Ng|ddlZddlmZmZddlZddlmZddlmZej e Z GddeeZ dS)N)ListOptional) Embeddings) BaseModelceZdZUdZdZeed< dZee ed< dede e fdZ d e ede e e fd Z dede e fd ZdS) LlamafileEmbeddingsaLlamafile lets you distribute and run large language models with a single file. To get started, see: https://github.com/Mozilla-Ocho/llamafile To use this class, you will need to first: 1. Download a llamafile. 2. Make the downloaded file executable: `chmod +x path/to/model.llamafile` 3. Start the llamafile in server mode with embeddings enabled: `./path/to/model.llamafile --server --nobrowser --embedding` Example: .. code-block:: python from langchain_community.embeddings import LlamafileEmbeddings embedder = LlamafileEmbeddings() doc_embeddings = embedder.embed_documents( [ "Alpha is the first letter of the Greek alphabet", "Beta is the second letter of the Greek alphabet", ] ) query_embedding = embedder.embed_query( "What is the second letter of the Greek alphabet" ) zhttp://localhost:8080base_urlNrequest_timeouttextreturnc tj|jdddid|i|j}n@#tjj$r)tjd|jdwxYw||}d|vrtd |d}t|d krtd |S) Nz /embeddingz Content-Typezapplication/jsoncontent)urlheadersjsontimeoutzTCould not connect to Llamafile server. Please make sure that a server is running at . embeddingzPUnexpected output from /embedding endpoint, output dict missing 'embedding' key.gz^Embedding sums to 0, did you start the llamafile server with the `--embedding` option enabled?) requestspostr r exceptionsConnectionErrorraise_for_statusrKeyErrorsum ValueError)selfr responsecontentsrs d/var/www/html/ai-engine/env/lib/python3.11/site-packages/langchain_community/embeddings/llamafile.py_embedzLlamafileEmbeddings._embed0s( }}000"$6t,   HH"2   %55@/3}@@@   !!###==?? h & &+  [)  y>>S 4  s ),=A)textscdg}|D]*}|||+|S)a6Embed documents using a llamafile server running at `self.base_url`. llamafile server should be started in a separate process before invoking this method. Args: texts: The list of texts to embed. Returns: List of embeddings, one for each text. )appendr!)rr"doc_embeddingsr s r embed_documentsz#LlamafileEmbeddings.embed_documents\sA 5 5D  ! !$++d"3"3 4 4 4 4c,||S)aEmbed a query using a llamafile server running at `self.base_url`. llamafile server should be started in a separate process before invoking this method. Args: text: The text to embed. Returns: Embeddings for the text. )r!)rr s r embed_queryzLlamafileEmbeddings.embed_queryls{{4   r')__name__ __module__ __qualname____doc__r str__annotations__r rintrfloatr!r&r)r'r rr s<,Hc+++;%)OXc])))%*3*4;****XT#Y4U 3D ! !U ! ! ! ! ! !r'r) loggingtypingrrrlangchain_core.embeddingsrpydanticr getLoggerr*loggerrr2r'r r9s!!!!!!!!000000  8 $ $l!l!l!l!l!)Zl!l!l!l!l!r'