ENH Support sample weights in PartialDependenceDisplay.from_estimator

[vitaliset]

This PR fixes #24872.
Built on top of #25209.

As partial dependence of a model at a point is defined as an expectation, as #24872 points out, it should respect sample_weight if someone wishes to use it (for instance, when you know your X does not follow the distribution you are interested in).

The prior PR (#25209) incorporated this logic into the partial_dependence_plot function. This current PR completes the integration by incorporating it into PartialDependenceDisplay.from_estimator.

[github-actions]

✔️ Linting Passed

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

Generated for commit: 956002d

[vitaliset]

Kindly pinging @glemaitre, @lorentzenchr, @jeremiedbb here as you might be interested and already familiar with the previous PR.
Maybe we can add a 1.3 milestone to attract reviewers. This PR should be easy to review and easy for me to edit. ;)

I wasn't creative in the test, as it was unclear to me what we want to evaluate. Therefore, any suggestions are welcome!

[glemaitre]

It looks good to me.
@jeremiedbb I think we need this one not to have a discrepancy between partial_dependence and the display in 1.3.

[jeremiedbb]

I'm okay with that

[thomasjpfan]

LGTM

[vitaliset]

I just updated the branch with main. CI is green. I'm ok with it being merged as is. :)

[jeremiedbb]

Thanks @vitaliset