BLD: Run flake8 check on Cython files in pre-commit

[xhochy]

• closes pre-commit check for flake8 doesn't check Cython code #30843
• tests added / passed
• passes black pandas
• passes git diff upstream/master -u -- "*.py" | flake8 --diff
• whatsnew entry

[datapythonista]

Not sure if it's a typo or it's intentional, but we used to validate only pyx files for these errors, and the CI is failing because there are pxd and pxi files that got errors.

[xhochy]

Yes this is a typo. I separated it out to another .flake8, pushed the fix.

[xhochy]

@datapythonista Fixed the configuration, CI now passes.

[datapythonista]

Looks good, I'd probably have the flake config files somewhere else, since we've already got too much in our root, but other than that looks good. I'll let someone else to review, since I'm not using pre-config myself.

@TomAugspurger if you're not busy with the RC, I guess you may want to have a look here.

[gfyoung]

Looks good, I'd probably have the flake config files somewhere else, since we've already got too much in our root, but other than that looks good.

I think a flake8 directory would make sense. Would also help to simplify filenames.

[xhochy]

flake8 or .flake8?

[gfyoung]

flake8 (no dot) should be fine I think. Keep things simple.

[datapythonista]

lgtm

[WillAyd]

Can the template files go in with the normal pyx files? I think they would be closer to that than the header files

[WillAyd]

@xhochy can you check this? I think good to go otherwise

[xhochy]

pxd can go with the normal pyx files but the pxi.in raise a lot of errors, especially with indentation. I guess that the current indentation is actually sometimes preferred to the flake8 one as it makes the templates more readable.

[jreback]

@jbrockmendel can you try this out

[jbrockmendel]

I'm missing something from the workflow. I've checked out this branch, added a flake-violation to groupby.pyx, but commiting it doesnt seem to cause any problem. What am I missing?

Can W293 for trailing whitespace be added?

Not a deal-breaker, but I like the more verbose way we list codes in statsmodels: https://github.com/statsmodels/statsmodels/blob/master/setup.cfg#L73

[xhochy]

I'm missing something from the workflow. I've checked out this branch, added a flake-violation to groupby.pyx, but commiting it doesnt seem to cause any problem. What am I missing?

Can W293 for trailing whitespace be added?

Not a deal-breaker, but I like the more verbose way we list codes in statsmodels: https://github.com/statsmodels/statsmodels/blob/master/setup.cfg#L73

The obvious one:
Did run pre-commit install once?

Otherwise: Can you tell me what you did change? Then I can try to reproduce locally.

[xhochy]

rebased and CI passed.

[WillAyd]

lgtm

[jbrockmendel]

Did run pre-commit install once?

Looks like brew install pre-commit followed by pre-commit install did the trick, thanks.

LGTM

W293 can be follow-up.

[WillAyd]

Thanks @xhochy