Error when getting the baselines with sequence_labeling

[jerbarnes]

Error when getting the baselines with sequence_labeling

[jerbarnes]

Hi Iago,

Looks like this error was due to an inconsistency in the the polarity labels. The baseline code expects the polarity label to be title case ('negative' -> 'Negative'). This error was also present in MPQA. I've corrected this now in the preprocessing, but you'll have to rerun the preprocessing commands:

cd data/mpqa
bash process_mpqa.sh
cd ../darmstadt_unis
bash process_darmstadt.sh

After that, you should be able to run get_baselines.sh without a problem. Let me know if it works on your end.

[MinionAttack]

Hi,

MPQA works fine but I'm having problems with darmstadt when reading the universities.zip.

extracting: universities/basedata/DeVry_University_27_12-08-2007.txt  
  inflating: universities/basedata/DeVry_University_27_12-08-2007_words.xml  
  inflating: universities/basedata/DeVry_University_29_11-15-2007.txt  
  inflating: universities/basedata/DeVry_University_29_11-15-2007_words.xml  
  inflating: universities/basedata/DeVry_University_30_10-26-2007.txt  
  ..........
  inflating: universities/markables/University_of_Phoenix_Online_189_07-25-2004_SentenceOpinionAnalysisResult_level.xml  
  inflating: universities/customization/OpinionExpression_customization.xml  
  inflating: universities/customization/SentenceOpinionAnalysisResult_customization.xml  
sed: can't read : No such file or directory

Regards.

[jerbarnes]

Ok, the darmstadt error came about because of a fix for OSX users. I've corrected it now.

[MinionAttack]

Yes, that error has disappeared but it seems to have been covering another error:

Traceback (most recent call last):
  File "process_darmstadt.py", line 475, in <module>
    o = get_opinions(bfile, mfile)
  File "process_darmstadt.py", line 113, in get_opinions
    text += token + " "
TypeError: unsupported operand type(s) for +: 'NoneType' and 'str'

I don't know if it's related with the data, I've printed the idx and token variable and before the error it outputs:

word_128
I
word_129
also
word_130
was
word_131
duped
word_132
with
word_133
my
word_134
financial
word_135
aid
word_136
None

[jerbarnes]

Sorry, there was a bug in the sed command :/ How about now?

[MinionAttack]

Now it works! Thanks.