def test_scalarizer_simple():
scalarizer = Scalarizer(simple_vector_valued_fun, simple_scalarizer)
xs = np.array([[1, 2, 3, 4], [9, 8, 7, 6], [1, 5, 7, 3]])
res = scalarizer.evaluate(xs)
assert np.array_equal(res, [14, 60, 25])
def test_scalarizer_asf():
asf = PointMethodASF(np.array([10, 10, 10]), np.array([-10, -10, -10]))
ref = np.atleast_2d([1, 5, 2.5])
scalarizer = Scalarizer(
simple_vector_valued_fun, asf, scalarizer_args={"reference_point": ref}
)
res = scalarizer.evaluate(np.atleast_2d([2, 1, 1, 1]))
assert np.allclose(res, 0.1000002)
def test_scalarizer_simple_with_arg():
scalarizer = Scalarizer(
simple_vector_valued_fun,
simple_scalarizer,
evaluator_args={"extra": 1},
scalarizer_args={"extra": 4},
)
xs = np.array([[1, 2, 3, 4], [9, 8, 7, 6], [1, 5, 7, 3]])
res = scalarizer.evaluate(xs)
assert np.array_equal(res, [-17, -63, -28])
dscalarizer = DiscreteScalarizer(lambda x: np.sum(x, axis=1))
res_1d = dscalarizer(vector)
assert np.array_equal(res_1d, [6.0])
def test_discrete_args():
vectors = np.array([[1, 1, 1], [2, 2, 2], [4, 5, 6.0]])
dscalarizer = DiscreteScalarizer(
lambda x, a=1: a * np.sum(x, axis=1), scalarizer_args={"a": 2}
)
res = dscalarizer(vectors)
assert np.array_equal(res, [6, 12, 30])
if __name__ == "__main__":
asf = PointMethodASF(np.array([10, 10, 10]), np.array([-10, -10, -10]))
ref = np.atleast_2d([2.5, 2.5, 2.5])
scalarizer = Scalarizer(
simple_vector_valued_fun, asf, scalarizer_args={"reference_point": ref}
)
res = scalarizer.evaluate(np.atleast_2d([2, 1, 1, 1]))
print(res)
asf.nadir = np.array([9, 9, 9])
res = scalarizer.evaluate(np.atleast_2d([2, 1, 1, 1]))
print(res)