AlgoKit CLI CI/CD

[robdmoore]

Options:

• GitHub flow (pull request -> main) and we can optionally release a build to PyPi (by clicking a button in GitHub actions); vs
• Git Flow (pull request -> develop -> main) and a merge to main automatically

GitHub flow is simpler, Git flow allows for longer-running experimental branches.

We decided to use GitHub flow for simplicity and to encourage regular releasing. If there is a need for an experimental branch then it can be released by running the GitHub action against the branch and thus published as a pre-release.

Acceptance Criteria:

• When a pull request is opened to AlgoKit CLI a GitHub action is run that checks: the code typings are valid, formatting is valid according to the Python Black configuration, any tests (present) pass
• When there's a new commit on the main branch of the AlgoKit CLI a GitHub workflow is run that:
  1. Checks the code typings are valid
  2. Checks the formatting is valid according to the Python Black configuration
  3. Checks the linting is valid according to the TBD configuration
  4. Checks all tests pass
  5. Builds a PyPi Wheel package ready for publishing
  6. Releases a pre-release package to PyPi
• If the branch is not main then the branch name will be used as a pre-release tag, but otherwise it will be beta
• There is an ability for a product manager to trigger a "stable" release, which will deploy a new non-pre-release package to PyPi
  (NOTE: These last criteria are being completed in Releasing AlgoKit #16)

[mzaatar]

Use Poetry for building, Ruff for linting, PyTest for unit testing, and upload Wheel.
We will start simple with Ubuntu images and assume it will work cross-platform
Tox for later in CI/CD: test with supported Python versions (3.10 and 3.11)

Use PyPI GitHub Action: https://github.com/marketplace/actions/pypi-publish

Later on, think of the OS variants for building and testing.

[mzaatar]

Team decided on GitHub flow