def initialize():
chk_file = raw_input("What is the file to spellcheck? ")
field = raw_input("What FIELD do you want to spellcheck? ")
s_file = raw_input("What is name of final file? ")
checker = enchant.checker.SpellChecker("en_US")
cmdln = CmdLineChecker()
file_data = pd.read_csv(chk_file)
fields = list(file_data.apply(lambda x:'%s' % (x[field]),axis=1))
# maybe i don't even need this...
#fields = strip_html(fields)
corrected_text = []
for data_field in fields:
checker.set_text(str(data_field))
for err in checker:
print err.word
print err.suggest()
correct = raw_input("provide 0-index int of correct word or i to ignore, e to edit ")
if correct == 'i':
pass
elif correct == 'e':
suggest = raw_input("")
err.replace(suggest)
else:
correct = int(correct)
suggest = err.suggest()[correct]
err.replace(suggest)
corrected_text.append(checker.get_text())
saved_file = write_fixed_file(corrected_text, s_file)
def do_check(checker,to_check):
for text in to_check:
checker.set_text(text)
cmdline_checker = CmdLineChecker()
cmdline_checker.set_checker(checker)
cmdline_checker.run()
to_check[to_check.index(text)] = checker.get_text()
def spellcheck(self, text, tld=''):
from scanner.models import BadWord
# guess language code
self.log.debug(' * guessing language...')
#lang_code, lang_num, lang_name = guess_language.guessLanguageInfo(text)
lang_name, lang_code, reliable, bytes_found, details = \
cld.detect(text.encode('utf-8'), hintTopLevelDomain=tld)
self.log.debug(' -> detected lang: %s (%s)' %
(lang_name, lang_code))
if lang_code.upper() == 'UNKNOWN' or lang_name.upper(
) == 'UNKNOWN' or not reliable:
self.log.warning(
' -> Cannot detect language of page - end : %s' % details)
return None, set()
self.log.debug(' * searching for dictionary')
try:
checker = enchant.checker.SpellChecker(
lang_code,
filters=[
EmailFilter,
URLFilter,
# BetterURLFilter,
])
except enchant.DictNotFoundError:
if lang_code in self.not_supported_lang:
self.log.debug(
" -> Cannot find language for spellchecker for %s - end (blacklisted)"
% lang_code)
else:
self.log.error(
" -> Cannot find language for spellchecker for %s - end"
% lang_code)
return None, set()
# checking page for bad words
self.log.debug(' * check spelling...')
checker.set_text(text)
self.log.debug(' -> ok')
self.log.debug(' * get errors...')
errors = [er.word for er in checker if len(er.word) < 128]
self.log.debug(' -> ok')
self.log.debug(' * found %d bad words and adding them to DB' %
len(errors))
BadWord.objects.bulk_create(
[BadWord(word=bad_word.strip().lower()) for bad_word in errors])
self.log.debug(' -> ok')
self.log.debug(' * call filtering bad words')
errors = BadWord.filter_bad_words(errors)
self.log.debug(' -> ok')
self.log.debug(' * after filtering out there is %d errors (%s)' %
(len(errors), errors))
return lang_name, set(errors)
def spell_check_split_words(self, indices, doc, language, register=-1):
"""
Split misspelled words based on spell-checker suggestions.
Using this is usually not a good idea unless you have an insane
dictionary that contains all possible compound words in `language`.
Raise :exc:`enchant.Error` if dictionary instatiation fails.
"""
new_indices = []
new_texts = []
re_multispace = re.compile(r" +")
checker = self._get_enchant_checker(language)
indices = indices or self.get_all_indices()
for index in indices:
subtitle = self.subtitles[index]
text = subtitle.get_text(doc)
text = re_multispace.sub(" ", text)
checker.set_text(text)
while True:
try:
next(checker)
except StopIteration:
break
if checker.word.capitalize() == checker.word:
# Skip capitalized words, which are usually names
# and thus not always found in dictionaries.
continue
suggestions = []
for i, suggestion in enumerate(checker.suggest()):
if suggestion.find(" ") > 0:
if suggestion.replace(" ", "") == checker.word:
suggestions.append(suggestion)
# Split word only if only one two-word suggestion found that
# has all the same characters as the original unsplit word.
if len(suggestions) != 1: continue
text = checker.get_text()
a = checker.wordpos
z = checker.wordpos + len(checker.word)
checker.set_text(text[:a] + suggestions[0] + text[z:])
new_text = checker.get_text()
if new_text != text:
new_indices.append(index)
new_texts.append(new_text)
if not new_indices: return
self.replace_texts(new_indices, doc, new_texts, register=register)
description = _("Splitting words by spell-check suggestions")
self.set_action_description(register, description)
def spell_check_split_words(self, indices, doc, language, register=-1):
"""
Split misspelled words based on spell-checker suggestions.
Using this is usually not a good idea unless you have an insane
dictionary that contains all possible compound words in `language`.
Raise :exc:`enchant.Error` if dictionary instatiation fails.
"""
new_indices = []
new_texts = []
re_multispace = re.compile(r" +")
checker = self._get_enchant_checker(language)
indices = indices or self.get_all_indices()
for index in indices:
subtitle = self.subtitles[index]
text = subtitle.get_text(doc)
text = re_multispace.sub(" ", text)
checker.set_text(text)
while True:
try: next(checker)
except StopIteration: break
if checker.word.capitalize() == checker.word:
# Skip capitalized words, which are usually names
# and thus not always found in dictionaries.
continue
suggestions = []
for i, suggestion in enumerate(checker.suggest()):
if suggestion.find(" ") > 0:
if suggestion.replace(" ", "") == checker.word:
suggestions.append(suggestion)
# Split word only if only one two-word suggestion found that
# has all the same characters as the original unsplit word.
if len(suggestions) != 1: continue
text = checker.get_text()
a = checker.wordpos
z = checker.wordpos + len(checker.word)
checker.set_text(text[:a] + suggestions[0] + text[z:])
new_text = checker.get_text()
if new_text != text:
new_indices.append(index)
new_texts.append(new_text)
if not new_indices: return
self.replace_texts(new_indices, doc, new_texts, register=register)
description = _("Splitting words by spell-check suggestions")
self.set_action_description(register, description)
def spell_check_join_words(self, indices, doc, language, register=-1):
"""
Join misspelled words based on spell-checker suggestions.
Raise :exc:`enchant.Error` if dictionary instatiation fails.
"""
new_indices = []
new_texts = []
re_multispace = re.compile(r" +")
checker = self._get_enchant_checker(language)
seeker = self._get_enchant_checker(language)
for index in indices or self.get_all_indices():
subtitle = self.subtitles[index]
text = subtitle.get_text(doc)
text = re_multispace.sub(" ", text)
checker.set_text(text)
while True:
try:
next(checker)
except StopIteration:
break
text = checker.get_text()
a = checker.wordpos
z = checker.wordpos + len(checker.word)
ok_with_prev = ok_with_next = False
if checker.leading_context(1) == " ":
seeker.set_text(text[:a - 1] + text[a:])
poss = self._get_misspelled_indices(seeker)
ok_with_prev = not (a - 1) in poss
if checker.trailing_context(1) == " ":
seeker.set_text(text[:z] + text[z + 1:])
poss = self._get_misspelled_indices(seeker)
ok_with_next = not a in poss
# Join backwards or forwards if only one direction,
# but not both, produce a correctly spelled result.
if ok_with_prev and not ok_with_next:
checker.set_text(text[:a - 1] + text[a:])
if ok_with_next and not ok_with_prev:
checker.set_text(text[:z] + text[z + 1:])
new_text = checker.get_text()
if new_text != text:
new_indices.append(index)
new_texts.append(new_text)
if not new_indices: return
self.replace_texts(new_indices, doc, new_texts, register=register)
description = _("Joining words by spell-check suggestions")
self.set_action_description(register, description)
def spell_check_join_words(self, indices, doc, language, register=-1):
"""
Join misspelled words based on spell-checker suggestions.
Raise :exc:`enchant.Error` if dictionary instatiation fails.
"""
new_indices = []
new_texts = []
re_multispace = re.compile(r" +")
checker = self._get_enchant_checker(language)
seeker = self._get_enchant_checker(language)
for index in indices or self.get_all_indices():
subtitle = self.subtitles[index]
text = subtitle.get_text(doc)
text = re_multispace.sub(" ", text)
checker.set_text(text)
while True:
try: next(checker)
except StopIteration: break
text = checker.get_text()
a = checker.wordpos
z = checker.wordpos + len(checker.word)
ok_with_prev = ok_with_next = False
if checker.leading_context(1) == " ":
seeker.set_text(text[:a - 1] + text[a:])
poss = self._get_misspelled_indices(seeker)
ok_with_prev = not (a - 1) in poss
if checker.trailing_context(1) == " ":
seeker.set_text(text[:z] + text[z + 1:])
poss = self._get_misspelled_indices(seeker)
ok_with_next = not a in poss
# Join backwards or forwards if only one direction,
# but not both, produce a correctly spelled result.
if ok_with_prev and not ok_with_next:
checker.set_text(text[:a - 1] + text[a:])
if ok_with_next and not ok_with_prev:
checker.set_text(text[:z] + text[z + 1:])
new_text = checker.get_text()
if new_text != text:
new_indices.append(index)
new_texts.append(new_text)
if not new_indices: return
self.replace_texts(new_indices, doc, new_texts, register=register)
description = _("Joining words by spell-check suggestions")
self.set_action_description(register, description)
def spell_check(pkg, str, fmt, lang, ignored):
dict_found = True
warned = set()
if enchant:
if lang == 'C':
lang = 'en_US'
checker = _enchant_checkers.get(lang)
if not checker and lang not in _enchant_checkers:
try:
checker = enchant.checker.SpellChecker(
lang, filters=[enchant.tokenize.EmailFilter,
enchant.tokenize.URLFilter,
enchant.tokenize.WikiWordFilter])
except enchant.DictNotFoundError:
printInfo(pkg, 'enchant-dictionary-not-found', lang)
pass
_enchant_checkers[lang] = checker
if checker:
# squeeze whitespace to ease leading context check
checker.set_text(re.sub(r'\s+', ' ', str))
if use_utf8:
uppername = Pkg.to_unicode(pkg.header[rpm.RPMTAG_NAME]).upper()
else:
uppername = pkg.name.upper()
upperparts = uppername.split('-')
if lang.startswith('en'):
ups = [x + "'S" for x in upperparts]
upperparts.extend(ups)
for err in checker:
# Skip already warned and ignored words
if err.word in warned or err.word in ignored:
continue
# Skip all capitalized words that do not start a sentence
if err.word[0].isupper() and not \
sentence_break_regex.search(checker.leading_context(3)):
continue
upperword = err.word.upper()
# Skip all uppercase words
if err.word == upperword:
continue
# Skip errors containing package name or equal to a
# "component" of it, case insensitively
if uppername in upperword or upperword in upperparts:
continue
# Work around enchant's digit tokenizing behavior:
# http://github.com/rfk/pyenchant/issues/issue/3
if checker.leading_context(1).isdigit() or \
checker.trailing_context(1).isdigit():
continue
# Warn and suggest
sug = ', '.join(checker.suggest()[:3])
if sug:
sug = '-> %s' % sug
printWarning(pkg, 'spelling-error', fmt % lang, err.word, sug)
warned.add(err.word)
else:
dict_found = False
if not enchant or not dict_found:
for seq in str.split():
for word in re.split(r'[^a-z]+', seq.lower()):
if len(word) == 0:
continue
correct = BAD_WORDS.get(word)
if not correct:
continue
if word[0] == '\'':
word = word[1:]
if word[-1] == '\'':
word = word[:-1]
if word in warned or word in ignored:
continue
printWarning(pkg, 'spelling-error', fmt % lang, word, '->',
correct)
warned.add(word)
def spell_check(pkg, str, fmt, lang, ignored):
dict_found = True
warned = set()
if enchant:
if lang == 'C':
lang = 'en_US'
checker = _enchant_checkers.get(lang)
if not checker and lang not in _enchant_checkers:
try:
checker = enchant.checker.SpellChecker(
lang, filters=[enchant.tokenize.EmailFilter,
enchant.tokenize.URLFilter,
enchant.tokenize.WikiWordFilter])
except enchant.DictNotFoundError:
printInfo(pkg, 'enchant-dictionary-not-found', lang)
pass
_enchant_checkers[lang] = checker
if checker:
# squeeze whitespace to ease leading context check
checker.set_text(re.sub(r'\s+', ' ', str))
if use_utf8:
uppername = Pkg.to_unicode(pkg.header[rpm.RPMTAG_NAME]).upper()
else:
uppername = pkg.name.upper()
upperparts = uppername.split('-')
if lang.startswith('en'):
ups = [x + "'S" for x in upperparts]
upperparts.extend(ups)
for err in checker:
# Skip already warned and ignored words
if err.word in warned or err.word in ignored:
continue
# Skip all capitalized words that do not start a sentence
if err.word[0].isupper() and not \
sentence_break_regex.search(checker.leading_context(3)):
continue
upperword = err.word.upper()
# Skip all uppercase words
if err.word == upperword:
continue
# Skip errors containing package name or equal to a
# "component" of it, case insensitively
if uppername in upperword or upperword in upperparts:
continue
# Work around enchant's digit tokenizing behavior:
# http://github.com/rfk/pyenchant/issues/issue/3
if checker.leading_context(1).isdigit() or \
checker.trailing_context(1).isdigit():
continue
# Warn and suggest
sug = ', '.join(checker.suggest()[:3])
if sug:
sug = '-> %s' % sug
printWarning(pkg, 'spelling-error', fmt % lang, err.word, sug)
warned.add(err.word)
else:
dict_found = False
if not enchant or not dict_found:
for seq in str.split():
for word in re.split('[^a-z]+', seq.lower()):
if len(word) == 0:
continue
correct = BAD_WORDS.get(word)
if not correct:
continue
if word[0] == '\'':
word = word[1:]
if word[-1] == '\'':
word = word[:-1]
if word in warned or word in ignored:
continue
printWarning(pkg, 'spelling-error', fmt % lang, word, '->',
correct)
warned.add(word)
def check_lang(t):
code, num, name = guess_language.guessLanguageInfo(t)
checker = enchant.checker.SpellChecker(code,
filters=[EmailFilter, URLFilter])
checker.set_text(t)
return code, [x.word for x in checker]
