Download & Installation¶

Dependencies¶

PyCS is developed using python 2.7 and might be fine with older versions, but currently not with python 3.

It requires numpy, scipy and matplotlib. As of March 2016, everything seems to work with the latest versions of these packages. Those are all you need to run the free-knot splines.

Some other estimators (e.g., the “regression difference” technique) might have furter dependencies:

Download¶

Get the latest PyCS by cloning it from GitHub:

git clone https://github.com/COSMOGRAIL/PyCS.git

If for some reason you want the exact version of PyCS that was used in the original paper, you can still find it here:

svn export https://svn.epfl.ch/svn/pycs/tags/PyCS-1.0 ./PyCS-1.0

Installation¶

PyCS is not exactly “production level code”, and you might want to update or tweak the sources. So to keep this simple, we suggest to just add your cloned repository to your PYTHONPATH.

For tcsh, add for instance something like this to your .tcshrc

#setenv PYTHONPATH /path/to/PyCS
# or, to add to the existing stuff :

setenv PYTHONPATH ${PYTHONPATH}:/path/to/PyCS

or, if you use bash, to your .bash_profile or equivalent:

export PYTHONPATH=${PYTHONPATH}:/path/to/PyCS/

… and then from any python you can simply import pycs.

If you don’t plan to tweak the code, you can also simply

python setup.py install

or maybe

python setup.py install --user

… if you don’t have write access to the global site-packages directory of your machine.