1 | /**************************************************************************
|
---|
2 | *
|
---|
3 | * text.pass1.c -- Text compression (Pass 1)
|
---|
4 | * Copyright (C) 1994 Neil Sharman, Gary Eddy and Alistair Moffat
|
---|
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 | * $Id: text.pass1.c 7399 2004-05-24 21:10:52Z kjdon $
|
---|
21 | *
|
---|
22 | **************************************************************************/
|
---|
23 |
|
---|
24 | #include "sysfuncs.h"
|
---|
25 |
|
---|
26 | #include "memlib.h"
|
---|
27 | #include "messages.h"
|
---|
28 | #include "huffman.h"
|
---|
29 | #include "netorder.h" /* [RPAP - Jan 97: Endian Ordering] */
|
---|
30 |
|
---|
31 |
|
---|
32 | #include "mg_files.h"
|
---|
33 | #include "mg.h"
|
---|
34 | #include "build.h"
|
---|
35 | #include "locallib.h"
|
---|
36 | #include "words.h"
|
---|
37 | #include "text.h"
|
---|
38 | #include "hash.h"
|
---|
39 | #include "local_strings.h"
|
---|
40 | /* for more meaningful messages - jrm21 (gsdl) */
|
---|
41 | #ifdef HAVE_STRERROR
|
---|
42 | #include <string.h>
|
---|
43 | #include <errno.h>
|
---|
44 | #endif
|
---|
45 |
|
---|
46 | #define POOL_SIZE 1024*1024
|
---|
47 | #define INITIAL_HASH_SIZE 7927
|
---|
48 |
|
---|
49 |
|
---|
50 | typedef struct hash_rec
|
---|
51 | {
|
---|
52 | unsigned long wcnt; /* word frequency */
|
---|
53 | unsigned long occurance_num;
|
---|
54 | u_char *word;
|
---|
55 | }
|
---|
56 | hash_rec;
|
---|
57 |
|
---|
58 | typedef struct dict_data
|
---|
59 | {
|
---|
60 | hash_rec *HashTable;
|
---|
61 | unsigned long HashSize;
|
---|
62 | unsigned long HashUsed;
|
---|
63 | unsigned long wordnum;
|
---|
64 | unsigned long words_read;
|
---|
65 | unsigned long bytes_diff;
|
---|
66 | huff_data hd;
|
---|
67 | }
|
---|
68 | dict_data;
|
---|
69 |
|
---|
70 |
|
---|
71 |
|
---|
72 | static unsigned long LongestDoc = 0;
|
---|
73 | static unsigned long occurance_num = 0;
|
---|
74 | static dict_data DictData[2];
|
---|
75 |
|
---|
76 | static u_char *Pool;
|
---|
77 | static int PoolLeft;
|
---|
78 | static double inputbytes = 0; /* [RJM 07/97: 4G limit] */
|
---|
79 | static unsigned long MaxMemInUse = 0;
|
---|
80 | static unsigned long MemInUse = 0;
|
---|
81 | static compression_stats_header csh =
|
---|
82 | {0, 0.0}; /* [RJM 07/97: 4G limit] */
|
---|
83 |
|
---|
84 |
|
---|
85 | static void
|
---|
86 | ChangeMem (int Change)
|
---|
87 | {
|
---|
88 | MemInUse += Change;
|
---|
89 | if (MemInUse > MaxMemInUse)
|
---|
90 | MaxMemInUse = MemInUse;
|
---|
91 | }
|
---|
92 |
|
---|
93 |
|
---|
94 |
|
---|
95 |
|
---|
96 | int
|
---|
97 | init_text_1 (char *FileName)
|
---|
98 | {
|
---|
99 | int which;
|
---|
100 | if (!(Pool = Xmalloc (POOL_SIZE)))
|
---|
101 | {
|
---|
102 | Message ("Unable to allocate memory for pool");
|
---|
103 | return (COMPERROR);
|
---|
104 | }
|
---|
105 | PoolLeft = POOL_SIZE;
|
---|
106 | ChangeMem (POOL_SIZE);
|
---|
107 |
|
---|
108 | for (which = 1; which >= 0; which--)
|
---|
109 | {
|
---|
110 | u_char *word;
|
---|
111 | hash_rec *ent;
|
---|
112 | dict_data *dd = &DictData[which];
|
---|
113 |
|
---|
114 | dd->wordnum = 0;
|
---|
115 | dd->words_read = 0;
|
---|
116 | dd->bytes_diff = 0;
|
---|
117 | dd->HashSize = INITIAL_HASH_SIZE;
|
---|
118 | dd->HashUsed = 0;
|
---|
119 | if (!(dd->HashTable = Xmalloc (sizeof (hash_rec) * dd->HashSize)))
|
---|
120 | {
|
---|
121 | Message ("Unable to allocate memory for table");
|
---|
122 | return (COMPERROR);
|
---|
123 | }
|
---|
124 | ChangeMem (sizeof (hash_rec) * dd->HashSize);
|
---|
125 | bzero ((char *) (dd->HashTable), sizeof (hash_rec) * dd->HashSize);
|
---|
126 |
|
---|
127 | word = Pool;
|
---|
128 | *Pool++ = '\0';
|
---|
129 | PoolLeft--;
|
---|
130 | {
|
---|
131 | register u_char *wptr;
|
---|
132 | register int hsize = dd->HashSize;
|
---|
133 | register unsigned long hashval, step;
|
---|
134 |
|
---|
135 | HASH (hashval, step, word, hsize);
|
---|
136 | wptr = (dd->HashTable + hashval)->word;
|
---|
137 | while (wptr)
|
---|
138 | {
|
---|
139 | hashval += step;
|
---|
140 | if (hashval >= hsize)
|
---|
141 | hashval -= hsize;
|
---|
142 | wptr = (dd->HashTable + hashval)->word;
|
---|
143 | }
|
---|
144 | ent = dd->HashTable + hashval;
|
---|
145 | }
|
---|
146 | ent->wcnt = 1;
|
---|
147 | ent->word = word;
|
---|
148 | dd->HashUsed = 1;
|
---|
149 | }
|
---|
150 | return (COMPALLOK);
|
---|
151 | }
|
---|
152 |
|
---|
153 |
|
---|
154 |
|
---|
155 |
|
---|
156 | int
|
---|
157 | process_text_1 (u_char * s_in, int l_in)
|
---|
158 | {
|
---|
159 | int which;
|
---|
160 | u_char *end = s_in + l_in - 1;
|
---|
161 | if (l_in > LongestDoc)
|
---|
162 | LongestDoc = l_in;
|
---|
163 |
|
---|
164 | csh.num_docs++;
|
---|
165 | csh.num_bytes += l_in;
|
---|
166 |
|
---|
167 | which = inaword (s_in, end);
|
---|
168 | /*
|
---|
169 | ** Alternately parse off words and non-words from the input
|
---|
170 | ** stream beginning with a non-word. Each token is then
|
---|
171 | ** inserted into the set if it does not exist or has it's
|
---|
172 | ** frequency count incremented if it does.
|
---|
173 | */
|
---|
174 | for (; s_in <= end; which = !which)
|
---|
175 | {
|
---|
176 | u_char Word[MAXWORDLEN + 1];
|
---|
177 | dict_data *dd = &DictData[which];
|
---|
178 |
|
---|
179 | /* First parse a word or non-word out of the string */
|
---|
180 | if (which)
|
---|
181 | PARSE_WORD (Word, s_in, end);
|
---|
182 | else
|
---|
183 | PARSE_NON_WORD (Word, s_in, end);
|
---|
184 |
|
---|
185 | dd->wordnum++;
|
---|
186 | inputbytes += *Word;
|
---|
187 | dd->words_read++;
|
---|
188 |
|
---|
189 | /* Search the hash table for Word */
|
---|
190 | {
|
---|
191 | register unsigned long hashval, step;
|
---|
192 | register int hsize = dd->HashSize;
|
---|
193 | HASH (hashval, step, Word, hsize);
|
---|
194 | for (;;)
|
---|
195 | {
|
---|
196 | register u_char *s1;
|
---|
197 | register u_char *s2;
|
---|
198 | register int len;
|
---|
199 | register hash_rec *ent;
|
---|
200 | ent = dd->HashTable + hashval;
|
---|
201 | if (!ent->word)
|
---|
202 | {
|
---|
203 | int len = *Word + 1;
|
---|
204 | if (len > PoolLeft)
|
---|
205 | {
|
---|
206 | if (!(Pool = Xmalloc (POOL_SIZE)))
|
---|
207 | {
|
---|
208 | Message ("Unable to allocate memory for pool");
|
---|
209 | return (COMPERROR);
|
---|
210 | }
|
---|
211 | PoolLeft = POOL_SIZE;
|
---|
212 | ChangeMem (POOL_SIZE);
|
---|
213 | }
|
---|
214 | ent->occurance_num = occurance_num++;
|
---|
215 | ent->wcnt = 1;
|
---|
216 | ent->word = Pool;
|
---|
217 | memcpy (Pool, Word, len);
|
---|
218 | Pool += len;
|
---|
219 | PoolLeft -= len;
|
---|
220 | dd->HashUsed++;
|
---|
221 | dd->bytes_diff += Word[0];
|
---|
222 | break;
|
---|
223 | }
|
---|
224 | /* Compare the words */
|
---|
225 | s1 = Word;
|
---|
226 | s2 = ent->word;
|
---|
227 | len = *s1 + 1;
|
---|
228 | for (; len; len--)
|
---|
229 | if (*s1++ != *s2++)
|
---|
230 | break;
|
---|
231 |
|
---|
232 | if (len)
|
---|
233 | {
|
---|
234 | hashval = (hashval + step);
|
---|
235 | if (hashval >= hsize)
|
---|
236 | hashval -= hsize;
|
---|
237 | }
|
---|
238 | else
|
---|
239 | {
|
---|
240 | ent->wcnt++;
|
---|
241 | break;
|
---|
242 | }
|
---|
243 | }
|
---|
244 | }
|
---|
245 |
|
---|
246 |
|
---|
247 | if (dd->HashUsed >= dd->HashSize >> 1)
|
---|
248 | {
|
---|
249 | hash_rec *ht;
|
---|
250 | unsigned long size;
|
---|
251 | unsigned long i;
|
---|
252 | size = prime (dd->HashSize * 2);
|
---|
253 | if (!(ht = Xmalloc (sizeof (hash_rec) * size)))
|
---|
254 | {
|
---|
255 | Message ("Unable to allocate memory for table");
|
---|
256 | return (COMPERROR);
|
---|
257 | }
|
---|
258 | ChangeMem (sizeof (hash_rec) * size);
|
---|
259 | bzero ((char *) ht, sizeof (hash_rec) * size);
|
---|
260 |
|
---|
261 | for (i = 0; i < dd->HashSize; i++)
|
---|
262 | if (dd->HashTable[i].word)
|
---|
263 | {
|
---|
264 | register u_char *wptr;
|
---|
265 | register unsigned long hashval, step;
|
---|
266 |
|
---|
267 | wptr = dd->HashTable[i].word;
|
---|
268 | HASH (hashval, step, wptr, size);
|
---|
269 | wptr = (ht + hashval)->word;
|
---|
270 | while (wptr)
|
---|
271 | {
|
---|
272 | hashval += step;
|
---|
273 | if (hashval >= size)
|
---|
274 | hashval -= size;
|
---|
275 | wptr = (ht + hashval)->word;
|
---|
276 | }
|
---|
277 | ht[hashval] = dd->HashTable[i];
|
---|
278 | }
|
---|
279 | Xfree (dd->HashTable);
|
---|
280 | ChangeMem (-sizeof (hash_rec) * dd->HashSize);
|
---|
281 | dd->HashTable = ht;
|
---|
282 | dd->HashSize = size;
|
---|
283 |
|
---|
284 |
|
---|
285 | }
|
---|
286 | }
|
---|
287 | return (COMPALLOK);
|
---|
288 | } /* encode */
|
---|
289 |
|
---|
290 |
|
---|
291 |
|
---|
292 | static int
|
---|
293 | PackHashTable (dict_data * dd)
|
---|
294 | {
|
---|
295 | int s, d;
|
---|
296 | for (s = d = 0; s < dd->HashSize; s++)
|
---|
297 | if (dd->HashTable[s].word)
|
---|
298 | dd->HashTable[d++] = dd->HashTable[s];
|
---|
299 | ChangeMem (-sizeof (hash_rec) * dd->HashSize);
|
---|
300 | ChangeMem (sizeof (hash_rec) * dd->HashUsed);
|
---|
301 | if (!(dd->HashTable = Xrealloc (dd->HashTable,
|
---|
302 | sizeof (hash_rec) * dd->HashUsed)))
|
---|
303 | {
|
---|
304 | Message ("Out of memory");
|
---|
305 | return COMPERROR;
|
---|
306 | }
|
---|
307 | dd->HashSize = dd->HashUsed;
|
---|
308 | return COMPALLOK;
|
---|
309 | }
|
---|
310 |
|
---|
311 |
|
---|
312 |
|
---|
313 |
|
---|
314 |
|
---|
315 | static int
|
---|
316 | ent_comp (const void *s1, const void *s2)
|
---|
317 | {
|
---|
318 | return casecompare (((hash_rec *) s1)->word, ((hash_rec *) s2)->word);
|
---|
319 | }
|
---|
320 |
|
---|
321 |
|
---|
322 |
|
---|
323 | static void
|
---|
324 | WriteHashTable (FILE * fp, dict_data * dd)
|
---|
325 | {
|
---|
326 | frags_stats_header fsh;
|
---|
327 | u_long j = 0;
|
---|
328 | u_char *curr;
|
---|
329 |
|
---|
330 | if (PackHashTable (dd) == COMPERROR)
|
---|
331 | return;
|
---|
332 |
|
---|
333 | qsort (dd->HashTable, dd->HashUsed, sizeof (hash_rec), ent_comp);
|
---|
334 |
|
---|
335 | fsh.num_frags = dd->HashSize;
|
---|
336 | fsh.mem_for_frags = dd->HashSize;
|
---|
337 | for (j = 0; j < dd->HashSize; j++)
|
---|
338 | fsh.mem_for_frags += dd->HashTable[j].word[0];
|
---|
339 |
|
---|
340 | /* [RPAP - Jan 97: Endian Ordering] */
|
---|
341 | HTONUL(fsh.num_frags);
|
---|
342 | HTONUL(fsh.mem_for_frags);
|
---|
343 |
|
---|
344 | fwrite (&fsh, sizeof (fsh), 1, fp);
|
---|
345 |
|
---|
346 | for (j = 0; j < dd->HashSize; j++)
|
---|
347 | {
|
---|
348 | curr = dd->HashTable[j].word;
|
---|
349 |
|
---|
350 | /* [RPAP - Jan 97: Endian Ordering] */
|
---|
351 | HTONUL(dd->HashTable[j].wcnt);
|
---|
352 | HTONUL(dd->HashTable[j].occurance_num);
|
---|
353 |
|
---|
354 | fwrite (&dd->HashTable[j].wcnt, sizeof (dd->HashTable[j].wcnt), 1, fp);
|
---|
355 | fwrite (&dd->HashTable[j].occurance_num,
|
---|
356 | sizeof (dd->HashTable[j].occurance_num), 1, fp);
|
---|
357 |
|
---|
358 | /* [RPAP - Jan 97: Endian Ordering] */
|
---|
359 | NTOHUL(dd->HashTable[j].wcnt);
|
---|
360 | NTOHUL(dd->HashTable[j].occurance_num);
|
---|
361 |
|
---|
362 | fwrite (curr, sizeof (u_char), curr[0] + 1, fp);
|
---|
363 | }
|
---|
364 | }
|
---|
365 |
|
---|
366 |
|
---|
367 | int
|
---|
368 | done_text_1 (char *file_name)
|
---|
369 | {
|
---|
370 | char *temp_str;
|
---|
371 | FILE *fp;
|
---|
372 |
|
---|
373 | if (!(fp = create_file (file_name, TEXT_STATS_DICT_SUFFIX, "wb",
|
---|
374 | MAGIC_STATS_DICT, MG_MESSAGE))) /* [RPAP - Feb 97: WIN32 Port] */
|
---|
375 | {
|
---|
376 | fprintf(stderr,"Couldn't create file %s%s:%s\n",
|
---|
377 | file_name, TEXT_STATS_DICT_SUFFIX,
|
---|
378 | #if defined(HAVE_STRERROR) || defined(__WIN32__)
|
---|
379 | strerror(errno)
|
---|
380 | #else
|
---|
381 | " "
|
---|
382 | #endif
|
---|
383 | );
|
---|
384 | return COMPERROR;
|
---|
385 | }
|
---|
386 |
|
---|
387 | temp_str = msg_prefix;
|
---|
388 | msg_prefix = "text.pass1";
|
---|
389 |
|
---|
390 | /* [RPAP - Jan 97: Endian Ordering] */
|
---|
391 | HTONUL(csh.num_docs);
|
---|
392 | HTOND(csh.num_bytes); /* [RJM 07/97: 4G limit] */
|
---|
393 |
|
---|
394 | fwrite (&csh, sizeof (csh), 1, fp);
|
---|
395 |
|
---|
396 | /* [RPAP - Jan 97: Endian Ordering] */
|
---|
397 | NTOHUL(csh.num_docs);
|
---|
398 | NTOHD(csh.num_bytes); /* [RJM 07/97: 4G limit] */
|
---|
399 |
|
---|
400 | WriteHashTable (fp, &DictData[0]);
|
---|
401 | WriteHashTable (fp, &DictData[1]);
|
---|
402 | msg_prefix = temp_str;
|
---|
403 | fclose(fp);
|
---|
404 | return COMPALLOK;
|
---|
405 | } /* done_encode */
|
---|