arne-cl
commented
3 years ago Problem group2tree expects assert len(children['nucleus']) == 1, but we have:
>>> children
defaultdict(<type 'list'>, {'satellite': ['99'], 'nucleus': ['95', '97']})Open arne-cl opened 3 years ago
arne-cl
commented
3 years ago Problem group2tree expects assert len(children['nucleus']) == 1, but we have:
>>> children
defaultdict(<type 'list'>, {'satellite': ['99'], 'nucleus': ['95', '97']})minimal input:
Szeryng subsequently focused on teaching before resuming his concert career in 1954.
The "Le Duc" was the instrument on which he performed and recorded mostly, while the latter ("King David" Strad) was donated to the State of Israel.feng-hirst-2014 output:
ParseTree('Elaboration[N][S]', [ParseTree('Temporal[N][S]', ['Szeryng subsequently focused on teaching', 'before resuming his concert career in 1954 .']), ParseTree('Elaboration[N][S]', ['The " Le Duc " was the instrument', ParseTree('Temporal[N][N]', ['on which he performed and recorded mostly ,', ParseTree('same-unit[N][N]', [ParseTree('same-unit[N][N]', [ParseTree('Elaboration[N][S]', ['while the latter', '( " King David "']), 'Strad )']), 'was donated to the State of Israel .'])])])])If we convert the original feng-hirst-2014 parser output to a tree,
$ curl -X POST -F "input=@feng-hirst-2014-result-minimal.fh2014" http://localhost:5000/convert/hilda/tree.prettyprint
Elaboration
_______________________________________________|________________
| S
| |
| Elaboration
| __________________________________|____________________________
| | S
| | |
| | Temporal
| | _____________________________________________|_____________
| | | N
| | | |
| | | same-unit
| | | _____________|____________________
| | | N |
| | | | |
| | | same-unit |
| | | _____________|______________________ |
N | | N | |
| | | | | |
Temporal | | Elaboration | |
____________|____________ | | ______________|_____________ | |
N S N N N S N N
| | | | | | | |
Szeryng subseque before resuming The " Le Duc " on which he while the latter ( " King David " Strad ) was donated to
ntly focused on his concert was the instrume performed and the State of
teaching career in 1954 . nt recorded mostly , Israel .we see that it has two Temporal relations with different nuclearity:
The first one is (N: ... focused on teaching, S: before resuming ...).
The second one is (N: on which he performed ... mostly, N: while the latter ... was donated).
The while should not be interpreted in a temporal sense, but that's probably not the issue here.
arne-cl
commented
3 years ago The rs3 files has some odd things, like relation names in different casing/spelling in the <relations> section,
while in the
Temporal (and not temporal)
<relations>
<rel name="Manner-Means" type="rst"/>
<rel name="Mannermeans" type="rst"/>
...
<rel name="Same-Unit" type="multinuc"/>
<rel name="Same-unit" type="multinuc"/>
<rel name="Temporal" type="multinuc"/>
<rel name="same-unit" type="multinuc"/>
<rel name="same_unit" type="multinuc"/>
<rel name="temporal" type="rst"/>
</relations>
...
<body>
<segment id="5" parent="3" relname="Temporal">Szeryng subsequently focused on teaching</segment>
<segment id="7" parent="3" relname="Temporal">before resuming his concert career in 1954 .</segment>
<segment id="15" parent="13" relname="Temporal">on which he performed and recorded mostly ,</segment>
<group id="17" type="multinuc" parent="13" relname="Temporal"/>
</body>
feng-hirst-2014-result.rs3.txt