jZ%SSKJr SSKrSSKrSSKrSSKJr SSKJr SSK J r SSK J r J r JrJrJrJrJr SSKJrJrJrJrJrJrJrJrJrJrJrJrJ r J!r!J"r"J#r# SSK$J%r% SS K&J'r' SS K(J)r)J*r* SS K+J,r, SS K-J.r.J/r/ SS K0J1r1 SSK2J3r3 SSK4J5r5J6r6 SSK7J8r8 SSK9J:r: \Rv"\<5r=Sr>S\?S'Sr@S\?S'SSKAJBrBJCrC SSKDJErEJFrF SSKGJHrH SSKIJJrJ Sr>SrMS\?S'SSKNrNSSK0JOrOJPrP SrMSrQSrRS rSS!rTS"\?S#'S$rUS"\?S%'S&rVS"\?S''\'R"SSSSS(9rXS"\?S)'S-S*jrY"S+S,\.\/5rZg!\Ka rL\Lr@SrLCLNiSrLCLff=f!\Ka Ncf=f).) annotationsN)BytesIO)Path)mkdtemp)AnyCallableFinalIterableIteratorOptionalUnion) BoundingBox CoordOrigin DocItemLabelDoclingDocumentDocumentOrigin GroupLabelImageRefPictureClassificationLabelPictureClassificationMetaFieldPictureClassificationPrediction PictureMetaProvenanceItemSize TableCell TableDataTabularChartMetaField) ContentLayer)etree)ImageUnidentifiedImageError)override)DeclarativeDocumentBackendPaginatedDocumentBackend)convert_to_modern_format)MsPowerpointBackendOptions)FormatToMimeType InputFormat) InputDocument)DocumentLoadErrorFbool_PPTX_AVAILABLEzImportError | None_PPTX_IMPORT_ERROR) Presentation presentation)MSO_SHAPE_TYPEPP_PLACEHOLDER)InvalidXmlError)CT_TextLineBreakT_CHART_RENDER_AVAILABLE)crop_whitespaceget_docx_to_pdf_converterz}The 'python-pptx' package is required to process PowerPoint files. Install it with `pip install 'docling-slim[format-pptx]'`.zLibreOffice is required to render PowerPoint charts as images (render_chart_images=True). Install LibreOffice and make sure `soffice` is on PATH. Charts still keep their classification and reconstructed tabular data without it.zLibreOffice is required to rasterize EMF/WMF pictures embedded in slides. Install LibreOffice and make sure `soffice` is on PATH. Such pictures are still recorded with their position on the slide without it, but carry no image data.sƚr _WMF_PLACEABLE_MAGICs EMF_EMF_SIGNATURE(_EMF_SIGNATURE_OFFSET)resolve_entitiesload_dtd no_networkdtd_validation_SAFE_XML_PARSERcVUSS[:H=(d U[[S-[:H$)aReturn True when the bytes are a Windows metafile (WMF or EMF). Pillow renders metafiles only on Windows, where it delegates to the GDI ``PlayEnhMetaFile`` API; the wheels for every other platform carry a stub that raises on load. Telling a metafile apart from a genuinely broken image decides whether a picture is worth rasterizing externally. Args: image_bytes: Raw bytes of an embedded picture. Returns: True when the bytes carry a placeable WMF or an EMF signature. N)r7r:r8) image_bytess a/home/mande/repo/quber/.venv/lib/python3.13/site-packages/docling/backend/mspowerpoint_backend.py _is_metafilerDrs3 r?2 2 ),AA,EF.Xc^\rSrSr%SrSSSS.rS\S'S rS \S 'S rS r S4S5U4Sjjjr S6Sjr \ S7Sj5r \\ S7Sj55r\ S5r\\ S8Sj55r\ S9Sj5rS\"SSS94SjrS6SjrS:SjrS;SjrSSjrS?S jrS@S!jrSAS"jrSBS#jrS$r SCS%jr!S&r"S'r#\$SDS(j5r%\$SES)j5r&\$SFS*j5r'SGS+jr(S,r)SHS-jr*SIS.jr+SJS/jr,SKSLS0jjr-SMS1jr.SNS2jr/S3r0U=r1$)OMsPowerpointDocumentBackendaBackend for parsing PowerPoint presentations (PPTX and PPT files). Converts presentations into structured DoclingDocument format, extracting text, tables, images, lists, and comments from slides. Legacy ``.ppt`` files (binary PowerPoint 97-2003 format) are first converted to ``.pptx`` via LibreOffice before parsing. Note: Comments are extracted and added to the NOTES content layer. Unlike Word documents, PPTX comments only contain position coordinates (x, y) and do not directly reference specific shapes or text ranges. Therefore, comments are associated with their parent slide group rather than specific document elements. z5http://schemas.openxmlformats.org/drawingml/2006/mainz6http://schemas.openxmlformats.org/drawingml/2006/chartz:http://schemas.openxmlformats.org/presentationml/2006/main)acpzFinal[dict[str, str]] NAMESPACESiz Final[int]_SHAPE_ROW_TOLERANCE_EMUzLhttp://schemas.openxmlformats.org/officeDocument/2006/relationships/commentszRhttp://schemas.openxmlformats.org/officeDocument/2006/relationships/commentAuthorsc>[(d[[5[eUR[ R :Xa [USS5nUc [5n[TU])XU5 X l URRUl SUlSUlSUlSUlSUlSUl[)UR[*5(a[-UR5UlOB[)UR[.5(a#[-[1UR55UlSUlg![2an[5SUR635UeSnAff=f)NpptpptxFTz>MsPowerpointDocumentBackend could not load document with hash )r, ImportError _INSTALL_HINTr-formatr(PPTr%r&super__init__path_or_streamlimits page_rangepptx_to_pdf_converterpptx_to_pdf_converter_init_render_charts_metafile_hint_emittedpptx_objvalid isinstancerr.rstr Exceptionr* document_hash)selfin_docrWoptionse __class__s rCrV$MsPowerpointDocumentBackend.__init__s+ m,2D D ==KOO +5neVTN ?02G 94B --229="05'$),1#=A    $--w77 ,T-@-@ A D//66 ,S1D1D-E F DJ   #PQUQcQcPde  s5BD99 E!EE!cUR5(a.URce[URR5$g)Nr)is_validr^lenslidesrds rC page_count&MsPowerpointDocumentBackend.page_counts5 ==??==, ,,t}}++, ,rEcUR$N)r_rns rCrk$MsPowerpointDocumentBackend.is_valids zzrEcg)NTclss rCsupports_pagination/MsPowerpointDocumentBackend.supports_paginationsrEc[UR[5(aURR5 SUlgrr)r`rWrcloserns rCunload"MsPowerpointDocumentBackend.unloads/ d))7 3 3    % % '"rEcB[R[R1$rr)r(PPTXrTrvs rCsupported_formats-MsPowerpointDocumentBackend.supported_formatss  +//22rEc[URR=(d S[URSUR S9n[ URR=(d SUS9nUR(aDURUR5nURupEURURX#XES9nU$)zTParse the PPTX into a structured document model. Returns: The parsed document. filer)filenamemimetype binary_hash)nameorigin) start_pageend_page) rrrr' input_formatrcrstemr^_build_comment_author_maprY _walk_linear)rdrdoc author_maprrs rCconvert#MsPowerpointDocumentBackend.converts YY^^-v%d&7&78;**  499>>#;VFK ==77 FJ#'?? J## s:$C rEwidthheightcSURURURUR4;a1URnURnURnURnOSnSnURnURnXVXW-Xh-/n [R "U [ RS9n [US-S[U5/U S9n U $)Nr)rrpage_nocharspanbbox) lefttoprrr from_tupler BOTTOMLEFTrrl) rdshape slide_indtext slide_sizerrrr shape_bboxprovs rC_generate_prov*MsPowerpointDocumentBackend._generate_provs  EIIu{{ELLI I::D))CKKE\\FDC$$E&&Fs|<  ++J{?U?UV MQD N  rEcURSURS9nUb+SUR;a[UR S55$g![ a gf=f)aReturn the indentation level of a paragraph XML element. Paragraphs can have different indentation levels (0-8). The level is stored in the `lvl` attribute of the `a:pPr` element (paragraph properties). Args: paragraph: Paragraph XML element whose level should be extracted. Returns: Paragraph level in the range (0, 8). Returns 0 when no `a:pPr` element is found, no `lvl` attribute exists, or the `lvl` attribute value is invalid. a:pPr namespaceslvlr)findrLattribintget ValueError)rd paragraphpPrs rC_get_paragraph_level0MsPowerpointDocumentBackend._get_paragraph_levels^nnWnA ?u 2 3775>**  sA AAcDUcgURSURS9bgURSURS9nUbSSURS54$URS URS9nUbSS URS 54$URS URS9nUbg g)aParse bullet or numbering information from a paragraph properties node. This inspects the `a:pPr` or `a:lvlXpPr` element and extracts information about the bullet character, automatic numbering, picture bullets, or explicit `buNone` markers. Args: pPr: Paragraph properties XML element (`a:pPr` or `a:lvlXpPr`). Returns: A 3-tuple (`is_list`, `kind`, `detail`) where: `is_list` is True/False/None indicating whether this is a list item; `kind` is one of `buChar`, `buAutoNum`, `buBlip`, `buNone` or None, describing the marker type; `detail` is the bullet character, numbering type string, or None if not applicable. NNNza:buNoner)FbuNoneNza:buCharTbuCharcharz a:buAutoNum buAutoNumtypeza:buBlip)TbuBlipimage)rrLr)rdrrbuAutors rC'_parse_bullet_from_paragraph_propertiesCMsPowerpointDocumentBackend._parse_bullet_from_paragraph_properties3s& ;% 88J4??8 ; G**A  (FJJv$67 7-DOOD  +vzz&'9: :*A  ,!rEcNUcgSUS-S3nURX0RS9$)aFind the level-specific paragraph properties node from a list style. This looks for an `a:lvl{lvl+1}pPr` node inside an `a:lstStyle` element, where `a:lvl1pPr` corresponds to level 0, `a:lvl2pPr` to level 1, and so on. Args: lstStyle: List style XML element `a:lstStyle`. lvl: Paragraph level in the range (0, 8). Returns: Matching `a:lvl{lvl+1}pPr` XML element, or None if no matching element is found. Nza:lvlrrr)rrL)rdlstStylertags rC$_find_level_properties_in_list_style@MsPowerpointDocumentBackend._find_level_properties_in_list_style^s4  cAgYc"}}S__}==rEcUcgURSURS9nURX25nURU5upVnXVU4$)aParse bullet or numbering information from a text body's list style. This searches for `a:lstStyle/a:lvl{lvl+1}pPr` under a `txBody` and uses the level-specific paragraph properties to deduce bullet or numbering information. Args: txBody: Text body XML element `p:txBody`. lvl: Paragraph level in the range (0, 8). Returns: A 3-tuple (`is_list`, `kind`, `detail`) where: `is_list` is True/False/None indicating whether this is a list item; `kind` is one of `buChar`, `buAutoNum`, `buBlip`, `buNone` or None, describing the marker type; `detail` is the bullet character, numbering type string, or None if not applicable. rz a:lstStyler)rrLrr)rdtxBodyrrlvl_pPris_listkinddetails rC'_parse_bullet_from_text_body_list_styleCMsPowerpointDocumentBackend._parse_bullet_from_text_body_list_styleqsR& >%;;|;H;;HJ $ L LW Uvv&&rEc^URRSURS9nUcgU[R[R 4;aURSURS9$U[R :XaURSURS9$URSURS9$)aGet the appropriate master text style node for a placeholder. Most content placeholders (BODY/OBJECT) use `p:bodyStyle`, while titles use `p:titleStyle`. All other placeholders default to `p:otherStyle`. Args: slide_master: Slide master object associated with the current slide. placeholder_type: Placeholder type enum from `PP_PLACEHOLDER`. Returns: Matching style node from master `p:txStyles` (`p:bodyStyle`, `p:titleStyle` or `p:otherStyle`) or None when no styles are defined. z .//p:txStylesrNz p:bodyStylez p:titleStylez p:otherStyle)_elementrrLr1BODYOBJECTTITLE)rd slide_masterplaceholder_typetxStyless rC_get_master_text_style_node7MsPowerpointDocumentBackend._get_master_text_style_nodes  ((-- .     3 3^5J5JK K==4??=K K ~33 3==DOO=L L}}^}HHrEcURX5nUcgURSUS-S3URS9nURU5upgnXgU4$)aParse bullet or numbering information from the slide master text styles. This looks up the appropriate style bucket in the slide master's `p:txStyles` (`titleStyle`, `bodyStyle` or `otherStyle`) and extracts bullet or numbering information for the given level. Args: slide_master: Slide master object associated with the current slide. placeholder_type: Placeholder type enum from `PP_PLACEHOLDER`. lvl: Paragraph level in the range (0, 8). Returns: A 3-tuple (`is_list`, `kind`, `detail`) where: `is_list` is True/False/None indicating whether this is a list item; `kind` is one of `buChar`, `buAutoNum`, `buBlip`, `buNone` or None, describing the marker type; `detail` is the bullet character, numbering type string, or None if not applicable. rz.//a:lvlrrr)rrrLr) rdrrrstylerrrrs rC%_parse_bullet_from_master_text_stylesAMsPowerpointDocumentBackend._parse_bullet_from_master_text_styless_*00P =%**xay4*Q $ L LW Uvv&&rEcURnSnURnURnUbRURX15nUSSLd USS;aUSS:XagUSS:Xag g USS Lag URS :ag g UR SSUR S0S9bgUR SSUR S0S9bg URS :ag g ![a Nf=f)a6Determine whether a paragraph should be treated as a list item. The method first tries to resolve list style information via the shape that owns the paragraph. If that is not possible, it falls back to simpler checks based on paragraph properties and level. Args: paragraph: `python-pptx` paragraph object to inspect. Returns: A 2-tuple (`is_list`, `bullet_type`) where: `is_list` is True if the paragraph is considered a list item, otherwise False; `bullet_type` is one of `Bullet`, `Numbered` or `None`, describing the list marker type. NrTrrrrr)TBulletr)TNumbered)TNoneF)Frrz .//a:buCharrIrz.//a:buAutoNum)r_parentAttributeError_get_effective_list_markerlevelrrL)rdrrKr text_frame marker_infos rC _is_list_item)MsPowerpointDocumentBackend._is_list_items)    #**J&&E  99%KK9%-V1DI2 v&(2+ (K7-)9%.&"%" 66-S$//#2F,G6 H T# FF#dooc6J0KF L X% __q !"Q   sC CCcURnURU5nURSURS9nUR U5upgnUbUUUUS.$URRSURS9n UR X5upgnUbUUUUS.$Sn UR (Ga URRn URRRn Sn U RRU 5n U bU RRSURS9nUR X5upgnUbUUUUS.n URRnURRRR nUR#UX5upgnUS;aSUUUS.$UbUUUUS.$U bU $SSSUS.$![a Sn Nf=f)aWReturn a dictionary describing the effective list marker for a paragraph. List marker information can come from several sources: direct paragraph properties, shape-level list styles, layout placeholders, or slide master text styles. This helper resolves all of these layers and returns a unified view of the effective marker. Args: shape: Shape object that contains the paragraph. paragraph: `python-pptx` paragraph object to inspect. Returns: Information about the list marker in a dictionary, where: `is_list` is True/False/None indicating if this is a list item; `kind` is one of `buChar`, `buAutoNum`, `buBlip`, `buNone` or None, describing the marker type; `detail` is the bullet character or numbering type string, or None if not applicable; `level` is the paragraph level in the range (0, 8). rrN)rrrrz .//p:txBodyrT)rrrrLrris_placeholderplaceholder_formatidxpartslide slide_layout placeholdersrrbrrr)rdrrrKrrrrrr layout_resultrlayout layout_ph layout_txph_typemasters rCr6MsPowerpointDocumentBackend._get_effective_list_marker s;(   ''*ffWf9 $ L LS Qv  "   $$]t$O $ L L ! v  "       **..CZZ%%22FI !"//33C8 $%..33!doo4 )-(T(T)%v &#* $"(!$ %M 2277))66CC(,(R(RG)%v <<#' $"(!$  (#* $"(!$ ($$   a !  !s%F== G  G cn[X5nUcg[ U5$![[[4a gf=f)a$Return a shape position attribute as an integer for ordering. Args: shape: A python-pptx shape object. Position attributes such as ``top`` and ``left`` are accessed via ``getattr``; property getters that raise are silently treated as absent. attr: Name of the position attribute to read, typically ``"top"`` or ``"left"``. Returns: The attribute value cast to ``int``, or ``None`` if the attribute is missing, ``None``-valued, or its getter raises ``AttributeError``, ``ValueError``, or ``TypeError``. N)getattrrr TypeErrorr)rdrattrvalues rC_get_shape_position/MsPowerpointDocumentBackend._get_shape_positionws@ E(E =5z  I6  s 44c #l# SSKJn "SSU5nURnSn/n[U5HKupxUR US5n UR US5n UR U"UUU bU OUU bU OUS 95 MM UR S S 9 /n /n Sn UHXnU bURU - U::aU R U5 URn M8U R U 5 U/n URn MZ U (aU R U 5 U H#n[US S 9HnURv M M% g7f) a%Iterate shapes in visual top-to-bottom, left-to-right order. PowerPoint stores shapes in creation/z-order, which can differ from the way content is visually read on a slide. Sorting by position keeps split text boxes, such as a subheading followed by its bullet text box, adjacent in the extracted document. Shapes that share a top-edge within ``row_tolerance`` EMUs of their immediate predecessor (after sorting by ``top``) are placed in the same row and then sorted left-to-right within that row. Adjacency is evaluated against the previous shape's ``top``, not the row's first anchor, so that a row can span more than one tolerance step when shapes form a contiguous band. Shapes whose position attributes are unavailable or raise an exception are placed at the end in their original relative order. Args: shapes: Iterable of python-pptx shape objects to sort. May be a slide's top-level shape collection or the children of a group shape. Yields: Shapes from ``shapes`` in visual reading order (top-to-bottom, left-to-right). r) NamedTuplec>\rSrSr%S\S'S\S'S\S'S\S'Srg ) HMsPowerpointDocumentBackend._iter_shapes_by_position.._ShapeInfoirindexobjectrrrruN)__name__ __module__ __qualname____firstlineno____annotations____static_attributes__rurErC _ShapeInforsJMHIrEr lrrN)rrrrc2URUR4$rr)rrsis rCFMsPowerpointDocumentBackend._iter_shapes_by_position..sRVVRXX$6rE)keyc2URUR4$rr)rrrs rCrrs"''2881DrE) typingrrM enumeraterappendsortrsortedr)rdshapesrr  row_tolerancefallback_positionentriesrrrrrows current_rowprev_topentryrowrs rC_iter_shapes_by_position4MsPowerpointDocumentBackend._iter_shapes_by_positions<2 &   55 %$&%f-LE**5%8C++E6:D NN"4E!%!17H  .  6 7')(* E599x#7=#H""5) 99 K($g  99  KK $CS&DEhhFsD2D4c FSnSnSn[Rn URXURR 5U5n UR R GHEn URU 5upU RnSnURH.n[U[5(aUS- nMUUR- nM0 U (aQSnU S:HnU(dURSUS9nSnSnU(aUS - n[U5S -nURUUUUU S 9 MU(aSnSnSn[Rn UR (aJUR"R$nU[&R([&R*4;a[R*n UR-U UUU S 9 GMH g) NFrr rlist)rparentTr.)marker enumeratedr(rrlabelr(rr)r LIST_ITEMrrstripr paragraphsrrcontent_childrenr`r3add_list_groupra add_list_item PARAGRAPHrrrr1 CENTER_TITLEradd_text)rdr parent_sliderrris_list_group_createdenum_list_item_valuenew_list doc_labelrr is_a_list bullet_typerKp_textrg enum_markerr+rs rC_handle_text_elements1MsPowerpointDocumentBackend._handle_text_elementss!&  ** ""5UZZ5E5E5GT))44I%)%7%7 %B "I""AF''a!122cMFaff$F (  (J6 ,"11#+ 2 H-1)+,((A-("%&:";c"AK!!&)# "),1)#H+,()22 ''','?'?'D'D$'&33&,,, %1$6$6  #' k5v rEc:URRnURR5nUR XU5n[ UR55S:aU[ R[ R4;aB[RSUR35 UR[RX&US9 gU[ R:XaB[RSUR35 UR[RUUUS9 g)Nrz Title found: r,zSubtitle found: )rrrr/rrlr1r5r_loginfor6rSUBTITLESECTION_HEADER)rdrr7rrrtxtrs rC _handle_title)MsPowerpointDocumentBackend._handle_titles 3388jj ""5S9 syy{ a N$?$?AUAU#VV M%**67 &,,\RV "^%<%<< ,UZZL9: &55'  rEctUR5nUc2UR(d SUl[R[5 gUSS[ :XaSOSn[ [55nUSU3- nUS- nURU5 U"XV5 UR5(d7[RSUR5 [R"USS 9 g[R"[!U55nUS n[#UR%S S 9R'55n UR)5 UR)5 U [R"USS 9 $![*a5n [RS U 5 Sn A [R"USS 9 gSn A ff=f![R"USS 9 f=f)aRasterize a raw EMF or WMF picture via LibreOffice. LibreOffice converts a standalone ``.emf``/``.wmf`` to PDF directly, so no wrapper document is needed; the first page is then rendered with pypdfium2. This is the same route ``_render_chart_image`` already takes for native charts. Args: image_bytes: Raw metafile data. Returns: A PIL Image, or None when LibreOffice is unavailable or the conversion fails. NTrAz.wmfz.emfrz image.pdfz)LibreOffice produced no PDF output for %s ignore_errorsrscalez0EMF/WMF rasterization via LibreOffice failed: %s)_get_libreoffice_converterr]rCwarning_IMAGE_RENDER_HINTr7rr write_bytesexistsdebugrshutilrmtree pypdfium2 PdfDocumentrar5renderto_pilr{rb) rdrB convertersuffixtemp_dir input_path output_pathpdfpage pil_imageexcs rC_rasterize_metafile/MsPowerpointDocumentBackend._rasterize_metafile5sj335  ...2+ /0&r?.BB ? 8!eF8$44J"[0K  " "; / j .%%'' F X MM($ 7''K(89Cq6D' ! (<(C(C(EFI JJL IIK MM($ 7   JJI3 O MM($ 7   MM($ 7s2+AEA+E F(F>F FF F7cURnURnURupURXSU5n Sn [R"[U55n [ R""XS9n UR-UU SU S9 g![[ [ [[[4a'n [R"SU 3[SS9 Sn A gSn A ff=f![[[4a{n [%U5(d"[R"SU 3[SS9 Sn A g[&R)SU 5 UR+U5nUb[ R""USS9n Sn A NSn A ff=f) Nrz"Skipping malformed picture shape: rM stacklevelrdpiz&Pillow cannot render this metafile: %sHr(rcaptionr)rblobrkrr!r2KeyErrorrrOSErrorwarningswarn UserWarningr openrrfrom_pilrDrCrUre add_picture)rdrr7rrrrrBim_dpi_rrg image_refrc rasterizeds rC_handle_pictures,MsPowerpointDocumentBackend._handle_picturesasX KKE**K IF&&uZHD")-  H 7;#78I )) FI(   Y #        MM4QC8   ('< H ,, 8<   JJ? C11+>J%%--JBG # Hs49B3C$C(C  CE"'-E?EE"cUR(GaURnURnURXSU5nSn [ UR 5n /n URn[ UR 5GH.up[ U R5U :a[ U R5n [ U R5HupURSU S-SUS-S35nU(dM,USnURS5nURS5nUcSnO [U5nUcSnO [U5n[URR5UUU U U-UUU-U S:HS S 9 n[ URR55S:dMU RU5 M GM1 [X/S 9nU HnUR RU5 M [ U 5S:aUR#UUUS 9 g) Nrrz.//a:tbl/a:tr[rz]/a:tc[]rowSpangridSpanF rrow_spancol_spanstart_row_offset_idxend_row_offset_idxstart_col_offset_idxend_col_offset_idx column_header row_headernum_rowsnum_cols table_cells)r(datar) has_tabletablerrrlrrcellsxpathrrrrr/rrr add_table)rdrr7rrrr table_xmlrrrtcellsrow_idxr"col_idxcellcell_xmlrricellrtcells rC_handle_tables*MsPowerpointDocumentBackend._handle_tabless ???KKEI&&uZHDH5::HFI )%** 5 syy>H,"399~H%.syy%9MG((1 WWq[MK H$ '{H'||I6H'||J7H'#$#&x='#$#&x=%!YY__.!)!)-4+2X+=-4+2X+=&-l#( E499??,-1 e,G&:!6PhrRD  ''. 6{Q \4 HrEc Ub UROSnSU;dSU;a[R$SU;a[R$SU;a[R$SU;dSU;a[R $[R $)aMap a python-pptx ``XL_CHART_TYPE`` to a docling classification label. ``chart.chart_type`` is a granular enum (``COLUMN_CLUSTERED``, ``XY_SCATTER_LINES``, ``THREE_D_PIE``, ...). Rather than enumerate every member, we match on the enum member's name by family, mirroring the Excel backend's tagname mapping: column/bar variants become BAR_CHART, line variants LINE_CHART, pie/doughnut PIE_CHART, scatter SCATTER_CHART, and everything else (area, radar, stock, surface, ...) OTHER_CHART. SCATTER is matched before LINE because ``XY_SCATTER_LINES`` contains both. Args: chart_type: An ``XL_CHART_TYPE`` member, or None when python-pptx cannot determine the type (e.g. combination charts). Returns: The matching PictureClassificationLabel. rPIEDOUGHNUTSCATTERLINEBARCOL)rr PIE_CHART SCATTER_CHART LINE_CHART BAR_CHART OTHER_CHART) chart_typers rC_chart_type_to_classification9MsPowerpointDocumentBackend._chart_type_to_classificationsz&#-"8zb D=J$.-77 7  -;; ; T>-88 8 D=ETM-77 7)555rEcUR(dgURRRR 5nU=(d S$)zReturn the chart's title text, or None when it has no title. Args: chart: A python-pptx ``Chart`` object. Returns: The stripped title text, or None. N) has_title chart_titlerrr/)chartrs rC_chart_title_text-MsPowerpointDocumentBackend._chart_title_texts8  ++00668|trEcUcg[U[5(a)UR5(a[[ U55$[U5$)aFormat a chart category label or data value as cell text. python-pptx returns numeric series values as floats and empty points as None. Integer-valued floats are rendered without a trailing ``.0`` so the reconstructed table reads like the source data (``120`` not ``120.0``). Args: value: A category label or data point (str, float, int, or None). Returns: The value as a string ("" for None). r)r`float is_integerrar)rs rC _cell_text&MsPowerpointDocumentBackend._cell_texts? = eU # #(8(8(:(:s5z? "5zrEc[UR5nU(dg[UR5nU(a.USRVs/sHo@R U5PM snO/n/nUHinUR bUR UR 5OSnUR V s/sHoR U 5PM n n URX45 Mk [[U5/UV V s/sHup[U 5PM sn n -5n U S:XagU S-n S[U5-n/nS/UVV s/sHupUPM sn n-n[U5H)unnUR[USSSSUUS-SSS9 5 M+ [U 5HnUS-nU[U5:aUUOSnU/UV V s/sHupU[U 5:aU UOSPM sn n -n[U5H/unnUR[USSUUS-UUS-SUS:HS9 5 M1 M [XUS9$s snfs sn fs sn n fs sn nfs sn n f) uReconstruct a chart's underlying data grid as a TableData. Layout produced (categories down the first column, one column per series): | | | | ... | cat_0 | val_0,0 | val_1,0 | ... | cat_1 | val_0,1 | val_1,1 | ... Unlike the Excel backend, python-pptx exposes the plotted numbers directly on each series (``series.values``) and the shared category labels on the plot (``plot.categories``) — no workbook reference resolution is needed. Args: chart: A python-pptx ``Chart`` object. Returns: A TableData, or None if the chart exposes no usable series. NrrrTFrr)r'seriesplots categoriesrrvaluesrmaxrlrrranger)rdr series_listrcatrcolumnsrrvrry num_data_rowsrrr header_labelsrr-data_rowrcategory row_textsrs rC_chart_to_table_data0MsPowerpointDocumentBackend._chart_to_table_datasT(5<<( U[[!EJU1X-@-@ A-@c__S !-@ APR 02!F39;;3J4??6;;/PRD28--@-Qooa(-F@ NND> *" S_-g0VgVg0VVW A  1$s7|#!#GR@UUU[BRDS9 TRGXX2U5 GME U$)NF rzslide-r)rr-r(r)rsizec|UR$![a# [RSUR5 gf=f)zReturn shape.shape_type, or None if unrecognized. python-pptx raises NotImplementedError for elements that don't match any known shape category (placeholder, freeform, autoshape, textbox). z)Skipping shape with unrecognized type: %sN) shape_typerrCrUr)rs rC_safe_shape_typeBMsPowerpointDocumentBackend._walk_linear.._safe_shape_typeGs8  +++* JJJEJJW s *;;c>T"XX#U5 UR(aTRXX#U5 UR(aTRXX#U5 T"U5[R :XaTR XX#U5 [US5(dgURcg[URR55S:XagUR(d[RS5 gTRXX#U5 g)Nrrz*Warning: shape has text but not text_frame)rr has_chartrr0PICTUREr|hasattrrrlr/has_text_framerCrQr@)rr7rrrr handle_groupsrds rC handle_shapes?MsPowerpointDocumentBackend._walk_linear..handle_shapesTse9:N??''YZX??&&uIJW#E*n.D.DD))YZuf--::%uzz'')*a/++LL!MN**rEc>T"U5[R:Xa-TRUR5H nT"XQX#U5 M ggrr)r0GROUPr#r) rr7rrr groupedshaperrrds rCr?MsPowerpointDocumentBackend._walk_linear..handle_groupsqsE#E*n.B.BB(,(E(Eell(S %( )TCrE)ltrbr)r-r(rr content_layer)$ slide_width slide_heightr`rfr&render_chart_imagesr\rPrCrQ_CHART_RENDER_HINTrr'rrm add_grouprCHAPTERradd_pager#rhas_notes_slide notes_slidenotes_text_framerr/rrrlr6rTEXTrNOTES_extract_slide_comments)rdr^rrrrr r  max_levelsparentsiselected_slidesrrr7rrr notes_textrrrrrs` @@@rCr(MsPowerpointDocumentBackend._walk_linear$s** ,,  t||%? @ 1 00    4#B#B#D#L LL+ ,"'D  z"AGJ#y9::>HU / I==i[)1C1CGTUJ)LKEJ LLQZL @  : 66u||De\cNE$$#// //;!,!=!=!B!B!H!H!JJ!*Q!qA>-$-M&'Z%9!%   "."3"3#/!+!%*6*<*< %  ( (*< e!0l rEc 0nURRR5HnURUR:XdM[ R "URR[S9nURSURS9Vs0sH8nURSS5URSS5URSS54_M: nnM U$s snf![a#n[RSU35 S nAU$S nAff=f) zBuild a map of comment author IDs to (name, initials). Args: pptx_obj: The PowerPoint presentation object. Returns: Dictionary mapping author ID to (name, initials) tuple. parserz p:cmAuthorridrrinitialsz&Could not parse PPTX comment authors: N)rrelsrreltypeCOMMENT_AUTHORS_RELr fromstring target_partror?findallrLrrbrCrU)rdr^rrelroot author_elrgs rCr5MsPowerpointDocumentBackend._build_comment_author_maps 24  E}}))002;;$":":: ++,,5ED*.(T__*6* " *I " dB/%MM&"5%MM*b92* "J 3 " E JJ?sC D D Es0?CA C?CCC D &DD c URnURR 5GHnUR UR:waM [R"URR[S9n U RSURS9GHtn U RSS 5n U RS S 5n U R!S URS9n U b#U R"=(d S R%5OS nU(dMsURU S 5unn/nU(a&S U3nU(a USUS3- nUR'U5 U (aUR'SU 35 SR)U5nU(aSUSU3OUnU RS[+U55nUR-[.R0SUS-SU3[2R4S9nUR7USUS9 [RSUS-SU35 GMw GM g![a(n[RSUS-SU35 SnAgSnAff=f![a'n[RSUSU35 SnAGMHSnAff=f)aDExtract and add comments for a specific slide. Args: slide: The slide object to extract comments from. slide_idx: Zero-based slide index. author_map: Dictionary mapping author ID to (name, initials). doc: The DoclingDocument to add comments to. parent_slide: The parent slide group to link comments to. z&Could not access slide part for slide rz, skipping comment extraction: Nrzp:cmrauthorIdrdtzp:text)rrzauthor: z ()ztime: z, [z]: rz comment-slide-)r-rr )rtargetsr(zAdded PPTX comment slide z idx=z#Could not parse comments for slide z: )rrbrCrUr#rr$ COMMENT_RELrr&r'ror?r(rLrrrr/rjoinrarrCOMMENT_SECTIONrr add_comment)rdr slide_idxrrr7 slide_partrgr)r*cm author_idr/text_elraw_textrr"metadata_parts author_strprefix full_text comment_idx comment_groups rCr3MsPowerpointDocumentBackend._extract_slide_commentss_" J??))+C{{d...' S''(<(&'' <'4I !I<'36' <':A#Fl \2ENE3BEN.*8X5n>@66:  &PRdKZ* ++.1+=A+ +Z-8-8.1-8 -8h"& q+qq/ q  q  q qf1 #>DSDS/ DS  DS DSDSrErG)rBrIrEr+)[ __future__rloggingrVrriorpathlibrtempfilerrrrr r r r r docling_core.types.docrrrrrrrrrrrrrrrrdocling_core.types.doc.documentrlxmlrPILr r!typing_extensionsr" docling.backend.abstract_backendr#r$$docling.backend.docx.drawingml.utilsr%!docling.datamodel.backend_optionsr&docling.datamodel.base_modelsr'r(docling.datamodel.documentr)docling.exceptionsr* getLoggerrrCr,r r-rPr.r/pptx.enum.shapesr0r1pptx.excr2pptx.oxml.textr3rQrgr4rXr5r6rRrrRr7r8r: XMLParserr?rDrGrurErCrfsq# LLL$9-&JHG40")-&-/?(/O!&%   # A 2e1!u! //  % U&uS"<>VuSW"  s*?D9%E9E ?EE EE