✔️ Linting Passed

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

_{Generated for commit: 035441f. Link to the linter CI: here}

In process_routing we say:

    Assuming this signature: ``fit(self, X, y, sample_weight=None, **fit_params)``,
    a call to this function would be:
    ``process_routing(self, sample_weight=sample_weight, **fit_params)``.

It looks incorrect then because we should not pass sample_weight when they are None.

Would we want to enforce this contract?

Only non-None metadata are routed. All metadata whose value is None, will be ignored.

I think it would be very odd for None to actually be a non-default value which has a side effect.

Oh, I see why CalibratedClassifierCV doesn't raise the same error. Here:

we only raise if anything is passed but is not either requested, or is a self accepted metadata. Therefore it doesn't matter if we pass sample_weight there or no, there will be no errors anyway. And passing it as None is of no consequence since it's not gonna be included in any routed param set, since it's not requested by the children. So the code is good as it is.

I think there should also be tests for any meta-estimator, like Pipeline, that takes several sub-estimators or sub-transformers, that might or might not support that metadata.

I think there should also be tests for any meta-estimator, like Pipeline, that takes several sub-estimators or sub-transformers, that might or might not support that metadata.

I think those are tested in the common metadata routing tests @StefanieSenger

@adrinjalali
I meant to add tests if Pipeline and other meta-estimators, that accept lists of several sub-estimators, will also accept a mix of consumers and non-consumers, for instance in test_pipeline.py.
Those newly added tests here in test_metaestimators_metadata_routing.py don't do that.

Yes, but I think those are already tested in the test_metadata_routing.py. We haven't had an issue yet which comes from having a mix of subestimators that are consumers and non-consumers.

cc @OmarManzoor @thomasjpfan for a quick review maybe? This is for the coming patch release.