ravendb/dist/commonjs/Documents/Session/VectorFieldFactory.d.ts

RavenDB client for Node.js

import { VectorEmbeddingType } from "../Queries/VectorSearch/VectorEmbeddingType.js";
import { Field } from "../../Types/index.js";
export interface IVectorFieldFactory<T> {
    /**
     * Defines the text field that vector search will be performed on
     * @param fieldName Name of the document field containing text data
     */
    withText(fieldName: Field<T>): IVectorEmbeddingTextField;
    /**
     * Defines the embedding field that vector search will be performed on
     * @param fieldName Name of the document field containing embedding data
     * @param storedEmbeddingQuantization Quantization that was performed on stored embeddings
     */
    withEmbedding(fieldName: Field<T>, storedEmbeddingQuantization?: VectorEmbeddingType): IVectorEmbeddingField;
    /**
     * Defines the embedding field (encoded as base64) that vector search will be performed on
     * @param fieldName Name of the document field containing base64 encoded embedding data
     * @param storedEmbeddingQuantization Quantization of stored embeddings
     */
    withBase64(fieldName: Field<T>, storedEmbeddingQuantization?: VectorEmbeddingType): IVectorEmbeddingField;
    /**
     * Defines the field (that's already indexed) that vector search will be performed on
     * @param fieldName Name of the index-field containing indexed data
     */
    withField(fieldName: Field<T>): IVectorField;
}
export interface IVectorEmbeddingTextField {
    /**
     * Defines quantization that will be performed on embeddings that are already in the database
     * @param targetEmbeddingQuantization Desired target quantization type
     */
    targetQuantization(targetEmbeddingQuantization: VectorEmbeddingType): IVectorEmbeddingTextField;
    /**
     * Defines which task will be used to get embeddings from
     * @param embeddingsGenerationTaskIdentifier Task identifier
     */
    usingTask(embeddingsGenerationTaskIdentifier: string): IVectorEmbeddingTextField;
}
export interface IVectorEmbeddingField {
    /**
     * Defines quantization that will be performed on embeddings that are already in the database
     * @param targetEmbeddingQuantization Desired target quantization type
     */
    targetQuantization(targetEmbeddingQuantization: VectorEmbeddingType): IVectorEmbeddingField;
}
export interface IVectorField {
}
/**
 * Interface for accessing embedding field factory properties
 * @internal
 */
export interface IVectorEmbeddingFieldFactoryAccessor<T> {
    fieldName: Field<T>;
    sourceQuantizationType: VectorEmbeddingType;
    destinationQuantizationType: VectorEmbeddingType;
    isBase64Encoded: boolean;
    embeddingsGenerationTaskIdentifier: string;
}
/**
 * Represents a RavenDB vector in TypeScript
 */
export interface IRavenVector<T> extends Array<T> {
}
/**
 * Factory for providing text field values for vector searches
 */
export interface IVectorEmbeddingTextFieldValueFactory {
    /**
     * Defines queried text.
     * @param text Queried text
     * @param embeddingsGenerationTaskIdentifier Task identifier for embeddings generation
     */
    byText(text: string, embeddingsGenerationTaskIdentifier?: string): void;
    /**
     * Query by the embedding(s) indexed from the specified document for the quried field.
     * @param documentId The unique identifier of the document to be processed.
     */
    forDocument(documentId: string): void;
    /**
     * Defines queried texts.
     * @param texts Queried texts
     * @param embeddingsGenerationTaskIdentifier Task identifier for embeddings generation
     */
    byTexts(texts: string[], embeddingsGenerationTaskIdentifier?: string): void;
}
export interface IVectorEmbeddingFieldValueFactory {
    /**
     * Defines queried embedding.
     * @param embedding Array containing embedding values
     */
    byEmbedding<T extends number>(embedding: T[]): void;
    /**
     * Defines queried embeddings.
     * @param embeddings Array containing embeddings values
     */
    byEmbeddings<T extends number>(embeddings: T[][]): void;
    /**
     * Defines queried embedding in base64 format.
     * @param base64Embedding Embedding encoded as base64 string
     */
    byBase64(base64Embedding: string): void;
    /**
     * Defines queried embedding.
     * @param embedding RavenVector containing embedding values
     */
    byEmbedding<T extends number>(embedding: {
        "@vector": IRavenVector<T>;
    }): void;
    /**
     * Query by the embedding(s) indexed from the specified document for the queried field.
     * @param documentId The unique identifier of the document to be processed.
     */
    forDocument(documentId: string): void;
}
export interface IVectorFieldValueFactory extends IVectorEmbeddingTextFieldValueFactory, IVectorEmbeddingFieldValueFactory {
}
export interface IVectorFieldValueFactoryAccessor {
    embeddings: number[][];
    embedding: number[];
    text: string;
    texts: string[];
    byId: string;
    embeddingsGenerationTaskIdentifier: string;
}
export declare class VectorEmbeddingFieldValueFactory implements IVectorEmbeddingFieldValueFactory, IVectorFieldValueFactoryAccessor {
    embedding: number[];
    embeddings: number[][];
    text: string;
    texts: string[];
    byId: string;
    embeddingsGenerationTaskIdentifier: string;
    byEmbedding<T extends number>(embedding: T[]): void;
    byEmbedding<T extends number>(embedding: {
        "@vector": IRavenVector<T>;
    }): void;
    /**
     * Defines queried embeddings.
     * @param embeddings Array containing embeddings values
     */
    byEmbeddings<T extends number>(embeddings: T[][]): void;
    /**
     * Defines queried embedding in base64 format.
     * @param base64Embedding Embedding encoded as base64 string
     */
    byBase64(base64Embedding: string): void;
    /**
     * Defines queried text.
     * @param text Queried text
     * @param embeddingsGenerationTaskIdentifier Task identifier for embeddings generation
     */
    byText(text: string, embeddingsGenerationTaskIdentifier?: string): void;
    /**
     * Defines queried texts.
     * @param texts Queried texts
     * @param embeddingsGenerationTaskIdentifier Task identifier for embeddings generation
     */
    byTexts(texts: string[], embeddingsGenerationTaskIdentifier?: string): void;
    forDocument(documentId: string): void;
}
//# sourceMappingURL=VectorFieldFactory.d.ts.map