Ok, so you want to do it the hard way?

For some people, the prepackaged pythons discussed above are not an option. That's OK, it's usually pretty easy to get a custom install working. You will first need to find out if you have python installed on your machine, and if not, install it. The official python builds are available for download here, but OS X users please read :ref:`which-python-for-osx`.

Once you have python up and running, you will need to install numpy. numpy provides high performance array data structures and mathematical functions, and is a requirement for matplotlib. You can test your progress:

>>> import numpy
>>> print numpy.__version__

matplotlib requires numpy version 1.1 or later. Although it is not a requirement to use matplotlib, we strongly encourage you to install ipython, which is an interactive shell for python that is matplotlib aware. Once you have ipython, numpy and matplotlib installed, in ipython's "pylab" mode you have a matlab-like environment that automatically handles most of the configuration details for you, so you can get up and running quickly:

johnh@flag:~> ipython -pylab
Python 2.4.5 (#4, Apr 12 2008, 09:09:16)
IPython 0.9.0 -- An enhanced Interactive Python.

  Welcome to pylab, a matplotlib-based Python environment.
  For more information, type 'help(pylab)'.

In [1]: x = randn(10000)

In [2]: hist(x, 100)

And a voila, a figure pops up. But we are putting the cart ahead of the horse -- first we need to get matplotlib installed. We provide prebuilt binaries for OS X and Windows on the matplotlib download page. Click on the latest release of the "matplotlib" package, choose your python version (2.4 or 2.5) and your platform (macosx or win32) and you should be good to go. If you have any problems, please check the :ref:`installing-faq`, google around a little bit, and post a question the mailing list.

Instructions for installing our OSX binaries are found in the FAQ ref:install_osx_binaries.

Note that when testing matplotlib installations from the interactive python console, there are some issues relating to user interface toolkits and interactive settings that are discussed in :ref:`mpl-shell`.

Installing from source

If you are interested perhaps in contributing to matplotlib development, or just like to build everything yourself, it is not difficult to build matplotlib from source. Grab the latest tar.gz release file from sourceforge, or if you want to develop matplotlib or just need the latest bugfixed version, grab the latest svn version :ref:`install-svn`.

Once you have satisfied the requirements detailed below (mainly python, numpy, libpng and freetype), you build matplotlib in the usual way:

cd matplotlib
python setup.py build
python setup.py install

We provide a setup.cfg file that lives along :file:`setup.py` which you can use to customize the build process, for example, which default backend to use, whether some of the optional libraries that matplotlib ships with are installed, and so on. This file will be particularly useful to those packaging matplotlib.

Build requirements

These are external packages which you will need to install before installing matplotlib. Windows users only need the first two (python and numpy) since the others are built into the matplotlib windows installers available for download at the sourceforge site. If you are building on OSX, see :ref:`build_osx`

:term:`python` 2.4 (or later but not python3)

matplotlib requires python 2.4 or later (download)

:term:`numpy` 1.1 (or later)

array support for python (download)

libpng 1.1 (or later)

library for loading and saving :term:`PNG` files (download). libpng requires zlib. If you are a windows user, you can ignore this since we build support into the matplotlib single click installer

:term:`freetype` 1.4 (or later)

library for reading true type font files. If you are a windows user, you can ignore this since we build support into the matplotlib single click installer.

Optional

These are optional packages which you may want to install to use matplotlib with a user interface toolkit. See :ref:`what-is-a-backend` for more details on the optional matplotlib backends and the capabilities they provide

:term:`tk` 8.3 or later

The TCL/Tk widgets library used by the TkAgg backend

:term:`pyqt` 3.1 or later

The Qt3 widgets library python wrappers for the QtAgg backend

:term:`pyqt` 4.0 or later

The Qt4 widgets library python wrappersfor the Qt4Agg backend

:term:`pygtk` 2.2 or later

The python wrappers for the GTK widgets library for use with the GTK or GTKAgg backend

:term:`wxpython` 2.6 or later

The python wrappers for the wx widgets library for use with the WXAgg backend

:term:`wxpython` 2.8 or later

The python wrappers for the wx widgets library for use with the WX backend

:term:`pyfltk` 1.0 or later

The python wrappers of the FLTK widgets library for use with FLTKAgg

Required libraries that ship with matplotlib

:term:`agg` 2.4

The antigrain C++ rendering engine. matplotlib links against the agg template source statically, so it will not affect anything on your system outside of matplotlib.

pytz 2007g or later

timezone handling for python datetime objects. By default, matplotlib will install pytz if it isn't already installed on your system. To override the default, use setup.cfg to force or prevent installation of pytz.

dateutil 1.1 or later

extensions to python datetime handling. By default, matplotlib will install dateutil if it isn't already installed on your system. To override the default, use setup.cfg to force or prevent installation of dateutil.

Building on OSX

The build situation on OSX is complicated by the various places one can get the png and freetype requirements from (darwinports, fink, /usr/X11R6) and the different architectures (x86, ppc, universal) and the different OSX version (10.4 and 10.5). We recommend that you build the way we do for the OSX release: by grabbing the tarbar or svn repository, cd-ing into the release/osx dir, and following the instruction in the README. This directory has a Makefile which will automatically grab the zlib, png and freetype dependencies from the web, build them with the right flags to make universal libraries, and then build the matplotlib source and binary installers.