Example7.py

from neuromllite import (
    Network,
    Cell,
    Population,
    Synapse,
    RectangularRegion,
    RandomLayout,
)
from neuromllite import Projection, RandomConnectivity, OneToOneConnector, Simulation

from neuromllite.NetworkGenerator import check_to_generate_or_run

import sys


def generate():
    dt = 0.025
    simtime = 1000

    ################################################################################
    ###   Build new network

    net = Network(id="Example7_Brunel2000")
    net.notes = "Example 7: based on network of Brunel 2000"

    net.parameters = {
        "g": 4,
        "eta": 1,
        "order": 5,
        "epsilon": 0.1,
        "J": 0.1,
        "delay": 1.5,
        "tauMem": 20.0,
        "tauSyn": 0.1,
        "tauRef": 2.0,
        "U0": 0.0,
        "theta": 20.0,
    }

    cell = Cell(id="ifcell", pynn_cell="IF_curr_alpha")

    cell.parameters = {
        "tau_m": "tauMem",
        "tau_refrac": "tauRef",
        "v_rest": "U0",
        "v_reset": "U0",
        "v_thresh": "theta",
        "cm": 0.001,
        "i_offset": 0,
    }

    # cell = Cell(id='hhcell', neuroml2_source_file='test_files/hhcell.cell.nml')
    net.cells.append(cell)

    poisson_input = Cell(id="poisson_input", pynn_cell="SpikeSourcePoisson")
    poisson_input.parameters = {
        "rate": "1000 * (eta*theta/(J*4*order*epsilon*tauMem)) * (4*order*epsilon)",
        "start": 0,
        "duration": 1e9,
    }
    net.cells.append(poisson_input)

    r1 = RectangularRegion(
        id="region1", x=0, y=0, z=0, width=1000, height=100, depth=1000
    )
    net.regions.append(r1)

    pE = Population(
        id="Epop",
        size="4*order",
        component=cell.id,
        properties={"color": ".9 0 0", "radius": 5},
        random_layout=RandomLayout(region=r1.id),
    )
    pEpoisson = Population(
        id="expoisson",
        size="4*order",
        component=poisson_input.id,
        properties={"color": "0.9 0.7 0.7", "radius": 3},
        random_layout=RandomLayout(region=r1.id),
    )
    pI = Population(
        id="Ipop",
        size="1*order",
        component=cell.id,
        properties={"color": "0 0 .9", "radius": 5},
        random_layout=RandomLayout(region=r1.id),
    )
    pIpoisson = Population(
        id="inpoisson",
        size="1*order",
        component=poisson_input.id,
        properties={"color": "0.7 0.7 0.9", "radius": 3},
        random_layout=RandomLayout(region=r1.id),
    )

    net.populations.append(pE)
    net.populations.append(pEpoisson)
    net.populations.append(pI)
    net.populations.append(pIpoisson)

    net.synapses.append(
        Synapse(
            id="ampa",
            pynn_receptor_type="excitatory",
            pynn_synapse_type="curr_alpha",
            parameters={"tau_syn": 0.1},
        )
    )

    net.synapses.append(
        Synapse(
            id="gaba",
            pynn_receptor_type="inhibitory",
            pynn_synapse_type="curr_alpha",
            parameters={"tau_syn": 0.1},
        )
    )

    delay_ext = dt

    downscale = 1
    J_eff = "J*%s" % (downscale)
    # synaptic weights, scaled for alpha functions, such that
    # for constant membrane potential, charge J would be deposited
    fudge = 0.00041363506632638  # ensures dV = J at V=0
    JE = "((%s)/tauSyn)*%s" % (J_eff, fudge)
    JI = "-1*g*%s" % (JE)

    net.projections.append(
        Projection(
            id="projEinput",
            presynaptic=pEpoisson.id,
            postsynaptic=pE.id,
            synapse="ampa",
            delay=delay_ext,
            weight=JE,
            one_to_one_connector=OneToOneConnector(),
        )
    )

    net.projections.append(
        Projection(
            id="projIinput",
            presynaptic=pIpoisson.id,
            postsynaptic=pI.id,
            synapse="ampa",
            delay=delay_ext,
            weight=JE,
            one_to_one_connector=OneToOneConnector(),
        )
    )

    net.projections.append(
        Projection(
            id="projEE",
            presynaptic=pE.id,
            postsynaptic=pE.id,
            synapse="ampa",
            delay="delay",
            weight=JE,
            random_connectivity=RandomConnectivity(probability="epsilon"),
        )
    )

    net.projections.append(
        Projection(
            id="projEI",
            presynaptic=pE.id,
            postsynaptic=pI.id,
            synapse="ampa",
            delay="delay",
            weight=JE,
            random_connectivity=RandomConnectivity(probability="epsilon"),
        )
    )

    net.projections.append(
        Projection(
            id="projIE",
            presynaptic=pI.id,
            postsynaptic=pE.id,
            synapse="gaba",
            delay="delay",
            weight=JI,
            random_connectivity=RandomConnectivity(probability="epsilon"),
        )
    )

    net.projections.append(
        Projection(
            id="projII",
            presynaptic=pI.id,
            postsynaptic=pI.id,
            synapse="gaba",
            delay="delay",
            weight=JI,
            random_connectivity=RandomConnectivity(probability="epsilon"),
        )
    )

    # print(net)
    # print(net.to_json())
    new_file = net.to_json_file("%s.json" % net.id)

    ################################################################################
    ###   Build Simulation object & save as JSON

    sim = Simulation(
        id="SimExample7",
        network=new_file,
        duration=simtime,
        dt=dt,
        seed=123,
        record_traces={pE.id: [0, 1], pI.id: [0, 1]},
        record_spikes={
            pE.id: "*",
            pI.id: "*",
            pEpoisson.id: [0, 1, 2, 3, 4],
            pIpoisson.id: [0, 1, 2, 3, 4],
        },
    )

    sim.to_json_file()

    return sim, net


if __name__ == "__main__":
    if "-sweep" in sys.argv:
        from neuromllite.sweep.ParameterSweep import *

        sim, net = generate()
        sim.record_traces = {}

        fixed = {"dt": 0.025, "order": 5}

        vary = {"eta": [0.5, 1, 1.5, 2, 3, 4, 5, 6, 7, 8, 9, 10]}
        # vary = {'eta':[1,2,5]}
        # vary['seed'] = [i for i in range(10)]
        vary["seed"] = [i for i in range(5)]

        simulator = "jNeuroML_NetPyNE"
        simulator = "jNeuroML_NEURON"
        simulator = "jNeuroML"
        simulator = "PyNN_NEST"
        simulator = "jNeuroML"

        nmllr = NeuroMLliteRunner("SimExample7.json", simulator=simulator)

        ps = ParameterSweep(
            nmllr,
            vary,
            fixed,
            num_parallel_runs=4,
            plot_all=False,
            heatmap_all=False,
            show_plot_already=False,
            peak_threshold=0,
        )

        report = ps.run()
        ps.print_report()

        #  ps.plotLines('weightInput','average_last_1percent',save_figure_to='average_last_1percent.png')
        # ps.plotLines('weightInput','mean_spike_frequency',save_figure_to='mean_spike_frequency.png')
        # ps.plotLines('eta','Einput[0]/spike:mean_spike_frequency',save_figure_to='mean_spike_frequency.png')
        ps.plotLines(
            "eta",
            "expoisson/0/poisson_input/spike:mean_spike_frequency",
            second_param="seed",
            save_figure_to="mean_spike_frequency_ein.png",
        )
        ps.plotLines(
            "eta",
            "inpoisson/0/poisson_input/spike:mean_spike_frequency",
            second_param="seed",
            save_figure_to="mean_spike_frequency_iin.png",
        )
        ps.plotLines(
            "eta",
            "Epop/0/ifcell/spike:mean_spike_frequency",
            second_param="seed",
            save_figure_to="mean_spike_frequency_e.png",
        )
        ps.plotLines(
            "eta",
            "Ipop/0/ifcell/spike:mean_spike_frequency",
            second_param="seed",
            save_figure_to="mean_spike_frequency_i.png",
        )

        import matplotlib.pyplot as plt

        plt.show()

    if "-sweep2" in sys.argv:
        sim, net = generate()
        sim.record_traces = {}

        fixed = {"dt": 0.025, "order": 5}

        vary = {"eta": [0.5, 1, 1.5, 2, 3]}
        vary = {"epsilon": [0.01, 0.1, 0.2, 0.5, 0.9]}
        vary = {"J": [0.01, 0.1, 0.2, 0.5, 0.9]}
        # vary = {'g':[0.5,1,1.5,2,3,20]}

        first = vary.keys()[0]

        # vary = {'eta':[1,2,5]}
        # vary['seed'] = [i for i in range(10)]
        vary["seed"] = [i for i in range(3)]

        simulator = "jNeuroML"
        simulator = "PyNN_NEST"
        simulator = "jNeuroML_NetPyNE"
        simulator = "jNeuroML"
        simulator = "jNeuroML_NEURON"

        nmllr = NeuroMLliteRunner("SimExample7.json", simulator=simulator)

        ps = ParameterSweep(
            nmllr,
            vary,
            fixed,
            num_parallel_runs=16,
            plot_all=False,
            heatmap_all=False,
            show_plot_already=False,
            peak_threshold=0,
        )

        report = ps.run()
        ps.print_report()

        #  ps.plotLines('weightInput','average_last_1percent',save_figure_to='average_last_1percent.png')
        # ps.plotLines('weightInput','mean_spike_frequency',save_figure_to='mean_spike_frequency.png')
        # ps.plotLines('eta','Einput[0]/spike:mean_spike_frequency',save_figure_to='mean_spike_frequency.png')

        second = "seed"
        ps.plotLines(
            first,
            "expoisson/0/poisson_input/spike:mean_spike_frequency",
            second_param=second,
            save_figure_to="mean_spike_frequency_ein.png",
        )
        ps.plotLines(
            first,
            "inpoisson/0/poisson_input/spike:mean_spike_frequency",
            second_param=second,
            save_figure_to="mean_spike_frequency_iin.png",
        )
        ps.plotLines(
            first,
            "Epop/0/ifcell/spike:mean_spike_frequency",
            second_param=second,
            save_figure_to="mean_spike_frequency_e.png",
        )
        ps.plotLines(
            first,
            "Ipop/0/ifcell/spike:mean_spike_frequency",
            second_param=second,
            save_figure_to="mean_spike_frequency_i.png",
        )

        import matplotlib.pyplot as plt

        plt.show()

    else:
        sim, net = generate()

        ################################################################################
        ###   Run in some simulators

        import sys

        check_to_generate_or_run(sys.argv, sim)