def test_unique():
n = 1024
#test k-nearest neighbors
for dim in range(1, 8):
data = np.random.rand(n, dim)
data_unique = np.unique(data, axis=1)
query_data = np.random.rand(dim)
mkdtree = cKDTree.cKDTree(dim)
mkdtree.build(data_unique.copy())
query = mkdtree.query(query_data, 10)
kdtree = scipyKDTree(data_unique, leafsize=1)
dist, index = kdtree.query(query_data, 10)
result = dict(zip(index, dist))
assert(query == result)
#test distance smaller than dist
for dim in range(1, 8):
data = np.random.rand(n, dim)
data_unique = np.unique(data, axis=1)
dist = np.random.rand() / 4
query_data = np.random.rand(dim)
mkdtree = cKDTree.cKDTree(dim)
mkdtree.build(data_unique.copy())
query = list(mkdtree.query_distance(query_data, dist).keys())
kdtree = scipyKDTree(data_unique, leafsize=1)
result = kdtree.query_ball_point(query_data, dist)
result.sort()
assert(query == result)
for dim in range(1, 8):
data = np.random.rand(n, dim)
data_unique = np.unique(data, axis=1)
query_data = np.random.rand(dim)
mkdtree = cKDTree.cKDTree(dim)
mkdtree.build(data_unique.copy())
query = mkdtree.query_nearest(query_data)[0]
kdtree = scipyKDTree(data_unique, leafsize=1)
result = kdtree.query(query_data, 1)[1]
assert(query == result)
def test_sparse_distance_matrix(coordinates, max_dist, p):
"""
Test that the `sparse_distance_matrix` method returns the same thing in the
redistributed KDTree and on the actual scipy one.
"""
original_tree = scipyKDTree(coordinates)
redis_tree = redisKDTree(coordinates)
original_output = original_tree.sparse_distance_matrix(original_tree, max_dist, p)
redis_output = redis_tree.sparse_distance_matrix(redis_tree, max_dist, p)
assert dict_close(redis_output, original_output)
def __init__(self, tree_type, point_list, nclose=0):
self.tree_type = tree_type
self.nclose = nclose
if tree_type not in ['node', 'element']:
# verifies you're calling the right
msg = 'Error! Invalid tree_type\n'
msg += "tree_type=%r valid='node','element'" % tree_type
raise RuntimeError(msg)
npoints = len(point_list)
nodes = array((npoints, 3), dtype='float64')
self.node_ids = array(npoints, dtype='int32')
i = 0
for node_id, xyz in sorted(iteritems(point_list)):
nodes[i, :] = xyz
self.node_ids[i] = node_id
self.tree = scipyKDTree(point_list)