uPj~dZddlmZmZmZddlmZddlmZddl m Z erddl m Z ddl mZddlmZGd d eZy ) z OCR Handler for Unified Guardrails Provides guardrail translation support for the OCR endpoint. Processes the extracted markdown text from OCR pages. ) TYPE_CHECKINGAnyFinal)verbose_proxy_logger)BaseTranslation)GenericGuardrailAPIInputs)CustomGuardrail)Logging) OCRResponsec `eZdZdZ ddedddddefd Z dd d ddddd edzd edzdef dZy) OCRHandlera Handler for processing OCR requests/responses with guardrails. Input: The OCR input is a document URL/reference - not text content. We pass the document URL as text for guardrails that may want to validate or filter document sources. Output: OCR responses contain extracted markdown text per page. The handler extracts all page markdown, applies guardrails, and maps the guardrailed text back to the pages. Ndataguardrail_to_applyr litellm_logging_objzLiteLLMLoggingObj | NonereturncK|jd}|t|tstjd|Sg}|jd}|dk(r5|jd}|r[t|t rK|j |n9|dk(r4|jd}|r!t|t r|j ||s|St|}|jd} | r| |d<|j||d | d{|S7w) a Process OCR input by applying guardrails to the document reference. The OCR input contains a document dict with a URL. We extract the URL and pass it to the guardrail for validation. Args: data: Request data containing 'document' parameter guardrail_to_apply: The guardrail instance to apply litellm_logging_obj: Optional logging object Returns: Modified data with guardrails applied documentNz6OCR guardrail: No valid document found in request datatype document_url image_urltextsmodelrequestinputs request_data input_type logging_obj) get isinstancedictrdebugstrappendrapply_guardrail) selfrrrrtexts_to_checkdoc_typeurlrrs {/home/oem/CyberBrain26-unified/.venv/lib/python3.12/site-packages/litellm/llms/mistral/ocr/guardrail_translation/handler.pyprocess_input_messagesz!OCRHandler.process_input_messages!s(((:.  :h#= & &'_ `K,.",,v. ~ %,,~.Cz#s+%%c*  $,,{+Cz#s+%%c*K1Gxx( #F7O 00 + 1     sDD D D responser user_api_key_dictrcKt|dr |jstjd|Sg}g}t |jD]K\}} t| ds| j s |j | j |j |M|s|St|} t|dd} | r| | d<|i}|-|j|} | r| j| d|vr| |d<|j| |d| d{} | jd g}t |D]+\}}|t|ks|||j|_-tjd t||S7pw) a% Process OCR output by applying guardrails to extracted page text. Extracts markdown text from each OCR page, applies guardrails, and maps the guardrailed text back to the pages. Args: response: OCRResponse with pages containing markdown text guardrail_to_apply: The guardrail instance to apply litellm_logging_obj: Optional logging object user_api_key_dict: User API key metadata Returns: Modified OCRResponse with guardrailed page text pagesz-OCR guardrail: No pages found in OCR responsemarkdownrrNlitellm_metadatar-rrz,OCR guardrail: Applied guardrail to %d pages)hasattrr0rr# enumerater1r%rgetattr'transform_user_api_key_dict_to_metadataupdater&r len)r'r-rrr.rr( page_indicesipagerr user_metadataguardrailed_inputsguardrailed_textsidxpage_idxs r+process_output_responsez"OCRHandler.process_output_responseWs.x) & &'V WO,.)+  0 'GAttZ(T]]%%dmm4##A& ' O1Gx$7 #F7O  L  (#'#O#OPa#bM m,%\97DL!34*<*L*L%!+ +M+ % $6#9#9'2#F&|4 KMCS*++4Ec4Jx(1 K "" : ! " %% s%AE; E;+BE;E9 5E;?;E;)N)NNN)__name__ __module__ __qualname____doc__r"rr,rAr+r r s  ;? 44.48 4 4t;?(,$( MM.M8 M : M Tk M MrGr N)rEtypingrrrlitellm._loggingrrPs5-,1X9EWDPPrG