fix: KeyError no longer occurs when using groupfolds for regression tasks.

[dannycg1996]

Why are these changes needed?

Currently KeyErrors are raised (under certain conditions outlined in the linked issue) when using AutoML for regression tasks, with GroupKFold splits.
This was caused by the index not being reset for group folds.

As part of this PR, I've modified the if statement in prepare_data to ensure that we now also reset the index, regardless of the split type. This will ensure that this error doesn't occur when group folds are used, or when custom splitters are passed in. I've also added a test, which checks that my example code from the linked issue can now run without errors.

Hopefully that's merge-able - please let me know if you have any suggestions/comments.
Thanks!

Related issue number

Close #1384
Closes #1254

Checks

• I've used pre-commit to lint the changes in this PR (note the same in integrated in our CI checks).
• I've included any doc changes needed for https://microsoft.github.io/FLAML/. See https://microsoft.github.io/FLAML/docs/Contribute#documentation to build and test documentation locally.
• I've added tests (if relevant) corresponding to the changes introduced in this PR.
• I've made sure all auto checks have passed.

[thinkall]

Thank you @dannycg1996 for the PR. Please see my comments in #1384 .

[thinkall]

Since we're allowing group for regression task. Do you mind update the related docstring as well? Thanks.

[dannycg1996]

Thanks @thinkall - I have updated the docs. Please let me know if I've missed anything!
Can I please ask what the plans are for future releases? We are building a project that depends on the latest FLAML changes.
Thanks again!

[thinkall]

Thanks @thinkall - I have updated the docs. Please let me know if I've missed anything! Can I please ask what the plans are for future releases? We are building a project that depends on the latest FLAML changes. Thanks again!

I can make a release once your PR is merged.

[dannycg1996]

That's great - thanks!