feat(freshness-monitor): historical-mode probe — daily gap detection across all artifacts#339
Merged
Merged
Conversation
…across all artifacts
Closes the gap surfaced 2026-05-28: current-state probe answers
'is the artifact present now?' but operators also need 'did it
land last weekend? are there gaps in the producer's history?'
Filed per the same feedback memory observe_mode_unconditional_gates
— absence-of-artifact is the failure mode, and a single-cycle
absence could be a false-positive where a multi-cycle gap is a
real producer regression.
Adds:
- event['mode']='historical' dispatch in handler(). Routes to a
new _handle_historical(s3, now, started_at, lookback_overrides)
path that walks the registry, probes the last N cycles per
artifact, and writes _freshness_monitor/history.json (page 26
will surface per-row history expanders + gap counts).
- New EB cron alpha-engine-freshness-monitor-historical-cron
(daily 04:00 UTC, off-peak) wired in deploy.sh --bootstrap.
- Default lookback: 12 saturday_sf + 30 weekday_sf/eod_sf cycles
(~3 months each). continuous skipped (current-state covers).
Tunable via event['lookback'] override.
403/404/NoSuchKey normalization: S3 returns 403 (not 404) for
missing keys when the Lambda lacks s3:ListBucket. Treat both as
cleanly-absent (no error_code in output) so page 26 doesn't show
spurious '403 errors' on legitimately-absent historical cycles.
9 new unit tests cover: saturday/weekday/eod cycle-date
resolution, continuous skip, zero-count short-circuit,
date/trading_day/no-placeholder template rendering, and handler
mode-dispatch.
Live smoke (post-deploy + manual invoke):
n_artifacts=51, n_cycles_probed=474, duration=10.08s
Surfaced 1 real finding for follow-up: several artifacts use
calendar-vs-trading-day-anchored templates that don't match
producer behavior. research_signals registered as
signals/{date}/signals.json with cadence=saturday_sf, but
producer writes to mostly Friday trading-day keys (2026-05-22,
2026-05-15, etc.). The historical probe correctly reports the
Saturday keys as absent — which IS the right answer given the
registry template. ROADMAP follow-up filed separately to audit
all registry templates for calendar-vs-trading-day mismatch.
Calendar-naive by design — NYSE holidays surface as
false-positive absent days but operators can interpret in
context. Calendar-aware backfill is a P3 follow-up if the
noise becomes worth the dependency lift.
Composes with the OBSERVATION_REGISTRY arc (#349/#351/#352/#355
+ #135/#136/#137).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Merged
3 tasks
cipher813
added a commit
that referenced
this pull request
May 28, 2026
…or historical probe (#341) Closes the calendar-vs-trading-day mismatch surfaced by PR #339's historical probe — research_signals was registered as signals/{date}/signals.json + cadence=saturday_sf but the producer writes to Friday trading-day keys (signals/2026-05-22/, /05-15/, /05-08/...). The pre-fix probe correctly reported the Saturday keys absent — but the operator-facing display showed 9/12 gaps when the true gap rate was lower; the registry just wasn't asking the right question. Restructure: - _iter_sf_firing_dates: returns the SF cron's calendar firing dates (last N Saturdays for saturday_sf, last N Mon-Fri for weekday_sf / eod_sf). No change from prior calendar-naive behavior at this layer. - _resolve_axis_dates: NEW. Translates firing dates to the date-axis the template actually uses. {date} → calendar firing date; {trading_day} → previous_trading_day(firing_date) for saturday_sf/weekday_sf, or firing_date itself for eod_sf (EOD writes today's close). alpha_engine_lib.trading_calendar.previous_trading_day IS NYSE- holiday-aware, so Memorial-Day-style holidays resolve correctly. - _iter_historical_cycle_dates: now takes optional template arg and composes the two helpers. Backward-compat: callers omitting template get calendar-axis (pre-PR behavior). - _handle_historical: passes spec.s3_key_template to the resolver. 6 new unit tests cover the trading_day-axis path for all three cadences + NYSE-holiday skip via the lib + backward-compat for the template-less call signature. Full suite: 27 passed (21 prior + 6 new). Live smoke (post-deploy + historical invoke after companion alpha-engine-config registry flip): research_signals — 12/12 gaps (under {date}) → 5/12 gaps (under {trading_day}). 4 cycles correctly recovered. Composes with alpha-engine-config registry PR (this branch's sibling): flips research_signals, research_consolidated_morning, scanner_candidates_json from {date} → {trading_day}. 4 backtest_* entries documented + held at {date} pending producer-side audit (backtester writes to ad-hoc current-date-of-write, neither {date} nor {trading_day}). Per the system's now_dual() convention codified in alpha-engine-docs/private/DATE_CONVENTIONS.md — trading_day = last_closed_trading_day(now); this PR brings the registry + historical probe into compliance with that convention. Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
cipher813
added a commit
that referenced
this pull request
May 28, 2026
…342) * feat(freshness-monitor): template-aware trading-day-axis resolution for historical probe Closes the calendar-vs-trading-day mismatch surfaced by PR #339's historical probe — research_signals was registered as signals/{date}/signals.json + cadence=saturday_sf but the producer writes to Friday trading-day keys (signals/2026-05-22/, /05-15/, /05-08/...). The pre-fix probe correctly reported the Saturday keys absent — but the operator-facing display showed 9/12 gaps when the true gap rate was lower; the registry just wasn't asking the right question. Restructure: - _iter_sf_firing_dates: returns the SF cron's calendar firing dates (last N Saturdays for saturday_sf, last N Mon-Fri for weekday_sf / eod_sf). No change from prior calendar-naive behavior at this layer. - _resolve_axis_dates: NEW. Translates firing dates to the date-axis the template actually uses. {date} → calendar firing date; {trading_day} → previous_trading_day(firing_date) for saturday_sf/weekday_sf, or firing_date itself for eod_sf (EOD writes today's close). alpha_engine_lib.trading_calendar.previous_trading_day IS NYSE- holiday-aware, so Memorial-Day-style holidays resolve correctly. - _iter_historical_cycle_dates: now takes optional template arg and composes the two helpers. Backward-compat: callers omitting template get calendar-axis (pre-PR behavior). - _handle_historical: passes spec.s3_key_template to the resolver. 6 new unit tests cover the trading_day-axis path for all three cadences + NYSE-holiday skip via the lib + backward-compat for the template-less call signature. Full suite: 27 passed (21 prior + 6 new). Live smoke (post-deploy + historical invoke after companion alpha-engine-config registry flip): research_signals — 12/12 gaps (under {date}) → 5/12 gaps (under {trading_day}). 4 cycles correctly recovered. Composes with alpha-engine-config registry PR (this branch's sibling): flips research_signals, research_consolidated_morning, scanner_candidates_json from {date} → {trading_day}. 4 backtest_* entries documented + held at {date} pending producer-side audit (backtester writes to ad-hoc current-date-of-write, neither {date} nor {trading_day}). Per the system's now_dual() convention codified in alpha-engine-docs/private/DATE_CONVENTIONS.md — trading_day = last_closed_trading_day(now); this PR brings the registry + historical probe into compliance with that convention. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * refactor(email): delegate SMTP+SES dispatch to alpha_engine_lib.email_sender (L4356) `emailer.py::send_step_email` now builds subject/body/HTML and delegates the Gmail SMTP + SES fallback dispatch to `alpha_engine_lib.email_sender.send_email` (the L4356 chokepoint). Drops the local 50-line dispatch boilerplate + the `smtplib`/`boto3` imports. Same secret-resolution semantics (`EMAIL_SENDER`/`EMAIL_RECIPIENTS`/`GMAIL_APP_PASSWORD`/`AWS_REGION`), same Gmail-primary-SES-fallback ordering, same never-raises contract. ROADMAP: **L4356 part 1/5** — sibling PRs follow in alpha-engine-backtester, alpha-engine, alpha-engine-research, and alpha-engine-predictor. Suite: 1675 passed. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
cipher813
added a commit
that referenced
this pull request
May 28, 2026
) deploy.sh --bootstrap was tripping on the put-targets shorthand when the Input contains JSON. From PR #339 the historical cron's target carries Input={"mode":"historical"}; the shorthand form Id=N,Arn=...,Input={...} confuses argparse on the embedded quotes + comma. Caught live 2026-05-28 when --bootstrap re-run partial-fired: ParamValidation: Error parsing parameter '--targets': Expected: '=', received: '"' for input Switch to a temp-file JSON spec via file://. The shorthand for the original 15-min cron stays unchanged because that target has no Input payload. The historical EB cron is wired live in AWS — target attached manually via direct CLI during this session with the same JSON. First daily firing at 04:00 UTC. This commit ensures the next operator who runs --bootstrap from main doesn't trip the same trap. Co-authored-by: Claude Opus 4.7 (1M context) <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
1 participant
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 gap surfaced 2026-05-28: current-state probe answers "is it present now?" but operators need "are there gaps in the producer's history?" The historical probe walks each artifact's last N cycles, HEAD-checks them, and writes
_freshness_monitor/history.jsonfor page 26 to surface per-row history + gap counts.Architecture
event['mode']='historical'dispatch inhandler(). Same Lambda, different event branch — separate code path; current-state probe stays fast (15-min cadence).alpha-engine-freshness-monitor-historical-cron(daily 04:00 UTC, off-peak) wired indeploy.sh --bootstrapwith target Input{"mode":"historical"}.continuousskipped (current-state already covers). Tunable viaevent['lookback']override.{cadence, severity, owner_repo, s3_key_template, is_latest_pointer, lookback_cycles, gap_count, continuous, history:[{date, present, size?, last_modified?, error_code?}, ...]}.s3:ListBucket. Both treated as cleanly-absent (noerror_codein output) so page 26 doesn't show spurious "403 errors".Test plan
n_artifacts=51, n_cycles_probed=474, duration=10.08serror_codein output after redeploy)Real finding surfaced by this probe
research_signalsis registered withs3_key_template: signals/{date}/signals.json+cadence: saturday_sf, but the producer writes to mostly Friday trading-day keys (2026-05-22, 2026-05-15, 2026-05-08...). The historical probe correctly reports the Saturday keys as absent — that IS the right answer given the registry template. ROADMAP follow-up: audit all registry templates for calendar-vs-trading-day mismatch. Multiple artifacts probably have the same issue.This is exactly the bug class the probe was designed to catch. Composes with [[feedback_observe_mode_unconditional_gates_govern_cutover]].
Composes with
Follow-on PR
alpha-engine-dashboard PR for page 26 per-row history expander + gap count summary (reads
history.jsonwritten by this Lambda).🤖 Generated with Claude Code