U NfnN@sUdZdZdZddlZddlZddlZddlZddlZddlm Z ddl m Z m Z ddl m Z mZddlmZdd lmZmZmZmZmZmZmZmZmZmZejdd d krdd lmZn dd lmZd dlm Z m!Z!m"Z"eee#e#e#fee#e#e#fee#e#e#ffZ$ee%d<ee#e#e#e#e#e#fZ&ee%d<eeZ'ee(e'fZ)dZ*dZ+dZ,dZ-ee(ee dddZ.e(e(e(dddZ/die'ee0e1dddZ2e'e1ddd Z3e'e4dd!d"Z5e1e4d#d$d%Z6e'ddd&d'Z7e'ee1e1d(d)d*Z8e'e0e1d+d,d-Z9e'e1dd.d/Z:e$e$e$d0d1d2Z;e'ddd3d4Zee(d9d:d;Z?ee(e0d9dd=Z@ee0e0d9d?d=Z@ee0e(e1fee0e1fd9d@d=Z@dAZAdBZBe0e0e0e0dCdDdEZCdje(e0ddGdHdIZDe(ddJdKdLZEe(e(e(ddMdNdOZFe(e(e(ddMdPdQZGe(e(ddRdSdTZHe(e(ddRdUdVZIe(e(ddWdXdYZJe(e(ddWdZd[ZKdke(ee(efee(e(fe4dd]d^d_ZLe0e(d`dadbZMe GdcddddZNe GdedfdfeNZOejPGdgdhdhZQdS)lz"Utility functions for PDF library.zMathieu Fenniakzbiziqe@mathieu.fenniak.netN) dataclass)datetimetimezone)DEFAULT_BUFFER_SIZEBytesIO)SEEK_CUR) IOAnyDictListOptionalPatternTupleUnioncastoverload) ) TypeAlias)STREAM_TRUNCATED_PREMATURELYDeprecationErrorPdfStreamErrorTransformationMatrixTypeCompressedTransformationMatrixz1{} is deprecated and will be removed in pypdf {}.z-{} is deprecated and was removed in pypdf {}.zA{} is deprecated and will be removed in pypdf {}. Use {} instead.z={} is deprecated and was removed in pypdf {}. Use {} instead.)textreturnc Cs|}|dkrdS|dr$d|}|dr6|d7}|dddddd }t|d|d }|dkr|t|d kr|d 7}d D]T}zt||}Wntk rYqYqX|dddkr|jt j d}|Sqtd|dS)NrzD:)ZzZ0000r+r'-Z00)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: ) isdigitendswithreplacemaxfindlenrstrptime ValueErrorrutc)rZorgtextifdr30/tmp/pip-unpacked-wheel-zevpxvmc/pypdf/_utils.pyparse_iso8824_dateRs(     r5)header1header2rcCsfd}g}||kr |||||kr8|||t|dkrZtd|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,r.r*)r6r7versionsZpdf_header_indicesr3r3r4_get_max_pdf_version_headerrs r;)streammaxcharsrcCs8d}|d}|s4|sq4||7}t||krq4q|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. r)readisspacer,)r<r=txttokr3r3r4read_until_whitespaces    rC)r<rcCs"|d}|tkr|d}q |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)r? WHITESPACESr<rBr3r3r4read_non_whitespaces  rFcCs0td}d}|tkr(|d}|d7}q |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)rDr?)r<rBZcntr3r3r4skip_over_whitespaces   rG)valuercCs4tt|D]"}|||d}|tkr dSq 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. rFT)ranger,rD)rHr9currentr3r3r4check_if_whitespace_onlys rKcCs6|d}|dd|dkr2|dkr2|d}qdS)Nr%)  )r?seekrEr3r3r4skip_over_comments   rQ)r<regexrcCspd}|d}|s|S|||}|dk rb||t|t|d||d|}ql||7}q|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. r>Nr)r?searchrPstartr,)r<rRnamerBmr3r3r4read_until_regexs    rX)r<to_readrcCs>||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)tellrrPrr?)r<rYr?r3r3r4read_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. Args: stream: StreamType: Returns: The data which was read. Frrs TNr>rL) rZrrminrr[r,r8rPrjoin)r<Z line_contentZ found_crlfrYblockidxr3r3r4read_previous_line s.      r`)abrcstfdd|DS)Nc3s(|] tfddtDVqdS)c3s&|]}tddt|DVqdS)css"|]\}}t|t|VqdSN)float).0r0jr3r3r4 Gsz6matrix_multiply....N)sumzip)recolrowr3r4rgGsz,matrix_multiply...N)tupleri)rerbrkr4rgFsz"matrix_multiply..)rm)rarbr3rnr4matrix_multiplyCs roc Csdd}|| dtdd0}||||d|||W5QRX|| ddS)z5Create text file showing current location in context.irzpypdf_pdfLocation.txtwbsHEREN)rPopenwriter?)r<ZradiusZ output_fhr3r3r4 mark_locationLs  rsB_CACHE)srcCst|tr|St}||kr"||Sz$|d}t|dkrB|||<|WStk r||d}t|dkrt|||<|YSXdS)Nlatin-1rzutf-8) isinstancebytesrtencoder, Exception)ruZbcrr3r3r4b_[s     r|)rbrcCs t|tr|dSt|SdS)Nrv)rwrxdecodestrrnr3r3r4str_ms  rcCsdSrcr3rnr3r3r4ord_tsrcCsdSrcr3rnr3r3r4ryscCsdSrcr3rnr3r3r4r~scCst|trt|S|Src)rwr~ordrnr3r3r4rs ) rNrO s[ ])leftupup_leftrcCsX|||}t||}t||}t||}||krD||krD|S||krP|S|SdSrc)abs)rrrpZ dist_leftZdist_upZ dist_up_leftr3r3r4paeth_predictors    rr)msg stacklevelrcCstj|t|ddS)N)r)warningswarnDeprecationWarning)rrr3r3r4 deprecatesr)rrcCs t|dSrc)r)rr3r3r4 deprecationsr)old_namenew_name removed_inrcCstt|||ddS)zIRaise an exception that a feature will be removed, but has a replacement.N)rDEPR_MSGformatrrrr3r3r4deprecate_with_replacementsrcCstt|||dS)zMRaise an exception that a feature was already removed, but has a replacement.N)rDEPR_MSG_HAPPENEDrrr3r3r4deprecation_with_replacementsr)rVrrcCstt||ddS)zFRaise an exception that a feature will be removed without replacement.rN)rDEPR_MSG_NO_REPLACEMENTrrVrr3r3r4deprecate_no_replacementsrcCstt||dS)zJRaise an exception that a feature was already removed without replacement.N)r DEPR_MSG_NO_REPLACEMENT_HAPPENEDrrr3r3r4deprecation_no_replacementsr)rsrcrcCst||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 getLoggererrorrrr3r3r4 logger_errors rcCst||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)rrwarningrr3r3r4logger_warningsrF) func_namekwargsaliasesfailrc Cs|D]~\}}||kr|r0t|d|d||kr^t|d|d|d|d|d ||||<tj|d|dtdqd 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)rrrrZold_termZnew_termr3r3r4 rename_kwargss  r)rxrcCsV|dkr|dS|dkr*|dddS|dkrB|dddS|dddSdS) Niz Bytei@Bz.1fz kBiʚ;z MBz GBr3)rxr3r3r4_human_readable_bytess rc@sfeZdZUddlmZeed<eed<dZe e ed<dZ e eed<edd d Z edd d Z dS) FilerIndirectObjectrVdataNimageindirect_reference)rcCs&|jjd|jdtt|jdS)Nz(name=z, data: )) __class____name__rVrr,rselfr3r3r4__str__ sz File.__str__cCs"|dddt|jdS)NrLz, hash: r)rhashrrr3r3r4__repr__sz File.__repr__)r __module__ __qualname__genericrr~__annotations__rxrr r rrrr3r3r3r4rs  rc@sLeZdZUddlmZdZeeed<dZ eeed<eeddddZ dS) ImageFilerrNrr) new_imagerrc Ks8ddlm}ddlm}ddlm}ddlm}m}|j dkrFt dt |j j d s\t d t ||jspt d t}|j|d f|||} | jdjdj dk st| jdjdj |j j j|j jd<|j t||j _ |t||j \} } } | dk s t|jd|jd | |_| |_| |_dS)a Replace the Image with a new PIL image. Args: new_image (PIL.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 imageZ_id_translatedz4Can not update an image not belonging to a PdfWriterznew_image shall be a PIL ImageZPDF.)ZPILrZ_readerrfiltersrrrrrrhasattrZpdfrwrsaveZpagesZimagesAssertionErrorZ get_object_objectsZidnumrrVrfindrr) rrrrrrrrrbreader extensionZ byte_streamimgr3r3r4r)s8      zImageFile.replace) rrrrrrr r rrr)r3r3r3r4rs  rc@sbeZdZedZeddddZeee e efdddZ e e dd d Zee dd d ZdS) Versionz ^(\d+)(.*)$N) version_strrcCs||_|||_dSrc)r_parse_version components)rrr3r3r4__init__RszVersion.__init__cCsn|d}g}|D]V}tj|}|s6|d|fq|d}|d}|dkrVd}|t||fq|S)Nrrrr)splitrCOMPONENT_PATTERNmatchr8groupint)rrrZparsed_components componentrZinteger_prefixsuffixr3r3r4rVs    zVersion._parse_version)otherrcCst|tsdS|j|jkS)NF)rwrr)rrr3r3r4__eq__es zVersion.__eq__cCst|tstdt|tt|jt|j}t|D]X}|j|\}}|j|\}}||krhdS||krvdS||krdS||kr:dSq:t|jt|jkS)Nz#Version cannot be compared against TF)rwrr.typer\r,rrI)rrZmin_lenr0Z self_valueZ self_suffixZ other_valueZ other_suffixr3r3r4__lt__js  zVersion.__lt__)rrrrecompilerr~rr rrrobjectboolrr rr3r3r3r4rNs  r)N)r)F)R__doc__ __author____author_email__ functoolsrrsysrZ dataclassesrrriorrosrtypingrr r r r r rrrr version_inforZtyping_extensionserrorsrrrrdrrrZ StreamTyper~Z StrByteTyperrrrr5r;rrxrCrFrrGrKrQrXr[r`rorsrtr|rrrDZWHITESPACES_AS_REGEXPrrrrrrrrrrrrrtotal_orderingrr3r3r3r4s  0  &    :  "   ! ;