def test_areame_different_size1(self):
"""
Test on datasets of different sizes
"""
d1 = [1]
d2 = [3, 3, 3]
result = am.areaMe(d1, d2)
self.assertEqual(result, 2)
def test_areame_point(self):
"""
Test that it can compute the area of two degenerate datasets
"""
d1 = [3]
d2 = [5]
result = am.areaMe(d1, d2)
self.assertEqual(result, 2)
def test_areame_same(self):
"""
Test that it can compute the area of two equal datasets
"""
d1 = [1, 2, 3]
d2 = [1, 2, 3]
result = am.areaMe(d1, d2)
self.assertEqual(result, 0)
def test_areame_tabular_2d(self):
"""
Test on tabular data sets
"""
X, Y = am.example_2d()
areas = am.areaMe(X, Y)
reference_solution = [7.00875, 6.38500, 14.73125, 8.12000, 11.41875]
for a, r in zip(areas, reference_solution):
self.assertAlmostEqual(a, r, places=5)
def test_areame_different_size2(self):
"""
Test on datasets of different sizes
"""
xx = am.skinny()
yy = am.puffy()
result = am.areaMe(xx, yy)
self.assertAlmostEqual(result[0], 1.26611, places=5)
self.assertAlmostEqual(result[1], 1.59278, places=5)
def test_areame_puffy_reverse(self):
"""
Test on the puffy dataset in reverse order
"""
dd = am.puffy()
d1 = [d[0] for d in dd]
d2 = [d[1] for d in dd]
result = am.areaMe(d2, d1)
self.assertAlmostEqual(result, 2.828, places=3)
def test_areame_puffy(self):
"""
Test on the puffy dataset
"""
dd = am.puffy()
d1 = [d[0] for d in dd]
d2 = [d[1] for d in dd]
result = am.areaMe(d1, d2)
self.assertAlmostEqual(result, 2.828, places=3)
def test_areame_skinny_reverse(self):
"""
Test on the skinny dataset in reverse order
"""
dd = am.skinny()
d1 = [d[0] for d in dd]
d2 = [d[1] for d in dd]
result = am.areaMe(d2, d1)
self.assertAlmostEqual(result, 0.51, places=3)
def test_areame_skinny(self):
"""
Test on the skinny dataset
"""
dd = am.skinny()
d1 = [d[0] for d in dd]
d2 = [d[1] for d in dd]
result = am.areaMe(d1, d2)
self.assertAlmostEqual(result, 0.51, places=3)
def test_areame_mixtures_1d(self):
dim = (7, )
X = am.example_random_Nd(n=179, dim=dim)
Y = am.example_random_Nd(n=179, dim=dim)
areas = am.areaMe(X, Y)
XY_m = am.mixture((X, Y))
area_mix = am.areame_mixture(XY_m)
for a, a_mix in zip(areas, area_mix):
self.assertAlmostEqual(a, a_mix, places=5)
def test_areame_vs_scipy_2d(self):
dim = (3, 5)
X = am.example_random_Nd(n=79, dim=dim)
Y = am.example_random_Nd(n=79, dim=dim)
areas = am.areaMe(X, Y)
J = am.map_index_flat_to_array(dim)
for i in range(np.prod(dim)):
j, k = J[i]
x, y = X[:, j, k], Y[:, j, k]
a_scipy = wasserstein_distance(x, y)
self.assertAlmostEqual(areas[j, k], a_scipy, places=5)
