def __init__(self, fname, dataset=None, **kwargs):
kwargs.setdefault('input_parser', audio.raw)
if kwargs.get('input_parser') is None:
raise ValueError("input_parser must be set")
if kwargs.get('label_parser') is None:
raise ValueError("label_parser must be set")
with codecs.open(fname, 'r', encoding='utf8') as f:
ld = json.load(f)
data = utils.ld2dl(ld)
if dataset:
inputs = np.array(
[d['audio'] for d in data if d['dataset'] == dataset])
labels = np.array(
[d['label'] for d in data if d['dataset'] == dataset])
else:
inputs = np.array(data['audio'])
labels = np.array(data['label'])
super(JSONIterator, self).__init__(inputs, labels, **kwargs)
self.durations = np.array(data['duration'])
def __init__(self, fname, dataset=None, **kwargs):
self._logger = logging.getLogger('%s.%s' % (__name__,
self.__class__.__name__))
kwargs.setdefault('input_parser', audio.raw)
if kwargs.get('input_parser') is None:
raise ValueError("input_parser must be set")
if kwargs.get('label_parser') is None:
raise ValueError("label_parser must be set")
with codecs.open(fname, 'r', encoding='utf8') as f:
ld = json.load(f)
data = utils.ld2dl(ld)
if dataset and 'dataset' not in data:
self._logger.warning('No dataset key found. Falling back to None')
dataset = None
if dataset:
inputs = np.array([i for i, d in zip(
data['input'], data['dataset']) if d == dataset])
labels = np.array([l for l, d in zip(
data['label'], data['dataset']) if d == dataset])
else:
inputs = np.array(data['input'])
labels = np.array(data['label'])
super(JSONIterator, self).__init__(inputs, labels, **kwargs)
self.durations = np.array(data['duration'])
def to_json(self, fname=None):
''' Parse the entire dataset to a list of dictionary containin at least
two keys:
`input`: path to audio file
`duration`: length of the audio
`label`: transcription of the audio
'''
fname = fname or os.path.join(self.default_output_dir, 'data.json')
if os.path.exists(fname) and override:
os.remove(fname)
if not os.path.isdir(os.path.split(fname)[0]):
safe_mkdirs(os.path.split(fname)[0])
data = self._to_ld()
with codecs.open(fname, 'w', encoding='utf8') as f:
json.dump(data, f)
self._logger.info(self._report(ld2dl(data)))
return fname