ValueError in model simulation due to wrong assumption

[msis]

When calling model.simulate, it is assumed that time 1 is can be interpolated:

pysindy/pysindy/pysindy.py

Line 738 in 0f7ca74

if ndim(u_fun(1)) == 1:

This assumption is wrong is t is a given array.

Example:

import numpy as np

import pysindy as ps

t = np.linspace(0, 1, 100)
u = np.array([0]*100)
x = np.power(t, 2)

model = ps.SINDy()
model.fit(
    x,
    t=t,
    u=u,
)
print(model.print())

def gen_u(t):
    return t

x0 = 6

t_test_success = np.arange(0, 1.1, 0.1)
u_test_success = gen_u(t_test_success)

t_test_fail = np.arange(0, 1, 0.1)
u_test_fail = gen_u(t_test_fail)

sim_success = model.simulate(
    [x0],
    t=t_test_success,
    u=u_test_success,
)
sim_fail = model.simulate(
    [x0],
    t=t_test_fail,
    u=u_test_fail,
)

In this example, sim_fail will raise a ValueError exception as 1 is out of the interpolation range (see linked line above).

Proposed fix:

Replace : if ndim(u_fun(1)) == 1: with if ndim(u_fun(t[0])) == 1: .

[briandesilva]

Thanks for pointing this out. I'll push a fix this afternoon.

[briandesilva]

Fixed in 9d58553. I've got another bug fix or two in the pipeline that I'd like to bundle with this in a new release. It should be out by the end of the week. If you urgently need this fixed, feel free to clone the master branch and install it in development mode:

git clone https://github.com/dynamicslab/pysindy.git
cd pysindy
pip install -e .

[msis]

Thanks @briandesilva !

I've been using the following work around for my situation:

sim_fail = model.simulate(
    [x0],
    t=t_test_fail-t_test_fail[0],
    u=u_test_fail,
)

[briandesilva]

You shifted the time to start at 0—clever!