feat(fit): add sample_weight

[tczhao]

Added sample_wegith to .fit()

Thanks for the fantastic work.

I found that half of the sample_weight is implemented but not expose to the final fit API.

The sample_weight can make a big difference (90% -> 95%) despite the surprising stablity of the deep forest on skewed classification.

[xuyxu]

Thanks for the PR @tczhao. I will take a look latter ;-)

In addition, how about adding two things in your PR?

• Numpy style docstrings below the function fit
• Your contribution and name in CHANGELOG.rst. I think the corresponding badge should be Feature.

I will stay tuned, and let me if you have any problem.

[xuyxu]

Great, we are close to the merge! I will check the entire workflow and merge this PR if there is no problem. Thanks for your nice codes!

[xuyxu]

Merged, thanks for your nice codes @tczhao ;-)