jvc fUdZdZdZddlZddlZddlZddlZddlZddlZddl m Z ddl m Z m Z ddl mZddlmZdd lmZdd lmZmZmZmZmZmZejdd d krdd lmZndd lmZejdkrddlmZnddlmZddlmZmZmZm Z e!e!e"e"e"fe!e"e"e"fe!e"e"e"ffZ#ee$d<e!e"e"e"e"e"e"fZ%ee$d<eeZ&ee'Z(ee)e&fZ*dee)dee fdZ+de de)fdZ,de)de)de)fdZ-dZ.d/e.Z0de0zdzZ1ddd!e&d"ee2d#e3de'fd$Z4d!e(de'fd%Z5d!e&de3fd&Z6d'e'de3fd(Z7d)Z8ded+e)d,e)de3fd-Z9d.Z:dfd+e)d/e)d0e)de3fd1Z;d!e&ddfd2Zd6d!e&d7ee'd8e2de'fd9Z?d!e(d:e2de'fd;Z@d!e&de'fd<ZAd=e#d>e#de#fd?ZBd!e&ddfd@ZCdgdBe)dCe2ddfdDZDdBe)defdEZEdFe)dGe)dHe)ddfdIZFdFe)dGe)dHe)defdJZGdKe)dHe)ddfdLZHdKe)dHe)defdMZIdNe)dOe)dPeddfdQZJdNe)dOe)dPeddfdRZK dhdSe)dTeLe)efdUeLe)e)fdVe3ddf dWZMdXe2de)fdYZNd!e&de2fdZZOd!e&de'fd[ZPGd\d]ZQe Gd^d_ZRejSGd`daZTe GdbdcZUdS)iz"Utility functions for PDF library.zMathieu Fenniakzbiziqe@mathieu.fenniak.netN) dataclass)datetimetimezone)DEFAULT_BUFFER_SIZE)SEEK_CUR)Pattern)IOAnyNoReturnOptionalUnioncast) ) TypeAlias)r )Self)STREAM_TRUNCATED_PREMATURELYDeprecationErrorLimitReachedErrorPdfStreamErrorTransformationMatrixTypeCompressedTransformationMatrixtextreturnc|}|sdS|drd|z}|dr|dz }|dddddd }t|d|d }|dkr|t |d z kr|d z }d D]`} t j||}|dr |tj }|cS#t$rY]wxYwtd|)NrzD:)Zz0000r +r'-00)zD:%YzD:%Y%mzD:%Y%m%dz D:%Y%m%d%Hz D:%Y%m%d%H%MD:%Y%m%d%H%M%SzD:%Y%m%d%H%M%S%zz+0000)tzinfozCan not convert date: ) isdigitendswithreplacemaxfindlenrstrptimerutc ValueError)rorgtextifds ~/Users/tedpeters/Documents/Projects/Claude Projects/TPC Houses/house-viewer/.venv/lib/python3.11/site-packages/pypdf/_utils.pyparse_iso8824_dater8RscG t Awd{ }}Z    <<S ! ! ) )#s 3 3 ; ;C D DD DIIcNNDIIcNN++A1uuc$ii!m##   !$**A}}W%% 3IIX\I22HHH     H  7g77 8 88sD(( D54D5dtc,|d}|jw|}|Jt|}t t |d\}}|dz}|dkrdnd}|||dd |dd z }|S) a Convert a datetime object to PDF date string format. Converts datetime to the PDF date format D:YYYYMMDDHHmmSSOHH'mm as specified in the PDF Reference. Args: dt: A datetime object to convert. Returns: A date string in PDF format. r(Ni<rr"r%02dr#)strftimer) utcoffsetint total_secondsdivmodabs)r9date_stroffsetr@hours remainderminutessigns r7format_iso8824_daterIrs{{+,,H y!!!F002233 !#m"4"4d;;yr/#q((ssct7U77777777 Oheader1header2c6d}g}||vr(|||||vr(|||t|dkrtd|d|d|t |S)N)z%PDF-1.3z%PDF-1.4z%PDF-1.5z%PDF-1.6z%PDF-1.7z%PDF-2.0rzNeither z nor z are proper headers)appendindexr/r2r-)rKrLversionspdf_header_indicess r7_get_max_pdf_version_headerrRsH(!!(.."9"9:::(!!(.."9"9::: !##RGRRGRRRSSS C*++ ,,rJ)     rJ[]Fstream max_bytesstrictc8t} |d}|s |tvs|snH|@t ||kr-|rt d|dt dt|n||z }}t|S)a Read non-whitespace characters and return them. Stops upon encountering whitespace or when max_bytes is reached. Args: stream: The data stream from which was read. max_bytes: The maximum number of bytes which are considered valid. One more byte will always be read from the stream. strict: Whether to raise an exception if the token exceeds max_bytes. If False, a warning is logged and only the first max_bytes are returned. Returns: The data which was read. TrNz Token exceeds maximum length of z bytes.z4Token exceeds maximum length of %(max_bytes)d bytes.)sourcer\) bytearrayreadisspace WHITESPACESr/rlogger_warning__name__bytes)r[r\r]txttoks r7read_until_whitespaceris" ++C kk!nn ;;== C;..c.   SXX%:%: _'(]9(](](]^^^ Fxcl      s   ::rJc~|d}|tvr|d}|tv|S)z Find and read the next non-whitespace character (ignores whitespace). Args: stream: The data stream from which was read. Returns: The data which was read. rrarcr[rhs r7read_non_whitespacerms? ++a..C   kk!nn    JrJc|d}d}|tvr#|dz }|d}|tv#|dkS)a  Similar to read_non_whitespace, but return a boolean if at least one whitespace character was read. Args: stream: The data stream from which was read. Returns: True if one or more whitespace was skipped, otherwise return False. rrrk)r[rhcnts r7skip_over_whitespacerpsS ++a..C C    qkk!nn    7NrJvaluec4td|DS)z Check if the given value consists of whitespace characters only. Args: value: The bytes to check. Returns: True if the value only has whitespace characters, otherwise return False. c3(K|] }|tvVdSN)WHITESPACES_AS_BYTES).0bs r7 z+check_if_whitespace_only..s(88Qq((888888rJ)all)rqs r7check_if_whitespace_onlyrzs! 88%888 8 88rJ))/):@)u٠u٭)u۰u۹)u u)u₠u⇿r$charcustom_special_characterscntfdtDrdSt|o|vS)z8Check if a character is part of neutral character rangesc3>K|]\}}|cxko|kncVdSrtrvstartendrs r7rxz"is_char_neutral..sC K KJE35D    C     K K K K K KrJT)anyNEUTRAL_CHARACTER_RANGESbool)rrs` r7is_char_neutralrsH K K K K2J K K KKKt )Od6O.O P PPrJ))u֐uࣿ)uיִu﷿)uﹰucustom_rtl_mincustom_rtl_maxctfdtDrdSt|o|o|cxko|kncS)z4Check if a character is part of RTL character rangesc3>K|]\}}|cxko|kncVdSrtrrs r7rxzis_char_rtl..sC G GJE35D    C     G G G G G GrJT)rRTL_CHARACTER_RANGESr)rrrs` r7 is_char_rtlrsc G G G G2F G G GGGt `>`~7_7_7_7_Q_7_7_7_7_ a aarJc|d}|dd|dkr4|dvr2|d}|dkrtd|dv.dSdSdS)Nr%)rUrWrJzFile ended unexpectedly.)raseekrrls r7skip_over_commentrs ++a..C KKA d{{''++a..Cczz$%?@@@''''{''rJ)limitrc@|d}|t| dd}|dkr\|||krtd|dt |t |}|dd|dz }|dkZdSdS)a Skip consecutive leading comments and the whitespace following each one. The stream must be positioned at the first comment, if any. If the next character is not ``%``, the stream position is unchanged. If only whitespace follows the last comment before EOF, leave the stream at the last whitespace byte, preserving the existing read_object behaviour for a missing object. Args: stream: The stream to read from. limit: The maximum number of comments to skip, or ``None`` for no limit. Raises: LimitReachedError: If another comment follows after ``limit`` comments; the stream remains at its ``%`` without reading its body. PdfStreamError: If a comment ends before a line ending. rrrNz,Maximum number of leading comments reached: .r)rarr/rrrm)r[rrh comment_counts r7skip_over_commentsr"s& ++a..C KKS 1M ++  %!7!7#$[SX$[$[$[\\ \&!!!!&)) B ++++++rJ)lengthregexrc*g}d}d}d} ||}|sd|St|}||zx} |krtd| d|d||z} || } ||| bt|} || z | z} || |z |z d d|d| S||z }|d d}|d kr|d z} ) a Read until the regular expression pattern matched (ignore the match). Treats EOF on the underlying stream as the end of the token to be matched. Args: stream: The stream to read from. regex: The pattern to search for. length: The (approximated) maximum number of bytes to read before raising an exception. Returns: The read bytes. rrJTzRead stream length of z# exceeds maximum allowed length of rNrii )rajoinr/rsearchrNrr)r[rrparts total_lengthtail chunk_sizetoken token_lengthcurrent_lengthcurrent_buffer search_matchoverlap actual_starts r7read_until_regexrAsSEL DJ J'' #88E?? "5zz *\9 9Nf D D#eee\beee  ||N33  U  #$iiG''1L4F4F4H4HHL KK |3lBA F F F88E??=L=1 1 $ STT{    1 J5rJto_readc||krtd|| t||}|| t|S)a 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. Args: stream: to_read: Returns: The data which was read. z!Could not read malformed PDF file)tellrrrra)r[rras r7read_block_backwardsrpsh{{}}w@AAA KK(### ;;w  D KK(### KrJcg}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. Args: stream: StreamType: Returns: The data which was read. FrTrs NrJr) rrrminrrr/rNrrr)r[ line_content found_crlfrblockidxs r7read_previous_liners"LJ {{}}9::: )6;;==99 a<< %VW55%jj1n "((uSz88q((uSz88axx!  '   cAgii 0 1 1 1((uSzW44q((uSzW44    & & & !88 KKa * * * A D 88L2& ' ''rJarwc:tfd|DS)Nc3^K|]&tfdtDV'dS)c3fK|]+}tdt|DV,dS)c3ZK|]&\}}t|t|zV'dSrt)float)rvr4js r7rxz6matrix_multiply....s7@@$!Q%((U1XX%@@@@@@rJN)sumzip)rvcolrows r7rxz,matrix_multiply...sB T TSc@@#c3--@@@@@ T T T T T TrJN)tupler)rvrrws @r7rxz"matrix_multiply..sZ   T T T TCQRG T T TTTrJ)r)rrws `r7matrix_multiplyrs=   rJcd}|| dtdd5}||||d|||dddn #1swxYwY|| ddS)z5Create text file showing current location in context.irzpypdf_pdfLocation.txtwbsHEREN)ropenwritera)r[radius output_fhs r7 mark_locationrsF KK %t , ,-  F++,,,    F++,,,--------------- KKsA&BB #B rmsg stacklevelc>tj|t|dS)N)r)warningswarnDeprecationWarning)rrs r7 deprecaters  M#)jAAAAAArJc t|rt)r)rs r7 deprecationrs 3  rJold_namenew_name removed_inc8t|d|d|dddS)zFIssue a warning that a feature will be removed, but has a replacement., is deprecated and will be removed in pypdf . Use instead.Nrrrrs r7deprecate_with_replacementrs> ffffS[fff rJc6t|d|d|ddS)zMRaise an exception that a feature was already removed, but has a replacement.( is deprecated and was removed in pypdf rrNrrs r7deprecation_with_replacementrs: bbZbbxbbbrJnamec2t|d|dddS)zCIssue a warning that a feature will be removed without replacement.rrrNrrrs r7deprecate_no_replacementrs* PP:PPPRSTTTTTrJc0t|d|ddS)zJRaise an exception that a feature was already removed without replacement.rrNrrs r7deprecation_no_replacementrs'4NNNNNOOOOOrJmessager_valuesc |r*tj|||dStj||dS)z Use this instead of logger.error directly. That allows people to overwrite it more easily. See the docs on when to use which: https://pypdf.readthedocs.io/en/latest/user/suppress-warnings.html N)logging getLoggererrorrr_rs r7 logger_errorrsX1&!!''88888&!!''00000rJc |r*tj|||dStj||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 pypdf, e.g. a non-compliant PDF being read in a way that pypdf could apply a robustness fix to still read it. This applies mainly to strict=False mode. N)rrwarningrs r7rdrdsZ 3&!!))'6::::: &!!))'22222rJ func_namekwargsaliasesfailc (|D]|\}}||vrs|rt|d|d||vrt|d|d|d|d|d ||||<t j|d|dt d }d S) z Helper function to deprecate arguments. Args: func_name: Name of the function to be deprecated kwargs: aliases: fail: z# is deprecated as an argument. Use z insteadz received both z and z as an argument. z is deprecated. Use rr)rcategoryrN)itemsr TypeErrorpoprrr)rrrrold_termnew_terms r7 rename_kwargsr!s&mmoo( v   &VVHVVV6!! //////$,//#/// &zz(33F8  MVVHVVV+     rJrfcf|dkr|dS|dkr |dz ddS|dkr |dz ddS|dz ddS) Niz Bytei@Bz.1fz kBiʚ;z MBz GBr)rfs r7_human_readable_bytesrDsk u}} u}}%-((((( u}}%-(((((em $ $ $ $$rJc8t|dr,tt|jS|}|dtj|}|||S)N getbufferr) hasattrrr?rnbytesrrosSEEK_END)r[positionsizes r7get_stream_sizer Nszv{##4C))++2333{{}}H KK2; ;;==D KK KrJcJt|dr!t|S|}|dt j|d}||tt|S)Nrrr) rrfrrrrSEEK_SETrar)r[rdatas r7get_stream_datarYsv{##)V%%''((({{}}H KK2; ;;r??D KK t  rJc6eZdZdZdddZddefdZdefdZdS) classpropertyz Decorator that converts a method with a single cls argument into a property that can be accessed directly from the class. Nrc||_dSrtfgetselfmethods r7__init__zclassproperty.__init__s  rJc,||Srtr)rinstanceclss r7__get__zclassproperty.__get__syy~~rJc||_|Srtrrs r7getterzclassproperty.getters  rJrt)rN) re __module__ __qualname____doc__rr rrrrrJr7rrsr SrJrcpeZdZUddlmZdZeed< dZe ed< dZ e eed< d efd Z d efd Z dS) Filer)IndirectObjectr$rrJr Nindirect_referencerc t|jjd|jdtt |jdS)Nz(name=z, data: )) __class__rerrr/r rs r7__str__z File.__str__s>.)llllDYZ]^b^gZhZhDiDillllrJcl|dddt|jdzS)Nrz, hash: r&)r)hashr r(s r7__repr__z File.__repr__s2||~~crc"%BTY%B%B%BBBrJ)rerrgenericr#rstr__annotations__r rfr$r r)r,rrJr7r"r"s''''''D#NNND%480777mmmmmC#CCCCCCrJr"ceZdZejdZdeddfdZdedee e effdZ de de fdZde fd Zdede fd ZdS) Versionz ^(\d+)(.*)$ version_strrNcH||_|||_dSrt)r2_parse_version components)rr2s r7rzVersion.__init__s#&--k::rJcT|d}g}|D]}tj|}|s|d|f;|d}|d}|d}|t ||f|S)Nrrrr)splitr1COMPONENT_PATTERNmatchrNgroupr?)rr2r5parsed_components componentr9integer_prefixsuffixs r7r4zVersion._parse_versions &&s++ # D DI-33I>>E !((!Y888"[[^^N[[^^F%!"  $ $c.&9&96%B C C C C  rJothercPt|tsdS|j|jkSNF) isinstancer1r5)rr?s r7__eq__zVersion.__eq__s)%)) 5%"222rJcRt|jt|jfSrt)r+r'rr5r(s r7__hash__zVersion.__hash__s!T^U4?%;%;<===rJcZt|tstdt|t |j|jD]3\}}|\}}|\}}||krdS||krdS||krdS||krdS4t |jt |jkS)Nz#Version cannot be compared against TF)rBr1r2typerr5r/)rr?self_componentother_component self_value self_suffix other_value other_suffixs r7__lt__zVersion.__lt__s%)) RP4;;PPQQ Q/24?EDT/U/U   +NO&4 #J (7 %KK''ttK''uu\))tt\))uu*4?##c%*:&;&;;;rJ)rerrrecompiler8r.rlistrr?r4objectrrCrEr rNrrJr7r1r1s" >22;C;D;;;; !# !$uS#X2G ! ! ! !3F3t3333 >#>>>>rhs|8)((  / !!!!!!''''''''""""""BQB7""       ++++++w&&&&&&', % eUE&9 :E%PUBURUQ\,,, ,75>,3,af,,,,^!1CE27(z7(e7(7(7(7(t$< *     BB3BCBBBBB S X    QU3#3S[U3UCUDUUUU PSPcPhPPPP 1# 1# 1 1 1 1 1 13C3C333433334SX    cN 59#s(^ KO      F%%%%%%J3J5\" CCCCCCC C. 1<1<1<1<1<1<1<1