NgddlmZmZddlmZddlmZddlmZdde de fdZ dde de d e d e d e d eee f dZ dZ dd e d e de fdZd S))ListOptional)logger)PartitionStrategy)dependency_existsFstrategyis_imagectjtjtjtjg}||vrt |d|tjkr|rt ddSdS)z?Determines if the strategy is valid for the specified filetype.z is not a valid strategy.z3The fast strategy is not available for image files.N)rAUTOFASTOCR_ONLYHI_RES ValueError)rr valid_strategiess ]/var/www/html/ai-engine/env/lib/python3.11/site-packages/unstructured/partition/strategies.pyvalidate_strategyrsz "  '''H???@@@$)))h)NOOO*)))Npdf_text_extractableinfer_table_structureextract_images_in_pdfextract_image_block_typesctd}td}|tjkr4|pt|}|rt }nt |||}t | | | grtd|tjkrX|sVtj d|r tj dtj Stj dtj S|tj krX|sVtj d|r tj dtj Stj d tjS|S) zDetermines what strategy to use for processing PDFs or images, accounting for fallback logic if some dependencies are not available.unstructured_pytesseractunstructured_inferencerrextract_elementzunstructured_inference is not installed, pytesseract is not installed and the text of the PDF is not extractable. To process this file, install unstructured_inference, install pytesseract, or remove copy protection from the PDF.zunstructured_inference is not installed. Cannot use the hi_res partitioning strategy. Falling back to partitioning with another strategy.z+Falling back to partitioning with ocr_only.z'Falling back to partitioning with fast.zpytesseract is not installed. Cannot use the ocr_only partitioning strategy. Falling back to partitioning with another strategy.z)Falling back to partitioning with hi_res.) rrr bool_determine_image_auto_strategy_determine_pdf_auto_strategyallrrrwarningr r ) rr rrrrpytesseract_installed unstructured_inference_installedrs rdetermine_pdf_or_image_strategyr$s..HII'89Q'R'R$$)))/R48Q3R3R  577HH3%9&; /H  - -3H/HNbJbc  6   $+++4T+ L   ! * NH I I I$- - ND E E E$) ) &/ / /8M / L     , ND E E E$) ) NF G G G$+ + OrctjS)zWIf "auto" is passed in as the strategy, determines what strategy to use for images.)rrrrrrWs   ##rrcV|s|r tjS|r tjStjS)zUIf "auto" is passed in as the strategy, determines what strategy to use for PDFs.)rrr r rs rrr^s6(( ''* %% ))r)F)FFFFN)FFF)typingrrunstructured.loggerr&unstructured.partition.utils.constantsrunstructured.utilsrstrrrr$rrr&rrr-sD!!!!!!!!&&&&&&DDDDDD000000 P P Pt P P P P$!&"'"'59 <<<<< <  < (S 2 <<<<~$$$"'"'!**********r