fix(ingest): recover textual-conflict gate failures; fix query-history adapter (#255)

* fix(ingest): recover textual-conflict gate failures; fix query-history adapter

Two latent gaps in the isolated-diff local-ingest pipeline that can abort an
otherwise-successful ingest:

- Metabase: when a work-unit patch hit both a textual conflict and a post-merge
  dangling sl_ref, the after-textual-resolution branch returned a hard
  semantic_conflict and rolled back the whole job. It now runs the same
  repairGateFailure recovery the clean-apply branch already uses (re-validate,
  then commit the union of resolved + repaired paths), reaching parity.

- Query history: the historic-sql adapter was registered only when ktx.yaml had
  context.queryHistory.enabled=true, so `--query-history` threw "Adapter not
  available for local ingest". Registration now resolves the dialect from driver
  capability, since the explicit --query-history request is itself the opt-in;
  the config-gated helper is unchanged for status/setup/probes.

Adds the previously-missing tests for both paths.

* chore: sync uv.lock to 0.8.0 (regenerated with pinned uv 0.11.11)

* fix(ingest): drop ktx's own scan probes and dedup tables in query history

Query history (historic-sql) mined two kinds of noise back into context:

- ktx's own warehouse scan emits relationship- and column-profiling probes
  (the relationship_profile_values aggregation and the child_values/parent_values
  FK-overlap CTEs) into pg_stat_statements. shouldDropBySql now filters these
  ktx-owned, dialect-stable signatures so ktx introspection is not ingested as
  usage history.

- The same physical table appears both bare (accounts, via search_path) and
  schema-qualified (orbit_raw.accounts), producing duplicate per-table work
  units. canonicalizeTableIdentifiers collapses a bare name into its unique
  qualified form before work-unit keying; ambiguous names are left untouched.

On the orbit demo this removes ~35% of sampled query templates (ktx self-probes)
and ~45 duplicate per-table work units.

* docs(agents): add Design Reasoning Defaults section
This commit is contained in:
Andrey Avtomonov 2026-06-03 13:05:59 +02:00 committed by GitHub
parent 9d3a0b751d
commit f5dea9a089
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
9 changed files with 437 additions and 12 deletions

View file

@ -401,4 +401,72 @@ describe('integrateWorkUnitPatch', () => {
});
await expect(readFile(join(configDir, 'wiki/global/a.md'), 'utf-8')).resolves.toBe('old\n');
});
it('repairs a semantic gate failure after a textual conflict is resolved', async () => {
const { homeDir, configDir, git } = await makeRepo();
await mkdir(join(configDir, 'wiki/global'), { recursive: true });
await writeFile(join(configDir, 'wiki/global/a.md'), 'base\n', 'utf-8');
await git.commitFiles(['wiki/global/a.md'], 'base page', 'System User', 'system@example.com');
const conflictBase = await git.revParseHead();
await writeFile(join(configDir, 'wiki/global/a.md'), 'accepted\n', 'utf-8');
await git.commitFiles(['wiki/global/a.md'], 'accepted edit', 'System User', 'system@example.com');
const childDir = join(homeDir, 'child-conflict-repair');
await git.addWorktree(childDir, 'child-conflict-repair', conflictBase);
const childGit = git.forWorktree(childDir);
await writeFile(join(childDir, 'wiki/global/a.md'), 'proposal\n', 'utf-8');
await childGit.commitFiles(['wiki/global/a.md'], 'proposal edit', 'System User', 'system@example.com');
const patchPath = join(homeDir, 'proposal-repair.patch');
await childGit.writeBinaryNoRenamePatch(conflictBase, 'HEAD', patchPath);
const trace = new FileIngestTraceWriter({
tracePath: join(homeDir, '.ktx/ingest-traces/job-resolver-repair/trace.jsonl'),
jobId: 'job-resolver-repair',
connectionId: 'warehouse',
sourceKey: 'metabase',
level: 'trace',
});
// Gate fails on the resolver's merged tree, then passes after the repair edit.
const validateAppliedTree = vi
.fn()
.mockRejectedValueOnce(
new Error('final artifact gates failed:\narr-definition: unknown sl_refs entity mart_arr_daily.arr_dollars'),
)
.mockResolvedValueOnce(undefined);
const repairGateFailure = vi.fn(async (context: { unitKey: string; touchedPaths: string[] }) => {
expect(context).toMatchObject({ unitKey: 'wu-conflict-repair', touchedPaths: ['wiki/global/a.md'] });
await writeFile(join(configDir, 'wiki/global/a.md'), 'accepted\nproposal repaired\n', 'utf-8');
return { status: 'repaired' as const, attempts: 1, changedPaths: ['wiki/global/a.md'] };
});
const result = await integrateWorkUnitPatch({
unitKey: 'wu-conflict-repair',
patchPath,
integrationGit: git,
trace,
author: { name: 'System User', email: 'system@example.com' },
slDisallowed: false,
allowedTargetConnectionIds: new Set(['warehouse']),
validateAppliedTree,
resolveTextualConflict: vi.fn(async () => {
await writeFile(join(configDir, 'wiki/global/a.md'), 'accepted\nproposal\n', 'utf-8');
return { status: 'repaired' as const, attempts: 1, changedPaths: ['wiki/global/a.md'] };
}),
repairGateFailure,
});
expect(result).toMatchObject({
status: 'accepted',
touchedPaths: ['wiki/global/a.md'],
textualResolution: { status: 'repaired' },
gateRepair: { status: 'repaired', attempts: 1, changedPaths: ['wiki/global/a.md'] },
});
expect(validateAppliedTree).toHaveBeenCalledTimes(2);
expect(repairGateFailure).toHaveBeenCalledOnce();
await expect(readFile(join(configDir, 'wiki/global/a.md'), 'utf-8')).resolves.toBe('accepted\nproposal repaired\n');
await expect(readFile(trace.tracePath, 'utf-8')).resolves.toContain('patch_accepted_after_textual_resolution');
});
});