=pjsYSSKJrJrJrJrJr SSKrSSKJ r SSK J r SSK J r SSKJr SS KJrJr SS KJr S S KJr S S KJr S SKJrJrJrJrJrJr SSK J!r! \RD"SS\55r#\ "SSS9"SS\55r$g))AnyDictListOptionalUnionN)logging)pipeline_requires_extra)ImageBatchSampler) ReadImage) HPIConfigPaddlePredictorOption) benchmark)(AutoParallelImageSimpleInferencePipeline) BasePipeline) CropByPolys SortPolyBoxes SortQuadBoxescal_ocr_word_boxconvert_points_to_boxes rotate_image) OCRResultcb^\rSrSrSrSSSSSSS.S\S\\S\\S \\\\4S \\ S \ S \\ \\\4\ 4S S4U4Sjjjr S\\R S\\S \\R 4SjrS\S \ 4SjrS\\ S\\ S\\ S \4SjrS$S\\S\\S\\S\\S\\S\\S \4SjjrS%S\ \\\\R \\R 4S\\ S\\ S\\ S\\S\\S\\S\\S\\S\\S \\S!\\ S \4S"jjrS#rU=r$)& _OCRPipeline&z OCR PipelineNFdeviceengine engine_config pp_optionuse_hpip hpi_configconfigr r!r"r#r$r%returnc >[T U]"S5UUUUUUS.UD6 URSS5UlUR(a:URS05RSSS05n UR U 5UlURSS5UlUR (a:URS 05RS S S 05n URU 5UlURS 05RS S S05n USUl URS:XaU RSS5Ul U RSS5Ul U RSS5Ul U RSS5Ul U RSS5UlU RSS5UlU RSS5Ul[#5Ul['SS 9UlOURS!:XaU RSS"5Ul U RSS#5Ul U RSS5Ul U RSS$5Ul U RSS5UlU RSS%5UlU RSS5Ul[+5Ul['S&S 9UlO$[-S'R/UR55eURU URURURURURUR URS(9UlURS 05RS)S S*05n U RS+S,5UlU RS-S.5UlU RSS5UlURXRS/9UlRS0S15S29Ul[=S3S49Ulg)6aInitializes the OCR 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`. ruse_doc_preprocessorT SubPipelinesDocPreprocessorpipeline_config_errorz+config error for doc_preprocessor_pipeline!use_textline_orientation SubModulesTextLineOrientationmodel_config_errorz,config error for textline_orientation_model! TextDetectionz config error for text_det_model! text_typegenerallimit_side_leni limit_typemaxmax_side_limitithreshg333333? box_threshg333333? input_shapeN unclip_ratiog@quad) det_box_typesealiming?g?polyzUnsupported text type {})r4r5r7r8r9r;r:TextRecognitionz config error for text_rec_model! score_threshrreturn_word_boxF)r: batch_sizer)rDBGR)format) super__init__getr)create_pipelinedoc_preprocessor_pipeliner- create_modeltextline_orientation_modelr2text_det_limit_side_lentext_det_limit_typetext_det_max_side_limittext_det_threshtext_det_box_threshr:text_det_unclip_ratior _sort_boxesr_crop_by_polysr ValueErrorrFtext_det_modeltext_rec_score_threshrCtext_rec_modelr batch_samplerr img_reader)selfr&r r!r"r#r$r%kwargsdoc_preprocessor_configtextline_orientation_configtext_det_configtext_rec_config __class__s d/var/www/html/pdf-tiff/venv/lib/python3.13/site-packages/paddlex/inference/pipelines/ocr/pipeline.pyrI_OCRPipeline.__init__*s2  '!   %+JJ/Et$L!  $ $&,jj&D&H&H!+-Z' # .2-A-A'.D *)/ 3Mt(T%  ( (*0**\2*F*J*J%%'UV+ '/3.?.?+/D +!**\26:: 24VW  , >>Y &+:+>+>?OQT+UD ('6':':<'OD $+:+>+>?OQU+VD (#2#6#6x#ED '6':':<'MD $.22=$GD )8)<)<^S)QD &,D "-6"BD  ^^v %+:+>+>?OQT+UD ('6':':<'OD $+:+>+>?OQU+VD (#2#6#6x#ED '6':':<'MD $)8)<)<^S)QD &.22=$GD ,D "-6"BD 7>>t~~NO O"// 77//77''//33((0  !**\26::  !#E F &5%8%8%K".223DeL*..}dC"// )9)90 /&**\ST:UV#51image_array_listrotate_angle_listc[U5[U5:Xd S[U5S[U5S35eUHnUS;aM SU35e /n[X5H&upVUS-n[XW5nURU5 M( U$)aX Rotate the given image arrays by their corresponding angles. 0 corresponds to 0 degrees, 1 corresponds to 180 degrees. Args: image_array_list (List[np.ndarray]): A list of input image arrays to be rotated. rotate_angle_list (List[int]): A list of rotation indicators (0 or 1). 0 means rotate by 0 degrees 1 means rotate by 180 degrees Returns: List[np.ndarray]: A list of rotated image arrays. Raises: AssertionError: If any rotate_angle is not 0 or 1. AssertionError: If the lengths of input lists don't match. zLength of image_array_list (z*) must match length of rotate_angle_list ())rrz&rotate_angle must be 0 or 1, now it's )lenziprappend) r]rgrhanglerotated_images image_arrayrotate_indicator rotate_angle rotated_images rdr_OCRPipeline.rotate_images(#$ )   E )#.>*?)@@jknpAlBkCCD E E 'EF? T&LUG$T T?'-01A-U )K+c1L(CM  ! !- 0 .V rfmodel_settingscUS(a(UR(d[R"S5 gUS(a(UR(d[R"S5 gg)a Check if the input parameters are valid based on the initialized models. Args: model_info_params(Dict): A dictionary containing input parameters. Returns: bool: True if all required models are initialized according to input parameters, False otherwise. r)zRSet use_doc_preprocessor, but the models for doc preprocessor are not initialized.Fr-z^Set use_textline_orientation, but the models for use_textline_orientation are not initialized.T)r)r errorr-)r]rvs rdcheck_model_settings_valid'_OCRPipeline.check_model_settings_validsT 0 1$:S:S MMd  5 611 MMp rfuse_doc_orientation_classifyuse_doc_unwarpingr-cxUcUc URnOUSLdUSLaSnOSnUc URn[UUS9$)a Get the model settings based on the provided parameters or default values. Args: use_doc_orientation_classify (Optional[bool]): Whether to use document orientation classification. use_doc_unwarping (Optional[bool]): Whether to use document unwarping. use_textline_orientation (Optional[bool]): Whether to use textline orientation. Returns: dict: A dictionary containing the model settings. TF)r)r-)r)r-dict)r]r{r|r-r)s rdget_model_settings_OCRPipeline.get_model_settingss]" ( /4E4M#'#<#< +t37HD7P'+$',$ # +'+'D'D $!5%=  rfrOrPrQrRrSrTc Uc URnUc URnUc URnUc URnUc URnUc UR n[ UUUUUUS9$)a Get text detection parameters. If a parameter is None, its default value from the instance will be used. Args: text_det_limit_side_len (Optional[int]): The maximum side length of the text box. text_det_limit_type (Optional[str]): The type of limit to apply to the text box. text_det_max_side_limit (Optional[int]): The maximum side length of the text box. text_det_thresh (Optional[float]): The threshold for text detection. text_det_box_thresh (Optional[float]): The threshold for the bounding box. text_det_unclip_ratio (Optional[float]): The ratio for unclipping the text box. Returns: dict: A dictionary containing the text detection parameters. )r4r5r8r7r9r;)rOrPrQrRrSrTr~)r]rOrPrQrRrSrTs rdget_text_det_params _OCRPipeline.get_text_det_paramss2 # *&*&B&B #  &"&":":  " *&*&B&B #  ""22O  &"&":":  ($($>$> !2*"2*.   rfinputrYrCc #( # URX#U5n URU 5(dSS0v URUUUUU U 5nU c URn U c URn [ UR U55GHWunnURUR5nU S(a[URUUUS95nOUVs/sHnSU0PM nnUVs/sHnUSPM nn[UR"U40UD65nUVs/sHnUSPM nnUVs/sHnURU5PM nn[URURUU5VVVVs/sH"unnnnUUUUU UUR U U ////S. PM$ nnnnnU (aUH n/US '/US 'M [[#[%U555nUVs/sHn[%UU5S :dMUPM nnU(Ga_/n S /n!UHn[UR'UUUU55n"/n#/n$[U"UU5Hdun%n&U%R(S :dMU%R*S S :dM-U%R*S S :dMBU#R-U%5 U$R-U&5 Mf U$UU'U$UUS'U R/U#5 U!R-U!S [%U#5-5 M U S(arUR1U 5V's/sHCn'[3[4R6"U'S[4R8S9R;5S 5PME n(n'UR=U U(5n OS /[%U 5-n([ U5Hun)nUUnU(U!U)U!U)S -US'M [ U5GHun)nU U!U)U!U)S -n"UUnUUn[ U"5V*V%s/sH1un*n%U*U%R*S [?U%R*S 5- S.PM3 n+n*n%[AU+SS9n,U,V-s/sH n-U"U-SPM n.n-[ URCU.U S95Hun)n/U,U)Sn0U/U+U0S'M [#[%U+55Hn1U+U1Sn/U/SU :dMU (ab[EU/SS UU1U/SS 5un2n3US R-U25 US R-U35 USR-U/SS 5 OUSR-U/S5 USR-U/S5 USR-U/S5 USR-UU15 M GM UHvnUR S:Xa=[GUS5n4U4US 'U (a"US V5s/sHn5[GU55PM sn5US!'O[4RH"/5US '[KU5v Mx GMZ gs snfs snfs snfs snfs snnnnfs snfs sn'fs sn%n*fs sn-fs sn5f7f)"a Predict OCR results based on input images or arrays with optional preprocessing steps. Args: input (Union[str, list[str], np.ndarray, list[np.ndarray]]): Input image of pdf path(s) or numpy array(s). use_doc_orientation_classify (Optional[bool]): Whether to use document orientation classification. use_doc_unwarping (Optional[bool]): Whether to use document unwarping. use_textline_orientation (Optional[bool]): Whether to use textline orientation prediction. text_det_limit_side_len (Optional[int]): Maximum side length for text detection. text_det_limit_type (Optional[str]): Type of limit to apply for text detection. text_det_max_side_limit (Optional[int]): Maximum side length for text detection. text_det_thresh (Optional[float]): Threshold for text detection. text_det_box_thresh (Optional[float]): Threshold for text detection boxes. text_det_unclip_ratio (Optional[float]): Ratio for unclipping text detection boxes. text_rec_score_thresh (Optional[float]): Score threshold for text recognition. return_word_box (Optional[bool]): Whether to return word boxes along with recognized texts. Returns: OCRResult: Generator yielding OCR results for each input image. rxz0the input params for model settings are invalid!Nr))r{r| output_imgdt_polys) input_path page_indexdoc_preprocessor_resrrvtext_det_paramsr2rYrC rec_texts rec_scores rec_polys vis_fonts text_wordtext_word_regionrrr- class_ids)dtypetextline_orientation_angles) sub_img_id sub_img_ratioc US$)NrrG)xs rd&_OCRPipeline.predict..s ?9Krf)keyr)rCrec_res rec_scorerec_textrrrvis_fontrr3 rec_boxestext_word_boxes)&rryrrYrC enumerater[r\ instanceslistrLrXrUrm input_paths page_indexesr2rangerlrVsizeshapernextendrNintnpasarrayint64ravelrfloatsortedrZrrarrayr)6r]rr{r|r-rOrPrQrRrSrTrYrCrvr_ batch_data image_arraysdoc_preprocessor_resultsarritemdoc_preprocessor_images det_results dt_polys_listrrrrresultsresindicesidxall_subs_of_imgs chunk_indicesall_subs_of_img filtered_subsfiltered_polyssub_imgr@textline_angle_infoanglesiimg_idsub_img_info_listsorted_subs_inforsorted_subs_of_imgrrsnoword_box_content_list word_box_listrlines6 rdpredict_OCRPipeline.predict$sF00 (=U ..~>>NO O22 #  #   !   ! ($($>$> !  ""22O&t'9'9%'@AMAz??:+?+?@L45+/22$5Q*;3,(LX+X?G&-Mgs]35G1H11LsgGM#% !" "C&*++3C8-:L'O %'M%'N),_mC>P)Q #LL1, ' a 01 4 ' a 01 4)009*11$7*R*8M#&/=GCL,$++M:!((r):S=O)OP'#,""<=483R3R,4  4/ JJ 3K @#ega) 4  (,'8'89I6'R$ TC(8$99F'0FAs!#,C9?%a(=Q+?:C561(0FAs&6%a(=Q+?'O"#,C,S1H 09/I ) 0JOFG+1-4]]1-=gmmTUFV@W-W0J &)(.)/K($CS*BRQ,8BR'*'0++.,' 7 &6a%8%F CJ)*5i@ ' %S):%;<"3C"8"C";/3HH.GW$+J$7$:$,SM$+J$7$:H" D 5} !$K 0 7 78M N #$6 7 > >} M #K 0 7 7 8KA8N O #K 0 7 7 8K L -44W[5IJ ,33GJ4GH ,33HSMB! =51V>>Y. 7K8H II'0C $&),,>(?2(?4D9(?2-. (*xx|C $n$gB,Y'GN<N8 4)*@2sCX WXW!-"XW&X$W+?)X()W0 AXW8+W81AXX&X;BXA W= BX8X XX-AX DXX %A-X)rVrUr[rLr\r:rCrSrOrPrQrXrRrTrZrYr2rNr)r-)NNNNNN) NNNNNNNNNNN)__name__ __module__ __qualname____firstlineno____doc__rrstrrrboolrrrIrrndarrayrrryr~rrrrr__static_attributes__ __classcell__)rcs@rdrr&s !% $2659AEj2j2 j2  j2  S#X/ j212j2j2U4S>9#<=>j2 j2j2X" $RZZ 0"EI#Y" bjj "H$8 &.tn $D> #+4.    D26-115+//315, !)#, &c], "*# , "% , &e_ ,  (,  , b8<,03715-115+//31515*.W%S$s)RZZbjj1AABW%'/tnW%$D> W% #+4. W% "*# W%&c]W%"*#W%"%W%&e_W% (W% (W%"$W% W%W%rfrocrzocr-core)altc.\rSrSrSr\S5rSrSrg) OCRPipelineiOCRc[$)N)r)r]s rd _pipeline_clsOCRPipeline._pipeline_clssrfc&URSS5$)NrDr)rJ)r]r&s rd_get_batch_sizeOCRPipeline._get_batch_sizeszz,**rfrGN) rrrrentitiespropertyrrrrGrfrdrrsH +rfr)%typingrrrrrnumpyrutilsr utils.depsr common.batch_samplerr common.readerr modelsrrutils.benchmarkr _parallelrbaser componentsrrrrrrresultr time_methodsrrrGrfrdrs4325&6(@ T%<T%T%nJ/+:+0+rf