Add causalml support

[alexander-pv]

Overview

Hi!
This PR is reopened #2654

Checklist

• All pre-commit checks pass.
• Unit tests added (if fixing a bug or adding a new feature)

[connortann]

The tests are failing due to an unrelated HuggingFace error: apologies. We'll get that sorted asap (on #3286), and then we can get this PR passing.

Edit: done.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Comparison is base (8130ec4) 57.79% compared to head (61a74f0) 57.75%.
Report is 71 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #3273      +/-   ##
==========================================
- Coverage   57.79%   57.75%   -0.05%     
==========================================
  Files          89       89              
  Lines       12540    12528      -12     
==========================================
- Hits         7248     7236      -12     
  Misses       5292     5292              

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

[connortann]

This PR has quite a large diff, as there are many unrelated style changes alongside the functional changes. Please would you remove the style changes, leaving just the functional changes?

We plan to introduce python black soon to remove this kind of hassle for contributors.

Nb. I inadvertently briefly closed this PR (currently on mobile), apologies.

[alexander-pv]

Apologies for style change. I will remove it.

Python black usage would be great!

[thatlittleboy]

Hi @alexander-pv , do you think you could also add in some tests to test_tree.py (just a simple function or two to demonstrate that the proposed changes here work as expected?)

Thanks!

[connortann]

I'll add the "stale" label for now as there hasn't been recent activity. Please feel free to comment if you're still interested in the PR and are happy to work on adding a minimal unit test.

[alexander-pv]

Hi, @connortann! I added several test to check if shap works as intended with causalml trees.

[alexander-pv]

About current scikit-learn building issue in recent tests: scikit-learn/scikit-learn#26858

[connortann]

Apologies that we haven't had time to look further at this PR, as many maintainers are away over the Christmas period. Thanks again for the PR - I hope to get to this early in the new year.

[alexander-pv]

Hi, @connortann! I updated the PR resolving a couple of conflicts. Tests work as expected on my computer.
So, awaiting feedback from label can be removed now. I will be glad to contribute if you think anything else is needed for this PR!

Thanks!

[jeongyoonlee]

Hi, @connortann. Can you take a look at this PR? This has been a long-standing issue (e.g., uber/causalml#265, uber/causalml#527, uber/causalml#615, uber/causalml#727, uber/causalml#735) for the causalml package, and we hope to resolve it soon.

Your support will be greatly appreciated.