path: root/kviewshell/plugins/djvu/libdjvu/DjVuText.cpp

//C-  -*- C++ -*-
//C- -------------------------------------------------------------------
//C- DjVuLibre-3.5
//C- Copyright (c) 2002  Leon Bottou and Yann Le Cun.
//C- Copyright (c) 2001  AT&T
//C-
//C- This software is subject to, and may be distributed under, the
//C- GNU General Public License, Version 2. The license should have
//C- accompanied the software or you may obtain a copy of the license
//C- from the Free Software Foundation at http://www.fsf.org .
//C-
//C- This program is distributed in the hope that it will be useful,
//C- but WITHOUT ANY WARRANTY; without even the implied warranty of
//C- MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
//C- GNU General Public License for more details.
//C- 
//C- DjVuLibre-3.5 is derived from the DjVu(r) Reference Library
//C- distributed by Lizardtech Software.  On July 19th 2002, Lizardtech 
//C- Software authorized us to replace the original DjVu(r) Reference 
//C- Library notice by the following text (see doc/lizard2002.djvu):
//C-
//C-  ------------------------------------------------------------------
//C- | DjVu (r) Reference Library (v. 3.5)
//C- | Copyright (c) 1999-2001 LizardTech, Inc. All Rights Reserved.
//C- | The DjVu Reference Library is protected by U.S. Pat. No.
//C- | 6,058,214 and patents pending.
//C- |
//C- | This software is subject to, and may be distributed under, the
//C- | GNU General Public License, Version 2. The license should have
//C- | accompanied the software or you may obtain a copy of the license
//C- | from the Free Software Foundation at http://www.fsf.org .
//C- |
//C- | The computer code originally released by LizardTech under this
//C- | license and unmodified by other parties is deemed "the LIZARDTECH
//C- | ORIGINAL CODE."  Subject to any third party intellectual property
//C- | claims, LizardTech grants recipient a worldwide, royalty-free, 
//C- | non-exclusive license to make, use, sell, or otherwise dispose of 
//C- | the LIZARDTECH ORIGINAL CODE or of programs derived from the 
//C- | LIZARDTECH ORIGINAL CODE in compliance with the terms of the GNU 
//C- | General Public License.   This grant only confers the right to 
//C- | infringe patent claims underlying the LIZARDTECH ORIGINAL CODE to 
//C- | the extent such infringement is reasonably necessary to enable 
//C- | recipient to make, have made, practice, sell, or otherwise dispose 
//C- | of the LIZARDTECH ORIGINAL CODE (or portions thereof) and not to 
//C- | any greater extent that may be necessary to utilize further 
//C- | modifications or combinations.
//C- |
//C- | The LIZARDTECH ORIGINAL CODE is provided "AS IS" WITHOUT WARRANTY
//C- | OF ANY KIND, EITHER EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED
//C- | TO ANY WARRANTY OF NON-INFRINGEMENT, OR ANY IMPLIED WARRANTY OF
//C- | MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE.
//C- +------------------------------------------------------------------
// 
// $Id: DjVuText.cpp,v 1.10 2004/07/07 19:23:36 leonb Exp $
// $Name: release_3_5_15 $

#ifdef HAVE_CONFIG_H
# include "config.h"
#endif
#if NEED_GNUG_PRAGMAS
# pragma implementation
#endif

#include "DjVuText.h"
#include "IFFByteStream.h"
#include "BSByteStream.h"
#include "debug.h"
#include <ctype.h>



#ifdef HAVE_NAMESPACES
namespace DJVU {
# ifdef NOT_DEFINED // Just to fool emacs c++ mode
}
#endif
#endif



#ifdef min
#undef min
#endif
template<class TYPE>
static inline TYPE min(TYPE a,TYPE b) { return (a<b)?a:b; }

//***************************************************************************
//******************************** DjVuTXT **********************************
//***************************************************************************

const char DjVuTXT::end_of_column    = 013;      // VT: Vertical Tab
const char DjVuTXT::end_of_region    = 035;      // GS: Group Separator
const char DjVuTXT::end_of_paragraph = 037;      // US: Unit Separator
const char DjVuTXT::end_of_line      = 012;      // LF: Line Feed

const int DjVuTXT::Zone::version  = 1;

DjVuTXT::Zone::Zone()
  : ztype(DjVuTXT::PAGE), text_start(0), text_length(0), zone_parent(0)
{
}

DjVuTXT::Zone *
DjVuTXT::Zone::append_child()
{
  Zone empty;
  empty.ztype = ztype;
  empty.text_start = 0;
  empty.text_length = 0;
  empty.zone_parent=this;
  children.append(empty);
  return & children[children.lastpos()];
}

void
DjVuTXT::Zone::cleartext()
{
  text_start = 0;
  text_length = 0;
  for (GPosition i=children; i; ++i)
    children[i].cleartext();
}

void
DjVuTXT::Zone::normtext(const char *instr, GUTF8String &outstr)
{
  if (text_length == 0)
    {
      // Descend collecting text below
      text_start = outstr.length();
      for (GPosition i=children; i; ++i)
        children[i].normtext(instr, outstr);
      text_length = outstr.length() - text_start;
      // Ignore empty zones
      if (text_length == 0)
        return;
    }
  else
    {
      // Collect text at this level
      int new_start = outstr.length();
      outstr = outstr + GUTF8String(instr+text_start, text_length);
      text_start = new_start;
      // Clear textual information on lower level nodes
      for (GPosition i=children; i; ++i)
        children[i].cleartext();
    }
  // Determine standard separator
  char sep;
  switch (ztype)
    {
    case COLUMN:
      sep = end_of_column; break;
    case REGION:
      sep = end_of_region; break;
    case PARAGRAPH: 
      sep = end_of_paragraph; break;
    case LINE:
      sep = end_of_line; break;
    case WORD:
      sep = ' '; break;
    default:
      return;
    }
  // Add separator if not present yet.
  if (outstr[text_start+text_length-1] != sep)
    {
      outstr = outstr + GUTF8String(&sep, 1);
      text_length += 1;
    }
}

unsigned int 
DjVuTXT::Zone::memuse() const
{
  int memuse = sizeof(*this);
  for (GPosition i=children; i; ++i)
    memuse += children[i].memuse();
  return memuse;
}


#ifndef NEED_DECODER_ONLY
void 
DjVuTXT::Zone::encode(
  const GP<ByteStream> &gbs, const Zone * parent, const Zone * prev) const
{
  ByteStream &bs=*gbs;
  // Encode type
  bs.write8(ztype);
  
  // Modify text_start and bounding rectangle based on the context
  // (whether there is a previous non-zero same-level-child or parent)
  int start=text_start;
  int x=rect.xmin, y=rect.ymin;
  int width=rect.width(), height=rect.height();
  if (prev)
  {
    if (ztype==PAGE || ztype==PARAGRAPH || ztype==LINE)
    {
      // Encode offset from the lower left corner of the previous
      // child in the coord system in that corner with x to the
      // right and y down
      x=x-prev->rect.xmin;
      y=prev->rect.ymin-(y+height);
    } else // Either COLUMN or WORD or CHARACTER
    {
      // Encode offset from the lower right corner of the previous
      // child in the coord system in that corner with x to the
      // right and y up
      x=x-prev->rect.xmax;
      y=y-prev->rect.ymin;
    }
    start-=prev->text_start+prev->text_length;
  } else if (parent)
  {
    // Encode offset from the upper left corner of the parent
    // in the coord system in that corner with x to the right and y down
    x=x-parent->rect.xmin;
    y=parent->rect.ymax-(y+height);
    start-=parent->text_start;
  }
  // Encode rectangle
  bs.write16(0x8000+x);
  bs.write16(0x8000+y);
  bs.write16(0x8000+width);
  bs.write16(0x8000+height);
  // Encode text info
  bs.write16(0x8000+start);
  bs.write24(text_length);
  // Encode number of children
  bs.write24(children.size());
  
  const Zone * prev_child=0;
  // Encode all children
  for (GPosition i=children; i; ++i)
  {
    children[i].encode(gbs, this, prev_child);
    prev_child=&children[i];
  }
}
#endif

void 
DjVuTXT::Zone::decode(const GP<ByteStream> &gbs, int maxtext,
		      const Zone * parent, const Zone * prev)
{
  ByteStream &bs=*gbs;
  // Decode type
  ztype = (ZoneType) bs.read8();
  if ( ztype<PAGE || ztype>CHARACTER )
    G_THROW( ERR_MSG("DjVuText.corrupt_text") );

  // Decode coordinates
  int x=(int) bs.read16()-0x8000;
  int y=(int) bs.read16()-0x8000;
  int width=(int) bs.read16()-0x8000;
  int height=(int) bs.read16()-0x8000;

  // Decode text info
  text_start = (int) bs.read16()-0x8000;
//  int start=text_start;
  text_length = bs.read24();
  if (prev)
  {
    if (ztype==PAGE || ztype==PARAGRAPH || ztype==LINE)
    {
      x=x+prev->rect.xmin;
      y=prev->rect.ymin-(y+height);
    } else // Either COLUMN or WORD or CHARACTER
    {
      x=x+prev->rect.xmax;
      y=y+prev->rect.ymin;
    }
    text_start+=prev->text_start+prev->text_length;
  } else if (parent)
  {
    x=x+parent->rect.xmin;
    y=parent->rect.ymax-(y+height);
    text_start+=parent->text_start;
  }
  rect=GRect(x, y, width, height);
  // Get children size
  int size = bs.read24();

  // Checks
  if (rect.isempty() || text_start<0 || text_start+text_length>maxtext )
    G_THROW( ERR_MSG("DjVuText.corrupt_text") );

  // Process children
  const Zone * prev_child=0;
  children.empty();
  while (size-- > 0) 
  {
    Zone *z = append_child();
    z->decode(gbs, maxtext, this, prev_child);
    prev_child=z;
  }
}

void 
DjVuTXT::normalize_text()
{
  GUTF8String newtextUTF8;
  page_zone.normtext( (const char*)textUTF8, newtextUTF8 );
  textUTF8 = newtextUTF8;
}

int 
DjVuTXT::has_valid_zones() const
{
  if (!textUTF8)
    return false;
  if (page_zone.children.isempty() || page_zone.rect.isempty()) 
    return false;
  return true;
}


#ifndef NEED_DECODER_ONLY
void 
DjVuTXT::encode(const GP<ByteStream> &gbs) const
{
  ByteStream &bs=*gbs;
  if (! textUTF8 )
    G_THROW( ERR_MSG("DjVuText.no_text") );
  // Encode text
  int textsize = textUTF8.length();
  bs.write24( textsize );
  bs.writall( (void*)(const char*)textUTF8, textsize );
  // Encode zones
  if (has_valid_zones())
  {
    bs.write8(Zone::version);
    page_zone.encode(gbs);
  }
}
#endif

void 
DjVuTXT::decode(const GP<ByteStream> &gbs)
{
  ByteStream &bs=*gbs;
  // Read text
  textUTF8.empty();
  int textsize = bs.read24();
  char *buffer = textUTF8.getbuf(textsize);
  int readsize = bs.read(buffer,textsize);
  buffer[readsize] = 0;
  if (readsize < textsize)
    G_THROW( ERR_MSG("DjVuText.corrupt_chunk") );
  // Try reading zones
  unsigned char version;
  if ( bs.read( (void*) &version, 1 ) == 1) 
  {
    if (version != Zone::version)
      G_THROW( ERR_MSG("DjVuText.bad_version") "\t" + GUTF8String(version) );
    page_zone.decode(gbs, textsize);
  }
}

GP<DjVuTXT> 
DjVuTXT::copy(void) const
{
  return new DjVuTXT(*this);
}


static inline bool
intersects_zone(GRect box, const GRect &zone)
{
  return
    ((box.xmin < zone.xmin)
      ?(box.xmax >= zone.xmin)
      :(box.xmin <= zone.xmax))
    &&((box.ymin < zone.ymin)
      ?(box.ymax >= zone.ymin)
      :(box.ymin <= zone.ymax));
}

void
DjVuTXT::Zone::get_text_with_rect(const GRect &box, 
                                  int &string_start, int &string_end) const
{
  GPosition pos=children;
  if(pos?box.contains(rect):intersects_zone(box,rect))
  {
    const int text_end=text_start+text_length;
    if(string_start == string_end)
    {
      string_start=text_start;
      string_end=text_end;
    }else
    {
      if (string_end < text_end)
        string_end=text_end;
      if(text_start < string_start)
        string_start=text_start;
    }
  }else if(pos&&intersects_zone(box,rect))
  {
    do
    {
      children[pos].get_text_with_rect(box,string_start,string_end);
    } while(++pos);
  }
}

void
DjVuTXT::Zone::find_zones(GList<Zone *> &list, 
                          const int string_start, const int string_end) const
{
  const int text_end=text_start+text_length;
  if(text_start >= string_start)
    {
      if(text_end <= string_end)
        {
          list.append(const_cast<Zone *>(this));
        }
      else if(text_start < string_end)
        {
          if (children.size())
            for (GPosition pos=children; pos; ++pos)
              children[pos].find_zones(list,string_start,string_end);
          else
            list.append(const_cast<Zone *>(this));
        }
    }
  else if( text_end > string_start)
    {
      if (children.size())
        for (GPosition pos=children; pos; ++pos)
          children[pos].find_zones(list,string_start,string_end);
      else
        list.append(const_cast<Zone *>(this));
    }
}

void
DjVuTXT::Zone::get_smallest(GList<GRect> &list) const
{
  GPosition pos=children;
  if(pos)
    {
      do {
        children[pos].get_smallest(list);
      } while (++pos);
    }
  else
    {
      list.append(rect);
    }
}

void
DjVuTXT::Zone::get_smallest(GList<GRect> &list, const int padding) const
{
  GPosition pos=children;
  if(pos)
    {
      do {
        children[pos].get_smallest(list,padding);
      } while (++pos);
    }
  else if(zone_parent && zone_parent->ztype >= PARAGRAPH)
    {
      const GRect &xrect=zone_parent->rect;
      if(xrect.height() < xrect.width())
        {
          list.append(GRect(rect.xmin-padding,xrect.ymin-padding,rect.width()
                            +2*padding,xrect.height()+2*padding));
        }
      else
        {
          list.append(GRect(xrect.xmin-padding,rect.ymin-padding,xrect.width()
                            +2*padding,rect.height()+2*padding));
        }
    }
  else
    {
      list.append(GRect(rect.xmin-padding,rect.ymin-padding,rect.width()
                        +2*padding,rect.height()+2*padding));
    }
}

void
DjVuTXT::get_zones(int zone_type, const Zone *parent, 
                   GList<Zone *> & zone_list) const 
   // get all the zones of  type zone_type under zone node parent
{
   // search all branches under parent
   const Zone *zone=parent;
   for( int cur_ztype=zone->ztype; cur_ztype<zone_type; ++cur_ztype )
   {
      GPosition pos;
      for(pos=zone->children; pos; ++pos)
      {
	 Zone *zcur=(Zone *)&zone->children[pos];
	 if ( zcur->ztype == zone_type )
	 {
	    GPosition zpos=zone_list;
	    if ( !zone_list.search(zcur,zpos) )
	       zone_list.append(zcur);
	 }
	 else if ( zone->children[pos].ztype < zone_type )
	    get_zones(zone_type, &zone->children[pos], zone_list);
      }
   }
}

GList<GRect>
DjVuTXT::find_text_with_rect(const GRect &box, GUTF8String &text, 
                             const int padding) const
{
  GList<GRect> retval;
  int text_start=0;
  int text_end=0;
  page_zone.get_text_with_rect(box,text_start,text_end);
  if(text_start != text_end)
  {
    GList<Zone *> zones;
    page_zone.find_zones(zones,text_start,text_end);
    GPosition pos=zones;
    if(pos)
    {
      do
      {
        if(padding >= 0)
        {
          zones[pos]->get_smallest(retval,padding);
        }else
        {
          zones[pos]->get_smallest(retval);
        }
      } while(++pos);
    }
  }
  text=textUTF8.substr(text_start,text_end-text_start);
  return retval;
}


GList<DjVuTXT::Zone *>
DjVuTXT::find_text_in_rect(GRect target_rect, GUTF8String &text) const
   // returns a list of zones of type WORD in the nearest/selected paragraph 
{
   GList<Zone *> zone_list;
   GList<Zone *> lines;

   get_zones((int)PARAGRAPH, &page_zone, zone_list);
   // it's possible that no paragraph structure exists for reasons that  
   // 1) ocr engine is not capable 2) file was modified by user. In such case, 
   // we can only make a rough guess, i.e., select all the lines intersected with
   // target_rect
   if (zone_list.isempty())
   {
      get_zones((int)LINE, &page_zone, zone_list);
      GPosition pos;
      for(pos=zone_list; pos; ++pos)
      {
	 GRect rect=zone_list[pos]->rect;
	 int h0=rect.height()/2;
	 if(rect.intersect(rect,target_rect) && rect.height()>h0)
	    lines.append(zone_list[pos]);
      }
   } else 
   {
      GPosition pos, pos_sel=zone_list;
      float ar=0;
      for(pos=zone_list; pos; ++pos)
      {
	 GRect rect=zone_list[pos]->rect;
	 int area=rect.area();
	 if (rect.intersect(rect, target_rect))
	 {
	    float ftmp=rect.area()/(float)area;
	    if ( !ar || ar<ftmp )
	    {
	       ar=ftmp;
	       pos_sel=pos;
	    }
	 }
      }
      Zone *parag = 0;
      if ( ar>0 ) parag=zone_list[pos_sel];
      zone_list.empty();
      if ( ar>0 ) 
      {
	 get_zones((int)LINE, parag, zone_list);
	 if ( !zone_list.isempty() )
	 {
	    for(GPosition pos=zone_list; pos; ++pos)
	    {
	       GRect rect=zone_list[pos]->rect;
	       int h0=rect.height()/2;
	       if(rect.intersect(rect,target_rect) && rect.height()>h0)
		  lines.append(zone_list[pos]);
	    }
	 }
      }
   }

   zone_list.empty();
   if (!lines.isempty()) 
   {
      int i=1, lsize=lines.size();

      GList<Zone *> words;
      for (GPosition pos=lines; pos; ++pos, ++i)
      {
	 words.empty();
	 get_zones((int)WORD, lines[pos], words);

	 if ( lsize==1 )
	 {
	    for(GPosition p=words;p;++p)
	    {
	       GRect rect=words[p]->rect;
	       if(rect.intersect(rect,target_rect))
	       //if (target_rect.contains(words[p]->rect))
		  zone_list.append(words[p]);
	    }
	 } else
	 {
	    if (i==1)
	    {
	       bool start=true;
	       for(GPosition p=words; p; ++p)
	       {
		  if ( start )
		  {
		     GRect rect=words[p]->rect;
		     if(rect.intersect(rect,target_rect))
			//if (target_rect.contains(words[p]->rect))
		     {
			start=false;
			zone_list.append(words[p]);
		     }
		  } else 
		     zone_list.append(words[p]);
	       }
	    } else if (i==lsize)
	    {
	       bool end=true;
	       for(GPosition p=words.lastpos();p;--p)
	       {
		  if ( end )
		  {
		     GRect rect=words[p]->rect;
		     if(rect.intersect(rect,target_rect))
			//if(target_rect.contains(words[p]->rect) )
		     {
			end=false;
			zone_list.append(words[p]);
		     }
		  } else 
		     zone_list.append(words[p]);
	       }
	    }

	    if (i!=1 && i!=lsize )
	    {
	       for(GPosition p=words;p;++p)
		  zone_list.append(words[p]);
	    }
	 }
      }
   } 

   return zone_list;
}

unsigned int 
DjVuTXT::get_memory_usage() const
{
  return sizeof(*this) + textUTF8.length() + page_zone.memuse() - sizeof(page_zone); 
}



//***************************************************************************
//******************************** DjVuText *********************************
//***************************************************************************

void
DjVuText::decode(const GP<ByteStream> &gbs)
{
  GUTF8String chkid;
  GP<IFFByteStream> giff=IFFByteStream::create(gbs);
  IFFByteStream &iff=*giff;
  while( iff.get_chunk(chkid) )
  {
    if (chkid == "TXTa")
    {
      if (txt)
        G_THROW( ERR_MSG("DjVuText.dupl_text") );
      txt = DjVuTXT::create();
      txt->decode(iff.get_bytestream());
    }
    else if (chkid == "TXTz")
    {
      if (txt)
        G_THROW( ERR_MSG("DjVuText.dupl_text") );
      txt = DjVuTXT::create();
      const GP<ByteStream> gbsiff=BSByteStream::create(iff.get_bytestream());
      txt->decode(gbsiff);
    }
    // Add decoding of other chunks here
    iff.close_chunk();
  }
}

void
DjVuText::encode(const GP<ByteStream> &gbs)
{
  if (txt)
  {
    const GP<IFFByteStream> giff=IFFByteStream::create(gbs);
    IFFByteStream &iff=*giff;
    iff.put_chunk("TXTz");
    {
      GP<ByteStream> gbsiff=BSByteStream::create(iff.get_bytestream(),50);
      txt->encode(gbsiff);
    }
    iff.close_chunk();
  }
  // Add encoding of other chunks here
}


GP<DjVuText>
DjVuText::copy(void) const
{
   GP<DjVuText> text= new DjVuText;
      // Copy any primitives (if any)
   *text=*this;
      // Copy each substructure
   if (txt)
     text->txt = txt->copy();
   return text;
}

static GUTF8String
indent ( int spaces)
{
  GUTF8String ret;
  for( int i = 0 ; i < spaces ; i++ )
    ret += ' ';
  return ret;
}

static const char *tags[8]=
{ 0,
  "HIDDENTEXT",
  "PAGECOLUMN",
  "REGION",
  "PARAGRAPH",
  "LINE",
  "WORD",
  "CHARACTER" };
static const int tags_size=sizeof(tags)/sizeof(const char *);

static GUTF8String
start_tag(const DjVuTXT::ZoneType zone)
{
  GUTF8String retval;
  if((tags_size > (int)zone)&&((int)zone > 0))
  {
    switch (zone)
    {
      case DjVuTXT::CHARACTER:
        retval="<"+GUTF8String(tags[zone])+">";
        break;
      case DjVuTXT::WORD:
        retval=indent(2*(int)zone+2)+"<"+tags[zone]+">";
        break;
      default:
        retval=indent(2*(int)zone+2)+"<"+tags[zone]+">\n";
        break;
    }
  }
  return retval;
}

static GUTF8String
start_tag(const DjVuTXT::ZoneType zone, const GUTF8String &attributes)
{
  GUTF8String retval;
  if((tags_size > (int)zone)&&((int)zone > 0))
  {
    switch (zone)
    {
      case DjVuTXT::CHARACTER:
        retval="<"+GUTF8String(tags[zone])+" "+attributes+">";
        break;
      case DjVuTXT::WORD:
        retval=indent(2*(int)zone+2)+"<"+tags[zone]+" "+attributes+">";
        break;
      default:
        retval=indent(2*(int)zone+2)+"<"+tags[zone]+" "+attributes+">\n";
        break;
    }
  }
  return retval;
}

static inline GUTF8String
start_tag(const int layer)
{
  return start_tag((const DjVuTXT::ZoneType)layer);
}


static GUTF8String
end_tag(const DjVuTXT::ZoneType zone)
{
  GUTF8String retval;
  if((tags_size > (int)zone)&&((int)zone >= 0))
  {
    switch (zone)
    {
      case DjVuTXT::CHARACTER:
        retval="</"+GUTF8String(tags[zone])+">";
        break;
      case DjVuTXT::WORD:
        retval="</"+GUTF8String(tags[zone])+">\n";
        break;
      default:
        retval=indent(2*(int)zone+2)+"</"+tags[zone]+">\n";
        break;
    }
  }
  return retval;
}

static inline GUTF8String
end_tag(const int layer)
{
  return end_tag((const DjVuTXT::ZoneType)layer);
}

static GUTF8String
tolayer(int &layer, const DjVuTXT::ZoneType next_layer)
{
  GUTF8String retval;
  for( ;layer < (int)next_layer;layer++ )
  {
    retval+=start_tag(layer);
  }
  while (layer > (int)next_layer )
  {
    retval+=end_tag(--layer);
  }
  return retval;
}

static void
writeText( ByteStream & str_out,
            const GUTF8String &textUTF8,
            const DjVuTXT::Zone &zone,
            const int WindowHeight );

static void
writeText( ByteStream & str_out,
           const GUTF8String &textUTF8,
           const DjVuTXT::ZoneType zlayer,
           const GList<DjVuTXT::Zone> &children,
           const int WindowHeight )
{
//  assert( txt->has_valid_zones() );
//  DEBUG_MSG( "--zonetype=" << txt->page_zone.ztype << "\n" );

  //  Beginning tags for missing layers
  int layer=(int)zlayer;
  //  Output the next layer
  for(GPosition pos=children ; pos ; ++pos )
  {
    str_out.writestring(tolayer(layer,children[pos].ztype));
    writeText( str_out,
                textUTF8,
                children[pos],
                WindowHeight );
  }
  str_out.writestring(tolayer(layer,zlayer));
}

static void
writeText( ByteStream & str_out,
            const GUTF8String &textUTF8,
            const DjVuTXT::Zone &zone,
            const int WindowHeight )
{
//  DEBUG_MSG( "--zonetype=" << zone.ztype << "\n" );

  const GUTF8String xindent(indent( 2 * zone.ztype + 2 ));
  GPosition pos=zone.children;
  // Build attribute string
  if( ! pos )
  {
    GUTF8String coords;
    coords.format("coords=\"%d,%d,%d,%d\"",
      zone.rect.xmin, WindowHeight - 1 - zone.rect.ymin,
      zone.rect.xmax, WindowHeight - 1 - zone.rect.ymax);
    const int start=zone.text_start;
    const int end=textUTF8.firstEndSpace(start,zone.text_length);
    str_out.writestring(start_tag(zone.ztype,coords));
    str_out.writestring(textUTF8.substr(start,end-start).toEscaped());
    str_out.writestring(end_tag(zone.ztype));
  } else
  {
    writeText(str_out,textUTF8,zone.ztype,zone.children,WindowHeight);
  }
}

void
DjVuTXT::writeText(ByteStream &str_out,const int height) const
{
  if(has_valid_zones())
  {
    ::writeText(str_out,textUTF8,DjVuTXT::PAGE,page_zone.children,height);
  }else
  {
    str_out.writestring(start_tag(DjVuTXT::PAGE));
    str_out.writestring(end_tag(DjVuTXT::PAGE));
  }
}

void
DjVuText::writeText(ByteStream &str_out,const int height) const
{
  if(txt)
  {
    txt->writeText(str_out,height);
  }else
  {
    str_out.writestring("<"+GUTF8String(tags[DjVuTXT::PAGE])+"/>\n");
  }
   
}
GUTF8String
DjVuTXT::get_xmlText(const int height) const
{
  GP<ByteStream> gbs(ByteStream::create());
  ByteStream &bs=*gbs;
  writeText(bs,height);
  bs.seek(0L);
  return bs.getAsUTF8();
}

GUTF8String
DjVuText::get_xmlText(const int height) const
{
  GUTF8String retval;
  if(txt)
  {
    retval=txt->get_xmlText(height);
  }else
  {
    retval="<"+GUTF8String(tags[DjVuTXT::PAGE])+"/>\n";
  }
  return retval;
}


#ifdef HAVE_NAMESPACES
}
# ifndef NOT_USING_DJVU_NAMESPACE
using namespace DJVU;
# endif
#endif