ValkeyVectorStore(
self,
valkey_url: str,
index_name: str,
embedding: Embeddings,
vector_schema| Name | Type | Description |
|---|---|---|
valkey_url* | str | Connection URL for Valkey server. |
index_name* | str | Name of the index. |
embedding* | Embeddings | Embeddings object. |
vector_schema | Optional[Dict[str, Union[str, int]]] | Default: None |
relevance_score_fn | Optional[Callable[[float], float]] | Default: None |
key_prefix | Optional[str] | Default: None |
cluster_mode | Optional[bool] | Default: None |
**kwargs | Any | Default: {} |
Valkey vector database.
To use, you should have the valkey-glide-sync python package installed:
pip install langchain-aws[valkey]Connection URL schemas:
Examples:
Initialize and load documents:
from langchain_aws.vectorstores import ValkeyVectorStore
from langchain_aws.embeddings import BedrockEmbeddings
embeddings = BedrockEmbeddings()
vds = ValkeyVectorStore.from_documents(
documents,
embeddings,
valkey_url="valkey://cluster_endpoint:6379",
)Initialize with texts and metadata:
vds = ValkeyVectorStore.from_texts(
texts,
metadata,
embeddings,
valkey_url="valkey://cluster_endpoint:6379",
)Vector schema configuration.
Function to compute relevance score.
Prefix for document keys.
If True, create cluster client. If False, create standalone client. If None (default), try cluster first and fall back to standalone.
Additional arguments to pass to GLIDE client.