bug in time shift?

I think there's a bug in https://git.ligo.org/Monash/tupak/blob/master/tupak/gw/detector.py#L1190

I think it should read

        dt = parameters['geocent_time'] + time_shift - self.strain_data.start_time