o ưiK,@sdZddlmZddlmZmZmZmZmZm Z m Z ddl m Z ddl mZmZddlmZddlmZerBddlmZdd lmZGd d d eZd S) z Gemini-specific RAG Ingestion implementation. Gemini handles embedding and chunking internally when files are uploaded to File Search stores, so this implementation skips the embedding step and directly uploads files. ) annotations) TYPE_CHECKINGAnyDictListOptionalTuplecast)verbose_logger)get_async_httpx_clienthttpxSpecialProvider)GeminiModelInfo)BaseRAGIngestion)Router)RAGIngestOptionscsbeZdZdZ d)d*fdd Zd+d dZd,ddZd-ddZd.d d!Zd/d$d%Z d0d'd(Z Z S)1GeminiRAGIngestionaj Gemini-specific RAG ingestion using File Search API. Key differences from base: - Embedding is handled by Gemini when files are uploaded to File Search stores - Files are uploaded using uploadToFileSearchStore API - Chunking is done by Gemini's File Search (supports custom white_space_config) - Supports custom metadata attachment Ningest_options'RAGIngestOptions'routerOptional['Router']cstj||dt|_dS)N)rr)super__init__r Z model_info)selfrr __class__]/home/app/Keep/.python/lib/python3.10/site-packages/litellm/rag/ingestion/gemini_ingestion.pyr$s zGeminiRAGIngestion.__init__chunks List[str]returnOptional[List[List[float]]]csdS)z Gemini handles embedding internally - skip this step. Returns: None (Gemini embeds when files are uploaded to File Search store) Nr)rrrrrembed,s zGeminiRAGIngestion.embed file_contentOptional[bytes]filename Optional[str] content_type embeddings#Tuple[Optional[str], Optional[str]]c s|jd}ttttf|j}ttt|dpt}ttt|dp,t } |s3t d| s9t dd} | d| } |sR|j || |j pLdd Id H}d } |rh|rh|rh|j || ||||d Id H} || fS) a Store content in Gemini File Search store. Gemini workflow: 1. Create File Search store (if not provided) 2. Upload file using uploadToFileSearchStore (Gemini handles chunking/embedding) Args: file_content: Raw file bytes filename: Name of the file content_type: MIME type chunks: Ignored - Gemini handles chunking embeddings: Ignored - Gemini handles embedding Returns: Tuple of (vector_store_id, file_id) vector_store_idapi_keyapi_basezCGEMINI_API_KEY or GOOGLE_API_KEY is required for Gemini File SearchzGEMINI_API_BASE is requiredZv1beta/zlitellm-rag-ingest)r*base_url display_nameN)r*r-r)r$r"r&)vector_store_configgetr rstrrrr Z get_api_keyZ get_api_base ValueError_create_file_search_storeZ ingest_name_upload_to_file_search_store) rr"r$r&rr'r)r/r*r+ api_versionr-Zresult_file_idrrrstore9s8     zGeminiRAGIngestion.storer*r1r-r.c s|d|}d|i}ttjddid}|j||ddidId H}|jd kr6d |j}t|t|| } | d d } t d| | S)a Create a Gemini File Search store. Args: api_key: Gemini API key base_url: Base URL for Gemini API display_name: Display name for the store Returns: Store name (format: fileSearchStores/xxxxxxx) z/fileSearchStores?key= displayNametimeoutN@Z llm_providerparams Content-Typeapplication/jsonjsonheadersNz$Failed to create File Search store: namezCreated File Search store: ) r r RAGpost status_codetextr error Exceptionr?r0debug) rr*r-r.url request_bodyclientresponse error_msg response_dataZ store_namerrrr3ys*    z,GeminiRAGIngestion._create_file_search_storer)bytesc s>|j||||t||p ddIdH}|j||dIdH}|S)a Upload a file to Gemini File Search store using resumable upload. Args: api_key: Gemini API key base_url: Base URL for Gemini API vector_store_id: File Search store name filename: Name of the file file_content: File content bytes content_type: MIME type Returns: File ID or document name zapplication/octet-stream)r*r-r)r$ file_sizer&N) upload_urlr")_initiate_resumable_uploadlen_upload_file_content) rr*r-r)r$r"r&rSfile_idrrrr4s   z/GeminiRAGIngestion._upload_to_file_search_storerRintcsN|dd}|d|d|}d|i} |j} | r7t| tr7| d} | r7d| dd | d d d i| d <ttttt t f|j d} | rN| | d<ddt ||dd} t d|ttjddid}|j|| | dIdH}|jdvrd|j}t |t|t d|j|jd}|stdt d||S) zy Initiate a resumable upload session. Returns: Upload URL for the resumable session z/v1betarCz/upload/v1beta/z:uploadToFileSearchStore?key=r7white_space_configZwhiteSpaceConfigZmax_tokens_per_chunki Zmax_overlap_tokensi)ZmaxTokensPerChunkZmaxOverlapTokensZchunkingConfigcustom_metadataZcustomMetadataZ resumablestartr=)zX-Goog-Upload-ProtocolX-Goog-Upload-Commandz#X-Goog-Upload-Header-Content-Lengthz!X-Goog-Upload-Header-Content-Typer<zInitiating resumable upload: r8r9r:r>NrAzFailed to initiate upload: z$Initiate resumable upload response: zx-goog-upload-urlz*No upload URL returned in response headerszGot upload URL: )replacechunking_strategy isinstancedictr0r rrrr1rr/r rJr r rDrErFrGrHrIr@)rr*r-r)r$rRr&r+rKrLr`rYrZr@rMrNrOrSrrrrTsV    "    z-GeminiRAGIngestion._initiate_resumable_uploadrSc stt|ddd}tdt|dttjddid}|j|||d Id H}|jd vr>d |j }t |t |z| }| d dpR| di d d}td||WSt yx} ztd| WYd } ~ dSd } ~ ww)z Upload file content to the resumable upload URL. Returns: File ID or document name from the response 0zupload, finalize)zContent-LengthzX-Goog-Upload-Offsetr\zUploading file content (z bytes)r8gr@r:)contentr@Nr]zFailed to upload file: rBrCfilezUpload complete. File ID: z!Could not parse upload response: Zuploaded)r1rUr rJr r rDputrFrGrHrIr?r0warning) rrSr"r@rMrNrOrPrWerrrrVs:     z'GeminiRAGIngestion._upload_file_content)N)rrrr)rrrr ) r"r#r$r%r&r%rrr'r rr()r*r1r-r1r.r1rr1)r*r1r-r1r)r1r$r1r"rQr&r%rr1)r*r1r-r1r)r1r$r1rRrXr&r1rr1)rSr1r"rQrr1) __name__ __module__ __qualname____doc__rr!r6r3r4rTrV __classcell__rrrrrs   @ , )LrN)rl __future__rtypingrrrrrrr Zlitellm._loggingr Z&litellm.llms.custom_httpx.http_handlerr r Z litellm.llms.gemini.common_utilsr Z$litellm.rag.ingestion.base_ingestionrZlitellmrZlitellm.types.ragrrrrrrs $