def test_bfs(self):
a = igl.adjacency_matrix(self.f1)
p, d = igl.bfs(a, 0)
self.assertEqual(p.shape, (self.v1.shape[0],))
self.assertEqual(p.shape, (self.v1.shape[0],))
try:
p, d, = igl.bfs(a, -1)
self.assertTrue(False)
except IndexError as e:
pass
a = csc.csc_matrix(np.zeros([0, 0], dtype=np.int32))
try:
p, d, = igl.bfs(a, 0)
self.assertTrue(False)
except ValueError as e:
pass
a = csc.csc_matrix(np.zeros([10, 11], dtype=np.int32))
try:
p, d, = igl.bfs(a, 0)
self.assertTrue(False)
except ValueError as e:
pass
a = csc.csc_matrix(np.zeros([10, 10], dtype=np.int32))
p, d, = igl.bfs(a, 0)
self.assertEqual(p.shape, ())
self.assertTrue(np.array_equal(d, -np.ones(10)))
self.assertTrue(p.flags.c_contiguous)
def test_vertex_components(self):
a = igl.adjacency_matrix(self.f1)
c, count = igl.vertex_components_from_adjacency_matrix(a)
self.assertEqual(c.shape[0], self.v1.shape[0])
c = igl.vertex_components(self.f1)
self.assertEqual(c.shape[0], self.v1.shape[0])
def get_graph_laplacian(faces):
"""
Return the Graph Laplacian of the surface described by faces
"""
A = -igl.adjacency_matrix(faces)
D = A.sum(axis=1)
A.setdiag(-np.squeeze(np.asarray(D)))
return A
def smooth_data(v, f, w):
# Smoothing = Minimize curvature
#
# min_{p*}{E(p*)}
#
# E(p*) = sum_{p in points on mesh}{A_p * ((Lp*)^2 + w(p* - p)^2)},
# where A_p is the voronoi region around point p
#
# Solve dE(p*)/dp* = 0
# => (L'ML + wM)p* = (wM)p ~= Ax = b
#
# L = M^(-1)L_w, w = cotangent[c]/uniform[u]/...
#
# L_c defined in slides 5 page 66
# M and L defined in slides 6 page 23
# Constructing the squared Laplacian and squared Hessian energy
# NOTE L_c is sparse
# L_c = igl.cotmatrix(v, f) # discrete laplacian
# L_c.data[np.isnan(L_c.data)] = 0
# L_c.data[L_c.data > 1e7] = 1e7
# L_c.data[L_c.data < -1e7] = -1e7
# Uniform laplacian
# adopted from here:
# https://libigl.github.io/libigl-python-bindings/igl_docs/#cotmatrix
adj = igl.adjacency_matrix(f)
# Sum each row = number of neighbours
adj_sum = np.squeeze(np.asarray(np.sum(adj, axis=1)))
# Convert row sums into diagonal of sparse matrix
adj_diag = sp.sparse.diags(adj_sum)
# Build uniform laplacian
L_u = adj - adj_diag
# NOTE M is sparse
M = igl.massmatrix(v, f, igl.MASSMATRIX_TYPE_VORONOI)
# Give corrupted triangles only very little weight
mask = np.logical_or(np.isnan(M.data), M.data == 0)
M.data[mask] = 1e-7
M_inv = sp.sparse.diags(1 / M.diagonal())
# NOTE L_c can become very weird when dealing with corrupted meshes
# L = M_inv @ L_c
L = M_inv @ L_u
A = L.T @ M @ L + w * M
# TODO lambda * ID ?
A += sp.sparse.identity(A.shape[0])
b = w * M @ v
v_star = sp.sparse.linalg.spsolve(A, b)
return v_star
def test_adjacency_matrix(self):
a = igl.adjacency_matrix(self.f)
self.assertTrue(a.shape == (self.v.shape[0], self.v.shape[0]))
self.assertTrue(a.dtype == self.f.dtype)
self.assertTrue(type(a) == csc.csc_matrix)
