formulae is a Python library that implements Wilkinson's formulas for mixed-effects models. The main difference with other implementations like Patsy or formulaic is that formulae can work with formulas describing a model with both common and group specific effects (a.k.a. fixed and random effects, respectively).
This package has been written to make it easier to specify models with group effects in Bambi, a package that makes it easy to work with Bayesian GLMMs in Python, but it could be used independently as a backend for another library. The approach in this library is to extend classical statistical formulas in a similar way than in R package lme4.
formulae requires a working Python interpreter (3.7+) and the libraries numpy, scipy and pandas with versions specified in the requirements.txt file.
Assuming a standard Python environment is installed on your machine (including pip), the latest release of formulae can be installed in one line using pip:
pip install formulae
Alternatively, if you want the development version of the package you can install from GitHub:
pip install git+https://github.com/bambinos/formulae.git
The official documentation can be found here
- The
dataargument only accepts objects of classpandas.DataFrame. y ~ .is not implemented and won't be implemented in a first version. However, it is planned to be included in the future.