j5jGdZddlmZddlZddlZddlmZmZmZm Z ddl m Z ddl m Z ddlmZddlmZdd lmZdd lmZmZdd lmZmZmZdd lmZmZdd lmZddl m!Z!ddl"m#Z#ddl$m%Z%m&Z&m'Z'ddl(m)Z)m*Z*ddl+m,Z,ddl-m.Z.ddl/m0Z0m1Z1m2Z2ejfZ4d dZ5d!dZ6 d"dZ7GddeZ8GddZ9eZ:GddZ;dZ.rects_intersect8sItad{Jqtad{Jqtad{Jqtad{Jr*FT)r.rr'rreturnbool)float) text_blocks page_width page_heightpwph margin_ratio interior_bboxr/has_textbboxs r(_page_has_textr<,s : k 2BLr \ R \ Rr MKH 4 /H  O  Or*c#K||D]l}|jd}|jd}t|ts2|jdk7}|j dk(}t |j ||nyw)z}Extract only limited content from text boxes. We do this to save memory and ensure that our objects are pickleable. rr%u�N)_objs isinstancer rendermodeget_textrr;) miner_pagetextbox_getterbox first_line first_charvisiblecorrupts r(simplify_textboxesrIIs}j)6YYq\ %%a( *&67 ''1,%%'83#((GW556sA9A;c@eZdZUdZded< ded< ded< ded<y)PageResolutionProfilez,Information about the resolutions of a page.r2 weighted_dpimax_dpiaverage_to_max_dpi_ratio area_ratioN)__name__ __module__ __qualname____doc____annotations__r-r*r(rKrKZs.6S N2##r*rKceZdZUdZded<ded<gZded< d d!dZ d!d Zed"d Z ed#d Z ed#d Z ed#d Z ed$dZ ed$dZed"dZed"dZed"dZej$dZed%dZed%dZed%dZed%dZed%dZed&dZd'd(dZed)dZed$dZed*dZd+dZdZy),PageInfoz9Information about type of contents on each page in a PDF. bool | None _has_text _has_vectorlist[ImageInfo]_imagesNcZ||_||_||_|j||||||y)zInitialize a PageInfo object.N)_pageno_infile_detailed_analysis_gather_pageinfo)selfpdfpagenoinfile check_pagesdetailed_analysis miner_states r(__init__zPageInfo.__init__{s6  "3  .? r*cZ|j|}t|jjDcgc]}t t |} }| d| dz } | d| dz } t |jj|_t |jj|_ t |jj|_ t |jj|_ t |jj|_||v} | ra|r_|j!|} | t#t%| t&|_ng|_d|j(D}t+|| | |_ng|_d|_|j/t0j2t d}t5|ts t |}||_| |zt dz |_| |zt dz |_t=t?|j@dd|_!|dd|ddf}| rd |_"d |_g|_#tI||| D]g}t5|tJrd |_"t5|tLrd |_3t5|tNr|jFjQ|_tSnd|_"d|_g|_#d|_*|jFrtWd d jYd |jFD}||_*t=t[|j\t_|j8z|_0t=t[|jbt_|j:z|_2yycc}w)Nr$rr%r#c34K|]}|jywNr;).0rDs r( z,PageInfo._gather_pageinfo..s:3chh:?gR@RotateF)rb container shorthandTc3NK|]}|js|jywrk) renderabledpi)rmimages r(rnz,PageInfo._gather_pageinfo..s!0#u7G7G 0s%%)3pagesrmediaboxas_listr strr)artbox_artboxbleedbox _bleedboxcropbox_cropbox _mediaboxtrimbox_trimboxget_page_analysislistrIr! _textboxesr<rXgetrUserUnitr? _userunit _width_inches_height_inchesintgetattrobj_rotaterYr[rrrrappendNotImplementedError_dpirtake_maxroundxr2 _width_pixelsy_height_pixels)rarbrcrdrerfrgpagedrzwidth_pt height_ptcheck_this_page page_analysisbboxesuserunituserunit_shorthandinforws r(r`zPageInfo._gather_pageinfos YYv&-7 8M8M8O-PQGCFOQQA;!,QK(1+-  !4!4!67 "4==#8#8#:;!$,,"6"6"89 "4==#8#8#:;!$,,"6"6"89  K/ 0'99&AM("&&}nE##%:$//:F+FHiHDN DO!DN88DMM73<8(G,x(H!%074=@'(2WT]B7488Xq9: &1h1= $D "DNDL043E 0dL1'+D$j1%)DNi0LL''--// 0 $D !DNDL <<S#&//0'+||0CDI!$U35559K9K3L+L%M!ND "%eCEEE$:M:M4N,N&O"PD  uRsN(c|jS)zReturn page number (0-based).)r]ras r(rczPageInfo.pageno||r*c,t|jS)z6Return True if page has text, False if not or unknown.)r1rXrs r(r:zPageInfo.has_textsDNN##r*ch|js tdtd|jDS)z>Return True if page has corrupt text, False if not or unknown.zDid not do detailed analysisc34K|]}|jywrk) is_corrupt)rmtboxs r(rnz,PageInfo.has_corrupt_text..s?t4???ro)r_ranyrrs r(has_corrupt_textzPageInfo.has_corrupt_texts.&&%&DE E?t???r*c,t|jS)zReturn True if page has vector graphics, False if not or unknown. Vector graphics are sometimes used to draw fonts, so it may not be obvious on visual inspection whether a page has text or not. )r1rYrs r( has_vectorzPageInfo.has_vectorsD$$%%r*c|jS)zReturn width of page in inches.)rrs r( width_incheszPageInfo.width_inches!!!r*c|jS)z Return height of page in inches.)rrs r( height_incheszPageInfo.height_inches"""r*c~ttt|j|jj zS)zReturn width of page in pixels.)rrr2rrwrrs r( width_pixelszPageInfo.width_pixelss,5t001DHHJJ>?@@r*c~ttt|j|jj zS)z Return height of page in pixels.)rrr2rrwrrs r( height_pixelszPageInfo.height_pixelss,5t112TXXZZ?@AAr*c|jS)zTReturn rotation of page in degrees. Will only be a multiple of 90. )rrs r(rotationzPageInfo.rotations ||r*c0|dvr||_ytd)N)rZiihiiLiz!rotation must be a cardinal angle)r ValueError)ravalues r(rzPageInfo.rotations ; ; DL@A Ar*c|jS)z*Return cropbox of page in PDF coordinates.)rrs r(rzPageInfo.cropbox}}r*c|jS)z+Return mediabox of page in PDF coordinates.)rrs r(rzzPageInfo.mediabox~~r*c|jS)z*Return trimbox of page in PDF coordinates.)rrs r(rzPageInfo.trimboxrr*c|jS)z)Return artbox of page in PDF coordinates.)r~rs r(r}zPageInfo.artbox$rr*c|jS)z+Return bleedbox of page in PDF coordinates.)rrs r(rzPageInfo.bleedbox)rr*c|jS)zReturn images.)r[rs r(imageszPageInfo.images.rr*c dd|js  td|jSfd|jDS)z?Return textareas bounding boxes in PDF coordinates on the page.cVd}||j|k7rd}||j|k7rd}|S)NTF) is_visibler)r want_visible want_corruptresults r( predicatez)PageInfo.get_textareas..predicate6s:F'CNNl,J'CNNl,JMr*z#Incomplete information on textboxesc3LK|]}|s|jywrkrl)rmrrHrrGs r(rnz)PageInfo.get_textareas..Es XSygw7WXs$$)rrrrWrrWr0r1)rr)rarGrHrs ``@r( get_textareaszPageInfo.get_textareas3sa  ,7 GR   "w':)*OPP?? "XDOOXXr*cJ|j tddS|jS)z3Return DPI needed to render all images on the page.rt)rrrs r(rwz PageInfo.dpiGs% 99 c3' 'yyr*c|jS)zReturn user unit of page.)rrs r(rzPageInfo.userunitNrr*c|jyy)z6Return minimum PDF version needed to render this page.z1.6z1.5rrs r( min_versionzPageInfo.min_versionSs == $r*cg}g}|jD]S}|js|j|jj |j|j Ut |}|dk(ry|Dcgc]}||z  }}tj||}t|}||z } |j|} || |z } t||| | Scc}w)aReturn information about the DPIs of the page. This is useful to detect pages with a small proportion of high-resolution content that is forcing us to use a high DPI for the whole page. The ratio is weighted by the area of each image. If images overlap, the overlapped area counts. Vector graphics and text are ignored. Returns None if there is no meaningful DPI for the page. rN) r[rvrrw to_scalar printed_areasum statistics harmonic_meanmaxindexrK) ra image_dpis image_areasrxtotal_drawn_areaareaweightsrLrMdpi_average_max_ratio arg_max_dpimax_area_ratios r(page_dpi_profilezPageInfo.page_dpi_profile[s  \\ 3E##   eii113 4   u11 2  3 {+ q 7BCt4**CC!// GD j/ ,w 6 &&w/ $[14DD$   !    Ds< Cc d|jd|jd|jd|jd|jd|j d S)Return string representation.z)rcrrrrwr:rs r(__repr__zPageInfo.__repr__sZkk]!D$5$5#6b9K9K8LM eDHH:Z a Q r*)FN) rbrrcrrdr rezContainer[int]rfr1rgzPdfMinerState | None)r0rr0r1)r0r )r0r)r0rZ)NN)rGrWrHrW)r0rr0r|)r0zPageResolutionProfile | None)rPrQrRrSrTr[rhr`propertyrcr:rrrrrrrsetterrrzrr}rrrrwrrrrr-r*r(rVrVtsIC!G_!#(,0     $    * "JQ JQJQ JQ $ JQ  JQ*JQX$$@@ &&""##AABB__BB Y( % N r*rVc.eZdZUdZdZded<dZded<dZded<dZded<dddd de d  dd Z e dd Z e dd Z e ddZe ddZe ddZe ddZe ddZe ddZe ddZddZdZdZy)PdfInfozExtract summary information about a PDF without retaining the PDF itself. Crucially this lets us get the information in a pure Python format so that it can be pickled and passed to a worker process. Fr1 _has_acroform_has_signature_needs_rendering_has_structure_treeNT)rfprogbar max_workers use_threadsreexecutorc ||_| tdd}tj|5}|jr t t |jjtjdjd} |r t|| n t|_|j5} t|||||||||  |_ddd|j"jtj$d|_tj(|j"vrt+|j"j(jtj,gdkDs&tj.|j"j(vrd|_t3|j"j(jtj4dd z|_t3|j"jtj8ijtj:d|_tj>|j"v|_ dddy#1swYxYw#1swYyxYw) z Initialize.Nriʚ;PScript5)rerfrgFTr#)!r^rangeropen is_encryptedrr|docinforrCreator startswithr r _miner_stater_pagesRootNeedsRenderingrAcroFormlenFieldsXFArr1SigFlagsrMarkInfoMarked _is_taggedStructTreeRootr) rardrfrrrrerrb pscript5_modergs r(rhzPdfInfo.__init__s  =1K XXf " G'))  b ABMMM %fm4 ]   "" k6 +&7 +   %(HHLL1D1De$LD !}}())--dkk2>?!Cxx388#4#44)-D&&*388+<+<+@+@PQ+RUV+V&W#" T]]B/33DKKGDO(,':':chh'FD $E" G" G  " G" Gs%BI-I E.sCd4##Cs)rryrs r(rzPdfInfo.min_versionsC CCCr*c:td|jDS)z(Return True if any page has a user unit.c3@K|]}|s|jdk7yw)rpNrrs r(rnz'PdfInfo.has_userunit..sGD$4==C'Gs)rryrs r( has_userunitzPdfInfo.has_userunitsGDJJGGGr*c|jS)z4Return True if the document catalog has an AcroForm.)rrs r( has_acroformzPdfInfo.has_acroformrr*c|jS)z@Return True if the document annotations has a digital signature.)rrs r( has_signaturezPdfInfo.has_signaturerr*c|jS)zCReturn True if the document catalog indicates this is a Tagged PDF.)rrs r( is_taggedzPdfInfo.is_taggedsr*c|jS)zAReturn True if the document catalog has a logical structure tree.)rrs r(has_structure_treezPdfInfo.has_structure_trees'''r*crt|jttzs t d|jS)zReturn filename of PDF.zcan't get filename from stream)r?r^r|r rrs r(filenamezPdfInfo.filenames,$,,d 3%&FG G||r*c|jS)zReturn True if PDF contains XFA forms. XFA forms are not supported by most standard PDF renderers, so we need to detect and suppress them. )rrs r(needs_renderingzPdfInfo.needs_renderings$$$r*c |j|S)z.Return PageInfo object for page number `item`.r)raitems r( __getitem__zPdfInfo.__getitem__s{{4  r*c,t|jS)zReturn number of pages in PDF.)r rrs r(__len__zPdfInfo.__len__s4;;r*c dt|dS)rzrS __future__rloggingrcollections.abcrrrr contextlibrdecimalr osr pathlibr typingr pdfminer.layoutr rpikepdfrrrocrmypdf._concurrentrrocrmypdf._pageboxesrocrmypdf.exceptionsrocrmypdf.helpersrocrmypdf.pdfinfo._contentstreamrrrocrmypdf.pdfinfo._imagerrocrmypdf.pdfinfo._typesrocrmypdf.pdfinfo._workerrocrmypdf.pdfinfo.layoutrr r! getLoggerloggerr)r<rIrKrVr.rr9rPr-r*r(rPs 6"CC"-##9*1'QQG-=    $ :66(O66"J4T T n"#~;~;B" zFr*