Interface to use R from Python


Keywords
cffi, data-science, interoperability, python, r, statistics
License
CNRI-Python-GPL-Compatible
Install
conda install -c conda-forge rpy2

Documentation

Python -> R bridge

pypi Codecov GH Actions

The project's webpage is here: https://rpy2.github.io/

Installation

pip should work out of the box:

pip install rpy2

The package has optional depencies providing specific functionalities not otherwise required to use the rest of rpy2.

For example, to be able to run the unit tests:

pip install rpy2[test]

To install all optional dependencies (numpy, pandas, ipython), use:

pip install rpy2[all]

The package is known to compile on Linux, MacOSX (provided that developper tools are installed, and you are ready figure out how by yourself). The situation is currently a little more complicated on Windows. Check the issue tracker.

In case you find yourself with this source without any idea of what it takes to compile anything on your platform, try first

python setup.py install

Issues loading shared C libraries

Whenever R is in not installed in a system location, the system might not know where to find the R shared library.

If R is in the PATH, that is entering R on the command line successfully starts an R terminal, but rpy2 does not work because of missing C libraries, try the following before starting Python:

export LD_LIBRARY_PATH="$(python -m rpy2.situation LD_LIBRARY_PATH)":${LD_LIBRARY_PATH}

Documentation

Documentation is available either in the source tree (doc/), or online.

Testing

rpy2 uses pytest, with the plugin pytest-cov for code coverage. To test the package from the source tree, either to check and installation on your system or before submitting a pull request, do:

pytest tests/

For code coverage, do:

pytest --cov=rpy2.rinterface_lib \
       --cov=rpy2.rinterface \
       --cov=rpy2.ipython \
       --cov=rpy2.robject \
       tests

For more options, such as how to run specify tests, please refer to the pytest documentation.

License

RPy2 can be used under the terms of the GNU General Public License Version 2 or later (see the file gpl-2.0.txt). This is the very same license R itself is released under.