Hi,
I have run the following commands, but I get an error in the last one:
Command for training: mwetoolkit/bin/train_crf.py mwetoolkit/EL/train.cupt
Command for applying the tagger to the trial blind test file: mwetoolkit/bin/annotate_crf.py mwetoolkit/EL/test.blind.cupt > mwetoolkit/EL/output_crf.cupt
Command for evaluation: python mwetoolkit/bin/evaluate.py --pred mwetoolkit/EL/output_crf.cupt --gold mwetoolkit/EL/test.cupt --train mwetoolkit/EL/train.cupt
The problem is that when I am running the last command I am getting the following error:
Traceback (most recent call last):
File "mwetoolkit/bin/evaluate.py", line 8, in <module>
import tsvlib
ImportError: No module named 'tsvlib'
However, I have downloaded the TSV-to-FoLiA tool from here: https://github.com/proycon/parseme-support , but the error still says that the tsvlibrary is not there. This happens when I put this "parseme-support" file outside the bin.
I also tried to put the tsvlib.py and the tsvtofolia.py alongside the evaluation script, and the error is the following:
===> ERROR when reading mwetoolkit/EL/train.cupt (line 135803)
Traceback (most recent call last):
File "mwetoolkit/bin/evaluate.py", line 491, in <module>
Main(parser.parse_args()).run()
File "mwetoolkit/bin/evaluate.py", line 67, in run
seen = SeenInfo(self.args.train_file)
File "mwetoolkit/bin/evaluate.py", line 277, in __init__
self.mwe_fieldindex_sets[field_name] = self._calc_mwe_fieldindex_sets(field_name, sents)
File "mwetoolkit/bin/evaluate.py", line 288, in _calc_mwe_fieldindex_sets
return set(fieldindex_set for sentence in sents for fieldindex_set
File "mwetoolkit/bin/evaluate.py", line 289, in <genexpr>
in sentence.iter_mwe_fields_and_normalizedindexes(field_name))
AttributeError: 'TSVSentence' object has no attribute 'iter_mwe_fields_and_normalizedindexes'
Could you please help me figure out what is going wrong?
Thanks in advance,
Rebecca Anagnostopoulou