scikit-learn/scikit-learn

About

scikits.learn is a python module for machine learning built on top of scipy.

The project was started in 2007 by David Cournapeu as a Google Summer of Code project, and since then many volunteers have contributed. See the AUTHORS file for a complete list of contributors.

It is currently maintained by a team of volunteers.

Download

You can download source code and Windows binaries from SourceForge:

http://sourceforge.net/projects/scikit-learn/files/

Dependencies

The required dependencies to build the software are python >= 2.5, NumPy >= 1.1, SciPy, the Boost libraries and a working C++ compiler.

Optional dependencies are scikits.optimization for module machine.manifold.

To run the tests you will also need nosetests and python-dap (http://pypi.python.org/pypi/dap/).

Install

This packages uses distutils, which is the default way of installing python modules. The install command is:

python setup.py install

Mailing list

There's a general and development mailing list, visit https://lists.sourceforge.net/lists/listinfo/scikit-learn-general to subscribe to the mailing list.

Development

Code

SVN

To check out the sources from the main repository, which is under subversion control, run the command:

svn co http://scikit-learn.svn.sourceforge.net/svnroot/scikit-learn/trunk scikit-learn

You can also browse the code online at http://scikit-learn.svn.sourceforge.net/viewvc/scikit-learn

GIT Mirror

There is also an automatically updated git mirror of scikit-learn available from github (http://github.com/yarikoptic/scikit-learn). So you could setup your development using GIT and commit results back to SVN using git svn. To get started using git mirror and link it to original SVN, use:

git clone --origin=svn git://github.com/yarikoptic/scikit-learn.git scikit-learn
cd scikit-learn
git svn init --prefix=svn/ -s https://scikit-learn.svn.sourceforge.net/svnroot/scikit-learn
git svn rebase -l -n                           # should be quick
git config svn.authorsfile .svnauthors

Such setup was inspired by GIT Mirror for NumPy and SciPy projects, which is well described (including some of git svn caveats) at http://projects.scipy.org/numpy/wiki/GitMirror .

If you are not a part of the team and/or have no commit rights to SVN, but want to share code/documentation changes (bug fixes, enhancements, etc.), you can either make use of git format-patch commands to prepare a bundle of patches to be posted with bugreport (see Bugs), or simply create your own fork of git repository on github ("Fork" button in github web UI) and point to corresponding commits within your bugreport.

Bugs

Please submit bugs you might encounter, as well as patches and feature requests to the tracker located at the address https://sourceforge.net/apps/trac/scikit-learn/report

Testing

To execute the test suite, run from the project's top directory (you will need to have nosetest installed):

python setup.py test