Adding GroupKFold and testing it

[hyun-seo]

Referring to #157, there is an issue about group cross-validation.

This pull request implementation this issue by adding new argument kfolds_group and
using sklearn.model_selection.GroupKFold.

[mandjevant]

Requesting these initial changes for it to go through the automated testing.
Once it gets through, Ill take a closer look

[hyun-seo]

@mandjevant
Thank you so much for taking the time to review my pr.
This is my first attempt at contributing to open source and I had a lot of issues, but I think I learned a lot from your kind help.
I have one question: what is the exact meaning of # noqa: E226 in the testcode? Why is this phrase needed in the test?

[mandjevant]

@mandjevant Thank you so much for taking the time to review my pr. This is my first attempt at contributing to open source and I had a lot of issues, but I think I learned a lot from your kind help. I have one question: what is the exact meaning of # noqa: E226 in the testcode? Why is this phrase needed in the test?

Thank you for your contribution!
The reason your PR did not go through automated testing was because of flake8 style guide enforcement. In most lines, flake was correct and notified us of style issues that we resolved. But in that one line, I deemed that flake8 was wrong.
It threw an inconsistency with rule E226 which, when enforced, would make the line unreadable.

Thus, I suppressed the rule on (only) that line by adding a comment stating noqa: E226.

[mandjevant]

In addition, please change

n_folds: int, default is None
            When n_folds is not None, use K-fold cross-validation when
            n_folds > 2. Or, use leave-one-out cross-validation when
            n_folds = -1.

to

n_folds: int, default is None
            When n_folds is not None, use K-fold cross-validation when
            n_folds > 2. Or, use leave-one-out cross-validation when
            n_folds = -1. For Group K-fold cross-validation, functions as
            `n_splits`.

In both places:

https://github.com/hyperopt/hyperopt-sklearn/blob/9938225c7d4dd85fbc45be12aa2f74cdde527ad6/hpsklearn/estimator/estimator.py#L244C30-L244C30

https://github.com/hyun-seo/hyperopt-sklearn/blob/9938225c7d4dd85fbc45be12aa2f74cdde527ad6/hpsklearn/estimator/_cost_fn.py#L60C26-L60C26

Function and test case look appropriate

[mandjevant]

Looks good! Thanks for your contribution!
@bjkomer Any thoughts?

[mandjevant]

Ill go ahead and merge this. Thank you for your contribution @hyun-seo