Homec4science
Diffusion akantu (features/cleaning-dead-code) Browse

rAKA/test/test_model/patch_tests/patch_test_linear_fixture.py
5671899d4a2afeatures/cleaning-dead…

Last Change
Raw File

patch_test_linear_fixture.py
View Options

#!/usr/bin/env python3
# ------------------------------------------------------------------------------
__author__ = "Guillaume Anciaux"
__copyright__ = "Copyright (C) 2016-2018, EPFL (Ecole Polytechnique Fédérale" \
" de Lausanne) Laboratory (LSMS - Laboratoire de Simulation" \
" en Mécanique des Solides)"
__credits__ = ["Guillaume Anciaux"]
__license__ = "L-GPLv3"
__maintainer__ = "Guillaume Anciaux"
__email__ = "guillaume.anciaux@epfl.ch"
# ------------------------------------------------------------------------------
from mpi4py import MPI
import akantu
import unittest
import numpy as np
class TestPatchTestLinear(unittest.TestCase):
alpha = np.array([[0.01, 0.02, 0.03, 0.04],
[0.05, 0.06, 0.07, 0.08],
[0.09, 0.10, 0.11, 0.12]])
gradient_tolerance = 1e-13
result_tolerance = 1e-13
dofs_tolerance = 1e-15
def __init__(self, test_name, elem_type_str, functor=None):
self.test_name = test_name
self.functor = functor
self.elem_type = akantu.getElementTypes()[elem_type_str]
self.elem_type_str = elem_type_str
self.dim = akantu.Mesh.getSpatialDimension(self.elem_type)
super().__init__(test_name)
def _internal_call(self):
self.functor(self)
def __getattr__(self, key):
if key == self.test_name:
return self._internal_call
def setUp(self):
self.mesh = akantu.Mesh(self.dim, self.elem_type_str)
self.mesh.read(str(self.elem_type_str) + ".msh")
akantu.MeshUtils.buildFacets(self.mesh)
self.mesh.createBoundaryGroupFromGeometry()
self.model = self.model_type(self.mesh)
def tearDown(self):
del self.model
del self.mesh
def initModel(self, method, material_file):
akantu.parseInput(material_file)
akantu.setDebugLevel(akantu.dblError)
self.model.initFull(method)
self.applyBC()
if method != akantu._static:
self.model.setTimeStep(0.8 * self.model.getStableTimeStep())
def applyBC(self):
boundary = self.model.getBlockedDOFs()
for eg in self.mesh.iterateElementGroups():
nodes = eg.getNodeGroup().getNodes()
boundary[nodes, :] = True
def applyBConDOFs(self, dofs):
coordinates = self.mesh.getNodes()
for eg in self.mesh.iterateElementGroups():
nodes = eg.getNodeGroup().getNodes().flatten()
dofs[nodes] = self.setLinearDOF(dofs[nodes],
coordinates[nodes])
def prescribed_gradient(self, dof):
gradient = self.alpha[:dof.shape[1], 1:self.dim + 1]
return gradient
def checkGradient(self, gradient, dofs):
pgrad = self.prescribed_gradient(dofs).T
gradient = gradient.reshape(gradient.shape[0], *pgrad.shape)
diff = gradient[:] - pgrad
norm = np.abs(pgrad).max()
gradient_error = np.abs(diff).max() / norm
self.assertAlmostEqual(0, gradient_error,
delta=self.gradient_tolerance)
def checkResults(self, presult_func, results, dofs):
presult = presult_func(self.prescribed_gradient(dofs)).flatten()
remaining_size = np.prod(np.array(results.shape[1:]))
results = results.reshape((results.shape[0], remaining_size))
for result in results:
diff = result - presult
norm = np.abs(result).max()
if norm == 0:
result_error = np.abs(diff).max()
else:
result_error = np.abs(diff).max() / norm
self.assertAlmostEqual(0., result_error,
delta=self.result_tolerance)
def setLinearDOF(self, dof, coord):
nb_dofs = dof.shape[1]
dof[:] = np.einsum('ik,ak->ai',
self.alpha[:nb_dofs, 1:self.dim + 1], coord)
for i in range(0, nb_dofs):
dof[:, i] += self.alpha[i, 0]
return dof
def checkDOFs(self, dofs):
coordinates = self.mesh.getNodes()
ref_dofs = np.zeros_like(dofs)
self.setLinearDOF(ref_dofs, coordinates)
diff = dofs - ref_dofs
dofs_error = np.abs(diff).max()
self.assertAlmostEqual(0., dofs_error, delta=self.dofs_tolerance)
@classmethod
def TYPED_TEST(cls, functor, label):
for type_name, _type in akantu.getElementTypes().items():
if type_name == "_point_1":
continue
method_name = type_name + '_' + label
test_case = cls(method_name, type_name, functor)
test_case.setUp()
functor(test_case)
test_case.tearDown()
c4science · Help