Ng\dZddlmZGddZdZdZdZdd Z d d Zd!d Z dZ dZ ddZ ddZ d"dZdZdZdZdZd#dZdZdZdZdZdS)$z* Copyright (C) 2021 Microsoft Corporation ) defaultdictc.eZdZddZdZdZdZdZdS)RectNc|d|_d|_d|_d|_dS|d|_|d|_|d|_|d|_dS)Nrx_miny_minx_maxy_max)selfbboxs k/var/www/html/ai-engine/env/lib/python3.11/site-packages/unstructured_inference/models/table_postprocess.py__init__z Rect.__init__ sW <DJDJDJDJJJaDJaDJaDJaDJJJcT|j|jz |j|jz z}|dkr|ndS)z$Calculates the area of the rectanglerg)r r rr )rareas rget_areaz Rect.get_areas0 TZ'DJ,CDaxxttS(rc8|dkr1|j|_|j|_|j|_|j|_nt |j|j|_t |j|j|_t |j|j|_t |j|j|_|j|jks(|j|jks|dkrd|_d|_d|_d|_|S)z2Calculates the intersection with another rectangler)rr r r rmaxmin)rothers r intersectzRect.intersects ==??a  DJDJDJDJJTZ55DJTZ55DJTZ55DJTZ55DJzDJ&&$*tz*A*AT]]__XYEYEY     rct|}|dkr2|j|_|j|_|j|_|j|_|St |j|j|_t |j|j|_t|j|j|_t|j|j|_|S)z4Calculates a rectangle that includes both rectanglesr)rrr r r rrr)rrrs r include_rectzRect.include_rect/sT  ==??a  DJDJDJDJKU[11 U[11 U[11 U[11  rc6|j|j|j|jgS)z1Returns the coordinates that define the rectangler )rs rget_bboxz Rect.get_bboxGs DJ DJ??r)N)__name__ __module__ __qualname__rrrrrrrrrsi ! ! ! !))) *0@@@@@rrc fd|DS)z3 Filter out objects below a certain score. c,g|]}|dk|S)scorer#).0obj thresholds r z#apply_threshold..Ps' @ @ @Cc'li&?&?C&?&?&?rr#)objectsr)s `rapply_thresholdr,Ls  A @ @ @7 @ @ @@rct|dkr#t||d}t||nt|ddd}t|dkrt |}|S)z^ Apply operations to the detected rows, such as thresholding, NMS, and alignment. r?overlap_thresholdobject2_overlapTmatch_criteriamatch_threshold keep_higherr)lennms_by_containmentremove_objects_without_contentnmssort_objects_top_to_bottom)rowstokensscore_thresholds r refine_rowsr>Sss  6{{Q!$#FFF&vt44444(93\`aaa 4yy1}})$// Krct|dkr#t||d}t||nt|ddd}t|dkrt |}|S) za Apply operations to the detected columns, such as thresholding, NMS, and alignment. rr.r/r1g?Tr2r)r6r7r8r9sort_objects_left_to_right)columnsr<r=s rrefine_columnsrBds  6{{Q$WfLLL&vw7777 ,       7||a,W55 Nrr.c t|}t|}d|D t|||d\}}}td|D]}t ||}t|dkrd |<t|D]J} |s@t ||} t|| dkrd |<K fdt |D} | S)z_ Non-maxima suppression (NMS) of objects based on shared containment of other objects. cg|]}dSFr#r'r(s rr*z&nms_by_containment..s666S5666rF)r0forced_assignmentrrTc*g|]\}}| |Sr#r#r'idxr( suppressions rr*z&nms_by_containment..s(^^^XS#[Y\M]^S^^^r)sort_objects_by_scorer6slot_into_containersrangeset intersection enumerate) container_objectspackage_objectsr0 num_objectspackages_by_container_ object2_numobject2_packages object1_numobject1_packages final_objectsrKs @rr7r7zsA..?@@'((K66$5666K"6+ ###1aQ ,,44 4[ABB  A % %'+K $ -- 4 4K{+ 4#&'<['I#J#J '445EFFGG!KK/3K ,  4 _^^^3D)E)E^^^M rFc|g}d|D}d|D}t|dkst|dkr|||fStt}t|D]N\}} g}t | d} | } t|D]x\} } t | d}|t | d}| dkr|| z }|| | |dyt|dkrt|}|d}||d|s |d|krB||d||||dP|||fS)z Slot a collection of objects into the container they occupy most (the container which holds the largest fraction of the object). cg|]}gSr#r#)r' containers rr*z(slot_into_containers..sCCCIRCCCrcg|]}gSr#r#)r'packages rr*z(slot_into_containers..s==='2===rrr)r^ container_numr&r&ra) r6rdictrQrrrappendrL)rRrSr0rGbest_match_scorescontainer_assignmentspackage_assignments match_scores package_numr` package_rect package_arearar^container_rectintersect_areaoverlap_fractionsorted_match_scoresbest_match_scores rrMrMsCC1BCCC==_=== ""c/&:&:a&?&?$&9;LLLt$$L )/ : :[[ W GFO,, #,,.. (12C(D(D   $M9!)F"344N+55d76?6K6KLLUUWWNa#1L#@ ##%.)6!1 |  q "7 "E"E 215   $ $%5g%> ? ? ?  [$4W$=AR$R$R%&6&GHOOP[\\\#K0778H8YZZZ "57H HHrTc(t|d|S)zG Put any set of objects in order from high score to low score. c|dS)Nr&r#ks rz'sort_objects_by_score..s 7r)keyreversesorted)r+rvs rrLrLs '33W E E EErc|ddD]U}t||d\}}t|dkr||VdS)z~ Remove any objects (these can be rows, columns, supercells, etc.) that don't have any text associated with them. Nrr)extract_text_inside_bboxr6stripremove) page_spansr+r( object_textrVs rr8r8sn qqqz  1*c&kJJ Q {  "" # #q ( ( NN3     rcLt||}t|d}||fS)z1 Extract the text inside a bounding box. T)remove_integer_superscripts)get_bbox_span_subsetextract_text_from_spans)spansr bbox_spans bbox_texts rrzrzs1&eT22J' PTUUUI j  rclg}|D].}t|d||r||/|S)z Reduce the set of spans to those that fall within a bounding box. threshold: the fraction of the span that must overlap with the bbox. r)overlapsrc)rrr) span_subsetspans rrrsK K%% DL$ 2 2 %   t $ $ $ rctt|}|}|dkrdS|tt||z |kS)zN Test if more than "threshold" fraction of bbox1 overlaps with bbox2. rF)rlistrr)bbox1bbox2r)rect1area1s rrrsj e  E NN  E zzu ??4U ,, - - 6 6 8 85 @I MMrcR|rdnd}|dd}|r^|D][}d|vr|d}|dzrG|dr||Vd|d<\t|d krdS|d |d |d g}|d dg}t |dd|ddD]\} } | d| dks| d| dkr||} t| d kr>| ddkr2t| dkr| ddkr | ddks|s| dz } || | dg}|| d||} || ||S)zT Convert a collection of page tokens/words/spans into a single text string.  NflagsrtextT superscriptrc|dS)Nspan_numr#rs rrtz)extract_text_from_spans.. T*%5rruc|dS)Nline_numr#rs rrtz)extract_text_from_spans..rrc|dS)N block_numr#rs rrtz)extract_text_from_spans..s T+%6rrr-)r{isdigitr|r6sortzipjoinrc) rjoin_with_spacer join_char spans_copyrr line_textsline_span_textsspan1span2 line_texts rrrsm '.BIqqqJ" / / /Dd""MEt| /<%%''//11/%%d++++*.D' :!rOO55O666OO55O666OO66O777J!!}V,-OJssOZ^<< 2 2 u  {!3 3 3uZ7HER\L]7]7]!77==??II""bMS((Y!++ " 0D0DSUZ]I]I]'J^S   i ( ( ($V}oOO  " "5= 1 1 1 1//Ii   >>* % % + + - --rc&t|dS)z6 Put the objects in order from left to right. c<|dd|ddzS)Nrrrr#rrs rrtz,sort_objects_left_to_right..3ailQvYq\&Arrrwobjss rr@r@/ $AA B B BBrc&t|dS)z6 Put the objects in order from top to bottom. c<|dd|ddzS)Nrrr r#rrs rrtz,sort_objects_top_to_bottom..:rrrrwrs rr:r:6rrc |D]$}|d|dd<|d|dd<%n)#t$r}td|Yd}~nd}~wwxYw|S)zd For every column, align the top and bottom boundaries to the final table bounding box. rrr zCould not align columns: N Exceptionprint)rArcolumnerrs r align_columnsr=s   ( (F $QF6N1  $QF6N1   (     /#//000   N'* AA  Ac |D]$}|d|dd<|d|dd<%n)#t$r}td|Yd}~nd}~wwxYw|S)za For every row, align the left and right boundaries to the final table bounding box. rrrzCould not align rows: Nr)r;rrowrs r align_rowsrMs   % %C!!WCKN!!WCKNN %     ,s,,---   Krr1皙?c t|dkrgSt||}t|}d|D td|D]}t||d}|}t|D]} |st||d} | } | |} |dkr| | z } n|dkr| |z } n|dkr | | |z| z z } | |krd |<n#t $rYwxYw fd t|DS) a2 A customizable version of non-maxima suppression (NMS). Default behavior: If a lower-confidence object overlaps more than 5% of its area with a higher-confidence object, remove the lower-confidence object. objects: set of dicts; each object dict must have a 'bbox' and a 'score' field match_criteria: how to measure how much two objects "overlap" match_threshold: the cutoff for determining that overlap requires suppression of one object keep_higher: if True, keep the object with the higher metric; otherwise, keep the lower r)rvcg|]}dSrEr#rFs rr*znms..os,,,S5,,,rrrobject1_overlapr1iouTc*g|]\}}| |Sr#r#rIs rr*znms..s' K K KHC+c:J KC K K Kr)r6rLrNrrrZeroDivisionErrorrQ)r+r3r4r5rTrW object2_rect object2_arearY object1_rect object1_arearlmetricrKs @rr9r9]s 7||q #G[AAAGg,,K,,G,,,KQ ,, GK0899 #,,..  --  K{+ #GK$8$@AA +4466 !-!7!7 !E!E!N!N!P!P %):::!/,!>'+<<'500!/<,3NQ_3_!`0037 K01)D $ L K K K ' 2 2 K K KKs44D++ D87D8c  g}|D]}d|d<d}d}t}t}tD]\} } | dd| ddz } |dd|ddz } t| dd|dd} t| dd|dd}|| z }d|vrt|| z || z }n|| z }|dkr7d| vr| dr|| || d|d<t |d krPt |d kr=t |t |krt}nt}t |d krd |d<nd|vr||}|D]A} |t| d} || d}B| g}tD]\}}|dd |dd z }|dd |dd z }t|dd |dd }t|dd |dd }||z }d|vr$t||z ||z }|dr|d z}n||z }|dkrH| ||t|d}||d}|4| | }||d<t |d krjt |d krVt |dkst |dkr/t||d <||d <| |d|vr|drt |d dkrtd t|d D]} | g|d |dd d}fd|d D}fd|d D}td|Dtd|Dtd|Dtd|Dg}||d<| ||S)z For each supercell, align it to the rows it intersects 50% of the height of, and the columns it intersects 50% of the width of. Eliminate supercells for which there are no rows and columns it intersects 50% with. FheaderNrr rrr.rTr row_numberscolumn_numbersr&)rrr& propagatedc g|] }| Sr#r#)r'rJrAs rr*z$align_supercells..s,a,a,acWS\,a,a,arc g|] }| Sr#r#)r'rJr;s rr*z$align_supercells..s)X)X)X$s))X)X)Xrc*g|]}|ddS)rrr#r'rs rr*z$align_supercells..!SSS6VF^A.SSSrc*g|]}|ddS)rrr#r'rs rr*z$align_supercells.. JJJS[^JJJrc*g|]}|ddS)rrr#rs rr*z$align_supercells..rrc*g|]}|ddS)rr r#rs rr*z$align_supercells..rr)rOrQrraddr6unionrrrcrrrrN) supercellsr;rAaligned_supercells supercell row_bbox_rect col_bbox_rectintersecting_header_rowsintersecting_data_rowsrow_numr row_heightsupercell_heightmin_row_overlapmax_row_overlapoverlap_heightrmintersecting_rowsintersecting_colscol_numcol col_widthsupercell_widthmin_col_overlapmax_col_overlap overlap_widthsupercell_bbox new_supercellnew_supercell_columnsnew_supercell_rowsrs `` ralign_supercellsrs e=e= # (  #&55 !$%dOO 8 8LGSVQ#f+a.8J(03i6G6JJ !#f+a.)F2CA2FGGO!#f+a.)F2CA2FGGO,>N""#&"Z/"%55$$  $2J#> 3&&s??s8}?,009999*..w777$ ( % & & * *s3K/L/Lq/P/P)**S1I-J-JJJ+.55((),& ' ( (1 , ,"&Ih   y 2889QRR( R RG$ $T']6%: ; ; - : :4=;P Q Q  %g.. L LLGSF AVQ7I'/2Yv5Fq5IIO!#f+a.)F2CA2FGGO!#f+a.)F2CA2FGGO+o=M""#&}y'@-RaBa#b#b X&<'7!';$#09#< 3&&!((111 ($(V$5$5MM$1$>$>s6{$K$KM  &00??HHJJ* & ! " "Q & &%&&**&''!++s3D/E/E/I/I'+,='>'>Im $*;I& '  % %i 0 0 0""y':"s9M]C^?_?_bc?c?c$QIm,D(E(EFF==G(/y*34D*E!*7!3&* %%M -b,a,a,aYO_E`,a,a,a))X)X)X)Xy?W)X)X)X&SS=RSSSTTJJ7IJJJKKSS=RSSSTTJJ7IJJJKK D -1M&)&--m<<<< rct|}t|}d|Dtd|D]}||}t|D]}||}t||t|ddkrt|ddks2t|ddkst|ddkrd|<fdt |DS) a A NMS scheme for supercells that first attempts to shrink supercells to resolve overlap. If two supercells overlap the same (sub)cell, shrink the lower confidence supercell to resolve the overlap. If shrunk supercell is empty, remove it. cg|]}dSrEr#r'rs rr*z"nms_supercells..s555Y5555rrrrrrTc*g|]\}}| |Sr#r#rIs rr*z"nms_supercells..s' N N NHC[=M NC N N Nr)rLr6rNremove_supercell_overlaprQ)rnum_supercellssupercell2_num supercell2supercell1_num supercell1rKs @rnms_supercellsrs'z22J__N55*555K>22 / // #N33 = =N#N3J $Z < < < < M* + +a / /C CS8T4U4UXY4Y4Y:m,--22:./00A55*.K ' N N N N * 5 5 N N NNrcd|D}t|}|ddD]}tt}t|d}|D]s}t |d}||krVt |dt |dr|dD]}||xxdz cc<ttd|D]%}||dkr||n&dS)aW Make sure no supercell in the header is below more than one supercell in any row above it. The cells in the header form a tree, but a supercell with more than one supercell in a row above it means that some cell has more than one parent, which is not allowed. Eliminate any supercell that would cause this to be violated. c*g|]}d|v|d|S)rr#rs rr*z)header_supercell_tree..s4Y1F1F9U]K^1F 1F1F1FrNrrrr) rLrintrrrOissubsetrNr|) rheader_supercellsheader_supercellancestors_by_rowmin_rowheader_supercell2max_row2row2rs rheader_supercell_treersY#-..?@@-aaa0  &s++&}566!2 0 0 ,];<\>\]]K$4566CC J' ())N k  Q  3~#6#6#:#: z-( ) )C ;K0L,M,M M MZ(89::JZ(89::J^++%%j111+,33J????~--%%j111+,33J????/1 +,!$*]344G*]344G+%%""7+++=)009999K''""7+++=)009999,. =)!ee 9 k  Q  3~#6#6#:#:#:#:  #:#:  rN)r.)r.F)T)TT)r1rT)__doc__ collectionsrrr,r>rBr7rMrLr8rzrrrr@r:rrr9rrrrr#rrrs$#####A@A@A@A@A@A@A@A@HAAA",@ /I/I/I/IdFFFF   !!!    NNNN/./././.dCCCCCC       *L*L*L*LZooodOOO681$1$1$1$1$r