kafe2 is an open-source Python package for the likelihood-based estimation of model parameters from measured data. As the spiritual successor to the original kafe package it aims to provide state-of-the-art statistical methods in a way that is still easy to use. More information here.
If you have installed pip just run
pip install kafe2to install the latest stable version and you're (mostly) ready to go. The Python package iminuit which kafe2 uses internally for numerical optimization may fail to be installed automatically if no C++ compiler is available on your system . While iminuit is strictly speaking not required its use is heavily recommended. Make sure to read the pip installation log. As of kafe2 v2.4.0 only Python 3 is supported. kafe2 works with matplotlib version 3.4 and newer.
The documentation under kafe2.readthedocs.io has more detailed installation instructions. It also explains kafe2 usage as well as the mathematical foundations upon which kafe2 is built.
If you prefer a more practical approach you can instead look at the various examples. In addition to the regular Python/kafe2go files there are also Jupyter notebook tutorials (in English and in German) that mostly cover the same topics.
If you encounter any bugs or have an improvement proposal, please let us know by opening an issue here.