Parent: [0860b5] (diff)

Child: [4eaf12] (diff)

Download this file

textsplit.cpp    988 lines (890 with data), 28.0 kB

  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
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
/* Copyright (C) 2004 J.F.Dockes
* 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.,
* 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
*/
#ifndef TEST_TEXTSPLIT
#include "autoconfig.h"
#include <assert.h>
#include <iostream>
#include <string>
#include <set>
#include <cstring>
#include "textsplit.h"
#include "debuglog.h"
//#define UTF8ITER_CHECK
#include "utf8iter.h"
#include "uproplist.h"
#ifndef NO_NAMESPACES
using namespace std;
#endif /* NO_NAMESPACES */
/**
* Splitting a text into words. The code in this file works with utf-8
* in a semi-clean way (see uproplist.h). Ascii still gets special treatment.
*/
// Character classes: we have three main groups, and then some chars
// are their own class because they want special handling.
//
// We have an array with 256 slots where we keep the character types.
// The array could be fully static, but we use a small function to fill it
// once.
// The array is actually a remnant of the original version which did no utf8.
// Only the lower 127 slots are now used, but keep it at 256
// because it makes some tests in the code simpler.
const unsigned int charclasses_size = 256;
enum CharClass {LETTER=256, SPACE=257, DIGIT=258, WILD=259,
A_ULETTER=260, A_LLETTER=261};
static int charclasses[charclasses_size];
// Real UTF-8 characters are handled with sets holding all characters
// with interesting properties. This is far from full-blown management
// of Unicode properties, but seems to do the job well enough in most
// common cases
static set<unsigned int> unicign;
static set<unsigned int> visiblewhite;
class CharClassInit {
public:
CharClassInit()
{
unsigned int i;
// Set default value for all: SPACE
for (i = 0 ; i < 256 ; i ++)
charclasses[i] = SPACE;
char digits[] = "0123456789";
for (i = 0; i < strlen(digits); i++)
charclasses[int(digits[i])] = DIGIT;
char upper[] = "ABCDEFGHIJKLMNOPQRSTUVWXYZ";
for (i = 0; i < strlen(upper); i++)
charclasses[int(upper[i])] = A_ULETTER;
char lower[] = "abcdefghijklmnopqrstuvwxyz";
for (i = 0; i < strlen(lower); i++)
charclasses[int(lower[i])] = A_LLETTER;
char wild[] = "*?[]";
for (i = 0; i < strlen(wild); i++)
charclasses[int(wild[i])] = WILD;
char special[] = ".@+-,#'_\n\r";
for (i = 0; i < strlen(special); i++)
charclasses[int(special[i])] = special[i];
for (i = 0; i < sizeof(uniign) / sizeof(int); i++) {
unicign.insert(uniign[i]);
}
unicign.insert((unsigned int)-1);
for (i = 0; i < sizeof(avsbwht) / sizeof(int); i++) {
visiblewhite.insert(avsbwht[i]);
}
}
};
static const CharClassInit charClassInitInstance;
static inline int whatcc(unsigned int c)
{
if (c <= 127) {
return charclasses[c];
} else {
if (unicign.find(c) != unicign.end())
return SPACE;
else
return LETTER;
}
}
// CJK Unicode character detection:
//
// 2E80..2EFF; CJK Radicals Supplement
// 3000..303F; CJK Symbols and Punctuation
// 3040..309F; Hiragana
// 30A0..30FF; Katakana
// 3100..312F; Bopomofo
// 3130..318F; Hangul Compatibility Jamo
// 3190..319F; Kanbun
// 31A0..31BF; Bopomofo Extended
// 31C0..31EF; CJK Strokes
// 31F0..31FF; Katakana Phonetic Extensions
// 3200..32FF; Enclosed CJK Letters and Months
// 3300..33FF; CJK Compatibility
// 3400..4DBF; CJK Unified Ideographs Extension A
// 4DC0..4DFF; Yijing Hexagram Symbols
// 4E00..9FFF; CJK Unified Ideographs
// A700..A71F; Modifier Tone Letters
// AC00..D7AF; Hangul Syllables
// F900..FAFF; CJK Compatibility Ideographs
// FE30..FE4F; CJK Compatibility Forms
// FF00..FFEF; Halfwidth and Fullwidth Forms
// 20000..2A6DF; CJK Unified Ideographs Extension B
// 2F800..2FA1F; CJK Compatibility Ideographs Supplement
// Note: the p > 127 test is not necessary, but optimizes away the ascii case
#define UNICODE_IS_CJK(p) \
((p) > 127 && \
(((p) >= 0x2E80 && (p) <= 0x2EFF) || \
((p) >= 0x3000 && (p) <= 0x9FFF) || \
((p) >= 0xA700 && (p) <= 0xA71F) || \
((p) >= 0xAC00 && (p) <= 0xD7AF) || \
((p) >= 0xF900 && (p) <= 0xFAFF) || \
((p) >= 0xFE30 && (p) <= 0xFE4F) || \
((p) >= 0xFF00 && (p) <= 0xFFEF) || \
((p) >= 0x20000 && (p) <= 0x2A6DF) || \
((p) >= 0x2F800 && (p) <= 0x2FA1F)))
bool TextSplit::isCJK(int c)
{
return UNICODE_IS_CJK(c);
}
bool TextSplit::o_processCJK = true;
unsigned int TextSplit::o_CJKNgramLen = 2;
bool TextSplit::o_noNumbers = false;
// Do some checking (the kind which is simpler to do here than in the
// main loop), then send term to our client.
inline bool TextSplit::emitterm(bool isspan, string &w, int pos,
int btstart, int btend)
{
LOGDEB3(("TextSplit::emitterm: [%s] pos %d\n", w.c_str(), pos));
unsigned int l = w.length();
if (l > 0 && l < (unsigned)m_maxWordLength) {
// 1 byte word: we index single ascii letters and digits, but
// nothing else. We might want to turn this into a test for a
// single utf8 character instead ?
if (l == 1) {
int c = (int)w[0];
if (charclasses[c] != A_ULETTER && charclasses[c] != A_LLETTER &&
charclasses[c] != DIGIT) {
//cerr << "ERASING single letter term " << c << endl;
return true;
}
}
if (pos != m_prevpos || l != m_prevlen) {
bool ret = takeword(w, pos, btstart, btend);
m_prevpos = pos;
m_prevlen = w.length();
return ret;
}
LOGDEB2(("TextSplit::emitterm:dup: [%s] pos %d\n", w.c_str(), pos));
}
return true;
}
/**
* A routine called from different places in text_to_words(), to
* adjust the current state of the parser, and call the word
* handler/emitter. Emit and reset the current word, possibly emit the current
* span (if different). In query mode, words are not emitted, only final spans
*
* This is purely for factoring common code from different places in
* text_to_words().
*
* @return true if ok, false for error. Splitting should stop in this case.
* @param spanerase Set if the current span is at its end. Reset it.
* @param bp The current BYTE position in the stream
* @param spanemit This is set for intermediate spans: glue char changed.
*/
inline bool TextSplit::doemit(bool spanerase, int bp, bool spanemit)
{
LOGDEB2(("TextSplit::doemit: sper %d bp %d spem %d. spp %d wS %d wL %d "
"inn %d span [%s]\n",
spanerase, bp, spanemit, m_spanpos, m_wordStart, m_wordLen,
m_inNumber, m_span.c_str()));
// Emit span? When splitting for query, we only emit final spans
// (spanerase)
bool spanemitted = false;
if (!(m_flags & TXTS_NOSPANS) &&
!((m_wordLen == m_span.length()) &&
(o_noNumbers) && m_inNumber) &&
((spanemit && !(m_flags & TXTS_ONLYSPANS)) || spanerase) ) {
// Check for an acronym/abbreviation ie I.B.M.
if (spanerase && m_wordLen != m_span.length() && m_span.length() > 2
&& m_span.length() <= 20) {
bool acron = true;
for (unsigned int i = 1 ; i < m_span.length(); i += 2) {
if (m_span[i] != '.') {
acron = false;
break;
}
}
if (acron) {
string acronym;
for (unsigned int i = 0; i < m_span.length(); i += 2) {
acronym += m_span[i];
}
if (!emitterm(false, acronym, m_spanpos, bp - m_span.length(),
bp))
return false;
}
}
// Maybe trim at end. These are chars that we would keep inside
// a span, but not at the end
while (m_span.length() > 0) {
switch (m_span[m_span.length()-1]) {
case '.':
case '-':
case ',':
case '@':
case '_':
case '\'':
m_span.resize(m_span.length()-1);
if (--bp < 0)
bp = 0;
break;
default:
goto breakloop1;
}
}
breakloop1:
spanemitted = true;
if (!emitterm(true, m_span, m_spanpos, bp - m_span.length(), bp))
return false;
}
// Emit word if different from span and not 'no words' mode
if (!(m_flags & TXTS_ONLYSPANS) && m_wordLen &&
!(o_noNumbers && m_inNumber) &&
(!spanemitted || m_wordLen != m_span.length())) {
string s(m_span.substr(m_wordStart, m_wordLen));
if (!emitterm(false, s, m_wordpos, bp - m_wordLen, bp))
return false;
}
// Adjust state
if (m_wordLen) {
m_wordpos++;
m_wordLen = 0;
}
if (spanerase) {
discardspan();
} else {
m_wordStart = m_span.length();
}
return true;
}
void TextSplit::discardspan()
{
m_span.erase();
m_spanpos = m_wordpos;
m_wordStart = 0;
m_wordLen = 0;
}
/**
* Splitting a text into terms to be indexed.
* We basically emit a word every time we see a separator, but some chars are
* handled specially so that special cases, ie, c++ and jfd@recoll.com etc,
* are handled properly,
*/
bool TextSplit::text_to_words(const string &in)
{
LOGDEB1(("TextSplit::text_to_words: docjk %d (%d) %s%s%s [%s]\n",
o_processCJK, o_CJKNgramLen,
m_flags & TXTS_NOSPANS ? " nospans" : "",
m_flags & TXTS_ONLYSPANS ? " onlyspans" : "",
m_flags & TXTS_KEEPWILD ? " keepwild" : "",
in.substr(0,50).c_str()));
m_span.erase();
m_inNumber = false;
m_wordStart = m_wordLen = m_prevpos = m_prevlen = m_wordpos = m_spanpos = 0;
int curspanglue = 0;
// Running count of non-alphanum chars. Reset when we see one;
int nonalnumcnt = 0;
Utf8Iter it(in);
for (; !it.eof(); it++) {
unsigned int c = *it;
nonalnumcnt++;
if (c == (unsigned int)-1) {
LOGERR(("Textsplit: error occured while scanning UTF-8 string\n"));
return false;
}
if (o_processCJK && UNICODE_IS_CJK(c)) {
// CJK character hit.
// Do like at EOF with the current non-cjk data.
if (m_wordLen || m_span.length()) {
if (!doemit(true, it.getBpos()))
return false;
}
// Hand off situation to the cjk routine.
if (!cjk_to_words(&it, &c)) {
LOGERR(("Textsplit: scan error in cjk handler\n"));
return false;
}
// Check for eof, else c contains the first non-cjk
// character after the cjk sequence, just go on.
if (it.eof())
break;
}
int cc = whatcc(c);
switch (cc) {
case DIGIT:
if (m_wordLen == 0)
m_inNumber = true;
m_wordLen += it.appendchartostring(m_span);
nonalnumcnt = 0;
break;
case SPACE:
SPACE:
curspanglue = 0;
nonalnumcnt = 0;
if (m_wordLen || m_span.length()) {
if (!doemit(true, it.getBpos()))
return false;
m_inNumber = false;
}
break;
case WILD:
if (m_flags & TXTS_KEEPWILD)
goto NORMALCHAR;
else
goto SPACE;
break;
case '-':
case '+':
curspanglue = cc;
if (m_wordLen == 0) {
// + or - don't start a term except if this looks like
// it's going to be to be a number
if (whatcc(it[it.getCpos()+1]) == DIGIT) {
// -10
m_inNumber = true;
m_wordLen += it.appendchartostring(m_span);
} else {
goto SPACE;
}
} else if (m_inNumber && (m_span[m_span.length() - 1] == 'e' ||
m_span[m_span.length() - 1] == 'E')) {
if (whatcc(it[it.getCpos()+1]) == DIGIT) {
m_wordLen += it.appendchartostring(m_span);
} else {
goto SPACE;
}
} else {
if (!doemit(false, it.getBpos()))
return false;
m_inNumber = false;
m_wordStart += it.appendchartostring(m_span);
}
break;
case '.':
case ',':
{
// Need a little lookahead here. At worse this gets the end null
int nextc = it[it.getCpos()+1];
int nextwhat = whatcc(nextc);
if (m_inNumber) {
// we're eliminating 132.jpg here. Good idea ?
if (nextwhat != DIGIT && nextc != 'e' && nextc != 'E')
goto SPACE;
m_wordLen += it.appendchartostring(m_span);
curspanglue = cc;
break;
} else {
// If . inside a word, it's spanglue, else, it's whitespace.
// We also keep an initial '.' for catching .net, but this adds
// quite a few spurious terms !
// Another problem is that something like .x-errs
// will be split as .x-errs, x, errs but not x-errs
// A final comma in a word will be removed by doemit
// Only letters and digits make sense after
if (nextwhat != A_LLETTER && nextwhat != A_ULETTER &&
nextwhat != DIGIT && nextwhat != LETTER)
goto SPACE;
if (cc == '.') {
// Check for number like .1
if (m_span.length() == 0 && nextwhat == DIGIT) {
m_inNumber = true;
m_wordLen += it.appendchartostring(m_span);
curspanglue = cc;
break;
}
if (m_wordLen) {
// Disputable special case: set spanemit to
// true when encountering a '.' while spanglue
// is '_'. Think of a_b.c Done to
// avoid breaking stuff after changing '_'
// from wordchar to spanglue
if (!doemit(false, it.getBpos(), curspanglue == '_'))
return false;
curspanglue = cc;
// span length could have been adjusted by trimming
// inside doemit
if (m_span.length())
m_wordStart += it.appendchartostring(m_span);
break;
} else {
m_wordStart += it.appendchartostring(m_span);
curspanglue = cc;
break;
}
}
}
goto SPACE;
}
break;
case '@':
if (m_wordLen) {
if (!doemit(false, it.getBpos()))
return false;
curspanglue = cc;
m_inNumber = false;
}
m_wordStart += it.appendchartostring(m_span);
break;
case '_':
if (m_wordLen) {
if (!doemit(false, it.getBpos()))
return false;
curspanglue = cc;
m_inNumber = false;
}
m_wordStart += it.appendchartostring(m_span);
break;
case '\'':
// If in word, potential span: o'brien, else, this is more
// whitespace
if (m_wordLen) {
if (!doemit(false, it.getBpos()))
return false;
curspanglue = cc;
m_inNumber = false;
m_wordStart += it.appendchartostring(m_span);
}
break;
case '#':
// Keep it only at end of word ... Special case for c# you see...
if (m_wordLen > 0) {
int w = whatcc(it[it.getCpos()+1]);
if (w == SPACE || w == '\n' || w == '\r') {
m_wordLen += it.appendchartostring(m_span);
break;
}
}
goto SPACE;
break;
case '\n':
case '\r':
if (m_span.length() && m_span[m_span.length() - 1] == '-') {
// if '-' is the last char before end of line, just
// ignore the line change. This is the right thing to
// do almost always. We'd then need a way to check if
// the - was added as part of the word hyphenation, or was
// there in the first place, but this would need a dictionary.
// Also we'd need to check for a soft-hyphen and remove it,
// but this would require more utf-8 magic
} else {
// Handle like a normal separator
goto SPACE;
}
break;
#ifdef RCL_SPLIT_CAMELCASE
// Camelcase handling.
// If we get uppercase ascii after lowercase ascii, emit word.
// This emits "camel" when hitting the 'C' of camelCase
// Not enabled by defaults as this makes phrase searches quite
// confusing.
// ie "MySQL manual" is matched by "MySQL manual" and
// "my sql manual" but not "mysql manual"
case A_ULETTER:
if (m_span.length() &&
charclasses[(unsigned char)m_span[m_span.length() - 1]] ==
A_LLETTER) {
if (m_wordLen) {
if (!doemit(false, it.getBpos()))
return false;
}
}
goto NORMALCHAR;
// CamelCase handling.
// If we get lowercase after uppercase and the current
// word length is bigger than one, it means we had a
// string of several upper-case letters: an
// acronym (readHTML) or a single letter article (ALittleHelp).
// Emit the uppercase word before proceeding
case A_LLETTER:
if (m_span.length() &&
charclasses[(unsigned char)m_span[m_span.length() - 1]] ==
A_ULETTER && m_wordLen > 1) {
// Multiple upper-case letters. Single letter word
// or acronym which we want to emit now
m_wordLen--;
if (!doemit(false, it.getBpos()))
return false;
m_wordStart--;
m_wordLen++;
}
goto NORMALCHAR;
#endif /* CAMELCASE */
default:
NORMALCHAR:
if (m_inNumber && c != 'e' && c != 'E') {
m_inNumber = false;
}
m_wordLen += it.appendchartostring(m_span);
nonalnumcnt = 0;
break;
}
}
if (m_wordLen || m_span.length()) {
if (!doemit(true, it.getBpos()))
return false;
}
return true;
}
// Using an utf8iter pointer just to avoid needing its definition in
// textsplit.h
//
// We output ngrams for exemple for char input a b c and ngramlen== 2,
// we generate: a ab b bc c as words
//
// This is very different from the normal behaviour, so we don't use
// the doemit() and emitterm() routines
//
// The routine is sort of a mess and goes to show that we'd probably
// be better off converting the whole buffer to utf32 on entry...
bool TextSplit::cjk_to_words(Utf8Iter *itp, unsigned int *cp)
{
LOGDEB1(("cjk_to_words: m_wordpos %d\n", m_wordpos));
Utf8Iter &it = *itp;
// We use an offset buffer to remember the starts of the utf-8
// characters which we still need to use.
assert(o_CJKNgramLen < o_CJKMaxNgramLen);
unsigned int boffs[o_CJKMaxNgramLen+1];
// Current number of valid offsets;
unsigned int nchars = 0;
unsigned int c = 0;
for (; !it.eof(); it++) {
c = *it;
if (!UNICODE_IS_CJK(c)) {
// Return to normal handler
break;
}
if (whatcc(c) == SPACE) {
// Flush the ngram buffer and go on
nchars = 0;
continue;
}
if (nchars == o_CJKNgramLen) {
// Offset buffer full, shift it. Might be more efficient
// to have a circular one, but things are complicated
// enough already...
for (unsigned int i = 0; i < nchars-1; i++) {
boffs[i] = boffs[i+1];
}
} else {
nchars++;
}
// Take note of byte offset for this character.
boffs[nchars-1] = it.getBpos();
// Output all new ngrams: they begin at each existing position
// and end after the new character. onlyspans->only output
// maximum words, nospans=> single chars
if (!(m_flags & TXTS_ONLYSPANS) || nchars == o_CJKNgramLen) {
unsigned int btend = it.getBpos() + it.getBlen();
unsigned int loopbeg = (m_flags & TXTS_NOSPANS) ? nchars-1 : 0;
unsigned int loopend = (m_flags & TXTS_ONLYSPANS) ? 1 : nchars;
for (unsigned int i = loopbeg; i < loopend; i++) {
if (!takeword(it.buffer().substr(boffs[i],
btend-boffs[i]),
m_wordpos - (nchars-i-1), boffs[i], btend)) {
return false;
}
}
if ((m_flags & TXTS_ONLYSPANS)) {
// Only spans: don't overlap: flush buffer
nchars = 0;
}
}
// Increase word position by one, other words are at an
// existing position. This could be subject to discussion...
m_wordpos++;
}
// If onlyspans is set, there may be things to flush in the buffer
// first
if ((m_flags & TXTS_ONLYSPANS) && nchars > 0 && nchars != o_CJKNgramLen) {
unsigned int btend = it.getBpos(); // Current char is out
if (!takeword(it.buffer().substr(boffs[0], btend-boffs[0]),
m_wordpos - nchars,
boffs[0], btend)) {
return false;
}
}
m_span.erase();
m_inNumber = false;
m_wordStart = m_wordLen = m_prevpos = m_prevlen = 0;
m_spanpos = m_wordpos;
*cp = c;
return true;
}
// Specialization for countWords
class TextSplitCW : public TextSplit {
public:
int wcnt;
TextSplitCW(Flags flags) : TextSplit(flags), wcnt(0) {}
bool takeword(const string &, int, int, int) {
wcnt++;
return true;
}
};
int TextSplit::countWords(const string& s, TextSplit::Flags flgs)
{
TextSplitCW splitter(flgs);
splitter.text_to_words(s);
return splitter.wcnt;
}
bool TextSplit::hasVisibleWhite(const string &in)
{
Utf8Iter it(in);
for (; !it.eof(); it++) {
unsigned int c = (unsigned char)*it;
LOGDEB3(("TextSplit::hasVisibleWhite: testing 0x%04x\n", c));
if (c == (unsigned int)-1) {
LOGERR(("hasVisibleWhite: error while scanning UTF-8 string\n"));
return false;
}
if (visiblewhite.find(c) != visiblewhite.end())
return true;
}
return false;
}
template <class T> bool u8stringToStrings(const string &s, T &tokens)
{
Utf8Iter it(s);
string current;
tokens.clear();
enum states {SPACE, TOKEN, INQUOTE, ESCAPE};
states state = SPACE;
for (; !it.eof(); it++) {
unsigned int c = *it;
if (visiblewhite.find(c) != visiblewhite.end())
c = ' ';
LOGDEB3(("TextSplit::stringToStrings: 0x%04x\n", c));
if (c == (unsigned int)-1) {
LOGERR(("TextSplit::stringToStrings: error while "
"scanning UTF-8 string\n"));
return false;
}
switch (c) {
case '"':
switch(state) {
case SPACE: state = INQUOTE; continue;
case TOKEN: goto push_char;
case ESCAPE: state = INQUOTE; goto push_char;
case INQUOTE: tokens.push_back(current);current.clear();
state = SPACE; continue;
}
break;
case '\\':
switch(state) {
case SPACE:
case TOKEN: state=TOKEN; goto push_char;
case INQUOTE: state = ESCAPE; continue;
case ESCAPE: state = INQUOTE; goto push_char;
}
break;
case ' ':
case '\t':
case '\n':
case '\r':
switch(state) {
case SPACE: continue;
case TOKEN: tokens.push_back(current); current.clear();
state = SPACE; continue;
case INQUOTE:
case ESCAPE: goto push_char;
}
break;
default:
switch(state) {
case ESCAPE: state = INQUOTE; break;
case SPACE: state = TOKEN; break;
case TOKEN:
case INQUOTE: break;
}
push_char:
it.appendchartostring(current);
}
}
// End of string. Process residue, and possible error (unfinished quote)
switch(state) {
case SPACE: break;
case TOKEN: tokens.push_back(current); break;
case INQUOTE:
case ESCAPE: return false;
}
return true;
}
bool TextSplit::stringToStrings(const string &s, list<string> &tokens)
{
return u8stringToStrings<list<string> >(s, tokens);
}
#else // TEST driver ->
#include <unistd.h>
#include <errno.h>
#include <fcntl.h>
#include <stdio.h>
#include <iostream>
#include "textsplit.h"
#include "readfile.h"
#include "debuglog.h"
#include "transcode.h"
#include "unacpp.h"
#include "termproc.h"
using namespace std;
class myTermProc : public Rcl::TermProc {
int first;
bool nooutput;
public:
myTermProc() : TermProc(0), first(1), nooutput(false) {}
void setNoOut(bool val) {nooutput = val;}
virtual bool takeword(const string &term, int pos, int bs, int be)
{
if (nooutput)
return true;
FILE *fp = stdout;
if (first) {
fprintf(fp, "%3s %-20s %4s %4s\n", "pos", "Term", "bs", "be");
first = 0;
}
fprintf(fp, "%3d %-20s %4d %4d\n", pos, term.c_str(), bs, be);
return true;
}
};
static string teststring =
"Un bout de texte \nnormal. 2eme phrase.3eme;quatrieme.\n"
"\"Jean-Francois Dockes\" <jfd@okyz.com>\n"
"n@d @net .net t@v@c c# c++ o'brien 'o'brien' l'ami\n"
"data123\n"
"134 +134 -14 0.1 .1 2. -1.5 +1.5 1,2 1.54e10 1,2e30 .1e10 1.e-8\n"
"@^#$(#$(*)\n"
"192.168.4.1 one\n\rtwo\r"
"Debut-\ncontinue\n"
"[olala][ululu] (valeur) (23)\n"
"utf-8 ucs-4�� \\nodef\n"
"A b C 2 . +"
"','this\n"
" ,able,test-domain "
" -wl,--export-dynamic "
" ~/.xsession-errors "
;
static string teststring1 = " nouvel-an ";
static string thisprog;
static string usage =
" textsplit [opts] [filename]\n"
" -q : no output\n"
" -s : only spans\n"
" -w : only words\n"
" -n : no numbers\n"
" -k : preserve wildcards (?*)\n"
" -c : just count words\n"
" -u : use unac\n"
" -C [charset] : input charset\n"
" -S [stopfile] : stopfile to use for commongrams\n"
" if filename is 'stdin', will read stdin for data (end with ^D)\n"
" \n\n"
;
static void
Usage(void)
{
cerr << thisprog << ": usage:\n" << usage;
exit(1);
}
static int op_flags;
#define OPT_s 0x1
#define OPT_w 0x2
#define OPT_q 0x4
#define OPT_c 0x8
#define OPT_k 0x10
#define OPT_C 0x20
#define OPT_n 0x40
#define OPT_S 0x80
#define OPT_u 0x100
int main(int argc, char **argv)
{
string charset, stopfile;
thisprog = argv[0];
argc--; argv++;
while (argc > 0 && **argv == '-') {
(*argv)++;
if (!(**argv))
/* Cas du "adb - core" */
Usage();
while (**argv)
switch (*(*argv)++) {
case 'c': op_flags |= OPT_c; break;
case 'C': op_flags |= OPT_C; if (argc < 2) Usage();
charset = *(++argv); argc--;
goto b1;
case 'k': op_flags |= OPT_k; break;
case 'n': op_flags |= OPT_n; break;
case 'q': op_flags |= OPT_q; break;
case 's': op_flags |= OPT_s; break;
case 'S': op_flags |= OPT_S; if (argc < 2) Usage();
stopfile = *(++argv); argc--;
goto b1;
case 'u': op_flags |= OPT_u; break;
case 'w': op_flags |= OPT_w; break;
default: Usage(); break;
}
b1: argc--; argv++;
}
DebugLog::getdbl()->setloglevel(DEBDEB1);
DebugLog::setfilename("stderr");
TextSplit::Flags flags = TextSplit::TXTS_NONE;
if (op_flags&OPT_s)
flags = TextSplit::TXTS_ONLYSPANS;
else if (op_flags&OPT_w)
flags = TextSplit::TXTS_NOSPANS;
if (op_flags & OPT_k)
flags = (TextSplit::Flags)(flags | TextSplit::TXTS_KEEPWILD);
if (op_flags & OPT_n)
TextSplit::noNumbers();
Rcl::StopList stoplist;
if (op_flags & OPT_S) {
if (!stoplist.setFile(stopfile)) {
cerr << "Can't read stopfile: " << stopfile << endl;
exit(1);
}
}
string odata, reason;
if (argc == 1) {
const char *filename = *argv++; argc--;
if (!strcmp(filename, "stdin")) {
char buf[1024];
int nread;
while ((nread = read(0, buf, 1024)) > 0) {
odata.append(buf, nread);
}
} else if (!file_to_string(filename, odata, &reason)) {
cerr << "Failed: file_to_string(" << filename << ") failed: "
<< reason << endl;
exit(1);
}
} else {
cout << endl << teststring << endl << endl;
odata = teststring;
}
string& data = odata;
string ndata;
if ((op_flags & OPT_C)) {
if (!transcode(odata, ndata, charset, "UTF-8")) {
cerr << "Failed: transcode error" << endl;
exit(1);
} else {
data = ndata;
}
}
if (op_flags & OPT_c) {
int n = TextSplit::countWords(data, flags);
cout << n << " words" << endl;
} else {
myTermProc printproc;
Rcl::TermProc *nxt = &printproc;
Rcl::TermProcCommongrams commonproc(nxt, stoplist);
if (op_flags & OPT_S)
nxt = &commonproc;
Rcl::TermProcPrep preproc(nxt);
if (op_flags & OPT_u)
nxt = &preproc;
Rcl::TextSplitP splitter(nxt, flags);
if (op_flags & OPT_q)
printproc.setNoOut(true);
splitter.text_to_words(data);
}
}
#endif // TEST