Release and publication
SigProfilerPlotting uses Git tags as its version source. setuptools-scm
derives the package version from tags matching vX.Y.Z; maintainers should
not manually edit a version constant for a release.
Before tagging
- Merge the intended release changes into the target branch, normally
master. - Confirm CI passes for Python 3.9 and Python 3.12.
- Confirm the documentation builds without warnings.
- For changes to SBS4608 static output, run the real renderer check:
- Choose a version that has not already been used on GitHub, TestPyPI, or PyPI. Published package versions cannot be replaced.
Create the version tag and TestPyPI release
- Open Actions > Create version tag > Run workflow.
- Enter the version without a leading
v, for example1.5.0. - Select the branch to tag, normally
master. - Run the workflow.
The workflow creates and pushes the annotated vX.Y.Z tag, then dispatches
the Release workflow. That workflow:
- builds the source distribution and wheel;
- validates both with Twine;
- installs the wheel and verifies the derived version;
- stores the distributions as a workflow artifact; and
- publishes the tagged build to TestPyPI.
TestPyPI publishing uses the protected GitHub environment named testpypi
and PyPI trusted publishing (OIDC), not a stored API token.
Publish to PyPI
After validating installation and basic plotting from TestPyPI, create and
publish a GitHub Release for the same vX.Y.Z tag. The published-release
event runs the verified build again and publishes it to PyPI through the
protected GitHub environment named pypi and PyPI trusted publishing.
The PyPI trusted publisher must match exactly:
| Setting | Value |
|---|---|
| Repository owner | SigProfilerSuite |
| Repository name | SigProfilerPlotting |
| Workflow | release.yml |
| Environment | pypi |
Configure the equivalent TestPyPI publisher with environment testpypi.
Post-release verification
- Confirm the GitHub Actions release run is green.
- Install the release into a clean environment:
python -m pip install --upgrade SigProfilerPlotting==X.Y.Z
python -c "import sigProfilerPlotting as spp; print(spp.__version__)"
- Generate at least one standard plot and, when applicable, one SBS4608 static PDF or dashboard plot.
- Confirm the GitHub Release, PyPI project, and documentation links point to the intended version.