o wi,'@sdZddlmZmZddlmZddlmZddlm Z ddl m Z m Z m Z mZddlmZmZeeZGd d d e d d ZGd dde d d ZGddde ZdgZdS)z Processor class for UDOP. )OptionalUnion)logging) BatchFeature) ImageInput)ProcessingKwargsProcessorMixin TextKwargsUnpack)PreTokenizedInput TextInputc@sReZdZUeeeeeeefed<eeeeeeeefed<dS)UdopTextKwargs word_labelsboxesN)__name__ __module__ __qualname__rrlistint__annotations__rre/home/ubuntu/sommelier/.venv/lib/python3.10/site-packages/transformers/models/udop/processing_udop.pyr s (rF)totalc @s4eZdZUeed<dddddddddd idZdS)UdopProcessorKwargs text_kwargsTFr) add_special_tokenspadding truncationstridereturn_overflowing_tokensreturn_special_tokens_maskreturn_offsets_mapping return_lengthverbose)r images_kwargsN)rrrrr _defaultsrrrrr%s  rc seZdZdZddgZdZdZdgZfddZ dd d d d e e d e e e ee ee fd eedefddZddZddZddZeddZZS) UdopProcessora Constructs a UDOP processor which combines a LayoutLMv3 image processor and a UDOP tokenizer into a single processor. [`UdopProcessor`] offers all the functionalities you need to prepare data for the model. It first uses [`LayoutLMv3ImageProcessor`] to resize, rescale and normalize document images, and optionally applies OCR to get words and normalized bounding boxes. These are then provided to [`UdopTokenizer`] or [`UdopTokenizerFast`], which turns the words and bounding boxes into token-level `input_ids`, `attention_mask`, `token_type_ids`, `bbox`. Optionally, one can provide integer `word_labels`, which are turned into token-level `labels` for token classification tasks (such as FUNSD, CORD). Additionally, it also supports passing `text_target` and `text_pair_target` to the tokenizer, which can be used to prepare labels for language modeling tasks. Args: image_processor (`LayoutLMv3ImageProcessor`): An instance of [`LayoutLMv3ImageProcessor`]. The image processor is a required input. tokenizer (`UdopTokenizer` or `UdopTokenizerFast`): An instance of [`UdopTokenizer`] or [`UdopTokenizerFast`]. The tokenizer is a required input. image_processor tokenizerLayoutLMv3ImageProcessor) UdopTokenizerUdopTokenizerFast text_paircst||dS)N)super__init__)selfr(r) __class__rrr/SszUdopProcessor.__init__N)audiovideosimagestextkwargsreturncOs|jtfd|jji||j|}|ddd}|ddd} |ddd} |ddd} |dd d} |dd d} |jjrN|durNt d |jjrZ| durZt d | rb| sbt d | durp|jdi|dS|jdd|i|d}|dd}|dd}|dd d|ddd| |dd<|dur|n||dd<| |dd<|dur|jjr| durt |t r|g}||dd<|jdd|dur|n|i|d}| dur| |d|d|d<| ||S)a~ This method first forwards the `images` argument to [`~UdopImageProcessor.__call__`]. In case [`UdopImageProcessor`] was initialized with `apply_ocr` set to `True`, it passes the obtained words and bounding boxes along with the additional arguments to [`~UdopTokenizer.__call__`] and returns the output, together with the prepared `pixel_values`. In case [`UdopImageProcessor`] was initialized with `apply_ocr` set to `False`, it passes the words (`text`/``text_pair`) and `boxes` specified by the user along with the additional arguments to [`~UdopTokenizer.__call__`] and returns the output, together with the prepared `pixel_values`. Alternatively, one can pass `text_target` and `text_pair_target` to prepare the targets of UDOP. Please refer to the docstring of the above two methods for more information. tokenizer_init_kwargsrrNrr-r Fr" text_targetzdYou cannot provide bounding boxes if you initialized the image processor with apply_ocr set to True.zaYou cannot provide word labels if you initialized the image processor with apply_ocr set to True.zKYou cannot return overflowing tokens without returning the offsets mapping.r5r%wordstext_pair_targetr6T pixel_valuesoverflow_to_sample_mappingr) _merge_kwargsrr) init_kwargs'prepare_and_validate_optional_call_argspopgetr( apply_ocr ValueError isinstancestrget_overflowing_imagesupdate)r0r5r6r3r4argsr7 output_kwargsrrr-r r"r:featuresfeatures_wordsfeatures_boxesencoded_inputsrrr__call__Vsh        zUdopProcessor.__call__cCsLg}|D] }|||qt|t|kr$tdt|dt||S)Nz`Expected length of images to be the same as the length of `overflow_to_sample_mapping`, but got z and )appendlenrE)r0r5r>images_with_overflow sample_idxrrrrHsz$UdopProcessor.get_overflowing_imagescO|jj|i|S)z This method forwards all its arguments to PreTrainedTokenizer's [`~PreTrainedTokenizer.batch_decode`]. Please refer to the docstring of this method for more information. )r) batch_decoder0rJr7rrrrVzUdopProcessor.batch_decodecOrU)z This method forwards all its arguments to PreTrainedTokenizer's [`~PreTrainedTokenizer.decode`]. Please refer to the docstring of this method for more information. )r)decoderWrrrrYrXzUdopProcessor.decodecCsgdS)N)r= input_idsbboxattention_maskr)r0rrrmodel_input_namesszUdopProcessor.model_input_names)NN)rrr__doc__ attributesimage_processor_classtokenizer_classoptional_call_argsr/rrrr r rr rrrPrHrVrYpropertyr] __classcell__rrr1rr'7s4    _r'N)r^typingrr transformersrimage_processing_utilsr image_utilsrprocessing_utilsrr r r tokenization_utils_baser r get_loggerrloggerrrr'__all__rrrrs     "