Class RankedDocument

Document represents document content along with its metadata that can be embedded, indexed or retrieved. Each document can contain multiple parts (for example text and an image)

Hierarchy (View Summary)

Implements

Constructors

constructor

Properties

content metadata

Accessors

data dataType media text

Methods

getEmbeddingDocuments score toJSON fromData fromMedia fromText

Constructors

constructor

  • new RankedDocument(
        data: {
            content: (
                | { media?: undefined; text: string }
                | { media: { contentType?: string; url: string }; text?: undefined }
            )[];
            metadata: { score: number } & { [k: string]: unknown };
        },
    ): RankedDocument

  • Parameters

    • data: {
          content: (
              | { media?: undefined; text: string }
              | { media: { contentType?: string; url: string }; text?: undefined }
          )[];
          metadata: { score: number } & { [k: string]: unknown };
      }

    Returns RankedDocument

Properties

content

content: (
    | { media?: undefined; text: string }
    | { media: { contentType?: string; url: string }; text?: undefined }
)[]

metadata

metadata: { score: number } & Record<string, any>

Accessors

data

  • get data(): string

  • Gets the first item in the document. Either text or media url.

    Returns string

dataType

  • get dataType(): undefined | string

  • Gets the contentType of the data that is returned by data()

    Returns undefined | string

media

  • get media(): { contentType?: string; url: string }[]

  • Media array getter.

    Returns { contentType?: string; url: string }[]

    the array of media parts.

text

  • get text(): string

  • Concatenates all text parts present in the document with no delimiter.

    Returns string

    A string of all concatenated text parts.

Methods

getEmbeddingDocuments

  • getEmbeddingDocuments(
        embeddings: { embedding: number[]; metadata?: Record<string, unknown> }[],
    ): Document[]

  • Embedders may return multiple embeddings for a single document. But storage still requires a 1:1 relationship. So we create an array of Documents from a single document - one per embedding.

    Parameters

    • embeddings: { embedding: number[]; metadata?: Record<string, unknown> }[]

      The embeddings to create the documents from.

    Returns Document[]

    an array of documents based on this document and the embeddings.

score

  • score(): number

  • Returns the score of the document.

    Returns number

    The score of the document.

toJSON

  • toJSON(): {
        content: (
            | { media?: undefined; text: string }
            | { media: { contentType?: string; url: string }; text?: undefined }
        )[];
        metadata?: Record<string, any>;
    }

  • Returns {
        content: (
            | { media?: undefined; text: string }
            | { media: { contentType?: string; url: string }; text?: undefined }
        )[];
        metadata?: Record<string, any>;
    }

StaticfromData

  • fromData(
        data: string,
        dataType?: string,
        metadata?: Record<string, unknown>,
    ): Document

  • Parameters

    • data: string
    • OptionaldataType: string
    • Optionalmetadata: Record<string, unknown>

    Returns Document

StaticfromMedia

  • fromMedia(
        url: string,
        contentType?: string,
        metadata?: Record<string, unknown>,
    ): Document

  • Parameters

    • url: string
    • OptionalcontentType: string
    • Optionalmetadata: Record<string, unknown>

    Returns Document

StaticfromText

  • fromText(text: string, metadata?: Record<string, any>): Document

  • Parameters

    • text: string
    • Optionalmetadata: Record<string, any>

    Returns Document

Settings

Member Visibility

On This Page

Constructors
Properties
Accessors
Methods