integrate deformetrica registrations with nat

[jefferis]

I guess that would be enhances nat, but might need to be imports because of the generic.

Gist would probably be to define a type like deformetricareg that referenced the registration folder on disk and then define an xformpoints.deformetricareg method. Question is whether we would need to import nat::xformpoints

Might also be worth thinking again about an xyzmatrix<-.neuronlist method to avoid repeated calls to ShootAndFlow3.

[alexanderbates]

Added code that achieves xyzmatrix<-.neuronlist, to deal with neuronlists for shootflow.neuronlist. Didn't make a specific "xyzmatrix<-.neuronlist" function, figured you'd push one to nat at some point.

[jefferis]

Added code that achieves xyzmatrix<-.neuronlist, to deal with neuronlists for shootflow.neuronlist.

Great.

Didn't make a specific "xyzmatrix<-.neuronlist" function, figured you'd push one to nat at some point.

Quite right.

Do push and I can review

[alexanderbates]

Done. Took 2 hours to "otherside" all the catmaid neurons with good somas (4,005) on my machine, running one core

[jefferis]

See #5 for my initial tidying.