Member Junction
    Preparing search index...

    Options for hybrid search combining vector similarity with keyword (BM25) search. Providers that support hybrid search implement VectorDBBase.HybridQuery().

    interface HybridQueryOptions {
        Alpha?: number;
        filter?: object;
        FusionMethod?: "rrf" | "weighted";
        includeMetadata?: boolean;
        includeValues?: boolean;
        KeywordQuery?: string;
        providerConfig?: Record<string, unknown>;
        topK: number;
        vector: RecordValues;
    }

    Hierarchy (View Summary)

    Index

    Properties

    Alpha?: number

    Balance between vector and keyword: 0.0 = pure keyword, 1.0 = pure vector (default: 0.7)

    filter?: object

    This parameter allows you to modify your query with a metadata filter.

    FusionMethod?: "rrf" | "weighted"

    Method for fusing vector and keyword result lists (default: 'rrf')

    includeMetadata?: boolean

    This boolean value specifies whether metadata values are returned with query results.

    By default, metadata values are not returned to reduce the size of the request payload.

    includeValues?: boolean

    This boolean value specifies whether embedding values are returned with query results.

    By default, values are not returned to reduce the size of the request payload.

    KeywordQuery?: string

    Keyword query string for the BM25/text search component

    providerConfig?: Record<string, unknown>

    Optional opaque configuration blob passed through to the vector DB driver. Each driver reads the keys it understands and ignores the rest — callers do not need to know which driver is in use. Populated from the rendered ExternalIndexConfig on the scope's external-index row.

    Example: Pinecone reads providerConfig.namespace for tenant isolation.

    topK: number

    The number of query results you would like returned.

    vector: RecordValues

    Vector values for the semantic similarity component