def test_gathering_scores():
np.random.seed(0)
num_cells = 3
pi = np.zeros(num_cells)
# Always leave middle region
pi[1] = 1
# region 2 is more attractive
s = np.array([1, 1, 2])
beta = 1
pos = np.arange(num_cells)
d = np.abs(pos[np.newaxis, :] - pos[:, np.newaxis])
K = 2
sim = forward.ForwardSimulator(pi, s, beta, d, K)
N_init = np.array([0, 1000, 0])
N_sim, M_sim = sim.simulate(N_init, 2)
N_end = N_sim[-1]
# Should have approximately twice as many people in region 2 as region 0
assert N_end[0] / N_end[2] < 0.60
assert N_end[0] / N_end[2] > 0.40
def test_forbiden_regions():
"""
Test that people don't move further than they are allowed
"""
np.random.seed(0)
num_cells = 10
pi = np.random.uniform(size=num_cells)
s = np.random.uniform(size=num_cells)
beta = 1
pos = np.arange(num_cells)
d = np.abs(pos[np.newaxis, :] - pos[:, np.newaxis])
K = 2
sim = forward.ForwardSimulator(pi, s, beta, d, K)
N_init = np.random.randint(10, 100, num_cells)
N_sim, M_sim = sim.simulate(N_init, 20)
assert np.array_equal(M_sim, M_sim * (d <= K))
def test_always_leave():
np.random.seed(0)
num_cells = 2
pi = np.zeros(num_cells)
# Always leave first region
pi[0] = 1
s = np.ones(num_cells)
beta = 1
pos = np.arange(num_cells)
d = np.abs(pos[np.newaxis, :] - pos[:, np.newaxis])
K = 1
sim = forward.ForwardSimulator(pi, s, beta, d, K)
N_init = np.array([10, 0])
N_sim, M_sim = sim.simulate(N_init, 4)
N_true = np.array([
[N_init[0], N_init[1]],
[0, N_init.sum()],
[0, N_init.sum()],
[0, N_init.sum()],
])
assert np.array_equal(N_sim, N_true)
def test_no_destinations_no_leave():
"""
An exception should NOT be raised if there is no possible destination for a
cell but leaving probability is zero.
"""
num_cells = 3
pi = np.zeros(num_cells)
s = np.ones_like(pi)
beta = 1
pos = np.arange(num_cells)
d = np.abs(pos[np.newaxis, :] - pos[:, np.newaxis])
# Distance cutoff is too short to move anywhere
K = np.diff(pos).min() / 2
_ = forward.ForwardSimulator(pi, s, beta, d, K)
def test_no_destinations():
"""
An exception should be raised if there is no possible destination for a cell
"""
num_cells = 3
pi = np.zeros(num_cells)
# Always leave middle region
pi[1] = 1
s = np.ones_like(pi)
beta = 1
pos = np.arange(num_cells)
d = np.abs(pos[np.newaxis, :] - pos[:, np.newaxis])
# Distance cutoff is too short to move anywhere
K = np.diff(pos).min() / 2
with pytest.raises(ValueError) as _:
_ = forward.ForwardSimulator(pi, s, beta, d, K)
def test_never_leave():
np.random.seed(0)
num_cells = 5
pi = np.zeros(num_cells)
s = np.ones(num_cells)
beta = 1
pos = np.arange(num_cells)
d = np.abs(pos[np.newaxis, :] - pos[:, np.newaxis])
K = 1
sim = forward.ForwardSimulator(pi, s, beta, d, K)
N_init = np.array([10, 0, 5, 6, 7])
N_sim, M_sim = sim.simulate(N_init, 4)
N_true = np.array([N_init, N_init, N_init, N_init])
assert np.array_equal(N_sim, N_true)