[DISC] Having IbisML closer to sklearn

[gtauzin]

Following discussion in #135, I thought I would give it a try and set up a wrapper of IbisML steps as fully-fledged sklearn transformers so that they follow more closely the sklearn API.

Running the adapted example notebook, you can already check that :

• An ibis step wrapped into a transformer within the sklearn pipeline does not eagerly execute;
• get_params works as intended.

I'll edit this msg later with a list of more detailed points.

[codecov-commenter]

Codecov Report

Attention: Patch coverage is 35.29412% with 22 lines in your changes missing coverage. Please review.

Project coverage is 84.08%. Comparing base (80881f8) to head (e2f99c6). Report is 1 commits behind head on main.

Files with missing lines	Patch %	Lines
ibis_ml/estimators.py	33.33%	22 Missing :warning:

Additional details and impacted files

```diff @@ Coverage Diff @@ ## main #141 +/- ## ========================================== - Coverage 84.94% 84.08% -0.87% ========================================== Files 26 27 +1 Lines 1920 1954 +34 ========================================== + Hits 1631 1643 +12 - Misses 289 311 +22 ```

:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.

[gtauzin]

Closing as @jcrist made it clear that having Steps as sklean transformer would not be a good idea in #135 .