break allen cell types into two files and follow the pattern from huang2017 about pyontutils

[tmsincomb]

What pattern are you referring to? Did you want me to utilize LogicalPhenotype? Also, did you want me to break the file back into 2 files where one creates the transgenic_line details and then the other the cell phenotypes?

[tgbugs]

1. In this case just the actual layout and scoping of the various parts of the code. Basically the top level control structure for neuron lang need to be in the top level of a file, not burried in a class. See for example.
   https://github.com/tgbugs/pyontutils/blob/master/pyontutils/neuron_models/huang2017.py#L14
   https://github.com/tgbugs/pyontutils/blob/master/pyontutils/neuron_models/huang2017.py#L19
   https://github.com/tgbugs/pyontutils/blob/master/pyontutils/neuron_models/huang2017.py#L337
   https://github.com/tgbugs/pyontutils/blob/master/pyontutils/neuron_models/huang2017.py#L338
   You can run all your conversion logic inside the class as you do now, its just that all the 'top level' stuff needs to actually be at the top level because of how I implemented it (sort of for a good reason to force that information into places where it is more visible).
2. No.
3. Yes please, name the transgenic lines file allen_transgenic_lines and stick it in the pyontutils/pyontutils folder.

[tmsincomb]

L19. Is NeuronEBM and to create 2 different Neurons necessary then?
L337. allen_cell_types has Neuron.write() already at the end after all the Neurons were created. Unless that's not what you meant.

[tgbugs]

Ah, I was using L337 as an example of the scope of the call, which is independent of the ordering in the program, but it was clearly a bad example. You can call Neuron.write() anywhere in the program and it will do the right thing. The issue is mostly with the fact that class NeuronACT needs to be declared after Config is called and also in global scope.

[tgbugs]

This is more or less resolved, any further fixes will be via the neurondm simple rework.