def test_rotated(method):
X, cells = circle_random(40, 1.0)
X = numpy.column_stack([X, numpy.zeros(X.shape[0])])
# Apply a robust method first to avoid too crazy meshes.
X, cells = optimesh.cpt.fixed_point_uniform(X, cells, 0.0, 2)
X_orig = X.copy()
cells_orig = cells.copy()
# Create reference solution
num_steps = 10
X_ref, cells_ref = method(X, cells, 0.0, num_steps)
# Create a rotated mesh
theta = numpy.pi / 4
k = numpy.array([1.0, 0.0, 0.0])
X_rot = _rotate(X_orig, theta, k)
cells_rot = cells_orig.copy()
X2, cells2 = method(X_rot, cells_rot, 0.0, num_steps)
# rotate back
X2 = _rotate(X2, -theta, k)
assert numpy.all(cells_ref == cells2)
assert numpy.all(numpy.abs(X_ref - X2) < 1.0e-12)
return
def test_comparison():
X, cells = circle_random(40, 1.0)
# Do a few steps of a robust method to avoid too crazy meshes.
tol = 0.0
n = 10
# X, cells = optimesh.cpt.fixed_point_uniform(X, cells, tol, n)
# X, cells = optimesh.odt.fixed_point_uniform(X, cells, tol, n)
X, cells = optimesh.cvt.quasi_newton_uniform_lloyd(X,
cells,
tol,
n,
omega=2.0)
# from meshplex import MeshTri
# mesh = MeshTri(X, cells)
# mesh.write("out.vtk")
# exit(1)
num_steps = 50
d = {
"cpt-uniform-fp": optimesh.cpt.fixed_point_uniform,
"cpt-uniform-qn": optimesh.cpt.quasi_newton_uniform,
#
"cvt-uniform-lloyd": optimesh.cvt.quasi_newton_uniform_lloyd,
"cvt-uniform-lloyd(2.0)": optimesh.cvt.quasi_newton_uniform_lloyd,
"cvt-uniform-qnb": optimesh.cvt.quasi_newton_uniform_blocks,
"cvt-uniform-qnf": optimesh.cvt.quasi_newton_uniform_full,
#
"odt-uniform-fp": optimesh.odt.fixed_point_uniform,
"odt-uniform-bfgs": optimesh.odt.nonlinear_optimization_uniform,
}
avg_quality = numpy.empty((len(d), num_steps + 1))
for i, (name, method) in enumerate(d.items()):
def callback(k, mesh):
avg_quality[i, k] = numpy.average(mesh.cell_quality)
return
X_in = X.copy()
cells_in = cells.copy()
if name == "cvt-uniform-lloyd(2.0)":
method(X_in,
cells_in,
0.0,
num_steps,
omega=2.0,
callback=callback)
else:
method(X_in, cells_in, 0.0, num_steps, callback=callback)
# sort by best final quality
idx = numpy.argsort(avg_quality[:, -1])[::-1]
labels = list(d.keys())
sorted_labels = [labels[i] for i in idx]
for label, values in zip(sorted_labels, avg_quality[idx]):
plt.plot(values, "-", label=label)
plt.xlim(0, num_steps)
plt.ylim(0.93, 1.0)
plt.grid()
plt.xlabel("step")
plt.ylabel("average cell quality")
plt.legend(bbox_to_anchor=(1.03, 1), loc="upper left", borderaxespad=0)
plt.savefig("comparison.svg", transparent=True, bbox_inches="tight")
# plt.show()
return