Owner: a stock with a NEGATIVE overall (combined) score must not be recommended
in the profit buckets, even with positive momentum + theme. BGRIM/TTB had
combined -0.029/-0.066 yet still landed in bucket 1.
- allocate_capital profit_pool now gates on combined_score > 0 (in addition to
momentum>0 + theme>0), still ranking by momentum within that pool.
- Verified: bucket 1 now picks PTTGC (+0.209, mom +1.85) instead of BGRIM/TTB;
BANPU in bucket 2; ADVANC/SCB/LH in bucket 3 by yield. Full suite 377 green.
Owner: momentum must enter the formula (a factor), not be bolted on outside it.
- Single declarative source _SIAMCHART_WEIGHTS = {eps_growth:1.5, dividend_yield:2.0, momentum:0.5}.
- New siamchart_raw_score(g,d,m) = single source of the formula; momentum is an
explicit term inside it. All 3 call sites (build_siamchart_score + both
symbol_breakdown spots) now share it — no duplicated arithmetic.
- Pure refactor: outputs unchanged (weights identical). Tests added for the
momentum-inside-formula rule + momentum raising the score. Full suite 376 green.
- load_price_snapshot picked the last snapshot by filename (lexicographic),
selecting a stale 9-symbol collection over the full 50-symbol universe. Now
picks the snapshot with the latest source.retrieved_at.
- allocate_capital profit buckets now also require momentum > 0 (a falling-price
name is not 'ทำกำไร'), while momentum/theme_signal stay Optional so the PIT
backtest path (which doesn't provide them) still allocates.
- Suggestion now allocates across all 50 SET50 names (B1: BGRIM,TTB; B2: BANPU;
B3: ADVANC,SCB,LH).
- Regression tests for both. Full suite 374 green.
Owner rule: a stock that should be bought for profit is one whose PRICE is
likely to rise in the next 3-6 months — not one with high EPS growth (BTS had
EPS +137% yet flat/falling price). The old selection ranked buckets 1/2 by
(60/40 theme+siamchart where siamchart was EPS-growth dominated).
- themes.price_trend_score(): blend of ~3/6/12-month price momentum, z-scored
across the universe (heavier 3/6m weight per the 3-6 month tenure).
- allocate_capital: buckets 1/2 rank by momentum, gated on theme_signal > 0
(mean surprise across the symbol's themes). theme_signal=None (backtest path)
is not gated so PIT backtest still allocates. Bucket 3 unchanged (yield top).
- suggestion endpoint passes real momentum + theme_signal from the live board.
- Verified: bucket 1 now picks CRC/BEM (dividend + rising price); falling-price
PTT/MINT go to bucket 3 by yield, not bucket 1. Full suite 372 green (3 new
momentum/theme-gate tests).
Owner's rule: backtest must run as soon as there's enough data to estimate an
investment — it must NOT be blocked just because some sources lack deep PIT
history. Scoring is deliberately flexible (a theme uses whatever subset of
factors was knowable that day).
- backtest_readiness: readiness = usable window (price + Siamchart + >=1 factor),
not all-factors-present. Missing factors still reported (transparency) but no
longer block the run. recommended_start = oldest executable price (oldest
history the system holds); recommended_end = last complete trading day.
- pit_scorer.theme_surprise_report: flexible — skips factors not released by
as_of; blocked only when NO factor has a value. pit_meta.partial_pit reflects
themes scored from a partial factor subset.
- Verified end-to-end: readiness ready=true (recommended 2024-01-03 -> 2026-08-29);
POST /api/v1/backtest/run default window returns 201 full result (1M -> final
equity 1,117,243.95), no 400 from missing factors.
- test_backtest_readiness updated to earliest-runnable semantics; full suite 369 green.
Q2 flexible scoring: _fetch_with_cache now degrades instead of raising
DashboardError — a source that fails with no cached value returns {} so the
theme scorer drops that source's factors; a previously-good value is kept as
stale by the daily cache. Verified: all-sources-down still builds 13 themes.
Q3 per-source audit: new themes.factor_source_breakdown(fetched, theme) shows
per factor source/raw/normalized/weight/contribution; dashboard exposes
fetch_data + factor_sources; per-symbol modal renders symbolDetail.factor_sources
(e.g. retail: te_thailand ยอดขายปลีก -14.5 -> -1.0 x 0.7 = -0.7).
Suite 369 green; independent review passed: true.
Q1 (HAR for deferred sources) spike recorded: method works, REIC needs deeper
interaction; NBTC 403 likely unbpassable without a session.
Committing the prior uncommitted working-tree state that predates this session's
data-source work (was already modified/untracked at session start) so the tree
is clean before push. Includes: event-study + research report integrity/forward
observation work, prices tests, research hash migration script, and the
2026-08-23/24 engineering-log + test-evidence notes. Verified green as part of
the full 362-test suite.
Append the PIT-work session to HANDOFF and engineering-log: factor
vintages store, partial PIT score provider, honest leakage gating, /
api/v1/backtest use_pit wiring, 255-test verification, and the honest
scope (no pre-2026-08-27 factor history; siamchart fundamental partial).
A. Cross-theme comparability:
- compute_theme_surprises now weight-normalizes by total |weight| (weighted
average), so every theme surprise on same [-1,1] scale regardless of factor
count/weight (retail 0.189->0.145; auto_credit 1.0->0.64).
B. Historical factor store (enables learning macro/demographic factors):
- New factor_history.py: append-only per-factor JSONL, dedupes unchanged
values, rejects non-finite, records every FACTORS value each scheduler run.
- scheduler.py: jobs carry fetch_module; refresh_all records factor history
(non-fatal); added bank_npl job.
- GET /api/v1/learning/factors?min_points= reports n_points/learnable per
factor so users see when P4 learning unlocks (validated query parsing).
- weight_learning: generic learn_factor_series() aggregator (momentum reuses).
Independent review deleg_5dd358e3 passed=true (empty security/logic arrays);
its two robustness suggestions applied (finite guard in record(), clean 400 on
bad min_points). 234 tests pass; Vite build passes.
P0-B (registry is the single source of truth for scoring):
- FACTORS now carries center/span normalization spec; unused hand-written
per-theme surprise blocks in dashboard.py replaced by one registry-driven
compute_theme_surprises() (themes.py).
- THEMES['banks'] adds bank_npl weight so NPL is genuinely blended.
- factor_value/normalize hardened against NaN/inf (finite guards).
- Board re-ranks (TRUE/GULF up, TOP->3) per registry weights; 3 new tests
incl. 'changing a registry weight changes output'.
P3 (point-in-time backtest):
- run_backtest is now a real multi-rebalance engine (reallocates every window,
reconciles holdings, marks to market) instead of allocate-once+break.
- Added leakage_guard (False unless a PIT score_fn is supplied), planned vs
actual rebalances, and momentum_at() true 12-1 (skips last month, PIT).
P4 (factor-weight learning):
- weight_learning.py: cross-sectional Spearman IC, forward-return builder,
IC aggregation + t-stat, and apply_weight_update (new = clip(old*(1+shrink*IC))).
- GET /api/v1/learning/momentum endpoint. Live result: momentum IC=0.012
t=0.132 over 22 periods -> momentum has no reliable predictive power here.
Macro/demographic factors blocked (no historical factor vintages yet).
Two independent review gates passed (deleg_fe6f45cd, deleg_718218f8): empty
security/logic arrays; their non-blocking suggestions applied (finite guards,
dedupe leakage_guard resolution). 226 tests pass; Vite build passes.
- P1: /api/v1/simulation now uses the canonical board score (default_scores)
instead of a divergent 3-theme recompute -> 'จำลอง' can't disagree with board
(live check: sim top pick PTT == top board combined 1.600). Removes binary
auto/en signs, restores quality+momentum+dividend screen consistency.
- P2: dashboard emits source_summary{factor_keys, rows}; frontend shows
'N ปัจจัย · M แหล่ง' so the 7-vs-5 count confusion is impossible.
- P5: removed dead themes.list_themes()/Theme/build_theme_scores/_map_index and
the tests that locked them; added tests/conftest.py so pytest needs no PYTHONPATH.
- docs: audit-and-plan-2026-08-26.md (full P0-P5 plan) + engineering-log entry.
- 203 backend tests pass; Vite build passes. Independent reviewer: no security or
logic blockers (minor error-leak suggestion applied: 503 message no longer leaks
exception detail).
- auto_npl.py: parse BOT Gross NPLs by business (reportID=794); extract auto loan NPL (20,602 mn THB, 3.95% of NPLs, 2.06% of loans)
- /api/v1/themes now exposes auto_npl_pct + auto_npl_amount alongside car-sales volume
- 3 new tests; full suite 188 OK; live verified (themes shows auto_npl_pct 3.95)
- energy_thai.py: scrape Thai Oil (TOP) investor financial-highlights -> quarterly + annual EBITDA/Net Profit/Sales (Million Baht), largest Thai refinery
- Thai-specific factor per user (energy must reflect Thai companies, not US EIA proxy); Krungsri was projection-only, TOP gives real quarterly actuals
- Frequency: quarterly (documented in research note)
- 4 tests; full backend suite 156 OK; compileall ok; static scan clean
