def test_make_blobs_ary_parameters(dtype, n_samples, n_features, centers,
cluster_std, center_box, shuffle,
random_state):
centers = centers.astype(np.dtype(dtype))
cluster_std = cluster_std.astype(np.dtype(dtype))
if centers.shape[1] != n_features or cluster_std.shape[0] != n_features:
with pytest.raises(ValueError):
out, labels = \
cuml.make_blobs(dtype=dtype, n_samples=n_samples,
n_features=n_features, centers=centers,
cluster_std=cluster_std,
center_box=center_box, shuffle=shuffle,
random_state=random_state)
else:
out, labels = \
cuml.make_blobs(dtype=dtype, n_samples=n_samples,
n_features=n_features, centers=centers,
cluster_std=cluster_std,
center_box=center_box, shuffle=shuffle,
random_state=random_state)
assert out.shape == (n_samples, n_features), "out shape mismatch"
assert labels.shape == (n_samples, ), "labels shape mismatch"
labels_np = labels.copy_to_host()
assert np.unique(labels_np).shape == (len(centers),), \
"unexpected number of clusters"
def generate_synthetic_dataset(config):
coilType = config['coilType']
nSamples = config['nSamples']
coilDensity = config['coilDensity']
coil1StDev = config['coil1StDev']
coil2StDev = config['coil2StDev']
nGuidePointsPerCoil = config['nGuidePointsPerCoil']
randomSeed = config['randomSeed']
shuffleFlag = config['shuffleFlag']
startTime = time.time()
coil1Centers, coil2Centers = gen_two_coils(nPoints=nGuidePointsPerCoil,
coilType=coilType,
coilDensity=coilDensity)
samplesPerCoil = nSamples // 2
nDims = 3
coil1Data, _ = cuml.make_blobs(n_samples=samplesPerCoil,
n_features=nDims,
centers=coil1Centers,
cluster_std=coil1StDev,
random_state=randomSeed,
dtype='float')
coil2Data, _ = cuml.make_blobs(n_samples=samplesPerCoil,
n_features=nDims,
centers=coil2Centers,
cluster_std=coil2StDev,
random_state=randomSeed,
dtype='float')
combinedData = cupy.empty(shape=(samplesPerCoil * 2, nDims),
dtype='float32',
order='F')
combinedData[0::2] = coil1Data
combinedData[1::2] = coil2Data
combinedLabels = cupy.empty(shape=(samplesPerCoil * 2, 1),
dtype='int',
order='F')
combinedLabels[0::2] = cupy.ones(shape=(samplesPerCoil, 1), dtype='int')
combinedLabels[1::2] = cupy.zeros(shape=(samplesPerCoil, 1), dtype='int')
if shuffleFlag:
cupy.random.seed(randomSeed)
shuffledInds = cupy.random.permutation(combinedData.shape[0])
combinedData = cupy.asfortranarray(combinedData[shuffledInds, :])
combinedLabels = cupy.asfortranarray(combinedLabels[shuffledInds])
data = cudf.DataFrame.from_gpu_matrix(combinedData,
columns=['x', 'y', 'z'])
labels = cudf.DataFrame.from_gpu_matrix(combinedLabels, columns=['labels'])
elapsedTime = time.time() - startTime
return data, labels, elapsedTime
def test_make_blobs_scalar_parameters(dtype, n_samples, n_features, centers,
cluster_std, center_box, shuffle,
random_state):
out, labels = cuml.make_blobs(dtype=dtype,
n_samples=n_samples,
n_features=n_features,
centers=centers,
cluster_std=0.001,
center_box=center_box,
shuffle=shuffle,
random_state=random_state)
# we can use cupy in the future
labels_np = labels.copy_to_host()
assert out.shape == (n_samples, n_features), "out shape mismatch"
assert labels.shape == (n_samples, ), "labels shape mismatch"
if centers is None:
assert np.unique(labels_np).shape == (3,), \
"unexpected number of clusters"
elif centers <= n_samples:
assert np.unique(labels_np).shape == (centers,), \
"unexpected number of clusters"
def test_make_blobs_scalar_parameters(dtype, n_samples, n_features, centers,
cluster_std, center_box, shuffle,
random_state, order):
out, labels = cuml.make_blobs(dtype=dtype,
n_samples=n_samples,
n_features=n_features,
centers=centers,
cluster_std=0.001,
center_box=center_box,
shuffle=shuffle,
random_state=random_state,
order=order)
assert out.shape == (n_samples, n_features), "out shape mismatch"
assert labels.shape == (n_samples, ), "labels shape mismatch"
if order == 'F':
assert out.flags['F_CONTIGUOUS']
elif order == 'C':
assert out.flags['C_CONTIGUOUS']
if centers is None:
assert cp.unique(labels).shape == (3,), \
"unexpected number of clusters"
elif centers <= n_samples:
assert cp.unique(labels).shape == (centers,), \
"unexpected number of clusters"
def test_make_blobs_ary_parameters(dtype, n_samples, n_features, centers,
cluster_std, center_box, shuffle,
random_state):
centers = centers.astype(np.dtype(dtype))
cluster_std = np.full(shape=(1, 10), fill_value=cluster_std, dtype=dtype)
if centers.shape[1] != n_features or \
cluster_std.shape[1] != centers.shape[0]:
with pytest.raises(ValueError):
out, labels = \
cuml.make_blobs(dtype=dtype, n_samples=n_samples,
n_features=n_features, centers=centers,
cluster_std=cluster_std,
center_box=center_box, shuffle=shuffle,
random_state=random_state)
else:
out, labels = \
cuml.make_blobs(dtype=dtype, n_samples=n_samples,
n_features=n_features, centers=centers,
cluster_std=cluster_std,
center_box=center_box, shuffle=shuffle,
random_state=random_state)
assert out.shape == (n_samples, n_features), "out shape mismatch"
assert labels.shape == (n_samples, ), "labels shape mismatch"
labels_np = labels.copy_to_host()
out_np = out.copy_to_host()
assert np.unique(labels_np).shape == (centers.shape[0],), \
"unexpected number of clusters"
# Use kmeans to verify k cluster centers
from sklearn.cluster import KMeans
model = KMeans(n_clusters=centers.shape[0])
model.fit(np.array(out_np))
assert adjusted_rand_score(model.labels_, labels_np)
def test_output_type(input_type: str):
# Set the output type and ensure its respected by the function
with cuml.using_output_type(input_type):
X, y = cuml.make_blobs(n_samples=10,
centers=3,
n_features=2,
random_state=0)
if (isinstance(test_output_types[input_type], tuple)):
assert (isinstance(X, test_output_types[input_type][0]))
assert (isinstance(y, test_output_types[input_type][1]))
else:
assert (isinstance(X, test_output_types[input_type]))
assert (isinstance(y, test_output_types[input_type]))