Member Junction
    Preparing search index...
    type VectorEmeddingData = {
        aiModelID: string;
        embedding?: BaseEmbeddings;
        embeddingAPIKey: string;
        embeddingDriverClass: string;
        embeddingModelAPIName: string;
        embeddingRunner: AIEmbeddingRunner;
        vectorDB: VectorDBBase;
        vectorDBAPIKey: string;
        vectorDBClassKey: string;
    }
    Index

    Properties

    aiModelID: string

    The entity document's embedding model (MJ: AI Models.ID).

    embedding?: BaseEmbeddings

    Use embeddingRunner. Embedding now goes through AIEmbeddingRunner, which brings credential resolution, failover and run rows. For older callers this is still a working driver, built on first read with the legacy environment-variable key.

    embeddingAPIKey: string

    The legacy AI_VENDOR_API_KEY__<DRIVER> value, or '' when unset (normal for keyless drivers). Informational only: AIEmbeddingRunner resolves credentials itself.

    embeddingDriverClass: string
    embeddingModelAPIName: string

    The AIModel's APIName (e.g. 'Xenova/gte-small') — required by some providers (LocalEmbedding throws if absent) to identify the underlying model.

    embeddingRunner: AIEmbeddingRunner

    Runs the embedding calls. Pass aiModelID as ModelID so every vector comes from the document's model.

    vectorDB: VectorDBBase
    vectorDBAPIKey: string
    vectorDBClassKey: string