def test_datatrainer(self): datatrainer = DataTrainer() datatrainer.create() self.assertEqual( datatrainer.check(), None ) dire = datatrainer.workdir self.assertTrue( os.path.exists( dire ) ) datatrainer.delete() self.assertFalse( os.path.exists( dire ) )
setup_logging(None,None) # --------------------------------- # 1. Create a Data Manager # it manages the data created at each step of the acm training procedure # create parameters: # - workdir=None (in) # - scriptsdir=DEFAULT_SCRIPTS_DIR (in) # - featsdir=DEFAULT_FEATURES_DIR (in) # - logdir=DEFAULT_LOG_DIR (in) # - protodir=None (in) # - protofilename=DEFAULT_PROTO_FILENAME (out) datatrainer = DataTrainer() datatrainer.create( workdir=args.t, protodir=args.p ) # --------------------------------- # 2. Create a Corpus Manager # it manages the set of training data: # - establishes the list of phonemes (from the dict); # - converts the input annotated data into the HTK-specific data format; # - codes the audio data. corpus = TrainingCorpus( datatrainer, lang=args.l ) corpus.fix_resources( dictfile=args.r, mappingfile=args.m ) if args.i: for entry in args.i: if os.path.isdir( entry ):