.i<\ vUdZddlZddlZddlZddlmZddlmZmZm Z m Z m Z m Z m Z mZmZmZmZmZmZmZmZmZddlmZmZerddlmZddlZdZeejeej fZ!eee fZ"Gdd Z#d ed e$fd Z%d e&d efdZ'dede(d efdZ) d\dee$efdeded efdZ*de(de(de(d e(fdZ+de(de(de(de$d e$f d Z,d!e(de(de(de(de$d e$f d"Z-ee.e.fZ/ee.e.e.e.fZ0ee.e.e.e.e.e.fZ1eeeeee.e.feee.e.e.e.feee.e.e.e.e.e.ffZ2d#Z3e1e4d$<d ed e0fd%Z5d&e1d'e1d e1fd(Z6d)e1d*e/d e1fd+Z7d)e1d*e/d e/fd,Z8d)e1d-e0d e0fd.Z9d)e1d*e/d e/fd/Z:d0e&d e;fd1Z<ed2Z=d3e e=d ee=fd4Z>d!e e=ge;fd3e e=d eee=ee=ffd5Z?d6e.d7e.d8e(d e@fd9ZAd:e e/d e0fd;ZB d]dee=d ee=fd?ZCd@e(ds1 < &   # # % % % % % & &r,) __name__ __module__ __qualname____doc__ FileOrNamerr+r'r0objectr6r,r*rr)s N NC N3 N4 N N N N!5!!!!&&&&&&T&&&&&&r,rin_strrct|ts$Jtt||S)z'Converts to bytes, encoding to unicode.)rr!r(encode)r>s r*make_compat_bytesrACs< fc " "55CV $5$5555 ==??r,oct|trOtj|} ||dS#t $rt |cYSwxYwt |S)z>Converts everything to string, if bytes guessing the encoding.encoding)rbytescharset_normalizerdetectdecodeUnicodeDecodeErrorr!)rBencs r*make_compat_strrKIsv!U '** 88C O,, ,!   q66MMM 1vv sAA"!A"ssizec|dkr |d|St||kr |dz dz}|d|d|| dS|S)Nz ... )len)rLrMlengths r* shorten_strrTUs_ axx$x 1vv}}(qGVG*001fWXX;000r,utf-8ignore bytesorstringrD erractionct|tr|St|ts$Jtt||||S)z^When Py2 str.encode is called, it often means bytes.encode in Py3. This does either. )rr!rEr(rH)rWrDrXs r*compatible_encode_methodrZ_s`-%% mU + +EESm1D1D-E-EEEE   ) 4 44r,leftabove upper_leftc||z|z }t||z }t||z }t||z }||kr||kr|S||kr|S|Sr.)abs)r[r\r]ppapbpcs r*paeth_predictorrdnsm u z!A QXB QYB Q^  B RxxB"HH r r,colorscolumnsbitspercomponentdatacx|dkrd|}t|||dzz}||z}g}tdt||D]d}g} t|D];} ||| z} | |kr| | | |z z } | dz} | | <|| et |S)zReverse the effect of the TIFF predictor 2 Documentation: https://www.itu.int/itudoc/itu-t/com16/tiff-fx/docs/tiff6.pdf (Section 14, page 64) z Unsupported `bitspercomponent': r)rrangerRappendextendrE) rerfrgrh error_msgbppnbytesbuf scanline_irawi new_values r*apply_tiff_predictorrws1I7GII I&&& $) *C s]FCAs4yy&11 v " "AZ!^,ICxxSS\) S JJy ! ! ! ! 3 ::r,predc|dvrd|z}t|||z|zdz}||zdz}g}td|z} tdt||dzD]} || } || dz| dz|z} g} | dkrt| } n| dkrXt | D]F\}}||z dkrd}nt | ||z }||zdz}| |Gnb| dkr5t| | D]"\}}||zdz}| |#n'| d krrt | D]a\}}||z dkrd}nt | ||z }t | |}|||zdzzdz}| |bn| d krt | D]\}}||z dkrd}d}n0t | ||z }t | ||z }t | |}t|||}||zdz}| |ntd | z| | | } t|S) zxReverse the effect of the PNG predictor Documentation: http://www.libpng.org/pub/png/spec/1.2/PNG-Filters.html )rjz"Unsupported `bitspercomponent': %drjrrzrQzUnsupported predictor value: %d) rlistrlrR enumerateintrmziprdrnrE)rxrerfrgrhmsgrqrprr line_abovers filter_type line_encodedrtjsub_x raw_x_bppraw_xup_xprior_x average_xpaeth_x prior_x_bpppaeths r*apply_png_predictorrs%v%%25EEC   g  0 0A 5F # #q (C Cg'((JAs4yy&1*55NN :& JNZ!^f-DDE  !  |$$CC A  &l33 " "5s7Q;; !II #CCL 1 1I*c1 5!!!!  "A  "%\:!>!> " " g3. 5!!!! "A  !*, 7 7 " " 9s7Q;; !II #CCL 1 1Ijm,,"i'&9a%??3F 5!!!! "A  ( 55 " " 7s7Q;; !I"#KK #CCL 1 1I"%jS&9":":Kjm,,' 7KHH 5C/ 5!!!! " AK OPP P 3 ::r,)rzrrrzrrMATRIX_IDENTITYc |\}}}}t|t|t|t|fS#t$rtdwxYw)NzCould not parse rectangle)float ValueErrorr)rBx0y0x1y1s r* parse_rectr sf9RRRyy%))U2YYb 99 99978889s AAAm1m0c|\}}}}}}|\}} } } } } ||z| |zz| |z| |zz||z| |zz| |z| |zz||z| |zz| z| |z| |zz| zfSr.r=)rra1b1c1d1e1f1a0b0c0d0e0f0s r* mult_matrixrs!RRR!RRR5 R"r' R"r' R"r' R"r' R"r'B R"r'B  r,mvc^|\}}}}}}|\}} ||||||z| |zz|z||z| |zz|zfS)zTranslates a matrix by (x, y) inside the projection. The matrix is changed so that its origin is at the specified point in its own coordinate system. Note that this is different from translating it within the original coordinate system.r= rrabcdefxys r*translate_matrixr#sT Q1aA FQ aAq1uq1u}q(!a%!a%-!*; ;;r,cV|\}}}}}}|\}} ||z|| zz|z||z|| zz|zfS)zApplies a matrix to a point.r=rs r*apply_matrix_ptr.sJQ1aA FQ q51q5=1 a!ea!ema/ //r,rectc T|\}}}}||f}||f}||f}||f} t||\} } t||\} } t||\}}t|| \}}t| || |t| | ||t| || |t| | ||fS)aApplies a matrix to a rectangle. Note that the result is not a rotated rectangle, but a rectangle with the same orientation that tightly fits the outside of the rotated content. :param m: The rotation matrix. :param rect: The rectangle coordinates (x0, y0, x1, y1), where x0 < x1 and y0 < y1. :returns a rectangle with the same orientation, but that would fit the rotated content. )rminmax)rrrrrr left_bottom right_bottom right_topleft_topleft1bottom1right1bottom2right2top1left2top2s r*apply_matrix_rectr5sRRr(K8LRIBxH&q+66UG'<88VW$Q 22NVT#Ax00MUD E5&&)) GWdD)) E5&&)) GWdD))  r,cJ|\}}}}}}|\}} ||z|| zz||z|| zzfS)zCEquivalent to apply_matrix_pt(M, (p,q)) - apply_matrix_pt(M, (0,0))r=) rrrrrrrrr`qs r*apply_matrix_normrSs@Q1aA FQ q51q5=!a%!a%- ''r,rc:t|ttfSr.)rrrrs r*isnumberr]s a#u & &&r,_Tobjsc#lKt}|D] }||vr|||V!dS)zEliminates duplicated elements.N)setadd)rdoneobjs r*uniqrdsN 55D $;;    r,cg}g}|D]8}||r||#||9||fS)z9Split a list into two classes according to the predicate.)rm)rxrtrrs r*fsplitrnsV A A 499  HHSMMMM HHSMMMM a4Kr,v0v1rchtt||zt||z|zS)zReturns a discrete range.)rlr)rrrs r*drangerzs+ RAs26{{a/ 0 00r,ptscttt t f}|\}}}}|D]E\}}t||}t||}t||}t||}F||||fS)z7Compute a minimal rectangle that covers all the points.)INFrr)rlimitrrrrrrs r* get_boundrsycTC4(ERR1 QZZ QZZ QZZ QZZ r2r>r,seqfuncmaxobjcBd}|D]}||}|||kr||}}|S)z;Picks the object obj where func(obj) has the highest value.Nr=)rrrmaxscorerscores r*pickrsD H..S   x%//"'vX Mr,n.c#Kg}|D]=}||t||krt|Vg}>dS)z$Groups every n elements of the list.N)rmrRtuple)rrrrs r*choplistrsW A   q66Q;;((NNNA r,defaultcbt|}|s|St|ddS)z7Unpacks variable-length unsigned integers (big endian).bigF) byteordersigned)rRr from_bytes)rLrrSs r*nunpackrs3 VVF @~~a5~???r,c#4K|]}t|VdSr.)chr).0rs r* rsDDD FFDDDDDDr,(rrzrQr}r~rPrOrj r iiiiiiii !"#$%&'()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\]^_`abcdefghijklmnopqrstuvwxyz{|}~ri" i i! i& i i iiD i9 i: i"i0 i i i i i i i"!iiiAiRi`ixi}i1iBiSiai~ri rr|c|drt|ddddSdd|DS)z+Decodes a PDFDocEncoding string to Unicode.srQNzutf-16berVrc30K|]}t|VdSr.)PDFDocEncoding)rrs r*rzdecode_text..s'44Q~a(444444r,) startswithr!join)rLs r* decode_textrsQ||K  51QRR5*h///ww44!444444r,cNt|trdSt|S)z"Encodes a string for SGML/XML/HTMLr)rrErrs r*rJrJs%!Ur !99r,bboxc2|\}}}}|dd|dd|dd|dS)Nz.3f,r=)rrrrrs r*bbox2strrs@RR 1 1 1r 1 1 1 1 1 1 1 1 11r,c J|\}}}}}}d|dd|dd|dd|dd|dd|dd S)N[z.2frz, (z)]r=)rrrrrrrs r* matrix2strrs_Q1aA Bq B B Bq B B Bq B B Bq B B B B B B B B B BBr,obj1robj2c*t|j|jt|j|j}}t|j|jt|j|j}}||z ||z }}||jz |jz ||jz |jz } }|dkrX| dkrR|j|jzdz |j|jzdz } } |j|jzdz |j|jzdz } } | | z | | z fStd|td| fS)aA distance function between two TextBoxes. Consider the bounding rectangle for obj1 and obj2. Return vector between 2 boxes boundaries if they don't overlap, otherwise returns vector betweeen boxes centers +------+..........+ (x1, y1) | obj1 | : +------+www+------+ : | obj2 | (x0, y0) +..........+------+ rrQ)rrrrrrwidthheight)rrrrrrowohiwihxc1yc1xc2yc2s r*vecBetweenBoxesrs DGTW%%s47DG'<'<RDGTW%%s47DG'<'<RRbRTZ$*,b4;.>.LR Avv"q&&w(A-$'0AQ/Fcw(A-$'0AQ/FcSy#)##1bzz3q"::%%r, LTComponentT)boundceZdZdZddededdfdZdefdZde e fd Z defd Z d e defd Zdede efd Zdee ddfdZd e ddfdZd e ddfdZdede e fdZdS)PlanezA set-like data structure for objects placed on a plane. Can efficiently find objects in a certain rectangular area. It maintains two parallel lists of objects, each of which is sorted by its x or y coordinate. rrgridsizerNcg|_t|_i|_||_|\|_|_|_|_dSr.) _seqr_objs_gridrrrrr)r)rrs r*r+zPlane.__init__s<(* (+ 68   /3,$'47DGGGr,c&dt|zS)Nz)rr/s r*__repr__zPlane.__repr__s 4::--r,c*fdjDS)Nc3.K|]}|jv |VdSr.r)rrr)s r*rz!Plane.__iter__..s/>>C4:,=,=,=,=,=,=>>r,)rr/s`r*__iter__zPlane.__iter__s>>>>ty>>>>r,c*t|jSr.)rRrr/s r*__len__z Plane.__len__s4:r,rc||jvSr.r)r)rs r* __contains__zPlane.__contains__sdj  r,c#K|\}}}}||jks!|j|ks||jks |j|krdSt |j|}t |j|}t |j|}t |j|}t |||jD]!}t |||jD]}||fV "dSr.)rrrrrrrr)r)rrrrrgrid_ygrid_xs r* _getrangezPlane._getrangesRR ==DGrMMR47]]dgmm F "   "   "   "  RT]33 ' 'F R77 ' 'v&&&&& ' ' 'r,rc:|D]}||dSr.)r)r)rrs r*rnz Plane.extends,  C HHSMMMM  r,c>||j|j|j|jfD]:}||jvr g}||j|<n |j|}||;|j||j |dS)zPlace an object.N) rrrrrrrmrrr)r)rkrs r*rz Plane.adds @AA  A ""(* ! 1 JqM HHSMMMM  sr,c ||j|j|j|jfD]:} |j||$#ttf$rY7wxYw|j |dS)zDisplace an object.N) rrrrrrremoveKeyErrorrr)r)rrs r*rz Plane.remove s @AA  A  1 $$S))))j)     #s AA&%A&c#.K|\}}}}t}||D]g}||jvr |j|D]M}||vr|||j|ks!||jks|j|ks ||jkrI|VNhdS)z)Finds objects that are in a certain area.N)rrrrrrrr) r)rrrrrrrrs r*findz Plane.find)sRRuu%%  A ""z!}  $;; 6R<<2<<36R<<2<<    r,)r)r7r8r9r:Rectrr+r!rr rrrr<boolrPointrr rnrrrr=r,r*rrs|44T4S4$4444.#....?(<0????!!4!!!! 'd 'x ' ' ' '8L1d |     ,4  (<"8      r,r)rurrr)rlrvaluecxd|cxkrdksnJg}d}|dkr t|d\}}|dkrF|dt||dt|dzn|dkrC|dt||dt|nV|dk}|r&|dt||dz}||rdndt||z|dz }|dk d|S) z,Format a number as lowercase Roman numerals.rirrrzr~rPr)divmodinsert ROMAN_ONES ROMAN_FIVESr)rresultindex remainder over_fives r*format_int_romanr=sN u    t      F E 1**!%,,y >> MM!Z. / / / MM!Z 2 3 3 3 3 !^^ MM!Z. / / / MM![/ 0 0 0 0!QI  aU!3444Q MMy/!!aE1BY1N O O O   1** 776??r,c$|dksJg}|dkrXt|dz ttj\}}|tj||dkX|d|S)z5Format a number as lowercase letters a-z, aa-zz, etc.rrzr)r rRstringascii_lowercasermreverser)rr rs r*format_int_alpharVs 19999F 1**!%!)S1G-H-HIIy f,Y7888 1** NN 776??r,paddedct|dkr|S|ddkr|St|kr|Stfd| dDr |d S|S)a1Remove block padding as described in PDF 1.7 section 7.6.2: > For an original message length of M, the pad shall consist of 16 - (M mod 16) bytes whose value shall also be 16 - (M mod 16). > Note that the pad is present when M is evenly divisible by 16; it contains 16 bytes of 0x10. rrc3$K|] }|kV dSr.r=)rrpaddings r*rzunpad_aes..us' 3 3A1< 3 3 3 3 3 3r,N)rRall)rrs @r* unpad_aesrcs 6{{a RjG|| V  3 3 3 3 !2 3 3 333!ixi  Mr,)rUrVr.)r)Tr:r%rrhtmlrtypingrrrrrr r r r r rrrrrrpdfminer.pdfexceptionsrrpdfminer.layoutrrFrr r!r&r;r'rrErAr<rKrrTrZrdrwrrrrMatrix PathSegmentr__annotations__rrrrrrrrrrrrlrrrrrrrrrJrrrrrr r rrrr=r,r*r%s  &?>>>>>>>,++++++ 7#S")3 4  fh&&&&&&&&4ce v #    3cc 5 5$ 5 5 5 5 5 5 5#css$ 14W    1u1%1C1E1111 8E? t        "  B4;   RL b\    (2,8E"c'N+C@@u@s@3@@@@DDADDDDDN555S555533242C2222 C&CSCCCC &-&}&&&&&4w~];;; MMMMMGL !MMM`" ! ! oo CC2 C C    er,