/j-j΃ SrSSKrSSKrSSKrSSKJr SSKJr SSKJ r SSK J r SSK J r JrJrJrJrJrJrJrJrJrJr SSKJr SS KJrJrJr SS KJrJ r J!r!J"r"J#r#J$r$J%r%J&r&J'r'J(r(J)r) SS K*J+r+J,r,J-r-J.r.J/r/ SS K0J1r1J2r2J3r3J4r4J5r5 \Rl"\75r8\"S 5r9\"S5r:\"S5r;\"S5r<\"S5r=\"S5r>\"S5r?\S/\@4rA\"S5rB\ (aSSKCJDrD SSKEJFrFJGrG SSKHJIrIJJrJ SSKHJKrL \"SS\S55rM\"SS55rKS;S\\N\\N\AS4S \A4S!jjrOS;S"\SS\\N\\N\AS4S \S4S#jjrP"S$S%\5rQ"S&S'\\\M\K45rR\"S(S\Q\\S\M\K455rS\RS)\S*\S+S,\1S \\S\M\K44S-j5rU\UR\W5S)\WS*\S+S,\1S \\S\M\K44S.j5rX\UR\Y5S)\YS*\S+S,\1S \\M4S/j5rZ\UR\[5S)\\N\4S*\S+S,\1S \\\M\KS44S0j5r\S)\\N\4S*\S+S,\1S \\M4S1jr]S)\\N\4S*\S+S,\1S \\\KS44S2jr^S)\\N\4S*\S+S,\1S \\2S44S3jr_S4S5S \\S\M\K44S6jr`"S7S8\Q\\S\M\K45ra"S9S:\\\SS45rbg) Lazy interface to PDF logical structure (PDF 1.7 sect 14.7). N)deque) dataclass)groupby) attrgetter) TYPE_CHECKINGCallableDequeDictIterableIteratorListPatternSequenceUnionoverload) NumberTree)LIT PDFObject PSLiteral) KWD ContentStreamObjRefRect dict_value int_value list_value literal_nameresolve1 rect_value stream_value)transform_bboxget_bound_rectsstring_propertychoplist decode_text) DocumentRefPageRef_deref_document _deref_page _ref_documentAnnotXObjectImageMCROBJRStructTreeRoot StructElemElementsBDC)Document) AnnotationPage) ImageObject XObjectObject) ContentObjectc\rSrSr%Sr\\S'\\S'\\ S4\S'Sr \\ S4\S'Sr \\ S4\S'\S \S 4S j5r\SS j5r\S \\ S44S j5r\S \ 4Sj5rS \S4SjrS \S4SjrSrg) ContentItemKzContent item in logical structure tree. This corresponds to an individual marked content section on a specific page, and can be used to (lazily) find that section if desired. One can also iterate over ContentObjects inside it, or extract text from it. _pagerefmcidNstream_bbox_textreturnr5NcHURcg[UR5$)z.Specific page for this structure tree, if any.N)r<r)selfs L/home/mande/repo/quber/.venv/lib/python3.13/site-packages/playa/structure.pypageContentItem.page]s  == 4==))c4URup[U5$z,The document containing this content object.r<r(rEdocref_s rFdocContentItem.docdMM v&&rIcURb UR$URb[SU55UlUR$)zFind the bounding box, if any, of this item, which is the smallest rectangle enclosing all objects in its marked content section. If the `page` attribute is `None`, then `bbox` will be `None`. c38# UHoRv M g7fN)bbox).0objs rF #ContentItem.bbox..vs(BTcTs)r?rGr"rDs rFrVContentItem.bboxjs> :: !::  99 ((BT(BBDJzzrIclSSKJn URb UR$UR5nUcSUlUR$/nURUR Hn[ XA5(dMURnUb UR cM3SUR;a=[ URS[5(de[URS5nO URnURSS5nURU5 M SRU5UlUR$)z,Unicode text contained in this content item.r) TextObject ActualText­) playa.contentr]r@_page_or_xobjectmarked_contentr= isinstancemcspropsbytesr%charsreplaceappendjoin)rEr]page_or_xobjecttextsrXrerhs rFtextContentItem.textys - :: !:: //1  "DJ:: "11$))xobjectsobjid)rErXs rFrbContentItem._page_or_xobjectsZ 99  ;; "yy))::##t{{'8'88J*yyrIPageContentObjectcUR5nUc [S5$URURnUc [S5$[U5$)zIterate over `playa.content.ContentObject` (not to be confused with `playa.structure.ContentObject`) in this marked content section. )rbiterrcr=)rErlcontentss rF__iter__ContentItem.__iter__sN//1  "8O"11$))<  8OH~rI)r?r@rAr3)__name__ __module__ __qualname____firstlineno____doc__r'__annotations__intrrr?rr@strpropertyrGrPrVrnrbr ry__static_attributes__rvrIrFr:r:Ks I -% &&#E5t #"E5d " *eL)** ''  eD$J'  c8%(E"F (#67 rIr:rtc\rSrSr%Sr\\S'\\\ \ \ 44\S'\ S\S4Sj5r \ S\4Sj5r\ SS j5r\ SS j5r\ S\\S 44S j5rS rg )r8aContent object in logical structure tree. This corresponds to a content item that is an entire PDF (X)Object (PDF 1.7 section 14.7.43), and can be used to (lazily) get that The standard is very unclear on what this could be aside from an `Annotation` or an `XObject` (presumably either a Form XObject or an image). An XObject must be a content stream, so that's clear enough... Otherwise, since the `Type` key is not required in an annotation dictionary we presume that this is an annotation if it's not present. Not to be confused with `playa.content.ContentObject`. While you *can* get there from here with the `obj` property, it may not be a great idea, because the only way to do that correctly in the case of an `XObject` (or image) is to interpret the containing page. Sometimes, but not always, you can nonetheless rapidly access the `bbox`, so this is also provided as a property here. r<rfrA)r7r6r4NcURnU[La-SSKJn UR"UR UR 5$U[La[UR [5(deUR RS5nU[LaUR ROUR RnUH5nURRUR R:XdM3Us $ g)z+Return an instantiated object, if possible.r)r4SubtypeN)type LITERAL_ANNOT playa.pager4 from_dictrfrGLITERAL_XOBJECTrdrget LITERAL_IMAGEimagesrqr>rr)rEobjtyper4subtypeitorrXs rFrXContentObject.objs)) m # -'' DII> > o %djj-88 88jjnnY/G'.-'?499##TYYEWEWD::##tzz'7'77JrIc[UR[5(a[$URR S5n[U[ 5(d[ $U$)z>Type of this object, usually LITERAL_ANNOT or LITERAL_XOBJECT.Type)rdrfrrrrr)rErs rFrContentObject.typesE djj- 0 0" "**..('9-- rIc,[UR5$)z(Containing page for this content object.)r)r<rDs rFrGContentObject.pages4==))rIc4URup[U5$rKrLrMs rFrPContentObject.docrRrINcZSUR;a8[URS5n[URRU5$SUR;a8[URS5n[URRU5$UR nUcgUR $)zqFind the bounding box, if any, of this object. If there is no bounding box (very unlikely) this will be `None`. BBoxrN)rfrr!rGctmrXrV)rErawboxrXs rFrVContentObject.bboxs TZZ  6 23F!$))--8 8 TZZ  6 23F!$))--8 8hh ;xxrIrv)rAr5r{)r|r}r~rrr'rrrr rrrrXrrrGrPrrVrrvrIrFr8r8s, S)^ 44 55 UMN$i**'' eD$J'rIr8matcherrAc^SSS[4SjnSSS[4U4SjjnSSS[4U4SjjnTcU$[T[5(aU$[T[R5(aU$T$)Nxr2rAcg)NTrv)rs rF match_all#_make_match_func..match_allsrIc">URT:H$)zMatch an element name.)rolerrs rF match_tag#_make_match_func..match_tag svv  rIc:>TRUR5$)z,Match an element name by regular expression.)matchrrs rF match_regex%_make_match_func..match_regexs}}QVV$$rI)boolrdrrer)rrrrs` rF_make_match_funcrsxY4!Y!4!%y%T% GS ! ! GRZZ ( (rIelementsc#,# [U5nUR5 U(anUR5nU"U5(aUv [[ U55H+n[ U[ 5(dMURU5 M- U(aMmgg7f)z5 Common code for `find_all()` in trees and elements. N)rreversepopreversedlistrdr2rj)rr match_funcelchilds rF _find_allrsl"'*J   \\^ b>>Hd2h'E%))&( (sA.B4BBc|\rSrSrSrS S\\\\\S4S\ S4Sjjr S S\\\\\S4S\S4S jjr S r g) Findablei.zNfind() and find_all() methods that can be inherited to avoid repeating oneselfNrrAr2c,[[U5U5$)aIterate depth-first over matching elements in subtree. The `matcher` argument is either a string, a regular expression, or a function taking a `Element` and returning `True` if the element matches, or `None` (default) to return all descendants in depth-first order. For compatibility with `pdfplumber` and consistent behaviour across documents, names and regular expressions are matched against the `role` attribute. If you wish to match the "raw" structure type from the `type` attribute, you can do this with a matching function. )rrrErs rFfind_allFindable.find_all2s"dW--rIr2Nc`[[[U5U55$![a gf=f)a#Find the first matching element in subtree. The `matcher` argument is either a string or a regular expression to be matched against the `role` attribute, or a function taking a `Element` and returning `True` if the element matches, or `None` (default) to just get the first child element. N)nextrr StopIterationrs rFfind Findable.findEs/  $t*g67 7  s  --rvrU) r|r}r~rrrrr MatchFuncr rrrrvrIrFrr.srDH.S'#, 4?@. ) .(DHS'#, 4?@  rIrcx\rSrSrSrS\\S\\44Sjr \ S\ \\\44Sj5r S\\\44Sjr S rg ) ContentIteratoriWz.Iterator over contents of a structure element.rr2c&[U5/UlgrU)rw_itors)rErs rF__init__ContentIterator.__init__ZsBxj rIrAc## [U5nUR[S5S9 [U[S55GH up#[U5n0nUHCn[ U[ 5(aUv M[ U[ 5(dM5XeUR'ME [/S5n[URRS055nURHn URU 5 U [LdMUupn [ U [5(aURU R 05n [U 5n SU ;dMh[#U S5nX;dM}X^v M GM# g7f)a#Iterate over content items and objects in page order as defined in PDF 2.0 section 14.8.2.5. Items are grouped by page, then ordered inside a page by their appearance in the content stream. Because content objects are outside the page's content stream, it is not clear how they should be ordered with respect to marked content sections. For lack of a better idea, we will place them at the start of the page ordering. This interface is not at all Lazy as it requires us to create the entire set of items/objects, then interpret all the pages. z page.page_idx)keyrG PropertiesMCIDN)rsortrrrdr8r:r=rr resourcesrrxrjKWD_BDCrnamer)rE all_itemsrGitems page_itemsby_mciditemstackproprsrcrXtagrfrOpropdictr=s rF page_orderContentIterator.page_order]s  J :o67"9j.@AKDeJ.0G"dM22Jk22)-DII& # ',BlE!$.."4"4\2"FGH}} S!'>$)MC!%33 ( UZZ <)%0H)(&)9:?")-/%Bs A?E!A*E!3AE!<E!E!cnUR(d[e[URS5n[ U[ 5(a&URR [U55 Ml[ U[[45(aU$M![a URR5 Mf=f)N) rrrrrdr2rjrwr:r8)rEkids rF__next__ContentIterator.__next__s;;## 4;;r?+#w'' ""49-# ];<<  !  ! sB $B43B4)rN)r|r}r~rrr rr:r8rrr rrrrvrIrFrrWsc8!8E)[-*O$PQ!&0HU; +E%FG&0&0P % ] :; rIrcr\rSrSr%Sr\\S'\\\ 4\S'Sr \ \S4\S'\ SSS \\\ 4S S4S j5r \S \4S j5r\S \4S j5r\S Sj5r\S \ S4Sj5r\S \ S4Sj5r\S \4Sj5r\S \4Sj5r\S \ \S44Sj5r\S \ \S44Sj5r\S \ \S44Sj5r\S \ \S44Sj5r\S \ \S44Sj5r\S \ \S44Sj5r\S \ \\\ 4S44Sj5r\S \ \S44Sj5rS \ \ S\!\"44Sjr#S \$4Sjr%Sr&g)!r2izcLogical structure element. Attributes: props: Structure element dictionary (PDF 1.7 table 323). _docrefrfN_rolerPr3rXrAc"U"[U5US9$)z0Construct from PDF structure element dictionary.rrf)r*)clsrPrXs rFrElement.from_dicts=-S99rIc2[URS5$)aStructure type for this element. Note: Raw and standard structure types This type is quite likely idiosyncratic and defined by whatever style sheets the author used in their word processor. Standard structure types (PDF 1.7 section 14.8.4) are accessible through the `role_map` attribute of the structure root, or, for convenience (this is slow) via the `role` attribute on elements. S)rrfrDs rFr Element.typesDJJsO,,rIcURb UR$URRnUc UR$URR URUR5$)aStandardized structure type. Note: Roles are always mapped Since it is common for documents to use standard types directly for some of their structure elements (typically ones with no content) and thus to omit them from the role map, `role` will always return a string in order to facilitate processing. If you must absolutely know whether an element's type has no entry in the role map then you will need to consult it directly. )rrP structurerrole_mapr)rEtrees rFr Element.rolesU :: !:: xx!! <99 }}  DII66rIc,[UR5$)z%Containing document for this element.r(rrDs rFrP Element.doct||,,rIrBclURRS5nUcg[U[5(a0URR R UR5$[RSUR5 g![a# [RSUR5 gf=f)z)Containing page for this element, if any.PgN$'Pg' entry not found in document: %r2'Pg' entry is not an indirect object reference: %r) rfrrdrrPpagesby_idrrKeyErrorLOGwarning)rEpgs rFrG Element.pagesZZ^^D ! : F # # Pxx~~++BHH55 KKDdjj   P BDJJO   Ps.B*B32B3)r2TreeNc[URRS55nUcg[U5nURS5[LaUR R $[RUR U5$)NPr) rrfrrLITERAL_STRUCTTREEROOTrPrr2r)rEps rFparentElement.parentsb TZZ^^C( ) 9 qM 55=2 288%% %  1--rIc[U5$)zIterate over all content items contained in an element. By default these are returned in depth-first order, which is "logical content order" as defined in PDF 2.0 section 14.8.2.5. rrDs rFrxElement.contentsst$$rIcFSRSUR55$)z2Unicode text of all content items in this element.r^c3h# UH(n[U[5(dMURv M* g7frU)rdr:rn)rWrs rFrYElement.text..s" "/$:dK3PIDII-s22)rkrxrDs rFrn Element.texts&ww "&--   rIc^URmTcgSUR;a.[URS5n[TRU5$U4SjUR 5n[ U5$)aThe bounding box, if any, of this element. Elements may explicitly define a `BBox` in default user space, in which case this is used. Otherwise, the bounding box is the smallest rectangle enclosing all of the content items contained by this element (which may take some time to compute). Note that this is currently quite inefficient as it involves interpreting the entire page for every.single.marked.content.section.omg.yikes Note: Elements may span multiple pages! In the case of an element (such as a `Document` for instance) that spans multiple pages, the bounding box cannot exist, and `None` will be returned. If the `page` attribute is `None`, then `bbox` will be `None`. Nrc3\># UH!oRTLdMURv M# g7frU)rGrV)rWrrGs rFrYElement.bbox..s N=4IIX%?%F&.O'9M 3U;rIc`[URRS55nUcg[U[5(a [ U5$[U[ 5(a;SnSn[SU5HupEUbXS:dM[U5nUnM [ U5$[RSU5 g)zAttribute class nameCNrr2zUnrecognizable C property: %r) rrfrrdrrrr$rr)rEclassesr5r6 classnamer8s rF class_nameElement.class_nameOs4::>>#./ ? gy ) )( ( gt $ $FO'/7';# >X%?%i0F&.O(< ' ' 3W=rIc## SUR;aA[URS5n[XRUR5ShvN ggN7f)NK)rfr _make_kidsrGr)rEkidss rFryElement.__iter__bsC $** DJJsO,D!$ 4<<@ @ @  @sA AAAcV[UR[UR545$rU)hashrreprrfrDs rF__hash__Element.__hash__gs!T\\4 #3455rIrvr{)'r|r}r~rrr&rr rrrr classmethodrrrrrPrGr rrxrnrrVrr#r'r+r.r9r?r r:r8ryrrIrrvrIrFr2r2sw  Y "E5d ":J:T#y.-A:i:: -c - -7c77&--eL) .56..%/%% c  *eD$J'**<0uS$Y'003%T *332uS$Y'7220c4i(8009U39-99E$sI~"6"<=8E#t),$A(5K)N#OPA 6#6rIkrGrBrNc#P# [RSU5 SShvN gN7f)a Make a child for this element from its K array. K in Element can be (PDF 1.7 Table 323): - a structure element (not a content item) - an integer marked-content ID - a marked-content reference dictionary - an object reference dictionary - an array of one or more of the above zUnrecognized 'K' element: %rrvN)rrrLrGrNs rFrCrCnsKK.2MMs &$&c#\# UH n[[U5X5ShvN M" gN 7frU)rCr)rLrGrNrs rF_make_kids_listrPs(hrlD9999s ,* ,c#n# Uc[RSU5 g[URUSS9v g7f)N.No page found for marked-content reference: %rr<r=r>)rrr:pagerefrNs rF_make_kids_intrUs1 | DaH t||!D AAs35c## URS5nU[La[XU5ShvN gU[La[ XU5ShvN g[ X S9v gN/N7f)Nrr)r LITERAL_MCR_make_kids_mcr LITERAL_OBJR_make_kids_objrr2)rLrGrNktypes rF_make_kids_dictr\sW EE&ME !!6222 , "1F333f.. 33s!*A AA  A A A c#Z# [URS55nUb[U[5(d[R SU5 gSn[ XU5nUcg[US5n[XSUS9v g![a N[a [R SU5 N:f=f7f)Nrz'MCID' entry is not an int: %rStmz''Stm' entry is not a content stream: %rrS) rrrdrrr_get_kid_pagerefr r TypeErrorr:)rLrGrNr=r>rTs rFrXrXs AEE&M "D |:dC00 4a8)-Fq/GBah' w& AA   B =qABs6AB+A:- B+: B(B+B(%B+'B((B+c## URS5n[U[5(d[R SU5 gUR 5n[U[ [45(d[R SU5 gURS5nU[La8[U[ 5(d[R SU5 g[X$S9v g[XU5nUcg[XdS9v g7f)NObjz3'Obj' entry is not an indirect object reference: %rz9'Obj' entry does not point to a dict or ContentStream: %rrz('Obj' entry does not point to a dict: %rr)r<rf) rrdrrrresolver3rLITERAL_STRUCTELEMr2r_r8)rLrGrNrefrXr[rTs rFrZrZs %%,C c6 " " I1M ++-C cD-0 1 1 OQTU GGFOE ""#t$$ KKBC H f00"1F3 ? W88sC.C0cURS5nUb][U[5(a2[U5nURR UR 5nO[RSU5 UcUc[RSU5 gUR$![a [RSU5 NKf=f)NrrrrR) rrdrr(rrrrrrrrT)rLrGrNrrPs rFr_r_s tB ~ b& ! ! G%f-yyrxx0 KKLa P | < KKH! L << G BAF Gs0B B>=B>rPr3c## [URRS55nUcg[U5n[URS55nUc[R SU5 g[ U[5(aU/nO-[ U[5(aO[R SU5 gUHn[U5n[ U[5(d[R SU5 O]URS5[La[R SU5 O.URS5[La[R SU5 [US[U55ShvN M gN 7f) Nr0rBz5'K' entry in StructTreeRoot could not be resolved: %rz<'K' entry in StructTreeRoot should be dict or list but is %rz3'K' entry in StructTreeRoot contains non-element %rrz8'K' entry in StructTreeRoot contains object reference %rz@'K' entry in StructTreeRoot contains marked content reference %r) rcatalogrrrrrdr3rrYrWrCr*)rProotrDrLs rF_iter_structurerjs' CKKOO$45 6D | d D DHHSM "D | KTR$v D$    JD    QK !T"" KKMq Q UU6]l * KKRTU V UU6]k ) KKRTU a}S'9::: ;sE E."E,# E.c\rSrSr%Sr\\S'\\\ 4\S'Sr \ \\\4S4\S'Sr \ \ S4\S'SrSrSrSrSrSS jrS \\ S \\44S jr\S \\\44S j5r\S \ 4Sj5r\S \4Sj5r\SSj5rSrg)ri a8Logical structure tree. A structure tree can be iterated over in the same fashion as its elements. Note that even though it is forbidden for structure tree root to contain content items, PLAYA is robust to this possibility, thus you should not presume that iterating over it will only yield `Element` instances. The various attributes (role map, class map, pronunciation dictionary, etc, etc) are accessible through `props` but currently have no particular interpretation aside from the role map which is accessible in normalized form through `role_map`. Attributes: props: Structure tree root dictionary (PDF 1.7 table 322). role_map: Mapping of structure element types (as strings) to standard structure types (as strings) (PDF 1.7 section 14.8.4) parent_tree: Parent tree linking marked content sections to structure elements (PDF 1.7 section 14.7.4.4) parent: A structure tree has no parent element, so this is `None` bbox: A structure tree has no bounding box so this is `None` type: This is "StructTreeRoot" role: This is also "StructTreeRoot" rrfN _role_map _parent_treer0rAc^[U5Ul[URS5Ulg)Nr0)r*rrrhrfrErPs rFr Tree.__init__-s$$S)  ,< => rIr2cB[UR5n[U5$rU)r(rrjros rFry Tree.__iter__1sdll+s##rIcrURb UR$0Ul[URRS55n[ U[ 5(aSUR 5H?up#[ U[5(a [U5nO [U5nX@RU'MA UR$)zZDictionary mapping some (not necessarily all) element types to their standard equivalents.RoleMap) rlrrfrrdr3rrrr)rErmrLvrs rFr Tree.role_map5s >> %>> ! djjnnY/ 0 b$   a++'?Dq6D$(q! # ~~rIcURb UR$SUR;a[05UlUR$[URS5UlUR$)aParent tree for this document. This is a somewhat obscure data structure that links marked content sections to their corresponding structure elements. If you don't know what that means, you probably don't need it, but if you do, here it is. Unlike the structure tree itself, if there is no parent tree, this will be an empty `NumberTree`, not `None`. This is because the parent tree is required by the spec in the case where structure elements contain marked content, which is nearly all the time. ParentTree)rmrfrrDs rF parent_treeTree.parent_treeFsf    ($$ $ tzz ) *2D    !+4::l+C DD    rIc[U5$)z+Iterate over all content items in the tree.rrDs rFrx Tree.contents^st$$rIc,[UR5$)z6Document with which this structure tree is associated.rrDs rFrPTree.doccrrI)rrmrlrf)rPr3rANr{)r|r}r~rrr&rr rrrlrrmrrGr rVrrrr r:r8ryrrrzrrxrPrrvrIrFrr s2 Y -1IuT#s(^T)*1,0L% D()0 D F D D D?$(5K)N#OP$$sCx. !Z!!.%/%%--rIrcD\rSrSr%Sr\\\S'\\ \ 4\S'S\ S\SS4Sjr \ SS j5rS\ 4S jr\S \ S\\ S44S j5r\S \SS4S j5rS \\ \4S\S\ S44SjrSS\\\\\S4S\S4SjjrSS\\\\\S4S\S4SjjrSrg) PageStructureiizE Sequence of structural content elements for a page or Form XObject. parentsrrTrANcRUuUlnXl[U5Ul0UlgrU)rNrTrrr)rErTrrOs rFrPageStructure.__init__qs%  Q !'*  rIc,[UR5$)z%Get associated document if it exists.)r(rNrDs rFrPPageStructure.docwst{{++rIc,[UR5$rU)lenrrDs rF__len__PageStructure.__len__|s4<<  rIidxcgrUrvrErs rF __getitem__PageStructure.__getitem__s=@rIcgrUrvrs rFrrs:=rIc P[U[5(a[[UR[ UR [ UR556Vs/sHo RUPM sn5$URUnUcg[U[5(asURUR;a@[RUR[U55URUR'URUR$[R!SU5 gs snf)Nz:ParentTree element is not an indirect object reference: %r)rdslicerrTrangeindicesrrrrrrr2rrPrrr)rErrobjrefs rFrrs c5 ! !  */S=N1O*PQ*PQa*PQ c" >  ' '||4==0.5.?.?HHj0/ fll+==. . KKLf #RsD# rr2c## [U5n[5nUHcnUcMU"U5(aXC;aUv URU5 M1[UR[ 5(aMRURnUbM[Me g7f)a!Search up depth-first for matching elements in the parent tree. The `matcher` argument is either a string, a regular expression, or a function taking a `Element` and returning `True` if the element matches, or `None` (default) to return all descendants in depth-first order. For compatibility with `pdfplumber` and consistent behaviour across documents, names and regular expressions are matched against the `role` attribute. If you wish to match the "raw" structure type from the `type` attribute, you can do this with a matching function. N)rsetaddrdr r)rErrseenelements rFrPageStructure.find_allsr"&g. uG%g&&*% HHW%gnnd33!..%sBAB=BrcX[URU55$![a gf=f)a+Find the first matching element in the parent tree. The `matcher` argument is either a string or a regular expression to be matched against the `role` attribute, or a function taking a `Element` and returning `True` if the element matches, or `None` (default) to just get the first child element. N)rrrrs rFrPageStructure.finds-  g./ /  s  )))rNrrTrr{rU)r|r}r~rrrrrr rr2r'rrrPrrrrrrrrr rrrrvrIrFrris.i  3<  ) ,,!!@s@uWd]';@@ =u===e$ - .4DH)S'#, 4?@) ) )@DHS'#, 4?@  rIrrU)cr functoolsloggingr collectionsr dataclassesr itertoolsroperatorrtypingrrr r r r r rrrrplaya.data_structuresr playa.parserrrrplaya.pdftypesrrrrrrrrrrr playa.utilsr!r"r#r$r% playa.workerr&r'r(r)r* getLoggerr|rrrrrWrYr rdrrrplaya.documentr3rr4r5rar6r7r8rtr:rrrrrr2singledispatchrCregisterrrPrrUr3r\rXrZr_rjrrrvrIrFrsa !    -22    !G i.G %j 6{ -.& i[$& ' f+'+8@ ](./] ]@ TT Tp:> 3 i5 64:>'9o' 3 i5 6'i'$&x&R<hu[-%?@A<~ T6hy+}'L!MNT6 T6n l+5@ eI{M9:;" T: :&:0;: eI{M9:;:: SB B %B/:B kBB T / CN /#(#6 /@K / eK 9:; / /B CNB#(#6B@KB kB*9 CN9#(#69@K9 eM9,-.92 CN#(#6@K 7D=(%; %; eI{M9:;%;P]-8XeI{M$IJK]-@cHU7D=12crI