Contributions are welcome when they improve validation, reproducibility, model robustness, or documentation.
python -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt
python -m pytest -q
python -m py_compile deconvolve.py external_validation.pyThe full deconvolution run downloads public data and may take longer than unit tests. Use synthetic tests for small pull requests.
python -m pytest -qpasses.python -m py_compile deconvolve.py external_validation.pypasses.- Model architecture, HVG selection, cell-type filtering, or validation changes are documented in the README.
- Claims about accuracy include the dataset, split strategy, metric, and limitation.
Pseudo-bulk validation is an upper bound. Do not present it as matched clinical validation. If you add or change external validation, report both positive and negative findings.
By contributing, you agree that your contributions are licensed under the MIT licence.