Handing failures due to eccentricity decay

[md-arif-shaikh]

PR #203 addresses #84 when no sufficient extrema are not found and we can not build interpolants through the pericenters and apocenters. However, there would be cases where we can find enough extrema in the initial part of the waveform and therefore build the interpolants but due to decay of eccentricity the last pericenters/apocenters may be far from the merger. In such cases, if tref_in is later than the last pericenters/apocenters, eccentricity measurement will fail. We need to handle such failures also.