ktx/packages/cli/test/context/ingest/context-candidates/candidate-dedup.service.test.ts
Andrey Avtomonov 56985b7e09
test: split cli tests from source tree (#216)
* feat(cli): define full warehouse dialect contract

* test(cli): keep dialect edge tests focused

* fix(cli): stabilize dialect contract foundation

* refactor(connectors): own read-only query preparation

* refactor(connectors): resolve dialects through registry

* refactor(connectors): keep concrete dialect classes internal

* chore(workspace): enforce dialect import boundary

* refactor(cli): resolve relationship dialect at scan boundary

* refactor(cli): use dialect display parsing for entity details

* refactor(cli): use dialect display parsing for warehouse catalog

* refactor(cli): use dialect SQL in relationship workflows

* test(cli): verify solid dialect scan workflow closure

* test: split cli tests from source tree

* refactor(cli): standardize BigQuery scope listing

* feat(sqlite): implement connector scope listing

* test(connectors): cover required table listing

* feat(cli): add warehouse driver registry

* refactor(setup): route scope discovery through driver registry

* refactor(cli): route local query execution through driver registry

* refactor(historic-sql): route dialect support through driver registry

* refactor(cli): test warehouse connections through driver registry

* fix(cli): close driver registry type export gaps

* Improve setup daemon diagnostics

* refactor(setup): centralize rail-prefixed diagnostics + query-history fallback

Extract errorMessage, writePrefixedLines, and flushPrefixedBufferedCommandOutput
into clack.ts so the setup wizard, managed daemons, and embedding/agent steps
share one rail-formatted writer. setup-databases.ts also adds a
"disable query history and retry" option when the schema-context build fails
and query history is the likely culprit, surfaced via a new
failed-query-history-unavailable status.

* fix(cli): carry catalog through the picker so BigQuery/Snowflake/SQL Server scope filters match

The setup picker's KtxTableListEntry was a 2-level { schema, name }, so
qualifiedTableId always wrote db.name into enabled_tables. When BigQuery,
Snowflake, or SQL Server later ran fast ingest, their introspect step filtered
the scope set with scopedTableNames(scope, { catalog: projectId|database, db })
— catalog was non-null on the introspect side but null in the scope refs, so
every entry was rejected, the live-database adapter staged zero table files,
and detect() failed with 'Adapter "live-database" did not recognize fetched
source output'.

Align the picker boundary with the canonical 3-level KtxTableRef:

- Add catalog: string | null to KtxTableListEntry.
- BigQuery/Snowflake/SQL Server listTables populate catalog from the
  resolved projectId / database; Postgres/MySQL/ClickHouse/SQLite set null.
- qualifiedTableId emits catalog.schema.name when catalog is non-null
  (resolveEnabledTables already accepts the 3-part shape) and
  schemasFromEnabledTables now goes through parseDottedTableEntry so it
  recovers the schema correctly from both 2-part and 3-part entries.
- Export parseDottedTableEntry from enabled-tables.ts (@internal) for picker
  reuse.

Update listTables expectations in all seven connector tests and the setup /
picker test fixtures. Add a picker regression test that covers the
catalog-bearing round-trip (save + refine).

* fix(cli): allow debug telemetry under opt-out env
2026-05-26 08:49:05 +02:00

268 lines
11 KiB
TypeScript

import { beforeEach, describe, expect, it, vi } from 'vitest';
import type { ContextCandidateForDedup } from '../../../../src/context/ingest/ports.js';
import { CandidateDedupService } from '../../../../src/context/ingest/context-candidates/candidate-dedup.service.js';
import type { ContextCandidateStorePort } from '../../../../src/context/ingest/context-candidates/store.js';
import type { ContextCandidateEmbeddingPort } from '../../../../src/context/ingest/context-candidates/types.js';
const vector = (...values: number[]): string => JSON.stringify(values);
const candidate = (
overrides: Partial<ContextCandidateForDedup> & { candidateKey: string },
): ContextCandidateForDedup => ({
id: `${overrides.candidateKey}-id`,
candidateKey: overrides.candidateKey,
topic: overrides.topic ?? overrides.candidateKey,
assertion: overrides.assertion ?? `Assertion for ${overrides.candidateKey}`,
promotionScore: overrides.promotionScore ?? 1,
createdAt: overrides.createdAt ?? new Date('2026-04-29T10:00:00.000Z'),
evidenceChunkIds: overrides.evidenceChunkIds ?? [],
evidenceRefs: overrides.evidenceRefs ?? [],
embedding: 'embedding' in overrides ? (overrides.embedding ?? null) : vector(1, 0, 0),
lane: overrides.lane ?? null,
});
function buildHarness(
overrides: {
enabled?: boolean;
threshold?: number;
scoreAggregation?: 'max' | 'mean' | 'sum';
candidates?: ContextCandidateForDedup[];
} = {},
) {
const store = {
listPendingCandidatesForDedup: vi.fn().mockResolvedValue(overrides.candidates ?? []),
updateCandidateEmbedding: vi.fn().mockResolvedValue(undefined),
markCandidatesAsMergedToCluster: vi.fn().mockResolvedValue(undefined),
listBudgetExhaustedCandidatesForCarryForward: vi.fn(),
listCurrentRunEvidenceChunksForCarryForward: vi.fn(),
insertCandidate: vi.fn(),
};
const embeddings = {
maxBatchSize: 100,
computeEmbedding: vi.fn(),
computeEmbeddingsBulk: vi.fn(),
};
const service = new CandidateDedupService({
store: store as unknown as ContextCandidateStorePort,
embeddings: embeddings as unknown as ContextCandidateEmbeddingPort,
settings: {
enabled: overrides.enabled ?? true,
topicSimilarityThreshold: overrides.threshold ?? 0.85,
scoreAggregation: overrides.scoreAggregation ?? 'max',
},
});
return { service, store, embeddings };
}
describe('CandidateDedupService', () => {
beforeEach(() => {
vi.restoreAllMocks();
});
it('returns raw representatives without writes when dedup is disabled', async () => {
const first = candidate({ candidateKey: 'first', embedding: vector(1, 0, 0) });
const duplicate = candidate({ candidateKey: 'duplicate', embedding: vector(0.99, 0.01, 0) });
const { service, store, embeddings } = buildHarness({
enabled: false,
candidates: [first, duplicate],
});
const result = await service.deduplicateRun('run-1');
expect(result).toMatchObject({
enabled: false,
candidatesIn: 2,
clustersOut: 2,
mergedCount: 0,
largestClusterSize: 1,
embeddingFailures: 0,
});
expect(result.representatives.map((item) => item.candidateKey)).toEqual(['first', 'duplicate']);
expect(store.markCandidatesAsMergedToCluster).not.toHaveBeenCalled();
expect(embeddings.computeEmbeddingsBulk).not.toHaveBeenCalled();
});
it('clusters near duplicates and persists representative evidence unions', async () => {
const rep = candidate({
candidateKey: 'icp-primary',
topic: 'ICP',
assertion: 'Finance operators are the ICP.',
promotionScore: 11,
evidenceChunkIds: ['00000000-0000-0000-0000-000000000001'],
evidenceRefs: [{ stableCitationKey: 'icp-a', rawPath: 'pages/a/page.md' }],
embedding: vector(1, 0, 0),
});
const duplicate = candidate({
candidateKey: 'icp-duplicate',
topic: 'Ideal customer profile',
assertion: 'The ICP is finance teams.',
promotionScore: 7,
evidenceChunkIds: ['00000000-0000-0000-0000-000000000002'],
evidenceRefs: [{ stableCitationKey: 'icp-b', rawPath: 'pages/b/page.md' }],
embedding: vector(0.99, 0.02, 0),
});
const unique = candidate({
candidateKey: 'pricing-policy',
promotionScore: 6,
evidenceChunkIds: ['00000000-0000-0000-0000-000000000003'],
evidenceRefs: [{ stableCitationKey: 'price-a', rawPath: 'pages/pricing/page.md' }],
embedding: vector(0, 1, 0),
});
const { service, store } = buildHarness({ candidates: [rep, duplicate, unique] });
const result = await service.deduplicateRun('run-1');
expect(result).toMatchObject({
enabled: true,
candidatesIn: 3,
clustersOut: 2,
mergedCount: 1,
largestClusterSize: 2,
embeddingFailures: 0,
});
expect(result.representatives.map((item) => item.candidateKey)).toEqual(['icp-primary', 'pricing-policy']);
expect(store.markCandidatesAsMergedToCluster).toHaveBeenCalledWith({
representativeId: rep.id,
memberIds: [duplicate.id],
evidenceChunkIds: ['00000000-0000-0000-0000-000000000001', '00000000-0000-0000-0000-000000000002'],
evidenceRefs: [
{ stableCitationKey: 'icp-a', rawPath: 'pages/a/page.md' },
{ stableCitationKey: 'icp-b', rawPath: 'pages/b/page.md' },
],
promotionScore: 11,
});
});
it('uses the configured similarity threshold', async () => {
const base = candidate({ candidateKey: 'base', embedding: vector(1, 0, 0), promotionScore: 5 });
const borderline = candidate({ candidateKey: 'borderline', embedding: vector(0.8, 0.6, 0), promotionScore: 4 });
const strict = buildHarness({ candidates: [base, borderline], threshold: 0.95 });
const strictResult = await strict.service.deduplicateRun('run-1');
expect(strictResult.clustersOut).toBe(2);
expect(strict.store.markCandidatesAsMergedToCluster).not.toHaveBeenCalled();
const loose = buildHarness({ candidates: [base, borderline], threshold: 0.75 });
const looseResult = await loose.service.deduplicateRun('run-1');
expect(looseResult.clustersOut).toBe(1);
expect(loose.store.markCandidatesAsMergedToCluster).toHaveBeenCalledTimes(1);
});
it('fills missing embeddings in batches and persists them before clustering', async () => {
const first = candidate({ candidateKey: 'missing-a', embedding: null });
const second = candidate({ candidateKey: 'missing-b', embedding: null });
const { service, store, embeddings } = buildHarness({ candidates: [first, second] });
embeddings.computeEmbeddingsBulk.mockResolvedValueOnce([
[1, 0, 0],
[0, 1, 0],
]);
const result = await service.deduplicateRun('run-1');
expect(result.embeddingFailures).toBe(0);
expect(embeddings.computeEmbeddingsBulk).toHaveBeenCalledWith([
'missing-a - Assertion for missing-a',
'missing-b - Assertion for missing-b',
]);
expect(store.updateCandidateEmbedding).toHaveBeenCalledWith(first.id, [1, 0, 0]);
expect(store.updateCandidateEmbedding).toHaveBeenCalledWith(second.id, [0, 1, 0]);
});
it('isolates a single embedding failure and keeps that candidate as a singleton', async () => {
const first = candidate({ candidateKey: 'embed-ok', embedding: null });
const second = candidate({ candidateKey: 'embed-fail', embedding: null });
const { service, store, embeddings } = buildHarness({ candidates: [first, second] });
embeddings.computeEmbeddingsBulk.mockRejectedValueOnce(new Error('bulk provider unavailable'));
embeddings.computeEmbedding
.mockResolvedValueOnce([1, 0, 0])
.mockRejectedValueOnce(new Error('single candidate failed'));
const result = await service.deduplicateRun('run-1');
expect(result.embeddingFailures).toBe(1);
expect(result.clustersOut).toBe(2);
expect(result.warnings).toEqual(
expect.arrayContaining([
expect.stringContaining(
'embedding bulk failed: bulk provider unavailable; falling back to per-candidate embedding for 2 candidates',
),
expect.stringContaining('Embedding failed for candidate embed-fail'),
]),
);
expect(store.updateCandidateEmbedding).toHaveBeenCalledTimes(1);
expect(store.updateCandidateEmbedding).toHaveBeenCalledWith(first.id, [1, 0, 0]);
});
it('applies mean and sum score aggregation modes', async () => {
const rep = candidate({ candidateKey: 'score-rep', promotionScore: 9, embedding: vector(1, 0, 0) });
const duplicate = candidate({
candidateKey: 'score-duplicate',
promotionScore: 3,
embedding: vector(0.99, 0.02, 0),
});
const mean = buildHarness({ candidates: [rep, duplicate], scoreAggregation: 'mean' });
await mean.service.deduplicateRun('run-1');
expect(mean.store.markCandidatesAsMergedToCluster).toHaveBeenCalledWith(
expect.objectContaining({ promotionScore: 6 }),
);
const sum = buildHarness({ candidates: [rep, duplicate], scoreAggregation: 'sum' });
await sum.service.deduplicateRun('run-1');
expect(sum.store.markCandidatesAsMergedToCluster).toHaveBeenCalledWith(
expect.objectContaining({ promotionScore: 12 }),
);
});
it('rounds mean score aggregation for the integer promotion score column', async () => {
const rep = candidate({ candidateKey: 'rounded-rep', promotionScore: 10, embedding: vector(1, 0, 0) });
const duplicate = candidate({
candidateKey: 'rounded-duplicate',
promotionScore: 7,
embedding: vector(0.99, 0.02, 0),
});
const { service, store } = buildHarness({ candidates: [rep, duplicate], scoreAggregation: 'mean' });
await service.deduplicateRun('run-1');
expect(store.markCandidatesAsMergedToCluster).toHaveBeenCalledWith(expect.objectContaining({ promotionScore: 9 }));
});
it('is a no-op on a rerun after non-representatives are already merged', async () => {
const rep = candidate({ candidateKey: 'rerun-rep', promotionScore: 9, embedding: vector(1, 0, 0) });
const duplicate = candidate({
candidateKey: 'rerun-duplicate',
promotionScore: 3,
embedding: vector(0.99, 0.02, 0),
});
const { service, store } = buildHarness();
store.listPendingCandidatesForDedup.mockResolvedValueOnce([rep, duplicate]).mockResolvedValueOnce([rep]);
const first = await service.deduplicateRun('run-1');
const second = await service.deduplicateRun('run-1');
expect(first.mergedCount).toBe(1);
expect(second.mergedCount).toBe(0);
expect(second.clustersOut).toBe(1);
expect(store.markCandidatesAsMergedToCluster).toHaveBeenCalledTimes(1);
});
it('returns raw candidates with a warning when cluster persistence throws', async () => {
const rep = candidate({ candidateKey: 'persist-rep', promotionScore: 9, embedding: vector(1, 0, 0) });
const duplicate = candidate({
candidateKey: 'persist-duplicate',
promotionScore: 3,
embedding: vector(0.99, 0.02, 0),
});
const { service, store } = buildHarness({ candidates: [rep, duplicate] });
store.markCandidatesAsMergedToCluster.mockRejectedValueOnce(new Error('database unavailable'));
const result = await service.deduplicateRun('run-1');
expect(result.clustersOut).toBe(2);
expect(result.mergedCount).toBe(0);
expect(result.representatives.map((item) => item.candidateKey)).toEqual(['persist-rep', 'persist-duplicate']);
expect(result.warnings).toEqual([expect.stringContaining('Dedup failed for run run-1')]);
});
});