j> xddlZddlZddlZddlZddlmZddlmZmZddlm Z ddl m Z m Z m Z mZmZmZmZmZmZmZmZddlmZmZddlmZmZdd lmZdd lm Z m!Z!m"Z"m#Z#m$Z$m%Z%m&Z&m'Z'm(Z(m)Z)m*Z*dd l+m,Z-dd l+m.Z/dd l+m0Z0m1Z1m2Z2ddl+m3Z4ddl+m5Z6ddl+m7Z8ddl+m9Z:ddl+m;Z<ddl+m=Z>ddl?m@Z@mAZAmBZBmCZCmDZDddlEmFZFmGZGmHZHmIZImJZJmKZKmLZLmMZMmNZNmOZOmPZPmQZQmRZRmSZSmTZTmUZUmVZVmWZWmXZXddlYmZZZm[Z[ddl\m]Z]de^de_dee_ee dfffdZ`GddeKZaGdd ZbGd!d"ee e fZcdS)#N)datetime)BytesIOUnsupportedOperation)Path) AnyCallableDictIterableIteratorListMappingOptionalTupleUnioncast) Encryption PasswordType) PageObject _VirtualList) index2label) StrByteType StreamTypeb_deprecate_with_replacementlogger_warningparse_iso8824_dateread_non_whitespaceread_previous_lineread_until_whitespaceskip_over_commentskip_over_whitespace)CatalogAttributes)CatalogDictionary)CheckboxRadioButtonAttributesGoToActionArgumentsUserAccessPermissions)Core)DocumentInformationAttributes)FieldDictionaryAttributes)PageAttributes)PagesAttributes) TrailerKeys)EmptyFileErrorFileNotDecryptedError PdfReadErrorPdfStreamErrorWrongPasswordError) ArrayObject BooleanObject ContentStreamDecodedStreamObject DestinationDictionaryObjectEncodedStreamObjectFieldFit FloatObjectIndirectObject NameObject NullObject NumberObject PdfObjectTextStringObject TreeObjectViewerPreferences read_object) OutlineType PagemodeType)XmpInformationdsizereturn.c|dkrtdd|z}|dd}tjd|dS)Nzinvalid size in convert_to_intsiz>qr)r0structunpack)rIrJs W/lsinfo/ai/hellotax_ai/base_platform/venv/lib/python3.11/site-packages/pypdf/_reader.pyconvert_to_intrQpsH axx;<<<+a/A "##A =q ! !! $$cfeZdZdZddZdedeefdZedeefdZ edeefdZ edeefd Z edeefd Z edeefd Z edeefd Zedeefd ZedeefdZedeefdZedeefdZedeefdZedeefdZedeefdZedeefdZdS)DocumentInformationa A class representing the basic document metadata provided in a PDF File. This class is accessible through :py:class:`PdfReader.metadata`. All text properties of the document metadata have *two* properties, eg. author and author_raw. The non-raw property will always return a ``TextStringObject``, making it ideal for a case where the metadata is being displayed. The raw property can sometimes return a ``ByteStringObject``, if pypdf was unable to decode the string's text encoding; this requires additional safety in the caller and therefore is not as commonly accessed. rKNc.tj|dSN)r8__init__selfs rPrWzDocumentInformation.__init__s!$'''''rRkeyc`||d}t|tr|SdSrV)get isinstancerB)rYrZretvals rP _get_textzDocumentInformation._get_texts2#t$$ f. / / MtrRc|tjrP|tjp0|tjndS)z Read-only property accessing the document's title. Returns a ``TextStringObject`` or ``None`` if the title is not specified. N)r\DITITLEr_ get_objectrXs rPtitlezDocumentInformation.titlesRxx!! DNN28 $ $ G(:(:(E(E(G(G rRc@|tjS)z>The "raw" version of title; can return a ``ByteStringObject``.)r\rarbrXs rP title_rawzDocumentInformation.title_rawsxx!!!rRc@|tjS)z Read-only property accessing the document's author. Returns a ``TextStringObject`` or ``None`` if the author is not specified. )r_raAUTHORrXs rPauthorzDocumentInformation.authors~~bi(((rRc@|tjS)z?The "raw" version of author; can return a ``ByteStringObject``.)r\rarhrXs rP author_rawzDocumentInformation.author_rawsxx """rRc@|tjS)z Read-only property accessing the document's subject. Returns a ``TextStringObject`` or ``None`` if the subject is not specified. )r_raSUBJECTrXs rPsubjectzDocumentInformation.subjects~~bj)))rRc@|tjS)z@The "raw" version of subject; can return a ``ByteStringObject``.)r\rarmrXs rP subject_rawzDocumentInformation.subject_rawxx ###rRc@|tjS)ac Read-only property accessing the document's creator. If the document was converted to PDF from another format, this is the name of the application (e.g. OpenOffice) that created the original document from which it was converted. Returns a ``TextStringObject`` or ``None`` if the creator is not specified. )r_raCREATORrXs rPcreatorzDocumentInformation.creators~~bj)))rRc@|tjS)z@The "raw" version of creator; can return a ``ByteStringObject``.)r\rarsrXs rP creator_rawzDocumentInformation.creator_rawrqrRc@|tjS)aG Read-only property accessing the document's producer. If the document was converted to PDF from another format, this is the name of the application (for example, OSX Quartz) that converted it to PDF. Returns a ``TextStringObject`` or ``None`` if the producer is not specified. )r_raPRODUCERrXs rPproducerzDocumentInformation.producers~~bk***rRc@|tjS)zAThe "raw" version of producer; can return a ``ByteStringObject``.)r\rarxrXs rP producer_rawz DocumentInformation.producer_rawsxx $$$rRcZt|tjS)z:Read-only property accessing the document's creation date.)rr_ra CREATION_DATErXs rP creation_datez!DocumentInformation.creation_dates!"$..1A"B"BCCCrRc@|tjS)z The "raw" version of creation date; can return a ``ByteStringObject``. Typically in the format ``D:YYYYMMDDhhmmss[+Z-]hh'mm`` where the suffix is the offset from UTC. )r\rar}rXs rPcreation_date_rawz%DocumentInformation.creation_date_rawsxx()))rRcZt|tjS)z Read-only property accessing the document's modification date. The date and time the document was most recently modified. )rr_raMOD_DATErXs rPmodification_datez%DocumentInformation.modification_dates "$.."="=>>>rRc@|tjS)z The "raw" version of modification date; can return a ``ByteStringObject``. Typically in the format ``D:YYYYMMDDhhmmss[+Z-]hh'mm`` where the suffix is the offset from UTC. )r\rarrXs rPmodification_date_rawz)DocumentInformation.modification_date_rawsxx $$$rR)rKN)__name__ __module__ __qualname____doc__rWstrrr_propertyrdrfrirkrnrprtrvryr{rr~rrrrRrPrTrTxs  ((((SXc]   x}    X  "8C="""X") )))X)#HSM###X#*#***X*$Xc]$$$X$ *# * * *X *$Xc]$$$X$ +(3- + + +X +%hsm%%%X%Dx1DDDXD*8C=***X*?8H#5???X?%x}%%%X%%%rRrTc eZdZdZedeefdZ dedee e fde dede e fddfd Z dfd edee fd edee fdee effd Zede fd ZedeefdZedeefdZdefdZdedefdZedee effdZ dgdeedeeeefdeedeee effdZdede fdZ deeefdeeefdededdf dZ!deeefdededdfdZ"dedededdfdZ#dhde dee effd Z$dee%e&e'fde(efd!Z) dfdeedfdeedee effd"Z*ede+fd#Z, dfd$eed%eede+fd&Z-edee.fd'Z/d(edee0e'fdeefd)Z1d*edeefd+Z2d,e3deefd-Z4d.e d/ee(ee5e'de0efde3fd0Z6d$edee3fd1Z7ede(efd2Z8ede(e fd3Z9edee fd4Z:edee;fd5Z< dgd6edeefd7eee efd(ee'ddfd8Z=d(e'deee&e ffd9Z>d:ed;edee&fd<Z?d(eee'fdee&fd=Z@deAdeBeeffd>ZCd?ed@edee&fdAZDd?ed@edBee&dee&fdCZEdeAddfdDZFdeAddfdEZGdeAddfdFZHdeAdefdGZIdeAddfdHZJdeAdIeedJeddfdKZKdeAdeefdLZLdeAdIedeefdMZMdeAdeeNeOePffdNZQeRdeAdIedefdOZSdeAddfdPZTdQe(edReUegeeeBedSfffdTeUeeeeBedSffge fddfdUZVd/e(edeeBeeffdVZWdee e fdeXfdWZYdXedee e ffdYZZedee[fdZZ\ede fd[Z]edeee effd\Z^d]e deefd^Z_d]e deefd_Z`edeae e(e ffd`Zbde(e fdaZcd]e de(e fdbZd didcee dee ee e(e fffddZedS)j PdfReadera Initialize a PdfReader object. This operation can take some time, as the PDF stream's cross-reference tables are read into memory. Args: stream: A File object or an object that supports the standard read and seek methods similar to a File object. Could also be a string representing a path to a PDF file. strict: Determines whether user should be warned of all problems and also causes some correctable problems to be fatal. Defaults to ``False``. password: Decrypt PDF file at initialization. If the password is None, the file will not be decrypted. Defaults to ``None`` rKctt|jdtjd}|dS|}t|tst|}|S)zCReturns the existing ViewerPreferences as an overloaded dictionary./RootN) rr8trailerr\CDVIEWER_PREFERENCESrcr]rD)rYos rPviewer_preferenceszPdfReader.viewer_preferencessq !4<#8 9 9 = =  !4   94 LLNN!.// %!!$$ArRFNstreamstrictpasswordc||_d|_i|_d|_d|_t |drd|jvrtdtt|ttfrIt|d5}t|}dddn #1swxYwY||||_d|_d|_|jrd|_|jt*j}|r|djnd}t3t4|jt*j}t9j|||_||nd}|j|t<jkr|tAd d|_dS|tCd dS) NrmodebzQPdfReader stream/file object is not in binary mode. It may not be read correctly.rbFTrRzWrong passwordNot encrypted file)"rflattened_pagesresolved_objects xref_index _page_id2numhasattrrrrr]rropenrreadr_override_encryption _encryption is_encryptedrr\TKIDrcoriginal_bytesrr8ENCRYPTrverifyr NOT_DECRYPTEDr2r0) rYrrrfhid_entry id1_entry encrypt_entrypwds rPrWzPdfReader.__init__'s   ;?LN   66 " " s&+'='= 0    fsDk * * ,fd## ,r ++ , , , , , , , , , , , , , , , & $)!15   5(,D %|''..HCKT ..00??QTI  $,rz":"E"E"G"GM *}iHHD '2((C '',, 0JJJ())9:::(-D % % %  !344 4" !s>"B,,B03B0includeexcludec|jd|j}d|i} fd|D} fd|D}|S)z Integration into Jupyter Notebooks. This method returns a dictionary that maps a mime-type to it's representation. See https://ipython.readthedocs.io/en/stable/config/integrating.html rzapplication/pdfNc$i|] \}}|v || Srr).0kvrs rP z/PdfReader._repr_mimebundle_..ps$BBBTQQ'\\Aq\\\rRc$i|] \}}|v || Srr)rrrrs rPrz/PdfReader._repr_mimebundle_..ts)FFFTQQg5E5EAq5E5E5ErR)rseekritems)rYrrpdf_datadatas `` rP_repr_mimebundle_zPdfReader._repr_mimebundle_[s ;##%% x   BBBBTZZ\\BBBD  FFFFTZZ\\FFFD rRc|j}|jdd|jddd}|j|d|S)z The first 8 bytes of the file. This is typically something like ``'%PDF-1.6'`` and can be used to detect if the file is actually a PDF file and which version it is. rrMzutf-8backslashreplace)rtellrrdecode)rYlocpdf_file_versions rP pdf_headerzPdfReader.pdf_headerxstk   A;++A..55g?QRR a   rRctj|jvrdS|jtj}t}t |t drt d|||S)a Retrieve the PDF file's document information dictionary, if it exists. Note that some PDF files use metadata streams instead of docinfo dictionaries, and these metadata streams will not be accessed by this function. NzEtrailer not found or does not point to document information directory)rINFOrrTr]typer0update)rYobjr^s rPmetadatazPdfReader.metadatasw 7$, & &4l27#$&& c4:: & & W   c rRcp d|_|jtjjd|_S#d|_wxYw)z(XMP (Extensible Metadata Platform) data.TF)rrrROOT xmp_metadatarXs rPrzPdfReader.xmp_metadatas= .(,D %<(5(-D % %D % - - - -s", 5c|jr#|jtjddS|j|t |jS)a Calculate the number of pages in this PDF file. Returns: The number of pages of the parsed PDF file Raises: PdfReadError: if file is encrypted and restrictions prevent this action. /Pages/Count)rrrrr_flattenlenrXs rP_get_num_pageszPdfReader._get_num_pagessO   -<(28< <#+ t+,, ,rR page_numberct|j||j Jd|j|S)z Retrieve a page by number from this PDF file. Args: page_number: The page number to retrieve (pages begin at zero) Returns: A :class:`PageObject` instance. N hint for mypy)rr)rYrs rP _get_pagezPdfReader._get_pages?   ' MMOOO#/////#K00rRc*|S)zu A read-only dictionary which maps names to :class:`Destinations` )_get_named_destinationsrXs rPnamed_destinationszPdfReader.named_destinationss ++---rRtreer^fileobjcrtj}|tj|mi}t t |jtj}tj |vr1t tt|tj }ndS||S| ||||D] }||vr|||||n!d|vrLt t|d}|D].}|} || |||/|S)aX Extract field data if this PDF contains interactive form fields. The *tree* and *retval* parameters are for recursive use. Args: tree: retval: fileobj: A file object (usually a text file) to write a report to on all interactive form fields found. Returns: A dictionary where each key is a field name, and each value is a :class:`Field` object. By default, the mapping name is used for keys. ``None`` if form data could not be located. N/Fields)FAattributes_dictrr%rr8rrrr ACRO_FORMrrC _check_kids _build_fieldr3rc) rYrr^rfield_attributescatalogattrfieldsffields rP get_fieldszPdfReader.get_fieldssC.-// = M O OPPP >F+T\"'-BCCG|w&&HZ0'",2GHHt <M vw///$  Dt||!!$9IJJJ   +tI77F L L !!%:JKKKK rRparentcd|vrtt|dSd|vrM|tt|ddztt|dzStt|dS)N/TM/Parent./T)rr_get_qualified_field_namer8)rYrs rPrz#PdfReader._get_qualified_field_names F??VE]++ + & ..)6)+<==sF4L)) * VD\** *rRrrc|||| tt|d}nz#t$rm d|vr2|tt |ddz}nd}|tt|dz }n#t$rYYdSwxYwYnwxYw|r,|||||dt|||<||j }| tj ddkr5|ttj||td<| tj dd krd |vrt!t#|d d ||td<d ||dvr=||tdtd dSdS| tj dd kry| tjd tjjzd krDg}t!|||td<| tjiD]}| }t#|d d D]} | |vr|| t!|||td<| tjd tjjzd krDd ||dvr6||d||dd =dSdSdSdSdS)Nrrrr /Chz /_States_/Btnz/AP/Nz/Offr)rrrKeyErrorrr8 _write_fieldwriter:indirect_referencercr\rFTr>Optr3listkeysappendFfFfBitsRadioKids NoToggleToOffindex) rYrr^rrrZrstatesrss rPrzPdfReader._build_fields 000 sE%L))CC    %%66 !153CDDCCtCt---         gu.> ? ? ? MM$   Ells Sk,7799 7725"   & &36z"&7I7I3JF3K ;// 0 7725"   ' 'ESLL3>SZ%**,,--44F3K ;// 0VC[555s J{334;;Jvv3F3FF3K ;// 0WWRWb)) K KLLNNahtn113344))A a(((7B67J7Js J{3344q!!BI$;;q@@fSk+6663K ,VC[-E-K-KF-S-STTT* ).V.VA@66s45 B,ABB, B&!B,%B&&B,+B,ctj|vr>|tjD]-}||||,dSdSrV)PAKIDSrrc)rYrr^rkids rPrzPdfReader._check_kidsKs\ 7d??BG} C C 0 0&'BBBB ? C CrRctj}|tjz}|D]}|tjtjfvr||} |tjkr9ddddd}|||vr'||d|||dn|tjkr^ ||tj}n(#t$r||tj }YnwxYw||d|dn!||d||d#t$rYwxYwdS)NButtonTextChoice Signature)r/Txrz/Sig: r) r attributesr%rAArrParentTMrT) rYrrrfield_attributes_tupler attr_nametypesnames rPrzPdfReader._write_fieldSs!# "%B%M%O%O O +  D(.I 25==!)%' + E T{e++ &L&LeE$K6H&L&L&LMMMRY&&1$T{251#111$T{2401MMY":":$":":":;;;;MMY"A"A%+"A"A"ABBB    7  s8AD--CD-"C+(D-*C++AD-- D;:D;full_qualified_namec dtdtttfdtfd}|}|iSi}|D]t\}}|ddkrV|r|d||<9|d||t t|d |<u|S) a Retrieve form fields from the document with textual data. Args: full_qualified_name: to get full name Returns: A dictionary. The key is the name of the form field, the value is the content of the field. If the document contains multiple form fields with the same name, the second and following will get the suffix .2, .3, ... rrrKcr|vrSdzttfd|DdzzS)NrcDg|]}|dzdS)rr) startswith)rkkrs rP zGPdfReader.get_form_text_fields..indexed_key..s.LLLRR]]1s75K5KLqLLLrR)rsum)rrs` rP indexed_keyz3PdfReader.get_form_text_fields..indexed_keysW#LLLL6LLLMMPQQRRSrRN/FTrz/Vr)rr rrrr\r)rYr r) formfieldsffrvalues rPget_form_text_fieldszPdfReader.get_form_text_fieldsxs 3 S#X 3    __&&  I &,,.. R RLE5yy5((&R % $BuIIBG))D//B{{4U4[#9#92>>? rRcdtdtdtffd  ttjn"#t $r}td|d}~wwxYwdtdg}d d d kr4d vrd g}nfd j D}ndd}|D]md d d kr=dvr9d vr|d gz }W|fdj Dz }nfd|DS)a Provides list of pages where the field is called. Args: field: Field Object, PdfObject or IndirectObject referencing a Field Returns: List of pages: - Empty list: The field has no widgets attached (either hidden field or ancestor field). - Single page list: Page where the widget is present (most common). - Multi-page list: Field with multiple kids widgets (example: radio buttons, field repeated on multiple pages). rrZrKc||vr||Sd|vr7tt|d|SdS)Nr)rr8rc)rrZ_get_inheriteds rPr1z9PdfReader.get_pages_showing_field.._get_inheriteds[czz3xc!!%~)3y>+D+D+F+FGGtrRzfield type is invalidNr*zfield is not validz/Subtyperz/Widgetz/PcNg|]!}j|ddv|"Sz/Annotsrrr\)rprs rPr&z5PdfReader.get_pages_showing_field..s?/155B3G3GGGGGGrR/KidsrrcNg|]!}j|ddv|"Sr3r4)rr5rs rPr&z5PdfReader.get_pages_showing_field..s?    ! 3quuY7K7KKKKKKrRcg|]>}t|tr|n$j|j?Sr)r]rpages_get_page_number_by_indirectr)rxrYs rPr&z5PdfReader.get_pages_showing_field..sZ   !Z(( WAA*T>>q?STTU   rR) r8rrrrrc Exception ValueErrorr\r9)rYrexcretkidsr1rs`` @@rPget_pages_showing_fieldz!PdfReader.get_pages_showing_fields,  0 s s       ?)5+C+N+N+P+PQQEE ? ? ?4553 > ? >% ' ' /122 2 99Z $ $ 1 1u}}T{--//0!Z 99Wb))D  LLNNEE*b))Y66T]]qyy$ 2 2 4 455    %)Z             s,A A-A((A-c|i}tt|jtj}t j|vr&tt|t j}nft j|vrXtt|t j}t j|vr%tt|t j}||Stj |vrRtt|tj D]*}| | |+nt j|vrtt|t j}d}|t|krtt|| }|dz }t!|ts[ || }n#t"$rYn\wxYw|dz }t!|trd|vr |d}n|||} | | ||<|t|kn|D]k\} } | } t!| tr d| vr| d } nN|| | } | | || <l|S)z Retrieve the named destinations present in the document. Args: tree: retval: Returns: A dictionary which maps names to :class:`Destinations`. Nrr/D)rr8rrrCADESTSrCNAMESr rr3rrcrrr] IndexError_build_destinationr) rYrr^rnamesrirZr-destk__v__vals rPrz!PdfReader._get_named_destinationss >F+T\"'-BCCGx7""J(9::W$$-wrx/@AA8u$$ E"(O<::EAc%jj..3a 3 3 5 566Q!#s++!!H//11EE!EQe%566!u}} %d  ..sE::#"&F3K#c%jj..&!JJLL ' 'Snn&&c#344!s{{!$i2244 ..sC88#"&F3K s?G G'&G'c*|S)z Read-only property for the outline present in the document. (i.e., a collection of 'outline items' which are also known as 'bookmarks') ) _get_outlinerXs rPoutlinezPdfReader.outline*s  """rRnoderQc|g}tt|jtj}t j|vr]tt|t j}t|tr|S|d|vrtt|d}| |_ ||S | |}|r| |d|vrHg}| tt|d||r| |d|vrntt|d}|S)N/FirstTz/Next)rr8rrrCOOUTLINESr]r?r _namedDests_build_outline_itemrrP)rYrRrQrlines outline_obj sub_outlines rPrPzPdfReader._get_outline4sY ?G+T\"'-BCCG{g%%-wr{/CDDeZ00#"N$U):): 0%/BBD#;;==D  <N 922488K ,{+++4)+ !!$'7h"H"H+VVV0NN;///d""($w-88D 9 rRctt|jtj}t j|vr td|t jSdS)u Read-only property for the list of threads. See §8.3.2 from PDF 1.7 spec. It's an array of dictionaries with "/F" and "/I" properties or None if there are no articles. r3N)rr8rrrrUTHREADS)rYrs rPthreadszPdfReader.threads]sF'bg)>?? :  wrz':;; ;4rRrc"|j#dt|jD|_|t|trdSt|t r|}n|j}|j Jd|j|d}|S)z Generate _page_id2num. Args: indirect_reference: Returns: The page number or None Nc.i|]\}}|jj|Sr)ridnum)rrJr;s rPrz:PdfReader._get_page_number_by_indirect..zs1!!!26!Q$*A!!!rRr)r enumerater9r]r?intrar\)rYrrar?s rPr:z&PdfReader._get_page_number_by_indirectms   $!!:CDJ:O:O!!!D   %4F )S)S %4 (# . . -&EE&,E ,,o,,,##E400 rRpagec6||jS)a Retrieve page number of a given PageObject. Args: page: The page to get page number. Should be an instance of :class:`PageObject` Returns: The page number or None if page is not found )r:r)rYrds rPget_page_numberzPdfReader.get_page_numbers001HIIIrR destinationc6||jS)z Retrieve page number of a given Destination object. Args: destination: The destination to get page number. Returns: The page number or None if page is not found )r:rd)rYrgs rPget_destination_page_numberz%PdfReader.get_destination_page_numbers001ABBBrRrdarraycZd\}}t|ttfs*t|trt |dks|0t}t ||t jS|dd\}}|dd} t ||t ||S#t$rltd|d|t|j r|j dj }|tn|}t ||t jcYSwxYw)NNNrr')fit_typefit_argszUnknown destination:  )r]r?rr3rr7r;fitr0rrrr9r)rYrdrjrdtyptmprs rPrHzPdfReader._build_destinations> c uz3/ 0 0 I5+.. I36u::??}<z0PdfReader._build_outline_item..s)<`<`PQ[^^<`<`<`<`<`<`rR/Frrz /%is_open%)rrrr0r8r>r&SDr]r3rHrrW dest_arrayrrr4r\rRrAttributeError)rYrRrKrd outline_itemaction action_types rPrXzPdfReader._build_outline_items$4!e\ X//EE   { X"#Vd#V#VWWWEEE  4<<*DJ77Fz62E2G+HIIKg%%134 __=D$ 011 "ddllDz dK ( ( @225$??LL c " " @ D#664+D1<    D D D#66udCC  D\ 225$??LL{ "#ET#E#EFFFOdOOO 225$??L  t||1<<`<`UYZ^U_<`<`<`1`1` Z--.t||26d Z--.46:(^ Z1125B1%%*66LL11 2!  .2.EL + +    D s3%AA&D== E E 0 I== J  J c6t|j|jS)zbRead-only property that emulates a list of :py:class:`PageObject` objects.)rrrrXs rPr9zPdfReader.pages sD/@@@rRc^fdttjDS)z A list of labels for the pages in this document. This property is read-only. The labels are in the order that the pages appear in the document. c0g|]}t|Sr)page_index2page_label)rrJrYs rPr&z)PdfReader.page_labels..s$OOO1%dA..OOOrR)rangerr9rXs`rP page_labelszPdfReader.page_labelss0POOOc$*oo8N8NOOOOrRctt|jtj}t j|vr%tt|t jSdS)a Get the page layout currently being used. .. list-table:: Valid ``layout`` values :widths: 50 200 * - /NoLayout - Layout explicitly not specified * - /SinglePage - Show one page at a time * - /OneColumn - Show one column at a time * - /TwoColumnLeft - Show pages in two columns, odd-numbered pages on the left * - /TwoColumnRight - Show pages in two columns, odd-numbered pages on the right * - /TwoPageLeft - Show two pages at a time, odd-numbered pages on the left * - /TwoPageRight - Show two pages at a time, odd-numbered pages on the right N)rr8rrrr PAGE_LAYOUTr>)rYrs rP page_layoutzPdfReader.page_layoutsF.'bg)>?? >W $ $ GBN$;<< <trRc` |jtjdS#t$rYdSwxYw)a2 Get the page mode currently being used. .. list-table:: Valid ``mode`` values :widths: 50 200 * - /UseNone - Do not show outline or thumbnails panels * - /UseOutlines - Show outline (aka bookmarks) panel * - /UseThumbs - Show page thumbnails panel * - /FullScreen - Fullscreen view * - /UseOC - Show Optional Content Group (OCG) panel * - /UseAttachments - Show attachments panel z /PageModeN)rrrrrXs rP page_modezPdfReader.page_mode7s=* <(5 5   44 s  --r9inheritcttjttjttjttjf}|i}|J|jtj }|d }g|_ tj |vr|tj }ntj |vrd}nd}|dkrl|D]}||vr ||||<|tj D]C}i} t|tr|| d<| } | r|j| |fi| DdS|dkrqt#|D]\} } | |vr| || <t'||} | ||j | dSdS)Nrz/Pager)r>PG RESOURCESMEDIABOXCROPBOXROTATErrrrcrr TYPErr]r=rrrrrr)rYr9rrinheritable_page_attributesrtrrdaddtrattr_inr-page_objs rPrzPdfReader._flattenQs r| $ $ r{ # # rz " " ry ! ! ' # ?G =l27+6688GH%0022E#%D 7e  bgAA WE ! !AAA ==3 0 05==$)$KGDMbg 8 8dN33615D-.oo''8!DM#w77$777 8 8'\\"&w}}"7"7 + +%''%*E'N!$(:;;H OOE " " "  ' ' 1 1 1 1 1\rRc |j|j\}}t|d|}t t |ddksJ||dksJt t|}t|dD]}t|| ddtj |}t|| ddtj |}t|| dd||jkr|jr||krtd| t!|d|zdt|| dd t#||} ni#t$$r\} t'd |d |jd |jd | t*|jrtd | t-} Yd} ~ nd} ~ wwxYw| cS|jrtdt-S)Nr/Typez/ObjStmrrzObject is in wrong index.rTzInvalid stream (index z) within object rorzCan't read object stream: z%This is a fatal error in strict mode.) xref_objStmrar=rcrrrrget_datarrrr@read_from_streamrr0rcrEr1r generationrr?) rYrstmnumidxobj_stm stream_datarJobjnumoffsetrr>s rP_get_object_from_streamz!PdfReader._get_object_from_streams &'9'?@ '5fa'F'F'Q'Q'S'SC)**i7777WT]""""b!1!1!3!34455 wt}%%$ $ A  , , ,   R # # #!2;??F  , , ,   R # # #!2;??F  , , ,   R # # #+111{ @saxx"#>???   S!2V!;<>>"#98DDD%f--D 2A2h(""""""rRct|} t|}t|}|dddkrtdnr#t$re|dstdt|dd}t dtYnwxYw|S)z Find startxref entry - the location of the xref table. Args: stream: Returns: The bytes offset N s startxrefzstartxref not foundz startxref on same line as offset)rrcr0r=r$striprr)rYrrrs rPrzPdfReader._find_startxref_poss"&)) :D I&f--DBQBx<''"#8999( I I I??<00 :"#8999DHNN,,--I =x H H H H H  IsA A,B<;B<c|d}|dkrtdt||ddd} t t t ||}|r)|dkr#||_|jrtdtd }t||ddt t t ||}t||ddd}||kr|d }|dd vr5|d d|d }|dd v5|dd vr|dd |dd d\}} |dd} t |t | } } n0#t$r"t|dr"t|} nT|}|dd|d} ||t#j|d| }|td|dtd} d} nOtd|dtt |d} |} YnwxYw| |jvri|j| <i|j| <||j| vrn\| |j| |< | dk|j| |<n#t$rYnwxYw | dk|jd|<n#t$rYnwxYw|dz }|dz }||kt||dd|d}|dkr|ddndS)Nrsrefzxref table read errorrrTrzFXref table not zero-indexed. ID numbers for objects will be corrected.F is 0123456789t rz\s+(\d+)\s+objzentry z( in Xref table invalid; object not foundrz' in Xref table invalid but object foundfstraileri)rr0rrrrcrErrrrsplitr<rrrrrrrgrouprrr)rYrref first_timerrJcntroffset_b generation_b entry_type_brrrr5r trailer_tags rP_read_standard_xref_tablez#PdfReader._read_standard_xref_tablesWkk!nn &==677 7F### B c sK5566C cQhh"%;"`  J  ' ' ' KKA   [6677D  ' ' ' KKA   C**{{21g,,KK+++!;;r??D1g,,8~--KKA&&&+-1#2#Y__T-B-B*Hl#'2;L),XL8I8IJFF +++v{33'#F$4$4$6$677"KKMM Aq)))$kk"oo A S";";";"B"B"D"DcJJAy&RSRRR$&+ !#&QSQQQ$&)__ !"1+4TY..,.DIj)79D(4$)J/// 17DIj)#.@LPT@T,Z8==$;G4;O,U3C88$qqS**T  ' ' ' KKA    ++a..Kj(( B""""Gc s8AGD*LLM M('M(,N N Nrrci|_i|_i|_t|_|@||d|d}|dvr|d}|dkr||}n|r+ ||dS#t$rd}YnwxYw| r | |}n_#t$rR}tj |jvr#td|jt Yd}~dSt#d|jd}~wwxYwtj tjtjtjtjf}|D]9}||vr3||jvr*|||jt/|<:d|vrq|} |t3t4|ddzd| ||| dd|vrt3t4|d}ndS|||}|>dSdS) Nrrrxz!Previous trailer can not be read ztrailer can not be read /XRefStm/Prev)rrrr8rrr _read_xref_rebuild_xref_tabler<isdigit_read_pdf15_xref_streamrrrargsrr0rrrSIZEraw_getr>rrrc_read_xref_other_error) rYrrrr; xrefstreame trailer_keysrZr5s rPrz(PdfReader._read_xref_tables_and_trailersLs02 :<79')) # KK 1 % % % AAG||KKNNDyy OOF33  K&,,V444E &&&$%MMM& K P!%!=!=f!E!EJJ PPPw$,..&HHH$*+Naf+N+NOOOP "w BGRUBGK 'PPCj((S -D-D8B8J8J38O8O Z__5++ AKKS*Z*@ A AA EqIII00888KK1%%%j(( $S*W*= > >IIE 77 JJ Q#####s0B'' B76B7C&& E00D=&D==Ec||t||ddttt t ft||}|D]\}}||j vr ||j |<d|vr| }|tt|ddzd | |n/#t$r"td|ddtYnwxYw||dd|vr |d}|SdS)NrrrrzXRef object at z, can not be read, some object may be missingr)rrrrr rrrErrrrcrr<rr)rYr new_trailerrZr-r5rs rPrzPdfReader._read_xref}sq &&v...F### B4S>;vt+D+DEE %++-- * *JC$,&&$) S!  $ $ A KKS+j"9::Q> B B B ,,V4444   kk*&=kkk  KK1    k ! !#G,I 4s(C>>)D*)D*c|dkr-|jrtdtdtdS|dd|d}|d}|dkr |d |z z}|S||dtd D]2}|dr ||z }|cS3d |j vrQ|jsJtd t | |dS#t$rtd wxYwtd)Nrz6/Prev=0 in the trailer (try opening with strict=False)zA/Prev=0 in the trailer - assuming there is no previous xref tableirrxrefr rz"Invalid parent xref., rebuild xrefzcan not rebuild xrefz/Could not find xref table at specified location) rr0rrrrfindrrrrr<)rYrrrrxref_loclooks rPrz PdfReader._read_xref_other_errors >>{ "L S   4  Ckk"oo88G$$ r>> h &I  Iq!!!"II ! !D{{1~~%%'' !T!      ! dl " "4; " ? J J J ;((000t ; ; ;"#9::: ;LMMMs DD3c |dd|\}}ttt |}tt |ddksJ|||tt|  | dd| dg}ttttf| dtd ksJjr%td krtd d t d t"t t$t d fff fd }dt dt"t t$t d ffd t&ffd }||||S)Nrrrz/XRefz/Indexrz/Sizez/WrzToo many entry sizes: rJrK.c|dkr1|}t||S|dkrdSdS)Nrr)rrQ)rJrI entry_sizesrs rP get_entryz4PdfReader._read_pdf15_xref_stream..get_entrysT1~!!$$[^44%aQ888AvvqqrRrrcP|j|gvp|jvSrV)rr\r)rrrYs rP used_beforez6PdfReader._read_pdf15_xref_stream..used_befores+$)-- B777R3$BR;R RrR)rrrr5rErrrrrr\r rrrr0rcrrbool_read_xref_subsections) rYrrarr idx_pairsrrrrs ` @@rPrz!PdfReader._read_pdf15_xref_streams  B 33F;;z-VT)B)BCC CG,--8888 "":ujAAAb!4!4!6!67788 NN8a1H1H-IJJ 4S>:>>$+?+?@@ ;1$$$$ ; G3{++a//E EEFF F  sE#s(O';!<        SS SeCsCx4H.I Sd S S S S S S ##Iy+FFFrRc||dz d|d}|dkr|d}|dvrdS|d}|dkrYd}|dvr!|d}|dkrd S|dv!||d z }|d krd SdS) z Return an int which indicates an issue. 0 means there is no issue. Args: stream: startxref: Returns: 0 means no issue, other values represent specific issues. rrjs rrRs 0123456789 r'sobjr)rrlower)rrrs rPrzPdfReader._get_xref_issuess  IM1%%%{{1~~ 4<<;;q>>D z ! !1{{1~~ 7??D***{{1~~3;;1*** FKKNN "Dzz||v%%qqrRci|_|dd|d}tjd|D]|}t |d}t |d}||jvr i|j|<|d|j||<}|ddtjd|D]}||ddttttft||}t| D]\}}||j|<dS)Nrrs([\r\n \t][ \t]*(\d+)[ \t]+(\d+)[ \t]+objrr's$[\r\n \t][ \t]*trailer[\r\n \t]*(<<))rrrrfinditerrcrrrr rrErrr) rYrf_rrarrrZr-s rPrzPdfReader._rebuild_xref_tables]  Aq [[__I2NN 6 6A OOEQWWQZZJ**(* *%+,771::DIj !% ( ( AqErJJ * *A KK A & & &tCH~{64/H/HIIK";#4#4#6#677 * * U$) S!! *  * *rRrr.rc||D]\}}t|||zD]}|d}|dkr|d}|d} *|dkrF|d} |d} | |jvr i|j| <||| s| |j| |<v|dkr1|d} |d} d} ||| s | | f|j|<|jrt d|njdS)Nrrr'zUnknown xref type: )_pairsrrrrr0)rYrrrrrJr xref_typenext_free_objectnext_generation byte_offsetr objstr_num obstr_idxs rPrz PdfReader._read_xref_subsectionssd  ;;y11 J JKE4UEDL11 J J%IaLL >>'0y||$&/illOO!^^"+)A,,K!*1J!2202 *-&;sJ77A5@ *-c2!^^!*1J ) ! I!"J&;sJ77H1;Y0G(-[J&'HY'H'HIIIJ/ J J JrRc#rKd} ||||dzfV|dz }|dzt|krdS3)NrTrr')r)rYrjrJs rPrzPdfReader._pairs;sU  (E!a%L( ( ( ( FAA#e**$$  rRcb|jstd|j|S)a When using an encrypted / secured PDF file with the PDF Standard encryption handler, this function will allow the file to be decrypted. It checks the given password against the document's user password and owner password, and then stores the resulting decryption key if either password is correct. It does not matter which password was matched. Both passwords provide the correct decryption key that will allow the document to be used with this library. Args: password: The password to match. Returns: An indicator if the document was decrypted and weather it was the owner password or the user password. r)rr0r)rYrs rPdecryptzPdfReader.decryptCs5& 5344 4&&x000rRpermissions_codec tdddtjtjtjtjtjtjtjtj d}fd| DS)z>Take the permissions as an integer, return the allowed access.decode_permissionsuser_access_permissionsz5.0.0)old_namenew_name removed_in)printmodifycopy annotationsforms accessabilityassembleprint_high_qualityc*i|]\}}||zdkSrr)rrZflagr's rPrz0PdfReader.decode_permissions..ns:   T !D(A-   rR) rr'PRINTMODIFYEXTRACT ADD_OR_MODIFYFILL_FORM_FIELDSEXTRACT_TEXT_AND_GRAPHICS ASSEMBLE_DOCPRINT_TO_REPRESENTATIONr)rYr'permissions_mappings ` rPr)zPdfReader.decode_permissions[s").    +0+2)10>*;2L-:"7"O       06688    rRcF|jdSt|jjS)zWGet the user access permissions for encrypted documents. Returns None if not encrypted.N)rr'PrXs rPr*z!PdfReader.user_access_permissionsss&   #4$T%5%7888rRc(tj|jvS)z Read-only boolean property showing whether this PDF file is encrypted. Note that this property, if true, will remain true even after the :meth:`decrypt()` method is called. )rrrrXs rPrzPdfReader.is_encryptedzszT\))rRc$d}i}tt|jtj}d|vs|dsdStt |d}d|vrtt |d}t|}|D]}|}t|}t|tr_ttt| }|r+tjt!|j} | ||<|S)N /AcroFormz/XFA)rr8rrrrCr3iternextr]r=rr9rczlib decompressr_data) rYrr^rrrJrtagress rPxfaz PdfReader.xfas%)!#'bg)>?? g % %W[-A %4J 455 T>>+tF|44FV A ) )GGa00) *=!> OOE)!_R __==&(s  rRrcRtt|jtj}d|vst |dtsdStt|t d}d|vrdSt}t||t d<|t d|t d<|dtd|j Ddz|t}| |j ||t d<tt|dD]O}|}d |vrtd |j d t |j |t d <P|S) z Add a top level form that groups all form fields below it. Args: name: text string of the "/T" Attribute of the created object Returns: The created object. ``None`` means no object was created. rFNrrr6rc$g|] \}}|dk |Sr7r)rgrJs rPr&z.PdfReader.add_form_topname..s!AAAv1!q&&&&&rRrrzTop Level Form Field z have a non-expected parent)rr8rrrr]r>rBrmaxrr3rrrcrr)rYrracroforminterimarrrrs rPadd_form_topnamezPdfReader.add_form_topnames'bg)>?? g % %Z K "2. . %4('*[2I2I*JKK H $ $4"$$$4T$:$: 4  !'/ 90E0E'F 7##$ "" AA!6AAA B BQ F    mm 7-...*-I&&'k77#344 D DA,,..CC_C,B___*1)CC 9%% & &rRc tt|jtj}d|vst |dtsdStt|t d}d|vrdStttt|t dd}t||t d<|S)z Rename top level form field that all form fields below it. Args: name: text string of the "/T" field of the created object Returns: The modified object. ``None`` means no object was modified. rFNrrr) rr8rrrr]r>r3rcrB)rYrrrSrTs rPrename_form_topnamezPdfReader.rename_form_topnames'bg)>?? g % %Z K "2. . %4('*[2I2I*JKK H $ $4  hz)'<'<= > >q A L L N N  %5T$:$: 4  !rRc^tfdDS)Nc$i|] }|j|f Sr)_get_attachment_list)rrrYs rPrz)PdfReader.attachments..s3   t0$7   rR)LazyDict_list_attachmentsrXs`rP attachmentszPdfReader.attachmentssG     2244      rRc  tt|jd} tttttt|ddd}n#t$rgcYSwxYwd|D}|S)zv Retrieves the list of filenames of file attachments. Returns: list of filenames r/Names/EmbeddedFilesc<g|]}t|t|Sr)r]r)rrs rPr&z/PdfReader._list_attachments..s'HHH1Z35G5GHQHHHrR)rr8rr3r)rYr filenamesattachments_namess rPr]zPdfReader._list_attachmentss'g)>?? $)78+<==>NOII   III HH HHH  A A00 A?>A?cl|||}t|tr|S|gSrV)_get_attachmentsr]r)rYrouts rPr[zPdfReader._get_attachment_lists8##D))$/ c4  Ju rRfilenamec tt|jd} tttttt|ddd}n#t$ricYSwxYwi}t t |D]}||}t|tr|||kr(|}||dz }|dd } ||vrCt||ts ||g||<|| | | ||<|S)a Retrieves all or selected file attachments of the PDF as a dictionary of file names and the file data as a bytestring. Args: filename: If filename is None, then a dictionary of all attachments will be returned, where the key is the filename and the value is the content. Otherwise, a dictionary with just a single key - the filename - and its content will be returned. Returns: dictionary of filename -> Union[bytestring or List[ByteString]] if the filename exists multiple times a List of the different version will be provided rr`raNrz/EFrx) rr8rr3rrrr]rrcrrr) rYrirrcr^rJrrf_dictf_datas rPrgzPdfReader._get_attachmentssv"'g)>?? $)78+<==>NOII   III <> s9~~&& / /A! A!S!! /'AMM"1q5)4466t,5577;&&%k$&7>>@-8->,? D)%,,V4444(.K%re)FNrlrt)FrV)frrrrrrrDrrrrrrrrWr r rrrrTrrHrrcrrrrrCrr8rrrrr.r:rAr=r rArrFrQrPr3r^r?r:rfr7rir@rHrXr9rrrGrrrrrcrrrrrrrrrrrrrr5r9r6r staticmethodrrrrrrr&r)r'r*rrNrVrXr r^r]r[rgrrRrPrrs $ H->$?   X ,0 2525k4'(2525c5() 25  25252525l/3.2tXc]*+tXc]*+ c3h :  C    X  (#67X&.h~6...X.-----*1S1Z1111 .DcN...X.&*+/!% 00z"0c3h(0# 0 $sCx. ! 0000d +0@ +S + + + +4UZ!1124US#X4U 4U  4U  4U4U4U4UlC*&667CADCORC CCCC#C##s#t####J$$$cSVh$$$$LD 5)^;<D j D D D D P)- $DDJ$%D D c3h DDDDL####X#QU''-.'@H ' ''''R +.   X "'c:~(M"N #6 JJ J8C= J J J J C{ CxPS} C C C CII lND*FVVW  I IIII@I(8IXk=RIIIIVAtJ'AAAXAPT#YPPPXPXc]X68L1X6<@,07; 3232T+Z7832$sCx.)32%^4 32  32323232j4"04 sIs" #4444l ; ;# ;(9:M ; ; ; ;@"'^(;"<@ ) @@@@D++c3h++++8>>&)> ) >>>>  &) 080C  )     $ :$ $$ $ $ $ L$ $t$$$$&.z.d....$*2k ktkkkkZ/K /K-5c]/KKN/K /K/K/K/Kb 2'N 'N-0'N #'N'N'N'NR& & }13FF G&&&&P\<*******& J9 JSE5eCHo)=#>>? JsE#uS#X*>$?@$FG J  J J J JDDI(5c?*C1c5j 11l11110 3 4T ?    092G)H999X9 *d***X*Xd38n-X.(S(X6F-G((((T9I0J6 WS$u+%56   X !49!!!!,e)---  - c5U +,, -------rRrcbeZdZdededdfdZdedefdZdeefdZde fd Z defd Z dS) r\rkwrKNc(t|i||_dSrV)dict _raw_dict)rYrros rPrWzLazyDict.__init__7st*r**rRrZcR|j|\}}||SrV)rr __getitem__)rYrZfuncargs rPrtzLazyDict.__getitem__:s)N..s33 ctCyyrRc*t|jSrV)rGrrrXs rP__iter__zLazyDict.__iter__>sDN###rRc*t|jSrV)rrrrXs rP__len__zLazyDict.__len__As4>"""rRcLdt|dS)NzLazyDict(keys=r)rrrXs rP__str__zLazyDict.__str__Ds#4TYY[[ 1 14444rR) rrrrrWrrtr rxrcrzr|rrRrPr\r\6s+c++++++ss$(3-$$$$#####5555555rRr\)drrrNrIriorrpathlibrtypingrrr r r r r rrrrrrr_pagerr _page_labelsrr_utilsrrrrrrrrr r!r" constantsr#rDr$rr%r&r'r(rUr)rar*rr+rr,r r-rerrorsr.r/r0r1r2genericr3r4r5r6r7r8r9r:r;r<r=r>r?r@rArBrCrDrErrFrGxmprHrrcrQrTrr\rrRrPrs< ,,,,,,,,                          21111111++++++++>>>>>>                          /........... "!!!!!::::::666666++++++,,,,,,((((((*-,,,,,,,%e%3%5eCHo1E+F%%%%L%L%L%L%L%*L%L%L%^llllllll^955555wsCx 55555rR