cola/spellcheck.py

   1 import codecs
   2 import collections
   3 import os
   4
   5 from . import resources
   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 {e2 for e1 in edits1(word) for e2 in edits1(e1) if e2 in words}
  32
  33
  34 def known(word, words):
  35     return {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:
  54     def __init__(
  55         self,
  56         words='dict/words',
  57         propernames='dict/propernames',
  58     ):
  59         data_dirs = resources.xdg_data_dirs()
  60         self.dictwords = resources.find_first(words, data_dirs)
  61         self.propernames = resources.find_first(propernames, data_dirs)
  62         self.words = collections.defaultdict(lambda: 1)
  63         self.extra_words = set()
  64         self.dictionary = None
  65         self.initialized = False
  66
  67     def set_dictionary(self, dictionary):
  68         self.dictionary = dictionary
  69
  70     def init(self):
  71         if self.initialized:
  72             return
  73         self.initialized = True
  74         train(self.read(), self.words)
  75         train(self.extra_words, self.words)
  76
  77     def add_word(self, word):
  78         self.extra_words.add(word)
  79
  80     def suggest(self, word):
  81         self.init()
  82         return suggest(word, self.words)
  83
  84     def check(self, word):
  85         self.init()
  86         return word.replace('.', '') in self.words
  87
  88     def read(self):
  89         """Read dictionary words"""
  90         paths = []
  91
  92         words = self.dictwords
  93         propernames = self.propernames
  94         cfg_dictionary = self.dictionary
  95
  96         if 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 codecs.open(
 108                     path, 'r', encoding='utf-8', errors='ignore'
 109                 ) as words_file:
 110                     for line in words_file:
 111                         word = line.rstrip()
 112                         yield word
 113                         if title:
 114                             yield word.title()
 115             except OSError:
 116                 pass