[ENH] set partial triangle through loc

[henrydingliu]

Summary of Changes

adding support to set triangle values a few origins or devs at a time

Related GitHub Issue(s)

closes #1069

Additional Context for Reviewers

the heavy lifting is done by the existing _contig_slice function, which is already used in the getter. not sure why it was never applied in the setter

• 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
Touches core 4-D indexing assignment in slice.py, which can silently mis-assign data if np.ix_ or contiguous-key logic is wrong; scope is numpy-only with explicit guards and new tests.

Overview
Extends Triangle.loc and Triangle.iloc assignment on the numpy backend so you can write partial slices of origin/development periods and non-contiguous index/column selections, matching behavior that getters already supported.

The setter now runs keys through _contig_slice (same as get_idx) before writing values. For fancy index lists on the index/column axes, writes use np.ix_ so assignments align with the getter’s Cartesian-product selection. Fancy indexing on origin or development when setting still raises ValueError; sparse backends remain limited to .at / .iat.

New tests cover partial origin/dev updates, non-contiguous loc/iloc writes, and the new error paths.

^{Reviewed by Cursor Bugbot for commit 64329cc. Bugbot is set up for automated code reviews on this repo. Configure here.}

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 91.79%. Comparing base (6c9ce99) to head (64329cc).
⚠️ Report is 81 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #1103      +/-   ##
==========================================
+ Coverage   90.48%   91.79%   +1.30%     
==========================================
  Files          91       91              
  Lines        5289     6252     +963     
  Branches      671      889     +218     
==========================================
+ Hits         4786     5739     +953     
- Misses        359      363       +4     
- Partials      144      150       +6     

Flag	Coverage Δ
unittests	91.77% <100.00%> (+1.28%)	⬆️

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.

[github-actions]

Pyright Type Completeness

View the full pyright --verifytypes output for this commit

Project (full chainladder package, at this PR's head): 14.5% of exported symbols fully typed (185 / 1274)

	Known	Ambiguous	Unknown	Total
Project (head)	185	106	983	1274

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: 313
• Functions without default param: 0
• Classes without docstring: 10

Patch (exported symbols added or changed by this PR): 0.0% fully typed (0 / 5)

	Known	Ambiguous	Unknown	Total
Patch	0	0	5	5

Patch symbol details

Symbol	Status	Change
chainladder.core.tests.test_slicing.test_iloc_setting_non_contiguous_index_and_columns	❌ unknown	new
chainladder.core.tests.test_slicing.test_loc_setitem_partial_triangles	❌ unknown	new
chainladder.core.tests.test_slicing.test_loc_setting_non_contiguous_index_and_columns	❌ unknown	new
chainladder.core.tests.test_slicing.test_set_fancy_development_raises	❌ unknown	new
chainladder.core.tests.test_slicing.test_set_fancy_origin_raises	❌ unknown	new

[genedan]

Looks like I missed filling out a docstring when I annotated the file. Could you fill it out?

[henrydingliu]

done

[kennethshsu]

How are we doing here with this PR? Can I help?

[genedan]

@kennethshsu that comment I made about the dense version of raa (and other fixtures, possibly) being run twice instead of alternating between dense/sparse as intended could definitely use assistance. It'll need it's own issue created and PR to remedy. I think it might reveal a lot more test failures too once we fix it.

[henrydingliu]

@kennethshsu pausing till the sparse test fixture is fixed

@genedan i'd like your feedback on

1. i'd like to continue to observe John's intent around no loc/iloc setting for sparse.
2. I'd like to enable non-contig setting if the non-contig dims are index and column.

[genedan]

1. Agreed, while theoretically we could enable the feature there are other things more worth our time and this can remain as-is unless a bunch of users ask for it.
2. Agreed.

[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.}

^{Reviewed by Cursor Bugbot for commit 64329cc. Configure here.}

[cursor]

Stepped slice step dropped

Medium Severity

The new setter runs _contig_slice and then _normalize_index, unlike get_idx, which uses the contig result directly. When an arithmetic index sequence spans a full axis, _contig_slice builds a stepped slice, and _normalize_index then clears both bounds and drops the step, so assignment can write the entire axis instead of the selected positions.

Additional Locations (1)

• chainladder/core/slice.py#L190-L194

 

^{Reviewed by Cursor Bugbot for commit 64329cc. Configure here.}