cola/spellcheck.py

   1 from __future__ import division, absolute_import, unicode_literals
   2 import collections
   3 import os
   4
   5 from cola import core
   6
   7 __copyright__ = """
   8 2012 Peter Norvig (http://norvig.com/spell-correct.html)
   9 2013-2018 David Aguilar <davvid@gmail.com>
  10 """
  11
  12 alphabet = 'abcdefghijklmnopqrstuvwxyz'
  13
  14
  15 def train(features, model):
  16     for f in features:
  17         model[f] += 1
  18     return model
  19
  20
  21 def edits1(word):
  22     splits = [(word[:i], word[i:]) for i in range(len(word) + 1)]
  23     deletes = [a + b[1:] for a, b in splits if b]
  24     transposes = [a + b[1] + b[0] + b[2:] for a, b in splits if len(b) > 1]
  25     replaces = [a + c + b[1:] for a, b in splits for c in alphabet if b]
  26     inserts = [a + c + b for a, b in splits for c in alphabet]
  27     return set(deletes + transposes + replaces + inserts)
  28
  29
  30 def known_edits2(word, words):
  31     return set(e2 for e1 in edits1(word) for e2 in edits1(e1) if e2 in words)
  32
  33
  34 def known(word, words):
  35     return set(w for w in word if w in words)
  36
  37
  38 def suggest(word, words):
  39     candidates = (
  40         known([word], words)
  41         or known(edits1(word), words)
  42         or known_edits2(word, words)
  43         or [word]
  44     )
  45     return candidates
  46
  47
  48 def correct(word, words):
  49     candidates = suggest(word, words)
  50     return max(candidates, key=words.get)
  51
  52
  53 class NorvigSpellCheck(object):
  54     def __init__(
  55         self,
  56         words='/usr/share/dict/words',
  57         cracklib='/usr/share/dict/cracklib-small',
  58         propernames='/usr/share/dict/propernames',
  59     ):
  60         self.dictwords = words
  61         self.cracklib = cracklib
  62         self.propernames = propernames
  63         self.words = collections.defaultdict(lambda: 1)
  64         self.extra_words = set()
  65         self.dictionary = None
  66         self.initialized = False
  67
  68     def set_dictionary(self, dictionary):
  69         self.dictionary = dictionary
  70
  71     def init(self):
  72         if self.initialized:
  73             return
  74         self.initialized = True
  75         train(self.read(), self.words)
  76         train(self.extra_words, self.words)
  77
  78     def add_word(self, word):
  79         self.extra_words.add(word)
  80
  81     def suggest(self, word):
  82         self.init()
  83         return suggest(word, self.words)
  84
  85     def check(self, word):
  86         self.init()
  87         return word.replace('.', '') in self.words
  88
  89     def read(self):
  90         """Read dictionary words"""
  91         paths = []
  92
  93         words = self.dictwords
  94         cracklib = self.cracklib
  95         propernames = self.propernames
  96         cfg_dictionary = self.dictionary
  97
  98         if cracklib and os.path.exists(cracklib):
  99             paths.append((cracklib, True))
 100         elif words and os.path.exists(words):
 101             paths.append((words, True))
 102
 103         if propernames and os.path.exists(propernames):
 104             paths.append((propernames, False))
 105
 106         if cfg_dictionary and os.path.exists(cfg_dictionary):
 107             paths.append((cfg_dictionary, False))
 108
 109         for (path, title) in paths:
 110             try:
 111                 with open(path, 'r') as f:
 112                     for word in f:
 113                         word = core.decode(word.rstrip())
 114                         yield word
 115                         if title:
 116                             yield word.title()
 117             except IOError:
 118                 pass