| 1 | #!/usr/bin/env python2
|
| 2 | """
|
| 3 | spelling.py
|
| 4 |
|
| 5 | Filter the output of 'lynx -dump' into a list of words to spell check.
|
| 6 | """
|
| 7 | from __future__ import print_function
|
| 8 |
|
| 9 | from collections import Counter
|
| 10 | import optparse
|
| 11 | import re
|
| 12 | import sys
|
| 13 |
|
| 14 | from doctools.util import log
|
| 15 |
|
| 16 |
|
| 17 | def SplitWords(contents):
|
| 18 | # Remove URLs so path components don't show up as words
|
| 19 | contents = re.sub(r'(http|https|file)://\S+', '', contents)
|
| 20 |
|
| 21 | # Take into account contractions with apostrophes
|
| 22 | #
|
| 23 | # - doesn't
|
| 24 | # - can't
|
| 25 |
|
| 26 | WORD_RE = re.compile(r'''
|
| 27 | [a-zA-Z]+
|
| 28 | (?:\'t\b)? # optional contraction
|
| 29 | ''', re.VERBOSE)
|
| 30 |
|
| 31 | words = WORD_RE.findall(contents)
|
| 32 |
|
| 33 | for w in words:
|
| 34 | yield w
|
| 35 |
|
| 36 |
|
| 37 | def WordList(f):
|
| 38 | for line in f:
|
| 39 | # no special characters allowed
|
| 40 | yield line.strip()
|
| 41 |
|
| 42 |
|
| 43 | def Options():
|
| 44 | """Returns an option parser instance."""
|
| 45 | p = optparse.OptionParser()
|
| 46 | p.add_option(
|
| 47 | '--known-words', dest='known_words',
|
| 48 | help='List of words like /usr/share/dict/words')
|
| 49 | p.add_option(
|
| 50 | '--more-than-bash', dest='more_than_bash', type=int, default=0,
|
| 51 | help='Expected number of cases where OSH starts more processes than bash')
|
| 52 | return p
|
| 53 |
|
| 54 |
|
| 55 | def main(argv):
|
| 56 | o = Options()
|
| 57 | opts, argv = o.parse_args(argv[1:])
|
| 58 |
|
| 59 | action = argv[0]
|
| 60 |
|
| 61 | if action == 'word-split':
|
| 62 | contents = sys.stdin.read()
|
| 63 | for w in SplitWords(contents):
|
| 64 | print(w)
|
| 65 |
|
| 66 | elif action == 'check':
|
| 67 | word_files = argv[1:]
|
| 68 |
|
| 69 | d = Counter()
|
| 70 |
|
| 71 | for path in word_files:
|
| 72 | with open(path) as f:
|
| 73 | for word in WordList(f):
|
| 74 | d[word] += 1
|
| 75 |
|
| 76 | print('')
|
| 77 | print('Most common words')
|
| 78 | print('')
|
| 79 | for word, count in d.most_common()[:20]:
|
| 80 | print('%10d %s' % (count, word))
|
| 81 |
|
| 82 | print('')
|
| 83 | print('Least common words')
|
| 84 | print('')
|
| 85 | for word, count in d.most_common()[-20:]:
|
| 86 | print('%10d %s' % (count, word))
|
| 87 |
|
| 88 | log('%d word files', len(word_files))
|
| 89 | log('%d unique words', len(d))
|
| 90 |
|
| 91 | known_words = {}
|
| 92 | with open(opts.known_words) as f:
|
| 93 | for w in WordList(f):
|
| 94 | known_words[w] = True
|
| 95 |
|
| 96 | print('')
|
| 97 | print('Potential Misspellings')
|
| 98 | print('')
|
| 99 |
|
| 100 | for path in word_files:
|
| 101 |
|
| 102 | print()
|
| 103 | print('\t%s' % path)
|
| 104 | print()
|
| 105 |
|
| 106 | with open(path) as f:
|
| 107 | unknown = {}
|
| 108 | for w in WordList(f):
|
| 109 | #if d.get(word) == 1:
|
| 110 | # print(word)
|
| 111 | if w.lower() not in known_words:
|
| 112 | unknown[w] = True
|
| 113 |
|
| 114 | if unknown:
|
| 115 | for u in sorted(unknown):
|
| 116 | # only occurs once
|
| 117 | if d.get(u) == 1:
|
| 118 | print(u)
|
| 119 | log('\t%d unknown words in %s', len(unknown), path)
|
| 120 |
|
| 121 |
|
| 122 | # Checking algorithms:
|
| 123 | #
|
| 124 | # - Does it appear in the dictionary? Problem: most computer terms
|
| 125 | # - Does it appear only once or twice in the whole corpus?
|
| 126 | # - Is the edit distance very close to a dictinoary word?
|
| 127 | # - e.g. substitutions is a typo
|
| 128 |
|
| 129 | else:
|
| 130 | raise RuntimeError('Invalid action %r' % action)
|
| 131 |
|
| 132 |
|
| 133 | if __name__ == '__main__':
|
| 134 | try:
|
| 135 | main(sys.argv)
|
| 136 | except RuntimeError as e:
|
| 137 | print('FATAL: %s' % e, file=sys.stderr)
|
| 138 | sys.exit(1)
|