Vector

OpenSearch implementation of vector search and CRUD capability.

OpenSearchVectorCapability(index_name, em_invoker, client, opensearch_url=None, query_field='text', vector_query_field='vector', retrieval_strategy=None, distance_strategy=None, connection_params=None, encryption=None, default_batch_size=None)

Bases: BaseVectorCapability, ElasticLikeCore

OpenSearch implementation of VectorCapability protocol.

This class provides document CRUD operations and vector search using OpenSearch. Uses LangChain's OpenSearchVectorSearch for create and retrieve operations, and direct OpenSearch client for update and delete operations.

Attributes:

Name	Type	Description
index_name	str	The name of the OpenSearch index.
vector_store	OpenSearchVectorSearch	The vector store instance.
client	AsyncOpenSearch	AsyncOpenSearch client for direct operations.
em_invoker	BaseEMInvoker	The embedding model to perform vectorization.

Initialize the OpenSearch vector capability.

OpenSearchVectorSearch creates its own sync and async clients internally based on the provided connection parameters. The async client is used for operations like update, delete, and clear.

Parameters:

Name	Type	Description	Default
index_name	str	The name of the OpenSearch index.	required
em_invoker	BaseEMInvoker	The embedding model to perform vectorization.	required
client	AsyncOpenSearch	The OpenSearch client for direct operations.	required
opensearch_url	str | None	The URL of the OpenSearch server. Used for LangChain's OpenSearchVectorSearch initialization. If None, will be extracted from client connection info. Defaults to None.	None
query_field	str	The field name for text queries. Defaults to "text".	'text'
vector_query_field	str	The field name for vector queries. Defaults to "vector".	'vector'
retrieval_strategy	Any	Not used with OpenSearchVectorSearch (kept for API compatibility).	None
distance_strategy	str | None	The distance strategy for retrieval. For example, "l2" for Euclidean distance, "l2squared" for squared Euclidean distance, "cosine" for cosine similarity, etc. Defaults to None.	None
connection_params	dict[str, Any] | None	Additional connection parameters to override defaults. These will be merged with automatically detected parameters (authentication, SSL settings). User-provided params take precedence. Defaults to None. Available parameters include: 1. http_auth (tuple[str, str] | None): HTTP authentication tuple (username, password). 2. use_ssl (bool): Whether to use SSL/TLS. Defaults to True for HTTPS URLs. 3. verify_certs (bool): Whether to verify SSL certificates. Defaults to True for HTTPS URLs. 4. ssl_show_warn (bool): Whether to show SSL warnings. Defaults to True for HTTPS URLs. 5. ssl_assert_hostname (str | None): SSL hostname assertion. Defaults to None. 6. max_retries (int): Maximum number of retries for requests. Defaults to 3. 7. retry_on_timeout (bool): Whether to retry on timeouts. Defaults to True. 8. client_cert (str | None): Path to the client certificate file. Defaults to None. 9. client_key (str | None): Path to the client private key file. Defaults to None. 10. root_cert (str | None): Path to the root certificate file. Defaults to None. 11. Additional kwargs: Any other parameters accepted by OpenSearch client constructor.	None
encryption	EncryptionCapability | None	Encryption capability. Defaults to None.	None
default_batch_size	int | None	Default batch size. Defaults to None.	None

delete_by_id(id, **kwargs) async

Delete records from the data store based on IDs.

Parameters:

Name	Type	Description	Default
id	str | list[str]	ID or list of IDs to delete.	required
**kwargs	Any	Additional arguments passed to the delete operation.	{}

ensure_index(mapping=None, index_settings=None, dimension=None, distance_strategy=None, **kwargs) async

Ensure OpenSearch index exists, creating it if necessary.

This method is idempotent - if the index already exists, it will skip creation and return early.

Parameters:

Name	Type	Description	Default
mapping	dict[str, Any] | None	Custom mapping dictionary to use for index creation. If provided, this mapping will be used directly. The mapping should follow OpenSearch mapping format. Defaults to None, in which default mapping will be used.	None
index_settings	dict[str, Any] | None	Custom index settings. These settings will be merged with any default settings. Defaults to None.	None
dimension	int | None	Vector dimension. If not provided and mapping is not provided, will be inferred from em_invoker by generating a test embedding.	None
distance_strategy	str | None	Distance strategy for vector similarity. Supported values: "l2", "l2squared", "cosine", "innerproduct", etc. Only used when building default mapping. Defaults to "l2" if not specified.	None
**kwargs	Any	Additional arguments.	{}

Raises:

Type	Description
ValueError	If mapping is invalid or required parameters are missing.
RuntimeError	If index creation fails due to backend errors.