if corpus_type == "semeval":
instances, labels = reader.get_instances(sys.argv[1], sys.argv[6])
else:
instances, labels = reader.get_instances(sys.argv[1])
corpus = Corpus(instances, labels)
pos_lexicon = sys.argv[3]
neg_lexicon = sys.argv[4]
feature_csv = sys.argv[5]
# generate feature csv based on model
if sys.argv[2] == "Majority":
pass
elif sys.argv[2] in ["OCC", "OCCRB"]:
for inst in corpus.get_instances():
inst.generate_deps()
new_feature_csv = FeatureGeneratorOCC(
corpus.get_instances()).generate_feature_csv(feature_csv, pos_lexicon,
neg_lexicon)
elif sys.argv[2] == "OCCSRL":
new_feature_csv = FeatureGeneratorSRL(
corpus.get_instances()).generate_feature_csv(feature_csv, pos_lexicon,
neg_lexicon)
elif sys.argv[2] == "DD":
new_feature_csv, postag_set = FeatureGeneratorDD(
corpus.get_instances()).generate_feature_csv(feature_csv, pos_lexicon,
neg_lexicon)
elif sys.argv[2] == "COMB":
