a h!@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__rrd/var/www/html/assistant/venv/lib/python3.9/site-packages/transformers/models/udop/processing_udop.pyr s  rF)totalc @s4eZdZUeed<dddddddddd idZdS)UdopProcessorKwargs text_kwargsTFr) Zadd_special_tokenspaddingZ truncationZstridereturn_overflowing_tokensZreturn_special_tokens_maskreturn_offsets_mappingZ return_lengthverbose)r images_kwargsN)rrrrr _defaultsrrrrr%s  rcsveZdZdZddgZdZdZfddZdee e e e e e e e feeed d d Zd d ZeddZZS) 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 tokenizerZLayoutLMv3ImageProcessor)Z UdopTokenizerZUdopTokenizerFastcst||dS)N)super__init__)selfr#r$ __class__rrr&QszUdopProcessor.__init__N)imagestextkwargsreturncKs|jtfd|jji|}|ddd}|ddd}|ddd} |ddd} |dd d} |dd d} |jjr|durtd |jjr|durtd | r| std | dur|jfi|dS|jfd|i|d} | dd}| dd}|dd d|ddd| |dd<|dur@|n||dd<||dd<|dur|jjr| durt |t r|g}||dd<|jfd|dur|n|i|d}| dur| | d|d| d<| || SdS)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. Ztokenizer_init_kwargsrrNr text_pairrFr 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.r*r wordsZtext_pair_targetr+TZ pixel_valuesoverflow_to_sample_mapping) Z _merge_kwargsrr$Z init_kwargspopgetr#Z apply_ocr ValueError isinstancestrget_overflowing_imagesupdate)r'r*r+ZaudioZvideosr,Z output_kwargsrrr.rrr/featuresZfeatures_wordsZfeatures_boxesZencoded_inputsrrr__call__Tsd         zUdopProcessor.__call__cCsLg}|D]}|||qt|t|krHtdt|dt||S)Nz`Expected length of images to be the same as the length of `overflow_to_sample_mapping`, but got z and )appendlenr4)r'r*r1Zimages_with_overflowZ sample_idxrrrr7sz$UdopProcessor.get_overflowing_imagescCs"|jj}|jj}t||dgS)NZbbox)r$model_input_namesr#r)r'Ztokenizer_input_namesZimage_processor_input_namesrrrr=szUdopProcessor.model_input_names)NNNN)rrr__doc__ attributesZimage_processor_classZtokenizer_classr&rrrr r rr rrr:r7propertyr= __classcell__rrr(rr"7s$  Xr"N)r>typingrrZ transformersrZimage_processing_utilsrZ image_utilsrZprocessing_utilsrr r r Ztokenization_utils_baser r Z get_loggerrloggerrrr"__all__rrrrs