Installation

pip install qsarkit-learn

Note

The distribution is qsarkit-learn; the package you import is qsarkit — the same split as scikit-learn and sklearn. The bare name qsarkit on PyPI belongs to an unrelated project.

To track the development version instead:

pip install git+https://github.com/omixlab/qsarkit-learn

Either installs the core: NumPy, SciPy, pandas, scikit-learn, RDKit, NetworkX and Plotly. Everything in qsarkit.chemistry, qsarkit.sar, qsarkit.applicability, qsarkit.neighbors and qsarkit.cluster works with just that.

Optional extras

Heavier features load their dependencies lazily, so you install only what you use. A missing dependency raises OptionalDependencyError, which names the extra to install.

Extra	Enables
embeddings	Mol2Vec (gensim)
nlp	ChemBERTa embeddings (transformers, torch)
explainability	SHAP and LIME attribution
boosting	XGBoost and LightGBM estimators
embedding_viz	UMAP chemical-space projections (umap-learn)
reporting	Static image export for reports (kaleido)

pip install "qsarkit-learn[explainability]"   # pick what you need
pip install "qsarkit-learn[all]"              # or take everything

A missing optional dependency raises an error naming the extra that provides it, rather than an ImportError you have to interpret.

Development install

git clone https://github.com/omixlab/qsarkit-learn
cd qsarkit-learn
pip install -e ".[dev]"

pytest                             # run the suite
pytest -m slow                     # execute the example notebooks too
pytest --cov=qsarkit --cov-branch  # with coverage
mypy qsarkit                       # strict type check
ruff check qsarkit                 # lint

Verifying a source distribution

The sdist ships the test suite, so a packager can verify a build without the repository:

tar xf qsarkit_learn-*.tar.gz
cd qsarkit_learn-*/
pip install ".[dev]"
pytest                             # the notebooks need `-m slow`

Building the documentation. The sources are in docs-sphinx/; the built site lives in docs/ and is committed, because that is what GitHub Pages serves:

pip install -e ".[docs]"

make docs        # rebuild and copy the site into docs/
make preview     # build into docs-sphinx/build/html, leaving docs/ alone
make doctest     # execute every example in this documentation
make linkcheck   # verify external links resolve

make docs builds from scratch with warnings treated as errors, so a broken cross-reference cannot reach the published site and a page deleted from the source cannot linger in it. The same targets are available from inside docs-sphinx/, where the one that writes to ../docs is called make publish.

Note

typings/ holds a local stub shadow for RDKit. The rdkit-stubs bundled with the RDKit wheel contain an auto-generation bug — a C++ enum member named None, which is an illegal annotation target in a .pyi — that otherwise aborts every mypy run and silently hides all real type errors. Keep it.