utils/diarization_pipeline.py

from simple_diarizer.diarizer import Diarizer


class DiarizationPipeline:
    def __init__(self, ):
        super(DiarizationPipeline, self).__init__()
        self.diar = Diarizer(
            embed_model='ecapa',  # supported types: ['xvec', 'ecapa']
            cluster_method='ahc',  # supported types: ['ahc', 'sc']
            window=1,  # size of window to extract embeddings (in seconds)
            period=0.1  # hop of window (in seconds)
        )

    def __call__(self, wav_file):
        segments = self.diar.diarize(wav_file,
                                     num_speakers=None,
                                     threshold=9e-1, )

        return segments


diarization = DiarizationPipeline()

if __name__ == '__main__':
    diarization('../converted.wav')