def test_shelve_save(mesh, casedir):
mtf = MockTupleField(dict(save=True, save_as="shelve"))
msf = MockScalarField(dict(save=True, save_as="shelve"))
pp = PostProcessor(dict(casedir=casedir))
pp.add_fields([mtf, msf])
pp.update_all({}, 0.0, 0)
pp.update_all({}, 0.1, 1)
pp.update_all({}, 0.2, 2)
pp.finalize_all()
for mf in [mtf, msf]:
assert os.path.isdir(pp.get_savedir(mf.name))
assert os.path.isfile(
os.path.join(pp.get_savedir(mf.name), "metadata.db"))
assert os.path.isfile(
os.path.join(pp.get_savedir(mf.name), mf.name + ".db"))
md = shelve.open(os.path.join(pp.get_savedir(mf.name), "metadata.db"),
'r')
assert 'shelve' in md["0"]
assert md['saveformats'] == ['shelve']
md.close()
# Read back
data = shelve.open(
os.path.join(pp.get_savedir(mf.name), mf.name + ".db"), 'r')
for i in ["0", "1", "2"]:
d = data[i]
data.close()
assert d == pp.get(mf.name)
def test_default_save(mesh, casedir):
spacepool = SpacePool(mesh)
Q = spacepool.get_space(1, 0)
V = spacepool.get_space(1, 1)
mff = MockFunctionField(Q, dict(save=True))
mvff = MockVectorFunctionField(V, dict(save=True))
mtf = MockTupleField(dict(save=True))
msf = MockScalarField(dict(save=True))
pp = PostProcessor(dict(casedir=casedir))
pp.add_fields([mff, mvff, mtf, msf])
pp.update_all({}, 0.0, 0)
pp.update_all({}, 0.1, 1)
pp.update_all({}, 0.2, 2)
pp.finalize_all()
for mf in [mff, mvff]:
assert os.path.isdir(pp.get_savedir(mf.name))
assert os.path.isfile(
os.path.join(pp.get_savedir(mf.name), "metadata.db"))
assert os.path.isfile(
os.path.join(pp.get_savedir(mf.name), mf.name + ".hdf5"))
assert os.path.isfile(
os.path.join(pp.get_savedir(mf.name), mf.name + ".h5"))
assert os.path.isfile(
os.path.join(pp.get_savedir(mf.name), mf.name + ".xdmf"))
assert len(os.listdir(pp.get_savedir(mf.name))) == 4
md = shelve.open(os.path.join(pp.get_savedir(mf.name), "metadata.db"),
'r')
assert 'hdf5' in md["0"]
assert 'hdf5' in md['saveformats']
assert 'xdmf' in md["0"]
assert 'xdmf' in md['saveformats']
assert set(md['saveformats']) == set(['hdf5', 'xdmf'])
md.close()
for mf in [mtf, msf]:
assert os.path.isdir(pp.get_savedir(mf.name))
assert os.path.isfile(
os.path.join(pp.get_savedir(mf.name), "metadata.db"))
assert os.path.isfile(
os.path.join(pp.get_savedir(mf.name), mf.name + ".db"))
assert os.path.isfile(
os.path.join(pp.get_savedir(mf.name), mf.name + ".txt"))
md = shelve.open(os.path.join(pp.get_savedir(mf.name), "metadata.db"),
'r')
assert 'txt' in md["0"]
assert 'txt' in md['saveformats']
assert 'shelve' in md["0"]
assert 'shelve' in md['saveformats']
assert set(md['saveformats']) == set(['txt', 'shelve'])
md.close()
def test_TimeIntegral(problem, pp, start_time, end_time, dt):
# Setup some mock scheme state
dt, timesteps, start_timestep = compute_regular_timesteps(problem)
spacepool = SpacePool(problem.mesh)
Q = spacepool.get_space(1,0)
V = spacepool.get_space(1,1)
params = dict(finalize=True, start_time=start_time, end_time=end_time)
pp.add_fields([
MockFunctionField(Q),
MockVectorFunctionField(V),
MockTupleField(),
])
pp.add_fields([
TimeIntegral("t", params),
TimeIntegral("MockFunctionField", params),
TimeIntegral("MockVectorFunctionField", params),
TimeIntegral("MockTupleField", params),
])
# Update postprocessor for a number of timesteps, this is where the main code under test is
for timestep, t in enumerate(timesteps, start_timestep):
# Run postprocessing step
pp.update_all({}, t, timestep)
pp.finalize_all()
assert abs( pp.get("TimeIntegral_t") - (0.5*(end_time**2-start_time**2)) ) < 1e-8
assert errornorm(
pp.get("TimeIntegral_MockFunctionField"),
interpolate(Expression("t1-t0+0.5*x[0]*x[1]*(t1*t1-t0*t0)", degree=1, t1=end_time, t0=start_time), Q)
) < 1e-8
D = problem.D
if D == 2:
assert errornorm(
pp.get("TimeIntegral_MockVectorFunctionField"),
interpolate(Expression(("t1-t0+0.5*x[0]*(t1*t1-t0*t0)", "3*(t1-t0)+0.5*x[1]*(t1*t1-t0*t0)"), degree=1, t1=end_time, t0=start_time), V)
) < 1e-8
elif D == 3:
assert errornorm(
pp.get("TimeIntegral_MockVectorFunctionField"),
interpolate(Expression(("t1-t0+0.5*x[0]*(t1*t1-t0*t0)", "3*(t1-t0)+0.5*x[1]*(t1*t1-t0*t0)", "10*(t1-t0)+0.5*x[2]*(t1*t1-t0*t0)"), degree=1, t1=end_time, t0=start_time), V)
) < 1e-8
I = 0.5*(end_time**2-start_time**2)
assert max( [ abs(x1-x0) for x1,x0 in zip(pp.get("TimeIntegral_MockTupleField"), (I, 3*I, end_time-start_time+5*I) ) ] ) < 1e-8
def test_Maximum(problem, pp, start_time, end_time, dt):
# Setup some mock scheme state
dt, timesteps, start_timestep = compute_regular_timesteps(problem)
spacepool = SpacePool(problem.mesh)
Q = spacepool.get_space(1,0)
V = spacepool.get_space(1,1)
pp.add_fields([
MockFunctionField(Q),
MockVectorFunctionField(V),
MockTupleField(),
])
pp.add_fields([
Maximum("t"),
Maximum("MockFunctionField"),
Maximum("MockVectorFunctionField"),
Maximum("MockTupleField"),
])
xmax = MPI.max(mpi_comm_world(), max(problem.mesh.coordinates()[:,0]))
ymax = MPI.max(mpi_comm_world(), max(problem.mesh.coordinates()[:,1]))
if problem.D > 2:
zmax = MPI.max(mpi_comm_world(), max(problem.mesh.coordinates()[:,2]))
# Update postprocessor for a number of timesteps, this is where the main code under test is
for timestep, t in enumerate(timesteps, start_timestep):
# Run postprocessing step
pp.update_all({}, t, timestep)
if start_time < t < end_time:
assert abs(pp.get("Maximum_t") - t) < 1e-8
assert abs(pp.get("Maximum_MockFunctionField") - (1+xmax*ymax*t)) < 1e-8
if problem.D == 2:
assert abs(pp.get("Maximum_MockVectorFunctionField") - (3+ymax*t)) < 1e-8
elif problem.D == 3:
assert abs(pp.get("Maximum_MockVectorFunctionField") - (10+zmax*t)) < 1e-8
assert abs(pp.get("Maximum_MockTupleField") - (1+5*t)) < 1e-8
pp.finalize_all()
assert abs(pp.get("Maximum_t") - t) < 1e-8
assert abs(pp.get("Maximum_MockFunctionField") - (1+xmax*ymax*t)) < 1e-8
if problem.D == 2:
assert abs(pp.get("Maximum_MockVectorFunctionField") - (3+ymax*t)) < 1e-8
elif problem.D == 3:
assert abs(pp.get("Maximum_MockVectorFunctionField") - (10+zmax*t)) < 1e-8
assert abs(pp.get("Maximum_MockTupleField") - (1+5*t)) < 1e-8
def test_TimeDerivative(problem, pp, start_time, end_time, dt):
# Setup some mock scheme state
dt, timesteps, start_timestep = compute_regular_timesteps(problem)
spacepool = SpacePool(problem.mesh)
Q = spacepool.get_space(1,0)
V = spacepool.get_space(1,1)
params = dict(finalize=True, start_time=start_time, end_time=end_time)
pp.add_fields([
MockFunctionField(Q),
MockVectorFunctionField(V),
MockTupleField(),
])
pp.add_fields([
TimeDerivative("t", params),
TimeDerivative("timestep", params),
TimeDerivative("MockFunctionField", params),
TimeDerivative("MockVectorFunctionField", params),
TimeDerivative("MockTupleField", params),
])
# Update postprocessor for a number of timesteps, this is where the main code under test is
for timestep, t in enumerate(timesteps, start_timestep):
# Run postprocessing step
pp.update_all({}, t, timestep)
pp.finalize_all()
# Get and check values from the final timestep
assert abs( (pp.get("TimeDerivative_t", compute=False)) - (1.0) ) < 1e-8
assert abs( (pp.get("TimeDerivative_timestep", compute=False)) - (1.0/dt) ) < 1e-8
assert errornorm(pp.get("TimeDerivative_MockFunctionField"),
interpolate(Expression("x[0]*x[1]", degree=1), Q)) < 1e-8
D = problem.D
if D == 2:
assert errornorm(pp.get("TimeDerivative_MockVectorFunctionField"),
interpolate(Expression(("x[0]", "x[1]"), degree=1), V)) < 1e-8
elif D == 3:
assert errornorm(pp.get("TimeDerivative_MockVectorFunctionField"),
interpolate(Expression(("x[0]", "x[1]", "x[2]"), degree=1), V)) < 1e-8
assert max( [ abs(x1-x0) for x1,x0 in zip(pp.get("TimeDerivative_MockTupleField"), (1,3,5)) ] ) < 1e-8
def test_Operators(problem, pp, start_time, end_time, dt):
# Setup some mock scheme state
dt, timesteps, start_timestep = compute_regular_timesteps(problem)
spacepool = SpacePool(problem.mesh)
Q = spacepool.get_space(1,0)
V = spacepool.get_space(1,1)
D = problem.mesh.geometry().dim()
pp.add_fields([
MockFunctionField(Q),
MockVectorFunctionField(V),
MockTupleField(),
])
fields = [
Norm("MockFunctionField"),
Norm("MockFunctionField", dict(norm_type='L2')),
Norm("MockFunctionField", dict(norm_type='H10')),
Norm("MockVectorFunctionField"),
Norm("MockVectorFunctionField", dict(norm_type='L2')),
Norm("MockVectorFunctionField", dict(norm_type='H10')),
]
pp.add_fields(fields)
exact = dict()
exact["Norm_MockFunctionField"] = lambda t: norm(interpolate(Expression("1+x[0]*x[1]*t", degree=1, t=t), Q))
exact["Norm_L2_MockFunctionField"] = lambda t: norm(interpolate(Expression("1+x[0]*x[1]*t", degree=1, t=t), Q), 'L2')
exact["Norm_H10_MockFunctionField"] = lambda t: norm(interpolate(Expression("1+x[0]*x[1]*t", degree=1, t=t), Q), 'H10')
if D == 2:
exact["Norm_MockVectorFunctionField"] = lambda t: norm(interpolate(Expression(("1+x[0]*t", "3+x[1]*t"), degree=1, t=t), V))
exact["Norm_L2_MockVectorFunctionField"] = lambda t: norm(interpolate(Expression(("1+x[0]*t", "3+x[1]*t"), degree=1, t=t), V), 'L2')
exact["Norm_H10_MockVectorFunctionField"] = lambda t: norm(interpolate(Expression(("1+x[0]*t", "3+x[1]*t"), degree=1, t=t), V), 'H10')
elif D == 3:
exact["Norm_MockVectorFunctionField"] = lambda t: norm(interpolate(Expression(("1+x[0]*t", "3+x[1]*t", "10+x[2]*t"), degree=1, t=t), V))
exact["Norm_L2_MockVectorFunctionField"] = lambda t: norm(interpolate(Expression(("1+x[0]*t", "3+x[1]*t", "10+x[2]*t"), degree=1, t=t), V), 'L2')
exact["Norm_H10_MockVectorFunctionField"] = lambda t: norm(interpolate(Expression(("1+x[0]*t", "3+x[1]*t", "10+x[2]*t"), degree=1, t=t), V), 'H10')
for f1 in fields:
pp.add_field(f1+2.0)
pp.add_field(2.0+f1)
pp.add_field(f1-2.0)
pp.add_field(2.0-f1)
pp.add_field(f1*2.0)
pp.add_field(2.0*f1)
pp.add_field(f1/2.0)
pp.add_field(2.0/f1)
for f2 in fields:
pp.add_field(f1+f2)
pp.add_field(f1*f2)
pp.add_field(f1-f2)
pp.add_field(f1/f2)
return
# Update postprocessor for a number of timesteps, this is where the main code under test is
for timestep, t in enumerate(timesteps, start_timestep):
# Run postprocessing step
pp.update_all({}, t, timestep)
"""
# Skip these time consuming tests
if start_time < t < end_time:
for f1 in fields:
E1 = exact[f1.name](t)
assert abs(pp.get("Add_%s_2.0" %f1.name) - (2.0+E1)) < 1e-14
assert abs(pp.get("Add_2.0_%s" %f1.name) - (2.0+E1)) < 1e-14
assert abs(pp.get("Subtract_%s_2.0" %f1.name) - (E1-2.0)) < 1e-14
assert abs(pp.get("Subtract_2.0_%s" %f1.name) - (2.0-E1)) < 1e-14
assert abs(pp.get("Multiply_%s_2.0" %f1.name) - (2.0*E1)) < 1e-14
assert abs(pp.get("Multiply_2.0_%s" %f1.name) - (2.0*E1)) < 1e-14
assert abs(pp.get("Divide_%s_2.0" %f1.name) - (E1/2.0)) < 1e-14
if abs(E1) > 1e-14:
assert abs(pp.get("Divide_2.0_%s" %f1.name) - (2.0/E1)) < 1e-14
for f2 in fields:
E2 = exact[f2.name](t)
assert abs(pp.get("Add_%s_%s" %(f1.name, f2.name))-(E1+E2)) < 1e-14
assert abs(pp.get("Multiply_%s_%s" %(f1.name, f2.name))-(E1*E2)) < 1e-14
assert abs(pp.get("Subtract_%s_%s" %(f1.name, f2.name))-(E1-E2)) < 1e-14
if abs(E2) > 1e-14:
assert abs(pp.get("Divide_%s_%s" %(f1.name, f2.name))-(E1/E2)) < 1e-14
"""
pp.finalize_all()
for f1 in fields:
E1 = exact[f1.name](t)
assert abs(pp.get("Add_%s_2.0" %f1.name) - (2.0+E1)) < 1e-14
assert abs(pp.get("Add_2.0_%s" %f1.name) - (2.0+E1)) < 1e-14
assert abs(pp.get("Subtract_%s_2.0" %f1.name) - (E1-2.0)) < 1e-14
assert abs(pp.get("Subtract_2.0_%s" %f1.name) - (2.0-E1)) < 1e-14
assert abs(pp.get("Multiply_%s_2.0" %f1.name) - (2.0*E1)) < 1e-14
assert abs(pp.get("Multiply_2.0_%s" %f1.name) - (2.0*E1)) < 1e-14
assert abs(pp.get("Divide_%s_2.0" %f1.name) - (E1/2.0)) < 1e-14
if abs(E1) > 1e-14:
assert abs(pp.get("Divide_2.0_%s" %f1.name) - (2.0/E1)) < 1e-14
for f2 in fields:
E2 = exact[f2.name](t)
assert abs(pp.get("Add_%s_%s" %(f1.name, f2.name))-(E1+E2)) < 1e-14
assert abs(pp.get("Multiply_%s_%s" %(f1.name, f2.name))-(E1*E2)) < 1e-14
assert abs(pp.get("Subtract_%s_%s" %(f1.name, f2.name))-(E1-E2)) < 1e-14
if abs(E2) > 1e-14:
assert abs(pp.get("Divide_%s_%s" %(f1.name, f2.name))-(E1/E2)) < 1e-14
def test_DomainAvg_DomainSD(problem, pp, start_time, end_time, dt):
# Setup some mock scheme state
dt, timesteps, start_timestep = compute_regular_timesteps(problem)
mesh = problem.mesh
spacepool = SpacePool(mesh)
Q = spacepool.get_space(2,0)
V = spacepool.get_space(2,1)
D = V.num_sub_spaces()
pp.add_fields([
MockFunctionField(Q),
MockVectorFunctionField(V),
MockTupleField(),
])
measures = []
facet_domains = FacetFunction("size_t", problem.mesh)
facet_domains.set_all(0)
cell_domains = CellFunction("size_t", problem.mesh)
cell_domains.set_all(0)
subdomains = AutoSubDomain(lambda x: x[0]<0.5)
subdomains.mark(facet_domains, 1)
subdomains.mark(cell_domains, 1)
measures = [dict(),
dict(measure=ds),
dict(cell_domains=cell_domains, indicator=0),
dict(cell_domains=cell_domains, indicator=1),
dict(facet_domains=facet_domains, indicator=0),
dict(facet_domains=facet_domains, indicator=1),]
for m in measures:
pp.add_field(DomainAvg("MockFunctionField", **m))
pp.add_field(DomainAvg("MockVectorFunctionField", **m))
pp.add_field(DomainSD("MockFunctionField", **m))
pp.add_field(DomainSD("MockVectorFunctionField", **m))
for timestep, t in enumerate(timesteps, start_timestep):
# Run postprocessing step
pp.update_all({}, t, timestep)
v = assemble(Constant(1)*dx(domain=mesh))
v_dx0 = assemble(Constant(1)*dx(0, domain=mesh, subdomain_data=cell_domains))
v_dx1 = assemble(Constant(1)*dx(1, domain=mesh, subdomain_data=cell_domains))
v_ds = assemble(Constant(1)*ds(domain=mesh))
v_ds0 = assemble(Constant(1)*ds(0, domain=mesh, subdomain_data=facet_domains))
v_ds1 = assemble(Constant(1)*ds(1, domain=mesh, subdomain_data=facet_domains))
u = pp.get("MockFunctionField")
uv = pp.get("MockVectorFunctionField")
avg = assemble(u*dx)/v
avg_dx0 = assemble(u*dx(0, subdomain_data=cell_domains))/v_dx0
avg_dx1 = assemble(u*dx(1, subdomain_data=cell_domains))/v_dx1
avg_ds = assemble(u*ds)/v_ds
avg_ds0 = assemble(u*ds(0, subdomain_data=facet_domains))/v_ds0
avg_ds1 = assemble(u*ds(1, subdomain_data=facet_domains))/v_ds1
assert abs(pp.get("DomainAvg_MockFunctionField") - avg) < 1e-8
assert abs(pp.get("DomainAvg_MockFunctionField-dx0") - avg_dx0) < 1e-8
assert abs(pp.get("DomainAvg_MockFunctionField-dx1") - avg_dx1) < 1e-8
assert abs(pp.get("DomainAvg_MockFunctionField-ds") - avg_ds) < 1e-8
assert abs(pp.get("DomainAvg_MockFunctionField-ds0") - avg_ds0) < 1e-8
assert abs(pp.get("DomainAvg_MockFunctionField-ds1") - avg_ds1) < 1e-8
std = sqrt(assemble((u-Constant(avg))**2*dx)/v)
std_dx0 = sqrt(assemble((u-Constant(avg_dx0))**2*dx(0, subdomain_data=cell_domains))/v_dx0)
std_dx1 = sqrt(assemble((u-Constant(avg_dx1))**2*dx(1, subdomain_data=cell_domains))/v_dx1)
std_ds = sqrt(assemble((u-Constant(avg_ds))**2*ds)/v_ds)
std_ds0 = sqrt(assemble((u-Constant(avg_ds0))**2*ds(0, subdomain_data=facet_domains))/v_ds0)
std_ds1 = sqrt(assemble((u-Constant(avg_ds1))**2*ds(1, subdomain_data=facet_domains))/v_ds1)
assert abs(pp.get("DomainSD_MockFunctionField") - std) < 1e-8
assert abs(pp.get("DomainSD_MockFunctionField-dx0") - std_dx0) < 1e-8
assert abs(pp.get("DomainSD_MockFunctionField-dx1") - std_dx1) < 1e-8
assert abs(pp.get("DomainSD_MockFunctionField-ds") - std_ds) < 1e-8
assert abs(pp.get("DomainSD_MockFunctionField-ds0") - std_ds0) < 1e-8
assert abs(pp.get("DomainSD_MockFunctionField-ds1") - std_ds1) < 1e-8
avg = [assemble(uv[i]*dx)/v for i in xrange(D)]
avg_dx0 = [assemble(uv[i]*dx(0, subdomain_data=cell_domains))/v_dx0 for i in xrange(D)]
avg_dx1 = [assemble(uv[i]*dx(1, subdomain_data=cell_domains))/v_dx1 for i in xrange(D)]
avg_ds = [assemble(uv[i]*ds)/v_ds for i in xrange(D)]
avg_ds0 = [assemble(uv[i]*ds(0, subdomain_data=facet_domains))/v_ds0 for i in xrange(D)]
avg_ds1 = [assemble(uv[i]*ds(1, subdomain_data=facet_domains))/v_ds1 for i in xrange(D)]
assert max(abs(x-y) for x,y in zip(avg, pp.get("DomainAvg_MockVectorFunctionField"))) < 1e-8
assert max(abs(x-y) for x,y in zip(avg_dx0, pp.get("DomainAvg_MockVectorFunctionField-dx0"))) < 1e-8
assert max(abs(x-y) for x,y in zip(avg_dx1, pp.get("DomainAvg_MockVectorFunctionField-dx1"))) < 1e-8
assert max(abs(x-y) for x,y in zip(avg_ds, pp.get("DomainAvg_MockVectorFunctionField-ds"))) < 1e-8
assert max(abs(x-y) for x,y in zip(avg_ds0, pp.get("DomainAvg_MockVectorFunctionField-ds0"))) < 1e-8
assert max(abs(x-y) for x,y in zip(avg_ds1, pp.get("DomainAvg_MockVectorFunctionField-ds1"))) < 1e-8
std = [sqrt(assemble((uv[i]-Constant(avg[i]))**2*dx)/v) for i in xrange(D)]
std_dx0 = [sqrt(assemble((uv[i]-Constant(avg_dx0[i]))**2*dx(0, subdomain_data=cell_domains))/v_dx0) for i in xrange(D)]
std_dx1 = [sqrt(assemble((uv[i]-Constant(avg_dx1[i]))**2*dx(1, subdomain_data=cell_domains))/v_dx1) for i in xrange(D)]
std_ds = [sqrt(assemble((uv[i]-Constant(avg_ds[i]))**2*ds)/v_ds) for i in xrange(D)]
std_ds0 = [sqrt(assemble((uv[i]-Constant(avg_ds0[i]))**2*ds(0, subdomain_data=facet_domains))/v_ds0) for i in xrange(D)]
std_ds1 = [sqrt(assemble((uv[i]-Constant(avg_ds1[i]))**2*ds(1, subdomain_data=facet_domains))/v_ds1) for i in xrange(D)]
assert max(abs(x-y) for x,y in zip(std, pp.get("DomainSD_MockVectorFunctionField"))) < 1e-8
assert max(abs(x-y) for x,y in zip(std_dx0, pp.get("DomainSD_MockVectorFunctionField-dx0"))) < 1e-8
assert max(abs(x-y) for x,y in zip(std_dx1, pp.get("DomainSD_MockVectorFunctionField-dx1"))) < 1e-8
assert max(abs(x-y) for x,y in zip(std_ds, pp.get("DomainSD_MockVectorFunctionField-ds"))) < 1e-8
assert max(abs(x-y) for x,y in zip(std_ds0, pp.get("DomainSD_MockVectorFunctionField-ds0"))) < 1e-8
assert max(abs(x-y) for x,y in zip(std_ds1, pp.get("DomainSD_MockVectorFunctionField-ds1"))) < 1e-8
def test_ErrorNorm(problem, pp, start_time, end_time, dt):
# Setup some mock scheme state
dt, timesteps, start_timestep = compute_regular_timesteps(problem)
spacepool = SpacePool(problem.mesh)
Q = spacepool.get_space(1,0)
V = spacepool.get_space(1,1)
class MockFunctionField2(MockFunctionField):
def before_first_compute(self, get):
t = get('t')
self.expr = Expression("1+x[0]*x[1]*t+0.4", degree=1, t=t)
class MockVectorFunctionField2(MockVectorFunctionField):
def before_first_compute(self, get):
t = get('t')
D = self.f.function_space().mesh().geometry().dim()
if D == 2:
self.expr = Expression(("1+x[0]*t+0.2", "3+x[1]*t+0.3"), degree=1, t=t)
elif D == 3:
self.expr = Expression(("1+x[0]*t+0.2", "3+x[1]*t+0.3", "10+x[2]*t+0.4"), degree=1, t=t)
class MockTupleField2(MockTupleField):
def compute(self, get):
return (t+0.2, 3*t+0.3, 1+5*t+0.4)
class MockScalarField2(MockScalarField):
def compute(self, get):
t = get('t')
return 3*t**0.5+0.3
pp.add_fields([
MockFunctionField(Q),
MockVectorFunctionField(V),
MockTupleField(),
MockScalarField(),
])
pp.add_fields([
MockFunctionField2(Q),
MockVectorFunctionField2(V),
MockTupleField2(),
MockScalarField2(),
])
pp.add_fields([
ErrorNorm("MockFunctionField", "MockFunctionField2"),
ErrorNorm("MockVectorFunctionField", "MockVectorFunctionField2"),
ErrorNorm("MockTupleField", "MockTupleField2"),
ErrorNorm("MockScalarField", "MockScalarField2"),
])
D = problem.D
# Update postprocessor for a number of timesteps, this is where the main code under test is
for timestep, t in enumerate(timesteps, start_timestep):
# Run postprocessing step
pp.update_all({}, t, timestep)
if start_time < t < end_time:
assert abs(pp.get("ErrorNorm_MockFunctionField_MockFunctionField2") - \
norm(interpolate(Expression("0.4", degree=1), Q))) < 1e-8
assert abs(pp.get("ErrorNorm_MockVectorFunctionField_MockVectorFunctionField2") - \
norm(interpolate(Expression(["0.2", "0.3", "0.4"][:D], degree=1), V))) < 1e-8
assert abs(pp.get("ErrorNorm_MockTupleField_MockTupleField2") - \
sum([0.2**2, 0.3**2, 0.4**2])**0.5) < 1e-8
assert abs(pp.get("ErrorNorm_MockScalarField_MockScalarField2")-0.3) < 1e-8
pp.finalize_all()
assert abs(pp.get("ErrorNorm_MockFunctionField_MockFunctionField2") - \
norm(interpolate(Expression("0.4", degree=1), Q))) < 1e-8
assert abs(pp.get("ErrorNorm_MockVectorFunctionField_MockVectorFunctionField2") - \
norm(interpolate(Expression(["0.2", "0.3", "0.4"][:D], degree=1), V))) < 1e-8
assert abs(pp.get("ErrorNorm_MockTupleField_MockTupleField2") - \
sum([0.2**2, 0.3**2, 0.4**2])**0.5) < 1e-8
assert abs(pp.get("ErrorNorm_MockScalarField_MockScalarField2")-0.3) < 1e-8
def test_PointEval(problem, pp, start_time, end_time, dt):
# Setup some mock scheme state
dt, timesteps, start_timestep = compute_regular_timesteps(problem)
spacepool = SpacePool(problem.mesh)
Q = spacepool.get_space(2,0)
V = spacepool.get_space(2,1)
pp.add_fields([
MockFunctionField(Q),
MockVectorFunctionField(V),
MockTupleField(),
])
D = problem.mesh.geometry().dim()
xmin = MPI.min(mpi_comm_world(), min(problem.mesh.coordinates()[:,0]))
ymin = MPI.min(mpi_comm_world(), min(problem.mesh.coordinates()[:,1]))
if problem.D > 2:
zmin = MPI.min(mpi_comm_world(), min(problem.mesh.coordinates()[:,2]))
xmax = MPI.max(mpi_comm_world(), max(problem.mesh.coordinates()[:,0]))
ymax = MPI.max(mpi_comm_world(), max(problem.mesh.coordinates()[:,1]))
if problem.D > 2:
zmax = MPI.max(mpi_comm_world(), max(problem.mesh.coordinates()[:,2]))
points = []
#for i in range(5):
bbtree = problem.mesh.bounding_box_tree()
while len(points) < 5:
p = [xmin+(xmax-xmin)*random(), ymin+(ymax-ymin)*random()]
if D == 3:
p += [zmin+(zmax-xmin)*random()]
points.append(p)
pp.add_fields([
PointEval("MockFunctionField", points),
PointEval("MockVectorFunctionField", points),
])
for timestep, t in enumerate(timesteps, start_timestep):
# Run postprocessing step
pp.update_all({}, t, timestep)
if start_time < t < end_time:
pevalfunction = pp.get("PointEval_MockFunctionField")
pevalvectorfunction = pp.get("PointEval_MockVectorFunctionField")
for i, p in enumerate(points):
if D == 2:
x,y = p
assert abs( pevalfunction[i] - (1+x*y*t) ) < 1e-10
assert sum( abs( pevalvectorfunction[i][k] - (1+x*t, 3+y*t)[k]) for k in range(D)) < 1e-10
elif D == 3:
x,y,z = p
assert abs( pevalfunction[i] - (1+x*y*t) ) < 1e-10
assert sum( abs( pevalvectorfunction[i][k] - (1+x*t, 3+y*t, 10+z*t)[k]) for k in range(D)) < 1e-10
pp.finalize_all()
pevalfunction = pp.get("PointEval_MockFunctionField")
pevalvectorfunction = pp.get("PointEval_MockVectorFunctionField")
for i, p in enumerate(points):
if D == 2:
x,y = p
assert abs( pevalfunction[i] - (1+x*y*t) ) < 1e-10
assert sum( abs( pevalvectorfunction[i][k] - (1+x*t, 3+y*t)[k]) for k in range(D)) < 1e-10
elif D == 3:
x,y,z = p
assert abs( pevalfunction[i] - (1+x*y*t) ) < 1e-10
assert sum( abs( pevalvectorfunction[i][k] - (1+x*t, 3+y*t, 10+z*t)[k]) for k in range(D)) < 1e-10
def test_Norm(problem, pp, start_time, end_time, dt):
# Setup some mock scheme state
dt, timesteps, start_timestep = compute_regular_timesteps(problem)
spacepool = SpacePool(problem.mesh)
Q = spacepool.get_space(1,0)
V = spacepool.get_space(1,1)
pp.add_fields([
MockFunctionField(Q),
MockVectorFunctionField(V),
MockTupleField(),
])
pp.add_fields([
Norm("t"),
Norm("t", dict(norm_type='l2')),
Norm("t", dict(norm_type='linf')),
Norm("MockFunctionField"),
Norm("MockFunctionField", dict(norm_type='L2')),
Norm("MockFunctionField", dict(norm_type='H10')),
Norm("MockVectorFunctionField"),
Norm("MockVectorFunctionField", dict(norm_type='L2')),
Norm("MockVectorFunctionField", dict(norm_type='H10')),
Norm("MockTupleField"),
Norm("MockTupleField", dict(norm_type='l4')),
Norm("MockTupleField", dict(norm_type='linf')),
])
D = problem.mesh.geometry().dim()
# Update postprocessor for a number of timesteps, this is where the main code under test is
for timestep, t in enumerate(timesteps, start_timestep):
# Run postprocessing step
pp.update_all({}, t, timestep)
if start_time < t < end_time:
assert abs(pp.get("Norm_t") - t) < 1e-14
assert abs(pp.get("Norm_l2_t") - t) < 1e-14
assert abs(pp.get("Norm_linf_t") - t) < 1e-14
assert abs(pp.get("Norm_MockFunctionField") - norm(interpolate(Expression("1+x[0]*x[1]*t", degree=1, t=t), Q))) < 1e-14
if D == 2:
assert abs(pp.get("Norm_MockVectorFunctionField") - norm(interpolate(Expression(("1+x[0]*t", "3+x[1]*t"), degree=1, t=t), V))) < 1e-14
assert abs(pp.get("Norm_L2_MockVectorFunctionField") - norm(interpolate(Expression(("1+x[0]*t", "3+x[1]*t"), degree=1, t=t), V), 'L2')) < 1e-14
assert abs(pp.get("Norm_H10_MockVectorFunctionField") - norm(interpolate(Expression(("1+x[0]*t", "3+x[1]*t"), degree=1, t=t), V), 'H10')) < 1e-14
elif D == 3:
assert abs(pp.get("Norm_MockVectorFunctionField") - norm(interpolate(Expression(("1+x[0]*t", "3+x[1]*t", "10+x[2]*t"), degree=1, t=t), V))) < 1e-14
assert abs(pp.get("Norm_L2_MockVectorFunctionField") - norm(interpolate(Expression(("1+x[0]*t", "3+x[1]*t", "10+x[2]*t"), degree=1, t=t), V), 'L2')) < 1e-14
assert abs(pp.get("Norm_H10_MockVectorFunctionField") - norm(interpolate(Expression(("1+x[0]*t", "3+x[1]*t", "10+x[2]*t"), degree=1, t=t), V), 'H10')) < 1e-14
assert abs(pp.get("Norm_MockTupleField") - sqrt(t**2+(3*t)**2+(1+5*t)**2)) < 1e-14
assert abs(pp.get("Norm_l4_MockTupleField") - (t**4+(3*t)**4+(1+5*t)**4)**(0.25)) < 1e-14
assert abs(pp.get("Norm_linf_MockTupleField") - (1+5*t)) < 1e-14
pp.finalize_all()
assert abs(pp.get("Norm_t") - t) < 1e-14
assert abs(pp.get("Norm_l2_t") - t) < 1e-14
assert abs(pp.get("Norm_linf_t") - t) < 1e-14
assert abs(pp.get("Norm_MockFunctionField") - norm(interpolate(Expression("1+x[0]*x[1]*t", degree=1, t=t), Q))) < 1e-14
if D == 2:
assert abs(pp.get("Norm_MockVectorFunctionField") - norm(interpolate(Expression(("1+x[0]*t", "3+x[1]*t"), degree=1, t=t), V))) < 1e-14
assert abs(pp.get("Norm_L2_MockVectorFunctionField") - norm(interpolate(Expression(("1+x[0]*t", "3+x[1]*t"), degree=1, t=t), V), 'L2')) < 1e-14
assert abs(pp.get("Norm_H10_MockVectorFunctionField") - norm(interpolate(Expression(("1+x[0]*t", "3+x[1]*t"), degree=1, t=t), V), 'H10')) < 1e-14
elif D == 3:
assert abs(pp.get("Norm_MockVectorFunctionField") - norm(interpolate(Expression(("1+x[0]*t", "3+x[1]*t", "10+x[2]*t"), degree=1, t=t), V))) < 1e-14
assert abs(pp.get("Norm_L2_MockVectorFunctionField") - norm(interpolate(Expression(("1+x[0]*t", "3+x[1]*t", "10+x[2]*t"), degree=1, t=t), V), 'L2')) < 1e-14
assert abs(pp.get("Norm_H10_MockVectorFunctionField") - norm(interpolate(Expression(("1+x[0]*t", "3+x[1]*t", "10+x[2]*t"), degree=1, t=t), V), 'H10')) < 1e-14
assert abs(pp.get("Norm_MockTupleField") - sqrt(t**2+(3*t)**2+(1+5*t)**2)) < 1e-14
assert abs(pp.get("Norm_l4_MockTupleField") - (t**4+(3*t)**4+(1+5*t)**4)**(0.25)) < 1e-14
assert abs(pp.get("Norm_linf_MockTupleField") - (1+5*t)) < 1e-14
def test_TimeIntegral_of_TimeDerivative(problem, pp, start_time, end_time, dt):
# Setup some mock scheme state
dt, timesteps, start_timestep = compute_regular_timesteps(problem)
spacepool = SpacePool(problem.mesh)
Q = spacepool.get_space(1,0)
V = spacepool.get_space(1,1)
pp.add_fields([
MockFunctionField(Q),
MockVectorFunctionField(V),
MockTupleField(),
])
pp.add_fields([
TimeDerivative("t"),
TimeDerivative("MockFunctionField"),
TimeDerivative("MockVectorFunctionField"),
TimeDerivative("MockTupleField"),
])
params = dict(start_time=start_time, end_time=end_time, finalize=False)
pp.add_fields([
TimeIntegral("TimeDerivative_t", params),
TimeIntegral("TimeDerivative_MockFunctionField", params),
TimeIntegral("TimeDerivative_MockVectorFunctionField", params),
TimeIntegral("TimeDerivative_MockTupleField", params),
])
# Because of backward finite differencing in TimeDerivative, a numerical error will be introduced if start_time<dt
err_factor = max(0.0, dt-start_time)
err_t = err_factor*0.5*(1-start_time/dt)
err_MockFunctionField = err_factor*norm(interpolate(Expression("0.5*x[0]*x[1]*(1-t0/dt)", degree=1, dt=dt, t0=start_time), Q))
D = problem.mesh.geometry().dim()
if D == 2:
err_MockVectorFunctionField = err_factor*norm(interpolate(Expression(("0.5*x[0]*(1-t0/dt)", "0.5*x[1]*(1-t0/dt)"), degree=1, dt=dt, t0=start_time), V))
elif D == 3:
err_MockVectorFunctionField = err_factor*norm(interpolate(Expression(("0.5*x[0]*(1-t0/dt)", "0.5*x[1]*(1-t0/dt)", "0.5*x[2]*(1-t0/dt)"), degree=1, dt=dt, t0=start_time), V))
else:
raise Exception("D must be 2 or 3")
err_MockTupleField = tuple([err_factor*(1-start_time/dt)*x/2.0 for x in [1.0, 3.0, 5.0]])
if start_time > dt:
assert err_factor == 0.0
# Update postprocessor for a number of timesteps, this is where the main code under test is
for timestep, t in enumerate(timesteps, start_timestep):
# Run postprocessing step
pp.update_all({}, t, timestep)
if start_time < t < end_time:
assert abs( pp.get("TimeIntegral_TimeDerivative_t") - ((t-start_time)-err_t) ) < 1e-8
pp.finalize_all()
assert err_t-1e-8 < abs( pp.get("TimeIntegral_TimeDerivative_t") - (end_time-start_time)) < err_t+1e-8
assert err_MockFunctionField-1e-8 < \
errornorm(pp.get("TimeIntegral_TimeDerivative_MockFunctionField"),
interpolate(Expression("x[0]*x[1]*(t1-t0)", degree=1, t0=start_time, t1=end_time), Q)
) < \
err_MockFunctionField+1e-8
if D == 2:
assert err_MockVectorFunctionField-1e-8 < \
errornorm(pp.get("TimeIntegral_TimeDerivative_MockVectorFunctionField"),
interpolate(Expression(("x[0]*(t1-t0)", "x[1]*(t1-t0)"), degree=1, t0=start_time, t1=end_time), V)
) < \
err_MockVectorFunctionField+1e-8
elif D == 3:
assert err_MockVectorFunctionField-1e-8 < \
errornorm(pp.get("TimeIntegral_TimeDerivative_MockVectorFunctionField"),
interpolate(Expression(("x[0]*(t1-t0)", "x[1]*(t1-t0)", "x[2]*(t1-t0)"), degree=1, t0=start_time, t1=end_time), V)
) < \
err_MockVectorFunctionField+1e-8
else:
raise Exception("D must be 2 or 3")
I = end_time-start_time
assert abs( abs(pp.get("TimeIntegral_TimeDerivative_MockTupleField")[0] - I) - err_MockTupleField[0]) < 1e-8
assert abs( abs(pp.get("TimeIntegral_TimeDerivative_MockTupleField")[1] - 3*I) - err_MockTupleField[1]) < 1e-8
assert abs( abs(pp.get("TimeIntegral_TimeDerivative_MockTupleField")[2] - 5*I) - err_MockTupleField[2]) < 1e-8