Skip to main content

Interface: HybridRetrieverOptions

Defined in: packages/agentos/src/memory/retrieval/hybrid/HybridRetriever.ts:63

Options for constructing a HybridRetriever.

Properties​

bm25Config?​

optional bm25Config: BM25Config

Defined in: packages/agentos/src/memory/retrieval/hybrid/HybridRetriever.ts:66

BM25 config (k1, b, optional tokenizer). Defaults match BM25Index.

defaultDenseWeight?​

optional defaultDenseWeight: number

Defined in: packages/agentos/src/memory/retrieval/hybrid/HybridRetriever.ts:94

Default dense weight in RRF.

Default​

0.7

defaultRrfK?​

optional defaultRrfK: number

Defined in: packages/agentos/src/memory/retrieval/hybrid/HybridRetriever.ts:98

Default RRF constant.

Default​

60

defaultSparseWeight?​

optional defaultSparseWeight: number

Defined in: packages/agentos/src/memory/retrieval/hybrid/HybridRetriever.ts:96

Default sparse weight in RRF.

Default​

0.3

factGraphQueryClassifier?​

optional factGraphQueryClassifier: FactGraphQueryClassifier

Defined in: packages/agentos/src/memory/retrieval/hybrid/HybridRetriever.ts:117

Step-9: optional query classifier. Given the user's query, returns (subject, predicate) pairs to look up in the factStore, plus whether the query is temporal (in which case ALL facts for a subject are prepended, not just the latest). When absent and factStore is set, a keyword-based default is used.

factStore?​

optional factStore: FactStore

Defined in: packages/agentos/src/memory/retrieval/hybrid/HybridRetriever.ts:109

Step-9: optional FactStore of (subject, predicate) → Fact[] tuples extracted at session-ingest time. When present AND the query classifier extracts at least one (subject, predicate) pair from the query, the latest fact per pair is prepended as a synthetic ScoredMemoryTrace (retrievalScore: 1.0, sourceType: 'fact_graph') to the merged pool BEFORE rerank. Preserves literal object tokens where summary-based approaches (Steps 5/7/8) paraphrased them away.

hydeRetriever?​

optional hydeRetriever: HydeRetriever

Defined in: packages/agentos/src/memory/retrieval/hybrid/HybridRetriever.ts:82

Optional HyDE retriever for query expansion (Step 4). When set, each retrieve() call generates a hypothesis and uses it as the query for BOTH dense (memoryStore.query) and sparse (bm25.search). The reranker continues to use the ORIGINAL user query so it scores documents against real user intent, not the hypothesis. HyDE generation is non-critical — errors fall back to the raw query without aborting retrieval.

memoryStore​

memoryStore: MemoryStore

Defined in: packages/agentos/src/memory/retrieval/hybrid/HybridRetriever.ts:64

rerankerService?​

optional rerankerService: RerankerService

Defined in: packages/agentos/src/memory/retrieval/hybrid/HybridRetriever.ts:72

Optional neural reranker. When provided, the merged pool is reranked before truncation. Passing the same reranker the baseline uses is the matched-ablation path.

splitAmbiguousThreshold?​

optional splitAmbiguousThreshold: number

Defined in: packages/agentos/src/memory/retrieval/hybrid/HybridRetriever.ts:92

Step-6: enable split-on-ambiguous rerank refinement. When set to a value in (0, 1], the bottom fraction of traces by first-pass rerank score are split at sentence boundaries, rescored with a second rerank call (same query), and replaced by their better half ONLY IF the better half outscores the original. Monotonic.

Default: undefined (no split, Step 3 behavior preserved).