docs(composite): Epic #150 Phase 1.3 smoke-test — composite.py docstring cross-ref

[dackclup]

Summary

Epic #150 Phase 1.3 — smoke-test for pre-merge-prod-sim workflow.

The pre-merge-prod-sim.yml workflow (shipped via PR #148 + PR #149) is gated by compute/scoring/** or compute/features/** path filters. Neither PR #148 nor PR #149 touched those paths, so the sticky-comment + diff-table composition has never run end-to-end in CI. This PR is the first real dogfood.

What this PR does

Adds a 3-line docstring cross-reference in compute/scoring/composite.py pointing to epic #150 Phase 3 (pillar correlation analysis + Quality + Profitability ROE double-counting at ~33% effective composite weight on ROE signals).

No compute logic change:

• PHASE3_WEIGHTS unchanged
• Sum-to-1.0 invariant lock at composite.py:43-45 unchanged
• All ACTIVE_PILLARS_PHASE3 / INACTIVE_PILLARS_PHASE3 unchanged
• Composite scoring math unchanged

The cross-reference is a real documentation improvement — points future readers to where the structural work lives. Useful regardless of smoke-test outcome.

Expected workflow behavior (verification matrix)

Step	Expected
Path filter	✅ Triggers (touches compute/scoring/composite.py)
Run PR-branch compute	✅ Warm-cache run from main's cache, ~10-15 min
Extract metadata	✅ universe_size, version, git_commit populated
Fetch main's compute output	✅ git show origin/main:frontend/public/data/* succeeds
Compute diff vs main	✅ Markdown diff with universe-size delta + Top-10 movers
Upload artifact	✅ pr-152-compute-output artifact (14-day retention)
Post sticky comment	✅ Summary table + diff section

Expected diff result: Composite scores should be identical between PR-branch and main (docstring-only change, no scoring logic). Top-10 movers table should show all-zero deltas (with ticker-asc tiebreaker — same pattern as the dogfood I ran locally earlier).

If anything else fires (non-zero deltas, missing diff section, failure-path comment) — that's the workflow telling us something. This PR is purely an observability test.

Plan after this PR

If workflow ✅ green and diff renders correctly:

• Close this PR without merging (throwaway by design)
• Tick Phase 1.3 in epic Foundation reconciliation roadmap — process + analytical scrutinize findings (2 rounds) #150
• Move to Phase 1.4 (test_verify_helper.py)

If workflow fails:

• Investigate the failure
• File fix PR
• Re-run smoke-test

Constraints honored

• ✅ No scoring logic change
• ✅ No schema change
• ✅ No workflow_dispatch trigger
• ✅ CLAUDE.md + AGENTS.md lockstep (Rule from PR docs(claude-md): token diet — auto-loaded file 236 → 172 lines (Optimization PR B) #142)
• ✅ Composite invariants preserved

---

Generated by Claude Code — Epic #150 Phase 1.3 smoke-test

---

Generated by Claude Code

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated (UTC)
quantrank	Ready	Preview, Comment	May 20, 2026 2:08pm

[github-actions]

Pre-merge production simulation

Field	Value
Duration	1001s
Universe size	502
Schema version	0.9.2-phase4h.2
Compute commit	1bc32ea9d97dc12aeca968d99b40fe43474bedf3
PR-branch output	pr-153-compute-output (14-day retention)

Diff vs main

Field	Main	PR	Δ
Universe size	502	502	+0
Schema version	0.9.1-phase4h.2	0.9.2-phase4h.2	⚠️ bumped

Main baseline: 2026-05-19T23:14:34Z (0.6 days old)

Top-10 movers (sorted by |Δcomposite_score|)

Ticker	PR rank	main rank	Δrank	PR score	main score	Δscore
ADI	385	435	+50	43.54	40.05	+3.49
HAS	392	350	-42	43.09	45.75	-2.66
AKAM	355	385	+30	45.60	43.86	+1.74
CRWD	440	454	+14	39.89	38.60	+1.29
HON	401	388	-13	42.46	43.71	-1.25
CEG	388	369	-19	43.42	44.66	-1.24
VMC	266	287	+21	50.15	49.13	+1.02
LYV	343	317	-26	46.23	47.24	-1.01
TGT	59	47	-12	61.64	62.59	-0.95
VICI	127	115	-12	56.53	57.46	-0.93