======================================
Requirements for the Python workshop
======================================
.. contents::
..
1 Core requirements
2 Basic configuration
3 Testing
4 Checking your versions
5 Platform specific instructions
6 Optional packages
Core requirements
=================
The four python packages you will need for this course are
- ipython_: an interactive python shell
- numpy_: high performance numerical arrays
- scipy_: algorithms and numerics
- matplotlib_: plotting
.. _ipython: http://ipython.scipy.org
.. _numpy: http://numpy.scipy.org
.. _scipy: http://scipy.org
.. _matplotlib: http://matplotlib.sf.net
With these four packages, you will be able to do about 90% of the
exercises in the course. In addition, to generate plot windows,
matplotlib requires a user interface toolkit, for example Tk, GTK,
wxPython or Qt. Tk comes with python, so that should work out of the
box, but some of the other code we will be demoing work best with
wxPython, so we also recommend you install wxPython_
.. _wxPython: http://www.wxpython.org
We will detail the optional packages that will be used in the workshop
below. It will be great if you can get all the optional packages as
well, but since most of the workshop depends on these four
requirements, make sure you have these four installed, configured and
tested before turning to the optional packages.
For those of you who do not succeed in getting a machine properly
configured with the packages above, we will be providing an X11 server
that you can log into to work on the exercises remotely. If your
machine supports ssh with X11 forwarding, you will be in good shape.
If not, we will provide ubuntu boot CDs so you can reboot into linux
and then ssh into the x11 server to run the exercises.
Basic configuration
===================
You need to configure matplotlib to work with the user interface you
have selected. The configuration file is called matplotlibrc and it
is installed in site-packages/matplotlib/mpl-data/matplotlibrc along
with the rest of your matplotlib install. Create a directory in your
home directory called .matplotlib and copy this file into it (or
simply edit in place in mpl-data) and change the line that starts with
'backend' to::
backend : WXAgg
for use with wxpython.
Testing
=======
If you can execute the following commands w/o error, and have a plot
window pop up, you have an installation that will work for 90% of the
exercises in the workshop::
> ipython -pylab
Python 2.5 (r25:51918, Sep 19 2006, 08:49:13)
Type "copyright", "credits" or "license" for more information.
IPython 0.8.3.svn.r2876 -- An enhanced Interactive Python.
In [1]: import numpy
In [2]: import scipy
In [3]: plot([1,2,3])
Out[3]: [<matplotlib.lines.Line2D instance at 0x4c9ae40>]
Checking your versions
======================
It helps to have the most recent packages for the workshop. We are
recommending at least the following
- ipython >= 0.7.0
- numpy >= 1.01
- matplotlib >= 0.90.0
- scipy >= 0.5.2
The example code below shows you how to check your versions::
In [4]: import numpy
In [5]: numpy.__version__
Out[5]: '1.0.4.dev4380'
In [6]: import scipy
In [7]: scipy.__version__
Out[7]: '0.7.0.dev3494'
In [8]: import IPython
In [9]: IPython.__version__
Out[9]: '0.8.3.svn.r2876'
In [10]: import matplotlib
In [11]: matplotlib.__version__
Out[11]: '0.91.1'
Platform specific instructions
==============================
Linux
=====
TODO
OS X
====
TODO
Microsoft Windows
=================
- Install the latest windows `python MSI installer`_ from the
python.org site.
- install the win32com_ extensions. IPython needs these to set up the
start menu properly and they have lots of useful goodies for windows
users anyhow, like the abiliy to read and write MS Office documents
via the COM API.
- install pyreadline_: ipython needs this for tab completion and other handy features
- ipython: grab the latest win32 exe installer from the `ipython distributions`_ page
- numpy: grab the latest exe or msi installer for your version of
python at `numpy download`_
- scipy: grab the latest exe or msi installer from the "Binary
installation for Windows" distribution from the `scipy download`_
page.
- matplotlib: grab the latest exe from the `matplotlib download`_
page. matplotlib requires the msvcp71.dll library, which you can
download from `dll_files`. You will need to unzip the archive and
drag the dll into c:\\windows\\system32
.. _`python MSI installer`: http://www.python.org/download/releases/2.5.1/
.. _win32com: http://sourceforge.net/project/showfiles.php?group_id=78018&package_id=79063&release_id=449591
.. _pyreadline: http://ipython.scipy.org/moin/PyReadline/Intro
.. _`ipython distributions`: http://ipython.scipy.org/dist/
.. _`numpy download`: http://sourceforge.net/project/showfiles.php?group_id=1369&package_id=175103
.. _`scipy download`: http://www.scipy.org/Download
.. _`matplotlib download`: http://sourceforge.net/project/showfiles.php?group_id=80706&package_id=82474&release_id=558083
.. _`dll files`: http://www.dll-files.com/dllindex/dll-files.shtml?msvcp71
Optional packages
=================
We will be doing demos and exercises with a nmber of additional
packages, which we will describe briefly below. Because there are so
many variations in operating systems and platforms, we will not
attempt to provide details so consult the web sites and other
resources if there is a package you are particularly interested in and
want to make sure you are configured properly to do an exercise on
your own machine. We will be providing a properly configured server,
so even if you do not have each of these packages installed, as long
as your machine supports remote login with X11 forwarding to our
server, you can long in to do the exercises if you are missing a
package on your local machine.
- pytables_ : a package for managing large datasets efficiently
(also requires hdf5)
- pydap_ : a python implementation of the Data Access Protocol
(DAP).
- `enthought tool suite`_ (ETS) : provides traits, a package for strong
typing in python that also facilitates automatic user interface
generation and more. Also provides tvtk and mayavi2 for 3D
visualization.
- weave : a package for inlining C and C++ in your python code. Comes
with scipy but requires a working compiler suite (eg, gcc and g++)
- f2py : comes with scipy but requires a working fortran compiler. See
http://scipy.org/Installing_SciPy for platform specific instructions
- pyrex_ : lets you mix c and python for writing high performance
python extensions
- basemap_ : a matplotlib toolkit for workin with geographic data and
map projections. Available for download at the matplotlib download
page
http://sourceforge.net/project/platformdownload.php?group_id=80706
- svn_ : we wil be teaching open source software development
processes, which require svn. Windows users should consider
`tortise svn`_
.. _pytables: http://www.pytables.org/moin
.. _basemap: http://matplotlib.sourceforge.net/matplotlib.toolkits.basemap.basemap.html
.. _pyrex: http://www.cosc.canterbury.ac.nz/greg.ewing/python/Pyrex/
.. _`enthought tool suite`: http://code.enthought.com/ets/
.. _pydap: http://pydap.org.
.. _svn: http://subversion.tigris.org/
.. `tortise svn`_: http://tortoisesvn.tigris.org/
