28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
|
r'(?P<PUNC>[][,.;:!?…«»“”‘’"(){}·–—¿¡])',
r'(?P<WORD_ACRONYM>[A-Z][.][A-Z][.](?:[A-Z][.])*)',
r'(?P<LINK>(?:https?://|www[.]|\w+[@.]\w\w+[@.])\w[\w./?&!%=+*"\'@$#-]+)',
r'(?P<HASHTAG>[#@][\w-]+)',
r'(?P<HTML><\w+.*?>|</\w+ *>)',
r'(?P<PSEUDOHTML>\[/?\w+\])',
r"(?P<WORD_ELIDED>(?:l|d|n|m|t|s|j|c|ç|lorsqu|puisqu|jusqu|quoiqu|qu)['’`])",
r'(?P<WORD_ORDINAL>\d+(?:ers?|nds?|es?|des?|ièmes?|èmes?|emes?|ᵉʳˢ?|ⁿᵈˢ?|ᵉˢ?|ᵈᵉˢ?)\b)',
r'(?P<HOUR>\d\d?[h:]\d\d(?:[m:]\d\ds?|)\b)',
r'(?P<NUM>\d+(?:[.,]\d+|))',
r'(?P<SIGN>[&%‰€$+±=*/<>⩾⩽#|×¥£¢§¬÷@-])',
r"(?P<WORD>\w+(?:[’'`-]\w+)*)"
)
}
|
|
|
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
|
r'(?P<PUNC>[][,.;:!?…«»“”‘’"(){}·–—¿¡])',
r'(?P<WORD_ACRONYM>[A-Z][.][A-Z][.](?:[A-Z][.])*)',
r'(?P<LINK>(?:https?://|www[.]|\w+[@.]\w\w+[@.])\w[\w./?&!%=+*"\'@$#-]+)',
r'(?P<HASHTAG>[#@][\w-]+)',
r'(?P<HTML><\w+.*?>|</\w+ *>)',
r'(?P<PSEUDOHTML>\[/?\w+\])',
r"(?P<WORD_ELIDED>(?:l|d|n|m|t|s|j|c|ç|lorsqu|puisqu|jusqu|quoiqu|qu)['’`])",
r'(?P<WORD_ORDINAL>\d+(?:ers?|res?|è[rm]es?|i[èe][mr]es?|de?s?|nde?s?|ès?|es?|ᵉʳˢ?|ʳᵉˢ?|ᵈᵉ?ˢ?|ⁿᵈᵉ?ˢ?|ᵉˢ?)\b)',
r'(?P<HOUR>\d\d?[h:]\d\d(?:[m:]\d\ds?|)\b)',
r'(?P<NUM>\d+(?:[.,]\d+|))',
r'(?P<SIGN>[&%‰€$+±=*/<>⩾⩽#|×¥£¢§¬÷@-])',
r"(?P<WORD>\w+(?:[’'`-]\w+)*)"
)
}
|