import corpus2
corpus = corpus2.Corpus()
tagset = corpus2.get_named_tagset('nkjp')
cr = corpus2.CorpusReader(tagset, 'document')
crfile = '/home/rk/tmp/corpus_file.txt'
readed_corp = cr.read(crfile)
while True:
  doc = readed_corp.next_document()
  if not doc:
    break
  print doc.path()