def test_load(sample):
parser = UEMParser()
timelines = parser.read(sample)
timeline1 = timelines(uri="uri1")
assert list(timeline1) == [Segment(1, 3.5),
Segment(3, 7.5),
Segment(6, 9)]
class EsterSpeakerDiarizationProtocol(SpeakerDiarizationProtocol):
"""Base speaker diarization protocol for ESTER database"""
def __init__(self, preprocessors={}, **kwargs):
super(EsterSpeakerDiarizationProtocol,
self).__init__(preprocessors=preprocessors, **kwargs)
self.uem_parser_ = UEMParser()
self.mdtm_parser_ = MDTMParser()
def _subset(self, protocol, subset):
data_dir = op.join(op.dirname(op.realpath(__file__)), 'data')
# load annotated parts
# e.g. /data/{tv|radio|all}.{train|dev|test}.uem
path = op.join(
data_dir, '{protocol}.{subset}.uem'.format(subset=subset,
protocol=protocol))
uems = self.uem_parser_.read(path)
# load annotations
path = op.join(
data_dir, '{protocol}.{subset}.mdtm'.format(subset=subset,
protocol=protocol))
mdtms = self.mdtm_parser_.read(path)
for uri in sorted(uems.uris):
annotated = uems(uri)
annotation = mdtms(uri)
current_file = {
'database': 'Ester',
'uri': uri,
'annotated': annotated,
'annotation': annotation
}
yield current_file
class EtapeSpeakerDiarizationProtocol(SpeakerDiarizationProtocol):
"""Base speaker diarization protocol for ETAPE database
This class should be inherited from, not used directly.
Parameters
----------
preprocessors : dict or (key, preprocessor) iterable
When provided, each protocol item (dictionary) are preprocessed, such
that item[key] = preprocessor(**item). In case 'preprocessor' is not
callable, it should be a string containing placeholder for item keys
(e.g. {'wav': '/path/to/{uri}.wav'})
"""
def __init__(self, preprocessors={}, **kwargs):
super(EtapeSpeakerDiarizationProtocol,
self).__init__(preprocessors=preprocessors, **kwargs)
self.uem_parser_ = UEMParser()
self.mdtm_parser_ = MDTMParser()
def _subset(self, protocol, subset):
data_dir = op.join(op.dirname(op.realpath(__file__)), 'data')
# load annotated parts
# e.g. /data/{tv|radio|all}.{train|dev|test}.uem
path = op.join(
data_dir, '{protocol}.{subset}.uem'.format(subset=subset,
protocol=protocol))
uems = self.uem_parser_.read(path)
# load annotations
path = op.join(
data_dir, '{protocol}.{subset}.mdtm'.format(subset=subset,
protocol=protocol))
mdtms = self.mdtm_parser_.read(path)
for uri in sorted(uems.uris):
annotated = uems(uri)
annotation = mdtms(uri)
current_file = {
'database': 'Etape',
'uri': uri,
'annotated': annotated,
'annotation': annotation
}
yield current_file