-
Notifications
You must be signed in to change notification settings - Fork 3
/
wordpict_relmap_ud_fi.tsv
We can make this file beautiful and searchable if this error is corrected: Illegal quoting in line 23.
95 lines (91 loc) · 2.12 KB
/
wordpict_relmap_ud_fi.tsv
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
# Map the Finnish UD1 and UD2 dependency relation labels (as produced
# by the Turku Neural Parser Pipeline (TNPP)) to those used in the
# Korp word picture: SS (subject), OBJ (object), ADV (adverbial), PA
# (preposition), AT ((pre)modifier), ET (postmodifier), CPL
# (complement). Additional, alternative labels, currently commented
# out and converted to other labels or ignored: APP (apposition), AUX
# (auxiliary verb), CPR (compared element), CRD (coordinated element),
# CPS (copular subject), OTH (other), MOD (modifier), OWN (owner), VPT
# (verb particle). Relations to be ignored are marked with a dash.
#
# In addition, the mapping contains in comments a few tentative
# relation transformations: constructs for specifying derived
# relations that do not occur directly in the data. They are not (yet)
# supported by vrt-extract-relations.py.
# For a name relation, add an extra head containing all the words from
# dependent to head:
# &name h = d..h
_ -
acl -
acl:relcl -
advcl ADV
# ADV if head POS is V or Adv; otherwise MOD
# advmod ADV: h.pos~"V|Adv"; MOD
# Could be converted to MOD (or ET) or maybe omitted for other than
# verbs and adverbs
advmod ADV
amod AT
# appos APP
appos -
# aux AUX
aux -
auxpass -
aux:pass -
# Could the following be PA, or maybe inverted PA?
case -
cc -
# Could be converted to ET for nouns
ccomp CPL
cc:preconj -
compound AT
compound:nn AT
compound:prt -
conj -
cop -
cop:own -
csubj SS
# csubj:cop CPS
csubj:cop -
# dep OTH
dep -
det -
discourse -
dobj OBJ
# Does it make sense to map fixed and flat(:*) (MWEs) to ET, or would
# it be better not to map them to anything?
fixed ET
flat ET
flat:foreign ET
flat:name ET
goeswith AT
mark -
mwe ET
# AT; inverse relation (dependent to head) ET; head covers all words
# from the dependent to head:
# name AT; ^ET; @h = d..h
name AT
neg -
nmod ADV
nmod:gobj AT
nmod:gsubj AT
# Add an OWN relation from the nsubj dependent of head to the
# dependent of nommod-own
# nmod:own OWN: -> h>nsubj d
nmod:own -
nmod:poss AT
nsubj SS
# nsubj:cop CPS
nsubj:cop -
nummod -
obj OBJ
obl ADV
orphan -
# parataxis CRD
parataxis -
punct -
remnant -
rel -
root -
vocative -
xcomp CPL
xcomp:ds CPL