Gap junction mapping from jNeuroML to NetPyNE

[pgleeson]

The mapping from NML2 models to/from NetPyNE is nearly complete. Significantly, this gives an easy way to get network models in NML2 running on Parallel NEURON. One missing feature is gap junctions, and support for these has recently been added to NetPyNE, and requires adding to the NML2 import.

[pgleeson]

Gap junction mapping working and tested nml->netpyne: https://github.com/NeuroML/NeuroML2/blob/master/LEMSexamples/test/.test.ex19a.jnmlnetpyne.omt