spaCy/spacy/lang/en/lex_attrs.py

60 lines
988 B
Python
Raw Normal View History

2017-05-09 02:09:52 +03:00
from ...attrs import LIKE_NUM
2017-05-08 16:47:25 +03:00
_num_words = [
"zero",
"one",
"two",
"three",
"four",
"five",
"six",
"seven",
"eight",
"nine",
"ten",
"eleven",
"twelve",
"thirteen",
"fourteen",
"fifteen",
"sixteen",
"seventeen",
"eighteen",
"nineteen",
"twenty",
"thirty",
"forty",
"fifty",
"sixty",
"seventy",
"eighty",
"ninety",
"hundred",
"thousand",
"million",
"billion",
"trillion",
"quadrillion",
"gajillion",
"bazillion",
]
2017-05-08 16:47:25 +03:00
2017-05-09 02:09:52 +03:00
def like_num(text):
if text.startswith(("+", "-", "±", "~")):
text = text[1:]
text = text.replace(",", "").replace(".", "")
2017-05-09 02:09:52 +03:00
if text.isdigit():
return True
if text.count("/") == 1:
num, denom = text.split("/")
2017-05-09 02:09:52 +03:00
if num.isdigit() and denom.isdigit():
return True
2018-01-08 05:25:08 +03:00
if text.lower() in _num_words:
2017-05-09 02:09:52 +03:00
return True
return False
LEX_ATTRS = {LIKE_NUM: like_num}