2026-05-10 23:12:26 +02:00
|
|
|
import Database from 'better-sqlite3';
|
|
|
|
|
import { afterEach, describe, expect, it } from 'vitest';
|
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
|
|
|
import { getDialectForDriver } from '../../../src/context/connections/dialects.js';
|
|
|
|
|
import type { KtxEnrichedColumn, KtxEnrichedSchema, KtxEnrichedTable } from '../../../src/context/scan/enrichment-types.js';
|
|
|
|
|
import { generateKtxRelationshipDiscoveryCandidates } from '../../../src/context/scan/relationship-candidates.js';
|
|
|
|
|
import type { KtxRelationshipProfileArtifact } from '../../../src/context/scan/relationship-profiling.js';
|
|
|
|
|
import { profileKtxRelationshipSchema } from '../../../src/context/scan/relationship-profiling.js';
|
|
|
|
|
import { validateKtxRelationshipDiscoveryCandidates } from '../../../src/context/scan/relationship-validation.js';
|
|
|
|
|
import type { KtxQueryResult, KtxReadOnlyQueryInput, KtxScanContext } from '../../../src/context/scan/types.js';
|
2026-05-10 23:12:26 +02:00
|
|
|
|
|
|
|
|
class InMemorySqliteExecutor {
|
|
|
|
|
readonly db = new Database(':memory:');
|
|
|
|
|
queryCount = 0;
|
|
|
|
|
|
2026-05-10 23:51:24 +02:00
|
|
|
executeReadOnly(input: KtxReadOnlyQueryInput, _ctx: KtxScanContext): Promise<KtxQueryResult> {
|
2026-05-10 23:12:26 +02:00
|
|
|
this.queryCount += 1;
|
|
|
|
|
const rows = this.db.prepare(input.sql).all() as Record<string, unknown>[];
|
|
|
|
|
const headers = Object.keys(rows[0] ?? {});
|
|
|
|
|
return Promise.resolve({
|
|
|
|
|
headers,
|
|
|
|
|
rows: rows.map((row) => headers.map((header) => row[header])),
|
|
|
|
|
totalRows: rows.length,
|
|
|
|
|
rowCount: rows.length,
|
|
|
|
|
});
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
close(): void {
|
|
|
|
|
this.db.close();
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2026-05-10 23:51:24 +02:00
|
|
|
function column(tableId: string, name: string, overrides: Partial<KtxEnrichedColumn> = {}): KtxEnrichedColumn {
|
2026-05-10 23:12:26 +02:00
|
|
|
const tableRef = overrides.tableRef ?? { catalog: null, db: null, name: tableId };
|
|
|
|
|
return {
|
|
|
|
|
id: `${tableId}.${name}`,
|
|
|
|
|
tableId,
|
|
|
|
|
tableRef,
|
|
|
|
|
name,
|
|
|
|
|
nativeType: overrides.nativeType ?? 'INTEGER',
|
|
|
|
|
normalizedType: overrides.normalizedType ?? 'integer',
|
|
|
|
|
dimensionType: overrides.dimensionType ?? 'number',
|
|
|
|
|
nullable: overrides.nullable ?? true,
|
|
|
|
|
primaryKey: overrides.primaryKey ?? false,
|
|
|
|
|
parentColumnId: null,
|
|
|
|
|
descriptions: {},
|
|
|
|
|
embedding: null,
|
|
|
|
|
sampleValues: null,
|
|
|
|
|
cardinality: null,
|
|
|
|
|
...overrides,
|
|
|
|
|
};
|
|
|
|
|
}
|
|
|
|
|
|
2026-05-10 23:51:24 +02:00
|
|
|
function table(name: string, columns: KtxEnrichedColumn[]): KtxEnrichedTable {
|
2026-05-10 23:12:26 +02:00
|
|
|
const ref = { catalog: null, db: null, name };
|
|
|
|
|
return {
|
|
|
|
|
id: name,
|
|
|
|
|
ref,
|
|
|
|
|
enabled: true,
|
|
|
|
|
descriptions: {},
|
|
|
|
|
columns: columns.map((item) => ({ ...item, tableId: name, tableRef: ref })),
|
|
|
|
|
};
|
|
|
|
|
}
|
|
|
|
|
|
2026-05-10 23:51:24 +02:00
|
|
|
function schema(tables?: KtxEnrichedTable[]): KtxEnrichedSchema {
|
2026-05-10 23:12:26 +02:00
|
|
|
return {
|
|
|
|
|
connectionId: 'warehouse',
|
|
|
|
|
tables: tables ?? [
|
|
|
|
|
table('accounts', [
|
|
|
|
|
column('accounts', 'id', { nullable: false }),
|
|
|
|
|
column('accounts', 'name', { nativeType: 'TEXT', normalizedType: 'text', dimensionType: 'string' }),
|
|
|
|
|
]),
|
|
|
|
|
table('users', [column('users', 'id', { nullable: false }), column('users', 'account_id', { nullable: false })]),
|
|
|
|
|
table('invoices', [
|
|
|
|
|
column('invoices', 'id', { nullable: false }),
|
|
|
|
|
column('invoices', 'account_id', { nullable: false }),
|
|
|
|
|
]),
|
|
|
|
|
],
|
|
|
|
|
relationships: [],
|
|
|
|
|
};
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
describe('relationship validation', () => {
|
|
|
|
|
let executor: InMemorySqliteExecutor | null = null;
|
|
|
|
|
|
|
|
|
|
afterEach(() => {
|
|
|
|
|
executor?.close();
|
|
|
|
|
executor = null;
|
|
|
|
|
});
|
|
|
|
|
|
|
|
|
|
it('accepts a relationship-discovery candidate with unique parent values and full source coverage', async () => {
|
|
|
|
|
executor = new InMemorySqliteExecutor();
|
|
|
|
|
executor.db.exec(`
|
|
|
|
|
CREATE TABLE accounts (id INTEGER, name TEXT);
|
|
|
|
|
CREATE TABLE users (id INTEGER, account_id INTEGER);
|
|
|
|
|
CREATE TABLE invoices (id INTEGER, account_id INTEGER);
|
|
|
|
|
INSERT INTO accounts (id, name) VALUES (1, 'Acme'), (2, 'Globex'), (3, 'Initech');
|
|
|
|
|
INSERT INTO users (id, account_id) VALUES (10, 1), (11, 2), (12, 3);
|
|
|
|
|
INSERT INTO invoices (id, account_id) VALUES (20, 1), (21, 2), (22, 999);
|
|
|
|
|
`);
|
|
|
|
|
const testSchema = schema();
|
2026-05-10 23:51:24 +02:00
|
|
|
const profiles = await profileKtxRelationshipSchema({
|
2026-05-10 23:12:26 +02:00
|
|
|
connectionId: 'warehouse',
|
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
|
|
|
dialect: getDialectForDriver('sqlite'),
|
2026-05-10 23:12:26 +02:00
|
|
|
schema: testSchema,
|
|
|
|
|
executor,
|
|
|
|
|
ctx: { runId: 'validate-test' },
|
|
|
|
|
});
|
2026-05-10 23:51:24 +02:00
|
|
|
const candidates = generateKtxRelationshipDiscoveryCandidates(testSchema).filter(
|
2026-05-10 23:12:26 +02:00
|
|
|
(candidate) => candidate.from.table.name === 'users',
|
|
|
|
|
);
|
|
|
|
|
|
2026-05-10 23:51:24 +02:00
|
|
|
const validated = await validateKtxRelationshipDiscoveryCandidates({
|
2026-05-10 23:12:26 +02:00
|
|
|
connectionId: 'warehouse',
|
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
|
|
|
dialect: getDialectForDriver('sqlite'),
|
2026-05-10 23:12:26 +02:00
|
|
|
candidates,
|
|
|
|
|
profiles,
|
|
|
|
|
executor,
|
|
|
|
|
ctx: { runId: 'validate-test' },
|
2026-05-21 02:38:18 +02:00
|
|
|
tableCount: testSchema.tables.length,
|
2026-05-10 23:12:26 +02:00
|
|
|
});
|
|
|
|
|
|
|
|
|
|
expect(validated).toHaveLength(1);
|
|
|
|
|
expect(validated[0]).toMatchObject({
|
|
|
|
|
from: { table: { name: 'users' }, columns: ['account_id'] },
|
|
|
|
|
to: { table: { name: 'accounts' }, columns: ['id'] },
|
|
|
|
|
status: 'accepted',
|
|
|
|
|
score: expect.any(Number),
|
|
|
|
|
validation: {
|
|
|
|
|
targetUniqueness: 1,
|
|
|
|
|
sourceCoverage: 1,
|
|
|
|
|
violationCount: 0,
|
|
|
|
|
violationRatio: 0,
|
|
|
|
|
reasons: expect.arrayContaining(['validation_passed']),
|
|
|
|
|
},
|
|
|
|
|
});
|
|
|
|
|
expect(validated[0]?.score).toBeGreaterThanOrEqual(0.85);
|
|
|
|
|
});
|
|
|
|
|
|
|
|
|
|
it('rejects a candidate with missing parent values and records the deterministic reason', async () => {
|
|
|
|
|
executor = new InMemorySqliteExecutor();
|
|
|
|
|
executor.db.exec(`
|
|
|
|
|
CREATE TABLE accounts (id INTEGER, name TEXT);
|
|
|
|
|
CREATE TABLE users (id INTEGER, account_id INTEGER);
|
|
|
|
|
CREATE TABLE invoices (id INTEGER, account_id INTEGER);
|
|
|
|
|
INSERT INTO accounts (id, name) VALUES (1, 'Acme'), (2, 'Globex');
|
|
|
|
|
INSERT INTO users (id, account_id) VALUES (10, 1), (11, 2);
|
|
|
|
|
INSERT INTO invoices (id, account_id) VALUES (20, 1), (21, 999), (22, 1000);
|
|
|
|
|
`);
|
|
|
|
|
const testSchema = schema();
|
2026-05-10 23:51:24 +02:00
|
|
|
const profiles = await profileKtxRelationshipSchema({
|
2026-05-10 23:12:26 +02:00
|
|
|
connectionId: 'warehouse',
|
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
|
|
|
dialect: getDialectForDriver('sqlite'),
|
2026-05-10 23:12:26 +02:00
|
|
|
schema: testSchema,
|
|
|
|
|
executor,
|
|
|
|
|
ctx: { runId: 'validate-test' },
|
|
|
|
|
});
|
2026-05-10 23:51:24 +02:00
|
|
|
const candidates = generateKtxRelationshipDiscoveryCandidates(testSchema).filter(
|
2026-05-10 23:12:26 +02:00
|
|
|
(candidate) => candidate.from.table.name === 'invoices',
|
|
|
|
|
);
|
|
|
|
|
|
2026-05-10 23:51:24 +02:00
|
|
|
const validated = await validateKtxRelationshipDiscoveryCandidates({
|
2026-05-10 23:12:26 +02:00
|
|
|
connectionId: 'warehouse',
|
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
|
|
|
dialect: getDialectForDriver('sqlite'),
|
2026-05-10 23:12:26 +02:00
|
|
|
candidates,
|
|
|
|
|
profiles,
|
|
|
|
|
executor,
|
|
|
|
|
ctx: { runId: 'validate-test' },
|
2026-05-21 02:38:18 +02:00
|
|
|
tableCount: testSchema.tables.length,
|
2026-05-10 23:12:26 +02:00
|
|
|
settings: {
|
|
|
|
|
minSourceCoverage: 0.9,
|
|
|
|
|
maxViolationRatio: 0.01,
|
|
|
|
|
},
|
|
|
|
|
});
|
|
|
|
|
|
|
|
|
|
expect(validated).toHaveLength(1);
|
|
|
|
|
expect(validated[0]).toMatchObject({
|
|
|
|
|
from: { table: { name: 'invoices' }, columns: ['account_id'] },
|
|
|
|
|
to: { table: { name: 'accounts' }, columns: ['id'] },
|
|
|
|
|
status: 'rejected',
|
|
|
|
|
validation: {
|
|
|
|
|
sourceCoverage: 1 / 3,
|
|
|
|
|
violationCount: 2,
|
|
|
|
|
violationRatio: 2 / 3,
|
|
|
|
|
reasons: expect.arrayContaining(['low_source_coverage', 'excessive_violations']),
|
|
|
|
|
},
|
|
|
|
|
});
|
|
|
|
|
});
|
|
|
|
|
|
|
|
|
|
it('keeps over-budget candidates review-only without executing coverage SQL for them', async () => {
|
|
|
|
|
executor = new InMemorySqliteExecutor();
|
|
|
|
|
executor.db.exec(`
|
|
|
|
|
CREATE TABLE accounts (id INTEGER, name TEXT);
|
|
|
|
|
CREATE TABLE users (id INTEGER, account_id INTEGER);
|
|
|
|
|
CREATE TABLE invoices (id INTEGER, account_id INTEGER);
|
|
|
|
|
INSERT INTO accounts (id, name) VALUES (1, 'Acme'), (2, 'Globex'), (3, 'Initech');
|
|
|
|
|
INSERT INTO users (id, account_id) VALUES (10, 1), (11, 2), (12, 3);
|
|
|
|
|
INSERT INTO invoices (id, account_id) VALUES (20, 1), (21, 2), (22, 3);
|
|
|
|
|
`);
|
|
|
|
|
const testSchema = schema();
|
2026-05-10 23:51:24 +02:00
|
|
|
const profiles = await profileKtxRelationshipSchema({
|
2026-05-10 23:12:26 +02:00
|
|
|
connectionId: 'warehouse',
|
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
|
|
|
dialect: getDialectForDriver('sqlite'),
|
2026-05-10 23:12:26 +02:00
|
|
|
schema: testSchema,
|
|
|
|
|
executor,
|
|
|
|
|
ctx: { runId: 'validate-budget-profile' },
|
|
|
|
|
});
|
|
|
|
|
executor.queryCount = 0;
|
2026-05-10 23:51:24 +02:00
|
|
|
const candidates = generateKtxRelationshipDiscoveryCandidates(testSchema).map((candidate) => ({
|
2026-05-10 23:12:26 +02:00
|
|
|
...candidate,
|
|
|
|
|
confidence: candidate.from.table.name === 'users' ? 0.99 : 0.5,
|
|
|
|
|
}));
|
|
|
|
|
|
2026-05-10 23:51:24 +02:00
|
|
|
const validated = await validateKtxRelationshipDiscoveryCandidates({
|
2026-05-10 23:12:26 +02:00
|
|
|
connectionId: 'warehouse',
|
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
|
|
|
dialect: getDialectForDriver('sqlite'),
|
2026-05-10 23:12:26 +02:00
|
|
|
candidates,
|
|
|
|
|
profiles,
|
|
|
|
|
executor,
|
|
|
|
|
ctx: { runId: 'validate-budget' },
|
|
|
|
|
tableCount: testSchema.tables.length,
|
|
|
|
|
settings: {
|
|
|
|
|
validationBudget: 1,
|
|
|
|
|
},
|
|
|
|
|
});
|
|
|
|
|
|
|
|
|
|
expect(executor.queryCount).toBe(1);
|
|
|
|
|
expect(validated).toHaveLength(2);
|
|
|
|
|
expect(validated.find((candidate) => candidate.from.table.name === 'users')).toMatchObject({
|
|
|
|
|
status: 'accepted',
|
|
|
|
|
validation: { reasons: expect.arrayContaining(['validation_passed']) },
|
|
|
|
|
});
|
|
|
|
|
expect(validated.find((candidate) => candidate.from.table.name === 'invoices')).toMatchObject({
|
|
|
|
|
status: 'review',
|
|
|
|
|
validation: {
|
|
|
|
|
reasons: ['validation_unattempted'],
|
|
|
|
|
},
|
|
|
|
|
});
|
|
|
|
|
});
|
|
|
|
|
|
|
|
|
|
it('treats validation budget zero as review-only validation without coverage SQL', async () => {
|
|
|
|
|
executor = new InMemorySqliteExecutor();
|
|
|
|
|
executor.db.exec(`
|
|
|
|
|
CREATE TABLE accounts (id INTEGER, name TEXT);
|
|
|
|
|
CREATE TABLE users (id INTEGER, account_id INTEGER);
|
|
|
|
|
INSERT INTO accounts (id, name) VALUES (1, 'Acme'), (2, 'Globex');
|
|
|
|
|
INSERT INTO users (id, account_id) VALUES (10, 1), (11, 2);
|
|
|
|
|
`);
|
|
|
|
|
const testSchema = schema([
|
|
|
|
|
table('accounts', [
|
|
|
|
|
column('accounts', 'id', { nullable: false }),
|
|
|
|
|
column('accounts', 'name', { nativeType: 'TEXT', normalizedType: 'text', dimensionType: 'string' }),
|
|
|
|
|
]),
|
|
|
|
|
table('users', [column('users', 'id', { nullable: false }), column('users', 'account_id', { nullable: false })]),
|
|
|
|
|
]);
|
2026-05-10 23:51:24 +02:00
|
|
|
const profiles = await profileKtxRelationshipSchema({
|
2026-05-10 23:12:26 +02:00
|
|
|
connectionId: 'warehouse',
|
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
|
|
|
dialect: getDialectForDriver('sqlite'),
|
2026-05-10 23:12:26 +02:00
|
|
|
schema: testSchema,
|
|
|
|
|
executor,
|
|
|
|
|
ctx: { runId: 'validate-zero-budget-profile' },
|
|
|
|
|
});
|
|
|
|
|
executor.queryCount = 0;
|
2026-05-10 23:51:24 +02:00
|
|
|
const candidates = generateKtxRelationshipDiscoveryCandidates(testSchema);
|
2026-05-10 23:12:26 +02:00
|
|
|
|
2026-05-10 23:51:24 +02:00
|
|
|
const validated = await validateKtxRelationshipDiscoveryCandidates({
|
2026-05-10 23:12:26 +02:00
|
|
|
connectionId: 'warehouse',
|
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
|
|
|
dialect: getDialectForDriver('sqlite'),
|
2026-05-10 23:12:26 +02:00
|
|
|
candidates,
|
|
|
|
|
profiles,
|
|
|
|
|
executor,
|
|
|
|
|
ctx: { runId: 'validate-zero-budget' },
|
|
|
|
|
tableCount: testSchema.tables.length,
|
|
|
|
|
settings: {
|
|
|
|
|
validationBudget: 0,
|
|
|
|
|
},
|
|
|
|
|
});
|
|
|
|
|
|
|
|
|
|
expect(executor.queryCount).toBe(0);
|
|
|
|
|
expect(validated).toHaveLength(1);
|
|
|
|
|
expect(validated[0]).toMatchObject({
|
|
|
|
|
status: 'review',
|
|
|
|
|
score: expect.any(Number),
|
|
|
|
|
validation: {
|
|
|
|
|
checkedValues: 0,
|
|
|
|
|
reasons: ['validation_unattempted'],
|
|
|
|
|
},
|
|
|
|
|
});
|
|
|
|
|
});
|
|
|
|
|
|
|
|
|
|
it('marks rejected LLM proposals with the spec rejection reason', async () => {
|
|
|
|
|
executor = new InMemorySqliteExecutor();
|
|
|
|
|
executor.db.exec(`
|
|
|
|
|
CREATE TABLE customers (id INTEGER);
|
|
|
|
|
CREATE TABLE orders (buyer_ref INTEGER);
|
|
|
|
|
INSERT INTO customers (id) VALUES (1), (2);
|
|
|
|
|
INSERT INTO orders (buyer_ref) VALUES (98), (99);
|
|
|
|
|
`);
|
|
|
|
|
const testSchema = schema([
|
|
|
|
|
table('customers', [column('customers', 'id', { nullable: false })]),
|
|
|
|
|
table('orders', [column('orders', 'buyer_ref')]),
|
|
|
|
|
]);
|
2026-05-10 23:51:24 +02:00
|
|
|
const profiles = await profileKtxRelationshipSchema({
|
2026-05-10 23:12:26 +02:00
|
|
|
connectionId: 'warehouse',
|
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
|
|
|
dialect: getDialectForDriver('sqlite'),
|
2026-05-10 23:12:26 +02:00
|
|
|
schema: testSchema,
|
|
|
|
|
executor,
|
|
|
|
|
ctx: { runId: 'llm-rejected-validation' },
|
|
|
|
|
});
|
2026-05-10 23:51:24 +02:00
|
|
|
const [candidate] = generateKtxRelationshipDiscoveryCandidates(
|
2026-05-10 23:12:26 +02:00
|
|
|
schema([
|
|
|
|
|
table('customers', [column('customers', 'id', { nullable: false })]),
|
|
|
|
|
table('orders', [column('orders', 'customer_id')]),
|
|
|
|
|
]),
|
|
|
|
|
);
|
|
|
|
|
if (!candidate) {
|
|
|
|
|
throw new Error('Expected base candidate');
|
|
|
|
|
}
|
|
|
|
|
const llmCandidate = {
|
|
|
|
|
...candidate,
|
|
|
|
|
id: 'orders:(orders.buyer_ref)->customers:(customers.id)',
|
|
|
|
|
from: { ...candidate.from, columnIds: ['orders.buyer_ref'], columns: ['buyer_ref'] },
|
|
|
|
|
source: 'llm_proposal' as const,
|
|
|
|
|
evidence: {
|
|
|
|
|
...candidate.evidence,
|
|
|
|
|
reasons: ['llm_proposal'],
|
|
|
|
|
llmConfidence: 0.84,
|
|
|
|
|
llmRationale: 'Buyer references should map to customers.',
|
|
|
|
|
},
|
|
|
|
|
};
|
|
|
|
|
|
2026-05-10 23:51:24 +02:00
|
|
|
const [validated] = await validateKtxRelationshipDiscoveryCandidates({
|
2026-05-10 23:12:26 +02:00
|
|
|
connectionId: 'warehouse',
|
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
|
|
|
dialect: getDialectForDriver('sqlite'),
|
2026-05-10 23:12:26 +02:00
|
|
|
candidates: [llmCandidate],
|
|
|
|
|
profiles,
|
|
|
|
|
executor,
|
|
|
|
|
ctx: { runId: 'llm-rejected-validation' },
|
2026-05-21 02:38:18 +02:00
|
|
|
tableCount: testSchema.tables.length,
|
2026-05-10 23:12:26 +02:00
|
|
|
});
|
|
|
|
|
|
|
|
|
|
expect(validated?.status).toBe('rejected');
|
|
|
|
|
expect(validated?.validation.reasons).toEqual(
|
|
|
|
|
expect.arrayContaining(['low_source_coverage', 'llm_proposed_but_validation_failed']),
|
|
|
|
|
);
|
|
|
|
|
});
|
|
|
|
|
|
|
|
|
|
it('limits validation query concurrency', async () => {
|
|
|
|
|
const executor = new InMemorySqliteExecutor();
|
|
|
|
|
executor.db.exec(`
|
|
|
|
|
CREATE TABLE accounts (id INTEGER NOT NULL);
|
|
|
|
|
CREATE TABLE orders (id INTEGER NOT NULL, account_id INTEGER NOT NULL);
|
|
|
|
|
CREATE TABLE invoices (id INTEGER NOT NULL, account_id INTEGER NOT NULL);
|
|
|
|
|
INSERT INTO accounts VALUES (1), (2);
|
|
|
|
|
INSERT INTO orders VALUES (10, 1), (11, 2);
|
|
|
|
|
INSERT INTO invoices VALUES (20, 1), (21, 2);
|
|
|
|
|
`);
|
|
|
|
|
|
|
|
|
|
let active = 0;
|
|
|
|
|
let maxActive = 0;
|
|
|
|
|
const throttled = {
|
2026-05-10 23:51:24 +02:00
|
|
|
executeReadOnly: async (input: KtxReadOnlyQueryInput, ctx: KtxScanContext) => {
|
2026-05-10 23:12:26 +02:00
|
|
|
active += 1;
|
|
|
|
|
maxActive = Math.max(maxActive, active);
|
|
|
|
|
await new Promise((resolve) => setTimeout(resolve, input.sql.includes('WITH child_values') ? 10 : 0));
|
|
|
|
|
const result = await executor.executeReadOnly(input, ctx);
|
|
|
|
|
active -= 1;
|
|
|
|
|
return result;
|
|
|
|
|
},
|
|
|
|
|
};
|
|
|
|
|
|
|
|
|
|
const testSchema = schema([
|
|
|
|
|
table('accounts', [column('accounts', 'id', { nullable: false })]),
|
|
|
|
|
table('orders', [column('orders', 'id', { nullable: false }), column('orders', 'account_id')]),
|
|
|
|
|
table('invoices', [column('invoices', 'id', { nullable: false }), column('invoices', 'account_id')]),
|
|
|
|
|
]);
|
2026-05-10 23:51:24 +02:00
|
|
|
const profiles = await profileKtxRelationshipSchema({
|
2026-05-10 23:12:26 +02:00
|
|
|
connectionId: 'warehouse',
|
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
|
|
|
dialect: getDialectForDriver('sqlite'),
|
2026-05-10 23:12:26 +02:00
|
|
|
schema: testSchema,
|
|
|
|
|
executor,
|
|
|
|
|
ctx: { runId: 'validation-concurrency-profile' },
|
|
|
|
|
});
|
2026-05-10 23:51:24 +02:00
|
|
|
const candidates = generateKtxRelationshipDiscoveryCandidates(testSchema);
|
2026-05-10 23:12:26 +02:00
|
|
|
|
2026-05-10 23:51:24 +02:00
|
|
|
await validateKtxRelationshipDiscoveryCandidates({
|
2026-05-10 23:12:26 +02:00
|
|
|
connectionId: 'warehouse',
|
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
|
|
|
dialect: getDialectForDriver('sqlite'),
|
2026-05-10 23:12:26 +02:00
|
|
|
candidates,
|
|
|
|
|
profiles,
|
|
|
|
|
executor: throttled,
|
|
|
|
|
ctx: { runId: 'validation-concurrency' },
|
2026-05-21 02:38:18 +02:00
|
|
|
tableCount: testSchema.tables.length,
|
2026-05-10 23:12:26 +02:00
|
|
|
settings: { concurrency: 1 },
|
|
|
|
|
});
|
|
|
|
|
|
|
|
|
|
expect(maxActive).toBe(1);
|
|
|
|
|
executor.close();
|
|
|
|
|
});
|
|
|
|
|
|
|
|
|
|
it('pins column_suffix_match validation scoring for plan-code suffix candidates', async () => {
|
|
|
|
|
const candidate = {
|
|
|
|
|
id: 'mart:(current_plan_code)->plans:(plan_code)',
|
|
|
|
|
from: {
|
|
|
|
|
tableId: 'mart-account-segments-id',
|
|
|
|
|
columnIds: ['current-plan-code-col'],
|
|
|
|
|
table: { catalog: null, db: null, name: 'mart_account_segments' },
|
|
|
|
|
columns: ['current_plan_code'],
|
|
|
|
|
},
|
|
|
|
|
to: {
|
|
|
|
|
tableId: 'plans-id',
|
|
|
|
|
columnIds: ['plan-code-col'],
|
|
|
|
|
table: { catalog: null, db: null, name: 'stg_plans' },
|
|
|
|
|
columns: ['plan_code'],
|
|
|
|
|
},
|
|
|
|
|
relationshipType: 'many_to_one' as const,
|
|
|
|
|
confidence: 0.902,
|
|
|
|
|
source: 'column_suffix_match' as const,
|
|
|
|
|
status: 'review' as const,
|
|
|
|
|
evidence: {
|
|
|
|
|
sourceColumnBase: 'current_plan',
|
|
|
|
|
targetTableBase: 'plan',
|
|
|
|
|
targetColumnBase: 'plan_code',
|
|
|
|
|
targetKeyScore: 0.86,
|
|
|
|
|
nameScore: 0.78,
|
|
|
|
|
reasons: ['column_suffix_match', 'profile_unique_target'],
|
|
|
|
|
},
|
|
|
|
|
};
|
|
|
|
|
const profiles = {
|
|
|
|
|
connectionId: 'warehouse',
|
|
|
|
|
driver: 'sqlite',
|
|
|
|
|
sqlAvailable: true,
|
|
|
|
|
queryCount: 0,
|
|
|
|
|
tables: [],
|
|
|
|
|
warnings: [],
|
|
|
|
|
columns: {
|
|
|
|
|
'mart_account_segments.current_plan_code': {
|
|
|
|
|
table: { catalog: null, db: null, name: 'mart_account_segments' },
|
|
|
|
|
column: 'current_plan_code',
|
|
|
|
|
nativeType: 'TEXT',
|
|
|
|
|
normalizedType: 'text',
|
|
|
|
|
rowCount: 4,
|
|
|
|
|
nullCount: 0,
|
|
|
|
|
distinctCount: 4,
|
|
|
|
|
uniquenessRatio: 1,
|
|
|
|
|
nullRate: 0,
|
|
|
|
|
sampleValues: ['basic', 'enterprise', 'free', 'pro'],
|
|
|
|
|
minTextLength: 4,
|
|
|
|
|
maxTextLength: 10,
|
|
|
|
|
},
|
|
|
|
|
'stg_plans.plan_code': {
|
|
|
|
|
table: { catalog: null, db: null, name: 'stg_plans' },
|
|
|
|
|
column: 'plan_code',
|
|
|
|
|
nativeType: 'TEXT',
|
|
|
|
|
normalizedType: 'text',
|
|
|
|
|
rowCount: 4,
|
|
|
|
|
nullCount: 0,
|
|
|
|
|
distinctCount: 4,
|
|
|
|
|
uniquenessRatio: 1,
|
|
|
|
|
nullRate: 0,
|
|
|
|
|
sampleValues: ['basic', 'enterprise', 'free', 'pro'],
|
|
|
|
|
minTextLength: 4,
|
|
|
|
|
maxTextLength: 10,
|
|
|
|
|
},
|
|
|
|
|
},
|
2026-05-10 23:51:24 +02:00
|
|
|
} satisfies KtxRelationshipProfileArtifact;
|
2026-05-10 23:12:26 +02:00
|
|
|
const executor = {
|
|
|
|
|
async executeReadOnly() {
|
|
|
|
|
return {
|
|
|
|
|
headers: ['child_distinct', 'parent_distinct', 'overlap', 'violation_count'],
|
|
|
|
|
rows: [[4, 4, 4, 0]],
|
|
|
|
|
rowCount: 1,
|
|
|
|
|
totalRows: 1,
|
|
|
|
|
};
|
|
|
|
|
},
|
|
|
|
|
};
|
|
|
|
|
|
2026-05-10 23:51:24 +02:00
|
|
|
const [validated] = await validateKtxRelationshipDiscoveryCandidates({
|
2026-05-10 23:12:26 +02:00
|
|
|
connectionId: 'warehouse',
|
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
|
|
|
dialect: getDialectForDriver('sqlite'),
|
2026-05-10 23:12:26 +02:00
|
|
|
candidates: [candidate],
|
|
|
|
|
profiles,
|
|
|
|
|
executor,
|
|
|
|
|
ctx: { runId: 'rule-b-validation-score' },
|
2026-05-21 02:38:18 +02:00
|
|
|
tableCount: 2,
|
2026-05-10 23:12:26 +02:00
|
|
|
});
|
|
|
|
|
|
|
|
|
|
expect(validated).toMatchObject({
|
|
|
|
|
status: 'accepted',
|
|
|
|
|
score: 0.98,
|
|
|
|
|
validation: {
|
|
|
|
|
targetUniqueness: 1,
|
|
|
|
|
sourceCoverage: 1,
|
|
|
|
|
violationRatio: 0,
|
|
|
|
|
reasons: ['validation_passed'],
|
|
|
|
|
},
|
|
|
|
|
});
|
|
|
|
|
});
|
|
|
|
|
});
|