def get_data(meetingNum=1, **kwargs): ''' Load data for specified single sequence. Args ---- meetingNum : int Identifies which sequence out of the 21 possible to use. Must be valid number in range [1,2,3, ... 21]. Returns ------- Data : GroupXData holding only the data for a single sequence. ''' if meetingNum <= 0 or meetingNum > len(fileNames): raise ValueError('Bad value for meetingNum: %s' % (meetingNum)) fName = fileNames[meetingNum - 1].replace(suffix, '') matfilepath = os.path.join(datasetdir, 'rawData', 'speakerDiarizationData', fName) if not os.path.isfile(matfilepath): raise ValueError('CANNOT FIND SPEAKDIAR DATASET MAT FILE:\n' + matfilepath) Data = GroupXData.read_from_mat(matfilepath) Data.summary = \ 'Pre-processed audio data from NIST file %s (meeting %d / 21)' \ % (fName.replace(suffix, ''), meetingNum) Data.name = 'SpeakerDiar' + str(meetingNum) Data.fileNames = [fName] return Data
def get_data(**kwargs): Data = GroupXData.read_from_mat(matfilepath) Data.summary = get_data_info() Data.name = get_short_name() # Verify that true state space is indexed starting at 0, not 1 # Violating this can cause bugs in the alignment code assert Data.TrueParams['Z'].min() == 0 assert Data.TrueParams['Z'].max() == 11 return Data
def get_data(**kwargs): Data = GroupXData.read_from_mat(matfilepath) Data.summary = get_data_info() Data.name = get_short_name() return Data