✔️ Linting Passed

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

_{Generated for commit: 3b3a034. Link to the linter CI: here}

OK so the reason is that we should not ravel the predictions and the true labels and only let the scorer the average. In this case, they all be "sample aligned" based on n_samples.

I added a non-regression test.

I don't think that we need to have an entry in the changelog since before we were not using sample_weight to compute the score (which was another bug).

ping @OmarManzoor @StefanieSenger @adrinjalali that already have a good understanding if this part of the code.

FYI the original error would have been something like:

E           ValueError: Found input variables with inconsistent numbers of samples: [200, 200, 100]

because the raveling would transform the predictions and true label from a shape of (100, 2) to (200,) while the sample_weight is always an array of (100,).

I can add an additional test by defining a custom scorer that would not do a simple uniform average between the target. In this case, it should fail in main but work in this PR.