AltimateAI / AltimateAI/altimate-code

feat(review): expand grain-key not_null detector to fire on any contracted-model column change

Offen
#1,034 0 Kommentare 0 Reaktionen 0 zugewiesene Personen Auf GitHub ansehen
Vorherrschende Sprache
TypeScript
Sterne
811
Forks
134
Ø Merge
3 T. 2 Std.
Gemergte PRs (30 T.)
50

Beschreibung

Deferred from the v0.9.3 pre-release review.

### Current behavior

`extractGrainKeyGaps` in `packages/opencode/src/altimate/review/dbt-patterns.ts` fires only when the grain declaration itself changes, filtered by `grainDeclChangedEntities` — which iterates `newMap` and compares against `oldMap` for additions.

### Gap

A PR that adds a new column to an existing contracted model without touching `unique_combination_of_columns` will not surface uncovered grain columns, even when new columns increase the exposure surface. The scoping is conservative and correct for the initial ship, but leaves recall on the table.

### Proposal

Extend the change-scoping filter to also fire when any column added/modified in a contracted model's schema.yml overlaps with existing grain-key coverage. Keep the current "grain declaration changed" path as one of two entry conditions.

### Why deferred

Scope expansion touches filter semantics and could shift the false-positive rate; the initial detector is intentionally conservative and documented, so a follow-up release is the right cadence for widening.

Beitragsleitfaden

Beitragsleitfaden öffnen

Rechercherichtung

Start in packages/opencode/src/altimate/review/dbt-patterns.ts at extractGrainKeyGaps and the grainDeclChangedEntities filter. Trace how newMap and oldMap identify additions, then inspect how contracted-model schema.yml column changes are represented. Done means the detector preserves the existing grain-declaration path and also evaluates added or modified columns that overlap existing grain-key coverage without unnecessarily broadening unrelated changes.

Vom Indexierungsmodell aus dem Issue-Text verfasst.

Bewertung

Tech-Stack
typescript
Bereich
data-engineering
Issue-Typ
Feature
Schwierigkeit
4/5
Geschätzter Aufwand
3-5 Tage
Aktivitätsstatus
Ruhig
Klarheit
Größtenteils klar
Anfängerfreundlichkeit
48/100

Neue Issues direkt in Ihr Postfach

Eine kurze Übersicht über anfängerfreundliche GitHub-Issues.