Hi, cdec users,
I hope this group is still active.
I've just started using cdec and only tried to reproduce an MT system building as written in tutorial (
http://www.cdec-decoder.org/guide/tutorial.html).
On step 6 - grammar extraction for dev and test sets - the process stops due to an error "Illegal key type <type 'long'> for IntList".
I type in this command:
python -m cdec.sa.extract -c extract.ini -g dev.grammars -j 1 < dev.lc-tok.es-en > dev.lc-tok.es-en.sgm
and get this output:
INFO:cdec.sa:Sampling strategy: uniform, max sample size = 300
INFO:cdec.sa:Reading precomputed data from file /home/varvara/workspace/cdec-examples/cdec-spanish-demo/
training.sa/precomp.5.2.15.1.100.10.bin...
INFO:cdec.sa:Converting 100 hash keys on precomputed inverted index...
INFO:cdec.sa:Converting 23000 hash keys on precomputed collocations...
INFO:cdec.sa:Processing precomputations took 0.332021 seconds
Traceback (most recent call last):
File "/usr/lib/python2.7/runpy.py", line 162, in _run_module_as_main
"__main__", fname, loader, pkg_name)
File "/usr/lib/python2.7/runpy.py", line 72, in _run_code
exec code in run_globals
File "/usr/local/lib/python2.7/dist-packages/cdec/sa/extract.py", line 113, in <module>
main()
File "/usr/local/lib/python2.7/dist-packages/cdec/sa/extract.py", line 106, in main
for output in map(extract, enumerate(sys.stdin)):
File "/usr/local/lib/python2.7/dist-packages/cdec/sa/extract.py", line 57, in extract
for rule in extractor.grammar(sentence):
File "rulefactory.pxi", line 1133, in input (src/sa/_sa.c:49363)
File "rulefactory.pxi", line 1681, in _sa.HieroCachingRuleFactory.extract (src/sa/_sa.c:55643)
File "rulefactory.pxi", line 1468, in _sa.HieroCachingRuleFactory.create_alignments (src/sa/_sa.c:53589)
File "int_list.pxi", line 109, in _sa.IntList.__getitem__ (src/sa/_sa.c:6308)
TypeError: Illegal key type <type 'long'> for IntList
What causes this error? How can I fix it?
Btw I saw just the same error reported:
https://github.com/redpony/cdec/issues/13but it got no answer.
Thank you,
Varvara.