dhSSKJr SSKJrJrJrJrJr SSKr SSK J r SSK J r SSKJr SSKJr SSKJr SS KJr SS KJr "S S \\5r"S S\5rg))Enum)AnyDictListOptionalUnionN)CallbackManagerForRetrieverRun)Document) Embeddings) BaseRetriever) get_fields) ConfigDict)maximal_marginal_relevancec \rSrSrSrSrSrSrg) SearchTypez-Enumerator of the types of search to perform. similaritymmrN)__name__ __module__ __qualname____firstlineno____doc__rr__static_attributes__r_/var/www/html/shao/venv/lib/python3.13/site-packages/langchain_community/retrievers/docarray.pyrrs7J Crrc t\rSrSr%SrSr\\S'\\S'\ \S'\ \S'\ Rr \ \S'S r \\S 'Sr\\\S '\"S S 9rS\ S\S\\4SjrS\R0S \S\\\\ \4\44SjrS\R0S\\4SjrS\R0S\\4SjrS\\\ \4\4S\4SjrSrg)DocArrayRetrievera`DocArray Document Indices` retriever. Currently, it supports 5 backends: InMemoryExactNNIndex, HnswDocumentIndex, QdrantDocumentIndex, ElasticDocIndex, and WeaviateDocumentIndex. Args: index: One of the above-mentioned index instances embeddings: Embedding model to represent text as vectors search_field: Field to consider for searching in the documents. Should be an embedding/vector/tensor. content_field: Field that represents the main content in your document schema. Will be used as a `page_content`. Everything else will go into `metadata`. search_type: Type of search to perform (similarity / mmr) filters: Filters applied for document retrieval. top_k: Number of documents to return Nindex embeddings search_field content_field search_typetop_kfiltersT)arbitrary_types_allowedquery run_managerreturncV[R"URRU55nUR[ R :XaURU5nU$UR[ R:XaURU5nU$[SURS35e)zGet documents relevant for a query. Args: query: string to find relevant documents for Returns: List of relevant documents z Search type z5 does not exist. Choose either 'similarity' or 'mmr'.) nparrayr" embed_queryr%rr_similarity_searchr _mmr_search ValueError)selfr*r+ query_embresultss r_get_relevant_documents)DocArrayRetriever._get_relevant_documents5sHHT__88?@   z44 4--i8G   /&&y1G t//0178 rr5cpSSKJnJn 0nURn[ UR U5(aUR US'SnO:[ UR U5(aUR US'OUR US'UR (aUR R5RXS9R"S 0UD6RUS9nUR RU5n[US 5(a URnUS UnU$UR RXUS 9RnU$) z Perform a search using the query embedding and return top_k documents. Args: query_emb: Query represented as an embedding top_k: Number of documents to return Returns: A list of top_k documents matching the query r)ElasticDocIndexWeaviateDocumentIndex where_filterr* filter_query)r*r#)limit documentsN)r*r#r?r)docarray.indexr:r;r# isinstancer!r( build_queryfindfilterbuild execute_queryhasattrr@) r4r5r'r:r; filter_argsr#r*docss r_searchDocArrayRetriever._searchQs, J (( djj"7 8 8*.,,K 'L  O 4 4#'< content_field metadata <-> all other fields excluding tensors and embeddings (so float, int, string) Args: doc: DocArray document Returns: Document in langchain format Raises: ValueError: If the document doesn't contain the content field z.Document does not contain the content field - .) page_content) rBrUkeysr r$r3r rVstrintfloatboolmetadata)r4rPfieldslc_docnamevalues rrO,DocArrayRetriever._docarray_to_langchain_docs, *#t44*S/   V +@ASAS@TTUV #t$$//0001  D!+C!6!6CIGC RZZDN2*eDcNC4G.H*X*rr)enumrtypingrrrrrnumpyr.langchain_core.callbacksr langchain_core.documentsr langchain_core.embeddingsr langchain_core.retrieversr langchain_core.utils.pydanticr pydanticr&langchain_community.vectorstores.utilsrr^rrrrrrusB33C-034Mdz zr