U h@sdddlZddlmZmZmZmZmZddlZddlm Z ddl m Z e e ZGddde e ZdS)N)AnyDictListMappingOptional) Embeddings) BaseModelc@seZdZUdZdZeed<dZeed<dZeed<dZ eed <d Z e e ed <d Z e eed <d Ze eed <d Ze e ed<d Ze e ed<d Ze e ed<d Ze e ed<d Ze eed<d Ze eed<d Ze eeed<d Ze eed<d Ze e ed<d Ze eed<dZeed<d Ze eed<ee ee!fdddZ"d Z#e eed<ee$ee!fddd Z%Gd!d"d"Z&eeed#d$d%Z'eeeeed#d&d'Z(eeeeed(d)d*Z)eeed+d,d-Z*d S).OllamaEmbeddingsaOllama locally runs large language models. To use, follow the instructions at https://ollama.ai/. Example: .. code-block:: python from langchain_community.embeddings import OllamaEmbeddings ollama_emb = OllamaEmbeddings( model="llama:7b", ) r1 = ollama_emb.embed_documents( [ "Alpha is the first letter of Greek alphabet", "Beta is the second letter of Greek alphabet", ] ) r2 = ollama_emb.embed_query( "What is the second letter of Greek alphabet" ) zhttp://localhost:11434base_urlZllama2modelz passage: embed_instructionzquery: query_instructionNmirostat mirostat_eta mirostat_taunum_ctxnum_gpu num_thread repeat_last_nrepeat_penalty temperaturestoptfs_ztop_ktop_pF show_progressheaders)returncCsB|j|j|j|j|j|j|j|j|j|j |j |j |j |j d dS)z.Get the default parameters for calling Ollama.) rrrrrrrrrrrrr)r options)r rrrrrrrrrrrrrselfr!I/tmp/pip-unpacked-wheel-9gdii04g/langchain_community/embeddings/ollama.py_default_paramsrs z OllamaEmbeddings._default_params model_kwargscCsd|ji|jS)zGet the identifying parameters.r )r r#rr!r!r"_identifying_paramssz$OllamaEmbeddings._identifying_paramsc@seZdZdZdS)zOllamaEmbeddings.ConfigZforbidN)__name__ __module__ __qualname__extrar!r!r!r"Configsr*)inputrc Csddi|jpi}z*tj|jd||j|d|jd}Wn4tjjk rp}ztd|W5d}~XYnX|j dkrtd |j |j fz| }|d WStjj k r}ztd |d |j W5d}~XYnXdS) zProcess a response from the API. Args: response: The response from the API. Returns: The response as a dictionary. z Content-Typezapplication/jsonz/api/embeddings)r prompt)rjsonz$Error raised by inference endpoint: Nz/Error raised by inference API HTTP code: %s, %s embeddingzError raised by inference API: z . Response: ) rrequestspostr r r# exceptionsRequestException ValueError status_codetextr-JSONDecodeError)r r+rresetr!r!r"_process_emb_responses4       z&OllamaEmbeddings._process_emb_responsecs^jrHzddlm}||dd}WqLtk rDtd|}YqLXn|}fdd|DS)Nr)tqdmr )desczgUnable to show progress bar because tqdm could not be imported. Please install with `pip install tqdm`.csg|]}|qSr!)r;).0r,rr!r" sz+OllamaEmbeddings._embed..)rr< ImportErrorloggerwarning)r r+r<Ziter_r!rr"_embeds  zOllamaEmbeddings._embed)textsrcs fdd|D}|}|S)zEmbed documents using an Ollama deployed embedding model. Args: texts: The list of texts to embed. Returns: List of embeddings, one for each text. csg|]}j|qSr!)r )r>r6rr!r"r?sz4OllamaEmbeddings.embed_documents..)rC)r rDZinstruction_pairsZ embeddingsr!rr"embed_documentss  z OllamaEmbeddings.embed_documents)r6rcCs"|j|}||gd}|S)zEmbed a query using a Ollama deployed embedding model. Args: text: The text to embed. Returns: Embeddings for the text. r)r rC)r r6Zinstruction_pairr/r!r!r" embed_querys zOllamaEmbeddings.embed_query)+r&r'r(__doc__r str__annotations__r r r rrintrfloatrrrrrrrrrrrrrboolrdictpropertyrrr#r$rr%r*r;rCrErFr!r!r!r"r s<      $ r )loggingtypingrrrrrr0Zlangchain_core.embeddingsrZlangchain_core.pydantic_v1r getLoggerr&rAr r!r!r!r"s