pyssa 0.7.0

Python package for stochastic simulations

pyssa : Python package for stochastic simulations

Introduction

pyssa is a Python package for stochastic simulations. It offers a simple api to define models, perform stochastic simulations on them and visualize the results in a convenient manner.

Install

Install with pip:

$ pip install pyssa

Documentation

• General: https://pyssa.readthedocs.io.

Usage

from pyssa.simulation import Simulation
V_r = np.array([[1, 0], [0, 1], [0, 0]])  # Reactant matrix
V_p = np.array([[0, 0], [1, 0], [0, 1]])  # Product matrix
X0 = np.array([100, 0, 0])  # Initial state
k = np.array([1.0, 1.0])  # Rate constants
sim = Simulation(V_r, V_p, X0, k)  # Declare the simulation object
# Run the simulation
sim.simulate(max_t=150, max_iter=1000, chem_flag=True, n_rep=10)

You can change the algorithm used to perform the simulation by changing the algorithm flag

sim.simulate(max_t=150, max_iter=1000, chem_flag=True, n_rep=10, algorithm="tau_leaping")

Plotting

sim.plot()

Accessing the results

results = sim.results

You can also access the final states of all the simulation runs by

final_times, final_states = results.final

Benchmarks

We chose numba after extensive testing and benchmarking against python and cython implementations.

Name (time in ms)	Min	Max	Mean	StdDev	Median	IQR	Outliers	OPS	Rounds	Iterations
test_numba_benchmark	314.1758 (1.0)	342.9915 (1.0)	322.9318 (1.0)	11.4590 (1.0)	318.7983 (1.0)	9.1533 (1.0)	1;1	3.0966 (1.0)	5	1
test_cy_benchmark	17,345.7698 (55.21)	19,628.3931 (57.23)	18,255.3931 (56.53)	862.4711 (75.27)	18,148.9358 (56.93)	1,030.3676 (112.57)	2;0	0.0548 (0.02)	5	1
test_py_benchmark	27,366.3681 (87.11)	28,417.8333 (82.85)	27,782.2482 (86.03)	387.2758 (33.80)	27,728.4224 (86.98)	347.3891 (37.95)	2;0	0.0360 (0.01)	5	1

License

Copyright (c) 2018-2019, Dileep Kishore, Srikiran Chandrasekaran. Released under: Apache Software License 2.0

Credits

• Numba
• pytest
• Cookiecutter
• audreyr/cookiecutter-pypackage
• black