summaryrefslogtreecommitdiff
path: root/languages/classes/LanguageKk.php
blob: 34c5c39a7447cd1644048162b42856183acde6c5 (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
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
<?php

require_once( dirname(__FILE__).'/../LanguageConverter.php' );
require_once( dirname(__FILE__).'/LanguageKk_cyrl.php' );

define( 'KK_C_UC', 'АӘБВГҒДЕЁЖЗИЙКҚЛМНҢОӨПРСТУҰҮФХҺЦЧШЩЪЫІЬЭЮЯ' ); # Kazakh Cyrillic uppercase
define( 'KK_C_LC', 'аәбвгғдеёжзийкқлмнңоөпрстуұүфхһцчшщъыіьэюя' ); # Kazakh Cyrillic lowercase
define( 'KK_L_UC', 'AÄBCÇDEÉFGĞHIİÏJKLMNÑOÖPQRSŞTUÜVWXYÝZ' ); # Kazakh Latin uppercase
define( 'KK_L_LC', 'aäbcçdeéfgğhıiïjklmnñoöpqrsştuüvwxyýz' ); # Kazakh Latin lowercase
//define( 'KK_A', 'ٴابپتجحدرزسشعفقكلمنڭەوۇۋۆىيچھ' ); # Kazakh Arabic
define( 'H_HAMZA', 'ٴ' ); # U+0674 ARABIC LETTER HIGH HAMZA
//define( 'ZWNJ', '‌' ); # U+200C ZERO WIDTH NON-JOINER

/** Kazakh (Қазақша)
 * converter routines
 *
 * @ingroup Language
 */
class KkConverter extends LanguageConverter {

	function __construct($langobj, $maincode,
								$variants=array(),
								$variantfallbacks=array(),
								$markup=array(),
								$flags = array()) {
		parent::__construct( $langobj, $maincode,
			$variants, $variantfallbacks, $markup, $flags );

		// No point delaying this since they're in code.
		// Waiting until loadDefaultTables() means they never get loaded
		// when the tables themselves are loaded from cache.
		$this->loadRegs();
	}

	function loadDefaultTables() {
		// require( dirname(__FILE__)."/../../includes/KkConversion.php" );
		// Placeholder for future implementing. Remove variables declarations
		// after generating KkConversion.php
		$kk2Cyrl = array();
		$kk2Latn = array();
		$kk2Arab = array();
		$kk2KZ   = array();
		$kk2TR   = array();
		$kk2CN   = array();

		$this->mTables = array(
			'kk-cyrl' => new ReplacementArray( $kk2Cyrl ),
			'kk-latn' => new ReplacementArray( $kk2Latn ),
			'kk-arab' => new ReplacementArray( $kk2Arab ),
			'kk-kz'   => new ReplacementArray( array_merge($kk2Cyrl, $kk2KZ) ),
			'kk-tr'   => new ReplacementArray( array_merge($kk2Latn, $kk2TR) ),
			'kk-cn'   => new ReplacementArray( array_merge($kk2Arab, $kk2CN) ),
			'kk'      => new ReplacementArray()
		);
	}

	function postLoadTables() {
		$this->mTables['kk-kz']->merge( $this->mTables['kk-cyrl'] );
		$this->mTables['kk-tr']->merge( $this->mTables['kk-latn'] );
		$this->mTables['kk-cn']->merge( $this->mTables['kk-arab'] );
	}

	function loadRegs() {

		$this->mCyrl2Latn = array(
			## Punctuation
			'//u' => 'No.',
			## Е after vowels
			'/([АӘЕЁИОӨҰҮЭЮЯЪЬ])Е/u' => '$1YE',
			'/([АӘЕЁИОӨҰҮЭЮЯЪЬ])е/ui' => '$1ye',
			## leading ЁЮЯЩ
			'/(['.KK_C_UC.']|$)/u' => 'YO$1', '/(['.KK_C_LC.']|$)/u' => 'Yo$1',
			'/(['.KK_C_UC.']|$)/u' => 'YU$1', '/(['.KK_C_LC.']|$)/u' => 'Yu$1',
			'/(['.KK_C_UC.']|$)/u' => 'YA$1', '/(['.KK_C_LC.']|$)/u' => 'Ya$1',
			'/(['.KK_C_UC.']|$)/u' => 'ŞÇ$1', '/(['.KK_C_LC.']|$)/u' => 'Şç$1',
			## other ЁЮЯ
			'/Ё/u' => 'YO', '/ё/u' => 'yo',
			'/Ю/u' => 'YU', '/ю/u' => 'yu',
			'/Я/u' => 'YA', '/я/u' => 'ya',
			'/Щ/u' => 'ŞÇ', '/щ/u' => 'şç',
			## soft and hard signs
			'/[ъЪ]/u' => 'ʺ', '/[ьЬ]/u' => 'ʹ',
			## other characters
			'/А/u' => 'A', '/а/u' => 'a', '/Ә/u' => 'Ä', '/ә/u' => 'ä',
			'/Б/u' => 'B', '/б/u' => 'b', '/В/u' => 'V', '/в/u' => 'v',
			'/Г/u' => 'G', '/г/u' => 'g', '/Ғ/u' => 'Ğ', '/ғ/u' => 'ğ',
			'/Д/u' => 'D', '/д/u' => 'd', '/Е/u' => 'E', '/е/u' => 'e',
			'/Ж/u' => 'J', '/ж/u' => 'j', '/З/u' => 'Z', '/з/u' => 'z',
			'/И/u' => 'Ï', '/и/u' => 'ï', '/Й/u' => 'Ý', '/й/u' => 'ý',
			'/К/u' => 'K', '/к/u' => 'k', '/Қ/u' => 'Q', '/қ/u' => 'q',
			'/Л/u' => 'L', '/л/u' => 'l', '/М/u' => 'M', '/м/u' => 'm',
			'/Н/u' => 'N', '/н/u' => 'n', '/Ң/u' => 'Ñ', '/ң/u' => 'ñ',
			'/О/u' => 'O', '/о/u' => 'o', '/Ө/u' => 'Ö', '/ө/u' => 'ö',
			'/П/u' => 'P', '/п/u' => 'p', '/Р/u' => 'R', '/р/u' => 'r',
			'/С/u' => 'S', '/с/u' => 's', '/Т/u' => 'T', '/т/u' => 't',
			'/У/u' => 'W', '/у/u' => 'w', '/Ұ/u' => 'U', '/ұ/u' => 'u',
			'/Ү/u' => 'Ü', '/ү/u' => 'ü', '/Ф/u' => 'F', '/ф/u' => 'f',
			'/Х/u' => 'X', '/х/u' => 'x', '/Һ/u' => 'H', '/һ/u' => 'h',
			'/Ц/u' => 'C', '/ц/u' => 'c', '/Ч/u' => 'Ç', '/ч/u' => 'ç',
			'/Ш/u' => 'Ş', '/ш/u' => 'ş', '/Ы/u' => 'I', '/ы/u' => 'ı',
			'/І/u' => 'İ', '/і/u' => 'i', '/Э/u' => 'É', '/э/u' => 'é',
		);

		$this->mLatn2Cyrl = array(
			## Punctuation
			'/#|No\./' => '№',
			## Şç
			'/ŞÇʹ/u'=> 'ЩЬ', '/Şçʹ/u'=> 'Щь', '/Şçʹ/u'=> 'Щь',
			'/Ş[Çç]/u' => 'Щ', '/şç/u' => 'щ',
			## soft and hard signs
			'/(['.KK_L_UC.'])ʺ(['.KK_L_UC.'])/u' => '$1Ъ$2',
			'/ʺ(['.KK_L_LC.'])/u' => $1',
			'/(['.KK_L_UC.'])ʹ(['.KK_L_UC.'])/u' => '$1Ь$2',
			'/ʹ(['.KK_L_LC.'])/u' => $1',
			'/ʺ/u' => 'ъ',
			'/ʹ/u' => 'ь',
			## Ye Yo Yu Ya.
			'/Y[Ee]/u' => 'Е', '/ye/u' => 'е',
			'/Y[Oo]/u' => 'Ё', '/yo/u' => 'ё',
			'/Y[UWuw]/u' => 'Ю', '/y[uw]/u' => 'ю',
			'/Y[Aa]/u' => 'Я', '/ya/u' => 'я',
			## other characters
			'/A/u' => 'А', '/a/u' => 'а', '/Ä/u' => 'Ә', '/ä/u' => 'ә',
			'/B/u' => 'Б', '/b/u' => 'б', '/C/u' => 'Ц', '/c/u' => 'ц',
			'/Ç/u' => 'Ч', '/ç/u' => 'ч', '/D/u' => 'Д', '/d/u' => 'д',
			'/E/u' => 'Е', '/e/u' => 'е', '/É/u' => 'Э', '/é/u' => 'э',
			'/F/u' => 'Ф', '/f/u' => 'ф', '/G/u' => 'Г', '/g/u' => 'г',
			'/Ğ/u' => 'Ғ', '/ğ/u' => 'ғ', '/H/u' => 'Һ', '/h/u' => 'һ',
			'/I/u' => 'Ы', '/ı/u' => 'ы', '/İ/u' => 'І', '/i/u' => 'і',
			'/Ï/u' => 'И', '/ï/u' => 'и', '/J/u' => 'Ж', '/j/u' => 'ж',
			'/K/u' => 'К', '/k/u' => 'к', '/L/u' => 'Л', '/l/u' => 'л',
			'/M/u' => 'М', '/m/u' => 'м', '/N/u' => 'Н', '/n/u' => 'н',
			'/Ñ/u' => 'Ң', '/ñ/u' => 'ң', '/O/u' => 'О', '/o/u' => 'о',
			'/Ö/u' => 'Ө', '/ö/u' => 'ө', '/P/u' => 'П', '/p/u' => 'п',
			'/Q/u' => 'Қ', '/q/u' => 'қ', '/R/u' => 'Р', '/r/u' => 'р',
			'/S/u' => 'С', '/s/u' => 'с', '/Ş/u' => 'Ш', '/ş/u' => 'ш',
			'/T/u' => 'Т', '/t/u' => 'т', '/U/u' => 'Ұ', '/u/u' => 'ұ',
			'/Ü/u' => 'Ү', '/ü/u' => 'ү', '/V/u' => 'В', '/v/u' => 'в',
			'/W/u' => 'У', '/w/u' => 'у', '/Ý/u' => 'Й', '/ý/u' => 'й',
			'/X/u' => 'Х', '/x/u' => 'х', '/Z/u' => 'З', '/z/u' => 'з',
		);

		$this->mCyLa2Arab = array(
			## Punctuation -> Arabic
			'/#|№|No\./u' => '؀', # &#x0600;
			'/\,/' => '،', # &#x060C;
			'/;/'  => '؛', # &#x061B;
			'/\?/' => '؟', # &#x061F;
			'/%/'  => '٪', # &#x066A;
			'/\*/' => '٭', # &#x066D;
			## Digits -> Arabic
			'/0/' => '۰', # &#x06F0;
			'/1/' => '۱', # &#x06F1;
			'/2/' => '۲', # &#x06F2;
			'/3/' => '۳', # &#x06F3;
			'/4/' => '۴', # &#x06F4;
			'/5/' => '۵', # &#x06F5;
			'/6/' => '۶', # &#x06F6;
			'/7/' => '۷', # &#x06F7;
			'/8/' => '۸', # &#x06F8;
			'/9/' => '۹', # &#x06F9;
			## Cyrillic -> Arabic
			'/Аллаһ/ui' => 'ﷲ',
			'/([АӘЕЁИОӨҰҮЭЮЯЪЬ])е/ui' => '$1يە',
			'/[еэ]/ui' => 'ە', '/[ъь]/ui' => '',
			'/[аә]/ui' => 'ا', '/[оө]/ui' => 'و', '/[ұү]/ui' => 'ۇ', '/[ыі]/ui' => 'ى',
			'/[и]/ui' => 'ىي', '/ё/ui' => 'يو', '/ю/ui' => 'يۋ', '/я/ui' => 'يا', '/[й]/ui' => 'ي',
			'/ц/ui' => 'تس', '/щ/ui' => 'شش',
			'/һ/ui' => 'ح', '/ч/ui' => 'تش',
			#'/һ/ui' => 'ھ', '/ч/ui' => 'چ',
			'/б/ui' => 'ب', '/в/ui' => 'ۆ', '/г/ui' => 'گ', '/ғ/ui' => 'ع',
			'/д/ui' => 'د', '/ж/ui' => 'ج', '/з/ui' => 'ز', '/к/ui' => 'ك',
			'/қ/ui' => 'ق', '/л/ui' => 'ل', '/м/ui' => 'م', '/н/ui' => 'ن',
			'/ң/ui' => 'ڭ', '/п/ui' => 'پ', '/р/ui' => 'ر', '/с/ui' => 'س',
			'/т/ui' => 'ت', '/у/ui' => 'ۋ', '/ф/ui' => 'ف', '/х/ui' => 'ح',
			'/ш/ui' => 'ش',
			## Latin -> Arabic // commented for now...
			/*'/Allah/ui' => 'ﷲ',
			'/[eé]/ui' => 'ە', '/[yý]/ui' => 'ي', '/[ʺʹ]/ui' => '',
			'/[aä]/ui' => 'ا', '/[oö]/ui' => 'و', '/[uü]/ui' => 'ۇ',
			'/[ï]/ui' => 'ىي', '/[ıIiİ]/u' => 'ى',
			'/c/ui' => 'تس',
			'/ç/ui' => 'تش', '/h/ui' => 'ح',
			#'/ç/ui' => 'چ', '/h/ui' => 'ھ',
			'/b/ui' => 'ب','/d/ui' => 'د',
			'/f/ui' => 'ف', '/g/ui' => 'گ', '/ğ/ui' => 'ع',
			'/j/ui' => 'ج', '/k/ui' => 'ك', '/l/ui' => 'ل', '/m/ui' => 'م',
			'/n/ui' => 'ن', '/ñ/ui' => 'ڭ', '/p/ui' => 'پ', '/q/ui' => 'ق',
			'/r/ui' => 'ر', '/s/ui' => 'س', '/ş/ui' => 'ش', '/t/ui' => 'ت',
			'/v/ui' => 'ۆ', '/w/ui' => 'ۋ', '/x/ui' => 'ح', '/z/ui' => 'ز',*/
		);
	}

	/* rules should be defined as -{ekavian | iyekavian-} -or-
		-{code:text | code:text | ...}-
		update: delete all rule parsing because it's not used
				currently, and just produces a couple of bugs
	*/
	function parseManualRule($rule, $flags=array()) {
		if(in_array('T',$flags)){
			return parent::parseManualRule($rule, $flags);
		}

		// otherwise ignore all formatting
		foreach($this->mVariants as $v) {
			$carray[$v] = $rule;
		}

		return $carray;
	}

	// Do not convert content on talk pages
	function parserConvert( $text, &$parser ){
		if(is_object($parser->getTitle() ) && $parser->getTitle()->isTalkPage())
			$this->mDoContentConvert=false;
		else
			$this->mDoContentConvert=true;

		return parent::parserConvert($text, $parser );
	}

	/*
	 * A function wrapper:
	 *  - if there is no selected variant, leave the link
	 *    names as they were
	 *  - do not try to find variants for usernames
	 */
	function findVariantLink( &$link, &$nt ) {
		// check for user namespace
		if(is_object($nt)){
			$ns = $nt->getNamespace();
			if($ns==NS_USER || $ns==NS_USER_TALK)
				return;
		}

		$oldlink=$link;
		parent::findVariantLink($link,$nt);
		if($this->getPreferredVariant()==$this->mMainLanguageCode)
			$link=$oldlink;
	}

	/*
	 * An ugly function wrapper for parsing Image titles
	 * (to prevent image name conversion)
	 */
	function autoConvert($text, $toVariant=false) {
		global $wgTitle;
		if(is_object($wgTitle) && $wgTitle->getNameSpace()==NS_IMAGE){
			$imagename = $wgTitle->getNsText();
			if(preg_match("/^$imagename:/",$text)) return $text;
		}
		return parent::autoConvert($text,$toVariant);
	}

	/**
	 *  It translates text into variant
	 */
	function translate( $text, $toVariant ){
		global $wgContLanguageCode;
		$text = parent::translate( $text, $toVariant );

		$letters = '';
		switch( $toVariant ) {
			case 'kk-cyrl':
			case 'kk-kz':
				$letters = KK_L_UC . KK_L_LC . 'ʺʹ#0123456789';
				$wgContLanguageCode = 'kk';
				break;
			case 'kk-latn':
			case 'kk-tr':
				$letters = KK_C_UC . KK_C_LC . '№0123456789';
				$wgContLanguageCode = 'kk-Latn';
				break;
			case 'kk-arab':
			case 'kk-cn':
				$letters = KK_C_UC.KK_C_LC./*KK_L_UC.KK_L_LC.'ʺʹ'.*/',;\?%\*№0123456789';
				$wgContLanguageCode = 'kk-Arab';
				break;
			default:
				$wgContLanguageCode = 'kk';
				return $text;
		}
		// disable conversion variables like $1, $2...
		$varsfix = '\$[0-9]';

		$matches = preg_split( '/' . $varsfix . '[^' . $letters . ']+/u', $text, -1, PREG_SPLIT_OFFSET_CAPTURE);
		$mstart = 0;
		$ret = '';
		foreach( $matches as $m ) {
			$ret .= substr( $text, $mstart, $m[1]-$mstart );
			$ret .= $this->regsConverter( $m[0], $toVariant );
			$mstart = $m[1] + strlen($m[0]);
		}
		return $ret;
	}

	function regsConverter( $text, $toVariant ) {
		if ($text == '') return $text;

		$pat = array();
		$rep = array();
		switch( $toVariant ) {
			case 'kk-arab':
			case 'kk-cn':
				$letters = KK_C_LC.KK_C_UC/*.KK_L_LC.KK_L_UC*/;
				$front = 'әөүіӘӨҮІ'/*.'äöüiÄÖÜİ'*/;
				$excludes = 'еэгғкқЕЭГҒКҚ'/*.'eégğkqEÉGĞKQ'*/;
				// split text to words
				$matches = preg_split( '/[\b\s\-\.:]+/', $text, -1, PREG_SPLIT_OFFSET_CAPTURE);
				$mstart = 0;
				$ret = '';
				foreach( $matches as $m ) {
					$ret .= substr( $text, $mstart, $m[1] - $mstart );
					// is matched the word to front vowels?
					// exclude a words matched to е, э, г, к, к, қ,
					// them should be without hamza
					if ( preg_match('/['.$front.']/u', $m[0]) && !preg_match('/['.$excludes.']/u', $m[0]) ) {
						$ret .= preg_replace('/['.$letters.']+/u', H_HAMZA.'$0', $m[0]);
					} else {
						$ret .= $m[0];
					}
					$mstart = $m[1] + strlen($m[0]);
				}
				$text =& $ret;
				foreach( $this->mCyLa2Arab as $pat => $rep ) {
					$text = preg_replace( $pat, $rep, $text );
				}
				return $text;
				break;
			case 'kk-latn':
			case 'kk-tr':
				foreach( $this->mCyrl2Latn as $pat => $rep ) {
					$text = preg_replace( $pat, $rep, $text );
				}
				return $text;
				break;
			case 'kk-cyrl':
			case 'kk-kz':
				foreach( $this->mLatn2Cyrl as $pat => $rep ) {
					$text = preg_replace( $pat, $rep, $text );
				}
				return $text;
				break;
			default:
				return $text;
		}
	}

	/*
	 * We want our external link captions to be converted in variants,
	 * so we return the original text instead -{$text}-, except for URLs
	 */
	function markNoConversion( $text, $noParse=false ) {
		if( $noParse || preg_match( "/^https?:\/\/|ftp:\/\/|irc:\/\//", $text ) )
			return parent::markNoConversion( $text );
		return $text;
	}

	function convertCategoryKey( $key ) {
		return $this->autoConvert( $key, 'kk' );
	}

}

/**
 * class that handles Cyrillic, Latin and Arabic scripts for Kazakh
 * right now it only distinguish kk_cyrl, kk_latn, kk_arab and kk_kz, kk_tr, kk_cn.
 *
 * @ingroup Language
 */
class LanguageKk extends LanguageKk_cyrl {

	function __construct() {
		global $wgHooks;
		parent::__construct();

		$variants = array( 'kk', 'kk-cyrl', 'kk-latn', 'kk-arab', 'kk-kz', 'kk-tr', 'kk-cn' );
		$variantfallbacks = array(
			'kk'      => 'kk-cyrl',
			'kk-cyrl' => 'kk',
			'kk-latn' => 'kk',
			'kk-arab' => 'kk',
			'kk-kz'   => 'kk-cyrl',
			'kk-tr'   => 'kk-latn',
			'kk-cn'   => 'kk-arab'
		);

		$this->mConverter = new KkConverter( $this, 'kk', $variants, $variantfallbacks );

		$wgHooks['ArticleSaveComplete'][] = $this->mConverter;
	}

	/**
	 * Work around for right-to-left direction support in kk-arab and kk-cn
	 *
	 * @return bool
	 */
	function isRTL() {
		$variant = $this->getPreferredVariant();
		if ( $variant == 'kk-arab' || $variant == 'kk-cn' ) {
			return true;
		} else {
			return parent::isRTL();
		}
	}

	/*
	 * It fixes issue with ucfirst for transforming 'i' to 'İ'
	 *
	 */
	function ucfirst ( $string ) {
		$variant = $this->getPreferredVariant();
		if ( ($variant == 'kk-latn' || $variant == 'kk-tr') && $string[0] == 'i' ) {
			$string = 'İ' . substr( $string, 1 );
		} else {
			$string = parent::ucfirst( $string );
		}
		return $string;
	}

	/*
	 * It fixes issue with  lcfirst for transforming 'I' to 'ı'
	 *
	 */
	function lcfirst ( $string ) {
		$variant = $this->getPreferredVariant();
		if ( ($variant == 'kk-latn' || $variant == 'kk-tr') && $string[0] == 'I' ) {
			$string = 'ı' . substr( $string, 1 );
		} else {
			$string = parent::lcfirst( $string );
		}
		return $string;
	}

	function convertGrammar( $word, $case ) {
		wfProfileIn( __METHOD__ );

		$variant = $this->getPreferredVariant();
		switch ( $variant ) {
			case 'kk-arab':
			case 'kk-cn':
				$word = parent::convertGrammarKk_arab( $word, $case );
				break;
			case 'kk-latn':
			case 'kk-tr':
				$word = parent::convertGrammarKk_latn( $word, $case );
				break;
			case 'kk-cyrl':
			case 'kk-kz':
			case 'kk':
			default:
				$word = parent::convertGrammarKk_cyrl( $word, $case );
		}

		wfProfileOut( __METHOD__ );
		return $word;
	}
}