HOI (Higher Order Interactions) is a Python package to go beyond pairwise interactions by quantifying the statistical dependencies between 2 or more units using information-theoretical metrics. The package is built on top of Jax allowing computations on CPU or GPU.
HOI requires :
- Python (>= 3.8)
- numpy(>=1.22)
- scipy (>=1.9)
- jax
- pandas
- scikit-learn
- jax-tqdm
- tqdm
To install Jax on GPU or CPU-only, please refer to Jax's documentation : https://jax.readthedocs.io/en/latest/installation.html
If you already have a working installation of NumPy, SciPy and Jax,
the easiest way to install hoi is using pip:
pip install -U hoiYou can also install the latest version of the software directly from Github :
pip install git+https://github.com/brainets/hoi.gitFor developers, you can install it in develop mode with the following commands :
git clone https://github.com/brainets/hoi.git
cd hoi
pip install -e .['full']The full installation of HOI includes additional packages to test the software and build the documentation :
- pytest
- pytest-cov
- codecov
- xarray
- sphinx!=4.1.0
- sphinx-gallery
- pydata-sphinx-theme
- sphinxcontrib-bibtex
- numpydoc
- matplotlib
- flake8
- pep8-naming
- black
- Link to the documentation: https://brainets.github.io/hoi/
- Overview of the mathematical background : https://brainets.github.io/hoi/theory.html
- List of implemented HOI metrics : https://brainets.github.io/hoi/api/modules.html
- Examples : https://brainets.github.io/hoi/auto_examples/index.html
For questions, please use the following link : https://github.com/brainets/hoi/discussions
HOI was mainly developed during the Google Summer of Code 2023 (https://summerofcode.withgoogle.com/archive/2023/projects/z6hGpvLS)