doc.add(Field("answer", ii['Answer'], t1))
doc.add(Field("qid", ii['Question ID'], t1))
doc.add(Field("category", ii['category'], t1))
doc.add(Field("position", ii['Sentence Position'], t1))
doc.add(Field("question", ii['Question Text'], t2))
doc.add(Field("wiki_plain",
self.wiki_reader.get_text(ii['Answer']), t2))
writer.addDocument(doc)
if __name__ == '__main__':
if len(sys.argv) < 2:
print IndexDocs.__doc__
sys.exit(1)
lucene.initVM(vmargs=['-Djava.awt.headless=true'])
print 'lucene', lucene.VERSION
start = datetime.now()
try:
train_path = sys.argv[1]
train_set = Corpus()
train_set.read(train_path)
train_bench, test_bench = train_set.train_test_split()
base_dir = os.path.dirname(os.path.abspath(sys.argv[0]))
IndexDocs(train_bench, os.path.join(base_dir, INDEX_DIR),
StandardAnalyzer(Version.LUCENE_CURRENT))
end = datetime.now()
print end - start
except Exception, e:
print "Failed: ", e
raise e
