Cutlery/calibre
1
0
Fork 0
mirror of https://github.com/kovidgoyal/calibre.git synced 2025-07-09 03:04:10 -04:00
Code Issues Packages Projects Releases Wiki Activity

Edit Book: Fix spell check treating the unicode hyphen (U+2010) differently from the normal hyphen (U+002d). Fixes #1656319 [Problems with hyphen in spell checking](https://bugs.launchpad.net/calibre/+bug/1656319)

Browse Source
This commit is contained in:
Kovid Goyal 2017-01-14 10:10:44 +05:30
parent 6011d45ef3
commit 97a1a92e1b
1 changed files with 13 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

14
src/calibre/spell/dictionary.py
View File

@ -44,6 +44,7 @@ class UserDictionary(object):
return {'name':self.name, 'is_active': self.is_active, 'words':[ return {'name':self.name, 'is_active': self.is_active, 'words':[
(w, l) for w, l in self.words]} (w, l) for w, l in self.words]}
_builtins = _custom = None _builtins = _custom = None
@ -82,6 +83,7 @@ def custom_dictionaries(reread=False):
_custom = frozenset(dics) _custom = frozenset(dics)
return _custom return _custom
default_en_locale = 'en-US' default_en_locale = 'en-US'
try: try:
ul = parse_lang_code(get_system_locale() or 'en-US') ul = parse_lang_code(get_system_locale() or 'en-US')
@ -368,7 +370,7 @@ class Dictionaries(object):
d = self.dictionary_for_locale(locale) d = self.dictionary_for_locale(locale)
if d is not None: if d is not None:
try: try:
ans = d.obj.recognized(word) ans = d.obj.recognized(word.replace('\u2010', '-'))
except ValueError: except ValueError:
pass pass
else: else:
@ -381,6 +383,7 @@ class Dictionaries(object):
def suggestions(self, word, locale=None): def suggestions(self, word, locale=None):
locale = locale or self.default_locale locale = locale or self.default_locale
d = self.dictionary_for_locale(locale) d = self.dictionary_for_locale(locale)
has_unicode_hyphen = '\u2010' in word
ans = () ans = ()
def add_suggestion(w, ans): def add_suggestion(w, ans):
@ -407,6 +410,8 @@ class Dictionaries(object):
fw = w1 + m.group() + ' ' + capitalize(w2) fw = w1 + m.group() + ' ' + capitalize(w2)
ans = add_suggestion(fw, ans) ans = add_suggestion(fw, ans)
if has_unicode_hyphen:
ans = tuple(w.replace('-', '\u2010') for w in ans)
return ans return ans
@ -418,6 +423,12 @@ def test_dictionaries():
sg = partial(dictionaries.suggestions, locale=eng) sg = partial(dictionaries.suggestions, locale=eng)
if not rec('recognized'): if not rec('recognized'):
raise ValueError('recognized not recognized') raise ValueError('recognized not recognized')
if not rec('one-half'):
raise ValueError('one-half not recognized')
if not rec('one\u2010half'):
raise ValueError('one\u2010half not recognized with unicode hyphen (U+2010)')
if 'one\u2010half' not in sg('oone\u2010half'):
raise ValueError('Unicode hyphen not preserved in suggestions')
if 'adequately' not in sg('ade-quately'): if 'adequately' not in sg('ade-quately'):
raise ValueError('adequately not in %s' % sg('ade-quately')) raise ValueError('adequately not in %s' % sg('ade-quately'))
if 'magic. Wand' not in sg('magic.wand'): if 'magic. Wand' not in sg('magic.wand'):
@ -425,5 +436,6 @@ def test_dictionaries():
d = load_dictionary(get_dictionary(parse_lang_code('es'))).obj d = load_dictionary(get_dictionary(parse_lang_code('es'))).obj
assert d.recognized('Achí') assert d.recognized('Achí')
if __name__ == '__main__': if __name__ == '__main__':
test_dictionaries() test_dictionaries()