import nlpio
if __name__ == '__main__':
#print nlpio.parseDocFile('data/duc2004/docs/d30001t/APW19981016.0240')
documents = nlpio.loadDocumentsFromFile('testset.txt')
predictions = [doc.peers[0] for doc in documents] #using the given predictions
print nlpio.evaluateRouge(documents,predictions)
predictions = ['lorem ipsum dolor' for doc in documents] #using own predictions
print nlpio.evaluateRouge(documents,predictions)
#TODO: make better
def __init__(self):
pass
def fit(self,documents,y=None):
return self
def transform(self,documents):
for doc in documents:
print 'Parsing: ' + doc.name
doc.ext['coreNLP'] = nlpio.stanfordParse(doc.text)
return documents
if __name__ == '__main__':
docs = nlpio.loadDocumentsFromFile(
'testset.txt',
# 'testset_all.txt',
'data/eval/models/1/', 'data/eval/peers/1/')
print 'Do Cleanup...'
textCleaner = SimpleTextCleaner()
textCleaner.transform(docs)
print 'Writing file and start coreNLP processing...'
# Writing out the individual files.
for doc in docs:
write_file('tmp/' + doc.name, doc.text)
# Writing the filelist
write_file('tmp/filelist.txt', '\n'.join(
['tmp/' + doc.name for doc in docs]))
