def test_t_span_input(self):
"""Ensure correct error handling for invalid t_span inputs"""
with self.assertRaises(ValueError): # t_span is 2D
list(se.solve_gillespie(self.m, self.initial, t_span=[0]))
# convert generator to list to force function to be evaluated#
with self.assertRaises(ValueError): # t_span must have range
list(se.solve_gillespie(self.m, self.initial, t_span=[0, 0]))
with self.assertRaises(ValueError): # t_stop > t_start
list(se.solve_gillespie(self.m, self.initial, t_span=[-2, 0]))
with self.assertRaises(TypeError): # time values must be floats
list(se.solve_gillespie(self.m, self.initial, t_span=[0, 'ten']))
def test_neg_propensity(self):
"""Test error handling of negative elements in propensity matrix"""
m_neg = MagicMock()
m_neg.return_value = np.array([[0, -1], [0, 0]])
solve = se.solve_gillespie(m_neg, self.initial, self.t_span)
with self.assertRaises(ValueError):
next(solve)
def test_gillespie_zeros(self, initial):
"""Ensure that zero propensity gives unchanged state for any initial"""
m_zeros = MagicMock()
m_zeros.return_value = np.array([[0, 0], [0, 0]])
solution = list(se.solve_gillespie(m_zeros, initial, [0, 10]))
final_sol = solution[-1][1:] # take only compartment nums at end
self.assertEqual(final_sol.tolist(), initial.tolist(),
'Unexpected output - changed state')
def test_propensity_call(self):
m_count = MagicMock()
m_count.return_value = np.array([[0, 1], [0, 0]])
solve = se.solve_gillespie(m_count, self.initial, self.t_span)
next(solve) # return a yield
self.assertEqual(m_count.call_count, 1,
'Propensity Func called unexpected number of times')
next(solve) # return a yield
self.assertEqual(m_count.call_count, 2,
'Propensity Func called unexpected number of times')
def test_population_conservation(self, initial, propensity_mat):
"""Ensure population is conserved for any initial and prop matrix"""
m_3dim = MagicMock()
m_3dim.return_value = propensity_mat
initial_pop = np.sum(initial)
solve = se.solve_gillespie(m_3dim, initial, [0, 1])
while True:
try:
output = next(solve)
except StopIteration:
break
self.assertEqual(output.shape, (len(initial) + 1, ),
'Unexpected number of output compartments')
self.assertAlmostEqual(np.sum(output[1:]), initial_pop,
'Unexpected output - pop. not conserved')
def test_gillespie_output(self):
def prop_func(x: np.ndarray):
return np.array([[0, x[1]], [0, 0]])
state = self.initial
solve = se.solve_gillespie(prop_func, state, [0, 100])
while True:
try:
output = next(solve)
except StopIteration:
break
state = output[1:]
self.assertTrue(np.all(state >= 0),
'Returned negative values in state array')
self.assertEqual(state.tolist(), [0, 10],
'Unexpected output - incomplete infection')
def test_intial_input(self):
with self.assertRaises(ValueError): # initial conditions must be +ve
list(se.solve_gillespie(self.m, np.array([-10, 0]), self.t_span))
def test_tspan_ordering(self, start, stop):
test_span = [start, stop]
if stop <= start:
with self.assertRaises(ValueError):
list(se.solve_gillespie(self.m, self.initial, test_span))