refactor(association,viz): bring cluster_tracks and render_track_list under S3776

[montge]

Why

SonarCloud has two open issues on the project, both rust:S3776 (cognitive complexity 17, limit 15): crates/thresh-association/src/jpda.rs::cluster_tracks and crates/thresh-viz/src/app.rs::render_track_list. The quality gate is still OK, but these are the only two open issues, and openspec/specs/cognitive-complexity and CONTRIBUTING both set 15 as the limit.

What

Same phase-helper decomposition as #78, no suppressions, no threshold changes.

cluster_tracks (17 -> 0). The nested find / union functions move to module level as find_root / union_sets: Sonar charges a nested fn's body to the enclosing function at +1 nesting, which alone was 4 of the 17. The three phases become union_tracks_sharing_detections (11), group_tracks_by_root (1) and collect_cluster_detections (1). Loop bodies are moved verbatim: same iteration order, same union direction (parent[rb] = ra), same HashMap, nothing cloned. This function is union-find over a boolean gating matrix, so no floating-point code is touched; the numerical JPDA functions in the file are unchanged.

render_track_list (17 -> 2). The per-row body becomes render_track_row (7), verbatim apart from taking id and confirmed by value and class as Option<&str>.

The counting model was checked against SonarCloud's own per-line increments for both issues, which it reproduces exactly; Sonar's number for the new code appears when it analyses this PR.

Evidence that behaviour is unchanged

• Old (git show HEAD:) and new cluster_tracks compared on 300,000 seeded random gating matrices (0-13 tracks and detections, 1-70 % density, extra rows, 278,355 non-square): identical after normalising cluster order. Run twice, independently, by the agent that made the change (200,000) and the one that tried to refute it (300,000). The harness is not in the tree, since a copy of the old function would itself be flagged.
• The three existing cluster tests only asserted the cluster count. New tests, written and passing before the refactor: the full output pinned on a 6x6 case (transitive chain, pair, a track gating nothing, a detection nobody gates), rows beyond n_tracks ignored, and no tracks.
• Mutation testing of the new code: 8 of 11 single-line mutants were caught, one is equivalent, and two escaped every test: swapping n_tracks / n_dets at the call site, and grouping by parent[i] instead of the root. Every case was square with a flat forest. test_cluster_tracks_non_square_two_level_forest (3 tracks, 2 detections, forest [0, 0, 1]) now fails on both, and passes on the original function.
• GUI: nothing exercised the click toggle being moved, so there is a headless egui_kittest test, clicking_track_row_toggles_selection, through a new ThreshVizApp::selected_track() accessor (same pattern as ellipses_shown() and the other test accessors). It passed before the refactor and fails when the toggle is broken.

Checks

• cargo fmt --all -- --check
• cargo clippy -p thresh-association -p thresh-viz --all-targets -- -D warnings, and again with -p thresh-viz --features gui, which is the run that actually compiles app.rs
• cargo test -p thresh-association: 55 passed. cargo test -p thresh-viz --features gui: 41 + 6 + 1 passed.
• RUSTDOCFLAGS=-Dwarnings cargo doc for both crates.
• Not run: a whole-workspace build or test. cluster_tracks has no callers outside its own tests.

For you to decide

• CI never lints or tests the GUI code. cargo clippy --workspace and cargo test --workspace run without --features gui, so app.rs, streaming.rs, theme.rs and all six kittest tests (the five existing ones too) are compiled out; viz-build only builds. The new click test therefore runs only locally. A cargo test -p thresh-viz --features gui step in viz-build, where Linux already installs the GUI packages, would close it. Not added here.
• The confirmation glyph in the track row is computed and then discarded (let _ = status;). Moved verbatim; deleting it or drawing the icon is a UI decision.
• The order of the clusters that cluster_tracks returns varies from run to run (HashMap with RandomState), before and after. Contents are deterministic and now pinned. A BTreeMap would make the order reproducible if a caller ever needs that.
• selected_track() is a small addition to the public API; it is in the CHANGELOG line.
• No OpenSpec delta: the cognitive-complexity spec lists six functions by name and Fix all 19 SonarCloud cognitive complexity violations #78 fixed nineteen without adding to it. Say if you want these two named there.

OpenSpec Validate is expected to be red until #155 merges.

🤖 Generated with Claude Code

https://claude.ai/code/session_01C53cGKw2kFHBrY6cy4tRnP

[Copilot]

Copilot was unable to review this pull request because the user who requested the review has reached their quota limit.

[coderabbitai]

Warning

Review limit reached

Next included review available in 59 minutes.

Check out review usage here.

View limit details

Limit details: You’ve used the included review currently available.

You've used all free OSS reviews for now. Wait for the free limit to reset to keep reviewing this public repository.

Learn how review limits work.

Review configuration:

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Advanced

Run ID: 31589f8d-5f01-43c7-903e-688d3960a118

📥 Commits

Reviewing files that changed from the base of the PR and between eb91f22 and 4087cb4.

📒 Files selected for processing (4)

• CHANGELOG.md
• crates/thresh-association/src/jpda.rs
• crates/thresh-viz/src/app.rs
• crates/thresh-viz/tests/app_kittest.rs

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.

📢 Thoughts on this report? Let us know!

[sonarqubecloud]

Quality Gate passed

Issues
0 New issues
0 Accepted issues

Measures
0 Security Hotspots
100.0% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube Cloud