o f@sddlZddlZddlmZddlmZmZmZmZm Z m Z m Z m Z ddl mZddlmZddlmZddlmZdd lmZdd lmZdd lmZdd lmZdd lmZddlmZddlmZddlmZddlmZddlmZddl m!Z!ddl"m#Z#ddl"m$Z$ddl"m%Z%ddl"m&Z&ddl"m'Z'ddl"m(Z(ddl"m)Z)ddl*m+Z+ddl*m,Z,ddl*m-Z-ddl*m.Z.ddl*m/Z/m0Z0dd l*m1Z1dd!l*m2Z2dd"l*m3Z3dd#l*m4Z4dd$l5m6Z6dd%l5m7Z7m8Z8m9Z9m:Z:dd&l5m;Z;dd'l5mZ?Gd(d)d)e#Z@Gd*d+d+e#ZAe,d,ZBe,d-ZCe,d.ZDe,d/ZEe,d0ZFGd1d2d2ZGe eHe eHeHeHfe eHeHeHeHffZIGd3d4d4ZJGd5d6d6ZKGd7d8d8e1e e.e%fZLe2e%ZM Gd9d:d:ZNdS);N)BytesIO)DictListMappingOptionalSequenceTupleUnioncast)settings)CMap)CMapBase)CMapDB) PDFColorSpace)PREDEFINED_COLORSPACE) PDFDevice) PDFTextSeq) PDFCIDFont)PDFFont) PDFFontError)PDFTrueTypeFont) PDFType1Font) PDFType3Font)PDFPage) PDFException) PDFObjRef) PDFStream) dict_value) list_value)resolve1) stream_value)KWD)LIT)PSEOF) PSKeyword) PSLiteral PSTypeError) PSStackParser) PSStackType) keyword_name literal_name)MATRIX_IDENTITY)MatrixPoint PathSegmentRect)choplist) mult_matrixc@ eZdZdS)PDFResourceErrorN__name__ __module__ __qualname__r:r:K/home/ubuntu/webapp/venv/lib/python3.10/site-packages/pdfminer/pdfinterp.pyr5-r5c@r4)PDFInterpreterErrorNr6r:r:r:r;r=1r<r=PDFTextFontFormImagec@sJeZdZUeed<eed<d ddZdefddZdd d Z d d d Z dS) PDFTextStatematrix linematrixreturnNcCs<d|_d|_d|_d|_d|_d|_d|_d|_|dS)Nrd) fontfontsize charspace wordspacescalingleadingrenderriseresetselfr:r:r;__init__@s zPDFTextState.__init__c Cs0d|j|j|j|j|j|j|j|j|j|j f S)Nz) rHrIrJrKrLrMrNrOrDrErQr:r:r;__repr__MszPDFTextState.__repr__cCsZt}|j|_|j|_|j|_|j|_|j|_|j|_|j|_|j|_|j |_ |j |_ |SN) rCrHrIrJrKrLrMrNrOrDrErRobjr:r:r;copy`szPDFTextState.copycCst|_d|_dS)Nrr)r-rDrErQr:r:r;rPn zPDFTextState.resetrFN)rFrC) r7r8r9r.__annotations__r/rSstrrTrXrPr:r:r:r;rC<s   rCc@s.eZdZd ddZd ddZdefddZdS) PDFGraphicStaterFNcCs:d|_d|_d|_d|_d|_d|_d|_d|_d|_dSNr linewidthlinecaplinejoin miterlimitdashintentflatnessscolorncolorrQr:r:r;rS{s zPDFGraphicState.__init__cCsRt}|j|_|j|_|j|_|j|_|j|_|j|_|j|_|j|_|j |_ |SrU) r^rarbrcrdrerfrgrhrirVr:r:r;rXszPDFGraphicState.copyc Cs,d|j|j|j|j|j|j|j|j|jf S)Nzr`rQr:r:r;rTszPDFGraphicState.__repr__r[)rFr^)r7r8r9rSrXr]rTr:r:r:r;r^zs   r^c@speZdZdZddeddfddZdeeddfd d Zdd e d ede fddZ dede e efde fddZdS)PDFResourceManagerzRepository of shared resources. ResourceManager facilitates reuse of shared resources such as fonts and images so that large objects are not allocated multiple times. TcachingrFNcCs||_i|_dSrU)rk _cached_fonts)rRrkr:r:r;rSrZzPDFResourceManager.__init__procscCs$|D] }|tur q|turq qdSrU) LITERAL_PDF LITERAL_TEXT)rRrmprocr:r:r; get_procsetszPDFResourceManager.get_procsetFcmapnamestrictcCs.zt|WStjy|rtYSwrU)rget_cmap CMapNotFoundr )rRrrrsr:r:r;rts  zPDFResourceManager.get_cmapobjidspeccCsL|r||jvr|j|}|Std||tjr"|dtur"tdd|vr-t|d}n tjr4tdd}|dvr@t||}nZ|dkrJt ||}nP|d krTt ||}nF|d vr^t ||}n<|d krt |d }|slJt |d }dD]}||vrt||||<qv|d|}ntjrtd|t||}|r|jr||j|<|S)Nz#get_font: create: objid=%r, spec=%rTypezType is not /FontSubtypezFont Subtype is not specified.Type1)rzMMType1TrueTypeType3) CIDFontType0 CIDFontType2Type0DescendantFontsr)Encoding ToUnicodezInvalid Font spec: %r)rllogdebugr STRICT LITERAL_FONTrr,rrrrrrrXr get_fontrk)rRrvrwrHsubtypedfontssubspeckr:r:r;rsF (          zPDFResourceManager.get_font)T)F)r7r8r9__doc__boolrSrobjectrqr]rrtrrrr:r:r:r;rjs  "rjc @seZdZdeeddfddZdddZdeddfd d Zdd d Z ddede de ee ffddZ dddZ edZedZed ZdededdfddZdS)PDFContentParserstreamsrFNcCs||_d|_t|ddSr_)ristreamr(rS)rRrr:r:r;rSszPDFContentParser.__init__cCsT|js(|jt|jkrt|j|j}|jd7_ntdt||_dSdS)Nr zUnexpected EOF, file truncated?)fprlenrr!r$rget_data)rRstrmr:r:r;fillfpszPDFContentParser.fillfpposcCs|t||dSrU)rr(seek)rRrr:r:r;rszPDFContentParser.seekcCsT|jt|jkr dS ||j|_|j|j|_|jr!nd|_q d|_dS)Nr r) charposrbufrrtellbufposreadBUFSIZrQr:r:r;fillbuf s  zPDFContentParser.fillbufEItargetcCsL||d}d}|t|kr||rQ|j|j}t|f}||7}|jd7_t||kr:|r:|d7}nP|t|krN|t||fkrN|d7}nEsz/PDFContentParser.do_keyword..sID ) KEYWORD_BI start_type KEYWORD_IDend_typerformatr'r2rrpush KEYWORD_EIr r) rRrr_objs error_msgdrrWr:r:r; do_keyword;s&     zPDFContentParser.do_keywordr[)r)r7r8r9rrrSrintrrrrrrr"rrrr%rr:r:r:r;rs   rc@seZdZdZdededdfddZddd Zd ee e fddfd d Z d e ddfddZ de ddfddZdedee fddZdee eeffddZdee eefddfddZdddZdddZde d e d!e d"e d#e d$e ddfd%d&Zd'e ddfd(d)Zd*e ddfd+d,Zd-e ddfd.d/Zd0e ddfd1d2Zd3e d4e ddfd5d6Zd7e ddfd8d9Zd:e ddfd;d<Z d=e ddfd>d?Z!d@e dAe ddfdBdCZ"d@e dAe ddfdDdEZ#dFe dGe dHe dIe dJe dKe ddfdLdMZ$dHe dIe dJe dKe ddf dNdOZ%dFe dGe dJe dKe ddf dPdQZ&ddRdSZ'd@e dAe dTe dUe ddf dVdWZ(ddXdYZ)ddZd[Z*dd\d]Z+dd^d_Z,dd`daZ-ddbdcZ.ddddeZ/ddfdgZ0ddhdiZ1ddjdkZ2ddldmZ3ddndoZ4d=e ddfdpdqZ5d=e ddfdrdsZ6dte ddfdudvZ7dte ddfdwdxZ8dye dze d{e ddfd|d}Z9dye dze d{e ddfd~dZ:de de dAe de ddf ddZ;de de dAe de ddf ddZdddZ?dddZ@d=e ddfddZAdddZBdddZCdddZDdddZEde ddfddZFde de ddfddZGde ddfddZHde de ddfddZIdddZJde ddfddZKde ddfddZLde ddfddZMde ddfddZNde de ddfddZOde ddfddZPde ddfddZQde de ddfddZRde de ddfddZSde d{e de de de de ddfddńZTdddDŽZUde ddfddʄZVde ddfdd̈́ZWde ddfddτZXde de de ddfddӄZYdddՄZZdddׄZ[de ddfddلZ\de ddfdd܄Z]de^ddfdd߄Z_e`fd ee e fdeae d e ddfddZbdeae ddfddZcdS)PDFPageInterpreterzhProcessor for the content of a PDF page Reference: PDF Reference, Appendix A, Operator Summary rsrcmgrdevicerFNcCs||_||_dSrU)rr)rRrrr:r:r;rS[szPDFPageInterpreter.__init__cCs||j|jSrU) __class__rrrQr:r:r;dup`szPDFPageInterpreter.dup resourcesc Cs2||_i|_i|_t|_|sdSdtdttfdd}t | D]s\}}t d|||dkrVt | D]\}}d}t |trF|j}t |}|j|||j|<q8q#|dkrut | D]\}}|t|} | durs| |j|<q`q#|d kr|jt|q#|d krt | D] \} } | |j| <qq#dS) z@Prepare the fonts and XObjects listed in the Resource attribute.NrwrFcSst|tr t|d}nt|}|dkr*t|tr*dt|kr*t|t|ddS|dkrDt|trDdt|krDt|tt|dSt|S)NrICCBasedrr NDeviceN) isinstancelistr,rrr!rrget)rwnamer:r:r;get_colorspacels  z9PDFPageInterpreter.init_resources..get_colorspacezResource: %r: %rr@ ColorSpaceProcSetXObject)rfontmapxobjmaprrXcsmaprrrritemsrrrrrvrrr rqr) rRrrrrfontidrwrvcsid colorspacexobjidxobjstrmr:r:r;init_resourcescs>      z!PDFPageInterpreter.init_resourcesctmcCsfg|_||_|j|jt|_t|_g|_g|_ d|_ d|_ |j r1t t|j |_ |_ dS)zClose, fill, and stroke path using nonzero winding number ruleN)r%r3rQr:r:r;do_b^r/zPDFPageInterpreter.do_bcCr-)z0Close, fill, and stroke path using even-odd ruleN)r%r4rQr:r:r;do_b_adr/zPDFPageInterpreter.do_b_acCs g|_dS)z$End path without filling or strokingN)rrQr:r:r;do_njszPDFPageInterpreter.do_ncCr )z3Set clipping path using nonzero winding number ruleNr:rQr:r:r;do_WozPDFPageInterpreter.do_WcCr )z%Set clipping path using even-odd ruleNr:rQr:r:r;do_W_asr9zPDFPageInterpreter.do_W_acC>z |jt||_WdStytjrtd|YdSw)zOSet color space for stroking operations Introduced in PDF 1.1 Undefined ColorSpace: %rN)rr,rKeyErrorr rr=r r:r:r;do_CSws  zPDFPageInterpreter.do_CScCr;)z*Set color space for nonstroking operationsr<N)rr,rr=r rr=r r:r:r;do_css  zPDFPageInterpreter.do_csgraycCtt||j_|jd|_dS)z&Set gray level for stroking operations DeviceGrayNr rrrhrrrRr@r:r:r;do_G zPDFPageInterpreter.do_GcCrA)z)Set gray level for nonstroking operationsrBNr rrrirrrDr:r:r;do_grFzPDFPageInterpreter.do_grgbcC0tt|tt|tt|f|j_|jd|_dS)z%Set RGB color for stroking operations DeviceRGBNrCrRrIrJrKr:r:r;do_RG  zPDFPageInterpreter.do_RGcCrL)z(Set RGB color for nonstroking operationsrMNrGrNr:r:r;do_rgrPzPDFPageInterpreter.do_rgrrrcC8tt|tt|tt|tt|f|j_|jd|_dS)z&Set CMYK color for stroking operations DeviceCMYKNrCrRrrrrr:r:r;do_K zPDFPageInterpreter.do_KcCrR)z)Set CMYK color for nonstroking operationsrSNrGrTr:r:r;do_krVzPDFPageInterpreter.do_kcC:|jr|jj}n tjrtdd}tt|||j_ dS)z"Set color for stroking operations.No colorspace specified!r N) r ncomponentsr rr=r ColorrrrhrRrr:r:r;do_SCN zPDFPageInterpreter.do_SCNcCrX)$Set color for nonstroking operationsrYr N) rrZr rr=r r[rrrir\r:r:r;do_scnr^zPDFPageInterpreter.do_scncC |dS)z!Set color for stroking operationsN)r]rQr:r:r;do_SCrzPDFPageInterpreter.do_SCcCra)r_N)r`rQr:r:r;do_scrzPDFPageInterpreter.do_sccCr )z%Paint area defined by shading patternNr:r r:r:r;do_shr9zPDFPageInterpreter.do_shcC|jdS)zBegin text object Initializing the text matrix, Tm, and the text line matrix, Tlm, to the identity matrix. Text objects cannot be nested; a second BT cannot appear before an ET. N)rrPrQr:r:r;do_BTs zPDFPageInterpreter.do_BTcCr )zEnd a text objectNr:rQr:r:r;do_ETr9zPDFPageInterpreter.do_ETcCr )zBegin compatibility sectionNr:rQr:r:r;do_BXr9zPDFPageInterpreter.do_BXcCr )zEnd compatibility sectionNr:rQr:r:r;do_EXr9zPDFPageInterpreter.do_EXtagcC|jtt|dS)zDefine marked-content pointNrdo_tagr r&rRrjr:r:r;do_MPzPDFPageInterpreter.do_MPpropscC|jtt||dS)z.Define marked-content point with property listNrlrRrjrqr:r:r;do_DPzPDFPageInterpreter.do_DPcCrk)zBegin marked-content sequenceNr begin_tagr r&rnr:r:r;do_BMCrpzPDFPageInterpreter.do_BMCcCrr)z0Begin marked-content sequence with property listNrvrsr:r:r;do_BDCruzPDFPageInterpreter.do_BDCcCre)zEnd marked-content sequenceN)rend_tagrQr:r:r;do_EMCs zPDFPageInterpreter.do_EMCspacecCr)zSet character spacing. Character spacing is used by the Tj, TJ, and ' operators. :param space: a number expressed in unscaled text space units. N)r rrrJrRr|r:r:r;do_Tc zPDFPageInterpreter.do_TccCr)zSet the word spacing. Word spacing is used by the Tj, TJ, and ' operators. :param space: a number expressed in unscaled text space units N)r rrrKr}r:r:r;do_TwrzPDFPageInterpreter.do_TwscalecCr)zuSet the horizontal scaling. :param scale: is a number specifying the percentage of the normal width N)r rrrL)rRrr:r:r;do_Tz!zPDFPageInterpreter.do_TzrMcCstt| |j_dS)zSet the text leading. Text leading is used only by the T*, ', and " operators. :param leading: a number expressed in unscaled text space units N)r rrrM)rRrMr:r:r;do_TL)szPDFPageInterpreter.do_TLrrIcCs`z |jt||j_Wnty&tjrtd||j di|j_Ynwt t ||j_ dS)zSet the text font :param fontid: the name of a font resource in the Font subdictionary of the current resource dictionary :param fontsize: size is a number representing a scale factor. zUndefined Font id: %rN) rr,rrHr=r rr=rrr rrI)rRrrIr:r:r;do_Tf3s  zPDFPageInterpreter.do_TfrNcCr)zSet the text rendering modeN)r rrrN)rRrNr:r:r;do_TrCrzPDFPageInterpreter.do_TrrOcCr)z`Set the text rise :param rise: a number expressed in unscaled text space units N)r rrrO)rRrOr:r:r;do_TsHrzPDFPageInterpreter.do_Tstxtyc Cshtt|}tt|}|jj\}}}}}}||||||||||||||f|j_d|j_dS)zMove text positionrYN)r rrrDrE rRrrarKrrefr:r:r;do_TdPs  4zPDFPageInterpreter.do_Tdc Csptt|}tt|}|jj\}}}}}}||||||||||||||f|j_||j_d|j_dS)z"Move text position and set leadingrYN)r rrrDrMrErr:r:r;do_TDYs  4zPDFPageInterpreter.do_TDrrrrcCs&tt||||||f|j_d|j_dS)z$Set text matrix and text line matrixrYN)r r.rrDrErRrrKrrrrr:r:r;do_Tmcs zPDFPageInterpreter.do_TmcCsL|jj\}}}}}}|||||jj|||jj||f|j_d|j_dS)zMove to start of next text linerYN)rrDrMrErr:r:r;do_T_aqszPDFPageInterpreter.do_T_aseqcCsR|jjdurtjr tddS|jdusJ|j|jtt ||j|j dS)z0Show text, allowing individual glyph positioningNzNo font specified!) rrHr rr=rr render_stringr rrrX)rRrr:r:r;do_TJs zPDFPageInterpreter.do_TJscCs||gdS)z Show textN)rrRrr:r:r;do_Tjr zPDFPageInterpreter.do_TjcCs|||gdS)zPMove to next line and show text The ' (single quote) operator. N)rrrr:r:r;do__qs zPDFPageInterpreter.do__qawaccCs$||||||gdS)zqSet word and character spacing, move to next line, and show text The " (double quote) operator. N)rr~r)rRrrrr:r:r;do__ws   zPDFPageInterpreter.do__wcCr )zBegin inline image objectNr:rQr:r:r;do_BIr9zPDFPageInterpreter.do_BIcCr )zBegin inline image dataNr:rQr:r:r;do_IDr9zPDFPageInterpreter.do_IDcCsTt|tr(d|vr(d|vr(tt|}|j|dt|j|||j|dS)zEnd inline image objectWHrrr r N) rrr]idr begin_figurer- render_image end_figure)rRrWiobjidr:r:r;do_EIs   zPDFPageInterpreter.do_EI xobjid_argc CsFttt|}z t|j|}Wnty#tjr td|YdSwt d|| d}|t ur}d|vr}| }ttt|d}ttt| dt}| d}|r[t|} n|j} |j||||j| |gt||jd|j|dS|turd |vrd |vr|j|d t|j|||j|dS dS) zInvoke named XObjectzUndefined xobject id: %rNzProcessing xobj: %rryBBoxr. ResourcesrWidthHeightr)r r]r,r!rr=r rr=rrr LITERAL_FORMrr1rr.r-rrrXrrrender_contentsr3rr LITERAL_IMAGEr) rRrrxobjr interpreterbboxrDxobjresrr:r:r;do_Dos>         zPDFPageInterpreter.do_DopagecCstd||j\}}}}|jdkrdddd| |f}n'|jdkr*dddd||f}n|jdkr9dddd|| f}n dddd| | f}|j|||j|j|j|d|j |dS) NzProcessing page: %rZrr ir) rrmediaboxrotater begin_pagerrcontentsend_page)rRrx0y0rrrr:r:r; process_pages     zPDFPageInterpreter.process_pagercCs6td||||||||t|dS)zTRender the content streams. This method may be called recursively. z1render_contents: resources=%r, streams=%r, ctm=%rN)rrrrexecuter)rRrrrr:r:r;rs   z"PDFPageInterpreter.render_contentsc Cszt|}Wn tyYdSw z|\}}Wn ty$YdSwt|tr}t|}d|dddddd}t||rqt||}|j j d}|rg| |} t d || t| |krf|| nt d ||ntjr|d |} t| n||q) Nr zdo_%s*_a"_w'_qz exec: %s %rzexec: %szUnknown operator: %r)rr$ nextobjectrr%r*replacehasattrgetattr__code__ co_argcountrrrrr rr=r) rRrparserrrWrmethodfuncnargsargsrr:r:r;rsD           zPDFPageInterpreter.execute)rFrr[)dr7r8r9rrjrrSrrrrr.r PDFStackTrrrrrrCr^rrrrrrrrrrr r rrrrr r"r%r'r+r.r0r1r2r3r4r5r6r7r8r:r>r?rErHrOrQrUrWr]r`rbrcrdrfrgrhrirortrxryr{r~rrrrrrrrrrrrrrrrrrrrr-rrrr:r:r:r;rUs *                                 %  r)Ologgingriortypingrrrrrrr r r cmapdbr rrpdfcolorrr pdfdevicerrpdffontrrrrrrpdfpagerpdftypesrrrrrr r!psparserr"r#r$r%r&r'r(r)r*r,utilsr-r.r/r0r1r2r3 getLoggerr7rr5r=rnrorrrrCrr[r^rjrrrr:r:r:r;sx (                                  7 0J]