kh dZddlmZ ddlmZddlmZddlm Z ddl m Z dd l m Z dd l mZGd d e Zy #e$r ddlmZY7wxYw)a1 Text Line objects based on PDF raw dict extracted with ``PyMuPDF``. Data structure of line in text block referring to this `link `_:: { 'bbox': (x0,y0,x1,y1), 'wmode': m, 'dir': [x,y], 'spans': [ spans ] } )Point)Iterable)TextSpan)Element) TextDirection)Spans) ImageSpanceZdZdZddeffd ZedZedZedZ edZ edZ d Z fd Z d Zd efd ZdZdZxZS)Linez)Object representing a line in text block.rawc|i}|jdd|_d|vr2tt|dtj z|_n ddg|_|jdd|_|jdd|_d|vr|jdt|-|t| j|jd g|_y) Nwmoderdir?g line_breaktab_stopbbox)parentspans)getrlistrr pure_rotation_matrixrrrpopsuper__init__r restorer)selfr __class__s N/var/www/teggl/fontify/venv/lib/python3.12/site-packages/pdf2docx/text/Line.pyrz Line.__init__!s ;bWWWa(  C<E#e*-d.G.G.IIJDHSzDH'',2 A.  S=#''&/ $'//0DE ct|jDcgc]}|j}}dj|Scc}w)zIJoining span text. Note image is translated to a placeholder ````.)rtextjoinrspan spans_texts r!r%z Line.text9s3-1JJ7Ddii7 7wwz""8s5c|jDcgc]}t|ts|j!}}dj |Scc}w)z%Joining span text with image ignored.r$)r isinstancerr%r&r's r!raw_textz Line.raw_text@s=-1JJUD*T8:TdiiU Uwwz""Vs AAc|jD]0}t|tsy|jj s0yy)zXIf this line contains only white space or not. If True, this line is safe to be removed.FT)rr+rr%striprr(s r!white_space_onlyzLine.white_space_onlyGs8JJ /DdH-eyy  /r"cBttd|jS)zGet image spans in this Line.c"t|tSN)r+r )r(s r!z"Line.image_spans..Ts D)4r")rfilterrrs r! image_spanszLine.image_spansPs!F 4djj   r"c|jddk(rtjS|jddk(rtjStjS)zGet text direction. Consider ``LEFT_RIGHT`` and ``LEFT_RIGHT`` only. Returns: TextDirection: Text direction of this line. rrrg)rr LEFT_RIGHT BOTTOM_TOPIGNOREr6s r!text_directionzLine.text_directionXsH 88A;#  ++ + XXa[D  ++ + '' 'r"c6|jjS)z.Remove redundant blanks at the begin/end span.)rr.r6s r!r.z Line.stripgszz!!r"c t|}|j|j|j|j |j |jDcgc]}|jc}d|Scc}w)N)rrrrr)rstoreupdaterrrrr)rresr(r s r!r?z Line.storemsdgmo **((//--)-!%     sA7cxt|tr|D]}|j|y|j|y)zAdd span list to current Line. Args: span_or_list (Span, Iterable): TextSpan or TextSpan list to add. N)r+radd_span)r span_or_listr(s r!addzLine.add|s7 lH -$ $ d# $ MM, 'r"r(c:|jj|y)zAdd span to current Line.N)rappendr/s r!rCz Line.add_spans $r"c|j|jr|jStd|ji}|j |_|j D]$}|j|}|j|&|S)zCreate new Line object with spans contained in given bbox. Args: rect (fitz.Rect): Target bbox. Returns: Line: The created Line instance. r) containsrcopyr rrr intersectsrE)rrectliner(contained_spans r!rKzLine.intersectssv == #99; Wdjj)*88JJ %D!__T2N HH^ $ % r"c |jr8t|jD] }|jj"|jD]}|j ||j r|jdyy)z0Create docx line, i.e. a run in ``python-docx``. N)rrangeadd_runadd_tabr make_docxr)rp_r(s r!rTzLine.make_docxsa ==4==) @199;+>+>+@ @JJ1Dq 11 ??AIIdO?r"r3)__name__ __module__ __qualname____doc__dictrpropertyr%r,r0r7r<r.r?rErrCrKrT __classcell__)r s@r!r r s3F4F0## ##    ( ("  ( G 2 ,r"r N)rZfitzr collectionsr ImportErrorcollections.abcrcommon.Elementr common.sharer r image.ImageSpanr r r"r!rfsJ )$$('S,7S,)()s: AA