CIRCL
/
AIL-framework
mirror of https://github.com/CIRCL/AIL-framework
2
1
Fork 0
Code Issues Releases Wiki Activity

fix: [languages] fix items language min probability

ocr
terrtia 2024-03-27 11:15:22 +01:00
parent ee563a79d3
commit a3a664b7f1
No known key found for this signature in database
GPG Key ID: 1E1B1F50D84613D0
1 changed files with 7 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

10
bin/lib/Language.py
View File

@ -434,7 +434,7 @@ def delete_obj_translation(obj_global_id, language, field=''):
class LanguagesDetector: class LanguagesDetector:
def __init__(self, nb_langs=3, min_proportion=0.2, min_probability=0.7, min_len=0): def __init__(self, nb_langs=3, min_proportion=0.2, min_probability=-1, min_len=0):
self.lt = LibreTranslateAPI(get_translator_instance()) self.lt = LibreTranslateAPI(get_translator_instance())
try: try:
self.lt.languages() self.lt.languages()
@ -461,9 +461,11 @@ class LanguagesDetector:
languages.append(lang.language) languages.append(lang.language)
return languages return languages
def detect_lexilang(self, content): # TODO clean text ??? - TODO REMOVE SEPARATOR def detect_lexilang(self, content):
language, prob = lexilang_detect(content) language, prob = lexilang_detect(content)
if prob > 0: if prob > 0 and self.min_probability == -1:
return [language]
elif prob > 0.4:
return [language] return [language]
else: else:
return [] return []
@ -506,6 +508,8 @@ class LanguagesDetector:
# libretranslate # libretranslate
# else: # else:
# languages = self.detect_libretranslate(content) # languages = self.detect_libretranslate(content)
if not languages:
languages = []
return languages return languages
class LanguageTranslator: class LanguageTranslator: