lkrsnik/cjvt-srl-tagging
1
0
Fork 0
forked from kristjan/cjvt-srl-tagging
Code Issues Pull Requests Projects Releases Wiki Activity

finished parse + tag toolchain -> TODO: tagger error

Browse Source
This commit is contained in:
voje
2019-02-18 08:49:04 +01:00
parent 9e9e1910a0
commit f251c912e3
20 changed files with 26489 additions and 54 deletions
Download Patch File Download Diff File Expand all files Collapse all files

25
tools/parser/README.md
View File

@@ -2,5 +2,30 @@
Help conversion between english and slovenian MSD.
Hardcoded values from online documentation (html tables).
## Tagging
Go to `./srl-29...` and run `./scripts/{learn...,parse...}`.
Change paths in the scripts.
## ERR
Getting this weird error:
```bash
Executing: java -cp srl.jar:lib/liblinear-1.51-with-deps.jar:lib/anna.jar -Xmx2g se.lth.cs.srl.Parse ger ./../../data/kres_example_out/F0006347.xml.parsed.tsv ./srl-ger.model -nopi ger-eval.out
Loading pipeline from ./srl-ger.model
Writing corpus to ger-eval.out...
Opening reader for ./../../data/kres_example_out/F0006347.xml.parsed.tsv...
Exception in thread "main" java.lang.IndexOutOfBoundsException: Index: 33, Size: 32
at java.util.ArrayList.rangeCheck(ArrayList.java:657)
at java.util.ArrayList.get(ArrayList.java:433)
at se.lth.cs.srl.corpus.Sentence.buildDependencyTree(Sentence.java:61)
at se.lth.cs.srl.corpus.Sentence.newSRLOnlySentence(Sentence.java:182)
at se.lth.cs.srl.io.SRLOnlyCoNLL09Reader.readNextSentence(SRLOnlyCoNLL09Reader.java:23)
at se.lth.cs.srl.io.AbstractCoNLL09Reader.open(AbstractCoNLL09Reader.java:43)
at se.lth.cs.srl.io.AbstractCoNLL09Reader.<init>(AbstractCoNLL09Reader.java:26)
at se.lth.cs.srl.io.SRLOnlyCoNLL09Reader.<init>(SRLOnlyCoNLL09Reader.java:11)
at se.lth.cs.srl.Parse.main(Parse.java:36)
root@9f69d66a0d39:/cjvt-srl-tagging/tools/srl-20131216#
```
## Sources
[1] (conll09 data format) https://nlpado.de/~sebastian/pub/papers/conll09_hajic.pdf

9
tools/parser/parser.py
View File

@@ -34,7 +34,7 @@ class Parser:
guess_corpus = None # SSJ | KRES
res_dict = {}
with open(filepath, "rb") as fp:
with filepath.open("rb") as fp:
# remove namespaces
bstr = fp.read()
@@ -135,8 +135,10 @@ class Parser:
print(sentence_entry["links"])
"""
# 1 3 4 5 6 7 8 9 10 11 12 13 14 15
out_str += "{}\t{}\t{}\t{}\t{}\t{}\t{}\t{}\t{}\t{}\t{}\t{}\t{}\t{}\n".format(
more_underscores = "".join(["\t_" for x in range(9)])
# 1 3 4 5 6 7 8 9 10 11 12 13 14 15 n
out_str += "{}\t{}\t{}\t{}\t{}\t{}\t{}\t{}\t{}\t{}\t{}\t{}\t{}\t{}{}\r\n".format(
t_id, # id
token[2], # form
token[3], # lemma
@@ -151,6 +153,7 @@ class Parser:
sentence_entry["links"][t_id][0], # pdeprel
"Y" if fprd else "_", # fillpred
token[3] if fprd else "_", # pred
more_underscores,
)
out_str += "\n"
# print(out_str)