Make dbt PR reviewer demo-parity reliable#919
Merged
Conversation
|
Note Reviews pausedIt looks like this branch is under active development. To avoid overwhelming you with review comments due to an influx of new commits, CodeRabbit has automatically paused this review. You can configure this behavior by changing the Use the following commands to manage reviews:
Use the checkboxes below for quick actions:
No actionable comments were generated in the recent review. 🎉 ℹ️ Recent review info⚙️ Run configurationConfiguration used: Repository UI Review profile: CHILL Plan: Pro Run ID: 📒 Files selected for processing (4)
🚧 Files skipped from review as they are similar to previous changes (4)
📝 WalkthroughWalkthroughThis PR extends dbt SQL review with equivalence dialect support, refines structural severity categorization for join risks, implements diff-scoped PII detection with intelligent fallback, hardens DuckDB connection callback safety, and adds comprehensive test coverage across these areas. Changesdbt PR Review Improvements
🎯 4 (Complex) | ⏱️ ~75 minutes Possibly related PRs
Suggested labels
Suggested reviewers
Poem
🚥 Pre-merge checks | ✅ 3 | ❌ 2❌ Failed checks (2 warnings)
✅ Passed checks (3 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. ✨ Finishing Touches📝 Generate docstrings
🧪 Generate unit tests (beta)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 1
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In `@packages/drivers/src/duckdb.ts`:
- Around line 59-81: The onOpen callback can run synchronously during new
duckdb.Database(...) so ensure the returned database object is assigned to
instance immediately so onOpen never sees instance as undefined: construct the
Database and assign its return value to instance before relying on onOpen, then
keep the existing onOpen logic (checking instance.close and resolve(instance));
update the instantiation site where instance, opts, onOpen, duckdb.Database and
dbPath are used so the assignment happens synchronously and onOpen always
resolves/rejects with the real database object.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: Repository UI
Review profile: CHILL
Plan: Pro
Run ID: 56052e0b-8e0c-4260-b629-2920d8df241d
📒 Files selected for processing (11)
docs/docs/usage/dbt-pr-review.mddocs/internal/2026-06-08-dbt-pr-review-self-improvement-loop.mdpackages/drivers/src/duckdb.tspackages/opencode/src/altimate/native/altimate-core.tspackages/opencode/src/altimate/native/types.tspackages/opencode/src/altimate/review/orchestrate.tspackages/opencode/src/altimate/review/runner.tspackages/opencode/src/cli/cmd/review.tspackages/opencode/test/altimate/data-diff-duckdb-e2e.test.tspackages/opencode/test/altimate/review-runner.test.tspackages/opencode/test/altimate/review.test.ts
![cubic-dev-ai[bot]](https://avatars.githubusercontent.com/in/1082092?s=60&v=4){kind=small}
There was a problem hiding this comment.
4 issues found across 11 files
Reply with feedback, questions, or to request a fix.
Re-trigger cubic
dev-punia-altimate
requested changes
Jun 8, 2026
dev-punia-altimate
left a comment
dev-punia-altimate
left a comment
Contributor
There was a problem hiding this comment.
Multi-Persona Review — Verdict: block
The PR introduces a critical compliance violation by instantiating Settings at module level, directly violating CLAUDE.md guidelines for distributed environments. This risk is confirmed by both code reviewer and tech lead, and is compounded by a high-confidence logging bug that could break observability. While the feature is well-designed and tested, these foundational issues must be fixed before shipping.
15/15 agents completed · 220s · 5 findings (1 critical, 1 high, 1 medium)
Critical
- [code-reviewer, tech-lead] Module-level instantiation of Settings via get_settings() violates CLAUDE.md policy to avoid import-time side effects in distributed environments. →
app/utils/mcp_modules/semantic_metrics.py:617- 💡 Move
settings = get_settings()inside each function that uses it, or defer access until runtime (e.g., inside list_semantic_metrics).
- 💡 Move
High
- [code-reviewer] logger.info in list_all_semantic_metrics uses positional args instead of structured keyword arguments, breaking consistent log parsing across the codebase. →
app/utils/mcp_modules/semantic_metrics.py:620- 💡 Change to: logger.info("list_all_semantic_metrics", tenant=tenant, filter_provided=model_filter is not None)
Medium
- [tech-lead] Configuration settings are imported and instantiated at module level in a utility module, which may cause issues in test isolation and dependency injection. →
app/utils/mcp_modules/semantic_metrics.py:36- 💡 Move settings instantiation to function scope or inject via dependency injection pattern, especially since this module is used in MCP tooling where test contexts may need mock settings.
Low
- [tech-lead] Function name list_all_semantic_metrics is misleading — it does not 'list all' by default; it lists only those passing the filter. →
app/utils/mcp_modules/semantic_metrics.py:620- 💡 Rename to filter_semantic_metrics or list_semantic_metrics_filtered to better reflect its behavior when a filter is provided, and update docstring to clarify default behavior.
- [tech-lead] Tenant-to-platform routing is tightly coupled to a hardcoded setting (DATABRICKS_TENANTS) without abstraction or extensibility. →
app/utils/mcp_modules/semantic_metrics.py:620- 💡 Consider extracting tenant-to-platform routing into a dedicated service (e.g., PlatformRouter) to improve testability and future extensibility.
Multi-Persona Review · vllm:qwen3-next-80b (waves) + vllm-fallback (synth) ·
anandgupta42
force-pushed
the
codex/dbt-review-demo-parity
branch
from
4a28ba0 to
49fc784
Compare
Contributor
Author
|
Addressed current review comments in 49fc784. Summary:
Validation:
|
There was a problem hiding this comment.
Actionable comments posted: 2
🧹 Nitpick comments (1)
packages/opencode/src/altimate/review/orchestrate.ts (1)
965-972: 💤 Low valueConsider memoizing column lineage to avoid duplicate calls.
introducedOutputColumnscomputes head/base column sets fromrunner.columnLineage, butpiiClassifyLane(lines 927-932) performs nearly identical computations. SincerunReviewcalls both for each model context (lines 1091 and 1102), this results in redundantcolumnLineagecalls—up to 4 per file when 2 would suffice.Consider extracting the lineage computation into the
ModelContextinitialization (lines 1026-1046) so both consumers share the result.🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@packages/opencode/src/altimate/review/orchestrate.ts` around lines 965 - 972, introducedOutputColumns repeats runner.columnLineage work that piiClassifyLane also does, causing duplicate lineage calls; move or memoize the lineage computation into the ModelContext so both functions reuse it: add a cached property (e.g., ctx.columnLineageCache or ctx.getColumnLineage) populated during ModelContext initialization (where ModelContext is created) by invoking runner.columnLineage once per SQL/dialect and storing head/base sets, then update introducedOutputColumns and piiClassifyLane to read from that cached result instead of calling runner.columnLineage directly; ensure the cache keys include dialect and engineNewSql/engineOldSql presence so deleted files or missing SQL still return undefined/empty sets consistently.
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In `@docs/internal/2026-06-08-dbt-pr-review-self-improvement-loop.md`:
- Line 196: Replace the hard-coded absolute path string
`/Users/anandgupta/codebase/altimate-core-internal` in the line containing
"latest core main was" with a portable placeholder or relative reference (for
example `<altimate-core-internal-checkout>` or `../altimate-core-internal`) so
the docs no longer contain a local filesystem path; update only the path text
and keep the surrounding sentence intact.
In `@packages/drivers/test/driver-security.test.ts`:
- Around line 146-152: The test currently only checks that two connection
attempts occurred but doesn't assert the access mode on retry; modify the test
around connectAttempts and opts?.access_mode to record each attempt's
access_mode (e.g., push opts?.access_mode into an array) when the mock callback
runs, and after the second attempt assert that the recorded second value equals
"READ_ONLY"; ensure you still simulate the first attempt failing with the lock
error (done(new Error("DUCKDB_LOCKED: file is locked"))) and the second calling
done(null) so the assertion verifies opts?.access_mode for the retry.
---
Nitpick comments:
In `@packages/opencode/src/altimate/review/orchestrate.ts`:
- Around line 965-972: introducedOutputColumns repeats runner.columnLineage work
that piiClassifyLane also does, causing duplicate lineage calls; move or memoize
the lineage computation into the ModelContext so both functions reuse it: add a
cached property (e.g., ctx.columnLineageCache or ctx.getColumnLineage) populated
during ModelContext initialization (where ModelContext is created) by invoking
runner.columnLineage once per SQL/dialect and storing head/base sets, then
update introducedOutputColumns and piiClassifyLane to read from that cached
result instead of calling runner.columnLineage directly; ensure the cache keys
include dialect and engineNewSql/engineOldSql presence so deleted files or
missing SQL still return undefined/empty sets consistently.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: Repository UI
Review profile: CHILL
Plan: Pro
Run ID: ff50f9aa-0721-4127-9b0b-5dc1be6b2e7d
📒 Files selected for processing (12)
docs/docs/usage/dbt-pr-review.mddocs/internal/2026-06-08-dbt-pr-review-self-improvement-loop.mdpackages/drivers/src/duckdb.tspackages/drivers/test/driver-security.test.tspackages/opencode/src/altimate/native/altimate-core.tspackages/opencode/src/altimate/native/types.tspackages/opencode/src/altimate/review/orchestrate.tspackages/opencode/src/altimate/review/runner.tspackages/opencode/src/cli/cmd/review.tspackages/opencode/test/altimate/data-diff-duckdb-e2e.test.tspackages/opencode/test/altimate/review-runner.test.tspackages/opencode/test/altimate/review.test.ts
✅ Files skipped from review due to trivial changes (1)
- docs/docs/usage/dbt-pr-review.md
🚧 Files skipped from review as they are similar to previous changes (7)
- packages/opencode/src/altimate/native/altimate-core.ts
- packages/opencode/test/altimate/data-diff-duckdb-e2e.test.ts
- packages/drivers/src/duckdb.ts
- packages/opencode/src/cli/cmd/review.ts
- packages/opencode/src/altimate/native/types.ts
- packages/opencode/test/altimate/review.test.ts
- packages/opencode/src/altimate/review/runner.ts
There was a problem hiding this comment.
1 issue found across 2 files (changes from recent commits).
Prompt for AI agents (unresolved issues)
Check if these issues are valid — if so, understand the root cause of each and fix them. If appropriate, use sub-agents to investigate and fix each issue separately.
<file name="docs/docs/usage/dbt-pr-review.md">
<violation number="1">
P2: Docs regression: proof-grade requirement for `target/catalog.json` was dropped. This can mislead CI setup and reduce equivalence/PII determinism versus the stated rollout policy.</violation>
</file>
Reply with feedback, questions, or to request a fix.
Re-trigger cubic
There was a problem hiding this comment.
♻️ Duplicate comments (1)
packages/drivers/test/driver-security.test.ts (1)
233-264:⚠️ Potential issue | 🟡 Minor | ⚡ Quick winAssert
READ_ONLYmode explicitly on retry (same pattern as async test).This test mirrors the async test at lines 137-174, which has a pending suggestion to assert
access_mode. Currently the test only proves there were two attempts; it can pass even if the second attempt doesn't actually useREAD_ONLY. Trackopts?.access_modeper attempt and assert the second value is"READ_ONLY".Consider addressing both tests together for consistency.
Suggested test hardening
test("retries read-only when native open synchronously reports a file lock", async () => { let connectAttempts = 0 + const accessModes: Array<string | undefined> = [] mock.module("duckdb", () => ({ default: { Database: class { constructor(_path: string, optsOrCb: any, cb?: (err: Error | null) => void) { const opts = typeof optsOrCb === "function" ? undefined : optsOrCb const done = openCallback(optsOrCb, cb) connectAttempts++ + accessModes.push(opts?.access_mode) done(connectAttempts === 1 && !opts?.access_mode ? new Error("DUCKDB_LOCKED: file is locked") : null) } @@ await connector.connect() expect(connectAttempts).toBe(2) + expect(accessModes[0]).toBeUndefined() + expect(accessModes[1]).toBe("READ_ONLY") expect(await connector.execute("SELECT 1")).toMatchObject({ columns: ["ok"], rows: [[1]], row_count: 1 })🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@packages/drivers/test/driver-security.test.ts` around lines 233 - 264, The test currently only asserts connectAttempts === 2 but does not verify the access_mode used on the retry; modify the test that mocks duckdb's Database constructor (the class in mock.module("duckdb")) to record the opts?.access_mode for each constructor invocation (e.g., push to an array per attempt) and then assert that the second recorded access_mode is "READ_ONLY"; update the same pattern in the async counterpart test as well so both tests explicitly verify that the retry uses READ_ONLY when the first synchronous open reports a file lock, referring to the mocked Database constructor and the connect/imported connect function from ../src/duckdb.
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Duplicate comments:
In `@packages/drivers/test/driver-security.test.ts`:
- Around line 233-264: The test currently only asserts connectAttempts === 2 but
does not verify the access_mode used on the retry; modify the test that mocks
duckdb's Database constructor (the class in mock.module("duckdb")) to record the
opts?.access_mode for each constructor invocation (e.g., push to an array per
attempt) and then assert that the second recorded access_mode is "READ_ONLY";
update the same pattern in the async counterpart test as well so both tests
explicitly verify that the retry uses READ_ONLY when the first synchronous open
reports a file lock, referring to the mocked Database constructor and the
connect/imported connect function from ../src/duckdb.
ℹ️ Review info
⚙️ Run configuration
Configuration used: Repository UI
Review profile: CHILL
Plan: Pro
Run ID: b197646e-f13b-43f6-9850-ba7a1a81a465
📒 Files selected for processing (4)
packages/drivers/test/driver-security.test.tspackages/opencode/src/altimate/review/orchestrate.tspackages/opencode/test/altimate/data-diff-duckdb-e2e.test.tspackages/opencode/test/altimate/review.test.ts
There was a problem hiding this comment.
1 issue found across 4 files (changes from recent commits).
Tip: Review your code locally with the cubic CLI to iterate faster.
Re-trigger cubic
Fixes raised by the multi-model review of #919, prioritizing the PII false-negative that 5 models flagged. - PII recall (MAJOR): the diff-scoped PII dedup keyed off whether the classifier *ran* and off ALL introduced columns, so a high-risk column core's classifier MISSED (e.g. `ssn`) was dropped from the broad `schema.detect_pii` fallback AND had its regex twin suppressed — silently. Now dedup against the columns core actually CLASSIFIED (`classifiedColumns`), and make the regex-twin suppression column-aware (`pii_into_mart` via its matched `result.line`; `select-pii-columns` only when a precise detector actually emitted a finding). A column core never classified now stays eligible for the broad detector and the regex net — including on trivial/lite tiers that don't run the `pii_exposure` lane. Low-confidence name/address suppression is preserved (core *considered* those columns). - Structural escalation (MAJOR): only `join_risk` (SC010 join-key regression) escalates a core `error` to critical/blocking. Other structural rules core marks `error` stay advisory (warning), so the lane never over-blocks on a rule it wasn't designed to gate. - DuckDB (MINOR): arm the 2s fallback timer BEFORE replaying a synchronous open callback so a sync resolve/reject can clear it (no dangling timer delaying process exit); release the half-open handle on open error. - Concurrency (MINOR): run the diff-scoped PII lane as a concurrent task instead of an inline `await`, so its native lineage/classify calls overlap the other per-file engine lanes. - Polish: align `ReviewRunner.equivalence` dialect param to optional (matches the impl); named constants for the PII confidence thresholds; document that PII exposure severity intentionally escalates only for `marts/reporting`; comment the `--no-ai`/`--ai=false` dual check. Adds regression tests: core-missed column surfaces via broad fallback; core-missed high-risk PII still caught by the regex twin when the broad PII lane is off; non-join `error` structural rule stays advisory; sub-error join-key regression stays a warning. Existing 76 review + 26 driver tests still pass. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
There was a problem hiding this comment.
Actionable comments posted: 1
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In `@packages/opencode/src/cli/cmd/review.ts`:
- Around line 61-64: The comment explaining how noAi is derived is misleading
because only the "no-ai" option is registered; yargs won't auto-create an "--ai"
flag, so remove the phrase about boolean auto-negation and clarify that args.ai
=== false is only a programmatic/explicit fallback. Update the comment above the
noAi assignment in review.ts (the line computing noAi from args.noAi and
args.ai) to state that args.noAi comes from the registered --no-ai flag and that
args.ai === false is handled only for callers that explicitly pass ai: false, or
alternatively register an "ai" boolean option if you want to support
--ai/--ai=false from the CLI.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: Repository UI
Review profile: CHILL
Plan: Pro
Run ID: 435a8bfa-ab5d-4e34-a9c1-0a15fa8775fc
📒 Files selected for processing (4)
packages/drivers/src/duckdb.tspackages/opencode/src/altimate/review/orchestrate.tspackages/opencode/src/cli/cmd/review.tspackages/opencode/test/altimate/review.test.ts
Resolve the three unresolved bot-review threads on PR #919: - `orchestrate.ts` (`piiClassifyLane`): when `columnLineage` resolves but derives zero output targets, return `empty` (`completed: false`) instead of `completed: true`. Zero head targets is lineage *uncertainty* (the classifier considered nothing), not a confident "no new PII", so the file must not be marked PII-complete — keeping the deterministic regex fallback alive. Distinct from the `!newCols.length` case, where lineage resolved columns but none are newly introduced (genuinely done). Defensive/intent fix: behavior-neutral for the current detectors (the tokenless suppression path is unreachable because `pii_into_mart` always carries a recoverable column token), but it makes the `completed` contract honest and removes a latent footgun. - `review.ts`: correct the misleading `noAi` comment. Only `--no-ai` is registered, so `--ai=false` is not a supported CLI flag; `args.ai === false` only covers programmatic callers. - `driver-security.test.ts`: assert the DuckDB lock retry actually requests `READ_ONLY` (record `access_mode` per attempt), not just that two attempts occurred. Tests: add two `review.test.ts` regressions — a zero-target-lineage guard and a token-gated suppression guard (core muted a low-confidence column → the coarse regex twin is suppressed). Review suite 82 pass, driver suite 26 pass. Reviewed via multi-model consensus (Claude + 7 models); the lone objection was test completeness, now incorporated. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
`core.checkEquivalence` (altimate-core@0.4.0) is typed `(sqlA, sqlB, schema)` — 3 params, no dialect (unlike `columnLineage`/`compareQueries`). A 4th `dialect` arg was added to this call earlier on the branch (b5060e0), which the napi binding silently ignores at runtime but which fails `tsgo` typecheck (TS2554) and blocks the pre-push hook. Align with the binding contract and the sibling call site in `native/sql/register.ts`. Behavior-neutral. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
|
👋 This PR was automatically closed by our quality checks. Common reasons:
If you believe this was a mistake, please open an issue explaining your intended contribution and a maintainer will help you. |
1 similar comment
|
👋 This PR was automatically closed by our quality checks. Common reasons:
If you believe this was a mistake, please open an issue explaining your intended contribution and a maintainer will help you. |
There was a problem hiding this comment.
1 issue found across 5 files (changes from recent commits).
Prompt for AI agents (unresolved issues)
Check if these issues are valid — if so, understand the root cause of each and fix them. If appropriate, use sub-agents to investigate and fix each issue separately.
<file name="packages/opencode/src/altimate/native/altimate-core.ts">
<violation number="1">
P1: Dialect is no longer forwarded to `checkEquivalence`, so equivalence checks can run with the wrong SQL dialect. This can misclassify refactors for warehouses with dialect-specific semantics.</violation>
</file>
Tip: Review your code locally with the cubic CLI to iterate faster.
Re-trigger cubic
Contributor
Author
|
Pushed
All CI checks pass and the branch is conflict-free. Note: the blocking |
Contributor
❌ Tests — Failures DetectedTypeScript — 15 failure(s)
Next StepPlease address the failing cases above and re-run verification. |
Contributor
Author
|
Re: cubic's P1 on The napi binding signature is
export declare function checkEquivalence(sqlA: string, sqlB: string, schema: Schema): Promise<EquivalenceResult>The 4th The three earlier coderabbit nitpicks are all addressed in HEAD: the |
5 tasks
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
--no-aihandling for deterministic demo runsCore dependency
@altimateai/altimate-corefrom that branch.Acceptance evidence
demo/safe-refactor:APPROVEdemo/join-key-breakage:REQUEST_CHANGESdemo/test-removal:COMMENTdemo/new-pii-exposure:REQUEST_CHANGESdemo/mart-select-star:COMMENTdemo/incremental-without-guard:COMMENT15/15caught bad cases and0/5false positives.demo/join-key-breakagehaddbt buildpass, while reviewer blocked pre-merge using deterministicaltimate_core.structural_diffevidence.Focused validation
bun test --timeout 30000 test/altimate/review-runner.test.tsbun test --timeout 30000 test/altimate/review.test.tsbun run typecheckbun run --conditions=browser script/review-realworld-eval.tsALTIMATE_RUN_WAREHOUSE_E2E=1 bun test --timeout 30000 test/altimate/data-diff-duckdb-e2e.test.tsbun test --timeout 30000 test/altimate/data-diff-duckdb-e2e.test.tsbun turbo typecheckpassedNotes
ALTIMATE_RUN_WAREHOUSE_E2E=1to run it.docs/internal/2026-06-08-dbt-pr-review-self-improvement-loop.md.Summary by cubic
Make the dbt PR reviewer deterministic and demo‑parity reliable with fixed PII recall, precise join‑risk gating, dialect‑aware equivalence, and sturdier DuckDB handling. All six public demo branches reach the expected verdicts, and the benchmark floor is preserved.
New Features
manifest.metadata.adapter_typeintoaltimate_core.equivalence(dialect now optional in the API).join_key_regression) tojoin_risk; escalate to critical only when core reportsseverity: error(other structural rules stay advisory).@altimateai/drivers(ALTIMATE_RUN_WAREHOUSE_E2E=1).Bug Fixes
schema.detect_piifallback only against columns core actually classified; treat zero‑target lineage as uncertainty (keep fallback alive); make regex twin suppression column‑aware (pii_into_martvia its matched column,select-pii-columnsonly when a precise detector emitted a PII finding); keep deterministic fallback when lineage/classifier fails or misses.--no-ai; programmaticai: falsealso disables the advisory lane.Written for commit 89f77a7. Summary will update on new commits.
Summary by CodeRabbit
New Features
Bug Fixes
Tests