phase_align can't interpolate when cycle contains one sample

When the length of cycle_inds is 1 (Does this mean the cycle only contains one sample?), there's an error at the interpolation stage.

This happened in the example below:

C = emd.cycles.Cycles(IP[nimf, :])

# Compute per-cycle metrics
C.compute_cycle_metric('max_amp', IA[nimf, :], np.max)
C.compute_cycle_metric('duration', IA[nimf, :], len)
C.compute_cycle_metric('mean_if', IF[nimf, :], np.mean)
C.compute_cycle_metric('has_replay', replay_ts, np.max)

Citer = C.iterate(conditions=['has_replay>0'])
pa_r, phasex = emd.cycles.phase_align(IP[nimf, :], replay_ts, cycles=Citer)

See below for a screenshot of where the when and where the error occurs, i.e. interp.interp1d is called on x_data, which has length one. Naively, one option to deal with this issue is to add to the iteration in line ~520 (but I realise this might be too crude):

if len(cycle_inds)<=1:
     continue`