Run parameter space

This example script shows how to iterate over parameter ranges and generate comparison plots.

Import parameters, the Python module used here to define parameter ranges.

from mesocircuit import mesocircuit_framework as mesoframe
from mesocircuit.plotting.paramspace_figure import parameterspace_overviews
import parametersets
import parameters as ps

Define a 2x3 parameter space for the downscaled Potjans-Diesmann microcircuit varying the relative inhibition and the Poisson background rate. Parameter ranges are only allowed for net_dict and sim_dict since only these two dictionaries are used to create a unique hash.

name = 'local_microcircuit_PD_g_vs_bg_rate'
params_key = 'local_microcircuit_PD'
custom_params = parametersets.ps_dicts[params_key]
custom_params['net_dict'].update({
    'g': ps.ParameterRange([-4., -4.5]),
    'bg_rate': ps.ParameterRange([7., 8., 9])})

Instantiate a MesocircuitExperiment and inspect the individual circuits of each parameter combination.

meso_exp = mesoframe.MesocircuitExperiment(name, custom_params)
print(meso_exp.parameterview)
print(meso_exp.circuits)

For each parameter combination, run the default simulation, analysis, and plotting scripts.

for circuit in meso_exp.circuits:
    circuit.run_jobs(
        jobs=[
            'network',
            'analysis_and_plotting',
        ],
        machine='local'
    )

For each default figure, create one overview figure that compares the data for different parameter combinations, using latex.

if 1:
    parameterspace_overviews(
        paramspace_key=name,
        data_dir=meso_exp.data_dir)