Handle different numbers of samples per window

[jhenin]

[radakb]

Similar to other issues, the lack of this feature seems to be baked into certain visualization functions and data structures, but is not actually imposed by the analysis routines.

[jhenin]

Looks like we worked past each other: deinterleave now outputs correct files with uniform samples, and parseFEP deals with non-uniform samples. Still, not all the details are sorted out on that front, so I think the new version of deinterleave is useful.