mirror of
https://github.com/rspeer/wordfreq.git
synced 2024-12-23 09:21:37 +00:00
forgot about Italian
This commit is contained in:
parent
5ef807117d
commit
8f6cd0e57b
@ -48,8 +48,8 @@ def cld2_surface_tokenizer(text):
|
||||
# Low-frequency languages tend to be detected incorrectly. Keep a limited
|
||||
# list of languages we're allowed to use here.
|
||||
KEEP_THESE_LANGUAGES = {
|
||||
'ar', 'de', 'el', 'en', 'es', 'fr', 'hr', 'id', 'ja', 'ko', 'ms', 'nl',
|
||||
'pl', 'pt', 'ro', 'ru', 'sv', 'th'
|
||||
'ar', 'de', 'el', 'en', 'es', 'fr', 'hr', 'id', 'it', 'ja', 'ko', 'ms',
|
||||
'nl', 'pl', 'pt', 'ro', 'ru', 'sv', 'th'
|
||||
}
|
||||
|
||||
|
||||
|
Loading…
Reference in New Issue
Block a user