Graphorin API reference v0.4.0
Graphorin API reference / @graphorin/memory / / FusionStrategy
Type Alias: FusionStrategy
ts
type FusionStrategy =
| {
strategy: "rrf";
}
| {
k?: number;
strategy: "weighted";
weights: FusionWeights;
};Defined in: packages/memory/src/tiers/semantic-memory.ts:88
Score-fusion strategy for SemanticMemory.search (X-2).
'rrf'(the default whenfusionis omitted) fuses the candidate lists through the configured reranker — the zero-tuning RRFReranker unless one was overridden.'weighted'fuses through WeightedRRFReranker, scaling each list's reciprocal-rank contribution by its FusionWeights, for callers who have calibrated retriever reliability against labels (the P0-1 eval harness). At equal weights it reproduces RRF.
Union Members
Type Literal
ts
{
strategy: "rrf";
}Type Literal
ts
{
k?: number;
strategy: "weighted";
weights: FusionWeights;
}| Name | Type | Description | Defined in |
|---|---|---|---|
k? | number | Override the RRF constant for the weighted fuse. Default 60. | packages/memory/src/tiers/semantic-memory.ts:94 |
strategy | "weighted" | - | packages/memory/src/tiers/semantic-memory.ts:91 |
weights | FusionWeights | - | packages/memory/src/tiers/semantic-memory.ts:92 |