sOiddlmZmZmZmZddlmZddlmZddl m Z m Z ddl m Z ddlmZddlmZddlmZGd d eZy ) )AnyIterableSequenceType)asdict)DenseModelDescription) NumpyArrayDevice) OnnxProvider)Colbert) JinaColbert) LateInteractionTextEmbeddingBaseceZdZUeegZeeee d<e dee e e ffdZe deefdZdddej$ddfde de dzd edzd eedzd eezd eedzd ede ffd ZedefdZe de defdZ dde ee zdededzde deef dZde ee zde deefdZ dde ee zdededede def dZxZ S)LateInteractionTextEmbeddingEMBEDDINGS_REGISTRYreturncZ|jDcgc] }t|c}Scc}w)a Lists the supported models. Returns: list[dict[str, Any]]: A list of dictionaries containing the model information. Example: ``` [ { "model": "colbert-ir/colbertv2.0", "dim": 128, "description": "Late interaction model", "license": "mit", "size_in_GB": 0.44, "sources": { "hf": "colbert-ir/colbertv2.0", }, "model_file": "model.onnx", }, ] ``` )_list_supported_modelsr)clsmodels /volume1/homes/robertsu/coba/app/.venv/lib/python3.12/site-packages/fastembed/late_interaction/late_interaction_text_embedding.pylist_supported_modelsz2LateInteractionTextEmbedding.list_supported_modelss+2,/+E+E+GH+G%u +GHHHs(cjg}|jD]!}|j|j#|SN)rextendr)rresult embeddings rrz3LateInteractionTextEmbedding._list_supported_models,s1.000I MM)::< =1 NF model_name cache_dirthreads providerscuda device_ids lazy_loadkwargsc t |||fi||jD]=} | j} t fd| Ds(| |f|||||d||_yt dd)Nc3tK|]/}j|jjk(1ywr)lowerr).0rrs r z8LateInteractionTextEmbedding.__init__..As/[JZ:##%):):)<>!,(  !@LM ;,, KM+J<8**58  NsB documents batch_sizeparallelc+^K|jj|||fi|Ed{y7w)a Encode a list of documents into list of embeddings. We use mean pooling with attention so that the model can handle variable-length inputs. Args: documents: Iterator of documents or single document to embed batch_size: Batch size for encoding -- higher values will use more memory, but be faster parallel: If > 1, data-parallel encoding will be used, recommended for offline encoding of large datasets. If 0, use all available cores. If None, don't use data-parallel processing, use default onnxruntime threading instead. Returns: List of embeddings, one per document N)rembed)r0r=r>r?r&s rrAz"LateInteractionTextEmbedding.embedus+,$4::##Iz8NvNNNs #-+-queryc+ZK|jj|fi|Ed{y7w)z Embeds queries Args: query (Union[str, Iterable[str]]): The query to embed, or an iterable e.g. list of queries. Returns: Iterable[NdArray]: The embeddings. N)r query_embed)r0rBr&s rrDz(LateInteractionTextEmbedding.query_embeds'*4::))%:6:::s !+)+textsis_docinclude_extensionc D|jj|f|||d|S)aReturns the number of tokens in the texts. Args: texts (str | Iterable[str]): The list of texts to embed. batch_size (int): Batch size for encoding is_doc (bool): Whether the texts are documents (disable embedding a query with include_mask=True). include_extension (bool): Turn on to count DOC / QUERY marker tokens, and [MASK] token in query mode. Returns: int: Sum of number of tokens in the texts. )r>rFrG)r token_count)r0rEr>rFrGr&s rrIz(LateInteractionTextEmbedding.token_counts8&&tzz%%  !/     r)N)iTF)!__name__ __module__ __qualname__r r rlistrr__annotations__ classmethoddictstrrrrrr AUTOintrr boolr-propertyr7r6rr rArDrI __classcell__)r3s@rrrsIPR]H^d#CDE^Id4S>&:II4t,A'B!%"37$kk'+  : t  L)D0  Vm  I$   @$$$ CC:# O#&OO* O  O *  O0 ;x}!4 ; ;Q[H\ ;""'  Xc]"          rrN)typingrrrr dataclassesr"fastembed.common.model_descriptionrfastembed.common.typesr r fastembed.commonr "fastembed.late_interaction.colbertr 'fastembed.late_interaction.jina_colbertr :fastembed.late_interaction.late_interaction_embedding_baserrrrras100D5)6? f #Cf r