def test_fwdwnn_kpoints_group():
"""Test FwdWNN with group of k same points provided : all request should return
point of the dataset."""
result = True
size = 30
for i in range(20):
n = random.randint(1, size)
m = random.randint(1, size)
k = random.randint(1, size)
sigma = random.uniform(0.1, 10)
model = WeightedNNForwardModel(n, m, sigma=sigma, k=k)
ygroup = set()
for i in range(random.randint(1, size)):
x = np.random.rand(n)
y = np.random.rand(m)
ygroup.add(tuple(y))
for i in range(k):
model.add_xy(x, y)
for i in range(10):
x = np.random.rand(n)
yp = model.predict_y(x)
check = min([tools.dist(yp, y) for y in ygroup]) < 1e-10
if not check:
print('Error:', n, m, k,
min([tools.dist(yp, y) for y in ygroup]), yp)
result = result and check
return result
def test_fwdwnn_samepoint():
"""Test FwdWNN with one point provided multiple times."""
result = True
for i in range(25):
n = random.randint(1, 3)
m = random.randint(1, 3)
k = random.randint(1, 20)
sigma = random.uniform(0.1, 10)
model = WeightedNNForwardModel(n, m, sigma=sigma, k=k)
x = np.random.rand(n)
y = np.random.rand(m)
for i in range(random.randint(1, 20)):
model.add_xy(x, y)
for i in range(10):
x = np.random.rand(n)
ye = y
yp = model.predict_y(x)
check = np.allclose(ye, yp, rtol=1e-5, atol=1e-5)
if not check:
print('Error:', x, ye, yp)
result = result and check
return result
def test_fwdwnn_onepoint():
"""Test FwdWNN with only one point provided."""
result = True
max_dim = 5
for i in range(5):
n = random.randint(1, max_dim)
m = random.randint(1, max_dim)
k = random.randint(1, max_dim)
sigma = random.uniform(0.1, 10)
model = WeightedNNForwardModel(n, m, sigma=sigma, k=k)
x = np.random.rand(n)
y = np.random.rand(m)
model.add_xy(x, y)
for i in range(10):
x = np.random.rand(n)
ye = y
yp = model.predict_y(x)
check = np.allclose(ye, yp, rtol=1e-10, atol=1e-10)
if not check:
print('Error:', x, ye, yp)
result = result and check
return result
def test_fwdwnn_kpoints_group():
"""Test FwdWNN with group of k same points provided : all request should return
point of the dataset."""
result = True
size = 30
for i in range(20):
n = random.randint(1, size)
m = random.randint(1, size)
k = random.randint(1, size)
sigma = random.uniform(0.1, 10)
model = WeightedNNForwardModel(n, m, sigma = sigma, k = k)
ygroup = set()
for i in range(random.randint(1, size)):
x = np.random.rand(n)
y = np.random.rand(m)
ygroup.add(tuple(y))
for i in range(k):
model.add_xy(x, y)
for i in range(10):
x = np.random.rand(n)
yp = model.predict_y(x)
check = min([tools.dist(yp, y) for y in ygroup]) < 1e-10
if not check:
print('Error:', n, m, k, min([tools.dist(yp, y) for y in ygroup]), yp)
result = result and check
return result
def test_fwdwnn_samepoint():
"""Test FwdWNN with one point provided multiple times."""
result = True
for i in range(25):
n = random.randint(1, 3)
m = random.randint(1, 3)
k = random.randint(1, 20)
sigma = random.uniform(0.1, 10)
model = WeightedNNForwardModel(n, m, sigma = sigma, k = k)
x = np.random.rand(n)
y = np.random.rand(m)
for i in range(random.randint(1, 20)):
model.add_xy(x, y)
for i in range(10):
x = np.random.rand(n)
ye = y
yp = model.predict_y(x)
check = np.allclose(ye, yp, rtol = 1e-5, atol = 1e-5)
if not check:
print('Error:', x, ye, yp)
result = result and check
return result
def test_fwdwnn_onepoint():
"""Test FwdWNN with only one point provided."""
result = True
max_dim = 5
for i in range(5):
n = random.randint(1, max_dim)
m = random.randint(1, max_dim)
k = random.randint(1, max_dim)
sigma = random.uniform(0.1, 10)
model = WeightedNNForwardModel(n, m, sigma = sigma, k = k)
x = np.random.rand(n)
y = np.random.rand(m)
model.add_xy(x, y)
for i in range(10):
x = np.random.rand(n)
ye = y
yp = model.predict_y(x)
check = np.allclose(ye, yp, rtol = 1e-10, atol = 1e-10)
if not check:
print('Error:', x, ye, yp)
result = result and check
return result
