U 3g7@sUdZdZdZddlZddlZddlZddlmZddlm Z ddl m Z ddl m Z dd lmZmZmZmZmZmZmZmZmZzdd lmZWn ek rdd lmZYnXd d lmZmZmZeeeeefeeeefeeeeffZ ee!d <eeeeeeefZ"ee!d<eZ#ee$e#fZ%dZ&dZ'dZ(dZ)e*e*e*dddZ+dee#ee,e*dddZ-e#e*dddZ.e#e/dddZ0e#ddddZ1dfe#ee*e/e*d!d"d#Z2e#e,e*d$d%d&Z3e#e*dd'd(Z4e e e d)d*d+Z5e#ddd,d-Z6iZ7eee$e*fe*fe!d.<ee$e*fe*d/d0d1Z8ee$e$d2d3d4Z9ee*e$d2d5d4Z9ee$e*fe$d2d6d4Z9ee$e,d2d7d8Z:ee*e*d2d9d8Z:ee,e,d2d:d8Z:ee,e$e*fee,e*fd2d;d8Z:e*e*d2dd?d@Zdge$e,ddFdGdHZ?e$ddIdJdKZ@dhe$e$e$ddMdNdOZAdie$e$e$ddMdPdQZBdje$e$ddRdSdTZCdke$e$ddRdUdVZDe$e$ddWdXdYZEe$edZd[d\ZFdle$ee$efee$e$fe/d]d^d_ZGe,e$d`dadbZHe GdcddddZIdS)mz"Utility functions for PDF library.zMathieu Fenniakzbiziqe@mathieu.fenniak.netN getencoder) dataclass)DEFAULT_BUFFER_SIZE)SEEK_CUR) IOAnyCallableDictOptionalPatternTupleUnionoverload) TypeAlias)STREAM_TRUNCATED_PREMATURELYDeprecationErrorPdfStreamErrorTransformationMatrixTypeCompressedTransformationMatrixz2{} is deprecated and will be removed in PyPDF2 {}.z.{} is deprecated and was removed in PyPDF2 {}.zE{} is deprecated and will be removed in PyPDF2 3.0.0. Use {} instead.z>{} is deprecated and was removed in PyPDF2 {}. Use {} instead.)header1header2returncCsfd}g}||kr |||||kr8|||t|dkrZtd|d|d|t|S)N)s%PDF-1.3s%PDF-1.4s%PDF-1.5s%PDF-1.6s%PDF-1.7s%PDF-2.0rzneither z nor z are proper headers)appendindexlen ValueErrormax)rrversionsZpdf_header_indicesr 1/tmp/pip-unpacked-wheel-ifm06wwb/PyPDF2/_utils.py_get_max_pdf_version_headerOs r")streammaxcharsrcCs8d}|d}|s4|sq4||7}t||krq4q|S)z~ Read non-whitespace characters and return them. Stops upon encountering whitespace or when maxchars is reached. r)readisspacer)r#r$txttokr r r!read_until_whitespacebs   r*)r#rcCs"|d}|tkr|d}q |S)zEFind and read the next non-whitespace character (ignores whitespace).r)r& WHITESPACESr#r)r r r!read_non_whitespacess  r-cCs0td}d}|tkr(|d}|d7}q |dkS)zr Similar to read_non_whitespace, but return a Boolean if more than one whitespace character was read. rr)r+r&)r#r)Zcntr r r!skip_over_whitespace{s   r.cCs6|d}|dd|dkr2|dkr2|d}qdS)Nr%)  )r&seekr,r r r!skip_over_comments   r4F)r#regex ignore_eofrcCspd}|d}|s"|r|Stt||}|dk rb||d|7}||t|dql||7}q|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%Nr)r&rrsearchstartr3r)r#r5r6namer)mr r r!read_until_regexs    r<)r#to_readrcCs>||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)tellrr3rr&)r#r=r&r r r!read_block_backwardss   r?cCsg}d}|dkrtttt|}|dkr4qt||}t|d}|sx|dkrl||dkrl|d8}qN|dkrxd}|r|||dd|dkr||dkr|d8}qn |||dkr||dt qqd |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. Frrs TNr%r/) r>rrminrr?rrr3rjoin)r#Z line_contentZ found_crlfr=blockidxr r r!read_previous_lines.       rD)abrcstfdd|DS)Nc3s(|] tfddtDVqdS)c3s&|]}tddt|DVqdS)css"|]\}}t|t|VqdSN)float).0ijr r r! sz6matrix_multiply....N)sumzip)rIcolrowr r!rLsz,matrix_multiply...N)tuplerN)rIrFrPr!rLsz"matrix_multiply..)rR)rErFr rSr!matrix_multiplys rTc Csdd}|| dtdd0}||||d|||W5QRX|| ddS)z5Create text file showing current location in context.irzPyPDF2_pdfLocation.txtwbsHEREN)r3openwriter&)r#ZradiusZ output_fhr r r! mark_locations  rXB_CACHE)srcCst}||kr||St|tr"|Sz$|d}t|dkrB|||<|WStk r||d}t|dkrt|||<|YSXdS)Nlatin-1zutf-8)rY isinstancebytesencoder Exception)rZZbcrr r r!b_s     rb)rFrcCsdSrGr rSr r r!str_srccCsdSrGr rSr r r!rcscCst|tr|dS|SdS)Nr[)r]r^decoderSr r r!rc"s  cCsdSrGr rSr r r!ord_)srecCsdSrGr rSr r r!re.scCsdSrGr rSr r r!re3scCst|trt|S|SrG)r]strordrSr r r!re8s cCstd}||}|dS)N hex_codecrr)rFZcoderZcodedr r r! hexencode>sri)numrcCst|ddS)NL)hexreplace)rjr r r!hex_strEsro) r1r2 )leftupup_leftrcCsX|||}t||}t||}t||}||krD||krD|S||krP|S|SdSrG)abs)rsrtrupZ dist_leftZdist_upZ dist_up_leftr r r!paeth_predictorLs    rx)msg stacklevelrcCstj|t|ddS)N)r{)warningswarnDeprecationWarning)rzr{r r r! deprecateZsr)rzrcCs t|dSrG)r)rzr r r! deprecation^sr3.0.0)old_namenew_name removed_inrcCstt|||ddS)zS Raise an exception that a feature will be removed, but has a replacement. N)rDEPR_MSGformatrrrr r r!deprecate_with_replacementbsrcCstt|||dS)zW Raise an exception that a feature was already removed, but has a replacement. N)rDEPR_MSG_HAPPENEDrrr r r!deprecation_with_replacementksr)r:rrcCstt||ddS)zP Raise an exception that a feature will be removed without replacement. rN)rDEPR_MSG_NO_REPLACEMENTrr:rr r r!deprecate_no_replacementtsrcCstt||dS)zT Raise an exception that a feature was already removed without replacement. N)r DEPR_MSG_NO_REPLACEMENT_HAPPENEDrrr r r!deprecation_no_replacement{sr)rzsrcrcCst||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)rzrr r r!logger_warningsr)aliasesrc stdfdd }|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 funccstfdd}|S)Ncstj|dd||S)NT)fail) rename_kwargs__name__)argskwargs)rrr r!wrappersz9deprecation_bookmark..decoration..wrapper) functoolswraps)rrrrr! decorationsz(deprecation_bookmark..decoration)r )rrr rr!deprecation_bookmarksr) func_namerrrc Cs|D]~\}}||kr|r0t|d|d||kr^t|d|d|d|d|d ||||<tj|d|dtdqd 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 TypeErrorpopr|r}r~)rrrrZold_termZnew_termr r r!rs r)r^rcCsV|dkr|dS|dkr*|dddS|dkrB|dddS|dddSdS) Niz Bytei@Bz.1fz kBiʚ;z MBz GBr )r^r r r!_human_readable_bytess rc@s:eZdZUeed<eed<edddZedddZdS) Filer:data)rcCsd|jdtt|jdS)N File(name=, data: ))r:rrrselfr r r!__str__sz File.__str__cCs*d|jdtt|jdt|jdS)Nrrz, hash: r)r:rrrhashrr r r!__repr__sz File.__repr__N)r __module__ __qualname__rf__annotations__r^rrr r r r!rs r)N)F)ry)r)r)r)r)F)J__doc__ __author____author_email__rrr|codecsrZ dataclassesriorosrtypingrrr r r r r rrr ImportErrorZtyping_extensionserrorsrrrrHrrrZ StreamTyperfZ StrByteTyperrrrr^r"intr*r-boolr.r4r<r?rDrTrXrYrbrcreriror+rxrrrrrrrrrrrr r r r!s    , &      4  "