initial implementation of TriangleSelector

[henrydingliu]

Summary of Changes

Related GitHub Issue(s)

Additional Context for Reviewers

• I passed tests locally for both code (uv run pytest) and documentation changes (uv run jb build docs --builder=custom --custom-builder=doctest)

---

Note

Medium Risk
Changes how voting ultimates are weighted and aggregated; incorrect weight slicing could affect ensemble results when pipelines return different column counts, though existing same-shape paths should behave similarly.

Overview
Adds TriangleSelector, a sklearn-style pipeline step that keeps a single triangle column (e.g. paid vs incurred) so VotingChainladder can blend estimators built on different columns from the same multi-column triangle. It is exported on chainladder and covered by public API tests.

VotingChainladder._get_ultimate now gathers each estimator’s ultimate_, raises ValueError when shapes disagree (with a hint about misconfigured TriangleSelector), and slices the weight tensor to the ultimate’s column size instead of assuming it matches the full input X shape.

New voting tests cover TriangleSelector.fit_transform and the mismatch case when one pipeline selects a column and another does not.

^{Reviewed by Cursor Bugbot for commit 4892df9. Bugbot is set up for automated code reviews on this repo. Configure here.}

[github-actions]

Pyright Type Completeness

View the full pyright --verifytypes output for this commit

Project (full chainladder package, at this PR's head): 14.6% of exported symbols fully typed (187 / 1280)

	Known	Ambiguous	Unknown	Total
Project (head)	187	106	987	1280

Other symbols referenced but not exported by chainladder: 13

	Known	Ambiguous	Unknown	Total
Other (head)	3	1	9	13

Symbols without documentation:

• Functions without docstring: 315
• Functions without default param: 0
• Classes without docstring: 10

Patch (exported symbols added or changed by this PR): 25.0% fully typed (2 / 8)

	Known	Ambiguous	Unknown	Total
Patch	2	0	6	8

Patch symbol details

Symbol	Status	Change
chainladder.workflow.TriangleSelector	❌ unknown	new
chainladder.workflow.tests.test_voting.test_mismatching_tri_sel	❌ unknown	new
chainladder.workflow.tests.test_voting.test_tri_sel	❌ unknown	new
chainladder.workflow.voting.TriangleSelector	❌ unknown	new
chainladder.workflow.voting.TriangleSelector.__init__	✅ known	new
chainladder.workflow.voting.TriangleSelector.col	✅ known	new
chainladder.workflow.voting.TriangleSelector.fit	❌ unknown	new
chainladder.workflow.voting.TriangleSelector.transform	❌ unknown	new

[cursor]

Cursor Bugbot has reviewed your changes and found 1 potential issue.

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

Want reviews to match your repository better? Bugbot Learning can learn team-specific rules from PR activity. A team admin can enable Learning in the Cursor dashboard.

^{Reviewed by Cursor Bugbot for commit 4892df9. Configure here.}

[cursor]

Weight slice wrong broadcast axis

High Severity

_get_ultimate trims the weight array with ultimate_.shape[1], treating that size as the value-column count. After _broadcast_weights expands 4D weights_ by repeating along X.shape[0], the second axis is index replication, not columns. Slicing it to 1 drops most index-level weights and skews the voted ultimate for multi-index triangles.

 

^{Reviewed by Cursor Bugbot for commit 4892df9. Configure here.}

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 90.55%. Comparing base (32ae098) to head (4892df9).
⚠️ Report is 1 commits behind head on TriangleSelector.

Additional details and impacted files

@@                 Coverage Diff                  @@
##           TriangleSelector    #1113      +/-   ##
====================================================
+ Coverage             90.50%   90.55%   +0.04%     
====================================================
  Files                    91       91              
  Lines                  5289     5306      +17     
  Branches                671      672       +1     
====================================================
+ Hits                   4787     4805      +18     
  Misses                  359      359              
+ Partials                143      142       -1     

Flag	Coverage Δ
unittests	90.55% <100.00%> (+0.04%)	⬆️

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

☔ View full report in Codecov by Harness.
📢 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.