Bz_LoadPhy should save spikes in same format as bz_getspikes

[brendonw1]

This way bz_getspikes can be called no matter which spike sorter was used to manually check - this goes along with diea that buzcode code can work on all data the same once it's in buzcode format.

This is a lesser goal than (and therefore separate from) merging bz_loadphy and bz_getspikes (issue#264)

[brendonw1]

I think it needs two things

• spikes.numcells field
• spikes.regions field
  That's all

[dlevenstein]

perhaps there's a way we can have bz_getspikes call bz_loadphy, that way they have to be piped through the same input/output (i.e. related issue #264, #261 )

[dlevenstein]

closed in lieu of #261