feat(validation): Phase 4.6 follow-on — wire universe_provider into pbo_dsr gates#275
Merged
Merged
Conversation
…bo_dsr gates Closes the next-task gap surfaced by Research Report v1.0 §7.4 follow-up: PR #274 landed `compute/ingest/historical_universe.members_at()` as a library but nothing called it. This PR threads an optional `universe_provider` callable through `factor_passes_gates()` so the returned metrics dict carries honest universe provenance — closing the loop between the historical-membership module and the validation gates that every Phase 4+ (OSAP / JKP / Qlib / IPCA) and Phase 5 (ML meta-learner) candidate factor must pass. ## What changed - `compute/validation/pbo_dsr.py::factor_passes_gates()` — 3 new optional kwargs (`universe_provider`, `as_of_date`, `current_universe`). When all 3 are passed, the function calls the provider once and enriches the metrics dict with `universe_as_of` (ISO date), `universe_size` (int), and `survivorship_bias_corrected` (bool). When the provider returns `is_complete=False` (e.g., pre-2020 date), the flag flips False AND a warning is logged. When the provider raises, behavior degrades gracefully (3 fields stay None, validation still completes from the supplied returns_matrix). - `compute/validation/pbo_dsr.py::today_utc_date()` — small helper for callers wiring `as_of_date=today_utc_date()` in forward-cron validation paths. - `tests/test_validation/test_pbo_dsr.py` — 6 new tests covering: * backward-compat (no universe kwargs → 3 new keys = None) * happy path (members_at() with date well inside coverage) * degraded path (pre-EARLIEST_EVENT_DATE → is_complete=False) * provider-raises (graceful degradation, validation still runs) * partial-kwargs warning (caller passes only some of the 3) * today_utc_date helper smoke ## Caller migration Pre-Phase-4.6 callers (osap-integration / jkp-integration / qlib scout / ipca scout PRs already merged) are byte-identical — all 3 new kwargs default to None and the function returns the legacy 10-key metrics dict augmented with 3 None values. No caller code needs to change. NEW callers (Phase 4i.1+ integration PRs + Phase 5 ML meta-learner) should pass: from compute.ingest.historical_universe import members_at from compute.ingest.universe import get_sp500_constituents current = frozenset(get_sp500_constituents().ticker) passes, metrics = factor_passes_gates( factor_returns, returns_matrix, n_trials=n_trials, universe_provider=members_at, as_of_date=as_of_date, # backtest cutoff or today_utc_date() current_universe=current, ) The metrics dict can then be threaded into `compute/output/schemas.py::Metadata.universe_membership_as_of` + `Metadata.survivorship_bias_corrected` at writer time (separate follow-up PR — `compute/main.py` wiring stays out of this scope). ## Why this matters (anchor) Hou-Xue-Zhang (2020) RFS replication-crisis evidence emphasizes survivorship as the PRIMARY failure mode in factor-zoo work. Every PBO/DSR number that doesn't carry universe provenance is suspect — pre-Phase-4.6 we couldn't tell, post-Phase-4.6 we can. This PR is the first place that knowledge surfaces in the validation metric dict. ## Hard rules preserved - Rule 9 (schema triple) — N/A; no Pydantic / TS / snapshot touched in this PR. Metadata wiring is deferred to a separate writer PR per scope discipline. - Rule 16 (composite formula sacred) — N/A; no scoring change. - Rule 18 (observability before wiring) — diagnostic surface (metrics dict keys) ships in same PR as the integration; consumer code (writer) lands next, after this metric surface lives in production for ≥ 1 cron cycle. - License: no new deps. Pure forward TYPE_CHECKING import for type hinting `MembershipResult` (zero runtime cost). ## Verification - `ruff check` — clean (linter moved `Callable` to `collections.abc` per modern convention; deliberate, not reverted) - `pytest tests/test_validation/test_pbo_dsr.py tests/test_ingest/test_historical_universe.py tests/test_config.py` — 59 passed in 9.34s - Test count delta: +6 (universe-provider integration cases) - Backward-compat case explicitly tested (no universe kwargs → legacy behavior preserved) ## NOT in this PR (deferred to follow-ups) - `compute/main.py` writer wiring to populate `Metadata.universe_membership_as_of` + `survivorship_bias_corrected` in the forward-cron output JSON - Re-validation of existing pillars + `manipulation_index` with the historical universe (likely revises some published baselines DOWNWARD — explicit honest-correction PR) - Verify-helper Section M for universe-provenance accounting equation
|
The latest updates on your projects. Learn more about Vercel for GitHub.
|
Contributor
Pre-merge production simulation
Diff vs main
Main baseline: Top-10 movers (sorted by |Δcomposite_score|)
|
4 tasks
dackclup
added a commit
that referenced
this pull request
May 27, 2026
…tadata in forward cron (#276) Closes the last leg of the Phase 4.6 chain. PR #274 landed the `historical_universe.members_at()` module + 2 nullable Metadata fields. PR #275 wired `universe_provider` into `pbo_dsr.factor_passes_gates()` so validation gates carry honest provenance. This PR makes the forward-cron `metadata.json` output ACTUALLY populate those fields instead of leaving them None. ## What changed - `compute/main.py` — `Metadata(...)` construction now passes: - `universe_membership_as_of=now.date().isoformat()` (today's date — forward cron scores as-of today) - `survivorship_bias_corrected=True` (today's S&P 500 IS the honest universe for an as-of-today query, per the PR #274 schema docstring semantic — True means "this output's universe assumption is honest for its as_of_date") - `tests/test_output/test_writer.py` — 2 new round-trip tests: - Phase 4.6 happy path: both fields survive Pydantic → JSON - Legacy snapshot back-compat: when neither field is passed (pre-0.10.7 caller pattern), Pydantic defaults to None and JSON writes nulls ## Hard rules preserved - ✅ Rule 9 (schema triple) — no schema change in this PR (fields already in schemas.py + types.ts + snapshot from PR #274) - ✅ Rule 16 — N/A (no scoring change) - ✅ Rule 18 — observability surface from PR #274 is now actually populated; consumers can branch on it - ✅ No new deps - ✅ No new env-vars ## Verification - `ruff check compute/main.py tests/test_output/test_writer.py` — clean - `python -m compute.output.schema_check` — Schema snapshot in sync - `python -m pytest tests/test_output/test_writer.py -k metadata` — 4 passed (2 existing + 2 new) ## What goes live on next cron Next weekday cron (Wed 2026-05-28 22:00 UTC) writes: metadata.json: ... universe_membership_as_of: "2026-05-28" survivorship_bias_corrected: true Backward compat: legacy snapshots (pre-0.10.7) still have these fields as null per the Pydantic optional default. ## Closes the Phase 4.6 chain | Layer | PR | Status | |---|---|---| | Module | #274 | members_at() + CSV + tests | | Schema | #274 | Metadata fields + types.ts + snapshot | | Validation gate | #275 | universe_provider kwarg in pbo_dsr | | **Writer** | **this PR** | **forward cron populates Metadata** | ## NOT in this PR (next follow-ups) - Honest re-validation of existing pillars + manipulation_index with historical universe (likely shifts PBO/DSR baselines DOWN 5-15%) - Verify-helper Section M for universe-provenance accounting equation - Backtest harness that consumes the new universe_provider end-to-end Co-authored-by: Claude <noreply@anthropic.com>
4 tasks
dackclup
added a commit
that referenced
this pull request
May 27, 2026
…er (#278) Closes follow-up #2a from ``docs/research/historical-revalidation-harness.md``. Pure-function + cached git-archive walker that reconstructs the daily ``composite_score`` time series per ticker from the cron's committed rankings.json snapshots. This is the **first data source** for the downstream Phase 4.6 task #2 chain: - #2c per-pillar IC at historical dates → consumes this DataFrame - #2e manipulation_index distribution shift report → consumes this DataFrame - #2f honest-baseline report → consumes the chain output ## What ships - `compute/validation/ranking_history.py` (NEW, ~190 LOC): - `list_ranking_commits(start, end, path)` — enumerates (sha, date) tuples from `git log` with optional ISO-date filter window - `dedupe_by_date(commits, prefer="latest"|"earliest")` — collapses multiple commits per day to one (default: latest) - `load_snapshot_at(sha, path)` — `git show SHA:path` → parsed JSON list. Returns empty list (not raise) when SHA pre-dates the file's existence, per the standard cron's chore-commit pattern. - `load_ranking_history(start, end, columns, dedupe_dates=True)` — high-level orchestrator returning `(date, ticker)` MultiIndex DataFrame with the requested columns. Default columns = `("composite_score",)`. Caller-extensible to grab `composite_score_adjusted` / `current_price` / etc. - `tests/test_validation/test_ranking_history.py` (NEW, 18 tests): - 5 dedupe_by_date helper cases (latest / earliest / invalid / empty / sort) - 3 list_ranking_commits live-git cases (smoke / date-filter / empty-window) - 2 load_snapshot_at live-git cases (HEAD parses / unknown-SHA → empty list) - 5 load_ranking_history live-git cases (smoke / default columns / custom columns / empty window / dedupe behavior) - 1 canonical-path constant pin - 2 monkeypatched edge cases: malformed-JSON snapshot skipped; rows missing required column skipped ## Subprocess safety Every `git` call goes through `subprocess.run` with list-arg argv (no `shell=True`). SHA + path inputs are caller-supplied but never interpolated into a shell string, so the standard shell-injection vector is closed. ## Hard rules preserved - ✅ Rule 9 — no schema change (read-only consumer of existing JSON) - ✅ Rule 16 — N/A (no scoring change) - ✅ Rule 18 — diagnostic ships in same PR (the loader IS the diagnostic surface for downstream IC / distribution-shift work) - ✅ License — pure stdlib + pandas + subprocess; no new deps - ✅ No frontend touched ## Schema-compat Handles every rankings.json shape from `0.5.x` onward (pre-LedgerCraft reskin + Phase 4.5e + Phase 4h.x + Phase 3c). Top-level keys `ticker` + `composite_score` are present in every historical snapshot per ``compute/output/writer.py``. Rows missing those default-required keys get a debug log + are silently skipped (rare; would indicate a pre-rankings.json commit or a schema-break we haven't shipped). ## Verification - `pytest tests/test_validation/test_ranking_history.py` — 18/18 pass in 3.95s (5 use real git; 13 use stdlib mocks) - Full `pytest tests/test_validation/` — 83/83 pass in 50.24s (no regressions on #2 first-unit drift tests / #2-precursor pbo_dsr universe-provider tests / pre-existing ic_decay + osap_validation) - `ruff check` — clean - Test count delta: +18 ## Next in the chain Per `docs/research/historical-revalidation-harness.md` Future-work TODO: - #2b forward-return computation per ticker from `compute/cache/prices/` (0.5d) — gitignored cache; needs warm CI run - #2c per-pillar IC at historical dates (1d, needs 2a + 2b) - #2d PBO/DSR re-baseline via PR #275's `universe_provider` kwarg (1d, needs 2c) - #2e `manipulation_index` distribution shift report (0.5d, needs 2a) - #2f `docs/research/honest-baseline-2026-05-27.md` (0.5d, needs 2d) ## NOT in this PR - The DataFrame is built from git history; no IC math, no PBO/DSR call sites, no chart rendering. Downstream PRs land that. - No CLI wrapper. Direct module import expected; if a CLI feels warranted later, it can wrap this in 30 LOC + 5 tests. Co-authored-by: Claude <noreply@anthropic.com>
dackclup
added a commit
that referenced
this pull request
May 27, 2026
…n shift (#279) Closes follow-up #2e from ``docs/research/historical-revalidation-harness.md``. Consumes the #2a rankings.json time-series loader (PR #278) and reports the manipulation_index distribution shift across the cron's lifetime: mean / std / quantiles + fire-rate by band (LOW [0,20), MODERATE [20,50), HIGH [50,∞)) per date + window-end deltas. This answers the honest question — **has the cohort of flagged stocks materially changed across the cron's history?** A universe-mean drift > ~5 pts would signal Phase 4.5e weight recalibration is needed (per Q3 2026-08-19 cohort-audit gate). ## What ships - `compute/validation/manipulation_distribution.py` (NEW, ~250 LOC): - `DistributionSummary` dataclass — per-date snapshot (n / mean / std / median / q25 / q75 / q95 / max + band counts + top-3 tickers) - `ShiftReport` dataclass — aggregate across window + first-to-last deltas (mean_delta / std_delta / high_count_delta) + note - `compute_manipulation_distribution_shift(start_date, end_date, repo=None)` — main entry; pure-function wrapping #2a's loader - `format_shift_report(report, max_dates=20)` — human-readable text rendering; truncates long windows with "..." marker - `tests/test_validation/test_manipulation_distribution.py` (NEW, 11 tests): - Band-boundary constants pin (LOW < 20 ≤ MODERATE < 50 ≤ HIGH) - `_summarize_one_date` band partition correctness + empty + top-3 ordering - `compute_shift` empty-window / all-null-window / single-date / two-date-with-deltas paths (monkeypatched loader) - `format_shift_report` header + delta + cap rendering - Live-git smoke against the repo's recent cron commits ## Real-world artifact (live repo, window 2026-05-01 → 2026-05-27) 3 cron dates available on main: date n mean std p75 p95 HIGH top 2026-05-22 502 4.38 9.28 5.00 25.00 2 SMCI=84.0, WAT=64.0, NVDA=48.0 2026-05-23 502 4.38 9.28 5.00 25.00 2 SMCI=84.0, WAT=64.0, NVDA=48.0 2026-05-26 502 4.38 9.28 5.00 25.00 2 SMCI=84.0, WAT=64.0, NVDA=48.0 Δmean=+0.00, Δstd=+0.00, ΔHIGH=+0 (2 → 2) Distribution is **stable** across the window (expected for 5-day horizon — pillar inputs change slowly). Top-3 invariant: SMCI 84, WAT 64, NVDA 48 — matches Phase 4.5f production-verified ``ManipulationRiskCard`` fire-rate snapshot. Universe mean 4.38 sits solidly in LOW band; only 2 tickers in HIGH band (Phase 4.5f spec target: 1-3 stocks). No recalibration signal. A longer window (≥ 90 days) would let this report detect drift; the chain is now ready when cron history accumulates. ## Hard rules preserved - ✅ Rule 9 — no schema change (read-only consumer of rankings.json) - ✅ Rule 16 — N/A (no scoring change) - ✅ Rule 18 — diagnostic surface ships in same PR - ✅ License — pure stdlib + pandas; no new deps - ✅ Universe S&P 500 only ## Verification - `pytest tests/test_validation/test_manipulation_distribution.py` — 11/11 pass in 0.70s - `ruff check` — clean (linter trimmed unused imports + sorted) - Live-git artifact above generated cleanly ## Next in the chain Per ``docs/research/historical-revalidation-harness.md`` Future-work TODO: - #2b forward-return computation from `compute/cache/prices/` (0.5d) — gitignored cache; CI-only data - #2c per-pillar IC at historical dates (1d, needs 2a + 2b) - #2d PBO/DSR re-baseline via PR #275's `universe_provider` kwarg (1d, needs 2c) - #2f honest-baseline report (0.5d, needs 2d) #2e (this PR) is independent of 2b/2c — could ship before, in parallel with, or after them. Co-authored-by: Claude <noreply@anthropic.com>
This was referenced May 27, 2026
Merged
dackclup
added a commit
that referenced
this pull request
May 27, 2026
…tes (#281) New `compute/validation/historical_ic.py` orchestrator pairs PR #278's `load_ranking_history` (ranking at T) with PR #280's `compute_forward_returns_batch` (realized return at T + horizon) and computes per-pillar Spearman IC across the historical window — closes the IC re-baseline half of the Phase 4.6 chain. API: - `compute_pillar_ic(scores, returns, *, method, min_tickers)` pure cross-sectional IC for one (pillar, date) pair - `compute_historical_ic_report(start, end, *, horizon_months, pillars, ...)` walks rankings.json snapshots + forward returns cache, aggregates into `HistoricalICReport` - `format_ic_report(report)` human-readable text rendering - `PillarICEntry` / `PillarICSummary` / `HistoricalICReport` three frozen-dataclass carriers Spearman computed as Pearson on rank-transformed series (Spearman 1904 + Conover 1999 §5.4) to avoid pulling scipy into the dep set (QuantRank ships without scipy; pandas' `Series.corr(method= 'spearman')` requires it transitively). Drops with descriptive notes: - cross-section < MIN_TICKERS_PER_DATE = 30 (Grinold-Kahn 2000 §4.2) - None / NaN / inf in either input - constant inputs (std=0 → correlation undefined) Aggregates per pillar: mean / std / median / min / max / IC IR / hit-rate. IC IR = mean/std × sqrt(n_dates) (Grinold-Kahn 2000 §4.4). Hit-rate = fraction of dates with strictly positive IC. Honest-baseline disclaimer per Research Report v1.0: - IC reported here is NAIVE — no costs / slippage / sector neutralization. Real net-of-cost IC typically 30-50% smaller per McLean-Pontiff 2016 JF post-publication decay - The historical universe MUST come from PR #274 members_at() to avoid survivorship bias (Hou-Xue-Zhang 2020 RFS); orchestrator reads the historical universe FROM rankings.json at as-of T which is correct by construction (snapshot itself is historical universe) - Report is a TIME SERIES + summary, not a single headline number Tests: 28 new (28 passing). Coverage: module constants, pure IC computation edge cases (perfect ±1.0, constant inputs, NaN drops, below-min cross-section, method validation), summary aggregation math (IC IR formula pinned, hit-rate semantics), orchestrator full-path (one date / multi-date / missing pillar / malformed JSON), text rendering, and a live-git smoke that auto-degrades gracefully when the gitignored price cache is absent. Schema impact: zero. No new Pydantic / TS / snapshot field. Production-wiring impact: zero. No compute/main.py import. The orchestrator is purely a validation / re-baseline tool. Downstream PRs (#2d PBO/DSR re-baseline + #2f honest-baseline report) consume the output. Phase 4.6 chain status: 5 of 6 items now landed (#1/#2 PR #277, #2a PR #278, #2b PR #280, #2c this PR, #2e PR #279; #2d gate kwarg shipped PR #275). #4 PBO/DSR re-baseline needs a warm-CI execution to publish actual numbers; #6 honest-baseline doc closes the chain. PHASE_STATUS_INFLIGHT.md updated per PR #237 side-file convention. Harness doc TODO list updated: 5 of 6 items now landed. https://claude.ai/code/session_01AGU8d6pm4u2fQQ5cebg9qa Co-authored-by: Claude <noreply@anthropic.com>
6 tasks
dackclup
added a commit
that referenced
this pull request
May 27, 2026
…on + CLI (closing the chain) (#282) Closes the Phase 4.6 honest re-validation harness structurally — 6 of 6 chain items now landed. The only remaining work is a warm-CI execution session that fills the TBD numeric cells. New artifacts: - docs/research/honest-baseline-2026-05-27.md — 10-section skeleton with TBD numeric cells in §2 (per-pillar IC) / §3 (PBO/DSR) / §4 (manipulation distribution) / §5 (survivorship-bias delta). Methodology + framing + honest-α ceiling + disclaimer ladder final-form. Citation block: Hou-Xue-Zhang 2020 RFS, McLean-Pontiff 2016 JF, Bailey-Lopez de Prado 2014 JPM, Bailey-Borwein-Lopez de Prado-Zhu 2014 AMS Notices, Grinold-Kahn 2000, Spearman 1904, Conover 1999, Kissell-Glantz 2003. - scripts/generate_honest_baseline.py — argparse CLI that wires compute_historical_ic_report (PR #281) + compute_manipulation_ distribution_shift (PR #279) end-to-end. Text mode emits the disclaimer banner to stderr; JSON mode embeds __banner__ in the payload. Exit codes: 0 (report produced), 1 (input validation), 2 (empty report — useful CI signal). - tests/test_validation/test_generate_honest_baseline_cli.py — 17 tests: argparse shape, _parse_date, exit codes, banner emission, JSON payload shape + α ceiling cells + disclaimer string, banner embedding, _report_to_payload with synthetic + populated manip reports, and a constant pin on the banner's 5 mandatory phrases (NAIVE / McLean-Pontiff / 2-5% / Rule 16 / S&P 500). Honest-baseline disclaimer per Research Report v1.0 autonomous mission: - IC / PBO / DSR figures NAIVE — no costs / slippage / sector neutralization - Real net-of-cost IC typically 30-50% smaller per McLean-Pontiff 2016 JF 32% post-publication decay - Honest net α ceiling: 2-5% per year (hard-coded into doc + JSON) - Composite formula sacred (Rule 16) — never replayed retroactively - Universe = S&P 500 (502) only - No trade recommendation of specific tickers — methodological report only Schema impact: zero. No new Pydantic / TS / snapshot field. Production-wiring impact: zero. No compute/main.py import. Smoke run against real repo's recent rankings.json (no live price cache) — orchestrator walks 3 commits, returns n_dates_with_ic=0, exit code 2 surfaces missing-cache signal cleanly. Phase 4.6 chain status — 6 of 6 items structurally landed: - #1/#2 universe-drift first unit: PR #277 - #2a ranking_history loader: PR #278 - #2b forward_returns loader: PR #280 - #2c per-pillar IC orchestrator: PR #281 - #2d PBO/DSR gate kwarg: PR #275 (warm-CI execution pending) - #2e manipulation_distribution shift: PR #279 - #2f honest-baseline skeleton + CLI: this PR Deferred follow-ups (NOT in this PR): warm-CI execution session, --markdown writer mode, --include-pbo-dsr factor-return wiring. PHASE_STATUS_INFLIGHT.md updated per PR #237 side-file convention. Harness doc TODO list: 6 of 6 items now landed. https://claude.ai/code/session_01AGU8d6pm4u2fQQ5cebg9qa Co-authored-by: Claude <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Closes the Phase 4.6 next-task gap surfaced by Research Report v1.0 §7.4 follow-up. PR #274 landed
compute/ingest/historical_universe.members_at()as a library but nothing called it. This PR threads an optionaluniverse_providercallable throughfactor_passes_gates()so the returned metrics dict carries honest universe provenance — closing the loop between the historical-membership module and the validation gates every Phase 4+ candidate factor must pass.Backward-compatible: every pre-Phase-4.6 caller (osap-integration / jkp-integration / qlib scout / ipca scout already-merged PRs) is byte-identical because all 3 new kwargs default to None.
What changed
compute/validation/pbo_dsr.pyfactor_passes_gates()gains 3 optional kwargs (universe_provider,as_of_date,current_universe). When all 3 passed, metrics dict carriesuniverse_as_of/universe_size/survivorship_bias_corrected. Graceful degradation when provider raises. New helpertoday_utc_date()for forward-cron callers.tests/test_validation/test_pbo_dsr.pyNew API surface (sample)
Hard rules preserved
TYPE_CHECKINGforward import forMembershipResult)Why this matters (anchor)
Hou-Xue-Zhang (2020) RFS replication-crisis evidence emphasizes survivorship as the primary failure mode in factor-zoo work. Pre-Phase-4.6 every PBO/DSR number was implicitly survivorship-biased without saying so; post-Phase-4.6 the metrics dict can carry honest provenance. This PR is the first place that knowledge surfaces in the validation metric dict.
Verification
ruff checkCallabletocollections.abcper modern convention; deliberate)pytest tests/test_validation/test_pbo_dsr.py tests/test_ingest/test_historical_universe.py tests/test_config.pytest_factor_passes_gates_backward_compat_no_universe_kwargsTest plan
ruff check .— cleanpytest tests/test_validation/— 29/29 pass (23 legacy + 6 new)NOT in this PR (next follow-ups per priority order)
compute/main.pywriter wiring — populateMetadata.universe_membership_as_of+survivorship_bias_correctedfrom the metrics dict at JSON write time. Small companion PR; would make the diagnostic visible in production JSON output.manipulation_indexwith the historical universe (likely revises some published PBO/DSR baselines DOWNWARD by 5-15%). Explicit honest-correction PR.<aff10b5One>plan_adoption_date extraction; needs minimal lxml parser first.Subscribe-after-open suggestion: same pattern as #271-#274 — subscribe me to PR activity for CI + review comments.
Generated by Claude Code