/**************************************************************************
 *
 * TextGet.cpp -- Decompressing the text
 * Copyright (C) 1999  Rodger McNab
 *
 * This program is free software; you can redistribute it and/or modify
 * it under the terms of the GNU General Public License as published by
 * the Free Software Foundation; either version 2 of the License, or
 * (at your option) any later version.
 *
 * This program is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 * GNU General Public License for more details.
 *
 * You should have received a copy of the GNU General Public License
 * along with this program; if not, write to the Free Software
 * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
 *
 **************************************************************************/

// is important to be first, so we escape the truncation warning on VC++
#include "TextGet.h"
// need this to avoid bizarre compiler problems under VC++ 6.0
#if defined (__WIN32__) && !defined (GSDL_USE_IOS_H)
# include <iostream>
#endif

#include "mg_files.h"
#include "netorder.h"
#include "mg_errors.h"
#include "locallib.h"
#include "words.h"
#include "local_strings.h"
#include "bitio_m_stdio.h"

typedef enum huff_type {lengths, chars};


static auxiliary_dict *LoadAuxDict (compression_dict &cd, FILE *text_aux_dict) {
  auxiliary_dict *ad;
  int i;

  if (!(ad = new auxiliary_dict))
    {
      mg_errno = MG_NOMEM;
      return (NULL);
    }

  memset (ad, '\0', sizeof (*ad));

  for (i = 0; i <= 1; ++i)
    {
      int j;
      u_char *pos;

      fread (&ad->afh[i], sizeof (aux_frags_header), 1, text_aux_dict);

      /* [RPAP - Jan 97: Endian Ordering] */
      NTOHUL(ad->afh[i].num_frags);
      NTOHUL(ad->afh[i].mem_for_frags);

      if (!(ad->word_data[i] = new u_char[ad->afh[i].mem_for_frags]))
	{
	  mg_errno = MG_NOMEM;
	  delete ad;
	  return (NULL);
	}
      if (!(ad->words[i] = new u_char* [ad->afh[i].num_frags]))
	{
	  mg_errno = MG_NOMEM;
	  delete ad;
	  return (NULL);
	}

      fread (ad->word_data[i], ad->afh[i].mem_for_frags, sizeof (u_char),
	     text_aux_dict);

      pos = ad->word_data[i];
      for (j = 0; j < (int)ad->afh[i].num_frags; ++j)
	{
	  ad->words[i][j] = pos;
	  pos += *pos + 1;
	}
      if (cd.cdh.novel_method == MG_NOVEL_HYBRID)
	{
	  int num;
	  num = 1;
	  ad->blk_start[i][0] = 0;
	  ad->blk_end[i][0] = cd.cdh.num_words[i] - 1;
	  while (num < 33)
	    {
	      ad->blk_start[i][num] = ad->blk_end[i][num - 1] + 1;
	      ad->blk_end[i][num] = ad->blk_start[i][num] +
		(ad->blk_end[i][num - 1] - ad->blk_start[i][num - 1]) * 2;
	      ++num;
	    }
	}
    }
  return (ad);
}


static u_char ***ReadInWords (FILE *dict, compression_dict &cd,
			      comp_frags_header *cfh, u_char **escape) {
  int i, lookback;
  int ptrs_reqd = 0;
  int mem_reqd = 0;
  int num_set[MAX_HUFFCODE_LEN + 1];
  u_char *next_word[MAX_HUFFCODE_LEN + 1];
  u_char **vals;
  u_char ***values;
  u_char word[MAXWORDLEN + 1];
  u_char last_word[MAX_HUFFCODE_LEN + 1][MAXWORDLEN + 1];

  lookback = cd.cdh.lookback;

  for (i = cfh->hd.mincodelen; i <= cfh->hd.maxcodelen; ++i) {
    ptrs_reqd += (cfh->hd.lencount[i] + ((1 << lookback) - 1)) >> lookback;
    mem_reqd += cfh->huff_words_size[i];
  }

  if (!(vals = new u_char* [ptrs_reqd]))
    return (NULL);

  if (!(values = new u_char** [MAX_HUFFCODE_LEN + 1]))
    return (NULL);

  if (!(next_word[0] = new u_char[mem_reqd])) 
    return (NULL);

  cd.MemForCompDict += ptrs_reqd * sizeof (*vals) +
    (MAX_HUFFCODE_LEN + 1) * sizeof (u_char **) +
    mem_reqd;

  values[0] = vals;
  values[0][0] = next_word[0];
  for (i = 1; i <= cfh->hd.maxcodelen; ++i)
    {
      int next_start = (values[i - 1] - vals) +
      ((cfh->hd.lencount[i - 1] + ((1 << lookback) - 1)) >> lookback);
      values[i] = &vals[next_start];
      next_word[i] = next_word[i - 1] + cfh->huff_words_size[i - 1];
      values[i][0] = next_word[i];
    }

  memset (num_set, '\0', sizeof (num_set));

  for (i = 0; i < cfh->hd.num_codes; ++i)
    {
      register int val, copy;
      register int len = cfh->hd.clens[i];
      val = getc (dict);
      copy = (val >> 4) & 0xf;
      val &= 0xf;

      fread (word + copy + 1, sizeof (u_char), val, dict);
      *word = val + copy;

      if ((num_set[len] & ((1 << lookback) - 1)) == 0)
	{
	  values[len][num_set[len] >> lookback] = next_word[len];
	  memcpy (next_word[len], word, *word + 1);
	  if (escape && i == cfh->hd.num_codes - 1)
	    *escape = next_word[len];
	  next_word[len] += *word + 1;
	}
      else
	{
	  copy = prefixlen (last_word[len], word);
	  memcpy (next_word[len] + 1, word + copy + 1, *word - copy);
	  *next_word[len] = (copy << 4) + (*word - copy);
	  if (escape && i == cfh->hd.num_codes - 1)
	    *escape = next_word[len];
	  next_word[len] += (*word - copy) + 1;
	}
      memcpy (last_word[len], word, *word + 1);
      ++num_set[len];
    }
  if (cfh->hd.clens)
    delete []cfh->hd.clens;
  cfh->hd.clens = NULL;
  return values;
}

static int Load_Comp_HuffData(compression_dict &cd, int which, FILE *dict, 
			      huff_type type) {
  huff_data * hd;
  mg_u_long **   vals;

  if (!(hd = new huff_data))
    return 1;
  cd.MemForCompDict += sizeof (huff_data);
  if (Read_Huffman_Data (dict, hd, &cd.MemForCompDict, NULL) == -1)
    return 2;
  if (!(vals = Generate_Huffman_Vals (hd, &cd.MemForCompDict)))
    return 3;
  if (hd->clens)
    delete []hd->clens;
  hd->clens = NULL;
  if (type == chars)
    {
      cd.chars_huff[which] = hd;
      cd.chars_vals[which] = vals;
    }
  else
    {
      cd.lens_huff[which] = hd;
      cd.lens_vals[which] = vals;
    }

  return 0;
}

static int Load_Comp_FragsHeader(compression_dict &cd, int which, int getEscape,
				 FILE *dict) {
  if (!(cd.cfh[which] = new comp_frags_header))
    return 1;
  cd.MemForCompDict += sizeof (*cd.cfh[which]);
  if (Read_cfh (dict, cd.cfh[which], &cd.MemForCompDict, NULL) == -1)
    return 2;

  if (!(cd.values[which] = ReadInWords (dict, cd, cd.cfh[which],
					getEscape == 0 ? NULL : &cd.escape[which])))
    return 3;
  
  return 0;
}

static bool LoadSlowCompDict (FILE *dict, FILE *aux_dict, compression_dict &cd) {
  if (dict == NULL) return false;

  int which;

  memset (&cd, '\0', sizeof (compression_dict));

  cd.MemForCompDict = sizeof (compression_dict);

  if (Read_cdh (dict, &cd.cdh, &cd.MemForCompDict, NULL) == -1)
    return false;

  for (which = 0; which < 2; ++which)
    switch (cd.cdh.dict_type)
      {
      case MG_COMPLETE_DICTIONARY:
	{
	  if (Load_Comp_FragsHeader(cd, which, 0, dict) != 0)
	    return false;
	  cd.escape[which] = NULL;

	}
	break;
      case MG_PARTIAL_DICTIONARY:
	{
	  if (cd.cdh.num_words[which])
	    {
	      if (Load_Comp_FragsHeader(cd, which, 1, dict) != 0)
		return false;
	    }

	  if (Load_Comp_HuffData(cd, which, dict, chars) != 0)
	    return false;

	  if (Load_Comp_HuffData(cd, which, dict, lengths) != 0)
	    return false;
	}
	break;
      case MG_SEED_DICTIONARY:
	{
	  if (cd.cdh.num_words[which])
	    {
	      if (Load_Comp_FragsHeader(cd, which, 1, dict) != 0)
		return false;
	    }
	  switch (cd.cdh.novel_method)
	    {
	    case MG_NOVEL_HUFFMAN_CHARS:
	      if (Load_Comp_HuffData(cd, which, dict, chars) != 0)
		return false;

	      if (Load_Comp_HuffData(cd, which, dict, lengths) != 0)
		return false;
	      break;
	    case MG_NOVEL_DELTA:
	      break;
	    case MG_NOVEL_HYBRID:
	      break;
	    }
	  break;
	}
      }

  if (cd.cdh.novel_method == MG_NOVEL_DELTA ||
      cd.cdh.novel_method == MG_NOVEL_HYBRID)
    {
      if (!aux_dict)
	{
	  mg_errno = MG_NOFILE;
	  cd.Clear();
	  return false;
	}

      if (!(cd.ad = LoadAuxDict (cd, aux_dict)))
	{
	  cd.Clear();
	  return false;
	}
    }

  mg_errno = MG_NOERROR;

  cd.fast_loaded = 0;
  
  return true;
}



#define WORDNO(p, base) ((((char*)(p))-((char*)(base)))/sizeof(u_char*))
#define IS_FIXUP(p) ((fixup[WORDNO(p,cd)/8] & (1<<(WORDNO(p,cd) & 7))) != 0)

// fast loading really needs to be totally re-writen. "Unloading" the
// text data will currently cause a crash because memory is being
// deleted multiple times (and probably a zillion other reasons).
static bool LoadFastCompDict (FILE *text_fast_comp_dict, compression_dict &_cd) {
  if (text_fast_comp_dict == NULL) return false;
  
  uintptr_t *p, *end;
  u_char *fixup;
  mg_u_long mem;
  mg_u_long fixup_mem;
  int i;  /* [RPAP - Jan 97: Endian Ordering] */

  fread (&mem, sizeof (mem), 1, text_fast_comp_dict);
  NTOHUL(mem);  /* [RPAP - Jan 97: Endian Ordering] */
  fread (&fixup_mem, sizeof (fixup_mem), 1, text_fast_comp_dict);
  NTOHUL(fixup_mem);  /* [RPAP - Jan 97: Endian Ordering] */

  compression_dict *cd;
  if (!(cd = (compression_dict *)malloc (mem))) {
    mg_errno = MG_NOMEM;
    return false;
  }

  end = (uintptr_t *) (((u_char *) cd) + mem);
  fread (cd, sizeof (u_char), mem, text_fast_comp_dict);

  if (!(fixup = new u_char[fixup_mem]))
    {
      mg_errno = MG_NOMEM;
      return false;
    }

  fread (fixup, fixup_mem, sizeof (u_char), text_fast_comp_dict);

  for (p = (uintptr_t *) cd; (uintptr_t) p < (uintptr_t) end; ++p)
    if (IS_FIXUP (p))
      {
	NTOHUL(*p);  /* [RPAP - Jan 97: Endian Ordering] */
	*p = *p + (uintptr_t) cd;
      }

  /* [RPAP - Jan 97: Endian Ordering] */
  /* cdh */
  NTOHUL(cd->cdh.dict_type);
  NTOHUL(cd->cdh.novel_method);
  for (i = 0; i < TEXT_PARAMS; ++i)
    NTOHUL(cd->cdh.params[i]);
  NTOHUL(cd->cdh.num_words[0]);
  NTOHUL(cd->cdh.num_words[1]);
  NTOHUL(cd->cdh.num_word_chars[0]);
  NTOHUL(cd->cdh.num_word_chars[1]);
  NTOHUL(cd->cdh.lookback);
  /* cfh */
  for (i = 0; i <= 1; ++i)
    {
      int j;

      NTOHSI(cd->cfh[i]->hd.num_codes);
      NTOHSI(cd->cfh[i]->hd.mincodelen);
      NTOHSI(cd->cfh[i]->hd.maxcodelen);
      for (j = 0; j < MAX_HUFFCODE_LEN + 1; ++j)
	{
	  NTOHSI(cd->cfh[i]->hd.lencount[j]);
	  NTOHUL(cd->cfh[i]->hd.min_code[j]);
	}
      NTOHUL(cd->cfh[i]->uncompressed_size);
      for (j = 0; j < MAX_HUFFCODE_LEN + 1; ++j)
	NTOHUL(cd->cfh[i]->huff_words_size[j]);
    }
  NTOHUL(cd->MemForCompDict);
  /* ad */
  if (cd->cdh.novel_method == MG_NOVEL_DELTA ||
      cd->cdh.novel_method == MG_NOVEL_HYBRID)
    for (i = 0; i <= 1; ++i)
      {
	int j;
	
	NTOHUL(cd->ad->afh[i].num_frags);
	NTOHUL(cd->ad->afh[i].mem_for_frags);
	for (j = 0; j < 33; ++j)
	  {
	    NTOHSI(cd->ad->blk_start[i][j]);
	    NTOHSI(cd->ad->blk_end[i][j]);
	  }
      }
  NTOHSI(cd->fast_loaded);

  delete []fixup;

  // the whole fast comp dict is a bit of a hack so I don't
  // feel too bad about the next line :-) -- Rodger.
  _cd = *cd;
  
  return true;
}


static bool LoadCompDict (FILE *compDictFile,
			  FILE *auxDictFile,
			  FILE *fastCompDictFile,
			  compression_dict &cd) {
  // see if we have a fast loading compression dictionary
  if (fastCompDictFile != NULL)
    return LoadFastCompDict (fastCompDictFile, cd);
  
  // slow compression dictionary
  return LoadSlowCompDict (compDictFile, auxDictFile, cd);
}


// try to open the dictionary files and load the dictionary
static bool OpenLoadCompDict (char *textname, compression_dict &cd) {
  FILE *compDictFile = NULL;
  FILE *auxDictFile = NULL;
  FILE *fastCompDictFile = NULL;

  fastCompDictFile = open_file (textname, TEXT_DICT_FAST_SUFFIX,
				"rb", MAGIC_FAST_DICT, MG_CONTINUE);
  
  if (fastCompDictFile == NULL) {
    compDictFile = open_file (textname, TEXT_DICT_SUFFIX,
			      "rb", MAGIC_DICT, MG_MESSAGE);
    auxDictFile =  open_file (textname, TEXT_DICT_AUX_SUFFIX,
			      "rb", MAGIC_AUX_DICT, MG_CONTINUE);
  }
  
  bool res = LoadCompDict (compDictFile, auxDictFile, fastCompDictFile, cd);

  if (compDictFile != NULL) fclose (compDictFile);
  if (auxDictFile != NULL) fclose (auxDictFile);
  if (fastCompDictFile != NULL) fclose (fastCompDictFile);
  
  return res;
}

static bool LoadLevels (char *textname, FTextLevel &levels) {
  FILE *levelFile = NULL;

  // open the text level file
  levelFile = open_file (textname, TEXT_LEVEL_SUFFIX,
			 "rb", MAGIC_TEXT_LEVELS, MG_CONTINUE);
  if (levelFile == NULL) return false;

  // seek to the appropriate place and read the level information
  bool res = ((fseek (levelFile, sizeof (mg_u_long), SEEK_SET) == 0) &&
	      levels.Read (levelFile));

  // close the file
  fclose (levelFile);
  
  return res;
}


TextData::TextData () {
  // put file pointers in known state first
  textFile = NULL;
  textIdxFile = NULL;
  Clear ();
}

void TextData::Clear () {
  cd.Clear(); 
  textFile = NULL;
  textIdxFile = NULL;
  cth.Clear();
  levels.Clear();
}

bool TextData::LoadData (char *basepath, char *textname) {
  
  if (textname[0] == '\0') return false;
  
  // set the basepath
  set_basepath(basepath);
  
  // load the compression dictionary
  if (!OpenLoadCompDict (textname, cd)) return false;

  // open the compressed text and text index file
  textFile = open_file (textname, TEXT_SUFFIX, "rb", MAGIC_TEXT, MG_CONTINUE);
  if (textFile == NULL) return false;

  textIdxFile = open_file (textname, TEXT_IDX_SUFFIX, "rb", MAGIC_TEXI, MG_CONTINUE);
  if (textIdxFile == NULL) return false;

  // read in the compressed text header
  if ((fseek (textFile, sizeof (mg_u_long), SEEK_SET) != 0) || !cth.Read (textFile))
    return false;

  // read in the level information
  if (!LoadLevels (textname, levels)) return false;
  
  return true;
}

bool TextData::UnloadData () {
  // close any open files
  if (textFile != NULL) {
    fclose (textFile);
    textFile = NULL;
  }
  if (textIdxFile != NULL) {
    fclose (textIdxFile);
    textIdxFile = NULL;
  }

  // do general clear
  Clear ();
  
  return true;  
}


bool GetDocIdx (TextData &td, const UCArray &docLevel,
		mg_u_long docNum, TextIdx &docIdx) {
  // make sure the text index file was opened successfully
  if (td.textIdxFile == NULL) return false;

  // read in the index
  TextLevelInfo &levelInfo = td.levels.levelInfo[docLevel];
  if (!docIdx.Read (td.textIdxFile, levelInfo, docNum)) return false;
  
  return true;
}




#define MY_HUFF_DECODE(len, code, mcodes)				\
  do {									\
    register mg_u_long *__min_code = (mcodes);			\
    register mg_u_long *__mclen = __min_code;			\
    register mg_u_long __code = 0;					\
    do									\
      {									\
        __code += __code + buffer.bit();	                  	\
      }									\
    while (__code < *++__mclen);					\
    (len) = __mclen - __min_code;					\
    (code) = __code - *__mclen;						\
  } while(0);


bool GetDocText (TextData &td, const UCArray &docLevel,
		 mg_u_long docNum, UCArray &docText) {
  // erase the current text
  docText.erase (docText.begin(), docText.end());

  // look up the information about this document
  TextIdx docIdx;
  if (!GetDocIdx (td, docLevel, docNum, docIdx)) return false;

  // do seek to appropriate position
  stdio_bitio_buffer buffer (td.textFile);
  buffer.seek (docIdx.start.byte, docIdx.start.bit);
  
  // decompress the document
  compression_dict &cd = td.cd;
  auxiliary_dict *ad = cd.ad;
  int which = docIdx.which;
  mg_u_long num_bits = (docIdx.end.byte*8+(8-docIdx.end.bit)) -
    (docIdx.start.byte*8+(8-docIdx.start.bit));
  mg_u_long bits = 0;

  if (docText.capacity() < docText.size() + num_bits + 1) {
    docText.reserve(docText.size() + num_bits + 1);
  }
  // keep decoding bits until enough bits have been decoded
  while (bits < num_bits) {
    register unsigned code, len;
    register int r;
    register u_char *t, *b = NULL;
    u_char word[MAXWORDLEN + 1];
    
    if (cd.cfh[which]) {
      MY_HUFF_DECODE (len, code, cd.cfh[which]->hd.min_code);
      bits += len;
      
      r = code & ((1 << cd.cdh.lookback) - 1);
      t = cd.values[which][len][code >> cd.cdh.lookback];
      
      /* step through from base pointer */
      b = word + 1;
      while (r--) {
	register int copy = *t >> 4;
	memcpy (word + copy + 1, t + 1, *t & 0xf);
	word[0] = copy + (*t & 0xf);
	t += ((*t) & 0xf) + 1;
      }
    } else t = NULL;
    
    if (t == cd.escape[which]) {
      switch (cd.cdh.novel_method) {
      case MG_NOVEL_HUFFMAN_CHARS:
	{
	  int len, i;
	  int c;
	  len = buffer.huff_decode(cd.lens_huff[which]->min_code,
				   cd.lens_vals[which], &bits);
	  for (i = 0; i < len; ++i) {
	    c = buffer.huff_decode(cd.chars_huff[which]->min_code,
				   cd.chars_vals[which], &bits);
	    docText.push_back (c);
	  }
	}
	break;
      case MG_NOVEL_DELTA:
      case MG_NOVEL_HYBRID:
	{
	  int idx = 0, len;
	  u_char *base;
	  switch (cd.cdh.novel_method)
	    {
	    case MG_NOVEL_DELTA:
	      {
		idx = buffer.delta_decode (&bits);
		--idx;
	      }
	      break;
	    case MG_NOVEL_HYBRID:
	      {
		int k;
		k = buffer.gamma_decode (&bits);
		--k;
		idx = buffer.binary_decode(ad->blk_end[which][k] -
					   ad->blk_start[which][k] + 1, 
					   &bits);
		idx += ad->blk_start[which][k] - 1;
	      }
	      break;
	    }
	  base = ad->words[which][idx];
	  len = *base++;
	  for (; len; --len)
	    {
	      docText.push_back (*base++);
	    }
	}
	break;
      }
    }
    else
      {
	/* copy over the matching prefix */
	r = (*t >> 4);
	while (r--) {
	  docText.push_back (*b++);
	}
	
	/* and the stored suffix */
	r = ((*t) & 0xf);
	while (r--) {
	  docText.push_back (*++t);
	}
      }
    which = !which;
  }
  
  buffer.done();
  
  return true;
}