explosion/spaCy
1
1
Fork 0
mirror of https://github.com/explosion/spaCy.git synced 2024-11-13 05:07:03 +03:00
Code Issues Packages Projects Releases Wiki Activity

Merge pull request #3539 from jeannefukumaru/master

Browse Source 
Added tags previously missing from Indonesian `tag_map.py`
This commit is contained in:
Ines Montani 2019-04-04 11:57:03 +02:00 committed by GitHub
commit cd21778bef
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
1 changed files with 85 additions and 26 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

83
spacy/lang/id/tag_map.py
View File

@ -2,28 +2,28 @@
from __future__ import unicode_literals from __future__ import unicode_literals
from ...symbols import POS, PUNCT, ADJ, CCONJ, NUM, DET, ADV, ADP, X, VERB from ...symbols import POS, PUNCT, ADJ, CCONJ, NUM, DET, ADV, ADP, X, VERB
from ...symbols import NOUN, PRON, AUX, SCONJ from ...symbols import NOUN, PRON, AUX, SCONJ, INTJ, PART, PROPN
# POS explanations for indonesian available from https://www.aclweb.org/anthology/Y12-1014 # POS explanations for indonesian available from https://www.aclweb.org/anthology/Y12-1014
TAG_MAP = { TAG_MAP = {
"NSD": {POS: NOUN}, "NSD": {POS: NOUN},
"Z–": {POS: PUNCT}, "Z–-": {POS: PUNCT},
"VSA": {POS: VERB}, "VSA": {POS: VERB},
"CC-": {POS: NUM}, "CC-": {POS: NUM},
"R–": {POS: ADP}, "R–-": {POS: ADP},
"D–": {POS: ADV}, "D–-": {POS: ADV},
"ASP": {POS: ADJ}, "ASP": {POS: ADJ},
"S–": {POS: SCONJ}, "S–-": {POS: SCONJ},
"VSP": {POS: VERB}, "VSP": {POS: VERB},
"H–": {POS: CCONJ}, "H–-": {POS: CCONJ},
"F–": {POS: X}, "F–-": {POS: X},
"B–": {POS: DET}, "B–-": {POS: DET},
"CO-": {POS: NUM}, "CO-": {POS: NUM},
"G–": {POS: ADV}, "G--": {POS: ADV},
"PS3": {POS: PRON}, "PS3": {POS: PRON},
"W–": {POS: ADV}, "W--": {POS: ADV},
"O–": {POS: AUX}, "O-–": {POS: AUX},
"PP1": {POS: PRON}, "PP1": {POS: PRON},
"ASS": {POS: ADJ}, "ASS": {POS: ADJ},
"PS1": {POS: PRON}, "PS1": {POS: PRON},
@ -31,4 +31,63 @@ TAG_MAP = {
"CD-": {POS: NUM}, "CD-": {POS: NUM},
"VPA": {POS: VERB}, "VPA": {POS: VERB},
"VPP": {POS: VERB}, "VPP": {POS: VERB},
} "X—-": {POS: X},
"CO-+PS3": {POS: NUM},
"NSD+PS3": {POS: NOUN},
"ASP+PS3": {POS: ADJ},
"M—-": {POS: AUX},
"VSA+PS3": {POS: VERB},
"R--+PS3": {POS: ADP},
"W--+T--": {POS: ADV},
"PS2": {POS:PRON},
"NSD+PS1": {POS:NOUN},
"PP3": {POS: PRON},
"VSA+T--": {POS: VERB},
"D--+T--": {POS: ADV},
"VSP+PS3": {POS: VERB},
"F--+PS3": {POS: X},
"M--+T—": {POS: AUX},
"F--+T—": {POS: X},
"PUNCT": {POS: PUNCT},
"PROPN": {POS: PROPN},
"I—-": {POS: INTJ},
"S--+PS3": {POS: SCONJ},
"ASP+T—": {POS: ADJ},
"CC-+PS3": {POS: NUM},
"NSD+PS2": {POS: NOUN},
"B--+T—": {POS: DET},
"H--+T--": {POS: CCONJ},
"VSA+PS2": {POS: VERB},
"NSF": {POS: NOUN},
"PS1+VSA": {POS: PRON},
"NPD": {POS: NOUN},
"PP2": {POS:PRON},
"VSA+PS1": {POS: VERB},
"T—": {POS: PART},
"NSM": {POS: NOUN},
"NUM": {POS: NUM},
"ASP+PS2": {POS: ADJ},
"G--+T—": {POS: PART},
"D--+PS3": {POS: ADV},
"R--+PS2": {POS: ADP},
"NSM+PS3": {POS: NOUN},
"VSP+T—": {POS: VERB},
"M--+PS3": {POS: AUX},
"ASS+PS3": {POS: ADJ},
"G--+PS3": {POS: PART},
"F--+PS1": {POS: X},
"NSD+T--": {POS: NOUN},
"PP1+T--": {POS: PRON},
"B--+PS3": {POS: DET},
"NOUN": {POS: NOUN},
"NPD+PS3": {POS: NOUN},
"R--+PS1": {POS: ADP},
"F--+PS2": {POS: X},
"CD-+PS3": {POS: NUM},
"PS1+VSA+T—": {POS: VERB},
"PS2+VSA": {POS: VERB},
"VERB": {POS: VERB},
"CC-+T—": {POS: NUM},
"NPD+PS2":{POS: NOUN},
"D--+PS2":{POS: ADV},
"PP3+T—": {POS: PRON}}