cola/spellcheck.py

   1 from __future__ import absolute_import, division, print_function, unicode_literals
   2 import codecs
   3 import collections
   4 import os
   5
   6 from . import resources
   7
   8 __copyright__ = """
   9 2012 Peter Norvig (http://norvig.com/spell-correct.html)
  10 2013-2018 David Aguilar <davvid@gmail.com>
  11 """
  12
  13 alphabet = 'abcdefghijklmnopqrstuvwxyz'
  14
  15
  16 def train(features, model):
  17     for f in features:
  18         model[f] += 1
  19     return model
  20
  21
  22 def edits1(word):
  23     splits = [(word[:i], word[i:]) for i in range(len(word) + 1)]
  24     deletes = [a + b[1:] for a, b in splits if b]
  25     transposes = [a + b[1] + b[0] + b[2:] for a, b in splits if len(b) > 1]
  26     replaces = [a + c + b[1:] for a, b in splits for c in alphabet if b]
  27     inserts = [a + c + b for a, b in splits for c in alphabet]
  28     return set(deletes + transposes + replaces + inserts)
  29
  30
  31 def known_edits2(word, words):
  32     return set(e2 for e1 in edits1(word) 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 = (
  41         known([word], words)
  42         or known(edits1(word), words)
  43         or known_edits2(word, words)
  44         or [word]
  45     )
  46     return candidates
  47
  48
  49 def correct(word, words):
  50     candidates = suggest(word, words)
  51     return max(candidates, key=words.get)
  52
  53
  54 class NorvigSpellCheck(object):
  55     def __init__(
  56         self,
  57         words='dict/words',
  58         propernames='dict/propernames',
  59     ):
  60         data_dirs = resources.xdg_data_dirs()
  61         self.dictwords = resources.find_first(words, data_dirs)
  62         self.propernames = resources.find_first(propernames, data_dirs)
  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         propernames = self.propernames
  95         cfg_dictionary = self.dictionary
  96
  97         if words and os.path.exists(words):
  98             paths.append((words, True))
  99
 100         if propernames and os.path.exists(propernames):
 101             paths.append((propernames, False))
 102
 103         if cfg_dictionary and os.path.exists(cfg_dictionary):
 104             paths.append((cfg_dictionary, False))
 105
 106         for (path, title) in paths:
 107             try:
 108                 with codecs.open(
 109                     path, 'r', encoding='utf-8', errors='ignore'
 110                 ) as words_file:
 111                     for line in words_file:
 112                         word = line.rstrip()
 113                         yield word
 114                         if title:
 115                             yield word.title()
 116             except (IOError, OSError):
 117                 pass