mirror of
https://github.com/explosion/spaCy.git
synced 2024-11-14 13:47:13 +03:00
172 lines
1.1 KiB
Python
172 lines
1.1 KiB
Python
|
# coding: utf8
|
||
|
|
from __future__ import unicode_literals
|
||
|
|
|
||
|
|
|
||
|
|
# Source: https://github.com/stopwords-iso/stopwords-lv
|
||
|
|
|
||
|
|
STOP_WORDS = set(
|
||
|
|
"""
|
||
|
|
aiz
|
||
|
|
ap
|
||
|
|
apakš
|
||
|
|
apakšpus
|
||
|
|
ar
|
||
|
|
arī
|
||
|
|
augšpus
|
||
|
|
bet
|
||
|
|
bez
|
||
|
|
bija
|
||
|
|
biji
|
||
|
|
biju
|
||
|
|
bijām
|
||
|
|
bijāt
|
||
|
|
būs
|
||
|
|
būsi
|
||
|
|
būsiet
|
||
|
|
būsim
|
||
|
|
būt
|
||
|
|
būšu
|
||
|
|
caur
|
||
|
|
diemžēl
|
||
|
|
diezin
|
||
|
|
droši
|
||
|
|
dēļ
|
||
|
|
esam
|
||
|
|
esat
|
||
|
|
esi
|
||
|
|
esmu
|
||
|
|
gan
|
||
|
|
gar
|
||
|
|
iekam
|
||
|
|
iekams
|
||
|
|
iekām
|
||
|
|
iekāms
|
||
|
|
iekš
|
||
|
|
iekšpus
|
||
|
|
ik
|
||
|
|
ir
|
||
|
|
it
|
||
|
|
itin
|
||
|
|
iz
|
||
|
|
ja
|
||
|
|
jau
|
||
|
|
jeb
|
||
|
|
jebšu
|
||
|
|
jel
|
||
|
|
jo
|
||
|
|
jā
|
||
|
|
ka
|
||
|
|
kamēr
|
||
|
|
kaut
|
||
|
|
kolīdz
|
||
|
|
kopš
|
||
|
|
kā
|
||
|
|
kļuva
|
||
|
|
kļuvi
|
||
|
|
kļuvu
|
||
|
|
kļuvām
|
||
|
|
kļuvāt
|
||
|
|
kļūs
|
||
|
|
kļūsi
|
||
|
|
kļūsiet
|
||
|
|
kļūsim
|
||
|
|
kļūst
|
||
|
|
kļūstam
|
||
|
|
kļūstat
|
||
|
|
kļūsti
|
||
|
|
kļūstu
|
||
|
|
kļūt
|
||
|
|
kļūšu
|
||
|
|
labad
|
||
|
|
lai
|
||
|
|
lejpus
|
||
|
|
līdz
|
||
|
|
līdzko
|
||
|
|
ne
|
||
|
|
nebūt
|
||
|
|
nedz
|
||
|
|
nekā
|
||
|
|
nevis
|
||
|
|
nezin
|
||
|
|
no
|
||
|
|
nu
|
||
|
|
nē
|
||
|
|
otrpus
|
||
|
|
pa
|
||
|
|
par
|
||
|
|
pat
|
||
|
|
pie
|
||
|
|
pirms
|
||
|
|
pret
|
||
|
|
priekš
|
||
|
|
pār
|
||
|
|
pēc
|
||
|
|
starp
|
||
|
|
tad
|
||
|
|
tak
|
||
|
|
tapi
|
||
|
|
taps
|
||
|
|
tapsi
|
||
|
|
tapsiet
|
||
|
|
tapsim
|
||
|
|
tapt
|
||
|
|
tapāt
|
||
|
|
tapšu
|
||
|
|
taču
|
||
|
|
te
|
||
|
|
tiec
|
||
|
|
tiek
|
||
|
|
tiekam
|
||
|
|
tiekat
|
||
|
|
tieku
|
||
|
|
tik
|
||
|
|
tika
|
||
|
|
tikai
|
||
|
|
tiki
|
||
|
|
tikko
|
||
|
|
tiklab
|
||
|
|
tiklīdz
|
||
|
|
tiks
|
||
|
|
tiksiet
|
||
|
|
tiksim
|
||
|
|
tikt
|
||
|
|
tiku
|
||
|
|
tikvien
|
||
|
|
tikām
|
||
|
|
tikāt
|
||
|
|
tikšu
|
||
|
|
tomēr
|
||
|
|
topat
|
||
|
|
turpretim
|
||
|
|
turpretī
|
||
|
|
tā
|
||
|
|
tādēļ
|
||
|
|
tālab
|
||
|
|
tāpēc
|
||
|
|
un
|
||
|
|
uz
|
||
|
|
vai
|
||
|
|
var
|
||
|
|
varat
|
||
|
|
varēja
|
||
|
|
varēji
|
||
|
|
varēju
|
||
|
|
varējām
|
||
|
|
varējāt
|
||
|
|
varēs
|
||
|
|
varēsi
|
||
|
|
varēsiet
|
||
|
|
varēsim
|
||
|
|
varēt
|
||
|
|
varēšu
|
||
|
|
vien
|
||
|
|
virs
|
||
|
|
virspus
|
||
|
|
vis
|
||
|
|
viņpus
|
||
|
|
zem
|
||
|
|
ārpus
|
||
|
|
šaipus
|
||
|
|
""".split()
|
||
|
|
)
|