✔️ Linting Passed

All linting checks passed. Your pull request is in excellent shape! ☀️

_{Generated for commit: c271ae5. Link to the linter CI: here}

CI error is related.

Also, please avoid force pushing to your branch here, since it makes review harder.

Sorry about that.

I'm not familiar with PR that much. If I'm not pushing my branch, then how could I modify the code and update the PR?
Also, I'm not sure how to solve CI error, since there's nothing much I committed to the code in relation to that issue.

You push to your branch, but don't force push to the branch. Make sure you merge with main instead of rebasing if that's the issue.

If you look at the failing CI, you see this message:

>               raise AssertionError(msg)
E               AssertionError: The type specification of Parameter 'average' is inconsistent between
E               ['precision_recall_fscore_support'] and ['f1_score', 'fbeta_score',
E               'precision_score', 'recall_score']:
E               
E               *** ['precision_recall_fscore_support']
E               --- ['f1_score', 'fbeta_score', 'precision_score', 'recall_score']
E               ***************
E               
E               *** 1,8 ****
E               
E                 {'micro', 'macro', 'samples', 'weighted', 'binary'} or None,
E               ! default=None
E               --- 1,8 ----
E               
E                 {'micro', 'macro', 'samples', 'weighted', 'binary'} or None,
E               ! default='binary'

That's a test which makes sure things are consistently documented. In this case, you need to check if the other functions also need a change, or if only this one needs a change and instead the test needs to be fixed.

@adrinjalali

Thanks for the detailed guide, but I still need some more help.

1. I just pushed to my remote branch using git push and I'm not sure how it did force-push, which hinders me from making another commit
2. Where could I get that error script? I looked at https://github.com/scikit-learn/scikit-learn/actions/runs/14876302969/job/41774196534?pr=31270 and elsewhere I could check, but I think I'm missing this bit. All the error message I can see is related to TypeError: ufunc 'isnan' not supported for the input types, and the inputs could not be safely coerced to any supported types according to the casting rule ''safe'

It seems that the other related functions set default average='binary' whereas for precision_recall_fscore_support it is None (and has been the case for the last 11 years).

What do you think about this difference in default values @adrinjalali ? Note 'binary' gives you the values for the pos_label class, whereas None gives you all the values.

What do you think about this difference in default values @adrinjalali ? Note 'binary' gives you the values for the pos_label class, whereas None gives you all the values.

hmm, not sure. I wouldn't mind changing the default via a deprecation cycle to make them all consistent.

I just pushed to my remote branch using git push and I'm not sure how it did force-push, which hinders me from making another commit

Just make sure you do git merge instead of git rebase and that you don't do git push -f

And the error is in this CI run: https://dev.azure.com/scikit-learn/scikit-learn/_build/results?buildId=76333&view=logs&j=f71949a9-f9d9-549e-cf45-2e99c7b412d1&t=d5baef2b-8bda-5a4c-e848-157b5caff279

I wouldn't mind changing the default via a deprecation cycle to make them all consistent.

I was thinking the same. Maybe cc @glemaitre or @lorentzenchr may be interested to weigh in?