numbakit-ode: leveraging numba to speed up ODE integration

numbakit-ode: **physical quantities**

numbakit-ode (nbkode) is a Python package to solve ordinary differential equations (ODE) that uses Numba to compile code and therefore speed up calculations.

The API is very similar to scipy’s integrate module therefore allowing for easy migration.

It runs in Python 3.7+ depending on NumPy, SciPy and Numba. It is licensed under BSD.

It is extremely easy and natural to use:

>>> import nbkode
>>> def func(t, y):
...     return -0.1 * y
>>> t0 = 0.
>>> y0 = 1.
>>> solver = nbkode.ForwardEuler(func, t0, y0)
>>> ts, ys =[0., 5., 10.])

You can get a list of all solvers:

>>> import nbkode
>>> nbkode.get_solvers() 

or filter by characteristics or group name (or names).

>>> nbkode.get_solvers(implicit=False, fixed_step=True) 
>>> nbkode.get_solvers('euler', 'adam-bashforth') 

Quick Installation

To install numbakit-ode, simply (soon):

$ pip install numbakit-ode

or utilizing conda, with the conda-forge channel (soon):

$ conda install -c conda-forge numbakit-ode

and then simply enjoy it!

User Guide

Design principles

Fast: We love Numba. It allows you to write clean Python code that translates to optimized machine code at runtime. We aim to be able to leverage this power to solve a system of ordinary differential equations.

Simple but useful API: Solvers are classes easy to instantiate, with sensible defaults and convenient methods.

Correctness: We check against established libraries like SciPy that our implementation match those of established libraries using automated testing.

Data driven development: We take decisions based on data, and for this purpose we measure the performance of each part of the package, and the effect of each change we make.

numbakit-ode is maintained by a community. See AUTHORS for a complete list.

To review an ordered list of notable changes for each version of a project, see CHANGES