def test_one_point():
pbounds = {'p1': (0, 1), 'p2': (1, 100)}
space = TargetSpace(target_func, pbounds)
x = space.random_points(1)[0]
space.observe_point(x)
space._assert_internal_invariants(fast=False)
assert space._n_alloc_rows > len(space)
def test_observe_m_nd_points(m, n):
pbounds = {'p{}'.format(i): (0, i) for i in range(n)}
space = TargetSpace(target_func, pbounds)
for x in space.random_points(m):
space.observe_point(x)
space._assert_internal_invariants(fast=False)
space._assert_internal_invariants(fast=False)
def test_two_points():
"""
pytest tests/test_target_space.py::test_two_points
"""
pbounds = {'p1': (0, 1), 'p2': (1, 100)}
space = TargetSpace(target_func, pbounds)
for x in space.random_points(2):
space.observe_point(x)
space._assert_internal_invariants(fast=False)
space._assert_internal_invariants(fast=False)
assert space._n_alloc_rows == len(space)
def test_m_random_nd_points(m, n):
pbounds = {'p{}'.format(i): (0, i) for i in range(n)}
space = TargetSpace(target_func, pbounds, random_state=0)
X = space.random_points(m)
assert X.shape[0] == m
assert X.shape[1] == n
for i in range(n):
lower, upper = pbounds[space.keys[i]]
assert np.all(X.T[i] >= lower)
assert np.all(X.T[i] <= upper)
def test_contains():
# Simply re-observing a non-unique values returns the cached result
pbounds = {'p1': (0, 1), 'p2': (1, 100)}
space = TargetSpace(target_func, pbounds)
# add 1000 random points
for x in space.random_points(1000):
space.observe_point(x)
# now all points should be unique, so test contains
space2 = TargetSpace(target_func, pbounds)
for x in space.X:
assert x not in space2
y = space2.observe_point(x)
assert x in space2
assert y == space2.observe_point(x)
space2._assert_internal_invariants(fast=False)