jP9ddlZddlZddlZddlZddlmZmZddlmZm Z m Z m Z m Z m Z ddlmZddlmZddlmZddlmZddlmZdd lmZej4eZd Zd Zd Zed ddGddeeZ y)N)ThreadPoolExecutorwait)AnyDictListLiteralOptionalTuple) deprecated) Embeddings)create_base_retry_decorator)pre_init)_VertexAICommon)raise_vertex_import_errori Nz0.0.12z1.0z,langchain_google_vertexai.VertexAIEmbeddings)sinceremovalalternative_importceZdZUdZiZeeefed<dZ e ed< e dedefdZ dded e ed ed ed ed e edeffd ZededeefdZedeededeeefdZ ddeede edeeefdZ ddeede edeeeeeeeffdZ ddeedede eddeeefdZ d deededeeefdZdedeefdZxZS)!VertexAIEmbeddingsz'Google Cloud VertexAI embedding models.instanceFshow_progress_barvaluesreturnc|j||ddk(rtjdd|d< ddlm}j|d|d<|S#t $r t Y.wxYw)z8Validates that the python package exists in environment. model_nametextembedding-gecko-defaultzModel_name will become a required arg for VertexAIEmbeddings starting from Feb-01-2024. Currently the default is set to textembedding-gecko@001ztextembedding-gecko@001r)TextEmbeddingModelclient)_try_init_vertexailoggerwarningvertexai.language_modelsr ImportErrorrfrom_pretrained)clsrrs q/var/www/html/Fitness-lenito-AI-main/venv/lib/python3.12/site-packages/langchain_community/embeddings/vertexai.pyvalidate_environmentz'VertexAIEmbeddings.validate_environment$s} v& , #@ @ NN*  $=F<  ( C.==f\>RSx  ( % ' (sAA*)A*rprojectlocationrequest_parallelism max_retries credentialskwargsc &t|d||||||d||jdt|jd<|jd|jd<|jdt |jd<|jd|jd<t j|jd<d|jd<t| |jd <|jjjd  |jd <y )z$Initialize the sentence_transformer.)r*r+r.r,r-rmax_batch_size batch_sizemin_batch_sizemin_good_batch_sizelockFbatch_size_validated) max_workers task_executorz/textembedding-gecko@001embeddings_task_type_supportedN) super__init__get_MAX_BATCH_SIZEr_MIN_BATCH_SIZE threadingLockrr _endpoint_nameendswith) selfrr*r+r,r-r.r/ __class__s r(r<zVertexAIEmbeddings.__init__6s  # 3#!   +1**5E*W &'&*mm4D&E l#*0**5E*W &'/3}}=M/N +, ) 0 f05 ,-);+*  o&  **334NO O , textctjdz}d|d}tj||Dcgc]}|s| c}Scc}w)z9Splits a string by punctuation and whitespace characters.z z([z]))string punctuationresplit)rGsplit_bypatternsegments r(_split_by_punctuationz(VertexAIEmbeddings._split_by_punctuationXsI%%/xj#')xx'>J'>G''>JJJs AAtextsr2cd}t|}d}g}g}|dk(rgS||kr||}ttj|dz}d} |tkDr*t|dkDr|j ||g}|dz }d} nB||ztkDst||k(rd} n%||dz k(rd} ||z }|j ||dz }| r|j |g}d}||kr|S)zlSplits texts in batches based on current maximum batch size and maximum tokens per request. rFT)lenrrP_MAX_TOKENS_PER_BATCHappend) rQr2 text_index texts_lenbatch_token_lenbatches current_batch current_textcurrent_text_token_cnt end_of_batchs r(_prepare_batchesz#VertexAIEmbeddings._prepare_batches`s!  J #%#% >I9$ ,L&<<\JKaO #!L%(==}%)NN=1!- a # "88;PP}%3# Q.$(L#99$$\2a }- " "#K9$LrFembeddings_typecddlm}m}m}m}||||g}t |j }|dttdtffd } | |S)z1Makes a Vertex AI model request with retry logic.r)AbortedDeadlineExceededResourceExhaustedServiceUnavailable) error_typesr-texts_to_processrcr-jdrddlm}|Dcgc] }||}}n|}jj |}|Dcgc]}|j c}Scc}wcc}w)Nr9r)TextEmbeddingInput)rG task_type)rr$rjr get_embeddingsr)rhrjtrequests embeddingsembsrarDs r(_completion_with_retryzMVertexAIEmbeddings._get_embeddings_with_retry.._completion_with_retrys|4==1Q#RG.-'AI- ,33H=J,67JDDKKJ7 78s A(A-) google.api_core.exceptionsrcrdrerfr r-rstrr) rDrQrarcrdrerferrorsretry_decoratorrqs ` ` r(_get_embeddings_with_retryz-VertexAIEmbeddings._get_embeddings_with_retrysm         6((   8T#Y 83 8  8&e,,rFc ddlm}tj||jd}t |d|jdkrg|fS|jd5|jdrYt |d|jdkr g|fcdddSgtj||jdfcdddS|d}g}d} |j ||} t |}t|jd||jd<|s||jd k(rW||jd<d |jd<||jd k7r,tj||d|jd}n|d d}ddd||fS#|$rMd }t |}||jd k(rt|jd t|d z }|d|}YnwxYw#1swY|fSxYw)aPrepares text batches with one-time validation of batch size. Batch size varies between GCP regions and individual project quotas. # Returns embeddings of the first text batch that went through, # and text batches for the rest of the texts. r)InvalidArgumentr2r4r5r6NFTr3rSr1rT) rrrxrr`rrUrvmaxint) rDrQrarxr[ first_batch first_result had_failurefirst_batch_lens r(_prepare_and_validate_batchesz0VertexAIEmbeddings._prepare_and_validate_batchess3 ?$55 4==.  wqz?dmm,AB Bw;  ]]6 "}}34wqz?dmmL&AAw; # "1BBt}}\: # ""!*KLK @#'#B#B#_$L"+.O36 34o4DMM/ 0 o?O1PP.= l+8< 45#dmm4D&EE0AAo./|1LG "!"+c#hW$$='@"&K&)+&6O&$--8H*II&) &67_q=P9Q'O#..>"?K@ !#hW$$s>1G&$G& G&F!B"G&AG!G& G!!G&&G2embeddings_task_type)RETRIEVAL_QUERYRETRIEVAL_DOCUMENTSEMANTIC_SIMILARITYCLASSIFICATION CLUSTERINGcBt|dk(rgSg}g}|dkDrtj||}n|j||\}}|j |g}|j r ddlm}||d} n|} | D]<} |j|jdj|j| |>t|dkDr t||D]!} |j | j#|S#t$rtjd|} YwxYw)a/Embed a list of strings. Args: texts: List[str] The list of strings to embed. batch_size: [int] The batch size of embeddings to send to the model. If zero, then the largest batch size will be detected dynamically at the first request, starting from 250, down to 5. embeddings_task_type: [str] optional embeddings task type, one of the following RETRIEVAL_QUERY - Text is a query in a search/retrieval setting. RETRIEVAL_DOCUMENT - Text is a document in a search/retrieval setting. SEMANTIC_SIMILARITY - Embeddings will be used for Semantic Textual Similarity (STS). CLASSIFICATION - Embeddings will be used for classification. CLUSTERING - Embeddings will be used for clustering. Returns: List of embeddings, one for each text. r)tqdmr)desczgUnable to show progress bar because tqdm could not be imported. Please install with `pip install tqdm`.r8)rQra)rUrr`rextendrrr%r"r#rWrsubmitrvrresult) rDrQr2rrofirst_batch_resultr[tasksriter_batchrms r(embedzVertexAIEmbeddings.embeds4F u:?I(* 02 >(99%LG+/*L*L++ '  ,-  ! ! %W+?@EE LL o.5533$86  u:> KA   ahhj )) >   s&C;; DDc(|j||dS)aEmbed a list of documents. Args: texts: List[str] The list of texts to embed. batch_size: [int] The batch size of embeddings to send to the model. If zero, then the largest batch size will be detected dynamically at the first request, starting from 250, down to 5. Returns: List of embeddings, one for each text. rr)rDrQr2s r(embed_documentsz"VertexAIEmbeddings.embed_documentsOszz%-ABBrFc4|j|gdd}|dS)zEmbed a text. Args: text: The text to embed. Returns: Embedding for the text. rTrrr)rDrGros r( embed_queryzVertexAIEmbeddings.embed_query_s#ZZ+<= !}rF)rNz us-central1rN)N)rN)r)__name__ __module__ __qualname____doc__rrrsr__annotations__rboolrr)r rzr< staticmethodrrPr`floatrvr rrrrr __classcell__)rEs@r(rrs@ 2 "Hd38n!#t#J $4(8!%%#$%) P P# P  P ! P Pc] P PDKCKDIKK1S 1s1tDI11hBF$-#Y$-19#$- d5k $-NBFE%#YE%19#E% tDK $tCy/1 2E%T LCyLL'   L d5k L^34C#YC,/C d5k C  U rFr)!loggingrKrIr@concurrent.futuresrrtypingrrrrr r langchain_core._api.deprecationr langchain_core.embeddingsr #langchain_core.language_models.llmsr langchain_core.utilsr!langchain_community.llms.vertexair&langchain_community.utilities.vertexair getLoggerrr"rVr>r?rr:rFr(rs| 7<<60K)=L   8 $  E M*M  MrF