spaCy/spacy/lang/sv/tokenizer_exceptions.py

161 lines
3.7 KiB
Python
Raw Normal View History

2017-03-12 15:07:28 +03:00
# coding: utf8
2016-12-23 00:45:18 +03:00
from __future__ import unicode_literals
from ...symbols import LEMMA, NORM, ORTH, PRON_LEMMA
2017-05-08 16:54:29 +03:00
_exc = {}
# Verbs
for verb_data in [
{ORTH: "driver"},
{ORTH: "kör"},
{ORTH: "hörr", LEMMA: "hör"},
{ORTH: "fattar"},
{ORTH: "hajar", LEMMA: "förstår"},
{ORTH: "lever"},
{ORTH: "serr", LEMMA: "ser"},
{ORTH: "fixar"},
]:
verb_data_tc = dict(verb_data)
verb_data_tc[ORTH] = verb_data_tc[ORTH].title()
for data in [verb_data, verb_data_tc]:
2017-05-08 16:54:29 +03:00
_exc[data[ORTH] + "u"] = [
dict(data),
{ORTH: "u", LEMMA: PRON_LEMMA, NORM: "du"},
]
2016-12-23 00:45:18 +03:00
# Abbreviations for weekdays "sön." (for "söndag" / "söner")
# are left out because they are ambiguous. The same is the case
# for abbreviations "jul." and "Jul." ("juli" / "jul").
2017-05-08 16:54:29 +03:00
for exc_data in [
{ORTH: "jan.", LEMMA: "januari"},
{ORTH: "febr.", LEMMA: "februari"},
{ORTH: "feb.", LEMMA: "februari"},
{ORTH: "apr.", LEMMA: "april"},
{ORTH: "jun.", LEMMA: "juni"},
{ORTH: "aug.", LEMMA: "augusti"},
{ORTH: "sept.", LEMMA: "september"},
{ORTH: "sep.", LEMMA: "september"},
{ORTH: "okt.", LEMMA: "oktober"},
{ORTH: "nov.", LEMMA: "november"},
{ORTH: "dec.", LEMMA: "december"},
{ORTH: "mån.", LEMMA: "måndag"},
{ORTH: "tis.", LEMMA: "tisdag"},
{ORTH: "ons.", LEMMA: "onsdag"},
{ORTH: "tors.", LEMMA: "torsdag"},
{ORTH: "fre.", LEMMA: "fredag"},
{ORTH: "lör.", LEMMA: "lördag"},
{ORTH: "Jan.", LEMMA: "Januari"},
{ORTH: "Febr.", LEMMA: "Februari"},
{ORTH: "Feb.", LEMMA: "Februari"},
{ORTH: "Apr.", LEMMA: "April"},
{ORTH: "Jun.", LEMMA: "Juni"},
{ORTH: "Aug.", LEMMA: "Augusti"},
{ORTH: "Sept.", LEMMA: "September"},
{ORTH: "Sep.", LEMMA: "September"},
{ORTH: "Okt.", LEMMA: "Oktober"},
{ORTH: "Nov.", LEMMA: "November"},
{ORTH: "Dec.", LEMMA: "December"},
{ORTH: "Mån.", LEMMA: "Måndag"},
{ORTH: "Tis.", LEMMA: "Tisdag"},
{ORTH: "Ons.", LEMMA: "Onsdag"},
{ORTH: "Tors.", LEMMA: "Torsdag"},
{ORTH: "Fre.", LEMMA: "Fredag"},
{ORTH: "Lör.", LEMMA: "Lördag"},
{ORTH: "sthlm", LEMMA: "Stockholm"},
{ORTH: "gbg", LEMMA: "Göteborg"},
]:
2017-11-02 01:02:45 +03:00
_exc[exc_data[ORTH]] = [exc_data]
2016-12-23 00:45:18 +03:00
# Specific case abbreviations only
for orth in ["AB", "Dr.", "H.M.", "H.K.H.", "m/s", "M/S", "Ph.d.", "S:t", "s:t"]:
_exc[orth] = [{ORTH: orth}]
ABBREVIATIONS = [
"ang",
"anm",
"bl.a",
"d.v.s",
"doc",
"dvs",
"e.d",
"e.kr",
2019-02-07 22:54:07 +03:00
"el.",
"eng",
"etc",
"exkl",
2019-02-07 22:54:07 +03:00
"ev",
"f.",
"f.d",
"f.kr",
"f.n",
"f.ö",
"fid",
"fig",
"forts",
"fr.o.m",
"förf",
"inkl",
2019-02-07 22:54:07 +03:00
"iofs",
"jur.",
"kap",
"kl",
2019-02-07 22:54:07 +03:00
"kor.",
"kr",
"kungl",
"lat",
"m.a.o",
"m.fl",
"m.m",
"max",
"milj",
2019-02-07 22:54:07 +03:00
"min.",
"mos",
"mt",
2019-02-07 22:54:07 +03:00
"mvh",
"o.d",
"o.s.v",
"obs",
"osv",
"p.g.a",
"proc",
"prof",
"ref",
"resp",
"s.a.s",
"s.k",
"s.t",
"sid",
"t.ex",
"t.h",
"t.o.m",
"t.v",
"tel",
2019-02-07 22:54:07 +03:00
"ung.",
"vol",
2019-02-07 22:54:07 +03:00
"v.",
"äv",
"övers",
]
# Add abbreviation for trailing punctuation too. If the abbreviation already has a trailing punctuation - skip it.
for abbr in ABBREVIATIONS:
2019-02-08 16:14:49 +03:00
if not abbr.endswith("."):
ABBREVIATIONS.append(abbr + ".")
for orth in ABBREVIATIONS:
2017-05-08 16:54:29 +03:00
_exc[orth] = [{ORTH: orth}]
capitalized = orth.capitalize()
_exc[capitalized] = [{ORTH: capitalized}]
# Sentences ending in "i." (as in "... peka i."), "m." (as in "...än 2000 m."),
# should be tokenized as two separate tokens.
for orth in ["i", "m"]:
_exc[orth + "."] = [{ORTH: orth, LEMMA: orth, NORM: orth}, {ORTH: "."}]
TOKENIZER_EXCEPTIONS = _exc