Description
After the C1 witness: per-value and per-metric series counts from cardinality/label_values for capped, prioritised labels; feeds M-12, M-13 and M-22 inputs. Bumps catalogue_version. Serial with the other E tasks (shared label_inventory.py, label_rules.json, budget.py, hydrate.py).
Acceptance Criteria
- #1 Contract tests over the witnessed response shape
- #2 catalogue_version is bumped and rules depending on this input move from route_parked to evaluated
Definition of Done
- #1 just test
- #2 just tf-validate
- #3 just check-identifiers and just no-em-dashes both return clean
Implementation Plan
Consume frozen accepted C1 witness and source/evaluator envelope. Add capped prioritized Mimir label_values series-per-value/per-metric reductions only, no raw values/metric names persisted; activate corresponding reserved verified rules and bump catalogue. Serial E-set owner, return reviewed candidate for root stable composed gate and exact CI.
Implementation Notes
loop14:firstE1candidate activates verifiedM5/catalogue2 and realM4 populations/head shapes only;146narrowtests/6171subtests, guarded adversarialPASS topN/atlimit/population/empty/error/privacy, CodeRabbitcomplete0criticalmajor. Minor completeempty samples retained because evaluator missinginput neverpasses. Root4syntheticCLIview refresh plus trivialfakeHTTPrequest dispatch integration (ALL assertions/floors/canaries unchanged); failedoldfixture fullgate then changeded293ce fullgate2330passed/2skips, exactCI37587541925 allrequiredleavesgreen. Source07 exhaustedceiling unchanged, no transportrepair or unverified policy.