Secure your code as it's written. Use Snyk Code to scan source code in minutes - no build needed - and fix issues immediately.
(options, args) = parser.parse_args()
if options.verbose:
logging.basicConfig(level=logging.DEBUG)
else:
logging.basicConfig(level=logging.INFO)
if len(args) != 1:
parser.print_usage()
sys.exit(1)
#
# load config, set up global variables
#
config = misc.load_config ('.speechrc')
#
# convert CMU dict to zamia-speech IPA format
#
lex_new = {}
with codecs.open(args[0], 'r', 'utf8') as dictf:
for line in dictf:
if '#' in line:
logging.debug('comment : %s' % line)
line = line.split('#')[0]
logging.debug('comment removed: %s' % line)
parser.add_option ("-v", "--verbose", action="store_true", dest="verbose",
help="verbose output")
(options, args) = parser.parse_args()
if options.verbose:
logging.basicConfig(level=logging.DEBUG)
else:
logging.basicConfig(level=logging.INFO)
#
# config
#
config = misc.load_config('.speechrc')
speech_arc = config.get("speech", "speech_arc")
speech_corpora = config.get("speech", "speech_corpora")
srcdir = '%s/m_ailabs' % speech_arc
#
# audio, prompts
#
all_utts = set()
cnt = 0
with open('tmp/run_parallel.sh', 'w') as scriptf:
for localedir in os.listdir(srcdir):
WORKDIR_CONT = 'data/dst/asr-models/cmusphinx_cont/%s'
WORKDIR_PTM = 'data/dst/asr-models/cmusphinx_ptm/%s'
NJOBS = 12
ENABLE_NOISE_FILLER = False # CMU Sphinx decoding seems to become unstable otherwise
NOISE_WORD = 'nspc'
#
# init
#
misc.init_app ('speech_sphinx_export')
config = misc.load_config ('.speechrc')
#
# commandline parsing
#
parser = OptionParser("usage: %prog [options] model_name dict lm corpus [corpus2 ...]")
parser.add_option ("-d", "--debug", dest="debug", type='int', default=0,
help="limit number of transcripts (debug purposes only), default: 0 (unlimited)")
parser.add_option ("-v", "--verbose", action="store_true", dest="verbose",
help="enable verbose logging")
(options, args) = parser.parse_args()
if options.verbose:
logging.basicConfig(level=logging.DEBUG)
# corpus_name = 'tedlium3'
# corpus_name = 'm_ailabs_en'
# corpus_name = 'voxforge_en'
if options.verbose:
logging.basicConfig(level=logging.DEBUG)
logging.getLogger("requests").setLevel(logging.WARNING)
else:
logging.basicConfig(level=logging.INFO)
#
# config
#
config = misc.load_config('.speechrc')
wav16_dir = config.get("speech", "wav16")
#
# load transcripts
#
logging.info("loading transcripts...")
transcripts = Transcripts(corpus_name=corpus_name)
logging.info("loading transcripts...done.")
#
# compute stats
#
def format_duration(duration):
parser.add_option ("-v", "--verbose", action="store_true", dest="verbose",
help="verbose output")
(options, args) = parser.parse_args()
if options.verbose:
logging.basicConfig(level=logging.DEBUG)
else:
logging.basicConfig(level=logging.INFO)
#
# config
#
config = misc.load_config('.speechrc')
speech_arc = config.get("speech", "speech_arc")
speech_corpora = config.get("speech", "speech_corpora")
srcdir = '%s/LibriSpeech' % speech_arc
destdir = '%s/librispeech' % speech_corpora
misc.mkdirs(destdir)
#
# speakers
#
with open ('spk2gender.txt', 'w') as genderf:
with open ('%s/SPEAKERS.TXT' % srcdir, 'r') as speakersf:
parser.add_option ("-v", "--verbose", action="store_true", dest="verbose",
help="verbose output")
(options, args) = parser.parse_args()
if options.verbose:
logging.basicConfig(level=logging.DEBUG)
else:
logging.basicConfig(level=logging.INFO)
#
# config
#
config = misc.load_config('.speechrc')
speech_arc = config.get("speech", "speech_arc")
speech_corpora = config.get("speech", "speech_corpora")
srcdir = '%s/LJSpeech-1.1' % speech_arc
destdir = '%s/ljspeech' % speech_corpora
#
# audio, prompts
#
all_utts = set()
cnt = 0
with open('tmp/run_parallel.sh', 'w') as scriptf:
parser.add_option("-v", "--verbose", action="store_true", dest="verbose",
help="enable debug output")
(options, args) = parser.parse_args()
if options.verbose:
logging.basicConfig(level=logging.DEBUG)
logging.getLogger("requests").setLevel(logging.WARNING)
else:
logging.basicConfig(level=logging.INFO)
#
# config
#
config = misc.load_config('.airc')
ai_model = config.get ('server', 'model')
lang = config.get ('server', 'lang')
vf_login = config.get ('server', 'vf_login')
rec_dir = config.get ('server', 'rec_dir')
kaldi_model_dir = config.get ('server', 'kaldi_model_dir')
kaldi_model = config.get ('server', 'kaldi_model')
loc = config.get ('vad', 'loc')
source = config.get ('vad', 'source')
volume = config.getint('vad', 'volume')
aggressiveness = config.getint('vad', 'aggressiveness')
#
# curses
#
from kaldiasr.nnet3 import KaldiNNet3OnlineModel, KaldiNNet3OnlineDecoder
DEFAULT_ASR_MODEL = 'kaldi-chain-voxforge-de-latest'
MODELDIR = 'data/models/%s'
SAVE_RATE = 10
FAILLOG = 'tmp/decoding_fails.txt'
#
# init
#
misc.init_app ('auto_review')
config = misc.load_config ('.speechrc')
#
# command line
#
parser = OptionParser("usage: %prog [options] corpus")
parser.add_option ("-a", "--all", action="store_true", dest="do_all",
help="do not use ASR but auto-rate all matching submissions")
parser.add_option ("-f", "--filter", dest="ts_filter", type = "str",
help="filter (default: no filtering)")
parser.add_option ("-l", "--lang", dest="lang", type = "str", default="de",
help="tokenizer language (default: de)")
def __init__(self):
cmdln.Cmdln.__init__(self)
self.config = misc.load_config('.airc')
toplevel = self.config.get('semantics', 'toplevel')
xsb_root = self.config.get('semantics', 'xsb_root')
db_url = self.config.get('db', 'url')
self.kernal = AIKernal(db_url, xsb_root, toplevel)