hammer-pathways 0.2.0 — sklearn integration

Add optional MeanRankTransformer for scikit-learn pipelines, cloning, cross-validation and named outputs. Named input columns are checked; changing the feature definition requires refitting. The base NumPy/SciPy functions remain unchanged.

25 package tests pass on Python 3.10, 3.12 and 3.14. Includes a synthetic pipeline example.

Install the wheel with python -m pip install hammer_pathways-0.2.0-py3-none-any.whl. For the adapter, then run python -m pip install "hammer-pathways[sklearn]==0.2.0" in the same environment. Verify the wheel with SHA256SUMS.

No models or datasets are included. This is feature engineering, not a clinical model or a guarantee against leakage. Gene-set selection and evaluation design remain the researcher's responsibility. Not published on PyPI.

hammer-pathways 0.1.0

Initial MIT-licensed release of specimen-local gene ranking and mean-rank gene-set features.

Includes an installable wheel, SHA-256 checksum, typed public functions, synthetic example, tests and contribution instructions. Runtime dependencies are NumPy and SciPy. GitHub checks passed on Python 3.10, 3.12 and 3.14.

Install the downloaded wheel with python -m pip install hammer_pathways-0.1.0-py3-none-any.whl. Verify it against SHA256SUMS first.

No model weights or datasets are included. These are numerical feature transforms, not a diagnostic model or a pathway enrichment test. Not published on PyPI.