Q iDddlZddlmZmZmZmZmZmZddlm Z ddl m Z m Z m Z ddlmZerddlmZddlmZGdd eZdS) N) TYPE_CHECKINGAnyDictListOptionalUnion) ConfigDict)retrystop_after_attemptwait_random_exponential)retry_if_not_exception_type)EmbeddingsCache)BaseVectorizerc .eZdZdZedZ d(dedeed ed ed ffd Z e d e fdZ ded e eeeffdZded e eeeffdZdeefdZdeefdZd efdZd efdZdeedeedee fdZded eefdZeedded ee! d)deed"eed eeefd#Z ded eefd$Z!eedded ee! d)deed"eed eeefd%Z"ded e eefffd& Z#e d efd'Z$xZ%S)*VoyageAIVectorizera The VoyageAIVectorizer class utilizes VoyageAI's API to generate embeddings for text and multimodal (text / image / video) data. This vectorizer is designed to interact with VoyageAI's /embed and /multimodal_embed APIs, requiring an API key for authentication. The key can be provided directly in the `api_config` dictionary or through the `VOYAGE_API_KEY` environment variable. User must obtain an API key from VoyageAI's website (https://dash.voyageai.com/). Additionally, the `voyageai` python client must be installed with `pip install voyageai`. For image embeddings, the Pillow library must also be installed with `pip install pillow`. The vectorizer supports both synchronous and asynchronous operations, allows for batch processing of content and flexibility in handling preprocessing tasks. You can optionally enable caching to improve performance when generating embeddings for repeated text inputs. .. code-block:: python from redisvl.utils.vectorize import VoyageAIVectorizer # Basic usage vectorizer = VoyageAIVectorizer( model="voyage-3-large", api_config={"api_key": "your-voyageai-api-key"} # OR set VOYAGE_API_KEY in your env ) query_embedding = vectorizer.embed( content="your input query text here", input_type="query" ) doc_embeddings = vectorizer.embed_many( contents=["your document text", "more document text"], input_type="document" ) # Multimodal usage - requires Pillow and voyageai>=0.3.6 vectorizer = VoyageAIVectorizer( model="voyage-multimodal-3.5", api_config={"api_key": "your-voyageai-api-key"} # OR set VOYAGE_API_KEY in your env ) image_embedding = vectorizer.embed_image( "path/to/your/image.jpg", input_type="query" ) video_embedding = vectorizer.embed_video( "path/to/your/video.mp4", input_type="document" ) # With caching enabled from redisvl.extensions.cache.embeddings import EmbeddingsCache cache = EmbeddingsCache(name="voyageai_embeddings_cache") vectorizer = VoyageAIVectorizer( model="voyage-3-large", api_config={"api_key": "your-voyageai-api-key"}, cache=cache ) # First call will compute and cache the embedding embedding1 = vectorizer.embed( content="your input query text here", input_type="query" ) # Second call will retrieve from cache embedding2 = vectorizer.embed( content="your input query text here", input_type="query" ) T)arbitrary_types_allowedvoyage-3-largeNfloat32model api_configdtypecacherc lt||||j|fi|dS)aoInitialize the VoyageAI vectorizer. Visit https://docs.voyageai.com/docs/embeddings to learn about embeddings and check the available models. Args: model (str): Model to use for embedding. Defaults to "voyage-3-large". api_config (Optional[Dict], optional): Dictionary containing the API key. Defaults to None. dtype (str): the default datatype to use when embedding content as byte arrays. Used when setting `as_buffer=True` in calls to embed() and embed_many(). Defaults to 'float32'. cache (Optional[EmbeddingsCache]): Optional EmbeddingsCache instance to cache embeddings for better performance with repeated items. Defaults to None. Raises: ImportError: If the voyageai library is not installed. ValueError: If the API key is not provided. Notes: - Multimodal models require voyageai>=0.3.6 to be installed for video embeddings, as well as ffmpeg installed on the system. Image embeddings require pillow to be installed. )rrrN)super__init___setup)selfrrrrkwargs __class__s C:\Users\Dell Inspiron 16\Desktop\tws\AgrotaPowerBi\back-agrota-powerbi\mcp-client-agrota\venv\Lib\site-packages\redisvl/utils/vectorize/voyageai.pyrzVoyageAIVectorizer.__init__^sD> uE??? J))&)))))returncd|jvS)z/Whether a multimodal model has been configured. multimodalrrs r is_multimodalz VoyageAIVectorizer.is_multimodalstz))r! image_pathc |jstd ddlm}n#t$rt dwxYw|j||fi|S)zhEmbed an image (from its path on disk) using VoyageAI's multimodal API. Requires pillow to be installed.z/Cannot embed image with a non-multimodal model.r)ImagezXPillow library is required for image embedding. Please install with `pip install pillow`)r' ValueErrorPILr* ImportErrorembedopen)rr(rr*s r embed_imagezVoyageAIVectorizer.embed_images! PNOO O  ! ! ! ! ! ! !   ;   tz%**Z00;;F;;;9 video_pathc |jstd ddlm}n#t$rt dwxYw|||j}|j|fi|S)zEmbed a video (from its path on disk) using VoyageAI's multimodal API. Requires voyageai>=0.3.6 to be installed, as well as ffmpeg to be installed on the system. z/Cannot embed video with a non-multimodal model.rVideobvoyageai>=0.3.6 is required for video embedding. Please install with `pip install voyageai>=0.3.6`r%)r'r+voyageai.video_utilsr5ModuleNotFoundError from_pathrr.)rr2rr5videos r embed_videozVoyageAIVectorizer.embed_videos ! PNOO O  2 2 2 2 2 2 2"   %D    *   tz%**6***r1c |j|fi||jr#|jj|_|jj|_n"|jj|_|jj|_||_ dS)zBSet up the VoyageAI client and determine the embedding dimensions.N) _initialize_clientr'_clientmultimodal_embed _embed_fn_aclient _aembed_fnr._set_model_dimsdims)rrrs r rzVoyageAIVectorizer._setupsu  55f555   2!\:DN"mrA)rrrrFrGrHs r r=z%VoyageAIVectorizer._initialize_clients  J  4 4 4 4 4 4 4 4 4   =  *4 TJNN9 % % %CS9T9T  [  v88g888 # >>G>v>> s)c |dd}t|S#ttf$r$}t dt |d}~wt $r$}t dt |d}~wwxYw)z Determine the dimensionality of the embedding model by making a test call. Returns: int: Dimensionality of the embedding model Raises: ValueError: If embedding dimensions cannot be determined zdimension checkdocument) input_typez+Unexpected response from the VoyageAI API: Nz*Error setting embedding model dimensions: )_embedlenKeyError IndexErrorr+str Exception)r embeddingkees r rCz"VoyageAIVectorizer._set_model_dimss T $5* MMIy>> !*% V V VT3r77TTUU U T T TR#a&&RRSS S Ts %(B A B %BB cH|jdvrdS|jdvrdS|jdvrdSdS)z Determine the appropriate batch size based on the model being used. Returns: int: Recommended batch size for the current model )zvoyage-2z voyage-02H)z voyage-3-litezvoyage-3.5-lite)zvoyage-3z voyage-3.5 r%r&s r _get_batch_sizez"VoyageAIVectorizer._get_batch_sizesB :2 2 22 Z? ? ?2 Z5 5 521r!contentsrQ truncationc.t|tstd|js,|r*t|dtstd||dvrtd|$t|t stddSdS)a+ Validate the inputs to the embedding methods. Args: contents: List of items to embed input_type: Type of input (document or query) truncation: Whether to truncate long inputs Raises: TypeError: If inputs are invalid z[Must pass in a list of str, PIL.Image.Image, or voyageai.video_utils.Video values to embed.rz+Must pass in a list of str values to embed.N)rPqueryznMust pass in a allowed value for voyageai embedding input_type. See https://docs.voyageai.com/docs/embeddings.z*Truncation (optional) parameter is a bool.) isinstancelist TypeErrorr'rVbool)rrarQrbs r _validate_inputz"VoyageAIVectorizer._validate_inputs(D)) m ! Kh Kz(1+s7S7S KIJJ J  !j8M&M&MA   !*Z*F*F !HII I " ! ! !r!contentc 0|j|gfi|}|dS)a" Generate a vector embedding for a single item using the VoyageAI API. Args: content: Item to embed - must be one of str, PIL.Image.Image, or voyageai.video_utils.Video. Images and video require a multimodal model to be configured. **kwargs: Additional parameters to pass to the VoyageAI API Returns: List[float]: Vector embedding as a list of floats Raises: TypeError: If parameters are invalid ValueError: If embedding fails r) _embed_manyrrjrresults r rRzVoyageAIVectorizer._embeds*""!7)66v66ayr!<)minmax)waitstopr batch_sizec ddlm}|dd}|dd}||||||} g}|||D]=}|j|jr|gn|f|j||d|} | | j >|S#|$r%} tdt| | d} ~ wt$r} td | d} ~ wwxYw) a Generate vector embeddings for a batch of items using the VoyageAI API. Args: contents: List of items to embed - each item must be one of str, PIL.Image.Image, or voyageai.video_utils.Video. Images and video require a multimodal model to be configured. batch_size: Number of items to process in each API call **kwargs: Additional parameters to pass to the VoyageAI API Returns: List[List[float]]: List of vector embeddings as lists of floats Raises: TypeError: If `contents` is not a list, or parameters are invalid ValueError: If embedding fails rInvalidRequestErrorrQNrbTrrQrbInvalid input for embedding: Embedding texts failed: )voyageai.errorrypoprir`batchifyr@r'rextend embeddingsrgrVrWr+ rrarvrryrQrbrbatchresponserZs r rlzVoyageAIVectorizer._embed_many3sk0 766666ZZ d33 ZZ d33  Xz:>>>  --//J =!Jx<< 7 7)4>#'#5@5*)) !!("56666 " M M MDCFFDDEE1 L = = =;;;<< < =s%!AB99D> C D+C==Dc@K|j|gfi|d{V}|dS)a1 Asynchronously generate a vector embedding for a single item using the VoyageAI API. Args: content: Item to embed - must be one of str, PIL.Image.Image, or voyageai.video_utils.Video. Images and video require a multimodal model to be configured. **kwargs: Additional parameters to pass to the VoyageAI API Returns: List[float]: Vector embedding as a list of floats Raises: TypeError: If parameters are invalid ValueError: If embedding fails Nr) _aembed_manyrms r _aembedzVoyageAIVectorizer._aembedjs@")t('==f========ayr!cKddlm}|dd}|dd}||||||} g}|||D]C}|j|jr|gn|f|j||d|d{V} | | j D|S#|$r%} tdt| | d} ~ wt$r} td | d} ~ wwxYw) a Asynchronously generate vector embeddings for a batch of items using the VoyageAI API. Args: contents: List of items to embed - each item must be one of str, PIL.Image.Image, or voyageai.video_utils.Video. Images and video require a multimodal model to be configured. batch_size: Number of texts to process in each API call **kwargs: Additional parameters to pass to the VoyageAI API Returns: List[List[float]]: List of vector embeddings as lists of floats Raises: TypeError: If `contents` is not a list, or parameters are invalid ValueError: If embedding fails rrxrQNrbTrzr{r|)r}ryr~rir`rrBr'rrrrgrVrWr+rs r rzVoyageAIVectorizer._aembed_many~s0 766666ZZ d33 ZZ d33  Xz:>>>  --//J =!Jx<< 7 7!0#'#5@5"*)) """"!!("56666 " M M MDCFFDDEE1 L = = =;;;<< < =s%#ACD  C&& D 3DD c ddlm}n#t$rtdwxYwt||r|St |S)z&Convert content to a cacheable format.rr4r6)r7r5r8reto_bytesr_serialize_for_cache)rrjr5rs r rz'VoyageAIVectorizer._serialize_for_caches  2 2 2 2 2 2 2"   %D   gu % % &##%% %ww++G444s $cdS)NrKrJr&s r typezVoyageAIVectorizer.typeszr!)rNrN)N)&__name__ __module__ __qualname____doc__r model_configrVrrrpropertyrhr'rrfloatbytesr0r;rr=intrCr`rrirRr r r r rgrlrrrr __classcell__)rs@r rrsHHT:d;;;L&%)-1 !*!*!*TN!* !* )* !*!*!*!*!*!*F*t***X* P8Q < < < <+c+d5k5>P8Q++++, +$ + + + +$?Xd^$?$?$?$?LTTTTT( JS J/7}JJRSW.JJJJ8cU ( U $ $ 3 3 3   " ")))44 @D0=0=S 0=/7}0= d5k 0=0=0=  0=dStE{( U $ $ 3 3 3   " ")))44 @D0=0=S 0=/7}0= d5k 0=0=0=  0=d 5C 5E%*4E 5 5 5 5 5 5cXr!r)rMtypingrrrrrrpydanticr tenacityr r r tenacity.retryr .redisvl.extensions.cache.embeddings.embeddingsrredisvl.utils.vectorize.baserrrJr!r rs BBBBBBBBBBBBBBBBGGGGGGGGGG666666ONNNNNN777777 ttttttttttr!