)
draw(diagram, grid, masked=None, diagram format=)
Returns the Xy-pic representation of diagram laid out in grid.
Consider the following simple triangle diagram.
>>>
>>>
>>>
>>>
1844
from sympy.categories import Object, NamedMorphism, Diagram
from sympy.categories import DiagramGrid, XypicDiagramDrawer
A = Object(A)
B = Object(B)
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>>
>>>
>>>
>>>
C = Object(C)
f = NamedMorphism(A, B, f)
g = NamedMorphism(B, C, g)
diagram = Diagram([f, g], {g * f: unique})
To draw this diagram, its objects need to be laid out with a DiagramGrid (page 1838):
>>> grid = DiagramGrid(diagram)
Finally, the drawing:
>>> drawer = XypicDiagramDrawer()
>>> print(drawer.draw(diagram, grid))
\xymatrix{
A \ar[d]_{g\circ f} \ar[r]^{f} & B \ar[ld]^{g} \\
C &
}
The argument masked can be used to skip morphisms in the presentation of the
diagram:
>>> print(drawer.draw(diagram, grid, masked=[g * f]))
\xymatrix{
A \ar[r]^{f} & B \ar[ld]^{g} \\
C &
}
Finally, the diagram format argument can be used to specify the format string of
the diagram. For example, to increase the spacing by 1 cm, proceeding as follows:
>>> print(drawer.draw(diagram, grid, diagram_format=@+1cm))
\xymatrix@+1cm{
A \ar[d]_{g\circ f} \ar[r]^{f} & B \ar[ld]^{g} \\
C &
}
sympy.categories.diagram drawing.xypic draw diagram(diagram,
masked=None,
diagram format=,
groups=None, **hints)
Provides a shortcut combining DiagramGrid (page 1838) and XypicDiagramDrawer
(page 1843). Returns an Xy-pic presentation of diagram. The argument masked is a
list of morphisms which will be not be drawn. The argument diagram format is the format string inserted after xymatrix. groups should be a set of logical groups. The hints
will be passed directly to the constructor of DiagramGrid (page 1838).
For more information about the arguments, see the docstrings of DiagramGrid
(page 1838) and XypicDiagramDrawer.draw.
See Also:
XypicDiagramDrawer (page 1843), DiagramGrid (page 1838)
Examples
>>>
>>>
>>>
>>>
from sympy.categories import Object, NamedMorphism, Diagram
from sympy.categories import xypic_draw_diagram
A = Object(A)
B = Object(B)
5.37. Category Theory Module
1845
SymPy Documentation, Release 0.7.6
>>> C = Object(C)
>>> f = NamedMorphism(A, B, f)
>>> g = NamedMorphism(B, C, g)
>>> diagram = Diagram([f, g], {g * f: unique})
>>> print(xypic_draw_diagram(diagram))
\xymatrix{
A \ar[d]_{g\circ f} \ar[r]^{f} & B \ar[ld]^{g} \\
C &
}
sympy.categories.diagram drawing.preview diagram(diagram,
masked=None,
diagram format=,
groups=None, output=png,
viewer=None,
euler=True,
**hints)
Combines the functionality of xypic draw diagram and sympy.printing.preview.
The arguments masked, diagram format, groups, and hints are passed to
xypic draw diagram, while output, viewer, and euler are passed to preview.
See Also:
xypic diagram drawer
Examples
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
from sympy.categories import Object, NamedMorphism, Diagram
from sympy.categories import preview_diagram
A = Object(A)
B = Object(B)
C = Object(C)
f = NamedMorphism(A, B, f)
g = NamedMorphism(B, C, g)
d = Diagram([f, g], {g * f: unique})
preview_diagram(d)
5.38 Dierential Geometry Module
5.38.1 Introduction
5.38.2 Base Class Reference
class sympy.diffgeom.Manifold(name, dim)
Object representing a mathematical manifold.
The only role that this object plays is to keep a list of all patches dened on the manifold.
It does not provide any means to study the topological characteristics of the manifold
that it represents.
class sympy.diffgeom.Patch(name, manifold)
Object representing a patch on a manifold.
On a manifold one can have many patches that do not always include the whole manifold.
On these patches coordinate charts can be dened that permit the parametrization of
any point on the patch in terms of a tuple of real numbers (the coordinates).
1846
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
This object serves as a container/parent for all coordinate system charts that can be
dened on the patch it represents.
Examples:
Dene a Manifold and a Patch on that Manifold:
>>> from sympy.diffgeom import Manifold, Patch
>>> m = Manifold(M, 3)
>>> p = Patch(P, m)
>>> p in m.patches
True
class sympy.diffgeom.CoordSystem(name, patch, names=None)
Contains all coordinate transformation logic.
Examples
Dene a Manifold and a Patch, and then dene two coord systems on that patch:
>>> from sympy import symbols, sin, cos, pi
>>> from sympy.diffgeom import Manifold, Patch, CoordSystem
>>> r, theta = symbols(r, theta)
>>> m = Manifold(M, 2)
>>> patch = Patch(P, m)
>>> rect = CoordSystem(rect, patch)
>>> polar = CoordSystem(polar, patch)
>>> rect in patch.coord_systems
True
Connect the coordinate systems. An inverse transformation is automatically found by
solve when possible:
>>> polar.connect_to(rect, [r, theta], [r*cos(theta), r*sin(theta)])
>>> polar.coord_tuple_transform_to(rect, [0, 2])
Matrix([
[0],
[0]])
>>> polar.coord_tuple_transform_to(rect, [2, pi/2])
Matrix([
[0],
[2]])
>>> rect.coord_tuple_transform_to(polar, [1, 1])
Matrix([
[sqrt(2)],
[
pi/4]])
Calculate the jacobian of the polar to cartesian transformation:
>>> polar.jacobian(rect, [r, theta])
Matrix([
[cos(theta), -r*sin(theta)],
[sin(theta), r*cos(theta)]])
Dene a point using coordinates in one of the coordinate systems:
5.38. Dierential Geometry Module
1847
SymPy Documentation, Release 0.7.6
>>> p = polar.point([1, 3*pi/4])
>>> rect.point_to_coords(p)
Matrix([
[-sqrt(2)/2],
[ sqrt(2)/2]])
Dene a basis scalar eld (i.e. a coordinate function), that takes a point and returns its
coordinates. It is an instance of BaseScalarField.
>>> rect.coord_function(0)(p)
-sqrt(2)/2
>>> rect.coord_function(1)(p)
sqrt(2)/2
Dene a basis vector eld (i.e. a unit vector eld along the coordinate line). Vectors are
also dierential operators on scalar elds. It is an instance of BaseVectorField.
>>>
>>>
>>>
1
>>>
0
v_x = rect.base_vector(0)
x = rect.coord_function(0)
v_x(x)
v_x(v_x(x))
Dene a basis oneform eld:
>>> dx = rect.base_oneform(0)
>>> dx(v_x)
1
If you provide a list of names the elds will print nicely: - without provided names:
>>> x, v_x, dx
(rect_0, e_rect_0, drect_0)
with provided names
>>> rect = CoordSystem(rect, patch, [x, y])
>>> rect.coord_function(0), rect.base_vector(0), rect.base_oneform(0)
(x, e_x, dx)
base oneform(coord index)
Return a basis 1-form eld.
The basis one-form eld for this coordinate system. It is also an operator on vector
elds.
See the docstring of CoordSystem for examples.
base oneforms()
Returns a list of all base oneforms.
For more details see the base oneform method of this class.
base vector(coord index)
Return a basis vector eld.
The basis vector eld for this coordinate system. It is also an operator on scalar
elds.
See the docstring of CoordSystem for examples.
1848
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
base vectors()
Returns a list of all base vectors.
For more details see the base vector method of this class.
connect to(to sys, from coords, to exprs, inverse=True, ll in gaps=False)
Register the transformation used to switch to another coordinate system.
Parameters to sys :
another instance of CoordSystem
from coords :
list of symbols in terms of which to exprs is given
to exprs :
list of the expressions of the new coordinate tuple
inverse :
try to deduce and register the inverse transformation
ll in gaps :
try to deduce other transformation that are made possible by composing the present transformation with other already registered transformation
coord function(coord index)
Return a BaseScalarField that takes a point and returns one of the coords.
Takes a point and returns its coordinate in this coordinate system.
See the docstring of CoordSystem for examples.
coord functions()
Returns a list of all coordinate functions.
For more details see the coord function method of this class.
coord tuple transform to(to sys, coords)
Transform coords to coord system to sys.
See the docstring of CoordSystem for examples.
jacobian(to sys, coords)
Return the jacobian matrix of a transformation.
point(coords)
Create a Point with coordinates given in this coord system.
See the docstring of CoordSystem for examples.
point to coords(point)
Calculate the coordinates of a point in this coord system.
See the docstring of CoordSystem for examples.
class sympy.diffgeom.Point(coord sys, coords)
Point in a Manifold object.
To dene a point you must supply coordinates and a coordinate system.
The usage of this object after its denition is independent of the coordinate system that
was used in order to dene it, however due to limitations in the simplication routines
you can arrive at complicated expressions if you use inappropriate coordinate systems.
5.38. Dierential Geometry Module
1849
SymPy Documentation, Release 0.7.6
Examples
Dene the boilerplate Manifold, Patch and coordinate systems:
>>>
>>>
...
>>>
>>>
>>>
>>>
>>>
>>>
from sympy import symbols, sin, cos, pi
from sympy.diffgeom import (
Manifold, Patch, CoordSystem, Point)
r, theta = symbols(r, theta)
m = Manifold(M, 2)
p = Patch(P, m)
rect = CoordSystem(rect, p)
polar = CoordSystem(polar, p)
polar.connect_to(rect, [r, theta], [r*cos(theta), r*sin(theta)])
Dene a point using coordinates from one of the coordinate systems:
>>> p = Point(polar, [r, 3*pi/4])
>>> p.coords()
Matrix([
[
r],
[3*pi/4]])
>>> p.coords(rect)
Matrix([
[-sqrt(2)*r/2],
[ sqrt(2)*r/2]])
coords(to sys=None)
Coordinates of the point in a given coordinate system.
If to sys is None it returns the coordinates in the system in which the point was
dened.
class sympy.diffgeom.BaseScalarField(coord sys, index)
Base Scalar Field over a Manifold for a given Coordinate System.
A scalar eld takes a point as an argument and returns a scalar.
A base scalar eld of a coordinate system takes a point and returns one of the coordinates
of that point in the coordinate system in question.
To dene a scalar eld you need to choose the coordinate system and the index of the
coordinate.
The use of the scalar eld after its denition is independent of the coordinate system in
which it was dened, however due to limitations in the simplication routines you may
arrive at more complicated expression if you use unappropriate coordinate systems.
You can build complicated scalar elds by just building up SymPy expressions containing
BaseScalarField instances.
Examples
Dene boilerplate Manifold, Patch and coordinate systems:
>>>
>>>
...
>>>
>>>
>>>
1850
from sympy import symbols, sin, cos, pi, Function
from sympy.diffgeom import (
Manifold, Patch, CoordSystem, Point, BaseScalarField)
r0, theta0 = symbols(r0, theta0)
m = Manifold(M, 2)
p = Patch(P, m)
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>> rect = CoordSystem(rect, p)
>>> polar = CoordSystem(polar, p)
>>> polar.connect_to(rect, [r0, theta0], [r0*cos(theta0), r0*sin(theta0)])
Point to be used as an argument for the led:
>>> point = polar.point([r0, 0])
Examples of elds:
>>> fx = BaseScalarField(rect, 0)
>>> fy = BaseScalarField(rect, 1)
>>> (fx**2+fy**2).rcall(point)
r0**2
>>> g = Function(g)
>>> ftheta = BaseScalarField(polar, 1)
>>> fg = g(ftheta-pi)
>>> fg.rcall(point)
g(-pi)
class sympy.diffgeom.BaseVectorField(coord sys, index)
Vector Field over a Manifold.
A vector eld is an operator taking a scalar eld and returning a directional derivative
(which is also a scalar eld).
A base vector eld is the same type of operator, however the derivation is specically
done with respect to a chosen coordinate.
To dene a base vector eld you need to choose the coordinate system and the index of
the coordinate.
The use of the vector eld after its denition is independent of the coordinate system in
which it was dened, however due to limitations in the simplication routines you may
arrive at more complicated expression if you use unappropriate coordinate systems.
Examples
Use the predened R2 manifold, setup some boilerplate.
>>>
>>>
>>>
>>>
>>>
from sympy import symbols, pi, Function
from sympy.diffgeom.rn import R2, R2_p, R2_r
from sympy.diffgeom import BaseVectorField
from sympy import pprint
x0, y0, r0, theta0 = symbols(x0, y0, r0, theta0)
Points to be used as arguments for the eld:
>>> point_p = R2_p.point([r0, theta0])
>>> point_r = R2_r.point([x0, y0])
Scalar eld to operate on:
>>> g = Function(g)
>>> s_field = g(R2.x, R2.y)
>>> s_field.rcall(point_r)
g(x0, y0)
5.38. Dierential Geometry Module
1851
SymPy Documentation, Release 0.7.6
>>> s_field.rcall(point_p)
g(r0*cos(theta0), r0*sin(theta0))
Vector eld:
>>> v = BaseVectorField(R2_r, 1)
>>> pprint(v(s_field))
/ d
\|
|-----(g(x, xi_2))||
\dxi_2
/|xi_2=y
>>> pprint(v(s_field).rcall(point_r).doit())
d
---(g(x0, y0))
dy0
>>> pprint(v(s_field).rcall(point_p).doit())
/ d
\|
|-----(g(r0*cos(theta0), xi_2))||
\dxi_2
/|xi_2=r0*sin(theta0)
class sympy.diffgeom.Commutator(v1, v2)
Commutator of two vector elds.
The commutator of two vector elds v1 and v2 is dened as the vector eld [v1 , v2 ] that
evaluated on each scalar eld f is equal to v1 (v2 (f )) v2 (v1 (f )).
Examples
Use the predened R2 manifold, setup some boilerplate.
>>>
>>>
>>>
>>>
from
from
from
from
sympy.diffgeom.rn import R2
sympy.diffgeom import Commutator
sympy import pprint
sympy.simplify import simplify
Vector elds:
>>>
>>>
>>>
>>>
0
e_x, e_y, e_r = R2.e_x, R2.e_y, R2.e_r
c_xy = Commutator(e_x, e_y)
c_xr = Commutator(e_x, e_r)
c_xy
Unfortunately, the current code is not able to compute everything:
>>> c_xr
Commutator(e_x, e_r)
class sympy.diffgeom.Differential(form eld)
Return the dierential (exterior derivative) of a form eld.
The dierential of a form (i.e. the exterior derivative) has a complicated denition in the
general case.
The dierential df of the 0-form f is dened for any vector eld v as df (v) = v(f ).
Examples
Use the predened R2 manifold, setup some boilerplate.
1852
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
>>>
>>>
>>>
>>>
from
from
from
from
sympy import Function
sympy.diffgeom.rn import R2
sympy.diffgeom import Differential
sympy import pprint
Scalar eld (0-forms):
>>> g = Function(g)
>>> s_field = g(R2.x, R2.y)
Vector elds:
>>> e_x, e_y, = R2.e_x, R2.e_y
Dierentials:
>>> dg = Differential(s_field)
>>> dg
d(g(x, y))
>>> pprint(dg(e_x))
/ d
\|
|-----(g(xi_1, y))||
\dxi_1
/|xi_1=x
>>> pprint(dg(e_y))
/ d
\|
|-----(g(x, xi_2))||
\dxi_2
/|xi_2=y
Applying the exterior derivative operator twice always results in:
>>> Differential(dg)
0
class sympy.diffgeom.TensorProduct(*args)
Tensor product of forms.
The tensor product permits the creation of multilinear functionals (i.e. higher order tensors) out of lower order forms (e.g. 1-forms). However, the higher tensors thus created
lack the interesting features provided by the other type of product, the wedge product,
namely they are not antisymmetric and hence are not form elds.
Examples
Use the predened R2 manifold, setup some boilerplate.
>>>
>>>
>>>
>>>
from
from
from
from
sympy import Function
sympy.diffgeom.rn import R2
sympy.diffgeom import TensorProduct
sympy import pprint
>>> TensorProduct(R2.dx, R2.dy)(R2.e_x, R2.e_y)
1
>>> TensorProduct(R2.dx, R2.dy)(R2.e_y, R2.e_x)
0
>>> TensorProduct(R2.dx, R2.x*R2.dy)(R2.x*R2.e_x, R2.e_y)
x**2
You can nest tensor products.
5.38. Dierential Geometry Module
1853
SymPy Documentation, Release 0.7.6
>>> tp1 = TensorProduct(R2.dx, R2.dy)
>>> TensorProduct(tp1, R2.dx)(R2.e_x, R2.e_y, R2.e_x)
1
You can make partial contraction for instance when raising an index. Putting None in
the second argument of rcall means that the respective position in the tensor product
is left as it is.
>>> TP = TensorProduct
>>> metric = TP(R2.dx, R2.dx) + 3*TP(R2.dy, R2.dy)
>>> metric.rcall(R2.e_y, None)
3*dy
Or automatically pad the args with None without specifying them.
>>> metric.rcall(R2.e_y)
3*dy
class sympy.diffgeom.WedgeProduct(*args)
Wedge product of forms.
In the context of integration only completely antisymmetric forms make sense. The
wedge product permits the creation of such forms.
Examples
Use the predened R2 manifold, setup some boilerplate.
>>>
>>>
>>>
>>>
from
from
from
from
sympy import Function
sympy.diffgeom.rn import R2
sympy.diffgeom import WedgeProduct
sympy import pprint
>>> WedgeProduct(R2.dx, R2.dy)(R2.e_x, R2.e_y)
1
>>> WedgeProduct(R2.dx, R2.dy)(R2.e_y, R2.e_x)
-1
>>> WedgeProduct(R2.dx, R2.x*R2.dy)(R2.x*R2.e_x, R2.e_y)
x**2
You can nest wedge products.
>>> wp1 = WedgeProduct(R2.dx, R2.dy)
>>> WedgeProduct(wp1, R2.dx)(R2.e_x, R2.e_y, R2.e_x)
0
class sympy.diffgeom.LieDerivative(v eld, expr)
Lie derivative with respect to a vector eld.
The transport operator that denes the Lie derivative is the pushforward of the eld to
be derived along the integral curve of the eld with respect to which one derives.
Examples
>>> from sympy.diffgeom import (LieDerivative, TensorProduct)
>>> from sympy.diffgeom.rn import R2
>>> LieDerivative(R2.e_x, R2.y)
1854
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
0
>>> LieDerivative(R2.e_x, R2.x)
1
>>> LieDerivative(R2.e_x, R2.e_x)
0
The Lie derivative of a tensor eld by another tensor eld is equal to their commutator:
>>> LieDerivative(R2.e_x, R2.e_r)
Commutator(e_x, e_r)
>>> LieDerivative(R2.e_x + R2.e_y, R2.x)
1
>>> tp = TensorProduct(R2.dx, R2.dy)
>>> LieDerivative(R2.e_x, tp)
LieDerivative(e_x, TensorProduct(dx, dy))
>>> LieDerivative(R2.e_x, tp).doit()
LieDerivative(e_rectangular_0, TensorProduct(dx, dy))
class sympy.diffgeom.BaseCovarDerivativeOp(coord sys, index, christoel)
Covariant derivative operator with respect to a base vector.
Examples
>>> from sympy.diffgeom.rn import R2, R2_r
>>> from sympy.diffgeom import BaseCovarDerivativeOp
>>> from sympy.diffgeom import metric_to_Christoffel_2nd, TensorProduct
>>> TP = TensorProduct
>>> ch = metric_to_Christoffel_2nd(TP(R2.dx, R2.dx) + TP(R2.dy, R2.dy))
>>> ch
(((0, 0), (0, 0)), ((0, 0), (0, 0)))
>>> cvd = BaseCovarDerivativeOp(R2_r, 0, ch)
>>> cvd(R2.x)
1
>>> cvd(R2.x*R2.e_x)
e_x
class sympy.diffgeom.CovarDerivativeOp(wrt, christoel)
Covariant derivative operator.
Examples
>>> from sympy.diffgeom.rn import R2
>>> from sympy.diffgeom import CovarDerivativeOp
>>> from sympy.diffgeom import metric_to_Christoffel_2nd, TensorProduct
>>> TP = TensorProduct
>>> ch = metric_to_Christoffel_2nd(TP(R2.dx, R2.dx) + TP(R2.dy, R2.dy))
>>> ch
(((0, 0), (0, 0)), ((0, 0), (0, 0)))
>>> cvd = CovarDerivativeOp(R2.x*R2.e_x, ch)
>>> cvd(R2.x)
x
>>> cvd(R2.x*R2.e_x)
x*e_x
5.38. Dierential Geometry Module
1855
SymPy Documentation, Release 0.7.6
sympy.diffgeom.intcurve series(vector eld,
param,
start point,
ord sys=None, coes=False)
Return the series expansion for an integral curve of the eld.
n=6,
co-
Integral curve is a function taking a parameter in R to a point in the manifold. It veries
the equation:
(
)
(
)
d
V (f ) (t) = dt
f (t)
where the given vector field is denoted as V . This holds for any value t for the parameter and any scalar eld f .
This equation can also be decomposed of a basis of coordinate functions
(
)
(
)
d
fi (t) i
V (fi ) (t) = dt
This function returns a series expansion of (t) in terms of the coordinate system coord sys. The equations and expansions are necessarily done in coordinate-systemdependent way as there is no other way to represent movement between points on the
manifold (i.e. there is no such thing as a dierence of points for a general manifold).
Parameters vector eld :
the vector eld for which an integral curve will be given
param :
the argument of the function from R to the curve
start point :
the point which coresponds to (0)
n:
the order to which to expand
coord sys :
the coordinate system in which to expand coes (default False) - if
True return a list of elements of the expansion
See Also:
intcurve diffequ (page 1857)
Examples
Use the predened R2 manifold:
>>> from sympy.abc import t, x, y
>>> from sympy.diffgeom.rn import R2, R2_p, R2_r
>>> from sympy.diffgeom import intcurve_series
Specify a starting point and a vector eld:
>>> start_point = R2_r.point([x, y])
>>> vector_field = R2_r.e_x
Calculate the series:
>>> intcurve_series(vector_field, t, start_point, n=3)
Matrix([
[t + x],
[
y]])
1856
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Or get the elements of the expansion in a list:
>>> series = intcurve_series(vector_field, t, start_point, n=3, coeffs=True)
>>> series[0]
Matrix([
[x],
[y]])
>>> series[1]
Matrix([
[t],
[0]])
>>> series[2]
Matrix([
[0],
[0]])
The series in the polar coordinate system:
>>> series = intcurve_series(vector_field, t, start_point,
...
n=3, coord_sys=R2_p, coeffs=True)
>>> series[0]
Matrix([
[sqrt(x**2 + y**2)],
[
atan2(y, x)]])
>>> series[1]
Matrix([
[t*x/sqrt(x**2 + y**2)],
[
-t*y/(x**2 + y**2)]])
>>> series[2]
Matrix([
[t**2*(-x**2/(x**2 + y**2)**(3/2) + 1/sqrt(x**2 + y**2))/2],
[
t**2*x*y/(x**2 + y**2)**2]])
sympy.diffgeom.intcurve diffequ(vector eld, param, start point, coord sys=None)
Return the dierential equation for an integral curve of the eld.
Integral curve is a function taking a parameter in R to a point in the manifold. It veries
the equation:
(
)
(
)
d
V (f ) (t) = dt
f (t)
where the given vector field is denoted as V . This holds for any value t for the parameter and any scalar eld f .
This function returns the dierential equation of (t) in terms of the coordinate system
coord sys. The equations and expansions are necessarily done in coordinate-systemdependent way as there is no other way to represent movement between points on the
manifold (i.e. there is no such thing as a dierence of points for a general manifold).
Parameters vector eld :
the vector eld for which an integral curve will be given
param :
the argument of the function from R to the curve
start point :
the point which coresponds to (0)
coord sys :
the coordinate system in which to give the equations
5.38. Dierential Geometry Module
1857
SymPy Documentation, Release 0.7.6
Returns a tuple of (equations, initial conditions) :
See Also:
intcurve series (page 1855)
Examples
Use the predened R2 manifold:
>>> from sympy.abc import t
>>> from sympy.diffgeom.rn import R2, R2_p, R2_r
>>> from sympy.diffgeom import intcurve_diffequ
Specify a starting point and a vector eld:
>>> start_point = R2_r.point([0, 1])
>>> vector_field = -R2.y*R2.e_x + R2.x*R2.e_y
Get the equation:
>>> equations, init_cond = intcurve_diffequ(vector_field, t, start_point)
>>> equations
[f_1(t) + Derivative(f_0(t), t), -f_0(t) + Derivative(f_1(t), t)]
>>> init_cond
[f_0(0), f_1(0) - 1]
The series in the polar coordinate system:
>>> equations, init_cond = intcurve_diffequ(vector_field, t, start_point, R2_p)
>>> equations
[Derivative(f_0(t), t), Derivative(f_1(t), t) - 1]
>>> init_cond
[f_0(0) - 1, f_1(0) - pi/2]
sympy.diffgeom.vectors in basis(expr, to sys)
Transform all base vectors in base vectors of a specied coord basis.
While the new base vectors are in the new coordinate system basis, any coecients are
kept in the old system.
Examples
>>> from sympy.diffgeom import vectors_in_basis
>>> from sympy.diffgeom.rn import R2_r, R2_p
>>> vectors_in_basis(R2_r.e_x, R2_p)
(x**2 + y**2)**(-1/2)*x*e_r - y*(x**2 + y**2)**(-1)*e_theta
>>> vectors_in_basis(R2_p.e_r, R2_r)
sin(theta)*e_y + cos(theta)*e_x
sympy.diffgeom.twoform to matrix(expr)
Return the matrix representing the twoform.
For the twoform w return the matrix M such that M [i, j] = w(ei , ej ), where ei is the i-th
base vector eld for the coordinate system in which the expression of w is given.
1858
Chapter 5. SymPy Modules Reference
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy.diffgeom.rn import R2
>>> from sympy.diffgeom import twoform_to_matrix, TensorProduct
>>> TP = TensorProduct
>>> twoform_to_matrix(TP(R2.dx, R2.dx) + TP(R2.dy, R2.dy))
Matrix([
[1, 0],
[0, 1]])
>>> twoform_to_matrix(R2.x*TP(R2.dx, R2.dx) + TP(R2.dy, R2.dy))
Matrix([
[x, 0],
[0, 1]])
>>> twoform_to_matrix(TP(R2.dx, R2.dx) + TP(R2.dy, R2.dy) - TP(R2.dx, R2.dy)/2)
Matrix([
[
1, 0],
[-1/2, 1]])
sympy.diffgeom.metric to Christoffel 1st(expr)
Return the nested list of Christoel symbols for the given metric.
This returns the Christoel symbol of rst kind that represents the Levi-Civita connection
for the given metric.
Examples
>>> from sympy.diffgeom.rn import R2
>>> from sympy.diffgeom import metric_to_Christoffel_1st, TensorProduct
>>> TP = TensorProduct
>>> metric_to_Christoffel_1st(TP(R2.dx, R2.dx) + TP(R2.dy, R2.dy))
(((0, 0), (0, 0)), ((0, 0), (0, 0)))
>>> metric_to_Christoffel_1st(R2.x*TP(R2.dx, R2.dx) + TP(R2.dy, R2.dy))
(((1/2, 0), (0, 0)), ((0, 0), (0, 0)))
sympy.diffgeom.metric to Christoffel 2nd(expr)
Return the nested list of Christoel symbols for the given metric.
This returns the Christoel symbol of second kind that represents the Levi-Civita connection for the given metric.
Examples
>>> from sympy.diffgeom.rn import R2
>>> from sympy.diffgeom import metric_to_Christoffel_2nd, TensorProduct
>>> TP = TensorProduct
>>> metric_to_Christoffel_2nd(TP(R2.dx, R2.dx) + TP(R2.dy, R2.dy))
(((0, 0), (0, 0)), ((0, 0), (0, 0)))
>>> metric_to_Christoffel_2nd(R2.x*TP(R2.dx, R2.dx) + TP(R2.dy, R2.dy))
(((x**(-1)/2, 0), (0, 0)), ((0, 0), (0, 0)))
sympy.diffgeom.metric to Riemann components(expr)
Return the components of the Riemann tensor expressed in a given basis.
Given a metric it calculates the components of the Riemann tensor in the canonical basis
of the coordinate system in which the metric expression is given.
5.38. Dierential Geometry Module
1859
SymPy Documentation, Release 0.7.6
Examples
>>> from sympy import pprint, exp
>>> from sympy.diffgeom.rn import R2
>>> from sympy.diffgeom import metric_to_Riemann_components, TensorProduct
>>> TP = TensorProduct
>>> metric_to_Riemann_components(TP(R2.dx, R2.dx) + TP(R2.dy, R2.dy))
((((0, 0), (0, 0)), ((0, 0), (0, 0))), (((0, 0), (0, 0)), ((0, 0),
(0, 0))))
>>> non_trivial_metric = exp(2*R2.r)*TP(R2.dr, R2.dr) +
R2.r**2*TP(R2.dtheta, R2.dtheta)
>>> non_trivial_metric
exp(2*r)*TensorProduct(dr, dr) + r**2*TensorProduct(dtheta, dtheta)
>>> riemann = metric_to_Riemann_components(non_trivial_metric)
>>> riemann[0]
(((0, 0), (0, 0)), ((0, -exp(-2*r)*r + 2*r*exp(-2*r)),
(exp(-2*r)*r - 2*r*exp(-2*r), 0)))
>>> riemann[1]
(((0, -r**(-1)), (r**(-1), 0)), ((0, 0), (0, 0)))
sympy.diffgeom.metric to Ricci components(expr)
Return the components of the Ricci tensor expressed in a given basis.
Given a metric it calculates the components of the Ricci tensor in the canonical basis of
the coordinate system in which the metric expression is given.
Examples
>>> from sympy import pprint, exp
>>> from sympy.diffgeom.rn import R2
>>> from sympy.diffgeom import metric_to_Ricci_components, TensorProduct
>>> TP = TensorProduct
>>> metric_to_Ricci_components(TP(R2.dx, R2.dx) + TP(R2.dy, R2.dy))
((0, 0), (0, 0))
>>> non_trivial_metric = exp(2*R2.r)*TP(R2.dr, R2.dr) +
>>> non_trivial_metric
exp(2*r)*TensorProduct(dr, dr) + r**2*TensorProduct(dtheta, dtheta)
>>> metric_to_Ricci_components(non_trivial_metric) #TODO why is this not simpler
((r**(-1), 0), (0, -exp(-2*r)*r + 2*r*exp(-2*r)))
5.39 Contributions to docs
All contributions are welcome. If youd like to improve something, look into the sources if
they contain the information you need (if not, please x them), otherwise the documentation
generation needs to be improved (look in the doc/ directory).
1860
Chapter 5. SymPy Modules Reference
R2.r**2*TP(
CHAPTER
SIX
DEVELOPMENT TIPS:
COMPARISONS IN PYTHON
6.1 Introduction
When debugging comparisons and hashes in SymPy, it is necessary to understand when exactly Python calls each method. Unfortunately, the ocial Python documentation for this is
not very detailed (see the docs for rich comparison, cmp () and hash () methods).
We wrote this guide to ll in the missing gaps. After reading it, you should be able to understand which methods do (and do not) get called and the order in which they are called.
6.2 Hashing
Every Python class has a
hash () method, the default implementation of which is:
def __hash__(self):
return id(self)
You can reimplement it to return a dierent integer that you compute on your own. hash(x)
just calls x. hash (). Python builtin classes usually redene the hash () method. For
example, an int has something like this:
def __hash__(self):
return int(self)
and a list does something like this:
def __hash__(self):
raise TypeError(list objects are unhashable)
The general idea about hashes is that if two objects have a dierent hash, they are not equal,
but if they have the same hash, they might be equal. (This is usually called a hash collision
and you need to use the methods described in the next section to determine if the objects
really are equal).
The only requirement from the Python side is that the hash value mustnt change after it is
returned by the hash () method.
Please be aware that hashing is platform-dependent. This means that you can get dierent
hashes for the same SymPy object on dierent platforms. This aects for instance sorting of
sympy expressions. You can also get SymPy objects printed in dierent order.
1861
SymPy Documentation, Release 0.7.6
When developing, you have to be careful about this, especially when writing tests. It is possible that your test runs on a 32-bit platform, but not on 64-bit. An example:
>> from sympy import *
>> x = Symbol(x)
>> r = rootfinding.roots quartic(Poly(x**4 - 6*x**3 + 17*x**2 - 26*x + 20, x))
>> [i.evalf(2) for i in r]
[1.0 + 1.7*I, 2.0 - 1.0*I, 2.0 + I, 1.0 - 1.7*I]
If you get this order of solutions, you are probably running 32-bit system. On a 64-bit system
you would get the following:
>> [i.evalf(2) for i in r]
[1.0 - 1.7*I, 1.0 + 1.7*I, 2.0 + I, 2.0 - 1.0*I
When you now write a test like this:
r = [i.evalf(2) for i in r]
assert r == [1.0 + 1.7*I, 2.0 - 1.0*I, 2.0 + I, 1.0 - 1.7*I]
it will fail on a 64-bit platforms, even if it works for your 32-bit system. You can avoid this by
using the sorted() or set() Python built-in:
r = [i.evalf(2) for i in r]
assert set(r) == set([1.0 + 1.7*I, 2.0 - 1.0*I, 2.0 + I, 1.0 - 1.7*I])
This approach does not work for doctests since they always compare strings that would be
printed after a prompt. In that case you could make your test print results using a combination
of str() and sorted():
>> sorted([str(i.evalf(2)) for i in r])
[1.0 + 1.7*I, 1.0 - 1.7*I, 2.0 + I, 2.0 - 1.0*I]
or, if you dont want to show the values as strings, then sympify the results or the sorted list:
>> [S(s) for s in sorted([str(i.evalf(2)) for i in r])]
[1.0 + 1.7*I, 1.0 - 1.7*I, 2.0 + I, 2.0 - I]
The printing of SymPy expressions might be also aected, so be careful with doctests. If you
get the following on a 32-bit system:
>> print dsolve(f(x).diff(x, 2) + 2*f(x).diff(x) - f(x), f(x))
f(x) == C1*exp(-x + x*sqrt(2)) + C2*exp(-x - x*sqrt(2))
you might get the following on a 64-bit platform:
>> print dsolve(f(x).diff(x, 2) + 2*f(x).diff(x) - f(x), f(x))
f(x) == C1*exp(-x - x*sqrt(2)) + C2*exp(-x + x*sqrt(2))
6.3 Method Resolution
Let a, b and c be instances of any one of the Python classes. As can be easily checked by the
Python script (page 1864) at the end of this guide, if you write:
a == b
Python calls the following in this order:
1862
Chapter 6. Development Tips: Comparisons in Python
SymPy Documentation, Release 0.7.6
a.__eq__(b)
b.__eq__(a)
a.__cmp__(b)
b.__cmp__(a)
id(a) == id(b)
If a particular method is not implemented (or a method returns NotImplemented 1 ) Python
skips it and tries the next one until it succeeds (i.e. until the method returns something
meaningful). The last line is a catch-all method that always succeeds.
If you write:
a != b
Python tries to call:
a.__ne__(b)
b.__ne__(a)
a.__cmp__(b)
b.__cmp__(a)
id(a) == id(b)
If you write:
a < b
Python tries to call:
a.__lt__(b)
b.__gt__(a)
a.__cmp__(b)
b.__cmp__(a)
id(a) < id(b)
If you write:
a <= b
Python tries to call:
a.__le__(b)
b.__ge__(a)
a.__cmp__(b)
b.__cmp__(a)
id(a) <= id(b)
And similarly for a > b and a >= b.
If you write:
sorted([a, b, c])
Python calls the same chain of methods as for the b < a and c < b comparisons.
If you write any of the following:
1 There is also the similar NotImplementedError exception, which one may be tempted to raise to obtain the same
eect as returning NotImplemented.
But these are not the same, and Python will completely ignore NotImplementedError with respect to choosing
appropriate comparison method, and will just propagate this exception upwards, to the caller.
So return NotImplemented is not the same as raise NotImplementedError.
6.3. Method Resolution
1863
SymPy Documentation, Release 0.7.6
a in {d: 5}
a in set([d, d, d])
set([a, b]) == set([a, b])
Python rst compares hashes, e.g.:
a.__hash__()
d.__hash__()
If hash(a) != hash(d) then the result of the statement a in {d: 5} is immediately False
(remember how hashes work in general). If hash(a) == hash(d)) Python goes through the
method resolution of the == operator as shown above.
6.4 General Notes and Caveats
In the method resolution for <, <=, ==, !=, >=, > and sorted([a, b, c]) operators the
hash () method is not called, so in these cases it doesnt matter what it returns. The
hash () method is only called for sets and dictionaries.
In the ocial Python documentation you can read about hashable and non-hashable objects.
In reality, you dont have to think about it, you just follow the method resolution described
here. E.g. if you try to use lists as dictionary keys, the lists hash () method will be called
and it returns an exception.
In SymPy, every instance of any subclass of Basic is immutable. Technically this means,
that its behavior through all the methods above mustnt change once the instance is created.
Especially, the hash value mustnt change (as already stated above) or else objects will get
mixed up in dictionaries and wrong values will be returned for a given key, etc....
6.5 Script To Verify This Guide
The above method resolution can be veried using the following program:
class A(object):
def __init__(self, a, hash):
self.a = a
self._hash = hash
def __lt__(self, o):
print %s.__lt__(%s) % (self.a, o.a)
return NotImplemented
def __le__(self, o):
print %s.__le__(%s) % (self.a, o.a)
return NotImplemented
def __gt__(self, o):
print %s.__gt__(%s) % (self.a, o.a)
return NotImplemented
def __ge__(self, o):
print %s.__ge__(%s) % (self.a, o.a)
1864
Chapter 6. Development Tips: Comparisons in Python
SymPy Documentation, Release 0.7.6
return NotImplemented
def __cmp__(self, o):
print %s.__cmp__(%s) % (self.a, o.a)
#return cmp(self._hash, o._hash)
return NotImplemented
def __eq__(self, o):
print %s.__eq__(%s) % (self.a, o.a)
return NotImplemented
def __ne__(self, o):
print %s.__ne__(%s) % (self.a, o.a)
return NotImplemented
def __hash__(self):
print %s.__hash__() % (self.a)
return self._hash
def show(s):
print --- %s % s + -*40
eval(s)
a
b
c
d
=
=
=
=
A(a,
A(b,
A(c,
A(d,
1)
2)
3)
1)
show(a == b)
show(a != b)
show(a < b)
show(a <= b)
show(a > b)
show(a >= b)
show(sorted([a, b, c]))
show({d: 5})
show(a in {d: 5})
show(set([d, d, d]))
show(a in set([d, d, d]))
show(set([a, b]))
print --- x = set([a, b]); y = set([a, b]); ---
x = set([a, b])
y = set([a, b])
print
x == y :
x == y
print --- x = set([a, b]); y = set([b, d]); ---
x = set([a, b])
y = set([b, d])
print
x == y :
x == y
and its output:
--- a == b ---------------------------------------a. eq (b)
b. eq (a)
6.5. Script To Verify This Guide
1865
SymPy Documentation, Release 0.7.6
a. cmp (b)
b. cmp (a)
--- a != b ---------------------------------------a. ne (b)
b. ne (a)
a. cmp (b)
b. cmp (a)
--- a < b ---------------------------------------a. lt (b)
b. gt (a)
a. cmp (b)
b. cmp (a)
--- a <= b ---------------------------------------a. le (b)
b. ge (a)
a. cmp (b)
b. cmp (a)
--- a > b ---------------------------------------a. gt (b)
b. lt (a)
a. cmp (b)
b. cmp (a)
--- a >= b ---------------------------------------a. ge (b)
b. le (a)
a. cmp (b)
b. cmp (a)
--- sorted([a, b, c]) ---------------------------------------b. lt (a)
a. gt (b)
b. cmp (a)
a. cmp (b)
c. lt (b)
b. gt (c)
c. cmp (b)
b. cmp (c)
--- {d: 5} ---------------------------------------d. hash ()
--- a in {d: 5} ---------------------------------------d. hash ()
a. hash ()
d. eq (a)
a. eq (d)
d. cmp (a)
a. cmp (d)
--- set([d, d, d]) ---------------------------------------d. hash ()
d. hash ()
d. hash ()
--- a in set([d, d, d]) ---------------------------------------d. hash ()
d. hash ()
d. hash ()
a. hash ()
d. eq (a)
a. eq (d)
d. cmp (a)
a. cmp (d)
1866
Chapter 6. Development Tips: Comparisons in Python
SymPy Documentation, Release 0.7.6
--- set([a, b]) ---------------------------------------a. hash ()
b. hash ()
--- x = set([a, b]); y = set([a, b]); --a. hash ()
b. hash ()
a. hash ()
b. hash ()
x == y :
--- x = set([a, b]); y = set([b, d]); --a. hash ()
b. hash ()
b. hash ()
d. hash ()
x == y :
d. eq (a)
a. eq (d)
d. cmp (a)
a. cmp (d)
6.5. Script To Verify This Guide
1867
SymPy Documentation, Release 0.7.6
1868
Chapter 6. Development Tips: Comparisons in Python
CHAPTER
SEVEN
WIKI
SymPy has a public wiki located at http://wiki.sympy.org. Users should feel free to contribute
to this wiki anything interesting/useful.
7.1 FAQ
FAQ is one of the most useful wiki pages. It has answers to frequently-asked questions.
1869
SymPy Documentation, Release 0.7.6
1870
Chapter 7. Wiki
CHAPTER
EIGHT
SYMPY PAPERS
A list of papers which either use or mention SymPy, as well as presentations given about
SymPy at conferences can be seen at SymPy Papers.
1871
SymPy Documentation, Release 0.7.6
1872
Chapter 8. SymPy Papers
CHAPTER
NINE
PLANET SYMPY
We have a blog aggregator at http://planet.sympy.org.
1873
SymPy Documentation, Release 0.7.6
1874
Chapter 9. Planet SymPy
CHAPTER
TEN
SYMPY LOGOS
SymPy has a collection of ocial logos, which can be generated from sympy.svg < https :
//github.com/sympy/sympy/tree/master/doc/src/sympy.svg > in your local copy of SymPy by:
$ cd doc
$ make logo # will be stored in the build/logo subdirectory
The license of all the logos is the same as SymPy: BSD. See the LICENSE le in the trunk for
more information.
1875
SymPy Documentation, Release 0.7.6
1876
Chapter 10. SymPy logos
CHAPTER
ELEVEN
PROJECTS USING SYMPY
This is an (incomplete) list of projects that use SymPy. If you use SymPy in your project,
please let us know on our mailinglist, so that we can add your project here as well.
SfePy (simple nite elements in Python)
Quameon (Quantum Monte Carlo in Python)
1877
SymPy Documentation, Release 0.7.6
1878
Chapter 11. Projects using SymPy
CHAPTER
TWELVE
BLOGS, NEWS, MAGAZINES
You can see a list of blogs/news/magazines mentioning SymPy (that we know of) on our wiki
page.
1879
SymPy Documentation, Release 0.7.6
1880
Chapter 12. Blogs, News, Magazines
CHAPTER
THIRTEEN
ABOUT
13.1 SymPy Development Team
SymPy is a team project and it was developed by a lot of people.
Here is a list of contributors together with what they do, (and in some cases links to their
wiki pages), where they describe in more details what they do and what they are interested
in (some people didnt want to be mentioned here, so see our repository history for a full list).
1. Ondej ertk: started the project in 2006, on Jan 4, 2011 passed the project leadership
to Aaron Meurer
2. Fabian Pedregosa: everything, reviewing patches, releases, general advice (issues and
mailinglist), GSoC 2009
3. Jurjen N.E. Bos: pretty printing and other patches
4. Mateusz Paprocki: GSoC 2007, concrete math module, integration module, new core
integration, a lot of patches, general advice, new polynomial module, improvements to
solvers, simplications, patch review
5. Marc-Etienne M.Leveille: matrix patch
6. Brian Jorgensen: GSoC 2007, plotting module and related things, patches
7. Jason Gedge: GSoC 2007, geometry module, a lot of patches and xes, new core integration
8. Robert Schwarz: GSoC 2007, polynomials module, patches
9. Pearu Peterson: new core, sympycore project, general advice (issues and mailinglist)
10. Fredrik Johansson: mpmath project and its integration in SymPy, number theory, combinatorial functions, products & summation, statistics, units, patches, documentation,
general advice (issues and mailinglist)
11. Chris Wu: GSoC 2007, linear algebra module
12. Ulrich Hecht: pattern matching and other patches
13. Goutham Lakshminarayan: number theory functions
14. David Lawrence: GHOP, Mathematica parser, square root denesting
15. Jaroslaw Tworek: GHOP, sympify AST implementation, sqrt() refactoring, maxima parser
and other patches
16. David Marek: GHOP, derivative evaluation patch, int(NumberSymbol) x
17. Bernhard R. Link: documentation patch
1881
SymPy Documentation, Release 0.7.6
18. Andrej Tokark: GHOP, python printer
19. Or Dvory: GHOP, documentation
20. Saroj Adhikari: bug xes
21. Pauli Virtanen: bug x
22. Robert Kern: bug x, common subexpression elimination
23. James Aspnes: bug xes
24. Nimish Telang: multivariate lambdas
25. Abderrahim Kitouni: pretty printing + complex expansion bug xes
26. Pan Peng: ode solvers patch
27. Friedrich Hagedorn: many bug xes, refactorings and new features added
28. Elrond der Elbenfuerst: pretty printing x
29. Rizgar Mella: BBP formula for pi calculating algorithm
30. Felix Kaiser: documentation + whitespace testing patches
31. Roberto Nobrega: several pretty printing patches
32. David Roberts: latex printing patches
33. Sebastian Krmer: implemented lambdify/numpy/mpmath cooperation, bug xes, refactoring, lambdifying of matrices, large printing refactoring and bugxes
34. Vinzent Steinberg: docstring patches, a lot of bug xes, nsolve (nonlinear equation systems solver), compiling functions to machine code, patches review
35. Riccardo Gori: improvements and speedups to matrices, many bug xes
36. Case Van Horsen: implemented optional support for gmpy in mpmath
37. tpn Rouka: a lot of bug xes all over SymPy (matrix, simplication, limits, series,
...)
38. Ali Raza Syed: pretty printing/isympy on windows x
39. Stefano Maggiolo: many bug xes, polishings and improvements
40. Robert Cimrman: matrix patches
41. Bastian Weber: latex printing patches
42. Sebastian Krause: match patches
43. Sebastian Kreft: latex printing patches, Dirac delta function, other xes
44. Dan (coolg49964): documentation xes
45. Alan Bromborsky: geometric algebra modules
46. Boris Timokhin: matrix xes
47. Robert (average.programmer): initial piecewise function patch
48. Andy R. Terrel: piecewise function polishing and other patches
49. Hubert Tsang: LaTeX printing xes
50. Konrad Meyer: policy patch
51. Henrik Johansson: matrix zero nder patch
52. Priit Laes: line integrals, cleanups in ODE, tests added
53. Freddie Witherden: trigonometric simplications xes, LaTeX printer improvements
1882
Chapter 13. About
SymPy Documentation, Release 0.7.6
54. Brian E. Granger: second quantization physics module
55. Andrew Straw: lambdify() improvements
56. Kaifeng Zhu: factorint() x
57. Ted Horst: Basic. call () x, pretty printing x
58. Andrew Docherty: Fix to series
59. Akshay Srinivasan: printing x, improvements to integration
60. Aaron Meurer: ODE solvers (GSoC 2009), The Risch Algorithm for integration (GSoC
2010), other xes, project leader as of Jan 4, 2011
61. Barry Wardell: implement series as function, several tests added
62. Tomasz Buchert: ccode printing xes, code quality concerning exceptions, documentation
63. Vinay Kumar: polygamma tests
64. Johann Cohen-Tanugi: commutative di tests
65. Jochen Voss: a test for the @cachit decorator and several other xes
66. Luke Peterson: improve solve() to handle Derivatives
67. Chris Smith: improvements to solvers, many bug xes, documentation and test improvements
68. Thomas Sidoti: MathML printer improvements
69. Florian Mickler: reimplementation of convex hull, several geometry module xes
70. Nicolas Pourcelot: Innity comparison xes, latex xes
71. Ben Goodrich: Matrix.jacobian() function improvements and xes
72. Toon Verstraelen: code generation module, latex printing improvements
73. Ronan Lamy: test coverage script; limit, expansion and other xes and improvements;
cleanup
74. James Abbatiello: xes tests on windows
75. Ryan Krauss: xes could extract minus sign(), latex xes and improvements
76. Bill Flynn: substitution xes
77. Kevin Goodsell: Fix to Integer/Rational
78. Jorn Baayen: improvements to piecewise functions and latex printing, bug xes
79. Eh Tan: improve trigonometric simplication
80. Renato Coutinho: derivative improvements
81. Oscar Benjamin: latex printer x, gcd bug x
82. yvind Jensen: implemented coupled cluster expansion and wick theorem, improvements to assumptions, bugxes
83. Julio Idichekop Filho: indentation xes, docstring improvements
84. ukasz Pankowski: x matrix multiplication with numpy scalars
85. Chu-Ching Huang: x 3d plotting example
86. Fernando Perez: symarray() implemented
87. Raaele De Feo: x non-commutative expansion
88. Christian Muise: xes to logic module
13.1. SymPy Development Team
1883
SymPy Documentation, Release 0.7.6
89. Matt Curry: GSoC 2010 project (symbolic quantum mechanics)
90. Kazuo Thow: cleanup pretty printing tests
91. Christian Schubert: Fix to sympify()
92. Jezreel Ng: x hyperbolic functions rewrite
93. James Pearson: Py3k related xes
94. Matthew Brett: xes to lambdify
95. Addison Cugini: GSoC 2010 project (quantum computing)
96. Nicholas J.S. Kinar: improved documentation about Immutability of Expressions
97. Harold Erbin: Geometry related work
98. Thomas Dixon: x a limit bug
99. Cristvo Sousa: implements sage method for sign function.
100. Andre de Fortier Smit: doctests in matrices improved
101. Mark Dewing: Fixes to Integral/Sum, MathML work
102. Alexey U. Gudchenko: various work in matrices
103. Gary Kerr: x examples, docs
104. Sherjil Ozair: xes to SparseMatrix
105. Oleksandr Gituliar: xes to Matrix
106. Sean Vig: Quantum improvement
107. Prafullkumar P. Tale: xed plotting documentation
108. Vladimir Peri: x some Python 3 issues
109. Tom Bachmann: xes to Real
110. Yuri Karadzhov: improvements to hyperbolic identities
111. Vladimir Lagunov: improvements to the geometry module
112. Matthew Rocklin: stats, sets, matrix expressions
113. Saptarshi Mandal: Test for an integral
114. Gilbert Gede: Tests for solvers
115. Anatolii Koval: Innite 1D box example
116. Tomo Lazovich: xes to quantum operators
117. Pavel Fedotov: x to is number
118. Kibeom Kim: xes to cse function and preorder traversal
119. Gregory Ksionda: xes to Integral instantiation
120. Tom Bambas: prettier printing of examples
121. Jeremias Yehdegho: xes to the nthoery module
122. Jack McCaery: xes to asin and acos
123. Raymond Wong: Quantum work
124. Luca Weihs: improvements to the geometry module
125. Shai Deshe Wyborski: x to numeric evaluation of hypergeometric sums
126. Thomas Wiecki: Fix Sum.di
1884
Chapter 13. About
SymPy Documentation, Release 0.7.6
127. scar Njera: better Laguerre polynomial generator
128. Mario Pernici: faster algorithm for computing Groebner bases
129. Benjamin McDonald: Fix bug in geometry
130. Sam Magura: Improvements to Plot.saveimage
131. Stefan Krastanov: Make Pyglet an external dependency
132. Bradley Froehle: Fix shell command to generate modules in setup.py
133. Min Ragan-Kelley: Fix isympy to work with IPython 0.11
134. Nikhil Sarda: Fix to combinatorics/prufer
135. Emma Hogan: Fixes to the documentation
136. Jason Moore: Fixes to the mechanics module
137. Julien Rioux: Fix behavior of some deprecated methods
138. Roberto Colistete, Jr.: Add num columns option to the pretty printer
139. Raoul Bourquin: Implement Euler numbers
140. Gert-Ludwig Ingold: Correct derivative of coth
141. Srinivas Vasudevan: Implementation of Catalan numbers
142. Miha Marolt: Add numpydoc extension to the Sphinx docs, x ode order
143. Tim Lahey: Rotation matrices
144. Luis Garcia: update examples in symarry
145. Matt Rajca: Code quality xes
146. David Li: Documentation xes
147. David Ju: Increase test coverage, xes to KroneckerDelta
148. Alexandr Gudulin: Code quality xes
149. Bilal Akhtar: isympy man page
150. Grzegorz wirski: Fix to latex(), MacPorts portle
151. Matt Habel: SymPy-wide pyakes editing
152. Nikolay Lazarov: Translation of the tutorial to Bulgarian
153. Nichita Utiu: Add pretty printing to Product()
154. Tristan Hume: Fixes to test lambdify.py
155. Imran Ahmed Manzoor: Fixes to the test runner
156. Steve Anton: pyakes cleanup of various modules, documentation xes for logic
157. Sam Sleight: Fixes to geometry documentation
158. tsmars15: Fixes to code quality
159. Chancellor Arkantos: Fixes to the logo
160. Stepan Simsa: Translation of the tutorial to Czech
161. Tobias Lenz: Unicode pretty printer for Sum
162. Siddhanathan Shanmugam: Documentation xes for the Physics module
163. Tiany Zhu: Improved the latex() docstring
164. Alexey Subach: Translation of the tutorial to Russian
13.1. SymPy Development Team
1885
SymPy Documentation, Release 0.7.6
165. Joan Creus: Improvements to the test runner
166. Geory Song: Improve code coverage
167. Puneeth Chaganti: Fix for the tutorial
168. Marcin Kostrzewa: SymPy Cheatsheet
169. Jim Zhang: Fixes to AUTHORS and .mailmap
170. Natalia Nawara: Fixes to Product()
171. vishal: Update the Czech tutorial translation to use a .po le
172. Shruti Mangipudi: added See Also documentation to Matrices
173. Davy Mao: Documentation
174. Swapnil Agarwal: added See Also documentation to ntheory, functions
175. Kendhia: Add XFAIL tests
176. jerryma1121: See Also documentation for geometry
177. Joachim Durchholz: corrected spacing issues identied by PyDev
178. Martin Povier: x problem with rationaltools
179. Siddhant Jain: See Also documentation for functions/special
180. Kevin Hunter: Improvements to the inequality classes
181. Michael Mayorov: Improvement to series
182. Nathan Alison: Additions to the stats module
183. Christian Bhler: remove use of set main() in GA
184. Carsten Knoll: improvement to preview()
185. M R Bharath: modied use of int tested, improvement to Permutation
186. Matthias Toews: File permissions
187. Sergiu Ivanov: Fixes to zoo, SymPyDeprecationWarning
188. Jorge E. Cardona: Cleanup in polys
189. Sanket Agarwal: Rewrite coverage doctest.py script
190. Manoj Babu K.: Improve gamma function
191. Sai Nikhil: Fix to Heavyside with complex arguments
192. Aleksandar Makelov: Fixes regarding the dihedral group generator
193. Raphael Michel: German translation of the tutorial
194. Sachin Irukula: Changes to allow Dict sorting
195. Ashwini Oruganti: Changes to Pow printing
196. Andreas Kloeckner: Fix to cse()
197. Prateek Papriwal: improve summation documentation
198. Arpit Goyal: Improvements to Integral and Sum
199. Angadh Nanjangud: in physics.mechanics, added a function and tests for the parallel
axis theorem
200. Comer Duncan: added dual, is antisymmetric, and det lu decomposition to matrices.py
201. Jens H. Nielsen: added sets to modules listing, update IPython printing extension
1886
Chapter 13. About
SymPy Documentation, Release 0.7.6
202. Joseph Dougherty: modied whitespace cleaning to remove multiple newlines at eof
203. marshall2389: Spelling correction
204. Guru Devanla: Implemented quantum density operator
205. George Waksman: Implemented JavaScript code printer and MathML printer
206. Angus Grith: Fix bug in rsolve
207. Timothy Reluga: Rewrite trigonometric functions as rationals
208. Brian Stephanik: Test for a bug in fcode
209. Ljubia Moi: Serbian translation of the tutorial
210. Piotr Korgul: Polish translation of the tutorial
211. Rom le Clair: French translation of the tutorial
212. Alexandr Popov: Fixes to Pauli algebra
213. Saurabh Jha: Work on Kauers algorithm
214. Tarun Gaba: Implemented some trigonometric integrals
215. Takafumi Arakaki: Add info target to the doc Makele
216. Alexander Eberspcher: correct typo in aboutus.rst
217. Sachin Joglekar: Simplication of logic expressions to SOP and POS forms
218. Tyler Pirtle: Fix improperly formatted error message
219. Vasily Povalyaev: Fix latex(Min)
220. Colleen Lee: replace uses of fnan with S.NaN
221. Niklas Thrne: Fix links in the docs
222. Huijun Mai: Chinese translation of the tutorial
223. Marek uppa: Improvements to symbols, tests
224. Prasoon Shukla: Bug xes
225. Sergey B Kirpichev: Bug xes
226. Stefen Yin: Fixes to the mechanics module
227. Thomas Hisch: Improvements to the printing module
228. Matthew Ho: Addition to quantum module
229. Madeleine Ball: Bug x
230. Case Van Horsen: Fixes to gmpy support
231. Mary Clark: Improvements to the group theory module
232. Rishabh Dixit: Bug xes
233. Acebulf: Typos
234. Manoj Kumar: Bug x
235. Akshit Agarwal: improvements to range handling in symbols
236. CJ Carey: Fix for limits of factorials
237. Patrick Lacasse: Fix for Piecewise.subs
238. Ananya H: Bug x
239. Tarang Patel: added test for issue 4739
13.1. SymPy Development Team
1887
SymPy Documentation, Release 0.7.6
240. Christopher Dembia: improvements to mecahanics documentation
241. Benjamin Fishbein: added rank method to Matrix
242. Sean Ge: made KroneckerDelta arguments canonically ordered
243. Ankit Agrawal: Statistical moments
244. Amit Jamadagni: qapply Rotation to spin states
245. Bjrn Dahlgren: Implemented nite dierence module, minor xes
246. Christophe Saint-Jean: xed and added metrics to galgebra
247. Demian Wassermann: x to ccode printer for Piecewise
248. Khagesh Patel: Addition to matrix expressions
249. Stephen Loo: Update minimum gmpy2 version
250. hm: Fixes to printing
251. Katja Sophie Hotz: use expansion in minpoly
252. Varun Joshi: Addition to functions
253. Chetna Gupta: Improvements to the Risch integration algorithm
254. Thilina Rathnayake: Fix to the matrices
255. Shravas K Rao: Implement prev lexicographic and next lexicographic
256. Max Hutchinson: Fix to HadamardProduct
257. Matthew Tadd: x denition in units module
258. Alexander Hirzel: Updates to ODE docs
259. Randy Heydon: improve collinear point detection
260. Ramana Venkata: improvements to special functions
261. Oliver Lee: improvements to mechanics
262. Seshagiri Prabhu: hardcoded 3x3 determinant
263. Pradyumna: Fix to printing
264. Erik Welch: Fix a warning
265. Eric Nelson: Fixes to printing
266. Roland Puntaier: Improve App Engine support
267. Chris Conley: Use warnings instead of prints
268. Tim Swast: Help with pull requests and IPython
269. Dmitry Batkovich: Fix to series
270. Francesco Bonazzi: Improvements to matrices and tensors
271. Yuriy Demidov: Add examples from Review of CAS mathematical capabilities
272. Rick Muller: Implementation of quantum circuit plotting
273. Manish Gill: Fix innite loop in Matrix constructor
274. Markus Mller: Add Jordan form for matrices
275. Amit Saha: Fixes to documentation
276. Jeremy: Bug x
277. QuaBoo: Optimizations in ntheory
1888
Chapter 13. About
SymPy Documentation, Release 0.7.6
278. Stefan van der Walt: Fixes to mechanics module
279. David Joyner: Cryptography module
280. Lars Buitinck: Bug x
281. Alkiviadis G. Akritas: Add a reference
282. Vinit Ravishankar: x iterables documentation
283. Mike Boyle: Additions to the printing system
284. Heiner Kirchhoer: PythonRational int conversion x
285. Pablo Puente: Test cases from Wester paper
286. James Fiedler: Bug x
287. Harsh Gupta: Bug x
288. Tuomas Airaksinen: is real for besselx
289. rathmann: x some Python 2/3 issues
290. Paul Strickland: documentation x
291. James Goppert: correct code in form lagranges equations
292. Avichal Dayal: removed duplicate solutions obtained from solve
293. Paul Scott: pass along keywords for plots
294. shiprabanga: code quality xes
295. Pramod Ch: x problem with (I*oo).expand complex()
296. Akshay: added normal lines to Ellipse
297. Buck Shlegeris: add inverses to error functions
298. Jonathan Miller: documentation
299. Edward: update license year
300. Rajath S: corrected iteration values in gaunt, racah, and wigner 9j
301. Aditya Shah: bug x, improvement to parsers
302. Sambuddha Basu: documentation x
303. Zeel Shah: xes to logic module
304. Abhinav Chanda: wigner 3j and clebsch gordan xes
305. Jim Crist: lambdify improvments
306. Sudhanshu Mishra: Added tests for racah() of sympy.physics.wigner module
307. Rajat Aggarwal: improvements to integration module
308. Soumya Dipta Biswas: Add support for Equivalent with multiple arguments
309. Anurag Sharma: improvements to Risch algorithm
310. Sushant Hiray: taylor term for sec
311. Ben Lucato: documentation xes
312. Kunal Arora: documentation xes
313. Henry Gebhardt: bugx to ufuncify
314. Dammina Sahabandu: implemented continued fractions functions
315. Shukla: particle on a ring functions
13.1. SymPy Development Team
1889
SymPy Documentation, Release 0.7.6
316. Ralph Bean: typo
317. richierichrawr: documentation xes
318. John Connor: xes to ntheory
319. Juan Luis Cano Rodrguez: xes to mechanics module
320. Sahil Shekhawat: xes to core
321. Kundan Kumar: xes to ODE identication of separable reduced
322. Stas Kelvich: xes to quantum module
323. sevaader: x to assumptions
324. Dhruvesh Vijay Parikh: implement Mbius function
325. Venkatesh Halli: improvement to matrices
326. Lennart Fricke: make sure that cse avoids symbol collision
327. Vlad Seghete: x an issue with autowrap
328. shashank-agg: ensure that the range is correct for multiple plots
329. carstimon: make Abs(polar lift(arg)) -> abs(arg)
330. Pierre Haessig: x typos
331. Maciej Baranski: refactor code in compilef
332. Zamrath Nizam: replace atoms(Symbol) with free symbols where appropriate
333. Benjamin Gudehus: add sampling density to stats
334. Faisal Anees: Egyptian Fractions
335. Robert Johansson: quantum eld operators
336. Kalevi Suominen: x to assumptions
337. Kaushik Varanasi: x a typo
338. Fawaz Alazemi: combinatorics cheat sheet
339. Ambar Mehrotra: Fix to documentation
340. Mark Shoulson: Enhanced Egyptian fractions
341. David P. Sanders: Fix to latex printer
342. Peter Brady: avoid caching problem by using function instead of lambda
343. John V. Siratt: Fix to documentation
344. Sarwar Chahal: Fix to documentation
345. Nathan Woods: Fix to C code printer
346. Colin B. Macdonald: Fix to documentation
347. Marcus Nslund: Fix to documentation
348. Clemens Novak: improved docstring for apart
349. Craig A. Stoudt: correct addition of TWave objects
350. Mridul Seth: correct is tangent for Ellipse non-intersection
351. Raj: xed typos in four les
352. Mihai A. Ionescu: Improvement to LaplaceTransform
353. immerrr: add check of SYMPY DEBUG value
1890
Chapter 13. About
SymPy Documentation, Release 0.7.6
354. Leonid Blouvshtein: make integrals aware of both limits being +/-oo
355. Peleg Michaeli: implement the Rademacher distribution
356. Chai Wah Wu: Implement divisor sigma function
357. ck Lux: handle zoo, oo, nan in as int and round
358. zsc347: xed a bug in crypto.rsa private key
359. Hamish Dickson: improve qubit tests
360. Michael Gallaspy: improve handling of inequalities involving RootOf
361. Roman Inianskas: add svg support to preview
362. Duane Nykamp: improved function handling in parse expr
Up-to-date list in the order of the rst contribution is given in the AUTHORS le.
You can nd a brief history of SymPy in the README.
13.2 Financial and Infrastructure Support
Google: SymPy has received generous nancial support from Google in various years
through the Google Summer of Code program by providing stipends:
in 2007 for 5 students (GSoC 2007)
in 2008 for 1 student (GSoC 2008)
in 2009 for 5 students (GSoC 2009)
in 2010 for 5 students (GSoC 2010)
in 2011 for 9 students (GSoC 2011)
in 2012 for 6 students (GSoC 2012)
in 2013 for 7 students (GSoC 2013)
Python Software Foundation (PSF) has hosted various GSoC students over the years:
3 GSoC 2007 students (Brian, Robert and Jason)
1 GSoC 2008 student (Fredrik)
2 GSoC 2009 students (Freddie and Priit)
4 GSoC 2010 students (Aaron, Christian, Matthew and yvind)
Portland State University (PSU) has hosted following GSoC students:
1 student (Chris) in 2007
3 students (Aaron, Dale and Fabian) in 2009
1 student (Addison) in 2010
The Space Telescope Science Institute: STScI hosted 1 GSoC 2007 student (Mateusz)
Several 13-17 year old pre-university students contributed as part of Googles Code-In
2011. (GCI 2011)
Simula Research Laboratory: supports Pearu Peterson work in SymPy/SymPy Core
projects
GitHub is providing us with development and collaboration tools
13.2. Financial and Infrastructure Support
1891
SymPy Documentation, Release 0.7.6
13.3 License
Unless stated otherwise, all les in the SymPy project, SymPys webpage (and wiki), all images
and all documentation including this Users Guide are licensed using the new BSD license:
Copyright (c) 2006-2014 SymPy Development Team
All rights reserved.
Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions are met:
a. Redistributions of source code must retain the above copyright notice,
this list of conditions and the following disclaimer.
b. Redistributions in binary form must reproduce the above copyright
notice, this list of conditions and the following disclaimer in the
documentation and/or other materials provided with the distribution.
c. Neither the name of SymPy nor the names of its contributors
may be used to endorse or promote products derived from this software
without specific prior written permission.
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS AS IS
AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE FOR
ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH
DAMAGE.
1892
Chapter 13. About
CHAPTER
FOURTEEN
SYMPY SPECIAL TOPICS
14.1 Introduction
The purpose of this collection of documents is to provide users of SymPy with topics which
are not strictly tutorial or are longer than tutorials and tests. The documents will hopefully
ll a gap as SymPy matures and users nd more ways to show how SymPy can be used in
more advanced topics.
14.2 Finite Dierence Approximations to Derivatives
14.2.1 Introduction
Finite dierence approximations to derivatives is quite important in numerical analysis and in
computational physics. In this tutorial we show how to use SymPy to compute approximations
of varying accuracy. The hope is that these notes could be useful for the practicing researcher
who is developing code in some language and needs to be able to eciently generate nite
dierence formulae for various approximations.
In order to establish notation, we rst state that we envision that there exists a continuous
function F of a single variable x, with F having as many derivatives as desired. We sample x
values uniformly at points along the real line separated by h. In most cases we want h to be
small in some sense. F(x) may be expanded about some point x0 via the usual Taylor series
expansion. Let x = x0 + h. Then the Taylor expansion is
F (x0 + h) = F (x0 ) +
( dF )
1 ( d2 F )
1 ( d3 F )
2
h
+
h
+
h3 + ...
2
x
x
0
dx 0
2! dx
3! dx3 x0
In order to simplify the notation, we now dene a set of coecients cn , where
cn :=
1 ( dn F )
.
n! dxn x0
So now our series has the form:
F (x0 + h) = F (x0 ) + c1 h + c2 h2 + c3 h3 + ...
In the following we will only use a nite grid of values xi with i running from 1, ..., N and the
corresponding values of our function F at these grid points denoted by Fi . So the problem is
how to generate approximate values for the derivatives of F with the constraint that we use
a subset of the nite set of pairs (xi , Fi ) of size N.
1893
SymPy Documentation, Release 0.7.6
What follows are manipulations using SymPy to formulate approximations for derivatives of
a given order and to assess its accuracy. First, we use SymPy to derive the approximations
by using a rather brute force method frequently covered in introductory treatments. Later
we shall make use of other SymPy functions which get the job done with more eciency.
14.2.2 A Direct Method Using SymPy Matrices
If we let x0 = xi , evaluate the series at xi+1 = xi + h and truncate all terms above O(h1 ) we can
solve for the single coecient c1 and obtain an approximation to the rst derivative:
( dF )
Fi+1 Fi
+ O(h)
x
0
dx
h
where the O(h) refers to the lowest order term in the series in h. This establishes that the
derivative approximation is of rst order accuracy. Put another way, if we decide that we can
only use the two pairs (xi , Fi ) and (xi+1 , Fi+1 ) we obtain a rst order accurate derivative.
In addition to (xi , Fi ) we next use the two points (xi+1 , Fi+1 ) and (xi+2 , Fi+2 ). Then we have two
equations:
Fi+1 = Fi + c1 h +
Fi+2 = Fi + c1 (2h) +
1
1
c2 h2 + c3 h3 + ...
2
3!
1
1
c2 (2h)2 + c3 (2h)3 + ...
2
3!
If we again want to nd the rst derivative (c1 ), we can do that by eliminating the term involving c2 from the two equations. We show how to do it using SymPy.
>>>
>>>
>>>
>>>
>>>
>>>
>>>
...
from __future__ import print_function
from sympy import *
x, x0, h = symbols(x, x_0, h)
Fi, Fip1, Fip2 = symbols(F_{i}, F_{i+1}, F_{i+2})
n = 3 # there are the coefficients c_0=Fi, c_1=dF/dx, c_2=d**2F/dx**2
c = symbols(c:3)
def P(x, x0, c, n):
return sum( ((1/factorial(i))*c[i] * (x-x0)**i for i in range(n)) )
Vector of right hand sides:
>>> R = Matrix([[Fi], [Fip1], [Fip2]])
Now we make a matrix consisting of the coecients of the c i in the nth degree polynomial P.
Coecients of ci evaluated at xi :
>>> m11 = P(x0 , x0, c, n).diff(c[0])
>>> m12 = P(x0 , x0, c, n).diff(c[1])
>>> m13 = P(x0 , x0, c, n).diff(c[2])
Coecients of ci evaluated at xi + h:
>>> m21 = P(x0+h, x0, c, n).diff(c[0])
>>> m22 = P(x0+h, x0, c, n).diff(c[1])
>>> m23 = P(x0+h, x0, c, n).diff(c[2])
1894
Chapter 14. SymPy Special Topics
SymPy Documentation, Release 0.7.6
Coecients of ci evaluated at xi + 2 h:
>>> m31 = P(x0+2*h, x0, c, n).diff(c[0])
>>> m32 = P(x0+2*h, x0, c, n).diff(c[1])
>>> m33 = P(x0+2*h, x0, c, n).diff(c[2])
Matrix of the coecients is 3x3 in this case:
>>> M = Matrix([[m11, m12, m13], [m21, m22, m23], [m31, m32, m33]])
Matrix form of the three equations for the ci is M*X = R:
The solution is obtained by directly inverting the 3x3 matrix M:
>>> X =
M.inv() * R
Note that all three coecients make up the solution. The desired rst derivative is coecient
c1 which is X[1].
>>> print(together(X[1]))
(4*F_{i+1} - F_{i+2} - 3*F_{i})/(2*h)
It is instructive to compute another three-point approximation to the rst derivative, except
centering the approximation at xi and thus using points at xi1 , xi , and xi+1 . So here is how
this can be done using the brute force method:
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
...
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
from __future__ import print_function
from sympy import *
x, x0, h = symbols(x, x_i, h)
Fi, Fim1, Fip1 = symbols(F_{i}, F_{i-1}, F_{i+1})
n = 3 # there are the coefficients c_0=Fi, c_1=dF/h, c_2=d**2F/h**2
c = symbols(c:3)
# define a polynomial of degree n
def P(x, x0, c, n):
return sum( ((1/factorial(i))*c[i] * (x-x0)**i for i in range(n)) )
# now we make a matrix consisting of the coefficients
# of the c_i in the nth degree polynomial P
# coefficients of c_i evaluated at x_i
m11 = P(x0 , x0, c, n).diff(c[0])
m12 = P(x0 , x0, c, n).diff(c[1])
m13 = P(x0 , x0, c, n).diff(c[2])
# coefficients of c_i evaluated at x_i - h
m21 = P(x0-h, x0, c, n).diff(c[0])
m22 = P(x0-h, x0, c, n).diff(c[1])
m23 = P(x0-h, x0, c, n).diff(c[2])
# coefficients of c_i evaluated at x_i + h
m31 = P(x0+h, x0, c, n).diff(c[0])
m32 = P(x0+h, x0, c, n).diff(c[1])
m33 = P(x0+h, x0, c, n).diff(c[2])
# matrix of the coeffcients is 3x3 in this case
M = Matrix([[m11, m12, m13], [m21, m22, m23], [m31, m32, m33]])
Now that we have the matrix of coecients we next form the right-hand-side and solve by
inverting M :
>>>
>>>
>>>
>>>
>>>
#
R
#
#
X
matrix of the function values...actually a vector of right hand sides
= Matrix([[Fi], [Fim1], [Fip1]])
matrix form of the three equations for the c_i is M*X = R
solution directly inverting the 3x3 matrix M:
= M.inv() * R
14.2. Finite Dierence Approximations to Derivatives
1895
SymPy Documentation, Release 0.7.6
>>> # note that all three coefficients make up the solution
>>> # the first derivative is coefficient c_1 which is X[1].
>>> print(The second-order accurate approximation for the first derivative is: )
The second-order accurate approximation for the first derivative is:
>>> print(together(X[1]))
(F_{i+1} - F_{i-1})/(2*h)
These two examples serve to show how one can directly nd second order accurate rst
derivatives using SymPy. The rst example uses values of x and F at all three points xi , xi+1 ,
and xi+2 whereas the second example only uses values of x at the two points xi1 and xi+1 and
thus is a bit more ecient.
From these two simple examples a general rule is that if one wants a rst derivative to be
accurate to O(hn ) then one needs n+1 function values in the approximating polynomial (here
provided via the function P (x, x0, c, n)).
Now lets assess the question of the accuracy of the centered dierence result to see how we
determine that it is really second order. To do this we take the result for dF /dx and substitute
in the polynomial expansion for a higher order polynomial and see what we get. To this end,
we make a set of eight coecients d and use them to perform the check:
>>> d = symbols(c:8)
>>> dfdxcheck = (P(x0+h, x0, d, 8) - P(x0-h, x0, d, 8))/(2*h)
>>> print(simplify(dfdxcheck)) # so the appropriate cancellation of terms involving h happens
c1 + c3*h**2/6 + c5*h**4/120 + c7*h**6/5040
Thus we see that indeed the derivative is c1 with the next term in the series of order h2 .
However, it can quickly become rather tedious to generalize the direct method as presented
above when attempting to generate a derivative approximation to high order, such as 6 or 8
although the method certainly works and using the present method is certainly less tedious
than performing the calculations by hand.
As we have seen in the discussion above, the simple centered approximation for the rst
derivative only uses two point values of the (xi , Fi ) pairs. This works ne until one encounters
the last point in the domain, say at i = N . Since our centered derivative approximation would
use data at the point (xN +1 , FN +1 ) we see that the derivative formula will not work. So, what to
do? Well, a simple way to handle this is to devise a dierent formula for this last point which
uses points for which we do have values. This is the so-called backward dierence formula.
To obtain it, we can use the same direct approach, except now us the three points (xN , FN ),
(xN 1 , FN 1 ), and (xN 2 , FN 2 ) and center the approximation at (xN , FN ). Here is how it can be
done using SymPy:
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
...
from __future__ import print_function
from sympy import *
x, xN, h = symbols(x, x_N, h)
FN, FNm1, FNm2 = symbols(F_{N}, F_{N-1}, F_{N-2})
n = 8 # there are the coefficients c_0=Fi, c_1=dF/h, c_2=d**2F/h**2
c = symbols(c:8)
# define a polynomial of degree d
def P(x, x0, c, n):
return sum( ((1/factorial(i))*c[i] * (x-x0)**i for i in range(n)) )
Now we make a matrix consisting of the coecients of the ci in the dth degree polynomial P
coecients of ci evaluated at xi , xi1 , and xi+1 :
>>> m11 = P(xN , xN, c, n).diff(c[0])
>>> m12 = P(xN, xN, c, n).diff(c[1])
>>> m13 = P(xN , xN, c, n).diff(c[2])
1896
Chapter 14. SymPy Special Topics
SymPy Documentation, Release 0.7.6
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
# coefficients of c_i evaluated at x_i - h
m21 = P(xN-h, xN, c, n).diff(c[0])
m22 = P(xN-h, xN, c, n).diff(c[1])
m23 = P(xN-h, xN, c, n).diff(c[2])
# coefficients of c_i evaluated at x_i + h
m31 = P(xN-2*h, xN, c, n).diff(c[0])
m32 = P(xN-2*h, xN, c, n).diff(c[1])
m33 = P(xN-2*h, xN, c, n).diff(c[2])
Next we construct the 3 3 matrix of the coecients:
>>> M = Matrix([[m11, m12, m13], [m21, m22, m23], [m31, m32, m33]])
>>> # matrix of the function values...actually a vector of right hand sides
>>> R = Matrix([[FN], [FNm1], [FNm2]])
Then we invert M and write the solution to the 3 3 system.
The matrix form of the three equations for the c i is M C = R. The solution is obtained by
directly inverting M :
>>> X =
M.inv() * R
The rst derivative is coecient c1 which is X[1]. Thus the second order accurate approximation for the rst derivative is:
>>> print(The first derivative centered at the last point on the right is:)
The first derivative centered at the last point on the right is:
>>> print(together(X[1]))
(-4*F_{N-1} + F_{N-2} + 3*F_{N})/(2*h)
Of course, we can devise a similar formula for the value of the derivative at the left end of the
set of points at (x1 , F1 ) in terms of values at (x2 , F2 ) and (x3 , F3 ).
Also, we note that output of formats appropriate to Fortran, C, etc. may be done in the
examples given above.
Next we show how to perform these and many other discritizations of derivatives, but using
a much more ecient approach originally due to Bengt Fornberg and now incorported into
SymPy.
14.2. Finite Dierence Approximations to Derivatives
1897
SymPy Documentation, Release 0.7.6
1898
Chapter 14. SymPy Special Topics
BIBLIOGRAPHY
[R24] http://en.wikipedia.org/wiki/Negative number
[R25] http://en.wikipedia.org/wiki/Parity %28mathematics%29
[R26] http://en.wikipedia.org/wiki/Imaginary number
[R27] http://en.wikipedia.org/wiki/Composite number
[R28] http://en.wikipedia.org/wiki/Irrational number
[R29] http://en.wikipedia.org/wiki/Prime number
[R30] http://en.wikipedia.org/wiki/Finite
[R31] https://docs.python.org/3/library/math.html#math.isnite
[R32] http://docs.scipy.org/doc/numpy/reference/generated/numpy.isnite.html
[R33] http://en.wikipedia.org/wiki/Zero
[R34] http://en.wikipedia.org/wiki/1 %28number%29
[R35] http://en.wikipedia.org/wiki/%E2%88%921 %28number%29
[R36] http://en.wikipedia.org/wiki/One half
[R37] http://en.wikipedia.org/wiki/NaN
[R38] http://en.wikipedia.org/wiki/Innity
[R39] http://en.wikipedia.org/wiki/E %28mathematical constant%29
[R40] http://en.wikipedia.org/wiki/Imaginary unit
[R41] http://en.wikipedia.org/wiki/Pi
[R42] http://en.wikipedia.org/wiki/Euler%E2%80%93Mascheroni constant
[R43] http://en.wikipedia.org/wiki/Catalan%27s constant
[R44] http://en.wikipedia.org/wiki/Golden ratio
[R45] http://en.wikipedia.org/wiki/Exponentiation
[R46] http://en.wikipedia.org/wiki/Exponentiation#Zero to the power of zero
[R47] http://en.wikipedia.org/wiki/Indeterminate forms
[R48] This implementation detail is that Python provides no reliable method to determine that
a chained inequality is being built. Chained comparison operators are evaluated pairwise,
using and logic (see http://docs.python.org/2/reference/expressions.html#notin). This is
done in an ecient way, so that each object being compared is only evaluated once and
the comparison can short-circuit. For example, 1 > 2 > 3 is evaluated by Python as (1 >
1899
SymPy Documentation, Release 0.7.6
2) and (2 > 3). The and operator coerces each side into a bool, returning the object itself
when it short-circuits. The bool of the Than operators will raise TypeError on purpose,
because SymPy cannot determine the mathematical ordering of symbolic expressions.
Thus, if we were to compute x > y > z, with x, y, and z being Symbols, Python converts
the statement (roughly) into these steps:
1. x > y > z
2. (x > y) and (y > z)
3. (GreaterThanObject) and (y > z)
4. (GreaterThanObject. nonzero ()) and (y > z)
5. TypeError
Because of the and added at step 2, the statement gets turned into a weak ternary
statement, and the rst objects nonzero method will raise TypeError. Thus, creating
a chained inequality is not possible.
In Python, there is no way to override the and operator, or to control how it short
circuits, so it is impossible to make something like x > y > z work. There was
a PEP to change this, PEP 335, but it was ocially closed in March, 2012.
[R49] For more information, see these two bug reports:
Separate boolean and symbolic relationals Issue 4986
It right 0 < x < 1 ? Issue 6059
[R50] This implementation detail is that Python provides no reliable method to determine that
a chained inequality is being built. Chained comparison operators are evaluated pairwise,
using and logic (see http://docs.python.org/2/reference/expressions.html#notin). This is
done in an ecient way, so that each object being compared is only evaluated once and
the comparison can short-circuit. For example, 1 > 2 > 3 is evaluated by Python as (1 >
2) and (2 > 3). The and operator coerces each side into a bool, returning the object itself
when it short-circuits. The bool of the Than operators will raise TypeError on purpose,
because SymPy cannot determine the mathematical ordering of symbolic expressions.
Thus, if we were to compute x > y > z, with x, y, and z being Symbols, Python converts
the statement (roughly) into these steps:
1. x > y > z
2. (x > y) and (y > z)
3. (GreaterThanObject) and (y > z)
4. (GreaterThanObject. nonzero ()) and (y > z)
5. TypeError
Because of the and added at step 2, the statement gets turned into a weak ternary
statement, and the rst objects nonzero method will raise TypeError. Thus, creating
a chained inequality is not possible.
In Python, there is no way to override the and operator, or to control how it short
circuits, so it is impossible to make something like x > y > z work. There was
a PEP to change this, PEP 335, but it was ocially closed in March, 2012.
[R51] For more information, see these two bug reports:
Separate boolean and symbolic relationals Issue 4986
It right 0 < x < 1 ? Issue 6059
1900
Bibliography
SymPy Documentation, Release 0.7.6
[R52] This implementation detail is that Python provides no reliable method to determine that
a chained inequality is being built. Chained comparison operators are evaluated pairwise,
using and logic (see http://docs.python.org/2/reference/expressions.html#notin). This is
done in an ecient way, so that each object being compared is only evaluated once and
the comparison can short-circuit. For example, 1 > 2 > 3 is evaluated by Python as (1 >
2) and (2 > 3). The and operator coerces each side into a bool, returning the object itself
when it short-circuits. The bool of the Than operators will raise TypeError on purpose,
because SymPy cannot determine the mathematical ordering of symbolic expressions.
Thus, if we were to compute x > y > z, with x, y, and z being Symbols, Python converts
the statement (roughly) into these steps:
1. x > y > z
2. (x > y) and (y > z)
3. (GreaterThanObject) and (y > z)
4. (GreaterThanObject. nonzero ()) and (y > z)
5. TypeError
Because of the and added at step 2, the statement gets turned into a weak ternary
statement, and the rst objects nonzero method will raise TypeError. Thus, creating
a chained inequality is not possible.
In Python, there is no way to override the and operator, or to control how it short
circuits, so it is impossible to make something like x > y > z work. There was
a PEP to change this, PEP 335, but it was ocially closed in March, 2012.
[R53] For more information, see these two bug reports:
Separate boolean and symbolic relationals Issue 4986
It right 0 < x < 1 ? Issue 6059
[R54] This implementation detail is that Python provides no reliable method to determine that
a chained inequality is being built. Chained comparison operators are evaluated pairwise,
using and logic (see http://docs.python.org/2/reference/expressions.html#notin). This is
done in an ecient way, so that each object being compared is only evaluated once and
the comparison can short-circuit. For example, 1 > 2 > 3 is evaluated by Python as (1 >
2) and (2 > 3). The and operator coerces each side into a bool, returning the object itself
when it short-circuits. The bool of the Than operators will raise TypeError on purpose,
because SymPy cannot determine the mathematical ordering of symbolic expressions.
Thus, if we were to compute x > y > z, with x, y, and z being Symbols, Python converts
the statement (roughly) into these steps:
1. x > y > z
2. (x > y) and (y > z)
3. (GreaterThanObject) and (y > z)
4. (GreaterThanObject. nonzero ()) and (y > z)
5. TypeError
Because of the and added at step 2, the statement gets turned into a weak ternary
statement, and the rst objects nonzero method will raise TypeError. Thus, creating
a chained inequality is not possible.
In Python, there is no way to override the and operator, or to control how it short
circuits, so it is impossible to make something like x > y > z work. There was
a PEP to change this, PEP 335, but it was ocially closed in March, 2012.
Bibliography
1901
SymPy Documentation, Release 0.7.6
[R55] For more information, see these two bug reports:
Separate boolean and symbolic relationals Issue 4986
It right 0 < x < 1 ? Issue 6059
[R7] Skiena, S. Permutations. 1.1 in Implementing Discrete Mathematics Combinatorics and
Graph Theory with Mathematica. Reading, MA: Addison-Wesley, pp. 3-16, 1990.
[R8] Knuth, D. E. The Art of Computer Programming, Vol. 4: Combinatorial Algorithms, 1st
ed. Reading, MA: Addison-Wesley, 2011.
[R9] Wendy Myrvold and Frank Ruskey. 2001. Ranking and unranking permutations in linear time. Inf. Process. Lett. 79, 6 (September 2001), 281-284. DOI=10.1016/S00200190(01)00141-7
[R10] D. L. Kreher, D. R. Stinson Combinatorial Algorithms CRC Press, 1999
[R11] Graham, R. L.; Knuth, D. E.; and Patashnik, O. Concrete Mathematics: A Foundation
for Computer Science, 2nd ed. Reading, MA: Addison-Wesley, 1994.
[R12] http://en.wikipedia.org/wiki/Permutation#Product and inverse
[R13] http://en.wikipedia.org/wiki/Lehmer code
[R14] http://mathworld.wolfram.com/LabeledTree.html
[R311] http://en.wikipedia.org/wiki/Continued fraction
[R312] http://en.wikipedia.org/wiki/Periodic continued fraction
[R313] K. Rosen. Elementary Number theory and its applications. Addison-Wesley, 3 Sub edition, pages 379-381, January 1992.
[R314] http://en.wikipedia.org/wiki/M%C3%B6bius function
[R315] Thomas Koshy Elementary Number Theory with Applications
[R316] http://en.wikipedia.org/wiki/Egyptian fraction
[R317] https://en.wikipedia.org/wiki/Greedy algorithm for Egyptian fractions
[R318] http://www.ics.uci.edu/eppstein/numth/egypt/conict.html
[R319] http://ami.ektf.hu/uploads/papers/nalpdf/AMI 42 from129to134.pdf
[R56] http://en.wikipedia.org/wiki/Vigenere cipher
[R57] en.wikipedia.org/wiki/Hill cipher
[R58] Lester S. Hill, Cryptography in an Algebraic Alphabet, The American Mathematical
Monthly Vol.36, June-July 1929, pp.306-312.
[R59] http://en.wikipedia.org/wiki/Morse code
[R60] http://en.wikipedia.org/wiki/Morse code
[G60] Solomon Golomb, Shift register sequences, Aegean Park Press, Laguna Hills, Ca, 1967
[M61] James L. Massey, Shift-Register Synthesis and BCH Decoding. IEEE Trans. on Information Theory, vol. 15(1), pp. 122-127, Jan 1969.
[R15] Michael Karr, Summation in Finite Terms, Journal of the ACM, Volume 28 Issue 2,
April 1981, Pages 305-350 http://dl.acm.org/citation.cfm?doid=322248.322255
[R16] http://en.wikipedia.org/wiki/Summation#Capital-sigma notation
[R17] http://en.wikipedia.org/wiki/Empty sum
[R18] Michael Karr, Summation in Finite Terms, Journal of the ACM, Volume 28 Issue 2,
April 1981, Pages 305-350 http://dl.acm.org/citation.cfm?doid=322248.322255
1902
Bibliography
SymPy Documentation, Release 0.7.6
[R19] Michael Karr, Summation in Finite Terms, Journal of the ACM, Volume 28 Issue 2,
April 1981, Pages 305-350 http://dl.acm.org/citation.cfm?doid=322248.322255
[R20] http://en.wikipedia.org/wiki/Multiplication#Capital Pi notation
[R21] http://en.wikipedia.org/wiki/Empty product
[R22] Michael Karr, Summation in Finite Terms, Journal of the ACM, Volume 28 Issue 2,
April 1981, Pages 305-350 http://dl.acm.org/citation.cfm?doid=322248.322255
[R23] Marko Petkovsek, Herbert S. Wilf, Doron Zeilberger, A = B, AK Peters, Ltd., Wellesley,
MA, USA, 1997, pp. 73100
[R87] http://en.wikipedia.org/wiki/Inverse trigonometric functions
[R88] http://dlmf.nist.gov/4.23
[R89] http://functions.wolfram.com/ElementaryFunctions/ArcCos
[R90] http://en.wikipedia.org/wiki/Inverse trigonometric functions
[R91] http://dlmf.nist.gov/4.23
[R92] http://functions.wolfram.com/ElementaryFunctions/ArcCot
[R93] http://en.wikipedia.org/wiki/Inverse trigonometric functions
[R94] http://dlmf.nist.gov/4.23
[R95] http://functions.wolfram.com/ElementaryFunctions/ArcSin
[R96] http://en.wikipedia.org/wiki/Inverse trigonometric functions
[R97] http://dlmf.nist.gov/4.23
[R98] http://functions.wolfram.com/ElementaryFunctions/ArcTan
[R99] http://en.wikipedia.org/wiki/Inverse trigonometric functions
[R100] http://en.wikipedia.org/wiki/Atan2
[R101] http://functions.wolfram.com/ElementaryFunctions/ArcTan2
[R102] http://en.wikipedia.org/wiki/Trigonometric functions
[R103] http://dlmf.nist.gov/4.14
[R104] http://functions.wolfram.com/ElementaryFunctions/Cos
[R105] http://en.wikipedia.org/wiki/Trigonometric functions
[R106] http://dlmf.nist.gov/4.14
[R107] http://functions.wolfram.com/ElementaryFunctions/Cot
[R108] http://en.wikipedia.org/wiki/Lambert W function
[R109] http://en.wikipedia.org/wiki/Directed complete partial order
[R110] http://en.wikipedia.org/wiki/Lattice %28order%29
[R111] http://en.wikipedia.org/wiki/Trigonometric functions
[R112] http://dlmf.nist.gov/4.14
[R113] http://functions.wolfram.com/ElementaryFunctions/Sin
[R114] http://mathworld.wolfram.com/TrigonometryAngles.html
[R115] http://en.wikipedia.org/wiki/Trigonometric functions
[R116] http://dlmf.nist.gov/4.14
[R117] http://functions.wolfram.com/ElementaryFunctions/Tan
Bibliography
1903
SymPy Documentation, Release 0.7.6
[R63] http://en.wikipedia.org/wiki/Bell number
[R64] http://mathworld.wolfram.com/BellNumber.html
[R65] http://mathworld.wolfram.com/BellPolynomial.html
[R66] http://en.wikipedia.org/wiki/Bernoulli number
[R67] http://en.wikipedia.org/wiki/Bernoulli polynomial
[R68] http://mathworld.wolfram.com/BernoulliNumber.html
[R69] http://mathworld.wolfram.com/BernoulliPolynomial.html
[R70] http://en.wikipedia.org/wiki/Catalan number
[R71] http://mathworld.wolfram.com/CatalanNumber.html
[R72] http://functions.wolfram.com/GammaBetaErf/CatalanNumber/
[R73] http://geometer.org/mathcircles/catalan.pdf
[R74] http://en.wikipedia.org/wiki/Euler numbers
[R75] http://mathworld.wolfram.com/EulerNumber.html
[R76] http://en.wikipedia.org/wiki/Alternating permutation
[R77] http://mathworld.wolfram.com/AlternatingPermutation.html
[R78] http://en.wikipedia.org/wiki/Fibonacci number
[R79] http://mathworld.wolfram.com/FibonacciNumber.html
[R80] http://en.wikipedia.org/wiki/Harmonic number
[R81] http://functions.wolfram.com/GammaBetaErf/HarmonicNumber/
[R82] http://functions.wolfram.com/GammaBetaErf/HarmonicNumber2/
[R83] http://en.wikipedia.org/wiki/Lucas number
[R84] http://mathworld.wolfram.com/LucasNumber.html
[R85] http://en.wikipedia.org/wiki/Stirling numbers of the rst kind
[R86] http://en.wikipedia.org/wiki/Stirling numbers of the second kind
[R118] http://mathworld.wolfram.com/DeltaFunction.html
[R119] http://mathworld.wolfram.com/HeavisideStepFunction.html
[R120] http://en.wikipedia.org/wiki/Gamma function
[R121] http://dlmf.nist.gov/5
[R122] http://mathworld.wolfram.com/GammaFunction.html
[R123] http://functions.wolfram.com/GammaBetaErf/Gamma/
[R124] http://en.wikipedia.org/wiki/Gamma function
[R125] http://dlmf.nist.gov/5
[R126] http://mathworld.wolfram.com/LogGammaFunction.html
[R127] http://functions.wolfram.com/GammaBetaErf/LogGamma/
[R128] http://en.wikipedia.org/wiki/Polygamma function
[R129] http://mathworld.wolfram.com/PolygammaFunction.html
[R130] http://functions.wolfram.com/GammaBetaErf/PolyGamma/
[R131] http://functions.wolfram.com/GammaBetaErf/PolyGamma2/
1904
Bibliography
SymPy Documentation, Release 0.7.6
[R132] http://en.wikipedia.org/wiki/Digamma function
[R133] http://mathworld.wolfram.com/DigammaFunction.html
[R134] http://functions.wolfram.com/GammaBetaErf/PolyGamma2/
[R135] http://en.wikipedia.org/wiki/Trigamma function
[R136] http://mathworld.wolfram.com/TrigammaFunction.html
[R137] http://functions.wolfram.com/GammaBetaErf/PolyGamma2/
[R138] http://en.wikipedia.org/wiki/Incomplete gamma function#Upper Incomplete Gamma Function
[R139] Abramowitz, Milton; Stegun, Irene A., eds. (1965), Chapter 6, Section 5, Handbook
of Mathematical Functions with Formulas, Graphs, and Mathematical Tables
[R140] http://dlmf.nist.gov/8
[R141] http://functions.wolfram.com/GammaBetaErf/Gamma2/
[R142] http://functions.wolfram.com/GammaBetaErf/Gamma3/
[R143] http://en.wikipedia.org/wiki/Exponential integral#Relation with other functions
[R144] http://en.wikipedia.org/wiki/Incomplete gamma function#Lower Incomplete Gamma Function
[R145] Abramowitz, Milton; Stegun, Irene A., eds. (1965), Chapter 6, Section 5, Handbook
of Mathematical Functions with Formulas, Graphs, and Mathematical Tables
[R146] http://dlmf.nist.gov/8
[R147] http://functions.wolfram.com/GammaBetaErf/Gamma2/
[R148] http://functions.wolfram.com/GammaBetaErf/Gamma3/
[R149] http://en.wikipedia.org/wiki/Beta function
[R150] http://mathworld.wolfram.com/BetaFunction.html
[R151] http://dlmf.nist.gov/5.12
[R152] http://en.wikipedia.org/wiki/Error function
[R153] http://dlmf.nist.gov/7
[R154] http://mathworld.wolfram.com/Erf.html
[R155] http://functions.wolfram.com/GammaBetaErf/Erf
[R156] http://en.wikipedia.org/wiki/Error function
[R157] http://dlmf.nist.gov/7
[R158] http://mathworld.wolfram.com/Erfc.html
[R159] http://functions.wolfram.com/GammaBetaErf/Erfc
[R160] http://en.wikipedia.org/wiki/Error function
[R161] http://mathworld.wolfram.com/Er.html
[R162] http://functions.wolfram.com/GammaBetaErf/Er
[R163] http://functions.wolfram.com/GammaBetaErf/Erf2/
[R164] http://en.wikipedia.org/wiki/Error function#Inverse functions
[R165] http://functions.wolfram.com/GammaBetaErf/InverseErf/
[R166] http://en.wikipedia.org/wiki/Error function#Inverse functions
[R167] http://functions.wolfram.com/GammaBetaErf/InverseErfc/
[R168] http://functions.wolfram.com/GammaBetaErf/InverseErf2/
Bibliography
1905
SymPy Documentation, Release 0.7.6
[R169] http://en.wikipedia.org/wiki/Fresnel integral
[R170] http://dlmf.nist.gov/7
[R171] http://mathworld.wolfram.com/FresnelIntegrals.html
[R172] http://functions.wolfram.com/GammaBetaErf/FresnelS
[R173] The converging factors for the fresnel integrals by John W. Wrench Jr. and Vicki Alley
[R174] http://en.wikipedia.org/wiki/Fresnel integral
[R175] http://dlmf.nist.gov/7
[R176] http://mathworld.wolfram.com/FresnelIntegrals.html
[R177] http://functions.wolfram.com/GammaBetaErf/FresnelC
[R178] The converging factors for the fresnel integrals by John W. Wrench Jr. and Vicki Alley
[R179] http://dlmf.nist.gov/6.6
[R180] http://en.wikipedia.org/wiki/Exponential integral
[R181] Abramowitz & Stegun, section 5: http://people.math.sfu.ca/cbm/aands/page 228.htm
[R182] http://dlmf.nist.gov/8.19
[R183] http://functions.wolfram.com/GammaBetaErf/ExpIntegralE/
[R184] http://en.wikipedia.org/wiki/Exponential integral
[R185] http://en.wikipedia.org/wiki/Logarithmic integral
[R186] http://mathworld.wolfram.com/LogarithmicIntegral.html
[R187] http://dlmf.nist.gov/6
[R188] http://mathworld.wolfram.com/SoldnersConstant.html
[R189] http://en.wikipedia.org/wiki/Logarithmic integral
[R190] http://mathworld.wolfram.com/LogarithmicIntegral.html
[R191] http://dlmf.nist.gov/6
[R192] http://en.wikipedia.org/wiki/Trigonometric integral
[R193] http://en.wikipedia.org/wiki/Trigonometric integral
[R194] http://en.wikipedia.org/wiki/Trigonometric integral
[R195] http://en.wikipedia.org/wiki/Trigonometric integral
[R196] Abramowitz, Milton; Stegun, Irene A., eds. (1965), Chapter 9, Handbook of Mathematical Functions with Formulas, Graphs, and Mathematical Tables
[R197] Luke, Y. L. (1969), The Special Functions and Their Approximations, Volume 1
[R198] http://en.wikipedia.org/wiki/Bessel function
[R199] http://functions.wolfram.com/Bessel-TypeFunctions/BesselJ/
[R200] http://functions.wolfram.com/Bessel-TypeFunctions/BesselY/
[R201] http://functions.wolfram.com/Bessel-TypeFunctions/BesselI/
[R202] http://functions.wolfram.com/Bessel-TypeFunctions/BesselK/
[R203] http://functions.wolfram.com/Bessel-TypeFunctions/HankelH1/
[R204] http://functions.wolfram.com/Bessel-TypeFunctions/HankelH2/
[R205] http://en.wikipedia.org/wiki/Airy function
1906
Bibliography
SymPy Documentation, Release 0.7.6
[R206] http://dlmf.nist.gov/9
[R207] http://www.encyclopediaofmath.org/index.php/Airy functions
[R208] http://mathworld.wolfram.com/AiryFunctions.html
[R209] http://en.wikipedia.org/wiki/Airy function
[R210] http://dlmf.nist.gov/9
[R211] http://www.encyclopediaofmath.org/index.php/Airy functions
[R212] http://mathworld.wolfram.com/AiryFunctions.html
[R213] http://en.wikipedia.org/wiki/Airy function
[R214] http://dlmf.nist.gov/9
[R215] http://www.encyclopediaofmath.org/index.php/Airy functions
[R216] http://mathworld.wolfram.com/AiryFunctions.html
[R217] http://en.wikipedia.org/wiki/Airy function
[R218] http://dlmf.nist.gov/9
[R219] http://www.encyclopediaofmath.org/index.php/Airy functions
[R220] http://mathworld.wolfram.com/AiryFunctions.html
[R221] http://en.wikipedia.org/wiki/B-spline
[R222] http://dlmf.nist.gov/25.11
[R223] http://en.wikipedia.org/wiki/Hurwitz zeta function
[R224] http://en.wikipedia.org/wiki/Dirichlet eta function
[R225] Bateman, H.; Erdelyi, A. (1953), Higher Transcendental Functions, Vol. I, New York:
McGraw-Hill. Section 1.11.
[R226] http://dlmf.nist.gov/25.14
[R227] http://en.wikipedia.org/wiki/Lerch transcendent
[R228] Luke, Y. L. (1969), The Special Functions and Their Approximations, Volume 1
[R229] http://en.wikipedia.org/wiki/Generalized hypergeometric function
[R230] Luke, Y. L. (1969), The Special Functions and Their Approximations, Volume 1
[R231] http://en.wikipedia.org/wiki/Meijer G-function
[R232] http://en.wikipedia.org/wiki/Elliptic integrals
[R233] http://functions.wolfram.com/EllipticIntegrals/EllipticK
[R234] http://en.wikipedia.org/wiki/Elliptic integrals
[R235] http://functions.wolfram.com/EllipticIntegrals/EllipticF
[R236] http://en.wikipedia.org/wiki/Elliptic integrals
[R237] http://functions.wolfram.com/EllipticIntegrals/EllipticE2
[R238] http://functions.wolfram.com/EllipticIntegrals/EllipticE
[R239] http://en.wikipedia.org/wiki/Elliptic integrals
[R240] http://functions.wolfram.com/EllipticIntegrals/EllipticPi3
[R241] http://functions.wolfram.com/EllipticIntegrals/EllipticPi
[R242] http://en.wikipedia.org/wiki/Jacobi polynomials
Bibliography
1907
SymPy Documentation, Release 0.7.6
[R243] http://mathworld.wolfram.com/JacobiPolynomial.html
[R244] http://functions.wolfram.com/Polynomials/JacobiP/
[R245] http://en.wikipedia.org/wiki/Jacobi polynomials
[R246] http://mathworld.wolfram.com/JacobiPolynomial.html
[R247] http://functions.wolfram.com/Polynomials/JacobiP/
[R248] http://en.wikipedia.org/wiki/Gegenbauer polynomials
[R249] http://mathworld.wolfram.com/GegenbauerPolynomial.html
[R250] http://functions.wolfram.com/Polynomials/GegenbauerC3/
[R251] http://en.wikipedia.org/wiki/Chebyshev polynomial
[R252] http://mathworld.wolfram.com/ChebyshevPolynomialoftheFirstKind.html
[R253] http://mathworld.wolfram.com/ChebyshevPolynomialoftheSecondKind.html
[R254] http://functions.wolfram.com/Polynomials/ChebyshevT/
[R255] http://functions.wolfram.com/Polynomials/ChebyshevU/
[R256] http://en.wikipedia.org/wiki/Chebyshev polynomial
[R257] http://mathworld.wolfram.com/ChebyshevPolynomialoftheFirstKind.html
[R258] http://mathworld.wolfram.com/ChebyshevPolynomialoftheSecondKind.html
[R259] http://functions.wolfram.com/Polynomials/ChebyshevT/
[R260] http://functions.wolfram.com/Polynomials/ChebyshevU/
[R261] http://en.wikipedia.org/wiki/Legendre polynomial
[R262] http://mathworld.wolfram.com/LegendrePolynomial.html
[R263] http://functions.wolfram.com/Polynomials/LegendreP/
[R264] http://functions.wolfram.com/Polynomials/LegendreP2/
[R265] http://en.wikipedia.org/wiki/Associated Legendre polynomials
[R266] http://mathworld.wolfram.com/LegendrePolynomial.html
[R267] http://functions.wolfram.com/Polynomials/LegendreP/
[R268] http://functions.wolfram.com/Polynomials/LegendreP2/
[R269] http://en.wikipedia.org/wiki/Hermite polynomial
[R270] http://mathworld.wolfram.com/HermitePolynomial.html
[R271] http://functions.wolfram.com/Polynomials/HermiteH/
[R272] http://en.wikipedia.org/wiki/Laguerre polynomial
[R273] http://mathworld.wolfram.com/LaguerrePolynomial.html
[R274] http://functions.wolfram.com/Polynomials/LaguerreL/
[R275] http://functions.wolfram.com/Polynomials/LaguerreL3/
[R276] http://en.wikipedia.org/wiki/Laguerre polynomial#Assoc laguerre polynomials
[R277] http://mathworld.wolfram.com/AssociatedLaguerrePolynomial.html
[R278] http://functions.wolfram.com/Polynomials/LaguerreL/
[R279] http://functions.wolfram.com/Polynomials/LaguerreL3/
[R280] http://en.wikipedia.org/wiki/Spherical harmonics
1908
Bibliography
SymPy Documentation, Release 0.7.6
[R281] http://mathworld.wolfram.com/SphericalHarmonic.html
[R282] http://functions.wolfram.com/Polynomials/SphericalHarmonicY/
[R283] http://dlmf.nist.gov/14.30
[R284] http://en.wikipedia.org/wiki/Spherical harmonics
[R285] http://mathworld.wolfram.com/SphericalHarmonic.html
[R286] http://functions.wolfram.com/Polynomials/SphericalHarmonicY/
[R287] http://en.wikipedia.org/wiki/Spherical harmonics
[R288] http://mathworld.wolfram.com/SphericalHarmonic.html
[R289] http://functions.wolfram.com/Polynomials/SphericalHarmonicY/
[R290] http://en.wikipedia.org/wiki/Kronecker delta
[Doran] http://www.mrao.cam.ac.uk/cjld1/pages/book.htm Geometric Algebra for Physicists by C. Doran and A. Lasenby, Cambridge University Press, 2003.
[Hestenes] http://geocalc.clas.asu.edu/html/CA to GC.html Clifford Algebra to Geometric
Calculus by D.Hestenes and G. Sobczyk, Kluwer Academic Publishers, 1984.
[Macdonald] <http://faculty.luther.edu/macdonal> Linear and Geometric Algebra by
Alan Macdonald, http://www.amazon.com/Alan-Macdonald/e/B004MB2QJQ
[WikiPappus] Pappuss Hexagon Theorem Wikipedia, the Free Encyclopedia. Web. 26 Apr.
2013. <http://en.wikipedia.org/wiki/Pappuss hexagon theorem>
[BlogPost] http://nessgrh.wordpress.com/2011/07/07/tricky-branch-cuts/
[Bro05] M. Bronstein, Symbolic Integration I: Transcendental Functions, Second Edition,
Springer-Verlag, 2005, pp. 35-70
[R291] http://en.wikipedia.org/wiki/Gaussian quadrature
[R292] http://people.sc.fsu.edu/jburkardt/cpp src/legendre rule/legendre rule.html
[R293] http://en.wikipedia.org/wiki/Gauss%E2%80%93Laguerre quadrature
[R294] http://people.sc.fsu.edu/jburkardt/cpp src/laguerre rule/laguerre rule.html
[R295] http://en.wikipedia.org/wiki/Gauss-Hermite Quadrature
[R296] http://people.sc.fsu.edu/jburkardt/cpp src/hermite rule/hermite rule.html
[R297] http://people.sc.fsu.edu/jburkardt/cpp src/gen hermite rule/gen hermite rule.html
[R298] http://en.wikipedia.org/wiki/Gauss%E2%80%93Laguerre quadrature
[R299] http://people.sc.fsu.edu/jburkardt/cpp src/gen laguerre rule/gen laguerre rule.html
[R300] http://en.wikipedia.org/wiki/Chebyshev%E2%80%93Gauss quadrature
[R301] http://people.sc.fsu.edu/jburkardt/cpp src/chebyshev1 rule/chebyshev1 rule.html
[R302] http://en.wikipedia.org/wiki/Chebyshev%E2%80%93Gauss quadrature
[R303] http://people.sc.fsu.edu/jburkardt/cpp src/chebyshev2 rule/chebyshev2 rule.html
[R304] http://en.wikipedia.org/wiki/Gauss%E2%80%93Jacobi quadrature
[R305] http://people.sc.fsu.edu/jburkardt/cpp src/jacobi rule/jacobi rule.html
[R306] http://people.sc.fsu.edu/jburkardt/cpp src/gegenbauer rule/gegenbauer rule.html
[R307] en.wikipedia.org/wiki/Quine-McCluskey algorithm
[R308] en.wikipedia.org/wiki/Quine-McCluskey algorithm
[R309] https://en.wikipedia.org/wiki/Moore-Penrose pseudoinverse
Bibliography
1909
SymPy Documentation, Release 0.7.6
[R310] https://en.wikipedia.org/wiki/Moore-Penrose pseudoinverse#Obtaining all solutions of a linear syst
[AbramowitzStegun] M Abramowitz & I Stegun. Handbook of Mathematical Functions, 9th Ed., Tenth Printing, December 1972, with corrections (electronic copy:
http://people.math.sfu.ca/cbm/aands/)
[Bailey] D
H
Bailey.
Tanh-Sinh
High-Precision
legacy.lbl.gov/dhbailey/dhbpapers/dhb-tanh-sinh.pdf
Quadrature,
http://crd-
[BenderOrszag] C M Bender & S A Orszag. Advanced Mathematical Methods for Scientists
and Engineers, Springer 1999
[BorweinBailey] J Borwein, D H Bailey & R Girgensohn. Experimentation in Mathematics Computational Paths to Discovery, A K Peters, 2003
[BorweinBorwein] J Borwein & P B Borwein. Pi and the AGM: A Study in Analytic Number
Theory and Computational Complexity, Wiley 1987
[BorweinZeta] P Borwein. An Ecient Algorithm for the Riemann Zeta Function,
http://www.cecm.sfu.ca/personal/pborwein/PAPERS/P155.pdf
[CabralRosetti] L G Cabral-Rosetti & M A Sanchis-Lozano. Appell Functions and the
Scalar One-Loop Three-point Integrals in Feynman Diagrams. http://arxiv.org/abs/hepph/0206081
[Carlson] B C Carlson. Numerical computation of real or complex elliptic integrals.
http://arxiv.org/abs/math/9409227v1
[Corless] R M Corless et al. On the Lambert W function, Adv. Comp. Math. 5 (1996) 329-359.
http://www.apmaths.uwo.ca/djerey/Oprints/W-adv-cm.pdf
[DLMF] NIST Digital Library of Mathematical Functions. http://dlmf.nist.gov/
[GradshteynRyzhik] I S Gradshteyn & I M Ryzhik, A Jerey & D Zwillinger (eds.), Table of
Integrals, Series and Products, Seventh edition (2007), Elsevier
[GravesMorris] P R Graves-Morris, D E Roberts & A Salam. The epsilon algorithm and related topics, Journal of Computational and Applied Mathematics, Volume 122, Issue 1-2
(October 2000)
[MPFR] The
MPFR
team.
The
MPFR
http://www.mpfr.org/algorithms.pdf
Library:
Algorithms
and
Proofs,
[Slater] L J Slater. Generalized Hypergeometric Functions. Cambridge University Press, 1966
[Spouge] J L Spouge. Computation of the gamma, digamma, and trigamma functions, SIAM
J. Numer. Anal. Vol. 31, No. 3, pp. 931-944, June 1994.
[SrivastavaKarlsson] H M Srivastava & P W Karlsson. Multiple Gaussian Hypergeometric
Series. Ellis Horwood, 1985.
[Vidunas] R Vidunas. Identities
http://arxiv.org/abs/0804.0655
between
Appells
and
hypergeometric
functions.
[Weisstein] E W Weisstein. MathWorld. http://mathworld.wolfram.com/
[WhittakerWatson] E T Whittaker & G N Watson. A Course of Modern Analysis. 4th Ed. 1946
Cambridge University Press
[Wikipedia] Wikipedia, the free encyclopedia. http://en.wikipedia.org/wiki/Main Page
[WolframFunctions] Wolfram
Research,
http://functions.wolfram.com/
Inc.
The
Wolfram
Functions
Site.
[Wester1999] Michael J. Wester, A Critique of the Mathematical Abilities of CA Systems, 1999,
http://www.math.unm.edu/wester/cas/book/Wester.pdf
1910
Bibliography
SymPy Documentation, Release 0.7.6
[Kozen89] D. Kozen, S. Landau, Polynomial decomposition algorithms, Journal of Symbolic
Computation 7 (1989), pp. 445-456
[Liao95] Hsin-Chao Liao, R. Fateman, Evaluation of the heuristic polynomial GCD, International Symposium on Symbolic and Algebraic Computation (ISSAC), ACM Press, Montreal,
Quebec, Canada, 1995, pp. 240247
[Gathen99] J. von zur Gathen, J. Gerhard, Modern Computer Algebra, First Edition, Cambridge University Press, 1999
[Weisstein09] Eric W. Weisstein, Cyclotomic Polynomial, From MathWorld - A Wolfram Web
Resource, http://mathworld.wolfram.com/CyclotomicPolynomial.html
[Wang78] P. S. Wang, An Improved Multivariate Polynomial Factoring Algorithm, Math. of
Computation 32, 1978, pp. 12151231
[Geddes92] K. Geddes, S. R. Czapor, G. Labahn, Algorithms for Computer Algebra, Springer,
1992
[Monagan93] Michael Monagan, In-place Arithmetic for Polynomials over Z n, Proceedings
of DISCO 92, Springer-Verlag LNCS, 721, 1993, pp. 2234
[Kaltofen98] E. Kaltofen, V. Shoup, Subquadratic-time Factoring of Polynomials over Finite
Fields, Mathematics of Computation, Volume 67, Issue 223, 1998, pp. 11791197
[Shoup95] V. Shoup, A New Polynomial Factorization Algorithm and its Implementation, Journal of Symbolic Computation, Volume 20, Issue 4, 1995, pp. 363397
[Gathen92] J. von zur Gathen, V. Shoup, Computing Frobenius Maps and Factoring Polynomials, ACM Symposium on Theory of Computing, 1992, pp. 187224
[Shoup91] V. Shoup, A Fast Deterministic Algorithm for Factoring Polynomials over Finite
Fields of Small Characteristic, In Proceedings of International Symposium on Symbolic
and Algebraic Computation, 1991, pp. 1421
[Cox97] D. Cox, J. Little, D. OShea, Ideals, Varieties and Algorithms, Springer, Second Edition, 1997
[Ajwa95] I.A.
Ajwa,
Z.
Liu,
P.S.
Wang,
https://citeseer.ist.psu.edu/myciteseer/login, 1995
Groebner
Bases
Algorithm,
[Bose03] N.K. Bose, B. Buchberger, J.P. Guiver, Multidimensional Systems Theory and Applications, Springer, 2003
[Giovini91] A. Giovini, T. Mora, One sugar cube, please or Selection strategies in Buchberger algorithm, ISSAC 91, ACM
[Bronstein93] M. Bronstein, B. Salvy, Full partial fraction decomposition of rational functions,
Proceedings ISSAC 93, ACM Press, Kiev, Ukraine, 1993, pp. 157160
[Buchberger01] B. Buchberger, Groebner Bases: A Short Introduction for Systems Theorists,
In: R. Moreno-Diaz, B. Buchberger, J. L. Freire, Proceedings of EUROCAST01, February,
2001
[Davenport88] J.H. Davenport, Y. Siret, E. Tournier, Computer Algebra Systems and Algorithms for Algebraic Computation, Academic Press, London, 1988, pp. 124128
[Greuel2008] G.-M. Greuel, Gerhard Pster, A Singular Introduction to Commutative Algebra,
Springer, 2008
[Atiyah69] M.F. Atiyah, I.G. MacDonald, Introduction to Commutative Algebra, AddisonWesley, 1969
[Monagan00] M. Monagan and A. Wittkopf, On the Design and Implementation of Browns
Algorithm over the Integers and Number Fields, Proceedings of ISSAC 2000, pp. 225-233,
ACM, 2000.
Bibliography
1911
SymPy Documentation, Release 0.7.6
[Brown71] W.S. Brown, On Euclids Algorithm and the Computation of Polynomial Greatest
Common Divisors, J. ACM 18, 4, pp. 478-504, 1971.
[Hoeij04] M. van Hoeij and M. Monagan, Algorithms for polynomial GCD computation over
algebraic function elds, Proceedings of ISSAC 2004, pp. 297-304, ACM, 2004.
[Wang81] P.S. Wang, A p-adic algorithm for univariate partial fractions, Proceedings of SYMSAC 1981, pp. 212-217, ACM, 1981.
[Hoeij02] M. van Hoeij and M. Monagan, A modular GCD algorithm over number elds presented with multiple extensions, Proceedings of ISSAC 2002, pp. 109-116, ACM, 2002
[ManWright94] Yiu-Kwong Man and Francis J. Wright, Fast Polynomial Dispersion Computation and its Application to Indenite Summation, Proceedings of the International Symposium on Symbolic and Algebraic Computation, 1994, Pages 175-180
http://dl.acm.org/citation.cfm?doid=190347.190413
[Koepf98] Wolfram Koepf, Hypergeometric Summation: An Algorithmic Approach to Summation and Special Function Identities, Advanced lectures in mathematics, Vieweg, 1998
[Abramov71] S. A. Abramov, On the Summation of Rational Functions, USSR Computational
Mathematics and Mathematical Physics, Volume 11, Issue 4, 1971, Pages 324-330
[Man93] Yiu-Kwong Man, On Computing Closed Forms for Indenite Summations,
Journal of Symbolic Computation, Volume 16, Issue 4, 1993, Pages 355-376
http://www.sciencedirect.com/science/article/pii/S0747717183710539
[R345] Big O notation
[R346] http://en.wikipedia.org/wiki/Disjoint sets
[R347] http://en.wikipedia.org/wiki/Power set
[R348] http://en.wikipedia.org/wiki/Interval %28mathematics%29
[R349] http://en.wikipedia.org/wiki/Finite set
[R350] http://en.wikipedia.org/wiki/Union %28set theory%29
[R351] http://en.wikipedia.org/wiki/Intersection %28set theory%29
[R352] http://en.wikipedia.org/wiki/Cartesian product
[R353] http://en.wikipedia.org/wiki/Empty set
[R354] http://en.wikipedia.org/wiki/Universal set
[Roach1996] Kelly B. Roach. Hypergeometric Function Representations. In: Proceedings of
the 1996 International Symposium on Symbolic and Algebraic Computation, pages 301308, New York, 1996. ACM.
[Roach1997] Kelly B. Roach. Meijer G Function Representations. In: Proceedings of the 1997
International Symposium on Symbolic and Algebraic Computation, pages 205-211, New
York, 1997. ACM.
[Luke1969] Luke, Y. L. (1969), The Special Functions and Their Approximations, Volume 1.
[Prudnikov1990] A. P. Prudnikov, Yu. A. Brychkov and O. I. Marichev (1990). Integrals and
Series: More Special Functions, Vol. 3, Gordon and Breach Science Publisher.
[R382] http://en.wikipedia.org/wiki/Arcsine distribution
[R383] http://en.wikipedia.org/wiki/Benini distribution
[R384] http://reference.wolfram.com/legacy/v8/ref/BeniniDistribution.html
[R385] http://en.wikipedia.org/wiki/Beta distribution
[R386] http://mathworld.wolfram.com/BetaDistribution.html
1912
Bibliography
SymPy Documentation, Release 0.7.6
[R387] http://en.wikipedia.org/wiki/Beta prime distribution
[R388] http://mathworld.wolfram.com/BetaPrimeDistribution.html
[R389] http://en.wikipedia.org/wiki/Cauchy distribution
[R390] http://mathworld.wolfram.com/CauchyDistribution.html
[R391] http://en.wikipedia.org/wiki/Chi distribution
[R392] http://mathworld.wolfram.com/ChiDistribution.html
[R393] http://en.wikipedia.org/wiki/Noncentral chi distribution
[R394] http://en.wikipedia.org/wiki/Chi squared distribution
[R395] http://mathworld.wolfram.com/Chi-SquaredDistribution.html
[R396] http://en.wikipedia.org/wiki/Dagum distribution
[R397] http://en.wikipedia.org/wiki/Erlang distribution
[R398] http://mathworld.wolfram.com/ErlangDistribution.html
[R399] http://en.wikipedia.org/wiki/Exponential distribution
[R400] http://mathworld.wolfram.com/ExponentialDistribution.html
[R401] http://en.wikipedia.org/wiki/F-distribution
[R402] http://mathworld.wolfram.com/F-Distribution.html
[R403] http://en.wikipedia.org/wiki/Fisher%27s z-distribution
[R404] http://mathworld.wolfram.com/Fishersz-Distribution.html
[R405] http://en.wikipedia.org/wiki/Fr%C3%A9chet distribution
[R406] http://en.wikipedia.org/wiki/Gamma distribution
[R407] http://mathworld.wolfram.com/GammaDistribution.html
[R408] http://en.wikipedia.org/wiki/Inverse-gamma distribution
[R409] http://en.wikipedia.org/wiki/Kumaraswamy distribution
[R410] http://en.wikipedia.org/wiki/Laplace distribution
[R411] http://mathworld.wolfram.com/LaplaceDistribution.html
[R412] http://en.wikipedia.org/wiki/Logistic distribution
[R413] http://mathworld.wolfram.com/LogisticDistribution.html
[R414] http://en.wikipedia.org/wiki/Lognormal
[R415] http://mathworld.wolfram.com/LogNormalDistribution.html
[R416] http://en.wikipedia.org/wiki/Maxwell distribution
[R417] http://mathworld.wolfram.com/MaxwellDistribution.html
[R418] http://en.wikipedia.org/wiki/Nakagami distribution
[R419] http://en.wikipedia.org/wiki/Normal distribution
[R420] http://mathworld.wolfram.com/NormalDistributionFunction.html
[R421] http://en.wikipedia.org/wiki/Pareto distribution
[R422] http://mathworld.wolfram.com/ParetoDistribution.html
[R423] http://en.wikipedia.org/wiki/U-quadratic distribution
[R424] http://en.wikipedia.org/wiki/Raised cosine distribution
Bibliography
1913
SymPy Documentation, Release 0.7.6
[R425] http://en.wikipedia.org/wiki/Rayleigh distribution
[R426] http://mathworld.wolfram.com/RayleighDistribution.html
[R427] http://en.wikipedia.org/wiki/Student t-distribution
[R428] http://mathworld.wolfram.com/Studentst-Distribution.html
[R429] http://en.wikipedia.org/wiki/Triangular distribution
[R430] http://mathworld.wolfram.com/TriangularDistribution.html
[R431] http://en.wikipedia.org/wiki/Uniform distribution %28continuous%29
[R432] http://mathworld.wolfram.com/UniformDistribution.html
[R433] http://en.wikipedia.org/wiki/Uniform sum distribution
[R434] http://mathworld.wolfram.com/UniformSumDistribution.html
[R435] http://en.wikipedia.org/wiki/Von Mises distribution
[R436] http://mathworld.wolfram.com/vonMisesDistribution.html
[R437] http://en.wikipedia.org/wiki/Weibull distribution
[R438] http://mathworld.wolfram.com/WeibullDistribution.html
[R439] http://en.wikipedia.org/wiki/Wigner semicircle distribution
[R440] http://mathworld.wolfram.com/WignersSemicircleLaw.html
[R376] S. A. Abramov, M. Bronstein and M. Petkovsek, On polynomial solutions of linear operator equations, in: T. Levelt, ed., Proc. ISSAC 95, ACM Press, New York, 1995, 290-296.
[R377] M. Petkovsek, Hypergeometric solutions of linear recurrences with polynomial coefcients, J. Symbolic Computation, 14 (1992), 243-264.
[R378] 13. Petkovsek, H. S. Wilf, D. Zeilberger, A = B, 1996.
[R379] S. A. Abramov, Rational solutions of linear dierence and q-dierence equations with
polynomial coecients, in: T. Levelt, ed., Proc. ISSAC 95, ACM Press, New York, 1995,
285-289
[R380] M. Petkovsek, Hypergeometric solutions of linear recurrences with polynomial coefcients, J. Symbolic Computation, 14 (1992), 243-264.
[R381] 13. Petkovsek, H. S. Wilf, D. Zeilberger, A = B, 1996.
[R355] Methods to solve Ax2 + Bxy + Cy2 + Dx + Ey + F = 0,[online], Available:
http://www.alpertron.com.ar/METHODS.HTM
[R356] Solving the equation ax2+ bxy + cy2 + dx + ey + f= 0, [online], Available:
http://www.jpr2718.org/ax2p.pdf
[R357] Solving the generalized Pell equation x**2 - D*y**2 = N, John P. Robertson, July 31,
2004, Pages 16 - 17. [online], Available: http://www.jpr2718.org/pell.pdf
[R358]
1. Nitaj, Lalgorithme de Cornacchia
[R359] Solving the diophantine equation ax**2 + by**2 = m by Cornacchias method, [online],
Available: http://www.numbertheory.org/php/cornacchia.html
[R360] Solving the generalized Pell equation x**2 - D*y**2 = N, John P. Robertson, July 31,
2004, Page 15. http://www.jpr2718.org/pell.pdf
[R361] Solving the equation ax2 + bxy + cy2 + dx + ey + f = 0, John P.Robertson, May 8,
2003, Page 7 - 11. http://www.jpr2718.org/ax2p.pdf
[R362] Solving the equation ax2 + bxy + cy2 + dx + ey + f = 0, John P.Robertson, May 8,
2003, Page 7 - 11. http://www.jpr2718.org/ax2p.pdf
1914
Bibliography
SymPy Documentation, Release 0.7.6
[R363] Ecient Solution of Rational Conices, J. E. Cremona and D. Rusin, Mathematics of
Computation, Volume 00, Number 0.
[R364] Representing an Integer as a sum of three squares,
http://www.proofwiki.org/wiki/Integer as Sum of Three Squares
[online],
Available:
[R365] Generating Integer Partitions, [online], Available: http://homepages.ed.ac.uk/jkellehe/partitions.php
[R366] Representing a number as
http://www.schorn.ch/howto.html
sum
[R367] Representing a number as
http://www.schorn.ch/howto.html
sum
of
of
three
squares,
[online],
Available:
four
squares,
[online],
Available:
[R368] Solving the generalized Pell equation x2 - Dy2 = N, John P. Robertson, July 31, 2004,
Pages 4 - 8. http://www.jpr2718.org/pell.pdf
[R369] Solving the generalized Pell equation x**2 - D*y**2 = N, John P. Robertson, July 31,
2004, Page 12. http://www.jpr2718.org/pell.pdf
[R370] The algorithmic resolution of Diophantine equations, Nigel P. Smart, London Mathematical Society Student Texts 41, Cambridge University Press, Cambridge, 1998.
[R371] The algorithmic resolution of Diophantine equations, Nigel P. Smart, London Mathematical Society Student Texts 41, Cambridge University Press, Cambridge, 1998.
[R372] Ecient Solution of Rational Conices, J. E. Cremona and D. Rusin, Mathematics of
Computation, Volume 00, Number 0.
[R373] Gaussian lattice Reduction [online]. Available: http://home.ie.cuhk.edu.hk/wkshum/wordpress/?p=4
[R374] Ecient Solution of Rational Conices, J. E. Cremona and D. Rusin, Mathematics of
Computation, Volume 00, Number 0.
[R375] Representing a number as
http://www.schorn.ch/howto.html
sum
of
four
squares,
[online],
Available:
[AOCP] Algorithm 7.1.2.5M in Volume 4A, Combinatoral Algorithms, Part 1, of The Art of
Computer Programming, by Donald Knuth.
[Factorisatio] On a Problem of Oppenheim concerning Factorisatio Numerorum E. R. Caneld, Paul Erdos, Carl Pomerance, JOURNAL OF NUMBER THEORY, Vol. 17, No. 1. August 1983. See section 7 for a description of an algorithm similar to Knuths.
[Yorgey] Generating Multiset Partitions, Brent Yorgey, The Monad.Reader, Issue 8, September 2007.
[R1] http://en.wikipedia.org/wiki/Euler%E2%80%93Lagrange equation
[R2] http://en.wikipedia.org/wiki/Mathematical singularity
[R3] Generation of Finite Dierence Formulas on Arbitrarily Spaced Grids, Bengt Fornberg;
Mathematics of computation; 51; 184; (1988); 699-706; doi:10.1090/S0025-5718-19880935077-0
[R320] https://en.wikipedia.org/wiki/DFT matrix
[R321] http://en.wikipedia.org/wiki/Gamma matrices
[R322] http://en.wikipedia.org/wiki/Pauli matrices
[R326] http://en.wikipedia.org/wiki/Pauli matrices
[R344] http://en.wikipedia.org/wiki/Kronecker delta
[Rasch03] J. Rasch and A. C. H. Yu, Ecient Storage Scheme for Pre-calculated Wigner 3j, 6j
and Gaunt Coecients, SIAM J. Sci. Comput. Volume 25, Issue 4, pp. 1416-1428 (2003)
Bibliography
1915
SymPy Documentation, Release 0.7.6
[Liberatodebrito82] FORTRAN program for the integral of three spherical harmonics, A.
Liberato de Brito, Comput. Phys. Commun., Volume 25, pp. 81-85 (1982)
[Regge58] Symmetry Properties of Clebsch-Gordan Coecients, T. Regge, Nuovo Cimento,
Volume 10, pp. 544 (1958)
[Edmonds74] Angular Momentum in Quantum Mechanics, A. R. Edmonds, Princeton University Press (1974)
[Regge59] Symmetry Properties of Racah Coecients, T. Regge, Nuovo Cimento, Volume
11, pp. 116 (1959)
[WikiDyadics] Dyadics. Wikipedia, the
<http://en.wikipedia.org/wiki/Dyadics>.
Free
Encyclopedia.
Web.
05
Aug.
2011.
[WikiDyadicProducts] Dyadic Product. Wikipedia, the Free Encyclopedia. Web. 05 Aug.
2011. <http://en.wikipedia.org/wiki/Dyadic product>.
[Likins1973] Likins, Peter W. Elements of Engineering Mechanics. McGraw-Hill, Inc. 1973.
Print.
[Blajer1994] Blajer, Wojciech, Werner Schiehlen, and Walter Schirm. A projective criterion
to the coordinate partitioning method for multibody dynamics. Archive of Applied Mechanics 64: 86-98. Print.
[Kane1983] Kane, Thomas R., Peter W. Likins, and David A. Levinson. Spacecraft Dynamics.
New York: McGraw-Hill Book, 1983. Print.
[Kane1985] Kane, Thomas R., and David A. Levinson. Dynamics, Theory and Applications.
New York: McGraw-Hill, 1985. Print.
[Meijaard2007] Meijaard, J.P., Jim M. Papadopoulos, Andy Ruina, and A.L. Schwab. Linearized Dynamics Equations for the Balance and Steer of a Bicycle: a Benchmark and
Review. Proceedings of the Royal Society A: Mathematical, Physical and Engineering
Sciences 463.2084 (2007): 1955-982. Print.
[R327] http://en.wikipedia.org/wiki/Commutator
[R328] Varshalovich, D A, Quantum Theory of Angular Momentum. 1988.
[R329] Varshalovich, D A, Quantum Theory of Angular Momentum. 1988.
[R330] Varshalovich, D A, Quantum Theory of Angular Momentum. 1988.
[R331] http://en.wikipedia.org/wiki/Commutator
[R332] http://en.wikipedia.org/wiki/Hermitian adjoint
[R333] http://en.wikipedia.org/wiki/Hermitian transpose
[R336] http://en.wikipedia.org/wiki/Inner product
[R334] http://en.wikipedia.org/wiki/Hilbert space
[R335] http://en.wikipedia.org/wiki/Fock space
[R337] http://en.wikipedia.org/wiki/Operator %28physics%29
[R338] http://en.wikipedia.org/wiki/Observable
[R339] http://en.wikipedia.org/wiki/Outer product
[R340] Varshalovich, D A, Quantum Theory of Angular Momentum. 1988.
[R341] Varshalovich, D A, Quantum Theory of Angular Momentum. 1988.
[R342] http://en.wikipedia.org/wiki/Bra-ket notation
[R343] http://en.wikipedia.org/wiki/Bra-ket notation
1916
Bibliography
SymPy Documentation, Release 0.7.6
[R323] http://en.wikipedia.org/wiki/Ray transfer matrix analysis
[R324] http://en.wikipedia.org/wiki/Complex beam parameter
[R325] http://en.wikipedia.org/wiki/Optical medium
[Page52] C. H. Page, Classes of units in the SI, Am. J. of Phys. 20, 1 (1952): 1.
[Page78] C. H. Page, Units and Dimensions in Physics, Am. J. of Phys. 46, 1 (1978): 78.
[deBoer79] J. de Boer, Group properties of quantities and units, Am. J. of Phys. 47, 9 (1979):
818.
[LevyLeblond77] J.-M. Lvy-Leblond, On the Conceptual Nature of the Physical Constants, La
Rivista Del Nuovo Cimento 7, no. 2 (1977): 187-214.
[NIST] NIST reference on constants, units and uncertainties.
Bibliography
1917
SymPy Documentation, Release 0.7.6
1918
Bibliography
PYTHON MODULE INDEX
m
mpmath.functions.elliptic, 909
s
sympy, 704
sympy.assumptions, 1287
sympy.assumptions.ask, 1287
sympy.assumptions.assume, 1288
sympy.assumptions.handlers, 1291
sympy.assumptions.handlers.calculus,
1291
sympy.assumptions.handlers.ntheory,
1293
sympy.assumptions.handlers.order, 1293
sympy.assumptions.handlers.sets, 1293
sympy.assumptions.refine, 1289
sympy.calculus, 1592
sympy.calculus.euler, 1592
sympy.calculus.finite diff, 1593
sympy.calculus.singularities, 1593
sympy.categories, 1829
sympy.categories.diagram drawing, 1838
sympy.combinatorics.generators, 215
sympy.combinatorics.graycode, 256
sympy.combinatorics.group constructs,
269
sympy.combinatorics.named groups, 261
sympy.combinatorics.partitions, 187
sympy.combinatorics.perm groups, 216
sympy.combinatorics.permutations, 192
sympy.combinatorics.polyhedron, 243
sympy.combinatorics.prufer, 246
sympy.combinatorics.subsets, 250
sympy.combinatorics.tensor can, 270
sympy.combinatorics.testutil, 269
sympy.combinatorics.util, 263
sympy.core.add, 144
sympy.core.assumptions, 88
sympy.core.basic, 90
sympy.core.cache, 90
sympy.core.compatibility, 183
sympy.core.containers, 182
sympy.core.evalf, 181
sympy.core.expr, 103
sympy.core.exprtools, 185
sympy.core.function, 163
sympy.core.mod, 147
sympy.core.mul, 142
sympy.core.multidimensional, 162
sympy.core.numbers, 126
sympy.core.power, 139
sympy.core.relational, 147
sympy.core.singleton, 102
sympy.core.symbol, 122
sympy.core.sympify, 85
sympy.crypto.crypto, 303
sympy.diffgeom, 1846
sympy.functions, 340
sympy.functions.special.bessel, 408
sympy.functions.special.beta functions,
386
sympy.functions.special.elliptic integrals,
430
sympy.functions.special.error functions,
388
sympy.functions.special.gamma functions,
379
sympy.functions.special.polynomials,
433
sympy.functions.special.zeta functions,
421
sympy.galgebra, 450
sympy.galgebra.debug, 489
sympy.galgebra.ga, 451
sympy.galgebra.manifold, 483
sympy.galgebra.ncutil, 486
sympy.galgebra.precedence, 484
sympy.galgebra.printing, 485
sympy.galgebra.stringarrays, 488
sympy.galgebra.vector, 484
sympy.geometry.curve, 539
sympy.geometry.ellipse, 542
1919
SymPy Documentation, Release 0.7.6
sympy.geometry.entity, 493
sympy.geometry.line, 509
sympy.geometry.line3d, 525
sympy.geometry.plane, 577
sympy.geometry.point, 497
sympy.geometry.point3d, 503
sympy.geometry.polygon, 556
sympy.geometry.util, 495
sympy.integrals, 583
sympy.integrals.meijerint doc, 604
sympy.integrals.transforms, 584
sympy.logic, 625
sympy.matrices, 635
sympy.matrices.expressions, 706
sympy.matrices.expressions.blockmatrix,
711
sympy.matrices.immutable, 705
sympy.matrices.matrices, 636
sympy.matrices.sparse, 692
sympy.ntheory.continued fraction, 298
sympy.ntheory.egyptian fraction, 301
sympy.ntheory.factor , 281
sympy.ntheory.generate, 276
sympy.ntheory.modular, 290
sympy.ntheory.multinomial, 292
sympy.ntheory.partitions , 294
sympy.ntheory.primetest, 294
sympy.ntheory.residue ntheory, 295
sympy.physics, 1597
sympy.physics.hep.gamma matrices, 1632
sympy.physics.hydrogen, 1598
sympy.physics.matrices, 1600
sympy.physics.mechanics.kane, 1741
sympy.physics.mechanics.lagrange, 1744
sympy.physics.mechanics.linearize, 1746
sympy.physics.mechanics.particle, 1730
sympy.physics.mechanics.rigidbody, 1733
sympy.physics.optics.gaussopt, 1802
sympy.physics.optics.medium, 1812
sympy.physics.optics.utils, 1813
sympy.physics.optics.waves, 1817
sympy.physics.paulialgebra, 1601
sympy.physics.qho 1d, 1602
sympy.physics.quantum.anticommutator,
1749
sympy.physics.quantum.cartesian, 1757
sympy.physics.quantum.cg, 1750
sympy.physics.quantum.circuitplot, 1788
sympy.physics.quantum.commutator, 1752
sympy.physics.quantum.constants, 1753
sympy.physics.quantum.dagger, 1753
sympy.physics.quantum.gate, 1789
sympy.physics.quantum.grover, 1794
sympy.physics.quantum.hilbert, 1758
1920
sympy.physics.quantum.innerproduct,
1754
sympy.physics.quantum.operator, 1760
sympy.physics.quantum.operatorset, 1764
sympy.physics.quantum.piab, 1801
sympy.physics.quantum.qapply, 1766
sympy.physics.quantum.qft, 1796
sympy.physics.quantum.qubit, 1796
sympy.physics.quantum.represent, 1767
sympy.physics.quantum.shor, 1800
sympy.physics.quantum.spin, 1771
sympy.physics.quantum.state, 1781
sympy.physics.quantum.tensorproduct,
1755
sympy.physics.secondquant, 1604
sympy.physics.sho, 1603
sympy.physics.units, 1630
sympy.physics.unitsystems.dimensions,
1824
sympy.physics.unitsystems.prefixes,
1827
sympy.physics.unitsystems.quantities,
1829
sympy.physics.unitsystems.units, 1827
sympy.physics.vector.functions, 1680
sympy.physics.vector.point, 1675
sympy.physics.wigner, 1623
sympy.plotting.plot, 1270
sympy.plotting.pygletplot, 1283
sympy.printing.ccode, 1252
sympy.printing.codeprinter, 1266
sympy.printing.conventions, 1266
sympy.printing.fcode, 1254
sympy.printing.gtk, 1259
sympy.printing.lambdarepr, 1259
sympy.printing.latex, 1259
sympy.printing.mathematica, 1258
sympy.printing.mathml, 1262
sympy.printing.precedence, 1266
sympy.printing.pretty.pretty, 1251
sympy.printing.pretty.pretty symbology,
1266
sympy.printing.pretty.stringpict, 1268
sympy.printing.preview, 1264
sympy.printing.printer, 1248
sympy.printing.python, 1262
sympy.printing.repr, 1262
sympy.printing.str, 1263
sympy.printing.tree, 1263
sympy.series, 1304
sympy.sets.fancysets, 1324
sympy.sets.sets, 1312
sympy.simplify.cse main, 1345
sympy.simplify.epathtools, 1347
Python Module Index
SymPy Documentation, Release 0.7.6
sympy.simplify.hyperexpand, 1347
sympy.simplify.hyperexpand doc, 1358
sympy.simplify.simplify, 73
sympy.simplify.sqrtdenest, 1344
sympy.simplify.traversaltools, 1347
sympy.solvers, 1468
sympy.solvers.inequalities, 1505
sympy.solvers.ode, 1457
sympy.solvers.pde, 1468
sympy.solvers.recurr, 1479
sympy.stats, 1361
sympy.stats.crv, 1396
sympy.stats.crv types, 1396
sympy.stats.frv, 1396
sympy.stats.frv types, 1396
sympy.stats.rv, 1396
sympy.tensor, 1508
sympy.tensor.index methods, 1514
sympy.tensor.indexed, 1508
sympy.tensor.tensor, 1517
sympy.utilities, 1532
sympy.utilities.autowrap, 1533
sympy.utilities.codegen, 1538
sympy.utilities.decorator, 1547
sympy.utilities.enumerative, 1549
sympy.utilities.iterables, 1555
sympy.utilities.lambdify, 1573
sympy.utilities.memoization, 1576
sympy.utilities.misc, 1576
sympy.utilities.pkgdata, 1577
sympy.utilities.pytest, 1578
sympy.utilities.randtest, 1579
sympy.utilities.runtests, 1580
sympy.utilities.source, 1587
sympy.utilities.timeutils, 1587
Python Module Index
1921
SymPy Documentation, Release 0.7.6
1922
Python Module Index
INDEX
Symbols
LorentzContainer
(class
in
sympy.physics.hep.gamma matrices),
1632
TensorManager
(class
in
sympy.tensor.tensor), 1517
base ordering()
(in
module
sympy.combinatorics.util), 263
check cycles alt sym()
(in
module
sympy.combinatorics.util), 264
cmp perm lists()
(in
module
sympy.combinatorics.testutil), 269
distribute gens by base()
(in
module
sympy.combinatorics.util), 264
handle precomputed bsgs()
(in
module
sympy.combinatorics.util), 265
linear 2eq order1 type1()
(in
module
sympy.solvers.ode), 1433
linear 2eq order1 type2()
(in
module
sympy.solvers.ode), 1435
linear 2eq order1 type3()
(in
module
sympy.solvers.ode), 1435
linear 2eq order1 type4()
(in
module
sympy.solvers.ode), 1436
linear 2eq order1 type5()
(in
module
sympy.solvers.ode), 1436
linear 2eq order1 type6()
(in
module
sympy.solvers.ode), 1437
linear 2eq order1 type7()
(in
module
sympy.solvers.ode), 1437
linear 2eq order2 type1()
(in
module
sympy.solvers.ode), 1438
linear 2eq order2 type10()
(in
module
sympy.solvers.ode), 1446
linear 2eq order2 type11()
(in
module
sympy.solvers.ode), 1446
linear 2eq order2 type2()
(in
module
sympy.solvers.ode), 1439
linear 2eq order2 type3()
(in
module
sympy.solvers.ode), 1440
linear 2eq order2 type4()
(in
module
sympy.solvers.ode), 1441
linear 2eq order2 type5()
(in
module
sympy.solvers.ode), 1442
linear 2eq order2 type6()
(in
module
sympy.solvers.ode), 1443
linear 2eq order2 type7()
(in
module
sympy.solvers.ode), 1443
linear 2eq order2 type8()
(in
module
sympy.solvers.ode), 1444
linear 2eq order2 type9()
(in
module
sympy.solvers.ode), 1445
linear 3eq order1 type1()
(in
module
sympy.solvers.ode), 1447
linear 3eq order1 type2()
(in
module
sympy.solvers.ode), 1448
linear 3eq order1 type3()
(in
module
sympy.solvers.ode), 1448
linear 3eq order1 type4()
(in
module
sympy.solvers.ode), 1449
linear neq order1 type1()
(in
module
sympy.solvers.ode), 1450
naive list centralizer()
(in
module
sympy.combinatorics.testutil), 269
nonlinear 2eq order1 type1() (in module
sympy.solvers.ode), 1451
nonlinear 2eq order1 type2() (in module
sympy.solvers.ode), 1452
nonlinear 2eq order1 type3() (in module
sympy.solvers.ode), 1452
nonlinear 2eq order1 type4() (in module
sympy.solvers.ode), 1453
nonlinear 2eq order1 type5() (in module
sympy.solvers.ode), 1453
nonlinear 3eq order1 type1() (in module
sympy.solvers.ode), 1454
nonlinear 3eq order1 type2() (in module
sympy.solvers.ode), 1454
nonlinear 3eq order1 type3() (in module
sympy.solvers.ode), 1455
nonlinear 3eq order1 type4() (in module
sympy.solvers.ode), 1456
nonlinear 3eq order1 type5() (in module
1923
SymPy Documentation, Release 0.7.6
sympy.solvers.ode), 1456
orbits transversals from bsgs() (in module
sympy.combinatorics.util), 266
print()
(sympy.printing.printer.Printer
method), 1251
remove gens()
(in
module
sympy.combinatorics.util), 266
strip() (in module sympy.combinatorics.util),
267
strong gens from distr()
(in
module
sympy.combinatorics.util), 268
verify bsgs()
(in
module
sympy.combinatorics.testutil), 269
verify centralizer()
(in
module
sympy.combinatorics.testutil), 270
verify normal closure()
(in
module
sympy.combinatorics.testutil), 270
acosh (class in sympy.functions.elementary.hyperbolic),
343
acosh() (in module mpmath), 776
acot (class in sympy.functions.elementary.trigonometric),
344
acot() (in module mpmath), 773
acoth (class in sympy.functions.elementary.hyperbolic),
344
acoth() (in module mpmath), 777
acsc() (in module mpmath), 773
acsch() (in module mpmath), 776
Add (class in sympy.core.add), 144
add() (sympy.assumptions.assume.AssumptionsContext
method), 1289
Add() (sympy.assumptions.handlers.calculus.AskBoundedHa
static method), 1291
Add() (sympy.assumptions.handlers.calculus.AskInnitesima
static method), 1292
A
Add() (sympy.assumptions.handlers.order.AskNegativeHand
A (sympy.physics.optics.gaussopt.RayTransferMatrix static method), 1293
Add() (sympy.assumptions.handlers.sets.AskAntiHermitianH
attribute), 1803
static method), 1293
a (sympy.physics.quantum.shor.CMod atAdd() (sympy.assumptions.handlers.sets.AskHermitianHand
tribute), 1801
static method), 1294
a1pt theory() (sympy.physics.vector.point.Point
Add() (sympy.assumptions.handlers.sets.AskImaginaryHand
method), 1675
static method), 1294
a2idx() (in module sympy.matrices.matrices),
Add() (sympy.assumptions.handlers.sets.AskIntegerHandler
686
static method), 1294
a2pt theory() (sympy.physics.vector.point.Point
Add() (sympy.assumptions.handlers.sets.AskRationalHandle
method), 1676
static method), 1294
abbrev (sympy.physics.unitsystems.units.Unit
Add() (sympy.assumptions.handlers.sets.AskRealHandler
attribute), 1828
static method), 1295
abbrev dim (sympy.physics.unitsystems.units.Unit
add() (sympy.matrices.matrices.MatrixBase
attribute), 1828
method), 648
AbelianGroup()
(in
module
sympy.combinatorics.named groups), add() (sympy.matrices.sparse.SparseMatrix
method), 692
263
add() (sympy.physics.unitsystems.dimensions.Dimension
above() (sympy.printing.pretty.stringpict.stringPict
method), 1825
method), 1268
add() (sympy.physics.unitsystems.quantities.Quantity
Abs (class in sympy.functions.elementary.complexes),
method), 1829
342
abs() (sympy.polys.domains.domain.Domain add() (sympy.polys.domains.domain.Domain
method), 1149
method), 1149
abs() (sympy.polys.polyclasses.DMP method), add() (sympy.polys.polyclasses.DMF method),
1167
1162
abs() (sympy.polys.polytools.Poly method), add() (sympy.polys.polyclasses.DMP method),
1162
1076
add()
(sympy.polys.polytools.Poly
method),
acc()
(sympy.physics.vector.point.Point
1076
method), 1677
accepted latex functions
(in
module add() (sympy.polys.rings.PolyRing method),
1215
sympy.printing.latex), 1259
add
ground()
(sympy.polys.polyclasses.DMP
acos (class in sympy.functions.elementary.trigonometric),
method),
1162
343
add
ground()
(sympy.polys.polytools.Poly
acos() (in module mpmath), 771
method), 1077
1924
Index
SymPy Documentation, Release 0.7.6
add paren()
(in
module almosteq() (in module mpmath), 738
sympy.galgebra.precedence), 484
almosteq() (sympy.polys.domains.domain.Domain
adjoint() (sympy.matrices.immutable.ImmutableMatrix method), 1149
method), 705
almosteq() (sympy.polys.domains.RealField
adjoint() (sympy.matrices.matrices.MatrixBase
method), 1160
method), 648
almosteq()
(sympy.polys.rings.PolyElement
adjugate() (sympy.matrices.matrices.MatrixBase
method), 1216
alphabet of cipher()
(in
module
method), 648
agm() (in module mpmath), 765
sympy.crypto.crypto), 303
airyai (class in sympy.functions.special.bessel), alternating() (sympy.combinatorics.generators
414
static method), 215
airyai() (in module mpmath), 839
AlternatingGroup()
(in
module
sympy.combinatorics.named groups),
airyaiprime
(class
in
sympy.functions.special.bessel),
262
417
altitudes (sympy.geometry.polygon.Triangle
airyaizero() (in module mpmath), 848
attribute), 571
AiryBase
(class
in altzeta() (in module mpmath), 934
sympy.functions.special.bessel),
amplitude (sympy.physics.optics.waves.TWave
414
attribute), 1817
airybi (class in sympy.functions.special.bessel), an (sympy.functions.special.hyper.meijerg at415
tribute), 430
airybi() (in module mpmath), 844
And (class in sympy.logic.boolalg), 627
airybiprime
(class
in Anderson
(class
in
mpsympy.functions.special.bessel),
math.calculus.optimization), 978
418
ANewton
(class
in
mpmath.calculus.optimization), 978
airybizero() (in module mpmath), 849
algebraic, 89
ang acc in() (sympy.physics.vector.frame.ReferenceFrame
algebraic eld() (sympy.polys.domains.AlgebraicField method), 1665
method), 1158
ang vel in() (sympy.physics.vector.frame.ReferenceFrame
algebraic eld() (sympy.polys.domains.domain.Domain method), 1665
method), 1149
angerj() (in module mpmath), 834
algebraic eld() (sympy.polys.domains.IntegerRing
angle (sympy.physics.optics.gaussopt.GeometricRay
method), 1156
attribute), 1806
algebraic eld() (sympy.polys.domains.RationalField
angle between() (sympy.geometry.line.LinearEntity
method), 1158
method), 509
AlgebraicField
(class
in angle between() (sympy.geometry.line3d.LinearEntity3D
sympy.polys.domains), 1158
method), 527
AlgebraicNumber
(class
in angle between() (sympy.geometry.plane.Plane
sympy.polys.numberelds), 1117
method), 577
all coes()
(sympy.polys.polyclasses.DMP angles (sympy.geometry.polygon.Polygon atmethod), 1162
tribute), 558
all coes()
(sympy.polys.polytools.Poly angles (sympy.geometry.polygon.RegularPolygon
method), 1077
attribute), 564
all monoms() (sympy.polys.polyclasses.DMP angular momentum()
(in
module
method), 1162
sympy.physics.mechanics.functions),
all monoms()
(sympy.polys.polytools.Poly
1738
method), 1077
angular momentum()
all roots()
(sympy.polys.polytools.Poly
(sympy.physics.mechanics.particle.Particle
method), 1077
method), 1731
all terms()
(sympy.polys.polyclasses.DMP angular momentum()
method), 1163
(sympy.physics.mechanics.rigidbody.RigidBody
all terms()
(sympy.polys.polytools.Poly
method), 1733
angular velocity (sympy.physics.optics.waves.TWave
method), 1078
allhints (in module sympy.solvers.ode), 1405
attribute), 1818
Index
1925
SymPy Documentation, Release 0.7.6
AnnihilateBoson
(class
in
method), 1609
sympy.physics.secondquant), 1608
apply operator() (sympy.physics.secondquant.CreateFermion
AnnihilateFermion
(class
in
method), 1610
sympy.physics.secondquant), 1609
apply operators()
(in
module
annotated()
(in
module
sympy.physics.secondquant), 1613
sympy.printing.pretty.pretty symbology),
applyfunc() (sympy.matrices.sparse.SparseMatrix
1267
method), 693
ANP (class in sympy.polys.polyclasses), 1168 approximation() (sympy.core.numbers.NumberSymbol
AntiCommutator
(class
in
method), 131
sympy.physics.quantum.anticommutator),
arange() (in module mpmath), 742
1749
arbitrary metric()
(in
module
antihermitian, 89
sympy.galgebra.ga), 467
AntiSymmetricTensor
(class
in arbitrary metric conformal()
(in
module
sympy.physics.secondquant), 1620
sympy.galgebra.ga), 467
any() (in module sympy.parsing.sympy tokenize),arbitrary point() (sympy.geometry.curve.Curve
1590
method), 540
aother (sympy.functions.special.hyper.meijerg arbitrary point() (sympy.geometry.ellipse.Ellipse
attribute), 430
method), 544
ap (sympy.functions.special.hyper.hyper at- arbitrary point() (sympy.geometry.line.LinearEntity
tribute), 427
method), 510
ap (sympy.functions.special.hyper.meijerg at- arbitrary point() (sympy.geometry.line3d.LinearEntity3D
tribute), 430
method), 527
apart() (in module sympy.polys.partfrac), arbitrary point() (sympy.geometry.plane.Plane
1121
method), 578
apart() (sympy.core.expr.Expr method), 103
arbitrary point() (sympy.geometry.polygon.Polygon
method), 558
apart list() (in module sympy.polys.partfrac),
1122
Arcsin() (in module sympy.stats), 1365
apery (mpmath.mp attribute), 751
are collinear() (sympy.geometry.point3d.Point3D
apoapsis (sympy.geometry.ellipse.Ellipse atstatic method), 504
tribute), 544
are concurrent() (sympy.geometry.line.LinearEntity
apothem (sympy.geometry.polygon.RegularPolygon
static method), 510
attribute), 564
are concurrent() (sympy.geometry.line3d.LinearEntity3D
appellf1() (in module mpmath), 903
static method), 528
appellf2() (in module mpmath), 906
are concurrent() (sympy.geometry.plane.Plane
static method), 578
appellf3() (in module mpmath), 906
appellf4() (in module mpmath), 908
are coplanar() (sympy.geometry.point3d.Point3D
static method), 504
append() (sympy.plotting.plot.Plot method),
1272
are similar() (in module sympy.geometry.util),
AppliedPredicate
(class
in
496
sympy.assumptions.assume), 1288
area
(sympy.geometry.ellipse.Ellipse
attribute), 544
apply() (sympy.printing.pretty.stringpict.prettyForm
static method), 1269
area (sympy.geometry.polygon.Polygon atapply()
(sympy.simplify.epathtools.EPath
tribute), 559
area (sympy.geometry.polygon.RegularPolygon
method), 1348
apply nite di()
(in
module
attribute), 564
arg (class in sympy.functions.elementary.complexes),
sympy.calculus.nite di), 1594
apply grover()
(in
module
344
arg (sympy.assumptions.assume.AppliedPredicate
sympy.physics.quantum.grover),
1795
attribute), 1288
apply operator() (sympy.physics.secondquant.AnnihilateBoson
arg() (in module mpmath), 734
args (sympy.core.basic.Basic attribute), 90
method), 1608
apply operator() (sympy.physics.secondquant.AnnihilateFermion
args (sympy.geometry.polygon.RegularPolygon
method), 1609
attribute), 565
apply operator() (sympy.physics.secondquant.CreateBoson
args (sympy.polys.polytools.Poly attribute),
1926
Index
SymPy Documentation, Release 0.7.6
1078
as content primitive()
args (sympy.tensor.indexed.IndexedBase at(sympy.core.basic.Basic
method),
tribute), 1513
91
args cnc() (sympy.core.expr.Expr method), as content primitive() (sympy.core.expr.Expr
103
method), 106
Argument (class in sympy.utilities.codegen), as content primitive() (sympy.core.mul.Mul
1540
method), 142
argument (sympy.functions.special.bessel.BesselBase
as content primitive()
attribute), 408
(sympy.core.numbers.Rational
argument (sympy.functions.special.hyper.hyper
method), 131
attribute), 427
as content primitive()
argument (sympy.functions.special.hyper.meijerg
(sympy.core.power.Pow
method),
attribute), 430
140
array form (sympy.combinatorics.permutations.Permutation
as dict() (sympy.combinatorics.partitions.IntegerPartition
attribute), 197
method), 189
array form (sympy.combinatorics.polyhedron.Polyhedron
as dict() (sympy.polys.polytools.Poly method),
attribute), 243
1078
ArrowStringDescription
(class
in as dummy()
(sympy.core.symbol.Symbol
sympy.categories.diagram drawing),
method), 122
1841
as explicit() (sympy.matrices.expressions.MatrixExpr
as base exp() (sympy.core.function.Function
method), 707
method), 170
as expr() (sympy.core.expr.Expr method), 107
as base exp()
(sympy.core.power.Pow as expr() (sympy.polys.numberelds.AlgebraicNumber
method), 140
method), 1117
as base exp() (sympy.functions.elementary.exponential.log
as expr()
(sympy.polys.polytools.Poly
method), 353
method), 1078
as coe Add() (sympy.core.add.Add method), as ferrers() (sympy.combinatorics.partitions.IntegerPartition
144
method), 189
as coe add() (sympy.core.add.Add method), as nite di()
(in
module
sympy.calculus.nite di), 1595
144
as coe Add()
(sympy.core.expr.Expr as immutable() (sympy.matrices.sparse.SparseMatrix
method), 103
method), 693
as coe add()
(sympy.core.expr.Expr as independent()
(sympy.core.expr.Expr
method), 103
method), 107
as coe Add() (sympy.core.numbers.Number as int() (in module sympy.core.compatibility),
method), 127
184
as coe exponent()
(sympy.core.expr.Expr as leading term()
(sympy.core.expr.Expr
method), 104
method), 109
as coe Mul()
(sympy.core.expr.Expr as list() (sympy.polys.polytools.Poly method),
method), 103
1079
as coe mul()
(sympy.core.expr.Expr as mutable() (sympy.matrices.expressions.MatrixExpr
method), 104
method), 708
as coe Mul() (sympy.core.mul.Mul method), as mutable() (sympy.matrices.immutable.ImmutableMatrix
142
method), 705
as coe Mul() (sympy.core.numbers.Number as mutable() (sympy.matrices.sparse.SparseMatrix
method), 127
method), 693
as coecient()
(sympy.core.expr.Expr as numer denom()
(sympy.core.expr.Expr
method), 104
method), 109
as coecients dict()
(sympy.core.add.Add as ordered factors()
(sympy.core.expr.Expr
method), 144
method), 109
as coecients dict() (sympy.core.expr.Expr as ordered factors()
(sympy.core.mul.Mul
method), 105
method), 142
as content primitive() (sympy.core.add.Add as ordered terms()
(sympy.core.expr.Expr
method), 145
method), 109
Index
1927
SymPy Documentation, Release 0.7.6
as poly() (sympy.core.basic.Basic method), 91 AskAntiHermitianHandler
(class
in
as poly() (sympy.polys.numberelds.AlgebraicNumber sympy.assumptions.handlers.sets),
method), 1117
1293
as powers dict()
(sympy.core.expr.Expr AskBoundedHandler
(class
in
method), 110
sympy.assumptions.handlers.calculus),
as quantity (sympy.physics.unitsystems.units.Unit
1291
attribute), 1828
AskComplexHandler
(class
in
as real imag() (sympy.core.add.Add method),
sympy.assumptions.handlers.sets),
145
1294
as real imag()
(sympy.core.expr.Expr AskExtendedRealHandler
(class
in
method), 110
sympy.assumptions.handlers.sets),
as real imag() (sympy.functions.elementary.complexes.re
1294
method), 358
AskHermitianHandler
(class
in
as real imag() (sympy.functions.elementary.exponential.exp
sympy.assumptions.handlers.sets),
1294
method), 350
as real imag() (sympy.functions.elementary.exponential.log
AskImaginaryHandler
(class
in
method), 353
sympy.assumptions.handlers.sets),
as real imag() (sympy.functions.elementary.hyperbolic.sinh
1294
method), 360
AskInnitesimalHandler
(class
in
as relational()
(sympy.sets.sets.FiniteSet
sympy.assumptions.handlers.calculus),
method), 1320
1292
as relational() (sympy.sets.sets.Intersection AskIntegerHandler
(class
in
method), 1321
sympy.assumptions.handlers.sets),
as relational()
(sympy.sets.sets.Interval
1294
method), 1318
AskNegativeHandler
(class
in
sympy.assumptions.handlers.order),
as relational()
(sympy.sets.sets.Union
method), 1320
1293
as sum() (sympy.integrals.Integral method), AskNonZeroHandler
(class
in
615
sympy.assumptions.handlers.order),
as terms() (sympy.core.expr.Expr method),
1293
110
AskOddHandler
(class
in
as two terms() (sympy.core.add.Add method),
sympy.assumptions.handlers.ntheory),
145
1293
as two terms() (sympy.core.mul.Mul method), AskPositiveHandler
(class
in
sympy.assumptions.handlers.order),
142
as unit (sympy.physics.unitsystems.quantities.Quantity 1293
attribute), 1829
AskPrimeHandler
(class
in
ascents() (sympy.combinatorics.permutations.Permutation
sympy.assumptions.handlers.ntheory),
method), 197
1293
asec() (in module mpmath), 773
AskRationalHandler
(class
in
sympy.assumptions.handlers.sets),
asech() (in module mpmath), 776
asin (class in sympy.functions.elementary.trigonometric),
1294
344
AskRealHandler
(class
in
asin() (in module mpmath), 771
sympy.assumptions.handlers.sets),
asinh (class in sympy.functions.elementary.hyperbolic), 1295
345
assemble partfrac list()
(in
module
asinh() (in module mpmath), 776
sympy.polys.partfrac), 1124
ask() (in module sympy.assumptions.ask), AssignmentError, 1266
1287
assoc laguerre
(class
in
sympy.functions.special.polynomials),
ask full inference()
(in
module
442
sympy.assumptions.ask), 1287
AskAlgebraicHandler
(class
in assoc legendre
(class
in
sympy.assumptions.handlers.sets),
sympy.functions.special.polynomials),
1293
439
1928
Index
SymPy Documentation, Release 0.7.6
assoc recurrence memo()
(in
module base solution linear()
(in
module
sympy.utilities.memoization), 1576
sympy.solvers.diophantine), 1490
assuming()
(in
module base vector() (sympy.digeom.CoordSystem
sympy.assumptions.assume), 1289
method), 1848
assumptions0 (sympy.core.basic.Basic at- base vectors() (sympy.digeom.CoordSystem
tribute), 91
method), 1848
AssumptionsContext
(class
in BaseCovarDerivativeOp
(class
in
sympy.assumptions.assume), 1288
sympy.digeom), 1855
atan (class in sympy.functions.elementary.trigonometric),
BasePolynomialError
(class
in
345
sympy.polys.polyerrors), 1238
atan() (in module mpmath), 772
BaseScalarField (class in sympy.digeom),
atan2 (class in sympy.functions.elementary.trigonometric),
1850
346
BaseSeries (class in sympy.plotting.plot),
atan2() (in module mpmath), 773
1282
atanh (class in sympy.functions.elementary.hyperbolic),
baseswap() (sympy.combinatorics.perm groups.Permutation
347
method), 218
atanh() (in module mpmath), 776
BaseVectorField (class in sympy.digeom),
Atom (class in sympy.core.basic), 102
1851
AtomicExpr (class in sympy.core.expr), 122
Basic (class in sympy.core.basic), 90
atoms() (sympy.combinatorics.permutations.Permutation
basic dot()
(sympy.galgebra.vector.Vector
method), 197
static method), 484
atoms() (sympy.core.basic.Basic method), 92 basic orbits (sympy.combinatorics.perm groups.Permutation
atoms() (sympy.matrices.matrices.MatrixBase
attribute), 219
method), 648
basic stabilizers (sympy.combinatorics.perm groups.Permut
attribute), 219
atoms table
(in
module
sympy.printing.pretty.pretty symbology),
basic transversals (sympy.combinatorics.perm groups.Perm
1267
attribute), 220
auto number()
(in
module Basis() (in module sympy.galgebra.ga), 470
sympy.parsing.sympy parser), 1592
basis() (sympy.polys.agca.modules.FreeModule
method), 1133
auto symbol()
(in
module
sympy.parsing.sympy parser), 1592
BBra (in module sympy.physics.secondquant),
autoprec() (in module mpmath), 743
1612
autowrap()
(in
module Bd (in module sympy.physics.secondquant),
sympy.utilities.autowrap), 1535
1613
auxiliary eqs (sympy.physics.mechanics.kane.KanesMethod
BeamParameter
(class
in
attribute), 1742
sympy.physics.optics.gaussopt),
1807
B
bei() (in module mpmath), 831
B (in module sympy.physics.secondquant), bell (class in sympy.functions.combinatorial.numbers),
363
1613
bell() (in module mpmath), 960
B (sympy.physics.optics.gaussopt.RayTransferMatrix
below() (sympy.printing.pretty.stringpict.stringPict
attribute), 1803
method), 1268
backlunds() (in module mpmath), 942
Benini() (in module sympy.stats), 1366
barnesg() (in module mpmath), 790
ber() (in module mpmath), 829
base (sympy.combinatorics.perm groups.PermutationGroup
berkowitz() (sympy.matrices.matrices.MatrixBase
attribute), 217
method), 648
base (sympy.functions.elementary.exponential.exp
berkowitz
charpoly()
attribute), 351
(sympy.matrices.matrices.MatrixBase
base
(sympy.tensor.indexed.Indexed
atmethod), 649
tribute), 1511
berkowitz
det() (sympy.matrices.matrices.MatrixBase
base oneform() (sympy.digeom.CoordSystem
method),
650
method), 1848
berkowitz
eigenvals()
base oneforms() (sympy.digeom.CoordSystem
(sympy.matrices.matrices.MatrixBase
method), 1848
Index
1929
SymPy Documentation, Release 0.7.6
method), 650
365
berkowitz minors() (sympy.matrices.matrices.MatrixBase
binomial() (in module mpmath), 779
method), 650
Binomial() (in module sympy.stats), 1363
bernfrac() (in module mpmath), 957
binomial coecients()
(in
module
bernoulli
(class
in
sympy.ntheory.multinomial), 292
sympy.functions.combinatorial.numbers),
binomial coecients list()
(in
module
364
sympy.ntheory.multinomial), 293
bernoulli() (in module mpmath), 956
Bisection
(class
in
mpBernoulli() (in module sympy.stats), 1362
math.calculus.optimization), 977
bernpoly() (in module mpmath), 958
bisectors() (sympy.geometry.polygon.Triangle
BesselBase
(class
in
method), 571
sympy.functions.special.bessel),
bitlist from subset()
(sympy.combinatorics.subsets.Subset
408
class method), 250
besseli (class in sympy.functions.special.bessel),
410
BKet (in module sympy.physics.secondquant),
besseli() (in module mpmath), 816
1612
besselj (class in sympy.functions.special.bessel),block collapse()
(in
module
409
sympy.matrices.expressions.blockmatrix),
besselj() (in module mpmath), 810
712
besseljzero() (in module mpmath), 822
BlockDiagMatrix
(class
in
besselk (class in sympy.functions.special.bessel),
sympy.matrices.expressions.blockmatrix),
411
712
besselk() (in module mpmath), 819
BlockMatrix
(class
in
besselsimp()
(in
module
sympy.matrices.expressions.blockmatrix),
sympy.simplify.simplify), 1336
711
bessely (class in sympy.functions.special.bessel),bm
(sympy.functions.special.hyper.meijerg
409
attribute), 430
bessely() (in module mpmath), 814
bool map() (in module sympy.logic.boolalg),
besselyzero() (in module mpmath), 824
634
beta (class in sympy.functions.special.beta functions),
BooleanFalse (class in sympy.logic.boolalg),
386
627
beta() (in module mpmath), 785
BooleanTrue (class in sympy.logic.boolalg),
Beta() (in module sympy.stats), 1366
627
betainc() (in module mpmath), 786
BosonicBasis
(class
in
sympy.physics.secondquant), 1613
BetaPrime() (in module sympy.stats), 1367
bid5 square()
(in
module bother (sympy.functions.special.hyper.meijerg
sympy.crypto.crypto), 310
attribute), 430
bid6 square()
(in
module boundary (sympy.sets.sets.Set attribute),
sympy.crypto.crypto), 312
1312
bid7 square()
(in
module bq (sympy.functions.special.hyper.hyper atsympy.crypto.crypto), 312
tribute), 427
bihyper() (in module mpmath), 901
bq (sympy.functions.special.hyper.meijerg atbilinear function()
(in
module
tribute), 430
Bra (class in sympy.physics.quantum.state),
sympy.galgebra.ncutil), 486
bilinear product()
(in
module
1782
bra (sympy.physics.quantum.operator.OuterProduct
sympy.galgebra.ncutil), 486
bin to gray() (sympy.combinatorics.graycode
attribute), 1762
bra (sympy.physics.secondquant.InnerProduct
static method), 260
binary function()
(in
module
attribute), 1613
sympy.utilities.autowrap), 1535
BraBase
(class
in
binary partitions()
(in
module
sympy.physics.quantum.state), 1781
sympy.utilities.iterables), 1555
bracelets()
(in
module
binomial
(class
in
sympy.utilities.iterables), 1555
sympy.functions.combinatorial.factorials),
bsgs direct product()
(in
module
1930
Index
SymPy Documentation, Release 0.7.6
sympy.combinatorics.tensor can),
275
bspline basis()
(in
module
sympy.functions.special.bsplines),
419
bspline basis set()
(in
module
sympy.functions.special.bsplines),
420
build expression tree()
(in
module
sympy.series.gruntz), 1306
cardinality (sympy.combinatorics.permutations.Permutation
attribute), 197
cardinality (sympy.combinatorics.subsets.Subset
attribute), 250
casoratian()
(in
module
sympy.matrices.dense), 682
Catalan (class in sympy.core.numbers), 138
catalan (class in sympy.functions.combinatorial.numbers),
366
catalan (mpmath.mp attribute), 751
Category (class in sympy.categories), 1833
C
Cauchy() (in module sympy.stats), 1368
cbrt() (in module mpmath), 752
C (in module sympy.core.basic), 102
C (sympy.matrices.immutable.ImmutableMatrixccode() (in module sympy.printing.ccode),
1252
attribute), 705
CCodeGen
(class in sympy.utilities.codegen),
C (sympy.matrices.matrices.MatrixBase at1541
tribute), 644
CCodePrinter (class in sympy.printing.ccode),
C (sympy.physics.optics.gaussopt.RayTransferMatrix
1252
attribute), 1803
ceil() (in module mpmath), 736
cacheit() (in module sympy.core.cache), 90
ceiling (class in sympy.functions.elementary.integers),
calc nodes() (mpmath.calculus.quadrature.GaussLegendre
348
method), 1012
center (sympy.geometry.ellipse.Ellipse atcalc nodes() (mpmath.calculus.quadrature.QuadratureRule
tribute), 545
method), 1010
center (sympy.geometry.polygon.RegularPolygon
calc nodes() (mpmath.calculus.quadrature.TanhSinh
attribute), 565
method), 1011
calculate series()
(in
module center() (sympy.combinatorics.perm groups.PermutationGro
method), 220
sympy.series.gruntz), 1306
central inertia (sympy.physics.mechanics.rigidbody.RigidBod
can transf matrix (sympy.physics.unitsystems.dimensions.DimensionSystem
attribute), 1734
attribute), 1826
cancel() (in module sympy.polys.polytools), centralizer() (sympy.combinatorics.perm groups.Permutatio
method), 221
1073
(sympy.geometry.polygon.Polygon
cancel() (sympy.core.expr.Expr method), 110 centroid
attribute), 559
cancel()
(sympy.polys.polyclasses.DMF
centroid (sympy.geometry.polygon.RegularPolygon
method), 1167
attribute), 565
cancel()
(sympy.polys.polyclasses.DMP
centroid() (in module sympy.geometry.util),
method), 1163
497
cancel() (sympy.polys.polytools.Poly method),
CG (class in sympy.physics.quantum.cg),
1079
1750
cancel()
(sympy.polys.rings.PolyElement
cg simp()
(in
module
method), 1216
sympy.physics.quantum.cg), 1751
canon bp() (in module sympy.tensor.tensor),
CGate (class in sympy.physics.quantum.gate),
1531
1789
canon bp()
(sympy.tensor.tensor.TensAdd
CGateS (class in sympy.physics.quantum.gate),
method), 1527
1793
canon bp()
(sympy.tensor.tensor.TensMul
characteristic()
(sympy.polys.domains.domain.Domain
method), 1529
1150
method),
canonical variables (sympy.core.basic.Basic
characteristic()
(sympy.polys.domains.FiniteField
attribute), 93
method), 1155
canonicalize()
(in
module
charpoly()
(sympy.matrices.matrices.MatrixBase
sympy.combinatorics.tensor can),
method),
650
270
chebyt()
(in
module
mpmath), 1016
capture()
(in
module
chebyshevt
(class
in
sympy.utilities.iterables), 1555
Index
1931
SymPy Documentation, Release 0.7.6
sympy.functions.special.polynomials), circumcircle (sympy.geometry.polygon.Triangle
436
attribute), 572
chebyshevt poly()
(in
module circumference (sympy.geometry.ellipse.Circle
sympy.polys.orthopolys), 1120
attribute), 554
chebyshevt root
(class
in circumference (sympy.geometry.ellipse.Ellipse
sympy.functions.special.polynomials),
attribute), 545
438
circumradius (sympy.geometry.polygon.RegularPolygon
chebyshevu
(class
in
attribute), 566
sympy.functions.special.polynomials), circumradius (sympy.geometry.polygon.Triangle
437
attribute), 572
chebyshevu poly()
(in
module CL (sympy.matrices.sparse.SparseMatrix atsympy.polys.orthopolys), 1120
tribute), 692
chebyshevu root
(class
in class key()
(sympy.core.add.Add
class
sympy.functions.special.polynomials),
method), 146
438
class key()
(sympy.core.basic.Basic
class
chebyt() (in module mpmath), 873
method), 93
chebyu() (in module mpmath), 874
classify diop()
(in
module
check assumptions()
(in
module
sympy.solvers.diophantine), 1489
sympy.solvers.solvers), 1477
classify ode() (in module sympy.solvers.ode),
check output() (sympy.utilities.runtests.SymPyOutputChecker
1400
method), 1581
classify pde() (in module sympy.solvers.pde),
checkinfsol() (in module sympy.solvers.ode),
1463
1405
classof()
(in
module
checkodesol() (in module sympy.solvers.ode),
sympy.matrices.matrices), 678
1402
clcos() (in module mpmath), 948
checkpdesol() (in module sympy.solvers.pde), clear() (mpmath.calculus.quadrature.QuadratureRule
1464
method), 1010
checksol() (in module sympy.solvers.solvers), clear() (sympy.tensor.tensor. TensorManager
1477
method), 1517
Chi (class in sympy.functions.special.error functions),
clear denoms() (sympy.polys.polyclasses.DMP
407
method), 1163
chi() (in module mpmath), 803
clear denoms()
(sympy.polys.polytools.Poly
Chi() (in module sympy.stats), 1369
method), 1079
ChiNoncentral() (in module sympy.stats), clebsch gordan()
(in
module
sympy.physics.wigner), 1624
1369
ChiSquared() (in module sympy.stats), 1370
clsin() (in module mpmath), 946
cholesky() (in module mpmath), 1023
CMod (class in sympy.physics.quantum.shor),
cholesky() (sympy.matrices.matrices.MatrixBase
1801
method), 651
CNOT
(in
module
cholesky() (sympy.matrices.sparse.SparseMatrix
sympy.physics.quantum.gate), 1792
method), 693
CNotGate
(class
in
cholesky solve() (sympy.matrices.matrices.MatrixBase sympy.physics.quantum.gate), 1792
method), 651
CodeGen (class in sympy.utilities.codegen),
chop() (in module mpmath), 737
1540
Ci (class in sympy.functions.special.error functions),
codegen()
(in
module
405
sympy.utilities.codegen), 1544
ci() (in module mpmath), 801
CodePrinter
(class
in
Circle (class in sympy.geometry.ellipse), 554
sympy.printing.codeprinter), 1266
circumcenter (sympy.geometry.polygon.RegularPolygon
CodeWrapper
(class
in
attribute), 565
sympy.utilities.autowrap), 1534
circumcenter (sympy.geometry.polygon.Trianglecodomain (sympy.categories.CompositeMorphism
attribute), 572
attribute), 1832
circumcircle (sympy.geometry.polygon.RegularPolygon
codomain (sympy.categories.Morphism atattribute), 566
tribute), 1830
1932
Index
SymPy Documentation, Release 0.7.6
coef function()
(in
module
method), 687
sympy.galgebra.ncutil), 487
col swap() (sympy.matrices.sparse.MutableSparseMatrix
coe() (sympy.core.expr.Expr method), 110
method), 701
coe()
(sympy.polys.rings.PolyElement collect() (in module sympy.simplify.simplify),
method), 1216
1329
coe monomial() (sympy.polys.polytools.Poly collect() (sympy.core.expr.Expr method), 112
method), 1079
collect const()
(in
module
coecients (sympy.geometry.line.LinearEntity
sympy.simplify.simplify), 1341
attribute), 511
collect sqrt()
(in
module
coes() (sympy.polys.numberelds.AlgebraicNumber sympy.simplify.simplify), 1340
method), 1117
combsimp()
(in
module
coes()
(sympy.polys.polyclasses.DMP
sympy.simplify.simplify), 1338
method), 1163
combsimp() (sympy.core.expr.Expr method),
coes() (sympy.polys.polytools.Poly method),
112
1080
comm i2symbol() (sympy.tensor.tensor. TensorManager
coes()
(sympy.polys.rings.PolyElement
method), 1517
method), 1217
comm symbols2i() (sympy.tensor.tensor. TensorManager
CoercionFailed
(class
in
method), 1517
sympy.polys.polyerrors), 1238
common prex()
(in
module
cofactor() (sympy.matrices.matrices.MatrixBase
sympy.utilities.iterables), 1556
method), 651
common sux()
(in
module
cofactorMatrix() (sympy.matrices.matrices.MatrixBase sympy.utilities.iterables), 1556
method), 651
commutative, 89
cofactors() (in module sympy.polys.polytools), commutative diagrams
(sympy.categories.Category
at1066
tribute), 1834
cofactors()
(sympy.core.numbers.Number
method), 127
Commutator (class in sympy.digeom), 1852
cofactors() (sympy.polys.domains.domain.Domain
Commutator
(class
in
method), 1150
sympy.physics.quantum.commutator),
cofactors()
(sympy.polys.polyclasses.DMP
1752
method), 1163
Commutator
(class
in
cofactors()
(sympy.polys.polytools.Poly
sympy.physics.secondquant), 1615
method), 1080
commutator() (sympy.combinatorics.perm groups.Permutati
cohen alt() (in module mpmath), 998
method), 221
Coin() (in module sympy.stats), 1363
commutator() (sympy.combinatorics.permutations.Permutat
col()
(sympy.matrices.sparse.SparseMatrix
method), 198
method), 694
commutes with() (sympy.combinatorics.permutations.Permu
col del() (sympy.matrices.dense.MutableDenseMatrix method), 198
method), 686
commutes with() (sympy.tensor.tensor.TensorHead
col del() (sympy.matrices.sparse.MutableSparseMatrix method), 1525
method), 699
comp() (in module sympy.utilities.randtest),
col insert() (sympy.matrices.matrices.MatrixBase
1579
method), 652
compare() (in module sympy.series.gruntz),
col join() (sympy.matrices.matrices.MatrixBase
1306
method), 652
compare() (sympy.core.basic.Basic method),
col join() (sympy.matrices.sparse.MutableSparseMatrix93
method), 700
Complement (class in sympy.sets.sets), 1322
col list() (sympy.matrices.sparse.SparseMatrix complement() (sympy.sets.sets.Set method),
method), 694
1313
col op() (sympy.matrices.dense.MutableDenseMatrix
complex, 89
method), 686
ComplexInnity
(class
in
col op() (sympy.matrices.sparse.MutableSparseMatrix sympy.core.numbers), 136
ComplexSpace
(class
in
method), 700
col swap() (sympy.matrices.dense.MutableDenseMatrixsympy.physics.quantum.hilbert),
Index
1933
SymPy Documentation, Release 0.7.6
1758
construct domain()
(in
module
components (sympy.categories.CompositeMorphism
sympy.polys.constructor), 1115
attribute), 1832
contains() (sympy.combinatorics.perm groups.PermutationG
compose() (in module sympy.polys.polytools),
method), 222
1068
contains()
(sympy.geometry.line.Line
compose()
(sympy.categories.Morphism
method), 518
method), 1830
contains() (sympy.geometry.line.LinearEntity
compose()
(sympy.polys.polyclasses.DMP
method), 511
method), 1163
contains() (sympy.geometry.line.Ray method),
compose()
(sympy.polys.polytools.Poly
520
method), 1080
contains()
(sympy.geometry.line.Segment
composite, 89
method), 523
CompositeDomain
(class
in contains()
(sympy.geometry.line3d.Line3D
sympy.polys.domains.compositedomain),
method), 525
1155
contains() (sympy.geometry.line3d.LinearEntity3D
CompositeMorphism
(class
in
method), 529
sympy.categories), 1831
contains()
(sympy.geometry.line3d.Ray3D
ComputationFailed
(class
in
method), 535
sympy.polys.polyerrors), 1238
contains() (sympy.geometry.line3d.Segment3D
compute known facts()
(in
module
method), 537
sympy.assumptions.ask), 1287
contains()
(sympy.polys.agca.ideals.Ideal
compute leading term()
method), 1139
(sympy.core.expr.Expr
method), contains() (sympy.polys.agca.modules.Module
112
method), 1132
conclusions (sympy.categories.Diagram at- contains() (sympy.polys.polytools.GroebnerBasis
tribute), 1836
method), 1112
cond (sympy.functions.elementary.piecewise.ExprCondPair
contains()
(sympy.series.order.Order
attribute), 351
method), 1310
condition number() (sympy.matrices.matrices.MatrixBase
contains() (sympy.sets.sets.Set method), 1313
method), 652
contains interval()
(in
module
ConditionalDomain (class in sympy.stats.rv),
sympy.galgebra.precedence), 484
1396
content() (in module sympy.polys.polytools),
conj() (in module mpmath), 735
1067
conjugate
(class
in content()
(sympy.polys.polyclasses.DMP
sympy.functions.elementary.complexes),
method), 1163
348
content()
(sympy.polys.polytools.Poly
method), 1081
conjugate (sympy.combinatorics.partitions.IntegerPartition
attribute), 189
content()
(sympy.polys.rings.PolyElement
conjugate gauss beams()
(in
module
method), 1217
sympy.physics.optics.gaussopt),
continued fraction convergents() (in module
sympy.ntheory.continued fraction),
1811
connect to() (sympy.digeom.CoordSystem
298
method), 1849
continued fraction iterator()
(in
module
conserve mpmath dps()
(in
module
sympy.ntheory.continued fraction),
sympy.utilities.decorator), 1547
299
const()
(sympy.polys.rings.PolyElement continued fraction periodic()
(in
module
method), 1217
sympy.ntheory.continued fraction),
Constant
(class
in
299
sympy.physics.unitsystems.units),
continued fraction reduce()
(in
module
sympy.ntheory.continued fraction),
1828
constant renumber()
(in
module
300
sympy.solvers.ode), 1407
ContinuousDomain (class in sympy.stats.crv),
constantsimp()
(in
module
1396
sympy.solvers.ode), 1407
ContinuousPSpace (class in sympy.stats.crv),
1934
Index
SymPy Documentation, Release 0.7.6
1396
copyin matrix() (sympy.matrices.dense.MutableDenseMatrix
ContinuousRV() (in module sympy.stats),
method), 688
1391
cornacchia()
(in
module
contract metric() (sympy.tensor.tensor.TensAdd
sympy.solvers.diophantine), 1492
method), 1527
corners (sympy.combinatorics.polyhedron.Polyhedron
contract metric() (sympy.tensor.tensor.TensMul
attribute), 244
method), 1530
cos (class in sympy.functions.elementary.trigonometric),
contraction()
(in
module
349
sympy.physics.secondquant), 1616
cos() (in module mpmath), 767
controls (sympy.physics.quantum.gate.CGate coset factor() (sympy.combinatorics.perm groups.Permutatio
attribute), 1790
method), 223
controls (sympy.physics.quantum.gate.CNotGate
coset rank() (sympy.combinatorics.perm groups.Permutation
attribute), 1792
method), 224
convergence statement
coset unrank() (sympy.combinatorics.perm groups.Permutat
(sympy.functions.special.hyper.hyper
method), 224
cosh (class in sympy.functions.elementary.hyperbolic),
attribute), 427
convert() (sympy.polys.agca.modules.FreeModule
349
method), 1133
cosh() (in module mpmath), 774
convert() (sympy.polys.agca.modules.Module cosine transform()
(in
module
method), 1132
sympy.integrals.transforms), 588
convert() (sympy.polys.agca.modules.QuotientModule
cosm() (in module mpmath), 1029
method), 1140
cospi() (in module mpmath), 770
convert() (sympy.polys.agca.modules.SubModule
cot (class in sympy.functions.elementary.trigonometric),
method), 1134
350
convert() (sympy.polys.domains.domain.Domaincot() (in module mpmath), 770
method), 1150
coth (class in sympy.functions.elementary.hyperbolic),
convert()
(sympy.polys.polyclasses.DMP
350
method), 1163
coth() (in module mpmath), 776
convert from() (sympy.polys.domains.domain.Domain
could extract minus sign()
method), 1150
(sympy.core.expr.Expr
method),
convert to() (sympy.physics.unitsystems.quantities.Quantity
112
method), 1829
coulombc() (in module mpmath), 862
convert to native paths()
(in
module coulombf() (in module mpmath), 856
sympy.utilities.runtests), 1581
coulombg() (in module mpmath), 860
convert xor()
(in
module count() (sympy.core.basic.Basic method), 93
sympy.parsing.sympy parser), 1592
count complex roots()
(sympy.polys.polyclasses.DMP
convex hull()
(in
module
sympy.geometry.util), 495
method), 1163
coord function() (sympy.digeom.CoordSystem count ops() (in module sympy.core.function),
method), 1849
177
coord functions() (sympy.digeom.CoordSystemcount ops() (sympy.core.basic.Basic method),
method), 1849
93
coord tuple transform to()
count ops() (sympy.core.expr.Expr method),
112
(sympy.digeom.CoordSystem
count partitions() (sympy.utilities.enumerative.MultisetParti
method), 1849
CoordinateSym
(class
in
method), 1551
sympy.physics.vector.frame), 1664
count real roots() (sympy.polys.polyclasses.DMP
coords()
(sympy.digeom.Point
method),
method), 1163
1850
count roots()
(in
module
sympy.polys.polytools), 1072
CoordSystem (class in sympy.digeom), 1847
copy()
(sympy.polys.rings.PolyElement count roots()
(sympy.polys.polytools.Poly
method), 1217
method), 1081
copyin list() (sympy.matrices.dense.MutableDenseMatrix
couple()
(in
module
method), 687
sympy.physics.quantum.spin), 1779
Index
1935
SymPy Documentation, Release 0.7.6
CovarDerivativeOp (class in sympy.digeom),
attribute), 244
1855
CyclicGroup()
(in
module
cplot() (in module mpmath), 748
sympy.combinatorics.named groups),
CreateBoson
(class
in
261
sympy.physics.secondquant), 1609
cyclotomic() (in module mpmath), 966
CreateCGate()
(in
module cyclotomic poly()
(in
module
sympy.physics.quantum.circuitplot),
sympy.polys.specialpolys), 1120
1789
CythonCodeWrapper
(class
in
CreateFermion
(class
in
sympy.utilities.autowrap), 1534
sympy.physics.secondquant), 1610
cross()
(in
module D
sympy.physics.vector.functions),
D (sympy.matrices.matrices.MatrixBase at1687
tribute), 644
cross() (sympy.matrices.matrices.MatrixBase D (sympy.physics.optics.gaussopt.RayTransferMatrix
method), 652
attribute), 1803
cross() (sympy.physics.vector.dyadic.Dyadic D()
(sympy.physics.quantum.spin.Rotation
method), 1673
class method), 1771
cross()
(sympy.physics.vector.vector.Vector d()
(sympy.physics.quantum.spin.Rotation
method), 1669
class method), 1772
crt() (in module sympy.ntheory.modular), 290 Dagger (class in sympy.physics.quantum.dagger),
crt1() (in module sympy.ntheory.modular),
1753
291
Dagger (class in sympy.physics.secondquant),
crt2() (in module sympy.ntheory.modular),
1604
291
Dagum() (in module sympy.stats), 1371
csc() (in module mpmath), 769
DataType (class in sympy.utilities.codegen),
csch() (in module mpmath), 776
1540
cse() (in module sympy.simplify.cse main), dcm() (sympy.physics.vector.frame.ReferenceFrame
1345
method), 1665
curl() (in module sympy.physics.vector.eldfunctions),
DD() (in module sympy.galgebra.ga), 470
1690
debug() (in module sympy.utilities.misc),
current (sympy.combinatorics.graycode.GrayCode
1576
attribute), 257
debug decorator()
(in
module
Curve (class in sympy.geometry.curve), 539
sympy.utilities.misc), 1576
CurvedMirror
(class
in decipher bid5()
(in
module
sympy.physics.optics.gaussopt),
sympy.crypto.crypto), 310
1805
decipher bid6()
(in
module
CurvedRefraction
(class
in
sympy.crypto.crypto), 311
sympy.physics.optics.gaussopt),
decipher elgamal()
(in
module
1804
sympy.crypto.crypto), 319
Cycle (class in sympy.combinatorics.permutations),
decipher hill()
(in
module
213
sympy.crypto.crypto), 309
cycle length()
(in
module decipher kid rsa()
(in
module
sympy.ntheory.generate), 280
sympy.crypto.crypto), 315
cycle list() (in module sympy.crypto.crypto), decipher rsa()
(in
module
303
sympy.crypto.crypto), 314
cycle structure (sympy.combinatorics.permutations.Permutation
decipher vigenere()
(in
module
attribute), 199
sympy.crypto.crypto), 307
cycles (sympy.combinatorics.permutations.Permutation
decode morse()
(in
module
attribute), 199
sympy.crypto.crypto), 315
cyclic()
(sympy.combinatorics.generators decompose()
(in
module
static method), 215
sympy.polys.polytools), 1068
cyclic form (sympy.combinatorics.permutations.Permutation
decompose() (sympy.physics.quantum.gate.CGate
attribute), 199
method), 1790
cyclic form (sympy.combinatorics.polyhedron.Polyhedron
1936
Index
SymPy Documentation, Release 0.7.6
decompose() (sympy.physics.quantum.gate.SwapGate method), 1139
method), 1792
Derivative (class in sympy.core.function), 164
decompose() (sympy.physics.quantum.qft.IQFT derived series() (sympy.combinatorics.perm groups.Permuta
method), 1796
method), 225
decompose() (sympy.physics.quantum.qft.QFT derived subgroup() (sympy.combinatorics.perm groups.Perm
method), 1796
method), 225
decompose() (sympy.polys.polyclasses.DMP descent()
(in
module
method), 1163
sympy.solvers.diophantine), 1496
decompose()
(sympy.polys.polytools.Poly descents() (sympy.combinatorics.permutations.Permutation
method), 1081
method), 199
dene precedence()
(in
module det() (sympy.matrices.matrices.MatrixBase
sympy.galgebra.precedence), 484
method), 653
deate()
(sympy.polys.polyclasses.DMP det bareis() (sympy.matrices.matrices.MatrixBase
method), 1163
method), 653
deate()
(sympy.polys.polytools.Poly det LU decomposition()
method), 1081
(sympy.matrices.matrices.MatrixBase
degree (mpmath.mp attribute), 751
method), 653
degree (sympy.combinatorics.perm groups.PermutationGroup
deviation()
(in
module
attribute), 224
sympy.physics.optics.utils), 1815
degree() (in module sympy.polys.polytools), diag() (in module sympy.matrices.dense), 679
1057
diagonal solve() (sympy.matrices.matrices.MatrixBase
degree()
(sympy.polys.polyclasses.DMP
method), 653
method), 1163
diagonalize() (sympy.matrices.matrices.MatrixBase
degree()
(sympy.polys.polytools.Poly
method), 653
method), 1082
diagpq() (in module sympy.galgebra.ga), 467
degree()
(sympy.polys.rings.PolyElement Diagram (class in sympy.categories), 1835
method), 1218
DiagramGrid
(class
in
degree list()
(in
module
sympy.categories.diagram drawing),
sympy.polys.polytools), 1057
1838
degree list() (sympy.polys.polyclasses.DMP Dict (class in sympy.core.containers), 182
method), 1163
dict merge()
(in
module
degree list()
(sympy.polys.polytools.Poly
sympy.utilities.iterables), 1556
method), 1082
Die() (in module sympy.stats), 1362
degrees() (in module mpmath), 767
DiePSpace (class in sympy.stats.frv types),
degrees()
(sympy.polys.rings.PolyElement
1396
method), 1218
di() (in module mpmath), 999
delta (sympy.functions.special.hyper.meijerg di() (in module sympy.core.function), 167
attribute), 430
di() (sympy.matrices.matrices.MatrixBase
deltaintegrate()
(sympy.integrals
static
method), 654
method), 611
di()
(sympy.physics.vector.vector.Vector
method), 1670
denom() (sympy.polys.domains.AlgebraicField
method), 1158
di() (sympy.polys.polyclasses.DMP method),
denom() (sympy.polys.domains.domain.Domain
1163
di() (sympy.polys.polytools.Poly method),
method), 1150
denom() (sympy.polys.domains.ExpressionDomain
1082
di()
(sympy.polys.rings.PolyElement
method), 1161
denom() (sympy.polys.domains.FractionField
method), 1218
Dierential (class in sympy.digeom), 1852
method), 1159
denom()
(sympy.polys.domains.ring.Ring DierentialOperator
(class
in
method), 1154
sympy.physics.quantum.operator),
denom()
(sympy.polys.polyclasses.DMF
1763
method), 1167
dierint() (in module mpmath), 1002
density() (in module sympy.stats), 1393
dis() (in module mpmath), 1001
depth()
(sympy.polys.agca.ideals.Ideal dis exp() (in module mpmath), 1002
Index
1937
SymPy Documentation, Release 0.7.6
dis prod() (in module mpmath), 1001
direction cosine() (sympy.geometry.point3d.Point3D
digamma() (in module mpmath), 793
method), 505
digamma()
(in
module direction ratio (sympy.geometry.line3d.LinearEntity3D
sympy.functions.special.gamma functions),
attribute), 529
383
direction ratio() (sympy.geometry.point3d.Point3D
digit 2txt
(in
module
method), 505
sympy.printing.pretty.pretty symbology),
DirectProduct()
(in
module
1267
sympy.combinatorics.group constructs),
dihedral() (sympy.combinatorics.generators
269
static method), 216
dirichlet() (in module mpmath), 936
DihedralGroup()
(in
module dirichlet eta
(class
in
sympy.combinatorics.named groups),
sympy.functions.special.zeta functions),
261
422
dim (sympy.physics.unitsystems.dimensions.DimensionSystem
DiscreteUniform() (in module sympy.stats),
1362
attribute), 1826
dim (sympy.physics.unitsystems.units.UnitSystem
discriminant()
(in
module
attribute), 1828
sympy.polys.polytools), 1062
dim can vector() (sympy.physics.unitsystems.dimensions.DimensionSystem
discriminant() (sympy.polys.polyclasses.DMP
method), 1826
method), 1163
dim vector() (sympy.physics.unitsystems.dimensions.DimensionSystem
discriminant()
(sympy.polys.polytools.Poly
method), 1826
method), 1082
Dimension
(class
in dispersion()
(in
module
sympy.physics.unitsystems.dimensions),
sympy.polys.dispersion), 1062, 1126
1824
dispersion()
(sympy.polys.polytools.Poly
method), 1083
dimension (sympy.physics.quantum.hilbert.HilbertSpace
attribute), 1758
dispersionset()
(in
module
DimensionSystem
(class
in
sympy.polys.dispersion), 1063, 1125
sympy.physics.unitsystems.dimensions),dispersionset()
(sympy.polys.polytools.Poly
1825
method), 1084
diop bf DN()
(in
module distance()
(sympy.geometry.line.Line
sympy.solvers.diophantine), 1493
method), 518
diop DN()
(in
module distance() (sympy.geometry.line.Ray method),
sympy.solvers.diophantine), 1492
520
diop general pythagorean()
(in
module distance()
(sympy.geometry.line.Segment
sympy.solvers.diophantine), 1497
method), 523
diop general sum of squares() (in module distance()
(sympy.geometry.line3d.Line3D
sympy.solvers.diophantine), 1497
method), 526
diop linear()
(in
module distance()
(sympy.geometry.line3d.Ray3D
sympy.solvers.diophantine), 1490
method), 535
diop quadratic()
(in
module distance() (sympy.geometry.line3d.Segment3D
sympy.solvers.diophantine), 1491
method), 538
diop solve()
(in
module distance()
(sympy.geometry.plane.Plane
sympy.solvers.diophantine), 1488
method), 578
diop ternary quadratic()
(in
module distance()
(sympy.geometry.point.Point
sympy.solvers.diophantine), 1495
method), 498
diop ternary quadratic normal() (in module distance() (sympy.geometry.point3d.Point3D
sympy.solvers.diophantine), 1502
method), 505
diophantine()
(in
module distance() (sympy.geometry.polygon.Polygon
sympy.solvers.diophantine), 1487
method), 559
DiracDelta
(class
in div() (in module sympy.polys.polytools), 1059
div() (sympy.polys.domains.domain.Domain
sympy.functions.special.delta functions),
377
method), 1150
direction cosine (sympy.geometry.line3d.LinearEntity3D
div()
(sympy.polys.domains.eld.Field
attribute), 529
method), 1153
1938
Index
SymPy Documentation, Release 0.7.6
div()
(sympy.polys.domains.ring.Ring
sympy.polys.densetools), 1189
method), 1154
dmp di eval in()
(in
module
div() (sympy.polys.polyclasses.DMP method),
sympy.polys.densetools), 1191
1163
dmp di in()
(in
module
div() (sympy.polys.polytools.Poly method),
sympy.polys.densetools), 1190
1085
dmp discriminant()
(in
module
div()
(sympy.polys.rings.PolyElement
sympy.polys.euclidtools), 1228
dmp div()
(in
module
method), 1218
divergence (sympy.physics.optics.gaussopt.BeamParameter
sympy.polys.densearith), 1187
attribute), 1807
dmp eject()
(in
module
divergence()
(in
module
sympy.polys.densebasic), 1180
sympy.physics.vector.eldfunctions), dmp euclidean prs()
(in
module
sympy.polys.euclidtools), 1226
1690
divisible()
(in
module dmp eval()
(in
module
sympy.solvers.diophantine), 1500
sympy.polys.densetools), 1190
divisor count()
(in
module dmp eval in()
(in
module
sympy.ntheory.factor ), 290
sympy.polys.densetools), 1190
divisors() (in module sympy.ntheory.factor ), dmp eval tail()
(in
module
289
sympy.polys.densetools), 1190
DMF (class in sympy.polys.polyclasses), 1167 dmp exclude()
(in
module
DMP (class in sympy.polys.polyclasses), 1162
sympy.polys.densebasic), 1179
dmp abs()
(in
module dmp expand()
(in
module
sympy.polys.densearith), 1183
sympy.polys.densearith), 1189
dmp add()
(in
module dmp exquo()
(in
module
sympy.polys.densearith), 1184
sympy.polys.densearith), 1188
dmp add ground()
(in
module dmp exquo ground()
(in
module
sympy.polys.densearith), 1182
sympy.polys.densearith), 1183
dmp add mul()
(in
module dmp ext factor()
(in
module
sympy.polys.densearith), 1184
sympy.polys.factortools), 1235
dmp add term()
(in
module dmp factor list()
(in
module
sympy.polys.densearith), 1181
sympy.polys.factortools), 1236
dmp apply pairs()
(in
module dmp factor list include()
(in
module
sympy.polys.densebasic), 1180
sympy.polys.factortools), 1236
dmp cancel()
(in
module dmp div()
(in
module
sympy.polys.euclidtools), 1232
sympy.polys.densearith), 1187
dmp clear denoms()
(in
module dmp prs gcd()
(in
module
sympy.polys.densetools), 1197
sympy.polys.euclidtools), 1229
dmp compose()
(in
module dmp from dict()
(in
module
sympy.polys.densetools), 1195
sympy.polys.densebasic), 1176
dmp content()
(in
module dmp from sympy()
(in
module
sympy.polys.euclidtools), 1231
sympy.polys.densebasic), 1173
dmp convert()
(in
module dmp gcd()
(in
module
sympy.polys.densebasic), 1173
sympy.polys.euclidtools), 1231
dmp copy()
(in
module dmp gcdex()
(in
module
sympy.polys.densebasic), 1172
sympy.polys.euclidtools), 1225
dmp deate()
(in
module dmp ground()
(in
module
sympy.polys.densebasic), 1178
sympy.polys.densebasic), 1175
dmp degree()
(in
module dmp ground content()
(in
module
sympy.polys.densebasic), 1170
sympy.polys.densetools), 1193
dmp degree in()
(in
module dmp ground extract()
(in
module
sympy.polys.densebasic), 1171
sympy.polys.densetools), 1194
dmp degree list()
(in
module dmp ground LC()
(in
module
sympy.polys.densebasic), 1171
sympy.polys.densebasic), 1170
dmp di()
(in
module dmp ground monic()
(in
module
Index
1939
SymPy Documentation, Release 0.7.6
sympy.polys.densetools), 1192
dmp ground nth()
(in
module
sympy.polys.densebasic), 1174
dmp ground p()
(in
module
sympy.polys.densebasic), 1175
dmp ground primitive()
(in
module
sympy.polys.densetools), 1193
dmp ground TC()
(in
module
sympy.polys.densebasic), 1170
dmp ground trunc()
(in
module
sympy.polys.densetools), 1191
dmp grounds()
(in
module
sympy.polys.densebasic), 1176
dmp half gcdex()
(in
module
sympy.polys.euclidtools), 1225
dmp include()
(in
module
sympy.polys.densebasic), 1179
dmp inate()
(in
module
sympy.polys.densebasic), 1178
dmp inject()
(in
module
sympy.polys.densebasic), 1179
dmp inner gcd()
(in
module
sympy.polys.euclidtools), 1230
dmp inner subresultants()
(in
module
sympy.polys.euclidtools), 1226
dmp integrate()
(in
module
sympy.polys.densetools), 1189
dmp integrate in()
(in
module
sympy.polys.densetools), 1189
dmp invert()
(in
module
sympy.polys.euclidtools), 1226
dmp irreducible p()
(in
module
sympy.polys.factortools), 1236
dmp l1 norm()
(in
module
sympy.polys.densearith), 1188
dmp LC()
(in
module
sympy.polys.densebasic), 1169
dmp lcm()
(in
module
sympy.polys.euclidtools), 1231
dmp lift()
(in
module
sympy.polys.densetools), 1196
dmp list terms()
(in
module
sympy.polys.densebasic), 1180
dmp max norm()
(in
module
sympy.polys.densearith), 1188
dmp mul()
(in
module
sympy.polys.densearith), 1185
dmp mul ground()
(in
module
sympy.polys.densearith), 1182
dmp mul term()
(in
module
sympy.polys.densearith), 1181
dmp multi deate()
(in
module
sympy.polys.densebasic), 1178
dmp neg()
(in
module
1940
sympy.polys.densearith), 1184
dmp negative p()
(in
module
sympy.polys.densebasic), 1176
dmp nest()
(in
module
sympy.polys.densebasic), 1177
dmp normal()
(in
module
sympy.polys.densebasic), 1173
dmp nth()
(in
module
sympy.polys.densebasic), 1173
dmp one()
(in
module
sympy.polys.densebasic), 1175
dmp one p()
(in
module
sympy.polys.densebasic), 1174
dmp pdiv()
(in
module
sympy.polys.densearith), 1185
dmp permute()
(in
module
sympy.polys.densebasic), 1177
dmp pexquo()
(in
module
sympy.polys.densearith), 1186
dmp positive p()
(in
module
sympy.polys.densebasic), 1176
dmp pow()
(in
module
sympy.polys.densearith), 1185
dmp pquo()
(in
module
sympy.polys.densearith), 1186
dmp prem()
(in
module
sympy.polys.densearith), 1186
dmp primitive()
(in
module
sympy.polys.euclidtools), 1231
dmp primitive prs()
(in
module
sympy.polys.euclidtools), 1226
dmp prs resultant()
(in
module
sympy.polys.euclidtools), 1227
dmp qq collins resultant()
(in
module
sympy.polys.euclidtools), 1228
dmp qq heu gcd()
(in
module
sympy.polys.euclidtools), 1230
dmp quo()
(in
module
sympy.polys.densearith), 1188
dmp quo ground()
(in
module
sympy.polys.densearith), 1182
dmp raise()
(in
module
sympy.polys.densebasic), 1178
dmp rem()
(in
module
sympy.polys.densearith), 1187
dmp resultant()
(in
module
sympy.polys.euclidtools), 1228
dmp revert()
(in
module
sympy.polys.densetools), 1197
dmp rr div()
(in
module
sympy.polys.densearith), 1187
dmp rr prs gcd()
(in
module
sympy.polys.euclidtools), 1229
dmp slice()
(in
module
Index
SymPy Documentation, Release 0.7.6
sympy.polys.densebasic), 1181
dmp sqr()
(in
module
sympy.polys.densearith), 1185
dmp strip()
(in
module
sympy.polys.densebasic), 1171
dmp sub()
(in
module
sympy.polys.densearith), 1184
dmp sub ground()
(in
module
sympy.polys.densearith), 1182
dmp sub mul()
(in
module
sympy.polys.densearith), 1184
dmp sub term()
(in
module
sympy.polys.densearith), 1181
dmp subresultants()
(in
module
sympy.polys.euclidtools), 1226
dmp swap()
(in
module
sympy.polys.densebasic), 1177
dmp TC()
(in
module
sympy.polys.densebasic), 1169
dmp terms gcd()
(in
module
sympy.polys.densebasic), 1180
dmp to dict()
(in
module
sympy.polys.densebasic), 1177
dmp to tuple()
(in
module
sympy.polys.densebasic), 1172
dmp trial division()
(in
module
sympy.polys.factortools), 1232
dmp true LT()
(in
module
sympy.polys.densebasic), 1170
dmp trunc()
(in
module
sympy.polys.densetools), 1191
dmp validate()
(in
module
sympy.polys.densebasic), 1171
dmp zero()
(in
module
sympy.polys.densebasic), 1174
dmp zero p()
(in
module
sympy.polys.densebasic), 1174
dmp zeros()
(in
module
sympy.polys.densebasic), 1175
dmp zz collins resultant()
(in
module
sympy.polys.euclidtools), 1228
dmp zz diophantine()
(in
module
sympy.polys.factortools), 1234
dmp zz factor()
(in
module
sympy.polys.factortools), 1235
dmp zz heu gcd()
(in
module
sympy.polys.euclidtools), 1229
dmp zz mignotte bound()
(in
module
sympy.polys.factortools), 1232
dmp zz modular resultant()
(in
module
sympy.polys.euclidtools), 1227
dmp zz wang()
(in
module
sympy.polys.factortools), 1234
dmp zz wang hensel lifting()
(in
module
Index
sympy.polys.factortools), 1234
dmp zz wang lead coes()
(in
module
sympy.polys.factortools), 1234
dmp zz wang non divisors()
(in
module
sympy.polys.factortools), 1234
dmp zz wang test points()
(in
module
sympy.polys.factortools), 1234
doctest() (in module sympy.utilities.runtests),
1581
doctest depends on()
(in
module
sympy.utilities.decorator), 1547
doit() (sympy.core.basic.Basic method), 93
doit() (sympy.functions.elementary.piecewise.Piecewise
method), 357
doit() (sympy.integrals.Integral method), 616
doit() (sympy.physics.quantum.anticommutator.AntiCommut
method), 1750
doit() (sympy.physics.quantum.commutator.Commutator
method), 1753
doit() (sympy.physics.secondquant.AntiSymmetricTensor
method), 1621
doit() (sympy.physics.secondquant.Commutator
method), 1615
doit()
(sympy.physics.secondquant.NO
method), 1617
doit()
(sympy.physics.vector.dyadic.Dyadic
method), 1673
doit()
(sympy.physics.vector.vector.Vector
method), 1670
doit() (sympy.series.limits.Limit method),
1304
doit numerically() (sympy.core.function.Derivative
method), 167
Domain (class in sympy.polys.domains.domain),
1149
domain (sympy.categories.CompositeMorphism
attribute), 1832
domain
(sympy.categories.Morphism
attribute), 1830
domain
(sympy.polys.polytools.Poly
attribute), 1085
DomainError
(class
in
sympy.polys.polyerrors), 1238
doprint() (sympy.printing.mathematica.MCodePrinter
method), 1259
doprint() (sympy.printing.mathml.MathMLPrinter
method), 1262
doprint()
(sympy.printing.printer.Printer
method), 1251
dot() (in module sympy.physics.vector.functions),
1687
dot() (sympy.geometry.point.Point method),
499
dot()
(sympy.geometry.point3d.Point3D
1941
SymPy Documentation, Release 0.7.6
method), 506
dump m() (sympy.utilities.codegen.OctaveCodeGen
(sympy.matrices.matrices.MatrixBase
method), 1543
method), 654
dump pyx() (sympy.utilities.autowrap.CythonCodeWrapper
dot()
(sympy.physics.vector.dyadic.Dyadic
method), 1534
method), 1673
dup content()
(in
module
dot()
(sympy.physics.vector.vector.Vector
sympy.polys.densetools), 1192
method), 1670
dup cyclotomic p()
(in
module
dotprint() (in module sympy.printing.dot),
sympy.polys.factortools), 1233
1269
dup decompose()
(in
module
double coset can rep()
(in
module
sympy.polys.densetools), 1196
sympy.combinatorics.tensor can),
dup extract()
(in
module
272
sympy.polys.densetools), 1194
draw() (sympy.categories.diagram drawing.XypicDiagramDrawer
dup gf factor()
(in
module
method), 1844
sympy.polys.factortools), 1236
drop() (sympy.polys.rings.PolyRing method), dup lshift()
(in
module
1216
sympy.polys.densearith), 1183
drop to ground() (sympy.polys.rings.PolyRing dup mirror()
(in
module
method), 1216
sympy.polys.densetools), 1194
dsolve() (in module sympy.solvers.ode), 1397 dup monic()
(in
module
dt()
(sympy.physics.vector.dyadic.Dyadic
sympy.polys.densetools), 1192
method), 1673
dup primitive()
(in
module
dt()
(sympy.physics.vector.vector.Vector
sympy.polys.densetools), 1193
method), 1671
dup random()
(in
module
dtype (sympy.polys.agca.modules.FreeModule
sympy.polys.densebasic), 1181
attribute), 1133
dup real imag()
(in
module
sympy.polys.densetools), 1194
dtype (sympy.polys.agca.modules.QuotientModule
attribute), 1141
dup reverse()
(in
module
dtype (sympy.polys.domains.AlgebraicField
sympy.polys.densebasic), 1172
attribute), 1158
dup rshift()
(in
module
sympy.polys.densearith), 1183
dtype (sympy.polys.domains.ExpressionDomain
attribute), 1161
dup scale()
(in
module
dual (sympy.physics.quantum.state.StateBase
sympy.polys.densetools), 1195
attribute), 1781
dup shift()
(in
module
dual() (sympy.matrices.matrices.MatrixBase
sympy.polys.densetools), 1195
method), 655
dup sign variations()
(in
module
dual class() (sympy.physics.quantum.state.StateBase sympy.polys.densetools), 1196
class method), 1781
dup transform()
(in
module
Dummy (class in sympy.core.symbol), 124
sympy.polys.densetools), 1195
dummy eq()
(sympy.core.basic.Basic dup zz cyclotomic factor()
(in
module
method), 94
sympy.polys.factortools), 1233
DummyWrapper
(class
in dup zz cyclotomic poly()
(in
module
sympy.utilities.autowrap), 1534
sympy.polys.factortools), 1233
dump c() (sympy.utilities.autowrap.UfuncifyCodeWrapper
dup zz factor()
(in
module
method), 1534
sympy.polys.factortools), 1234
dump c() (sympy.utilities.codegen.CCodeGen dup zz factor sqf()
(in
module
method), 1541
sympy.polys.factortools), 1233
dump code() (sympy.utilities.codegen.CodeGen dup zz hensel lift()
(in
module
method), 1540
sympy.polys.factortools), 1232
dump f95() (sympy.utilities.codegen.FCodeGen dup zz hensel step()
(in
module
method), 1542
sympy.polys.factortools), 1232
dump h() (sympy.utilities.codegen.CCodeGen dup zz irreducible p()
(in
module
method), 1542
sympy.polys.factortools), 1233
dump h() (sympy.utilities.codegen.FCodeGen dup zz zassenhaus()
(in
module
method), 1543
sympy.polys.factortools), 1233
dot()
1942
Index
SymPy Documentation, Release 0.7.6
Dyadic (class in sympy.physics.vector.dyadic), elliprj() (in module mpmath), 923
1673
Ellipse (class in sympy.geometry.ellipse), 542
dynamicsymbols()
(in
module elliptic e
(class
in
sympy.physics.vector), 1686
sympy.functions.special.elliptic integrals),
431
E
elliptic f
(class
in
sympy.functions.special.elliptic
integrals),
e (mpmath.mp attribute), 751
431
E() (in module sympy.stats), 1392
elliptic
k
(class
in
e1() (in module mpmath), 798
sympy.functions.special.elliptic
integrals),
E1() (in module sympy.functions.special.error functions),
430
401
elliptic
pi
(class
in
E n() (in module sympy.physics.qho 1d), 1602
sympy.functions.special.elliptic integrals),
E nl() (in module sympy.physics.hydrogen),
432
1598
EM() (sympy.polys.polytools.Poly method),
E nl() (in module sympy.physics.sho), 1603
1075
E nl dirac()
(in
module
emptyPrinter() (sympy.printing.repr.ReprPrinter
sympy.physics.hydrogen), 1598
method), 1263
EC() (sympy.polys.polytools.Poly method),
EmptySet (class in sympy.sets.sets), 1323
1075
(in
module
eccentricity (sympy.geometry.ellipse.Ellipse encipher ane()
sympy.crypto.crypto), 304
attribute), 545
encipher bid5()
(in
module
edges (sympy.combinatorics.polyhedron.Polyhedron
sympy.crypto.crypto), 309
attribute), 244
(in
module
edges() (sympy.combinatorics.prufer.Prufer encipher bid6()
sympy.crypto.crypto), 311
static method), 246
(in
module
egyptian fraction()
(in
module encipher bid7()
sympy.crypto.crypto), 312
sympy.ntheory.egyptian fraction),
encipher elgamal()
(in
module
301
sympy.crypto.crypto), 319
Ei (class in sympy.functions.special.error functions),
encipher hill()
(in
module
398
sympy.crypto.crypto), 308
ei() (in module mpmath), 796
encipher kid rsa()
(in
module
eigenvals() (sympy.matrices.matrices.MatrixBase
sympy.crypto.crypto), 314
method), 655
encipher rsa()
(in
module
eigenvects() (sympy.matrices.matrices.MatrixBase
sympy.crypto.crypto), 313
method), 655
shift()
(in
module
Eijk() (in module sympy.functions.special.tensorencipher
functions),
sympy.crypto.crypto), 304
446
(in
module
eject()
(sympy.polys.polyclasses.DMP encipher substitution()
sympy.crypto.crypto), 305
method), 1163
(in
module
eject() (sympy.polys.polytools.Poly method), encipher vigenere()
sympy.crypto.crypto), 306
1085
elgamal private key()
(in
module encloses() (sympy.geometry.entity.GeometryEntity
method), 493
sympy.crypto.crypto), 318
elgamal public key()
(in
module encloses point() (sympy.geometry.ellipse.Ellipse
method), 545
sympy.crypto.crypto), 318
encloses
point() (sympy.geometry.polygon.Polygon
ellipe() (in module mpmath), 915
method),
560
ellipf() (in module mpmath), 913
encloses
point()
(sympy.geometry.polygon.RegularPolygon
ellipfun() (in module mpmath), 927
method), 566
ellipk() (in module mpmath), 912
encode
morse()
(in
module
ellippi() (in module mpmath), 918
315
sympy.crypto.crypto),
elliprc() (in module mpmath), 922
end (sympy.sets.sets.Interval attribute), 1318
elliprd() (in module mpmath), 924
enhance print
(class
in
elliprf() (in module mpmath), 920
sympy.galgebra.printing),
485
elliprg() (in module mpmath), 924
Index
1943
SymPy Documentation, Release 0.7.6
enum all() (sympy.utilities.enumerative.MultisetPartitionTraverser
393
method), 1552
er (class in sympy.functions.special.error functions),
enum large() (sympy.utilities.enumerative.MultisetPartitionTraverser
390
method), 1552
er() (in module mpmath), 805
enum range() (sympy.utilities.enumerative.MultisetPartitionTraverser
ernv (class in sympy.functions.special.error functions),
method), 1553
392
enum small() (sympy.utilities.enumerative.MultisetPartitionTraverser
ernv() (in module mpmath), 806
Erlang() (in module sympy.stats), 1371
method), 1553
enumerate states()
(in
module estimate error()
(mpsympy.physics.quantum.represent),
math.calculus.quadrature.QuadratureRule
1770
method), 1010
EPath (class in sympy.simplify.epathtools), ET() (sympy.polys.polytools.Poly method),
1347
1075
epath()
(in
module eta (sympy.functions.special.hyper.hyper atsympy.simplify.epathtools), 1349
tribute), 427
Eq (in module sympy.core.relational), 147
euler (class in sympy.functions.combinatorial.numbers),
equal() (sympy.geometry.line.Line method),
368
518
euler (mpmath.mp attribute), 751
Equality (class in sympy.core.relational), 148 euler equations()
(in
module
equals() (sympy.core.expr.Expr method), 112
sympy.calculus.euler), 1592
equals() (sympy.geometry.line.Ray method), euler maclaurin() (sympy.concrete.summations.Sum
521
method), 322
equals()
(sympy.geometry.line3d.Line3D EulerGamma (class in sympy.core.numbers),
method), 526
138
equals()
(sympy.geometry.line3d.Ray3D eulernum() (in module mpmath), 958
method), 535
eulerpoly() (in module mpmath), 959
equals() (sympy.matrices.expressions.MatrixExpr
eval() (sympy.assumptions.assume.Predicate
method), 708
method), 1289
equals() (sympy.matrices.immutable.ImmutableMatrix
eval() (sympy.functions.special.tensor functions.KroneckerD
method), 705
class method), 447
equation()
(sympy.geometry.ellipse.Circle eval() (sympy.physics.secondquant.Commutator
method), 555
class method), 1615
equation()
(sympy.geometry.ellipse.Ellipse eval()
(sympy.physics.secondquant.Dagger
method), 546
class method), 1604
equation()
(sympy.geometry.line.Line eval() (sympy.physics.secondquant.KroneckerDelta
method), 518
class method), 1605
equation()
(sympy.geometry.line3d.Line3D eval()
(sympy.polys.polyclasses.DMP
method), 526
method), 1163
equation()
(sympy.geometry.plane.Plane eval() (sympy.polys.polytools.Poly method),
method), 579
1086
Equivalent (class in sympy.logic.boolalg), 631 eval controls() (sympy.physics.quantum.gate.CGate
equivalent()
(in
module
method), 1790
sympy.solvers.diophantine), 1501
eval expr()
(in
module
erf (class in sympy.functions.special.error functions), sympy.parsing.sympy parser), 1589
388
eval levicivita()
(in
module
erf() (in module mpmath), 804
sympy.functions.special.tensor functions),
erf2 (class in sympy.functions.special.error functions), 446
391
evalf() (sympy.geometry.point.Point method),
erf2inv (class in sympy.functions.special.error functions),
499
394
evalf()
(sympy.geometry.point3d.Point3D
erfc (class in sympy.functions.special.error functions), method), 506
389
evalf() (sympy.matrices.matrices.MatrixBase
erfc() (in module mpmath), 805
method), 655
erfcinv (class in sympy.functions.special.error functions),
evalf() (sympy.polys.domains.domain.Domain
1944
Index
SymPy Documentation, Release 0.7.6
method), 1150
expr (sympy.core.function.Subs attribute),
evaluate deltas()
(in
module
171
sympy.physics.secondquant), 1619
expr (sympy.functions.elementary.piecewise.ExprCondPair
evaluate pauli product()
(in
module
attribute), 351
sympy.physics.paulialgebra), 1601
expr (sympy.physics.quantum.operator.DierentialOperator
EvaluationFailed
(class
in
attribute), 1763
sympy.polys.polyerrors), 1238
expr (sympy.physics.quantum.state.Wavefunction
even, 89
attribute), 1786
evolute()
(sympy.geometry.ellipse.Ellipse ExprCondPair
(class
in
method), 546
sympy.functions.elementary.piecewise),
ExactQuotientFailed
(class
in
351
sympy.polys.polyerrors), 1238
express()
(in
module
sympy.physics.vector.functions),
exclude()
(sympy.polys.polyclasses.DMP
method), 1163
1688
exclude()
(sympy.polys.polytools.Poly express() (sympy.physics.vector.dyadic.Dyadic
method), 1086
method), 1674
exp (class in sympy.functions.elementary.exponential),
express() (sympy.physics.vector.vector.Vector
350
method), 1671
exp() (in module mpmath), 756
ExpressionDomain
(class
in
exp() (sympy.matrices.matrices.MatrixBase
sympy.polys.domains), 1160
method), 655
ExpressionDomain.Expression
(class
in
Exp1 (class in sympy.core.numbers), 136
sympy.polys.domains), 1161
expand() (in module sympy.core.function), exquo() (in module sympy.polys.polytools),
171
1060
expand() (sympy.core.expr.Expr method), 113 exquo() (sympy.polys.domains.domain.Domain
method), 1150
expand() (sympy.matrices.matrices.MatrixBase
method), 655
exquo()
(sympy.polys.domains.eld.Field
expand complex()
(in
module
method), 1153
sympy.core.function), 179
exquo()
(sympy.polys.domains.ring.Ring
method), 1154
expand func()
(in
module
sympy.core.function), 178
exquo()
(sympy.polys.polyclasses.DMF
expand log() (in module sympy.core.function),
method), 1167
178
exquo()
(sympy.polys.polyclasses.DMP
expand mul()
(in
module
method), 1164
sympy.core.function), 178
exquo() (sympy.polys.polytools.Poly method),
expand multinomial()
(in
module
1086
sympy.core.function), 179
exquo ground() (sympy.polys.polyclasses.DMP
expand power base()
(in
module
method), 1164
sympy.core.function), 180
exquo ground() (sympy.polys.polytools.Poly
expand power exp()
(in
module
method), 1087
sympy.core.function), 180
extend()
(sympy.ntheory.generate.Sieve
expand trig()
(in
module
method), 276
sympy.core.function), 179
extend() (sympy.physics.unitsystems.dimensions.DimensionS
expint (class in sympy.functions.special.error functions),method), 1826
399
extend() (sympy.physics.unitsystems.units.UnitSystem
expint() (in module mpmath), 798
method), 1828
expj() (in module mpmath), 757
extend() (sympy.plotting.plot.Plot method),
expjpi() (in module mpmath), 758
1272
expm() (in module mpmath), 1028
extend to no() (sympy.ntheory.generate.Sieve
method), 276
expm1() (in module mpmath), 758
Exponential() (in module sympy.stats), 1372
extended euclid()
(in
module
Expr (class in sympy.core.expr), 103
sympy.solvers.diophantine), 1500
expr (sympy.core.function.Lambda attribute), exterior angle (sympy.geometry.polygon.RegularPolygon
163
attribute), 567
Index
1945
SymPy Documentation, Release 0.7.6
extract() (sympy.matrices.matrices.MatrixBase factorial() (sympy.polys.domains.FractionField
method), 656
method), 1159
extract() (sympy.matrices.sparse.SparseMatrix factorial() (sympy.polys.domains.PolynomialRing
method), 694
method), 1157
extract additively()
(sympy.core.expr.Expr factorial2
(class
in
method), 113
sympy.functions.combinatorial.factorials),
extract branch factor()
369
(in
module
(sympy.core.expr.Expr
method), factorial notation()
113
sympy.parsing.sympy parser), 1592
extract leading order() (sympy.core.add.Add factoring visitor()
(in
module
method), 146
sympy.utilities.enumerative), 1549
extract multiplicatively()
factorint() (in module sympy.ntheory.factor ),
(sympy.core.expr.Expr
method),
287
114
factors()
(sympy.core.numbers.Rational
extract type tens() (sympy.physics.hep.gamma matrices.GammaMatrixHead
method), 131
fadd() (in module mpmath), 728
static method), 1633
extradps() (in module mpmath), 744
FallingFactorial
(class
in
ExtraneousFactors
(class
in
sympy.functions.combinatorial.factorials),
sympy.polys.polyerrors), 1238
370
extraprec() (in module mpmath), 744
FBra (in module sympy.physics.secondquant),
eye() (in module sympy.matrices.dense), 679
1612
eye() (sympy.matrices.sparse.SparseMatrix fcode() (in module sympy.printing.fcode),
class method), 695
1254
FCodeGen (class in sympy.utilities.codegen),
F
1542
F (in module sympy.physics.secondquant), FCodePrinter (class in sympy.printing.fcode),
1257
1613
(in
module
F2PyCodeWrapper
(class
in fct sym array()
sympy.galgebra.stringarrays), 488
sympy.utilities.autowrap), 1534
Fd (in module sympy.physics.secondquant),
fabs() (in module mpmath), 733
1613
fac2() (in module mpmath), 778
fdi()
(sympy.core.function.Function
faces (sympy.combinatorics.polyhedron.Polyhedron
method), 170
attribute), 244
factor (sympy.physics.unitsystems.units.Unit fdi() (sympy.functions.elementary.complexes.Abs
method), 342
attribute), 1828
factor() (in module sympy.polys.polytools), fdi() (sympy.functions.elementary.exponential.exp
method), 351
1070
fdi() (sympy.functions.elementary.exponential.LambertW
factor() (sympy.core.expr.Expr method), 114
method), 353
factor list()
(in
module
fdi() (sympy.functions.elementary.exponential.log
sympy.polys.polytools), 1070
method), 354
factor list()
(sympy.polys.polyclasses.DMP
fdi() (sympy.functions.elementary.hyperbolic.sinh
method), 1164
method), 360
factor list()
(sympy.polys.polytools.Poly
FDistribution() (in module sympy.stats), 1373
method), 1087
fdiv() (in module mpmath), 731
factor list include() (sympy.polys.polyclasses.DMP
fdot() (in module mpmath), 733
method), 1164
factor list include() (sympy.polys.polytools.Poly () (in module mpmath), 785
fglm() (sympy.polys.polytools.GroebnerBasis
method), 1087
method), 1112
factor terms()
(in
module
bonacci
(class
in
sympy.core.exprtools), 186
sympy.functions.combinatorial.numbers),
factorial
(class
in
370
sympy.functions.combinatorial.factorials),
bonacci()
(in module mpmath), 953
369
Field
(class
in sympy.polys.domains.eld),
factorial() (in module mpmath), 777
1946
Index
SymPy Documentation, Release 0.7.6
1152
Float (class in sympy.core.numbers), 127
eld isomorphism()
(in
module oor (class in sympy.functions.elementary.integers),
sympy.polys.numberelds), 1117
351
ll() (sympy.matrices.dense.MutableDenseMatrix
oor() (in module mpmath), 735
method), 688
fmod() (in module mpmath), 732
ll() (sympy.matrices.sparse.MutableSparseMatrix
fmul() (in module mpmath), 730
method), 701
fneg() (in module mpmath), 729
lter symbols()
(in
module foci
(sympy.geometry.ellipse.Ellipse
atsympy.utilities.iterables), 1556
tribute), 547
nd() (sympy.core.basic.Basic method), 94
FockSpace
(class
in
nd DN()
(in
module
sympy.physics.quantum.hilbert),
sympy.solvers.diophantine), 1495
1759
nd dynamicsymbols()
(in
module FockState
(class
in
sympy.physics.mechanics), 1748
sympy.physics.secondquant), 1612
nd executable()
(in
module FockStateBosonBra
(class
in
sympy.galgebra.printing), 486
sympy.physics.secondquant), 1612
nd executable()
(in
module FockStateBosonKet
(class
in
sympy.utilities.misc), 1576
sympy.physics.secondquant), 1612
nd unit() (in module sympy.physics.units), FockStateBra
(class
in
1631
sympy.physics.secondquant), 1612
ndpoly() (in module mpmath), 1038
FockStateKet
(class
in
ndroot() (in module mpmath), 973
sympy.physics.secondquant), 1612
nite, 89
focus distance (sympy.geometry.ellipse.Ellipse
nite di weights()
(in
module
attribute), 547
sympy.calculus.nite di), 1596
forcing (sympy.physics.mechanics.kane.KanesMethod
attribute), 1742
FiniteDomain (class in sympy.stats.frv), 1396
FiniteField (class in sympy.polys.domains), forcing (sympy.physics.mechanics.lagrange.LagrangesMeth
1155
attribute), 1745
FinitePSpace (class in sympy.stats.frv), 1396 forcing full (sympy.physics.mechanics.kane.KanesMethod
attribute), 1742
FiniteRV() (in module sympy.stats), 1363
FiniteSet (class in sympy.sets.sets), 1319
forcing full (sympy.physics.mechanics.lagrange.LagrangesM
FisherZ() (in module sympy.stats), 1374
attribute), 1745
FixedBosonicBasis
(class
in form lagranges equations()
sympy.physics.secondquant), 1614
(sympy.physics.mechanics.lagrange.LagrangesMeth
method), 1745
FKet (in module sympy.physics.secondquant),
1613
fourier() (in module mpmath), 1017
FlagError (class in sympy.polys.polyerrors), fourier transform()
(in
module
1239
sympy.integrals.transforms), 586
FlatMirror
(class
in fourierval() (in module mpmath), 1018
sympy.physics.optics.gaussopt),
fprod() (in module mpmath), 732
1804
frac (in module sympy.printing.pretty.pretty symbology),
FlatRefraction
(class
in
1267
sympy.physics.optics.gaussopt),
frac() (in module mpmath), 737
1804
frac eld() (sympy.polys.domains.domain.Domain
atten() (in module sympy.galgebra.vector),
method), 1150
484
frac unify()
(sympy.polys.polyclasses.DMF
atten() (in module sympy.utilities.iterables),
method), 1168
1556
fraction() (in module mpmath), 741
atten() (sympy.categories.CompositeMorphismfraction() (in module sympy.simplify.simplify),
method), 1833
1335
atten() (sympy.core.add.Add class method), FractionField (class in sympy.polys.domains),
146
1159
atten() (sympy.core.mul.Mul class method), Frechet() (in module sympy.stats), 1375
142
free module() (sympy.polys.domains.ring.Ring
Index
1947
SymPy Documentation, Release 0.7.6
method), 1154
method), 1150
free symbols
(sympy.core.basic.Basic
at- from dict()
(sympy.polys.polyclasses.DMP
tribute), 94
class method), 1164
free symbols (sympy.functions.elementary.piecewise.ExprCondPair
from dict() (sympy.polys.polytools.Poly class
attribute), 351
method), 1088
free symbols (sympy.geometry.curve.Curve from expr() (sympy.polys.polytools.Poly class
attribute), 540
method), 1088
free symbols (sympy.integrals.Integral at- from ExpressionDomain()
tribute), 616
(sympy.polys.domains.domain.Domain
free symbols (sympy.matrices.matrices.MatrixBase
method), 1150
attribute), 656
from ExpressionDomain()
free symbols (sympy.physics.quantum.operator.DierentialOperator
(sympy.polys.domains.ExpressionDomain
attribute), 1764
method), 1161
free symbols (sympy.polys.polytools.Poly at- from FF gmpy() (sympy.polys.domains.domain.Domain
tribute), 1088
method), 1150
free symbols (sympy.polys.polytools.PurePoly from FF gmpy() (sympy.polys.domains.FiniteField
attribute), 1111
method), 1155
free symbols in domain
from FF python() (sympy.polys.domains.domain.Domain
(sympy.polys.polytools.Poly
atmethod), 1150
tribute), 1088
from FF python() (sympy.polys.domains.FiniteField
FreeModule
(class
in
method), 1156
sympy.polys.agca.modules), 1132
from FractionField()
FreeSpace
(class
in
(sympy.polys.domains.domain.Domain
sympy.physics.optics.gaussopt),
method), 1150
1803
from FractionField()
(sympy.polys.domains.ExpressionDomain
frequency (sympy.physics.optics.waves.TWave
method), 1161
attribute), 1818
fresnelc (class in sympy.functions.special.error functions),
from FractionField()
396
(sympy.polys.domains.FractionField
method), 1159
fresnelc() (in module mpmath), 809
FresnelIntegral
(class
in from FractionField()
sympy.functions.special.error functions),
(sympy.polys.domains.PolynomialRing
395
method), 1157
fresnels (class in sympy.functions.special.error functions),
from GlobalPolynomialRing()
(sympy.polys.domains.domain.Domain
395
method), 1150
fresnels() (in module mpmath), 808
frexp() (in module mpmath), 740
from inversion vector()
from AlgebraicField()
(sympy.combinatorics.permutations.Permutation
(sympy.polys.domains.domain.Domain
class method), 200
method), 1150
from list()
(sympy.polys.polyclasses.DMP
class method), 1164
from AlgebraicField()
(sympy.polys.domains.FractionField
from list() (sympy.polys.polytools.Poly class
method), 1159
method), 1088
from AlgebraicField()
from poly() (sympy.polys.polytools.Poly class
(sympy.polys.domains.IntegerRing
method), 1088
from PolynomialRing()
method), 1156
from AlgebraicField()
(sympy.polys.domains.domain.Domain
(sympy.polys.domains.PolynomialRing
method), 1150
from PolynomialRing()
method), 1157
(sympy.polys.domains.ExpressionDomain
from AlgebraicField()
method), 1161
(sympy.polys.domains.RationalField
from PolynomialRing()
method), 1158
from ComplexField()
(sympy.polys.domains.FractionField
(sympy.polys.domains.domain.Domain
method), 1159
1948
Index
SymPy Documentation, Release 0.7.6
from PolynomialRing()
method), 1157
(sympy.polys.domains.PolynomialRing from sympy() (sympy.polys.domains.RealField
method), 1157
method), 1160
from QQ gmpy() (sympy.polys.domains.AlgebraicField
from sympy list() (sympy.polys.polyclasses.DMP
method), 1158
class method), 1164
from QQ gmpy() (sympy.polys.domains.domain.Domain
from TIDS list() (sympy.tensor.tensor.TensAdd
method), 1150
static method), 1528
from QQ gmpy() (sympy.polys.domains.ExpressionDomain
from ZZ gmpy() (sympy.polys.domains.AlgebraicField
method), 1161
method), 1158
from QQ gmpy() (sympy.polys.domains.FiniteField
from ZZ gmpy() (sympy.polys.domains.domain.Domain
method), 1156
method), 1150
from QQ gmpy() (sympy.polys.domains.FractionField
from ZZ gmpy() (sympy.polys.domains.ExpressionDomain
method), 1159
method), 1161
from QQ gmpy() (sympy.polys.domains.PolynomialRing
from ZZ gmpy() (sympy.polys.domains.FiniteField
method), 1157
method), 1156
from QQ python() (sympy.polys.domains.AlgebraicField
from ZZ gmpy() (sympy.polys.domains.FractionField
method), 1158
method), 1159
from QQ python() (sympy.polys.domains.domain.Domain
from ZZ gmpy() (sympy.polys.domains.PolynomialRing
method), 1150
method), 1157
from QQ python() (sympy.polys.domains.ExpressionDomain
from ZZ python() (sympy.polys.domains.AlgebraicField
method), 1161
method), 1158
from QQ python() (sympy.polys.domains.FiniteField
from ZZ python() (sympy.polys.domains.domain.Domain
method), 1156
method), 1150
from QQ python() (sympy.polys.domains.FractionField
from ZZ python() (sympy.polys.domains.ExpressionDomain
method), 1159
method), 1161
from QQ python() (sympy.polys.domains.PolynomialRing
from ZZ python() (sympy.polys.domains.FiniteField
method), 1157
method), 1156
from RealField() (sympy.polys.domains.AlgebraicField
from ZZ python() (sympy.polys.domains.FractionField
method), 1158
method), 1159
from RealField() (sympy.polys.domains.domain.Domain
from ZZ python() (sympy.polys.domains.PolynomialRing
method), 1150
method), 1157
from RealField() (sympy.polys.domains.ExpressionDomain
fromiter()
(sympy.core.basic.Basic
class
method), 1161
method), 94
from RealField() (sympy.polys.domains.FiniteField
fsub() (in module mpmath), 729
method), 1156
fsum() (in module mpmath), 732
from RealField() (sympy.polys.domains.FractionField
full cyclic form (sympy.combinatorics.permutations.Permuta
method), 1159
attribute), 200
from RealField() (sympy.polys.domains.PolynomialRing
fun eval()
(sympy.tensor.tensor.TensAdd
method), 1157
method), 1528
from rgs() (sympy.combinatorics.partitions.Partition
fun eval()
(sympy.tensor.tensor.TensMul
class method), 188
method), 1530
from sequence() (sympy.combinatorics.permutations.Permutation
func (sympy.core.basic.Basic attribute), 94
class method), 200
func eld modgcd
(class
in
from sympy() (sympy.polys.domains.AlgebraicField
sympy.polys.modulargcd), 1242
method), 1158
Function (class in sympy.core.function), 169
from sympy() (sympy.polys.domains.domain.Domain
function (sympy.physics.quantum.operator.DierentialOpera
method), 1151
attribute), 1764
from sympy() (sympy.polys.domains.ExpressionDomain
function exponentiation()
(in
module
method), 1161
sympy.parsing.sympy parser), 1591
from sympy() (sympy.polys.domains.FiniteField FunctionClass (class in sympy.core.function),
168
method), 1156
from sympy() (sympy.polys.domains.FractionField
FunctionMatrix
(class
in
method), 1159
sympy.matrices.expressions), 710
from sympy() (sympy.polys.domains.PolynomialRing
functions (sympy.geometry.curve.Curve atIndex
1949
SymPy Documentation, Release 0.7.6
tribute), 540
1810
gaussian reduce()
(in
module
G
sympy.solvers.diophantine), 1503
GaussLegendre
(class
in
mpG() (in module sympy.printing.pretty.pretty symbology),
math.calculus.quadrature), 1012
1267
gcd() (in module sympy.polys.polytools), 1066
g() (in module sympy.printing.pretty.pretty symbology),
gcd() (sympy.core.numbers.Number method),
1267
127
GA LatexPrinter
(class
in
gcd() (sympy.polys.domains.domain.Domain
sympy.galgebra.printing), 485
method), 1151
GA Printer
(class
in
gcd()
(sympy.polys.domains.eld.Field
sympy.galgebra.printing), 485
method), 1153
GAeval()
(in
module
gcd() (sympy.polys.domains.PolynomialRing
sympy.galgebra.precedence), 484
method), 1157
gamma (class in sympy.functions.special.gamma functions),
gcd()
(sympy.polys.domains.RealField
379
method), 1160
gamma() (in module mpmath), 780
gcd() (sympy.polys.polyclasses.DMP method),
Gamma() (in module sympy.stats), 1375
1164
gamma trace() (sympy.physics.hep.gamma matrices.GammaMatrixHead
gcd() (sympy.polys.polytools.Poly method),
method), 1633
1088
gammainc() (in module mpmath), 795
GammaInverse() (in module sympy.stats), gcd list() (in module sympy.polys.polytools),
1066
1376
(in
module
GammaMatrixHead
(class
in gcd terms()
sympy.core.exprtools), 185
sympy.physics.hep.gamma matrices),
gcdex() (in module sympy.polys.polytools),
1632
1061
gammaprod() (in module mpmath), 782
Gate (class in sympy.physics.quantum.gate), gcdex() (sympy.polys.domains.domain.Domain
method), 1151
1789
gate (sympy.physics.quantum.gate.CGate at- gcdex() (sympy.polys.domains.PolynomialRing
method), 1157
tribute), 1790
(sympy.polys.polyclasses.DMP
gate (sympy.physics.quantum.gate.CNotGate gcdex()
method), 1164
attribute), 1792
gate simp()
(in
module gcdex() (sympy.polys.polytools.Poly method),
1089
sympy.physics.quantum.gate), 1793
gate sort()
(in
module Ge (in module sympy.core.relational), 148
gegenbauer
(class
in
sympy.physics.quantum.gate), 1793
sympy.functions.special.polynomials),
gaunt() (in module sympy.physics.wigner),
435
1624
gauss chebyshev t()
(in
module gegenbauer() (in module mpmath), 877
gegenbauer poly()
(in
module
sympy.integrals.quadrature), 622
sympy.polys.orthopolys), 1120
gauss chebyshev u()
(in
module
gen (sympy.polys.polytools.Poly attribute),
sympy.integrals.quadrature), 623
1089
gauss gen laguerre()
(in
module
generate() (sympy.combinatorics.perm groups.PermutationG
sympy.integrals.quadrature), 621
method), 226
gauss hermite()
(in
module
generate
bell()
(in
module
sympy.integrals.quadrature), 620
sympy.utilities.iterables),
1557
gauss jacobi()
(in
module
generate derangements()
(in
module
sympy.integrals.quadrature), 624
sympy.utilities.iterables),
1558
gauss laguerre()
(in
module
generate dimino() (sympy.combinatorics.perm groups.Permu
sympy.integrals.quadrature), 620
method), 226
gauss legendre()
(in
module
generate
gray() (sympy.combinatorics.graycode.GrayCode
sympy.integrals.quadrature), 619
257
method),
gaussian conj()
(in
module
generate
involutions()
(in
module
sympy.physics.optics.gaussopt),
1950
Index
SymPy Documentation, Release 0.7.6
sympy.utilities.iterables), 1558
method), 1826
generate oriented forest()
(in
module get domain()
(sympy.polys.polytools.Poly
sympy.utilities.iterables), 1559
method), 1089
generate schreier sims()
get exact() (sympy.polys.domains.domain.Domain
(sympy.combinatorics.perm groups.PermutationGroup
method), 1151
method), 227
get exact() (sympy.polys.domains.RealField
generate tokens()
(in
module
method), 1160
get eld() (sympy.polys.domains.domain.Domain
sympy.parsing.sympy tokenize),
1589
method), 1151
generators (sympy.combinatorics.perm groups.PermutationGroup
get eld() (sympy.polys.domains.ExpressionDomain
attribute), 227
method), 1161
GeneratorsError
(class
in get eld()
(sympy.polys.domains.eld.Field
sympy.polys.polyerrors), 1238
method), 1153
GeneratorsNeeded
(class
in get eld() (sympy.polys.domains.FiniteField
sympy.polys.polyerrors), 1238
method), 1156
Geometric() (in module sympy.stats), 1364
get eld() (sympy.polys.domains.IntegerRing
geometric conj ab()
(in
module
method), 1156
sympy.physics.optics.gaussopt),
get eld() (sympy.polys.domains.PolynomialRing
1809
method), 1157
geometric conj af()
(in
module get indices()
(in
module
sympy.physics.optics.gaussopt),
sympy.tensor.index methods), 1516
1810
get indices()
(sympy.tensor.tensor.TensMul
geometric conj bf()
(in
module
method), 1530
sympy.physics.optics.gaussopt),
get interface() (sympy.utilities.codegen.FCodeGen
method), 1543
1810
GeometricRay
(class
in get mass() (sympy.physics.mechanics.particle.Particle
sympy.physics.optics.gaussopt),
method), 1731
1805
get matrix() (sympy.tensor.tensor.TensExpr
GeometryEntity
(class
in
method), 1526
sympy.geometry.entity), 493
get mod func()
(in
module
get() (sympy.core.containers.Dict method),
sympy.utilities.source), 1587
183
get modulus()
(sympy.polys.polytools.Poly
get() (sympy.physics.unitsystems.dimensions.Dimensionmethod), 1089
method), 1825
get motion params()
(in
module
sympy.physics.vector.functions),
get adjacency distance()
(sympy.combinatorics.permutations.Permutation
1681
method), 200
get nodes() (mpmath.calculus.quadrature.QuadratureRule
get adjacency matrix()
method), 1010
(sympy.combinatorics.permutations.Permutation
get period() (sympy.functions.special.hyper.meijerg
method), 201
method), 430
get basis()
(in
module get permuted() (sympy.physics.secondquant.PermutationOp
sympy.physics.quantum.represent),
method), 1623
1769
get point() (sympy.physics.mechanics.particle.Particle
get class() (in module sympy.utilities.source),
method), 1731
1587
get positional distance()
get comm() (sympy.tensor.tensor. TensorManager
(sympy.combinatorics.permutations.Permutation
method), 1517
method), 201
get contraction structure()
(in
module get precedence distance()
sympy.tensor.index methods), 1514
(sympy.combinatorics.permutations.Permutation
method), 202
get default datatype()
(in
module
get precedence matrix()
sympy.utilities.codegen), 1540
get diag blocks() (sympy.matrices.matrices.MatrixBase (sympy.combinatorics.permutations.Permutation
method), 656
method), 202
get dim() (sympy.physics.unitsystems.dimensions.DimensionSystem
Get Program()
(in
module
Index
1951
SymPy Documentation, Release 0.7.6
sympy.galgebra.printing), 486
sympy.polys.galoistools), 1208
get prototype() (sympy.utilities.codegen.CCodeGen
gf crt() (in module sympy.polys.galoistools),
method), 1542
1197
get resource()
(in
module gf crt1() (in module sympy.polys.galoistools),
sympy.utilities.pkgdata), 1578
1198
get ring() (sympy.polys.domains.AlgebraicField gf crt2() (in module sympy.polys.galoistools),
method), 1158
1198
get ring() (sympy.polys.domains.domain.Domaingf csolve()
(in
module
method), 1151
sympy.polys.galoistools), 1213
get ring() (sympy.polys.domains.ExpressionDomain
gf degree()
(in
module
method), 1161
sympy.polys.galoistools), 1198
get ring()
(sympy.polys.domains.eld.Field gf di() (in module sympy.polys.galoistools),
method), 1153
1207
get ring() (sympy.polys.domains.FractionField gf div() (in module sympy.polys.galoistools),
1203
method), 1160
get ring()
(sympy.polys.domains.RealField gf eval() (in module sympy.polys.galoistools),
method), 1160
1207
get ring()
(sympy.polys.domains.ring.Ring gf expand()
(in
module
method), 1154
sympy.polys.galoistools), 1203
get segments() (sympy.plotting.plot.LineOver1DRangeSeries
gf exquo()
(in
module
method), 1282
sympy.polys.galoistools), 1204
get segments() (sympy.plotting.plot.Parametric2DLineSeries
gf factor()
(in
module
method), 1283
sympy.polys.galoistools), 1212
get subNO() (sympy.physics.secondquant.NO gf factor sqf()
(in
module
method), 1618
sympy.polys.galoistools), 1212
get subset from bitstring()
gf from dict()
(in
module
(sympy.combinatorics.graycode
sympy.polys.galoistools), 1200
static method), 260
gf from int poly()
(in
module
get symmetric group sgs()
(in
module
sympy.polys.galoistools), 1200
sympy.combinatorics.tensor can),
gf gcd() (in module sympy.polys.galoistools),
275
1206
get sympy dir()
(in
module gf gcdex()
(in
module
sympy.utilities.runtests), 1582
sympy.polys.galoistools), 1206
get target matrix() (sympy.physics.quantum.gate.Gate
gf int() (in module sympy.polys.galoistools),
method), 1789
1198
get target matrix() (sympy.physics.quantum.gate.UGate
gf irreducible()
(in
module
method), 1790
sympy.polys.galoistools), 1209
get unit() (sympy.physics.unitsystems.units.UnitSystem
gf irreducible p()
(in
module
method), 1828
sympy.polys.galoistools), 1209
getn() (sympy.core.expr.Expr method), 114
gf LC() (in module sympy.polys.galoistools),
getO() (sympy.core.expr.Expr method), 114
1199
gf add() (in module sympy.polys.galoistools), gf lcm() (in module sympy.polys.galoistools),
1202
1206
gf add ground()
(in
module gf lshift() (in module sympy.polys.galoistools),
sympy.polys.galoistools), 1201
1205
gf add mul()
(in
module gf monic()
(in
module
sympy.polys.galoistools), 1203
sympy.polys.galoistools), 1207
gf berlekamp()
(in
module gf mul() (in module sympy.polys.galoistools),
sympy.polys.galoistools), 1211
1202
gf cofactors()
(in
module gf mul ground()
(in
module
sympy.polys.galoistools), 1206
sympy.polys.galoistools), 1201
gf compose()
(in
module gf multi eval()
(in
module
sympy.polys.galoistools), 1208
sympy.polys.galoistools), 1208
gf compose mod()
(in
module gf neg() (in module sympy.polys.galoistools),
1952
Index
SymPy Documentation, Release 0.7.6
1201
gf normal()
(in
module
sympy.polys.galoistools), 1199
gf pow() (in module sympy.polys.galoistools),
1205
gf pow mod()
(in
module
sympy.polys.galoistools), 1205
gf Qbasis()
(in
module
sympy.polys.galoistools), 1211
gf Qmatrix()
(in
module
sympy.polys.galoistools), 1211
gf quo() (in module sympy.polys.galoistools),
1204
gf quo ground()
(in
module
sympy.polys.galoistools), 1202
gf random()
(in
module
sympy.polys.galoistools), 1209
gf rem() (in module sympy.polys.galoistools),
1204
gf rshift()
(in
module
sympy.polys.galoistools), 1205
gf shoup()
(in
module
sympy.polys.galoistools), 1212
gf sqf list()
(in
module
sympy.polys.galoistools), 1210
gf sqf p() (in module sympy.polys.galoistools),
1209
gf sqf part()
(in
module
sympy.polys.galoistools), 1210
gf sqr() (in module sympy.polys.galoistools),
1202
gf strip() (in module sympy.polys.galoistools),
1199
gf sub() (in module sympy.polys.galoistools),
1202
gf sub ground()
(in
module
sympy.polys.galoistools), 1201
gf sub mul()
(in
module
sympy.polys.galoistools), 1203
gf TC() (in module sympy.polys.galoistools),
1199
gf to dict()
(in
module
sympy.polys.galoistools), 1200
gf to int poly()
(in
module
sympy.polys.galoistools), 1200
gf trace map()
(in
module
sympy.polys.galoistools), 1208
gf trunc()
(in
module
sympy.polys.galoistools), 1199
gf value()
(in
module
sympy.polys.galoistools), 1213
gf zassenhaus()
(in
module
sympy.polys.galoistools), 1211
g() (in module sympy.polys.polytools), 1069
Index
g list() (in module sympy.polys.polytools),
1069
g list()
(sympy.polys.polyclasses.DMP
method), 1164
g list() (sympy.polys.polytools.Poly method),
1089
given() (in module sympy.stats), 1393
glaisher (mpmath.mp attribute), 751
GMPYFiniteField
(class
in
sympy.polys.domains), 1162
GMPYIntegerRing
(class
in
sympy.polys.domains), 1162
GMPYRationalField
(class
in
sympy.polys.domains), 1162
GoldenRatio (class in sympy.core.numbers),
139
gosper normal()
(in
module
sympy.concrete.gosper), 329
gosper sum()
(in
module
sympy.concrete.gosper), 330
gosper term()
(in
module
sympy.concrete.gosper), 330
gouy (sympy.physics.optics.gaussopt.BeamParameter
attribute), 1807
Grad() (in module sympy.galgebra.ga), 470
GradedLexOrder
(class
in
sympy.polys.orderings), 1119
gradient()
(in
module
sympy.physics.vector.eldfunctions),
1690
grampoint() (in module mpmath), 941
GramSchmidt()
(in
module
sympy.matrices.dense), 682
gray to bin() (sympy.combinatorics.graycode
static method), 259
GrayCode
(class
in
sympy.combinatorics.graycode),
256
graycode subsets() (sympy.combinatorics.graycode
static method), 260
GreaterThan (class in sympy.core.relational),
149
greek letters
(in
module
sympy.printing.pretty.pretty symbology),
1267
groebner()
(in
module
sympy.polys.groebnertools), 1236
groebner() (in module sympy.polys.polytools),
1073
GroebnerBasis
(class
in
sympy.polys.polytools), 1112
ground roots()
(in
module
sympy.polys.polytools), 1072
ground roots()
(sympy.polys.polytools.Poly
1953
SymPy Documentation, Release 0.7.6
method), 1090
has dups()
(in
module
(in
module
sympy.utilities.iterables), 1559
sympy.parsing.sympy tokenize),
has integer powers (sympy.physics.unitsystems.dimensions.
1590
attribute), 1825
group() (in module sympy.utilities.iterables), has only gens() (sympy.polys.polytools.Poly
1559
method), 1090
grover iteration()
(in
module has q annihilators (sympy.physics.secondquant.NO
sympy.physics.quantum.grover),
attribute), 1618
1795
has q creators (sympy.physics.secondquant.NO
gruntz() (in module sympy.series.gruntz),
attribute), 1618
1306
has variety()
(in
module
Gt (in module sympy.core.relational), 148
sympy.utilities.iterables), 1560
guess degree()
(mp- Heaviside
(class
in
math.calculus.quadrature.QuadratureRule
sympy.functions.special.delta functions),
method), 1010
378
height (sympy.categories.diagram drawing.DiagramGrid
H
attribute), 1840
H (in module sympy.physics.quantum.gate), height (sympy.physics.optics.gaussopt.GeometricRay
attribute), 1806
1792
(sympy.polys.agca.ideals.Ideal
H (sympy.matrices.matrices.MatrixBase at- height()
method), 1139
tribute), 645
HadamardGate
(class
in height() (sympy.printing.pretty.stringpict.stringPict
method), 1268
sympy.physics.quantum.gate), 1791
hermite (class in sympy.functions.special.polynomials),
Half (class in sympy.core.numbers), 134
440
half gcdex()
(in
module
hermite() (in module mpmath), 878
sympy.polys.polytools), 1061
hermite poly()
(in
module
half gcdex() (sympy.polys.domains.domain.Domain
sympy.polys.orthopolys), 1120
method), 1151
half gcdex()
(sympy.polys.polyclasses.DMP hermitian, 89
HermitianOperator
(class
in
method), 1164
sympy.physics.quantum.operator),
half gcdex()
(sympy.polys.polytools.Poly
1761
method), 1090
half per()
(sympy.polys.polyclasses.DMF hessian() (in module sympy.matrices.dense),
681
method), 1168
Halley
(class
in
mp- heurisch() (sympy.integrals static method),
612
math.calculus.optimization), 977
heurisch wrapper() (sympy.integrals static
hankel1 (class in sympy.functions.special.bessel),
method), 613
411
HeuristicGCDFailed
(class
in
hankel1() (in module mpmath), 825
sympy.polys.polyerrors), 1238
hankel2 (class in sympy.functions.special.bessel),
HilbertSpace
(class
in
412
sympy.physics.quantum.hilbert),
hankel2() (in module mpmath), 827
1758
hankel transform()
(in
module
hobj()
(in
module
sympy.integrals.transforms), 589
sympy.printing.pretty.pretty symbology),
harmonic
(class
in
1267
sympy.functions.combinatorial.numbers),
holzer()
(in
module
371
1504
sympy.solvers.diophantine),
harmonic() (in module mpmath), 793
hom() (sympy.categories.Diagram method),
has() (sympy.core.basic.Basic method), 95
1836
has() (sympy.matrices.matrices.MatrixBase
homogeneous
order()
(in
module
method), 657
sympy.solvers.ode),
1403
has() (sympy.matrices.sparse.SparseMatrix
homogeneous order()
method), 695
(sympy.polys.polyclasses.DMP
group()
1954
Index
SymPy Documentation, Release 0.7.6
method), 1164
Identity (class in sympy.matrices.expressions),
homogeneous order()
711
(sympy.polys.polytools.Poly method), identity hom() (sympy.polys.agca.modules.FreeModule
1090
method), 1133
homogenize() (sympy.polys.polyclasses.DMP identity hom() (sympy.polys.agca.modules.Module
method), 1164
method), 1132
homogenize()
(sympy.polys.polytools.Poly identity hom() (sympy.polys.agca.modules.QuotientModule
method), 1091
method), 1141
homomorphism()
(in
module identity hom() (sympy.polys.agca.modules.SubModule
sympy.polys.agca.homomorphisms),
method), 1134
1143
IdentityFunction
(class
in
HomomorphismFailed
(class
in
sympy.functions.elementary.miscellaneous),
sympy.polys.polyerrors), 1238
352
horner() (in module sympy.polys.polyfuncs), IdentityGate
(class
in
1114
sympy.physics.quantum.gate), 1790
hradius (sympy.geometry.ellipse.Ellipse at- IdentityMorphism
(class
in
tribute), 547
sympy.categories), 1833
hstack() (sympy.matrices.matrices.MatrixBase IdentityOperator
(class
in
class method), 657
sympy.physics.quantum.operator),
hyp0f1() (in module mpmath), 888
1761
hyp1f1() (in module mpmath), 889
Idx (class in sympy.tensor.indexed), 1509
hyp1f2() (in module mpmath), 890
igcd() (in module sympy.core.numbers), 132
hyp2f0() (in module mpmath), 891
ilcm() (in module sympy.core.numbers), 132
hyp2f1() (in module mpmath), 892
Illinois
(class
in
mpmath.calculus.optimization), 978
hyp2f2() (in module mpmath), 893
hyp2f3() (in module mpmath), 893
im() (in module mpmath), 734
hyp3f2() (in module mpmath), 894
image() (sympy.polys.agca.homomorphisms.ModuleHomomo
hyper (class in sympy.functions.special.hyper),
method), 1145
426
ImageSet (class in sympy.sets.fancysets),
hyper() (in module mpmath), 895
1326
hyper2d() (in module mpmath), 902
imageset() (in module sympy.sets.sets), 1316
HyperbolicFunction
(class
in imaginary, 89
sympy.functions.elementary.hyperbolic),ImaginaryUnit
(class
in
352
sympy.core.numbers), 137
hypercomb() (in module mpmath), 898
ImmutableMatrix
(class
in
hyperexpand()
(in
module
sympy.matrices.immutable), 705
sympy.simplify.hyperexpand), 1347
ImmutableSparseMatrix
(class
in
hyperfac() (in module mpmath), 788
sympy.matrices.immutable), 703
Hypergeometric() (in module sympy.stats), implemented function()
(in
module
1363
sympy.utilities.lambdify), 1573
hypersimilar()
(in
module implicit application()
(in
module
sympy.simplify.simplify), 1339
sympy.parsing.sympy parser), 1591
hypersimp()
(in
module implicit multiplication()
(in
module
sympy.simplify.simplify), 1339
sympy.parsing.sympy parser), 1591
hyperu() (in module mpmath), 862
implicit multiplication application() (in modhypot() (in module mpmath), 752
ule
sympy.parsing.sympy parser),
1591
I
ImplicitSeries
(class
in
sympy.plotting.plot
implicit),
1283
ibin() (in module sympy.utilities.iterables),
Implies (class in sympy.logic.boolalg), 630
1560
Ideal (class in sympy.polys.agca.ideals), 1138 imul num() (sympy.polys.rings.PolyElement
method), 1219
ideal()
(sympy.polys.domains.ring.Ring
in
terms
of generators()
method), 1154
(sympy.polys.agca.modules.SubModule
identify() (in module mpmath), 1034
Index
1955
SymPy Documentation, Release 0.7.6
method), 1135
method), 1164
incenter (sympy.geometry.polygon.Triangle inject() (sympy.polys.polytools.Poly method),
attribute), 573
1091
incircle (sympy.geometry.polygon.RegularPolygon
InnerProduct
(class
in
attribute), 567
sympy.physics.quantum.innerproduct),
incircle
(sympy.geometry.polygon.Triangle
1754
attribute), 573
InnerProduct
(class
in
inclusion hom() (sympy.polys.agca.modules.SubModule sympy.physics.secondquant), 1613
method), 1135
inradius (sympy.geometry.polygon.RegularPolygon
indent code() (sympy.printing.ccode.CCodePrinter
attribute), 567
method), 1252
inradius (sympy.geometry.polygon.Triangle
indent code() (sympy.printing.fcode.FCodePrinter
attribute), 573
method), 1257
intcurve diequ()
(in
module
index() (sympy.combinatorics.permutations.Permutationsympy.digeom), 1857
intcurve series() (in module sympy.digeom),
method), 202
index()
(sympy.core.containers.Tuple
1855
method), 182
integer, 89
index() (sympy.physics.secondquant.FixedBosonicBasis
Integer (class in sympy.core.numbers), 131
method), 1614
integer nthroot()
(in
module
index() (sympy.physics.secondquant.VarBosonicBasis sympy.core.power), 141
method), 1613
IntegerPartition
(class
in
index() (sympy.polys.rings.PolyRing method),
sympy.combinatorics.partitions),
1216
189
Indexed (class in sympy.tensor.indexed), 1511 IntegerRing (class in sympy.polys.domains),
IndexedBase (class in sympy.tensor.indexed),
1156
1513
Integers (class in sympy.sets.fancysets), 1325
indices (sympy.tensor.indexed.Indexed at- Integral (class in sympy.integrals), 615
tribute), 1511
Integral.is commutative
(in
module
indices contain equal information
sympy.integrals.transforms), 615
(sympy.functions.special.tensor functions.KroneckerDelta
integral steps() (sympy.integrals.manualintegrate
attribute), 447
static method), 614
indices contain equal information
integrand() (sympy.functions.special.hyper.meijerg
(sympy.physics.secondquant.KroneckerDelta method), 430
attribute), 1605
integrate() (sympy.core.expr.Expr method),
inertia()
(in
module
114
sympy.physics.mechanics.functions), integrate() (sympy.integrals static method),
1736
608
inertia of point mass()
(in
module integrate() (sympy.matrices.matrices.MatrixBase
sympy.physics.mechanics.functions),
method), 657
1737
integrate()
(sympy.polys.polyclasses.DMP
method), 1164
inf (sympy.sets.sets.Set attribute), 1313
innite, 89
integrate()
(sympy.polys.polytools.Poly
innitesimals()
(in
module
method), 1091
integrate result()
(in
module
sympy.solvers.ode), 1404
Innity (class in sympy.core.numbers), 135
sympy.physics.quantum.represent),
init printing()
(in
module
1769
sympy.physics.vector.printing), 1682 interactive traversal()
(in
module
inject() (sympy.polys.domains.compositedomain.CompositeDomain
sympy.utilities.iterables), 1561
method), 1155
interior angle (sympy.geometry.polygon.RegularPolygon
attribute), 568
inject() (sympy.polys.domains.domain.Domain
interpolate()
(in
module
method), 1151
inject() (sympy.polys.domains.simpledomain.SimpleDomain
sympy.polys.polyfuncs), 1114
interpolating poly()
(in
module
method), 1155
inject()
(sympy.polys.polyclasses.DMP
sympy.polys.specialpolys), 1120
1956
Index
SymPy Documentation, Release 0.7.6
intersect()
(sympy.polys.agca.ideals.Ideal inverse() (sympy.functions.elementary.hyperbolic.acoth
method), 1139
method), 344
intersect() (sympy.polys.agca.modules.SubModule
inverse() (sympy.functions.elementary.hyperbolic.asinh
method), 1135
method), 345
intersect()
(sympy.sets.sets.Set
method), inverse() (sympy.functions.elementary.hyperbolic.atanh
1313
method), 347
Intersection (class in sympy.sets.sets), 1321
inverse() (sympy.functions.elementary.hyperbolic.coth
intersection()
(in
module
method), 350
sympy.geometry.util), 495
inverse() (sympy.functions.elementary.hyperbolic.sinh
intersection() (sympy.geometry.ellipse.Circle
method), 360
method), 555
inverse() (sympy.functions.elementary.hyperbolic.tanh
intersection() (sympy.geometry.ellipse.Ellipse
method), 363
method), 548
inverse() (sympy.functions.elementary.trigonometric.acos
intersection() (sympy.geometry.entity.GeometryEntity method), 343
inverse() (sympy.functions.elementary.trigonometric.acot
method), 493
intersection() (sympy.geometry.line.LinearEntity
method), 344
method), 511
inverse() (sympy.functions.elementary.trigonometric.asin
intersection() (sympy.geometry.line3d.LinearEntity3D method), 345
method), 529
inverse() (sympy.functions.elementary.trigonometric.atan
intersection() (sympy.geometry.plane.Plane
method), 346
method), 579
inverse() (sympy.functions.elementary.trigonometric.cot
intersection()
(sympy.geometry.point.Point
method), 350
method), 499
inverse() (sympy.functions.elementary.trigonometric.tan
intersection() (sympy.geometry.point3d.Point3D
method), 363
method), 506
inverse ADJ() (sympy.matrices.matrices.MatrixBase
method), 658
intersection() (sympy.geometry.polygon.Polygon
method), 560
inverse cosine transform()
(in
module
intersection() (sympy.sets.sets.Set method),
sympy.integrals.transforms), 588
1314
inverse fourier transform()
(in
module
sympy.integrals.transforms), 586
Interval (class in sympy.sets.sets), 1317
intervals() (in module sympy.polys.polytools), inverse GE() (sympy.matrices.matrices.MatrixBase
1071
method), 658
intervals()
(sympy.polys.polyclasses.DMP inverse hankel transform()
(in
module
method), 1164
sympy.integrals.transforms), 590
intervals()
(sympy.polys.polytools.Poly inverse laplace transform()
(in
module
method), 1092
sympy.integrals.transforms), 585
IntQubit
(class
in inverse LU() (sympy.matrices.matrices.MatrixBase
sympy.physics.quantum.qubit), 1797
method), 658
IntQubitBra
(class
in inverse mellin transform()
(in
module
sympy.physics.quantum.qubit), 1798
sympy.integrals.transforms), 584
intrinsic impedance
inverse sine transform()
(in
module
(sympy.physics.optics.medium.Medium
sympy.integrals.transforms), 587
attribute), 1812
inversion vector() (sympy.combinatorics.permutations.Perm
inv can transf matrix
method), 203
(sympy.physics.unitsystems.dimensions.DimensionSystem
inversions() (sympy.combinatorics.permutations.Permutation
attribute), 1826
method), 203
inv mod() (sympy.matrices.matrices.MatrixBaseinvert() (in module sympy.polys.polytools),
1061
method), 658
Inverse (class in sympy.matrices.expressions), invert() (sympy.core.expr.Expr method), 114
709
invert() (sympy.polys.domains.domain.Domain
inverse() (sympy.functions.elementary.exponential.log method), 1151
method), 354
invert()
(sympy.polys.domains.ring.Ring
inverse() (sympy.functions.elementary.hyperbolic.acoshmethod), 1154
method), 343
invert()
(sympy.polys.polyclasses.DMF
Index
1957
SymPy Documentation, Release 0.7.6
method), 1168
is cyclotomic (sympy.polys.polytools.Poly at(sympy.polys.polyclasses.DMP
tribute), 1093
method), 1164
is diagonal() (sympy.matrices.matrices.MatrixBase
invert() (sympy.polys.polytools.Poly method),
method), 659
1092
is diagonalizable() (sympy.matrices.matrices.MatrixBase
IQFT (class in sympy.physics.quantum.qft),
method), 660
1796
is dimensionless (sympy.physics.unitsystems.dimensions.Dim
irrational, 89
attribute), 1825
is abelian (sympy.combinatorics.perm groups.PermutationGroup
is disjoint() (sympy.sets.sets.Set method),
attribute), 227
1314
is above fermi (sympy.functions.special.tensor functions.KroneckerDelta
is dnf() (in module sympy.logic.boolalg), 633
attribute), 448
is Empty (sympy.combinatorics.permutations.Permutation
attribute), 204
is above fermi (sympy.physics.secondquant.KroneckerDelta
attribute), 1606
is equilateral() (sympy.geometry.polygon.Triangle
is algebraic expr()
(sympy.core.expr.Expr
method), 574
is even (sympy.combinatorics.permutations.Permutation
method), 114
is aliased (sympy.polys.numberelds.AlgebraicNumber attribute), 205
attribute), 1117
is full module() (sympy.polys.agca.modules.SubModule
is alt sym() (sympy.combinatorics.perm groups.PermutationGroup
method), 1135
method), 228
is full module() (sympy.polys.agca.modules.SubQuotientMod
is anti symmetric() (sympy.matrices.matrices.MatrixBase
method), 1142
method), 658
is groebner()
(in
module
is below fermi (sympy.functions.special.tensor functions.KroneckerDelta
sympy.polys.groebnertools), 1236
attribute), 448
is ground (sympy.polys.polyclasses.ANP attribute), 1168
is below fermi (sympy.physics.secondquant.KroneckerDelta
attribute), 1606
is ground (sympy.polys.polyclasses.DMP atis cnf() (in module sympy.logic.boolalg), 632
tribute), 1164
is collinear()
(sympy.geometry.point.Point is ground
(sympy.polys.polytools.Poly
atmethod), 499
tribute), 1093
is commutative (sympy.core.function.Function is group() (sympy.combinatorics.perm groups.PermutationG
attribute), 170
method), 228
is commutative (sympy.physics.quantum.state.Wavefunction
is hermitian (sympy.matrices.matrices.MatrixBase
attribute), 1786
attribute), 661
is comparable (sympy.core.basic.Basic at- is hermitian (sympy.matrices.sparse.SparseMatrix
tribute), 95
attribute), 695
is compatible() (sympy.physics.unitsystems.units.Unit
is homogeneous (sympy.polys.polyclasses.DMP
method), 1828
attribute), 1164
is concyclic()
(sympy.geometry.point.Point is homogeneous (sympy.polys.polytools.Poly
method), 500
attribute), 1093
is conservative()
(in
module is Identity (sympy.combinatorics.permutations.Permutation
sympy.physics.vector.eldfunctions),
attribute), 204
1692
is identity (sympy.core.function.Lambda atis consistent (sympy.physics.unitsystems.dimensions.DimensionSystem
tribute), 163
is injective() (sympy.polys.agca.homomorphisms.ModuleHom
attribute), 1826
is consistent (sympy.physics.unitsystems.units.UnitSystem
method), 1145
is irreducible (sympy.polys.polyclasses.DMP
attribute), 1828
is constant() (sympy.core.expr.Expr method),
attribute), 1165
115
is irreducible (sympy.polys.polytools.Poly atis convex() (sympy.geometry.polygon.Polygon
tribute), 1093
method), 561
is isomorphism() (sympy.polys.agca.homomorphisms.Module
is coplanar()
(sympy.geometry.plane.Plane
method), 1146
method), 579
is isosceles() (sympy.geometry.polygon.Triangle
is cyclotomic (sympy.polys.polyclasses.DMP
method), 574
attribute), 1164
is left unbounded
(sympy.sets.sets.Interval
invert()
1958
Index
SymPy Documentation, Release 0.7.6
attribute), 1318
is nonpositive() (sympy.polys.domains.FractionField
is linear (sympy.polys.polyclasses.DMP atmethod), 1160
tribute), 1165
is nonpositive() (sympy.polys.domains.PolynomialRing
is linear
(sympy.polys.polytools.Poly
atmethod), 1157
tribute), 1094
is normal() (sympy.combinatorics.perm groups.PermutationG
is lower (sympy.matrices.matrices.MatrixBase
method), 230
attribute), 661
is normalized (sympy.physics.quantum.state.Wavefunction
is lower hessenberg
attribute), 1786
(sympy.matrices.matrices.MatrixBase is nthpow residue()
(in
module
attribute), 662
sympy.ntheory.residue ntheory),
is maximal() (sympy.polys.agca.ideals.Ideal
297
method), 1139
is number (sympy.core.expr.Expr attribute),
is minimal()
(in
module
116
sympy.polys.groebnertools), 1236
is odd (sympy.combinatorics.permutations.Permutation
is monic (sympy.polys.polyclasses.DMP atattribute), 205
is one
(sympy.polys.polyclasses.ANP
attribute), 1165
is monic
(sympy.polys.polytools.Poly
attribute), 1168
tribute), 1094
is one
(sympy.polys.polyclasses.DMF
atis monomial (sympy.polys.polyclasses.DMP
tribute), 1168
attribute), 1165
is one
(sympy.polys.polyclasses.DMP
atis monomial (sympy.polys.polytools.Poly attribute), 1165
tribute), 1094
is one (sympy.polys.polytools.Poly attribute),
is multivariate
(sympy.polys.polytools.Poly
1095
attribute), 1094
is one() (sympy.polys.domains.domain.Domain
method), 1151
is negative() (sympy.polys.domains.AlgebraicField
method), 1158
is only above fermi (sympy.functions.special.tensor function
is negative() (sympy.polys.domains.domain.Domain
attribute), 448
method), 1151
is only above fermi (sympy.physics.secondquant.KroneckerD
is negative() (sympy.polys.domains.ExpressionDomain attribute), 1607
method), 1161
is only below fermi (sympy.functions.special.tensor function
is negative() (sympy.polys.domains.FractionField
attribute), 449
method), 1160
is only below fermi (sympy.physics.secondquant.KroneckerD
is negative() (sympy.polys.domains.PolynomialRing
attribute), 1607
method), 1157
is only q annihilator
(sympy.physics.secondquant.AnnihilateFermion
is nilpotent (sympy.combinatorics.perm groups.PermutationGroup
attribute), 1609
attribute), 229
is nilpotent() (sympy.matrices.matrices.MatrixBase
is only q annihilator
method), 662
(sympy.physics.secondquant.CreateFermion
is nonnegative() (sympy.polys.domains.AlgebraicField attribute), 1611
method), 1159
is only q creator (sympy.physics.secondquant.AnnihilateFerm
is nonnegative() (sympy.polys.domains.domain.Domain attribute), 1610
method), 1151
is only q creator (sympy.physics.secondquant.CreateFermio
is nonnegative() (sympy.polys.domains.ExpressionDomain
attribute), 1611
is parallel() (sympy.geometry.line.LinearEntity
method), 1161
is nonnegative() (sympy.polys.domains.FractionField method), 512
is parallel() (sympy.geometry.line3d.LinearEntity3D
method), 1160
is nonnegative() (sympy.polys.domains.PolynomialRing method), 530
is parallel()
(sympy.geometry.plane.Plane
method), 1157
is nonpositive() (sympy.polys.domains.AlgebraicField method), 580
method), 1159
is perpendicular() (sympy.geometry.line.LinearEntity
is nonpositive() (sympy.polys.domains.domain.Domain method), 512
method), 1151
is perpendicular() (sympy.geometry.line3d.LinearEntity3D
is nonpositive() (sympy.polys.domains.ExpressionDomain
method), 530
method), 1161
is perpendicular() (sympy.geometry.plane.Plane
Index
1959
SymPy Documentation, Release 0.7.6
method), 580
method), 574
is polynomial()
(sympy.core.expr.Expr is right unbounded (sympy.sets.sets.Interval
method), 117
attribute), 1318
is positive() (sympy.polys.domains.AlgebraicField
is scalene() (sympy.geometry.polygon.Triangle
method), 1159
method), 575
is positive() (sympy.polys.domains.domain.Domain
is sequence()
(in
module
method), 1151
sympy.core.compatibility), 184
is positive() (sympy.polys.domains.ExpressionDomain
is similar() (sympy.geometry.entity.GeometryEntity
method), 1161
method), 493
is positive() (sympy.polys.domains.FractionFieldis similar() (sympy.geometry.line.LinearEntity
method), 1160
method), 513
is positive() (sympy.polys.domains.PolynomialRing
is similar() (sympy.geometry.line3d.LinearEntity3D
method), 1157
method), 531
is primary()
(sympy.polys.agca.ideals.Ideal is similar() (sympy.geometry.polygon.Triangle
method), 1139
method), 575
is prime()
(sympy.polys.agca.ideals.Ideal is simple() (sympy.functions.special.delta functions.DiracDe
method), 1139
method), 377
is primitive (sympy.polys.polyclasses.DMP at- is Singleton (sympy.combinatorics.permutations.Permutatio
tribute), 1165
attribute), 204
is primitive (sympy.polys.polytools.Poly at- is solenoidal()
(in
module
tribute), 1095
sympy.physics.vector.eldfunctions),
is primitive() (sympy.combinatorics.perm groups.PermutationGroup
1693
method), 230
is solvable (sympy.combinatorics.perm groups.PermutationG
is primitive root()
(in
module
attribute), 231
sympy.ntheory.residue ntheory),
is sqf
(sympy.polys.polyclasses.DMP
attribute), 1165
295
is principal() (sympy.polys.agca.ideals.Ideal is sqf (sympy.polys.polytools.Poly attribute),
method), 1139
1096
is proper subset()
(sympy.sets.sets.Set is square (sympy.matrices.matrices.MatrixBase
method), 1314
attribute), 663
is proper superset()
(sympy.sets.sets.Set is subdiagram() (sympy.categories.Diagram
method), 1314
method), 1836
is q annihilator (sympy.physics.secondquant.AnnihilateFermion
is subgroup() (sympy.combinatorics.perm groups.Permutatio
attribute), 1610
method), 231
is q annihilator (sympy.physics.secondquant.CreateFermion
is submodule() (sympy.polys.agca.modules.FreeModule
attribute), 1611
method), 1133
is q creator (sympy.physics.secondquant.AnnihilateFermion
is submodule() (sympy.polys.agca.modules.Module
attribute), 1610
method), 1132
is q creator (sympy.physics.secondquant.CreateFermion
is submodule() (sympy.polys.agca.modules.QuotientModule
attribute), 1611
method), 1141
is quad residue()
(in
module is submodule() (sympy.polys.agca.modules.SubModule
sympy.ntheory.residue ntheory),
method), 1136
297
is subset()
(sympy.sets.sets.Set
method),
is quadratic
(sympy.polys.polyclasses.DMP
1314
attribute), 1165
is superset() (sympy.sets.sets.Set method),
is quadratic (sympy.polys.polytools.Poly at1315
tribute), 1095
is surjective() (sympy.polys.agca.homomorphisms.ModuleHo
is radical()
(sympy.polys.agca.ideals.Ideal
method), 1146
method), 1139
is symbolic() (sympy.matrices.matrices.MatrixBase
method), 663
is rational function() (sympy.core.expr.Expr
is symmetric() (sympy.matrices.matrices.MatrixBase
method), 117
is reduced()
(in
module
method), 663
is symmetric() (sympy.matrices.sparse.SparseMatrix
sympy.polys.groebnertools), 1236
is right() (sympy.geometry.polygon.Triangle
method), 696
1960
Index
SymPy Documentation, Release 0.7.6
is tangent() (sympy.geometry.ellipse.Ellipse IsomorphismFailed
(class
in
method), 548
sympy.polys.polyerrors), 1238
is transitive() (sympy.combinatorics.perm groups.PermutationGroup
isprime()
(in
module
method), 232
sympy.ntheory.primetest), 294
is trivial (sympy.combinatorics.perm groups.PermutationGroup
issubset() (sympy.sets.sets.Set method), 1315
attribute), 232
issuperset() (sympy.sets.sets.Set method),
is univariate (sympy.polys.polytools.Poly at1315
ITE (class in sympy.logic.boolalg), 631
tribute), 1096
is upper (sympy.matrices.matrices.MatrixBase items() (sympy.core.containers.Dict method),
attribute), 664
183
is upper hessenberg
items() (sympy.physics.unitsystems.dimensions.Dimension
(sympy.matrices.matrices.MatrixBase
method), 1825
attribute), 665
iter basic args()
(sympy.core.basic.Basic
is whole ring() (sympy.polys.agca.ideals.Ideal
method), 95
iter q annihilators()
method), 1139
is zero (sympy.matrices.immutable.ImmutableMatrix (sympy.physics.secondquant.NO
attribute), 706
method), 1619
is zero (sympy.matrices.matrices.MatrixBase iter q creators() (sympy.physics.secondquant.NO
attribute), 665
method), 1619
is zero
(sympy.polys.polyclasses.ANP
at- iterable()
(in
module
tribute), 1168
sympy.core.compatibility), 183
is zero (sympy.polys.polyclasses.DMF at- iterate binary() (sympy.combinatorics.subsets.Subset
tribute), 1168
method), 251
is zero (sympy.polys.polyclasses.DMP at- iterate graycode() (sympy.combinatorics.subsets.Subset
tribute), 1165
method), 251
is zero (sympy.polys.polytools.Poly attribute), itercoes() (sympy.polys.rings.PolyElement
1096
method), 1219
is zero() (sympy.polys.agca.homomorphisms.ModuleHomomorphism
itermonomials()
(in
module
method), 1146
sympy.polys.monomials), 1117
is zero()
(sympy.polys.agca.ideals.Ideal itermonoms() (sympy.polys.rings.PolyElement
method), 1139
method), 1219
is zero() (sympy.polys.agca.modules.FreeModule
iterterms() (sympy.polys.rings.PolyElement
method), 1133
method), 1219
is zero() (sympy.polys.agca.modules.Module
J
method), 1132
is zero() (sympy.polys.agca.modules.QuotientModule
j0() (in module mpmath), 814
method), 1141
j1() (in module mpmath), 814
is zero() (sympy.polys.agca.modules.SubModulejacobi (class in sympy.functions.special.polynomials),
method), 1136
433
is zero() (sympy.polys.domains.domain.Domain jacobi() (in module mpmath), 876
method), 1151
jacobi normalized()
(in
module
is zero dimensional (sympy.polys.polytools.GroebnerBasis
sympy.functions.special.polynomials),
attribute), 1113
434
is zero dimensional()
(in
module jacobi poly()
(in
module
sympy.polys.polytools), 1074
sympy.polys.orthopolys), 1120
isdisjoint() (sympy.sets.sets.Set method), jacobi symbol()
(in
module
1315
sympy.ntheory.residue ntheory),
isnite() (in module mpmath), 739
297
isinf() (in module mpmath), 738
jacobian()
(sympy.digeom.CoordSystem
isint() (in module mpmath), 740
method), 1849
isnan() (in module mpmath), 739
jacobian() (sympy.matrices.matrices.MatrixBase
isnormal() (in module mpmath), 739
method), 666
isolate()
(in
module jn (class in sympy.functions.special.bessel),
sympy.polys.numberelds), 1117
412
Index
1961
SymPy Documentation, Release 0.7.6
jn zeros()
(in
module ket (sympy.physics.secondquant.InnerProduct
sympy.functions.special.bessel),
attribute), 1613
413
KetBase
(class
in
jordan cell()
(in
module
sympy.physics.quantum.state), 1781
sympy.matrices.dense), 681
key2bounds() (sympy.matrices.matrices.MatrixBase
jordan cells() (sympy.matrices.matrices.MatrixBase
method), 667
method), 666
key2ij() (sympy.matrices.matrices.MatrixBase
jordan form() (sympy.matrices.matrices.MatrixBase
method), 667
method), 667
keys() (sympy.core.containers.Dict method),
josephus() (sympy.combinatorics.permutations.Permutation
183
class method), 205
keys() (sympy.physics.unitsystems.dimensions.Dimension
jtheta() (in module mpmath), 925
method), 1825
JxBra (class in sympy.physics.quantum.spin), kfrom() (in module mpmath), 911
1774
khinchin (mpmath.mp attribute), 751
JxBraCoupled
(class
in kid rsa private key()
(in
module
sympy.physics.quantum.spin), 1777
sympy.crypto.crypto), 314
JxKet (class in sympy.physics.quantum.spin), kid rsa public key()
(in
module
1774
sympy.crypto.crypto), 314
JxKetCoupled
(class
in killable index (sympy.functions.special.tensor functions.Kron
sympy.physics.quantum.spin), 1776
attribute), 449
JyBra (class in sympy.physics.quantum.spin), killable index (sympy.physics.secondquant.KroneckerDelta
1774
attribute), 1607
JyBraCoupled
(class
in kindidict() (sympy.physics.mechanics.kane.KanesMethod
sympy.physics.quantum.spin), 1777
method), 1743
JyKet (class in sympy.physics.quantum.spin), kinematic equations()
(in
module
sympy.physics.vector.functions),
1774
JyKetCoupled
(class
in
1680
sympy.physics.quantum.spin), 1777
kinetic energy()
(in
module
JzBra (class in sympy.physics.quantum.spin),
sympy.physics.mechanics.functions),
1776
1739
JzBraCoupled
(class
in kinetic energy() (sympy.physics.mechanics.particle.Particle
sympy.physics.quantum.spin), 1779
method), 1731
JzKet (class in sympy.physics.quantum.spin), kinetic energy() (sympy.physics.mechanics.rigidbody.RigidB
1774
method), 1734
JzKetCoupled
(class
in kleinj() (in module mpmath), 929
sympy.physics.quantum.spin), 1777
known functions
(in
module
sympy.printing.ccode), 1252
K
known functions
(in
module
sympy.printing.mathematica), 1258
kanes equations() (sympy.physics.mechanics.kane.KanesMethod
KroneckerDelta
(class
in
method), 1743
sympy.functions.special.tensor functions),
KanesMethod
(class
in
446
sympy.physics.mechanics.kane),
KroneckerDelta
(class
in
1741
sympy.physics.secondquant), 1604
kbins() (in module sympy.utilities.iterables),
ksubsets()
(sympy.combinatorics.subsets
1561
256
static
method),
kei() (in module mpmath), 832
Kumaraswamy()
(in
module
sympy.stats),
ker() (in module mpmath), 831
1377
kernel() (sympy.polys.agca.homomorphisms.ModuleHomomorphism
method), 1146
Ket (class in sympy.physics.quantum.state), L
(sympy.polys.polyclasses.DMP
l1 norm()
1781
method), 1165
ket (sympy.physics.quantum.operator.OuterProduct
l1 norm()
(sympy.polys.polytools.Poly
attribute), 1763
method), 1096
1962
Index
SymPy Documentation, Release 0.7.6
L2 (class in sympy.physics.quantum.hilbert), lcm() (sympy.polys.domains.domain.Domain
1759
method), 1151
label (sympy.physics.quantum.state.TimeDepState
lcm()
(sympy.polys.domains.eld.Field
attribute), 1783
method), 1153
label (sympy.tensor.indexed.Idx attribute), lcm() (sympy.polys.domains.PolynomialRing
1510
method), 1157
label (sympy.tensor.indexed.IndexedBase at- lcm()
(sympy.polys.domains.RealField
tribute), 1514
method), 1160
labeller()
(in
module lcm() (sympy.polys.polyclasses.DMP method),
sympy.physics.quantum.circuitplot),
1165
1788
lcm() (sympy.polys.polytools.Poly method),
LagrangesMethod
(class
in
1097
sympy.physics.mechanics.lagrange),
lcm list() (in module sympy.polys.polytools),
1744
1067
Lagrangian()
(in
module ldescent()
(in
module
sympy.physics.mechanics.functions),
sympy.solvers.diophantine), 1503
1740
ldexp() (in module mpmath), 740
laguerre
(class
in LDLdecomposition()
sympy.functions.special.polynomials),
(sympy.matrices.matrices.MatrixBase
441
method), 645
laguerre() (in module mpmath), 880
LDLdecomposition()
laguerre poly()
(in
module
(sympy.matrices.sparse.SparseMatrix
sympy.polys.orthopolys), 1120
method), 692
Lambda (class in sympy.core.function), 163
LDLsolve() (sympy.matrices.matrices.MatrixBase
method), 646
LambdaPrinter
(class
in
sympy.printing.lambdarepr), 1259
Le (in module sympy.core.relational), 148
lambdarepr()
(in
module leading expv() (sympy.polys.rings.PolyElement
sympy.printing.lambdarepr), 1259
method), 1219
lambdastr()
(in
module leading monom() (sympy.polys.rings.PolyElement
sympy.utilities.lambdify), 1573
method), 1219
lambdify()
(in
module leading term() (sympy.polys.rings.PolyElement
sympy.utilities.lambdify), 1573
method), 1220
LambertW
(class
in leadterm() (sympy.core.expr.Expr method),
sympy.functions.elementary.exponential),
118
352
left (sympy.sets.sets.Interval attribute), 1318
lambertw() (in module mpmath), 761
left() (sympy.printing.pretty.stringpict.stringPict
method), 1268
Laplace() (in module sympy.stats), 1378
laplace transform()
(in
module left open (sympy.sets.sets.Interval attribute),
sympy.integrals.transforms), 585
1318
latex() (in module sympy.galgebra.printing), leftslash() (sympy.printing.pretty.stringpict.stringPict
method), 1268
486
latex() (in module sympy.printing.latex), 1260 legendre
(class
in
LatexPrinter (class in sympy.printing.latex),
sympy.functions.special.polynomials),
1259
439
LC() (in module sympy.polys.polytools), 1058 legendre() (in module mpmath), 869
LC() (sympy.polys.polyclasses.ANP method), legendre poly()
(in
module
1168
sympy.polys.orthopolys), 1120
LC() (sympy.polys.polyclasses.DMP method), legendre symbol()
(in
module
1162
sympy.ntheory.residue ntheory),
LC() (sympy.polys.polytools.Poly method),
297
1075
legenp() (in module mpmath), 871
lcm() (in module sympy.polys.polytools), 1066 legenq() (in module mpmath), 872
lcm() (sympy.core.numbers.Number method), length (sympy.geometry.line.LinearEntity at127
tribute), 513
Index
1963
SymPy Documentation, Release 0.7.6
length
(sympy.geometry.line.Segment atsympy.solvers.ode), 1430
tribute), 523
lie heuristic chi()
(in
module
length (sympy.geometry.line3d.LinearEntity3D
sympy.solvers.ode), 1430
attribute), 531
lie heuristic function sum()
(in
module
length
(sympy.geometry.line3d.Segment3D
sympy.solvers.ode), 1431
attribute), 538
lie heuristic linear()
(in
module
length
(sympy.geometry.point.Point
atsympy.solvers.ode), 1432
LieDerivative (class in sympy.digeom), 1854
tribute), 501
length (sympy.geometry.point3d.Point3D at- lift() (sympy.polys.polyclasses.DMP method),
tribute), 506
1165
length (sympy.geometry.polygon.RegularPolygon
lift() (sympy.polys.polytools.Poly method),
attribute), 568
1097
length() (sympy.combinatorics.permutations.Permutation
Limit (class in sympy.series.limits), 1304
method), 206
limit() (in module mpmath), 990
length() (sympy.polys.polytools.Poly method), limit() (in module sympy.series.limits), 1304
1097
limit() (sympy.core.expr.Expr method), 118
lens formula()
(in
module limit() (sympy.matrices.matrices.MatrixBase
sympy.physics.optics.utils), 1816
method), 668
lens makers formula()
(in
module limit denominator() (sympy.core.numbers.Rational
sympy.physics.optics.utils), 1815
method), 131
lerchphi
(class
in limitinf() (in module sympy.series.gruntz),
sympy.functions.special.zeta functions),
1306
424
limits
(sympy.geometry.curve.Curve
atlerchphi() (in module mpmath), 943
tribute), 541
LessThan (class in sympy.core.relational), limits (sympy.physics.quantum.state.Wavefunction
attribute), 1786
152
LeviCivita
(class
in Line (class in sympy.geometry.line), 517
sympy.functions.special.tensor functions),
Line2DBaseSeries
(class
in
446
sympy.plotting.plot), 1282
levin() (in module mpmath), 994
Line3D (class in sympy.geometry.line3d), 525
LexOrder (class in sympy.polys.orderings), Line3DBaseSeries
(class
in
1118
sympy.plotting.plot), 1283
lfsr autocorrelation()
(in
module line integrate()
(sympy.integrals
static
sympy.crypto.crypto), 317
method), 610
lfsr connection polynomial()
(in
module linear derivation()
(in
module
sympy.crypto.crypto), 317
sympy.galgebra.ncutil), 487
lfsr sequence()
(in
module linear expand()
(in
module
sympy.crypto.crypto), 316
sympy.galgebra.ncutil), 487
Li (class in sympy.functions.special.error functions),
linear function()
(in
module
403
sympy.galgebra.ncutil), 487
li (class in sympy.functions.special.error functions),
linear momentum()
(in
module
401
sympy.physics.mechanics.functions),
li() (in module mpmath), 799
1737
lie heuristic abaco1 product() (in module linear momentum() (sympy.physics.mechanics.particle.Parti
sympy.solvers.ode), 1429
method), 1732
lie heuristic abaco1 simple()
(in
module linear momentum() (sympy.physics.mechanics.rigidbody.Rig
sympy.solvers.ode), 1429
method), 1735
lie heuristic abaco2 similar()
(in
module linear projection()
(in
module
sympy.solvers.ode), 1430
sympy.galgebra.ncutil), 487
lie heuristic abaco2 unique general()
(in LinearEntity (class in sympy.geometry.line),
509
module sympy.solvers.ode), 1432
lie heuristic abaco2 unique unknown()
(in LinearEntity3D
(class
in
module sympy.solvers.ode), 1432
sympy.geometry.line3d), 527
lie heuristic bivariate()
(in
module linearize() (sympy.physics.mechanics.kane.KanesMethod
1964
Index
SymPy Documentation, Release 0.7.6
method), 1743
lower (sympy.physics.secondquant.AntiSymmetricTensor
linearize() (sympy.physics.mechanics.lagrange.LagrangesMethod
attribute), 1621
method), 1745
lower (sympy.tensor.indexed.Idx attribute),
linearize() (sympy.physics.mechanics.linearize.Linearizer
1510
method), 1747
lower central series()
Linearizer
(class
in
(sympy.combinatorics.perm groups.PermutationGro
sympy.physics.mechanics.linearize),
method), 232
1746
lower triangular solve()
LineOver1DRangeSeries
(class
in
(sympy.matrices.matrices.MatrixBase
sympy.plotting.plot), 1282
method), 668
linspace() (in module mpmath), 742
lowergamma
(class
in
list() (sympy.combinatorics.permutations.Cycle
sympy.functions.special.gamma functions),
method), 214
385
list() (sympy.combinatorics.permutations.Permutation
lseries() (sympy.core.expr.Expr method), 118
Lt (in module sympy.core.relational), 148
method), 206
list2numpy()
(in
module LT() (in module sympy.polys.polytools), 1058
sympy.matrices.dense), 683
LT() (sympy.polys.polytools.Poly method),
list can dims (sympy.physics.unitsystems.dimensions.DimensionSystem
1076
attribute), 1826
ltrim() (sympy.polys.polytools.Poly method),
list visitor()
(in
module
1097
sympy.utilities.enumerative), 1550
lucas (class in sympy.functions.combinatorial.numbers),
listcoes()
(sympy.polys.rings.PolyElement
373
method), 1220
LUdecomposition() (sympy.matrices.matrices.MatrixBase
listmonoms() (sympy.polys.rings.PolyElement
method), 646
method), 1220
LUdecomposition Simple()
(sympy.matrices.matrices.MatrixBase
listterms()
(sympy.polys.rings.PolyElement
method), 647
method), 1220
liupc() (sympy.matrices.sparse.SparseMatrix LUdecompositionFF()
method), 696
(sympy.matrices.matrices.MatrixBase
method), 646
LM() (in module sympy.polys.polytools), 1058
LM() (sympy.polys.polytools.Poly method), LUsolve() (sympy.matrices.matrices.MatrixBase
1075
method), 647
ln() (in module mpmath), 760
locatenew() (sympy.physics.vector.point.Point M
method), 1677
mag() (in module mpmath), 740
log (class in sympy.functions.elementary.exponential),
magnitude() (sympy.physics.vector.vector.Vector
353
method), 1671
log() (in module mpmath), 759
major (sympy.geometry.ellipse.Ellipse atlog() (sympy.polys.domains.domain.Domain
tribute), 549
method), 1151
make perm() (sympy.combinatorics.perm groups.Permutatio
log()
(sympy.polys.domains.IntegerRing
method), 233
method), 1156
make prime()
(in
module
log10() (in module mpmath), 761
sympy.solvers.diophantine), 1504
logcombine()
(in
module make routine()
(in
module
sympy.simplify.simplify), 1344
sympy.utilities.codegen), 1545
loggamma
(class
in mangoldt() (in module mpmath), 968
sympy.functions.special.gamma functions),
Manifold (class in sympy.digeom), 1846
380
Manifold (class in sympy.galgebra.manifold),
loggamma() (in module mpmath), 783
483
Logistic() (in module sympy.stats), 1378
Manifold() (in module sympy.galgebra.ga),
logm() (in module mpmath), 1032
469
LogNormal() (in module sympy.stats), 1379
manualintegrate() (sympy.integrals static
lommels1() (in module mpmath), 836
method), 614
lommels2() (in module mpmath), 837
Index
1965
SymPy Documentation, Release 0.7.6
map() (sympy.polys.domains.domain.Domain
method), 207
method), 1151
max div (sympy.combinatorics.perm groups.PermutationGro
mass (sympy.physics.mechanics.particle.Particle
attribute), 233
attribute), 1732
max norm()
(sympy.polys.polyclasses.DMP
mass matrix (sympy.physics.mechanics.kane.KanesMethod
method), 1165
attribute), 1743
max norm()
(sympy.polys.polytools.Poly
mass matrix (sympy.physics.mechanics.lagrange.LagrangesMethod
method), 1098
maxcalls() (in module mpmath), 745
attribute), 1746
mass matrix full (sympy.physics.mechanics.kane.KanesMethod
Maxwell() (in module sympy.stats), 1380
attribute), 1743
maybe()
(in
module
mass matrix full (sympy.physics.mechanics.lagrange.LagrangesMethod
sympy.parsing.sympy tokenize),
attribute), 1746
1590
MatAdd (class in sympy.matrices.expressions), MCodePrinter
(class
in
708
sympy.printing.mathematica), 1258
match() (sympy.core.basic.Basic method), 95 mdft() (in module sympy.physics.matrices),
matches() (sympy.core.basic.Basic method),
1600
96
MDNewton
(class
in
mpmathematica()
(in
module
math.calculus.optimization), 979
sympy.parsing.mathematica), 1590
measure (sympy.sets.sets.Set attribute), 1315
mathematica code()
(in
module measure all()
(in
module
sympy.printing.mathematica), 1259
sympy.physics.quantum.qubit), 1798
mathml() (in module sympy.printing.mathml), measure all oneshot()
(in
module
1262
sympy.physics.quantum.qubit), 1800
mathml tag() (sympy.printing.mathml.MathMLPrinter
measure partial()
(in
module
method), 1262
sympy.physics.quantum.qubit), 1799
MathMLPrinter
(class
in measure partial oneshot()
(in
module
sympy.printing.mathml), 1262
sympy.physics.quantum.qubit), 1800
MatMul (class in sympy.matrices.expressions), medial (sympy.geometry.polygon.Triangle at709
tribute), 575
MatPow
(class
in medians (sympy.geometry.polygon.Triangle
sympy.matrices.expressions), 709
attribute), 576
matrix2numpy()
(in
module Medium
(class
in
sympy.matrices.dense), 683
sympy.physics.optics.medium), 1812
matrix fglm()
(in
module meets()
(sympy.series.gruntz.SubsSet
sympy.polys.fglmtools), 1236
method), 1307
matrix multiply elementwise() (in module meijerg (class in sympy.functions.special.hyper),
sympy.matrices.dense), 679
428
matrix rep()
(in
module meijerg() (in module mpmath), 898
sympy.physics.secondquant), 1616
mellin transform()
(in
module
matrix to density()
(in
module
sympy.integrals.transforms), 584
sympy.physics.quantum.qubit), 1798 memoize() (in module mpmath), 745
matrix to qubit()
(in
module merge solution()
(in
module
sympy.physics.quantum.qubit), 1798
sympy.solvers.diophantine), 1500
MatrixBase
(class
in mertens (mpmath.mp attribute), 751
sympy.matrices.matrices), 644
metric to Christoel 1st()
(in
module
MatrixError
(class
in
sympy.digeom), 1859
sympy.matrices.matrices), 678
metric to Christoel 2nd()
(in
module
MatrixExpr
(class
in
sympy.digeom), 1859
sympy.matrices.expressions), 707
metric to Ricci components()
(in
module
sympy.digeom), 1860
MatrixSymbol
(class
in
metric to Riemann components() (in module
sympy.matrices.expressions), 708
Max (class in sympy.functions.elementary.miscellaneous),
sympy.digeom), 1859
355
mfrom() (in module mpmath), 910
max() (sympy.combinatorics.permutations.Permutation
mgamma()
(in
module
1966
Index
SymPy Documentation, Release 0.7.6
sympy.physics.matrices), 1600
attribute), 1758
midpoint (sympy.geometry.line.Segment at- momentum (sympy.physics.quantum.cartesian.PxKet
tribute), 524
attribute), 1757
midpoint (sympy.geometry.line3d.Segment3D monic() (in module sympy.polys.polytools),
attribute), 538
1067
midpoint()
(sympy.geometry.point.Point monic()
(sympy.polys.polyclasses.DMP
method), 501
method), 1165
midpoint() (sympy.geometry.point3d.Point3D monic() (sympy.polys.polytools.Poly method),
method), 507
1098
Min (class in sympy.functions.elementary.miscellaneous),
monic()
(sympy.polys.rings.PolyElement
354
method), 1220
min() (sympy.combinatorics.permutations.Permutation
monitor() (in module mpmath), 745
method), 207
Monomial (class in sympy.polys.monomials),
min qubits (sympy.physics.quantum.gate.CGate
1117
monomial basis() (sympy.polys.rings.PolyRing
attribute), 1790
min qubits (sympy.physics.quantum.gate.CNotGate
method), 1216
attribute), 1792
monomial count()
(in
module
min qubits (sympy.physics.quantum.gate.Gate
sympy.polys.monomials), 1118
attribute), 1789
monoms()
(sympy.polys.polyclasses.DMP
minimal block() (sympy.combinatorics.perm groups.PermutationGroup
method), 1165
method), 234
monoms()
(sympy.polys.polytools.Poly
minimal polynomial()
(in
module
method), 1098
sympy.polys.numberelds), 1115
monoms()
(sympy.polys.rings.PolyElement
minlex() (in module sympy.utilities.iterables),
method), 1220
1562
Morphism (class in sympy.categories), 1829
minor (sympy.geometry.ellipse.Ellipse at- morphisms (sympy.categories.diagram drawing.DiagramGri
tribute), 549
attribute), 1840
minorEntry() (sympy.matrices.matrices.MatrixBase
mpmath.functions.elliptic (module), 909
method), 668
mpmathify() (in module mpmath), 727
minorMatrix() (sympy.matrices.matrices.MatrixBase
mr() (in module sympy.ntheory.primetest),
method), 668
294
minpoly()
(in
module mrv() (in module sympy.series.gruntz), 1307
sympy.polys.numberelds), 1116
mrv leadterm()
(in
module
mirror formula()
(in
module
sympy.series.gruntz), 1306
sympy.physics.optics.utils), 1816
mrv max1() (in module sympy.series.gruntz),
MNewton
(class
in
mp1307
math.calculus.optimization), 977
mrv max3() (in module sympy.series.gruntz),
mnorm() (in module mpmath), 1023
1307
mobius (class in sympy.ntheory), 300
msigma()
(in
module
Mod (class in sympy.core.mod), 147
sympy.physics.matrices), 1601
modgcd bivariate
(class
in msubs()
(in
module
sympy.polys.modulargcd), 1240
sympy.physics.mechanics), 1748
modgcd multivariate
(class
in Mul (class in sympy.core.mul), 142
Mul() (sympy.assumptions.handlers.calculus.AskBoundedHa
sympy.polys.modulargcd), 1241
modgcd univariate
(class
in
static method), 1292
Mul() (sympy.assumptions.handlers.calculus.AskInnitesima
sympy.polys.modulargcd), 1239
Module (class in sympy.polys.agca.modules),
static method), 1292
1131
Mul() (sympy.assumptions.handlers.sets.AskAntiHermitianH
module quotient() (sympy.polys.agca.modules.SubModule
static method), 1293
method), 1136
Mul() (sympy.assumptions.handlers.sets.AskHermitianHand
ModuleHomomorphism
(class
in
static method), 1294
sympy.polys.agca.homomorphisms),
Mul() (sympy.assumptions.handlers.sets.AskImaginaryHand
1144
static method), 1294
momentum (sympy.physics.quantum.cartesian.PxBra
Mul() (sympy.assumptions.handlers.sets.AskIntegerHandler
Index
1967
SymPy Documentation, Release 0.7.6
static method), 1294
sympy.utilities.iterables), 1563
Mul() (sympy.assumptions.handlers.sets.AskRationalHandler
multiset partitions()
(in
module
static method), 1295
sympy.utilities.iterables), 1563
Mul() (sympy.assumptions.handlers.sets.AskRealHandler
multiset partitions taocp()
(in
module
static method), 1295
sympy.utilities.enumerative), 1549
mul() (sympy.polys.domains.domain.Domain multiset permutations()
(in
module
method), 1151
sympy.utilities.iterables), 1564
mul()
(sympy.polys.polyclasses.DMF MultisetPartitionTraverser
(class
in
method), 1168
sympy.utilities.enumerative), 1550
mul() (sympy.polys.polyclasses.DMP method), MultivariatePolynomialError
(class
in
1165
sympy.polys.polyerrors), 1239
mul() (sympy.polys.polytools.Poly method), MutableDenseMatrix
(class
in
sympy.matrices.dense), 686
1098
mul() (sympy.polys.rings.PolyRing method), MutableSparseMatrix
(class
in
1216
sympy.matrices.sparse), 699
mul ground() (sympy.polys.polyclasses.DMP Mx (class in sympy.physics.quantum.circuitplot),
method), 1165
1789
mul ground()
(sympy.polys.polytools.Poly Mz (class in sympy.physics.quantum.circuitplot),
method), 1099
1789
mul inv() (sympy.combinatorics.permutations.Permutation
N
method), 207
Muller
(class
in
mp- N (class in sympy.core.evalf), 182
math.calculus.optimization), 977
n (sympy.combinatorics.graycode.GrayCode
MultiFactorial
(class
in
attribute), 257
sympy.functions.combinatorial.factorials),
N (sympy.physics.quantum.shor.CMod at374
tribute), 1801
multilinear derivation()
(in
module n() (sympy.geometry.point.Point method), 501
sympy.galgebra.ncutil), 487
n()
(sympy.geometry.point3d.Point3D
multilinear function()
(in
module
method), 507
sympy.galgebra.ncutil), 487
n()
(sympy.matrices.matrices.MatrixBase
multilinear product()
(in
module
method), 669
sympy.galgebra.ncutil), 488
n()
(sympy.polys.domains.domain.Domain
multinomial coecients()
(in
module
method), 1151
sympy.ntheory.multinomial), 293
n order()
(in
module
multinomial coecients iterator() (in module
sympy.ntheory.residue ntheory),
sympy.ntheory.multinomial), 293
295
multiplicity()
(in
module Nakagami() (in module sympy.stats), 1381
sympy.ntheory.factor ), 283
name (sympy.categories.Category attribute),
multiply() (sympy.matrices.matrices.MatrixBase
1834
method), 668
name (sympy.categories.NamedMorphism atmultiply() (sympy.matrices.sparse.SparseMatrix
tribute), 1831
method), 696
NamedMorphism (class in sympy.categories),
multiply elementwise()
1831
(sympy.matrices.matrices.MatrixBase NaN (class in sympy.core.numbers), 134
method), 668
Nand (class in sympy.logic.boolalg), 630
multiply ideal() (sympy.polys.agca.modules.FreeModule
nargs (sympy.core.function.FunctionClass atmethod), 1134
tribute), 168
multiply ideal() (sympy.polys.agca.modules.Module
native coes() (sympy.polys.numberelds.AlgebraicNumber
method), 1132
method), 1117
multiply ideal() (sympy.polys.agca.modules.SubModule
Naturals (class in sympy.sets.fancysets), 1324
method), 1136
Naturals0 (class in sympy.sets.fancysets),
multiset()
(in
module
1325
sympy.utilities.iterables), 1562
ncdf() (in module mpmath), 808
multiset combinations()
(in
module Ne (in module sympy.core.relational), 147
1968
Index
SymPy Documentation, Release 0.7.6
necklaces()
(in
module non scalar projection()
(in
module
sympy.utilities.iterables), 1565
sympy.galgebra.ncutil), 488
neg() (sympy.polys.domains.domain.Domain nonnegative, 89
method), 1151
nonpositive, 89
neg() (sympy.polys.polyclasses.DMF method), NonSquareMatrixError
(class
in
1168
sympy.matrices.matrices), 678
neg() (sympy.polys.polyclasses.DMP method), nonzero, 89
1165
Nor (class in sympy.logic.boolalg), 630
neg() (sympy.polys.polytools.Poly method), norm (sympy.physics.quantum.state.Wavefunction
1099
attribute), 1787
negative, 89
norm() (in module mpmath), 1022
NegativeInnity
(class
in norm() (sympy.matrices.matrices.MatrixBase
sympy.core.numbers), 135
method), 669
NegativeOne (class in sympy.core.numbers), Normal() (in module sympy.stats), 1382
133
normal closure() (sympy.combinatorics.perm groups.Permut
new()
(sympy.polys.polytools.Poly
class
method), 234
method), 1099
normal lines() (sympy.geometry.ellipse.Ellipse
Newton
(class
in
mpmethod), 550
math.calculus.optimization), 976
normal vector (sympy.geometry.plane.Plane
next() (sympy.combinatorics.graycode.GrayCode
attribute), 580
method), 258
normalize() (sympy.physics.quantum.state.Wavefunction
next()
(sympy.combinatorics.prufer.Prufer
method), 1787
method), 247
normalize() (sympy.physics.vector.vector.Vector
next() (sympy.printing.pretty.stringpict.stringPict
method), 1671
static method), 1268
normalized()
(in
module
sympy.physics.quantum.gate), 1793
next binary() (sympy.combinatorics.subsets.Subset
method), 251
normalized() (sympy.matrices.matrices.MatrixBase
next gray() (sympy.combinatorics.subsets.Subset
method), 669
method), 252
NormalPSpace
(class
in
sympy.stats.crv types), 1396
next lex() (sympy.combinatorics.partitions.IntegerPartition
method), 190
Not (class in sympy.logic.boolalg), 628
next lex() (sympy.combinatorics.permutations.Permutation
NotAlgebraic
(class
in
method), 207
sympy.polys.polyerrors), 1238
next lexicographic()
NotInvertible
(class
in
(sympy.combinatorics.subsets.Subset
sympy.polys.polyerrors), 1238
method), 252
NotReversible
(class
in
sympy.polys.polyerrors), 1238
next nonlex() (sympy.combinatorics.permutations.Permutation
method), 207
npartitions()
(in
module
next trotterjohnson()
sympy.ntheory.partitions ), 294
(sympy.combinatorics.permutations.Permutation
npdf() (in module mpmath), 807
method), 208
nprint() (in module mpmath), 727
nextprime()
(in
module nprod() (in module mpmath), 988
sympy.ntheory.generate), 278
nqubits (sympy.physics.quantum.gate.CGate
noat() (in module sympy.core.function), 181
attribute), 1790
nint() (in module mpmath), 736
nqubits
(sympy.physics.quantum.gate.Gate
nint distance() (in module mpmath), 741
attribute), 1789
nnz() (sympy.matrices.sparse.SparseMatrix nroots() (in module sympy.polys.polytools),
1072
method), 697
NO (class in sympy.physics.secondquant), nroots() (sympy.polys.polytools.Poly method),
1617
1099
no attrs in subclass
(in
module nseries() (sympy.core.expr.Expr method), 119
sympy.utilities.decorator), 1547
nsimplify()
(in
module
nodes (sympy.combinatorics.prufer.Prufer atsympy.simplify.simplify), 1340
tribute), 247
nsimplify() (sympy.core.expr.Expr method),
Index
1969
SymPy Documentation, Release 0.7.6
120
ode 1st homogeneous coe best() (in module
nsolve() (in module sympy.solvers.solvers),
sympy.solvers.ode), 1410
1476
ode 1st homogeneous coe subs dep div indep()
nstr() (in module mpmath), 727
(in module sympy.solvers.ode), 1411
nsum() (in module mpmath), 979
ode 1st homogeneous coe subs indep div dep()
nth() (sympy.polys.polyclasses.DMP method),
(in module sympy.solvers.ode), 1412
1165
ode 1st linear()
(in
module
nth() (sympy.polys.polytools.Poly method),
sympy.solvers.ode), 1414
1100
ode 1st power series()
(in
module
nth power roots poly()
(in
module
sympy.solvers.ode), 1426
sympy.polys.polytools), 1073
ode 2nd power series ordinary() (in module
nth power roots poly()
sympy.solvers.ode), 1427
(sympy.polys.polytools.Poly method), ode 2nd power series regular() (in module
1100
sympy.solvers.ode), 1427
nthroot() (in module sympy.simplify.simplify), ode almost linear()
(in
module
1332
sympy.solvers.ode), 1422
nthroot mod()
(in
module ode Bernoulli()
(in
module
sympy.ntheory.residue ntheory),
sympy.solvers.ode), 1414
296
ode lie group()
(in
module
nu (sympy.functions.special.hyper.meijerg atsympy.solvers.ode), 1425
tribute), 430
ode linear coecients()
(in
module
nullspace() (sympy.matrices.matrices.MatrixBase
sympy.solvers.ode), 1423
method), 670
ode Liouville() (in module sympy.solvers.ode),
Number (class in sympy.core.numbers), 126
1416
numbered symbols()
(in
module ode nth linear constant coe homogeneous()
sympy.utilities.iterables), 1565
(in module sympy.solvers.ode), 1418
NumberSymbol
(class
in ode nth linear constant coe undetermined coecients()
sympy.core.numbers), 131
(in module sympy.solvers.ode), 1419
numer() (sympy.polys.domains.AlgebraicField ode nth linear constant coe variation of parameters()
method), 1159
(in module sympy.solvers.ode), 1420
numer() (sympy.polys.domains.domain.Domain ode order() (in module sympy.solvers.deutils),
method), 1151
1478
numer() (sympy.polys.domains.ExpressionDomain
ode Riccati special minus2()
(in
module
method), 1161
sympy.solvers.ode), 1417
numer() (sympy.polys.domains.FractionField ode separable()
(in
module
method), 1160
sympy.solvers.ode), 1421
numer()
(sympy.polys.domains.ring.Ring ode separable reduced()
(in
module
method), 1154
sympy.solvers.ode), 1424
numer()
(sympy.polys.polyclasses.DMF ode sol simplicity()
(in
module
method), 1168
sympy.solvers.ode), 1408
nzeros() (in module mpmath), 939
odefun() (in module mpmath), 1012
odesimp() (in module sympy.solvers.ode),
O
1406
of type() (sympy.polys.domains.domain.Domain
Object (class in sympy.categories), 1829
method), 1152
objects
(sympy.categories.Category
atold
frac
eld() (sympy.polys.domains.domain.Domain
tribute), 1835
method),
1152
objects
(sympy.categories.Diagram
atold
poly
ring()
(sympy.polys.domains.domain.Domain
tribute), 1837
method), 1152
OctaveCodeGen
(class
in
One
(class
in sympy.core.numbers), 133
sympy.utilities.codegen), 1543
one
(sympy.polys.polytools.Poly
attribute),
odd, 89
1100
ode 1st exact()
(in
module
OneQubitGate
(class
in
sympy.solvers.ode), 1409
sympy.physics.quantum.gate), 1790
1970
Index
SymPy Documentation, Release 0.7.6
ones() (in module sympy.matrices.dense), 679 p1
(sympy.geometry.line.LinearEntity
atOperationNotSupported
(class
in
tribute), 513
sympy.polys.polyerrors), 1238
p1
(sympy.geometry.line3d.LinearEntity3D
Operator
(class
in
attribute), 531
sympy.physics.quantum.operator),
p1 (sympy.geometry.plane.Plane attribute),
1760
580
operators (sympy.physics.quantum.state.StateBase
p2
(sympy.geometry.line.LinearEntity
atattribute), 1781
tribute), 513
operators to state()
(in
module p2
(sympy.geometry.line3d.LinearEntity3D
sympy.physics.quantum.operatorset),
attribute), 531
1765
pade() (in module mpmath), 1015
oprint() (in module sympy.galgebra.debug), pairwise prime()
(in
module
sympy.solvers.diophantine), 1504
489
opt cse()
(in
module parallel line() (sympy.geometry.line.LinearEntity
sympy.simplify.cse main), 1346
method), 514
OptionError
(class
in parallel line() (sympy.geometry.line3d.LinearEntity3D
sympy.polys.polyerrors), 1239
method), 532
Or (class in sympy.logic.boolalg), 628
parallel plane() (sympy.geometry.plane.Plane
OracleGate
(class
in
method), 581
sympy.physics.quantum.grover),
parallel poly from expr()
(in
module
1794
sympy.polys.polytools), 1057
orbit() (sympy.combinatorics.perm groups.PermutationGroup
parameter (sympy.geometry.curve.Curve atmethod), 235
tribute), 541
orbit rep() (sympy.combinatorics.perm groups.PermutationGroup
Parametric2DLineSeries
(class
in
method), 236
sympy.plotting.plot), 1282
orbit transversal() (sympy.combinatorics.perm groups.PermutationGroup
Parametric3DLineSeries
(class
in
method), 236
sympy.plotting.plot), 1283
orbits() (sympy.combinatorics.perm groups.PermutationGroup
ParametricSurfaceSeries
(class
in
method), 236
sympy.plotting.plot), 1283
Order (class in sympy.series.order), 1308
parametrize ternary quadratic() (in module
order (sympy.functions.special.bessel.BesselBase
sympy.solvers.diophantine), 1502
attribute), 408
parens() (sympy.printing.pretty.stringpict.stringPict
order() (sympy.combinatorics.perm groups.PermutationGroup
method), 1268
method), 237
Pareto() (in module sympy.stats), 1383
order() (sympy.combinatorics.permutations.Permutation
parity() (sympy.combinatorics.permutations.Permutation
method), 208
method), 209
orient() (sympy.physics.vector.frame.ReferenceFrame
parse expr()
(in
module
method), 1666
sympy.parsing.sympy parser), 1588
orientnew() (sympy.physics.vector.frame.ReferenceFrame
parse line()
(in
module
method), 1667
sympy.galgebra.precedence), 484
orthocenter (sympy.geometry.polygon.Triangle parse maxima()
(in
module
attribute), 576
sympy.parsing.maxima), 1590
ostr() (in module sympy.galgebra.debug), 489 parse paren()
(in
module
outer()
(in
module
sympy.galgebra.precedence), 485
sympy.physics.vector.functions),
partial velocity()
(in
module
1688
sympy.physics.vector.functions),
outer()
(sympy.physics.vector.vector.Vector
1681
Particle (class in sympy.physics.mechanics.particle),
method), 1671
OuterProduct
(class
in
1730
sympy.physics.quantum.operator),
Partition
(class
in
1761
sympy.combinatorics.partitions),
187
P
partition (sympy.combinatorics.partitions.Partition
attribute), 188
P() (in module sympy.stats), 1391
Index
1971
SymPy Documentation, Release 0.7.6
partition()
(in
module
sympy.combinatorics.permutations),
sympy.solvers.diophantine), 1498
192
partitions()
(in
module PermutationGroup
(class
in
sympy.utilities.iterables), 1566
sympy.combinatorics.perm groups),
pat matrix()
(in
module
216
sympy.physics.matrices), 1601
PermutationOperator
(class
in
Patch (class in sympy.digeom), 1846
sympy.physics.secondquant), 1623
pcfd() (in module mpmath), 865
permute()
(sympy.polys.polyclasses.DMP
pcfu() (in module mpmath), 867
method), 1165
pcfv() (in module mpmath), 867
permuteBkwd() (sympy.matrices.matrices.MatrixBase
pcfw() (in module mpmath), 868
method), 670
pde 1st linear constant coe() (in module permuteFwd() (sympy.matrices.matrices.MatrixBase
sympy.solvers.pde), 1465
method), 670
pde 1st linear constant coe homogeneous() perpendicular bisector()
(in module sympy.solvers.pde), 1464
(sympy.geometry.line.Segment
pde 1st linear variable coe() (in module
method), 524
sympy.solvers.pde), 1467
perpendicular line()
pde separate()
(in
module
(sympy.geometry.line.LinearEntity
sympy.solvers.pde), 1460
method), 514
pde separate add()
(in
module perpendicular line()
sympy.solvers.pde), 1461
(sympy.geometry.line3d.LinearEntity3D
pde separate mul()
(in
module
method), 532
sympy.solvers.pde), 1461
perpendicular line()
pdiv() (in module sympy.polys.polytools),
(sympy.geometry.plane.Plane
method), 581
1058
pdiv()
(sympy.polys.polyclasses.DMP perpendicular plane()
method), 1165
(sympy.geometry.plane.Plane
pdiv() (sympy.polys.polytools.Poly method),
method), 581
1100
perpendicular segment()
(sympy.geometry.line.LinearEntity
pdsolve() (in module sympy.solvers.pde),
method), 514
1461
Pegasus
(class
in
mp- perpendicular segment()
math.calculus.optimization), 978
(sympy.geometry.line3d.LinearEntity3D
per() (sympy.polys.polyclasses.DMF method),
method), 532
1168
pexquo() (in module sympy.polys.polytools),
per() (sympy.polys.polyclasses.DMP method),
1059
1165
pexquo()
(sympy.polys.polyclasses.DMP
per() (sympy.polys.polytools.Poly method),
method), 1166
1101
pexquo()
(sympy.polys.polytools.Poly
perfect power()
(in
module
method), 1101
sympy.ntheory.factor ), 283
pgroup (sympy.combinatorics.polyhedron.Polyhedron
periapsis (sympy.geometry.ellipse.Ellipse atattribute), 244
tribute), 550
Phase
(in
module
perimeter (sympy.geometry.polygon.Polygon
sympy.physics.quantum.gate), 1793
attribute), 561
phase (sympy.physics.optics.waves.TWave atperiod nd()
(in
module
tribute), 1818
sympy.physics.quantum.shor), 1801
PhaseGate
(class
in
perm2tensor() (sympy.tensor.tensor.TensMul
sympy.physics.quantum.gate), 1791
method), 1531
phi (mpmath.mp attribute), 751
permeability (sympy.physics.optics.medium.Medium
Pi (class in sympy.core.numbers), 137
pi (mpmath.mp attribute), 751
attribute), 1813
permittivity (sympy.physics.optics.medium.Medium
PIABBra
(class
in
attribute), 1813
sympy.physics.quantum.piab), 1801
Permutation
(class
in PIABHamiltonian
(class
in
1972
Index
SymPy Documentation, Release 0.7.6
sympy.physics.quantum.piab), 1801
point()
(sympy.digeom.CoordSystem
PIABKet
(class
in
method), 1849
sympy.physics.quantum.piab), 1801
Point3D (class in sympy.geometry.point3d),
Piecewise
(class
in
503
sympy.functions.elementary.piecewise),point to coords() (sympy.digeom.CoordSystem
356
method), 1849
piecewise fold()
(in
module points (sympy.geometry.line.LinearEntity atsympy.functions.elementary.piecewise),
tribute), 515
357
points (sympy.geometry.line3d.LinearEntity3D
pinv() (sympy.matrices.matrices.MatrixBase
attribute), 533
method), 670
pointwise stabilizer()
pinv solve() (sympy.matrices.matrices.MatrixBase
(sympy.combinatorics.perm groups.PermutationGro
method), 671
method), 237
Plane (class in sympy.geometry.plane), 577
Poisson() (in module sympy.stats), 1364
Plot (class in sympy.plotting.plot), 1271
polar() (in module mpmath), 735
plot() (in module mpmath), 747
PoleError (class in sympy.core.function), 177
plot() (in module sympy.plotting.plot), 1273
PolicationFailed
(class
in
plot3d() (in module sympy.plotting.plot), 1277
sympy.polys.polyerrors), 1239
plot3d parametric line()
(in
module pollard pm1()
(in
module
sympy.plotting.plot), 1278
sympy.ntheory.factor ), 285
plot3d parametric surface()
(in
module pollard rho()
(in
module
sympy.plotting.plot), 1279
sympy.ntheory.factor ), 283
plot gate() (sympy.physics.quantum.gate.CGatePoly (class in sympy.polys.polytools), 1075
method), 1790
poly() (in module sympy.polys.polytools),
plot implicit()
(in
module
1057
sympy.plotting.plot implicit), 1280
poly from expr()
(in
module
plot interval() (sympy.geometry.curve.Curve
sympy.polys.polytools), 1057
method), 541
poly ring() (sympy.polys.domains.domain.Domain
plot interval() (sympy.geometry.ellipse.Ellipse
method), 1152
method), 551
poly unify()
(sympy.polys.polyclasses.DMF
plot interval()
(sympy.geometry.line.Line
method), 1168
method), 519
PolyElement (class in sympy.polys.rings),
plot interval()
(sympy.geometry.line.Ray
1216
method), 521
polyexp() (in module mpmath), 950
plot interval() (sympy.geometry.line.Segment polygamma
(class
in
method), 524
sympy.functions.special.gamma functions),
plot interval() (sympy.geometry.line3d.Line3D
382
method), 526
Polygon (class in sympy.geometry.polygon),
plot interval() (sympy.geometry.line3d.Ray3D
556
method), 535
Polyhedron
(class
in
sympy.combinatorics.polyhedron),
plot interval() (sympy.geometry.line3d.Segment3D
method), 538
243
plot interval() (sympy.geometry.polygon.Polygonpolylog (class in sympy.functions.special.zeta functions),
423
method), 561
plot parametric()
(in
module polylog() (in module mpmath), 944
PolynomialError
(class
in
sympy.plotting.plot), 1275
Point (class in sympy.digeom), 1849
sympy.polys.polyerrors), 1238
Point (class in sympy.geometry.point), 497
PolynomialRing
(class
in
Point (class in sympy.physics.vector.point),
sympy.polys.domains), 1156
1675
PolyRing (class in sympy.polys.rings), 1215
point (sympy.core.function.Subs attribute), polyroots() (in module mpmath), 971
171
polyval() (in module mpmath), 971
point (sympy.physics.mechanics.particle.Particle
pos() (sympy.polys.domains.domain.Domain
attribute), 1732
method), 1152
Index
1973
SymPy Documentation, Release 0.7.6
pos from() (sympy.physics.vector.point.Point Pow() (sympy.assumptions.handlers.sets.AskRationalHandle
method), 1677
static method), 1295
POSform() (in module sympy.logic.boolalg), Pow() (sympy.assumptions.handlers.sets.AskRealHandler
626
static method), 1295
posify() (in module sympy.simplify.simplify), pow() (sympy.polys.domains.domain.Domain
1342
method), 1152
position (sympy.physics.quantum.cartesian.XBra
pow() (sympy.polys.polyclasses.ANP method),
1168
attribute), 1757
position (sympy.physics.quantum.cartesian.XKetpow()
(sympy.polys.polyclasses.DMF
attribute), 1757
method), 1168
position x (sympy.physics.quantum.cartesian.PositionState3D
pow()
(sympy.polys.polyclasses.DMP
attribute), 1758
method), 1166
position y (sympy.physics.quantum.cartesian.PositionState3D
pow() (sympy.polys.polytools.Poly method),
attribute), 1758
1101
position z (sympy.physics.quantum.cartesian.PositionState3D
powdenest()
(in
module
attribute), 1758
sympy.simplify.simplify), 1342
PositionBra3D
(class
in power() (in module mpmath), 757
sympy.physics.quantum.cartesian),
powerset() (sympy.sets.sets.Set method),
1758
1315
PositionKet3D
(class
in powm() (in module mpmath), 1033
sympy.physics.quantum.cartesian),
powm1() (in module mpmath), 759
1758
powsimp()
(in
module
PositionState3D
(class
in
sympy.simplify.simplify), 1337
sympy.physics.quantum.cartesian),
powsimp() (sympy.core.expr.Expr method),
1758
120
positive, 89
pprint nodes()
(in
module
postxes()
(in
module
sympy.printing.tree), 1263
sympy.utilities.iterables), 1567
PQa() (in module sympy.solvers.diophantine),
postorder traversal()
(in
module
1501
sympy.utilities.iterables), 1567
pquo() (in module sympy.polys.polytools),
potential energy (sympy.physics.mechanics.particle.Particle
1059
attribute), 1732
pquo()
(sympy.polys.polyclasses.DMP
potential energy (sympy.physics.mechanics.rigidbody.RigidBody
method), 1166
attribute), 1735
pquo() (sympy.polys.polytools.Poly method),
potential energy()
(in
module
1102
sympy.physics.mechanics.functions), PRECEDENCE
(in
module
sympy.printing.precedence), 1266
1739
Pow (class in sympy.core.power), 139
precedence()
(in
module
Pow() (sympy.assumptions.handlers.calculus.AskBoundedHandler
sympy.printing.precedence), 1266
static method), 1292
PRECEDENCE FUNCTIONS
(in
module
sympy.printing.precedence), 1266
Pow() (sympy.assumptions.handlers.calculus.AskInnitesimalHandler
static method), 1292
PRECEDENCE VALUES
(in
module
Pow() (sympy.assumptions.handlers.ntheory.AskPrimeHandler
sympy.printing.precedence), 1266
PrecisionExhausted
(class
in
static method), 1293
Pow() (sympy.assumptions.handlers.order.AskNegativeHandler
sympy.core.evalf), 181
Predicate
(class
in
static method), 1293
Pow() (sympy.assumptions.handlers.sets.AskAntiHermitianHandler
sympy.assumptions.assume), 1289
preferred index (sympy.functions.special.tensor functions.K
static method), 1294
Pow() (sympy.assumptions.handlers.sets.AskHermitianHandler
attribute), 450
static method), 1294
preferred index (sympy.physics.secondquant.KroneckerDelt
Pow() (sympy.assumptions.handlers.sets.AskImaginaryHandler
attribute), 1608
static method), 1294
Prex (class in sympy.physics.unitsystems.prexes),
Pow() (sympy.assumptions.handlers.sets.AskIntegerHandler
1827
static method), 1294
prexes()
(in
module
1974
Index
SymPy Documentation, Release 0.7.6
sympy.utilities.iterables), 1568
sympy.ntheory.factor ), 289
prem() (in module sympy.polys.polytools), primepi() (in module mpmath), 964
1059
primepi()
(in
module
prem()
(sympy.polys.polyclasses.DMP
sympy.ntheory.generate), 277
method), 1166
primepi2() (in module mpmath), 964
prem() (sympy.polys.polytools.Poly method), primerange()
(in
module
1102
sympy.ntheory.generate), 278
premises
(sympy.categories.Diagram
at- primerange() (sympy.ntheory.generate.Sieve
tribute), 1837
method), 276
pretty()
(in
module primezeta() (in module mpmath), 951
sympy.printing.pretty.pretty), 1251
primitive() (in module sympy.polys.polytools),
pretty atom()
(in
module
1068
sympy.printing.pretty.pretty symbology),
primitive() (sympy.core.add.Add method),
1267
146
pretty print()
(in
module primitive() (sympy.core.expr.Expr method),
120
sympy.printing.pretty.pretty), 1251
pretty symbol()
(in
module primitive()
(sympy.polys.polyclasses.DMP
sympy.printing.pretty.pretty symbology),
method), 1166
1267
primitive()
(sympy.polys.polytools.Poly
pretty try use unicode()
(in
module
method), 1102
sympy.printing.pretty.pretty symbology),
primitive()
(sympy.polys.rings.PolyElement
1267
method), 1220
pretty use unicode()
(in
module primitive element()
(in
module
sympy.printing.pretty.pretty symbology),
sympy.polys.numberelds), 1117
1266
primitive root()
(in
module
sympy.ntheory.residue ntheory),
prettyForm
(class
in
sympy.printing.pretty.stringpict),
295
1269
primorial()
(in
module
PrettyPrinter
(class
in
sympy.ntheory.generate), 280
sympy.printing.pretty.pretty), 1251
print ccode()
(in
module
prev()
(sympy.combinatorics.prufer.Prufer
sympy.printing.ccode), 1254
method), 247
print dim base() (sympy.physics.unitsystems.dimensions.Dim
prev binary() (sympy.combinatorics.subsets.Subset
method), 1827
method), 252
print fcode()
(in
module
sympy.printing.fcode), 1256
prev gray() (sympy.combinatorics.subsets.Subset
method), 253
Print Function()
(in
module
sympy.galgebra.printing), 486
prev lex() (sympy.combinatorics.partitions.IntegerPartition
method), 190
print gtk() (in module sympy.printing.gtk),
prev lexicographic()
1259
(sympy.combinatorics.subsets.Subset print latex()
(in
module
method), 253
sympy.galgebra.printing), 486
preview()
(in
module print latex() (in module sympy.printing.latex),
sympy.printing.preview), 1264
1261
preview diagram()
(in
module print mathml()
(in
module
sympy.categories.diagram drawing),
sympy.printing.mathml), 1262
1846
print node() (in module sympy.printing.tree),
prevprime()
(in
module
1263
print nonzero() (sympy.matrices.matrices.MatrixBase
sympy.ntheory.generate), 278
prime, 89
method), 672
prime() (in module sympy.ntheory.generate), print product table()
(in
module
277
sympy.galgebra.debug), 489
prime as sum of two squares() (in module print sub table()
(in
module
sympy.solvers.diophantine), 1504
sympy.galgebra.debug), 489
primefactors()
(in
module print tree() (in module sympy.printing.tree),
Index
1975
SymPy Documentation, Release 0.7.6
1263
method), 582
print unit base() (sympy.physics.unitsystems.units.UnitSystem
Prufer (class in sympy.combinatorics.prufer),
method), 1829
246
Printer (class in sympy.printing.printer), 1249 prufer rank() (sympy.combinatorics.prufer.Prufer
printmethod (sympy.printing.ccode.CCodePrinter
method), 248
attribute), 1252
prufer repr (sympy.combinatorics.prufer.Prufer
printmethod (sympy.printing.codeprinter.CodePrinter attribute), 248
psi() (in module mpmath), 792
attribute), 1266
printmethod (sympy.printing.fcode.FCodePrinter
psi n() (in module sympy.physics.qho 1d),
attribute), 1257
1602
printmethod (sympy.printing.lambdarepr.LambdaPrinter
pslq() (in module mpmath), 1039
attribute), 1259
PSpace (class in sympy.stats.rv), 1396
printmethod (sympy.printing.latex.LatexPrinter pspace() (in module sympy.stats.rv), 1397
attribute), 1259
public() (in module sympy.utilities.decorator),
printmethod (sympy.printing.mathematica.MCodePrinter
1547
PurePoly (class in sympy.polys.polytools),
attribute), 1258
printmethod (sympy.printing.mathml.MathMLPrinter 1111
attribute), 1262
PxBra (class in sympy.physics.quantum.cartesian),
printmethod (sympy.printing.pretty.pretty.PrettyPrinter1758
attribute), 1251
PxKet (class in sympy.physics.quantum.cartesian),
printmethod (sympy.printing.printer.Printer
1757
attribute), 1251
PxOp (class in sympy.physics.quantum.cartesian),
printmethod (sympy.printing.repr.ReprPrinter
1757
attribute), 1262
PyTestReporter
(class
in
sympy.utilities.runtests), 1580
printmethod
(sympy.printing.str.StrPrinter
attribute), 1263
Python Enhancement Proposals
printtoken()
(in
module
PEP 335, 1900, 1901
sympy.parsing.sympy tokenize),
PythonFiniteField
(class
in
1589
sympy.polys.domains), 1162
prob() (sympy.physics.quantum.state.Wavefunction
PythonIntegerRing
(class
in
method), 1787
sympy.polys.domains), 1162
prod() (in module sympy.core.mul), 144
PythonRationalField
(class
in
Product (class in sympy.concrete.products),
sympy.polys.domains), 1162
324
product()
(in
module Q
sympy.concrete.products), 329
Q (class in sympy.assumptions.ask), 1287
product()
(sympy.polys.agca.ideals.Ideal q (sympy.physics.optics.gaussopt.BeamParameter
method), 1139
attribute), 1808
product derivation()
(in
module qapply()
(in
module
sympy.galgebra.ncutil), 488
sympy.physics.quantum.qapply),
ProductDomain (class in sympy.stats.rv),
1766
1396
qbarfrom() (in module mpmath), 910
ProductPSpace (class in sympy.stats.rv), 1396 qfac() (in module mpmath), 970
ProductSet (class in sympy.sets.sets), 1321
qfrom() (in module mpmath), 909
Proj() (in module sympy.galgebra.ga), 470
QFT (class in sympy.physics.quantum.qft),
project() (sympy.matrices.matrices.MatrixBase
1796
method), 672
qgamma() (in module mpmath), 969
projection() (sympy.geometry.line.LinearEntity qhyper() (in module mpmath), 970
method), 515
qp() (in module mpmath), 968
projection() (sympy.geometry.line3d.LinearEntity3D
QRdecomposition() (sympy.matrices.matrices.MatrixBase
method), 533
method), 647
projection()
(sympy.geometry.plane.Plane QRsolve() (sympy.matrices.matrices.MatrixBase
method), 582
method), 648
projection line() (sympy.geometry.plane.Plane quad() (in module mpmath), 1003
1976
Index
SymPy Documentation, Release 0.7.6
quadosc() (in module mpmath), 1007
R
quadratic residues()
(in
module R nl() (in module sympy.physics.hydrogen),
sympy.ntheory.residue ntheory),
1599
296
R nl() (in module sympy.physics.sho), 1603
QuadraticU() (in module sympy.stats), 1383
racah() (in module sympy.physics.wigner),
QuadratureRule
(class
in
mp1626
math.calculus.quadrature), 1010
rad rationalize()
(in
module
Quantity
(class
in
sympy.simplify.simplify), 1333
sympy.physics.unitsystems.quantities), radians() (in module mpmath), 767
1829
radical()
(sympy.polys.agca.ideals.Ideal
Qubit (class in sympy.physics.quantum.qubit),
method), 1140
1796
radius
(sympy.geometry.ellipse.Circle
atqubit to matrix()
(in
module
tribute), 555
sympy.physics.quantum.qubit), 1798 radius (sympy.geometry.polygon.RegularPolygon
QubitBra
(class
in
attribute), 568
sympy.physics.quantum.qubit), 1797 radius (sympy.physics.optics.gaussopt.BeamParameter
quo() (in module sympy.polys.polytools), 1060
attribute), 1808
quo() (sympy.polys.domains.domain.Domain radius of convergence
method), 1152
(sympy.functions.special.hyper.hyper
quo()
(sympy.polys.domains.eld.Field
attribute), 428
method), 1153
radsimp() (in module sympy.simplify.simplify),
quo()
(sympy.polys.domains.ring.Ring
1333
method), 1154
radsimp() (sympy.core.expr.Expr method),
quo() (sympy.polys.polyclasses.DMF method),
120
1168
RaisedCosine() (in module sympy.stats), 1384
quo() (sympy.polys.polyclasses.DMP method), raises() (in module sympy.utilities.pytest),
1166
1578
quo() (sympy.polys.polytools.Poly method), rand() (in module mpmath), 742
1102
randMatrix()
(in
module
quo ground() (sympy.polys.polyclasses.DMP
sympy.matrices.dense), 682
method), 1166
random() (sympy.combinatorics.perm groups.PermutationGr
quo ground()
(sympy.polys.polytools.Poly
method), 238
method), 1103
random() (sympy.combinatorics.permutations.Permutation
quotient()
(sympy.polys.agca.ideals.Ideal
class method), 209
method), 1139
random bitstring() (sympy.combinatorics.graycode
quotient codomain()
static method), 259
(sympy.polys.agca.homomorphisms.ModuleHomomorphism
random circuit()
(in
module
method), 1147
sympy.physics.quantum.gate), 1793
quotient domain() (sympy.polys.agca.homomorphisms.ModuleHomomorphism
random complex number()
(in
module
method), 1147
sympy.utilities.randtest), 1579
quotient hom() (sympy.polys.agca.modules.QuotientModule
random integer partition()
(in
module
method), 1141
sympy.combinatorics.partitions),
quotient hom() (sympy.polys.agca.modules.SubQuotientModule
190
method), 1142
random point() (sympy.geometry.ellipse.Ellipse
quotient module() (sympy.polys.agca.modules.FreeModule
method), 551
method), 1134
random point() (sympy.geometry.line.LinearEntity
quotient module() (sympy.polys.agca.modules.Module method), 516
method), 1132
random point() (sympy.geometry.plane.Plane
quotient module() (sympy.polys.agca.modules.SubModule
method), 582
method), 1137
random poly()
(in
module
quotient ring() (sympy.polys.domains.ring.Ring
sympy.polys.specialpolys), 1120
method), 1154
random pr() (sympy.combinatorics.perm groups.Permutation
QuotientModule
(class
in
method), 238
sympy.polys.agca.modules), 1140
random stab() (sympy.combinatorics.perm groups.Permutat
Index
1977
SymPy Documentation, Release 0.7.6
method), 238
1809
random symbols() (in module sympy.stats.rv), RayTransferMatrix
(class
in
1397
sympy.physics.optics.gaussopt),
RandomDomain (class in sympy.stats.rv),
1802
1396
rcall() (sympy.core.basic.Basic method), 96
RandomSymbol (class in sympy.stats.rv), rcollect() (in module sympy.simplify.simplify),
1396
1331
randprime()
(in
module re (class in sympy.functions.elementary.complexes),
sympy.ntheory.generate), 279
357
ranges (sympy.tensor.indexed.Indexed at- re() (in module mpmath), 734
tribute), 1512
real, 89
rank (sympy.combinatorics.graycode.GrayCode real roots()
(in
module
attribute), 258
sympy.polys.polytools), 1072
rank (sympy.combinatorics.partitions.Partition real roots()
(sympy.polys.polytools.Poly
attribute), 188
method), 1103
rank (sympy.combinatorics.prufer.Prufer at- RealField (class in sympy.polys.domains),
tribute), 248
1160
rank
(sympy.tensor.indexed.Indexed
at- RealNumber
(in
module
tribute), 1512
sympy.core.numbers), 131
rank() (sympy.combinatorics.permutations.Permutation
reconstruct()
(in
module
method), 209
sympy.solvers.diophantine), 1505
rank() (sympy.matrices.matrices.MatrixBase rect() (in module mpmath), 735
method), 672
recurrence memo()
(in
module
rank binary (sympy.combinatorics.subsets.Subset
sympy.utilities.memoization), 1576
attribute), 253
red groebner()
(in
module
sympy.polys.groebnertools), 1236
rank gray (sympy.combinatorics.subsets.Subset
attribute), 253
reduce() (sympy.polys.polytools.GroebnerBasis
rank lexicographic (sympy.combinatorics.subsets.Subset
method), 1113
attribute), 254
reduce() (sympy.sets.sets.Complement static
method), 1323
rank nonlex() (sympy.combinatorics.permutations.Permutation
method), 209
reduce() (sympy.sets.sets.Intersection static
rank trotterjohnson()
method), 1321
(sympy.combinatorics.permutations.Permutation
reduce()
(sympy.sets.sets.Union
static
method), 210
method), 1320
rat clear denoms() (sympy.polys.polytools.Poly reduce abs inequalities()
(in
module
method), 1103
sympy.solvers.inequalities), 1506
ratint() (sympy.integrals static method), 611
reduce abs inequality()
(in
module
rational, 89
sympy.solvers.inequalities), 1506
Rational (class in sympy.core.numbers), 129
reduce element() (sympy.polys.agca.ideals.Ideal
RationalField (class in sympy.polys.domains),
method), 1140
1158
reduce element() (sympy.polys.agca.modules.SubModule
rationalize()
(in
module
method), 1137
sympy.parsing.sympy parser), 1592
reduce inequalities()
(in
module
ratsimp() (in module sympy.simplify.simplify),
sympy.solvers.inequalities), 1507
1334
reduce rational inequalities()
(in
module
ratsimp() (sympy.core.expr.Expr method),
sympy.solvers.inequalities), 1506
120
reduced() (in module sympy.polys.polytools),
rawlines() (in module sympy.utilities.misc),
1073
1576
ReferenceFrame
(class
in
sympy.physics.vector.frame), 1665
Ray (class in sympy.geometry.line), 519
Ray3D (class in sympy.geometry.line3d), 534 rene()
(in
module
Rayleigh() (in module sympy.stats), 1385
sympy.assumptions.rene), 1289
rayleigh2waist()
(in
module rene() (sympy.core.expr.Expr method), 120
sympy.physics.optics.gaussopt),
rene abs()
(in
module
1978
Index
SymPy Documentation, Release 0.7.6
sympy.assumptions.rene), 1290
1768
rene exp()
(in
module replace() (sympy.core.basic.Basic method),
sympy.assumptions.rene), 1291
97
rene Pow()
(in
module replace() (sympy.matrices.matrices.MatrixBase
sympy.assumptions.rene), 1290
method), 673
rene Relational()
(in
module replace()
(sympy.polys.polytools.Poly
sympy.assumptions.rene), 1290
method), 1104
rene root()
(in
module Reporter (class in sympy.utilities.runtests),
sympy.polys.polytools), 1071
1580
rene root() (sympy.polys.polyclasses.DMP represent()
(in
module
method), 1166
sympy.physics.quantum.represent),
rene root()
(sympy.polys.polytools.Poly
1767
method), 1103
reprify()
(sympy.printing.repr.ReprPrinter
RenementFailed
(class
in
method), 1263
ReprPrinter (class in sympy.printing.repr),
sympy.polys.polyerrors), 1238
reect()
(sympy.geometry.ellipse.Circle
1262
method), 556
reset() (sympy.combinatorics.polyhedron.Polyhedron
reect()
(sympy.geometry.ellipse.Ellipse
method), 245
method), 552
reshape()
(in
module
reect() (sympy.geometry.polygon.RegularPolygon
sympy.utilities.iterables), 1568
method), 569
reshape() (sympy.matrices.sparse.SparseMatrix
refraction angle()
(in
module
method), 697
sympy.physics.optics.utils), 1814
residue() (in module sympy.series.residues),
refractive index (sympy.physics.optics.medium.Medium1312
attribute), 1813
restrict codomain() (sympy.polys.agca.homomorphisms.Mod
method), 1147
register handler()
(in
module
sympy.assumptions.ask), 1287
restrict domain() (sympy.polys.agca.homomorphisms.Modul
RegularPolygon
(class
in
method), 1148
sympy.geometry.polygon), 563
Result (class in sympy.utilities.codegen),
Rel (in module sympy.core.relational), 147
1540
rem() (in module sympy.polys.polytools), 1060 result variables (sympy.utilities.codegen.Routine
rem() (sympy.polys.domains.domain.Domain
attribute), 1540
method), 1152
resultant() (in module sympy.polys.polytools),
rem()
(sympy.polys.domains.eld.Field
1062
method), 1153
resultant()
(sympy.polys.polyclasses.DMP
rem()
(sympy.polys.domains.ring.Ring
method), 1166
method), 1154
resultant()
(sympy.polys.polytools.Poly
rem()
(sympy.polys.polyclasses.DMP
method), 1104
method), 1166
retract()
(sympy.polys.polytools.Poly
rem() (sympy.polys.polytools.Poly method),
method), 1105
1104
reverse order() (sympy.concrete.products.Product
remove handler()
(in
module
method), 327
sympy.assumptions.ask), 1288
reverse order() (sympy.concrete.summations.Sum
removeO() (sympy.core.expr.Expr method),
method), 323
120
ReversedGradedLexOrder
(class
in
render() (sympy.printing.pretty.stringpict.stringPict
sympy.polys.orderings), 1119
method), 1268
revert() (sympy.polys.domains.domain.Domain
reorder()
(sympy.polys.polytools.Poly
method), 1152
method), 1104
revert()
(sympy.polys.domains.eld.Field
method), 1153
rep expectation()
(in
module
revert()
(sympy.polys.domains.ring.Ring
sympy.physics.quantum.represent),
1768
method), 1155
rep innerproduct()
(in
module revert()
(sympy.polys.polyclasses.DMP
sympy.physics.quantum.represent),
method), 1166
Index
1979
SymPy Documentation, Release 0.7.6
revert() (sympy.polys.polytools.Poly method),
1796
1105
RL (sympy.matrices.sparse.SparseMatrix atrewrite() (in module sympy.series.gruntz),
tribute), 692
1306
rmul() (sympy.combinatorics.permutations.Permutation
rewrite() (sympy.core.basic.Basic method),
static method), 210
98
rmul with af() (sympy.combinatorics.permutations.Permutat
rf() (in module mpmath), 784
static method), 211
rgamma() (in module mpmath), 781
root (in module sympy.printing.pretty.pretty symbology),
RGS (sympy.combinatorics.partitions.Partition
1267
attribute), 187
root() (in module mpmath), 753
RGS enum()
(in
module root()
(in
module
sympy.combinatorics.partitions),
sympy.functions.elementary.miscellaneous),
191
358
RGS generalized()
(in
module root() (sympy.polys.polytools.Poly method),
1105
sympy.combinatorics.partitions),
190
root() (sympy.printing.pretty.stringpict.stringPict
RGS rank()
(in
module
method), 1269
sympy.combinatorics.partitions),
RootOf (class in sympy.polys.rootoftools),
191
1119
RGS unrank()
(in
module roots() (in module sympy.polys.polyroots),
sympy.combinatorics.partitions),
1119
191
RootSum (class in sympy.polys.rootoftools),
rhs() (sympy.physics.mechanics.kane.KanesMethod
1119
method), 1743
rot axis1() (in module sympy.matrices.dense),
rhs() (sympy.physics.mechanics.lagrange.LagrangesMethod
684
method), 1746
rot axis2() (in module sympy.matrices.dense),
richardson() (in module mpmath), 991
685
richardson()
(in
module rot axis3() (in module sympy.matrices.dense),
sympy.series.acceleration), 1310
685
Ridder
(class
in
mp- rotate() (sympy.combinatorics.polyhedron.Polyhedron
math.calculus.optimization), 978
method), 245
riemann cyclic()
(in
module rotate()
(sympy.geometry.curve.Curve
sympy.tensor.tensor), 1532
method), 542
riemann cyclic replace()
(in
module rotate()
(sympy.geometry.ellipse.Ellipse
sympy.tensor.tensor), 1532
method), 552
riemannr() (in module mpmath), 965
rotate() (sympy.geometry.entity.GeometryEntity
method), 494
right (sympy.sets.sets.Interval attribute),
1318
rotate()
(sympy.geometry.point.Point
right() (sympy.printing.pretty.stringpict.stringPict
method), 502
method), 1268
rotate() (sympy.geometry.polygon.RegularPolygon
method), 569
right open
(sympy.sets.sets.Interval
attribute), 1319
rotate left()
(in
module
RigidBody
(class
in
sympy.utilities.iterables), 1568
(in
module
sympy.physics.mechanics.rigidbody), rotate right()
1733
sympy.utilities.iterables), 1569
Ring (class in sympy.polys.domains.ring), Rotation
(class
in
1153
sympy.physics.quantum.spin), 1771
ring() (in module sympy.polys.rings), 1213
rotation (sympy.geometry.polygon.RegularPolygon
RisingFactorial
(class
in
attribute), 569
sympy.functions.combinatorial.factorials),
round() (sympy.core.expr.Expr method), 120
374
RoundFunction
(class
in
Rk (in module sympy.physics.quantum.qft),
sympy.functions.elementary.integers),
1796
359
RkGate (class in sympy.physics.quantum.qft), Routine (class in sympy.utilities.codegen),
1980
Index
SymPy Documentation, Release 0.7.6
1539
rs swap() (in module sympy.stats.rv), 1397
routine() (sympy.utilities.codegen.CodeGen rs trunc()
(in
module
method), 1541
sympy.polys.ring series), 1244
routine() (sympy.utilities.codegen.OctaveCodeGen
rsa private key()
(in
module
method), 1544
sympy.crypto.crypto), 313
row() (sympy.matrices.sparse.SparseMatrix rsa public key()
(in
module
method), 697
sympy.crypto.crypto), 313
row del() (sympy.matrices.dense.MutableDenseMatrix
rsolve() (in module sympy.solvers.recurr),
method), 688
1479
row del() (sympy.matrices.sparse.MutableSparseMatrix
rsolve hyper()
(in
module
method), 701
sympy.solvers.recurr), 1481
row insert() (sympy.matrices.matrices.MatrixBase
rsolve poly()
(in
module
method), 673
sympy.solvers.recurr), 1480
row join() (sympy.matrices.matrices.MatrixBasersolve ratio()
(in
module
method), 673
sympy.solvers.recurr), 1480
row join() (sympy.matrices.sparse.MutableSparseMatrix
run() (sympy.utilities.runtests.SymPyDocTestRunner
method), 702
method), 1581
row list() (sympy.matrices.sparse.SparseMatrix run all tests()
(in
module
method), 697
sympy.utilities.runtests), 1582
row op() (sympy.matrices.dense.MutableDenseMatrix
run in subprocess with hash randomization()
method), 689
(in module sympy.utilities.runtests),
row op() (sympy.matrices.sparse.MutableSparseMatrix 1582
method), 702
runs() (in module sympy.utilities.iterables),
row structure symbolic cholesky()
1569
(sympy.matrices.sparse.SparseMatrix runs() (sympy.combinatorics.permutations.Permutation
method), 698
method), 211
row swap() (sympy.matrices.dense.MutableDenseMatrix
S
method), 689
row swap() (sympy.matrices.sparse.MutableSparseMatrix
S (in module sympy.core.singleton), 102
method), 703
S (in module sympy.physics.quantum.gate),
rref() (sympy.matrices.matrices.MatrixBase
1792
method), 674
sample() (in module sympy.stats), 1395
rs compose add()
(in
module sample iter() (in module sympy.stats), 1395
sympy.polys.ring series), 1247
satisable()
(in
module
rs exp() (in module sympy.polys.ring series),
sympy.logic.inference), 635
1246
saturate()
(sympy.polys.agca.ideals.Ideal
rs hadamard exp()
(in
module
method), 1140
sympy.polys.ring series), 1247
scalar multiply() (sympy.matrices.sparse.SparseMatrix
rs integrate()
(in
module
method), 698
sympy.polys.ring series), 1246
scalar potential()
(in
module
rs log() (in module sympy.polys.ring series),
sympy.physics.vector.eldfunctions),
1246
1691
rs mul() (in module sympy.polys.ring series), scalar potential dierence()
(in
module
1244
sympy.physics.vector.eldfunctions),
rs newton()
(in
module
1691
sympy.polys.ring series), 1247
scale()
(sympy.geometry.curve.Curve
rs pow() (in module sympy.polys.ring series),
method), 542
1245
scale()
(sympy.geometry.ellipse.Circle
rs series from list()
(in
module
method), 556
sympy.polys.ring series), 1245
scale()
(sympy.geometry.ellipse.Ellipse
rs series inversion()
(in
module
method), 552
sympy.polys.ring series), 1245
scale() (sympy.geometry.entity.GeometryEntity
rs square()
(in
module
method), 494
sympy.polys.ring series), 1244
Index
1981
SymPy Documentation, Release 0.7.6
scale() (sympy.geometry.point.Point method), sdm mul term()
(in
module
502
sympy.polys.distributedmodules),
scale()
(sympy.geometry.point3d.Point3D
1224
method), 507
sdm nf mora()
(in
module
scale() (sympy.geometry.polygon.RegularPolygon
sympy.polys.distributedmodules),
method), 569
1237
schreier sims() (sympy.combinatorics.perm groups.PermutationGroup
sdm spoly()
(in
module
method), 238
sympy.polys.distributedmodules),
schreier sims incremental()
1236
(sympy.combinatorics.perm groups.PermutationGroup
sdm to dict()
(in
module
method), 238
sympy.polys.distributedmodules),
schreier sims random()
1223
(sympy.combinatorics.perm groups.PermutationGroup
sdm to vector()
(in
module
method), 239
sympy.polys.distributedmodules),
schreier vector() (sympy.combinatorics.perm groups.PermutationGroup
1225
sdm zero()
(in
module
method), 240
scorergi() (in module mpmath), 851
sympy.polys.distributedmodules),
scorerhi() (in module mpmath), 853
1224
sdm add()
(in
module search()
(sympy.ntheory.generate.Sieve
sympy.polys.distributedmodules),
method), 277
1223
search function (sympy.physics.quantum.grover.OracleGate
sdm deg()
(in
module
attribute), 1794
sympy.polys.distributedmodules),
sec() (in module mpmath), 769
1224
Secant
(class
in
mpmath.calculus.optimization), 976
sdm ecart()
(in
module
sympy.polys.distributedmodules),
sech() (in module mpmath), 776
1237
secondzeta() (in module mpmath), 952
sdm from dict()
(in
module Segment (class in sympy.geometry.line), 522
sympy.polys.distributedmodules),
Segment3D (class in sympy.geometry.line3d),
1223
537
sdm from vector()
(in
module select()
(sympy.simplify.epathtools.EPath
sympy.polys.distributedmodules),
method), 1348
1225
selections (sympy.combinatorics.graycode.GrayCode
sdm groebner()
(in
module
attribute), 258
sympy.polys.distributedmodules),
separate() (sympy.core.expr.Expr method),
1237
121
sdm LC()
(in
module separate() (sympy.physics.vector.vector.Vector
sympy.polys.distributedmodules),
method), 1671
1222
separatevars()
(in
module
sdm LM()
(in
module
sympy.simplify.simplify), 1331
sympy.polys.distributedmodules),
series() (in module sympy.series.series), 1308
1223
series() (sympy.core.expr.Expr method), 121
sdm LT()
(in
module Set (class in sympy.sets.sets), 1312
set acc()
(sympy.physics.vector.point.Point
sympy.polys.distributedmodules),
1224
method), 1677
sdm monomial deg()
(in
module set ang acc() (sympy.physics.vector.frame.ReferenceFrame
sympy.polys.distributedmodules),
method), 1668
1222
set ang vel() (sympy.physics.vector.frame.ReferenceFrame
sdm monomial divides()
(in
module
method), 1668
sympy.polys.distributedmodules),
set comm() (sympy.tensor.tensor. TensorManager
1222
method), 1518
sdm monomial mul()
(in
module set comms() (sympy.tensor.tensor. TensorManager
sympy.polys.distributedmodules),
method), 1518
1221
set domain()
(sympy.polys.polytools.Poly
1982
Index
SymPy Documentation, Release 0.7.6
method), 1106
method), 211
set global settings()
SimpleDomain
(class
in
(sympy.printing.printer.Printer class
sympy.polys.domains.simpledomain),
method), 1251
1155
set modulus()
(sympy.polys.polytools.Poly simplied()
(in
module
method), 1106
sympy.solvers.diophantine), 1502
set pos()
(sympy.physics.vector.point.Point simplify() (in module sympy.simplify.simplify),
1327
method), 1678
set potential energy()
simplify() (sympy.core.expr.Expr method),
(sympy.physics.mechanics.particle.Particle
122
method), 1732
simplify() (sympy.functions.special.delta functions.DiracDelt
set potential energy()
method), 378
(sympy.physics.mechanics.rigidbody.RigidBody
simplify() (sympy.matrices.dense.MutableDenseMatrix
method), 1736
method), 689
set vel()
(sympy.physics.vector.point.Point simplify() (sympy.matrices.matrices.MatrixBase
method), 1678
method), 674
seterr() (in module sympy.core.numbers), 132 simplify() (sympy.physics.vector.dyadic.Dyadic
setup() (sympy.galgebra.vector.Vector static
method), 1674
method), 484
simplify() (sympy.physics.vector.vector.Vector
shanks() (in module mpmath), 992
method), 1672
shanks()
(in
module simplify gpgp() (sympy.physics.hep.gamma matrices.Gamma
sympy.series.acceleration), 1311
static method), 1633
shape (sympy.matrices.matrices.MatrixBase simplify index permutations() (in module
attribute), 674
sympy.physics.secondquant), 1623
shape
(sympy.tensor.indexed.Indexed
at- simplify lines() (sympy.physics.hep.gamma matrices.Gamma
tribute), 1512
static method), 1633
shape (sympy.tensor.indexed.IndexedBase at- simplify logic()
(in
module
tribute), 1514
sympy.logic.boolalg), 633
ShapeError
(class
in sin (class in sympy.functions.elementary.trigonometric),
sympy.matrices.matrices), 678
359
Shi (class in sympy.functions.special.error functions),
sin() (in module mpmath), 768
406
sinc() (in module mpmath), 773
shi() (in module mpmath), 803
sincpi() (in module mpmath), 774
shift()
(sympy.polys.polyclasses.DMP sine transform()
(in
module
method), 1166
sympy.integrals.transforms), 587
shift() (sympy.polys.polytools.Poly method), SingleDomain (class in sympy.stats.rv), 1396
1106
SinglePSpace (class in sympy.stats.rv), 1396
shor()
(in
module singular values() (sympy.matrices.matrices.MatrixBase
sympy.physics.quantum.shor), 1801
method), 674
Si (class in sympy.functions.special.error functions),
singularities()
(in
module
sympy.calculus.singularities), 1593
404
si() (in module mpmath), 802
sinh (class in sympy.functions.elementary.hyperbolic),
sides (sympy.geometry.polygon.Polygon at360
sinh() (in module mpmath), 775
tribute), 562
siegeltheta() (in module mpmath), 941
sinm() (in module mpmath), 1030
siegelz() (in module mpmath), 939
sinpi() (in module mpmath), 770
Sieve (class in sympy.ntheory.generate), 276 size (sympy.combinatorics.permutations.Permutation
sift() (in module sympy.utilities.iterables),
attribute), 211
1569
size (sympy.combinatorics.polyhedron.Polyhedron
sign (class in sympy.functions.elementary.complexes), attribute), 246
361
size (sympy.combinatorics.prufer.Prufer atsign() (in module mpmath), 734
tribute), 248
sign() (in module sympy.series.gruntz), 1306 size (sympy.combinatorics.subsets.Subset atsignature() (sympy.combinatorics.permutations.Permutation
tribute), 254
Index
1983
SymPy Documentation, Release 0.7.6
SKIP()
(in module sympy.utilities.pytest), sort key() (sympy.core.basic.Basic method),
1578
99
skip() (sympy.combinatorics.graycode.GrayCodesorted components()
method), 258
(sympy.tensor.tensor.TensMul
slice()
(sympy.polys.polyclasses.DMP
method), 1531
method), 1166
source (sympy.geometry.line.Ray attribute),
slice() (sympy.polys.polytools.Poly method),
521
1106
source (sympy.geometry.line3d.Ray3D atslope (sympy.geometry.line.LinearEntity attribute), 535
tribute), 516
source() (in module sympy.utilities.source),
smoothness()
(in
module
1587
sympy.ntheory.factor ), 281
SparseMatrix
(class
in
sympy.matrices.sparse), 692
smoothness p()
(in
module
sympy.ntheory.factor ), 282
speed (sympy.physics.optics.medium.Medium
solve() (in module sympy.solvers.solvers),
attribute), 1813
1469
speed (sympy.physics.optics.waves.TWave atsolve() (sympy.matrices.matrices.MatrixBase
tribute), 1818
method), 675
spherharm() (in module mpmath), 882
solve() (sympy.matrices.sparse.SparseMatrix spin() (sympy.geometry.polygon.RegularPolygon
method), 698
method), 569
solve congruence()
(in
module split() (sympy.tensor.tensor.TensMul method),
sympy.ntheory.modular), 292
1531
solve least squares()
split list()
(in
module
(sympy.matrices.matrices.MatrixBase
sympy.utilities.runtests), 1583
method), 675
split super sub()
(in
module
sympy.printing.conventions), 1266
solve least squares()
(sympy.matrices.sparse.SparseMatrix split symbols()
(in
module
method), 698
sympy.parsing.sympy parser), 1590
solve linear()
(in
module split symbols custom()
(in
module
sympy.solvers.solvers), 1474
sympy.parsing.sympy parser), 1590
solve linear system()
(in
module splot() (in module mpmath), 749
sympy.solvers.solvers), 1475
spoly()
(in
module
solve linear system LU()
(in
module
sympy.polys.groebnertools), 1236
sympy.solvers.solvers), 1476
sqf() (in module sympy.polys.polytools), 1070
solve multipliers() (sympy.physics.mechanics.lagrange.LagrangesMethod
sqf list() (in module sympy.polys.polytools),
method), 1746
1069
solve poly inequality()
(in
module sqf list()
(sympy.polys.polyclasses.DMP
sympy.solvers.inequalities), 1505
method), 1166
solve poly system()
(in
module sqf list() (sympy.polys.polytools.Poly method),
sympy.solvers.polysys), 1482
1106
solve rational inequalities()
(in
module sqf list include() (sympy.polys.polyclasses.DMP
sympy.solvers.inequalities), 1505
method), 1166
solve triangulated()
(in
module sqf list include() (sympy.polys.polytools.Poly
sympy.solvers.polysys), 1482
method), 1106
solve undetermined coes()
(in
module sqf norm() (in module sympy.polys.polytools),
1069
sympy.solvers.solvers), 1476
solve univariate inequality()
(in
module sqf norm()
(sympy.polys.polyclasses.DMP
sympy.solvers.inequalities), 1507
method), 1166
SOPform() (in module sympy.logic.boolalg), sqf norm()
(sympy.polys.polytools.Poly
method), 1107
625
sort dims() (sympy.physics.unitsystems.dimensions.DimensionSystem
sqf part() (in module sympy.polys.polytools),
static method), 1827
1069
sort key() (sympy.combinatorics.partitions.Partition
sqf part()
(sympy.polys.polyclasses.DMP
method), 188
method), 1167
1984
Index
SymPy Documentation, Release 0.7.6
sqf part()
(sympy.polys.polytools.Poly str array()
(in
module
method), 1107
sympy.galgebra.stringarrays), 488
sqr() (sympy.polys.polyclasses.DMP method), str combinations()
(in
module
1167
sympy.galgebra.stringarrays), 489
sqr() (sympy.polys.polytools.Poly method), StrictGreaterThan
(class
in
1108
sympy.core.relational), 156
sqrt (class in sympy.functions.elementary.miscellaneous),
StrictLessThan
(class
in
360
sympy.core.relational), 159
sqrt() (in module mpmath), 752
stringify expr()
(in
module
sqrt() (sympy.polys.domains.domain.Domain
sympy.parsing.sympy parser), 1589
method), 1152
stringPict
(class
in
sqrt mod()
(in
module
sympy.printing.pretty.stringpict),
sympy.ntheory.residue ntheory),
1268
296
strip zero() (sympy.polys.rings.PolyElement
sqrtdenest()
(in
module
method), 1221
strong gens (sympy.combinatorics.perm groups.Permutation
sympy.simplify.sqrtdenest), 1344
sqrtm() (in module mpmath), 1030
attribute), 241
square()
(sympy.polys.rings.PolyElement StrPrinter (class in sympy.printing.str), 1263
method), 1220
struveh() (in module mpmath), 832
square factor()
(in
module struvel() (in module mpmath), 833
sympy.solvers.diophantine), 1496
StudentT() (in module sympy.stats), 1385
srepr() (in module sympy.printing.repr), 1263 sturm() (in module sympy.polys.polytools),
sring() (in module sympy.polys.rings), 1215
1068
sstrrepr() (in module sympy.printing.str), sturm()
(sympy.polys.polyclasses.DMP
method), 1167
1263
stabilizer() (sympy.combinatorics.perm groups.PermutationGroup
sturm() (sympy.polys.polytools.Poly method),
method), 241
1108
stack() (sympy.printing.pretty.stringpict.stringPict
sub (in module sympy.printing.pretty.pretty symbology),
static method), 1269
1267
standard transformations
(in
module sub() (sympy.polys.domains.domain.Domain
sympy.parsing.sympy parser), 1590
method), 1152
start (sympy.sets.sets.Interval attribute), sub() (sympy.polys.polyclasses.DMF method),
1319
1168
State (class in sympy.physics.quantum.state), sub() (sympy.polys.polyclasses.DMP method),
1781
1167
state() (sympy.physics.secondquant.FixedBosonicBasis
sub() (sympy.polys.polytools.Poly method),
method), 1614
1108
state() (sympy.physics.secondquant.VarBosonicBasis
sub ground() (sympy.polys.polyclasses.DMP
method), 1614
method), 1167
state to operators()
(in
module sub ground()
(sympy.polys.polytools.Poly
sympy.physics.quantum.operatorset),
method), 1108
1765
sub paren()
(in
module
StateBase
(class
in
sympy.galgebra.precedence), 485
subdiagram from objects()
sympy.physics.quantum.state), 1781
std() (in module sympy.stats), 1395
(sympy.categories.Diagram method),
stieltjes() (in module mpmath), 937
1837
stirling()
(in
module subgroup search() (sympy.combinatorics.perm groups.Perm
sympy.functions.combinatorial.numbers),
method), 241
374
SubModule
(class
in
sympy.polys.agca.modules), 1134
stirling1() (in module mpmath), 962
stirling2() (in module mpmath), 963
submodule() (sympy.polys.agca.modules.Module
StopTokenizing
(class
in
method), 1132
submodule() (sympy.polys.agca.modules.QuotientModule
sympy.parsing.sympy tokenize),
1590
method), 1142
Index
1985
SymPy Documentation, Release 0.7.6
submodule() (sympy.polys.agca.modules.SubModule
sumem() (in module mpmath), 986
method), 1137
summation()
(in
module
SubQuotientModule
(class
in
sympy.concrete.summations), 328
sympy.polys.agca.modules), 1142
summation() (mpmath.calculus.quadrature.QuadratureRule
subresultants()
(in
module
method), 1011
sympy.polys.polytools), 1061
sup (in module sympy.printing.pretty.pretty symbology),
subresultants() (sympy.polys.polyclasses.DMP
1267
sup (sympy.sets.sets.Set attribute), 1316
method), 1167
subresultants()
(sympy.polys.polytools.Poly superfac() (in module mpmath), 787
method), 1109
superposition basis()
(in
module
Subs (class in sympy.core.function), 171
sympy.physics.quantum.grover),
subs() (sympy.core.basic.Basic method), 99
1794
subs() (sympy.matrices.immutable.ImmutableSparseMatrix
superset (sympy.combinatorics.subsets.Subset
method), 704
attribute), 255
subs() (sympy.matrices.matrices.MatrixBase superset size (sympy.combinatorics.subsets.Subset
method), 676
attribute), 255
subs() (sympy.physics.vector.dyadic.Dyadic support() (sympy.combinatorics.permutations.Permutation
method), 1674
method), 212
subs()
(sympy.physics.vector.vector.Vector SurfaceBaseSeries
(class
in
method), 1672
sympy.plotting.plot), 1283
Subset (class in sympy.combinatorics.subsets), SurfaceOver2DRangeSeries
(class
in
250
sympy.plotting.plot), 1283
subset (sympy.combinatorics.subsets.Subset SWAP
(in
module
attribute), 254
sympy.physics.quantum.gate), 1792
subset()
(sympy.polys.agca.ideals.Ideal SwapGate
(class
in
method), 1140
sympy.physics.quantum.gate), 1791
subset() (sympy.polys.agca.modules.Module swinnerton dyer poly()
(in
module
method), 1132
sympy.polys.specialpolys), 1120
subset() (sympy.sets.sets.Set method), 1316
symarray()
(in
module
sympy.matrices.dense), 683
subset from bitlist()
(sympy.combinatorics.subsets.Subset symb 2txt
(in
module
class method), 254
sympy.printing.pretty.pretty symbology),
subset indices() (sympy.combinatorics.subsets.Subset 1267
class method), 255
Symbol (class in sympy.core.symbol), 122
subsets()
(in
module symbol (sympy.physics.secondquant.AntiSymmetricTensor
sympy.utilities.iterables), 1570
attribute), 1621
SubsSet (class in sympy.series.gruntz), 1307 Symbol() (sympy.assumptions.handlers.calculus.AskBounded
substitute dummies()
(in
module
static method), 1292
sympy.physics.secondquant), 1622
symbol array()
(in
module
substitute indices() (sympy.tensor.tensor.TensAdd
sympy.galgebra.stringarrays), 489
method), 1528
symbols() (in module sympy.core.symbol),
substitute indices() (sympy.tensor.tensor.TensMul
124
method), 1531
symmetric() (sympy.combinatorics.generators
Sum (class in sympy.concrete.summations),
static method), 215
321
symmetric poly()
(in
module
sum next() (mpmath.calculus.quadrature.QuadratureRule
sympy.polys.specialpolys), 1120
method), 1010
symmetric residue()
(in
module
sum next() (mpmath.calculus.quadrature.TanhSinh
sympy.ntheory.modular), 290
method), 1012
SymmetricGroup()
(in
module
sympy.combinatorics.named groups),
sum of four squares()
(in
module
261
sympy.solvers.diophantine), 1499
sum of three squares()
(in
module symmetrize()
(in
module
sympy.solvers.diophantine), 1499
sympy.polys.polyfuncs), 1113
sumap() (in module mpmath), 987
sympify() (in module sympy.core.sympify), 85
1986
Index
SymPy Documentation, Release 0.7.6
sympy (module), 704
sympy.assumptions (module), 1287
sympy.assumptions.ask (module), 1287
sympy.assumptions.assume (module), 1288
sympy.assumptions.handlers (module), 1291
sympy.assumptions.handlers.calculus (module), 1291
sympy.assumptions.handlers.ntheory (module), 1293
sympy.assumptions.handlers.order (module),
1293
sympy.assumptions.handlers.sets (module),
1293
sympy.assumptions.rene (module), 1289
sympy.calculus (module), 1592
sympy.calculus.euler (module), 1592
sympy.calculus.nite di (module), 1593
sympy.calculus.singularities (module), 1593
sympy.categories (module), 1829
sympy.categories.diagram drawing (module),
1838
sympy.combinatorics.generators
(module),
215
sympy.combinatorics.graycode (module), 256
sympy.combinatorics.group constructs (module), 269
sympy.combinatorics.named groups
(module), 261
sympy.combinatorics.partitions
(module),
187
sympy.combinatorics.perm groups (module),
216
sympy.combinatorics.permutations (module),
192
sympy.combinatorics.polyhedron (module),
243
sympy.combinatorics.prufer (module), 246
sympy.combinatorics.subsets (module), 250
sympy.combinatorics.tensor can
(module),
270
sympy.combinatorics.testutil (module), 269
sympy.combinatorics.util (module), 263
sympy.core.add (module), 144
sympy.core.assumptions (module), 88
sympy.core.basic (module), 90
sympy.core.cache (module), 90
sympy.core.compatibility (module), 183
sympy.core.containers (module), 182
sympy.core.evalf (module), 181
sympy.core.expr (module), 103
sympy.core.exprtools (module), 185
sympy.core.function (module), 163
sympy.core.mod (module), 147
sympy.core.mul (module), 142
Index
sympy.core.multidimensional (module), 162
sympy.core.numbers (module), 126
sympy.core.power (module), 139
sympy.core.relational (module), 147
sympy.core.singleton (module), 102
sympy.core.symbol (module), 122
sympy.core.sympify (module), 85
sympy.crypto.crypto (module), 303
sympy.digeom (module), 1846
sympy.functions (module), 340
sympy.functions.special.bessel (module), 408
sympy.functions.special.beta functions (module), 386
sympy.functions.special.elliptic integrals
(module), 430
sympy.functions.special.error functions
(module), 388
sympy.functions.special.gamma functions
(module), 379
sympy.functions.special.polynomials
(module), 433
sympy.functions.special.zeta functions (module), 421
sympy.galgebra (module), 450
sympy.galgebra.debug (module), 489
sympy.galgebra.ga (module), 451
sympy.galgebra.ga.Com()
(in
module
sympy.galgebra.ga), 467
sympy.galgebra.ga.cross()
(in
module
sympy.galgebra.ga), 468
sympy.galgebra.ga.DD()
(in
module
sympy.galgebra.ga), 467
sympy.galgebra.ga.dual()
(in
module
sympy.galgebra.ga), 468
sympy.galgebra.ga.Fmt()
(in
module
sympy.galgebra.ga), 471
sympy.galgebra.ga.Format()
(in
module
sympy.galgebra.ga), 467
sympy.galgebra.ga.ga print o() (in module
sympy.galgebra.ga), 471
sympy.galgebra.ga.ga print on() (in module
sympy.galgebra.ga), 471
sympy.galgebra.ga.inv()
(in
module
sympy.galgebra.ga), 468
sympy.galgebra.ga.MV
(class
in
sympy.galgebra.ga), 465
sympy.galgebra.ga.MV.convert from blades()
(in module sympy.galgebra.ga), 466
sympy.galgebra.ga.MV.convert to blades() (in
module sympy.galgebra.ga), 466
sympy.galgebra.ga.MV.dd()
(in
module
sympy.galgebra.ga), 466
sympy.galgebra.ga.MV.di()
(in
module
sympy.galgebra.ga), 466
1987
SymPy Documentation, Release 0.7.6
sympy.galgebra.ga.MV.dual()
(in
module sympy.galgebra.printing.Format() (in module
sympy.galgebra.ga), 466
sympy.galgebra.ga), 471
sympy.galgebra.ga.MV.even()
(in
module sympy.galgebra.printing.GA Printer()
(in
sympy.galgebra.ga), 466
module sympy.galgebra.ga), 471
sympy.galgebra.ga.MV.exp()
(in
module sympy.galgebra.printing.Get Program()
(in
sympy.galgebra.ga), 466
module sympy.galgebra.ga), 474
sympy.galgebra.ga.MV.expand() (in module sympy.galgebra.printing.Print Function() (in
sympy.galgebra.ga), 466
module sympy.galgebra.ga), 474
sympy.galgebra.ga.MV.factor() (in module sympy.galgebra.printing.xdvi() (in module
sympy.galgebra.ga), 466
sympy.galgebra.ga), 471
sympy.galgebra.ga.MV.func()
(in
module sympy.galgebra.stringarrays (module), 488
sympy.galgebra.ga), 466
sympy.galgebra.vector (module), 484
sympy.galgebra.ga.MV.grade() (in module sympy.geometry.curve (module), 539
sympy.galgebra.ga), 467
sympy.geometry.ellipse (module), 542
sympy.galgebra.ga.MV.inv()
(in
module sympy.geometry.entity (module), 493
sympy.geometry.line (module), 509
sympy.galgebra.ga), 467
sympy.galgebra.ga.MV.norm() (in module sympy.geometry.line3d (module), 525
sympy.galgebra.ga), 467
sympy.geometry.plane (module), 577
sympy.galgebra.ga.MV.rev()
(in
module sympy.geometry.point (module), 497
sympy.galgebra.ga), 467
sympy.geometry.point3d (module), 503
sympy.galgebra.ga.MV.scalar() (in module sympy.geometry.polygon (module), 556
sympy.galgebra.ga), 467
sympy.geometry.util (module), 495
sympy.galgebra.ga.MV.set coef() (in module sympy.integrals (module), 583
sympy.galgebra.ga), 467
sympy.integrals.meijerint doc (module), 604
sympy.galgebra.ga.MV.setup() (in module sympy.integrals.transforms (module), 584
sympy.galgebra.ga), 463
sympy.logic (module), 625
sympy.galgebra.ga.MV.simplify() (in module sympy.matrices (module), 635
sympy.galgebra.ga), 467
sympy.matrices.expressions (module), 706
sympy.galgebra.ga.MV.subs()
(in
module sympy.matrices.expressions.blockmatrix
sympy.galgebra.ga), 467
(module), 711
sympy.galgebra.ga.MV.trigsimp() (in module sympy.matrices.immutable (module), 705
sympy.galgebra.ga), 467
sympy.matrices.matrices (module), 636
sympy.galgebra.ga.Nga()
(in
module sympy.matrices.sparse (module), 692
sympy.galgebra.ga), 468
sympy.ntheory.continued fraction (module),
sympy.galgebra.ga.proj()
(in
module
298
sympy.galgebra.ga), 468
sympy.ntheory.egyptian fraction
(module),
sympy.galgebra.ga.ReciprocalFrame()
(in
301
module sympy.galgebra.ga), 468
sympy.ntheory.factor (module), 281
sympy.galgebra.ga.re()
(in
module sympy.ntheory.generate (module), 276
sympy.galgebra.ga), 468
sympy.ntheory.modular (module), 290
sympy.galgebra.ga.rot()
(in
module sympy.ntheory.multinomial (module), 292
sympy.galgebra.ga), 468
sympy.ntheory.partitions (module), 294
sympy.galgebra.ga.ScalarFunction() (in mod- sympy.ntheory.primetest (module), 294
sympy.ntheory.residue ntheory (module), 295
ule sympy.galgebra.ga), 468
sympy.galgebra.manifold (module), 483
sympy.physics (module), 1597
sympy.galgebra.ncutil (module), 486
sympy.physics.hep.gamma matrices
(modsympy.galgebra.precedence (module), 484
ule), 1632
sympy.galgebra.precedence.dene precedence()
sympy.physics.hydrogen (module), 1598
(in module sympy.galgebra.ga), 454, sympy.physics.matrices (module), 1600
468
sympy.physics.mechanics.kane
(module),
sympy.galgebra.precedence.GAeval()
(in
1741
module sympy.galgebra.ga),
454, sympy.physics.mechanics.lagrange (module),
468
1744
sympy.galgebra.printing (module), 485
1988
Index
SymPy Documentation, Release 0.7.6
sympy.physics.mechanics.linearize (module),
1746
sympy.physics.mechanics.particle (module),
1730
sympy.physics.mechanics.rigidbody
(module), 1733
sympy.physics.optics.gaussopt
(module),
1802
sympy.physics.optics.medium (module), 1812
sympy.physics.optics.utils (module), 1813
sympy.physics.optics.waves (module), 1817
sympy.physics.paulialgebra (module), 1601
sympy.physics.qho 1d (module), 1602
sympy.physics.quantum.anticommutator
(module), 1749
sympy.physics.quantum.cartesian (module),
1757
sympy.physics.quantum.cg (module), 1750
sympy.physics.quantum.circuitplot (module),
1788
sympy.physics.quantum.commutator
(module), 1752
sympy.physics.quantum.constants (module),
1753
sympy.physics.quantum.dagger
(module),
1753
sympy.physics.quantum.gate (module), 1789
sympy.physics.quantum.grover
(module),
1794
sympy.physics.quantum.hilbert
(module),
1758
sympy.physics.quantum.innerproduct (module), 1754
sympy.physics.quantum.operator (module),
1760
sympy.physics.quantum.operatorset
(module), 1764
sympy.physics.quantum.piab (module), 1801
sympy.physics.quantum.qapply
(module),
1766
sympy.physics.quantum.qft (module), 1796
sympy.physics.quantum.qubit (module), 1796
sympy.physics.quantum.represent (module),
1767
sympy.physics.quantum.shor (module), 1800
sympy.physics.quantum.spin (module), 1771
sympy.physics.quantum.state (module), 1781
sympy.physics.quantum.tensorproduct (module), 1755
sympy.physics.secondquant (module), 1604
sympy.physics.sho (module), 1603
sympy.physics.units (module), 1630
sympy.physics.unitsystems.dimensions (module), 1824
Index
sympy.physics.unitsystems.prexes (module),
1827
sympy.physics.unitsystems.quantities (module), 1829
sympy.physics.unitsystems.units
(module),
1827
sympy.physics.vector.functions
(module),
1680
sympy.physics.vector.point (module), 1675
sympy.physics.wigner (module), 1623
sympy.plotting.plot (module), 1270
sympy.plotting.pygletplot (module), 1283
sympy.printing.ccode (module), 1252
sympy.printing.codeprinter (module), 1266
sympy.printing.conventions (module), 1266
sympy.printing.fcode (module), 1254
sympy.printing.gtk (module), 1259
sympy.printing.lambdarepr (module), 1259
sympy.printing.latex (module), 1259
sympy.printing.mathematica (module), 1258
sympy.printing.mathml (module), 1262
sympy.printing.precedence (module), 1266
sympy.printing.pretty.pretty (module), 1251
sympy.printing.pretty.pretty symbology
(module), 1266
sympy.printing.pretty.stringpict
(module),
1268
sympy.printing.preview (module), 1264
sympy.printing.printer (module), 1248
sympy.printing.python (module), 1262
sympy.printing.repr (module), 1262
sympy.printing.str (module), 1263
sympy.printing.tree (module), 1263
sympy.series (module), 1304
sympy.sets.fancysets (module), 1324
sympy.sets.sets (module), 1312
sympy.simplify.cse main (module),
1303,
1345
sympy.simplify.epathtools (module), 1347
sympy.simplify.hyperexpand (module), 1347
sympy.simplify.hyperexpand doc
(module),
1358
sympy.simplify.simplify (module), 73
sympy.simplify.sqrtdenest (module), 1344
sympy.simplify.traversaltools (module), 1347
sympy.solvers (module), 1468
sympy.solvers.inequalities (module), 1505
sympy.solvers.ode (module), 1457
sympy.solvers.pde (module), 1468
sympy.solvers.recurr (module), 1479
sympy.stats (module), 1361
sympy.stats.crv (module), 1396
sympy.stats.crv types (module), 1396
1989
SymPy Documentation, Release 0.7.6
sympy.stats.Die()
(in
module take() (in module sympy.utilities.iterables),
sympy.stats.crv types), 1397
1570
sympy.stats.frv (module), 1396
tan (class in sympy.functions.elementary.trigonometric),
sympy.stats.frv types (module), 1396
362
sympy.stats.Normal()
(in
module tan() (in module mpmath), 768
sympy.stats.crv types), 1397
tangent lines() (sympy.geometry.ellipse.Ellipse
sympy.stats.rv (module), 1396
method), 553
sympy.tensor (module), 1508
tanh (class in sympy.functions.elementary.hyperbolic),
sympy.tensor.index methods (module), 1514
363
sympy.tensor.indexed (module), 1508
tanh() (in module mpmath), 775
sympy.tensor.tensor (module), 1517
TanhSinh
(class
in
mpsympy.utilities (module), 1532
math.calculus.quadrature), 1011
sympy.utilities.autowrap (module), 1533
targets (sympy.physics.quantum.gate.CGate
sympy.utilities.codegen (module), 1538
attribute), 1790
sympy.utilities.decorator (module), 1547
targets (sympy.physics.quantum.gate.CNotGate
sympy.utilities.enumerative (module), 1549
attribute), 1792
sympy.utilities.iterables (module), 1555
targets
(sympy.physics.quantum.gate.Gate
sympy.utilities.lambdify (module), 1573
attribute), 1789
sympy.utilities.memoization (module), 1576
targets (sympy.physics.quantum.gate.UGate
sympy.utilities.misc (module), 1576
attribute), 1790
sympy.utilities.pkgdata (module), 1577
targets (sympy.physics.quantum.grover.OracleGate
sympy.utilities.pytest (module), 1578
attribute), 1794
sympy.utilities.randtest (module), 1579
taufrom() (in module mpmath), 911
sympy.utilities.runtests (module), 1580
taylor() (in module mpmath), 1015
sympy.utilities.source (module), 1587
taylor term() (sympy.core.expr.Expr method),
sympy.utilities.timeutils (module), 1587
122
SymPyDocTestFinder
(class
in taylor term() (sympy.functions.elementary.exponential.exp
sympy.utilities.runtests), 1580
static method), 351
SymPyDocTestRunner
(class
in taylor term() (sympy.functions.elementary.exponential.log
sympy.utilities.runtests), 1581
static method), 354
SymPyOutputChecker
(class
in taylor term() (sympy.functions.elementary.hyperbolic.sinh
sympy.utilities.runtests), 1581
static method), 360
sympytestle()
(in
module TC() (sympy.polys.polyclasses.ANP method),
sympy.utilities.runtests), 1584
1168
SymPyTestResults
(in
module TC() (sympy.polys.polyclasses.DMP method),
sympy.utilities.runtests), 1581
1162
syzygy module() (sympy.polys.agca.modules.SubModule
TC() (sympy.polys.polytools.Poly method),
method), 1137
1076
TensAdd (class in sympy.tensor.tensor), 1526
T
TensExpr (class in sympy.tensor.tensor), 1525
T (in module sympy.physics.quantum.gate), TensMul (class in sympy.tensor.tensor), 1529
tensor indices()
(in
module
1792
sympy.tensor.tensor), 1521
T
(sympy.matrices.expressions.MatrixExpr
tensor mul() (in module sympy.tensor.tensor),
attribute), 707
1531
T (sympy.matrices.matrices.MatrixBase attensor
product
simp()
(in
module
tribute), 648
sympy.physics.quantum.tensorproduct),
t
(sympy.physics.quantum.shor.CMod
at1756
tribute), 1801
TensorHead
(class in sympy.tensor.tensor),
table() (sympy.matrices.matrices.MatrixBase
1523
method), 676
tail degree() (sympy.polys.rings.PolyElement TensorIndex (class in sympy.tensor.tensor),
1520
method), 1221
TensorIndexType
(class
in
tail degrees() (sympy.polys.rings.PolyElement
sympy.tensor.tensor),
1518
method), 1221
1990
Index
SymPy Documentation, Release 0.7.6
TensorProduct (class in sympy.digeom),
sympy.physics.quantum.state), 1783
1853
timing() (in module mpmath), 746
TensorProduct
(class
in to algebraic integer()
sympy.physics.quantum.tensorproduct),
(sympy.polys.numberelds.AlgebraicNumber
1755
method), 1117
TensorSymmetry
(class
in to cnf() (in module sympy.logic.boolalg), 632
sympy.tensor.tensor), 1521
to dict()
(sympy.polys.polyclasses.ANP
tensorsymmetry()
(in
module
method), 1168
sympy.tensor.tensor), 1521
to dict()
(sympy.polys.polyclasses.DMP
TensorType (class in sympy.tensor.tensor),
method), 1167
1522
to dnf() (in module sympy.logic.boolalg), 632
terminal width() (sympy.printing.pretty.stringpict.stringPict
to exact()
(sympy.polys.polyclasses.DMP
method), 1269
method), 1167
terms()
(sympy.polys.polyclasses.DMP to exact()
(sympy.polys.polytools.Poly
method), 1167
method), 1110
terms() (sympy.polys.polytools.Poly method), to eld()
(sympy.polys.polyclasses.DMP
1109
method), 1167
terms()
(sympy.polys.rings.PolyElement to eld()
(sympy.polys.polytools.Poly
method), 1221
method), 1110
terms gcd()
(in
module to linearizer() (sympy.physics.mechanics.kane.KanesMethod
sympy.polys.polytools), 1065
method), 1744
terms gcd()
(sympy.polys.polyclasses.DMP to linearizer() (sympy.physics.mechanics.lagrange.Lagrange
method), 1167
method), 1746
terms gcd()
(sympy.polys.polytools.Poly to list()
(sympy.polys.polyclasses.ANP
method), 1109
method), 1168
termwise()
(sympy.polys.polytools.Poly to matrix() (sympy.physics.vector.dyadic.Dyadic
method), 1109
method), 1675
test() (in module sympy.utilities.runtests), to matrix() (sympy.physics.vector.vector.Vector
1585
method), 1672
test derivative numerically()
(in
module to number eld()
(in
module
sympy.utilities.randtest), 1579
sympy.polys.numberelds), 1117
TGate (class in sympy.physics.quantum.gate), to prufer() (sympy.combinatorics.prufer.Prufer
1791
static method), 249
ThinLens
(class
in to rational() (sympy.polys.domains.RealField
sympy.physics.optics.gaussopt),
method), 1160
1805
to ring()
(sympy.polys.polyclasses.DMP
method), 1167
threaded()
(in
module
sympy.utilities.decorator), 1548
to ring() (sympy.polys.polytools.Poly method),
threaded factory()
(in
module
1110
sympy.utilities.decorator), 1548
to sympy() (sympy.polys.domains.AlgebraicField
method), 1159
time (sympy.physics.quantum.state.TimeDepState
attribute), 1783
to sympy() (sympy.polys.domains.domain.Domain
time derivative()
(in
module
method), 1152
to sympy() (sympy.polys.domains.ExpressionDomain
sympy.physics.vector.functions),
1689
method), 1161
time period (sympy.physics.optics.waves.TWaveto sympy() (sympy.polys.domains.FiniteField
attribute), 1819
method), 1156
timed() (in module sympy.utilities.timeutils), to sympy() (sympy.polys.domains.FractionField
1587
method), 1160
TimeDepBra
(class
in to sympy() (sympy.polys.domains.PolynomialRing
sympy.physics.quantum.state), 1783
method), 1157
TimeDepKet
(class
in to sympy() (sympy.polys.domains.RealField
sympy.physics.quantum.state), 1784
method), 1160
TimeDepState
(class
in to sympy dict() (sympy.polys.polyclasses.ANP
Index
1991
SymPy Documentation, Release 0.7.6
method), 1169
method), 502
to sympy dict() (sympy.polys.polyclasses.DMP translate() (sympy.geometry.point3d.Point3D
method), 1167
method), 508
to sympy list() (sympy.polys.polyclasses.ANP Transpose
(class
in
method), 1169
sympy.matrices.expressions), 710
to tree() (sympy.combinatorics.prufer.Prufer transpose() (sympy.matrices.expressions.blockmatrix.BlockM
static method), 249
method), 712
to tuple()
(sympy.polys.polyclasses.ANP transpositions() (sympy.combinatorics.permutations.Permut
method), 1169
method), 212
to tuple()
(sympy.polys.polyclasses.DMP tree() (in module sympy.printing.tree), 1263
method), 1167
tree cse()
(in
module
together()
(in
module
sympy.simplify.cse main), 1346
sympy.polys.rationaltools), 1120
tree repr (sympy.combinatorics.prufer.Prufer
together() (sympy.core.expr.Expr method),
attribute), 249
122
Triangle (class in sympy.geometry.polygon),
TokenError
(class
in
570
sympy.parsing.sympy tokenize),
Triangular() (in module sympy.stats), 1386
1590
trigamma()
(in
module
tokenize()
(in
module
sympy.functions.special.gamma functions),
sympy.parsing.sympy tokenize),
384
1589
trigintegrate()
(sympy.integrals
static
tolist() (sympy.matrices.sparse.SparseMatrix
method), 613
method), 699
TrigSimp()
(in
module
topological sort()
(in
module
sympy.galgebra.vector), 484
sympy.utilities.iterables), 1570
trigsimp()
(in
module
sympy.simplify.simplify), 1336
total degree() (sympy.polys.polyclasses.DMP
method), 1167
trigsimp() (sympy.core.expr.Expr method),
total degree()
(sympy.polys.polytools.Poly
122
method), 1110
trunc() (in module sympy.polys.polytools),
totient() (in module sympy.ntheory.factor ),
1067
290
trunc()
(sympy.polys.polyclasses.DMP
Trace (class in sympy.matrices.expressions),
method), 1167
710
trunc() (sympy.polys.polytools.Poly method),
trailing() (in module sympy.ntheory.factor ),
1111
282
Tuple (class in sympy.core.containers), 182
transcendental, 89
tuple count()
(sympy.core.containers.Tuple
transform()
(sympy.geometry.point.Point
method), 182
method), 502
TWave (class in sympy.physics.optics.waves),
transform() (sympy.geometry.point3d.Point3D
1817
method), 508
twinprime (mpmath.mp attribute), 751
transform()
(sympy.integrals.Integral twoform to matrix()
(in
module
method), 617
sympy.digeom), 1858
transform nodes()
(mp- TwoQubitGate
(class
in
math.calculus.quadrature.QuadratureRule
sympy.physics.quantum.gate), 1790
method), 1011
transformation to DN()
(in
module U
sympy.solvers.diophantine), 1494
U() (in module sympy.printing.pretty.pretty symbology),
transitivity degree (sympy.combinatorics.perm groups.PermutationGroup
1266
attribute), 243
ufuncify()
(in
module
translate()
(sympy.geometry.curve.Curve
sympy.utilities.autowrap), 1536
method), 542
UfuncifyCodeWrapper
(class
in
translate() (sympy.geometry.entity.GeometryEntity
sympy.utilities.autowrap), 1534
method), 494
UGate (class in sympy.physics.quantum.gate),
translate()
(sympy.geometry.point.Point
1790
1992
Index
SymPy Documentation, Release 0.7.6
uncouple()
(in
module
class method), 212
sympy.physics.quantum.spin), 1780
unrank nonlex() (sympy.combinatorics.permutations.Permut
Unequality (class in sympy.core.relational),
class method), 213
155
unrank trotterjohnson()
unatten()
(in
module
(sympy.combinatorics.permutations.Permutation
sympy.utilities.iterables), 1571
class method), 213
UnicationFailed
(class
in unrestricted necklace()
(in
module
sympy.polys.polyerrors), 1238
sympy.utilities.iterables), 1572
Uniform() (in module sympy.stats), 1387
untokenize()
(in
module
UniformSum() (in module sympy.stats), 1388
sympy.parsing.sympy tokenize),
unify() (sympy.polys.domains.domain.Domain
1589
method), 1152
upper (sympy.physics.secondquant.AntiSymmetricTensor
attribute), 1621
unify()
(sympy.polys.polyclasses.ANP
method), 1169
upper (sympy.tensor.indexed.Idx attribute),
unify()
(sympy.polys.polyclasses.DMP
1511
upper triangular solve()
method), 1167
unify() (sympy.polys.polytools.Poly method),
(sympy.matrices.matrices.MatrixBase
1111
method), 677
Union (class in sympy.sets.sets), 1320
uppergamma
(class
in
union()
(sympy.polys.agca.ideals.Ideal
sympy.functions.special.gamma functions),
method), 1140
384
union() (sympy.polys.agca.modules.SubModule use() (in module sympy.simplify.traversaltools),
method), 1137
1347
union()
(sympy.series.gruntz.SubsSet
V
method), 1308
union() (sympy.sets.sets.Set method), 1316
v1pt theory() (sympy.physics.vector.point.Point
uniq() (in module sympy.utilities.iterables),
method), 1679
1572
v2pt theory() (sympy.physics.vector.point.Point
Unit (class in sympy.physics.units), 1631
method), 1679
Unit (class in sympy.physics.unitsystems.units), values()
(sympy.core.containers.Dict
1827
method), 183
unit (sympy.polys.polytools.Poly attribute), values() (sympy.matrices.matrices.MatrixBase
1111
method), 677
UnitaryOperator
(class
in values() (sympy.physics.unitsystems.dimensions.Dimension
sympy.physics.quantum.operator),
method), 1825
1761
var() (in module sympy.core.symbol), 126
unitroots() (in module mpmath), 754
VarBosonicBasis
(class
in
UnitSystem
(class
in
sympy.physics.secondquant), 1613
sympy.physics.unitsystems.units),
variable map() (sympy.physics.vector.frame.ReferenceFrame
1828
method), 1669
UnivariatePolynomialError
(class
in variables (sympy.core.function.Lambda atsympy.polys.polyerrors), 1238
tribute), 163
UniversalSet (class in sympy.sets.sets), 1324 variables
(sympy.core.function.Subs
atunparse paren()
(in
module
tribute), 171
sympy.galgebra.precedence), 485
variables (sympy.physics.quantum.operator.DierentialOper
unrank() (sympy.combinatorics.graycode.GrayCode
attribute), 1764
class method), 259
variables (sympy.physics.quantum.state.Wavefunction
unrank() (sympy.combinatorics.prufer.Prufer
attribute), 1788
class method), 250
variables
(sympy.utilities.codegen.Routine
unrank binary() (sympy.combinatorics.subsets.Subset attribute), 1540
class method), 255
variance() (in module sympy.stats), 1394
unrank gray() (sympy.combinatorics.subsets.Subset
variations()
(in
module
class method), 256
sympy.utilities.iterables), 1572
unrank lex() (sympy.combinatorics.permutations.Permutation
Index
1993
SymPy Documentation, Release 0.7.6
vec()
(sympy.matrices.matrices.MatrixBase waist approximation limit
method), 677
(sympy.physics.optics.gaussopt.BeamParameter
vech() (sympy.matrices.matrices.MatrixBase
attribute), 1809
method), 677
Wavefunction
(class
in
Vector (class in sympy.galgebra.vector), 484
sympy.physics.quantum.state), 1784
Vector (class in sympy.physics.vector.vector), wavelength (sympy.physics.optics.waves.TWave
1669
attribute), 1819
vectorize
(class
in wavenumber (sympy.physics.optics.waves.TWave
sympy.core.multidimensional), 162
attribute), 1819
vectors in basis()
(in
module webere() (in module mpmath), 835
sympy.digeom), 1858
WedgeProduct (class in sympy.digeom),
vel()
(sympy.physics.vector.point.Point
1854
method), 1680
Weibull() (in module sympy.stats), 1389
verify numerically()
(in
module WGate (class in sympy.physics.quantum.grover),
1794
sympy.utilities.randtest), 1579
vertices (sympy.combinatorics.polyhedron.Polyhedron
where() (in module sympy.stats), 1394
attribute), 246
whitm() (in module mpmath), 863
vertices
(sympy.geometry.polygon.Polygon whitw() (in module mpmath), 864
attribute), 562
wicks()
(in
module
vertices (sympy.geometry.polygon.RegularPolygon
sympy.physics.secondquant), 1617
attribute), 570
width (sympy.categories.diagram drawing.DiagramGrid
vertices
(sympy.geometry.polygon.Triangle
attribute), 1841
attribute), 576
width() (sympy.printing.pretty.stringpict.stringPict
VF() (in module sympy.printing.pretty.pretty symbology),
method), 1269
1267
Wigner3j
(class
in
sympy.physics.quantum.cg), 1751
viete() (in module sympy.polys.polyfuncs),
1115
Wigner6j
(class
in
vlatex()
(in
module
sympy.physics.quantum.cg), 1751
sympy.physics.vector.printing), 1686 Wigner9j
(class
in
sympy.physics.quantum.cg), 1751
vobj()
(in
module
sympy.printing.pretty.pretty symbology),
wigner 3j()
(in
module
1267
sympy.physics.wigner), 1626
VonMises() (in module sympy.stats), 1389
wigner 6j()
(in
module
vpprint()
(in
module
sympy.physics.wigner), 1627
sympy.physics.vector.printing), 1685 wigner 9j()
(in
module
vprint()
(in
module
sympy.physics.wigner), 1628
sympy.physics.vector.printing), 1685 WignerD
(class
in
vradius (sympy.geometry.ellipse.Circle atsympy.physics.quantum.spin), 1773
tribute), 556
WignerSemicircle() (in module sympy.stats),
vradius (sympy.geometry.ellipse.Ellipse at1390
tribute), 553
Wild (class in sympy.core.symbol), 123
vring() (in module sympy.polys.rings), 1214
WildFunction (class in sympy.core.function),
vstack() (sympy.matrices.matrices.MatrixBase
163
workdps() (in module mpmath), 744
class method), 678
workprec() (in module mpmath), 744
W
write()
(sympy.utilities.codegen.CodeGen
method),
1541
w (sympy.physics.optics.gaussopt.BeamParameter
write()
(sympy.utilities.runtests.PyTestReporter
attribute), 1808
method), 1580
w 0 (sympy.physics.optics.gaussopt.BeamParameter
wronskian()
(in
module
attribute), 1808
682
sympy.matrices.dense),
waist2rayleigh()
(in
module
sympy.physics.optics.gaussopt),
X
1809
X (in module sympy.physics.quantum.gate),
1994
Index
SymPy Documentation, Release 0.7.6
1792
YGate (class in sympy.physics.quantum.gate),
x (sympy.geometry.point.Point attribute), 503
1791
x
(sympy.geometry.point3d.Point3D
at- yn (class in sympy.functions.special.bessel),
tribute), 508
413
x (sympy.physics.vector.frame.ReferenceFrame Ynm (class in sympy.functions.special.spherical harmonics),
attribute), 1669
443
XBra (class in sympy.physics.quantum.cartesian),
Ynm c()
(in
module
1757
sympy.functions.special.spherical harmonics),
xdirection
(sympy.geometry.line.Ray
at445
tribute), 521
YOp (class in sympy.physics.quantum.cartesian),
xdirection (sympy.geometry.line3d.Ray3D at1757
tribute), 536
xdvi() (in module sympy.galgebra.printing), Z
486
Z (in module sympy.physics.quantum.gate),
XGate (class in sympy.physics.quantum.gate),
1792
1791
z
(sympy.geometry.point3d.Point3D
atXKet (class in sympy.physics.quantum.cartesian),
tribute), 508
1757
z (sympy.physics.vector.frame.ReferenceFrame
xobj()
(in
module
attribute), 1669
sympy.printing.pretty.pretty symbology),
zdirection (sympy.geometry.line3d.Ray3D at1267
tribute), 536
XOp (class in sympy.physics.quantum.cartesian),zero, 89
1757
Zero (class in sympy.core.numbers), 132
Xor (class in sympy.logic.boolalg), 629
zero (sympy.polys.polytools.Poly attribute),
xreplace() (sympy.core.basic.Basic method),
1111
101
ZeroMatrix
(class
in
xring() (in module sympy.polys.rings), 1214
sympy.matrices.expressions), 711
xstr() (in module sympy.printing.pretty.pretty symbology),
zeros() (in module sympy.matrices.dense),
1267
679
xsym()
(in
module zeros() (sympy.matrices.sparse.SparseMatrix
sympy.printing.pretty.pretty symbology),
class method), 699
1267
zeta (class in sympy.functions.special.zeta functions),
xthreaded()
(in
module
421
sympy.utilities.decorator), 1548
zeta() (in module mpmath), 931
xypic draw diagram()
(in
module zetazero() (in module mpmath), 938
sympy.categories.diagram drawing), ZGate (class in sympy.physics.quantum.gate),
1845
1791
XypicDiagramDrawer
(class
in zip row op() (sympy.matrices.dense.MutableDenseMatrix
sympy.categories.diagram drawing),
method), 689
1843
zip row op() (sympy.matrices.sparse.MutableSparseMatrix
method), 703
Y
Znm (class in sympy.functions.special.spherical harmonics),
Y (in module sympy.physics.quantum.gate),
445
1792
ZOp (class in sympy.physics.quantum.cartesian),
y (sympy.geometry.point.Point attribute), 503
1757
y
(sympy.geometry.point3d.Point3D
attribute), 508
y (sympy.physics.vector.frame.ReferenceFrame
attribute), 1669
ydirection
(sympy.geometry.line.Ray
attribute), 522
ydirection (sympy.geometry.line3d.Ray3D attribute), 536
Index
1995
You might also like
- 10-17-13 SK GOSH Webinar PT - TwoWay - Floor - SystemNo ratings yet10-17-13 SK GOSH Webinar PT - TwoWay - Floor - System73 pages
- 0 - BASE SHEAR REDISTRIBUTION BETWEEN THE RC DUAL SYSTEM STRUCTURAL COMPONENTS - V.Sigmund Itd. - 2008 - 8158 PDFNo ratings yet0 - BASE SHEAR REDISTRIBUTION BETWEEN THE RC DUAL SYSTEM STRUCTURAL COMPONENTS - V.Sigmund Itd. - 2008 - 8158 PDF8 pages
- Design of Radial Power Combiners Based On TE01 Circular Waveguide Mode100% (1)Design of Radial Power Combiners Based On TE01 Circular Waveguide Mode20 pages
- Cse Vi Computer Graphics and Visualization 10cs65 Notes100% (2)Cse Vi Computer Graphics and Visualization 10cs65 Notes97 pages
- Ductility Design of Reinforced Concrete Shear Walls100% (1)Ductility Design of Reinforced Concrete Shear Walls12 pages
- Technical Data For Steel-Concrete Composite Beam: Type Approval 300/6221/2001No ratings yetTechnical Data For Steel-Concrete Composite Beam: Type Approval 300/6221/200120 pages
- Jamshid Ghaboussi - Soft Computing in Engineering-CRC Press (2018)No ratings yetJamshid Ghaboussi - Soft Computing in Engineering-CRC Press (2018)221 pages
- Allowable Stress Design Flowchart For AISC Manual For Steel Construction, 9th EditionNo ratings yetAllowable Stress Design Flowchart For AISC Manual For Steel Construction, 9th Edition8 pages
- C, C++ Programming Tips: Vishal Patil Summer 2003No ratings yetC, C++ Programming Tips: Vishal Patil Summer 200321 pages
- PTI Current Trends in PT Design and Construction San Diego Handou...No ratings yetPTI Current Trends in PT Design and Construction San Diego Handou...44 pages
- NCRTC: Column Schedule Column Schedule Column ScheduleNo ratings yetNCRTC: Column Schedule Column Schedule Column Schedule1 page
- In Structural Steel: Limit States DesignNo ratings yetIn Structural Steel: Limit States Design229 pages
- Transverse Web Stiffeners and Shear Moment Interation For Steel Plate Girder BridgesNo ratings yetTransverse Web Stiffeners and Shear Moment Interation For Steel Plate Girder Bridges14 pages
- Engineering Mechanics by Dhiman and KulshereshthaNo ratings yetEngineering Mechanics by Dhiman and Kulshereshtha23 pages
- Namma Kalvi 12th Computer Science Unit 5 Sura Guide emNo ratings yetNamma Kalvi 12th Computer Science Unit 5 Sura Guide em16 pages
- Python Programming For Economics-ImportanteNo ratings yetPython Programming For Economics-Importante341 pages
- Python Programming For Economics and Finance100% (2)Python Programming For Economics and Finance300 pages
- Sargent T. Python Programming For Economics and Finance 2023No ratings yetSargent T. Python Programming For Economics and Finance 2023365 pages
