1
|
|
|
<?php |
2
|
|
|
/** |
3
|
|
|
* Webino (http://webino.sk) |
4
|
|
|
* |
5
|
|
|
* @link https://github.com/webino/WebinoI18nSanitizeLib for the canonical source repository |
6
|
|
|
* @copyright Copyright (c) 2017 Webino, s. r. o. (http://webino.sk) |
7
|
|
|
* @author Peter Bačinský <[email protected]> |
8
|
|
|
* @license BSD-3-Clause |
9
|
|
|
*/ |
10
|
|
|
|
11
|
|
|
namespace WebinoI18nSanitizeLib; |
12
|
|
|
|
13
|
|
|
use Zend\Filter; |
14
|
|
|
|
15
|
|
|
/** |
16
|
|
|
* Class Transliteration |
17
|
|
|
* |
18
|
|
|
* @author Martin Hujer [email protected] |
19
|
|
|
*/ |
20
|
|
|
class Transliteration implements Filter\FilterInterface |
21
|
|
|
{ |
22
|
|
|
/** |
23
|
|
|
* Returns $value translitered to ASCII |
24
|
|
|
* |
25
|
|
|
* @param string $value |
26
|
|
|
* @return string |
27
|
|
|
*/ |
28
|
|
|
public function filter($value) |
29
|
|
|
{ |
30
|
|
|
// transliterate specific chars |
31
|
|
|
$value = $this->transliterateCzech($value); |
32
|
|
|
$value = $this->transliterateSlovak($value); |
33
|
|
|
$value = $this->transliterateRussian($value); |
34
|
|
|
$value = $this->transliterateGerman($value); |
35
|
|
|
$value = $this->transliterateFrench($value); |
36
|
|
|
$value = $this->transliterateHungarian($value); |
37
|
|
|
$value = $this->transliteratePolish($value); |
38
|
|
|
$value = $this->transliterateDanish($value); |
39
|
|
|
$value = $this->transliterateCroatian($value); |
40
|
|
|
$value = $this->transliterateDiacritic($value); |
41
|
|
|
|
42
|
|
|
// split string to single characters |
43
|
|
|
$characters = mb_split('~(.)~', $value); |
44
|
|
|
|
45
|
|
|
$return = ''; |
46
|
|
|
foreach ($characters as $character) { |
47
|
|
|
/* maybe should contain also //IGNORE */ |
48
|
|
|
$converted = iconv('utf-8', 'ASCII//TRANSLIT', $character); |
49
|
|
|
|
50
|
|
|
// if character was converted, strip out wrong marks |
51
|
|
|
if ($character !== $converted) { |
52
|
|
|
$return .= preg_replace('~["\'^]+~', '', $converted); |
53
|
|
|
} else { |
54
|
|
|
$return .= $converted; |
55
|
|
|
} |
56
|
|
|
} |
57
|
|
|
|
58
|
|
|
return $return; |
59
|
|
|
} |
60
|
|
|
|
61
|
|
|
/** |
62
|
|
|
* Transliterate Russian chars (Cyrillic) |
63
|
|
|
* |
64
|
|
|
* @param string $s |
65
|
|
|
* @return string |
66
|
|
|
*/ |
67
|
|
|
private function transliterateRussian($s) |
68
|
|
|
{ |
69
|
|
|
$map = [ |
70
|
|
|
'А' => 'A', |
71
|
|
|
'Б' => 'B', |
72
|
|
|
'В' => 'V', |
73
|
|
|
'Г' => 'G', |
74
|
|
|
'Д' => 'D', |
75
|
|
|
'Є' => 'E', |
76
|
|
|
'Е' => 'JE', |
77
|
|
|
'Ё' => 'JO', |
78
|
|
|
'Ж' => 'ZH', |
79
|
|
|
'З' => 'Z', |
80
|
|
|
'И' => 'I', |
81
|
|
|
'Й' => 'J', |
82
|
|
|
'К' => 'K', |
83
|
|
|
'Л' => 'L', |
84
|
|
|
'М' => 'M', |
85
|
|
|
'Н' => 'N', |
86
|
|
|
'О' => 'O', |
87
|
|
|
'П' => 'P', |
88
|
|
|
'Р' => 'R', |
89
|
|
|
'С' => 'S', |
90
|
|
|
'Т' => 'T', |
91
|
|
|
'У' => 'U', |
92
|
|
|
'Ф' => 'F', |
93
|
|
|
'Х' => 'KH', |
94
|
|
|
'Ц' => 'TS', |
95
|
|
|
'Ч' => 'CH', |
96
|
|
|
'Ш' => 'SH', |
97
|
|
|
'Щ' => 'SHCH', |
98
|
|
|
'Ъ' => '', |
99
|
|
|
'Ы' => 'Y', |
100
|
|
|
'Ь' => '', |
101
|
|
|
'Э' => 'E', |
102
|
|
|
'Ю' => 'JU', |
103
|
|
|
'Я' => 'JA', |
104
|
|
|
'Ґ' => 'G', |
105
|
|
|
'Ї' => 'I', |
106
|
|
|
'а' => 'a', |
107
|
|
|
'б' => 'b', |
108
|
|
|
'в' => 'v', |
109
|
|
|
'г' => 'g', |
110
|
|
|
'д' => 'd', |
111
|
|
|
'є' => 'e', |
112
|
|
|
'е' => 'je', |
113
|
|
|
'ё' => 'jo', |
114
|
|
|
'ж' => 'zh', |
115
|
|
|
'з' => 'z', |
116
|
|
|
'и' => 'i', |
117
|
|
|
'й' => 'j', |
118
|
|
|
'к' => 'k', |
119
|
|
|
'л' => 'l', |
120
|
|
|
'м' => 'm', |
121
|
|
|
'н' => 'n', |
122
|
|
|
'о' => 'o', |
123
|
|
|
'п' => 'p', |
124
|
|
|
'р' => 'r', |
125
|
|
|
'с' => 's', |
126
|
|
|
'т' => 't', |
127
|
|
|
'у' => 'u', |
128
|
|
|
'ф' => 'f', |
129
|
|
|
'х' => 'kh', |
130
|
|
|
'ц' => 'ts', |
131
|
|
|
'ч' => 'ch', |
132
|
|
|
'ш' => 'sh', |
133
|
|
|
'щ' => 'shch', |
134
|
|
|
'ъ' => '', |
135
|
|
|
'ы' => 'y', |
136
|
|
|
'ь' => '', |
137
|
|
|
'э' => 'e', |
138
|
|
|
'ю' => 'ju', |
139
|
|
|
'я' => 'ja', |
140
|
|
|
'ґ' => 'g', |
141
|
|
|
'ї' => 'i', |
142
|
|
|
]; |
143
|
|
|
return strtr($s, $map); |
144
|
|
|
} |
145
|
|
|
|
146
|
|
|
/** |
147
|
|
|
* Transliterate Czech chars |
148
|
|
|
* |
149
|
|
|
* @param string $s |
150
|
|
|
* @return string |
151
|
|
|
*/ |
152
|
|
|
private function transliterateCzech($s) |
153
|
|
|
{ |
154
|
|
|
$map = [ |
155
|
|
|
'á' => 'a', |
156
|
|
|
'č' => 'c', |
157
|
|
|
'ď' => 'd', |
158
|
|
|
'é' => 'e', |
159
|
|
|
'ě' => 'e', |
160
|
|
|
'í' => 'i', |
161
|
|
|
'ň' => 'n', |
162
|
|
|
'ó' => 'o', |
163
|
|
|
'ř' => 'r', |
164
|
|
|
'š' => 's', |
165
|
|
|
'ť' => 't', |
166
|
|
|
'ú' => 'u', |
167
|
|
|
'ů' => 'u', |
168
|
|
|
'ý' => 'y', |
169
|
|
|
'ž' => 'z', |
170
|
|
|
'Á' => 'A', |
171
|
|
|
'Č' => 'C', |
172
|
|
|
'Ď' => 'D', |
173
|
|
|
'É' => 'E', |
174
|
|
|
'Ě' => 'E', |
175
|
|
|
'Í' => 'I', |
176
|
|
|
'Ň' => 'N', |
177
|
|
|
'Ó' => 'O', |
178
|
|
|
'Ř' => 'R', |
179
|
|
|
'Š' => 'S', |
180
|
|
|
'Ť' => 'T', |
181
|
|
|
'Ú' => 'U', |
182
|
|
|
'Ů' => 'U', |
183
|
|
|
'Ý' => 'Y', |
184
|
|
|
'Ž' => 'Z', |
185
|
|
|
]; |
186
|
|
|
return strtr($s, $map); |
187
|
|
|
} |
188
|
|
|
|
189
|
|
|
/** |
190
|
|
|
* Transliterate German chars |
191
|
|
|
* |
192
|
|
|
* @param string $s |
193
|
|
|
* @return string |
194
|
|
|
*/ |
195
|
|
View Code Duplication |
private function transliterateGerman($s) |
|
|
|
|
196
|
|
|
{ |
197
|
|
|
$map = [ |
198
|
|
|
//'ä' => 'ae', //messes up with slovak -> they have ä -> a |
|
|
|
|
199
|
|
|
'ë' => 'e', |
200
|
|
|
'ï' => 'i', |
201
|
|
|
'ö' => 'oe', |
202
|
|
|
'ü' => 'ue', |
203
|
|
|
'Ä' => 'Ae', |
204
|
|
|
'Ë' => 'E', |
205
|
|
|
'Ï' => 'I', |
206
|
|
|
'Ö' => 'Oe', |
207
|
|
|
'Ü' => 'Ue', |
208
|
|
|
'ß' => 'ss', |
209
|
|
|
]; |
210
|
|
|
return strtr($s, $map); |
211
|
|
|
} |
212
|
|
|
|
213
|
|
|
/** |
214
|
|
|
* Transliterate French chars |
215
|
|
|
* |
216
|
|
|
* @param string $s |
217
|
|
|
* @return string |
218
|
|
|
*/ |
219
|
|
View Code Duplication |
private function transliterateFrench($s) |
|
|
|
|
220
|
|
|
{ |
221
|
|
|
$map = [ |
222
|
|
|
'â' => 'a', |
223
|
|
|
'ê' => 'e', |
224
|
|
|
'î' => 'i', |
225
|
|
|
'ô' => 'o', |
226
|
|
|
'û' => 'u', |
227
|
|
|
'Â' => 'A', |
228
|
|
|
'Ê' => 'E', |
229
|
|
|
'Î' => 'I', |
230
|
|
|
'Ô' => 'O', |
231
|
|
|
'Û' => 'U', |
232
|
|
|
'œ' => 'oe', |
233
|
|
|
'æ' => 'ae', |
234
|
|
|
'Ÿ' => 'Y', |
235
|
|
|
'ç' => 'c', |
236
|
|
|
'Ç' => 'C', |
237
|
|
|
]; |
238
|
|
|
return strtr($s, $map); |
239
|
|
|
} |
240
|
|
|
|
241
|
|
|
/** |
242
|
|
|
* Transliterate Hungarian chars |
243
|
|
|
* |
244
|
|
|
* @param string $s |
245
|
|
|
* @return string |
246
|
|
|
*/ |
247
|
|
View Code Duplication |
private function transliterateHungarian($s) |
|
|
|
|
248
|
|
|
{ |
249
|
|
|
$map = [ |
250
|
|
|
'á' => 'a', |
251
|
|
|
'é' => 'e', |
252
|
|
|
'í' => 'i', |
253
|
|
|
'ó' => 'o', |
254
|
|
|
'ö' => 'o', |
255
|
|
|
'ő' => 'o', |
256
|
|
|
'ú' => 'u', |
257
|
|
|
'ü' => 'u', |
258
|
|
|
'ű' => 'u', |
259
|
|
|
]; |
260
|
|
|
return strtr($s, $map); |
261
|
|
|
} |
262
|
|
|
|
263
|
|
|
/** |
264
|
|
|
* Transliterate Polish chars |
265
|
|
|
* |
266
|
|
|
* @param string $s |
267
|
|
|
* @return string |
268
|
|
|
*/ |
269
|
|
View Code Duplication |
private function transliteratePolish($s) |
|
|
|
|
270
|
|
|
{ |
271
|
|
|
$map = [ |
272
|
|
|
'ą' => 'a', |
273
|
|
|
'ę' => 'e', |
274
|
|
|
'ó' => 'o', |
275
|
|
|
'ć' => 'c', |
276
|
|
|
'ł' => 'l', |
277
|
|
|
'ń' => 'n', |
278
|
|
|
'ś' => 's', |
279
|
|
|
'ż' => 'z', |
280
|
|
|
'ź' => 'z', |
281
|
|
|
'Ó' => 'O', |
282
|
|
|
'Ć' => 'C', |
283
|
|
|
'Ł' => 'L', |
284
|
|
|
'Ś' => 'S', |
285
|
|
|
'Ż' => 'Z', |
286
|
|
|
'Ź' => 'Z', |
287
|
|
|
]; |
288
|
|
|
return strtr($s, $map); |
289
|
|
|
} |
290
|
|
|
|
291
|
|
|
/** |
292
|
|
|
* Transliterate Danish chars |
293
|
|
|
* |
294
|
|
|
* @param string $s |
295
|
|
|
* @return string |
296
|
|
|
*/ |
297
|
|
|
private function transliterateDanish($s) |
298
|
|
|
{ |
299
|
|
|
$map = [ |
300
|
|
|
'æ' => 'ae', |
301
|
|
|
'ø' => 'oe', |
302
|
|
|
'å' => 'aa', |
303
|
|
|
'Æ' => 'Ae', |
304
|
|
|
'Ø' => 'Oe', |
305
|
|
|
'Å' => 'Aa', |
306
|
|
|
]; |
307
|
|
|
return strtr($s, $map); |
308
|
|
|
} |
309
|
|
|
|
310
|
|
|
/** |
311
|
|
|
* Transliterate Croatian chars |
312
|
|
|
* |
313
|
|
|
* @param string $s |
314
|
|
|
* @return string |
315
|
|
|
*/ |
316
|
|
|
private function transliterateCroatian($s) |
317
|
|
|
{ |
318
|
|
|
$map = [ |
319
|
|
|
'Č' => 'C', |
320
|
|
|
'Ć' => 'C', |
321
|
|
|
'Ž' => 'Z', |
322
|
|
|
'Š' => 'S', |
323
|
|
|
'Đ' => 'D', |
324
|
|
|
'č' => 'c', |
325
|
|
|
'ć' => 'c', |
326
|
|
|
'ž' => 'z', |
327
|
|
|
'š' => 's', |
328
|
|
|
'đ' => 'd', |
329
|
|
|
]; |
330
|
|
|
return strtr($s, $map); |
331
|
|
|
} |
332
|
|
|
|
333
|
|
|
/** |
334
|
|
|
* Transliterate Slovak chars |
335
|
|
|
* |
336
|
|
|
* @param string $s |
337
|
|
|
* @return string |
338
|
|
|
*/ |
339
|
|
|
private function transliterateSlovak($s) |
340
|
|
|
{ |
341
|
|
|
$map = [ |
342
|
|
|
'á' => 'a', |
343
|
|
|
'Á' => 'A', |
344
|
|
|
'ä' => 'a', |
345
|
|
|
'Ä' => 'A', |
346
|
|
|
'č' => 'c', |
347
|
|
|
'Č' => 'C', |
348
|
|
|
'ď' => 'd', |
349
|
|
|
'Ď' => 'D', |
350
|
|
|
'é' => 'e', |
351
|
|
|
'É' => 'E', |
352
|
|
|
'í' => 'i', |
353
|
|
|
'Í' => 'I', |
354
|
|
|
'ĺ' => 'l', |
355
|
|
|
'Ĺ' => 'L', |
356
|
|
|
'ľ' => 'l', |
357
|
|
|
'Ľ' => 'L', |
358
|
|
|
'ň' => 'n', |
359
|
|
|
'Ň' => 'N', |
360
|
|
|
'ó' => 'o', |
361
|
|
|
'Ó' => 'O', |
362
|
|
|
'ô' => 'o', |
363
|
|
|
'Ô' => 'O', |
364
|
|
|
'ŕ' => 'r', |
365
|
|
|
'Ŕ' => 'R', |
366
|
|
|
'š' => 's', |
367
|
|
|
'Š' => 'S', |
368
|
|
|
'ť' => 't', |
369
|
|
|
'Ť' => 'T', |
370
|
|
|
'ú' => 'u', |
371
|
|
|
'Ú' => 'U', |
372
|
|
|
'Ý' => 'Y', |
373
|
|
|
'ý' => 'y', |
374
|
|
|
'ž' => 'z', |
375
|
|
|
'Ž' => 'Z', |
376
|
|
|
]; |
377
|
|
|
return strtr($s, $map); |
378
|
|
|
} |
379
|
|
|
|
380
|
|
|
/** |
381
|
|
|
* Transliterate diacritic |
382
|
|
|
* |
383
|
|
|
* @param string $s |
384
|
|
|
* @return string |
385
|
|
|
*/ |
386
|
|
|
private function transliterateDiacritic($s) |
387
|
|
|
{ |
388
|
|
|
$map = [ |
389
|
|
|
'ˇ' => '', |
390
|
|
|
'´' => '', |
391
|
|
|
]; |
392
|
|
|
return strtr($s, $map); |
393
|
|
|
} |
394
|
|
|
} |
395
|
|
|
|
Duplicated code is one of the most pungent code smells. If you need to duplicate the same code in three or more different places, we strongly encourage you to look into extracting the code into a single class or operation.
You can also find more detailed suggestions in the “Code” section of your repository.