A class that represents a connection to a Google Vertex AI Matching Engine instance.

Hierarchy

Implements

Constructors

Properties

FilterType: Restriction[]
apiEndpoint: string

The host to connect to for queries and upserts.

apiVersion: string

The version of the API functions. Part of the path.

authOptions: GoogleAuthOptions<JSONClient>

Explicitly set Google Auth credentials if you cannot get them from google auth application-default login This is useful for serverless or autoscaling environments like Fargate

caller: AsyncCaller
callerOptions: AsyncCallerCallOptions
callerParams: AsyncCallerParams
deployedIndexId: string

The id for the "deployed index", which is an identifier in the index endpoint that references the index (but is not the index id)

docstore: Docstore

Docstore that retains the document, stored by ID

embeddings: Embeddings
endpoint: string

Hostname for the API call

index: string

The id for the index

indexEndpoint: string

The id for the index endpoint

indexEndpointClient: IndexEndpointConnection
location: string

Region where the LLM is stored

removeDatapointClient: RemoveDatapointConnection
upsertDatapointClient: UpsertDatapointConnection

Methods

  • Parameters

    • documents: Document<Record<string, any>>[]

    Returns Promise<void>

  • Parameters

    • vectors: number[][]
    • documents: Document<Record<string, any>>[]

    Returns Promise<void>

  • Create an index datapoint for the vector and document id. If an id does not exist, create it and set the document to its value.

    Parameters

    Returns IndexDatapoint

  • Parameters

    • documentMetadata: Record<string, any>

    Returns {
        [key: string]: string | number | boolean | string[] | null;
    }

    • [key: string]: string | number | boolean | string[] | null
  • For this index endpoint, figure out what API Endpoint URL and deployed index ID should be used to do upserts and queries. Also sets the apiEndpoint and deployedIndexId property for future use.

    Returns Promise<PublicAPIEndpointInfo>

    The URL

  • Returns Promise<string>

  • Returns Promise<string>

  • Given the metadata from a document, convert it to an array of Restriction objects that may be passed to the Matching Engine and stored. The default implementation flattens any metadata and includes it as an "allowList". Subclasses can choose to convert some of these to "denyList" items or to add additional restrictions (for example, to format dates into a different structure or to add additional restrictions based on the date).

    Parameters

    • documentMetadata: Record<string, any>

      The metadata from a document

    Returns Restriction[]

    a Restriction[] (or an array of a subclass, from the FilterType)

  • Parameters

    Returns Promise<Document<Record<string, any>>[]>

  • Parameters

    • query: number[]
    • k: number
    • Optional filter: Restriction[]

    Returns Promise<[Document<Record<string, any>>, number][]>

  • Parameters

    Returns Promise<[Document<Record<string, any>>, number][]>

  • Return documents selected using the maximal marginal relevance. Maximal marginal relevance optimizes for similarity to the query AND diversity among selected documents.

    Parameters

    Returns Promise<Document<Record<string, any>>[]>

    • List of documents selected by maximal marginal relevance.

Generated using TypeDoc