Continuous enhancement: market-relative quantamental resilience

[ryun6249]

Continuous Enhancement Report

1. Current Status

• Project purpose: Local FastAPI financial research workstation with market, macro, Quant Lab, Quantamental, ML Forecast, AI Portfolio, and grounded AI briefing surfaces.
• Existing structure preserved: Yes. Static UI, FastAPI routers, Quantamental services, deterministic AI guardrails, All-default dashboard behavior, and top-right quality summary are preserved.
• Main issues found: Quantamental score-screen keys were accumulating with duplicated label/row-field mapping, and users lacked a benchmark-aware diagnostic for screening market-relative resilience.

2. Changes Made

Compatibility

• Added market_resilience as an additive Quantamental score key only.
• Did not change composite score weights, strategy entry/exit, order/trading code, provider selection, API defaults, secrets, or env files.

Data Consistency

• Added market_relative_resilience_v1 using existing aligned asset/benchmark returns, beta/correlation, volatility, drawdown, and liquidity vectors.
• Emits required/available observations, 63-day asset/benchmark/active return, downside capture, active positive-share, component scores, input provenance, warnings, and used_in_composite_score=false.
• Added SCORE_SCREEN_REGISTRY for centralized score-screen label/row/algorithm metadata.

UI/UX

• Added compact MRR score/class rows to the Quantamental overview and algorithm summary.
• Added Market Resilience / ?? ??? to the Score Threshold Screener while preserving existing options.

Visualization

• Kept the existing compact Quantamental overview pattern; no new crowded chart was added.
• MRR appears alongside QAM/VAB/DRS/LPS/TES in the same score/class hierarchy.

AI Briefing

• Added MRR to quant_snapshot and deterministic fallback key_changes.market_resilience_algorithm.
• AI remains interpretation-only and cannot override deterministic scores/signals.

Translation

• Added Korean/English labels and smoke coverage for the new score-screen option.

Performance

• Reused already-loaded vectors; no new provider call, cache path, polling loop, or LLM call was added.

Tests / Validation

• Updated engine/API/UI/module/contract tests and local smoke scripts for MRR and v17/v7 cache-busted assets.

3. Changed Files

File	Change	Reason
pipelines/quantamental/quant_engine.py	Added market_relative_resilience_v1 algorithm and component score	Add benchmark-aware quant diagnostic
pipelines/quantamental/service.py	Added score-screen registry, MRR health metadata, MRR row mapping	Reduce score-key drift and expose new screen key
pipelines/quantamental/ai_service.py	Added MRR to AI context/fallback key changes	Keep AI briefing grounded in deterministic data
core/schemas/quantamental.py	Added market_resilience score key	Preserve typed API contract
app/web/index.html	Added Market Resilience option and bumped bundle versions	Expose user-facing screening path
app/web/app.js	Added score-key labels and allowlist	Keep selector behavior stable
app/web/modules/quantamental-ui.js	Added MRR labels, overview metrics, and summary row	Show diagnostic without a new panel
scripts/check_ui_contract.py	Added v17/MRR contract markers	Static contract validation
scripts/quantamental_ui_smoke.py	Smoke Market Resilience score screen	Browser flow coverage
scripts/ai_portfolio_ui_smoke.py	Updated bundle version expectations	Cross-dashboard smoke alignment
tests/test_quantamental_api.py	Added API/score-screen/health assertions	Contract coverage
tests/test_quantamental_engines.py	Added engine/AI-context assertions	Deterministic algorithm coverage
tests/test_ui_modules.py	Added module render assertions	UI rendering coverage
tests/test_ui_routing_contract.py	Updated bundle/MRR markers	Static routing contract coverage
docs/CONTINUOUS_ENHANCEMENT_LOG.md	Added run analysis, plan, changes, validation	Automation traceability

4. Validation Results

Check	Command / Tool	Result	Notes
Lint	python -m ruff check ...changed Python/test surfaces...	Passed	No ruff issues
Build	node --check app/web/modules/quantamental-ui.js; node --check app/web/app.js	Passed	Static JS syntax; no npm build manifest exists
Test	python -m pytest tests/test_quantamental_engines.py tests/test_quantamental_api.py tests/test_ui_modules.py tests/test_ui_routing_contract.py -q	Passed	84 passed, 4 subtests passed
Test	python -m pytest -q	Passed	696 passed, 9 subtests passed
UI validation	Browser at http://127.0.0.1:8414/ui/?range=1Y#quantamental	Passed	Desktop and 390px mobile: All view, quality badge, MRR visible, no horizontal overflow
Data validation	Live API smoke on port 8414	Passed	Health lists MRR; AAPL MRR 69.54; Market Resilience screen returned rows
AI briefing validation	Live deterministic AI report	Passed	key_changes includes market_relative_resilience_v1; no LLM-only claims
Smoke	python scripts/quantamental_ui_smoke.py --base-url http://127.0.0.1:8414 --output reports/quantamental_ui_smoke_continuous_20260519_1802.json	Passed	Market Resilience score screen and Quantamental flows passed
Smoke	python scripts/ai_portfolio_ui_smoke.py --base-url http://127.0.0.1:8414 --timeout-s 240 --output reports/ai_portfolio_ui_smoke_continuous_20260519_1802.json	Passed	Cross-dashboard actions passed; console errors empty

5. Remaining Risks

• MRR is diagnostic-only and should not be promoted into composite scoring without explicit approval.
• Score-screen keys are now centralized in service metadata, but frontend labels still live in the static UI copy tables.
• npm/pnpm commands remain excluded because the repo root has no frontend package manifest.

6. Next Enhancement Suggestions

1. Consider serving score-screen metadata from /quantamental/health into the UI selector to remove the remaining static label drift.
2. Add a compact algorithm detail drawer for QAM/VAB/DRS/LPS/TES/MRR inputs without cluttering the normal overview.
3. Add a targeted AI briefing assertion that all diagnostic algorithms included in quant_snapshot.algorithms are either summarized or explicitly marked unavailable.

7. Safety Notes

• main/master was not modified directly.
• Auto-merge was not enabled.
• Secrets were not touched.
• Trading/order execution logic was not modified.
• Unsupported facts were not presented as verified.