rap2hpoutre /
convert-accent-characters
| Conditions | 10 |
| Paths | 6 |
| Total Lines | 360 |
| Code Lines | 335 |
| Lines | 0 |
| Ratio | 0 % |
| Changes | 0 | ||
Small methods make your code easier to understand, in particular if combined with a good name. Besides, if your method is small, finding a good name is usually much easier.
For example, if you find yourself adding comments to a method's body, this is usually a good sign to extract the commented part to a new method, and use the comment as a starting point when coming up with a good name for this new method.
Commonly applied refactorings include:
If many parameters/temporary variables are present:
| 1 | <?php |
||
| 15 | function convert_accent_characters($string, $locale = null) |
||
| 16 | { |
||
| 17 | if (!preg_match('/[\x80-\xff]/', $string)) { |
||
| 18 | return $string; |
||
| 19 | } |
||
| 20 | |||
| 21 | $chars = [ |
||
| 22 | // Decompositions for Latin-1 Supplement |
||
| 23 | 'ª' => 'a', |
||
| 24 | 'º' => 'o', |
||
| 25 | 'À' => 'A', |
||
| 26 | 'Á' => 'A', |
||
| 27 | 'Â' => 'A', |
||
| 28 | 'Ã' => 'A', |
||
| 29 | 'Ä' => 'A', |
||
| 30 | 'Å' => 'A', |
||
| 31 | 'Æ' => 'AE', |
||
| 32 | 'Ç' => 'C', |
||
| 33 | 'È' => 'E', |
||
| 34 | 'É' => 'E', |
||
| 35 | 'Ê' => 'E', |
||
| 36 | 'Ë' => 'E', |
||
| 37 | 'Ì' => 'I', |
||
| 38 | 'Í' => 'I', |
||
| 39 | 'Î' => 'I', |
||
| 40 | 'Ï' => 'I', |
||
| 41 | 'Ð' => 'D', |
||
| 42 | 'Ñ' => 'N', |
||
| 43 | 'Ò' => 'O', |
||
| 44 | 'Ó' => 'O', |
||
| 45 | 'Ô' => 'O', |
||
| 46 | 'Õ' => 'O', |
||
| 47 | 'Ö' => 'O', |
||
| 48 | 'Ù' => 'U', |
||
| 49 | 'Ú' => 'U', |
||
| 50 | 'Û' => 'U', |
||
| 51 | 'Ü' => 'U', |
||
| 52 | 'Ý' => 'Y', |
||
| 53 | 'Þ' => 'TH', |
||
| 54 | 'ß' => 's', |
||
| 55 | 'à' => 'a', |
||
| 56 | 'á' => 'a', |
||
| 57 | 'â' => 'a', |
||
| 58 | 'ã' => 'a', |
||
| 59 | 'ä' => 'a', |
||
| 60 | 'å' => 'a', |
||
| 61 | 'æ' => 'ae', |
||
| 62 | 'ç' => 'c', |
||
| 63 | 'è' => 'e', |
||
| 64 | 'é' => 'e', |
||
| 65 | 'ê' => 'e', |
||
| 66 | 'ë' => 'e', |
||
| 67 | 'ì' => 'i', |
||
| 68 | 'í' => 'i', |
||
| 69 | 'î' => 'i', |
||
| 70 | 'ï' => 'i', |
||
| 71 | 'ð' => 'd', |
||
| 72 | 'ñ' => 'n', |
||
| 73 | 'ò' => 'o', |
||
| 74 | 'ó' => 'o', |
||
| 75 | 'ô' => 'o', |
||
| 76 | 'õ' => 'o', |
||
| 77 | 'ö' => 'o', |
||
| 78 | 'ø' => 'o', |
||
| 79 | 'ù' => 'u', |
||
| 80 | 'ú' => 'u', |
||
| 81 | 'û' => 'u', |
||
| 82 | 'ü' => 'u', |
||
| 83 | 'ý' => 'y', |
||
| 84 | 'þ' => 'th', |
||
| 85 | 'ÿ' => 'y', |
||
| 86 | 'Ø' => 'O', |
||
| 87 | // Decompositions for Latin Extended-A |
||
| 88 | 'Ā' => 'A', |
||
| 89 | 'ā' => 'a', |
||
| 90 | 'Ă' => 'A', |
||
| 91 | 'ă' => 'a', |
||
| 92 | 'Ą' => 'A', |
||
| 93 | 'ą' => 'a', |
||
| 94 | 'Ć' => 'C', |
||
| 95 | 'ć' => 'c', |
||
| 96 | 'Ĉ' => 'C', |
||
| 97 | 'ĉ' => 'c', |
||
| 98 | 'Ċ' => 'C', |
||
| 99 | 'ċ' => 'c', |
||
| 100 | 'Č' => 'C', |
||
| 101 | 'č' => 'c', |
||
| 102 | 'Ď' => 'D', |
||
| 103 | 'ď' => 'd', |
||
| 104 | 'Đ' => 'D', |
||
| 105 | 'đ' => 'd', |
||
| 106 | 'Ē' => 'E', |
||
| 107 | 'ē' => 'e', |
||
| 108 | 'Ĕ' => 'E', |
||
| 109 | 'ĕ' => 'e', |
||
| 110 | 'Ė' => 'E', |
||
| 111 | 'ė' => 'e', |
||
| 112 | 'Ę' => 'E', |
||
| 113 | 'ę' => 'e', |
||
| 114 | 'Ě' => 'E', |
||
| 115 | 'ě' => 'e', |
||
| 116 | 'Ĝ' => 'G', |
||
| 117 | 'ĝ' => 'g', |
||
| 118 | 'Ğ' => 'G', |
||
| 119 | 'ğ' => 'g', |
||
| 120 | 'Ġ' => 'G', |
||
| 121 | 'ġ' => 'g', |
||
| 122 | 'Ģ' => 'G', |
||
| 123 | 'ģ' => 'g', |
||
| 124 | 'Ĥ' => 'H', |
||
| 125 | 'ĥ' => 'h', |
||
| 126 | 'Ħ' => 'H', |
||
| 127 | 'ħ' => 'h', |
||
| 128 | 'Ĩ' => 'I', |
||
| 129 | 'ĩ' => 'i', |
||
| 130 | 'Ī' => 'I', |
||
| 131 | 'ī' => 'i', |
||
| 132 | 'Ĭ' => 'I', |
||
| 133 | 'ĭ' => 'i', |
||
| 134 | 'Į' => 'I', |
||
| 135 | 'į' => 'i', |
||
| 136 | 'İ' => 'I', |
||
| 137 | 'ı' => 'i', |
||
| 138 | 'IJ' => 'IJ', |
||
| 139 | 'ij' => 'ij', |
||
| 140 | 'Ĵ' => 'J', |
||
| 141 | 'ĵ' => 'j', |
||
| 142 | 'Ķ' => 'K', |
||
| 143 | 'ķ' => 'k', |
||
| 144 | 'ĸ' => 'k', |
||
| 145 | 'Ĺ' => 'L', |
||
| 146 | 'ĺ' => 'l', |
||
| 147 | 'Ļ' => 'L', |
||
| 148 | 'ļ' => 'l', |
||
| 149 | 'Ľ' => 'L', |
||
| 150 | 'ľ' => 'l', |
||
| 151 | 'Ŀ' => 'L', |
||
| 152 | 'ŀ' => 'l', |
||
| 153 | 'Ł' => 'L', |
||
| 154 | 'ł' => 'l', |
||
| 155 | 'Ń' => 'N', |
||
| 156 | 'ń' => 'n', |
||
| 157 | 'Ņ' => 'N', |
||
| 158 | 'ņ' => 'n', |
||
| 159 | 'Ň' => 'N', |
||
| 160 | 'ň' => 'n', |
||
| 161 | 'ʼn' => 'n', |
||
| 162 | 'Ŋ' => 'N', |
||
| 163 | 'ŋ' => 'n', |
||
| 164 | 'Ō' => 'O', |
||
| 165 | 'ō' => 'o', |
||
| 166 | 'Ŏ' => 'O', |
||
| 167 | 'ŏ' => 'o', |
||
| 168 | 'Ő' => 'O', |
||
| 169 | 'ő' => 'o', |
||
| 170 | 'Œ' => 'OE', |
||
| 171 | 'œ' => 'oe', |
||
| 172 | 'Ŕ' => 'R', |
||
| 173 | 'ŕ' => 'r', |
||
| 174 | 'Ŗ' => 'R', |
||
| 175 | 'ŗ' => 'r', |
||
| 176 | 'Ř' => 'R', |
||
| 177 | 'ř' => 'r', |
||
| 178 | 'Ś' => 'S', |
||
| 179 | 'ś' => 's', |
||
| 180 | 'Ŝ' => 'S', |
||
| 181 | 'ŝ' => 's', |
||
| 182 | 'Ş' => 'S', |
||
| 183 | 'ş' => 's', |
||
| 184 | 'Š' => 'S', |
||
| 185 | 'š' => 's', |
||
| 186 | 'Ţ' => 'T', |
||
| 187 | 'ţ' => 't', |
||
| 188 | 'Ť' => 'T', |
||
| 189 | 'ť' => 't', |
||
| 190 | 'Ŧ' => 'T', |
||
| 191 | 'ŧ' => 't', |
||
| 192 | 'Ũ' => 'U', |
||
| 193 | 'ũ' => 'u', |
||
| 194 | 'Ū' => 'U', |
||
| 195 | 'ū' => 'u', |
||
| 196 | 'Ŭ' => 'U', |
||
| 197 | 'ŭ' => 'u', |
||
| 198 | 'Ů' => 'U', |
||
| 199 | 'ů' => 'u', |
||
| 200 | 'Ű' => 'U', |
||
| 201 | 'ű' => 'u', |
||
| 202 | 'Ų' => 'U', |
||
| 203 | 'ų' => 'u', |
||
| 204 | 'Ŵ' => 'W', |
||
| 205 | 'ŵ' => 'w', |
||
| 206 | 'Ŷ' => 'Y', |
||
| 207 | 'ŷ' => 'y', |
||
| 208 | 'Ÿ' => 'Y', |
||
| 209 | 'Ź' => 'Z', |
||
| 210 | 'ź' => 'z', |
||
| 211 | 'Ż' => 'Z', |
||
| 212 | 'ż' => 'z', |
||
| 213 | 'Ž' => 'Z', |
||
| 214 | 'ž' => 'z', |
||
| 215 | 'ſ' => 's', |
||
| 216 | // Decompositions for Latin Extended-B |
||
| 217 | 'Ș' => 'S', |
||
| 218 | 'ș' => 's', |
||
| 219 | 'Ț' => 'T', |
||
| 220 | 'ț' => 't', |
||
| 221 | // Euro Sign |
||
| 222 | '€' => 'E', |
||
| 223 | // GBP (Pound) Sign |
||
| 224 | '£' => '', |
||
| 225 | // Vowels with diacritic (Vietnamese) |
||
| 226 | // unmarked |
||
| 227 | 'Ơ' => 'O', |
||
| 228 | 'ơ' => 'o', |
||
| 229 | 'Ư' => 'U', |
||
| 230 | 'ư' => 'u', |
||
| 231 | // grave accent |
||
| 232 | 'Ầ' => 'A', |
||
| 233 | 'ầ' => 'a', |
||
| 234 | 'Ằ' => 'A', |
||
| 235 | 'ằ' => 'a', |
||
| 236 | 'Ề' => 'E', |
||
| 237 | 'ề' => 'e', |
||
| 238 | 'Ồ' => 'O', |
||
| 239 | 'ồ' => 'o', |
||
| 240 | 'Ờ' => 'O', |
||
| 241 | 'ờ' => 'o', |
||
| 242 | 'Ừ' => 'U', |
||
| 243 | 'ừ' => 'u', |
||
| 244 | 'Ỳ' => 'Y', |
||
| 245 | 'ỳ' => 'y', |
||
| 246 | // hook |
||
| 247 | 'Ả' => 'A', |
||
| 248 | 'ả' => 'a', |
||
| 249 | 'Ẩ' => 'A', |
||
| 250 | 'ẩ' => 'a', |
||
| 251 | 'Ẳ' => 'A', |
||
| 252 | 'ẳ' => 'a', |
||
| 253 | 'Ẻ' => 'E', |
||
| 254 | 'ẻ' => 'e', |
||
| 255 | 'Ể' => 'E', |
||
| 256 | 'ể' => 'e', |
||
| 257 | 'Ỉ' => 'I', |
||
| 258 | 'ỉ' => 'i', |
||
| 259 | 'Ỏ' => 'O', |
||
| 260 | 'ỏ' => 'o', |
||
| 261 | 'Ổ' => 'O', |
||
| 262 | 'ổ' => 'o', |
||
| 263 | 'Ở' => 'O', |
||
| 264 | 'ở' => 'o', |
||
| 265 | 'Ủ' => 'U', |
||
| 266 | 'ủ' => 'u', |
||
| 267 | 'Ử' => 'U', |
||
| 268 | 'ử' => 'u', |
||
| 269 | 'Ỷ' => 'Y', |
||
| 270 | 'ỷ' => 'y', |
||
| 271 | // tilde |
||
| 272 | 'Ẫ' => 'A', |
||
| 273 | 'ẫ' => 'a', |
||
| 274 | 'Ẵ' => 'A', |
||
| 275 | 'ẵ' => 'a', |
||
| 276 | 'Ẽ' => 'E', |
||
| 277 | 'ẽ' => 'e', |
||
| 278 | 'Ễ' => 'E', |
||
| 279 | 'ễ' => 'e', |
||
| 280 | 'Ỗ' => 'O', |
||
| 281 | 'ỗ' => 'o', |
||
| 282 | 'Ỡ' => 'O', |
||
| 283 | 'ỡ' => 'o', |
||
| 284 | 'Ữ' => 'U', |
||
| 285 | 'ữ' => 'u', |
||
| 286 | 'Ỹ' => 'Y', |
||
| 287 | 'ỹ' => 'y', |
||
| 288 | // acute accent |
||
| 289 | 'Ấ' => 'A', |
||
| 290 | 'ấ' => 'a', |
||
| 291 | 'Ắ' => 'A', |
||
| 292 | 'ắ' => 'a', |
||
| 293 | 'Ế' => 'E', |
||
| 294 | 'ế' => 'e', |
||
| 295 | 'Ố' => 'O', |
||
| 296 | 'ố' => 'o', |
||
| 297 | 'Ớ' => 'O', |
||
| 298 | 'ớ' => 'o', |
||
| 299 | 'Ứ' => 'U', |
||
| 300 | 'ứ' => 'u', |
||
| 301 | // dot below |
||
| 302 | 'Ạ' => 'A', |
||
| 303 | 'ạ' => 'a', |
||
| 304 | 'Ậ' => 'A', |
||
| 305 | 'ậ' => 'a', |
||
| 306 | 'Ặ' => 'A', |
||
| 307 | 'ặ' => 'a', |
||
| 308 | 'Ẹ' => 'E', |
||
| 309 | 'ẹ' => 'e', |
||
| 310 | 'Ệ' => 'E', |
||
| 311 | 'ệ' => 'e', |
||
| 312 | 'Ị' => 'I', |
||
| 313 | 'ị' => 'i', |
||
| 314 | 'Ọ' => 'O', |
||
| 315 | 'ọ' => 'o', |
||
| 316 | 'Ộ' => 'O', |
||
| 317 | 'ộ' => 'o', |
||
| 318 | 'Ợ' => 'O', |
||
| 319 | 'ợ' => 'o', |
||
| 320 | 'Ụ' => 'U', |
||
| 321 | 'ụ' => 'u', |
||
| 322 | 'Ự' => 'U', |
||
| 323 | 'ự' => 'u', |
||
| 324 | 'Ỵ' => 'Y', |
||
| 325 | 'ỵ' => 'y', |
||
| 326 | // Vowels with diacritic (Chinese, Hanyu Pinyin) |
||
| 327 | 'ɑ' => 'a', |
||
| 328 | // macron |
||
| 329 | 'Ǖ' => 'U', |
||
| 330 | 'ǖ' => 'u', |
||
| 331 | // acute accent |
||
| 332 | 'Ǘ' => 'U', |
||
| 333 | 'ǘ' => 'u', |
||
| 334 | // caron |
||
| 335 | 'Ǎ' => 'A', |
||
| 336 | 'ǎ' => 'a', |
||
| 337 | 'Ǐ' => 'I', |
||
| 338 | 'ǐ' => 'i', |
||
| 339 | 'Ǒ' => 'O', |
||
| 340 | 'ǒ' => 'o', |
||
| 341 | 'Ǔ' => 'U', |
||
| 342 | 'ǔ' => 'u', |
||
| 343 | 'Ǚ' => 'U', |
||
| 344 | 'ǚ' => 'u', |
||
| 345 | // grave accent |
||
| 346 | 'Ǜ' => 'U', |
||
| 347 | 'ǜ' => 'u', |
||
| 348 | ]; |
||
| 349 | // Used for locale-specific rules |
||
| 350 | if ('de_DE' == $locale || 'de_DE_formal' == $locale || 'de_CH' == $locale || 'de_CH_informal' == $locale) { |
||
| 351 | $chars['Ä'] = 'Ae'; |
||
| 352 | $chars['ä'] = 'ae'; |
||
| 353 | $chars['Ö'] = 'Oe'; |
||
| 354 | $chars['ö'] = 'oe'; |
||
| 355 | $chars['Ü'] = 'Ue'; |
||
| 356 | $chars['ü'] = 'ue'; |
||
| 357 | $chars['ß'] = 'ss'; |
||
| 358 | } elseif ('da_DK' === $locale) { |
||
| 359 | $chars['Æ'] = 'Ae'; |
||
| 360 | $chars['æ'] = 'ae'; |
||
| 361 | $chars['Ø'] = 'Oe'; |
||
| 362 | $chars['ø'] = 'oe'; |
||
| 363 | $chars['Å'] = 'Aa'; |
||
| 364 | $chars['å'] = 'aa'; |
||
| 365 | } elseif ('ca' === $locale) { |
||
| 366 | $chars['l·l'] = 'll'; |
||
| 367 | } elseif ('sr_RS' === $locale || 'bs_BA' === $locale) { |
||
| 368 | $chars['Đ'] = 'DJ'; |
||
| 369 | $chars['đ'] = 'dj'; |
||
| 370 | } |
||
| 371 | $string = strtr($string, $chars); |
||
| 372 | |||
| 373 | return $string; |
||
| 374 | } |
