Resetting the kernel in g4py when running multiple simulations

Particles, Track, Event, Run and Biasing
1

I’m working on using the pytest package to establish tests for some Geant4 code built using g4py. Normally in pytest one would do something like the following to test multiple geometries:

@pytest.mark.parametrize('geometry', ['det0', 'det1', 'det2'])
def test_geometries(geometry):
	run_example(geometry)

However, each time run_example() is called we also do import geant4py as g4py, so we end up reinitializing the kernel without resetting it properly (if I understand correctly). This leads to a segfault on the second geometry tested, det1.

My question: is there a way to properly reset the kernel in this g4py framework, or should I restructure how I sweep over multiple geometries?

1 Like