0.3.0

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

JVMFileDocumentEmbeddingStorage

class JVMFileDocumentEmbeddingStorage(embedder: DocumentEmbedder<Path>, root: Path) : FileDocumentEmbeddingStorage<Path, Path> (source)

A file-system-based storage implementation for managing and embedding documents represented by file paths.

This class extends EmbeddingBasedDocumentStorage and is specialized for JVM-based systems where documents are represented as file paths (Path). It combines a DocumentEmbedder for embedding the file content into vectors and a JVMFileVectorStorage for managing the storage and retrieval of these embeddings along with their associated documents.

The primary responsibility of this class is to facilitate:

Storing and embedding documents housed in a file system using a specified DocumentEmbedder.
Ranking documents based on similarity to query embeddings.
Managing file-based vector storage via JVMFileVectorStorage.

Parameters

embedder

The embedder responsible for generating vector representations of file-based documents.

root

The root directory path used as the base for file-based vector storage.

Constructors

JVMFileDocumentEmbeddingStorage

constructor(embedder: DocumentEmbedder<Path>, root: Path)

Creates an instance of JVMFileDocumentEmbeddingStorage.

Functions

open override fun allDocuments(): Flow<Path>

Retrieves a flow of all documents stored in the system.

allDocumentsWithPayload

open override fun allDocumentsWithPayload(): Flow<DocumentWithPayload<Path, Unit>>

open suspend override fun delete(documentId: String): Boolean

Deletes the document with the specified ID from the storage.

open suspend override fun getPayload(documentId: String)

open override fun rankDocuments(query: String): Flow<RankedDocument<Path>>

Ranks documents based on their similarity to a given query string.

open suspend override fun read(documentId: String): Path?

Reads a document by its unique identifier.

readWithPayload

open suspend override fun readWithPayload(documentId: String): DocumentWithPayload<Path, Unit>?

open suspend fun store(document: Path): String

open suspend override fun store(document: Path, data: Unit): String

Stores the given document after embedding it into a vector representation.