def __getitem__()

in sagemaker-voice-classification/notebook/coswara_dataset.py [0:0]

• 27 lines of code
• 6 McCabe index (conditional complexity)

    def __getitem__(self, index):
        self.curfile = self.file_names[index]
        try:
            waveform, sample_rate = torchaudio.load(self.curfile)
            waveform = torchaudio.transforms.Resample(sample_rate, self.new_sr)(waveform[self.channel, :].view(1, -1))

            # Zero padding to keep desired audio length in seconds
            const_len = self.new_sr * self.audio_len
            tempData = torch.zeros([1, const_len])
            if waveform.shape[1] < const_len:
                tempData[0, : waveform.shape[1]] = waveform[:]
            else:
                tempData[0, :] = waveform[0, :const_len]
            sound = tempData

            # Resampling, need to cater for audio shorter than new length
            const_len = self.new_sr * self.audio_len
            tempData = torch.zeros([1, const_len])
            if sound.shape[1] < const_len:
                tempData[0, : sound.shape[1]] = sound[:]
            else:
                tempData[0, :] = sound[0, :const_len]
            sound = tempData
            # Resampling
            new_const_len = const_len // self.sampling_ratio
            soundFormatted = torch.zeros([1, new_const_len])
            soundFormatted[0, :] = sound[0, ::5]
            
            return soundFormatted, self.labels[index]
        
        except RuntimeError:
            print("Get Item RuntimeError for file: {}".format(self.curfile))
        except ExecuteUserScriptError: 
            print("Get Item ExecuteUserScriptError for file: {}".format(self.curfile))