Release Checklist¶
Use this before tagging a DatasetLint release.
Run the automated release gate first:
make release-check
mkdocs build --strict
Use PYTHON=/path/to/python when the dev dependencies are installed in a specific
environment, and WHEEL_DIR=/tmp/datasetlint-wheel when you want wheel artifacts outside
the worktree.
- Run
pytest. - Run
ruff check .. - Run
mypy datasetlint. - Run
mkdocs build --strict. - Build distributions with
python -m build. - Check distributions with
python -m twine check dist/*. - Confirm
datasetlint --version. - Confirm
datasetlint --help. - Confirm
datasetlint examples/minimal_dataset. - Confirm
datasetlint examples/bad_dataset. - Confirm
datasetlint diff examples/minimal_dataset examples/bad_dataset --fail-on-regression. - Confirm generated sample reports are current:
examples/reports/minimal_report.jsonexamples/reports/bad_report.jsonexamples/reports/bad_report.md- Confirm README limitations are near the top and still match adapter coverage.
- Lint
examples/minimal_datasetand confirm it passes. - Lint
examples/bad_datasetand confirm it reports errors. - Review
README.md,docs/quickstart.md, anddocs/checks.mdfor CLI/API drift. - Confirm
pyproject.tomlversion, Python requirement, dependencies, and console script. - Confirm
LICENSEis present and uses MIT terms. - Confirm
SECURITY.md,CITATION.cff, andPUBLISHING.mdare current. - Confirm the public GitHub repository is reachable and matches the release commit.
- Confirm the
testpypiGitHub environment and TestPyPI trusted publisher are configured for.github/workflows/publish.yml. - Run the
Publishworkflow manually withtarget: testpypifrom the release commit. - Install the release version from TestPyPI in a fresh virtual environment.
- Confirm the
pypiGitHub environment and PyPI trusted publisher are configured for.github/workflows/publish.yml. - Confirm the
datasetlintPyPI project is owned by the maintainer before announcingpip install datasetlintas the default install path. - Confirm GitHub Pages is enabled for the docs workflow.