CI: Improve API of --ignore_errors in validate_docstrings.py

[datapythonista]

Follow up of #57879, to use --ignore_errors PR02,SA01 instead of --ignore_errors * PR02,SA01.

I also make the script ignore the specified errors when running for a single document (i.e. validate_docstrings.py pandas.Series -i SA01 was reporting the SA01 error before this PR, and it won't with the changes here).

CC: @mroeschke

[dontgoto]

It's probably overly safe, but one could verify that only one CLI parameter is specifying globally ignored errors, otherwise one set would overwrite the other

[datapythonista]

It's probably overly safe, but one could verify that only one CLI parameter is specifying globally ignored errors, otherwise one set would overwrite the other

It shouldn't overwrite it, I'm not sure if I explicitly tested it, but the way I implemented was to allow to do validate_docstrings.py -i PR01 -i SA01 and ignore each of them:

ignore_errors[None].update(set(error_codes.split(",")))

What I don't check is if there are duplicates, but that's not likely going to be useful since we just have one error being fully ignored, and I don't think we'll be adding any or many more errors to the script.

What I'm not sure is what happens if an object (e.g. pandas.Series.head) contains a typo, maybe it's ignored silently.

[mroeschke]

Thanks @datapythonista

[jordan-d-murphy]

Thanks for all the work on this @datapythonista! code_checks.sh is looking awesome with thse new improvements. Just wanted to check and make sure this was the last change you had pending. If this is the final update, I can go ahead and open new issues to address these.

[datapythonista]

Yes, no more changes to the ignore functions in code_checks.sh. You should also have the triage permissions now.

[jordan-d-murphy]

Opened DOC: Enforce Numpy Docstring Validation (Parent Issue) #58063 as a parent issue for fixing docstrings based on the refactoring in code_checks.sh

Feel free to swing by and help out! 🙂