def test_align_basic_errors(basic_corpus_dir, large_dataset_dictionary,
generated_dir, temp_dir):
args = DummyArgs()
args.quiet = True
args.clean = True
args.acoustic_model_path = 'english'
args.corpus_directory = basic_corpus_dir
args.dictionary_path = large_dataset_dictionary
args.output_directory = os.path.join(generated_dir, 'basic_output')
args.temp_directory = temp_dir
run_align_corpus(args)
def test_nnet_use_model(basic_corpus_dir, generated_dir,
large_dataset_dictionary, temp_dir):
args = DummyArgs()
args.debug = True
args.clean = True
args.acoustic_model_path = os.path.join(generated_dir,
'nnet_test_model.zip')
args.corpus_directory = basic_corpus_dir
args.dictionary_path = large_dataset_dictionary
args.output_directory = os.path.join(generated_dir, 'nnet_basic_output')
args.temp_directory = temp_dir
run_align_corpus(args)
def test_align_large_textgrid(large_textgrid_format_directory,
textgrid_output_directory,
large_dataset_dictionary, temp_dir,
basic_align_config, skip_large):
if skip_large:
pytest.skip('Large testsets disabled')
args = DummyArgs()
args.clean = True
args.acoustic_model_path = 'english'
args.corpus_directory = large_textgrid_format_directory
args.output_directory = textgrid_output_directory
args.dictionary_path = large_dataset_dictionary
args.temp_directory = temp_dir
args.config_path = basic_align_config
run_align_corpus(args)
def test_align_basic(basic_corpus_dir, sick_dict_path, generated_dir,
large_dataset_dictionary, temp_dir, basic_align_config,
english_acoustic_model):
args = DummyArgs()
args.acoustic_model_path = 'english'
args.corpus_directory = basic_corpus_dir
args.dictionary_path = sick_dict_path
args.output_directory = os.path.join(generated_dir, 'basic_output')
args.quiet = True
args.clean = True
args.temp_directory = temp_dir
args.config_path = basic_align_config
with pytest.raises(PronunciationAcousticMismatchError):
run_align_corpus(args)
args.acoustic_model_path = 'english'
args.corpus_directory = basic_corpus_dir
args.dictionary_path = large_dataset_dictionary
args.output_directory = os.path.join(generated_dir, 'basic_output')
run_align_corpus(args)
download_parser = subparsers.add_parser('download')
download_parser.add_argument(
"model_type",
help="Type of model to download, one of 'acoustic', 'g2p', or 'dictionary'"
)
download_parser.add_argument(
"language",
help="Name of language code to download, if not specified, "
"will list all available languages",
nargs='?')
if __name__ == '__main__':
mp.freeze_support()
args = parser.parse_args()
fix_path()
if args.subcommand == 'align':
run_align_corpus(args, acoustic_languages)
elif args.subcommand == 'train':
run_train_corpus(args)
elif args.subcommand == 'g2p':
run_g2p(args, g2p_languages)
elif args.subcommand == 'train_g2p':
run_train_g2p(args)
elif args.subcommand == 'validate':
run_validate_corpus(args)
elif args.subcommand == 'download':
run_download(args)
unfix_path()