chrislit /
abydos
| Total Complexity | 628 |
| Total Lines | 2346 |
| Duplicated Lines | 1.62 % |
| Changes | 0 | ||
Duplicate code is one of the most pungent code smells. A rule that is often used is to re-structure code once it is duplicated in three or more places.
Common duplication problems, and corresponding solutions are:
Complex classes like abydos.stemmer 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 -*- |
||
|
coding-style
introduced
by
|
|||
| 2 | |||
| 3 | # Copyright 2014-2018 by Christopher C. Little. |
||
| 4 | # This file is part of Abydos. |
||
| 5 | # |
||
| 6 | # Abydos is free software: you can redistribute it and/or modify |
||
| 7 | # it under the terms of the GNU General Public License as published by |
||
| 8 | # the Free Software Foundation, either version 3 of the License, or |
||
| 9 | # (at your option) any later version. |
||
| 10 | # |
||
| 11 | # Abydos is distributed in the hope that it will be useful, |
||
| 12 | # but WITHOUT ANY WARRANTY; without even the implied warranty of |
||
| 13 | # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
||
| 14 | # GNU General Public License for more details. |
||
| 15 | # |
||
| 16 | # You should have received a copy of the GNU General Public License |
||
| 17 | # along with Abydos. If not, see <http://www.gnu.org/licenses/>. |
||
| 18 | |||
| 19 | """abydos.stemmer. |
||
| 20 | |||
| 21 | The stemmer module defines word stemmers including: |
||
| 22 | |||
| 23 | - the Lovins stemmer |
||
| 24 | - the Porter and Porter2 (Snowball English) stemmers |
||
| 25 | - Snowball stemmers for German, Dutch, Norwegian, Swedish, and Danish |
||
| 26 | - CLEF German, German plus, and Swedish stemmers |
||
| 27 | - Caumann's German stemmer |
||
| 28 | - UEA-Lite Stemmer |
||
| 29 | - Paice-Husk Stemmer |
||
| 30 | """ |
||
| 31 | |||
| 32 | from __future__ import unicode_literals |
||
| 33 | |||
| 34 | import re |
||
| 35 | import unicodedata |
||
| 36 | |||
| 37 | from six import text_type |
||
| 38 | from six.moves import range |
||
| 39 | |||
| 40 | |||
| 41 | def lovins(word): |
||
| 42 | """Return Lovins stem. |
||
| 43 | |||
| 44 | Lovins stemmer |
||
| 45 | |||
| 46 | The Lovins stemmer is described in Julie Beth Lovins's article at: |
||
| 47 | http://www.mt-archive.info/MT-1968-Lovins.pdf |
||
| 48 | |||
| 49 | :param word: the word to stem |
||
| 50 | :returns: word stem |
||
| 51 | :rtype: string |
||
| 52 | |||
| 53 | >>> lovins('reading') |
||
| 54 | 'read' |
||
| 55 | >>> lovins('suspension') |
||
| 56 | 'suspens' |
||
| 57 | >>> lovins('elusiveness') |
||
| 58 | 'elus' |
||
| 59 | """ |
||
| 60 | # pylint: disable=too-many-branches, too-many-locals |
||
| 61 | |||
| 62 | # lowercase, normalize, and compose |
||
| 63 | word = unicodedata.normalize('NFC', text_type(word.lower())) |
||
| 64 | |||
| 65 | def cond_b(word, suffix_len): |
||
| 66 | """Return Lovins' condition B.""" |
||
| 67 | return len(word)-suffix_len >= 3 |
||
| 68 | |||
| 69 | def cond_c(word, suffix_len): |
||
| 70 | """Return Lovins' condition C.""" |
||
| 71 | return len(word)-suffix_len >= 4 |
||
| 72 | |||
| 73 | def cond_d(word, suffix_len): |
||
| 74 | """Return Lovins' condition D.""" |
||
| 75 | return len(word)-suffix_len >= 5 |
||
| 76 | |||
| 77 | def cond_e(word, suffix_len): |
||
| 78 | """Return Lovins' condition E.""" |
||
| 79 | return word[-suffix_len-1] != 'e' |
||
| 80 | |||
| 81 | def cond_f(word, suffix_len): |
||
| 82 | """Return Lovins' condition F.""" |
||
| 83 | return (len(word)-suffix_len >= 3 and |
||
| 84 | word[-suffix_len-1] != 'e') |
||
| 85 | |||
| 86 | def cond_g(word, suffix_len): |
||
| 87 | """Return Lovins' condition G.""" |
||
| 88 | return (len(word)-suffix_len >= 3 and |
||
| 89 | word[-suffix_len-1] == 'f') |
||
| 90 | |||
| 91 | def cond_h(word, suffix_len): |
||
| 92 | """Return Lovins' condition H.""" |
||
| 93 | return (word[-suffix_len-1] == 't' or |
||
| 94 | word[-suffix_len-2:-suffix_len] == 'll') |
||
| 95 | |||
| 96 | def cond_i(word, suffix_len): |
||
| 97 | """Return Lovins' condition I.""" |
||
| 98 | return word[-suffix_len-1] not in {'e', 'o'} |
||
| 99 | |||
| 100 | def cond_j(word, suffix_len): |
||
| 101 | """Return Lovins' condition J.""" |
||
| 102 | return word[-suffix_len-1] not in {'a', 'e'} |
||
| 103 | |||
| 104 | def cond_k(word, suffix_len): |
||
| 105 | """Return Lovins' condition K.""" |
||
| 106 | return (len(word)-suffix_len >= 3 and |
||
| 107 | (word[-suffix_len-1] in {'i', 'l'} or |
||
| 108 | (word[-suffix_len-3] == 'u' and word[-suffix_len-1] == 'e'))) |
||
| 109 | |||
| 110 | def cond_l(word, suffix_len): |
||
| 111 | """Return Lovins' condition L.""" |
||
| 112 | return (word[-suffix_len-1] not in {'s', 'u', 'x'} or |
||
| 113 | word[-suffix_len-1] == 'os') |
||
| 114 | |||
| 115 | def cond_m(word, suffix_len): |
||
| 116 | """Return Lovins' condition M.""" |
||
| 117 | return word[-suffix_len-1] not in {'a', 'c', 'e', 'm'} |
||
| 118 | |||
| 119 | def cond_n(word, suffix_len): |
||
| 120 | """Return Lovins' condition N.""" |
||
| 121 | if len(word)-suffix_len >= 3: |
||
| 122 | if word[-suffix_len-3] == 's': |
||
| 123 | if len(word)-suffix_len >= 4: |
||
| 124 | return True |
||
| 125 | else: |
||
| 126 | return True |
||
| 127 | return False |
||
| 128 | |||
| 129 | def cond_o(word, suffix_len): |
||
| 130 | """Return Lovins' condition O.""" |
||
| 131 | return word[-suffix_len-1] in {'i', 'l'} |
||
| 132 | |||
| 133 | def cond_p(word, suffix_len): |
||
| 134 | """Return Lovins' condition P.""" |
||
| 135 | return word[-suffix_len-1] != 'c' |
||
| 136 | |||
| 137 | def cond_q(word, suffix_len): |
||
| 138 | """Return Lovins' condition Q.""" |
||
| 139 | return (len(word)-suffix_len >= 3 and |
||
| 140 | word[-suffix_len-1] not in {'l', 'n'}) |
||
| 141 | |||
| 142 | def cond_r(word, suffix_len): |
||
| 143 | """Return Lovins' condition R.""" |
||
| 144 | return word[-suffix_len-1] in {'n', 'r'} |
||
| 145 | |||
| 146 | def cond_s(word, suffix_len): |
||
| 147 | """Return Lovins' condition S.""" |
||
| 148 | return (word[-suffix_len-2:-suffix_len] == 'dr' or |
||
| 149 | (word[-suffix_len-1] == 't' and |
||
| 150 | word[-suffix_len-2:-suffix_len] != 'tt')) |
||
| 151 | |||
| 152 | def cond_t(word, suffix_len): |
||
| 153 | """Return Lovins' condition T.""" |
||
| 154 | return (word[-suffix_len-1] in {'s', 't'} and |
||
| 155 | word[-suffix_len-2:-suffix_len] != 'ot') |
||
| 156 | |||
| 157 | def cond_u(word, suffix_len): |
||
| 158 | """Return Lovins' condition U.""" |
||
| 159 | return word[-suffix_len-1] in {'l', 'm', 'n', 'r'} |
||
| 160 | |||
| 161 | def cond_v(word, suffix_len): |
||
| 162 | """Return Lovins' condition V.""" |
||
| 163 | return word[-suffix_len-1] == 'c' |
||
| 164 | |||
| 165 | def cond_w(word, suffix_len): |
||
| 166 | """Return Lovins' condition W.""" |
||
| 167 | return word[-suffix_len-1] not in {'s', 'u'} |
||
| 168 | |||
| 169 | def cond_x(word, suffix_len): |
||
| 170 | """Return Lovins' condition X.""" |
||
| 171 | return (word[-suffix_len-1] in {'i', 'l'} or |
||
| 172 | (word[-suffix_len-3:-suffix_len] == 'u' and |
||
| 173 | word[-suffix_len-1] == 'e')) |
||
| 174 | |||
| 175 | def cond_y(word, suffix_len): |
||
| 176 | """Return Lovins' condition Y.""" |
||
| 177 | return word[-suffix_len-2:-suffix_len] == 'in' |
||
| 178 | |||
| 179 | def cond_z(word, suffix_len): |
||
| 180 | """Return Lovins' condition Z.""" |
||
| 181 | return word[-suffix_len-1] != 'f' |
||
| 182 | |||
| 183 | def cond_aa(word, suffix_len): |
||
| 184 | """Return Lovins' condition AA.""" |
||
| 185 | return (word[-suffix_len-1] in {'d', 'f', 'l', 't'} or |
||
| 186 | word[-suffix_len-2:-suffix_len] in {'ph', 'th', 'er', 'or', |
||
| 187 | 'es'}) |
||
| 188 | |||
| 189 | def cond_bb(word, suffix_len): |
||
| 190 | """Return Lovins' condition BB.""" |
||
| 191 | return (len(word)-suffix_len >= 3 and |
||
| 192 | word[-suffix_len-3:-suffix_len] != 'met' and |
||
| 193 | word[-suffix_len-4:-suffix_len] != 'ryst') |
||
| 194 | |||
| 195 | def cond_cc(word, suffix_len): |
||
| 196 | """Return Lovins' condition CC.""" |
||
| 197 | return word[-suffix_len-1] == 'l' |
||
| 198 | |||
| 199 | suffix = {'alistically': cond_b, 'arizability': None, |
||
| 200 | 'izationally': cond_b, 'antialness': None, |
||
| 201 | 'arisations': None, 'arizations': None, 'entialness': None, |
||
| 202 | 'allically': cond_c, 'antaneous': None, 'antiality': None, |
||
| 203 | 'arisation': None, 'arization': None, 'ationally': cond_b, |
||
| 204 | 'ativeness': None, 'eableness': cond_e, 'entations': None, |
||
| 205 | 'entiality': None, 'entialize': None, 'entiation': None, |
||
| 206 | 'ionalness': None, 'istically': None, 'itousness': None, |
||
| 207 | 'izability': None, 'izational': None, 'ableness': None, |
||
| 208 | 'arizable': None, 'entation': None, 'entially': None, |
||
| 209 | 'eousness': None, 'ibleness': None, 'icalness': None, |
||
| 210 | 'ionalism': None, 'ionality': None, 'ionalize': None, |
||
| 211 | 'iousness': None, 'izations': None, 'lessness': None, |
||
| 212 | 'ability': None, 'aically': None, 'alistic': cond_b, |
||
| 213 | 'alities': None, 'ariness': cond_e, 'aristic': None, |
||
| 214 | 'arizing': None, 'ateness': None, 'atingly': None, |
||
| 215 | 'ational': cond_b, 'atively': None, 'ativism': None, |
||
| 216 | 'elihood': cond_e, 'encible': None, 'entally': None, |
||
| 217 | 'entials': None, 'entiate': None, 'entness': None, |
||
| 218 | 'fulness': None, 'ibility': None, 'icalism': None, |
||
| 219 | 'icalist': None, 'icality': None, 'icalize': None, |
||
| 220 | 'ication': cond_g, 'icianry': None, 'ination': None, |
||
| 221 | 'ingness': None, 'ionally': None, 'isation': None, |
||
| 222 | 'ishness': None, 'istical': None, 'iteness': None, |
||
| 223 | 'iveness': None, 'ivistic': None, 'ivities': None, |
||
| 224 | 'ization': cond_f, 'izement': None, 'oidally': None, |
||
| 225 | 'ousness': None, 'aceous': None, 'acious': cond_b, |
||
| 226 | 'action': cond_g, 'alness': None, 'ancial': None, |
||
| 227 | 'ancies': None, 'ancing': cond_b, 'ariser': None, |
||
| 228 | 'arized': None, 'arizer': None, 'atable': None, |
||
| 229 | 'ations': cond_b, 'atives': None, 'eature': cond_z, |
||
| 230 | 'efully': None, 'encies': None, 'encing': None, |
||
| 231 | 'ential': None, 'enting': cond_c, 'entist': None, |
||
| 232 | 'eously': None, 'ialist': None, 'iality': None, |
||
| 233 | 'ialize': None, 'ically': None, 'icance': None, |
||
| 234 | 'icians': None, 'icists': None, 'ifully': None, |
||
| 235 | 'ionals': None, 'ionate': cond_d, 'ioning': None, |
||
| 236 | 'ionist': None, 'iously': None, 'istics': None, |
||
| 237 | 'izable': cond_e, 'lessly': None, 'nesses': None, |
||
| 238 | 'oidism': None, 'acies': None, 'acity': None, |
||
| 239 | 'aging': cond_b, 'aical': None, 'alist': None, |
||
| 240 | 'alism': cond_b, 'ality': None, 'alize': None, |
||
| 241 | 'allic': cond_bb, 'anced': cond_b, 'ances': cond_b, |
||
| 242 | 'antic': cond_c, 'arial': None, 'aries': None, |
||
| 243 | 'arily': None, 'arity': cond_b, 'arize': None, |
||
| 244 | 'aroid': None, 'ately': None, 'ating': cond_i, |
||
| 245 | 'ation': cond_b, 'ative': None, 'ators': None, |
||
| 246 | 'atory': None, 'ature': cond_e, 'early': cond_y, |
||
| 247 | 'ehood': None, 'eless': None, 'elity': None, |
||
| 248 | 'ement': None, 'enced': None, 'ences': None, |
||
| 249 | 'eness': cond_e, 'ening': cond_e, 'ental': None, |
||
| 250 | 'ented': cond_c, 'ently': None, 'fully': None, |
||
| 251 | 'ially': None, 'icant': None, 'ician': None, |
||
| 252 | 'icide': None, 'icism': None, 'icist': None, |
||
| 253 | 'icity': None, 'idine': cond_i, 'iedly': None, |
||
| 254 | 'ihood': None, 'inate': None, 'iness': None, |
||
| 255 | 'ingly': cond_b, 'inism': cond_j, 'inity': cond_cc, |
||
| 256 | 'ional': None, 'ioned': None, 'ished': None, |
||
| 257 | 'istic': None, 'ities': None, 'itous': None, |
||
| 258 | 'ively': None, 'ivity': None, 'izers': cond_f, |
||
| 259 | 'izing': cond_f, 'oidal': None, 'oides': None, |
||
| 260 | 'otide': None, 'ously': None, 'able': None, 'ably': None, |
||
| 261 | 'ages': cond_b, 'ally': cond_b, 'ance': cond_b, 'ancy': cond_b, |
||
| 262 | 'ants': cond_b, 'aric': None, 'arly': cond_k, 'ated': cond_i, |
||
| 263 | 'ates': None, 'atic': cond_b, 'ator': None, 'ealy': cond_y, |
||
| 264 | 'edly': cond_e, 'eful': None, 'eity': None, 'ence': None, |
||
| 265 | 'ency': None, 'ened': cond_e, 'enly': cond_e, 'eous': None, |
||
| 266 | 'hood': None, 'ials': None, 'ians': None, 'ible': None, |
||
| 267 | 'ibly': None, 'ical': None, 'ides': cond_l, 'iers': None, |
||
| 268 | 'iful': None, 'ines': cond_m, 'ings': cond_n, 'ions': cond_b, |
||
| 269 | 'ious': None, 'isms': cond_b, 'ists': None, 'itic': cond_h, |
||
| 270 | 'ized': cond_f, 'izer': cond_f, 'less': None, 'lily': None, |
||
| 271 | 'ness': None, 'ogen': None, 'ward': None, 'wise': None, |
||
| 272 | 'ying': cond_b, 'yish': None, 'acy': None, 'age': cond_b, |
||
| 273 | 'aic': None, 'als': cond_bb, 'ant': cond_b, 'ars': cond_o, |
||
| 274 | 'ary': cond_f, 'ata': None, 'ate': None, 'eal': cond_y, |
||
| 275 | 'ear': cond_y, 'ely': cond_e, 'ene': cond_e, 'ent': cond_c, |
||
| 276 | 'ery': cond_e, 'ese': None, 'ful': None, 'ial': None, |
||
| 277 | 'ian': None, 'ics': None, 'ide': cond_l, 'ied': None, |
||
| 278 | 'ier': None, 'ies': cond_p, 'ily': None, 'ine': cond_m, |
||
| 279 | 'ing': cond_n, 'ion': cond_q, 'ish': cond_c, 'ism': cond_b, |
||
| 280 | 'ist': None, 'ite': cond_aa, 'ity': None, 'ium': None, |
||
| 281 | 'ive': None, 'ize': cond_f, 'oid': None, 'one': cond_r, |
||
| 282 | 'ous': None, 'ae': None, 'al': cond_bb, 'ar': cond_x, |
||
| 283 | 'as': cond_b, 'ed': cond_e, 'en': cond_f, 'es': cond_e, |
||
| 284 | 'ia': None, 'ic': None, 'is': None, 'ly': cond_b, |
||
| 285 | 'on': cond_s, 'or': cond_t, 'um': cond_u, 'us': cond_v, |
||
| 286 | 'yl': cond_r, '\'s': None, 's\'': None, 'a': None, |
||
| 287 | 'e': None, 'i': None, 'o': None, 's': cond_w, 'y': cond_b} |
||
| 288 | |||
| 289 | for suffix_len in range(11, 0, -1): |
||
| 290 | ending = word[-suffix_len:] |
||
| 291 | if (ending in suffix and |
||
| 292 | len(word)-suffix_len >= 2 and |
||
| 293 | (suffix[ending] is None or |
||
| 294 | suffix[ending](word, suffix_len))): |
||
| 295 | word = word[:-suffix_len] |
||
| 296 | break |
||
| 297 | |||
| 298 | def recode9(stem): |
||
| 299 | """Return Lovins' conditional recode rule 9.""" |
||
| 300 | if stem[-3:-2] in {'a', 'i', 'o'}: |
||
| 301 | return stem |
||
| 302 | return stem[:-2]+'l' |
||
| 303 | |||
| 304 | def recode24(stem): |
||
| 305 | """Return Lovins' conditional recode rule 24.""" |
||
| 306 | if stem[-4:-3] == 's': |
||
| 307 | return stem |
||
| 308 | return stem[:-1]+'s' |
||
| 309 | |||
| 310 | def recode28(stem): |
||
| 311 | """Return Lovins' conditional recode rule 28.""" |
||
| 312 | if stem[-4:-3] in {'p', 't'}: |
||
| 313 | return stem |
||
| 314 | return stem[:-1]+'s' |
||
| 315 | |||
| 316 | def recode30(stem): |
||
| 317 | """Return Lovins' conditional recode rule 30.""" |
||
| 318 | if stem[-4:-3] == 'm': |
||
| 319 | return stem |
||
| 320 | return stem[:-1]+'s' |
||
| 321 | |||
| 322 | def recode32(stem): |
||
| 323 | """Return Lovins' conditional recode rule 32.""" |
||
| 324 | if stem[-3:-2] == 'n': |
||
| 325 | return stem |
||
| 326 | return stem[:-1]+'s' |
||
| 327 | |||
| 328 | if word[-2:] in {'bb', 'dd', 'gg', 'll', 'mm', 'nn', 'pp', 'rr', 'ss', |
||
| 329 | 'tt'}: |
||
| 330 | word = word[:-1] |
||
| 331 | |||
| 332 | recode = (('iev', 'ief'), |
||
| 333 | ('uct', 'uc'), |
||
| 334 | ('umpt', 'um'), |
||
| 335 | ('rpt', 'rb'), |
||
| 336 | ('urs', 'ur'), |
||
| 337 | ('istr', 'ister'), |
||
| 338 | ('metr', 'meter'), |
||
| 339 | ('olv', 'olut'), |
||
| 340 | ('ul', recode9), |
||
| 341 | ('bex', 'bic'), |
||
| 342 | ('dex', 'dic'), |
||
| 343 | ('pex', 'pic'), |
||
| 344 | ('tex', 'tic'), |
||
| 345 | ('ax', 'ac'), |
||
| 346 | ('ex', 'ec'), |
||
| 347 | ('ix', 'ic'), |
||
| 348 | ('lux', 'luc'), |
||
| 349 | ('uad', 'uas'), |
||
| 350 | ('vad', 'vas'), |
||
| 351 | ('cid', 'cis'), |
||
| 352 | ('lid', 'lis'), |
||
| 353 | ('erid', 'eris'), |
||
| 354 | ('pand', 'pans'), |
||
| 355 | ('end', recode24), |
||
| 356 | ('ond', 'ons'), |
||
| 357 | ('lud', 'lus'), |
||
| 358 | ('rud', 'rus'), |
||
| 359 | ('her', recode28), |
||
| 360 | ('mit', 'mis'), |
||
| 361 | ('ent', recode30), |
||
| 362 | ('ert', 'ers'), |
||
| 363 | ('et', recode32), |
||
| 364 | ('yt', 'ys'), |
||
| 365 | ('yz', 'ys')) |
||
| 366 | |||
| 367 | for ending, replacement in recode: |
||
| 368 | if word.endswith(ending): |
||
| 369 | if callable(replacement): |
||
| 370 | word = replacement(word) |
||
| 371 | else: |
||
| 372 | word = word[:-len(ending)] + replacement |
||
| 373 | |||
| 374 | return word |
||
| 375 | |||
| 376 | |||
| 377 | def _m_degree(term, vowels): |
||
| 378 | """Return Porter helper function _m_degree value. |
||
| 379 | |||
| 380 | m-degree is equal to the number of V to C transitions |
||
| 381 | |||
| 382 | :param term: the word for which to calculate the m-degree |
||
| 383 | :param vowels: the set of vowels in the language |
||
| 384 | :returns: the m-degree as defined in the Porter stemmer definition |
||
| 385 | """ |
||
| 386 | mdeg = 0 |
||
| 387 | last_was_vowel = False |
||
| 388 | for letter in term: |
||
| 389 | if letter in vowels: |
||
| 390 | last_was_vowel = True |
||
| 391 | else: |
||
| 392 | if last_was_vowel: |
||
| 393 | mdeg += 1 |
||
| 394 | last_was_vowel = False |
||
| 395 | return mdeg |
||
| 396 | |||
| 397 | |||
| 398 | def _sb_has_vowel(term, vowels): |
||
| 399 | """Return Porter helper function _sb_has_vowel value. |
||
| 400 | |||
| 401 | :param term: the word to scan for vowels |
||
| 402 | :param vowels: the set of vowels in the language |
||
| 403 | :returns: true iff a vowel exists in the term (as defined in the Porter |
||
| 404 | stemmer definition) |
||
| 405 | """ |
||
| 406 | for letter in term: |
||
| 407 | if letter in vowels: |
||
| 408 | return True |
||
| 409 | return False |
||
| 410 | |||
| 411 | |||
| 412 | def _ends_in_doubled_cons(term, vowels): |
||
| 413 | """Return Porter helper function _ends_in_doubled_cons value. |
||
| 414 | |||
| 415 | :param term: the word to check for a final doubled consonant |
||
| 416 | :param vowels: the set of vowels in the language |
||
| 417 | :returns: true iff the stem ends in a doubled consonant (as defined in the |
||
| 418 | Porter stemmer definition) |
||
| 419 | """ |
||
| 420 | if len(term) > 1 and term[-1] not in vowels and term[-2] == term[-1]: |
||
| 421 | return True |
||
| 422 | return False |
||
| 423 | |||
| 424 | |||
| 425 | def _ends_in_cvc(term, vowels): |
||
| 426 | """Return Porter helper function _ends_in_cvc value. |
||
| 427 | |||
| 428 | :param term: the word to scan for cvc |
||
| 429 | :param vowels: the set of vowels in the language |
||
| 430 | :returns: true iff the stem ends in cvc (as defined in the Porter stemmer |
||
| 431 | definition) |
||
| 432 | """ |
||
| 433 | if len(term) > 2 and (term[-1] not in vowels and |
||
| 434 | term[-2] in vowels and |
||
| 435 | term[-3] not in vowels and |
||
| 436 | term[-1] not in tuple('wxY')): |
||
| 437 | return True |
||
| 438 | return False |
||
| 439 | |||
| 440 | |||
| 441 | def porter(word, early_english=False): |
||
| 442 | """Return Porter stem. |
||
| 443 | |||
| 444 | The Porter stemmer is defined at: |
||
| 445 | http://snowball.tartarus.org/algorithms/porter/stemmer.html |
||
| 446 | |||
| 447 | :param word: the word to calculate the stem of |
||
| 448 | :param early_english: set to True in order to remove -eth & -est (2nd & 3rd |
||
| 449 | person singular verbal agreement suffixes) |
||
| 450 | :returns: word stem |
||
| 451 | :rtype: str |
||
| 452 | |||
| 453 | >>> porter('reading') |
||
| 454 | 'read' |
||
| 455 | >>> porter('suspension') |
||
| 456 | 'suspens' |
||
| 457 | >>> porter('elusiveness') |
||
| 458 | 'elus' |
||
| 459 | |||
| 460 | >>> porter('eateth', early_english=True) |
||
| 461 | 'eat' |
||
| 462 | """ |
||
| 463 | # pylint: disable=too-many-branches |
||
| 464 | |||
| 465 | # lowercase, normalize, and compose |
||
| 466 | word = unicodedata.normalize('NFC', text_type(word.lower())) |
||
| 467 | |||
| 468 | # Return word if stem is shorter than 2 |
||
| 469 | if len(word) < 3: |
||
| 470 | return word |
||
| 471 | |||
| 472 | _vowels = {'a', 'e', 'i', 'o', 'u', 'y'} |
||
| 473 | # Re-map consonantal y to Y (Y will be C, y will be V) |
||
| 474 | if word[0] == 'y': |
||
| 475 | word = 'Y' + word[1:] |
||
| 476 | for i in range(1, len(word)): |
||
| 477 | if word[i] == 'y' and word[i-1] in _vowels: |
||
| 478 | word = word[:i] + 'Y' + word[i+1:] |
||
| 479 | |||
| 480 | # Step 1a |
||
| 481 | if word[-1] == 's': |
||
| 482 | if word[-4:] == 'sses': |
||
| 483 | word = word[:-2] |
||
| 484 | elif word[-3:] == 'ies': |
||
| 485 | word = word[:-2] |
||
| 486 | elif word[-2:] == 'ss': |
||
| 487 | pass |
||
| 488 | else: |
||
| 489 | word = word[:-1] |
||
| 490 | |||
| 491 | # Step 1b |
||
| 492 | step1b_flag = False |
||
| 493 | if word[-3:] == 'eed': |
||
| 494 | if _m_degree(word[:-3], _vowels) > 0: |
||
| 495 | word = word[:-1] |
||
| 496 | elif word[-2:] == 'ed': |
||
| 497 | if _sb_has_vowel(word[:-2], _vowels): |
||
| 498 | word = word[:-2] |
||
| 499 | step1b_flag = True |
||
| 500 | elif word[-3:] == 'ing': |
||
| 501 | if _sb_has_vowel(word[:-3], _vowels): |
||
| 502 | word = word[:-3] |
||
| 503 | step1b_flag = True |
||
| 504 | elif early_english: |
||
| 505 | if word[-3:] == 'est': |
||
| 506 | if _sb_has_vowel(word[:-3], _vowels): |
||
| 507 | word = word[:-3] |
||
| 508 | step1b_flag = True |
||
| 509 | elif word[-3:] == 'eth': |
||
| 510 | if _sb_has_vowel(word[:-3], _vowels): |
||
| 511 | word = word[:-3] |
||
| 512 | step1b_flag = True |
||
| 513 | |||
| 514 | if step1b_flag: |
||
| 515 | if word[-2:] in {'at', 'bl', 'iz'}: |
||
| 516 | word += 'e' |
||
| 517 | elif (_ends_in_doubled_cons(word, _vowels) and |
||
| 518 | word[-1] not in {'l', 's', 'z'}): |
||
| 519 | word = word[:-1] |
||
| 520 | elif _m_degree(word, _vowels) == 1 and _ends_in_cvc(word, _vowels): |
||
| 521 | word += 'e' |
||
| 522 | |||
| 523 | # Step 1c |
||
| 524 | if word[-1] in {'Y', 'y'} and _sb_has_vowel(word[:-1], _vowels): |
||
| 525 | word = word[:-1] + 'i' |
||
| 526 | |||
| 527 | # Step 2 |
||
| 528 | if len(word) > 1: |
||
| 529 | if word[-2] == 'a': |
||
| 530 | if word[-7:] == 'ational': |
||
| 531 | if _m_degree(word[:-7], _vowels) > 0: |
||
| 532 | word = word[:-5] + 'e' |
||
| 533 | elif word[-6:] == 'tional': |
||
| 534 | if _m_degree(word[:-6], _vowels) > 0: |
||
| 535 | word = word[:-2] |
||
| 536 | elif word[-2] == 'c': |
||
| 537 | if word[-4:] in {'enci', 'anci'}: |
||
| 538 | if _m_degree(word[:-4], _vowels) > 0: |
||
| 539 | word = word[:-1] + 'e' |
||
| 540 | elif word[-2] == 'e': |
||
| 541 | if word[-4:] == 'izer': |
||
| 542 | if _m_degree(word[:-4], _vowels) > 0: |
||
| 543 | word = word[:-1] |
||
| 544 | elif word[-2] == 'g': |
||
| 545 | if word[-4:] == 'logi': |
||
| 546 | if _m_degree(word[:-4], _vowels) > 0: |
||
| 547 | word = word[:-1] |
||
| 548 | elif word[-2] == 'l': |
||
| 549 | if word[-3:] == 'bli': |
||
| 550 | if _m_degree(word[:-3], _vowels) > 0: |
||
| 551 | word = word[:-1] + 'e' |
||
| 552 | elif word[-4:] == 'alli': |
||
| 553 | if _m_degree(word[:-4], _vowels) > 0: |
||
| 554 | word = word[:-2] |
||
| 555 | elif word[-5:] == 'entli': |
||
| 556 | if _m_degree(word[:-5], _vowels) > 0: |
||
| 557 | word = word[:-2] |
||
| 558 | elif word[-3:] == 'eli': |
||
| 559 | if _m_degree(word[:-3], _vowels) > 0: |
||
| 560 | word = word[:-2] |
||
| 561 | elif word[-5:] == 'ousli': |
||
| 562 | if _m_degree(word[:-5], _vowels) > 0: |
||
| 563 | word = word[:-2] |
||
| 564 | elif word[-2] == 'o': |
||
| 565 | if word[-7:] == 'ization': |
||
| 566 | if _m_degree(word[:-7], _vowels) > 0: |
||
| 567 | word = word[:-5] + 'e' |
||
| 568 | elif word[-5:] == 'ation': |
||
| 569 | if _m_degree(word[:-5], _vowels) > 0: |
||
| 570 | word = word[:-3] + 'e' |
||
| 571 | elif word[-4:] == 'ator': |
||
| 572 | if _m_degree(word[:-4], _vowels) > 0: |
||
| 573 | word = word[:-2] + 'e' |
||
| 574 | elif word[-2] == 's': |
||
| 575 | if word[-5:] == 'alism': |
||
| 576 | if _m_degree(word[:-5], _vowels) > 0: |
||
| 577 | word = word[:-3] |
||
| 578 | elif word[-7:] in {'iveness', 'fulness', 'ousness'}: |
||
| 579 | if _m_degree(word[:-7], _vowels) > 0: |
||
| 580 | word = word[:-4] |
||
| 581 | elif word[-2] == 't': |
||
| 582 | if word[-5:] == 'aliti': |
||
| 583 | if _m_degree(word[:-5], _vowels) > 0: |
||
| 584 | word = word[:-3] |
||
| 585 | elif word[-5:] == 'iviti': |
||
| 586 | if _m_degree(word[:-5], _vowels) > 0: |
||
| 587 | word = word[:-3] + 'e' |
||
| 588 | elif word[-6:] == 'biliti': |
||
| 589 | if _m_degree(word[:-6], _vowels) > 0: |
||
| 590 | word = word[:-5] + 'le' |
||
| 591 | |||
| 592 | # Step 3 |
||
| 593 | if word[-5:] == 'icate': |
||
| 594 | if _m_degree(word[:-5], _vowels) > 0: |
||
| 595 | word = word[:-3] |
||
| 596 | elif word[-5:] == 'ative': |
||
| 597 | if _m_degree(word[:-5], _vowels) > 0: |
||
| 598 | word = word[:-5] |
||
| 599 | elif word[-5:] in {'alize', 'iciti'}: |
||
| 600 | if _m_degree(word[:-5], _vowels) > 0: |
||
| 601 | word = word[:-3] |
||
| 602 | elif word[-4:] == 'ical': |
||
| 603 | if _m_degree(word[:-4], _vowels) > 0: |
||
| 604 | word = word[:-2] |
||
| 605 | elif word[-3:] == 'ful': |
||
| 606 | if _m_degree(word[:-3], _vowels) > 0: |
||
| 607 | word = word[:-3] |
||
| 608 | elif word[-4:] == 'ness': |
||
| 609 | if _m_degree(word[:-4], _vowels) > 0: |
||
| 610 | word = word[:-4] |
||
| 611 | |||
| 612 | # Step 4 |
||
| 613 | if word[-2:] == 'al': |
||
| 614 | if _m_degree(word[:-2], _vowels) > 1: |
||
| 615 | word = word[:-2] |
||
| 616 | elif word[-4:] == 'ance': |
||
| 617 | if _m_degree(word[:-4], _vowels) > 1: |
||
| 618 | word = word[:-4] |
||
| 619 | elif word[-4:] == 'ence': |
||
| 620 | if _m_degree(word[:-4], _vowels) > 1: |
||
| 621 | word = word[:-4] |
||
| 622 | elif word[-2:] == 'er': |
||
| 623 | if _m_degree(word[:-2], _vowels) > 1: |
||
| 624 | word = word[:-2] |
||
| 625 | elif word[-2:] == 'ic': |
||
| 626 | if _m_degree(word[:-2], _vowels) > 1: |
||
| 627 | word = word[:-2] |
||
| 628 | elif word[-4:] == 'able': |
||
| 629 | if _m_degree(word[:-4], _vowels) > 1: |
||
| 630 | word = word[:-4] |
||
| 631 | elif word[-4:] == 'ible': |
||
| 632 | if _m_degree(word[:-4], _vowels) > 1: |
||
| 633 | word = word[:-4] |
||
| 634 | elif word[-3:] == 'ant': |
||
| 635 | if _m_degree(word[:-3], _vowels) > 1: |
||
| 636 | word = word[:-3] |
||
| 637 | elif word[-5:] == 'ement': |
||
| 638 | if _m_degree(word[:-5], _vowels) > 1: |
||
| 639 | word = word[:-5] |
||
| 640 | elif word[-4:] == 'ment': |
||
| 641 | if _m_degree(word[:-4], _vowels) > 1: |
||
| 642 | word = word[:-4] |
||
| 643 | elif word[-3:] == 'ent': |
||
| 644 | if _m_degree(word[:-3], _vowels) > 1: |
||
| 645 | word = word[:-3] |
||
| 646 | elif word[-4:] in {'sion', 'tion'}: |
||
| 647 | if _m_degree(word[:-3], _vowels) > 1: |
||
| 648 | word = word[:-3] |
||
| 649 | elif word[-2:] == 'ou': |
||
| 650 | if _m_degree(word[:-2], _vowels) > 1: |
||
| 651 | word = word[:-2] |
||
| 652 | elif word[-3:] == 'ism': |
||
| 653 | if _m_degree(word[:-3], _vowels) > 1: |
||
| 654 | word = word[:-3] |
||
| 655 | elif word[-3:] == 'ate': |
||
| 656 | if _m_degree(word[:-3], _vowels) > 1: |
||
| 657 | word = word[:-3] |
||
| 658 | elif word[-3:] == 'iti': |
||
| 659 | if _m_degree(word[:-3], _vowels) > 1: |
||
| 660 | word = word[:-3] |
||
| 661 | elif word[-3:] == 'ous': |
||
| 662 | if _m_degree(word[:-3], _vowels) > 1: |
||
| 663 | word = word[:-3] |
||
| 664 | elif word[-3:] == 'ive': |
||
| 665 | if _m_degree(word[:-3], _vowels) > 1: |
||
| 666 | word = word[:-3] |
||
| 667 | elif word[-3:] == 'ize': |
||
| 668 | if _m_degree(word[:-3], _vowels) > 1: |
||
| 669 | word = word[:-3] |
||
| 670 | |||
| 671 | # Step 5a |
||
| 672 | if word[-1] == 'e': |
||
| 673 | if _m_degree(word[:-1], _vowels) > 1: |
||
| 674 | word = word[:-1] |
||
| 675 | elif (_m_degree(word[:-1], _vowels) == 1 and |
||
| 676 | not _ends_in_cvc(word[:-1], _vowels)): |
||
| 677 | word = word[:-1] |
||
| 678 | |||
| 679 | # Step 5b |
||
| 680 | if word[-2:] == 'll' and _m_degree(word, _vowels) > 1: |
||
| 681 | word = word[:-1] |
||
| 682 | |||
| 683 | # Change 'Y' back to 'y' if it survived stemming |
||
| 684 | for i in range(len(word)): |
||
| 685 | if word[i] == 'Y': |
||
| 686 | word = word[:i] + 'y' + word[i+1:] |
||
| 687 | |||
| 688 | return word |
||
| 689 | |||
| 690 | |||
| 691 | def _sb_r1(term, vowels, r1_prefixes=None): |
||
| 692 | """Return the R1 region, as defined in the Porter2 specification.""" |
||
| 693 | vowel_found = False |
||
| 694 | if hasattr(r1_prefixes, '__iter__'): |
||
| 695 | for prefix in r1_prefixes: |
||
| 696 | if term[:len(prefix)] == prefix: |
||
| 697 | return len(prefix) |
||
| 698 | |||
| 699 | for i in range(len(term)): |
||
| 700 | if not vowel_found and term[i] in vowels: |
||
| 701 | vowel_found = True |
||
| 702 | elif vowel_found and term[i] not in vowels: |
||
| 703 | return i + 1 |
||
| 704 | return len(term) |
||
| 705 | |||
| 706 | |||
| 707 | def _sb_r2(term, vowels, r1_prefixes=None): |
||
| 708 | """Return the R2 region, as defined in the Porter2 specification.""" |
||
| 709 | r1_start = _sb_r1(term, vowels, r1_prefixes) |
||
| 710 | return r1_start + _sb_r1(term[r1_start:], vowels) |
||
| 711 | |||
| 712 | |||
| 713 | def _sb_ends_in_short_syllable(term, vowels, codanonvowels): |
||
| 714 | """Return True iff term ends in a short syllable. |
||
| 715 | |||
| 716 | (...according to the Porter2 specification.) |
||
| 717 | |||
| 718 | NB: This is akin to the CVC test from the Porter stemmer. The description |
||
| 719 | is unfortunately poor/ambiguous. |
||
| 720 | """ |
||
| 721 | if not term: |
||
| 722 | return False |
||
| 723 | if len(term) == 2: |
||
| 724 | if term[-2] in vowels and term[-1] not in vowels: |
||
| 725 | return True |
||
| 726 | elif len(term) >= 3: |
||
| 727 | if ((term[-3] not in vowels and term[-2] in vowels and |
||
| 728 | term[-1] in codanonvowels)): |
||
| 729 | return True |
||
| 730 | return False |
||
| 731 | |||
| 732 | |||
| 733 | def _sb_short_word(term, vowels, codanonvowels, r1_prefixes=None): |
||
| 734 | """Return True iff term is a short word. |
||
| 735 | |||
| 736 | (...according to the Porter2 specification.) |
||
| 737 | """ |
||
| 738 | if ((_sb_r1(term, vowels, r1_prefixes) == len(term) and |
||
| 739 | _sb_ends_in_short_syllable(term, vowels, codanonvowels))): |
||
| 740 | return True |
||
| 741 | return False |
||
| 742 | |||
| 743 | |||
| 744 | def porter2(word, early_english=False): |
||
| 745 | """Return the Porter2 (Snowball English) stem. |
||
| 746 | |||
| 747 | The Porter2 (Snowball English) stemmer is defined at: |
||
| 748 | http://snowball.tartarus.org/algorithms/english/stemmer.html |
||
| 749 | |||
| 750 | :param word: the word to calculate the stem of |
||
| 751 | :param early_english: set to True in order to remove -eth & -est (2nd & 3rd |
||
| 752 | person singular verbal agreement suffixes) |
||
| 753 | :returns: word stem |
||
| 754 | :rtype: str |
||
| 755 | |||
| 756 | >>> porter2('reading') |
||
| 757 | 'read' |
||
| 758 | >>> porter2('suspension') |
||
| 759 | 'suspens' |
||
| 760 | >>> porter2('elusiveness') |
||
| 761 | 'elus' |
||
| 762 | |||
| 763 | >>> porter2('eateth', early_english=True) |
||
| 764 | 'eat' |
||
| 765 | """ |
||
| 766 | # pylint: disable=too-many-branches |
||
| 767 | # pylint: disable=too-many-return-statements |
||
| 768 | |||
| 769 | _vowels = {'a', 'e', 'i', 'o', 'u', 'y'} |
||
| 770 | _codanonvowels = {"'", 'b', 'c', 'd', 'f', 'g', 'h', 'j', 'k', 'l', 'm', |
||
| 771 | 'n', 'p', 'q', 'r', 's', 't', 'v', 'z'} |
||
| 772 | _doubles = {'bb', 'dd', 'ff', 'gg', 'mm', 'nn', 'pp', 'rr', 'tt'} |
||
| 773 | _li = {'c', 'd', 'e', 'g', 'h', 'k', 'm', 'n', 'r', 't'} |
||
| 774 | |||
| 775 | # R1 prefixes should be in order from longest to shortest to prevent |
||
| 776 | # masking |
||
| 777 | _r1_prefixes = ('commun', 'gener', 'arsen') |
||
| 778 | _exception1dict = { # special changes: |
||
| 779 | 'skis': 'ski', 'skies': 'sky', 'dying': 'die', |
||
| 780 | 'lying': 'lie', 'tying': 'tie', |
||
| 781 | # special -LY cases: |
||
| 782 | 'idly': 'idl', 'gently': 'gentl', 'ugly': 'ugli', |
||
| 783 | 'early': 'earli', 'only': 'onli', 'singly': 'singl'} |
||
| 784 | _exception1set = {'sky', 'news', 'howe', 'atlas', 'cosmos', 'bias', |
||
| 785 | 'andes'} |
||
| 786 | _exception2set = {'inning', 'outing', 'canning', 'herring', 'earring', |
||
| 787 | 'proceed', 'exceed', 'succeed'} |
||
| 788 | |||
| 789 | # lowercase, normalize, and compose |
||
| 790 | word = unicodedata.normalize('NFC', text_type(word.lower())) |
||
| 791 | # replace apostrophe-like characters with U+0027, per |
||
| 792 | # http://snowball.tartarus.org/texts/apostrophe.html |
||
| 793 | word = word.replace('’', '\'') |
||
| 794 | word = word.replace('’', '\'') |
||
| 795 | |||
| 796 | # Exceptions 1 |
||
| 797 | if word in _exception1dict: |
||
| 798 | return _exception1dict[word] |
||
| 799 | elif word in _exception1set: |
||
| 800 | return word |
||
| 801 | |||
| 802 | # Return word if stem is shorter than 3 |
||
| 803 | if len(word) < 3: |
||
| 804 | return word |
||
| 805 | |||
| 806 | # Remove initial ', if present. |
||
| 807 | while word and word[0] == '\'': |
||
| 808 | word = word[1:] |
||
| 809 | # Return word if stem is shorter than 2 |
||
| 810 | if len(word) < 2: |
||
| 811 | return word |
||
| 812 | |||
| 813 | # Re-map vocalic Y to y (Y will be C, y will be V) |
||
| 814 | if word[0] == 'y': |
||
| 815 | word = 'Y' + word[1:] |
||
| 816 | for i in range(1, len(word)): |
||
| 817 | if word[i] == 'y' and word[i-1] in _vowels: |
||
| 818 | word = word[:i] + 'Y' + word[i+1:] |
||
| 819 | |||
| 820 | r1_start = _sb_r1(word, _vowels, _r1_prefixes) |
||
| 821 | r2_start = _sb_r2(word, _vowels, _r1_prefixes) |
||
| 822 | |||
| 823 | # Step 0 |
||
| 824 | if word[-3:] == '\'s\'': |
||
| 825 | word = word[:-3] |
||
| 826 | elif word[-2:] == '\'s': |
||
| 827 | word = word[:-2] |
||
| 828 | elif word[-1:] == '\'': |
||
| 829 | word = word[:-1] |
||
| 830 | # Return word if stem is shorter than 2 |
||
| 831 | if len(word) < 3: |
||
| 832 | return word |
||
| 833 | |||
| 834 | # Step 1a |
||
| 835 | if word[-4:] == 'sses': |
||
| 836 | word = word[:-2] |
||
| 837 | elif word[-3:] in {'ied', 'ies'}: |
||
| 838 | if len(word) > 4: |
||
| 839 | word = word[:-2] |
||
| 840 | else: |
||
| 841 | word = word[:-1] |
||
| 842 | elif word[-2:] in {'us', 'ss'}: |
||
| 843 | pass |
||
| 844 | elif word[-1] == 's': |
||
| 845 | if _sb_has_vowel(word[:-2], _vowels): |
||
| 846 | word = word[:-1] |
||
| 847 | |||
| 848 | # Exceptions 2 |
||
| 849 | if word in _exception2set: |
||
| 850 | return word |
||
| 851 | |||
| 852 | # Step 1b |
||
| 853 | step1b_flag = False |
||
| 854 | if word[-5:] == 'eedly': |
||
| 855 | if len(word[r1_start:]) >= 5: |
||
| 856 | word = word[:-3] |
||
| 857 | elif word[-5:] == 'ingly': |
||
| 858 | if _sb_has_vowel(word[:-5], _vowels): |
||
| 859 | word = word[:-5] |
||
| 860 | step1b_flag = True |
||
| 861 | elif word[-4:] == 'edly': |
||
| 862 | if _sb_has_vowel(word[:-4], _vowels): |
||
| 863 | word = word[:-4] |
||
| 864 | step1b_flag = True |
||
| 865 | elif word[-3:] == 'eed': |
||
| 866 | if len(word[r1_start:]) >= 3: |
||
| 867 | word = word[:-1] |
||
| 868 | elif word[-3:] == 'ing': |
||
| 869 | if _sb_has_vowel(word[:-3], _vowels): |
||
| 870 | word = word[:-3] |
||
| 871 | step1b_flag = True |
||
| 872 | elif word[-2:] == 'ed': |
||
| 873 | if _sb_has_vowel(word[:-2], _vowels): |
||
| 874 | word = word[:-2] |
||
| 875 | step1b_flag = True |
||
| 876 | elif early_english: |
||
| 877 | if word[-3:] == 'est': |
||
| 878 | if _sb_has_vowel(word[:-3], _vowels): |
||
| 879 | word = word[:-3] |
||
| 880 | step1b_flag = True |
||
| 881 | elif word[-3:] == 'eth': |
||
| 882 | if _sb_has_vowel(word[:-3], _vowels): |
||
| 883 | word = word[:-3] |
||
| 884 | step1b_flag = True |
||
| 885 | |||
| 886 | if step1b_flag: |
||
| 887 | if word[-2:] in {'at', 'bl', 'iz'}: |
||
| 888 | word += 'e' |
||
| 889 | elif word[-2:] in _doubles: |
||
| 890 | word = word[:-1] |
||
| 891 | elif _sb_short_word(word, _vowels, _codanonvowels, _r1_prefixes): |
||
| 892 | word += 'e' |
||
| 893 | |||
| 894 | # Step 1c |
||
| 895 | if ((len(word) > 2 and word[-1] in {'Y', 'y'} and |
||
| 896 | word[-2] not in _vowels)): |
||
| 897 | word = word[:-1] + 'i' |
||
| 898 | |||
| 899 | # Step 2 |
||
| 900 | if word[-2] == 'a': |
||
| 901 | if word[-7:] == 'ational': |
||
| 902 | if len(word[r1_start:]) >= 7: |
||
| 903 | word = word[:-5] + 'e' |
||
| 904 | elif word[-6:] == 'tional': |
||
| 905 | if len(word[r1_start:]) >= 6: |
||
| 906 | word = word[:-2] |
||
| 907 | elif word[-2] == 'c': |
||
| 908 | if word[-4:] in {'enci', 'anci'}: |
||
| 909 | if len(word[r1_start:]) >= 4: |
||
| 910 | word = word[:-1] + 'e' |
||
| 911 | elif word[-2] == 'e': |
||
| 912 | if word[-4:] == 'izer': |
||
| 913 | if len(word[r1_start:]) >= 4: |
||
| 914 | word = word[:-1] |
||
| 915 | elif word[-2] == 'g': |
||
| 916 | if word[-3:] == 'ogi': |
||
| 917 | if ((r1_start >= 1 and len(word[r1_start:]) >= 3 and |
||
| 918 | word[-4] == 'l')): |
||
| 919 | word = word[:-1] |
||
| 920 | elif word[-2] == 'l': |
||
| 921 | if word[-6:] == 'lessli': |
||
| 922 | if len(word[r1_start:]) >= 6: |
||
| 923 | word = word[:-2] |
||
| 924 | elif word[-5:] in {'entli', 'fulli', 'ousli'}: |
||
| 925 | if len(word[r1_start:]) >= 5: |
||
| 926 | word = word[:-2] |
||
| 927 | elif word[-4:] == 'abli': |
||
| 928 | if len(word[r1_start:]) >= 4: |
||
| 929 | word = word[:-1] + 'e' |
||
| 930 | elif word[-4:] == 'alli': |
||
| 931 | if len(word[r1_start:]) >= 4: |
||
| 932 | word = word[:-2] |
||
| 933 | elif word[-3:] == 'bli': |
||
| 934 | if len(word[r1_start:]) >= 3: |
||
| 935 | word = word[:-1] + 'e' |
||
| 936 | elif word[-2:] == 'li': |
||
| 937 | if ((r1_start >= 1 and len(word[r1_start:]) >= 2 and |
||
| 938 | word[-3] in _li)): |
||
| 939 | word = word[:-2] |
||
| 940 | elif word[-2] == 'o': |
||
| 941 | if word[-7:] == 'ization': |
||
| 942 | if len(word[r1_start:]) >= 7: |
||
| 943 | word = word[:-5] + 'e' |
||
| 944 | elif word[-5:] == 'ation': |
||
| 945 | if len(word[r1_start:]) >= 5: |
||
| 946 | word = word[:-3] + 'e' |
||
| 947 | elif word[-4:] == 'ator': |
||
| 948 | if len(word[r1_start:]) >= 4: |
||
| 949 | word = word[:-2] + 'e' |
||
| 950 | elif word[-2] == 's': |
||
| 951 | if word[-7:] in {'fulness', 'ousness', 'iveness'}: |
||
| 952 | if len(word[r1_start:]) >= 7: |
||
| 953 | word = word[:-4] |
||
| 954 | elif word[-5:] == 'alism': |
||
| 955 | if len(word[r1_start:]) >= 5: |
||
| 956 | word = word[:-3] |
||
| 957 | elif word[-2] == 't': |
||
| 958 | if word[-6:] == 'biliti': |
||
| 959 | if len(word[r1_start:]) >= 6: |
||
| 960 | word = word[:-5] + 'le' |
||
| 961 | elif word[-5:] == 'aliti': |
||
| 962 | if len(word[r1_start:]) >= 5: |
||
| 963 | word = word[:-3] |
||
| 964 | elif word[-5:] == 'iviti': |
||
| 965 | if len(word[r1_start:]) >= 5: |
||
| 966 | word = word[:-3] + 'e' |
||
| 967 | |||
| 968 | # Step 3 |
||
| 969 | if word[-7:] == 'ational': |
||
| 970 | if len(word[r1_start:]) >= 7: |
||
| 971 | word = word[:-5] + 'e' |
||
| 972 | elif word[-6:] == 'tional': |
||
| 973 | if len(word[r1_start:]) >= 6: |
||
| 974 | word = word[:-2] |
||
| 975 | elif word[-5:] in {'alize', 'icate', 'iciti'}: |
||
| 976 | if len(word[r1_start:]) >= 5: |
||
| 977 | word = word[:-3] |
||
| 978 | elif word[-5:] == 'ative': |
||
| 979 | if len(word[r2_start:]) >= 5: |
||
| 980 | word = word[:-5] |
||
| 981 | elif word[-4:] == 'ical': |
||
| 982 | if len(word[r1_start:]) >= 4: |
||
| 983 | word = word[:-2] |
||
| 984 | elif word[-4:] == 'ness': |
||
| 985 | if len(word[r1_start:]) >= 4: |
||
| 986 | word = word[:-4] |
||
| 987 | elif word[-3:] == 'ful': |
||
| 988 | if len(word[r1_start:]) >= 3: |
||
| 989 | word = word[:-3] |
||
| 990 | |||
| 991 | # Step 4 |
||
| 992 | for suffix in ('ement', 'ance', 'ence', 'able', 'ible', 'ment', 'ant', |
||
| 993 | 'ent', 'ism', 'ate', 'iti', 'ous', 'ive', 'ize', 'al', 'er', |
||
| 994 | 'ic'): |
||
| 995 | if word[-len(suffix):] == suffix: |
||
| 996 | if len(word[r2_start:]) >= len(suffix): |
||
| 997 | word = word[:-len(suffix)] |
||
| 998 | break |
||
| 999 | else: |
||
| 1000 | if word[-3:] == 'ion': |
||
| 1001 | if ((len(word[r2_start:]) >= 3 and len(word) >= 4 and |
||
| 1002 | word[-4] in tuple('st'))): |
||
| 1003 | word = word[:-3] |
||
| 1004 | |||
| 1005 | # Step 5 |
||
| 1006 | if word[-1] == 'e': |
||
| 1007 | if (len(word[r2_start:]) >= 1 or |
||
| 1008 | (len(word[r1_start:]) >= 1 and |
||
| 1009 | not _sb_ends_in_short_syllable(word[:-1], _vowels, |
||
| 1010 | _codanonvowels))): |
||
| 1011 | word = word[:-1] |
||
| 1012 | elif word[-1] == 'l': |
||
| 1013 | if len(word[r2_start:]) >= 1 and word[-2] == 'l': |
||
| 1014 | word = word[:-1] |
||
| 1015 | |||
| 1016 | # Change 'Y' back to 'y' if it survived stemming |
||
| 1017 | for i in range(0, len(word)): |
||
| 1018 | if word[i] == 'Y': |
||
| 1019 | word = word[:i] + 'y' + word[i+1:] |
||
| 1020 | |||
| 1021 | return word |
||
| 1022 | |||
| 1023 | |||
| 1024 | def sb_german(word, alternate_vowels=False): |
||
| 1025 | """Return Snowball German stem. |
||
| 1026 | |||
| 1027 | The Snowball German stemmer is defined at: |
||
| 1028 | http://snowball.tartarus.org/algorithms/german/stemmer.html |
||
| 1029 | |||
| 1030 | :param word: the word to calculate the stem of |
||
| 1031 | :param alternate_vowels: composes ae as ä, oe as ö, and ue as ü before |
||
| 1032 | running the algorithm |
||
| 1033 | :returns: word stem |
||
| 1034 | :rtype: str |
||
| 1035 | |||
| 1036 | >>> sb_german('lesen') |
||
| 1037 | 'les' |
||
| 1038 | >>> sb_german('graues') |
||
| 1039 | 'grau' |
||
| 1040 | >>> sb_german('buchstabieren') |
||
| 1041 | 'buchstabi' |
||
| 1042 | """ |
||
| 1043 | # pylint: disable=too-many-branches |
||
| 1044 | |||
| 1045 | _vowels = {'a', 'e', 'i', 'o', 'u', 'y', 'ä', 'ö', 'ü'} |
||
| 1046 | _s_endings = {'b', 'd', 'f', 'g', 'h', 'k', 'l', 'm', 'n', 'r', 't'} |
||
| 1047 | _st_endings = {'b', 'd', 'f', 'g', 'h', 'k', 'l', 'm', 'n', 't'} |
||
| 1048 | |||
| 1049 | # lowercase, normalize, and compose |
||
| 1050 | word = unicodedata.normalize('NFC', word.lower()) |
||
| 1051 | word = word.replace('ß', 'ss') |
||
| 1052 | |||
| 1053 | if len(word) > 2: |
||
| 1054 | for i in range(2, len(word)): |
||
| 1055 | if word[i] in _vowels and word[i-2] in _vowels: |
||
| 1056 | if word[i-1] == 'u': |
||
| 1057 | word = word[:i-1] + 'U' + word[i:] |
||
| 1058 | elif word[i-1] == 'y': |
||
| 1059 | word = word[:i-1] + 'Y' + word[i:] |
||
| 1060 | |||
| 1061 | if alternate_vowels: |
||
| 1062 | word = word.replace('ae', 'ä') |
||
| 1063 | word = word.replace('oe', 'ö') |
||
| 1064 | word = word.replace('que', 'Q') |
||
| 1065 | word = word.replace('ue', 'ü') |
||
| 1066 | word = word.replace('Q', 'que') |
||
| 1067 | |||
| 1068 | r1_start = max(3, _sb_r1(word, _vowels)) |
||
| 1069 | r2_start = _sb_r2(word, _vowels) |
||
| 1070 | |||
| 1071 | # Step 1 |
||
| 1072 | niss_flag = False |
||
| 1073 | if word[-3:] == 'ern': |
||
| 1074 | if len(word[r1_start:]) >= 3: |
||
| 1075 | word = word[:-3] |
||
| 1076 | elif word[-2:] == 'em': |
||
| 1077 | if len(word[r1_start:]) >= 2: |
||
| 1078 | word = word[:-2] |
||
| 1079 | elif word[-2:] == 'er': |
||
| 1080 | if len(word[r1_start:]) >= 2: |
||
| 1081 | word = word[:-2] |
||
| 1082 | elif word[-2:] == 'en': |
||
| 1083 | if len(word[r1_start:]) >= 2: |
||
| 1084 | word = word[:-2] |
||
| 1085 | niss_flag = True |
||
| 1086 | elif word[-2:] == 'es': |
||
| 1087 | if len(word[r1_start:]) >= 2: |
||
| 1088 | word = word[:-2] |
||
| 1089 | niss_flag = True |
||
| 1090 | elif word[-1:] == 'e': |
||
| 1091 | if len(word[r1_start:]) >= 1: |
||
| 1092 | word = word[:-1] |
||
| 1093 | niss_flag = True |
||
| 1094 | elif word[-1:] == 's': |
||
| 1095 | if ((len(word[r1_start:]) >= 1 and len(word) >= 2 and |
||
| 1096 | word[-2] in _s_endings)): |
||
| 1097 | word = word[:-1] |
||
| 1098 | |||
| 1099 | if niss_flag and word[-4:] == 'niss': |
||
| 1100 | word = word[:-1] |
||
| 1101 | |||
| 1102 | # Step 2 |
||
| 1103 | if word[-3:] == 'est': |
||
| 1104 | if len(word[r1_start:]) >= 3: |
||
| 1105 | word = word[:-3] |
||
| 1106 | elif word[-2:] == 'en': |
||
| 1107 | if len(word[r1_start:]) >= 2: |
||
| 1108 | word = word[:-2] |
||
| 1109 | elif word[-2:] == 'er': |
||
| 1110 | if len(word[r1_start:]) >= 2: |
||
| 1111 | word = word[:-2] |
||
| 1112 | elif word[-2:] == 'st': |
||
| 1113 | if ((len(word[r1_start:]) >= 2 and len(word) >= 6 and |
||
| 1114 | word[-3] in _st_endings)): |
||
| 1115 | word = word[:-2] |
||
| 1116 | |||
| 1117 | # Step 3 |
||
| 1118 | if word[-4:] == 'isch': |
||
| 1119 | if len(word[r2_start:]) >= 4 and word[-5] != 'e': |
||
| 1120 | word = word[:-4] |
||
| 1121 | elif word[-4:] in {'lich', 'heit'}: |
||
| 1122 | if len(word[r2_start:]) >= 4: |
||
| 1123 | word = word[:-4] |
||
| 1124 | if ((word[-2:] in {'er', 'en'} and |
||
| 1125 | len(word[r1_start:]) >= 2)): |
||
| 1126 | word = word[:-2] |
||
| 1127 | elif word[-4:] == 'keit': |
||
| 1128 | if len(word[r2_start:]) >= 4: |
||
| 1129 | word = word[:-4] |
||
| 1130 | if word[-4:] == 'lich' and len(word[r2_start:]) >= 4: |
||
| 1131 | word = word[:-4] |
||
| 1132 | elif word[-2:] == 'ig' and len(word[r2_start:]) >= 2: |
||
| 1133 | word = word[:-2] |
||
| 1134 | elif word[-3:] in {'end', 'ung'}: |
||
| 1135 | if len(word[r2_start:]) >= 3: |
||
| 1136 | word = word[:-3] |
||
| 1137 | if ((word[-2:] == 'ig' and len(word[r2_start:]) >= 2 and |
||
| 1138 | word[-3] != 'e')): |
||
| 1139 | word = word[:-2] |
||
| 1140 | elif word[-2:] in {'ig', 'ik'}: |
||
| 1141 | if len(word[r2_start:]) >= 2 and word[-3] != 'e': |
||
| 1142 | word = word[:-2] |
||
| 1143 | |||
| 1144 | # Change 'Y' and 'U' back to lowercase if survived stemming |
||
| 1145 | for i in range(0, len(word)): |
||
| 1146 | if word[i] == 'Y': |
||
| 1147 | word = word[:i] + 'y' + word[i+1:] |
||
| 1148 | elif word[i] == 'U': |
||
| 1149 | word = word[:i] + 'u' + word[i+1:] |
||
| 1150 | |||
| 1151 | # Remove umlauts |
||
| 1152 | _umlauts = dict(zip((ord(_) for _ in 'äöü'), 'aou')) |
||
| 1153 | word = word.translate(_umlauts) |
||
| 1154 | |||
| 1155 | return word |
||
| 1156 | |||
| 1157 | |||
| 1158 | def sb_dutch(word): |
||
| 1159 | """Return Snowball Dutch stem. |
||
| 1160 | |||
| 1161 | The Snowball Dutch stemmer is defined at: |
||
| 1162 | http://snowball.tartarus.org/algorithms/dutch/stemmer.html |
||
| 1163 | |||
| 1164 | :param word: the word to calculate the stem of |
||
| 1165 | :returns: word stem |
||
| 1166 | :rtype: str |
||
| 1167 | |||
| 1168 | >>> sb_dutch('lezen') |
||
| 1169 | 'lez' |
||
| 1170 | >>> sb_dutch('opschorting') |
||
| 1171 | 'opschort' |
||
| 1172 | >>> sb_dutch('ongrijpbaarheid') |
||
| 1173 | 'ongrijp' |
||
| 1174 | """ |
||
| 1175 | # pylint: disable=too-many-branches |
||
| 1176 | |||
| 1177 | _vowels = {'a', 'e', 'i', 'o', 'u', 'y', 'è'} |
||
| 1178 | _not_s_endings = {'a', 'e', 'i', 'j', 'o', 'u', 'y', 'è'} |
||
| 1179 | |||
| 1180 | def _undouble(word): |
||
| 1181 | """Undouble endings -kk, -dd, and -tt.""" |
||
| 1182 | if ((len(word) > 1 and word[-1] == word[-2] and |
||
| 1183 | word[-1] in {'d', 'k', 't'})): |
||
| 1184 | return word[:-1] |
||
| 1185 | return word |
||
| 1186 | |||
| 1187 | # lowercase, normalize, decompose, filter umlauts & acutes out, and compose |
||
| 1188 | word = unicodedata.normalize('NFC', text_type(word.lower())) |
||
| 1189 | _accented = dict(zip((ord(_) for _ in 'äëïöüáéíóú'), 'aeiouaeiou')) |
||
| 1190 | word = word.translate(_accented) |
||
| 1191 | |||
| 1192 | for i in range(len(word)): |
||
| 1193 | if i == 0 and word[0] == 'y': |
||
| 1194 | word = 'Y' + word[1:] |
||
| 1195 | elif word[i] == 'y' and word[i-1] in _vowels: |
||
| 1196 | word = word[:i] + 'Y' + word[i+1:] |
||
| 1197 | elif (word[i] == 'i' and word[i-1] in _vowels and i+1 < len(word) and |
||
| 1198 | word[i+1] in _vowels): |
||
| 1199 | word = word[:i] + 'I' + word[i+1:] |
||
| 1200 | |||
| 1201 | r1_start = max(3, _sb_r1(word, _vowels)) |
||
| 1202 | r2_start = _sb_r2(word, _vowels) |
||
| 1203 | |||
| 1204 | # Step 1 |
||
| 1205 | if word[-5:] == 'heden': |
||
| 1206 | if len(word[r1_start:]) >= 5: |
||
| 1207 | word = word[:-3] + 'id' |
||
| 1208 | elif word[-3:] == 'ene': |
||
| 1209 | if ((len(word[r1_start:]) >= 3 and |
||
| 1210 | (word[-4] not in _vowels and word[-6:-3] != 'gem'))): |
||
| 1211 | word = _undouble(word[:-3]) |
||
| 1212 | elif word[-2:] == 'en': |
||
| 1213 | if ((len(word[r1_start:]) >= 2 and |
||
| 1214 | (word[-3] not in _vowels and word[-5:-2] != 'gem'))): |
||
| 1215 | word = _undouble(word[:-2]) |
||
| 1216 | elif word[-2:] == 'se': |
||
| 1217 | if len(word[r1_start:]) >= 2 and word[-3] not in _not_s_endings: |
||
| 1218 | word = word[:-2] |
||
| 1219 | elif word[-1:] == 's': |
||
| 1220 | if len(word[r1_start:]) >= 1 and word[-2] not in _not_s_endings: |
||
| 1221 | word = word[:-1] |
||
| 1222 | |||
| 1223 | # Step 2 |
||
| 1224 | e_removed = False |
||
| 1225 | if word[-1:] == 'e': |
||
| 1226 | if len(word[r1_start:]) >= 1 and word[-2] not in _vowels: |
||
| 1227 | word = _undouble(word[:-1]) |
||
| 1228 | e_removed = True |
||
| 1229 | |||
| 1230 | # Step 3a |
||
| 1231 | if word[-4:] == 'heid': |
||
| 1232 | if len(word[r2_start:]) >= 4 and word[-5] != 'c': |
||
| 1233 | word = word[:-4] |
||
| 1234 | if word[-2:] == 'en': |
||
| 1235 | if ((len(word[r1_start:]) >= 2 and |
||
| 1236 | (word[-3] not in _vowels and word[-5:-2] != 'gem'))): |
||
| 1237 | word = _undouble(word[:-2]) |
||
| 1238 | |||
| 1239 | # Step 3b |
||
| 1240 | if word[-4:] == 'lijk': |
||
| 1241 | if len(word[r2_start:]) >= 4: |
||
| 1242 | word = word[:-4] |
||
| 1243 | # Repeat step 2 |
||
| 1244 | if word[-1:] == 'e': |
||
| 1245 | if len(word[r1_start:]) >= 1 and word[-2] not in _vowels: |
||
| 1246 | word = _undouble(word[:-1]) |
||
| 1247 | elif word[-4:] == 'baar': |
||
| 1248 | if len(word[r2_start:]) >= 4: |
||
| 1249 | word = word[:-4] |
||
| 1250 | elif word[-3:] in ('end', 'ing'): |
||
| 1251 | if len(word[r2_start:]) >= 3: |
||
| 1252 | word = word[:-3] |
||
| 1253 | if ((word[-2:] == 'ig' and len(word[r2_start:]) >= 2 and |
||
| 1254 | word[-3] != 'e')): |
||
| 1255 | word = word[:-2] |
||
| 1256 | else: |
||
| 1257 | word = _undouble(word) |
||
| 1258 | elif word[-3:] == 'bar': |
||
| 1259 | if len(word[r2_start:]) >= 3 and e_removed: |
||
| 1260 | word = word[:-3] |
||
| 1261 | elif word[-2:] == 'ig': |
||
| 1262 | if len(word[r2_start:]) >= 2 and word[-3] != 'e': |
||
| 1263 | word = word[:-2] |
||
| 1264 | |||
| 1265 | # Step 4 |
||
| 1266 | if ((len(word) >= 4 and |
||
| 1267 | word[-3] == word[-2] and word[-2] in {'a', 'e', 'o', 'u'} and |
||
| 1268 | word[-4] not in _vowels and |
||
| 1269 | word[-1] not in _vowels and word[-1] != 'I')): |
||
| 1270 | word = word[:-2] + word[-1] |
||
| 1271 | |||
| 1272 | # Change 'Y' and 'U' back to lowercase if survived stemming |
||
| 1273 | for i in range(0, len(word)): |
||
| 1274 | if word[i] == 'Y': |
||
| 1275 | word = word[:i] + 'y' + word[i+1:] |
||
| 1276 | elif word[i] == 'I': |
||
| 1277 | word = word[:i] + 'i' + word[i+1:] |
||
| 1278 | |||
| 1279 | return word |
||
| 1280 | |||
| 1281 | |||
| 1282 | def sb_norwegian(word): |
||
| 1283 | """Return Snowball Norwegian stem. |
||
| 1284 | |||
| 1285 | The Snowball Norwegian stemmer is defined at: |
||
| 1286 | http://snowball.tartarus.org/algorithms/norwegian/stemmer.html |
||
| 1287 | |||
| 1288 | :param word: the word to calculate the stem of |
||
| 1289 | :returns: word stem |
||
| 1290 | :rtype: str |
||
| 1291 | |||
| 1292 | >>> sb_norwegian('lese') |
||
| 1293 | 'les' |
||
| 1294 | >>> sb_norwegian('suspensjon') |
||
| 1295 | 'suspensjon' |
||
| 1296 | >>> sb_norwegian('sikkerhet') |
||
| 1297 | 'sikker' |
||
| 1298 | """ |
||
| 1299 | _vowels = {'a', 'e', 'i', 'o', 'u', 'y', 'å', 'æ', 'ø'} |
||
| 1300 | _s_endings = {'b', 'c', 'd', 'f', 'g', 'h', 'j', 'l', 'm', 'n', 'o', 'p', |
||
| 1301 | 'r', 't', 'v', 'y', 'z'} |
||
| 1302 | # lowercase, normalize, and compose |
||
| 1303 | word = unicodedata.normalize('NFC', text_type(word.lower())) |
||
| 1304 | |||
| 1305 | r1_start = min(max(3, _sb_r1(word, _vowels)), len(word)) |
||
| 1306 | |||
| 1307 | # Step 1 |
||
| 1308 | _r1 = word[r1_start:] |
||
| 1309 | if _r1[-7:] == 'hetenes': |
||
| 1310 | word = word[:-7] |
||
| 1311 | elif _r1[-6:] in {'hetene', 'hetens'}: |
||
| 1312 | word = word[:-6] |
||
| 1313 | elif _r1[-5:] in {'heten', 'heter', 'endes'}: |
||
| 1314 | word = word[:-5] |
||
| 1315 | elif _r1[-4:] in {'ande', 'ende', 'edes', 'enes', 'erte'}: |
||
| 1316 | if word[-4:] == 'erte': |
||
| 1317 | word = word[:-2] |
||
| 1318 | else: |
||
| 1319 | word = word[:-4] |
||
| 1320 | elif _r1[-3:] in {'ede', 'ane', 'ene', 'ens', 'ers', 'ets', 'het', 'ast', |
||
| 1321 | 'ert'}: |
||
| 1322 | if word[-3:] == 'ert': |
||
| 1323 | word = word[:-1] |
||
| 1324 | else: |
||
| 1325 | word = word[:-3] |
||
| 1326 | elif _r1[-2:] in {'en', 'ar', 'er', 'as', 'es', 'et'}: |
||
| 1327 | word = word[:-2] |
||
| 1328 | elif _r1[-1:] in {'a', 'e'}: |
||
| 1329 | word = word[:-1] |
||
| 1330 | elif _r1[-1:] == 's': |
||
| 1331 | if (((len(word) > 1 and word[-2] in _s_endings) or |
||
| 1332 | (len(word) > 2 and word[-2] == 'k' and word[-3] not in _vowels))): |
||
| 1333 | word = word[:-1] |
||
| 1334 | |||
| 1335 | # Step 2 |
||
| 1336 | if word[r1_start:][-2:] in {'dt', 'vt'}: |
||
| 1337 | word = word[:-1] |
||
| 1338 | |||
| 1339 | # Step 3 |
||
| 1340 | _r1 = word[r1_start:] |
||
| 1341 | if _r1[-7:] == 'hetslov': |
||
| 1342 | word = word[:-7] |
||
| 1343 | elif _r1[-4:] in {'eleg', 'elig', 'elov', 'slov'}: |
||
| 1344 | word = word[:-4] |
||
| 1345 | elif _r1[-3:] in {'leg', 'eig', 'lig', 'els', 'lov'}: |
||
| 1346 | word = word[:-3] |
||
| 1347 | elif _r1[-2:] == 'ig': |
||
| 1348 | word = word[:-2] |
||
| 1349 | |||
| 1350 | return word |
||
| 1351 | |||
| 1352 | |||
| 1353 | def sb_swedish(word): |
||
| 1354 | """Return Snowball Swedish stem. |
||
| 1355 | |||
| 1356 | The Snowball Swedish stemmer is defined at: |
||
| 1357 | http://snowball.tartarus.org/algorithms/swedish/stemmer.html |
||
| 1358 | |||
| 1359 | :param word: the word to calculate the stem of |
||
| 1360 | :returns: word stem |
||
| 1361 | :rtype: str |
||
| 1362 | |||
| 1363 | >>> sb_swedish('undervisa') |
||
| 1364 | 'undervis' |
||
| 1365 | >>> sb_swedish('suspension') |
||
| 1366 | 'suspension' |
||
| 1367 | >>> sb_swedish('visshet') |
||
| 1368 | 'viss' |
||
| 1369 | """ |
||
| 1370 | _vowels = {'a', 'e', 'i', 'o', 'u', 'y', 'ä', 'å', 'ö'} |
||
| 1371 | _s_endings = {'b', 'c', 'd', 'f', 'g', 'h', 'j', 'k', 'l', 'm', 'n', |
||
| 1372 | 'o', 'p', 'r', 't', 'v', 'y'} |
||
| 1373 | |||
| 1374 | # lowercase, normalize, and compose |
||
| 1375 | word = unicodedata.normalize('NFC', text_type(word.lower())) |
||
| 1376 | |||
| 1377 | r1_start = min(max(3, _sb_r1(word, _vowels)), len(word)) |
||
| 1378 | |||
| 1379 | # Step 1 |
||
| 1380 | _r1 = word[r1_start:] |
||
| 1381 | View Code Duplication | if _r1[-7:] == 'heterna': |
|
| 1382 | word = word[:-7] |
||
| 1383 | elif _r1[-6:] == 'hetens': |
||
| 1384 | word = word[:-6] |
||
| 1385 | elif _r1[-5:] in {'anden', 'heten', 'heter', 'arnas', 'ernas', 'ornas', |
||
| 1386 | 'andes', 'arens', 'andet'}: |
||
| 1387 | word = word[:-5] |
||
| 1388 | elif _r1[-4:] in {'arna', 'erna', 'orna', 'ande', 'arne', 'aste', 'aren', |
||
| 1389 | 'ades', 'erns'}: |
||
| 1390 | word = word[:-4] |
||
| 1391 | elif _r1[-3:] in {'ade', 'are', 'ern', 'ens', 'het', 'ast'}: |
||
| 1392 | word = word[:-3] |
||
| 1393 | elif _r1[-2:] in {'ad', 'en', 'ar', 'er', 'or', 'as', 'es', 'at'}: |
||
| 1394 | word = word[:-2] |
||
| 1395 | elif _r1[-1:] in {'a', 'e'}: |
||
| 1396 | word = word[:-1] |
||
| 1397 | elif _r1[-1:] == 's': |
||
| 1398 | if len(word) > 1 and word[-2] in _s_endings: |
||
| 1399 | word = word[:-1] |
||
| 1400 | |||
| 1401 | # Step 2 |
||
| 1402 | if word[r1_start:][-2:] in {'dd', 'gd', 'nn', 'dt', 'gt', 'kt', 'tt'}: |
||
| 1403 | word = word[:-1] |
||
| 1404 | |||
| 1405 | # Step 3 |
||
| 1406 | _r1 = word[r1_start:] |
||
| 1407 | if _r1[-5:] == 'fullt': |
||
| 1408 | word = word[:-1] |
||
| 1409 | elif _r1[-4:] == 'löst': |
||
| 1410 | word = word[:-1] |
||
| 1411 | elif _r1[-3:] in {'lig', 'els'}: |
||
| 1412 | word = word[:-3] |
||
| 1413 | elif _r1[-2:] == 'ig': |
||
| 1414 | word = word[:-2] |
||
| 1415 | |||
| 1416 | return word |
||
| 1417 | |||
| 1418 | |||
| 1419 | def sb_danish(word): |
||
| 1420 | """Return Snowball Danish stem. |
||
| 1421 | |||
| 1422 | The Snowball Danish stemmer is defined at: |
||
| 1423 | http://snowball.tartarus.org/algorithms/danish/stemmer.html |
||
| 1424 | |||
| 1425 | :param word: the word to calculate the stem of |
||
| 1426 | :returns: word stem |
||
| 1427 | :rtype: str |
||
| 1428 | |||
| 1429 | >>> sb_danish('underviser') |
||
| 1430 | 'undervis' |
||
| 1431 | >>> sb_danish('suspension') |
||
| 1432 | 'suspension' |
||
| 1433 | >>> sb_danish('sikkerhed') |
||
| 1434 | 'sikker' |
||
| 1435 | """ |
||
| 1436 | _vowels = {'a', 'e', 'i', 'o', 'u', 'y', 'å', 'æ', 'ø'} |
||
| 1437 | _s_endings = {'a', 'b', 'c', 'd', 'f', 'g', 'h', 'j', 'k', 'l', 'm', 'n', |
||
| 1438 | 'o', 'p', 'r', 't', 'v', 'y', 'z', 'å'} |
||
| 1439 | |||
| 1440 | # lowercase, normalize, and compose |
||
| 1441 | word = unicodedata.normalize('NFC', text_type(word.lower())) |
||
| 1442 | |||
| 1443 | r1_start = min(max(3, _sb_r1(word, _vowels)), len(word)) |
||
| 1444 | |||
| 1445 | # Step 1 |
||
| 1446 | _r1 = word[r1_start:] |
||
| 1447 | View Code Duplication | if _r1[-7:] == 'erendes': |
|
| 1448 | word = word[:-7] |
||
| 1449 | elif _r1[-6:] in {'erende', 'hedens'}: |
||
| 1450 | word = word[:-6] |
||
| 1451 | elif _r1[-5:] in {'ethed', 'erede', 'heden', 'heder', 'endes', 'ernes', |
||
| 1452 | 'erens', 'erets'}: |
||
| 1453 | word = word[:-5] |
||
| 1454 | elif _r1[-4:] in {'ered', 'ende', 'erne', 'eren', 'erer', 'heds', 'enes', |
||
| 1455 | 'eres', 'eret'}: |
||
| 1456 | word = word[:-4] |
||
| 1457 | elif _r1[-3:] in {'hed', 'ene', 'ere', 'ens', 'ers', 'ets'}: |
||
| 1458 | word = word[:-3] |
||
| 1459 | elif _r1[-2:] in {'en', 'er', 'es', 'et'}: |
||
| 1460 | word = word[:-2] |
||
| 1461 | elif _r1[-1:] == 'e': |
||
| 1462 | word = word[:-1] |
||
| 1463 | elif _r1[-1:] == 's': |
||
| 1464 | if len(word) > 1 and word[-2] in _s_endings: |
||
| 1465 | word = word[:-1] |
||
| 1466 | |||
| 1467 | # Step 2 |
||
| 1468 | if word[r1_start:][-2:] in {'gd', 'dt', 'gt', 'kt'}: |
||
| 1469 | word = word[:-1] |
||
| 1470 | |||
| 1471 | # Step 3 |
||
| 1472 | if word[-4:] == 'igst': |
||
| 1473 | word = word[:-2] |
||
| 1474 | |||
| 1475 | _r1 = word[r1_start:] |
||
| 1476 | repeat_step2 = False |
||
| 1477 | if _r1[-4:] == 'elig': |
||
| 1478 | word = word[:-4] |
||
| 1479 | repeat_step2 = True |
||
| 1480 | elif _r1[-4:] == 'løst': |
||
| 1481 | word = word[:-1] |
||
| 1482 | elif _r1[-3:] in {'lig', 'els'}: |
||
| 1483 | word = word[:-3] |
||
| 1484 | repeat_step2 = True |
||
| 1485 | elif _r1[-2:] == 'ig': |
||
| 1486 | word = word[:-2] |
||
| 1487 | repeat_step2 = True |
||
| 1488 | |||
| 1489 | if repeat_step2: |
||
| 1490 | if word[r1_start:][-2:] in {'gd', 'dt', 'gt', 'kt'}: |
||
| 1491 | word = word[:-1] |
||
| 1492 | |||
| 1493 | # Step 4 |
||
| 1494 | if ((len(word[r1_start:]) >= 1 and len(word) >= 2 and |
||
| 1495 | word[-1] == word[-2] and word[-1] not in _vowels)): |
||
| 1496 | word = word[:-1] |
||
| 1497 | |||
| 1498 | return word |
||
| 1499 | |||
| 1500 | |||
| 1501 | def clef_german(word): |
||
| 1502 | """Return CLEF German stem. |
||
| 1503 | |||
| 1504 | The CLEF German stemmer is defined at: |
||
| 1505 | http://members.unine.ch/jacques.savoy/clef/germanStemmer.txt |
||
| 1506 | |||
| 1507 | :param word: the word to calculate the stem of |
||
| 1508 | :returns: word stem |
||
| 1509 | :rtype: str |
||
| 1510 | |||
| 1511 | >>> clef_german('lesen') |
||
| 1512 | 'lese' |
||
| 1513 | >>> clef_german('graues') |
||
| 1514 | 'grau' |
||
| 1515 | >>> clef_german('buchstabieren') |
||
| 1516 | 'buchstabier' |
||
| 1517 | """ |
||
| 1518 | # lowercase, normalize, and compose |
||
| 1519 | word = unicodedata.normalize('NFC', text_type(word.lower())) |
||
| 1520 | |||
| 1521 | # remove umlauts |
||
| 1522 | _umlauts = dict(zip((ord(_) for _ in 'äöü'), 'aou')) |
||
| 1523 | word = word.translate(_umlauts) |
||
| 1524 | |||
| 1525 | # remove plurals |
||
| 1526 | wlen = len(word)-1 |
||
| 1527 | |||
| 1528 | if wlen > 3: |
||
| 1529 | if wlen > 5: |
||
| 1530 | if word[-3:] == 'nen': |
||
| 1531 | return word[:-3] |
||
| 1532 | if wlen > 4: |
||
| 1533 | if word[-2:] in {'en', 'se', 'es', 'er'}: |
||
| 1534 | return word[:-2] |
||
| 1535 | if word[-1] in {'e', 'n', 'r', 's'}: |
||
| 1536 | return word[:-1] |
||
| 1537 | return word |
||
| 1538 | |||
| 1539 | |||
| 1540 | def clef_german_plus(word): |
||
| 1541 | """Return 'CLEF German stemmer plus' stem. |
||
| 1542 | |||
| 1543 | The CLEF German stemmer plus is defined at: |
||
| 1544 | http://members.unine.ch/jacques.savoy/clef/germanStemmerPlus.txt |
||
| 1545 | |||
| 1546 | :param word: the word to calculate the stem of |
||
| 1547 | :returns: word stem |
||
| 1548 | :rtype: str |
||
| 1549 | |||
| 1550 | >>> clef_german_plus('lesen') |
||
| 1551 | 'les' |
||
| 1552 | >>> clef_german_plus('graues') |
||
| 1553 | 'grau' |
||
| 1554 | >>> clef_german_plus('buchstabieren') |
||
| 1555 | 'buchstabi' |
||
| 1556 | """ |
||
| 1557 | _st_ending = {'b', 'd', 'f', 'g', 'h', 'k', 'l', 'm', 'n', 't'} |
||
| 1558 | |||
| 1559 | # lowercase, normalize, and compose |
||
| 1560 | word = unicodedata.normalize('NFC', text_type(word.lower())) |
||
| 1561 | |||
| 1562 | # remove umlauts |
||
| 1563 | _accents = dict(zip((ord(_) for _ in 'äàáâöòóôïìíîüùúû'), |
||
| 1564 | 'aaaaooooiiiiuuuu')) |
||
| 1565 | word = word.translate(_accents) |
||
| 1566 | |||
| 1567 | # Step 1 |
||
| 1568 | wlen = len(word)-1 |
||
| 1569 | if wlen > 4 and word[-3:] == 'ern': |
||
| 1570 | word = word[:-3] |
||
| 1571 | elif wlen > 3 and word[-2:] in {'em', 'en', 'er', 'es'}: |
||
| 1572 | word = word[:-2] |
||
| 1573 | elif wlen > 2 and (word[-1] == 'e' or |
||
| 1574 | (word[-1] == 's' and word[-2] in _st_ending)): |
||
| 1575 | word = word[:-1] |
||
| 1576 | |||
| 1577 | # Step 2 |
||
| 1578 | wlen = len(word)-1 |
||
| 1579 | if wlen > 4 and word[-3:] == 'est': |
||
| 1580 | word = word[:-3] |
||
| 1581 | elif wlen > 3 and (word[-2:] in {'er', 'en'} or |
||
| 1582 | (word[-2:] == 'st' and word[-3] in _st_ending)): |
||
| 1583 | word = word[:-2] |
||
| 1584 | |||
| 1585 | return word |
||
| 1586 | |||
| 1587 | |||
| 1588 | def clef_swedish(word): |
||
| 1589 | """Return CLEF Swedish stem. |
||
| 1590 | |||
| 1591 | The CLEF Swedish stemmer is defined at: |
||
| 1592 | http://members.unine.ch/jacques.savoy/clef/swedishStemmer.txt |
||
| 1593 | |||
| 1594 | :param word: the word to calculate the stem of |
||
| 1595 | :returns: word stem |
||
| 1596 | :rtype: str |
||
| 1597 | |||
| 1598 | >>> clef_swedish('undervisa') |
||
| 1599 | 'undervis' |
||
| 1600 | >>> clef_swedish('suspension') |
||
| 1601 | 'suspensio' |
||
| 1602 | >>> clef_swedish('visshet') |
||
| 1603 | 'viss' |
||
| 1604 | """ |
||
| 1605 | wlen = len(word)-1 |
||
| 1606 | |||
| 1607 | if wlen > 3 and word[-1] == 's': |
||
| 1608 | word = word[:-1] |
||
| 1609 | wlen -= 1 |
||
| 1610 | |||
| 1611 | if wlen > 6: |
||
| 1612 | if word[-5:] in {'elser', 'heten'}: |
||
| 1613 | return word[:-5] |
||
| 1614 | if wlen > 5: |
||
| 1615 | if word[-4:] in {'arne', 'erna', 'ande', 'else', 'aste', 'orna', |
||
| 1616 | 'aren'}: |
||
| 1617 | return word[:-4] |
||
| 1618 | if wlen > 4: |
||
| 1619 | if word[-3:] in {'are', 'ast', 'het'}: |
||
| 1620 | return word[:-3] |
||
| 1621 | if wlen > 3: |
||
| 1622 | if word[-2:] in {'ar', 'er', 'or', 'en', 'at', 'te', 'et'}: |
||
| 1623 | return word[:-2] |
||
| 1624 | if wlen > 2: |
||
| 1625 | if word[-1] in {'a', 'e', 'n', 't'}: |
||
| 1626 | return word[:-1] |
||
| 1627 | return word |
||
| 1628 | |||
| 1629 | |||
| 1630 | def caumanns(word): |
||
| 1631 | """Return Caumanns German stem. |
||
| 1632 | |||
| 1633 | Jörg Caumanns' stemmer is described in his article at: |
||
| 1634 | https://refubium.fu-berlin.de/bitstream/handle/fub188/18405/tr-b-99-16.pdf |
||
| 1635 | |||
| 1636 | This implementation is based on the GermanStemFilter described at: |
||
| 1637 | http://www.evelix.ch/unternehmen/Blog/evelix/2013/11/11/inner-workings-of-the-german-analyzer-in-lucene |
||
| 1638 | |||
| 1639 | :param word: the word to calculate the stem of |
||
| 1640 | :returns: word stem |
||
| 1641 | :rtype: str |
||
| 1642 | |||
| 1643 | >>> caumanns('lesen') |
||
| 1644 | 'les' |
||
| 1645 | >>> caumanns('graues') |
||
| 1646 | 'grau' |
||
| 1647 | >>> caumanns('buchstabieren') |
||
| 1648 | 'buchstabier' |
||
| 1649 | """ |
||
| 1650 | if not word: |
||
| 1651 | return '' |
||
| 1652 | |||
| 1653 | upper_initial = word[0].isupper() |
||
| 1654 | word = unicodedata.normalize('NFC', text_type(word.lower())) |
||
| 1655 | |||
| 1656 | # # Part 2: Substitution |
||
| 1657 | # 1. Change umlauts to corresponding vowels & ß to ss |
||
| 1658 | _umlauts = dict(zip((ord(_) for _ in 'äöü'), 'aou')) |
||
| 1659 | word = word.translate(_umlauts) |
||
| 1660 | word = word.replace('ß', 'ss') |
||
| 1661 | |||
| 1662 | # 2. Change second of doubled characters to * |
||
| 1663 | newword = word[0] |
||
| 1664 | for i in range(1, len(word)): |
||
| 1665 | if newword[i-1] == word[i]: |
||
| 1666 | newword += '*' |
||
| 1667 | else: |
||
| 1668 | newword += word[i] |
||
| 1669 | word = newword |
||
| 1670 | |||
| 1671 | # 3. Replace sch, ch, ei, ie with $, §, %, & |
||
| 1672 | word = word.replace('sch', '$') |
||
| 1673 | word = word.replace('ch', '§') |
||
| 1674 | word = word.replace('ei', '%') |
||
| 1675 | word = word.replace('ie', '&') |
||
| 1676 | word = word.replace('ig', '#') |
||
| 1677 | word = word.replace('st', '!') |
||
| 1678 | |||
| 1679 | # # Part 1: Recursive Context-Free Stripping |
||
| 1680 | # 1. Remove the following 7 suffixes recursively |
||
| 1681 | while len(word) > 3: |
||
| 1682 | if (((len(word) > 4 and word[-2:] in {'em', 'er'}) or |
||
| 1683 | (len(word) > 5 and word[-2:] == 'nd'))): |
||
| 1684 | word = word[:-2] |
||
| 1685 | elif ((word[-1] in {'e', 's', 'n'}) or |
||
| 1686 | (not upper_initial and word[-1] in {'t', '!'})): |
||
| 1687 | word = word[:-1] |
||
| 1688 | else: |
||
| 1689 | break |
||
| 1690 | |||
| 1691 | # Additional optimizations: |
||
| 1692 | if len(word) > 5 and word[-5:] == 'erin*': |
||
| 1693 | word = word[:-1] |
||
| 1694 | if word[-1] == 'z': |
||
| 1695 | word = word[:-1] + 'x' |
||
| 1696 | |||
| 1697 | # Reverse substitutions: |
||
| 1698 | word = word.replace('$', 'sch') |
||
| 1699 | word = word.replace('§', 'ch') |
||
| 1700 | word = word.replace('%', 'ei') |
||
| 1701 | word = word.replace('&', 'ie') |
||
| 1702 | word = word.replace('#', 'ig') |
||
| 1703 | word = word.replace('!', 'st') |
||
| 1704 | |||
| 1705 | # Expand doubled |
||
| 1706 | word = ''.join([word[0]] + [word[i-1] if word[i] == '*' else word[i] for |
||
| 1707 | i in range(1, len(word))]) |
||
| 1708 | |||
| 1709 | # Finally, convert gege to ge |
||
| 1710 | if len(word) > 4: |
||
| 1711 | word = word.replace('gege', 'ge', 1) |
||
| 1712 | |||
| 1713 | return word |
||
| 1714 | |||
| 1715 | |||
| 1716 | def uealite(word, max_word_length=20, max_acro_length=8, return_rule_no=False, |
||
| 1717 | var=None): |
||
| 1718 | """Return UEA-Lite stem. |
||
| 1719 | |||
| 1720 | The UEA-Lite stemmer is discussed in: |
||
| 1721 | Jenkins, Marie-Claire and Dan Smith. 2005. "Conservative stemming for |
||
| 1722 | search and indexing." |
||
| 1723 | http://lemur.cmp.uea.ac.uk/Research/stemmer/stemmer25feb.pdf |
||
| 1724 | |||
| 1725 | This is chiefly based on the Java implementation of the algorithm, with |
||
| 1726 | variants based on the Perl implementation and Jason Adams' Ruby port. |
||
| 1727 | |||
| 1728 | Java version: http://lemur.cmp.uea.ac.uk/Research/stemmer/UEAstem.java |
||
| 1729 | Perl version: http://lemur.cmp.uea.ac.uk/Research/stemmer/UEAstem.pl |
||
| 1730 | Ruby version: https://github.com/ealdent/uea-stemmer |
||
| 1731 | |||
| 1732 | :param word: the word to calculate the stem of |
||
| 1733 | :param max_word_length: the maximum word length allowed |
||
| 1734 | :param return_rule_no: if True, returns the stem along with rule number |
||
| 1735 | :param var: variant to use (set to 'Adams' to use Jason Adams' rules, |
||
| 1736 | or 'Perl' to use the original Perl set of rules) |
||
| 1737 | :returns: word stem |
||
| 1738 | :rtype: str or tuple(str, int) |
||
| 1739 | """ |
||
| 1740 | problem_words = {'is', 'as', 'this', 'has', 'was', 'during'} |
||
| 1741 | |||
| 1742 | # rule table format: |
||
| 1743 | # top-level dictionary: length-of-suffix: dict-of-rules |
||
| 1744 | # dict-of-rules: suffix: (rule_no, suffix_length_to_delete, |
||
| 1745 | # suffix_to_append) |
||
| 1746 | rule_table = {7: {'titudes': (30, 1, None), |
||
| 1747 | 'fulness': (34, 4, None), |
||
| 1748 | 'ousness': (35, 4, None), |
||
| 1749 | 'eadings': (40.7, 4, None), |
||
| 1750 | 'oadings': (40.6, 4, None), |
||
| 1751 | 'ealings': (42.4, 4, None), |
||
| 1752 | 'ailings': (42.2, 4, None), |
||
| 1753 | }, |
||
| 1754 | 6: {'aceous': (1, 6, None), |
||
| 1755 | 'aining': (24, 3, None), |
||
| 1756 | 'acting': (25, 3, None), |
||
| 1757 | 'ttings': (26, 5, None), |
||
| 1758 | 'viding': (27, 3, 'e'), |
||
| 1759 | 'ssings': (37, 4, None), |
||
| 1760 | 'ulting': (38, 3, None), |
||
| 1761 | 'eading': (40.7, 3, None), |
||
| 1762 | 'oading': (40.6, 3, None), |
||
| 1763 | 'edings': (40.5, 4, None), |
||
| 1764 | 'ddings': (40.4, 5, None), |
||
| 1765 | 'ldings': (40.3, 4, None), |
||
| 1766 | 'rdings': (40.2, 4, None), |
||
| 1767 | 'ndings': (40.1, 4, None), |
||
| 1768 | 'llings': (41, 5, None), |
||
| 1769 | 'ealing': (42.4, 3, None), |
||
| 1770 | 'olings': (42.3, 4, None), |
||
| 1771 | 'ailing': (42.2, 3, None), |
||
| 1772 | 'elings': (42.1, 4, None), |
||
| 1773 | 'mmings': (44.3, 5, None), |
||
| 1774 | 'ngings': (45.2, 4, None), |
||
| 1775 | 'ggings': (45.1, 5, None), |
||
| 1776 | 'stings': (47, 4, None), |
||
| 1777 | 'etings': (48.4, 4, None), |
||
| 1778 | 'ntings': (48.2, 4, None), |
||
| 1779 | 'irings': (54.4, 4, 'e'), |
||
| 1780 | 'urings': (54.3, 4, 'e'), |
||
| 1781 | 'ncings': (54.2, 4, 'e'), |
||
| 1782 | 'things': (58.1, 1, None), |
||
| 1783 | }, |
||
| 1784 | 5: {'iases': (11.4, 2, None), |
||
| 1785 | 'ained': (13.6, 2, None), |
||
| 1786 | 'erned': (13.5, 2, None), |
||
| 1787 | 'ifted': (14, 2, None), |
||
| 1788 | 'ected': (15, 2, None), |
||
| 1789 | 'vided': (16, 1, None), |
||
| 1790 | 'erred': (19, 3, None), |
||
| 1791 | 'urred': (20.5, 3, None), |
||
| 1792 | 'lored': (20.4, 2, None), |
||
| 1793 | 'eared': (20.3, 2, None), |
||
| 1794 | 'tored': (20.2, 1, None), |
||
| 1795 | 'noted': (22.4, 1, None), |
||
| 1796 | 'leted': (22.3, 1, None), |
||
| 1797 | 'anges': (23, 1, None), |
||
| 1798 | 'tting': (26, 4, None), |
||
| 1799 | 'ulted': (32, 2, None), |
||
| 1800 | 'uming': (33, 3, 'e'), |
||
| 1801 | 'rabed': (36.1, 1, None), |
||
| 1802 | 'rebed': (36.1, 1, None), |
||
| 1803 | 'ribed': (36.1, 1, None), |
||
| 1804 | 'robed': (36.1, 1, None), |
||
| 1805 | 'rubed': (36.1, 1, None), |
||
| 1806 | 'ssing': (37, 3, None), |
||
| 1807 | 'vings': (39, 4, 'e'), |
||
| 1808 | 'eding': (40.5, 3, None), |
||
| 1809 | 'dding': (40.4, 4, None), |
||
| 1810 | 'lding': (40.3, 3, None), |
||
| 1811 | 'rding': (40.2, 3, None), |
||
| 1812 | 'nding': (40.1, 3, None), |
||
| 1813 | 'dings': (40, 4, 'e'), |
||
| 1814 | 'lling': (41, 4, None), |
||
| 1815 | 'oling': (42.3, 3, None), |
||
| 1816 | 'eling': (42.1, 3, None), |
||
| 1817 | 'lings': (42, 4, 'e'), |
||
| 1818 | 'mming': (44.3, 4, None), |
||
| 1819 | 'rming': (44.2, 3, None), |
||
| 1820 | 'lming': (44.1, 3, None), |
||
| 1821 | 'mings': (44, 4, 'e'), |
||
| 1822 | 'nging': (45.2, 3, None), |
||
| 1823 | 'gging': (45.1, 4, None), |
||
| 1824 | 'gings': (45, 4, 'e'), |
||
| 1825 | 'aning': (46.6, 3, None), |
||
| 1826 | 'ening': (46.5, 3, None), |
||
| 1827 | 'gning': (46.4, 3, None), |
||
| 1828 | 'nning': (46.3, 4, None), |
||
| 1829 | 'oning': (46.2, 3, None), |
||
| 1830 | 'rning': (46.1, 3, None), |
||
| 1831 | 'sting': (47, 3, None), |
||
| 1832 | 'eting': (48.4, 3, None), |
||
| 1833 | 'pting': (48.3, 3, None), |
||
| 1834 | 'nting': (48.2, 3, None), |
||
| 1835 | 'cting': (48.1, 3, None), |
||
| 1836 | 'tings': (48, 4, 'e'), |
||
| 1837 | 'iring': (54.4, 3, 'e'), |
||
| 1838 | 'uring': (54.3, 3, 'e'), |
||
| 1839 | 'ncing': (54.2, 3, 'e'), |
||
| 1840 | 'sings': (54, 4, 'e'), |
||
| 1841 | # 'lling': (55, 3, None), # masked by 41 |
||
| 1842 | 'ating': (57, 3, 'e'), |
||
| 1843 | 'thing': (58.1, 0, None), |
||
| 1844 | }, |
||
| 1845 | 4: {'eeds': (7, 1, None), |
||
| 1846 | 'uses': (11.3, 1, None), |
||
| 1847 | 'sses': (11.2, 2, None), |
||
| 1848 | 'eses': (11.1, 2, 'is'), |
||
| 1849 | 'tled': (12.5, 1, None), |
||
| 1850 | 'pled': (12.4, 1, None), |
||
| 1851 | 'bled': (12.3, 1, None), |
||
| 1852 | 'eled': (12.2, 2, None), |
||
| 1853 | 'lled': (12.1, 2, None), |
||
| 1854 | 'ened': (13.7, 2, None), |
||
| 1855 | 'rned': (13.4, 2, None), |
||
| 1856 | 'nned': (13.3, 3, None), |
||
| 1857 | 'oned': (13.2, 2, None), |
||
| 1858 | 'gned': (13.1, 2, None), |
||
| 1859 | 'ered': (20.1, 2, None), |
||
| 1860 | 'reds': (20, 2, None), |
||
| 1861 | 'tted': (21, 3, None), |
||
| 1862 | 'uted': (22.2, 1, None), |
||
| 1863 | 'ated': (22.1, 1, None), |
||
| 1864 | 'ssed': (28, 2, None), |
||
| 1865 | 'umed': (31, 1, None), |
||
| 1866 | 'beds': (36, 3, None), |
||
| 1867 | 'ving': (39, 3, 'e'), |
||
| 1868 | 'ding': (40, 3, 'e'), |
||
| 1869 | 'ling': (42, 3, 'e'), |
||
| 1870 | 'nged': (43.2, 1, None), |
||
| 1871 | 'gged': (43.1, 3, None), |
||
| 1872 | 'ming': (44, 3, 'e'), |
||
| 1873 | 'ging': (45, 3, 'e'), |
||
| 1874 | 'ning': (46, 3, 'e'), |
||
| 1875 | 'ting': (48, 3, 'e'), |
||
| 1876 | # 'ssed': (49, 2, None), # masked by 28 |
||
| 1877 | # 'lled': (53, 2, None), # masked by 12.1 |
||
| 1878 | 'zing': (54.1, 3, 'e'), |
||
| 1879 | 'sing': (54, 3, 'e'), |
||
| 1880 | 'lves': (60.1, 3, 'f'), |
||
| 1881 | 'aped': (61.3, 1, None), |
||
| 1882 | 'uded': (61.2, 1, None), |
||
| 1883 | 'oded': (61.1, 1, None), |
||
| 1884 | # 'ated': (61, 1, None), # masked by 22.1 |
||
| 1885 | 'ones': (63.6, 1, None), |
||
| 1886 | 'izes': (63.5, 1, None), |
||
| 1887 | 'ures': (63.4, 1, None), |
||
| 1888 | 'ines': (63.3, 1, None), |
||
| 1889 | 'ides': (63.2, 1, None), |
||
| 1890 | }, |
||
| 1891 | 3: {'ces': (2, 1, None), |
||
| 1892 | 'sis': (4, 0, None), |
||
| 1893 | 'tis': (5, 0, None), |
||
| 1894 | 'eed': (7, 0, None), |
||
| 1895 | 'ued': (8, 1, None), |
||
| 1896 | 'ues': (9, 1, None), |
||
| 1897 | 'ees': (10, 1, None), |
||
| 1898 | 'ses': (11, 1, None), |
||
| 1899 | 'led': (12, 2, None), |
||
| 1900 | 'ned': (13, 1, None), |
||
| 1901 | 'ved': (17, 1, None), |
||
| 1902 | 'ced': (18, 1, None), |
||
| 1903 | 'red': (20, 1, None), |
||
| 1904 | 'ted': (22, 2, None), |
||
| 1905 | 'sed': (29, 1, None), |
||
| 1906 | 'bed': (36, 2, None), |
||
| 1907 | 'ged': (43, 1, None), |
||
| 1908 | 'les': (50, 1, None), |
||
| 1909 | 'tes': (51, 1, None), |
||
| 1910 | 'zed': (52, 1, None), |
||
| 1911 | 'ied': (56, 3, 'y'), |
||
| 1912 | 'ies': (59, 3, 'y'), |
||
| 1913 | 'ves': (60, 1, None), |
||
| 1914 | 'pes': (63.8, 1, None), |
||
| 1915 | 'mes': (63.7, 1, None), |
||
| 1916 | 'ges': (63.1, 1, None), |
||
| 1917 | 'ous': (65, 0, None), |
||
| 1918 | 'ums': (66, 0, None), |
||
| 1919 | }, |
||
| 1920 | 2: {'cs': (3, 0, None), |
||
| 1921 | 'ss': (6, 0, None), |
||
| 1922 | 'es': (63, 2, None), |
||
| 1923 | 'is': (64, 2, 'e'), |
||
| 1924 | 'us': (67, 0, None), |
||
| 1925 | }} |
||
| 1926 | |||
| 1927 | if var == 'Perl': |
||
| 1928 | perl_deletions = {7: ['eadings', 'oadings', 'ealings', 'ailings'], |
||
| 1929 | 6: ['ttings', 'ssings', 'edings', 'ddings', |
||
| 1930 | 'ldings', 'rdings', 'ndings', 'llings', |
||
| 1931 | 'olings', 'elings', 'mmings', 'ngings', |
||
| 1932 | 'ggings', 'stings', 'etings', 'ntings', |
||
| 1933 | 'irings', 'urings', 'ncings', 'things'], |
||
| 1934 | 5: ['vings', 'dings', 'lings', 'mings', 'gings', |
||
| 1935 | 'tings', 'sings'], |
||
| 1936 | 4: ['eeds', 'reds', 'beds']} |
||
| 1937 | |||
| 1938 | # Delete the above rules from rule_table |
||
| 1939 | for del_len in perl_deletions: |
||
| 1940 | for term in perl_deletions[del_len]: |
||
| 1941 | del rule_table[del_len][term] |
||
| 1942 | |||
| 1943 | elif var == 'Adams': |
||
| 1944 | adams_additions = {6: {'chited': (22.8, 1, None)}, |
||
| 1945 | 5: {'dying': (58.2, 4, 'ie'), |
||
| 1946 | 'tying': (58.2, 4, 'ie'), |
||
| 1947 | 'vited': (22.6, 1, None), |
||
| 1948 | 'mited': (22.5, 1, None), |
||
| 1949 | 'vided': (22.9, 1, None), |
||
| 1950 | 'mided': (22.10, 1, None), |
||
| 1951 | 'lying': (58.2, 4, 'ie'), |
||
| 1952 | 'arred': (19.1, 3, None), |
||
| 1953 | }, |
||
| 1954 | 4: {'ited': (22.7, 2, None), |
||
| 1955 | 'oked': (31.1, 1, None), |
||
| 1956 | 'aked': (31.1, 1, None), |
||
| 1957 | 'iked': (31.1, 1, None), |
||
| 1958 | 'uked': (31.1, 1, None), |
||
| 1959 | 'amed': (31, 1, None), |
||
| 1960 | 'imed': (31, 1, None), |
||
| 1961 | 'does': (31.2, 2, None), |
||
| 1962 | }, |
||
| 1963 | 3: {'oed': (31.3, 1, None), |
||
| 1964 | 'oes': (31.2, 1, None), |
||
| 1965 | 'kes': (63.1, 1, None), |
||
| 1966 | 'des': (63.10, 1, None), |
||
| 1967 | 'res': (63.9, 1, None), |
||
| 1968 | }} |
||
| 1969 | |||
| 1970 | # Add the above additional rules to rule_table |
||
| 1971 | for del_len in adams_additions: |
||
| 1972 | rule_table[del_len] = dict(rule_table[del_len], |
||
| 1973 | **adams_additions[del_len]) |
||
| 1974 | # Add additional problem word |
||
| 1975 | problem_words.add('menses') |
||
| 1976 | |||
| 1977 | def _stem_with_duplicate_character_check(word, del_len): |
||
| 1978 | if word[-1] == 's': |
||
| 1979 | del_len += 1 |
||
| 1980 | stemmed_word = word[:-del_len] |
||
| 1981 | if re.match(r'.*(\w)\1$', stemmed_word): |
||
| 1982 | stemmed_word = stemmed_word[:-1] |
||
| 1983 | return stemmed_word |
||
| 1984 | |||
| 1985 | def _stem(word): |
||
| 1986 | stemmed_word = word |
||
| 1987 | rule_no = 0 |
||
| 1988 | |||
| 1989 | if not word: |
||
| 1990 | return word, 0 |
||
| 1991 | if word in problem_words: |
||
| 1992 | return word, 90 |
||
| 1993 | if max_word_length and len(word) > max_word_length: |
||
| 1994 | return word, 95 |
||
| 1995 | |||
| 1996 | if "'" in word: |
||
| 1997 | if word[-2:] in {"'s", "'S"}: |
||
| 1998 | stemmed_word = word[:-2] |
||
| 1999 | if word[-1:] == "'": |
||
| 2000 | stemmed_word = word[:-1] |
||
| 2001 | stemmed_word = stemmed_word.replace("n't", 'not') |
||
| 2002 | stemmed_word = stemmed_word.replace("'ve", 'have') |
||
| 2003 | stemmed_word = stemmed_word.replace("'re", 'are') |
||
| 2004 | stemmed_word = stemmed_word.replace("'m", 'am') |
||
| 2005 | return stemmed_word, 94 |
||
| 2006 | |||
| 2007 | if word.isdigit(): |
||
| 2008 | return word, 90.3 |
||
| 2009 | else: |
||
| 2010 | hyphen = word.find('-') |
||
| 2011 | if hyphen > 0 and hyphen < len(word): |
||
| 2012 | if word[:hyphen].isalpha() and word[hyphen+1:].isalpha(): |
||
| 2013 | return word, 90.2 |
||
| 2014 | else: |
||
| 2015 | return word, 90.1 |
||
| 2016 | elif '_' in word: |
||
| 2017 | return word, 90 |
||
| 2018 | elif word[-1] == 's' and word[:-1].isupper(): |
||
| 2019 | if var == 'Adams' and len(word)-1 > max_acro_length: |
||
| 2020 | return word, 96 |
||
| 2021 | return word[:-1], 91.1 |
||
| 2022 | elif word.isupper(): |
||
| 2023 | if var == 'Adams' and len(word) > max_acro_length: |
||
| 2024 | return word, 96 |
||
| 2025 | return word, 91 |
||
| 2026 | elif re.match(r'^.*[A-Z].*[A-Z].*$', word): |
||
| 2027 | return word, 92 |
||
| 2028 | elif word[0].isupper(): |
||
| 2029 | return word, 93 |
||
| 2030 | elif var == 'Adams' and re.match(r'^[a-z]{1}(|[rl])(ing|ed)$', |
||
| 2031 | word): |
||
| 2032 | return word, 97 |
||
| 2033 | |||
| 2034 | for n in range(7, 1, -1): |
||
| 2035 | if word[-n:] in rule_table[n]: |
||
| 2036 | rule_no, del_len, add_str = rule_table[n][word[-n:]] |
||
| 2037 | if del_len: |
||
| 2038 | stemmed_word = word[:-del_len] |
||
| 2039 | else: |
||
| 2040 | stemmed_word = word |
||
| 2041 | if add_str: |
||
| 2042 | stemmed_word += add_str |
||
| 2043 | break |
||
| 2044 | |||
| 2045 | if not rule_no: |
||
| 2046 | if re.match(r'.*\w\wings?$', word): # rule 58 |
||
| 2047 | stemmed_word = _stem_with_duplicate_character_check(word, 3) |
||
| 2048 | rule_no = 58 |
||
| 2049 | elif re.match(r'.*\w\weds?$', word): # rule 62 |
||
| 2050 | stemmed_word = _stem_with_duplicate_character_check(word, 2) |
||
| 2051 | rule_no = 62 |
||
| 2052 | elif word[-1] == 's': # rule 68 |
||
| 2053 | stemmed_word = word[:-1] |
||
| 2054 | rule_no = 68 |
||
| 2055 | |||
| 2056 | return stemmed_word, rule_no |
||
| 2057 | |||
| 2058 | stem, rule_no = _stem(word) |
||
| 2059 | if return_rule_no: |
||
| 2060 | return stem, rule_no |
||
| 2061 | return stem |
||
| 2062 | |||
| 2063 | |||
| 2064 | def paice_husk(word): |
||
| 2065 | """Return Paice-Husk stem. |
||
| 2066 | |||
| 2067 | Implementation of the Paice-Husk Stemmer, also known as the Lancaster |
||
| 2068 | Stemmer, developed by Chris Paice, with the assistance of Gareth Husk |
||
| 2069 | |||
| 2070 | This is based on the algorithm's description in: |
||
| 2071 | Paice, Chris D. 1990. "Another stemmer." ACM SIGIR Forum 24:3, Fall 1990. |
||
| 2072 | 56-61. doi:10.1145/101306.101310. |
||
| 2073 | |||
| 2074 | :param word: the word to calculate the stem of |
||
| 2075 | :returns: word stem |
||
| 2076 | :rtype: str |
||
| 2077 | """ |
||
| 2078 | rule_table = {6: {'ifiabl': (False, 6, None, True), |
||
| 2079 | 'plicat': (False, 4, 'y', True)}, |
||
| 2080 | 5: {'guish': (False, 5, 'ct', True), |
||
| 2081 | 'sumpt': (False, 2, None, True), |
||
| 2082 | 'istry': (False, 5, None, True)}, |
||
| 2083 | 4: {'ytic': (False, 3, 's', True), |
||
| 2084 | 'ceed': (False, 2, 'ss', True), |
||
| 2085 | 'hood': (False, 4, None, False), |
||
| 2086 | 'lief': (False, 1, 'v', True), |
||
| 2087 | 'verj': (False, 1, 't', True), |
||
| 2088 | 'misj': (False, 2, 't', True), |
||
| 2089 | 'iabl': (False, 4, 'y', True), |
||
| 2090 | 'iful': (False, 4, 'y', True), |
||
| 2091 | 'sion': (False, 4, 'j', False), |
||
| 2092 | 'xion': (False, 4, 'ct', True), |
||
| 2093 | 'ship': (False, 4, None, False), |
||
| 2094 | 'ness': (False, 4, None, False), |
||
| 2095 | 'ment': (False, 4, None, False), |
||
| 2096 | 'ript': (False, 2, 'b', True), |
||
| 2097 | 'orpt': (False, 2, 'b', True), |
||
| 2098 | 'duct': (False, 1, None, True), |
||
| 2099 | 'cept': (False, 2, 'iv', True), |
||
| 2100 | 'olut': (False, 2, 'v', True), |
||
| 2101 | 'sist': (False, 0, None, True)}, |
||
| 2102 | 3: {'ied': (False, 3, 'y', False), |
||
| 2103 | 'eed': (False, 1, None, True), |
||
| 2104 | 'ing': (False, 3, None, False), |
||
| 2105 | 'iag': (False, 3, 'y', True), |
||
| 2106 | 'ish': (False, 3, None, False), |
||
| 2107 | 'fuj': (False, 1, 's', True), |
||
| 2108 | 'hej': (False, 1, 'r', True), |
||
| 2109 | 'abl': (False, 3, None, False), |
||
| 2110 | 'ibl': (False, 3, None, True), |
||
| 2111 | 'bil': (False, 2, 'l', False), |
||
| 2112 | 'ful': (False, 3, None, False), |
||
| 2113 | 'ial': (False, 3, None, False), |
||
| 2114 | 'ual': (False, 3, None, False), |
||
| 2115 | 'ium': (False, 3, None, True), |
||
| 2116 | 'ism': (False, 3, None, False), |
||
| 2117 | 'ion': (False, 3, None, False), |
||
| 2118 | 'ian': (False, 3, None, False), |
||
| 2119 | 'een': (False, 0, None, True), |
||
| 2120 | 'ear': (False, 0, None, True), |
||
| 2121 | 'ier': (False, 3, 'y', False), |
||
| 2122 | 'ies': (False, 3, 'y', False), |
||
| 2123 | 'sis': (False, 2, None, True), |
||
| 2124 | 'ous': (False, 3, None, False), |
||
| 2125 | 'ent': (False, 3, None, False), |
||
| 2126 | 'ant': (False, 3, None, False), |
||
| 2127 | 'ist': (False, 3, None, False), |
||
| 2128 | 'iqu': (False, 3, None, True), |
||
| 2129 | 'ogu': (False, 1, None, True), |
||
| 2130 | 'siv': (False, 3, 'j', False), |
||
| 2131 | 'eiv': (False, 0, None, True), |
||
| 2132 | 'bly': (False, 1, None, False), |
||
| 2133 | 'ily': (False, 3, 'y', False), |
||
| 2134 | 'ply': (False, 0, None, True), |
||
| 2135 | 'ogy': (False, 1, None, True), |
||
| 2136 | 'phy': (False, 1, None, True), |
||
| 2137 | 'omy': (False, 1, None, True), |
||
| 2138 | 'opy': (False, 1, None, True), |
||
| 2139 | 'ity': (False, 3, None, False), |
||
| 2140 | 'ety': (False, 3, None, False), |
||
| 2141 | 'lty': (False, 2, None, True), |
||
| 2142 | 'ary': (False, 3, None, False), |
||
| 2143 | 'ory': (False, 3, None, False), |
||
| 2144 | 'ify': (False, 3, None, True), |
||
| 2145 | 'ncy': (False, 2, 't', False), |
||
| 2146 | 'acy': (False, 3, None, False)}, |
||
| 2147 | 2: {'ia': (True, 2, None, True), |
||
| 2148 | 'bb': (False, 1, None, True), |
||
| 2149 | 'ic': (False, 2, None, False), |
||
| 2150 | 'nc': (False, 1, 't', False), |
||
| 2151 | 'dd': (False, 1, None, True), |
||
| 2152 | 'ed': (False, 2, None, False), |
||
| 2153 | 'if': (False, 2, None, False), |
||
| 2154 | 'ag': (False, 2, None, False), |
||
| 2155 | 'gg': (False, 1, None, True), |
||
| 2156 | 'th': (True, 2, None, True), |
||
| 2157 | 'ij': (False, 1, 'd', True), |
||
| 2158 | 'uj': (False, 1, 'd', True), |
||
| 2159 | 'oj': (False, 1, 'd', True), |
||
| 2160 | 'nj': (False, 1, 'd', True), |
||
| 2161 | 'cl': (False, 1, None, True), |
||
| 2162 | 'ul': (False, 2, None, True), |
||
| 2163 | 'al': (False, 2, None, False), |
||
| 2164 | 'll': (False, 1, None, True), |
||
| 2165 | 'um': (True, 2, None, True), |
||
| 2166 | 'mm': (False, 1, None, True), |
||
| 2167 | 'an': (False, 2, None, False), |
||
| 2168 | 'en': (False, 2, None, False), |
||
| 2169 | 'nn': (False, 1, None, True), |
||
| 2170 | 'pp': (False, 1, None, True), |
||
| 2171 | 'er': (False, 2, None, False), |
||
| 2172 | 'ar': (False, 2, None, True), |
||
| 2173 | 'or': (False, 2, None, False), |
||
| 2174 | 'ur': (False, 2, None, False), |
||
| 2175 | 'rr': (False, 1, None, True), |
||
| 2176 | 'tr': (False, 1, None, False), |
||
| 2177 | 'is': (False, 2, None, False), |
||
| 2178 | 'ss': (False, 0, None, True), |
||
| 2179 | 'us': (True, 2, None, True), |
||
| 2180 | 'at': (False, 2, None, False), |
||
| 2181 | 'tt': (False, 1, None, True), |
||
| 2182 | 'iv': (False, 2, None, False), |
||
| 2183 | 'ly': (False, 2, None, False), |
||
| 2184 | 'iz': (False, 2, None, False), |
||
| 2185 | 'yz': (False, 1, 's', True)}, |
||
| 2186 | 1: {'a': (True, 1, None, True), |
||
| 2187 | 'e': (False, 1, None, False), |
||
| 2188 | 'i': ((True, 1, None, True), (False, 1, 'y', False)), |
||
| 2189 | 'j': (False, 1, 's', True), |
||
| 2190 | 's': ((True, 1, None, False), (False, 0, None, True))}} |
||
| 2191 | |||
| 2192 | def _has_vowel(word): |
||
| 2193 | for char in word: |
||
| 2194 | if char in {'a', 'e', 'i', 'o', 'u', 'y'}: |
||
| 2195 | return True |
||
| 2196 | return False |
||
| 2197 | |||
| 2198 | def _acceptable(word): |
||
| 2199 | if word and word[0] in {'a', 'e', 'i', 'o', 'u'}: |
||
| 2200 | return len(word) > 1 |
||
| 2201 | return len(word) > 2 and _has_vowel(word[1:]) |
||
| 2202 | |||
| 2203 | def _apply_rule(word, rule, intact): |
||
| 2204 | old_word = word |
||
| 2205 | only_intact, del_len, add_str, set_terminate = rule |
||
| 2206 | # print(word, word[-n:], rule) |
||
| 2207 | |||
| 2208 | if (not only_intact) or (intact and only_intact): |
||
| 2209 | if del_len: |
||
| 2210 | word = word[:-del_len] |
||
| 2211 | if add_str: |
||
| 2212 | word += add_str |
||
| 2213 | else: |
||
| 2214 | return word, False, intact, terminate |
||
| 2215 | |||
| 2216 | if _acceptable(word): |
||
| 2217 | return word, True, False, set_terminate |
||
| 2218 | else: |
||
| 2219 | return old_word, False, intact, terminate |
||
| 2220 | |||
| 2221 | terminate = False |
||
| 2222 | intact = True |
||
| 2223 | while not terminate: |
||
| 2224 | for n in range(6, 0, -1): |
||
| 2225 | if word[-n:] in rule_table[n]: |
||
| 2226 | accept = False |
||
| 2227 | if len(rule_table[n][word[-n:]]) < 4: |
||
| 2228 | for rule in rule_table[n][word[-n:]]: |
||
| 2229 | (word, accept, intact, |
||
| 2230 | terminate) = _apply_rule(word, rule, intact) |
||
| 2231 | if accept: |
||
| 2232 | break |
||
| 2233 | else: |
||
| 2234 | rule = rule_table[n][word[-n:]] |
||
| 2235 | (word, accept, intact, |
||
| 2236 | terminate) = _apply_rule(word, rule, intact) |
||
| 2237 | |||
| 2238 | if accept: |
||
| 2239 | break |
||
| 2240 | else: |
||
| 2241 | break |
||
| 2242 | |||
| 2243 | return word |
||
| 2244 | |||
| 2245 | |||
| 2246 | def schinke(word): |
||
| 2247 | """Return the stem of a word according to the Schinke stemmer. |
||
| 2248 | |||
| 2249 | Source: |
||
| 2250 | Schinke, Robyn, Mark Greengrass, Alexander M. Robertson, and Peter Willett. |
||
| 2251 | 1996. "A Stemming Algorithm for Latin Text Databases." Journal of |
||
| 2252 | Documentation, 52(2). 172--187. |
||
| 2253 | doi:10.1108/eb026966 |
||
| 2254 | |||
| 2255 | :param word: |
||
| 2256 | :return: |
||
| 2257 | """ |
||
| 2258 | word = unicodedata.normalize('NFKD', text_type(word.lower())) |
||
| 2259 | word = ''.join(c for c in word if c in |
||
| 2260 | {'a', 'b', 'c', 'd', 'e', 'f', 'g', 'h', 'i', 'j', 'k', 'l', |
||
| 2261 | 'm', 'n', 'o', 'p', 'q', 'r', 's', 't', 'u', 'v', 'w', 'x', |
||
| 2262 | 'y', 'z'}) |
||
| 2263 | |||
| 2264 | # Rule 2 |
||
| 2265 | word = word.replace('j', 'i').replace('v', 'u') |
||
| 2266 | |||
| 2267 | # Rule 3 |
||
| 2268 | keep_que = {'at', 'quo', 'ne', 'ita', 'abs', 'aps', 'abus', 'adae', 'adus', |
||
| 2269 | 'deni', 'de', 'sus', 'obli', 'perae', 'plenis', 'quando', |
||
| 2270 | 'quis', 'quae', 'cuius', 'cui', 'quem', 'quam', 'qua', 'qui', |
||
| 2271 | 'quorum', 'quarum', 'quibus', 'quos', 'quas', 'quotusquis', |
||
| 2272 | 'quous', 'ubi', 'undi', 'us', 'uter', 'uti', 'utro', 'utribi', |
||
| 2273 | 'tor', 'co', 'conco', 'contor', 'detor', 'deco', 'exco', |
||
| 2274 | 'extor', 'obtor', 'optor', 'retor', 'reco', 'attor', 'inco', |
||
| 2275 | 'intor', 'praetor'} |
||
| 2276 | if word[-3:] == 'que': |
||
| 2277 | # This diverges from the paper by also returning 'que' itself unstemmed |
||
| 2278 | if word[:-3] in keep_que or word == 'que': |
||
| 2279 | return {'n': word, 'v': word} |
||
| 2280 | else: |
||
| 2281 | word = word[:-3] |
||
| 2282 | |||
| 2283 | # Base case will mean returning the words as is |
||
| 2284 | noun = word |
||
| 2285 | verb = word |
||
| 2286 | |||
| 2287 | # Rule 4 |
||
| 2288 | n_endings = {4: {'ibus'}, |
||
| 2289 | 3: {'ius'}, |
||
| 2290 | 2: {'is', 'nt', 'ae', 'os', 'am', 'ud', 'as', 'um', 'em', 'us', |
||
| 2291 | 'es', 'ia'}, |
||
| 2292 | 1: {'a', 'e', 'i', 'o', 'u'}} |
||
| 2293 | for endlen in range(4, 0, -1): |
||
| 2294 | if word[-endlen:] in n_endings[endlen]: |
||
| 2295 | if len(word)-2 >= endlen: |
||
| 2296 | noun = word[:-endlen] |
||
| 2297 | else: |
||
| 2298 | noun = word |
||
| 2299 | break |
||
| 2300 | |||
| 2301 | v_endings_strip = {6: {}, |
||
| 2302 | 5: {}, |
||
| 2303 | 4: {'mini', 'ntur', 'stis'}, |
||
| 2304 | 3: {'mur', 'mus', 'ris', 'sti', 'tis', 'tur'}, |
||
| 2305 | 2: {'ns', 'nt', 'ri'}, |
||
| 2306 | 1: {'m', 'r', 's', 't'}} |
||
| 2307 | v_endings_alter = {6: {'iuntur'}, |
||
| 2308 | 5: {'beris', 'erunt', 'untur'}, |
||
| 2309 | 4: {'iunt'}, |
||
| 2310 | 3: {'bor', 'ero', 'unt'}, |
||
| 2311 | 2: {'bo'}, |
||
| 2312 | 1: {}} |
||
| 2313 | for endlen in range(6, 0, -1): |
||
| 2314 | if word[-endlen:] in v_endings_strip[endlen]: |
||
| 2315 | addlen = 0 |
||
| 2316 | if len(word)-2 >= endlen: |
||
| 2317 | verb = word[:-endlen] |
||
| 2318 | else: |
||
| 2319 | verb = word |
||
| 2320 | break |
||
| 2321 | if word[-endlen:] in v_endings_alter[endlen]: |
||
| 2322 | if word[-endlen:] in {'iuntur', 'erunt', 'untur', 'iunt', 'unt'}: |
||
| 2323 | new_word = word[:-endlen]+'i' |
||
| 2324 | addlen = 1 |
||
| 2325 | elif word[-endlen:] in {'beris', 'bor', 'bo'}: |
||
| 2326 | new_word = word[:-endlen]+'bi' |
||
| 2327 | addlen = 2 |
||
| 2328 | else: |
||
| 2329 | new_word = word[:-endlen]+'eri' |
||
| 2330 | addlen = 3 |
||
| 2331 | |||
| 2332 | # Technically this diverges from the paper by considering the |
||
| 2333 | # length of the stem without the new suffix |
||
| 2334 | if len(new_word) >= 2+addlen: |
||
| 2335 | verb = new_word |
||
| 2336 | else: |
||
| 2337 | verb = word |
||
| 2338 | break |
||
| 2339 | |||
| 2340 | return {'n': noun, 'v': verb} |
||
| 2341 | |||
| 2342 | |||
| 2343 | if __name__ == '__main__': |
||
| 2344 | import doctest |
||
| 2345 | doctest.testmod() |
||
| 2346 |
Loading history...