Improve SmartCorrelatedSelector consistency with ties

[ClaudioSalvatoreArcidiacono]

Fixes #824

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 98.04%. Comparing base (2738032) to head (d0f85e3).
⚠️ Report is 1 commits behind head on main.

Additional details and impacted files

@@           Coverage Diff           @@
##             main     #825   +/-   ##
=======================================
  Coverage   98.04%   98.04%           
=======================================
  Files         111      111           
  Lines        4653     4653           
  Branches      742      742           
=======================================
  Hits         4562     4562           
  Misses         56       56           
  Partials       35       35           

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

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[solegalli]

Hi @ClaudioSalvatoreArcidiacono

Hope you are doing well and thanks for the patience regarding this PR.

We are releasing a new version of feature-engine soon and we could try and make this PR part of it.

I am happy to accept the change in the sorting algorithm, which will resolve the issue for most cases. But I don't want to pass from sets to lists. That will break backward compatibility.

Could you please rebase main and update the PR with these changes?

[ClaudioSalvatoreArcidiacono]

Hey @solegalli, I have updated the MR as you requested, could you take a second look?

The inconsistencies with ties will still remain in case we select "model_performance" as a selection metric but it should be solved for all other selection methods.

[solegalli]

Thanks a ton!