Run all commands from the repository root.
When you need a runnable Virne environment, install the runtime dependencies using the installation guide, then register the checkout in editable mode without changing the selected CPU/CUDA packages:
python -m pip install --no-deps --editable .Then install the documentation toolchain:
python -m pip install -r docs/requirements.txtFor a reproducible CPU development environment, use the committed lock file:
uv sync --locked --group docsThis installs Virne, the test tools, and the documentation toolchain from
uv.lock. Use install.sh instead when selecting a CUDA build or installing
the optional pyg_lib acceleration wheel.
For a documentation-only preview, installing docs/requirements.txt is sufficient. Sphinx mocks optional machine-learning and optimization packages; NumPy and PyYAML remain real because Virne uses them while importing the registry and configuration utilities.
Read the Docs performs the project and documentation installation automatically through .readthedocs.yaml.
sphinx-build -W --keep-going -b html docs/source docs/build/htmlpython -m http.server 8000 --bind 127.0.0.1 --directory ./docs/build/html