Extends the local-LLM layer from news to TRADES, reusing the proven enrichment
pattern (structured JSON, Postgres cache, off-hot-path scheduler, graceful
degradation). Advisory only — never gates a trade.
strategy/evaluation/trade_analysis.py: four analyses keyed by analysis_type —
* armed — does a loaded setup still make sense to pursue? (pursue/avoid/
monitor + confidence + concerns), weighing numbers + catalyst
* weak — why a too-soft setup failed and whether a thesis remains
* postmortem — closed-trade lesson (uses r_multiple / exit_reason)
* eod — once-a-day session narrative
Schema-constrained decoding (ANALYSIS_SCHEMA), tolerant _parse, and a
context_hash so an unchanged board isn't re-sent to the GPU. run_* helpers
upsert into ai_trade_analysis_cache; trade_analysis_map / query_trade_analysis
read it back. NEVER raise.
Wiring:
- ai_trade_analysis_cache table (storage/pg_schema.sql)
- OllamaConfig.trade_analysis_{enabled,interval_seconds,batch_limit} + env
- run_live_paper step_analyze on the Scheduler, gated by ollama.enabled AND
TRADE_ANALYSIS_ENABLED (ships OFF); reads the in-memory trigger book + day P&L
- query_trade_analysis projection; api snapshots() attaches the verdict to armed
triggers + approval rows and exposes trade_analysis
- dashboard: per-name AI verdict chip (pursue=green/avoid=red/monitor=amber) +
an "AI session note" panel
- .env.example documents the toggles
Tests: parse/enum/clamp, prompt builders, persistence + context_hash dedupe,
closed-trade + EOD paths, config from_env, and the dashboard projection. 230 pass.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01M5HTMkaadGDnAPtXe2hWwg
Completes the Phase 2 loop: the catalyst score now actually reaches the runtime
decision, not just the pure functions. Gated behind its own flag so enabling the
Phase 1 advisory/dashboard never silently changes trade scoring.
- OllamaConfig.catalyst_score_enabled (NEWS_CATALYST_SCORE_ENABLED, default off).
- Watcher accepts an optional catalyst_score_provider (symbol -> 0..1 | None) and
passes it into evaluate_setup; a None provider reproduces today's behavior. The
provider is injected so runtime/ stays free of research/ imports.
- run_live_paper: one TTL-cached catalyst provider shared by the watcher and the
execution service; the watcher gets the score provider only when the blend is
enabled. step_arm gains a gated graded ranking multiplier (bullish catalyst
boosts by 1+gain*score; a confirmed dilutive offering is sunk below the armed
cut), falling back to the existing binary news boost when off.
- Add the enrich task to the run_pass pipeline order so a single --once pass also
runs enrichment (previously only the continuous loop did).
- .env.example documents the new flags; watcher tests cover the blend + the
no-op None-provider case.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01M5HTMkaadGDnAPtXe2hWwg
Leverage a local Ollama model to interpret news catalysts for the small-cap
momentum strategy — distinguishing bullish catalysts (FDA approval, M&A) from
dilutive offerings the technicals alone can't see. Advisory-first, then automate.
Phase 1 (advisory; no change to trade behavior):
- research/ingestion/news_enrichment.py: stdlib urllib call to Ollama with
structured JSON output, classifying each headline into a catalyst type,
sentiment, conviction and dilution flag. Graceful degradation (None on any
failure) so trading is never blocked.
- news_catalyst_cache table + populate the previously-dead news_events.sentiment.
- Runs on the existing interval Scheduler (off the hot path), gated by
OllamaConfig; deduped by the raw_news_items payload hash.
- Surface the advisory on the dashboard approval queue (projection + api + UI).
Phase 2 (automation; ships gated OFF):
- Optional catalyst_score blended into calculate_setup_quality (None reproduces
the exact legacy weights); pure pass-through via evaluate_setup.
- Dilution veto in TradingExecutionService (slow + fast paths) emitting a
catalyst_dilution_veto risk event, behind NEWS_DILUTION_VETO_ENABLED.
Config (OllamaConfig.from_env), .env.example docs, and tests for parsing,
persistence/dedupe, the projection, the score blend, the veto, and an
end-to-end enrich->advisory flow (Ollama mocked). Default model moves to
qwen2.5:7b-instruct (best small-model JSON, fits a 12GB GPU).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01M5HTMkaadGDnAPtXe2hWwg
The ONLY entry change that survived rigorous validation on 14 SIP-backfilled
sessions (53 trades). Gap-size buckets: 15-30% gap = +0.65R/trade, but 30-60%
blow-offs = +0.10R. Train/test split (70/30 by date): lowering TRIGGER_GAP_MAX
60 -> 35 lifts HELD-OUT TEST expectancy 0.194 -> 0.462 R/trade AND total R
(18.7 -> 19.9), better win rate, on both splits. Principled (huge gaps
fade/buy-the-top) and corroborated by the entry-timing audit (losers were
wide-range blow-off gappers).
Rejected after validation (do NOT ship): open-delay (forgoes total R for
marginal expectancy), breakout-HOLD/close-confirmation (hurts on the larger
sample), rvol floor (no-op: all gappers already rvol>=6). Confirmed KEEP: the
$-vol floor is fine on SIP volume (the earlier "it hurts" was corrupted IEX
data); profit-lock ladder stays.
.env TRIGGER_GAP_MAX 60 -> 35. NB: 14-day sample — re-validate as data grows.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
From the multi-agent audit (ship-now, low-overfit-risk cluster):
INGESTION RESILIENCY
- client._request: 429/5xx now retry with Retry-After + exponential backoff
(previously hard-failed, feeding silent-stale ingest).
- SIP backfill feed (settings.backfill_feed='sip'): live realtime stays IEX, but
backtest/backfill ingestion uses the full consolidated tape — IEX sees only
~2-5% of true volume (verified: AAPL 22,797 vs 452,704), which was silently
corrupting every rvol/$-volume decision and backtest. eod_replay wired to it.
- step_ingest freshness watchdog: emits a STALE event + board note when the
newest bar is older than INGEST_STALE_SECONDS during RTH (45% of passes
historically returned 0 rows with no error). Optional hard arm-block
(INGEST_STALE_BLOCK, off by default).
RISK
- Portfolio gross-notional cap (TRADING_MAX_GROSS_NOTIONAL_PCT=0.60): both sizing
paths now cap each entry to the remaining gross budget, so 6 x 16% can't pile
the book into ~96% gross of correlated small-caps. Blocks/shrinks late entries.
- Fixed the duplicate TRADING_MAX_CONCURRENT_POSITIONS in .env.example.
MONITORING (UI)
- Naked/unprotected-position alarm: _unprotected_positions joins the latest
position + order snapshots (order sync now flattens legs) and the TUI board
shows a red UNPROTECTED banner the instant a held long lacks stop coverage —
the exact blind spot when APWC/NOWL went naked. Verified it flags today's
3 overnight positions.
Tests: gross-notional cap blocks when budget full; 50 pass.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
- Extract the cancel-protective-legs-then-close routine (duplicated in the exit
manager and the execution service after fixing the held-qty 403 in both) into
one shared, heavily-documented helper: runtime/flatten.py. Both callers now
delegate, so the fix can't drift out of sync. Behaviour unchanged; 121 tests
pass.
- Refresh .env.example: document every key added this session (liquidity gate +
sizing, backout cooldown, re-entry guard, managed-exit ladder, EOD flatten)
with the why, and flag MAX_DAILY_LOSS_PCT to revert to 0.03 before live.
- Remove throwaway data/_watch_*.py monitoring scripts (gitignored clutter).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Sweep over 172 real gapper sessions (16 days) with a per-setup breakdown found
the opening-range breakout is the ONLY profitable setup (+0.25R); every slow
pattern loses (bull_flag -$78k @ ~42min is the worst). Running ORB-ONLY flips
the book from -$95k to +$7.4k AND enters ~11 min earlier (23 vs 34 min avg).
- classify_setup prefers the ORB when valid and honors STRATEGY_SETUPS env
(comma-sep setup values) to restrict active setups; .env defaults to
STRATEGY_SETUPS=opening_range_break (ORB-only).
- BacktestEngine Trade.setup_type captured for per-setup attribution.
- learn_params load_sessions filters to gapper sessions (mimics live scanner) so
the backtest trades the same universe the system would.
75/75 tests green.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
- grafana/: docker-compose + provisioned Postgres datasource + 'Momentum — Live'
dashboard (evaluation board with status colours, ready signals, risk events,
bars + equity timeseries) reading the event store via JSONB extraction.
Run: cd grafana && docker compose up -d -> http://localhost:3000.
- RSS news-catalyst discovery wired: extract_tickers also matches $CASHTAGS;
run_discovery pulls tickers from recent news into the universe when NEWS_FEEDS
is set (no-op / safe otherwise).
- docs: .env.example NEWS_FEEDS; GO_LIVE monitoring adds 'momentum watch' + Grafana.
67/67 tests green. Live board (momentum watch) verified rendering the evaluation
board with real determinations.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
- Drop duckdb dependency (pyproject, requirements) -> psycopg2-binary.
- run_live_paper: boot log shows the real Postgres target (masked); add
_db_target_desc() helper.
- .env.example: document DATABASE_URL / PG* vars; mark WATCHER_EVENT_DB_PATH legacy.
- evaluation/storage.py: INSERT OR REPLACE -> INSERT (setup_id is a fresh uuid).
Verified: 67/67 tests green on Postgres with duckdb UNINSTALLED; app boots
against the live endpoint (127.0.0.1:5432 proxy) and wrote 205 daily bars.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Snapshot before the DuckDB->PostgreSQL migration. data/, archive/, the 232MB
rebuild zip, the Windows .venv, and .env are excluded via .gitignore.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude