mirror of
https://github.com/garrytan/gbrain.git
synced 2026-08-16 09:52:22 +00:00
Compare commits
2
Commits
| Author | SHA1 | Date | |
|---|---|---|---|
|
|
3fa01a4538 | ||
|
|
ded4aeaeae |
@@ -186,6 +186,18 @@ export function isSourceStale(src: SourceRow, now = Date.now(), floorMin = FULL_
|
||||
return ageMin >= floorMin;
|
||||
}
|
||||
|
||||
/**
|
||||
* #2060: count sources past the per-source cycle freshness floor. Consumed
|
||||
* by autopilot's dispatch decision — a stale source forces the fanout path
|
||||
* even when the doctor plan is small (score 70–94, plan ≤ 3, est < 300s),
|
||||
* so targeted mode can't leave cycle_freshness stale indefinitely.
|
||||
* dispatchPerSource's own throttles (skipped_fresh / fanoutMax / failure
|
||||
* cooldown) bound the resulting work.
|
||||
*/
|
||||
export function countStaleSources(sources: SourceRow[], now = Date.now(), floorMin = FULL_CYCLE_FLOOR_MIN): number {
|
||||
return sources.filter((s) => isSourceStale(s, now, floorMin)).length;
|
||||
}
|
||||
|
||||
/**
|
||||
* Most recent SUCCESSFUL cycle for a source. Prefers `last_source_cycle_at`
|
||||
* (per-source phases, written by the split cycle) and falls back to the legacy
|
||||
|
||||
@@ -901,13 +901,27 @@ export async function runAutopilot(engine: BrainEngine, args: string[]) {
|
||||
const FULL_CYCLE_FLOOR_MIN = 60;
|
||||
const minutesSinceLastFull = (Date.now() - lastFullCycleAt) / 60000;
|
||||
|
||||
// #2060: stale per-source cycle freshness is a dispatch input. Without
|
||||
// it, a brain sitting at score 70–94 with a small targeted plan (≤3
|
||||
// steps, <300s) stays in targeted mode indefinitely and no per-source
|
||||
// cycle is ever dispatched — cycle_freshness never advances. A stale
|
||||
// source forces the fanout path; dispatchPerSource's throttles
|
||||
// (skipped_fresh / fanoutMax / failure cooldown) bound the work.
|
||||
// Fail-open to 0: a read failure must not block dispatch.
|
||||
let staleCycleSources = 0;
|
||||
try {
|
||||
const { countStaleSources } = await import('./autopilot-fanout.ts');
|
||||
staleCycleSources = countStaleSources(await engine.listAllSources({ localPathOnly: true }));
|
||||
} catch { /* fail-open: freshness is a dispatch hint, not a gate */ }
|
||||
|
||||
const shouldFullCycle =
|
||||
(score >= 95 && plan.length === 0 && minutesSinceLastFull >= FULL_CYCLE_FLOOR_MIN) ||
|
||||
plan.length > 3 ||
|
||||
estTotal >= 300 ||
|
||||
score < 70;
|
||||
score < 70 ||
|
||||
staleCycleSources > 0;
|
||||
|
||||
const shouldSleep = score >= 95 && plan.length === 0 && minutesSinceLastFull < FULL_CYCLE_FLOOR_MIN;
|
||||
const shouldSleep = score >= 95 && plan.length === 0 && minutesSinceLastFull < FULL_CYCLE_FLOOR_MIN && staleCycleSources === 0;
|
||||
|
||||
if (shouldSleep) {
|
||||
if (jsonMode) {
|
||||
|
||||
+1
-14
@@ -1,5 +1,5 @@
|
||||
import type { BrainEngine } from '../core/engine.ts';
|
||||
import { embedBatch, currentEmbeddingSignature, resolveEmbeddingModelLabel } from '../core/embedding.ts';
|
||||
import { embedBatch, currentEmbeddingSignature } from '../core/embedding.ts';
|
||||
import type { ChunkInput } from '../core/types.ts';
|
||||
import { chunkText } from '../core/chunkers/recursive.ts';
|
||||
import { createProgress, type ProgressReporter } from '../core/progress.ts';
|
||||
@@ -581,16 +581,11 @@ async function embedPage(
|
||||
for (let j = 0; j < toEmbed.length; j++) {
|
||||
embeddingMap.set(toEmbed[j].chunk_index, embeddings[j]);
|
||||
}
|
||||
// #1717: label each (re)embedded chunk with the model that actually
|
||||
// produced its vector. Preserved chunks (not re-embedded this pass) keep
|
||||
// their existing model so a mixed-model page isn't relabeled wholesale.
|
||||
const embedModelLabel = resolveEmbeddingModelLabel();
|
||||
const updated: ChunkInput[] = chunks.map(c => ({
|
||||
chunk_index: c.chunk_index,
|
||||
chunk_text: c.chunk_text,
|
||||
chunk_source: c.chunk_source,
|
||||
embedding: embeddingMap.get(c.chunk_index),
|
||||
model: embeddingMap.has(c.chunk_index) && embedModelLabel ? embedModelLabel : c.model,
|
||||
token_count: c.token_count || Math.ceil(c.chunk_text.length / 4),
|
||||
}));
|
||||
|
||||
@@ -722,16 +717,12 @@ async function embedAll(
|
||||
for (let j = 0; j < toEmbed.length; j++) {
|
||||
embeddingMap.set(toEmbed[j].chunk_index, embeddings[j]);
|
||||
}
|
||||
// #1717: stamp the resolved embedding model on (re)embedded chunks;
|
||||
// preserve the existing model on chunks left untouched.
|
||||
const embedModelLabel = resolveEmbeddingModelLabel();
|
||||
// Preserve ALL chunks, only update embeddings for stale ones
|
||||
const updated: ChunkInput[] = chunks.map(c => ({
|
||||
chunk_index: c.chunk_index,
|
||||
chunk_text: c.chunk_text,
|
||||
chunk_source: c.chunk_source,
|
||||
embedding: embeddingMap.get(c.chunk_index) ?? undefined,
|
||||
model: embeddingMap.has(c.chunk_index) && embedModelLabel ? embedModelLabel : c.model,
|
||||
token_count: c.token_count || Math.ceil(c.chunk_text.length / 4),
|
||||
}));
|
||||
await observed(pacer, () => engine.upsertChunks(page.slug, updated, pageOpts));
|
||||
@@ -1021,15 +1012,11 @@ async function embedAllStale(
|
||||
for (let j = 0; j < stale.length; j++) {
|
||||
staleIdxToEmbedding.set(stale[j].chunk_index, embeddings[j]);
|
||||
}
|
||||
// #1717: label the re-embedded (stale) chunks with the resolved
|
||||
// model; preserve the existing model on the non-stale chunks.
|
||||
const embedModelLabel = resolveEmbeddingModelLabel();
|
||||
const merged: ChunkInput[] = existing.map(c => ({
|
||||
chunk_index: c.chunk_index,
|
||||
chunk_text: c.chunk_text,
|
||||
chunk_source: c.chunk_source,
|
||||
embedding: staleIdxToEmbedding.get(c.chunk_index) ?? undefined,
|
||||
model: staleIdxToEmbedding.has(c.chunk_index) && embedModelLabel ? embedModelLabel : c.model,
|
||||
token_count: c.token_count || Math.ceil(c.chunk_text.length / 4),
|
||||
}));
|
||||
await observed(pacer, () => engine.upsertChunks(slug, merged, { sourceId: keySourceId }));
|
||||
|
||||
+19
-10
@@ -854,7 +854,10 @@ interface SyncPhaseResult extends PhaseResult {
|
||||
/**
|
||||
* Resolve the source id for a brain directory by looking up the sources
|
||||
* table. Returns undefined when no registered source matches (falls back
|
||||
* to pre-v0.18 global config.sync.* keys).
|
||||
* to pre-v0.18 global config.sync.* keys) OR when MORE than one source
|
||||
* claims the path — an ambiguous match must not scope phases or stamp
|
||||
* last_full_cycle_at for an arbitrarily-picked source (the "freshness
|
||||
* stamp that lies" this resolution exists to prevent).
|
||||
*/
|
||||
async function resolveSourceForDir(
|
||||
engine: BrainEngine,
|
||||
@@ -865,10 +868,10 @@ async function resolveSourceForDir(
|
||||
if (brainDir === null) return undefined;
|
||||
try {
|
||||
const rows = await engine.executeRaw<{ id: string }>(
|
||||
`SELECT id FROM sources WHERE local_path = $1 LIMIT 1`,
|
||||
`SELECT id FROM sources WHERE local_path = $1 LIMIT 2`,
|
||||
[brainDir],
|
||||
);
|
||||
return rows[0]?.id;
|
||||
return rows.length === 1 ? rows[0]!.id : undefined;
|
||||
} catch {
|
||||
// sources table might not exist on very old brains — fall through.
|
||||
return undefined;
|
||||
@@ -2365,17 +2368,23 @@ export async function runCycle(
|
||||
}
|
||||
|
||||
// v0.38 (codex r1 P0-5): persist per-source cycle completion timestamp
|
||||
// when the cycle ran successfully against an explicit source. Read by
|
||||
// autopilot's per-source freshness gate next tick. Skipped when:
|
||||
// - opts.sourceId is unset (legacy callers — autopilot still here)
|
||||
// - engine is null (no-DB path)
|
||||
// when the cycle ran successfully against a resolvable source. Read by
|
||||
// autopilot's per-source freshness gate next tick.
|
||||
//
|
||||
// #1993: keyed off `cycleSourceId` (opts.sourceId ?? the source resolved
|
||||
// from brainDir) — the SAME id the cycle locked + scoped its phases to —
|
||||
// NOT raw opts.sourceId. The autopilot's inline cycle sets brainDir but
|
||||
// passes no explicit sourceId, so keying off opts.sourceId alone never
|
||||
// advanced last_full_cycle_at and cycle_freshness stayed stale even while
|
||||
// the autopilot cycled every interval. Skipped when:
|
||||
// - no source resolves (engine null, or no checkout AND no opts.sourceId)
|
||||
// - status is 'failed' or 'skipped' (don't mark a non-run as fresh)
|
||||
// - dryRun (writes are out of scope)
|
||||
//
|
||||
// Best-effort: a write failure does NOT change the CycleReport status.
|
||||
// The cost of writing the wrong timestamp post-failure is higher than
|
||||
// the cost of missing a successful write (next cycle will redo work).
|
||||
if (opts.sourceId && engine && !dryRun && !aborted && (status === 'ok' || status === 'clean' || status === 'partial')) {
|
||||
if (cycleSourceId && engine && !dryRun && !aborted && (status === 'ok' || status === 'clean' || status === 'partial')) {
|
||||
try {
|
||||
const nowIso = new Date().toISOString();
|
||||
// #2194 fix #3 (the cycle split): `last_source_cycle_at` is the NEW gate
|
||||
@@ -2385,13 +2394,13 @@ export async function runCycle(
|
||||
// phases (those gate on autopilot.last_global_at), so writing it on a
|
||||
// source-only cycle does not re-introduce the freshness poisoning codex
|
||||
// flagged in the rejected skip-based design.
|
||||
await engine.updateSourceConfig(opts.sourceId, {
|
||||
await engine.updateSourceConfig(cycleSourceId, {
|
||||
last_source_cycle_at: nowIso,
|
||||
last_full_cycle_at: nowIso,
|
||||
});
|
||||
} catch (e) {
|
||||
// Best-effort; cycle already succeeded by the time we get here.
|
||||
console.warn(`[cycle] failed to write last_source_cycle_at for source ${opts.sourceId}: ${e instanceof Error ? e.message : String(e)}`);
|
||||
console.warn(`[cycle] failed to write last_source_cycle_at for source ${cycleSourceId}: ${e instanceof Error ? e.message : String(e)}`);
|
||||
}
|
||||
}
|
||||
|
||||
|
||||
@@ -20,7 +20,6 @@
|
||||
import type { BrainEngine } from './engine.ts';
|
||||
import type { ChunkInput } from './types.ts';
|
||||
import { embedBatchWithBackoff } from '../commands/embed.ts';
|
||||
import { resolveEmbeddingModelLabel } from './embedding.ts';
|
||||
import { type DbPacer, createNoopPacer, observed } from './db-pacer.ts';
|
||||
import { AbortError } from './abort-check.ts';
|
||||
|
||||
@@ -201,17 +200,11 @@ export async function embedStaleForSource(
|
||||
for (let j = 0; j < stale.length; j++) {
|
||||
staleIdxToEmbedding.set(stale[j].chunk_index, embeddings[j]);
|
||||
}
|
||||
// #1717: label re-embedded chunks with the model that produced the
|
||||
// vector; preserved chunks keep their existing model. Without this,
|
||||
// upsertChunks falls back to DEFAULT_EMBEDDING_MODEL for every chunk
|
||||
// (the same mislabel the embed.ts paths fixed).
|
||||
const embedModelLabel = resolveEmbeddingModelLabel();
|
||||
const merged: ChunkInput[] = existing.map((c) => ({
|
||||
chunk_index: c.chunk_index,
|
||||
chunk_text: c.chunk_text,
|
||||
chunk_source: c.chunk_source,
|
||||
embedding: staleIdxToEmbedding.get(c.chunk_index) ?? undefined,
|
||||
model: staleIdxToEmbedding.has(c.chunk_index) && embedModelLabel ? embedModelLabel : c.model,
|
||||
token_count: c.token_count || Math.ceil(c.chunk_text.length / 4),
|
||||
// Carry through per-chunk metadata. upsertChunks writes these as
|
||||
// EXCLUDED.<col> (not COALESCE), so omitting them here resets image
|
||||
|
||||
@@ -113,21 +113,6 @@ export async function embedBatch(
|
||||
return results;
|
||||
}
|
||||
|
||||
/**
|
||||
* Resolve the embedding model label (`provider:model`) to stamp onto
|
||||
* `content_chunks.model`, so each chunk records the model that actually
|
||||
* produced its vector instead of the engine's hardcoded default (#1717).
|
||||
* Returns undefined if the gateway is unconfigured; callers then fall back
|
||||
* to the chunk's existing model rather than mislabeling it.
|
||||
*/
|
||||
export function resolveEmbeddingModelLabel(): string | undefined {
|
||||
try {
|
||||
return gatewayGetModel();
|
||||
} catch {
|
||||
return undefined;
|
||||
}
|
||||
}
|
||||
|
||||
/** Currently-configured embedding model (short form without provider prefix). */
|
||||
export function getEmbeddingModelName(): string {
|
||||
return gatewayGetModel().split(':').slice(1).join(':') || 'text-embedding-3-large';
|
||||
|
||||
+1
-11
@@ -8,7 +8,7 @@ import { chunkText } from './chunkers/recursive.ts';
|
||||
import { chunkCodeText, chunkCodeTextFull, detectCodeLanguage, CHUNKER_VERSION } from './chunkers/code.ts';
|
||||
import { findChunkForOffset } from './chunkers/edge-extractor.ts';
|
||||
import { extractCodeRefs, imageOfCandidates } from './link-extraction.ts';
|
||||
import { embedBatch, embedMultimodal, currentEmbeddingSignature, resolveEmbeddingModelLabel } from './embedding.ts';
|
||||
import { embedBatch, embedMultimodal, currentEmbeddingSignature } from './embedding.ts';
|
||||
import { slugifyPath, slugifyCodePath, isCodeFilePath } from './sync.ts';
|
||||
import type { ChunkInput, PageInput, PageType } from './types.ts';
|
||||
import { computeEffectiveDate } from './effective-date.ts';
|
||||
@@ -716,12 +716,8 @@ export async function importFromContent(
|
||||
? chunks.map((c) => wrapChunkForEmbedding(c.chunk_text, prefix, c.chunk_source))
|
||||
: chunks.map((c) => c.chunk_text);
|
||||
const embeddings = await embedBatch(wrappedTexts);
|
||||
// #1717: label each chunk with the model that actually produced its
|
||||
// vector, not the engine's hardcoded default.
|
||||
const embedModelLabel = resolveEmbeddingModelLabel();
|
||||
for (let i = 0; i < chunks.length; i++) {
|
||||
chunks[i].embedding = embeddings[i];
|
||||
if (embedModelLabel) chunks[i].model = embedModelLabel;
|
||||
// token_count tracks the wrapped string length so cost reporting
|
||||
// reflects what we actually sent to the embedder.
|
||||
chunks[i].token_count = Math.ceil(wrappedTexts[i].length / 4);
|
||||
@@ -1145,10 +1141,7 @@ export async function importCodeFile(
|
||||
const matched = existingByKey.get(key);
|
||||
if (matched && matched.embedding) {
|
||||
// Reuse the existing embedding verbatim. No API call, no cost.
|
||||
// #1717: carry the existing model label along with the reused vector
|
||||
// so the upsert doesn't relabel it with the engine default.
|
||||
chunks[i]!.embedding = matched.embedding as Float32Array;
|
||||
chunks[i]!.model = matched.model ?? undefined;
|
||||
chunks[i]!.token_count = matched.token_count ?? undefined;
|
||||
} else {
|
||||
needsEmbedIndexes.push(i);
|
||||
@@ -1160,12 +1153,9 @@ export async function importCodeFile(
|
||||
try {
|
||||
const textsToEmbed = needsEmbedIndexes.map((i) => chunks[i]!.chunk_text);
|
||||
const embeddings = await embedBatch(textsToEmbed);
|
||||
// #1717: stamp the model that produced these vectors.
|
||||
const embedModelLabel = resolveEmbeddingModelLabel();
|
||||
for (let j = 0; j < needsEmbedIndexes.length; j++) {
|
||||
const i = needsEmbedIndexes[j]!;
|
||||
chunks[i]!.embedding = embeddings[j]!;
|
||||
if (embedModelLabel) chunks[i]!.model = embedModelLabel;
|
||||
chunks[i]!.token_count = Math.ceil(chunks[i]!.chunk_text.length / 4);
|
||||
}
|
||||
} catch (e: unknown) {
|
||||
|
||||
@@ -54,6 +54,20 @@ describe('autopilot.ts ↔ dispatchPerSource wiring', () => {
|
||||
expect(AUTOPILOT_SRC).toMatch(/lastFullCycleAt\s*=\s*Date\.now\(\)/);
|
||||
});
|
||||
|
||||
test('stale per-source cycle freshness is a shouldFullCycle input (#2060)', () => {
|
||||
// Targeted mode (score 70–94, plan ≤3, est <300s) must not be able to
|
||||
// starve per-source cycle dispatch: a stale source (per countStaleSources
|
||||
// over listAllSources) forces the fanout path, and the sleep gate must
|
||||
// not fire while stale sources exist. Without these terms, cycle
|
||||
// freshness never advances for a brain that always lands in targeted mode.
|
||||
expect(AUTOPILOT_SRC).toMatch(/countStaleSources/);
|
||||
const fullCycleDeclIdx = AUTOPILOT_SRC.indexOf('const shouldFullCycle');
|
||||
expect(fullCycleDeclIdx).toBeGreaterThan(-1);
|
||||
const decl = AUTOPILOT_SRC.slice(fullCycleDeclIdx, fullCycleDeclIdx + 700);
|
||||
expect(decl).toMatch(/staleCycleSources\s*>\s*0/);
|
||||
expect(decl).toMatch(/const shouldSleep[^;]*staleCycleSources\s*===\s*0/);
|
||||
});
|
||||
|
||||
test('does NOT regress to the single-job dispatch on the full-cycle path', () => {
|
||||
// Pre-PR: the shouldFullCycle branch did:
|
||||
// const job = await queue.add('autopilot-cycle', { repoPath }, {
|
||||
|
||||
@@ -14,6 +14,7 @@ import { describe, test, expect } from 'bun:test';
|
||||
import {
|
||||
readLastFullCycleAt,
|
||||
isSourceStale,
|
||||
countStaleSources,
|
||||
selectSourcesForDispatch,
|
||||
resolveFanoutMax,
|
||||
dispatchPerSource,
|
||||
@@ -74,6 +75,23 @@ describe('isSourceStale', () => {
|
||||
});
|
||||
});
|
||||
|
||||
describe('countStaleSources (#2060 dispatch-decision input)', () => {
|
||||
const NOW = Date.parse('2026-05-22T12:00:00.000Z');
|
||||
test('counts never-cycled + past-floor sources, ignores fresh', () => {
|
||||
const sources = [
|
||||
src('never-cycled'), // stale (null)
|
||||
src('old', new Date(NOW - 2 * 60 * 60_000).toISOString()), // stale (2h)
|
||||
src('fresh', new Date(NOW - 30 * 60_000).toISOString()), // fresh (30min)
|
||||
];
|
||||
expect(countStaleSources(sources, NOW)).toBe(2);
|
||||
});
|
||||
test('returns 0 for all-fresh and for empty list', () => {
|
||||
const fresh = src('a', new Date(NOW - 10 * 60_000).toISOString());
|
||||
expect(countStaleSources([fresh], NOW)).toBe(0);
|
||||
expect(countStaleSources([], NOW)).toBe(0);
|
||||
});
|
||||
});
|
||||
|
||||
describe('selectSourcesForDispatch', () => {
|
||||
const NOW = Date.parse('2026-05-22T12:00:00.000Z');
|
||||
const fresh = (id: string, agoMin: number) =>
|
||||
|
||||
@@ -3,8 +3,10 @@
|
||||
* cycles. Closes codex round-1 P0-5 (write site for last_full_cycle_at
|
||||
* was unspecified pre-PR).
|
||||
*
|
||||
* Conditions for write:
|
||||
* - opts.sourceId is set (legacy callers without sourceId skip the write)
|
||||
* Conditions for write (keyed off `cycleSourceId` = opts.sourceId ?? the
|
||||
* source resolved from brainDir, so the autopilot's inline cycle — brainDir
|
||||
* set, no explicit sourceId — also advances the timestamp, #1993):
|
||||
* - a source resolves (explicit sourceId, or brainDir matches a source)
|
||||
* - engine is non-null (no-DB path skips)
|
||||
* - status is 'ok' | 'clean' | 'partial' (failed/skipped don't mark fresh)
|
||||
* - dryRun is false
|
||||
@@ -90,17 +92,45 @@ describe('runCycle last_full_cycle_at exit hook', () => {
|
||||
});
|
||||
});
|
||||
|
||||
test('legacy caller (no sourceId) does NOT write any source timestamp', async () => {
|
||||
test('no explicit sourceId but brainDir resolves a source → writes the resolved source timestamp', async () => {
|
||||
await withEnv({ GBRAIN_HOME: gbrainHome }, async () => {
|
||||
await seedSource('default-like');
|
||||
// No sourceId passed; should remain untouched.
|
||||
// The autopilot's inline cycle sets brainDir but passes no sourceId.
|
||||
// runCycle resolves the source from brainDir (local_path match) into
|
||||
// cycleSourceId and stamps last_full_cycle_at for it — otherwise
|
||||
// cycle_freshness reports the brain stale even while the autopilot
|
||||
// cycles every interval (#1993).
|
||||
await seedSource('resolved-from-dir'); // local_path = brainDir
|
||||
expect(await readLastFullCycleAt('resolved-from-dir')).toBeNull();
|
||||
|
||||
const t0 = Date.now();
|
||||
const report = await runCycle(engine, {
|
||||
brainDir,
|
||||
phases: ['lint'],
|
||||
});
|
||||
expect(['ok', 'clean']).toContain(report.status);
|
||||
|
||||
const after = await readLastFullCycleAt('resolved-from-dir');
|
||||
expect(after).not.toBeNull();
|
||||
expect(new Date(after!).getTime()).toBeGreaterThanOrEqual(t0);
|
||||
});
|
||||
});
|
||||
|
||||
test('no sourceId and brainDir matches no source → does not write', async () => {
|
||||
await withEnv({ GBRAIN_HOME: gbrainHome }, async () => {
|
||||
// A source exists but its local_path does NOT match brainDir, so
|
||||
// resolveSourceForDir returns undefined, cycleSourceId is undefined,
|
||||
// and no per-source timestamp is written.
|
||||
await engine.executeRaw(
|
||||
`INSERT INTO sources (id, name, local_path, config, archived, created_at)
|
||||
VALUES ('unmatched', 'unmatched', '/no/such/repo', '{}'::jsonb, false, NOW())
|
||||
ON CONFLICT (id) DO UPDATE SET local_path = EXCLUDED.local_path`,
|
||||
[],
|
||||
);
|
||||
await runCycle(engine, {
|
||||
brainDir,
|
||||
phases: ['lint'],
|
||||
});
|
||||
// No per-source write happens; default source's config stays empty.
|
||||
const after = await readLastFullCycleAt('default-like');
|
||||
expect(after).toBeNull();
|
||||
expect(await readLastFullCycleAt('unmatched')).toBeNull();
|
||||
});
|
||||
});
|
||||
|
||||
|
||||
@@ -15,7 +15,6 @@ import { describe, test, expect, beforeAll, afterAll, beforeEach } from 'bun:tes
|
||||
import { PGLiteEngine } from '../src/core/pglite-engine.ts';
|
||||
import { resetPgliteState } from './helpers/reset-pglite.ts';
|
||||
import { embedStaleForSource } from '../src/core/embed-stale.ts';
|
||||
import { configureGateway, resetGateway } from '../src/core/ai/gateway.ts';
|
||||
import type { ChunkInput } from '../src/core/types.ts';
|
||||
|
||||
let engine: PGLiteEngine;
|
||||
@@ -277,49 +276,4 @@ describe('embedStaleForSource', () => {
|
||||
// The stale text row actually got its embedding.
|
||||
expect(txtRow.embedded_at).not.toBeNull();
|
||||
});
|
||||
|
||||
// #1717: the backfill path must label re-embedded chunks with the model
|
||||
// that produced the vector, and preserve the existing label on chunks it
|
||||
// did not touch (before the fix, both were reset to the engine default).
|
||||
test('labels re-embedded chunks with the gateway model, preserves untouched labels (#1717)', async () => {
|
||||
configureGateway({
|
||||
embedding_model: 'openai:text-embedding-3-large',
|
||||
env: { OPENAI_API_KEY: 'sk-test-embed-stale-1717' },
|
||||
});
|
||||
try {
|
||||
await engine.putPage('notes/model-label', {
|
||||
type: 'note',
|
||||
title: 'model-label',
|
||||
compiled_truth: '# model-label\n\nseeded',
|
||||
});
|
||||
await engine.upsertChunks('notes/model-label', [
|
||||
{
|
||||
chunk_index: 0,
|
||||
chunk_text: 'already embedded elsewhere',
|
||||
chunk_source: 'compiled_truth',
|
||||
embedding: new Float32Array(1536).fill(0.01),
|
||||
model: 'voyage:voyage-3',
|
||||
token_count: 4,
|
||||
},
|
||||
{
|
||||
chunk_index: 1,
|
||||
chunk_text: 'stale chunk needing embed',
|
||||
chunk_source: 'compiled_truth',
|
||||
token_count: 5,
|
||||
embedding: undefined, // stale
|
||||
},
|
||||
]);
|
||||
|
||||
const result = await embedStaleForSource(engine, 'default', { embedFn: fakeEmbedFn });
|
||||
expect(result.embedded).toBe(1);
|
||||
|
||||
const after = await engine.getChunks('notes/model-label');
|
||||
const preserved = after.find((c) => c.chunk_index === 0)!;
|
||||
const reembedded = after.find((c) => c.chunk_index === 1)!;
|
||||
expect(reembedded.model).toBe('openai:text-embedding-3-large');
|
||||
expect(preserved.model).toBe('voyage:voyage-3');
|
||||
} finally {
|
||||
resetGateway();
|
||||
}
|
||||
});
|
||||
});
|
||||
|
||||
@@ -37,8 +37,6 @@ mock.module('../src/core/embedding.ts', () => ({
|
||||
// setPageEmbeddingSignature / invalidateStaleSignatureEmbeddings resolve to
|
||||
// null via the Proxy default, so the signature value is inert here.
|
||||
currentEmbeddingSignature: () => 'test:model:1536',
|
||||
// #1717: embed paths stamp this label on (re)embedded chunks.
|
||||
resolveEmbeddingModelLabel: () => 'openai:text-embedding-3-large',
|
||||
}));
|
||||
|
||||
// Import AFTER mocking.
|
||||
@@ -805,34 +803,3 @@ describe('embedAllStale --source threading (D7)', () => {
|
||||
expect((firstCallOpts as { sourceId?: string }).sourceId).toBe('media-corpus');
|
||||
});
|
||||
});
|
||||
|
||||
// #1717: content_chunks.model must record the model that actually produced
|
||||
// each vector, not the gateway/engine default.
|
||||
describe('content_chunks.model labeling (#1717)', () => {
|
||||
test('stamps the resolved embedding model on re-embedded chunks, preserves it on untouched chunks', async () => {
|
||||
let upserted: any[] | undefined;
|
||||
// Chunk 0 is stale (no embedded_at) → gets re-embedded this pass.
|
||||
// Chunk 1 is already embedded with a DIFFERENT model → must be preserved,
|
||||
// not relabeled to the current model.
|
||||
const chunks = [
|
||||
{ chunk_index: 0, chunk_text: 'a', chunk_source: 'compiled_truth', embedded_at: null, model: 'zeroentropyai:zembed-1', token_count: 1 },
|
||||
{ chunk_index: 1, chunk_text: 'b', chunk_source: 'compiled_truth', embedded_at: '2026-01-01', embedding: new Float32Array(1536), model: 'voyage:voyage-3', token_count: 1 },
|
||||
];
|
||||
const engine = mockEngine({
|
||||
getPage: async () => ({ slug: 'notes/x', compiled_truth: 'a', timeline: '', source_id: 'default' }),
|
||||
getChunks: async () => chunks,
|
||||
upsertChunks: async (_slug: string, c: any[]) => { upserted = c; },
|
||||
setPageEmbeddingSignature: async () => null,
|
||||
});
|
||||
|
||||
await runEmbedCore(engine, { slugs: ['notes/x'] });
|
||||
|
||||
expect(upserted).toBeDefined();
|
||||
const byIdx = Object.fromEntries(upserted!.map(c => [c.chunk_index, c]));
|
||||
// Re-embedded chunk carries the model that produced its vector (was
|
||||
// mislabeled with the default before the fix).
|
||||
expect(byIdx[0].model).toBe('openai:text-embedding-3-large');
|
||||
// Untouched chunk keeps its original model — no wholesale relabel.
|
||||
expect(byIdx[1].model).toBe('voyage:voyage-3');
|
||||
});
|
||||
});
|
||||
|
||||
@@ -73,21 +73,4 @@ describe('importFromContent embedding_signature stamping (F1)', () => {
|
||||
await importFromContent(engine, 'concepts/unstamped', '# Unstamped\n\nbody content.', { noEmbed: true });
|
||||
expect(await signatureOf('concepts/unstamped')).toBeNull();
|
||||
});
|
||||
|
||||
// #1717: content_chunks.model must record the model that produced the
|
||||
// vector (the configured gateway model), not the engine's hardcoded
|
||||
// default. The gateway here is configured to openai:text-embedding-3-large,
|
||||
// which differs from DEFAULT_EMBEDDING_MODEL — so this fails without the
|
||||
// import-path model stamping.
|
||||
test('inline embed labels content_chunks.model with the configured model (#1717)', async () => {
|
||||
await importFromContent(engine, 'concepts/labeled', '# Labeled\n\nsome body content to chunk and embed.', {});
|
||||
const rows = await engine.executeRaw<{ model: string }>(
|
||||
`SELECT cc.model FROM content_chunks cc
|
||||
JOIN pages p ON p.id = cc.page_id
|
||||
WHERE p.slug = $1 AND p.source_id = 'default'`,
|
||||
['concepts/labeled'],
|
||||
);
|
||||
expect(rows.length).toBeGreaterThan(0);
|
||||
for (const r of rows) expect(r.model).toBe('openai:text-embedding-3-large');
|
||||
});
|
||||
});
|
||||
|
||||
Reference in New Issue
Block a user