rag-vector/ai.koog.rag.vector.storage/VectorStorage

VectorStorage

interface VectorStorage<Document, in Request : SearchRequest> : WriteStorage<Document> , LookupStorage<Document> , SearchStorage<Document, Request> , DeletionStorage(source)

Interface for a vector storage that combines document ingestion and retrieval.

This is the primary user-facing abstraction for working with vector-based document storage. Implementations handle embedding documents into vectors and storing them for similarity-based retrieval.

Type Parameters

Document

The type representing the document being stored.

Request

The type of search requests accepted by this storage.

Inheritors

EmbeddingStorage

Functions

add

abstract suspend fun add(documents: List<Document>, namespace: String?): List<String>

delete

abstract suspend fun delete(ids: List<String>, namespace: String?): List<String>

get

abstract suspend fun get(ids: List<String>, namespace: String?): List<Document>

rankDocuments

open fun ~~rankDocuments~~(query: String): Flow<SearchResult<Document>>

abstract suspend fun search(request: Request, namespace: String?): List<SearchResult<Document>>

update

abstract suspend fun update(documents: Map<String, Document>, namespace: String?): List<String>