{-jpUfddlZddlmZmZmZmZmZmZddlZ ddl m Z ddl m Z ddlmZddlmZdd lmZmZdd lmZdd lmZd d lmZd dlmZd dlmZd dlm Z d dl!m"Z"ddl#m$Z$m%Z%ddl&m'Z'ddl m(Z(ej)GddeZ*e dGddeZ+dS)N)AnyDictListOptionalTupleUnion)logging)pipeline_requires_extra)ImageBatchSampler) ReadImage) HPIConfigPaddlePredictorOption) DetResult) benchmark)(AutoParallelImageSimpleInferencePipeline) BasePipeline) CropByBoxes)DocPreprocessorResult) OCRResult)SingleTableRecognitionResultTableRecognitionResult)get_table_recognition_res)get_neighbor_boxes_idxc!eZdZdZddddddddedeedeedeeeefd eed e d ee eeefe fd dffd Z dee dee dee dee d e f dZdededed e fdZdejde d eeejffdZdZdZ d*dejdedede de d e d efd!Z d+d"e eeeejeejfdee dee dee dee deedeed#eed$eed%eed&eed'eed(eede d ee d ef d)ZxZ S),_TableRecognitionPipelinezTable Recognition PipelineNFdeviceengine engine_config pp_optionuse_hpip hpi_configconfigr!r"r#r$r%r&returnc tjd||||||d||dd|_|jrF|didddi} || |_|dd|_|jrF|d id d d i} || |_|d id d di} || |_ |dd|_ |j rG|didddi} || |_ n/|didd|_ t|_td|_t#d|_dS)aInitializes the table recognition pipeline. Args: config (Dict): Configuration dictionary containing various settings. device (Optional[str], optional): The device to use for prediction. Defaults to `None`. engine (Optional[str], optional): Inference engine. Defaults to `None`. engine_config (Optional[Dict[str, Any]], optional): Engine-specific config. Defaults to `None`. pp_option (Optional[PaddlePredictorOption], optional): Paddle predictor options. Defaults to `None`. use_hpip (bool, optional): Whether to use HPIP. Defaults to `False`. hpi_config (Optional[Union[Dict[str, Any], HPIConfig]], optional): HPIP configuration. Defaults to `None`. r use_doc_preprocessorT SubPipelinesDocPreprocessorpipeline_config_errorz+config error for doc_preprocessor_pipeline!use_layout_detection SubModulesLayoutDetectionmodel_config_errorz"config error for layout_det_model!TableStructureRecognitionz'config error for table_structure_model! use_ocr_model GeneralOCRz&config error for general_ocr_pipeline!Nr) batch_sizeBGR)format)super__init__getr*create_pipelinedoc_preprocessor_pipeliner. create_modellayout_det_modeltable_structure_modelr3general_ocr_pipelinegeneral_ocr_config_bakr_crop_by_boxesr batch_samplerr img_reader)selfr'r!r"r#r$r%r&kwargsdoc_preprocessor_configlayout_det_configtable_structure_configgeneral_ocr_config __class__s v/var/www/html/banglarbhumi/venv/lib/python3.11/site-packages/paddlex/inference/pipelines/table_recognition/pipeline.pyr:z"_TableRecognitionPipeline.__init__)s?2  '!      %+JJ/Et$L$L!  $ &,jj&D&D&H&H!+-Z'' # .2-A-A'..D *%+JJ/Et$L$L!  $ I & < < < @ @!%'KL!! %)$5$56G$H$HD !!'L"!=!=!A!A ' !#L M" " &*%6%67M%N%N"#ZZ>>   !'NB!?!?!C!C(*RS"" )-(<(<=O(P(PD % %*0**^R*H*H*L*Ld++D '*mm.!<<<#5111use_doc_orientation_classifyuse_doc_unwarpingr.r3c|| ||j}n |dus|durd}nd}||j}||j}t|||S)a Get the model settings based on the provided parameters or default values. Args: use_doc_orientation_classify (Optional[bool]): Whether to use document orientation classification. use_doc_unwarping (Optional[bool]): Whether to use document unwarping. use_layout_detection (Optional[bool]): Whether to use layout detection. use_ocr_model (Optional[bool]): Whether to use OCR model. Returns: dict: A dictionary containing the model settings. NTF)r*r.r3)r*r.r3dict)rFrOrPr.r3r*s rMget_model_settingsz,_TableRecognitionPipeline.get_model_settingsws~& ( /4E4M#'#< +t337HD7P7P'+$$',$  '#'#<   .M!5!5'    rNmodel_settingsoverall_ocr_reslayout_det_rescv|dr|jstjddS|dr5|tjddS|jstjddS|dr6|tjd dS|jstjd dSn|tjd dSd S) a0 Check if the input parameters are valid based on the initialized models. Args: model_settings (Dict): A dictionary containing input parameters. overall_ocr_res (OCRResult): Overall OCR result obtained after running the OCR pipeline. The overall OCR result with convert_points_to_boxes information. layout_det_res (DetResult): The layout detection result. Returns: bool: True if all required models are initialized according to input parameters, False otherwise. r*zRSet use_doc_preprocessor, but the models for doc preprocessor are not initialized.Fr.Nz^The layout detection model has already been initialized, please set use_layout_detection=FalsezRSet use_layout_detection, but the models for layout detection are not initialized.r3zLThe OCR models have already been initialized, please set use_ocr_model=Falsez>Set use_ocr_model, but the models for OCR are not initialized.z7Set use_ocr_model=False, but no OCR results were found.T)r*r errorr.r3)rFrTrUrVs rMcheck_model_settings_validz4_TableRecognitionPipeline.check_model_settings_valids$ 0 1 $:S  Md   5 0 1 ) tu,  hu / * * bu%  Tu   & WXXXutrN image_array input_paramsc|drD|d}|d}t||||d}|d}ni}|}||fS)a Preprocess the document image based on input parameters. Args: image_array (np.ndarray): The input image array. input_params (dict): Dictionary containing preprocessing parameters. Returns: tuple[DocPreprocessorResult, np.ndarray]: A tuple containing the preprocessing result dictionary and the processed image array. r*rOrPrOrPr output_img)listr=)rFrZr[rOrPdoc_preprocessor_resdoc_preprocessor_images rMpredict_doc_preprocessor_resz6_TableRecognitionPipeline.predict_doc_preprocessor_ress . / 1+78V+W ( ,-@ A #'..1M&7/$$  $ &:,%G " "#% %0 "#%;;;rNc t|ts|}g}tt |D]~}d||D\}}}}t||||||ddfd} |d| d|S)t Splits OCR bounding boxes by table cells and retrieves text. Args: ori_img (ndarray): The original image from which text regions will be extracted. cells_bboxes (list or ndarray): Detected cell bounding boxes to extract text from. Returns: list: A list containing the recognized texts from each cell. c6g|]}tj|Sr8mathceil.0ks rM zM_TableRecognitionPipeline.split_ocr_bboxes_by_table_cells.. DDDqdillDDDrNNr rec_texts isinstancer_tolistrangelenrAappendjoin rFori_img cells_bboxes texts_listix1y1x2y2rec_tes rMsplit_ocr_bboxes_by_table_cellsz9_TableRecognitionPipeline.split_ocr_bboxes_by_table_cells,-- 1'..00L s<(()) < .!rmrNNrrnrorprws rMrz9_TableRecognitionPipeline.split_ocr_bboxes_by_table_cellsrrNT table_box use_ocr_results_with_table_cellsflag_find_nei_textcell_sort_by_y_projectionc|t||d}|dkr+|d}d|D}|||} ng} t|||| ||} d} |rBt |d|} t | dkr| D]} | |d| d zz } | | d <| S) a+ Predict table recognition results from an image array, layout detection results, and OCR results. Args: image_array (np.ndarray): The input image represented as a numpy array. overall_ocr_res (OCRResult): Overall OCR result obtained after running the OCR pipeline. The overall OCR results containing text recognition information. table_box (list): The table box coordinates. use_ocr_results_with_table_cells (bool): whether to use OCR results with cells. flag_find_nei_text (bool): Whether to find neighboring text. cell_sort_by_y_projection (bool): Whether to sort the matched OCR boxes by y-projection. Returns: SingleTableRecognitionResult: single table recognition result. rTbboxcJg|] }|d|d|d|dg!S)rrr r8)rjrects rMrlzR_TableRecognitionPipeline.predict_single_table_recognition_res..Cs>"""9=a$q'47DG4"""rNrrn rec_boxesroz; neighbor_texts)r_r@rrrrt)rFrZrUrrrrtable_structure_predtable_cells_resultcells_texts_listsingle_table_recognition_res neighbor_textmatch_idx_listidxs rM$predict_single_table_recognition_resz>_TableRecognitionPipeline.predict_single_table_recognition_res)s&. $D$>$>{$K$KLLQO +t 3 3!5f!= ""AS"""  $CC/     " '@    ,&? ( ( ( $  N3 ,iN>""Q&&)NNC!_[%A#%F%MMMM9F$%56++rNinputtext_det_limit_side_lentext_det_limit_typetext_det_threshtext_det_box_threshtext_det_unclip_ratiotext_rec_score_threshc +K|||||}|d}||||sddiVt||D]'\}}||jd}|dr,t ||||d}nd|i}|d}|d r0t |||| | | | | d}n2|d kr,|j dksJ| |j |_g}d }|d s]|[i}|j dd\}}dd|d z |d z g}| ||||d|}||d<| ||d z }n|d r(t ||d}|dD]}|ddvrf|||g}|d}|d}| |d||||}||d<| ||d z }|jd|jd|||||d}t'|V)dS)az This function predicts the layout parsing result for the given input. Args: input (Union[str, list[str], np.ndarray, list[np.ndarray]]): The input image(s) of pdf(s) to be processed. use_layout_detection (bool): Whether to use layout detection. use_doc_orientation_classify (bool): Whether to use document orientation classification. use_doc_unwarping (bool): Whether to use document unwarping. overall_ocr_res (OCRResult): The overall OCR result with convert_points_to_boxes information. It will be used if it is not None and use_ocr_model is False. layout_det_res (DetResult): The layout detection result. It will be used if it is not None and use_layout_detection is False. use_ocr_results_with_table_cells (bool): whether to use OCR results with cells. cell_sort_by_y_projection (bool): Whether to sort the matched OCR boxes by y-projection. **kwargs: Additional keyword arguments. Returns: TableRecognitionResult: The predicted table recognition result. NFrXz0the input params for model settings are invalid!rr*r]r^r3)rrrrrrTrr.r)rrtable_region_idboxeslabel)tableboximgr) input_path page_indexr`rVrUtable_res_listrT)rSrY enumeraterDrE instancesr_r=rArBr<shaperrur?lowerrC input_paths page_indexesr) rFrrOrPr.r3rUrVrrrrrrrrrGrTimg_id batch_datarZr`rarr img_height img_widthrsingle_table_rec_resbox_info crop_img_infosingle_img_ress rMpredictz!_TableRecognitionPipeline.predict^sN00 (      % ,(- %.. O^   PNO O O O"+D,>,>u,E,E"F"FT 9T 9 FJ//**>??BK45 C'+22#5Q*;3((  ($$)5k'B$%9,%G "o. "&--.0G,?(7,?.C.C. # # #2T992d::::,0,@,@/--) NO!"89% -n>T!#(>(DRaR(H% I9q=*q.A '+'P'P*#4',.G (Q(($;J$%67%%&:;;;1$!"89%)--.DEE&&&N!/w 7--H(..00I==(,(;(;K((T(T (5a(8 $1%$8  EE -e 4 / ) @:S F-CR,->?&--.BCCC'1,)4Q7(5a8(<"0#2"0"0N)88 8 8 8 8iT 9T 9rN)FTF)NNNNNNNNNNNNFN)!__name__ __module__ __qualname____doc__rrstrrrboolrrr:rRrSrrrYnpndarrayrrrbrr_rrrintfloatrr __classcell__)rLs@rMrr%s$$ !% $2659AEL2L2L2L2 L2  L2  S#X/ L212L2L2U4S>9#<=>L2 L2L2L2L2L2L2\% &.tn% $D>% 'tn %  ~ %  % % % % N55#5" 5  5555n<:<59< $bj0 1<<<<<6@27#'*/3,3,Z3,#3, 3, +/ 3, ! 3,$(3, &3,3,3,3,p8<,0/3(,/3.215-1+//315151648!I9I9S$s)RZbj1AABI9'/tnI9$D> I9 'tn I9  ~ I9"),I9!+I9"*#I9&c]I9"%I9&e_I9 (I9 (I9+/I9 $,D>!I9$ %I9I9I9I9I9I9I9I9rNrocrc0eZdZdgZedZdZdS)TableRecognitionPipelinetable_recognitionctS)N)r)rFs rM _pipeline_clsz&TableRecognitionPipeline._pipeline_clss((rNcdS)Nrr8)rFr's rM_get_batch_sizez(TableRecognitionPipeline._get_batch_sizesqrNN)rrrentitiespropertyrrr8rNrMrrsD#$H ))X)rNr),rgtypingrrrrrrnumpyrutilsr utils.depsr common.batch_samplerr common.readerrmodelsrrmodels.object_detection.resultrutils.benchmarkr _parallelrbaser componentsrdoc_preprocessor.resultr ocr.resultrresultrr!table_recognition_post_processingrr time_methodsrrr8rNrMrs ::::::::::::::::222222555555&&&&&&66666666777777((((((@@@@@@$$$$$$;;;;;;""""""HHHHHHHHHHHHHH)))))) A9A9A9A9A9 A9A9A9HG rN