summaryrefslogtreecommitdiff
path: root/Build/source/libs/icu/icu-4.4/common/dictbe.cpp
blob: 9698893b9dc8e370b646e66d80b4080b51c7ad32 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
/**
 *******************************************************************************
 * Copyright (C) 2006-2008, International Business Machines Corporation and others. *
 * All Rights Reserved.                                                        *
 *******************************************************************************
 */

#include "unicode/utypes.h"

#if !UCONFIG_NO_BREAK_ITERATION

#include "brkeng.h"
#include "dictbe.h"
#include "unicode/uniset.h"
#include "unicode/chariter.h"
#include "unicode/ubrk.h"
#include "uvector.h"
#include "triedict.h"

U_NAMESPACE_BEGIN

/*
 ******************************************************************
 */

/*DictionaryBreakEngine::DictionaryBreakEngine() {
    fTypes = 0;
}*/

DictionaryBreakEngine::DictionaryBreakEngine(uint32_t breakTypes) {
    fTypes = breakTypes;
}

DictionaryBreakEngine::~DictionaryBreakEngine() {
}

UBool
DictionaryBreakEngine::handles(UChar32 c, int32_t breakType) const {
    return (breakType >= 0 && breakType < 32 && (((uint32_t)1 << breakType) & fTypes)
            && fSet.contains(c));
}

int32_t
DictionaryBreakEngine::findBreaks( UText *text,
                                 int32_t startPos,
                                 int32_t endPos,
                                 UBool reverse,
                                 int32_t breakType,
                                 UStack &foundBreaks ) const {
    int32_t result = 0;

    // Find the span of characters included in the set.
    int32_t start = (int32_t)utext_getNativeIndex(text);
    int32_t current;
    int32_t rangeStart;
    int32_t rangeEnd;
    UChar32 c = utext_current32(text);
    if (reverse) {
        UBool   isDict = fSet.contains(c);
        while((current = (int32_t)utext_getNativeIndex(text)) > startPos && isDict) {
            c = utext_previous32(text);
            isDict = fSet.contains(c);
        }
        rangeStart = (current < startPos) ? startPos : current+(isDict ? 0 : 1);
        rangeEnd = start + 1;
    }
    else {
        while((current = (int32_t)utext_getNativeIndex(text)) < endPos && fSet.contains(c)) {
            utext_next32(text);         // TODO:  recast loop for postincrement
            c = utext_current32(text);
        }
        rangeStart = start;
        rangeEnd = current;
    }
    if (breakType >= 0 && breakType < 32 && (((uint32_t)1 << breakType) & fTypes)) {
        result = divideUpDictionaryRange(text, rangeStart, rangeEnd, foundBreaks);
        utext_setNativeIndex(text, current);
    }
    
    return result;
}

void
DictionaryBreakEngine::setCharacters( const UnicodeSet &set ) {
    fSet = set;
    // Compact for caching
    fSet.compact();
}

/*void
DictionaryBreakEngine::setBreakTypes( uint32_t breakTypes ) {
    fTypes = breakTypes;
}*/

/*
 ******************************************************************
 */


// Helper class for improving readability of the Thai word break
// algorithm. The implementation is completely inline.

// List size, limited by the maximum number of words in the dictionary
// that form a nested sequence.
#define POSSIBLE_WORD_LIST_MAX 20

class PossibleWord {
 private:
  // list of word candidate lengths, in increasing length order
  int32_t   lengths[POSSIBLE_WORD_LIST_MAX];
  int       count;      // Count of candidates
  int32_t   prefix;     // The longest match with a dictionary word
  int32_t   offset;     // Offset in the text of these candidates
  int       mark;       // The preferred candidate's offset
  int       current;    // The candidate we're currently looking at

 public:
  PossibleWord();
  ~PossibleWord();
  
  // Fill the list of candidates if needed, select the longest, and return the number found
  int       candidates( UText *text, const TrieWordDictionary *dict, int32_t rangeEnd );
  
  // Select the currently marked candidate, point after it in the text, and invalidate self
  int32_t   acceptMarked( UText *text );
  
  // Back up from the current candidate to the next shorter one; return TRUE if that exists
  // and point the text after it
  UBool     backUp( UText *text );
  
  // Return the longest prefix this candidate location shares with a dictionary word
  int32_t   longestPrefix();
  
  // Mark the current candidate as the one we like
  void      markCurrent();
};

inline
PossibleWord::PossibleWord() {
    offset = -1;
}

inline
PossibleWord::~PossibleWord() {
}

inline int
PossibleWord::candidates( UText *text, const TrieWordDictionary *dict, int32_t rangeEnd ) {
    // TODO: If getIndex is too slow, use offset < 0 and add discardAll()
    int32_t start = (int32_t)utext_getNativeIndex(text);
    if (start != offset) {
        offset = start;
        prefix = dict->matches(text, rangeEnd-start, lengths, count, sizeof(lengths)/sizeof(lengths[0]));
        // Dictionary leaves text after longest prefix, not longest word. Back up.
        if (count <= 0) {
            utext_setNativeIndex(text, start);
        }
    }
    if (count > 0) {
        utext_setNativeIndex(text, start+lengths[count-1]);
    }
    current = count-1;
    mark = current;
    return count;
}

inline int32_t
PossibleWord::acceptMarked( UText *text ) {
    utext_setNativeIndex(text, offset + lengths[mark]);
    return lengths[mark];
}

inline UBool
PossibleWord::backUp( UText *text ) {
    if (current > 0) {
        utext_setNativeIndex(text, offset + lengths[--current]);
        return TRUE;
    }
    return FALSE;
}

inline int32_t
PossibleWord::longestPrefix() {
    return prefix;
}

inline void
PossibleWord::markCurrent() {
    mark = current;
}

// How many words in a row are "good enough"?
#define THAI_LOOKAHEAD 3

// Will not combine a non-word with a preceding dictionary word longer than this
#define THAI_ROOT_COMBINE_THRESHOLD 3

// Will not combine a non-word that shares at least this much prefix with a
// dictionary word, with a preceding word
#define THAI_PREFIX_COMBINE_THRESHOLD 3

// Ellision character
#define THAI_PAIYANNOI 0x0E2F

// Repeat character
#define THAI_MAIYAMOK 0x0E46

// Minimum word size
#define THAI_MIN_WORD 2

// Minimum number of characters for two words
#define THAI_MIN_WORD_SPAN (THAI_MIN_WORD * 2)

ThaiBreakEngine::ThaiBreakEngine(const TrieWordDictionary *adoptDictionary, UErrorCode &status)
    : DictionaryBreakEngine((1<<UBRK_WORD) | (1<<UBRK_LINE)),
      fDictionary(adoptDictionary)
{
    fThaiWordSet.applyPattern(UNICODE_STRING_SIMPLE("[[:Thai:]&[:LineBreak=SA:]]"), status);
    if (U_SUCCESS(status)) {
        setCharacters(fThaiWordSet);
    }
    fMarkSet.applyPattern(UNICODE_STRING_SIMPLE("[[:Thai:]&[:LineBreak=SA:]&[:M:]]"), status);
    fMarkSet.add(0x0020);
    fEndWordSet = fThaiWordSet;
    fEndWordSet.remove(0x0E31);             // MAI HAN-AKAT
    fEndWordSet.remove(0x0E40, 0x0E44);     // SARA E through SARA AI MAIMALAI
    fBeginWordSet.add(0x0E01, 0x0E2E);      // KO KAI through HO NOKHUK
    fBeginWordSet.add(0x0E40, 0x0E44);      // SARA E through SARA AI MAIMALAI
    fSuffixSet.add(THAI_PAIYANNOI);
    fSuffixSet.add(THAI_MAIYAMOK);

    // Compact for caching.
    fMarkSet.compact();
    fEndWordSet.compact();
    fBeginWordSet.compact();
    fSuffixSet.compact();
}

ThaiBreakEngine::~ThaiBreakEngine() {
    delete fDictionary;
}

int32_t
ThaiBreakEngine::divideUpDictionaryRange( UText *text,
                                                int32_t rangeStart,
                                                int32_t rangeEnd,
                                                UStack &foundBreaks ) const {
    if ((rangeEnd - rangeStart) < THAI_MIN_WORD_SPAN) {
        return 0;       // Not enough characters for two words
    }

    uint32_t wordsFound = 0;
    int32_t wordLength;
    int32_t current;
    UErrorCode status = U_ZERO_ERROR;
    PossibleWord words[THAI_LOOKAHEAD];
    UChar32 uc;
    
    utext_setNativeIndex(text, rangeStart);
    
    while (U_SUCCESS(status) && (current = (int32_t)utext_getNativeIndex(text)) < rangeEnd) {
        wordLength = 0;

        // Look for candidate words at the current position
        int candidates = words[wordsFound%THAI_LOOKAHEAD].candidates(text, fDictionary, rangeEnd);
        
        // If we found exactly one, use that
        if (candidates == 1) {
            wordLength = words[wordsFound%THAI_LOOKAHEAD].acceptMarked(text);
            wordsFound += 1;
        }
        
        // If there was more than one, see which one can take us forward the most words
        else if (candidates > 1) {
            // If we're already at the end of the range, we're done
            if ((int32_t)utext_getNativeIndex(text) >= rangeEnd) {
                goto foundBest;
            }
            do {
                int wordsMatched = 1;
                if (words[(wordsFound+1)%THAI_LOOKAHEAD].candidates(text, fDictionary, rangeEnd) > 0) {
                    if (wordsMatched < 2) {
                        // Followed by another dictionary word; mark first word as a good candidate
                        words[wordsFound%THAI_LOOKAHEAD].markCurrent();
                        wordsMatched = 2;
                    }
                    
                    // If we're already at the end of the range, we're done
                    if ((int32_t)utext_getNativeIndex(text) >= rangeEnd) {
                        goto foundBest;
                    }
                    
                    // See if any of the possible second words is followed by a third word
                    do {
                        // If we find a third word, stop right away
                        if (words[(wordsFound+2)%THAI_LOOKAHEAD].candidates(text, fDictionary, rangeEnd)) {
                            words[wordsFound%THAI_LOOKAHEAD].markCurrent();
                            goto foundBest;
                        }
                    }
                    while (words[(wordsFound+1)%THAI_LOOKAHEAD].backUp(text));
                }
            }
            while (words[wordsFound%THAI_LOOKAHEAD].backUp(text));
foundBest:
            wordLength = words[wordsFound%THAI_LOOKAHEAD].acceptMarked(text);
            wordsFound += 1;
        }
        
        // We come here after having either found a word or not. We look ahead to the
        // next word. If it's not a dictionary word, we will combine it withe the word we
        // just found (if there is one), but only if the preceding word does not exceed
        // the threshold.
        // The text iterator should now be positioned at the end of the word we found.
        if ((int32_t)utext_getNativeIndex(text) < rangeEnd && wordLength < THAI_ROOT_COMBINE_THRESHOLD) {
            // if it is a dictionary word, do nothing. If it isn't, then if there is
            // no preceding word, or the non-word shares less than the minimum threshold
            // of characters with a dictionary word, then scan to resynchronize
            if (words[wordsFound%THAI_LOOKAHEAD].candidates(text, fDictionary, rangeEnd) <= 0
                  && (wordLength == 0
                      || words[wordsFound%THAI_LOOKAHEAD].longestPrefix() < THAI_PREFIX_COMBINE_THRESHOLD)) {
                // Look for a plausible word boundary
                //TODO: This section will need a rework for UText.
                int32_t remaining = rangeEnd - (current+wordLength);
                UChar32 pc = utext_current32(text);
                int32_t chars = 0;
                for (;;) {
                    utext_next32(text);
                    uc = utext_current32(text);
                    // TODO: Here we're counting on the fact that the SA languages are all
                    // in the BMP. This should get fixed with the UText rework.
                    chars += 1;
                    if (--remaining <= 0) {
                        break;
                    }
                    if (fEndWordSet.contains(pc) && fBeginWordSet.contains(uc)) {
                        // Maybe. See if it's in the dictionary.
                        // NOTE: In the original Apple code, checked that the next
                        // two characters after uc were not 0x0E4C THANTHAKHAT before
                        // checking the dictionary. That is just a performance filter,
                        // but it's not clear it's faster than checking the trie.
                        int candidates = words[(wordsFound+1)%THAI_LOOKAHEAD].candidates(text, fDictionary, rangeEnd);
                        utext_setNativeIndex(text, current+wordLength+chars);
                        if (candidates > 0) {
                            break;
                        }
                    }
                    pc = uc;
                }
                
                // Bump the word count if there wasn't already one
                if (wordLength <= 0) {
                    wordsFound += 1;
                }
                
                // Update the length with the passed-over characters
                wordLength += chars;
            }
            else {
                // Back up to where we were for next iteration
                utext_setNativeIndex(text, current+wordLength);
            }
        }
        
        // Never stop before a combining mark.
        int32_t currPos;
        while ((currPos = (int32_t)utext_getNativeIndex(text)) < rangeEnd && fMarkSet.contains(utext_current32(text))) {
            utext_next32(text);
            wordLength += (int32_t)utext_getNativeIndex(text) - currPos;
        }
        
        // Look ahead for possible suffixes if a dictionary word does not follow.
        // We do this in code rather than using a rule so that the heuristic
        // resynch continues to function. For example, one of the suffix characters
        // could be a typo in the middle of a word.
        if ((int32_t)utext_getNativeIndex(text) < rangeEnd && wordLength > 0) {
            if (words[wordsFound%THAI_LOOKAHEAD].candidates(text, fDictionary, rangeEnd) <= 0
                && fSuffixSet.contains(uc = utext_current32(text))) {
                if (uc == THAI_PAIYANNOI) {
                    if (!fSuffixSet.contains(utext_previous32(text))) {
                        // Skip over previous end and PAIYANNOI
                        utext_next32(text);
                        utext_next32(text);
                        wordLength += 1;            // Add PAIYANNOI to word
                        uc = utext_current32(text);     // Fetch next character
                    }
                    else {
                        // Restore prior position
                        utext_next32(text);
                    }
                }
                if (uc == THAI_MAIYAMOK) {
                    if (utext_previous32(text) != THAI_MAIYAMOK) {
                        // Skip over previous end and MAIYAMOK
                        utext_next32(text);
                        utext_next32(text);
                        wordLength += 1;            // Add MAIYAMOK to word
                    }
                    else {
                        // Restore prior position
                        utext_next32(text);
                    }
                }
            }
            else {
                utext_setNativeIndex(text, current+wordLength);
            }
        }
        
        // Did we find a word on this iteration? If so, push it on the break stack
        if (wordLength > 0) {
            foundBreaks.push((current+wordLength), status);
        }
    }
    
    // Don't return a break for the end of the dictionary range if there is one there.
    if (foundBreaks.peeki() >= rangeEnd) {
        (void) foundBreaks.popi();
        wordsFound -= 1;
    }

    return wordsFound;
}

U_NAMESPACE_END

#endif /* #if !UCONFIG_NO_BREAK_ITERATION */