Python Mapper does not contain any platform-specific code and depends only on cross-platform packages. Hence, it should run on almost all modern operating systems.

So far, installation has been tested under:

  • Arch Linux
  • Ubuntu
  • Microsoft Windows 8.1
  • OS X Mountain Lion

See the OS-specific pages here:

If you install Python Mapper on a certain platform not in the list above, please let me (Daniel) know so that I can extend the list. Especially if you needed to tweak or modify something, I am interested to know about it.


  • Python 2.6 or higher. The GUI needs Python 2 since it depends on wxPython and PyOpenGL; Python Mapper itself can be run under Python 2 and Python 3. I also recommend to install pip if it is not already included in your Python distribution. (Without pip, replace pip install with easy_install in the command lines below.)
  • NumPy and SciPy
  • Matplotlib
  • Graphviz
  • Optionally cmappertools. Python Mapper will run without this module, but with limited functionality.

For the GUI:

Highly recommended:

  • cmappertools

    This is Daniel Müllner’s Python module, written in C++, which replaces some slower Python routines by fast, parallelized C++ algorithms. Currently, it is in the cmappertools subdirectory of the Python Mapper distribution. Cmappertools are hosted on PyPI, so they can be installed with:

    $ pip install cmappertools --user
  • fastcluster

    This is Daniel Müllner’s C++ library for fast hierarchical clustering, again wrapped as a Python module. Again, it can be installed from PyPI by:

    $ pip install fastcluster --user

    If this does not work, please refer to the detailed installation instructions in the fastcluster distribution. You need the Python interface; the R interface can be ignored.

Standard installation

Simply type

$ pip install mapper --user

on a command line.

If everything worked, you may stop here and start using Mapper. The steps below describe alternatives and optional steps.

Source distribution

If the simple method above does not work, the source distribution of Python Mapper can be downloaded here:

Since Python Mapper is not stable yet and under active development, the distribution will be updated frequently. If you want to follow updates more easily and avoid to install the same package over and over again, it is recommended to use the Mercurial repository. Create a local copy of the repository with:

$ hg clone

To update the repository, type:

$ cd mapper
$ hg pull
$ hg up

Installation from source

The Python Mapper archive can be extracted anywhere. There is a setup script in the mapper directory, which can be run with:

$ python install --user

Alternatively, no real installation is necessary. Python just needs to know the location of the package. For this, add the directory where the files were extracted to Python’s search path. (Ie., add the directory which contains mapper as a subdirectory to the Python path.)

Users may also want to add a link to the mapper/bin/ script in a directory which is searched for executables. For example, my .bashrc contains a line

export PATH="${PATH+$PATH:}$HOME/.local/bin"

so I can add a link to the GUI by:

$ cd ~/.local/bin
$ ln -s (MAPPER PATH)/bin/


If the GUI refuses to start with an error message like

/usr/bin/env: python2: No such file or directory

there are three ways to deal with the problem:

  • Do not call MapperGUI as an executable script but (your Python 2 interpreter) (your path)/, eg.:

    $ python mapper/bin/
  • Create a symbolic link like:

    $ sudo ln -s (path to the Python 2 interpreter) /usr/local/bin/python2
  • Change the first line in from

    #!/usr/bin/env python2


    #!/usr/bin/env (your Python 2 interpreter)

    With the last method, however, changes will be lost when Python Mapper is updated.

Compiling the documentation

The HTML documentation (this page!) can be compiled with Sphinx:

$ cd mapper/doc
$ make html

If you get an error like

make: sphinx-build2: No such file or directory


$ make html SPHINXBUILD=sphinx-build