KhO UdZdZdZddlZddlZddlZddlZddlZddlm Z ddl m Z m Z ddl m Z mZddlmZdd lmZmZmZmZmZmZmZmZmZmZmZejdd d krdd lmZndd lmZd dl m!Z!m"Z"m#Z#eee$e$e$fee$e$e$fee$e$e$ffZ%ee&d<ee$e$e$e$e$e$fZ'ee&d<eeZ(ee)e(fZ*dZ+dZ,dZ-dZ.dee)dee fdZ/de0de0de0fdZ1d\de(dee2de0fdZ3de(de0fdZ4de(de5fdZ6d e0de5fd!Z7de(ddfd"Z8de(d#ee0de0fd$Z9de(d%e2de0fd&Z:de(de0fd'Z;d(e%d)e%de%fd*Zeee)e0fe0fe&d,<d-ee)e0fde0fd.Z?ed)e)de)fd/Z@ed)e0de)fd0Z@d)ee)e0fde)fd1Z@ed)e)de2fd2ZAed)e0de0fd3ZAed)e2de2fd4ZAd)ee2e)e0fdee2e0ffd5ZAd6ZBd7ZCd8e2d9e2d:e2de2fd;ZDd]d=e)d>e2ddfd?ZEd=e)ddfd@ZF d^dBe)dCe)dDe)ddfdEZG d^dBe)dCe)dDe)ddfdFZHd^dGe)dDe)ddfdHZId^dGe)dDe)ddfdIZJd=e)dJe)ddfdKZKdLe)dedMeffdNZL d_dPe)dQee)efdLee)e)fdRe5ddf dSZMdTe2de)fdUZNe GdVdWZOe GdXdYeOZPejQGdZd[ZRdS)`z"Utility functions for PDF library.zMathieu Fenniakzbiziqe@mathieu.fenniak.netN) dataclass)datetimetimezone)DEFAULT_BUFFER_SIZEBytesIO)SEEK_CUR) IOAnyCallableDictListOptionalPatternTupleUnioncastoverload) ) TypeAlias)STREAM_TRUNCATED_PREMATURELYDeprecationErrorPdfStreamErrorTransformationMatrixTypeCompressedTransformationMatrixz1{} is deprecated and will be removed in pypdf {}.z-{} is deprecated and was removed in pypdf {}.zA{} is deprecated and will be removed in pypdf {}. Use {} instead.z={} is deprecated and was removed in pypdf {}. Use {} instead.textreturnc|}|dS|drd|z}|dr|dz }|dddddd }t|d|d }|dkr|t |d z kr|d z }d D]Y} t j||}|dddkr |tj }|cS#t$rYVwxYwtd|)NrzD:)Zz0000r"+r!'-00)zD:%YzD:%Y%mzD:%Y%m%dz D:%Y%m%d%Hz D:%Y%m%d%H%MzD:%Y%m%d%H%M%SzD:%Y%m%d%H%M%S%zz+0000)tzinfozCan not convert date: ) isdigitendswithreplacemaxfindlenrstrptimerutc ValueError)rorgtextifds T/var/www/html/services/rag-service/venv/lib/python3.11/site-packages/pypdf/_utils.pyparse_iso8824_dater:Ss_G |t Awd{ }}Z    <<S ! ! ) )#s 3 3 ; ;C D DD DIIcNNDIIcNN++A1uuc$ii!m##   !$**ABCCyG##IIX\I22HHH     H  7g77 8 88sD!! D.-D.header1header2c6d}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)appendindexr1r4r/)r;r<versionspdf_header_indicess r9_get_max_pdf_version_headerrBssH(!!(.."9"9:::(!!(.."9"9::: !##RGRRGRRRSSS C*++ ,,streammaxcharscd} |d}|s|sn||z }t||krnF|S)a; Read non-whitespace characters and return them. Stops upon encountering whitespace or when maxchars is reached. Args: stream: The data stream from which was read. maxchars: The maximum number of bytes returned; by default unlimited. Returns: The data which was read. rCTr)readisspacer1)rDrEtxttoks r9read_until_whitespacerKs` Ckk!nn ;;==    s  s88x     JrCc~|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. r)rG WHITESPACESrDrJs r9read_non_whitespacerOs? ++a..C   kk!nn    JrCctd}d}|tvr#|d}|dz }|tv#|dkS)a Similar to read_non_whitespace, but return a boolean if more than one whitespace character was read. Args: stream: The data stream from which was read. Returns: True if more than one whitespace was skipped, otherwise return False. rr)rMrG)rDrJcnts r9skip_over_whitespacerRsM a.C C   kk!nn q    7NrCvaluecvtt|D]}|||dz}|tvrdSdS)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. rFT)ranger1rM)rSr?currents r9check_if_whitespace_onlyrWsOs5zz""eai( + % %55 & 4rCc|d}|dd|dkr|dvr|d}|dvdSdSdS)Nr%)  )rGseekrNs r9skip_over_commentr^sh ++a..C KKA d{{''++a..C''''{''rCregexcNd} |d}|s|S|||z}|h||t |t |zz d||zd|}n||z }|S)z 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: regex: re.Pattern Returns: The read bytes. rCTNr)rGsearchr]startr1)rDr_namerJms r9read_until_regexrfs D kk"oo K LL $ $ = KK SYYS%9:A > > >3J !'')) ,D     KrCto_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)tellrr]rrG)rDrgrGs r9read_block_backwardsrjsh{{}}w@AAA KK(### ;;w  D KK(### KrCcg}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 NrCrY) rirrminrrjr1r>r]rjoin)rD line_content found_crlfrgblockidxs r9read_previous_linerr s LJ {{}}9::: )6;;==99 a<< %VW55%jj1n "((uSz88q((uSz88axx!  '   cAgii 0 1 1 1((uSzW44q((uSzW44    & & & !88 KKa * * * A D 88L2& ' ''rCabc:tfd|DS)Nc3^K|]&tfdtDV'dS)c3fK|]+}tdt|DV,dS)c3ZK|]&\}}t|t|zV'dSN)float).0r6js r9 z6matrix_multiply....Hs7@@$!Q%((U1XX%@@@@@@rCN)sumzip)r{colrows r9r}z,matrix_multiply...HsB T TSc@@#c3--@@@@@ T T T T T TrCN)tupler)r{rrts @r9r}z"matrix_multiply..GsZ   T T T TCQRG T T TTTrC)r)rsrts `r9matrix_multiplyrDs=   rCcd}|| dtdd5}||||d|||dddn #1swxYwY|| ddS)z5Create text file showing current location in context.irzpypdf_pdfLocation.txtwbsHEREN)r]openwriterG)rDradius output_fhs r9 mark_locationrMsF KK %t , ,-  F++,,,    F++,,,--------------- KKsA&BB #B B_CACHEsc4t|tr|St}||vr||S |d}t |dkr|||<|S#t $r2|d}t |dkr|||<|cYSwxYw)Nlatin-1rzutf-8) isinstancebytesrencoder1 Exception)rbcrs r9b_r\s!U BBww!u   HHY   q66A::BqE  HHW   q66A::BqE s.A9BBcdSryrts r9str_rnCrCcdSryrrs r9rrsrrCcZt|tr|dS|S)Nr)rrdecoders r9rrxs+!Uxx """rCcdSryrrs r9ord_rrrCcdSryrrs r9rrrrCcdSryrrs r9rrrrCcNt|trt|S|Sry)rstrordrs r9rrs$!S1vv HrC) r[r\ s[ ]leftupup_leftc||z|z }t||z }t||z }t||z }||kr||kr|S||kr|S|Sry)abs)rrrp dist_leftdist_up dist_up_lefts r9paeth_predictorrsn r GAAH I!b&kkGq7{##LG \ 9 9 L  rCrmsg stacklevelc>tj|t|dS)N)r)warningswarnDeprecationWarning)rrs r9 deprecaters  M#)jAAAAAArCc t|ry)r)rs r9 deprecationrs 3  rC3.0.0old_namenew_name removed_incZtt|||ddS)zIRaise an exception that a feature will be removed, but has a replacement.N)rDEPR_MSGformatrrrs r9deprecate_with_replacementrs*hooh H==qAAAAArCcXtt|||dS)zMRaise an exception that a feature was already removed, but has a replacement.N)rDEPR_MSG_HAPPENEDrrs r9deprecation_with_replacementrs+!((:xHHIIIIIrCrdcXtt||ddS)zFRaise an exception that a feature will be removed without replacement.rN)rDEPR_MSG_NO_REPLACEMENTrrdrs r9deprecate_no_replacementrs) %,,T:>>BBBBBrCcVtt||dS)zJRaise an exception that a feature was already removed without replacement.N)r DEPR_MSG_NO_REPLACEMENT_HAPPENEDrrs r9deprecation_no_replacementrs'077jIIJJJJJrCsrccTtj||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)logging getLoggerwarning)rrs r9logger_warningrs)  c""3'''''rCaliases.c Jdtdtfdtffd }|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. func.rcvtjdtdtdtffd }|S)NargskwargsrcDtj|d|i|S)NT)fail) rename_kwargs__name__)rrrrs r9wrapperz9deprecation_bookmark..decoration..wrappers1 $-t D D D D4((( (rC) functoolswrapsr )rrrs` r9 decorationz(deprecation_bookmark..decorationsX    )3 )# )# ) ) ) ) ) )   )rC)r r )rrs` r9deprecation_bookmarkrs@#s(+ rCF func_namerrc &|D]{\}}||vrr|rt|d|d||vrt|d|d|d|d|d ||||<t j|d|dt |d S) z Helper function to deprecate arguments. Args: func_name: Name of the function to be deprecated kwargs: aliases: fail: z# is deprecated as an argument. Use z insteadz received both z and z as an argument. z is deprecated. Use z instead.)messagecategoryN)itemsr TypeErrorpoprrr)rrrrold_termnew_terms r9rrs&mmoo( v   &VVHVVV6!! //////$,//#/// &zz(33F8  MVVHVVV+     rCrcf|dkr|dS|dkr |dz ddS|dkr |dz ddS|dz ddS) Niz Bytei@Bz.1fz kBiʚ;z MBz GBr)rs r9_human_readable_bytesrsk u}} %-((((( %-(((((%-(((((rCc|eZdZUddlmZeed<eed<dZe e ed<dZ e eed<defd Z defd Z dS) FilerIndirectObjectrddataNimageindirect_referencerc t|jjd|jdtt |jdS)Nz(name=z, data: )) __class__rrdrr1rselfs r9__str__z File.__str__"s>.)llllDYZ]^b^gZhZhDiDillllrCcl|dddt|jdzS)NrYz, hash: r)rhashrrs r9__repr__z File.__repr__%s2||~~crc"%BTY%B%B%BBBrC)r __module__ __qualname__genericrr__annotations__rrrr rrrrrCr9rrs'''''' III KKKE8C=370777mmmmmC#CCCCCCrCrcdeZdZUddlmZdZeeed<dZ eeed<dededdfd Z dS) ImageFilerrNrr new_imagerrc vddlm}ddlm}ddlm}ddlm}m}|j tdt|j j d std t||jstd t}|j|d fi|||} | jdjdj J| jdjdj |j j j|j jdz <|j t+||j _ |t+||j \} } } | J|jd|jd | z|_| |_| |_dS)a Replace the Image with a new PIL image. Args: new_image (Image.Image): The new PIL image to replace the existing image. **kwargs: Additional keyword arguments to pass to `Image.Image.save()`. Raises: TypeError: If the image is inline or in a PdfReader. TypeError: If the image does not belong to a PdfWriter. TypeError: If `new_image` is not a PIL Image. Note: This method replaces the existing image with a new image. It is not allowed for inline images or images within a PdfReader. The `kwargs` parameter allows passing additional parameters to `Image.Image.save()`, such as quality. r)Imager) PdfReader)_xobj_to_image)DictionaryObject PdfObjectNzCan not update an inline image_id_translatedz4Can not update an image not belonging to a PdfWriterznew_image shall be a PIL ImagePDF.)PILr_readerrfiltersr rr r rrhasattrpdfrrsavepagesimages get_object_objectsidnumrrdrfindrr) rrrrrr r r rtreader extension byte_streamimgs r9r.zImageFile.replace0s& &&&&&& ,+++++88888888  " *<== =t.24DEE TRSS S)U[11 ><== = II q%**6***1|A%a(;GGG LO "1 % 8 C C E E #,T-D-JQ-NO $6 t.99;;  &4n !4#:#E#E#G#G H H' ' # ;$$$I4  4 445 A   rC) rrrrrrrr rrr.rrCr9rr)s{''''''E8C=370777222222222rCrceZdZejdZdeddfdZdedee e effdZ de de fdZdede fd ZdS) Versionz ^(\d+)(.*)$ version_strrNcH||_|||_dSry)r!_parse_version components)rr!s r9__init__zVersion.__init__is#&--k::rCcT|d}g}|D]}tj|}|s|d|f;|d}|d}|d}|t ||f|S)Nrrrr)splitr COMPONENT_PATTERNmatchr>groupint)rr!r$parsed_components componentr)integer_prefixsuffixs r9r#zVersion._parse_versionms &&s++ # D DI-33I>>E !((!Y888"[[^^N[[^^F%!"  $ $c.&9&96%B C C C C  rCothercPt|tsdS|j|jkS)NF)rr r$)rr0s r9__eq__zVersion.__eq__|s)%)) 5%"222rCct|tstdt|t t |jt |j}t|D]F}|j|\}}|j|\}}||krdS||krdS||krdS||krdSGt |jt |jkS)Nz#Version cannot be compared against TF)rr r4typerlr1r$rU)rr0min_lenr6 self_value self_suffix other_value other_suffixs r9__lt__zVersion.__lt__s%)) RP4;;PPQQ Qc$/**C0@,A,ABBw  A&*oa&8 #J (-(8(; %KK''ttk))uu\))tt|++uu,4?##c%*:&;&;;;rC)rrrrecompiler(rr%r rr+r#objectboolr2r r:rrCr9r r es" >22;C;D;;;; !# !$uS#X2G ! ! ! !3F3t3333 rRrWr^rfrjrrrrrrrrrMWHITESPACES_AS_REGEXPrrrrrrrrrrrrrtotal_orderingr rrCr9rNs8)((  / !!!!!!''''''''++++++++                          BQB7""       ++++++ ', % eUE&9 :E%PUBU*SX cN59#s(^KO B))))))  C C C C C C C  C 8888888 8v -<-<-<-<-<-<-<-<-<-