[ENH]: Optimise step wrapping

[fraimondo]

Currently, if a step does apply to anything different that "continuous", it will be wrapped in a column transformer.

This is unoptimal as there is an overhead (which also invoves joblib) when this happens.

This PR fixes that by really checking the apply_to of the step with the X_types of the data. If the step really applies to a subset of features, then it will wrapped in a column transformer.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 89.93%. Comparing base (eac0462) to head (f42ed99).
Report is 9 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #274      +/-   ##
==========================================
+ Coverage   89.89%   89.93%   +0.04%     
==========================================
  Files          56       56              
  Lines        2483     2493      +10     
  Branches      504      443      -61     
==========================================
+ Hits         2232     2242      +10     
- Misses        164      165       +1     
+ Partials       87       86       -1     

Flag	Coverage Δ
docs	100.00% <ø> (ø)
julearn	89.92% <100.00%> (+0.04%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Files with missing lines	Coverage Δ
julearn/config.py	100.00% <100.00%> (ø)
julearn/inspect/_pipeline.py	100.00% <100.00%> (ø)
julearn/pipeline/pipeline_creator.py	86.95% <100.00%> (+0.48%)	⬆️
julearn/transformers/ju_column_transformer.py	93.18% <100.00%> (+0.15%)	⬆️

... and 1 file with indirect coverage changes

[synchon]

LGTM, the docs are failing, once fixed, good to go!

[github-actions]

PR Preview Action v1.4.8
Preview removed because the pull request was closed.
2024-10-17 09:45 UTC