source: trunk/gsdl/src/mgpp/text/invf.cpp@ 1300

/**************************************************************************
 *
 * invf.cpp -- Data structures for inverted files
 * 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.
 *
 * $Id: invf.cpp 1300 2000-07-24 02:46:11Z kjm18 $
 *
 **************************************************************************/

#include "invf.h"
#include "UCArray.h"


invf_dict_header::invf_dict_header () {
  Clear();
}

invf_dict_header::~invf_dict_header () {
}

void invf_dict_header::Clear() {
  lookback = 0;
  word_dict_start = 0;
  word_dict_size = 0;
  tag_dict_start = 0;
  tag_dict_size = 0;
  num_docs = 0;
  num_frags = 0;
  num_words = 0;
  total_bytes = 0;
  index_string_bytes = 0;
  num_levels = 0;
}

bool invf_dict_header::Read (FILE *f) {
  return (ReadUL (f, lookback) &&
      ReadUL (f, word_dict_start) &&
      ReadUL (f, word_dict_size) &&
      ReadUL (f, tag_dict_start) &&
      ReadUL (f, tag_dict_size) &&
      ReadUL (f, num_docs) &&
      ReadUL (f, num_frags) &&
      ReadUL (f, num_words) &&
      ReadUL (f, total_bytes) &&
      ReadUL (f, index_string_bytes) &&
      ReadUL (f, num_levels));
}

bool invf_dict_header::Write (FILE *f) const {
  return (WriteUL (f, lookback) &&
      WriteUL (f, word_dict_start) &&
      WriteUL (f, word_dict_size) &&
      WriteUL (f, tag_dict_start) &&
      WriteUL (f, tag_dict_size) &&
      WriteUL (f, num_docs) &&
      WriteUL (f, num_frags) &&
      WriteUL (f, num_words) &&
      WriteUL (f, total_bytes) &&
      WriteUL (f, index_string_bytes) &&
      WriteUL (f, num_levels));
}



void dict_el::Clear () {
  el.erase (el.begin(), el.end());
  frag_occur = 0;
  freq = 0;
}

bool dict_el::Read (FILE *f) {
  return (ReadPreSufStr (f, el) &&
      ReadUL (f, frag_occur) &&
      ReadUL (f, freq));
}

bool dict_el::Write (FILE *f, const UCArray *lastEl) const {
  return (WritePreSufStr (f, lastEl, el) &&
      WriteUL (f, frag_occur) &&
      WriteUL (f, freq));
}


void word_dict_el::Clear () {
  dict_el::Clear();
  if (levelFreqs != NULL) delete [] levelFreqs;
  levelFreqs = NULL;
}

word_dict_el::~word_dict_el () {
  if (levelFreqs != NULL) delete [] levelFreqs;
}

void word_dict_el::SetNumLevels (unsigned long numLevels) {
  if (levelFreqs != NULL) delete [] levelFreqs;
  levelFreqs = new unsigned long [numLevels];
}
  
bool word_dict_el::Read (FILE *f, unsigned long numLevels) {
  if (!dict_el::Read (f)) return false;

  if (levelFreqs == NULL) return false;
  
  unsigned long i;
  for (i=0; i<numLevels; i++) {
    if (!ReadUL (f, levelFreqs[i])) return false;
  }

  return true;
}

bool word_dict_el::Write (FILE *f, const UCArray *lastEl,
              unsigned long numLevels) const {
  if (!dict_el::Write (f, lastEl)) return false;

  if (levelFreqs == NULL) return false;

  unsigned long i;
  for (i=0; i<numLevels; i++) {
    if (!WriteUL (f, levelFreqs[i])) return false;
  }

  return true;
}



block_dict_header::block_dict_header () {
  Clear();
}

void block_dict_header::Clear () {
  invf_dict_header::Clear();

  entries_per_wblk = 0;
  num_wblks = 0;
  max_wblk_size = 0;
  wblk_start = 0;
  wblk_idx_start = 0;
  
  entries_per_tblk = 0;
  num_tblks = 0;
  max_tblk_size = 0;
  tblk_start = 0;
  tblk_idx_start = 0;
}

bool block_dict_header::Read (FILE *f) {
  return (invf_dict_header::Read (f) &&

      ReadUL (f, entries_per_wblk) &&
      ReadUL (f, num_wblks) &&
      ReadUL (f, max_wblk_size) &&
      ReadUL (f, wblk_start) &&
      ReadUL (f, wblk_idx_start) &&
  
      ReadUL (f, entries_per_tblk) &&
      ReadUL (f, num_tblks) &&
      ReadUL (f, max_tblk_size) &&
      ReadUL (f, tblk_start) &&
      ReadUL (f, tblk_idx_start));
}

bool block_dict_header::Write (FILE *f) const {
  return (invf_dict_header::Write (f) &&

      WriteUL (f, entries_per_wblk) &&
      WriteUL (f, num_wblks) &&
      WriteUL (f, max_wblk_size) &&
      WriteUL (f, wblk_start) &&
      WriteUL (f, wblk_idx_start) &&
  
      WriteUL (f, entries_per_tblk) &&
      WriteUL (f, num_tblks) &&
      WriteUL (f, max_tblk_size) &&
      WriteUL (f, tblk_start) &&
      WriteUL (f, tblk_idx_start));
}





void block_dict_el::Clear () {
  el.erase (el.begin(), el.end());
  frag_occur = 0;
  freq = 0;
  invf_ptr = 0;
}

bool block_dict_el::Read (FILE *f) {
  return (ReadPreSufStr (f, el) &&
      ReadUL (f, frag_occur) &&
      ReadUL (f, freq) &&
      ReadUL (f, invf_ptr));
}

bool block_dict_el::Write (FILE *f, const UCArray *lastEl) const {
  return (WritePreSufStr (f, lastEl, el) &&
      WriteUL (f, frag_occur) &&
      WriteUL (f, freq) &&
      WriteUL (f, invf_ptr));
}





void word_block_dict_el::Clear () {
  block_dict_el::Clear();
  if (levelFreqs != NULL) delete [] levelFreqs;
  levelFreqs = NULL;
}

word_block_dict_el::~word_block_dict_el () {
  if (levelFreqs != NULL) delete [] levelFreqs;
}

void word_block_dict_el::SetNumLevels (unsigned long numLevels) {
  if (levelFreqs != NULL) delete [] levelFreqs;
  levelFreqs = new unsigned long [numLevels];
}

bool word_block_dict_el::Read (FILE *f, unsigned long numLevels) {
  if (!block_dict_el::Read (f)) return false;
  
  if (levelFreqs == NULL) return false;
  
  unsigned long i;
  for (i=0; i<numLevels; i++) {
    if (!ReadUL (f, levelFreqs[i])) return false;
  }

  return true;
}

bool word_block_dict_el::Write (FILE *f, const UCArray *lastEl,
                unsigned long numLevels) const {
  if (!block_dict_el::Write (f, lastEl)) return false;
  
  if (levelFreqs == NULL) return false;
  
  unsigned long i;
  for (i=0; i<numLevels; i++) {
    if (!WriteUL (f, levelFreqs[i])) return false;
  }
  
  return true;
}




block_idx_info::block_idx_info () {
  Clear ();
}

void block_idx_info::Clear () {
  el.erase (el.begin(), el.end());
  block_ptr = 0;
}

bool block_idx_info::Read (FILE *f) {
  return (ReadUCArray (f, el) &&
      ReadUL (f, block_ptr));
}

bool block_idx_info::Write (FILE *f) const {
  return (WriteUCArray (f, el) &&
      WriteUL (f, block_ptr));
}


bool ReadBlockIdx (FILE *f, block_idx &blockIdx) {
  blockIdx.erase (blockIdx.begin(), blockIdx.end());
  
  // read in the array size
  unsigned long arraySize = 0;
  if (!ReadVarLenUL (f, arraySize)) return false;
  
  // read in the array
  block_idx_info bi;
  while (arraySize > 0) {
    if (!bi.Read (f)) return false;
    blockIdx.push_back (bi);
    
    arraySize--;
  }
  
  return true;
}

bool WriteBlockIdx (FILE *f, const block_idx &blockIdx) {
  // write out the array size
  if (!WriteVarLenUL (f, blockIdx.size())) return false;

  block_idx::const_iterator here = blockIdx.begin();
  block_idx::const_iterator end = blockIdx.end();
  while (here != end) {
    if (!(*here).Write (f)) return false;
    here++;
  }

  return true;
}




stem_idx_header::stem_idx_header () {
  Clear ();
}

void stem_idx_header::Clear () {
  lookback = 0;
  dict_size = 0;

  entries_per_block = 0;
  num_blocks = 0;
  max_block_size = 0;
  blocks_start = 0;
  block_idx_start = 0;
  
  stemmer_num = 0;
  stem_method = 0;
}

bool stem_idx_header::Read (FILE *f) {
  return (ReadUL (f, lookback) &&
      ReadUL (f, dict_size) &&
      
      ReadUL (f, entries_per_block) &&
      ReadUL (f, num_blocks) &&
      ReadUL (f, max_block_size) &&
      ReadUL (f, blocks_start) &&
      ReadUL (f, block_idx_start) &&
      
      ReadUL (f, stemmer_num) &&
      ReadUL (f, stem_method));
}

bool stem_idx_header::Write (FILE *f) const {
  return (WriteUL (f, lookback) &&
      WriteUL (f, dict_size) &&
      
      WriteUL (f, entries_per_block) &&
      WriteUL (f, num_blocks) &&
      WriteUL (f, max_block_size) &&
      WriteUL (f, blocks_start) &&
      WriteUL (f, block_idx_start) &&
      
      WriteUL (f, stemmer_num) &&
      WriteUL (f, stem_method));
}



stem_block_dict_el::stem_block_dict_el () {
  Clear ();
}

void stem_block_dict_el::Clear () {
  el.erase (el.begin(), el.end());
  equivWords.erase (equivWords.begin(), equivWords.end());
}

bool stem_block_dict_el::Read (FILE *f) {
  equivWords.erase (equivWords.begin(), equivWords.end());
  
  if (!ReadPreSufStr (f, el)) return false;
  
  // read in the array size
  unsigned long arraySize = 0;
  if (!ReadVarLenUL (f, arraySize)) return false;
  
  // read in the array
  unsigned long wordNum;
  while (arraySize > 0) {
    if (!ReadUL (f, wordNum)) return false;
    equivWords.push_back (wordNum);
    
    arraySize--;
  }
  
  return true;
}

bool stem_block_dict_el::Write (FILE *f, const UCArray *lastEl) const {
  if (!WritePreSufStr (f, lastEl, el)) return false;

  // write out the array size
  if (!WriteVarLenUL (f, equivWords.size())) return false;

  vector<unsigned long>::const_iterator here = equivWords.begin();
  vector<unsigned long>::const_iterator end = equivWords.end();
  while (here != end) {
    if (!WriteUL (f, (*here))) return false;
    here++;
  }

  return true;
}




invf_file_header::invf_file_header () {
  Clear ();
}

void invf_file_header::Clear () {
  no_of_words = 0;
  no_of_tags = 0;
  skip_mode = SKIP_MODE_NO_SKIPS;
  word_level_index = 0;
  
  int i;
  for (i=0; i<16; i++) params[i] = 0;
}
  
bool invf_file_header::Read (FILE *f) {
  if (!ReadUL (f, no_of_words) ||
      !ReadUL (f, no_of_tags) ||
      !ReadUL (f, skip_mode) ||
      !ReadUL (f, word_level_index)) return false;

  int i;
  for (i=0; i<16; i++) {
    if (!ReadUL (f, params[i])) return false;
  }

  return true;
}

bool invf_file_header::Write (FILE *f) const {
  if (!WriteUL (f, no_of_words) ||
      !WriteUL (f, no_of_tags) ||
      !WriteUL (f, skip_mode) ||
      !WriteUL (f, word_level_index)) return false;

  int i;
  for (i=0; i<16; i++) {
    if (!WriteUL (f, params[i])) return false;
  }

  return true;
}





bool SearchElNum (const block_idx &bIdx,
          unsigned long entriesPerBlock,
          unsigned long elNum,
          unsigned long &blockIdxNum,
          unsigned long &blockStartElNum) {
  blockIdxNum = 0;
  blockStartElNum = 0;

  // make sure the element number isn't out of range
  if (elNum > bIdx.size()*entriesPerBlock) return false;

  blockIdxNum = elNum / entriesPerBlock;
  blockStartElNum = blockIdxNum * entriesPerBlock;
  
  return true;
}

bool SearchEl (const block_idx &bIdx,
           unsigned long entriesPerBlock,
           const UCArray &el,
           unsigned long &blockIdxNum,
           unsigned long &blockStartElNum) {
  blockIdxNum = 0;
  blockStartElNum = 0;

  unsigned long begin = 0;
  unsigned long bIdxEnd = bIdx.size();
  unsigned long end = bIdxEnd;
  unsigned long mid;
  while (begin < end) {
    mid = (begin+end)/2;
    if (DictCompare (el, bIdx[mid].el) < 0) {
      end = mid;
    } else if (mid+1 < bIdxEnd &&
           DictCompare (el, bIdx[mid+1].el) >= 0) {
      begin = mid+1;
    } else {
      blockIdxNum = mid;
      blockStartElNum = blockIdxNum * entriesPerBlock;
      return true;
    }
  }
  
  return false;
}



// use the block dictionary functions for tag entries, and word block dict
// functions for word entries.


bool SearchBlockDictElNum (FILE *dictFile,
               const block_idx &bIdx,
               unsigned long entriesPerBlock,
               unsigned long dictSize,
               unsigned long elNum,
               block_dict_el &dictEl) {
  UCArrayClear (dictEl.el);
  if (elNum >= dictSize) return false;

  // find the block that contains the element
  unsigned long blockIdxNum, curElNum;
  if (!SearchElNum (bIdx, entriesPerBlock, elNum,
            blockIdxNum, curElNum))
    return false;

  // look for the block
  fseek (dictFile, bIdx[blockIdxNum].block_ptr, SEEK_SET);
  do {
    dictEl.Read (dictFile);
  } while (curElNum++ < elNum);

  return true;
}

bool SearchBlockDictEl (FILE *dictFile,
            const block_idx &bIdx,
            unsigned long entriesPerBlock,
            unsigned long dictSize,
            const UCArray &el,
            block_dict_el &dictEl,
            unsigned long &elNum) {
  UCArrayClear (dictEl.el);

  // find the block that contains the element
  unsigned long blockIdxNum;
  if (!SearchEl (bIdx, entriesPerBlock, el,
         blockIdxNum, elNum))
    return false;

  unsigned long blockEndElNum = elNum + entriesPerBlock;
  if (blockEndElNum > dictSize) blockEndElNum = dictSize;

  // look for the block
  fseek (dictFile, bIdx[blockIdxNum].block_ptr, SEEK_SET);
  while (elNum < blockEndElNum) {
    dictEl.Read (dictFile);
    int res = DictCompare (dictEl.el, el);
    if (res == 0) return true; // found it
    else if (res > 0) break; // not here
    
    elNum++;
  }

  return false;
}




bool SearchWordBlockDictElNum (FILE *dictFile,
                   const block_idx &bIdx,
                   unsigned long entriesPerBlock,
                   unsigned long dictSize,
                   unsigned long numLevels,
                   unsigned long elNum,
                   word_block_dict_el &dictEl) {
  UCArrayClear (dictEl.el);
  if (elNum >= dictSize) return false;

  // find the block that contains the element
  unsigned long blockIdxNum, curElNum;
  if (!SearchElNum (bIdx, entriesPerBlock, elNum,
            blockIdxNum, curElNum))
    return false;

  // look for the block
  fseek (dictFile, bIdx[blockIdxNum].block_ptr, SEEK_SET);
  do {
    dictEl.Read (dictFile, numLevels);
  } while (curElNum++ < elNum);

  return true;
}

bool SearchWordBlockDictEl (FILE *dictFile,
                const block_idx &bIdx,
                unsigned long entriesPerBlock,
                unsigned long dictSize,
                unsigned long numLevels,
                const UCArray &el,
                word_block_dict_el &dictEl,
                unsigned long &elNum) {
  UCArrayClear (dictEl.el);

  // find the block that contains the element
  unsigned long blockIdxNum;
  if (!SearchEl (bIdx, entriesPerBlock, el,
         blockIdxNum, elNum))
    return false;

  unsigned long blockEndElNum = elNum + entriesPerBlock;
  if (blockEndElNum > dictSize) blockEndElNum = dictSize;

  // look for the block
  fseek (dictFile, bIdx[blockIdxNum].block_ptr, SEEK_SET);
  while (elNum < blockEndElNum) {
    dictEl.Read (dictFile, numLevels);
    int res = DictCompare (dictEl.el, el);
    if (res == 0) return true; // found it
    else if (res > 0) break; // not here
    
    elNum++;
  }

  return false;
}




bool SearchStemBlockDictElNum (FILE *dictFile,
                   const block_idx &bIdx,
                   unsigned long entriesPerBlock,
                   unsigned long dictSize,
                   unsigned long elNum,
                   stem_block_dict_el &dictEl) {
  UCArrayClear (dictEl.el);
  if (elNum >= dictSize) return false;

  // find the block that contains the element
  unsigned long blockIdxNum, curElNum;
  if (!SearchElNum (bIdx, entriesPerBlock, elNum,
            blockIdxNum, curElNum))
    return false;

  // look for the block
  fseek (dictFile, bIdx[blockIdxNum].block_ptr, SEEK_SET);
  do {
    dictEl.Read (dictFile);
  } while (curElNum++ < elNum);

  return true;
}

bool SearchStemBlockDictEl (FILE *dictFile,
                const block_idx &bIdx,
                unsigned long entriesPerBlock,
                unsigned long dictSize,
                const UCArray &el,
                stem_block_dict_el &dictEl,
                unsigned long &elNum) {
  UCArrayClear (dictEl.el);

  // find the block that contains the element
  unsigned long blockIdxNum;
  if (!SearchEl (bIdx, entriesPerBlock, el,
         blockIdxNum, elNum))
    return false;

  unsigned long blockEndElNum = elNum + entriesPerBlock;
  if (blockEndElNum > dictSize) blockEndElNum = dictSize;

  // look for the block
  fseek (dictFile, bIdx[blockIdxNum].block_ptr, SEEK_SET);
  while (elNum < blockEndElNum) {
    dictEl.Read (dictFile);
    int res = DictCompare (dictEl.el, el);
    if (res == 0) return true; // found it
    else if (res > 0) break; // not here
    
    elNum++;
  }

  return false;
}

//----------------------------------------------------------------
// functions for full text browse

bool NearestSearchWordBlockDictEl (FILE *dictFile,
                   const block_idx &bIdx,
                   unsigned long entriesPerBlock,
                   unsigned long dictSize,
                   unsigned long numLevels,
                   const UCArray &el,
                   word_block_dict_el &dictEl,
                   unsigned long &elNum) {

  UCArrayClear (dictEl.el);

  // find the block that contains the element
  unsigned long blockIdxNum;
  if (!SearchEl (bIdx, entriesPerBlock, el,
         blockIdxNum, elNum))
    return false;

  unsigned long blockEndElNum = elNum + entriesPerBlock;
  if (blockEndElNum > dictSize) blockEndElNum = dictSize;

  // look for the block
  fseek (dictFile, bIdx[blockIdxNum].block_ptr, SEEK_SET);
  while (elNum < blockEndElNum) {
    dictEl.Read (dictFile, numLevels);
    int res = DictCompare (el, dictEl.el); // look for the first word that is
                                     // greater or equal to the el
    if (res <= 0) {
      return true; // found one
   } 
    
    elNum++;
  }
  // it must be the last term
  return true;


}


bool SearchWordBlockDictElNumRange (FILE *dictFile,
                    const block_idx &bIdx,
                    unsigned long entriesPerBlock,
                    unsigned long dictSize,
                    unsigned long numLevels,
                    unsigned long elNum,
                    unsigned long numWords,
                    UCArrayVector &terms) {

  word_block_dict_el dictEl;  
  dictEl.SetNumLevels (numLevels);
  UCArrayClear(dictEl.el);

  terms.erase(terms.begin(), terms.end());

  if (elNum >= dictSize) return false;
  
  // find the block that contains the element
  unsigned long blockIdxNum, curElNum;
  if (!SearchElNum (bIdx, entriesPerBlock, elNum,
            blockIdxNum, curElNum))
    return false;

  unsigned long lastElNum = elNum + numWords - 1;
  if (lastElNum > dictSize) lastElNum = dictSize;

  // look for the block
  fseek (dictFile, bIdx[blockIdxNum].block_ptr, SEEK_SET);
  
  // get the first term
  do {
    dictEl.Read (dictFile, numLevels);
  } while (curElNum++ < elNum);

  terms.push_back(dictEl.el);
  while (curElNum <= lastElNum ) {
    dictEl.Read(dictFile, numLevels);
    terms.push_back(dictEl.el);
    curElNum++;
  }
  

  return true;
}

// NOte: before each addition of dictEl to the array, the level freqs array
// is deleted, as this was causing problems - generating a seg fault, I think if
// the vector had to be reallocated or something.
// setNumLevels has to be called each time before a read, now, to set up the level
//freqs array. this is necessary.
bool SearchWordBlockDictElNumRange (FILE *dictFile,
                    const block_idx &bIdx,
                    unsigned long entriesPerBlock,
                    unsigned long dictSize,
                    unsigned long numLevels,
                    unsigned long elNum,
                    unsigned long numWords,
                    word_block_dict_el_array &terms) {

  word_block_dict_el dictEl;
  dictEl.SetNumLevels (numLevels);
  UCArrayClear(dictEl.el);

  block_dict_el elem;
  terms.erase(terms.begin(), terms.end());

  if (elNum >= dictSize) return false;
  
  // find the block that contains the element
  unsigned long blockIdxNum, curElNum;
  if (!SearchElNum (bIdx, entriesPerBlock, elNum,
            blockIdxNum, curElNum))
    return false;

  unsigned long lastElNum = elNum + numWords - 1;
  if (lastElNum > dictSize) lastElNum = dictSize;

  // look for the block
  fseek (dictFile, bIdx[blockIdxNum].block_ptr, SEEK_SET);
  // get the first term
  do {
    dictEl.Read (dictFile, numLevels);
  } while (curElNum++ < elNum);

  dictEl.levelFreqs = NULL;
  terms.push_back(dictEl);

  while (curElNum <= lastElNum ) { 
    dictEl.SetNumLevels(numLevels);
    dictEl.Read(dictFile, numLevels);
    dictEl.levelFreqs = NULL;
    terms.push_back(dictEl);
    curElNum++;
  }

  return true;
}