Currently main group of poetry dependencies is not installed within CI, it should be added.

Currently when Python projects are run, poetry.lock is automatically
updated during linting (and other tests) as the pipeline is using
pre-commit.

We should disable tests (as these are run by separate workflow)
and poetry lock automated check (as these should be mostly local).

Currently __init__.py is exempt from the import checks, #noqa is redundant and actually remove by ruff during first commit.

Feedback from the first failing test suites might not be enough, hence we should switch to
to running all test runners even if some of them fail.

Currently all of the test have Python 3.x (3.11) for the test runners, while these should be tailored to specific Python version within os.matrix during CI.

Check whether scraping flake and flakeheaven for ruff would not be the best idea to proceed.

Difficulty will emerge when choosing rules as most of that has to be hand verified.

Currently precommit runs on all files in some cases, which is unnecessary (as only the ones affected
should be checked).

pytest should be run on the whole test suite, as it is currently impossible to infer whether
any changes in src influenced existing tests.

yamllint has no files provided during CI run, simply add .

Issue:

The style in which violation messages are presented by Ruff is indicated by the keyword "format" inside pyproject.toml, as shown below:

[tool.ruff]
format = "grouped"

Previously, this configuration used to work in our template. However, recently, it causes a failure with the following error:

Cause: TOML parse error at line 126, column 10
        |
126  | format = "grouped"
        |
invalid type: string "grouped", expected struct FormatOptions

Upon investigating the documentation, I found that the correct keyword should be "output-format" instead of "format".

Proposal:

I suggest updating the configuration in pyproject.toml to reflect the correct keyword "output-format" as follows:

[tool.ruff]
output-format = "grouped"

This change should resolve the parsing error and align with the documentation's requirements.

• Do not fetch all commits for code linting & testing
• Specify folders to lint explicitly
• Make tests verbose

Define a few aliases which might come handy when working with py-template.

Currently coverage is bugged and calculated across test folder instead of src

Currently some actions are run only when /tests are changed, these references should be changed to /test.

Only pytype does not support Python 3.11 currently, we can
specify this dependency explicitly.

Adjustments to installed packages within tests might have to be
required due to above (e.g. installation of specific additional
dependencies instead of --all-extras)

Currently code-linter checks docstrings coverage, we can safely
remove this action.

Currently Python version support is not unified across py-template.

In order to align with most of the (scientific) Python, we should settle on
Python 3.9 or above, according to NEP29

Remove leftover rules from flake8

Forward existing files and check their compliance with current YAML standard

Currently pytest tests will raise error if any assert is used.

Exclude ruff S101 for these files.