chrislit /
abydos
| Conditions | 18 |
| Total Lines | 703 |
| Code Lines | 458 |
| Lines | 0 |
| Ratio | 0 % |
| Tests | 98 |
| CRAP Score | 18 |
| 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:
Complex classes like abydos.stemmer._lovins.lovins() often do a lot of different things. To break such a class down, we need to identify a cohesive component within that class. A common approach to find such a component is to look for fields/methods that share the same prefixes, or suffixes.
Once you have determined the fields that belong together, you can apply the Extract Class refactoring. If the component makes sense as a sub-class, Extract Subclass is also a candidate, and is often faster.
| 1 | # -*- coding: utf-8 -*- |
||
| 34 | 1 | def lovins(word): |
|
|
Comprehensibility
introduced
by
|
|||
| 35 | """Return Lovins stem. |
||
| 36 | |||
| 37 | Lovins stemmer |
||
| 38 | |||
| 39 | The Lovins stemmer is described in Julie Beth Lovins's article |
||
| 40 | :cite:`Lovins:1968`. |
||
| 41 | |||
| 42 | :param str word: the word to stem |
||
| 43 | :returns: word stem |
||
| 44 | :rtype: str |
||
| 45 | |||
| 46 | >>> lovins('reading') |
||
| 47 | 'read' |
||
| 48 | >>> lovins('suspension') |
||
| 49 | 'suspens' |
||
| 50 | >>> lovins('elusiveness') |
||
| 51 | 'elus' |
||
| 52 | """ |
||
| 53 | # lowercase, normalize, and compose |
||
| 54 | 1 | word = normalize('NFC', text_type(word.lower())) |
|
| 55 | |||
| 56 | 1 | def cond_b(word, suffix_len): |
|
| 57 | """Return Lovins' condition B. |
||
| 58 | |||
| 59 | :param str word: word to check |
||
| 60 | :param int suffix_len: suffix length |
||
| 61 | :rtype: bool |
||
| 62 | """ |
||
| 63 | 1 | return len(word) - suffix_len >= 3 |
|
| 64 | |||
| 65 | 1 | def cond_c(word, suffix_len): |
|
| 66 | """Return Lovins' condition C. |
||
| 67 | |||
| 68 | :param str word: word to check |
||
| 69 | :param int suffix_len: suffix length |
||
| 70 | :rtype: bool |
||
| 71 | """ |
||
| 72 | 1 | return len(word) - suffix_len >= 4 |
|
| 73 | |||
| 74 | 1 | def cond_d(word, suffix_len): |
|
| 75 | """Return Lovins' condition D. |
||
| 76 | |||
| 77 | :param str word: word to check |
||
| 78 | :param int suffix_len: suffix length |
||
| 79 | :rtype: bool |
||
| 80 | """ |
||
| 81 | 1 | return len(word) - suffix_len >= 5 |
|
| 82 | |||
| 83 | 1 | def cond_e(word, suffix_len): |
|
| 84 | """Return Lovins' condition E. |
||
| 85 | |||
| 86 | :param str word: word to check |
||
| 87 | :param int suffix_len: suffix length |
||
| 88 | :rtype: bool |
||
| 89 | """ |
||
| 90 | 1 | return word[-suffix_len - 1] != 'e' |
|
| 91 | |||
| 92 | 1 | def cond_f(word, suffix_len): |
|
| 93 | """Return Lovins' condition F. |
||
| 94 | |||
| 95 | :param str word: word to check |
||
| 96 | :param int suffix_len: suffix length |
||
| 97 | :rtype: bool |
||
| 98 | """ |
||
| 99 | 1 | return len(word) - suffix_len >= 3 and word[-suffix_len - 1] != 'e' |
|
| 100 | |||
| 101 | 1 | def cond_g(word, suffix_len): |
|
| 102 | """Return Lovins' condition G. |
||
| 103 | |||
| 104 | :param str word: word to check |
||
| 105 | :param int suffix_len: suffix length |
||
| 106 | :rtype: bool |
||
| 107 | """ |
||
| 108 | 1 | return len(word) - suffix_len >= 3 and word[-suffix_len - 1] == 'f' |
|
| 109 | |||
| 110 | 1 | def cond_h(word, suffix_len): |
|
| 111 | """Return Lovins' condition H. |
||
| 112 | |||
| 113 | :param str word: word to check |
||
| 114 | :param int suffix_len: suffix length |
||
| 115 | :rtype: bool |
||
| 116 | """ |
||
| 117 | 1 | return ( |
|
| 118 | word[-suffix_len - 1] == 't' |
||
| 119 | or word[-suffix_len - 2 : -suffix_len] == 'll' |
||
| 120 | ) |
||
| 121 | |||
| 122 | 1 | def cond_i(word, suffix_len): |
|
| 123 | """Return Lovins' condition I. |
||
| 124 | |||
| 125 | :param str word: word to check |
||
| 126 | :param int suffix_len: suffix length |
||
| 127 | :rtype: bool |
||
| 128 | """ |
||
| 129 | 1 | return word[-suffix_len - 1] not in {'e', 'o'} |
|
| 130 | |||
| 131 | 1 | def cond_j(word, suffix_len): |
|
| 132 | """Return Lovins' condition J. |
||
| 133 | |||
| 134 | :param str word: word to check |
||
| 135 | :param int suffix_len: suffix length |
||
| 136 | :rtype: bool |
||
| 137 | """ |
||
| 138 | 1 | return word[-suffix_len - 1] not in {'a', 'e'} |
|
| 139 | |||
| 140 | 1 | def cond_k(word, suffix_len): |
|
| 141 | """Return Lovins' condition K. |
||
| 142 | |||
| 143 | :param str word: word to check |
||
| 144 | :param int suffix_len: suffix length |
||
| 145 | :rtype: bool |
||
| 146 | """ |
||
| 147 | 1 | return len(word) - suffix_len >= 3 and ( |
|
| 148 | word[-suffix_len - 1] in {'i', 'l'} |
||
| 149 | or (word[-suffix_len - 3] == 'u' and word[-suffix_len - 1] == 'e') |
||
| 150 | ) |
||
| 151 | |||
| 152 | 1 | def cond_l(word, suffix_len): |
|
| 153 | """Return Lovins' condition L. |
||
| 154 | |||
| 155 | :param str word: word to check |
||
| 156 | :param int suffix_len: suffix length |
||
| 157 | :rtype: bool |
||
| 158 | """ |
||
| 159 | 1 | return ( |
|
| 160 | word[-suffix_len - 1] not in {'s', 'u', 'x'} |
||
| 161 | or word[-suffix_len - 1] == 'os' |
||
| 162 | ) |
||
| 163 | |||
| 164 | 1 | def cond_m(word, suffix_len): |
|
| 165 | """Return Lovins' condition M. |
||
| 166 | |||
| 167 | :param str word: word to check |
||
| 168 | :param int suffix_len: suffix length |
||
| 169 | :rtype: bool |
||
| 170 | """ |
||
| 171 | 1 | return word[-suffix_len - 1] not in {'a', 'c', 'e', 'm'} |
|
| 172 | |||
| 173 | 1 | def cond_n(word, suffix_len): |
|
| 174 | """Return Lovins' condition N. |
||
| 175 | |||
| 176 | :param str word: word to check |
||
| 177 | :param int suffix_len: suffix length |
||
| 178 | :rtype: bool |
||
| 179 | """ |
||
| 180 | 1 | if len(word) - suffix_len >= 3: |
|
| 181 | 1 | if word[-suffix_len - 3] == 's': |
|
| 182 | 1 | if len(word) - suffix_len >= 4: |
|
| 183 | 1 | return True |
|
| 184 | else: |
||
| 185 | 1 | return True |
|
| 186 | 1 | return False |
|
| 187 | |||
| 188 | 1 | def cond_o(word, suffix_len): |
|
| 189 | """Return Lovins' condition O. |
||
| 190 | |||
| 191 | :param str word: word to check |
||
| 192 | :param int suffix_len: suffix length |
||
| 193 | :rtype: bool |
||
| 194 | """ |
||
| 195 | 1 | return word[-suffix_len - 1] in {'i', 'l'} |
|
| 196 | |||
| 197 | 1 | def cond_p(word, suffix_len): |
|
| 198 | """Return Lovins' condition P. |
||
| 199 | |||
| 200 | :param str word: word to check |
||
| 201 | :param int suffix_len: suffix length |
||
| 202 | :rtype: bool |
||
| 203 | """ |
||
| 204 | 1 | return word[-suffix_len - 1] != 'c' |
|
| 205 | |||
| 206 | 1 | def cond_q(word, suffix_len): |
|
| 207 | """Return Lovins' condition Q. |
||
| 208 | |||
| 209 | :param str word: word to check |
||
| 210 | :param int suffix_len: suffix length |
||
| 211 | :rtype: bool |
||
| 212 | """ |
||
| 213 | 1 | return len(word) - suffix_len >= 3 and word[-suffix_len - 1] not in { |
|
| 214 | 'l', |
||
| 215 | 'n', |
||
| 216 | } |
||
| 217 | |||
| 218 | 1 | def cond_r(word, suffix_len): |
|
| 219 | """Return Lovins' condition R. |
||
| 220 | |||
| 221 | :param str word: word to check |
||
| 222 | :param int suffix_len: suffix length |
||
| 223 | :rtype: bool |
||
| 224 | """ |
||
| 225 | 1 | return word[-suffix_len - 1] in {'n', 'r'} |
|
| 226 | |||
| 227 | 1 | def cond_s(word, suffix_len): |
|
| 228 | """Return Lovins' condition S. |
||
| 229 | |||
| 230 | :param str word: word to check |
||
| 231 | :param int suffix_len: suffix length |
||
| 232 | :rtype: bool |
||
| 233 | """ |
||
| 234 | 1 | return word[-suffix_len - 2 : -suffix_len] == 'dr' or ( |
|
| 235 | word[-suffix_len - 1] == 't' |
||
| 236 | and word[-suffix_len - 2 : -suffix_len] != 'tt' |
||
| 237 | ) |
||
| 238 | |||
| 239 | 1 | def cond_t(word, suffix_len): |
|
| 240 | """Return Lovins' condition T. |
||
| 241 | |||
| 242 | :param str word: word to check |
||
| 243 | :param int suffix_len: suffix length |
||
| 244 | :rtype: bool |
||
| 245 | """ |
||
| 246 | 1 | return ( |
|
| 247 | word[-suffix_len - 1] in {'s', 't'} |
||
| 248 | and word[-suffix_len - 2 : -suffix_len] != 'ot' |
||
| 249 | ) |
||
| 250 | |||
| 251 | 1 | def cond_u(word, suffix_len): |
|
| 252 | """Return Lovins' condition U. |
||
| 253 | |||
| 254 | :param str word: word to check |
||
| 255 | :param int suffix_len: suffix length |
||
| 256 | :rtype: bool |
||
| 257 | """ |
||
| 258 | 1 | return word[-suffix_len - 1] in {'l', 'm', 'n', 'r'} |
|
| 259 | |||
| 260 | 1 | def cond_v(word, suffix_len): |
|
| 261 | """Return Lovins' condition V. |
||
| 262 | |||
| 263 | :param str word: word to check |
||
| 264 | :param int suffix_len: suffix length |
||
| 265 | :rtype: bool |
||
| 266 | """ |
||
| 267 | 1 | return word[-suffix_len - 1] == 'c' |
|
| 268 | |||
| 269 | 1 | def cond_w(word, suffix_len): |
|
| 270 | """Return Lovins' condition W. |
||
| 271 | |||
| 272 | :param str word: word to check |
||
| 273 | :param int suffix_len: suffix length |
||
| 274 | :rtype: bool |
||
| 275 | """ |
||
| 276 | 1 | return word[-suffix_len - 1] not in {'s', 'u'} |
|
| 277 | |||
| 278 | 1 | def cond_x(word, suffix_len): |
|
| 279 | """Return Lovins' condition X. |
||
| 280 | |||
| 281 | :param str word: word to check |
||
| 282 | :param int suffix_len: suffix length |
||
| 283 | :rtype: bool |
||
| 284 | """ |
||
| 285 | 1 | return word[-suffix_len - 1] in {'i', 'l'} or ( |
|
| 286 | word[-suffix_len - 3 : -suffix_len] == 'u' |
||
| 287 | and word[-suffix_len - 1] == 'e' |
||
| 288 | ) |
||
| 289 | |||
| 290 | 1 | def cond_y(word, suffix_len): |
|
| 291 | """Return Lovins' condition Y. |
||
| 292 | |||
| 293 | :param str word: word to check |
||
| 294 | :param int suffix_len: suffix length |
||
| 295 | :rtype: bool |
||
| 296 | """ |
||
| 297 | 1 | return word[-suffix_len - 2 : -suffix_len] == 'in' |
|
| 298 | |||
| 299 | 1 | def cond_z(word, suffix_len): |
|
| 300 | """Return Lovins' condition Z. |
||
| 301 | |||
| 302 | :param str word: word to check |
||
| 303 | :param int suffix_len: suffix length |
||
| 304 | :rtype: bool |
||
| 305 | """ |
||
| 306 | 1 | return word[-suffix_len - 1] != 'f' |
|
| 307 | |||
| 308 | 1 | def cond_aa(word, suffix_len): |
|
| 309 | """Return Lovins' condition AA. |
||
| 310 | |||
| 311 | :param str word: word to check |
||
| 312 | :param int suffix_len: suffix length |
||
| 313 | :rtype: bool |
||
| 314 | """ |
||
| 315 | 1 | return word[-suffix_len - 1] in {'d', 'f', 'l', 't'} or word[ |
|
| 316 | -suffix_len - 2 : -suffix_len |
||
| 317 | ] in {'ph', 'th', 'er', 'or', 'es'} |
||
| 318 | |||
| 319 | 1 | def cond_bb(word, suffix_len): |
|
| 320 | """Return Lovins' condition BB. |
||
| 321 | |||
| 322 | :param str word: word to check |
||
| 323 | :param int suffix_len: suffix length |
||
| 324 | :rtype: bool |
||
| 325 | """ |
||
| 326 | 1 | return ( |
|
| 327 | len(word) - suffix_len >= 3 |
||
| 328 | and word[-suffix_len - 3 : -suffix_len] != 'met' |
||
| 329 | and word[-suffix_len - 4 : -suffix_len] != 'ryst' |
||
| 330 | ) |
||
| 331 | |||
| 332 | 1 | def cond_cc(word, suffix_len): |
|
| 333 | """Return Lovins' condition CC. |
||
| 334 | |||
| 335 | :param str word: word to check |
||
| 336 | :param int suffix_len: suffix length |
||
| 337 | :rtype: bool |
||
| 338 | """ |
||
| 339 | 1 | return word[-suffix_len - 1] == 'l' |
|
| 340 | |||
| 341 | 1 | suffix = { |
|
| 342 | 'alistically': cond_b, |
||
| 343 | 'arizability': None, |
||
| 344 | 'izationally': cond_b, |
||
| 345 | 'antialness': None, |
||
| 346 | 'arisations': None, |
||
| 347 | 'arizations': None, |
||
| 348 | 'entialness': None, |
||
| 349 | 'allically': cond_c, |
||
| 350 | 'antaneous': None, |
||
| 351 | 'antiality': None, |
||
| 352 | 'arisation': None, |
||
| 353 | 'arization': None, |
||
| 354 | 'ationally': cond_b, |
||
| 355 | 'ativeness': None, |
||
| 356 | 'eableness': cond_e, |
||
| 357 | 'entations': None, |
||
| 358 | 'entiality': None, |
||
| 359 | 'entialize': None, |
||
| 360 | 'entiation': None, |
||
| 361 | 'ionalness': None, |
||
| 362 | 'istically': None, |
||
| 363 | 'itousness': None, |
||
| 364 | 'izability': None, |
||
| 365 | 'izational': None, |
||
| 366 | 'ableness': None, |
||
| 367 | 'arizable': None, |
||
| 368 | 'entation': None, |
||
| 369 | 'entially': None, |
||
| 370 | 'eousness': None, |
||
| 371 | 'ibleness': None, |
||
| 372 | 'icalness': None, |
||
| 373 | 'ionalism': None, |
||
| 374 | 'ionality': None, |
||
| 375 | 'ionalize': None, |
||
| 376 | 'iousness': None, |
||
| 377 | 'izations': None, |
||
| 378 | 'lessness': None, |
||
| 379 | 'ability': None, |
||
| 380 | 'aically': None, |
||
| 381 | 'alistic': cond_b, |
||
| 382 | 'alities': None, |
||
| 383 | 'ariness': cond_e, |
||
| 384 | 'aristic': None, |
||
| 385 | 'arizing': None, |
||
| 386 | 'ateness': None, |
||
| 387 | 'atingly': None, |
||
| 388 | 'ational': cond_b, |
||
| 389 | 'atively': None, |
||
| 390 | 'ativism': None, |
||
| 391 | 'elihood': cond_e, |
||
| 392 | 'encible': None, |
||
| 393 | 'entally': None, |
||
| 394 | 'entials': None, |
||
| 395 | 'entiate': None, |
||
| 396 | 'entness': None, |
||
| 397 | 'fulness': None, |
||
| 398 | 'ibility': None, |
||
| 399 | 'icalism': None, |
||
| 400 | 'icalist': None, |
||
| 401 | 'icality': None, |
||
| 402 | 'icalize': None, |
||
| 403 | 'ication': cond_g, |
||
| 404 | 'icianry': None, |
||
| 405 | 'ination': None, |
||
| 406 | 'ingness': None, |
||
| 407 | 'ionally': None, |
||
| 408 | 'isation': None, |
||
| 409 | 'ishness': None, |
||
| 410 | 'istical': None, |
||
| 411 | 'iteness': None, |
||
| 412 | 'iveness': None, |
||
| 413 | 'ivistic': None, |
||
| 414 | 'ivities': None, |
||
| 415 | 'ization': cond_f, |
||
| 416 | 'izement': None, |
||
| 417 | 'oidally': None, |
||
| 418 | 'ousness': None, |
||
| 419 | 'aceous': None, |
||
| 420 | 'acious': cond_b, |
||
| 421 | 'action': cond_g, |
||
| 422 | 'alness': None, |
||
| 423 | 'ancial': None, |
||
| 424 | 'ancies': None, |
||
| 425 | 'ancing': cond_b, |
||
| 426 | 'ariser': None, |
||
| 427 | 'arized': None, |
||
| 428 | 'arizer': None, |
||
| 429 | 'atable': None, |
||
| 430 | 'ations': cond_b, |
||
| 431 | 'atives': None, |
||
| 432 | 'eature': cond_z, |
||
| 433 | 'efully': None, |
||
| 434 | 'encies': None, |
||
| 435 | 'encing': None, |
||
| 436 | 'ential': None, |
||
| 437 | 'enting': cond_c, |
||
| 438 | 'entist': None, |
||
| 439 | 'eously': None, |
||
| 440 | 'ialist': None, |
||
| 441 | 'iality': None, |
||
| 442 | 'ialize': None, |
||
| 443 | 'ically': None, |
||
| 444 | 'icance': None, |
||
| 445 | 'icians': None, |
||
| 446 | 'icists': None, |
||
| 447 | 'ifully': None, |
||
| 448 | 'ionals': None, |
||
| 449 | 'ionate': cond_d, |
||
| 450 | 'ioning': None, |
||
| 451 | 'ionist': None, |
||
| 452 | 'iously': None, |
||
| 453 | 'istics': None, |
||
| 454 | 'izable': cond_e, |
||
| 455 | 'lessly': None, |
||
| 456 | 'nesses': None, |
||
| 457 | 'oidism': None, |
||
| 458 | 'acies': None, |
||
| 459 | 'acity': None, |
||
| 460 | 'aging': cond_b, |
||
| 461 | 'aical': None, |
||
| 462 | 'alist': None, |
||
| 463 | 'alism': cond_b, |
||
| 464 | 'ality': None, |
||
| 465 | 'alize': None, |
||
| 466 | 'allic': cond_bb, |
||
| 467 | 'anced': cond_b, |
||
| 468 | 'ances': cond_b, |
||
| 469 | 'antic': cond_c, |
||
| 470 | 'arial': None, |
||
| 471 | 'aries': None, |
||
| 472 | 'arily': None, |
||
| 473 | 'arity': cond_b, |
||
| 474 | 'arize': None, |
||
| 475 | 'aroid': None, |
||
| 476 | 'ately': None, |
||
| 477 | 'ating': cond_i, |
||
| 478 | 'ation': cond_b, |
||
| 479 | 'ative': None, |
||
| 480 | 'ators': None, |
||
| 481 | 'atory': None, |
||
| 482 | 'ature': cond_e, |
||
| 483 | 'early': cond_y, |
||
| 484 | 'ehood': None, |
||
| 485 | 'eless': None, |
||
| 486 | 'elity': None, |
||
| 487 | 'ement': None, |
||
| 488 | 'enced': None, |
||
| 489 | 'ences': None, |
||
| 490 | 'eness': cond_e, |
||
| 491 | 'ening': cond_e, |
||
| 492 | 'ental': None, |
||
| 493 | 'ented': cond_c, |
||
| 494 | 'ently': None, |
||
| 495 | 'fully': None, |
||
| 496 | 'ially': None, |
||
| 497 | 'icant': None, |
||
| 498 | 'ician': None, |
||
| 499 | 'icide': None, |
||
| 500 | 'icism': None, |
||
| 501 | 'icist': None, |
||
| 502 | 'icity': None, |
||
| 503 | 'idine': cond_i, |
||
| 504 | 'iedly': None, |
||
| 505 | 'ihood': None, |
||
| 506 | 'inate': None, |
||
| 507 | 'iness': None, |
||
| 508 | 'ingly': cond_b, |
||
| 509 | 'inism': cond_j, |
||
| 510 | 'inity': cond_cc, |
||
| 511 | 'ional': None, |
||
| 512 | 'ioned': None, |
||
| 513 | 'ished': None, |
||
| 514 | 'istic': None, |
||
| 515 | 'ities': None, |
||
| 516 | 'itous': None, |
||
| 517 | 'ively': None, |
||
| 518 | 'ivity': None, |
||
| 519 | 'izers': cond_f, |
||
| 520 | 'izing': cond_f, |
||
| 521 | 'oidal': None, |
||
| 522 | 'oides': None, |
||
| 523 | 'otide': None, |
||
| 524 | 'ously': None, |
||
| 525 | 'able': None, |
||
| 526 | 'ably': None, |
||
| 527 | 'ages': cond_b, |
||
| 528 | 'ally': cond_b, |
||
| 529 | 'ance': cond_b, |
||
| 530 | 'ancy': cond_b, |
||
| 531 | 'ants': cond_b, |
||
| 532 | 'aric': None, |
||
| 533 | 'arly': cond_k, |
||
| 534 | 'ated': cond_i, |
||
| 535 | 'ates': None, |
||
| 536 | 'atic': cond_b, |
||
| 537 | 'ator': None, |
||
| 538 | 'ealy': cond_y, |
||
| 539 | 'edly': cond_e, |
||
| 540 | 'eful': None, |
||
| 541 | 'eity': None, |
||
| 542 | 'ence': None, |
||
| 543 | 'ency': None, |
||
| 544 | 'ened': cond_e, |
||
| 545 | 'enly': cond_e, |
||
| 546 | 'eous': None, |
||
| 547 | 'hood': None, |
||
| 548 | 'ials': None, |
||
| 549 | 'ians': None, |
||
| 550 | 'ible': None, |
||
| 551 | 'ibly': None, |
||
| 552 | 'ical': None, |
||
| 553 | 'ides': cond_l, |
||
| 554 | 'iers': None, |
||
| 555 | 'iful': None, |
||
| 556 | 'ines': cond_m, |
||
| 557 | 'ings': cond_n, |
||
| 558 | 'ions': cond_b, |
||
| 559 | 'ious': None, |
||
| 560 | 'isms': cond_b, |
||
| 561 | 'ists': None, |
||
| 562 | 'itic': cond_h, |
||
| 563 | 'ized': cond_f, |
||
| 564 | 'izer': cond_f, |
||
| 565 | 'less': None, |
||
| 566 | 'lily': None, |
||
| 567 | 'ness': None, |
||
| 568 | 'ogen': None, |
||
| 569 | 'ward': None, |
||
| 570 | 'wise': None, |
||
| 571 | 'ying': cond_b, |
||
| 572 | 'yish': None, |
||
| 573 | 'acy': None, |
||
| 574 | 'age': cond_b, |
||
| 575 | 'aic': None, |
||
| 576 | 'als': cond_bb, |
||
| 577 | 'ant': cond_b, |
||
| 578 | 'ars': cond_o, |
||
| 579 | 'ary': cond_f, |
||
| 580 | 'ata': None, |
||
| 581 | 'ate': None, |
||
| 582 | 'eal': cond_y, |
||
| 583 | 'ear': cond_y, |
||
| 584 | 'ely': cond_e, |
||
| 585 | 'ene': cond_e, |
||
| 586 | 'ent': cond_c, |
||
| 587 | 'ery': cond_e, |
||
| 588 | 'ese': None, |
||
| 589 | 'ful': None, |
||
| 590 | 'ial': None, |
||
| 591 | 'ian': None, |
||
| 592 | 'ics': None, |
||
| 593 | 'ide': cond_l, |
||
| 594 | 'ied': None, |
||
| 595 | 'ier': None, |
||
| 596 | 'ies': cond_p, |
||
| 597 | 'ily': None, |
||
| 598 | 'ine': cond_m, |
||
| 599 | 'ing': cond_n, |
||
| 600 | 'ion': cond_q, |
||
| 601 | 'ish': cond_c, |
||
| 602 | 'ism': cond_b, |
||
| 603 | 'ist': None, |
||
| 604 | 'ite': cond_aa, |
||
| 605 | 'ity': None, |
||
| 606 | 'ium': None, |
||
| 607 | 'ive': None, |
||
| 608 | 'ize': cond_f, |
||
| 609 | 'oid': None, |
||
| 610 | 'one': cond_r, |
||
| 611 | 'ous': None, |
||
| 612 | 'ae': None, |
||
| 613 | 'al': cond_bb, |
||
| 614 | 'ar': cond_x, |
||
| 615 | 'as': cond_b, |
||
| 616 | 'ed': cond_e, |
||
| 617 | 'en': cond_f, |
||
| 618 | 'es': cond_e, |
||
| 619 | 'ia': None, |
||
| 620 | 'ic': None, |
||
| 621 | 'is': None, |
||
| 622 | 'ly': cond_b, |
||
| 623 | 'on': cond_s, |
||
| 624 | 'or': cond_t, |
||
| 625 | 'um': cond_u, |
||
| 626 | 'us': cond_v, |
||
| 627 | 'yl': cond_r, |
||
| 628 | '\'s': None, |
||
| 629 | 's\'': None, |
||
| 630 | 'a': None, |
||
| 631 | 'e': None, |
||
| 632 | 'i': None, |
||
| 633 | 'o': None, |
||
| 634 | 's': cond_w, |
||
| 635 | 'y': cond_b, |
||
| 636 | } |
||
| 637 | |||
| 638 | 1 | for suffix_len in range(11, 0, -1): |
|
| 639 | 1 | ending = word[-suffix_len:] |
|
| 640 | 1 | if ( |
|
| 641 | ending in suffix |
||
| 642 | and len(word) - suffix_len >= 2 |
||
| 643 | and (suffix[ending] is None or suffix[ending](word, suffix_len)) |
||
| 644 | ): |
||
| 645 | 1 | word = word[:-suffix_len] |
|
| 646 | 1 | break |
|
| 647 | |||
| 648 | 1 | def recode9(stem): |
|
| 649 | """Return Lovins' conditional recode rule 9.""" |
||
| 650 | 1 | if stem[-3:-2] in {'a', 'i', 'o'}: |
|
| 651 | 1 | return stem |
|
| 652 | 1 | return stem[:-2] + 'l' |
|
| 653 | |||
| 654 | 1 | def recode24(stem): |
|
| 655 | """Return Lovins' conditional recode rule 24.""" |
||
| 656 | 1 | if stem[-4:-3] == 's': |
|
| 657 | 1 | return stem |
|
| 658 | 1 | return stem[:-1] + 's' |
|
| 659 | |||
| 660 | 1 | def recode28(stem): |
|
| 661 | """Return Lovins' conditional recode rule 28.""" |
||
| 662 | 1 | if stem[-4:-3] in {'p', 't'}: |
|
| 663 | 1 | return stem |
|
| 664 | 1 | return stem[:-1] + 's' |
|
| 665 | |||
| 666 | 1 | def recode30(stem): |
|
| 667 | """Return Lovins' conditional recode rule 30.""" |
||
| 668 | 1 | if stem[-4:-3] == 'm': |
|
| 669 | 1 | return stem |
|
| 670 | 1 | return stem[:-1] + 's' |
|
| 671 | |||
| 672 | 1 | def recode32(stem): |
|
| 673 | """Return Lovins' conditional recode rule 32.""" |
||
| 674 | 1 | if stem[-3:-2] == 'n': |
|
| 675 | 1 | return stem |
|
| 676 | 1 | return stem[:-1] + 's' |
|
| 677 | |||
| 678 | 1 | if word[-2:] in { |
|
| 679 | 'bb', |
||
| 680 | 'dd', |
||
| 681 | 'gg', |
||
| 682 | 'll', |
||
| 683 | 'mm', |
||
| 684 | 'nn', |
||
| 685 | 'pp', |
||
| 686 | 'rr', |
||
| 687 | 'ss', |
||
| 688 | 'tt', |
||
| 689 | }: |
||
| 690 | 1 | word = word[:-1] |
|
| 691 | |||
| 692 | 1 | recode = ( |
|
| 693 | ('iev', 'ief'), |
||
| 694 | ('uct', 'uc'), |
||
| 695 | ('umpt', 'um'), |
||
| 696 | ('rpt', 'rb'), |
||
| 697 | ('urs', 'ur'), |
||
| 698 | ('istr', 'ister'), |
||
| 699 | ('metr', 'meter'), |
||
| 700 | ('olv', 'olut'), |
||
| 701 | ('ul', recode9), |
||
| 702 | ('bex', 'bic'), |
||
| 703 | ('dex', 'dic'), |
||
| 704 | ('pex', 'pic'), |
||
| 705 | ('tex', 'tic'), |
||
| 706 | ('ax', 'ac'), |
||
| 707 | ('ex', 'ec'), |
||
| 708 | ('ix', 'ic'), |
||
| 709 | ('lux', 'luc'), |
||
| 710 | ('uad', 'uas'), |
||
| 711 | ('vad', 'vas'), |
||
| 712 | ('cid', 'cis'), |
||
| 713 | ('lid', 'lis'), |
||
| 714 | ('erid', 'eris'), |
||
| 715 | ('pand', 'pans'), |
||
| 716 | ('end', recode24), |
||
| 717 | ('ond', 'ons'), |
||
| 718 | ('lud', 'lus'), |
||
| 719 | ('rud', 'rus'), |
||
| 720 | ('her', recode28), |
||
| 721 | ('mit', 'mis'), |
||
| 722 | ('ent', recode30), |
||
| 723 | ('ert', 'ers'), |
||
| 724 | ('et', recode32), |
||
| 725 | ('yt', 'ys'), |
||
| 726 | ('yz', 'ys'), |
||
| 727 | ) |
||
| 728 | |||
| 729 | 1 | for ending, replacement in recode: |
|
| 730 | 1 | if word.endswith(ending): |
|
| 731 | 1 | if callable(replacement): |
|
| 732 | 1 | word = replacement(word) |
|
| 733 | else: |
||
| 734 | 1 | word = word[: -len(ending)] + replacement |
|
| 735 | |||
| 736 | 1 | return word |
|
| 737 | |||
| 743 |
Loading history...