2022-07-23__Record_many.jl

# -*- coding: utf-8 -*-
# ---
# jupyter:
#   jupytext:
#     formats: ipynb,jl:light
#     text_representation:
#       extension: .jl
#       format_name: light
#       format_version: '1.5'
#       jupytext_version: 1.13.7
#   kernelspec:
#     display_name: Julia 1.7.1
#     language: julia
#     name: julia-1.7
# ---

# # 2022-07-23 • Record many neurons

# ## Imports

# +
#
# -

using Revise

using MyToolbox

using VoltoMapSim

# ## Params

# Based on Roxin; same as previous nb's.

d = 6
p = get_params(
    duration = 10minutes,
    #duration = 10seconds,
    p_conn = 0.04,
    g_EE = 1   / d,
    g_EI = 18  / d,
    g_IE = 36  / d,
    g_II = 31  / d,
    ext_current = Normal(-0.5 * pA/√seconds, 5 * pA/√seconds),
    E_inh = -80 * mV,
    record_v = [1:40; 801:810;],
);

# ## Run sim

s = cached(sim, [p.sim]);

# Takes ~11' on work pc.

s = augment_simdata(s, p);

# ## Eval conntest perf for all v recorded

recorded = p.sim.network.record_v;

# +
# using Base.Threads

# +
detrates = Dict()

# @threads for m in recorded
for m in recorded
    ii = get_input_info(m, s, p);
    perf = cached(evaluate_conntest_perf, [ii.v, ii.spiketrains, p], key=[p, m])
    detrates[m] = perf.detection_rates
end;
# -

# Connection tests take ~1:20 per neuron.
# So for 50 neurons → 66'.

# Testing can happen multithreaded. So we'll set the JULIA_NUM_THREADS env var (to one less than the num logical cores).
#
# ~~Ah seems you can use all cores, as it doesn't utilize fully, so there's left for you.~~
# Nope, using all cores clogs pc.
#
# With the threading it was sometimes up to 7:30 per neuron..

# ## Plot distributions

using PyPlot

using VoltoMapSim.Plot

exc_rec = [1:40;]
inh_rec = [801:810;];

function ydistplot(pairs...; setkw...)
    labels = [p[1] for p in pairs]
    datas = [p[2] for p in pairs]
    fig, ax = plt.subplots()
    for (x, ys) in enumerate(datas)
        xs = x .- 0.15 .+ 0.1*rand(length(ys))
        ax.plot(xs, ys, "k.", clip_on=false, alpha=0.4)
        ax.boxplot(
            ys, whis=(5,95), positions=[x+0.1], showfliers=false, showmeans=true,
            medianprops=Dict(:color=>"black"),
            meanprops=Dict(:marker=>"D", :ms=>3, :mfc=>"black", :mec=>"none")
        )
    end
    N = length(labels)
    set(ax; xtype=:categorical, xlim=[0.5, N+0.5], setkw...)
    ax.set_xticks(1:N)  # can't use `set` as that sets xticks
    ax.set_xticklabels(labels)
    return ax
end;

function detplot(coll, name)
    N = length(coll)
    ax = ydistplot(
        "Exc. inputs" => [detrates[n].TPR_exc for n in coll],
        "Inh. inputs" => [detrates[n].TPR_inh for n in coll],
        "Unconn." => [detrates[n].FPR for n in coll],
        ylim = [0,1],
        hylabel = "Detection rates for $(name) neurons (n = $N)",
    )
    add_α_line(ax, p.evaluation.α)
    return nothing
end;

detplot(exc_rec, "excitatory")

detplot(inh_rec, "inhibitory")

# Whiskers are the 5 and 95 percentiles.