spaCy/spacy/lang/ky/punctuation.py

25 lines
904 B
Python
Raw Normal View History

2021-01-23 15:31:31 +00:00
# coding: utf8
from __future__ import unicode_literals
from ..char_classes import ALPHA, ALPHA_LOWER, ALPHA_UPPER, CONCAT_QUOTES, HYPHENS
from ..char_classes import LIST_ELLIPSES, LIST_ICONS
_hyphens_no_dash = HYPHENS.replace("-", "").strip("|").replace("||", "")
_infixes = (
LIST_ELLIPSES
+ LIST_ICONS
+ [
r"(?<=[{al}])\.(?=[{au}])".format(al=ALPHA_LOWER, au=ALPHA_UPPER),
r"(?<=[{a}])[,!?/()]+(?=[{a}])".format(a=ALPHA),
r"(?<=[{a}{q}])[:<>=](?=[{a}])".format(a=ALPHA, q=CONCAT_QUOTES),
r"(?<=[{a}])--(?=[{a}])".format(a=ALPHA),
r"(?<=[{a}]),(?=[{a}])".format(a=ALPHA),
r"(?<=[{a}])([{q}\)\]\(\[])(?=[\-{a}])".format(a=ALPHA, q=CONCAT_QUOTES),
r"(?<=[{a}])(?:{h})(?=[{a}])".format(a=ALPHA, h=_hyphens_no_dash),
2021-01-24 14:54:22 +00:00
r"(?<=[0-9])-(?=[{a}])".format(a=ALPHA),
2021-01-23 15:31:31 +00:00
r"(?<=[0-9])-(?=[0-9])",
]
)
TOKENIZER_INFIXES = _infixes