Sort out the GPS timeshift
The current implementation (which works and gives informative posteriors for the geocent_time on GW150914.py) has the following line in detector.py
signal_ifo = signal_ifo * np.exp(-1j * 2 * np.pi * (time_shift + parameters['geocent_time'])
@rory-smith suggests that this should instead be handled by subtracting the geocent_time
(or some reference epoch) from the input time. I'm not sure what the best way to handle this is, but I want to flag it before it gets forgotten.
@colm.talbot @paul-lasky @eric.thrane @marcus.lower @rory-smith