feat(8.0): vector allowedIds predicate-pushdown into find() (#46)

Filtered semantic search now keeps its recall. A find({ query, where }) that
pairs vector search with a metadata filter restricts the HNSW beam walk to the
matching candidates INSIDE the traversal (walk-all, collect-allowed) rather than
filtering the top-k afterward — so a query whose nearest vectors are all filtered
out still returns the best matches that DO pass the filter, instead of empty.

- JS HNSW search() honors the 8.0 `allowedIds: OpaqueIdSet | ReadonlySet<string>`
  contract param (ANDs with candidateIds; ignores the opaque Buffer form it can't
  decode — only a native provider consumes that).
- MetadataIndexProvider gains an OPTIONAL `getIdSetForFilter(filter): OpaqueIdSet`
  producer — the native (cor) metadata index returns its roaring filter result as
  a serialized buffer; the JS index does not implement it.
- find() forwards the matched universe to the vector walk: the opaque buffer
  (zero id materialization) when the native producer is present, alongside the
  materialized visibility-precise candidateIds for the JS index. Visibility stays
  correct via the existing post-search hard filter.

Tested: JS recall/restriction/AND/opaque-ignore on the index directly, plus
find() forwarding the opaque universe through to the beam walk via a stubbed
native producer.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
This commit is contained in:
David Snelling 2026-06-23 13:18:34 -07:00
parent 632d90aac5
commit dd325f2f94
5 changed files with 279 additions and 15 deletions

View file

@ -130,6 +130,22 @@ export interface MetadataIndexProvider {
getIds(field: string, value: any): Promise<string[]>
getIdsForFilter(filter: any): Promise<string[]>
/**
* @description OPTIONAL: resolve a `where` filter to its matching id universe as
* an {@link OpaqueIdSet} (a serialized roaring `Buffer`) WITHOUT materializing
* the ids in TypeScript the producer half of the predicate-pushdown
* (`CTX-PUSHDOWN-ALLOWEDIDS`) and graph queryexpand fusion. Brainy forwards the
* returned set opaquely to `VectorIndexProvider.search({ allowedIds })` and
* `GraphAccelerationProvider.traverse(seeds)`; it NEVER inspects it. A native
* (cor) metadata index returns its roaring filter result directly (zero
* crossing); the JS index does not implement this Brainy falls back to
* {@link MetadataIndexProvider.getIdsForFilter} + a `ReadonlySet<string>` when
* absent. Present the native stack is the active provider, so the matching
* native vector/graph engines can decode the same envelope.
* @param filter - The same filter shape accepted by `getIdsForFilter`.
* @returns The matching id universe as an opaque set.
*/
getIdSetForFilter?(filter: any): Promise<OpaqueIdSet>
getIdsForTextQuery(query: string): Promise<Array<{ id: string; matchCount: number }>>
getSortedIdsForFilter(filter: any, orderBy: string, order?: 'asc' | 'desc'): Promise<string[]>
getFilterValues(field: string): Promise<string[]>