qiMrdZddlmZmZddlmZmZmZddlm Z ddl m Z m Z m Z e Gdde ZdgZy ) z Processor class for Nougat. )OptionalUnion)PreTokenizedInput TextInputTruncationStrategy)ProcessorMixin)PaddingStrategy TensorTypeauto_docstringcB*eZdZfdZe d'dedzdedzdeeefdzdddedzd edzd edzd edzd ee zdzd edzde e e zdzde e e zdzde dde edfdzde eze e ze ezdzde eze e ze ezdzde eze e ze ezdzdedeezezdeezezdzdedzdedededzdeezdzdedzdedzd ed!ed"ed#ed$ef@d%Zd&ZxZS)(NougatProcessorc&t|||y)N)super__init__)selfimage_processor tokenizer __class__s ^/opt/pipecat/venv/lib/python3.12/site-packages/transformers/models/nougat/processing_nougat.pyrzNougatProcessor.__init__s )4Ndo_crop_margin do_resizesizeresamplePILImageResampling do_thumbnaildo_align_long_axisdo_pad do_rescalerescale_factor do_normalize image_mean image_std data_formatChannelDimensioninput_data_format text_pair text_targettext_pair_targetadd_special_tokenspadding truncation max_lengthstrideis_split_into_wordspad_to_multiple_ofreturn_tensorsreturn_token_type_idsreturn_attention_maskreturn_overflowing_tokensreturn_special_tokens_maskreturn_offsets_mapping return_lengthverbosec#| | td|!|j|||||||| | | | | ||||}#|$|j||||||||||||||||| |!|"}$|#S|$S$d#d<|#S)ag do_crop_margin (`bool`, *optional*): Whether to automatically crop white margins from document images. When enabled, the processor detects and removes white space around the edges of document pages, which is useful for processing scanned documents or PDFs with large margins. do_thumbnail (`bool`, *optional*): Whether to create a thumbnail version of the image. When enabled, a smaller version of the image is generated alongside the main processed image, which can be useful for preview or faster processing. do_align_long_axis (`bool`, *optional*): Whether to automatically align images so that the longer axis is horizontal. When enabled, portrait images are rotated to landscape orientation, which is typically better for document processing tasks. zBYou need to specify either an `images` or `text` input to process.)rrrrrrrr r!r"r#r$r2r%r')r(r)r*r+r,r-r.r/r0r1r2r3r4r5r6r7r8r9 input_idslabels) ValueErrorrr)%rimagestextrrrrrrrr r!r"r#r$r%r'r(r)r*r+r,r-r.r/r0r1r2r3r4r5r6r7r8r9inputs encodingss% r__call__zNougatProcessor.__call__sd >dlab b  ))-#!)#5%-)%#-'"3!*F$  #'!1#5%%$7#5-&;&;*C+E'=+''I, <M ^ (5F8 Mrc:|jj|i|S)z This method forwards all its arguments to NougatTokenizer's [`~PreTrainedTokenizer.post_process_generation`]. Please refer to the docstring of this method for more information. )rpost_process_generation)rargskwargss rrDz'NougatProcessor.post_process_generations 6t~~55tFvFFr)"NNNNNNNNNNNNNNchannels_firstNNNNTFNNrFNNNNFFFFT)__name__ __module__ __qualname__rr booldictstrintfloatlistrrrrr rr rBrD __classcell__)rs@rrrs5 &*!%&*)-$(*.""&-1$(15044DCGfjhlmq#'05=A!%$))-26-1-1*/+0',#Gdt d $; d 38nt# d'dTkd!4Kdt d4Kde d*dTkdDK'$.d4;&-d 01!d"!&8!89D@#d$004 ?BTJ[E\\_cc%d&!22T)_DtL]G^^aee'd($&77$y/IDQbLccfjj)d*!+d,o--d.3J!33d:/d0$J1d23d4"5d6 $J7d8j(4/9d: $d{;d< $d{=d>$(?d@%)AdB!%CdDEdFGddLGrrN)__doc__typingrr$transformers.tokenization_utils_baserrrprocessing_utilsr utilsr r r r__all__rrrYsJ#aa.@@pGnpGpGf  r