Ng E>ddlmZddlZddlZddlZddlZddlmZmZm Z m Z m Z m Z m Z mZddlZddlmZddlmZddlmZddlmZddlmZdd lmZdd lmZmZerdd l m!Z!ej"e#Z$ddZ%ddZ&edddGddeZ'dS)) annotationsN) TYPE_CHECKINGAnyCallableIterableListOptionalTupleUnion) deprecated)Document) Embeddings) batch_iterate) VectorStore)version)DistanceStrategymaximal_marginal_relevance)IndexreturnrcV ddl}n"#t$r}td|d}~wwxYw|S)Nrz_Could not import pinecone python package. Please install it with `pip install pinecone-client`.)pinecone ImportError)res e/var/www/html/ai-engine/env/lib/python3.11/site-packages/langchain_community/vectorstores/pinecone.py_import_pineconersU  D    Os &!&boolct}|j}tj|tjdkS)Nz 3.0.0.dev)r __version__rparse)rpinecone_client_versions r_is_pinecone_v3r!'s6!!H&2 =0 1 1W];5O5O OOz0.0.18z1.0zlangchain_pinecone.Pinecone)sinceremovalalternative_importcZeZdZdZdejfdGd ZedHdZdIdZ dJdZ dKdLd#Z dMdNd*Z d$ddd+dOd,Z dMdPd.ZdQd0ZedRd3Z dSdTd8Z dSdUd9Ze dVdWd=Ze dXdYdAZe dZd[dBZ d\d]dFZdS)^Pineconez`Pinecone` vector store. To use, you should have the ``pinecone-client`` python package installed. This version of Pinecone is deprecated. Please use `langchain_pinecone.Pinecone` instead. Nindexr embeddingUnion[Embeddings, Callable]text_keystr namespace Optional[str]distance_strategyOptional[DistanceStrategy]c"t}t|tstjdt||jst dt|||_||_ ||_ ||_ ||_ dS)z Initialize with Pinecone client.z`Passing in `embedding` as a Callable is deprecated. Please pass in an Embeddings object instead.z4client should be an instance of pinecone.Index, got N) r isinstancerwarningswarnr ValueErrortype_index _embedding _text_key _namespacer/)selfr(r)r+r-r/rs r__init__zPinecone.__init__9s$%%)Z00  M.   %00 X4PU;;XX  #!#!2r"rOptional[Embeddings]cHt|jtr|jSdS)z/Access the query embedding object if available.N)r2r8rr;s r embeddingszPinecone.embeddingsRs% doz 2 2 #? "tr"texts Iterable[str]List[List[float]]ctjtr'jt |Sfd|DS)zEmbed search docs.c:g|]}|S)r8).0tr;s r z-Pinecone._embed_documents..]s%222q""222r")r2r8rembed_documentslist)r;rAs` r_embed_documentszPinecone._embed_documentsYsN doz 2 2 @?224;;?? ?2222E2222r"text List[float]ct|jtr|j|S||S)zEmbed query text.)r2r8r embed_query)r;rMs r _embed_queryzPinecone._embed_query_s> doz 2 2 5?..t44 4t$$$r"  metadatasOptional[List[dict]]idsOptional[List[str]] batch_sizeintembedding_chunk_sizekwargs List[str]c jt|}|p d|D}|p d|D}t||D]\}} | |j<t dt ||D]w} || | |z} || | |z} || | |z} | }fdt|t| || D}d|Dx|S)aRun more texts through the embeddings and add to the vectorstore. Upsert optimization is done by chunking the embeddings and upserting them. This is done to avoid memory issues and optimize using HTTP based embeddings. For OpenAI embeddings, use pool_threads>4 when constructing the pinecone.Index, embedding_chunk_size>1000 and batch_size~64 for best performance. Args: texts: Iterable of strings to add to the vectorstore. metadatas: Optional list of metadatas associated with the texts. ids: Optional list of ids to associate with the texts. namespace: Optional pinecone namespace to add the texts to. batch_size: Batch size to use when adding the texts to the vectorstore. embedding_chunk_size: Chunk size to use when embedding the texts. Returns: List of ids from adding the texts into the vectorstore. NcNg|]"}ttj#SrF)r,uuiduuid4rG_s rrIz&Pinecone.add_texts..s&777Ac$*,,''777r"cg|]}iSrFrFras rrIz&Pinecone.add_texts..s!4!4!4"!4!4!4r"rc <g|]}jjd|ddS)T)vectorsr- async_reqrF)r7upsert)rGbatchr[r-r;s rrIz&Pinecone.add_texts..sY    # "!'"    r"c6g|]}|SrF)get)rGress rrIz&Pinecone.add_texts..s , , ,3SWWYY , , ,r")r:rKzipr9rangelenrLr)r;rArTrVr-rXrZr[metadatarMi chunk_texts chunk_idschunk_metadatasr@ async_ress` ` ` r add_textszPinecone.add_textses`8  IU 7777774!4!4e!4!4!4 !)U33 , ,NHd'+HT^ $ $ q#e**&:;; - -AA(<$< <=KA$8 889I'A0D,D(DEO..{;;J      +Iz? K K   I - ,) , , , , , r"querykfilterOptional[dict]List[Tuple[Document, float]]cZ||||||S)aReturn pinecone documents most similar to query, along with scores. Args: query: Text to look up documents similar to. k: Number of Documents to return. Defaults to 4. filter: Dictionary of argument(s) to filter on metadata namespace: Namespace to search in. Default will search in '' namespace. Returns: List of Documents most similar to the query and score for each rxryr-)&similarity_search_by_vector_with_scorerQ)r;rwrxryr-s rsimilarity_search_with_scorez%Pinecone.similarity_search_with_scores8$::   e $ $&I;   r"r}cl||j}g}|j|g|d||}|dD]}|d}|j|vrI||j} |d} |t | || f\td|jd |S) zGReturn pinecone documents most similar to embedding, along with scores.NT)vectortop_kinclude_metadatar-rymatchesroscore page_contentrozFound document with no `z` key. Skipping.) r:r7rwr9popappendr loggerwarning) r;r)rxryr-docsresultsrkrorMrs rr~z/Pinecone.similarity_search_by_vector_with_scores  I+##;! $  9%  C:H~))||DN33G  X4(KKKUSTTTTOt~OOO r"List[Document]c >|j|f|||d|}d|DS)aReturn pinecone documents most similar to query. Args: query: Text to look up documents similar to. k: Number of Documents to return. Defaults to 4. filter: Dictionary of argument(s) to filter on metadata namespace: Namespace to search in. Default will search in '' namespace. Returns: List of Documents most similar to the query and score for each r}cg|]\}}|SrFrF)rGdocrbs rrIz.Pinecone.similarity_search..s222Q222r")r)r;rwrxryr-r[docs_and_scoress rsimilarity_searchzPinecone.similarity_searchsJ&<$;  v  >D  32/2222r"Callable[[float], float]c|jtjkr|jS|jtjkr|jS|jtjkr|jStd)a8 The 'correct' relevance function may differ depending on a few things, including: - the distance / similarity metric used by the VectorStore - the scale of your embeddings (OpenAI's are unit normed. Many others are not!) - embedding dimensionality - etc. zXUnknown distance strategy, must be cosine, max_inner_product (dot product), or euclidean) r/rCOSINE_cosine_relevance_score_fnMAX_INNER_PRODUCT%_max_inner_product_relevance_score_fnEUCLIDEAN_DISTANCE_euclidean_relevance_score_fnr5r?s r_select_relevance_score_fnz#Pinecone._select_relevance_score_fnsl  !%5%< < <2 2  #'7'I I I= =  #'7'J J J5 5. r"rfloatc|dzdz S)z8Pinecone returns cosine similarity scores between [-1,1]rF)rs rrz#Pinecone._cosine_relevance_score_fns Qr"?fetch_k lambda_multc  |j}j|g|dd|| tt j|gtjd dD||} fd|D} fd | DS) 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. Returns: List of Documents selected by maximal marginal relevance. NT)rrinclude_valuesrr-ry)dtypecg|] }|d S)valuesrF)rGitems rrIzDPinecone.max_marginal_relevance_search_by_vector..0s ; ; ;T(^ ; ; ;r"r)rxrc8g|]}d|dS)rrorF)rGrprs rrIzDPinecone.max_marginal_relevance_search_by_vector..4s(LLL!GI&q)*5LLLr"cbg|]+}t|j|,S)r)r rr9)rGror;s rrIzDPinecone.max_marginal_relevance_search_by_vector..5sD    (,,"@"@8 T T T   r")r:r7rwrnparrayfloat32) r;r)rxrrryr-r[ mmr_selectedselectedrs ` @r'max_marginal_relevance_search_by_vectorz0Pinecone.max_marginal_relevance_search_by_vector s4  I+##;! $  2 Hi[ 3 3 3 ; ; (: ; ; ;#    MLLL|LLL    $    r"c `||}|||||||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. Returns: List of Documents selected by maximal marginal relevance. )rQr) r;rwrxrrryr-r[r)s rmax_marginal_relevance_searchz&Pinecone.max_marginal_relevance_search:s<4%%e,, ;; q';    r" index_name pool_threadsrc8t}tr`|tjd|}|}d|jdD}n|}||vr;tr||n|||}nJt|dkrtdtd|d d ||S) zReturn a Pinecone Index instance. Args: index_name: Name of the index to use. pool_threads: Number of threads to use for index upsert. Returns: Pinecone Index instance.PINECONE_API_KEY)api_keyrcg|] }|j SrF)name)rGrps rrIz/Pinecone.get_pinecone_index..nsIIIa16IIIr"indexes)rrzNo active indexes found in your Pinecone project, are you sure you're using the right Pinecone API key and Environment? Please double check your Pinecone dashboard.zIndex 'zQ' not found in your Pinecone project. Did you mean one of the following indexes: z, ) rr!r'osenvironrj list_indexes index_listrrnr5join)clsrrrpinecone_instancer index_namesr(s rget_pinecone_indexzPinecone.get_pinecone_indexYsM$%%    2 ( 1 1 '9::!2!! (4466GII7+=i+HIIIKK"//11K  $ $#$$K!'' 333^^J\^JJ E    " "?  W*WW>Bii >T>TWW  r"r upsert_kwargsembeddings_chunk_sizec v||| } || |||fi| }|j|f||||| d| pi|S)a DEPRECATED: use langchain_pinecone.PineconeVectorStore.from_texts instead: Construct Pinecone wrapper from raw documents. This is a user friendly interface that: 1. Embeds documents. 2. Adds the documents to a provided Pinecone index This is intended to be a quick way to get started. The `pool_threads` affects the speed of the upsert operations. Example: .. code-block:: python from langchain_pinecone import PineconeVectorStore from langchain_openai import OpenAIEmbeddings embeddings = OpenAIEmbeddings() index_name = "my-index" namespace = "my-namespace" vectorstore = Pinecone( index_name=index_name, embedding=embedding, namespace=namespace, ) )rTrVr-rXrZ)rru)rrAr)rTrVrXr+r-rrrrr[pinecone_indexrs r from_textszPinecone.from_textssV// LII3~y(IPPPP  !!6   "   r"cJ|||}|||||S)z*Load pinecone vectorstore from index name.)r)rrr)r+r-rrs rfrom_existing_indexzPinecone.from_existing_indexs/// LIIs>9h BBBr" delete_allOptional[bool]Nonec &||j}|r|jjd d|d|no|Fd}tdt ||D]$}||||z}|jjd ||d|%n'||jjd ||d|nt ddS) zDelete by vector IDs or filter. Args: ids: List of ids to delete. filter: Dictionary of conditions to filter vectors to delete. NT)rr-rSr)rVr-)ryr-z3Either ids, delete_all, or filter must be provided.rF)r:r7deletermrnr5) r;rVrr-ryr[ chunk_sizerpchunks rrzPinecone.deletes  I  T DK  N$) N Nv N N N N _J1c#hh 33 M MAJ./" "Lu LLVLLLL M  DK  Lf L LV L L L LRSS Str") r(rr)r*r+r,r-r.r/r0)rr=)rArBrrC)rMr,rrN)NNNrRrS)rArBrTrUrVrWr-r.rXrYrZrYr[rrr\)rvNN) rwr,rxrYryrzr-r.rr{) r)rNrxrYryrzr-r.rr{) rwr,rxrYryrzr-r.r[rrr)rr)rrrr)rvrrNN)r)rNrxrYrrYrrryrzr-r.r[rrr)rwr,rxrYrrYrrryrzr-r.r[rrr)rv)rr.rrYrr) NNrRrMNNNrvrS)rAr\r)rrTrUrVrWrXrYr+r,r-r.rr.rrzrrYrrYr[rrr')rMNrv) rr,r)rr+r,r-r.rrYrr')NNNN) rVrWrrr-r.ryrzr[rrr)__name__ __module__ __qualname____doc__rrr<propertyr@rLrQrurr~rr staticmethodrrr classmethodrrrrrFr"rr'r'-sz$(8H8O 333332X 3333 %%%%+/#'#'$(:::::~!%#'      4!%#' F!%#' 333330,\ !%#'. . . . . f !%#'     >))))[)V +/#'#'$((,%)6666[6p #' C C C C[ C$(%)#'!% r"r')rr)rr)( __future__rloggingrr_r3typingrrrrrr r r numpyrlangchain_core._api.deprecationr langchain_core.documentsr langchain_core.embeddingsrlangchain_core.utils.iterrlangchain_core.vectorstoresr packagingr&langchain_community.vectorstores.utilsrrrr getLoggerrrrr!r'rFr"rrs"""""" WWWWWWWWWWWWWWWWWWWW666666------000000333333333333   8 $ $PPPP   E6Sxxxxx{xxxxxr"