opensourcebrain/netpyneshowcase

NetPyNE & NeuroML interactions

NetPyNE Showcase

Files for testing NetPyNE & NeuroML interactions. See here for more info.

netpyneshowcase's Issues

Test pipeline NetPyNE-> NeuroML-> other simulators...

@salvadord was asking about options for examples of models built in NetPyNE, converted to NeuroML and then run successfully in another simulator. For various reasons EDEN is currently the best option for trying this.

I've added various OMV tests for running the generated NeuroML equivalents of NetPyNE models in EDEN, and the results are good so far. See test files ending jnmleden.omt here, the results of these are at https://github.com/OpenSourceBrain/NetPyNEShowcase/runs/6374237602?check_suite_focus=true.

As an example this: https://github.com/OpenSourceBrain/NetPyNEShowcase/blob/master/NetPyNE/HybridSmall/HybridSmall.py builds a small net with HH & Izhikevich cells, it gets generated to NML here: https://github.com/OpenSourceBrain/NetPyNEShowcase/blob/master/NeuroML2/HybridSmall.net.nml (see here for generating script).
The LEMS simulation file generated also (https://github.com/OpenSourceBrain/NetPyNEShowcase/blob/master/NeuroML2/LEMS_HybridSmall.xml) can be used to generate a Python script which loads the NML using the latest jnml or pynml:

jnml LEMS_HybridSmall.xml  -eden
# or ...
pynml LEMS_HybridSmall.xml  -eden

# then ...
python LEMS_HybridSmall_eden.py

This is what gets tested in OMV files like: https://github.com/OpenSourceBrain/NetPyNEShowcase/blob/master/NeuroML2/.test.hyizh.jnmleden.omt

cc @spanag

Point process cells connections working but can't test in showcase due to error

@pgleeson I tested connecting the Izhi07a (artificial cell) to HH and viceversa and it works. I added the code to the tut_import.py so you can try it yourself: https://github.com/Neurosim-lab/netpyne/blob/development/doc/source/code/tut_import.py

The key so that cells that don't use the section voltage work, is to specify in the 'vref' param what is the point process variable that will be used as voltage, eg. in the Izhi07a the variables is called 'V':
cellRule['secs']['soma']['pointps']['Izhi2007a_0']['vref'] = 'V' # specify that uses its own voltage V

I then changed my branch to the neuroml_export and pulled, to try to test your code here: https://github.com/OpenSourceBrain/NetPyNEShowcase/blob/a239a2818a361610b64cccb92028b42c04948f45/NeuroML2/LEMS_Spikers_netpyne.py

However, I got the following error:

Salvador-Duras-MacBook-Pro% ipython -i LEMS_Spikers_netpyne.py
Python 2.7.12 |Anaconda 2.0.1 (x86_64)| (default, Jul 2 2016, 17:43:17)
Type "copyright", "credits" or "license" for more information.

IPython 5.1.0 -- An enhanced Interactive Python.
? -> Introduction and overview of IPython's features.
%quickref -> Quick reference.
help -> Python's own help system.
object? -> Details about 'object', use 'object??' for extra details.
NEURON -- VERSION 7.5 (1454:2350fc838a79) 2016-08-01
Duke, Yale, and the BlueBrain Project -- Copyright 1984-2016
See http://neuron.yale.edu/neuron/credits

Running a NetPyNE based simulation for 1000.0ms (dt: 0.024999999ms) at 6.3 degC
Importing NeuroML 2 network from: Spikers.net.nml
Parsing: Spikers.net.nml
Loaded: Spikers.net.nml as NeuroMLDocument
INFO:DefaultNetworkHandler:Document: Spikers
INFO:DefaultNetworkHandler:Network: Spikers

TypeError Traceback (most recent call last)
/u/salvadord/Documents/ISB/Models/NetPyNEShowcase/NeuroML2/LEMS_Spikers_netpyne.py in ()
85 print("Running a NetPyNE based simulation for %sms (dt: %sms) at %s degC"%(simConfig.duration, simConfig.dt, h.celsius))
86
---> 87 gids = sim.importNeuroML2SimulateAnalyze(nml2_file_name,simConfig)
88
89 print("Finished simulation")

/u/salvadord/Documents/ISB/Models/netpyne_repo/netpyne/wrappers.pyc in importNeuroML2SimulateAnalyze(fileName, simConfig)
141 def importNeuroML2SimulateAnalyze(fileName, simConfig):
142
--> 143 return sim.importNeuroML2(fileName, simConfig)

/u/salvadord/Documents/ISB/Models/netpyne_repo/netpyne/simFuncs.py in importNeuroML2(fileName, simConfig)
2196 currParser = NeuroMLXMLParser(nmlHandler) # The XML handler knows of the structure of NeuroML and calls appropriate functions in NetworkHandler
2197
-> 2198 currParser.parse(fileName)
2199
2200 nmlHandler.finalise()

/u/salvadord/anaconda/lib/python2.7/site-packages/neuroml/hdf5/NeuroMLXMLParser.pyc in parse(self, filename)
53 self.netHandler.handlePopulation(population.id,
54 population.component,
---> 55 len(population.instances))
56 for inst in population.instances:
57

TypeError: handlePopulation() takes exactly 5 arguments (4 given)