3fiQ6ddlmZddlZddlZddlmZmZmZmZm Z ddl Z ddl m Z ddlmZddlmZddlmZddlmZej,eZGd d eZy) ) annotationsN)AnyIterableListOptionalTuple)Document) Embeddings) get_from_env) VectorStore)maximal_marginal_relevanceceZdZdZ ddZddZ d ddZ d ddZ d ddZ d ddZ d dd Z d dd Z d dd Z d dd Z e d dd Zy) DashVectorad`DashVector` vector store. To use, you should have the ``dashvector`` python package installed. Example: .. code-block:: python from langchain_community.vectorstores import DashVector from langchain_community.embeddings.openai import OpenAIEmbeddings import dashvector client = dashvector.Client(api_key="***") client.create("langchain", dimension=1024) collection = client.get("langchain") embeddings = OpenAIEmbeddings() vectorstore = DashVector(collection, embeddings.embed_query, "text") c ddl}t||jst dt |||_||_||_y#t$r tdwxYw)z&Initialize with DashVector collection.rN\Could not import dashvector python package. Please install it with `pip install dashvector`.zCcollection should be an instance of dashvector.Collection, bug got ) dashvector ImportError isinstance Collection ValueErrortype _collection _embedding _text_field)self collection embedding text_fieldrs i/var/www/auto_recruiter/arenv/lib/python3.12/site-packages/langchain_community/vectorstores/dashvector.py__init__zDashVector.__init__+s{  *j&;&;< +,.  &#% C  s A Ac:|jj|y)z)Create a Partition in current Collection.N)rcreate_partition)r partitions r_create_partition_if_not_existsz*DashVector._create_partition_if_not_existsEs )))4Nc@|jj||||}|s"td|jjg}|D]S}|j}|j |j } |j} |jt| || fU|S)z;Return docs most similar to query vector, along with scores)topkfilterr#$Fail to query docs by vector, error  page_contentmetadata) rqueryrmessagefieldspoprscoreappendr ) rrkr(r#retdocsdocr,textr1s r'_similarity_search_with_score_by_vectorz2DashVector._similarity_search_with_score_by_vectorIs$$ Af % 6t7G7G7O7O6PQ  QCzzH<< 0 01DIIE KKthGO P  Q  r%c |j||xs5|Dcgc])}ttjj+c}}t |}t dt||D]} t| |zt|} || | } || | } |jjt | } |r|| | }nt | | Dcgc]}i}}t|| D]\}}|||j<t t| | |}|jj||}|rtd|j |Scc}wcc}w)a2Run more texts through the embeddings and add to the vectorstore. Args: texts: Iterable of strings to add to the vectorstore. metadatas: Optional list of metadatas associated with the texts. ids: Optional list of ids associated with the texts. batch_size: Optional batch size to upsert docs. partition: a partition name in collection. [optional]. kwargs: vectorstore specific parameters Returns: List of ids from adding the texts into the vectorstore. rr#z9Fail to upsert docs to dashvector vector database,Error: )r$struuiduuid4hexlistrangelenminrembed_documentsziprrupsertrr.)rtexts metadatasids batch_sizer#kwargs_ text_listiend batch_texts batch_idsbatch_embeddingsbatch_metadatasr,r7r5r4s r add_textszDashVector.add_textscs[, ,,Y7;U;c$**,**+;K q#i.*5 Aa*nc)n5C#Ac*KAc I#>>tK?PQ "+Ac"2/4Q}"=!2"="="%o{"C 2$-1))* 2I'7IJD""))$))DC !kk],) 0 5<#>s .E  Ec Nt|jj||S)zDelete by vector ID. Args: ids: List of ids to delete. partition: a partition name in collection. [optional]. Returns: True if deletion is successful, False otherwise. r:)boolrdelete)rrHr#rJs rrVzDashVector.deletes%D$$++C9+EFFr%c `|j||||}|Dcgc]\}}| c}}Scc}}w)aReturn docs most similar to query. Args: query: Text to search documents similar to. k: Number of documents to return. Default to 4. filter: Doc fields filter conditions that meet the SQL where clause specification. partition: a partition name in collection. [optional]. Returns: List of Documents most similar to the query text. )'similarity_search_with_relevance_scores) rr-r3r(r#rJdocs_and_scoresr6rKs rsimilarity_searchzDashVector.similarity_searchs7*FF 1fi #22Q222 *c b|jj|}|j||||S)aReturn docs most similar to query text , alone with relevance scores. Less is more similar, more is more dissimilar. Args: query: input text k: Number of Documents to return. Defaults to 4. filter: Doc fields filter conditions that meet the SQL where clause specification. partition: a partition name in collection. [optional]. Returns: List of Tuples of (doc, similarity_score) )r3r(r#)r embed_queryr8)rr-r3r(r#rJrs rrXz2DashVector.similarity_search_with_relevance_scoress9.OO//6 ;; 6Y<  r%c `|j||||}|Dcgc]\}}| c}}Scc}}w)aReturn docs most similar to embedding vector. Args: embedding: Embedding to look up documents similar to. k: Number of Documents to return. Defaults to 4. filter: Doc fields filter conditions that meet the SQL where clause specification. partition: a partition name in collection. [optional]. Returns: List of Documents most similar to the query vector. )r8) rrr3r(r#rJrYr6rKs rsimilarity_search_by_vectorz&DashVector.similarity_search_by_vectors7(FF q&) #22Q222r[c d|jj|}|j||||||S)aReturn docs selected using the maximal marginal relevance. Maximal marginal relevance optimizes for similarity to query AND diversity among selected documents. Args: query: Text to look up documents similar to. k: Number of Documents to return. Defaults to 4. fetch_k: Number of Documents to fetch to pass to MMR algorithm. lambda_mult: Number between 0 and 1 that determines the degree of diversity among the results with 0 corresponding to maximum diversity and 1 to minimum diversity. Defaults to 0.5. filter: Doc fields filter conditions that meet the SQL where clause specification. partition: a partition name in collection. [optional]. Returns: List of Documents selected by maximal marginal relevance. )rr]'max_marginal_relevance_search_by_vector) rr-r3fetch_k lambda_multr(r#rJrs rmax_marginal_relevance_searchz(DashVector.max_marginal_relevance_searchs8<OO//6 ;; q';   r%c |jj||||d}|s"td|jj|D cgc]} | j} } t t j|| ||} | D cgc]} |j| j} } | Dcgc](}t|j|j|*c}Scc} wcc} wcc}w)aReturn docs selected using the maximal marginal relevance. Maximal marginal relevance optimizes for similarity to query AND diversity among selected documents. Args: embedding: Embedding to look up documents similar to. k: Number of Documents to return. Defaults to 4. fetch_k: Number of Documents to fetch to pass to MMR algorithm. lambda_mult: Number between 0 and 1 that determines the degree of diversity among the results with 0 corresponding to maximum diversity and 1 to minimum diversity. Defaults to 0.5. filter: Doc fields filter conditions that meet the SQL where clause specification. partition: a partition name in collection. [optional]. Returns: List of Documents selected by maximal marginal relevance. T)r'r(r#include_vectorr)r*) rr-rr.vectorr nparrayoutputr/r r0r)rrr3rbrcr(r#rJr4r6candidate_embeddings mmr_selectedrMrGr,s rraz2DashVector.max_marginal_relevance_search_by_vectors@$$  % 6t7G7G7O7O6PQ 7::s ::1 HHY !5{A 4@@aSZZ]))@ @&  (,,t/?/?"@8 T  ; A s C C)-C#c   ddl} |xs tdd}|xstddd }| j|| } | j || j |} | s]t |j|d}| j|| }|r| j |} ntd |jd || ||}|j||| ||S#t$r tdwxYw)aTReturn DashVector VectorStore initialized from texts and embeddings. This is the quick way to get started with dashvector vector store. Example: .. code-block:: python from langchain_community.vectorstores import DashVector from langchain_community.embeddings import OpenAIEmbeddings import dashvector embeddings = OpenAIEmbeddings() dashvector = DashVector.from_documents( docs, embeddings, dashvector_api_key="{DASHVECTOR_API_KEY}" ) rNrdashvector_api_keyDASHVECTOR_API_KEYdashvector_endpointDASHVECTOR_ENDPOINTz#dashvector.cn-hangzhou.aliyuncs.com)default)api_keyendpoint) dimensionz"Fail to create collection. Error: .) rrr ClientrVgetrAr]createrr.rS)clsrFrrGrnrpcollection_namerrIrHrJrdashvector_clientrdimrespdashvector_vector_dbs r from_textszDashVector.from_textsNs+@  0 < "64 2 \ ! !96  '--&1D.    1&**?; i++E!H56C$++Os+KD.22?C  #Edll^ST!UVV":y*E&&uijI##? C  s CC1)rrrr rr;)r#r;returnNone)Nrr) r List[float]r3intr( Optional[str]r#r;rList[Tuple[Document, float]])NNrr)rFz Iterable[str]rGOptional[List[dict]]rHOptional[List[str]]rIrr#r;rJrr List[str])Nrr)rHrr#r;rJrrrU) r-r;r3rr(rr#r;rJrrList[Document]) r-r;r3rr(rr#r;rJrrr) rrr3rr(rr#r;rJrrr)rg?Nrr)r-r;r3rrbrrcfloatr(Optional[dict]r#r;rJrrr)rrr3rrbrrcrr(rr#r;rJrrr)NNN langchainr7rN)rFrrr rGrrnrrprr{r;rr;rIrrHrrJrrr)__name__ __module__ __qualname____doc__r r$r8rSrVrZrXr_rdra classmethodrr%rrrsB$&&& &45 $"      & :+/#'" 11(1! 1  1  11 1hAJ G& G:= GUX G  G$ $" 33 3 3  3  3 3: $"            & > $" 33 3 3  3  3 38 !%"! !  !  !  !  ! ! !  ! L !%"5 5  5  5  5  5 5 5  5 n +/,0-1* #'@$@$@$( @$ * @$ + @$@$@$@$!@$@$ @$@$r%r) __future__rloggingr<typingrrrrrnumpyrhlangchain_core.documentsr langchain_core.embeddingsr langchain_core.utilsr langchain_core.vectorstoresr &langchain_community.vectorstores.utilsr getLoggerrloggerrrr%rrsL" -0-3M   8 $w$w$r%