//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, either Version 2 of the license,
//C- or (at your option) any later version. 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 from
//C- Lizardtech Software.  Lizardtech Software has authorized us to
//C- replace the original DjVu(r) Reference Library notice by the following
//C- text (see doc/lizard2002.djvu and doc/lizardtech2007.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, either Version 2 of the license,
//C- | or (at your option) any later version. 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- +------------------------------------------------------------------

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

/** @name csepdjvu

    {\bf Synopsis}
    \begin{verbatim}
        csepdjvu <....options-or-separated_files...> <outputdjvufile>
    \end{verbatim}

    {\bf Description}
    
    File #"csepdjvu.cpp"# demonstrates a complete back-end encoder that takes
    {\em separated files} as input and produces a djvu file as output.
    Each {\em separated file} contains a concatenation of pages. 
    Each page contains the following components:
    \begin{itemize}
    \item A run-length encoded file representing the foreground.
          Two kind of run-length encoded files are accepted.
          The Black-And-White RLE format is described in section
          \Ref{PNM and RLE file formats}.  The Color RLE format
          is described below.
    \item An optional PPM image representing the background.
          The size (width and height) of the PPM image must be obtained by
          rounding up the quotient of the foreground image size by
          an integer reduction factor ranging from 1 to 12.
    \item An arbitrary number of comment lines starting with 
          character '#'.
    \end{itemize}
    All the provided pages will be converted to Compound DjVu Images.
    Foreground colors will be encoded using a single solid color per component
    (see \Ref{DjVu Image files}).  Multiple pages will be gathered into a
    single bundled file.  Use \Ref{djvmcvt} or \Ref{djvujoin} to create an
    indirect file.
    
    {\bf Options}

    \begin{description}
    \item[-d n] Resolution written into the output file (default: 300).
    \item[-q <spec>] Quality level for background (default: 72+11+10+10).
                     See option #"-slice"# in program \Ref{c44} for details.
    \item[-v] Displays a brief message per page.
    \item[-vv] Displays lots of additional messages.
    \end{description}
          
    {\bf Color RLE Images}

    The Color RLE file format is a simple run-length encoding scheme for color
    images with a limited number of colors. Color RLE files always begin with
    a text header composed of:
    - the two characters #"R6"#,
    - the number of columns in decimal,\\
    - the number of rows in decimal,\\
    - the number of palette entries in decimal.\\
    These four items are separated by blank characters (space, tab, cr, or nl)
    or by comment lines introduced by character ``\#''.  The last number is
    followed by exactly one character (usually a nl character).  This header
    is followed by a palette containing three bytes per color.  The bytes
    represent the red, green, and blue components of the color.  

    The palette is followed by four bytes integers (MSB first) representing
    runs.  The twelve upper bits of this integer indicate the index of the run
    color in the palette entry.  The twenty lower bits of the integer indicate
    the run length.  Color indices greater than 0xff0 are reserved for pixels
    belonging to the background layer.  Color index 0xfff is used for
    transparent runs. Color index 0xffe is used for don't-care runs
    (i.e. pixels whose values should be taken by smoothly completing the
    background using the wavelet masking algorithm).  Each row is represented
    by a sequence of runs whose lengths add up to the image width.  Rows are
    encoded starting with the top row and progressing towards the bottom row.

    @memo
    Creates DjVu files from Separated files.
    @author
    L\'eon Bottou <leonb@research.att.com>
*/
//@{
//@}


#include "DjVuGlobal.h"
#include "GException.h"
#include "GSmartPointer.h"
#include "GContainer.h"
#include "ByteStream.h"
#include "IFFByteStream.h"
#include "GRect.h"
#include "GBitmap.h"
#include "JB2Image.h"
#include "DjVuPalette.h"
#include "IW44Image.h"
#include "DjVuInfo.h"
#include "DjVmDoc.h"
#include "DjVmNav.h"
#include "GOS.h"
#include "GURL.h"
#include "DjVuMessage.h"
#include "DjVuText.h"
#include "BSByteStream.h"

#include "miniexp.h"
#include "jb2tune.h"
#include "common.h"

#undef MIN
#undef MAX
inline int MIN(int a, int b) { return ( a<b ?a :b); }
inline int MAX(int a, int b) { return ( a>b ?a :b); }



// --------------------------------------------------
// OPTIONS
// --------------------------------------------------


struct csepdjvuopts
{
  int dpi;                  // resolution
  int verbose;              // verbosity level
  DjVuTXT::ZoneType text;   // level of text detail
  unsigned char slice[16];  // background quality spec
  csepdjvuopts();
};

csepdjvuopts::csepdjvuopts()
{
      dpi = 300;
      verbose = 0;
      text = DjVuTXT::WORD;
      slice[0] =  72;
      slice[1] =  83;      
      slice[2] =  93;
      slice[3] = 103;
      slice[4] =   0;
}


// --------------------------------------------------
// BUFFERED BYTESTREAM
// --------------------------------------------------

// -- A bytestream that performs buffering and 
//    offers a stdio-like interface for parsing files.

class BufferByteStream : public ByteStream 
{
public:
  enum {bufsize=512};
private:
  ByteStream &bs;
  unsigned char buffer[bufsize];
  int bufpos;
  int bufend;
public:
  BufferByteStream(ByteStream &lbs);
  size_t read(void *buffer, size_t size);
  size_t write(const void *buffer, size_t size);
  virtual long tell(void) const;
  int eof(void);
  int unget(int c);
  inline int get(void);
  // parsing helpers
  bool skip(const char *s = " \t\n\r");
  bool expect(int &c, const char *s); 
  bool read_integer(int &x);
  bool read_pair(int &x, int &y);
  bool read_geometry(GRect &r);
  bool read_ps_string(GUTF8String &s);
};

BufferByteStream::BufferByteStream(ByteStream &bs)
  : bs(bs), bufpos(1), bufend(1)
{ 
}

int 
BufferByteStream::eof(void) // aka. feof
{
  if (bufpos < bufend) 
    return false;
  bufend = bufpos = 1;
  bufend += bs.read(buffer+bufend, bufsize-bufend);
  return (bufend == bufpos);
}

size_t 
BufferByteStream::read(void *buf, size_t size)
{
  if (size < 1)
    return 0;
  if (bufend == bufpos) 
    {
      if (size >= bufsize)
        return bs.read(buf, size);
      if (eof())
        return 0;
    }
  if (bufpos + (int)size > bufend)
    size = bufend - bufpos;
  memcpy(buf, buffer+bufpos, size);
  bufpos += size;
  return size;
}

size_t 
BufferByteStream::write(const void *, size_t )
{
  G_THROW("Cannot write into a BufferByteStream");
  return 0;
}

long 
BufferByteStream::tell(void) const
{ 
  return bs.tell() + bufpos - bufend; 
}
    
inline int 
BufferByteStream::get(void) // aka. getc()
{
  if (bufpos < bufend || !eof())
    return buffer[bufpos++];
  return EOF;
}

int  
BufferByteStream::unget(int c) // aka. ungetc()
{
  if (bufpos > 0 && c != EOF) 
    return buffer[--bufpos] = (unsigned char)c;
  return EOF;
}

bool 
BufferByteStream::expect(int &c, const char *s)
{
  c = get();
  if (strchr(s, c))
    return true;
  unget(c);
  return false;
}

bool
BufferByteStream::skip(const char *s)
{
  int c;
  while (expect(c, s)) { }
  return true;
}

bool 
BufferByteStream::read_integer(int &x)
{
  x = 0;
  int c = get();
  if (c<'0' || c>'9')
    return false;
  while (c>='0' && c<='9') 
    {
      x = x*10 + c - '0';
      c = get();
    }
  unget(c);
  return true;
}

bool 
BufferByteStream::read_pair(int &x, int &y)
{
  int c;
  x = y = 0;
  expect(c, "-");
  if (! read_integer(x)) 
    return false;
  if (c == '-') 
    x = -x;
  if (! expect(c, ":"))
    return false;
  expect(c, "-");
  if (! read_integer(y)) 
    return false;
  if (c == '-')
    y = -y;
  return true;
}

bool 
BufferByteStream::read_geometry(GRect &r)
{
  int x,y,w,h,c;
  x = y = w = h = 0;
  if (read_integer(w) && expect(c, "x") && read_integer(h))
    {
      if (expect(c,"+-"))
        {
          if (c == '+')
            expect(c,"-");
          if (! read_integer(x))
            return false;
          if (c == '-')
            x = -x;
        }
      if (expect(c,"+-"))
        {
          if (c == '+')
            expect(c,"-");
          if (! read_integer(y))
            return false;
          if (c == '-')
            y = -y;
        }
      r = GRect(x,y,w,h);
      return true;
    }
  return false;
}

static void
add_to_string(GUTF8String &s, char *buffer, int len, int &bom)
{
  if (!s && !bom && len>=2)
    {
      if (buffer[0]==(char)0xfe && buffer[1]==(char)0xff)
        bom = 0xfeff;
      if (buffer[0]==(char)0xff && buffer[1]==(char)0xfe)
        bom = 0xfffe;
      if (bom)
        {
          buffer += 2; 
          len -= 2;
        }
    }
  if (bom == 0xfeff)
    for (int i=0; i<len; i+=2)
      *(uint16_t*)(buffer+i) = ((buffer[i]<<8) | buffer[i+1]);
  if (bom == 0xfffe)
    for (int i=0; i<len; i+=2)
      *(uint16_t*)(buffer+i) = ((buffer[i+1]<<8) | buffer[i]);
  if (bom)
    s += GUTF8String((const uint16_t*)buffer, len/2);
  else
    s += GUTF8String((const char*)buffer, len);
}

bool 
BufferByteStream::read_ps_string(GUTF8String &s)
{
  int bom = 0;
  unsigned int pos = 0;
  char buffer[512];
  if (get() != '(') 
    return false;
  s = "";
  for(;;)
    {
      int c = get();
      if (c == '\n' || c == '\r')
        return false;
      if (c == ')')
        break;
      if (c == '\\')
        {
          c = get();
          switch (c) 
            {
            case 'n': 
              c='\n' ; break;
            case 'r': 
              c='\r' ; break;
            case 't': 
              c='\t' ; break;
            case 'b': 
              c='\b' ; break;
            case 'f': 
              c='\f' ; break;
            default:
              if (c>='0' && c<='7') 
                {
                  int n = 0;
                  int x = 0;
                  while (c>='0' && c<='7' && n<3)
                    {
                      x = (x * 8) + c - '0';
                      c = get();
                      n++;
                    }
                  unget(c);
                  c = x;
                }
              break;
            }
        }
      if (c == EOF)
        return false;
      if (pos >= (int)sizeof(buffer))
        {
          add_to_string(s, buffer, pos, bom);
          pos = 0;
        }
      buffer[pos++] = c;
    }
  add_to_string(s, buffer, pos, bom);
  return true;
}

// --------------------------------------------------
// COLOR CONNECTED COMPONENT ANALYSIS
// --------------------------------------------------


// -- A run of pixels with the same color
struct Run    
{ 
  short y;       // vertical coordinate
  short x1;      // first horizontal coordinate
  short x2;      // last horizontal coordinate
  short color;   // color id
  int ccid;      // component id
};


// -- Compares runs for y-x sorting
static inline bool
operator <= (const Run &a, const Run &b)
{
  return (a.y<b.y) || (a.y==b.y && a.x1<=b.x1);
}


// -- Color component descriptor
struct CC    
{
  GRect bb;      // bounding box
  int npix;      // number of black pixels
  int nrun;      // number of runs
  int frun;      // first run in cc ordered array of runs
  int color;     // color id
};


// -- An image composed of runs
class CRLEImage 
{
public:
  int height;            // Height of the image in pixels
  int width;             // Width of the image in pixels
  GP<DjVuPalette> pal;   // Color palette
  GTArray<Run>    runs;  // Array of runs
  GTArray<CC>     ccs;   // Array of component descriptors
  int  nregularccs;      // Number of regular ccs (set by merge_and_split_ccs)
  char bg_flags;         // Comment flags about background.
  char fg_flags;         // Comment flags about foreground.
  CRLEImage(BufferByteStream &bs);
  GP<GBitmap> get_bitmap_for_cc(int ccid) const;
  void make_ccids_by_analysis();
  void make_ccs_from_ccids();
  void merge_and_split_ccs(int smallsize, int largesize);
  void sort_in_reading_order(); 
private:
  unsigned int read_integer(BufferByteStream &bs);
  void insert_runs(int y, const short *x1x2color, int nruns);
};


// -- Helper for CRLEImage::CRLEImage(ByteStream &bs)
unsigned int 
CRLEImage::read_integer(BufferByteStream &bs)
{
  int c, x;
  while (bs.skip() && bs.expect(c, "#"))
    {
      char buffer[256];
      char *s = buffer;
      while (c != EOF && c != '\n' && c != '\r')
        {
          if (s - buffer < (int)sizeof(buffer) - 1)
            *s++ = c;
          c = bs.get();
        }
      *s = 0;
      for(s = buffer; *s; s++)
        {
          if (!strncmp(s, "bg-", 3))
            {
              if (!strncmp(s+3,"bw",2)   ||
                  !strncmp(s+3,"gray",4) ||
                  !strncmp(s+3,"color",5)  )
                bg_flags = s[3];
            }
          if (!strncmp(s, "fg-", 3))
            {
              if (!strncmp(s+3,"bw",2)   ||
                  !strncmp(s+3,"gray",4) ||
                  !strncmp(s+3,"color",5)  )
                fg_flags = s[3];
            }
        }
    }
  if (! bs.read_integer(x) )
    G_THROW("csepdjvu: corrupted input file (bad file header)");
  return x;
}


// -- Helper for CRLEImage::CRLEImage(ByteStream &bs)
void
CRLEImage::insert_runs(int y, const short *x1x2color, int count)
{
  if (count > 0)
    {
      int index = runs.lbound() - count;
      runs.resize(index, runs.hbound());
      Run *run = &runs[index];
      while (--count>=0) {
        run->y     = y;
        run->x1    = *x1x2color++;
        run->x2    = *x1x2color++;
        run->color = *x1x2color++;
        run->ccid  = 0;
        run++;
      }
      runs.shift(-index);
    }
}


// -- Constructs CRLEImage from a run lenght encoded file,
//    making sure that runs are properly sorted.
CRLEImage::CRLEImage(BufferByteStream &bs)
  : height(0), width(0), nregularccs(0), bg_flags(0), fg_flags(0)
{
  unsigned int magic = bs.read16();
  width = read_integer(bs);
  height = read_integer(bs);
  if (width<1 || height<1)
    G_THROW("csepdjvu: corrupted input file (bad image size)");
  // An array for the runs and the buffered data
  GTArray<short> ax(3*width+3);
  // File format switch
  if (magic == 0x5234) // Black&White RLE data
    {
      // Skip one character
      bs.get(); 
      // Setup palette with one color
      pal = DjVuPalette::create();
      static char zeroes[4];
      GP<ByteStream> gzbs=ByteStream::create_static(zeroes,4);
      ByteStream &zbs=*gzbs;
      pal->decode_rgb_entries(zbs, 1);
      // RLE format
      int x, c, n;
      unsigned char p = 0;
      short *px = ax;
      n = height - 1;
      c = 0;
      while (n >= 0)
        {
          if (bs.eof())
            G_THROW( ByteStream::EndOfFile );
          x = bs.get();
          if (x >= 0xc0)
            x = (bs.get()) + ((x - 0xc0) << 8);
          if (c+x > width)
            G_THROW("csepdjvu: corrupted input file (lost RLE sync.)");
          if (p)
            {
              px[0] = c;
              px[1] = c+x-1;
              px[2] = 0;
              px += 3;
            }
          c += x;
          p = 1 - p;
          if (c >= width)
            {
              insert_runs(n, ax, (px-ax)/3);
              c = 0;
              p = 0;
              n -= 1; 
              px = ax;
            }
        }
    } else if (magic == 0x5236) { // Color-RLE data 
      // Read ncolors and skip one character.
      int ncolors = read_integer(bs);
      bs.get();
      // Setup palette 
      if (ncolors<1 || ncolors>4095) 
        G_THROW("csepdjvu: corrupted input file (bad number of colors)");

      pal = DjVuPalette::create();
      pal->decode_rgb_entries(bs, ncolors);
      // RLE format
      int x, c, n, p;
      n = height - 1;
      c = 0;
      short *px = ax;
      while (n >= 0)
        {
          if (bs.eof())
            G_THROW( ByteStream::EndOfFile );
          x  = (bs.get() << 24);
          x  |= (bs.get() << 16);
          x  |= (bs.get() << 8);
          x  |= (bs.get());
          p = (x >> 20) & 0xfff;
          x = (x & 0xfffff);
          if (c+x > width)
            G_THROW("csepdjvu: corrupted input file (lost RLE sync.)");
          if (p >= 0 && p < ncolors)
            {
              px[0] = c;
              px[1] = c+x-1;
              px[2] = p;
              px += 3;
            }
          c += x;
          if (c >= width)
            {
              insert_runs(n, ax, (px-ax)/3);
              c = 0;
              p = 0;
              n -= 1; 
              px = ax;
            }
        }
    } else { // Unrecognized file
      G_THROW("csepdjvu: corrupted input file (bad file header)");
    }
}


// -- Performs color connected component analysis
//    assuming that runs are already y-x sorted.
void
CRLEImage::make_ccids_by_analysis()
{
  // runs.sort(); (we know that runs are 
  // Single Pass Connected Component Analysis (with unodes)
  int n;
  int p=0;
  GTArray<int> umap;
  for (n=0; n<=runs.hbound(); n++)
    {
      int y = runs[n].y;
      int x1 = runs[n].x1 - 1;
      int x2 = runs[n].x2 + 1;
      int color = runs[n].color;
      int id = (umap.hbound() + 1);
      // iterate over previous line runs
      if (p>0) p--;
      for(;runs[p].y < y-1;p++);
      for(;(runs[p].y < y) && (runs[p].x1 <= x2);p++ )
        {
          if ( runs[p].x2 >= x1 )
            {
              if (runs[p].color == color)
                {
                  // previous run touches current run and has same color
                  int oid = runs[p].ccid;
                  while (umap[oid] < oid)
                    oid = umap[oid];
                  if ((int)id > umap.hbound()) {
                    id = oid;
                  } else if (id < oid) {
                    umap[oid] = id;
                  } else {
                    umap[id] = oid;
                    id = oid;
                  }
                  // freshen previous run id
                  runs[p].ccid = id;
                }
              // stop if previous run goes past current run
              if (runs[p].x2 >= x2)
                break;
            }
        }
      // create new entry in umap
      runs[n].ccid = id;
      if (id > umap.hbound())
        {
          umap.touch(id);
          umap[id] = id;
        }
    }
  // Update umap and ccid
  for (n=0; n<=runs.hbound(); n++)
    {
      Run &run = runs[n];
      int ccid = run.ccid;
      while (umap[ccid] < ccid)
        ccid = umap[ccid];
      umap[run.ccid] = ccid;
      run.ccid = ccid;
    }
}


// -- Constructs the ``ccs'' array from run's ccids.
void
CRLEImage::make_ccs_from_ccids()
{
  int n;
  Run *pruns = runs;
  // Find maximal ccid
  int maxccid = -1;
  for (n=0; n<=runs.hbound(); n++)
    if (pruns[n].ccid > maxccid)
      maxccid = runs[n].ccid;
  GTArray<int> armap(0,maxccid);
  int *rmap = armap;
  // Renumber ccs 
  for (n=0; n<=maxccid; n++)
    armap[n] = -1;
  for (n=0; n<=runs.hbound(); n++)
    if (pruns[n].ccid >= 0)
      rmap[ pruns[n].ccid ] = 1;
  int nid = 0;
  for (n=0; n<=maxccid; n++)
    if (rmap[n] > 0)
      rmap[n] = nid++;
  // Adjust nregularccs (since ccs are renumbered)
  while (nregularccs>0 && rmap[nregularccs-1]<0)
    nregularccs -= 1;
  if (nregularccs>0)
    nregularccs = 1 + rmap[nregularccs-1];
  // Prepare cc descriptors
  ccs.resize(0,nid-1);
  for (n=0; n<nid; n++)
    ccs[n].nrun = 0;
  // Relabel runs
  for (n=0; n<=runs.hbound(); n++)
    {
      Run &run = pruns[n];
      if (run.ccid < 0) continue;  // runs with negative ccids are destroyed
      int oldccid = run.ccid;
      int newccid = rmap[oldccid];
      CC &cc = ccs[newccid];
      run.ccid = newccid;
      cc.nrun += 1;
    }
  // Compute positions for runs of cc
  int frun = 0;
  for (n=0; n<nid; n++) 
    {
      ccs[n].frun = rmap[n] = frun;
      frun += ccs[n].nrun;
    }
  // Copy runs
  GTArray<Run> rtmp;
  rtmp.steal(runs);
  Run *ptmp = rtmp;
  runs.resize(0,frun-1);
  pruns = runs;
  for (n=0; n<=rtmp.hbound(); n++)
    {
      int id = ptmp[n].ccid;
      if (id < 0) continue;
      int pos = rmap[id]++;
      pruns[pos] = ptmp[n];
    }
  // Finalize ccs
  for (n=0; n<nid; n++)
    {
      CC &cc = ccs[n];
      int npix = 0;
      runs.sort(cc.frun, cc.frun+cc.nrun-1);
      Run *run = &runs[cc.frun];
      int xmin = run->x1;
      int xmax = run->x2;
      int ymin = run->y;
      int ymax = run->y;
      cc.color = run->color;
      for (int i=0; i<cc.nrun; i++, run++)
        {
          if (run->x1 < xmin)  xmin = run->x1;
          if (run->x2 > xmax)  xmax = run->x2;
          if (run->y  < ymin)  ymin = run->y;
          if (run->y  > ymax)  ymax = run->y;
          npix += run->x2 - run->x1 + 1;
        }
      cc.npix = npix;
      cc.bb.xmin = xmin;
      cc.bb.ymin = ymin;
      cc.bb.xmax = xmax + 1;
      cc.bb.ymax = ymax + 1;
    }
}


// -- Helper for merge_and_split_ccs
struct Grid_x_Color 
{
  short gridi;
  short gridj;
  int color;
};


// -- Helper for merge_and_split_ccs
static inline unsigned int
hash(const Grid_x_Color &x) 
{
  return (x.gridi<<16) ^ (x.gridj<<8) ^ x.color;
}


// -- Helper for merge_and_split_ccs
static inline bool
operator==(const Grid_x_Color &x, const Grid_x_Color &y)
{
  return (x.gridi==y.gridi) && (x.gridj==y.gridj) && (x.color==y.color);
}


// -- Helper for merge_and_split_ccs
static int
makeccid(const Grid_x_Color &x, GMap<Grid_x_Color,int> &map, int &ncc)
{
  GPosition p = map.contains(x);
  if (p) return map[p];
  return map[x] = ncc++;
}


// -- Merges small ccs of similar color and splits large ccs
void
CRLEImage::merge_and_split_ccs(int smallsize, int largesize)
{
  int ncc = ccs.size();
  int nruns = runs.size();
  int splitsize = largesize;
  if (ncc <= 0) return;
  // Associative map for storing merged ccids
  GMap<Grid_x_Color,int> map;
  nregularccs = ncc;
  // Set the correct ccids for the runs
  for (int ccid=0; ccid<ccs.size(); ccid++)
    {
      CC* cc = &ccs[ccid];
      if (cc->nrun <= 0) continue;
      Grid_x_Color key;
      key.color = cc->color;
      int ccheight = cc->bb.height();
      int ccwidth = cc->bb.width();
      if (ccheight<=smallsize && ccwidth<=smallsize)
        {
          key.gridi = (cc->bb.ymin+cc->bb.ymax)/splitsize/2;
          key.gridj = (cc->bb.xmin+cc->bb.xmax)/splitsize/2;
          int newccid = makeccid(key, map, ncc);
          for(int runid=cc->frun; runid<cc->frun+cc->nrun; runid++)
            runs[runid].ccid = newccid;
        }
      else if (ccheight>=largesize || ccwidth>=largesize)
        {
          for(int runid=cc->frun; runid<cc->frun+cc->nrun; runid++)
            {
              Run *r = & runs[runid];
              key.gridi = r->y/splitsize;
              key.gridj = r->x1/splitsize;
              int gridj_end = r->x2/splitsize;
              int gridj_span = gridj_end - key.gridj;
              r->ccid = makeccid(key, map, ncc);
              if (gridj_span>0)
                {
                  // truncate current run 
                  runs.touch(nruns+gridj_span-1);
                  r = &runs[runid];
                  int x = key.gridj*splitsize + splitsize;
                  int x_end = r->x2;
                  r->x2 = x-1;
                  // append additional runs to the runs array
                  while (++key.gridj < gridj_end)
                    {
                      Run& newrun = runs[nruns++];
                      newrun.y = r->y;
                      newrun.x1 = x;
                      x += splitsize;
                      newrun.x2 = x-1;
                      newrun.color = key.color;
                      newrun.ccid = makeccid(key, map, ncc);
                    }
                  // append last run to the run array
                  Run& newrun = runs[nruns++];
                  newrun.y = r->y;
                  newrun.x1 = x;
                  newrun.x2 = x_end;
                  newrun.color = key.color;
                  newrun.ccid = makeccid(key, map, ncc);
                }
            }
        }
    }
  // Recompute cc descriptors
  make_ccs_from_ccids();
}


// -- Helps sorting cc
static int 
top_edges_descending (const void *pa, const void *pb)
{
  if (((CC*) pa)->bb.ymax != ((CC*) pb)->bb.ymax)
    return (((CC*) pb)->bb.ymax - ((CC*) pa)->bb.ymax);
  if (((CC*) pa)->bb.xmin != ((CC*) pb)->bb.xmin)
    return (((CC*) pa)->bb.xmin - ((CC*) pb)->bb.xmin);
  return (((CC*) pa)->frun - ((CC*) pb)->frun);
}


// -- Helps sorting cc
static int 
left_edges_ascending (const void *pa, const void *pb)
{
  if (((CC*) pa)->bb.xmin != ((CC*) pb)->bb.xmin)
    return (((CC*) pa)->bb.xmin - ((CC*) pb)->bb.xmin);
  if (((CC*) pb)->bb.ymax != ((CC*) pa)->bb.ymax)
    return (((CC*) pb)->bb.ymax - ((CC*) pa)->bb.ymax);
  return (((CC*) pa)->frun - ((CC*) pb)->frun);
}


// -- Helps sorting cc
static int 
integer_ascending (const void *pa, const void *pb)
{
  return ( *(int*)pb - *(int*)pa );
}


// -- Sort ccs in approximate reading order
void 
CRLEImage::sort_in_reading_order()
{
  if (nregularccs<2) return;
  CC *ccarray = new CC[nregularccs];
  // Copy existing ccarray (but segregate special ccs)
  int ccid;
  for(ccid=0; ccid<nregularccs; ccid++)
    ccarray[ccid] = ccs[ccid];
  // Sort the ccarray list into top-to-bottom order.
  qsort (ccarray, nregularccs, sizeof(CC), top_edges_descending);
  // Subdivide the ccarray list roughly into text lines
  int maxtopchange = width / 40;
  if (maxtopchange < 32) 
    maxtopchange = 32;
  // - Loop until processing all ccs
  int ccno = 0;
  int *bottoms = new int[nregularccs];
  while (ccno < nregularccs)
    {
      // - Gather first line approximation
      int nccno;
      int sublist_top = ccarray[ccno].bb.ymax-1;
      int sublist_bottom = ccarray[ccno].bb.ymin;
      for (nccno=ccno; nccno < nregularccs; nccno++)
        {
          if (ccarray[nccno].bb.ymax-1 < sublist_bottom) break;
          if (ccarray[nccno].bb.ymax-1 < sublist_top - maxtopchange) break;
          int bottom = ccarray[nccno].bb.ymin;
          bottoms[nccno-ccno] = bottom;
          if (bottom < sublist_bottom)
            sublist_bottom = bottom;
        }
      // - If more than one candidate cc for the line
      if (nccno > ccno + 1)
        {
          // - Compute median bottom
          qsort(bottoms, nccno-ccno, sizeof(int), integer_ascending);
          int bottom = bottoms[ (nccno-ccno-1)/2 ];
          // - Compose final line
          for (nccno=ccno; nccno < nregularccs; nccno++)
            if (ccarray[nccno].bb.ymax-1 < bottom)
              break;
          // - Sort final line
          qsort (ccarray+ccno, nccno-ccno, sizeof(CC), left_edges_ascending);
        }
      // - Next line
      ccno = nccno;
    }
  // Copy ccarray back and renumber the runs
  for(ccid=0; ccid<nregularccs; ccid++)
    {
      CC& cc = ccarray[ccid];
      ccs[ccid] = cc;
      for(int r=cc.frun; r<cc.frun+cc.nrun; r++)
        runs[r].ccid = ccid;
    }
  // Free memory
  delete [] bottoms;
  delete[] ccarray;
}


// -- Creates a bitmap for a particular component
GP<GBitmap>   
CRLEImage::get_bitmap_for_cc(const int ccid) const
{
  const CC &cc = ccs[ccid];
  const GRect &bb = cc.bb;
  GP<GBitmap> bits = GBitmap::create(bb.height(), bb.width());
  const Run *prun = & runs[(int)cc.frun];
  for (int i=0; i<cc.nrun; i++,prun++)
    {
      if (prun->y<bb.ymin || prun->y>=bb.ymax)
        G_THROW("Internal error (y bounds)");
      if (prun->x1<bb.xmin || prun->x2>=bb.xmax)
        G_THROW("Internal error (x bounds)");
      unsigned char *row = (*bits)[prun->y - bb.ymin];
      for (int x=prun->x1; x<=prun->x2; x++)
        row[x - bb.xmin] = 1;
    }
  return bits;
}



// --------------------------------------------------
// PROCESS BACKGROUND PIXMAP
// --------------------------------------------------


// -- Tries to read a background pixmap
GP<GPixmap>
read_background(BufferByteStream &bs, int w, int h, int &bgred)
{
  // Skip null bytes (why?)
  int lookahead;
  while (! (lookahead = bs.get())) { }
  bs.unget(lookahead);
  // Check pixmap
  if (lookahead != 'P') 
    return 0;
  GP<GPixmap> pix = GPixmap::create(bs);
  // Check background reduction
  for (bgred=1; bgred<=12; bgred++) 
    {
      int subw = (w + bgred - 1) / bgred;
      int subh = (h + bgred - 1) / bgred;
      if (subh == (int)pix->rows() && subw == (int)pix->columns())
        // Found reduction factor
        return pix;
    }
  // Failure
  G_THROW("Background pixmap size does not match foreground");
  return 0;
}


// --------------------------------------------------
// HANDLE COMMENTS IN SEP FILES
// --------------------------------------------------

class Comments
{
public:
  Comments(int w, int h, const csepdjvuopts &opts);
  void process_comments(BufferByteStream &bs, int verbose=0);
  bool parse_comment_line(BufferByteStream &bs);
  void make_chunks(IFFByteStream &iff);
  GP<DjVmNav> get_djvm_nav();
protected:
  int w;
  int h;
  GRectMapper mapper;
  DjVuTXT::ZoneType detail;
  int lastx, lasty;
  int lastdirx, lastdiry;
  int lastsize[3];
  struct TxtMark : public GPEnabled {
    int x,y,dx,dy;
    int inter;
    GRect r;
    GUTF8String s;
  };
  GPList<TxtMark> lastline;
  GP<DjVuTXT> txt;
  struct LnkMark : public GPEnabled {
    GRect r;
    GUTF8String s;
  };
  GPList<LnkMark> links;
  GP<DjVmNav> nav;
protected:
  bool allspace(const TxtMark *mark);
  void textmark(GP<TxtMark> mark);
  void textflush(void);
};

Comments::Comments(int w, int h, const csepdjvuopts &opts)
  : w(w), h(h), detail(opts.text)
{
  GRect pagerect(0,0,w,h);
  mapper.set_input(pagerect);
  mapper.set_output(pagerect);
  mapper.mirrory();
}

void 
Comments::process_comments(BufferByteStream &bs, int verbose)
{
  int c;
  // Skip null bytes
  while (! (c = bs.get())) { }
  // Process comment lines
  while (c == '#')
    {
      const char *message = 0;
      bs.skip(" \t");
      G_TRY
        {
          if (! parse_comment_line(bs) && verbose > 1)
            message = "csepdjvu: unrecognized comment '# ";
          else if (bs.skip(" \t") && bs.expect(c, "\n\r"))
            bs.unget(c);
          else if (verbose > 1)
            message = "csepdjvu: garbage in comments: '";
        }
      G_CATCH(ex)
      {
        message = 0;
        GUTF8String str = DjVuMessageLite::LookUpUTF8(ex.get_cause());
        if (verbose > 1)
          DjVuPrintErrorUTF8("%s\n",(const char *)str);
      } 
      G_ENDCATCH;
      if (message)
        DjVuPrintErrorUTF8(message);
      c = bs.get();
      while (c != EOF && c != '\r' && c != '\n')
        {
          if (message)
            DjVuPrintErrorUTF8("%c", c);
          c = bs.get();
        }
      if (message)
        DjVuPrintErrorUTF8("'\n");
      bs.skip();
      c = bs.get();
    }
  bs.unget(c);
}

bool
Comments::parse_comment_line(BufferByteStream &bs)
{
  int c = bs.get();
  // Text comments
  if (c == 'T')
    {
      GP<TxtMark> mark = new TxtMark;
      if (! (bs.skip(" \t") && bs.read_pair(mark->x,mark->y) &&
             bs.skip(" \t") && bs.read_pair(mark->dx,mark->dy) &&
             bs.skip(" \t") && bs.read_geometry(mark->r) &&
             bs.skip(" \t") && bs.read_ps_string(mark->s) ) )
        G_THROW("csepdjvu: corrupted file (syntax error in text comment)");
      if (mark->r.isempty())
        G_THROW("csepdjvu: corrupted file (empty rectangle in text comment)");
      textmark(mark);
      return true;
    }
  // Link comments
  if (c == 'L')
    {
      GP<LnkMark> mark = new LnkMark;
      if (! (bs.skip(" \t") && bs.read_geometry(mark->r) &&
             bs.skip(" \t") && bs.read_ps_string(mark->s) ) )
        G_THROW("csepdjvu: corrupted file (syntax error in link comment)");
      if (mark->r.isempty())
        G_THROW("csepdjvu: corrupted file (empty rectangle in link comment)");
      int ymax = h - mark->r.ymin - 1; // reversed in gsdjvu ?
      int ymin = h - mark->r.ymax - 1; // reversed in gsdjvu ?
      mark->r.ymax = ymax;
      mark->r.ymin = ymin;
      links.append(mark);
      return true;
    }
  // Bookmark comments
  if (c == 'B')
    {
      int count; 
      GUTF8String url;
      GUTF8String title;
      if (! (bs.skip(" \t") && bs.read_integer(count) &&
             bs.skip(" \t") && bs.read_ps_string(title) &&
             bs.skip(" \t") && bs.read_ps_string(url) ) )
        G_THROW("csepdjvu: corrupted file (syntax error in outline comment)");
      GP<DjVmNav::DjVuBookMark> b = 
        DjVmNav::DjVuBookMark::create(count, title, url);
      if (b && ! nav)
        nav = DjVmNav::create();
      if (b)
        nav->append(b);
      return true;
    }
  // Unrecognized
  bs.unget(c);
  return false;
}

static int 
median3(int *p)
{
  if (p[0] > p[1])
    return MAX(p[1],MIN(p[0],p[2]));
  else
    return MIN(p[1],MAX(p[0],p[2]));
}

static bool 
allspaces(const GUTF8String &s)
{
  bool ok = true;
  for (int i=0; ok && i<(int)s.length(); i++)
    if (s[i] != ' ')
      ok = false;
  return ok;
}

void 
Comments::textmark(GP<TxtMark> mark)
{
  // determine direction
  int dirx = 0;
  int diry = 0;
  int size = 0;
  if (abs(mark->dx) > 8*abs(mark->dy))
    {
      dirx = (mark->dx > 0) ? +1 : -1;
      size = mark->r.height();
    }
  else if (abs(mark->dy) > 8*abs(mark->dy))
    {
      diry = (mark->dy > 0) ? +1 : -1;
      size = mark->r.width();
    }
  // make mark
  mark->inter = 0;
  // flush previous line
  if (lastline.size())
    {
      if (size != lastsize[0])
        {
          lastsize[2] = lastsize[1];
          lastsize[1] = lastsize[0];
          lastsize[0] = size;
        }
      int fontsize = median3(lastsize) + 1;
      int shx = (mark->x - lastx) * 100 / fontsize;
      int shy = (mark->y - lasty) * 100 / fontsize;
      int inter = dirx * shx + diry * shy;
      if ( (dirx || diry) && (dirx == lastdirx) && (diry == lastdiry) &&
           (inter > -150) && (inter < 300) && 
           abs(diry * shx + dirx * shy) < 80 )
        mark->inter = inter;
      else
        textflush();
    }
  if (! lastline.size())
    lastsize[0] = lastsize[1] = lastsize[2] = size;
  lastline.append(mark);
  lastdirx = dirx;
  lastdiry = diry;
  lastx = mark->x + mark->dx;
  lasty = mark->y + mark->dy;
}

void 
Comments::textflush(void)
{
  int size = lastline.size();
  if (size > 0)
    {
      // compute word spacing
      int i = 0;
      GTArray<int> inter(0,size-1);
      for (GPosition p=lastline; p; ++p)
        inter[i++] = lastline[p]->inter;
      inter.sort();
      int wordsep = MAX(10, 2 * inter[(2*size)/3]);
      // compute word list
      GP<TxtMark> word;
      GPList<TxtMark> words;
	  { // extra nesting for windows
        for (GPosition p=lastline; p; ++p)
        {
          TxtMark *mark = lastline[p];
          if (word && mark->inter > wordsep) 
            {
              if (! allspaces(word->s))
                words.append(word);
              word = 0;
            }
          if (! word) 
            {
              word = mark;
            } 
          else 
            {
              word->dx += mark->dx;
              word->dy += mark->dy;
              word->s += mark->s;
              word->r.recthull(word->r, mark->r);
            }
        }
	  }
      if (word)
        {
          if (! allspaces(word->s))
            words.append(word);
          word = 0;
        }
      // create text data
      int size = words.size();
      if (size)
        {
          DjVuTXT::Zone *lzone = 0;
          for (GPosition p = words; p; ++p)
            {
              word = words[p];
              mapper.map(word->r);
              if (! lzone)
                {
                  if (! txt)
                    {
                      txt = DjVuTXT::create();
                      txt->page_zone.ztype = DjVuTXT::PAGE;
                      txt->page_zone.rect = GRect(0,0,w,h);
                      txt->page_zone.text_start = 0;
                      txt->page_zone.text_length = 0;
                    }
                  lzone = txt->page_zone.append_child();
                  lzone->ztype = DjVuTXT::LINE;
                  lzone->text_start = txt->textUTF8.length();
                  lzone->text_length = 0;
                }
              if (detail >= DjVuTXT::WORD)
                {
                  DjVuTXT::Zone *wzone = lzone->append_child();
                  wzone->ztype = DjVuTXT::WORD;
                  wzone->text_start = txt->textUTF8.length();
                  txt->textUTF8 += word->s;
                  wzone->text_length = 
                    txt->textUTF8.length() - wzone->text_start;
                  wzone->rect = word->r;
                  lzone->rect.recthull(lzone->rect, word->r);
                }
              else
                {
                  if (lzone->text_length > 0) txt->textUTF8 += " ";
                  txt->textUTF8 += word->s;
                  lzone->text_length = 
                    txt->textUTF8.length() - lzone->text_start;
                  lzone->rect.recthull(lzone->rect, word->r);
                }
            }
        }
    }
  lastline.empty();
}

static int 
bytestream_fputs(miniexp_io_t *io, const char *s)
{
  ByteStream *outbs = (ByteStream*)io->data[0];
  return (outbs) ? outbs->write((const void*)s, strlen(s)) : -1;
}

void 
Comments::make_chunks(IFFByteStream &iff)
{
  // Write text chunk
  textflush();
  if (txt)
    {
      txt->normalize_text();
      iff.put_chunk("TXTz");
      {
        GP<ByteStream> bsb = BSByteStream::create(iff.get_bytestream(), 50);
        txt->encode(bsb);
      }
      iff.close_chunk();
    }
  // Create annotation chunk
  if (links.size() > 0)
    {
      iff.put_chunk("ANTz");
      {
        GP<ByteStream> bsb = BSByteStream::create(iff.get_bytestream(), 50);
        miniexp_io_t io;
        miniexp_io_init(&io);
        io.fputs = bytestream_fputs;
        io.data[0] = (void*)(ByteStream*)bsb;
        minivar_t exor = miniexp_cons(miniexp_symbol("xor"),miniexp_nil);
        minivar_t zstr = miniexp_string("");
        for (GPosition p = links; p; ++p)
          {
            GP<LnkMark> mark = links[p];
            minivar_t url = miniexp_string((const char*)(mark->s));
            minivar_t expr = miniexp_cons(exor, miniexp_nil);
            minivar_t area;
            area = miniexp_cons(miniexp_number(mark->r.height()), area);
            area = miniexp_cons(miniexp_number(mark->r.width()), area);
            area = miniexp_cons(miniexp_number(mark->r.ymin), area);
            area = miniexp_cons(miniexp_number(mark->r.xmin), area);
            area = miniexp_cons(miniexp_symbol("rect"),area);
            expr = miniexp_cons(area, expr);
            expr = miniexp_cons(zstr, expr);
            expr = miniexp_cons(url, expr);
            expr = miniexp_cons(miniexp_symbol("maparea"), expr);
            miniexp_pprint_r(&io, expr, 72);
          }
      }
      iff.close_chunk();
    }
}

GP<DjVmNav>
Comments::get_djvm_nav()
{
  if (nav && nav->getBookMarkCount() && nav->isValidBookmark())
    return nav;
  if (nav)
    DjVuPrintErrorUTF8("%s", "csepdjvu: corrupted outline comments.\n");
  return 0;
}


// --------------------------------------------------
// MAIN COMPRESSION ROUTINE
// --------------------------------------------------


// -- Compresses one page:
//    - bytestream bs contains the input separated file.
//    - bytestream obs will receive the output djvu file.
void 
csepdjvu_page(BufferByteStream &bs, 
              GP<ByteStream> obs, 
              GP<DjVmNav> &nav,
              const csepdjvuopts &opts)
{
  // Read rle data from separation file
  CRLEImage rimg(bs);
  int w = rimg.width;
  int h = rimg.height;
  if (opts.verbose > 1)
    DjVuFormatErrorUTF8( "%s\t%d\t%d\t%d\t%d",
                     ERR_MSG("csepdjvu.summary"), 
                     w, h, rimg.pal->size(), rimg.runs.size());
  
  // Perform Color Connected Component Analysis
  rimg.make_ccids_by_analysis();                  // Obtain ccids
  rimg.make_ccs_from_ccids();                     // Compute cc descriptors
  if (opts.verbose > 1)
    DjVuFormatErrorUTF8("%s\t%d", ERR_MSG("csepdjvu.analyzed"), 
                        rimg.ccs.size());
  
  // Post-process Color Connected Components
  int largesize = MIN(500, MAX(64, opts.dpi));
  int smallsize = MAX(2, opts.dpi/150);
  rimg.merge_and_split_ccs(smallsize,largesize);  // Eliminates gross ccs
  if (opts.verbose > 1)
    DjVuFormatErrorUTF8( "%s\t%d",
                     ERR_MSG("csepdjvu.merge_split"), 
                     rimg.ccs.size());
  rimg.sort_in_reading_order();                   // Sort cc descriptors
  
  // Create JB2Image and fill colordata
  GP<JB2Image> gjimg=JB2Image::create(); 
  JB2Image &jimg=*gjimg;
  jimg.set_dimension(w, h);
  int nccs = rimg.ccs.size();
  for (int ccid=0; ccid<nccs; ccid++)
    {
      JB2Shape shape;
      JB2Blit  blit;
      shape.parent = -1;
      shape.userdata = 0;
      if (ccid >= rimg.nregularccs)
        shape.userdata |= JB2SHAPE_SPECIAL;
      shape.bits = rimg.get_bitmap_for_cc(ccid);
      shape.bits->compress();
      CC& cc = rimg.ccs[ccid];
      blit.shapeno = jimg.add_shape(shape);
      blit.left = cc.bb.xmin;
      blit.bottom = cc.bb.ymin;
      int blitno = jimg.add_blit(blit);
      rimg.pal->colordata.touch(blitno);
      rimg.pal->colordata[blitno] = cc.color;
    }
  
  // Organize JB2Image
  tune_jb2image_lossless(&jimg);
  if (opts.verbose> 1)
    {
      int nshape=0, nrefine=0;
      for (int i=0; i<jimg.get_shape_count(); i++) 
        {
          if (!jimg.get_shape(i).bits) continue;
          if (jimg.get_shape(i).parent >= 0) nrefine++; 
          nshape++; 
        }
      DjVuFormatErrorUTF8( "%s\t%d\t%d",
                       ERR_MSG("csepdjvu.cross_code"), 
                       nshape, nrefine);
    }
  
  // Obtain background image
  int bgred;
  GP<GPixmap> bgpix = read_background(bs, w, h, bgred);
  if (opts.verbose > 1 && bgpix)
    DjVuFormatErrorUTF8( "%s\t%d", ERR_MSG("csepdjvu.reduction"), bgred);

  // Process comments
  Comments coms(w, h, opts);
  coms.process_comments(bs, opts.verbose);
  
  // Compute flags for simplifying output format
  bool white_background = (bgpix ? false : true);
  bool gray_background = white_background;
  if (rimg.bg_flags == 'g' || rimg.bg_flags=='b')
    gray_background = true;
  bool bitonal = false;
  if (white_background && rimg.pal->size() == 1) 
    {
      GPixel fgcolor;
      rimg.pal->index_to_color(0, fgcolor);
      if (fgcolor == GPixel::BLACK)
        bitonal = true;
    }
  if (opts.verbose > 1)
    {
      if (bitonal)
        DjVuWriteError( ERR_MSG("csepdjvu.bilevel") );
      else if (white_background) 
        DjVuWriteError( ERR_MSG("csepdjvu.white_bg") );
      else if (gray_background) 
        DjVuWriteError( ERR_MSG("csepdjvu.gray_bg") );
      else 
        DjVuWriteError( ERR_MSG("csepdjvu.color") );
    }
  
  // Create background image
  GP<IW44Image> iw;
  if (! white_background)
    {
      /* Perform masked compression */
      GP<GBitmap> mask = jimg.get_bitmap(bgred);
      mask->binarize_grays(bgred*bgred-1);
      IW44Image::CRCBMode mode = IW44Image::CRCBnormal;
      if (gray_background) mode = IW44Image::CRCBnone;
      iw = IW44Image::create_encode(*bgpix, mask, mode);
      bgpix = 0;
    } 
  else if (! bitonal) 
    {
      /* Compute white background */
      GPixel bgcolor = GPixel::WHITE;
      GP<GPixmap> inputsub=GPixmap::create((h+11)/12, (w+11)/12, &bgcolor);
      iw = IW44Image::create_encode(*inputsub, 0, IW44Image::CRCBnone);
    }
  
  // Assemble DJVU file
  GP<IFFByteStream> giff=IFFByteStream::create(obs);
  IFFByteStream &iff=*giff;
  // -- main composite chunk
  iff.put_chunk("FORM:DJVU", 1);
  // -- ``INFO'' chunk
  iff.put_chunk("INFO");
  GP<DjVuInfo> ginfo=DjVuInfo::create();
  DjVuInfo &info=*ginfo;
  info.height = h;
  info.width = w;
  info.dpi = opts.dpi;
  info.encode(*iff.get_bytestream());
  iff.close_chunk();
  // -- ``Sjbz'' chunk
  iff.put_chunk("Sjbz");
  jimg.encode(iff.get_bytestream());
  iff.close_chunk();
  // -- Color stuff
  if (! bitonal) 
    {
      // -- ``FGbz'' chunk
      iff.put_chunk("FGbz");
      rimg.pal->encode(iff.get_bytestream());
      iff.close_chunk();
      // -- ``BG44'' chunk
      IWEncoderParms iwparms;
      if (white_background) 
        {
          iff.put_chunk("BG44");
          iwparms.slices = 97;
          iw->encode_chunk(iff.get_bytestream(), iwparms);
          iff.close_chunk();
        }
      else
        {
          const unsigned char *slice = opts.slice;
          while ((iwparms.slices = *slice++))
            {
              iff.put_chunk("BG44");
              iw->encode_chunk(iff.get_bytestream(), iwparms);
              iff.close_chunk();
            }
        }
    }
  // -- terminate main composite chunk
  coms.make_chunks(iff);
  iff.close_chunk();
  // -- store outline
  if (! nav) 
    nav = coms.get_djvm_nav();
}  

// -- Checks whether there is another page in the same file
bool
check_for_another_page(BufferByteStream &bs, const csepdjvuopts &opts)
{
  // Skip null bytes (why?)
  int lookahead;
  while (! (lookahead = bs.get())) { }
  bs.unget(lookahead);
  // Check next header
  if (lookahead == 'R') 
    return true;
  if (lookahead != EOF) 
    DjVuPrintErrorUTF8("%s","csepdjvu: found corrupted data\n");
  return false;
}


// -- Prints usage message
void
usage()
{
  const char *msg = 
#ifdef DJVULIBRE_VERSION
    "CSEPDJVU --- DjVuLibre-" DJVULIBRE_VERSION "\n"
#endif
    "DjVu encoder working with \"separated\" files\n\n"
    "Usage: csepdjvu <...options_or_separatedfiles...> <outputdjvufile>\n"
    "Options are:\n"
    "   -v, -vv    Select verbosity level.\n"
    "   -d <n>     Set resolution to <n> dpi (default: 300).\n"
    "   -t         Restricts text information to lines only.\n"
    "   -q <spec>  Select quality for background (default: 72+11+10+10);\n"
    "              see option -slice in program c44 for more information.\n"
    "Each separated files contain one or more pages\n"
    "Each page is composed of:\n"
    " (1) a B&W-RLE or Color-RLE image representing the foreground,\n"
    " (2) an optional PPM image representing the background layer.\n";
  DjVuPrintErrorUTF8(msg);
  exit(10);
}


// -- Parsing quality spec (borrowed from c44)
void 
parse_slice(const char *q, csepdjvuopts &opts)
{
  int count = 0;
  int lastx = 0;
  while (*q)
    {
      char *ptr; 
      int x = strtol(q, &ptr, 10);
      if (ptr == q)
        G_THROW("csepdjvu: "
                "illegal quality specification (number expected)");
      if (lastx>0 && q[-1]=='+')
        x += lastx;
      if (x<1 || x>1000 || x<lastx)
        G_THROW("csepdjvu: "
                "illegal quality specification (number out of range)");
      lastx = x;
      if (*ptr && *ptr!='+' && *ptr!=',')
        G_THROW("csepdjvu: "
                "illegal quality specification (comma expected)");        
      q = (*ptr ? ptr+1 : ptr);
      if (count+1 >= (int)(sizeof(opts.slice)/sizeof(opts.slice[0])))
        G_THROW("csepdjvu: "
                "illegal quality specification (too many chunks)");
      opts.slice[count++] = x;
      opts.slice[count] = 0;
    }
  if (count < 1)
    G_THROW("csepdjvu: "
            "illegal quality specification (no chunks)");
}


// -- Main routine
int 
main(int argc, const char **argv)
{
  DJVU_LOCALE;
  GArray<GUTF8String> dargv(0,argc-1);
  for(int i=0;i<argc;++i)
    dargv[i]=GNativeString(argv[i]);
  G_TRY
    {
      GP<DjVmDoc> gdoc=DjVmDoc::create();
      GP<DjVmNav> gnav;
      DjVmDoc &doc=*gdoc;
      GURL outputurl;
      GP<ByteStream> goutputpage=ByteStream::create();
      csepdjvuopts opts;
      int pageno = 0;
      // Read outputurl name
      if (argc < 3) usage();
      outputurl = GURL::Filename::UTF8(dargv[--argc]);
      // Process arguments
      for (int i=1; i<argc; i++)
        {
          GUTF8String arg = dargv[i];
          if (arg == "-v")
            opts.verbose = 1;
          else if (arg == "-vv")
            opts.verbose = 2;
          else if (arg == "-t")
            opts.text = DjVuTXT::LINE;
          else if (arg == "-d" && i+1<argc)
            {
              // Specify resolution
              char *end;
              opts.dpi = strtol(dargv[++i], &end, 10);
              if (*end || opts.dpi<25 || opts.dpi>6000)
                usage();
            }
          else if (arg == "-q" && i+1 < argc)
            {
              // Specify background quality
              parse_slice(dargv[++i], opts);
            }
          else if (arg == "-l" || arg == "-t" || arg == "-h")
            {
              DjVuPrintErrorUTF8("csepdjvu: option %s not yet supported\n",
                                 (const char *)arg );
            }
          else if ((arg == "-j" || arg == "-p") && i+1 < argc)
            {
              i += 1;
              DjVuPrintErrorUTF8("csepdjvu: option %s not yet supported\n",
                                 (const char *)arg );
            }
          else 
            {
              // Process separation file
              GP<ByteStream> fbs = 
                ByteStream::create(GURL::Filename::UTF8(arg),"rb");
              BufferByteStream ibs(*fbs);
              do {
                char pagename[16];
                sprintf(pagename, "p%04d.djvu", ++pageno);
                if (opts.verbose > 1)
                  DjVuPrintErrorUTF8("%s","--------------------\n");
                // Compress page 
                goutputpage=ByteStream::create();
                ByteStream &outputpage=*goutputpage;
                csepdjvu_page(ibs, goutputpage, gnav, opts);
                if (opts.verbose) {
                  DjVuPrintErrorUTF8("csepdjvu: %d bytes for page %d",
                                     outputpage.size(), pageno);
                  if (arg == "-")
                    DjVuPrintErrorUTF8("%s"," (from stdin)\n");
                  else
                    DjVuPrintErrorUTF8(" (from file '%s')\n", 
                                       (const char*)arg);
                }
                // Insert page into document
                outputpage.seek(0);
                doc.insert_file(outputpage, DjVmDir::File::PAGE, 
                                pagename, pagename);
              } while (check_for_another_page(ibs, opts));
            }
        } 
      // Save file
      if (pageno == 1 && ! gnav) 
        {
          ByteStream &outputpage=*goutputpage;
          // Save as a single page 
          outputpage.seek(0);
          ByteStream::create(outputurl,"wb")->copy(outputpage);
        }
      else if (pageno >= 1)
        {
          // Save as a bundled file
          doc.set_djvm_nav(gnav);
          doc.write(ByteStream::create(outputurl,"wb"));
        }
      else 
        usage();
    }
  G_CATCH(ex)
    {
      ex.perror();
      exit(1);
    }
  G_ENDCATCH;
  return 0;
}