KhvddlZddlmZmZmZddlmZddlmZddl m Z m Z m Z m Z mZmZmZmZmZddlmZddlmZdd lmZdd lmZmZmZmZmZdd lm Z dd l!m"Z"m#Z#m$Z$dd l!m%Z&ddl'm(Z(m)Z)m*Z*m+Z+m,Z,m-Z-m.Z.m/Z/m0Z0m1Z1m2Z2m3Z3m4Z4ddl5m6Z6m7Z7ddlm8Z8m9Z9m:Z:m;Z;mGddZ?Gdde?Z@dS)N)BytesIOFileIOIOBase)Path) TracebackType) AnyDictIterableListOptionalTupleTypeUnioncast) Encryption) PageObject) PdfReader) StrByteTypedeprecate_with_replacementdeprecation_bookmarkdeprecation_with_replacementstr_) PdfWriter)GoToActionArguments TypArgumentsTypFitArguments)PagesAttributes) PAGE_FIT ArrayObject DestinationDictionaryObjectFit FloatObjectIndirectObject NameObject NullObject NumberObject OutlineItemTextStringObject TreeObject) PageRange PageRangeSpec)FitType LayoutType OutlineType PagemodeType ZoomArgTypez=close() was called and thus the writer cannot be used anymorec*eZdZdZdedededdfdZdS) _MergedPagez@Collect necessary information on each page that is being merged.pagedatasrcidreturnNc>||_||_d|_||_dSN)r6r5 out_pagedatar7)selfr5r6r7s U/var/www/html/services/rag-service/venv/lib/python3.11/site-packages/pypdf/_merger.py__init__z_MergedPage.__init__Qs#   )__name__ __module__ __qualname____doc__rrintr>r?r=r4r4NsJJJ)r?r4cteZdZdZed dAdedeeefdd fd Z dBd Z d e e e d e e de edd fdZedd dCde eded eeefde ede edede edd fdZdeeeefdeee effdZedd dDdeeeefde eded eeeefeeeefeefdedd f dZdeeefdd fdZdEdZdeeefdd fdZ deeefdd fdZ!de"dd fd Z#de"dd fd!Z$d"e%dd fd#Z&d"e%dd fd$Z'd%ed&eeeeeffdeeeefeeeefeefdeeeeffd'Z(d%ede)deeeefeeeefeefde)fd(Z*dEd)Z+ed dFde e,e-d*e e.dd fd+Z/ed,dee-e0fd-e1dd fd.Z2dee1dd fd/Z3ed dGdee1de e,e-dd fd0Z4ed, dGdeeefd1e e)de eefd2Z5ed, dGdeeefd1e e)de eefd3Z6d d d dde7d fd4ede ed*ed e.e8fd5e ee9e9e9fd6ed7ed8e:d9e ede8fd:Z; dHd4ed9ed*ed e.e8fd5e ee9e9e9fd6ed7ed8e dHd4ed9ed*ed e.e8fd5e ee9e9e9fd6ed7ed8eZ?d4ed9edd fd?Z@ dFd4ede ed9e edd fd@ZAd S)I PdfMergerzD Use :class:`PdfWriter` instead. .. deprecated:: 5.0.0 outline) bookmarksFstrictfileobjr8Nctdddg|_g|_t|_g|_g|_d|_||_||_ dS)NrGr5.0.0r) rinputspagesroutputrH named_destsid_countrLrK)r<rKrLs r=r>zPdfMerger.__init___sW #; WEEE35 " +4;; $& &(   r?c(tddd|S)NrGrrN)r)r<s r= __enter__zPdfMerger.__enter__ms"; WEEE r?exc_typeexc tracebackcp|jr||j|dS)z*Write to the fileobj and close the merger.N)rLwriteclose)r<rVrWrXs r=__exit__zPdfMerger.__exit__rs3 < % JJt| $ $ $ r? outline_itemimport_outline)bookmarkimport_bookmarksT page_numberrPpositionc |8|'|}d}d}tj|d|dtntd|td|td||\} } t | |j } |j| | f| | | _ |d t| j f}nwt|tr(|t| j }n:t|trn$t|t st#d g} g} |r| j} || | |} |rTt)t+|t-|jt1j}|xj|| gz c_n|xj| z c_| j}|| ||}|xj|z c_t;|D]L}| j |}|j}|xjd z c_t=|| |}| |M|| | | | |j ||<dS) a  Merge the pages from the given file into the output file at the specified page number. Args: page_number: The *page number* to insert this file. File will be inserted after the given number. fileobj: A File Object or an object that supports the standard read and seek methods similar to a File Object. Could also be a string representing a path to a PDF file. None as an argument is deprecated. outline_item: Optionally, you may specify an outline item (previously referred to as a 'bookmark') to be applied at the beginning of the included file by supplying the text of the outline item. pages: can be a :class:`PageRange` or a ``(start, stop[, step])`` tuple to merge only the specified range of pages from the source document into the output document. Can also be a list of pages to merge. import_outline: You may prevent the source document's outline (collection of outline items, previously referred to as 'bookmarks') from being imported by specifying this as ``False``. NrbrazF is deprecated as an argument and will be removed in pypdf=4.0.0. Use  insteadzCThe argument position of merge is deprecated. Use page_number only.page_number may not be Nonezfileobj may not be None)rKrz0"pages" must be a tuple of (start, stop[, step])r)!warningswarnDeprecationWarning ValueError_create_streamrrKrOappend _encryptionlenrP isinstancer,indiceslisttuple TypeErrorrH _trim_outliner)r*r(rSr#fitnamed_destinations _trim_destsrRranger4_associate_dests_to_pages!_associate_outline_items_to_pages)r<rarLr]rPr^rbold_termnew_termstreamencryption_objreadersrcpagesrHoutline_item_typdests trimmed_destsipager7mps r=mergezPdfMerger.merge}sB  "& %( #JJ7?JJJ' !,  :;; ; ?677 7!%!4!4W!=!=6$+666 FF+,,,  %!/F  =FL))*EE y ) ) PMM#fl"3"344EE t $ $ P E5)) PNOO O  AnG((%@@G  $* ..T]++     LL-w7 7LLL LLG #LL)((>>  M)  A<?DB MMQ MMT62..B OOB     &&x000 ..x888/7 ;{*+++r?cfd}t|ttfrt|d}nt|tr|jr|j}|j}|jdt|j }|j|nht|drIt|dr9|d| }t|}ntd||fS)NrbrseekreadzPdfMerger.merge requires an object that PdfReader can parse. Typically, that is a Path or a string representing a Path, a file object, or an object implementing .seek and .read. Passing a PdfReader directly works as well.) rnstrrrrrlr|tellrrrhasattrNotImplementedError)r<rLr}r| orig_tell file_contents r=rjzPdfMerger._create_streams& gT{ + + GT**FF  + + " 5!(!4++--I N   " " "W^002233F N   * * * * Wf % % ''6*B*B  LLOOO"<<>>L\**FF%>  ~%%r?c\|t|j||||dS)a Identical to the :meth:`merge()` method, but assumes you want to concatenate all pages onto the end of the file instead of specifying a position. Args: fileobj: 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. outline_item: Optionally, you may specify an outline item (previously referred to as a 'bookmark') to be applied at the beginning of the included file by supplying the text of the outline item. pages: can be a :class:`PageRange` or a ``(start, stop[, step])`` tuple to merge only the specified range of pages from the source document into the output document. Can also be a list of pages to append. import_outline: You may prevent the source document's outline (collection of outline items, previously referred to as 'bookmarks') from being imported by specifying this as ``False``. N)rrmrP)r<rLr]rPr^s r=rkzPdfMerger.appends,> 3tz??G\5.QQQQQr?c\|jtt|jD]}|j|jt tttf|jj }|j |tjd |_|||j|\}}|r|dSdS)z Write all data that has been merged to the given output file. Args: fileobj: Output file. Can be a filename or any kind of file-like object. N)rQ RuntimeErrorERR_CLOSED_WRITERrPadd_pager5rr rr_pages get_object get_referencePAKIDSr; _write_dests_write_outlinerZr[)r<rLr pages_objmy_file ret_fileobjs r=rZzPdfMerger.write6s ; 011 1 J  D K  / / /T#s(^T[-?-J-J-L-LMMI $ 9 9"'"2&1133!!D     ${0099           r?crg|_|jD]\}}|g|_d|_dS)zHShut all file descriptors (input and output) and clear all memory usage.N)rPrOr[rQ)r<fo_readers r=r[zPdfMerger.closeUs? ;  KB HHJJJJ  r?infoscp|jtt|j|dS)z Add custom metadata to the output. Args: infos: a Python dictionary where each key is a field and each value is your new metadata. An example is ``{'/Title': 'My title'}`` N)rQrr add_metadatar<rs r=rzPdfMerger.add_metadata^s7 ; 011 1   '''''r?cPtdd||dS)zS Use :meth:`add_metadata` instead. .. deprecated:: 1.28.0 addMetadatarN)rrrs r=rzPdfMerger.addMetadataks. %]NCCC %     r?layoutcPtdd||dS)zV Use :meth:`set_page_layout` instead. .. deprecated:: 1.28.0 setPageLayoutset_page_layoutN)rrr<rs r=rzPdfMerger.setPageLayoutts/ %_6GHHH V$$$$$r?cp|jtt|j|dS)a Set the page layout. Args: layout: The page layout to be used .. list-table:: Valid ``layout`` arguments :widths: 50 200 * - /NoLayout - Layout explicitly not specified * - /SinglePage - Show one page at a time * - /OneColumn - Show one column at a time * - /TwoColumnLeft - Show pages in two columns, odd-numbered pages on the left * - /TwoColumnRight - Show pages in two columns, odd-numbered pages on the right * - /TwoPageLeft - Show two pages at a time, odd-numbered pages on the left * - /TwoPageRight - Show two pages at a time, odd-numbered pages on the right N)rQrr_set_page_layoutrs r=rzPdfMerger.set_page_layout}s72 ; 011 1 $$V,,,,,r?modecRtddd||dS)zT Use :meth:`set_page_mode` instead. .. deprecated:: 1.28.0 setPageMode set_page_mode3.0.0N)rrr<rs r=rzPdfMerger.setPageModes0 %]OWMMM 4     r?cp|jtt|j|dS)aW Set the page mode. Args: mode: The page mode to use. .. list-table:: Valid ``mode`` arguments :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 N)rQrrrrs r=rzPdfMerger.set_page_modes7. ; 011 1 !!$'''''r?pdfrcg}t|tr|ntt|}|D]\}}|D]}|j||dkri|d|t d<t|t|dksJ||n|S)z Remove named destinations that are not a part of the specified page set. Args: pdf: dests: pages: /Page/Title) rnrprwitemsrPrr&rrk) r<rrrP new_destslstkeyobjjs r=rvzPdfMerger._trim_destss !%..GeeD4G4G   HC  9Q<**,,G 0G0G0I0III/27|/F/F/H/HC 7++,99S](;(;;;;;$$S)))E J r?cvg}d}t|tr|ntt|}t|D]\}}t|trO||||} | r5|s|||dz || id}|D]} |d |j| |dkrC|d|td<||d}n|S)z Remove outline item entries that are not a part of the specified page set. Args: pdf: outline: pages: Returns: An outline type TrFr) rnrprw enumeratersrkrPrr&) r<rrHrP new_outlineprev_header_addedrrr]subrs r=rszPdfMerger._trim_outlines_"  !%..GeeD4G4G(11  OA|,-- ((lC@@,,;#**71q5>:::&&s+++$)!  A#G,4 y|..00L4I4T4T4V4VVV:K :g#6#676% 88DDD E Er?parentc|jtt||j}| Jdd}|D]}t |t r|||.d}d|vrCt|jD].\}}|j |dkr| ||n/|!|d=|d=|j ||}dS)N hint for mypyr/Type) rQrrrHrnrprrrPr7_write_outline_item_on_pageadd_outline_item_dict)r<rHr last_addedr]page_nors r=rzPdfMerger._write_outlines ; 011 1 ?lG""O""" # U UL,-- ##L*===G,&&%.tz%:%:MGTw,w"77788tLLL8" )<+@![>>|VTT  U Ur?)r_rctt|d}t|jt |g}t jtjft j tjft j tj ft j tj ft j tj tjdft jtj tjtjtjfi}||dD]o}||vrDt%||t&s)|t+||n"|t+d||=pt-t t.jt dt t.jt5|i|t d<dS)Nrz/ZoomrErz/GoToz/A)rrr(r7r&rFIT_HrTOPFIT_BHFIT_VLEFTFIT_BVXYZFIT_RBOTTOMRIGHTgetrnr'rkr$r"rSDr )r<r]roi_typeargs fit2arg_keysarg_keys r=rz%PdfMerger._write_outline_item_on_page*ssL122TW%%z'':':;  !L$4#6  "\%5$7  !L$5#7  "\%6$8  ,"3\5Ew!O  !!#" $ 4 $''44 & &G,&&zW%z00& K W(=>>???? KNN+++W%%)9.011:g3F3F.011;t3D3D * * Z%%&&&r?cD|jD]}d}|d}t|tr"|D]8}||jkr|j}9|t d|ddt||td<dS)NrzUnresolved named destination 'r')rRrnr(rr5r7rir&)r<rPrrnprs r=rxz#PdfMerger._associate_dests_to_pagesMs* G GJJG$B"l++  ) )==??dm&>&>&@&@@@!%J! LZ5ILLL/;:.F.FJz'** + + G Gr?c||j}| Jd|D]}t|tr|||.d}|d}t|trN|D]8}||jkr|j}9|t ||td<dS)Nrr) rHrnrpryr(rr5r7r&)r<rPrHr]routline_item_pageps r=ryz+PdfMerger._associate_outline_items_to_pages_s ?lG""O"""# M ML,-- 66ulKKKJ ,W 5 +\::  & &$//11QZ5J5J5L5LLL!"J%4@4L4L Z001! M Mr?rootc&||j}t|D]w\}}t|tr!|||}|r|g|zcS;||ks1t t ttf|d|kr|gcSxdS)Nr)rHrrnrpfind_outline_itemrr r)r<r]rroi_enumress r=rzPdfMerger.find_outline_itemxs <011 1&&           r?/Fitrc xtddd|||||||t||S)zZ .. deprecated:: 1.28.0 Use :meth:`add_outline_item` instead. addBookmarkrrfit_typefit_argsrrr# r<rrrrrrrtrs r=rzPdfMerger.addBookmarkQ %]4FPPP$$       t , , ,   r?c xtddd|||||||t||S)zY .. deprecated:: 2.9.0 Use :meth:`add_outline_item` instead. rrrrrrs r= add_bookmarkzPdfMerger.add_bookmarkrr?cPtddd|||S)z_ .. deprecated:: 1.28.0 Use :meth:`add_named_destination` instead. addNamedDestinationadd_named_destinationr)rr)r<rrs r=rzPdfMerger.addNamedDestination s6 % !#:G   ))%999r?cN||td|&d}d}tj|d|dt|}|tdt t |t |tjd }|j |dS) z Add a destination to the output. Args: title: Title to use page_number: Page number this destination points at. NzRThe argument pagenum of add_named_destination is deprecated. Use page_number only.rrarrdrei:)top) rirfrgrhr!r*r(r#fit_horizontallyrRrk)r<rrarrzr{dests r=rzPdfMerger.add_named_destinations  "w':(    H$H MGG4<GGG#    "K  :;; ; U # #  % %  S ) ) )   %%%%%r?)FrJ)r8rG)NNNNTN)NNT)r8N)NNr:)NNFFr)Br@rArBrCrboolrrrr>rUr r BaseExceptionrr\rDrrr-rr rrrjr,r rkrZr[r rrrr/rrr1rrrvr0rsrr r)r+rr!r4rrxryrrrr%floatr#rr.r2rrrrrEr?r=rGrGXs+  I...HJ   -243D-E     /.  4 ./ m $ M*      >DTUUU&*=A&*)-#"&r7r7c]r7tT; 9:r7sm r7  & r7  r73-r7 r7r7r7VUr7h!&T; 9:!& vx ++ ,!&!&!&!&F>DTUUU'+ #RR{It34RsmR )U38_eCcM.BDI M RR RRRVUR@ U4#45 $    > ($sCx. (T ( ( ( (!c3h!D!!!!%J%4%%%%-j-T----:! !!!!!(,(4((((6 Cc3h'(U38_eCcM&:DIEF  d38n  2' ''U38_eCcM&:DIEF '  ''''R E E E EI...48'+UU(;/0U$U  UUU/.U6>222  !+{":;  CN      32  DGtK/@GTGGGG$I...SWMM+&M19(;:O1PM MMM/.M0>222'+38n{# $s)  320>222'+ : :38n :{# : $s)  : : :32 :&*:>6:!%7 7 7 c]7 dJ67 7 eUE123 7  7 7 7 #7  7 7 7 7 z;?6:    dJ67  eUE123            <;?6:    dJ67  eUE123            4::s:t::::&*!% $&$&$&c]$&# $&  $&$&$&$&$&$&r?rGc,eZdZdededdffd ZxZS) PdfFileMergerrkwargsr8Nctdddd|vrt|dkrd|d<tj|i|dS)NrrGrrKrT)rrmsuperr>)r<rr __class__s r=r>zPdfFileMerger.__init__<sZ$_k7KKK 6 ! !c$ii!mm#F8 $)&)))))r?)r@rArBrr> __classcell__)rs@r=rr;sR*c*S*T**********r?r)Arfiorrrpathlibrtypesrtypingrr r r r r rrrrlr_pagerrr_utilsrrrrr_writerr constantsrrrrrgenericrr r!r"r#r$r%r&r'r(r)r*r+ pageranger,r-r.r/r0r1r2rr4rGrrEr?r=r!s8&&&&&&&&&&                      $#####IIIIIIIIII,,,,,,0///////NNNNNNNNNNNNNNS` &` &` &` &` &` &` &` &F*****I*****r?