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)
  32                for e2 in edits1(e1) if e2 in words)
  33
  34
  35 def known(word, words):
  36     return set(w for w in word if w in words)
  37
  38
  39 def suggest(word, words):
  40     candidates = (known([word], words) or
  41                   known(edits1(word), words) or
  42                   known_edits2(word, words) or [word])
  43     return candidates
  44
  45
  46 def correct(word, words):
  47     candidates = suggest(word, words)
  48     return max(candidates, key=words.get)
  49
  50
  51 class NorvigSpellCheck(object):
  52
  53     def __init__(self, words='/usr/share/dict/words',
  54                  cracklib='/usr/share/dict/cracklib-small',
  55                  propernames='/usr/share/dict/propernames'):
  56         self.dictwords = words
  57         self.cracklib = cracklib
  58         self.propernames = propernames
  59         self.words = collections.defaultdict(lambda: 1)
  60         self.extra_words = set()
  61         self.dictionary = None
  62         self.initialized = False
  63
  64     def set_dictionary(self, dictionary):
  65         self.dictionary = dictionary
  66
  67     def init(self):
  68         if self.initialized:
  69             return
  70         self.initialized = True
  71         train(self.read(), self.words)
  72         train(self.extra_words, self.words)
  73
  74     def add_word(self, word):
  75         self.extra_words.add(word)
  76
  77     def suggest(self, word):
  78         self.init()
  79         return suggest(word, self.words)
  80
  81     def check(self, word):
  82         self.init()
  83         return word.replace('.', '') in self.words
  84
  85     def read(self):
  86         """Read dictionary words"""
  87         paths = []
  88
  89         words = self.dictwords
  90         cracklib = self.cracklib
  91         propernames = self.propernames
  92         cfg_dictionary = self.dictionary
  93
  94         if cracklib and os.path.exists(cracklib):
  95             paths.append((cracklib, True))
  96         elif words and os.path.exists(words):
  97             paths.append((words, True))
  98
  99         if propernames and os.path.exists(propernames):
 100             paths.append((propernames, False))
 101
 102         if cfg_dictionary and os.path.exists(cfg_dictionary):
 103             paths.append((cfg_dictionary, False))
 104
 105         for (path, title) in paths:
 106             try:
 107                 with open(path, 'r') as f:
 108                     for word in f:
 109                         word = core.decode(word.rstrip())
 110                         yield word
 111                         if title:
 112                             yield word.title()
 113             except IOError:
 114                 pass