spaCy/spacy/lang/el/punctuation.py

96 lines
3.3 KiB
Python
Raw Normal View History

from ..char_classes import LIST_PUNCT, LIST_ELLIPSES, LIST_QUOTES, LIST_CURRENCY
from ..char_classes import LIST_ICONS, ALPHA_LOWER, ALPHA_UPPER, ALPHA, HYPHENS
from ..char_classes import CONCAT_QUOTES, CURRENCY
_units = (
"km km² km³ m m² m³ dm dm² dm³ cm cm² cm³ mm mm² mm³ ha µm nm yd in ft "
"kg g mg µg t lb oz m/s km/h kmh mph hPa Pa mbar mb MB kb KB gb GB tb "
"TB T G M K км км² км³ м м² м³ дм дм² дм³ см см² см³ мм мм² мм³ нм "
"кг г мг м/с км/ч кПа Па мбар Кб КБ кб Мб МБ мб Гб ГБ гб Тб ТБ тб"
)
def merge_chars(char):
return char.strip().replace(" ", "|")
UNITS = merge_chars(_units)
_prefixes = (
[
"''",
"§",
"%",
"=",
r"\+[0-9]+%", # 90%
r"\'([0-9]){2}([\-]\'([0-9]){2})*", # '12'-13
r"\-([0-9]){1,9}\.([0-9]){1,9}", # -12.13
r"\'([Α-Ωα-ωίϊΐόάέύϋΰήώ]+)\'", # 'αβγ'
r"([Α-Ωα-ωίϊΐόάέύϋΰήώ]){1,3}\'", # αβγ'
r"http://www.[A-Za-z]+\-[A-Za-z]+(\.[A-Za-z]+)+(\/[A-Za-z]+)*(\.[A-Za-z]+)*",
r"[ΈΆΊΑ-Ωα-ωίϊΐόάέύϋΰήώ]+\*", # όνομα*
r"\$([0-9])+([\,\.]([0-9])+){0,1}",
]
+ LIST_PUNCT
+ LIST_ELLIPSES
+ LIST_QUOTES
+ LIST_CURRENCY
+ LIST_ICONS
)
_suffixes = (
LIST_PUNCT
+ LIST_ELLIPSES
+ LIST_QUOTES
+ LIST_ICONS
+ [
r"(?<=[0-9])\+", # 12+
r"([0-9])+\'", # 12'
r"([A-Za-z])?\'", # a'
r"^([0-9]){1,2}\.", # 12.
r" ([0-9]){1,2}\.", # 12.
r"([0-9]){1}\) ", # 12)
r"^([0-9]){1}\)$", # 12)
r"(?<=°[FfCcKk])\.",
r"([0-9])+\&", # 12&
r"(?<=[0-9])(?:{c})".format(c=CURRENCY),
r"(?<=[0-9])(?:{u})".format(u=UNITS),
2019-08-20 18:36:34 +03:00
r"(?<=[0-9{al}{e}(?:{q})])\.".format(
al=ALPHA_LOWER, e=r"²\-\+", q=CONCAT_QUOTES
),
r"(?<=[{au}][{au}])\.".format(au=ALPHA_UPPER),
r"(?<=[Α-Ωα-ωίϊΐόάέύϋΰήώ])\-", # όνομα-
r"(?<=[Α-Ωα-ωίϊΐόάέύϋΰήώ])\.",
r"^[Α-Ω]{1}\.",
r"\ [Α-Ω]{1}\.",
# πρώτος-δεύτερος , πρώτος-δεύτερος-τρίτος
r"[ΈΆΊΑΌ-Ωα-ωίϊΐόάέύϋΰήώ]+([\-]([ΈΆΊΑΌ-Ωα-ωίϊΐόάέύϋΰήώ]+))+",
r"([0-9]+)mg", # 13mg
r"([0-9]+)\.([0-9]+)m", # 1.2m
]
)
_infixes = (
LIST_ELLIPSES
+ LIST_ICONS
+ [
r"(?<=[0-9])[+\/\-\*^](?=[0-9])", # 1/2 , 1-2 , 1*2
r"([a-zA-Z]+)\/([a-zA-Z]+)\/([a-zA-Z]+)", # name1/name2/name3
r"([0-9])+(\.([0-9]+))*([\-]([0-9])+)+", # 10.9 , 10.9.9 , 10.9-6
r"([0-9])+[,]([0-9])+[\-]([0-9])+[,]([0-9])+", # 10,11,12
r"([0-9])+[ης]+([\-]([0-9])+)+", # 1ης-2
# 15/2 , 15/2/17 , 2017/2/15
r"([0-9]){1,4}[\/]([0-9]){1,2}([\/]([0-9]){0,4}){0,1}",
r"[A-Za-z]+\@[A-Za-z]+(\-[A-Za-z]+)*\.[A-Za-z]+", # abc@cde-fgh.a
r"([a-zA-Z]+)(\-([a-zA-Z]+))+", # abc-abc
r"(?<=[{al}])\.(?=[{au}])".format(al=ALPHA_LOWER, au=ALPHA_UPPER),
r"(?<=[{a}]),(?=[{a}])".format(a=ALPHA),
2019-08-20 18:36:34 +03:00
r"(?<=[{a}])(?:{h})(?=[{a}])".format(a=ALPHA, h=HYPHENS),
r"(?<=[{a}])[:<>=/](?=[{a}])".format(a=ALPHA),
]
)
TOKENIZER_PREFIXES = _prefixes
TOKENIZER_SUFFIXES = _suffixes
TOKENIZER_INFIXES = _infixes