Skip to content

dbt PR review: precision, coverage and AI verifier stack #1792

Open

Description

@RaghwendraSingh

Tracking issue for a stack of pull requests that improve altimate review (the dbt PR review) on three fronts:

  • Precision: review only the PR's own changes (real base, merge-base diff, project scope), fewer and better-placed comments, no false blocks from engine errors or noise rules.
  • Coverage: incremental/snapshot semantics, contracts and dialect-aware checks, data-test adequacy, deterministic rules for macro, test and yml defects, and reviewing every dbt project a PR touches.
  • AI lane: an opt-in verifier that confirms or refutes each AI comment from dbt evidence, wider AI input coverage, and focused batches. All AI features are opt-in; the default review adds no model calls.

The PRs are stacked; each is based on the previous one and should be merged in order:

  1. fix(review): review the PR's own changes: real base, merge-base diff, project scope, renames, model-paths, no raw Jinja #1719 review the PR's own changes
  2. fix(review): an honest PR comment: real lint-only banner, AI lane status, PR intent, project_dir #1720 an honest PR comment
  3. fix(review): a readable PR comment: diff-anchored inline comments, no reposts, grouped rules, size budget, new-only PII, no portability noise #1721 a readable PR comment
  4. fix(review): verdict trust: only the engine blocks, policy from the base branch, allowForceTier #1722 verdict trust
  5. fix(review): failures are never clean: engine errors reported, per-file isolation, PII abstention, gate dismissal #1723 failures are never clean
  6. feat(review): consumers outside dbt, engine-aware integer division, opt-in contract and column-list checks #1724 consumers outside dbt, engine-aware integer division, contract checks
  7. feat(review): incremental and snapshot semantics that fit any dbt project #1725 incremental and snapshot semantics
  8. feat(review): close remaining review gaps: action dbt_prepare, data-test adequacy, parse-only column changes #1726 remaining review gaps
  9. fix(review): precision and new rules from a replay of human-caught defects on public dbt PRs #1739 precision and new rules from a replay of public dbt PRs
  10. feat(review): post fewer, better-placed comments: rule tiers, inline budget, pre-existing tags, coverage ledger #1758 posting policy
  11. feat(review): opt-in AI verifier that confirms or refutes each AI comment from dbt evidence #1769 opt-in AI verifier
  12. feat(review): 11 deterministic rules for macro, test and yml defects #1770 deterministic rules for macro, test and yml defects
  13. feat(review): opt-in AI input coverage, hunk excerpts and a dbt manifest context card #1771 opt-in AI input coverage
  14. feat(review): opt-in test-gap check and focused AI batches; verifier reads every changed file #1772 opt-in test-gap check and focused AI batches
  15. feat(review): --all-projects reviews every dbt project a PR touches in one run #1773 --all-projects

Side draft: #1774 (AI evidence line, suggestions, base-branch repo instructions).

Related: #1240, #864, #894.

Activity

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions