langchain.js
    Preparing search index...

    Class AmazonKendraRetriever

    Class for interacting with Amazon Kendra, an intelligent search service provided by AWS. Extends the BaseRetriever class.

    const retriever = new AmazonKendraRetriever({
      topK: 10,
      indexId: "YOUR_INDEX_ID",
      region: "us-east-2",
      clientOptions: {
        credentials: {
          accessKeyId: "YOUR_ACCESS_KEY_ID",
          secretAccessKey: "YOUR_SECRET_ACCESS_KEY",
        },
      },
    });

    const docs = await retriever.getRelevantDocuments("How are clouds formed?");

    Hierarchy (View Summary)

    Index

    Constructors

    constructor

    Properties

    attributeFilter? indexId kendraClient lc_namespace topK

    Methods

    _getRelevantDocuments cleanResult combineText convertQueryItem convertRetrieverItem getDocAttributes getDocAttributeValue getQueryDocs getQueryItemExcerpt getRetrieverDocs queryKendra lc_name

    Constructors

    Properties

    attributeFilter?: AttributeFilter
    indexId: string
    kendraClient: KendraClient
    lc_namespace: string[] = ...
    topK: number

    Methods

    _getRelevantDocuments

    • _getRelevantDocuments(query: string): Promise<Document[]>

      Parameters

      • query: string

      Returns Promise<Document[]>

    • Cleans the result text by replacing sequences of whitespace with a single space and removing ellipses.

      Parameters

      • resText: string

        The result text to clean.

      Returns string

      The cleaned result text.

    • Combines title and excerpt into a single string.

      Parameters

      • Optionaltitle: string

        The title of the document.

      • Optionalexcerpt: string

        An excerpt from the document.

      Returns string

      A single string combining the title and excerpt.

    • Converts a QueryResultItem object into a Document object.

      Parameters

      • item: QueryResultItem

        The QueryResultItem object to convert.

      Returns any

      A Document object.

    • Converts a RetrieveResultItem object into a Document object.

      Parameters

      • item: RetrieveResultItem

        The RetrieveResultItem object to convert.

      Returns any

      A Document object.

    • Extracts the attribute key-value pairs from an array of DocumentAttribute objects.

      Parameters

      • OptionaldocumentAttributes: DocumentAttribute[]

        The array of DocumentAttribute objects to extract the key-value pairs from.

      Returns { [key: string]: unknown }

      An object containing the extracted attribute key-value pairs.

    • Extracts the attribute value from a DocumentAttributeValue object.

      Parameters

      • value: DocumentAttributeValue

        The DocumentAttributeValue object to extract the value from.

      Returns string | number | string[] | Date

      The extracted attribute value.

    • Extracts the top-k documents from a QueryCommandOutput object.

      Parameters

      • response: QueryCommandOutput

        The QueryCommandOutput object to extract the documents from.

      • pageSize: number

        The number of documents to extract.

      Returns any[]

      An array of Document objects.

    • Extracts the excerpt text from a QueryResultItem object.

      Parameters

      • item: QueryResultItem

        The QueryResultItem object to extract the excerpt text from.

      Returns string

      The extracted excerpt text.

    • Extracts the top-k documents from a RetrieveCommandOutput object.

      Parameters

      • response: RetrieveCommandOutput

        The RetrieveCommandOutput object to extract the documents from.

      • pageSize: number

        The number of documents to extract.

      Returns Document[]

      An array of Document objects.

    • Sends a retrieve or query request to Kendra and returns the top-k documents.

      Parameters

      • query: string

        The query to send to Kendra.

      • topK: number

        The number of top documents to return.

      • OptionalattributeFilter: AttributeFilter

        Optional filter to apply when retrieving documents.

      Returns Promise<any[]>

      A Promise that resolves to an array of Document objects.

    • Returns string

    Settings

    Member Visibility

    On This Page

    Constructors
    Properties
    Methods