[856] | 1 | /**************************************************************************
|
---|
| 2 | *
|
---|
| 3 | * mg_stem_idx.cpp -- stem index builder
|
---|
| 4 | * Copyright (C) 1999 Rodger McNab
|
---|
| 5 | *
|
---|
| 6 | * This program is free software; you can redistribute it and/or modify
|
---|
| 7 | * it under the terms of the GNU General Public License as published by
|
---|
| 8 | * the Free Software Foundation; either version 2 of the License, or
|
---|
| 9 | * (at your option) any later version.
|
---|
| 10 | *
|
---|
| 11 | * This program is distributed in the hope that it will be useful,
|
---|
| 12 | * but WITHOUT ANY WARRANTY; without even the implied warranty of
|
---|
| 13 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
---|
| 14 | * GNU General Public License for more details.
|
---|
| 15 | *
|
---|
| 16 | * You should have received a copy of the GNU General Public License
|
---|
| 17 | * along with this program; if not, write to the Free Software
|
---|
| 18 | * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
|
---|
| 19 | *
|
---|
| 20 | **************************************************************************/
|
---|
| 21 |
|
---|
| 22 |
|
---|
| 23 | #include "sysfuncs.h"
|
---|
| 24 | #include "messages.h"
|
---|
| 25 |
|
---|
| 26 | #include "mg_files.h"
|
---|
| 27 | #include "invf.h"
|
---|
| 28 | #include "UCArray.h"
|
---|
| 29 | #include "words.h"
|
---|
| 30 |
|
---|
| 31 | #include "stemmer.h"
|
---|
| 32 |
|
---|
| 33 |
|
---|
| 34 | #if defined(GSDL_USE_OBJECTSPACE)
|
---|
| 35 | # include <ospace\std\map>
|
---|
| 36 | # include <ospace\std\vector>
|
---|
| 37 | #elif defined(GSDL_USE_STL_H)
|
---|
| 38 | # include <map.h>
|
---|
| 39 | # include <vector.h>
|
---|
| 40 | #else
|
---|
| 41 | # include <map>
|
---|
| 42 | # include <vector>
|
---|
| 43 | #endif
|
---|
| 44 |
|
---|
| 45 |
|
---|
| 46 | /*
|
---|
| 47 | $Log$
|
---|
[858] | 48 | Revision 1.2 2000/01/14 02:45:51 sjboddie
|
---|
| 49 | fixed compiler warning
|
---|
| 50 |
|
---|
[856] | 51 | Revision 1.1 2000/01/14 02:26:20 sjboddie
|
---|
| 52 | Rodgers new C++ mg
|
---|
| 53 |
|
---|
| 54 | */
|
---|
| 55 |
|
---|
| 56 |
|
---|
| 57 | typedef vector<unsigned long> WordNumList;
|
---|
| 58 | typedef map<UCArray, WordNumList, DictLTUCArray> StemMapDict;
|
---|
| 59 |
|
---|
| 60 |
|
---|
| 61 | void CreateStemDict (char *filename,
|
---|
| 62 | StemMapDict &stemDict,
|
---|
| 63 | int stemMethod,
|
---|
| 64 | int stemmerNum) {
|
---|
| 65 | stemDict.erase (stemDict.begin(), stemDict.end());
|
---|
| 66 |
|
---|
| 67 | // open the dictionary
|
---|
| 68 | FILE *dictFile = open_file (filename, INVF_DICT_SUFFIX, "rb",
|
---|
| 69 | MAGIC_STEM_BUILD, MG_ABORT);
|
---|
| 70 | invf_dict_header idh;
|
---|
| 71 | idh.Read (dictFile);
|
---|
| 72 |
|
---|
| 73 | fseek (dictFile, idh.word_dict_start, SEEK_SET);
|
---|
| 74 |
|
---|
| 75 | unsigned long wordNum;
|
---|
| 76 | u_char mgWord[MAXSTEMLEN + 1];
|
---|
| 77 | word_dict_el wordEl;
|
---|
| 78 | UCArray stemEl;
|
---|
| 79 | wordEl.SetNumLevels (idh.num_levels);
|
---|
| 80 | for (wordNum=0; wordNum<idh.word_dict_size; wordNum++) {
|
---|
| 81 | // read in the next word
|
---|
| 82 | wordEl.Read (dictFile, idh.num_levels);
|
---|
| 83 |
|
---|
| 84 | // convert the word to an "mg word"
|
---|
| 85 | mgWord[0] = wordEl.el.size();
|
---|
| 86 | bcopy ((char *)wordEl.el.begin(), (char *)&mgWord[1], wordEl.el.size());
|
---|
| 87 |
|
---|
| 88 | // stem the word
|
---|
| 89 | stemmer (stemMethod, stemmerNum, mgWord);
|
---|
| 90 |
|
---|
| 91 | // convert the result back to a UCArray
|
---|
| 92 | stemEl.erase (stemEl.begin(), stemEl.end());
|
---|
| 93 | stemEl.insert (stemEl.end(), &mgWord[1], &mgWord[1] + mgWord[0]);
|
---|
| 94 |
|
---|
| 95 | // cout << "\"" << stemEl << "\" -> \"" << wordEl.el << "\"\n";
|
---|
| 96 |
|
---|
| 97 | // add this word number to the list of word numbers for this word
|
---|
| 98 | stemDict[stemEl].push_back (wordNum);
|
---|
| 99 | }
|
---|
| 100 |
|
---|
| 101 | fclose (dictFile);
|
---|
| 102 | }
|
---|
| 103 |
|
---|
| 104 |
|
---|
| 105 | void WriteStemDict (char *filename,
|
---|
| 106 | StemMapDict &stemDict,
|
---|
| 107 | int stemMethod,
|
---|
| 108 | int stemmerNum,
|
---|
| 109 | unsigned long entriesPerBlock) {
|
---|
| 110 | // Create appropriate stem index file
|
---|
[858] | 111 | FILE *stemDictFile = NULL;
|
---|
[856] | 112 | if (stemMethod == 1) {
|
---|
| 113 | stemDictFile = create_file (filename, INVF_DICT_BLOCKED_1_SUFFIX,
|
---|
| 114 | "wb", MAGIC_STEM_1, MG_ABORT);
|
---|
| 115 | } else if (stemMethod == 2) {
|
---|
| 116 | stemDictFile = create_file (filename, INVF_DICT_BLOCKED_2_SUFFIX,
|
---|
| 117 | "wb", MAGIC_STEM_2, MG_ABORT);
|
---|
| 118 | } else if (stemMethod == 3) {
|
---|
| 119 | stemDictFile = create_file (filename, INVF_DICT_BLOCKED_3_SUFFIX,
|
---|
| 120 | "wb", MAGIC_STEM_3, MG_ABORT);
|
---|
| 121 | } else {
|
---|
| 122 | FatalError (1, "Unknown stem method %d", stemMethod);
|
---|
| 123 | }
|
---|
| 124 |
|
---|
| 125 | stem_idx_header sih;
|
---|
| 126 | sih.lookback = 0;
|
---|
| 127 | sih.dict_size = stemDict.size();
|
---|
| 128 | sih.entries_per_block = entriesPerBlock;
|
---|
| 129 | sih.max_block_size = 0;
|
---|
| 130 |
|
---|
| 131 | sih.stemmer_num = stemmerNum;
|
---|
| 132 | sih.stem_method = stemMethod;
|
---|
| 133 |
|
---|
| 134 | // write out a place-holder version of the header
|
---|
| 135 | sih.Write (stemDictFile);
|
---|
| 136 |
|
---|
| 137 | sih.blocks_start = ftell (stemDictFile);
|
---|
| 138 |
|
---|
| 139 | block_idx stemIdx;
|
---|
| 140 | unsigned long stemNum = 0;
|
---|
| 141 | stem_block_dict_el stemEl;
|
---|
| 142 | UCArray lastEl;
|
---|
| 143 |
|
---|
| 144 | StemMapDict::const_iterator here = stemDict.begin();
|
---|
| 145 | StemMapDict::const_iterator end = stemDict.end();
|
---|
| 146 | while (here != end) {
|
---|
| 147 | // remember this stem (and position) if this is the start
|
---|
| 148 | // of a new block
|
---|
| 149 | if (stemNum % entriesPerBlock == 0) {
|
---|
| 150 | block_idx_info elIdx;
|
---|
| 151 | elIdx.el = (*here).first;
|
---|
| 152 | elIdx.block_ptr = ftell (stemDictFile);
|
---|
| 153 |
|
---|
| 154 | // see if this block is the longest so far
|
---|
| 155 | if (stemIdx.size() > 0) {
|
---|
| 156 | unsigned long blockLen = elIdx.block_ptr -
|
---|
| 157 | (*(stemIdx.end()-1)).block_ptr;
|
---|
| 158 | if (blockLen > sih.max_block_size) sih.max_block_size = blockLen;
|
---|
| 159 | }
|
---|
| 160 |
|
---|
| 161 | stemIdx.push_back (elIdx);
|
---|
| 162 | lastEl.erase (lastEl.begin(), lastEl.end()); // output full word
|
---|
| 163 | }
|
---|
| 164 |
|
---|
| 165 | // copy the information for this stem
|
---|
| 166 | stemEl.el = (*here).first;
|
---|
| 167 | stemEl.equivWords = (*here).second;
|
---|
| 168 |
|
---|
| 169 | // write out the stem
|
---|
| 170 | stemEl.Write (stemDictFile, &lastEl);
|
---|
| 171 |
|
---|
| 172 | here++; stemNum++;
|
---|
| 173 | }
|
---|
| 174 |
|
---|
| 175 |
|
---|
| 176 | // write out the element indexes
|
---|
| 177 | sih.num_blocks = stemIdx.size();
|
---|
| 178 | sih.block_idx_start = ftell (stemDictFile);
|
---|
| 179 | WriteBlockIdx (stemDictFile, stemIdx);
|
---|
| 180 |
|
---|
| 181 | // write out the stem dictionary header
|
---|
| 182 | fseek (stemDictFile, sizeof(unsigned long), SEEK_SET);
|
---|
| 183 | sih.Write (stemDictFile);
|
---|
| 184 |
|
---|
| 185 |
|
---|
| 186 | // close open files
|
---|
| 187 | fclose (stemDictFile);
|
---|
| 188 |
|
---|
| 189 | // print out information
|
---|
| 190 | Message ("Num word stems = %d\n", sih.dict_size);
|
---|
| 191 | Message ("Max stem block size = %d\n", sih.max_block_size);
|
---|
| 192 | Message ("Number of stem blocks written = %d\n", sih.num_blocks);
|
---|
| 193 | }
|
---|
| 194 |
|
---|
| 195 |
|
---|
| 196 | int main (int argc, char **argv) {
|
---|
| 197 | unsigned long entriesPerBlock = 16;
|
---|
| 198 | char *filename = "";
|
---|
| 199 | int ch;
|
---|
| 200 | int stemMethod = 0; // illegal value (no translation)
|
---|
| 201 | int stemmerNum = 0; // English stemmer
|
---|
| 202 | msg_prefix = argv[0];
|
---|
| 203 | opterr = 0;
|
---|
| 204 |
|
---|
| 205 | while ((ch = getopt (argc, argv, "f:d:b:s:h")) != -1) {
|
---|
| 206 | switch (ch) {
|
---|
| 207 | case 'f': // input file
|
---|
| 208 | filename = optarg;
|
---|
| 209 | break;
|
---|
| 210 | case 'd':
|
---|
| 211 | set_basepath (optarg);
|
---|
| 212 | break;
|
---|
| 213 | case 'b':
|
---|
| 214 | entriesPerBlock = atoi (optarg);
|
---|
| 215 | break;
|
---|
| 216 | case 's':
|
---|
| 217 | stemMethod = atoi (optarg);
|
---|
| 218 | break;
|
---|
| 219 | case 'a':
|
---|
| 220 | stemmerNum = stemmernumber ((unsigned char *) optarg);
|
---|
| 221 | break;
|
---|
| 222 | case 'h':
|
---|
| 223 | case '?':
|
---|
| 224 | fprintf (stderr, "usage: %s [-d directory] "
|
---|
| 225 | "[-b entries-per-block] [-h] -s 1|2|3 "
|
---|
| 226 | "[-a stemmer-method] -f name\n", argv[0]);
|
---|
| 227 | exit (1);
|
---|
| 228 | }
|
---|
| 229 | }
|
---|
| 230 |
|
---|
| 231 | if (stemMethod < 1 || stemMethod > 3)
|
---|
| 232 | FatalError (1, "Stem method must be 1, 2 or 3");
|
---|
| 233 |
|
---|
| 234 | // read in the dictionary and create the in memory dictionary
|
---|
| 235 | StemMapDict stemDict;
|
---|
| 236 | CreateStemDict (filename, stemDict, stemMethod, stemmerNum);
|
---|
| 237 |
|
---|
| 238 | // write out the dictionary as a blocked file
|
---|
| 239 | WriteStemDict (filename, stemDict, stemMethod, stemmerNum, entriesPerBlock);
|
---|
| 240 |
|
---|
| 241 | return 0;
|
---|
| 242 | }
|
---|