Add fit parameters passing

[sami-ka]

Description

In the MAPIE code, the fit function does not allow to pass parameters usable at fit time for base estimator. This is the case for early stopping in LightGBM for instance.

Fixes partially #212 (fit_params only)

Type of change

• New feature (non-breaking change which adds functionality)

How Has This Been Tested?

Please describe the tests that you ran to verify your changes. Provide instructions so we can reproduce. Please also list any relevant details for your test configuration

• Add test with passing monitor parameter in fit function of GradientBoostingRegressor
• Check that the other tests still work

Checklist

• I have read the contributing guidelines
• I have updated the HISTORY.rst and AUTHORS.rst files
• Linting passes successfully : make lint
• Typing passes successfully : make type-check
• Unit tests pass successfully : make tests
• Coverage is 100% : make coverage
• Documentation builds successfully : make doc

[codecov-commenter]

Codecov Report

All modified and coverable lines are covered by tests ✅

Comparison is base (614293e) 100.00% compared to head (a4adcb6) 100.00%.
Report is 177 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff             @@
##            master      #391    +/-   ##
==========================================
  Coverage   100.00%   100.00%            
==========================================
  Files           39        39            
  Lines         4616      4815   +199     
  Branches       487       800   +313     
==========================================
+ Hits          4616      4815   +199     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[LacombeLouis]

Hi @sami-ka,
This is great! Thank you for contributing to MAPIE.
I see that you've added for all the instances for **fit_params.
One small issue is that you've currently only implemented one test. I suggest adding tests for classification and regression.
Thank you!

[thibaultcordier]

Thank you very much for this proposal, which will extend the use of MAPIE to more estimators! Here is my feedback to complete this PR:

• add unit tests for calibration, classification and regression (as you did for quantile regression). No need for time series.
• some writing style suggestion.

[thibaultcordier]

All right, then! We're about to wrap up this PR. I'd like to suggest a few changes to help us maintain your code in the future.

I see that we've instantiated the gb object at the start of the test_quantile_regression.py file, but I think it would be better to do this directly in your tests in the other files (as it won't be used elsewhere except for test_quantile_regression.py).

[sami-ka]

All right, then! We're about to wrap up this PR. I'd like to suggest a few changes to help us maintain your code in the future.

I see that we've instantiated the gb object at the start of the test_quantile_regression.py file, but I think it would be better to do this directly in your tests in the other files (as it won't be used elsewhere except for test_quantile_regression.py).

You are right, it would definitely be more relevant inside the tests. Thanks for the suggestion!

[thibaultcordier]

Thank you very much @sami-ka for your qualitative contribution to MAPIE. As far as I'm concerned, everything's fine, even if I added a comment (thanks in advance if you can add this point). @vincentblot28 , @LacombeLouis if you want to see before the merge.

[sami-ka]

@thibaultcordier Is there something else to do? Could this PR be merged?

[thibaultcordier]

@sami-ka I'll let @LacombeLouis confirm that he has no more change requests. He will then merge your PR.