nat.retriever.milvus.retriever#

Attributes#

logger

Exceptions#

CollectionNotFoundError

Common base class for all non-exit exceptions.

Classes#

MilvusRetriever

Client for retrieving document chunks from a Milvus vectorstore

Functions#

`_wrap_milvus_results`(res, content_field)
`_wrap_milvus_single_results`(...)

Module Contents#

logger#

exception CollectionNotFoundError#

Bases: nat.retriever.models.RetrieverError

Common base class for all non-exit exceptions.

Initialize self. See help(type(self)) for accurate signature.

class MilvusRetriever( client: MilvusClient | AsyncMilvusClient, embedder: langchain_core.embeddings.Embeddings, content_field: str = 'text', use_iterator: bool = False, )#

Bases: nat.retriever.interface.Retriever

Client for retrieving document chunks from a Milvus vectorstore

Initialize the Milvus Retriever using a preconfigured MilvusClient or AsyncMilvusClient

Args:

_client: pymilvus.MilvusClient | pymilvus.AsyncMilvusClient#

_embedder#

_is_async#

_search_func#

_default_params = None#

_bound_params = []#

content_field = 'text'#

bind(\*\*kwargs) → None#

Bind default values to the search method. Cannot bind the ‘query’ parameter.

Args:: kwargs (dict): Key value pairs corresponding to the default values of search parameters.

get_unbound_params() → list[str]#: Returns a list of unbound parameters which will need to be passed to the search function.

async _validate_collection(collection_name: str) → bool#: Validate that a collection exists.

async search(query: str, \*\*kwargs)#: Retireve max(top_k) items from the data store based on vector similarity search (implementation dependent).

async _search_with_iterator( query: str, *, collection_name: str, top_k: int, filters: str | None = None, output_fields: list[str] | None = None, search_params: dict | None = None, timeout: float | None = None, vector_field_name: str | None = 'vector', distance_cutoff: float | None = None, \*\*kwargs, )#: Retrieve document chunks from a Milvus vectorstore using a search iterator, allowing for the retrieval of more results.

async _search( query: str, *, collection_name: str, top_k: int, filters: str | None = None, output_fields: list[str] | None = None, search_params: dict | None = None, timeout: float | None = None, vector_field_name: str | None = 'vector', \*\*kwargs, )#: Retrieve document chunks from a Milvus vectorstore

_wrap_milvus_results( res: list[pymilvus.client.abstract.Hit], content_field: str, )#

_wrap_milvus_single_results( res: pymilvus.client.abstract.Hit | dict, content_field: str, ) → nat.retriever.models.Document#