ZepVectorStore#

class llama_index.vector_stores.ZepVectorStore(collection_name: str, api_url: str, api_key: Optional[str] = None, collection_description: Optional[str] = None, collection_metadata: Optional[Dict[str, Any]] = None, embedding_dimensions: Optional[int] = None, is_auto_embedded: bool = False, **kwargs: Any)#

Bases: VectorStore

Zep Vector Store for storing and retrieving embeddings.

Zep supports both normalized and non-normalized embeddings. Cosine similarity is used to compute distance and the returned score is normalized to be between 0 and 1.

Parameters

  • collection_name (str) – Name of the Zep collection in which to store embeddings.

  • api_url (str) – URL of the Zep API.

  • api_key (str, optional) – Key for the Zep API. Defaults to None.

  • collection_description (str, optional) – Description of the collection. Defaults to None.

  • collection_metadata (dict, optional) – Metadata of the collection. Defaults to None.

  • embedding_dimensions (int, optional) – Dimensions of the embeddings. Defaults to None.

  • is_auto_embedded (bool, optional) – Whether the embeddings are auto-embedded. Defaults to False.

Attributes Summary

client

Get client.

flat_metadata

stores_text

Methods Summary

add(nodes, **add_kwargs)

Add nodes to the collection.

adelete([ref_doc_id])

Asynchronously delete a document from the collection.

aquery(query, **kwargs)

Asynchronously query the index for the top k most similar nodes to the

async_add(nodes, **add_kwargs)

Asynchronously add nodes to the collection.

delete([ref_doc_id])

Delete a document from the collection.

query(query, **kwargs)

Query the index for the top k most similar nodes to the given query.

Attributes Documentation

client#

Get client.

flat_metadata = False#
stores_text: bool = True#

Methods Documentation

add(nodes: List[BaseNode], **add_kwargs: Any) → List[str]#

Add nodes to the collection.

Parameters

nodes (List[BaseNode]) – List of nodes with embeddings.

Returns

List of IDs of the added documents.

Return type

List[str]

async adelete(ref_doc_id: Optional[str] = None, **delete_kwargs: Any) → None#

Asynchronously delete a document from the collection.

Parameters

  • ref_doc_id (Optional[str]) – ID of the document to delete. Not currently supported.

  • delete_kwargs – Must contain “uuid” key with UUID of the document to delete.

async aquery(query: VectorStoreQuery, **kwargs: Any) → VectorStoreQueryResult#
Asynchronously query the index for the top k most similar nodes to the

given query.

Parameters

query (VectorStoreQuery) – Query object containing either a query string or a query embedding.

Returns

Result of the query, containing the most similar

nodes, their similarities, and their IDs.

Return type

VectorStoreQueryResult

async async_add(nodes: List[BaseNode], **add_kwargs: Any) → List[str]#

Asynchronously add nodes to the collection.

Parameters

nodes (List[BaseNode]) – List of nodes with embeddings.

Returns

List of IDs of the added documents.

Return type

List[str]

delete(ref_doc_id: Optional[str] = None, **delete_kwargs: Any) → None#

Delete a document from the collection.

Parameters

  • ref_doc_id (Optional[str]) – ID of the document to delete. Not currently supported.

  • delete_kwargs – Must contain “uuid” key with UUID of the document to delete.

query(query: VectorStoreQuery, **kwargs: Any) → VectorStoreQueryResult#

Query the index for the top k most similar nodes to the given query.

Parameters

query (VectorStoreQuery) – Query object containing either a query string or a query embedding.

Returns

Result of the query, containing the most similar

nodes, their similarities, and their IDs.

Return type

VectorStoreQueryResult