Installation

sheaf_mpnn requires Python >= 3.13 and uses uv for dependency management.

Base environment

git clone https://github.com/alessioborgi/pytorch-SheafNeuralNetworks
cd pytorch-SheafNeuralNetworks
uv sync

This installs both sheaf_mpnn (the core library) and exp (the experiment runner). Activate the venv, then verify:

python -c "import sheaf_mpnn; print(sheaf_mpnn.__version__)"

W&B logging and the Optuna-W&B sweep integration are part of the base environment, so no extra install step is needed for sweeps.

Optional groups

Group	Command	Provides
dev group	uv sync --dev	tests, ruff, mypy, pre-commit
docs group	uv sync --group docs	Sphinx, pydata-sphinx-theme, MyST, autodoc extensions

The docs group is what CI uses to build this site; see the docs CI workflow.