Example repo: using GitHub Actions for CI/CD for a Python project
This is a minimal repo demonstrating the use of GitHub Actions for Continuous Integration (CI) & Continous Deployment (CD) for a Python project.
The GitHub Actions are:
Tests: Automatically unit test code using
unittestfrom Python's standard library.
- Code coverage: Generate code coverage reports using coverage.py. The reports are automatically uploaded to codecov. The @codecov-commenter bot adds a comment to PR on code coverage status.
- Lint and format code: Lint using pylint. Check code formatting using black. Check type hints using mypy.
- Check docs build: Check Sphinx docs build successfully.
Deploy docs to GitHub pages: Automatically deploy docs to a
GitHub Pages repo upon
- Upload release to PyPI: Publish the latest version of the package on PyPI when a new GitHub Release is created.
Each of these actions is stored in a YAML file in the .github/workflows directory.
File system structure
Push Token for Deploying Docs
Action #5 requires setting up a Personal Access Token with full repo access via
the Developer Settings page. Store this token as an Actions secret under the
PUSH_TOKEN in the main repo (i.e. the repo where the docs source code is
held). See more here.
Useful resources on the topic
- mCoding's YouTube video on using GitHub actions for automated testing (see associated code repo).
- Alex Damiani's YouTube videos on automated testing ,  &  and associated code repos:  , & .
librosa is a real-world example of
- Pharmpy uses Actions for building docs then deploying them to a separate GitHub Pages repo.
- Vinod Kurup's blog post on automating PyPI releases with GitHub Actions.