Loop r5: re-quarantine lab fixture (IDENTITY-flood diagnosis was wrong; observation not IDENTITY) pending B re-diagnosis; start C in parallel

Author: nullflux

utilities/comparison-fixtures/LOOP_round5.md

@@ -171,3 +171,7 @@ fixture answers the guard's (act,Q) at a different group.)
   zz_var_datamart_enrich intentionally left hardcoded (below flood, collision-safe). L10 hazard
   eliminated for all flood-range nbs_case_answer blocks. d_place -6 = FLAKY (3rd timing-variant dim
   with covid_contact/d_contact_record; went 37->37->31 across runs independent of fixture changes).
+- Phase B BLOCKED on re-diagnosis: the tick5 lab-fixture regression (6 obs tables emptied) was
+  MIS-attributed to IDENTITY flood, but observation/act are NOT identity columns (survey) -> real cause
+  unknown. Re-quarantined zz_lab100_101_fill.sql (.regresses-6-obs-tables-cause-TBD); spawned read-only
+  B-diagnosis agent. Proceeding with Phase C (COVID dedicated patient+PHC, 22055xxx) in parallel.

utilities/comparison-fixtures/catalog/uid_ranges.md

@@ -1620,3 +1620,4 @@ tail-EXEC sequences these correctly for standalone verification.
 | 22052000 - 22052999 | R4-M TB tail debug/fix (tb_datamart 2nd row) |
 | 22053000 - 22053999 | R4-N lab100 + lab101 — **allocated 2026-06-03**, `zz_lab100_101_fill.sql`. ODSE-ONLY (NO-SHORTCUT). Part A (LAB100 36/69→demographics): new fully-attributed Order `22053010` + Result child `22053011` (LOINC 13950-1, cond 10110 baseline-seeded) with participations PATSBJ→20000000 / ORD,VRF→20000010 / AUT,ORD(org)→20000020 / PRF→20000020 + role SPP(PSN)→20000010 (specimen collector) → CDC sets nrt_observation.patient_id/ordering_person_id/author_organization_id/ordering_organization_id/specimen_collector_id → sp_d_labtest_result_postprocessing (017) resolves D_PATIENT 4 / D_PROVIDER 12 / D_ORGANIZATION 7 → fills PERSON_*/PATIENT_*/PROVIDER_*/ORDERING_FACILITY/REPORTING_FACILITY*/ACCESSION_NBR. ROOT CAUSE of NULLs: v2 lab 20070010 had ZERO participations. Part B (LAB101 0/46→unblock): full ODSE I_Order `22053500` + I_Result `22053501` + 'Result' `22053502` + 35 LABxxx I_Result children `22053600-22053634` (cd LAB329a,LAB330..LAB363) each + obs_value_coded(display_name→TEST_RESULT_VAL_CD_DESC) + obs_value_txt, wired child→I_Order via COMP act_relationship so sp_observation_event (055) emits them in the I_Order followup_observations JSON → service sets nrt_observation.followup_observation_uid → sp_lab101 (020) step 2 #tmp_I_Result_vals resolves → 35-col trtdN pivot (LAB1..LAB35) lands. ROOT CAUSE of 0/46: the shortcut-era zz_lab101_unblock.sql injected RDB LAB_TEST/LAB_RESULT_VAL directly but its nrt_observation INSERT was stripped on no-shortcut → root-order nrt_observation never existed → chain starved. UIDs consumed: 22053010/22053011 (Part A acts/obs), 22053500/22053501/22053502 (Part B I_Order/I_Result/Result), 22053600-22053634 (35 LABxxx children). **ORCH_TODO:** (1) add `22053010,22053011,22053500,22053501,22053502` to BOTH the `sp_observation_event @obs_id_list` and the two postprocessing `@obs_ids` lists in `scripts/merge_and_verify.sh run_lab_chain()` (~lines 312-315); (2) add `22053011,22053501` to `LAB_OBS_UIDS` (~line 473). SEED note: LAB100/LAB101 are NOT subject to the covid_lab 11065-LOINC seed gate (bug #16) — the chain uses baseline-seeded LOINC 13950-1→cond 10110 and culture codes, so CONDITION_CD/PROGRAM_AREA fill without seed edits. |
 | 22054000 - 22054999 | R4-O hepatitis_datamart remainder |
+| 22055000 - 22055999 | R5-C COVID dedicated patient/provider/org + enriched PHC |