Improve precommit config to apply fixes locally

[wjbenfold]

🚀 Pull Request

Description

See #205 for context. Implements something similar to SciTools/iris#4235 for cf-units

[codecov]

Codecov Report

Merging #209 (2e69fd6) into main (63d5b2c) will not change coverage.
The diff coverage is n/a.

@@           Coverage Diff           @@
##             main     #209   +/-   ##
=======================================
  Coverage   91.05%   91.05%           
=======================================
  Files           6        6           
  Lines         816      816           
  Branches      121      121           
=======================================
  Hits          743      743           
  Misses         61       61           
  Partials       12       12           

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 63d5b2c...2e69fd6. Read the comment docs.

[trexfeathers]

@wjbenfold have you been able to work out that cf-units doesn't have any paths to exclude like Iris does? That was the main thing I was worried about.

#209 (comment)

[wjbenfold]

I'm not confident of that, no. So far just happy that it appears to be doing the same things at the same time as I'd expect from Iris.

Are there any other things you'd like checking that I might not have thought of (by which I mean anything at all that you'd check)?

[trexfeathers]

Scratch what I said before! It wasn't excluded files it was excluded hooks that I was thinking of.

Are there any other things you'd like checking that I might not have thought of (by which I mean anything at all that you'd check)?

1. Renaming stuff that says linting to pre-commit, to make it clear that it's doing whatever pre-commit is doing.
2. Make sure the list of pre-commit hooks that get run excludes those that would be harmful/non-sensical.

Sorry I see now that I jumped the gun and hadn't realised you were still working through the PR 😳

[wjbenfold]

Definitely my bad for not making it draft. I'll get on those things now

[wjbenfold]

Make sure the list of pre-commit hooks that get run excludes those that would be harmful/non-sensical.

The only one that Iris skips and that thus seems necessary to skip is no-commit-to-branch. I've told the CI to do that, and in the CI pre-commit check it shows up as skipped, but the pre-commit.ci check shows is Passed rather than Skipped. Is the pre-commit stuff running twice?

I've just checked, and the same things happens in Iris' CI, so I don't think it's a bad thing here

[trexfeathers]

Great stuff @wjbenfold , thanks so much!

And I can confirm that this is doing what I would expect, too.