CSB is distributed as a pure python package, which renders it cross-platform. Installation is therefore a very trivial task. The source code is also kept compatible across Python versions, so we guarantee that every official release package will work on both Python 2 and 3 without any change. We test all nightly builds against Python versions 2.7 and 3.6, but technically CSB should run on Python 2.6 and all subsequent versions except probably 3.0.
Install csb
from PyPi:
$ pip install csb
pip
will take care of all necessary dependencies. However, on Windows this might
fail. If this is the case, install numpy
and scipy
manually and rerun the pip
installation shown above. The easiest way to get numpy
and scipy
on Windows is
to get them from this site. Or if
you are using Anaconda:
$ conda install numpy scipy
$ pip install csb
ABOUT DEPENDENCIES
CSB depends on 2 very well-known and readily available python packages:
csb.io.plots
epydoc
is required to build the API documentation.
INSTALLATION
First get the source code from github:
$ git clone https://github.com/csb-toolbox/CSB.git
$ cd CSB
Then install the project in editable mode:
$ pip install --editable .[dev]
pip
will take care of all runtime and development dependencies and make the source
code location importable. No need to modify your $PYTHONPATH
.
If dependencies cannot be resolved on your system - this is often the case on Windows - the best way to go from here is to install Anaconda. Then use the conda package manager to install all missing packages:
$ conda install --file requirements.txt
$ pip install --editable .[dev]
Running the CSB test suite may be useful in order to check if your installation works.
All CSB tests are executed with the csb.test.Console
. A typical way to run the
console is:
$ csb-test "csb.test.cases.*"