TAi#~SrSSKJr SSKrSSKrSSKrSSKrSSKrSSKrSSK J r SSK J r SSK J r SSK Jr SSKrSSKr\R$S:a SS K Jr SS K Jr OSS KJrJr S S KJr S SKJr S SKJr S SKJr SrSr/SQr/SQr\\-r "SS5r!"SS\!5r""SS5r#"SS\#5r$"SS5r%"SS 5r&"S!S"\5r'"S#S$5r(g)%z6Contains the core functions to parse tables from PDFs.) annotationsN) itemgetter)Any)Iterable)Iterator) ) TypedDict)Unpack)r r )ImageConversionBackend)build_file_path_in_temp_dir)get_index_closest_point)get_textline_coords leftrightmiddle)topbottomcenterc*\rSrSrSrSrSrSrSrg) TextAlignment,aRepresents a list of textlines sharing an alignment on a coordinate. The alignment can be left/right/middle or top/bottom/center. (PDF coordinate space) Parameters ---------- coord : float coordinate of the initial text edge. Depending on the alignment it could be a vertical or horizontal coordinate. textline : obj the original textline to start the alignment align : str Name of the alignment (e.g. "left", "top", etc) Attributes ---------- coord : float The coordinate aligned averaged out across textlines. It can be along the x or y axis. textlines : array Array of textlines that demonstrate this alignment. align : str Name of the alignment (e.g. "left", "top", etc) c,XlU/UlX0lgN)coord textlinesalignselfrtextliner!s H/var/www/html/land-ocr/venv/lib/python3.13/site-packages/camelot/core.py__init__TextAlignment.__init__Gs " cSR[SURSS55RSS5nSURS[ UR5SUS 3$) Nz | c"UR5$r)get_textxs r%(TextAlignment.__repr__..Ns !**,r( z)joinmapr replacerlen)r# text_insides r%__repr__TextAlignment.__repr__Lsijj &r(: ; '$  tzzl$s4>>/B.CD*m5 2 r(cUR[UR5-U-[[UR5S-5- UlURR U5 g)z%  ! #C    h'r()r!rr N) __name__ __module__ __qualname____firstlineno____doc__r&r8r=__static_attributes__r(r%rr,s4  (r(rc<^\rSrSrSrU4SjrSrSSjrSrU=r $)TextEdge_aDefines a text edge coordinates relative to a left-bottom origin. (PDF coordinate space) An edge is an alignment bounded over a segment. Parameters ---------- coord : float coordinate of the text edge. Can be x or y. y0 : float y-coordinate of bottommost point. y1 : float y-coordinate of topmost point. align : string, optional (default: 'left') {'left', 'right', 'middle'} Attributes ---------- is_valid: bool A text edge is valid if it intersects with at least TEXTEDGE_REQUIRED_ELEMENTS horizontal text rows. cx>[TU]XU5 URUlURUlSUlg)NF)superr&y0y1is_valid)r#rr$r! __class__s r%r&TextEdge.__init__xs/ %0++++ r(c [URS5n[URS5n[URS5nSUSUSUSURSUR S3 $)Nr0z )roundrrKrLr!rM)r#r-rKrLs r%r8TextEdge.__repr__~sd $**a  477A  477A 1#T"T"WTZZL VW X r(c[R"URURUS9(aHURX!5 URUl[ UR 5[ :aSUlggg)zrUpdate text edge coordinates. Update the text edge's x and bottom y coordinates and sets the is_valid attribute. abs_tolTN)mathiscloserKr=r6r TEXTEDGE_REQUIRED_ELEMENTSrM)r#r-r$edge_tols r% update_coordsTextEdge.update_coordss\ <<h ?  * *8 7kkDG4>>"%?? $ @ @r()rMrKrL2) r?r@rArBrCr&r8r\rD __classcell__rNs@r%rGrG_s0   % %r(rGc:\rSrSrSrSr\S5rSrSr Sr g) TextAlignmentsz9Defines a dict of text edges across reference alignments.cB0UlUHn/URU'M gr_text_alignments)r#alignment_namesalignment_names r%r&TextAlignments.__init__s$ "-N46D ! !. 1.r(c[XU5$r)r)rr$r!s r%_create_new_text_alignment)TextAlignments._create_new_text_alignmentsUe44r(c[$r)NotImplementedr# alignmentrr$s r%_update_alignment TextAlignments._update_alignmentsr(cj[U5nURR5Hup4X#n[XTSS9nSnUcSnOFXFRnUS- Us=:a US-:aO OUR XFXQ5 O X:aUS-nOUnUcMgUR XQU5n URXy5 M g)z1Update an existing text edge in the current dict.cUR$r)rr,s r%r.3TextAlignments._register_textline..sQWWr()fnNrg?r )rrgitemsrrrrrlinsert) r#r$coords alignment_idalignment_arrayr idx_closest idx_insert coord_closest new_alignments r%_register_textline!TextAlignments._register_textlines$X.-1-B-B-H-H-J )L(E2+<K J" / < B B 3;<<**'4e#*!,qJ!,J% $ ? ?\!  &&zA9.Kr(rfN) r?r@rArBrCr& staticmethodrlrrrrDrEr(r%rcrcs)C7 55Br(rccr^\rSrSrSrSU4SjjrSrSrSrSr Sr S r S r S r S rS rSrSrU=r$) TextEdgeszDefines a dict text edges on the PDF page. The dict contains the left, right and middle text edges found on the PDF page. The dict has three keys based on the alignments, and each key's value is a list of camelot.core.TextEdge objects. c8>[TU][5 Xlgr)rJr&HORIZONTAL_ALIGNMENTSr[)r#r[rNs r%r&TextEdges.__init__s ./ r(c[XU5$r)rGr"s r%rl$TextEdges._create_new_text_alignments//r(cdURXU5nURURU5 g)z(Add a new text edge to the current dict.N)rlrgr<)r#rr$r!tes r%add TextEdges.adds-  , ,Ue D e$++B/r(c<URX#UR5 gr)r\r[rps r%rrTextEdges._update_alignments?r(cUHAn[UR5R55S:dM0URU5 MC g)z)TextEdges.get_relevant..s',Ib[[!BLL!!,Is66rc3r# UH-nUR(dM[UR5v M/ g7frrrs r%rrs,8B;;"BLL!!877rc3r# UH-nUR(dM[UR5v M/ g7frrrs r%rrs,9B;;"BLL!!9rrrr keyrcUR$r)rMrs r%r.(TextEdges.get_relevant..sbkkr()sumrgmaxrxrlistfilter)r#intersections_sumrelevant_aligns r% get_relevantTextEdges.get_relevants,0,A,A&,I//8 //9 &.446JqMJ1M )4+@+@+P Q  r(cURSS9 URU5nURX15 URU5nUVs0sHoPR XT5S_M nnU$s snf)a Return a dict of interesting table areas on the PDF page. The table areas are calculated using relevant text edges. Parameters ---------- textlines : list List of text line objects that are relevant for determining table areas. relevant_textedges : list List of relevant text edge objects used to identify table areas. Returns ------- dict A dictionary with padded table areas as keys and None as values. c4UR*UR4$r)rKrrs r%r.+TextEdges.get_table_areas..s/Ar(rN)sort_initialize_table_areas"_extend_table_areas_with_textlines"_calculate_average_textline_height_pad)r#r relevant_textedges table_areasaverage_textline_heightareapadded_table_areass r%get_table_areasTextEdges.get_table_areass}& $AB223EF  // G#'"I"I)"TGR GRtIId 4d :{  "!  sA'c0nUHMnU(d2SX#RURURUR4'M<URX#5 MO U$)a  Initialize table areas based on relevant text edges. Parameters ---------- relevant_textedges : list List of relevant text edge objects used to initialize table areas. Returns ------- dict A dictionary of table areas initialized from relevant text edges. N)rrKrL_update_table_areas)r#rrrs r%r!TextEdges._initialize_table_areas#sM $BBF XXruubhh>?((9 % r(cSnUH/nURUS:dMURUS::dM-Un O Uc1SXRURURUR4'gURU5 US[ URUS5[ USUR5[ USUR54nSX'g)z Update table areas by checking for overlaps with new text edges. Parameters ---------- table_areas : dict Current table areas to be updated. te : object The new text edge object to check for overlaps. Returns ------- None Nr rrr0)rLrKrpopminr)r#rrfoundr updated_areas r%rTextEdges._update_table_areas:sDuuQBEET!W$4  =>BK255"((BEE: ; OOE "aBEE58$E!Hbhh'E!Hbee$ L )-K %r(c|UHnSnUH/nURUS:dMURUS::dM-Un O UcM?URU5 [URUS5[URUS5[ USUR 5[ USUR54nSX'M g)z Extend table areas based on text lines that overlap vertically. Parameters ---------- table_areas : dict Current table areas to be extended. textlines : list List of text line objects relevant for extending table areas. Returns ------- None Nr rrr0)rKrLrrx0rx1)r#rr rrrrs r%r,TextEdges._extend_table_areas_with_textlines\sBE#55DG#a(8 E $  &uQx(uQx(a"%%(a"%%(   -1 )!r(cf[SU55nU(aU[[U55- $S$)z Calculate the average height of text lines. Parameters ---------- textlines : list List of text line objects. Returns ------- float The average height of the text lines. c3R# UHoRUR- v M g7fr)rLrK)rrs r%r?TextEdges._calculate_average_textline_height..s!D)B%%"%%-)s%'r)rr;r6)r#r sum_textline_heights r%r,TextEdges._calculate_average_textline_height}s1"!D)!DD>G"U3y>%::NQNr(chUS[- nUS[- nUS[-nUSUS--nX4XV4$)z Pad a given area by a constant value. Parameters ---------- area : tuple The area to be padded defined as (x0, y0, x1, y1). average_row_height : float The average height of rows to use for padding. Returns ------- tuple The padded area. rr r0r)TABLE_AREA_PADDING)r#raverage_row_heightrrKrrLs r%rTextEdges._padsP !W) ) !W) ) !W) ) !W)A- -r()r[r^)r?r@rArBrCr&rlrrrrrrrrrrrrDr`ras@r%rrsN!00 @,  <"@. -D1BO"  r(rc\rSrSrSrSrSr\S5r\RS5r\S Sj5r \S Sj5r \S 5r S r g ) Cellia^Defines a cell in a table. With coordinates relative to a left-bottom origin. (PDF coordinate space) Parameters ---------- x1 : float x-coordinate of left-bottom point. y1 : float y-coordinate of left-bottom point. x2 : float x-coordinate of right-top point. y2 : float y-coordinate of right-top point. Attributes ---------- lb : tuple Tuple representing left-bottom coordinates. lt : tuple Tuple representing left-top coordinates. rb : tuple Tuple representing right-bottom coordinates. rt : tuple Tuple representing right-top coordinates. left : bool Whether or not cell is bounded on the left. right : bool Whether or not cell is bounded on the right. top : bool Whether or not cell is bounded on the top. bottom : bool Whether or not cell is bounded on the bottom. text : string Text assigned to cell. cXlX lX0lX@lX4UlX4UlX24UlX44UlSUlSUl SUl SUl SUl g)NFr2) rrLx2y2lbltrbrtrrrr_textr#rrLrrs r%r& Cell.__init__s]((((    r(c [UR5n[UR5n[UR5n[UR5nSUSUSUSUS3 $)Nz >r() rrrrrrrrrrrrLrN)returnbool)r?r@rArBrCr&r8propertyrsetterrrrrDrEr(r%rrsz%N 9 [[..////??r(rc\rSrSrSrSrSrSr\S5r \S5r Sr S r SS jr S rS rS rSrSrSrSSjrSrSrSrSrSrSrSrSrSrg)Tableia|Defines a table with coordinates relative to a left-bottom origin. (PDF coordinate space) Parameters ---------- cols : list List of tuples representing column x-coordinates in increasing order. rows : list List of tuples representing row y-coordinates in decreasing order. Attributes ---------- df : :class:`pandas.DataFrame` shape : tuple Shape of the table. accuracy : float Accuracy with which text was assigned to the cell. whitespace : float Percentage of whitespace in the table. filename : str Path of the original PDF order : int Table number on PDF page. page : int PDF page number. cXlX lUVVs/sH-o1Vs/sHn[USUSUSUS5PM snPM/ snnUl[R "5UlSUlSUlSUl SUl SUl SUl SUl SUlSUlSUlSUlSUlSUlgs snfs snnf)Nrr )rr)colsrowsrcellspd DataFramedfshapeaccuracy whitespacefilenameorderpageflavorpdf_size_bboxparse parse_details_image _image_path)r#rrrcs r%r&Table.__init__s  LPQDqTBTtAaD!A$!ad3TBDQ ,,.         ! CQs C $C C C cPSURRSURS3$)zXReturn a string representation of the class . Returns ------- [type] [description] >**+74::,a@@r(cURUR:XaURUR:agURUR:agg)zReturn True if the two pages are less than the current page . Parameters ---------- other : [type] [description] Returns ------- [type] [description] TN)rr)r#others r%__lt__ Table.__lt__:s? 99 "zzEKK' 99uzz ! "r(c/nURH<nURUVs/sHo3RR5PM sn5 M> U$s snf)z1Returns two-dimensional list of strings in table.)rr<rr)r#drowcells r%data Table.dataMsE ::C HHC8CDiioo'C8 99s#A c[URS5[URS5URURS.nU$)z`Returns a parsing report. with % accuracy, % whitespace, table number on page and page number. r0)rrrr)rSrrrr)r#reports r%parsing_reportTable.parsing_reportUs<dmmQ/3ZZII    r(cfURcURcg[[RR UR 5S5Ul[SS9nURUR UR5 [R"UR5UlUR$)zCompute pdf image and cache it.z.pngT) use_fallback) r r rospathbasenamerr convertcv2imread)r#backends r% get_pdf_imageTable.get_pdf_imagees ;; '#>GG$$T]]3V$ 1dC t/?/?@**T%5%56DK{{r(cvURH(nUHnS=Ul=Ul=UlUlM! M* U$)zSet all table edges to True.T)rrrrr)r#rrs r% set_all_edgesTable.set_all_edgesqs=::CBFF FDJFDK r(cJURX5 URX#5 U$)aqSet the edges of the joint. Set a cell's edges to True depending on whether the cell's coordinates overlap with the line's coordinates within a tolerance. Parameters ---------- vertical : list List of detected vertical lines. horizontal : list List of detected horizontal lines. joint_tol : int, optional Tolerance for determining proximity, by default 2 )_set_vertical_edges_set_horizontal_edges)r#vertical horizontal joint_tols r% set_edgesTable.set_edgesxs%   5 "":9 r(cn[U5H&upE[R"X%SUS9(dM$Us $ g)NrrV) enumeraterXrY)r#rzrr2irs r%_find_close_pointTable._find_close_points2f%DA||EQ4;;&r(c4UHnURURUSU5nUcM(URURUSU5nUc[UR5nURURUSU5nUR XEU5 M g)Nrr r)r8rr6r_update_vertical_edges)r#r0r2vstartendr7s r%r.Table._set_vertical_edgessA**499adIFE}((AaD)DC{$))n&&tyy!A$ BA  ' 'A 6r(cUcI[UR5S- nUS:a*[X5HnSURUUlM ggUS:Xa*[X5HnSURUSlM g[X5H4nSURUUlSURUUS- lM6 gNr rT)r6rrangerrrr#r=r>indexjs r%r;Table._update_vertical_edgess = NQ&Ezu*A15DJJqM%(.+aZ5&(, 1 a %'5&,0 1 e$)15 1 eai(.'r(c4UHnURURUSU5nUcM(URURUSU5nUc[UR5nURURUSU5nUR XEU5 M g)Nrr0r )r8rr6r_update_horizontal_edges)r#r1r2hr=r>r7s r%r/Table._set_horizontal_edgessA**499adIFE}((AaD)DC{$))n&&tyy!A$ BA  ) )%a 8r(cUcI[UR5S- nUS:a*[X5HnSURUUlM ggUS:Xa*[X5HnSURSUlM g[X5H4nSURUUlSURUS- UlM6 grA)r6rrBrrrrCs r%rHTable._update_horizontal_edgess = NQ&Ezu*A26DJJu%a(/+aZ5&'+ 1 a $'5&+/ 5!!$(26 519%a(/'r(c^[UR5n[UR5mUS:XdTS:XaU$[UR5U:wd$[ U4SjUR55(a [ S5e[ U5H4nSURUSlSURUTS- lM6 [ T5H4nSURSUl SURUS- Ul M6 U$)z Sets table border edges to True.rc3@># UHn[U5T:gv M g7frr6)rrnum_colss r%r#Table.set_border..s. ,6SCH JszQInconsistent cells structure: cells should match the dimensions of rows and cols.Tr ) r6rrrany ValueErrorrBrrrr)r#num_rows row_index col_indexrPs @r% set_borderTable.set_borderstyy>tyy> q=HMK tzz?h &#. ,0JJ. + + c  xI,0DJJy !! $ )8 (A&   ( 7 6 9 7:6F(F( #!">" ! r(rcB\rSrSr%SrS\S'S\S'S\S'S\S'S rg ) _Kwiz.Helper class to define file related arguments.zos.PathLike[Any] | strr"strdirnamerootextrEN)r?r@rArBrC__annotations__rDrEr(r%rrs8  L I Hr(rc\rSrSrSrSSjrSrSrSrSSjr SSjr \ S 5r \ SS j5rSSS jjrSS jrSSSjjrSrg ) TableListizDefines a list of camelot.core.Table objects. Each table can be accessed using its index. Attributes ---------- n : int Number of tables in the list. cXlgr_tables)r#tabless r%r&TableList.__init__s(. r(cPSURRSURS3$)Nrz n=rR)rNr?nrs r%r8TableList.__repr__s&4>>**+3tvvha88r(c,[UR5$r)r6rrs r%__len__TableList.__len__s4<<  r(c URU$rr)r#idxs r% __getitem__TableList.__getitem__s||C  r(c,[UR5$r)iterrrs r%__iter__TableList.__iter__sDLL!!r(c[U5$r)nextrs r%__next__TableList.__next__s Dzr(c [USU35$)Nto_)getattr)tablerws r% _format_funcTableList._format_funcsuA3i((r(c[U5$)z!The number of tables in the list.rOrs r%r TableList.ns4yr(Nc USnUSnUSnURH[nUSURSURU3n[RR X75nUR Xa5n U "U5 M] g)Nrrr-page-r{)rrrr!r"r3r) r#rwrlrrrrrfilepath to_formats r% _write_fileTableList._write_filesx#f~Um\\Evejj\ cUKHww||G6H))%3I h  "r(c USnUSnUSnUSn[RR[RRU5U5S-n[R "USSS9nUR HqnUS URS URU3n [RRX95n URU [RRU 55 Ms SSS5 g!,(df  g=f) Nr"rrr.ziprrT allowZip64rr{) r!r"r3rzipfileZipFilerrrrur#) r#rlr"rrrzipnamezrrrs r% _compress_dirTableList._compress_dirsf~#f~Um'',,rwwt4d;fD __Wcd ;q"V6%**WU[[M#O77<<:"''"2"28"<=&< ; ;s +BC66 Dc[RRU5n[RRU5n[RR U5upgU(a[ R "5nXXgS.nUS;a/UR"SSU0UD6 U(aUR"S0UD6 g g US:XGa$[RRXE5n [R"U 5n URH8n SU RSU R3n U RR!XS9 M: U R#5 U(a[RR[RRU5U5S-n [$R&"U S S S 9nUR)U [RRU 55 S S S 5 g g US :Xa[RRXE5n URHn U R+U 5 M U(a[RR[RRU5U5S-n [$R&"U S S S 9nUR)U [RRU 55 S S S 5 g g g !,(df  g =f!,(df  g =f)zExport the list of tables to specified file format. Parameters ---------- path : str Output filepath. f : str File format. Can be csv, excel, html, json, markdown or sqlite. compress : bool Whether or not to add files to a ZIP archive. )r"rrr)csvhtmljsonmarkdownrwexcelrzr{)r|rrrTrNsqliterE)r!r"rr#splitexttempfilemkdtemprrr3rr}rrrrr~rrrrur)r#r"rwcompressrr#rrrlrrrr|rrs r%exportTableList.exports''//$'77##D)GG$$X.  &&(G#R 3 3    +q +F +"",V, '\ww||G6H^^H-F$UZZL }E !!&!@& LLN'',,rwwt'+B+Br(r))rC __future__rrXr!rsysrroperatorrtypingrrrr%pandasr version_infor r typing_extensionsbackendsr utilsrrrrZrrVERTICAL_ALIGNMENTSALL_ALIGNMENTSrrGrcrrrrrrEr(r%rs<"   w 3,.*& 41&)<<0(0(f3%}3%l/B/Bd\ \ ~S?S?lLL^  ) hBhBr(