Fix incorrect t0 in saved NN timeseries

[wbenoit26]

Uses the sequence t0 instead of the postprocessor t0, because the latter accounts for the offset from PSD settle-in time, etc.

[deepchatterjeeligo]

Can you give an example highlighting the difference in the t0 @wbenoit26 ?

[wbenoit26]

The t0 of sequence corresponds to the starting GPS time of the background segment, while the t0 of postprocessor has an offset because we crop some time at the start to account for PSD burn-in. However, the output timeseries that's being saved doesn't have that cropping, so there's a mismatch between t0 and the time of the first sample.