From 723dfe7bdc99b5db12151a614c14ffca31ba11e2 Mon Sep 17 00:00:00 2001 From: root Date: Fri, 17 Jul 2026 19:58:13 +0000 Subject: [PATCH] fix(equivalences): one-time triage of the 289,681-row pending backlog Set-based port of scheduler.ts evaluateEquivalenceResearch(), applied once live to every pending row with re_research_due_at IS NULL (the root cause this pairs with a code fix for, previous commit). Result: 41,963 approved, 187,599 rejected as technical mismatch, 57,683 rejected as insufficient evidence (re-research in 21 days), 2,436 rejected for no recent competitor price. Backup of pre-triage state in equivalence_pending_triage_backup_20260713. --- sql/139-equivalence-pending-triage.sql | 130 +++++++++++++++++++++++++ 1 file changed, 130 insertions(+) create mode 100644 sql/139-equivalence-pending-triage.sql diff --git a/sql/139-equivalence-pending-triage.sql b/sql/139-equivalence-pending-triage.sql new file mode 100644 index 0000000..b44be55 --- /dev/null +++ b/sql/139-equivalence-pending-triage.sql @@ -0,0 +1,130 @@ +-- sql/139-equivalence-pending-triage.sql +-- Set-based port of scheduler.ts's evaluateEquivalenceResearch() (line ~85), +-- applied once to the 289,681-row pending backlog in transceiver_equivalences. +-- +-- Root cause of the backlog: the daily "maintenance:re-research-equivalences" job +-- (scheduler.ts:3117) only evaluates rows where re_research_due_at IS NOT NULL AND +-- <= NOW(). catalog-reconcile.ts inserts new 'pending' rows WITHOUT setting +-- re_research_due_at (NULL by default) -> the job that was designed to drain the +-- queue never saw fresh pending rows. This migration (a) triages the existing +-- backlog once, set-based, using the exact same scoring rules as the JS +-- evaluator, and (b) is paired with code fixes (catalog-reconcile.ts sets +-- re_research_due_at on insert; scheduler.ts's reject branch preserves it for +-- "insufficient evidence" rejects) so the daily job keeps draining future inserts +-- on its own. +-- +-- Decision rules (identical to evaluateEquivalenceResearch, ported 1:1): +-- no recent competitor price observation (45d) -> reject, no re-research +-- form_factor or speed_gbps mismatch/missing -> reject, no re-research +-- wavelength diff >15nm / fiber mismatch / reach -> reject, no re-research +-- ratio <0.85 (all three "hard" if evidence exists) +-- wavelength / fiber_type / reach_meters MISSING -> reject, WITH re-research +-- (on either side, and no hard mismatch above) in 21 days (data-gap, +-- not a wrong match -- +-- give scrapers time to +-- enrich, then recheck) +-- confidence (score/115) >= 0.73 -> auto_approved +-- else -> reject, no re-research +-- +-- Idempotent for its WHERE clause (status='pending' AND re_research_due_at IS +-- NULL); re-running after the code fix lands is a no-op since new pending rows +-- will already carry a re_research_due_at. + +create table if not exists equivalence_pending_triage_backup_20260713 as +select *, now() as backed_up_at +from transceiver_equivalences +where status = 'pending' and re_research_due_at is null; + +with normalized as ( + select eq.id, + nullif(upper(trim(fx.form_factor)),'') as fx_form, nullif(upper(trim(cp.form_factor)),'') as cp_form, + fx.speed_gbps as fx_speed, cp.speed_gbps as cp_speed, + nullif(upper(trim(fx.standard_name)),'') as fx_std, nullif(upper(trim(cp.standard_name)),'') as cp_std, + nullif(upper(trim(fx.fiber_type)),'') as fx_fiber, nullif(upper(trim(cp.fiber_type)),'') as cp_fiber, + fx.reach_meters as fx_reach, cp.reach_meters as cp_reach, + (regexp_match(fx.wavelengths, '(\d{3,4})'))[1]::numeric as fx_nm, + (regexp_match(cp.wavelengths, '(\d{3,4})'))[1]::numeric as cp_nm, + exists(select 1 from price_observations po where po.transceiver_id=eq.competitor_id and po.time > now() - interval '45 days') as has_recent_price + from transceiver_equivalences eq + join transceivers fx on fx.id = eq.flexoptix_id + join transceivers cp on cp.id = eq.competitor_id + where eq.status = 'pending' and eq.re_research_due_at is null +), +scored as ( + select *, + (case when fx_form is not null and cp_form is not null and fx_form=cp_form then 25 else 0 end + + case when fx_speed is not null and cp_speed is not null and fx_speed=cp_speed then 20 else 0 end + + case when fx_std is not null and cp_std is not null and fx_std=cp_std then 30 else 0 end + + case when fx_nm is not null and cp_nm is not null then + case when abs(fx_nm-cp_nm)<=15 then 20 else -20 end + else 0 end + + case when fx_fiber is not null and cp_fiber is not null then + case when fx_fiber=cp_fiber then 10 else -15 end + else 0 end + + case when fx_reach is not null and cp_reach is not null and fx_reach>0 and cp_reach>0 then + case when least(fx_reach,cp_reach)/greatest(fx_reach,cp_reach) >= 0.85 then 10 else -15 end + else 0 end) as raw_score, + (fx_nm is null or cp_nm is null) as wl_missing, + (fx_fiber is null or cp_fiber is null) as fiber_missing, + (fx_reach is null or cp_reach is null or fx_reach<=0 or cp_reach<=0) as reach_missing, + (fx_nm is not null and cp_nm is not null and abs(fx_nm-cp_nm)>15) as wl_mismatch, + (fx_fiber is not null and cp_fiber is not null and fx_fiber<>cp_fiber) as fiber_mismatch, + (fx_reach is not null and cp_reach is not null and fx_reach>0 and cp_reach>0 and least(fx_reach,cp_reach)/greatest(fx_reach,cp_reach) < 0.85) as reach_mismatch, + (fx_form is null or cp_form is null or fx_form<>cp_form) as form_bad, + (fx_speed is null or cp_speed is null or fx_speed<>cp_speed) as speed_bad + from normalized +), +decided as ( + select *, + greatest(0, least(1, raw_score::numeric/115)) as confidence, + (not has_recent_price) as no_price, + (wl_mismatch or fiber_mismatch or reach_mismatch or form_bad or speed_bad) as critical_mismatch, + (wl_missing or fiber_missing or reach_missing) as missing_evidence + from scored +) +update transceiver_equivalences eq +set status = case + when d.no_price then 'rejected' + when d.critical_mismatch then 'rejected' + when d.missing_evidence then 'rejected' + when d.confidence >= 0.73 then 'auto_approved' + else 'rejected' + end, + confidence = case when d.no_price then 0 else d.confidence end, + reviewed_by = 'automated-research-bulk-139', + reviewed_at = now(), + reject_reason = case + when d.no_price then 'automated research: competitor has no recent price observation' + when d.critical_mismatch then 'automated research: technical mismatch' + when d.missing_evidence then 'automated research: insufficient technical evidence' + when d.confidence >= 0.73 then null + else 'automated research: confidence below approval threshold' + end, + re_research_due_at = case + when (not d.no_price) and (not d.critical_mismatch) and d.missing_evidence and d.confidence < 0.73 + then now() + interval '21 days' + else null + end, + re_researched_at = now(), + match_notes = case when eq.match_notes is null or eq.match_notes = '' + then 'bulk-triage(139) ' || now()::date + else eq.match_notes || ' | bulk-triage(139) ' || now()::date end, + updated_at = now() +from decided d +where eq.id = d.id; + +-- Mirror the daily job's side effect on approve: mark the Flexoptix product as +-- competitor_verified (matches scheduler.ts's confirm branch). +update transceivers t + set competitor_verified = true, + competitor_verified_at = coalesce(t.competitor_verified_at, now()) + from transceiver_equivalences eq + where eq.flexoptix_id = t.id + and eq.status = 'auto_approved' + and eq.reviewed_by = 'automated-research-bulk-139' + and eq.reviewed_at > now() - interval '5 minutes' + and t.competitor_verified = false; + +insert into _migrations (filename, applied_at) +select '139-equivalence-pending-triage.sql', now() +where not exists (select 1 from _migrations where filename = '139-equivalence-pending-triage.sql');