NEXTNetR (Next-reaction-based Epidemics eXtended to Temporal Networks) is a Python package for the efficient simulation of epidemics on complex networks (including weighted and temporal networks) with arbitrary transmission and recovery time distributions. NEXTNetPy is a pybind11-based Python wrapper around the C++ library NEXTNet.
See the NEXTNetPy website for a reference and usage examples.
Download the latest released version of NEXTNetPy-v<version>-pkg.tar.gz and install with
pip install NEXTNetPy-v<version>-pkg.tar.gz
Since NEXT-Net is implemented in C++, a C++ compiler is required to install NEXTNetPy. Alternatively, if Git is available, the latest released version of NEXTNetPy can be downloaded, built and installed with
git clone --recurse-submodules --branch latest-release https://github.com/oist/NEXTNetPy.git
cd NEXTNetPy
pip install .
Here is an example of how to run a SIR simulation on a Barabasi-Albert random graph using networkx.
import nextnet as nn
import networkx as nx
import matplotlib.pyplot as plt
n = 10**5 # size of the network
m = 1 # number of edges added per node
graph_nx = nx.barabasi_albert_graph(n,m)
# convert network to a NextNet object.
graph = nn.networkx(graph_nx)
# Define the distribution for the infection times (Gamma distributed)
MEAN_INFECTION = 5
VARIANCE_INFECTION = 1
psi = nn.transmission_time_gamma(MEAN_INFECTION,VARIANCE_INFECTION)
# Define the distribution for the recovery times
MEAN_RECOVERY = 14
VARIANCE_RECOVERY= 3
rho = nn.transmission_time_lognormal(MEAN_RECOVERY,VARIANCE_RECOVERY)
# Define the simulation object
sim = nn.simulation(graph,psi,rho,SIR=True)
# Add initial infections (node 0 infected at time t=0)
initial_infected = [(0,0)]
sim.add_infections(initial_infected)
# Set up random generator
seed = 0
rng = nn.rng(seed)
# Run simulation
results = sim.run(engine=rng)
# Plot results
plt.plot(results["time"],results["infected"])
plt.xlabel("time")
plt.ylabel("number of infected")
plt.show()
