added hashtag rule and fixed currency rules

This commit is contained in:
Jim Geovedi 2017-07-30 21:23:28 +07:00
parent e9af79a803
commit bb08d696f9

View File

@ -25,7 +25,7 @@ HTML_SUFFIX = r'</(b|strong|i|em|p|span|div|a)>'
MONTHS = merge_chars(_months)
LIST_CURRENCY = split_chars(_currency)
TOKENIZER_PREFIXES.remove('#') # hashtag
_prefixes = TOKENIZER_PREFIXES + LIST_CURRENCY + [HTML_PREFIX] + ['/', '']
_suffixes = TOKENIZER_SUFFIXES + [r'\-[Nn]ya', '-[KkMm]u', '[—-]'] + [