def test_output_names(self):
model = se.StochasticSEIRModel(
['S0', 'E0', 'I0', 'R0', 'beta', 'kappa', 'gamma'])
self.assertEqual(model.output_names(), ['S', 'E', 'I', 'R'])
model.set_outputs(['I'])
self.assertEqual(model.output_names(), ['I'])
def test_simulate(self):
model = se.StochasticSEIRModel(
['S0', 'E0', 'I0', 'R0', 'beta', 'kappa', 'gamma'])
initial_values = [9, 0, 1, 0]
constants = [1, 1, 1]
test_parameters = initial_values + constants
n_times = 10
test_times = np.linspace(0, 10, num=n_times)
model._parameters.configure_parameters(np.array(test_parameters))
model.set_outputs(['S', 'I'])
def inner(*args, **kwargs):
for i in range(n_times + 5):
yield np.zeros((5, )) + i
with patch('seirmo._stoch_model.solve_gillespie', side_effect=inner):
output = model.simulate(np.array(test_parameters), test_times)
# Check output shape
self.assertEqual(output.shape, (n_times, 2))
# Check positivity
pos_matrix = (output >= 0)
assert np.all(pos_matrix), 'One of the compartments has negative pop'
def test__init__(self):
model = se.StochasticSEIRModel(
['S0', 'E0', 'I0', 'R0', 'beta', 'kappa', 'gamma'])
self.assertEqual(model.output_names(), ['S', 'E', 'I', 'R'])
self.assertEqual(model.parameter_names(),
['S0', 'E0', 'I0', 'R0', 'beta', 'kappa', 'gamma'])
self.assertEqual(model.n_outputs(), 4)
self.assertEqual(model.n_parameters(), 7)
np.testing.assert_array_equal(model._output_collector._output_indices,
np.arange(4))
def test_set_outputs(self):
model = se.StochasticSEIRModel(
['S0', 'E0', 'I0', 'R0', 'beta', 'kappa', 'gamma'])
# Check ValueError will be raised when some output names
# are not as required
with self.assertRaises(ValueError):
model.set_outputs(['incidence number'])
model.set_outputs(['I', 'R'])
# Check the outputs names and number are as expected
self.assertEqual(model._output_collector._output_indices, [2, 3])
self.assertEqual(model.n_outputs(), 2)
def test_propens_func(self):
model = se.StochasticSEIRModel(
['S0', 'E0', 'I0', 'R0', 'beta', 'kappa', 'gamma'])
#check it works
initial_values = [0.9, 0, 0.1, 0]
constants = [1, 1, 1]
test_parameters = initial_values + constants
model._parameters.configure_parameters(np.array(test_parameters))
current_state = np.ones(5)
model_prop = model.update_propensity(current_state)
expected_propensity = np.array([[0, 1, 0, 0], [0, 0, 1, 0],
[0, 0, 0, 1], [0, 0, 0, 0]])
nptest.assert_array_equal(model_prop, expected_propensity)
self.assertEqual(model_prop.shape, (4, 4))
#check zeros
current_state = np.zeros(5)
model_prop = model.update_propensity(current_state)
expected_propensity = np.zeros((4, 4))
nptest.assert_array_equal(model_prop, expected_propensity)
def test_parameter_names(self):
model = se.StochasticSEIRModel(
['S0', 'E0', 'I0', 'R0', 'beta', 'kappa', 'gamma'])
self.assertEqual(model.parameter_names(),
['S0', 'E0', 'I0', 'R0', 'beta', 'kappa', 'gamma'])
def test_n_outputs(self):
model = se.StochasticSEIRModel(np.zeros((4, 4)))
self.assertEqual(model.n_outputs(), 4)