h7 UdZdZdZddlZddlZddlZddlmZddlm Z ddl m Z ddl m Z dd lmZmZmZmZmZmZmZmZmZ dd lmZn#e$r dd lmZYnwxYwd d lmZmZmZeeeeefeeeefeeeeffZ ee!d <eeeeeeefZ"ee!d<eZ#ee$e#fZ%dZ&dZ'dZ(dZ)de*de*de*fdZ+dTde#dee,de*fdZ-de#de*fdZ.de#de/fdZ0de#ddfdZ1 dUde#dee*de/de*fd Z2de#d!e,de*fd"Z3de#de*fd#Z4d$e d%e de fd&Z5de#ddfd'Z6iZ7eee$e*fe*fe!d(<d)ee$e*fde*fd*Z8ed%e$de$fd+Z9ed%e*de$fd,Z9d%ee$e*fde$fd-Z9ed%e$de,fd.Z:ed%e*de*fd/Z:ed%e,de,fd0Z:d%ee,e$e*fdee,e*ffd1Z:d%e*de*fd2Z;d3e,de$fd4ZdVd;e$dZ@ dWd@e$dAe$dBe$ddfdCZA dWd@e$dAe$dBe$ddfdDZBdWdEe$dBe$ddfdFZCdWdEe$dBe$ddfdGZDd;e$dHe$ddfdIZEdJe$defdKZF dUdLe$dMee$efdJee$e$fdNe/fdOZGdPe,de$fdQZHe GdRdSZIdS)Xz"Utility functions for PDF library.zMathieu Fenniakzbiziqe@mathieu.fenniak.netN getencoder) dataclass)DEFAULT_BUFFER_SIZE)SEEK_CUR) IOAnyCallableDictOptionalPatternTupleUnionoverload) TypeAlias)STREAM_TRUNCATED_PREMATURELYDeprecationErrorPdfStreamErrorTransformationMatrixTypeCompressedTransformationMatrixz2{} is deprecated and will be removed in PyPDF2 {}.z.{} is deprecated and was removed in PyPDF2 {}.zE{} is deprecated and will be removed in PyPDF2 3.0.0. Use {} instead.z>{} is deprecated and was removed in PyPDF2 {}. Use {} instead.header1header2returnc6d}g}||vr(|||||vr(|||t|dkrtd|d|d|t |S)N)s%PDF-1.3s%PDF-1.4s%PDF-1.5s%PDF-1.6s%PDF-1.7s%PDF-2.0rzneither z nor z are proper headers)appendindexlen ValueErrormax)rrversionspdf_header_indicess R/var/www/flask_pdf_extract_app/myenv/lib/python3.11/site-packages/PyPDF2/_utils.py_get_max_pdf_version_headerr$OsH(!!(.."9"9:::(!!(.."9"9::: !##RGRRGRRRSSS C*++ ,,streammaxcharscd} |d}|s|sn||z }t||krnF|S)z~ Read non-whitespace characters and return them. Stops upon encountering whitespace or when maxchars is reached. r%Tr)readisspacer)r&r'txttoks r#read_until_whitespacer-bs` Ckk!nn ;;==    s  s88x     Jr%c~|d}|tvr|d}|tv|S)zEFind and read the next non-whitespace character (ignores whitespace).r)r) WHITESPACESr&r,s r#read_non_whitespacer1ss= ++a..C   kk!nn    Jr%ctd}d}|tvr#|d}|dz }|tv#|dkS)zr Similar to read_non_whitespace, but return a Boolean if more than one whitespace character was read. rr)r/r))r&r,cnts r#skip_over_whitespacer4{sM a.C C   kk!nn q    7Nr%c|d}|dd|dkr|dvr|d}|dvdSdSdS)Nr%)  )r)seekr0s r#skip_over_commentr;sh ++a..C KKA d{{''++a..C''''{''r%Fregex ignore_eofcTd} |d}|s|r|Stt||}|X||d|z }||t |z dn||z }|S)z Read until the regular expression pattern matched (ignore the match). :raises PdfStreamError: on premature end-of-file :param bool ignore_eof: If true, ignore end-of-line and return immediately :param regex: re.Pattern r%TNr)r)rrsearchstartr:r)r&r<r=namer,ms r#read_until_regexrDs D kk"oo ?   !=>> > LL   = C !'')) $ $D KK CHH,a 0 0 0     Kr%to_readc||krtd|| t||}|| t|S)z Given a stream at position X, read a block of size to_read ending at position X. This changes the stream's position to the beginning of where the block was read. z!Could not read malformed PDF file)tellrr:rr))r&rEr)s r#read_block_backwardsrHsh{{}}w@AAA KK(### ;;w  D KK(### Kr%cg}d}|dkrtt tt|}|dkrnt ||}t |dz }|s-|dkr||dvr|dz}|dkr ||dv|dkrd}|rF|||dzd|dkr||dvr|dz}|dkr ||dvn|||dkr||dztnd |dddS) a Given a byte stream with current position X, return the previous line. All characters between the first CR/LF byte found before X (or, the start of the file, if no such byte is found) and position X After this call, the stream will be positioned one byte after the first non-CRLF character found beyond the first CR/LF byte before X, or, if no such byte is found, at the beginning of the stream. FrTrs Nr%r6) rGrrminrrHrrr:rjoin)r& line_content found_crlfrEblockidxs r#read_previous_linerPsLJ {{}}9::: )6;;==99 a<< %VW55%jj1n "((uSz88q((uSz88axx!  '   cAgii 0 1 1 1((uSzW44q((uSzW44    & & & !88 KKa * * * A D 88L2& ' ''r%abc:tfd|DS)Nc3^K|]&tfdtDV'dS)c3fK|]+}tdt|DV,dS)c3ZK|]&\}}t|t|zV'dSN)float).0ijs r# z6matrix_multiply....s7@@$!Q%((U1XX%@@@@@@r%N)sumzip)rYcolrows r#r\z,matrix_multiply...sB T TSc@@#c3--@@@@@ T T T T T Tr%N)tupler^)rYr`rRs @r#r\z"matrix_multiply..sZ   T T T TCQRG T T TTTr%)ra)rQrRs `r#matrix_multiplyrbs=   r%cd}|| dtdd5}||||d|||dddn #1swxYwY|| ddS)z5Create text file showing current location in context.irzPyPDF2_pdfLocation.txtwbsHEREN)r:openwriter))r&radius output_fhs r# mark_locationrisF KK & - -- F++,,,    F++,,,--------------- KKsA&BB #B B_CACHEsc4t}||vr||St|tr|S |d}t |dkr|||<|S#t $r2|d}t |dkr|||<|cYSwxYw)Nlatin-1zutf-8)rj isinstancebytesencoder Exception)rkbcrs r#b_rus BBww!u !U  HHY   q66A::BqE  HHW   q66A::BqE s.A9BBcdSrWrRs r#str_ryCr%cdSrWrwrxs r#ryryrzr%cZt|tr|dS|S)Nrm)rorpdecoderxs r#ryry"s+!Uxx """r%cdSrWrwrxs r#ord_r)rzr%cdSrWrwrxs r#rr.rzr%cdSrWrwrxs r#rr3rzr%cNt|trt|S|SrW)rostrordrxs r#rr8s$!S1vv Hr%cFtd}||}|dS)N hex_codecrr)rRcodercodeds r# hexencoder>s% { # #E E!HHE 8Or%numcHt|ddS)NL)hexreplace)rs r#hex_strrEs s88  C $ $$r%) r8r9 leftupup_leftc||z|z }t||z }t||z }t||z }||kr||kr|S||kr|S|SrW)abs)rrrp dist_leftdist_up dist_up_lefts r#paeth_predictorrLsn r GAAH I!b&kkGq7{##LG \ 9 9 L  r%msg stacklevelc>tj|t|dS)N)r)warningswarnDeprecationWarning)rrs r# deprecaterZs  M#)jAAAAAAr%c t|rW)r)rs r# deprecationr^s 3  r%3.0.0old_namenew_name removed_incZtt|||ddS)zS Raise an exception that a feature will be removed, but has a replacement. N)rDEPR_MSGformatrrrs r#deprecate_with_replacementrbs* hooh*==qAAAAAr%cXtt|||dS)zW Raise an exception that a feature was already removed, but has a replacement. N)rDEPR_MSG_HAPPENEDrrs r#deprecation_with_replacementrks+ !((:xHHIIIIIr%rBcXtt||ddS)zP Raise an exception that a feature will be removed without replacement. rN)rDEPR_MSG_NO_REPLACEMENTrrBrs r#deprecate_no_replacementrts+%,,T:>>BBBBBr%cVtt||dS)zT Raise an exception that a feature was already removed without replacement. N)r DEPR_MSG_NO_REPLACEMENT_HAPPENEDrrs r#deprecation_no_replacementr{s)077jIIJJJJJr%srccTtj||dS)a Use this instead of logger.warning directly. That allows people to overwrite it more easily. ## Exception, warnings.warn, logger_warning - Exceptions should be used if the user should write code that deals with an error case, e.g. the PDF being completely broken. - warnings.warn should be used if the user needs to fix their code, e.g. DeprecationWarnings - logger_warning should be used if the user needs to know that an issue was handled by PyPDF2, e.g. a non-compliant PDF being read in a way that PyPDF2 could apply a robustness fix to still read it. This applies mainly to strict=False mode. N)logging getLoggerwarning)rrs r#logger_warningrs)  c""3'''''r%aliasesc "dtffd }|S)z Decorator for deprecated term "bookmark" To be used for methods and function arguments outline_item = a bookmark outline = a collection of outline items funccJtjfd}|S)NcDtj|d|i|S)NT)fail) rename_kwargs__name__)argskwargsrrs r#wrapperz9deprecation_bookmark..decoration..wrappers1 $-t D D D D4((( (r%) functoolswraps)rrrs` r# decorationz(deprecation_bookmark..decorations>    ) ) ) ) )   )r%)r )rrs` r#deprecation_bookmarkrs0 r% func_namerrc &|D]{\}}||vrr|rt|d|d||vrt|d|d|d|d|d ||||<t j|d|dt |d S) z1 Helper function to deprecate arguments. z# is deprecated as an argument. Use z insteadz received both z and z as an argument. z is deprecated. Use z instead.)messagecategoryN)itemsr TypeErrorpoprrr)rrrrold_termnew_terms r#rrs&mmoo( v   &VVHVVV6!! IIIIIIII5=III &zz(33F8  MVVHVVV+     r%rpcf|dkr|dS|dkr |dz ddS|dkr |dz ddS|dz ddS) Niz Bytei@Bz.1fz kBiʚ;z MBz GBrw)rps r#_human_readable_bytesrsk u}} %-((((( %-(((((%-(((((r%c<eZdZUeed<eed<defdZdefdZdS)FilerBdatarc\d|jdtt|jdS)N File(name=, data: ))rBrrrselfs r#__str__z File.__str__s.WDIWW/DS^^/T/TWWWWr%cd|jdtt|jdt |jdS)Nrrz, hash: r)rBrrrhashrs r#__repr__z File.__repr__sCpDIpp/DS^^/T/Tpp^bcgcl^m^mppppr%N)r __module__ __qualname__r__annotations__rprrrwr%r#rrsj III KKKXXXXXq#qqqqqqr%rrW)F)r)r)J__doc__ __author____author_email__rrrcodecsr dataclassesriorosrtypingrr r r r r rrrr ImportErrortyping_extensionserrorsrrrrXrrr StreamTyper StrByteTyperrrrrpr$intr-r1boolr4r;rDrHrPrbrirjruryrrrr/rrrrrrrrrrrrrwr%r#rs8)((  /!!!!!!""""""                      ,       ,,,++++++++, ', % eUE&9 :E%PUBU