mirror of
https://github.com/garrytan/gbrain.git
synced 2026-07-29 16:39:15 +00:00
85286a556c
* fix(search): stop boosting compiled_truth at default detail (#3430) COMPILED_TRUTH_BOOST = 2.0 is applied AFTER RRF normalization, and RRF's whole dynamic range over a 100-deep pool is 1/60 -> 1/160 (a factor of 2.67). So a 2.0x multiplier consumes roughly three quarters of the range: break-even is `2/(60+r) >= 1/60`, i.e. r <= 60, which means ANY boosted chunk inside the first 60 ranks outranks an unboosted rank-1 chunk. That is a categorical filter, not a tilt. Measured against master's own rrfFusion, with the correct answer in a fenced_code chunk at vector rank 0: compiled_truth chunks in pool | final rank | in top-20 10 | 10 | yes 20 | 20 | NO 40 | 40 | NO 80 | 59 | NO With the boost off the answer stays at rank 0 in every case. The gate was spelled `detail !== 'high'` -- written as though `high` were the special case. The documented contract in src/core/operations.ts is "low (compiled truth only), medium (default, all with dedup), high (all chunks)", which makes LOW the special one: `low` already restricts to compiled_truth, so a boost there is a no-op among equals, while `medium` and `high` are both meant to see everything. So the default detail was silently compiled-truth-only, contradicting the op's own description. Three changes: 1. The three fusion call sites now route through a named predicate, `shouldBoostCompiledTruth(detail)`, returning true only for 'low'. Extracted rather than left inline precisely because an inline expression is only reachable through a full hybridSearch round trip -- which is why the inversion went unnoticed. The predicate is directly unit-testable. 2. KNOBS_HASH_VERSION 13 -> 14. Results are cached AFTER fusion, so rows ranked under the old semantics would otherwise be served under the new ones for the whole TTL (3600s default). One-time miss spike on upgrade. 3. test/search-compiled-truth-boost-scope.test.ts pins both the mapping and the arithmetic, and documents the displacement it prevents. Verified the tests discriminate: stubbing the OLD predicate body into master (so the failure is behavioral rather than a missing export) gives 4 fail / 3 pass; with the fix, 7 pass. typecheck clean, verify 32/32, and 144 pass / 0 fail across the search + fusion + cache suites. * fix(test): update the three remaining KNOBS_HASH_VERSION pins to 14 (#3430) Missed in the first pass because I ran a targeted set of test files instead of the full suite. CI shards 3, 8 and 10 caught them: test/search/knobs-hash-reranker.test.ts:67 test/cross-modal-phase1.test.ts:139,149 test/search-alias-resolved-boost.test.ts:93 Each carries the running history of why the version moved, so each gets the 13→14 rationale appended rather than just the number swapped. No pins at 13 remain anywhere in test/. --------- Co-authored-by: Garry Tan <garrytan@gmail.com>
96 lines
3.8 KiB
TypeScript
96 lines
3.8 KiB
TypeScript
// v0.42 Type Unification (T32) — alias_resolved search boost stage.
|
|
//
|
|
// Coverage: pages whose slug is a canonical_slug in slug_aliases get 1.05x
|
|
// score multiplier; non-alias-canonical pages unchanged; stage stamps
|
|
// alias_resolved_boost field for --explain; KNOBS_HASH_VERSION bumped.
|
|
|
|
import { afterAll, beforeAll, beforeEach, describe, expect, it } from 'bun:test';
|
|
import { PGLiteEngine } from '../src/core/pglite-engine.ts';
|
|
import { resetPgliteState } from './helpers/reset-pglite.ts';
|
|
import { runPostFusionStages, type PostFusionOpts } from '../src/core/search/hybrid.ts';
|
|
import { KNOBS_HASH_VERSION } from '../src/core/search/mode.ts';
|
|
import type { SearchResult } from '../src/core/types.ts';
|
|
|
|
let engine: PGLiteEngine;
|
|
|
|
beforeAll(async () => {
|
|
engine = new PGLiteEngine();
|
|
await engine.connect({});
|
|
await engine.initSchema();
|
|
});
|
|
|
|
afterAll(async () => {
|
|
await engine.disconnect();
|
|
});
|
|
|
|
beforeEach(async () => {
|
|
await resetPgliteState(engine);
|
|
});
|
|
|
|
const noopPostFusionOpts: PostFusionOpts = {
|
|
applyBacklinks: false,
|
|
salience: 'off',
|
|
recency: 'off',
|
|
graphSignalsEnabled: false,
|
|
};
|
|
|
|
describe('alias_resolved boost stage', () => {
|
|
it('applies 1.05x multiplier to pages that are canonicals of aliases', async () => {
|
|
// Insert an alias pointing at canonical-page
|
|
await engine.executeRaw(
|
|
`INSERT INTO slug_aliases (source_id, alias_slug, canonical_slug) VALUES ('default', 'old-name', 'canonical-page')`,
|
|
);
|
|
const results: SearchResult[] = [
|
|
{
|
|
slug: 'canonical-page', source_id: 'default', score: 1.0,
|
|
chunk_id: 1, page_id: 1, chunk_text: '', chunk_index: 0,
|
|
title: 'Canonical', type: 'concept' as never, slug_lower: 'canonical-page',
|
|
} as unknown as SearchResult,
|
|
{
|
|
slug: 'plain-page', source_id: 'default', score: 1.0,
|
|
chunk_id: 2, page_id: 2, chunk_text: '', chunk_index: 0,
|
|
title: 'Plain', type: 'concept' as never, slug_lower: 'plain-page',
|
|
} as unknown as SearchResult,
|
|
];
|
|
await runPostFusionStages(engine, results, noopPostFusionOpts);
|
|
// canonical-page gets 1.05x boost
|
|
expect(results[0].score).toBeCloseTo(1.05, 5);
|
|
expect(results[0].alias_resolved_boost).toBe(1.05);
|
|
// plain-page unchanged
|
|
expect(results[1].score).toBeCloseTo(1.0, 5);
|
|
expect(results[1].alias_resolved_boost).toBeUndefined();
|
|
});
|
|
|
|
it('does not boost when no aliases exist', async () => {
|
|
const results: SearchResult[] = [{
|
|
slug: 'plain', source_id: 'default', score: 1.0,
|
|
chunk_id: 1, page_id: 1, chunk_text: '', chunk_index: 0,
|
|
title: 'p', type: 'concept' as never, slug_lower: 'plain',
|
|
} as unknown as SearchResult];
|
|
await runPostFusionStages(engine, results, noopPostFusionOpts);
|
|
expect(results[0].score).toBeCloseTo(1.0, 5);
|
|
expect(results[0].alias_resolved_boost).toBeUndefined();
|
|
});
|
|
|
|
it('is source-scoped (F9): alias in source A does not boost in source B', async () => {
|
|
await engine.executeRaw(`INSERT INTO sources (id, name) VALUES ('alt', 'alt') ON CONFLICT DO NOTHING`);
|
|
await engine.executeRaw(
|
|
`INSERT INTO slug_aliases (source_id, alias_slug, canonical_slug) VALUES ('alt', 'old', 'shared')`,
|
|
);
|
|
// Same slug, different source — should NOT be boosted (alias is in 'alt')
|
|
const results: SearchResult[] = [{
|
|
slug: 'shared', source_id: 'default', score: 1.0,
|
|
chunk_id: 1, page_id: 1, chunk_text: '', chunk_index: 0,
|
|
title: 's', type: 'concept' as never, slug_lower: 'shared',
|
|
} as unknown as SearchResult];
|
|
await runPostFusionStages(engine, results, noopPostFusionOpts);
|
|
expect(results[0].alias_resolved_boost).toBeUndefined();
|
|
});
|
|
});
|
|
|
|
describe('KNOBS_HASH_VERSION', () => {
|
|
it('is 14 (13→14 compiled_truth boost no longer applies at detail=medium, so pre-fix rankings must be unreachable, #3430)', () => {
|
|
expect(KNOBS_HASH_VERSION).toBe(14);
|
|
});
|
|
});
|