how to handle very long waveforms

[md-arif-shaikh]

Currently, to measure eccentricity even at a single instance of time, we construct the e(t) interpolant using the full waveform. For very long waveforms, e.g. EMRI waveforms, this might be expensive.

Possible solution: We consider only the part of the waveform around the given reference time. We may want to use a number of orbits on either side of the reference time. We can choose the default to be the full waveform.