View Scores NoneType not subscriptable

[smart-patrol]

When running the following for scores like Rouge and others like:

vizseq.view_scores(ref, hypo, ['metric that's not bleu'], tags=tag)

I am getting:

~/.local/lib/python3.6/site-packages/vizseq/scorers/__init__.py in _score_multiprocess_averaged(self, hypothesis, references, tags, sent_score_func)
    170             for t in tag_set:
    171                 indices = [i for i, cur in enumerate(tags) if t in cur]
--> 172                 group_scores[t] = np.mean([sent_scores[i] for i in indices])
    173 
    174         return VizSeqScore.make(

~/.local/lib/python3.6/site-packages/vizseq/scorers/__init__.py in <listcomp>(.0)
    170             for t in tag_set:
    171                 indices = [i for i, cur in enumerate(tags) if t in cur]
--> 172                 group_scores[t] = np.mean([sent_scores[i] for i in indices])
    173 
    174         return VizSeqScore.make(

TypeError: 'NoneType' object is not subscriptable

Any idea why that's going on? Using text data this is at least 3 tokens or more.

This works fine when running view_examples.

[kahne]

@smart-patrol

Sorry for late reply and thank you very much for reporting this issue!

This bug has been fixed in #31 . Please upgrade VizSeq with pip install --upgrade vizseq or pull the latest master branch and reinstall VizSeq with pip install -e <github repo root>.

Thanks again for your support on VizSeq!