def test_copy(self):
l1 = Library(Dimension('x', np.linspace(0, 1, 10)),
Dimension('y', np.linspace(-1, 1, 4)),
Dimension('z', np.logspace(-1, 1, 7)))
l1['f'] = np.exp(l1.x_grid) * np.cos(l1.y_grid) * np.log(l1.z_grid)
l1.extra_attributes['name'] = 'my_library_name'
l2 = copy(l1)
l3 = Library.copy(l1)
self.assertTrue(
np.all(np.abs(l1.x_grid - l2.x_grid) < 10. * machine_epsilon))
self.assertTrue(
np.all(np.abs(l1.y_grid - l2.y_grid) < 10. * machine_epsilon))
self.assertTrue(
np.all(np.abs(l1.z_grid - l2.z_grid) < 10. * machine_epsilon))
self.assertTrue(
np.all(np.abs(l1['f'] - l2['f']) < 10. * machine_epsilon))
self.assertTrue(
np.all(np.abs(l1.x_grid - l3.x_grid) < 10. * machine_epsilon))
self.assertTrue(
np.all(np.abs(l1.y_grid - l3.y_grid) < 10. * machine_epsilon))
self.assertTrue(
np.all(np.abs(l1.z_grid - l3.z_grid) < 10. * machine_epsilon))
self.assertTrue(
np.all(np.abs(l1['f'] - l3['f']) < 10. * machine_epsilon))
self.assertTrue(
l1.extra_attributes['name'] == l2.extra_attributes['name'])
self.assertTrue(
l1.extra_attributes['name'] == l3.extra_attributes['name'])
def test_single_internal(self):
l1 = Library(Dimension('x', np.linspace(0, 1, 10)),
Dimension('y', np.linspace(-1, 1, 4)),
Dimension('z', np.logspace(-1, 1, 7)))
l1['f'] = np.exp(l1.x_grid) * np.cos(l1.y_grid) * np.log(l1.z_grid)
l1.extra_attributes['name'] = 'my_library_name'
n1x = 2
n2x = 8
n1y = 1
n2y = n1y + 1
n1z = 3
n2z = 6
l2 = l1[n1x:n2x, n1y:n2y, n1z:n2z]
self.assertTrue(
np.all(
np.abs(l1.x_grid[n1x:n2x, n1y:n2y, n1z:n2z] -
l2.x_grid) < 10. * machine_epsilon))
self.assertTrue(
np.all(
np.abs(l1['f'][n1x:n2x, n1y:n2y, n1z:n2z] - l2['f']) < 10. *
machine_epsilon))
self.assertTrue(l2.shape == l1.x_grid[n1x:n2x, n1y:n2y, n1z:n2z].shape)
self.assertTrue(l2.size == l1.x_grid[n1x:n2x, n1y:n2y, n1z:n2z].size)
self.assertTrue(
l1.extra_attributes['name'] == l2.extra_attributes['name'])
def test_full(self):
l1 = Library(Dimension('x', np.linspace(0, 1, 2)),
Dimension('y', np.linspace(-1, 1, 3)),
Dimension('z', np.logspace(-1, 1, 4)))
l1['f'] = np.exp(l1.x_grid) * np.cos(l1.y_grid) * np.log(l1.z_grid)
l1.extra_attributes['name'] = 'my_library_name'
l2 = l1[:, :, :]
self.assertTrue(
np.all(np.abs(l1.x_grid - l2.x_grid) < 10. * machine_epsilon))
self.assertTrue(
np.all(np.abs(l1.y_grid - l2.y_grid) < 10. * machine_epsilon))
self.assertTrue(
np.all(np.abs(l1.z_grid - l2.z_grid) < 10. * machine_epsilon))
self.assertTrue(
np.all(np.abs(l1['f'] - l2['f']) < 10. * machine_epsilon))
self.assertTrue(l2.shape == l1.x_grid.shape)
self.assertTrue(l2.size == l1.x_grid.size)
self.assertTrue(
l1.extra_attributes['name'] == l2.extra_attributes['name'])
l3 = l1[:]
self.assertTrue(
np.all(np.abs(l1.x_grid - l3.x_grid) < 10. * machine_epsilon))
self.assertTrue(
np.all(np.abs(l1.y_grid - l3.y_grid) < 10. * machine_epsilon))
self.assertTrue(
np.all(np.abs(l1.z_grid - l3.z_grid) < 10. * machine_epsilon))
self.assertTrue(
np.all(np.abs(l1['f'] - l3['f']) < 10. * machine_epsilon))
self.assertTrue(l3.shape == l1.x_grid.shape)
self.assertTrue(l3.size == l1.x_grid.size)
self.assertTrue(
l1.extra_attributes['name'] == l3.extra_attributes['name'])
def test_full(self):
l1 = Library(Dimension('x', np.linspace(0, 1, 16)))
l1['f'] = np.exp(l1.x_grid)
l1.extra_attributes['name'] = 'my_library_name'
l2 = l1[:]
self.assertTrue(
np.all(np.abs(l1.x_grid - l2.x_grid) < 10. * machine_epsilon))
self.assertTrue(
np.all(np.abs(l1['f'] - l2['f']) < 10. * machine_epsilon))
self.assertTrue(l2.shape == l1.x_grid.shape)
self.assertTrue(l2.size == l1.x_grid.size)
self.assertTrue(
l1.extra_attributes['name'] == l2.extra_attributes['name'])
def test_multiple_nonslice_args_2(self):
l1 = Library(Dimension('x', np.linspace(0, 1, 16)))
try:
l2 = l1[:, 'g']
self.assertTrue(False)
except LibraryIndexError:
self.assertTrue(True)
def test_invalid_number(self):
l1 = Library(Dimension('x', np.linspace(0, 1, 16)))
try:
l2 = l1[:, :]
self.assertTrue(False)
except LibraryIndexError:
self.assertTrue(True)
def test_single(self):
l1 = Library(Dimension('x', np.linspace(0, 1, 16)))
l1['f'] = np.exp(l1.x_grid)
l1.extra_attributes['name'] = 'my_library_name'
n1 = 2
n2 = n1 + 1
l2 = l1[n1:n2]
self.assertTrue(
np.all(
np.abs(l1.x_grid[n1:n2] - l2.x_grid) < 10. * machine_epsilon))
self.assertTrue(
np.all(np.abs(l1['f'][n1:n2] - l2['f']) < 10. * machine_epsilon))
self.assertTrue(l2.shape == l1.x_grid[n1:n2].shape)
self.assertTrue(l2.size == l1.x_grid[n1:n2].size)
self.assertTrue(
l1.extra_attributes['name'] == l2.extra_attributes['name'])
def test_save_to_text(self):
xvalues = np.linspace(0, 1, 16)
yvalues = np.linspace(1, 2, 8)
zvalues = np.linspace(2, 3, 4)
l1 = Library(Dimension('x', xvalues),
Dimension('y', yvalues),
Dimension('z', zvalues))
fvalues = l1.x_grid + l1.y_grid + l1.z_grid
gvalues = np.exp(l1.x_grid) * np.cos(np.pi * 2. * l1.y_grid) * l1.z_grid
lib_shape = fvalues.shape
l1['f'] = fvalues
l1['g'] = gvalues
lib_name = 'my_library_name'
l1.extra_attributes['name'] = lib_name
dir_name = 'out'
l1.save_to_text_directory(dir_name, ravel_order='F')
xread = np.loadtxt(dir_name + f'/bulkdata_ivar_x.txt')
yread = np.loadtxt(dir_name + f'/bulkdata_ivar_y.txt')
zread = np.loadtxt(dir_name + f'/bulkdata_ivar_z.txt')
fread = np.loadtxt(dir_name + f'/bulkdata_dvar_f.txt').reshape(lib_shape, order='F')
gread = np.loadtxt(dir_name + f'/bulkdata_dvar_g.txt').reshape(lib_shape, order='F')
with open(dir_name + '/metadata_user_defined_attributes.txt', 'r') as f:
ea_read = f.readline()
with open(dir_name + '/metadata_independent_variables.txt', 'r') as f:
iv_lines = f.readlines()
with open(dir_name + '/metadata_dependent_variables.txt', 'r') as f:
dv_lines = f.readlines()
rmtree(dir_name)
self.assertTrue(np.all(np.abs(xvalues - xread) < 100. * machine_epsilon))
self.assertTrue(np.all(np.abs(yvalues - yread) < 100. * machine_epsilon))
self.assertTrue(np.all(np.abs(zvalues - zread) < 100. * machine_epsilon))
self.assertTrue(np.all(np.abs(fvalues - fread) < 100. * machine_epsilon))
self.assertTrue(np.all(np.abs(gvalues - gread) < 100. * machine_epsilon))
self.assertTrue(ea_read, str(l1.extra_attributes))
self.assertTrue(all([ivf.strip() == ivn for (ivf, ivn) in zip(iv_lines, [d.name for d in l1.dims])]))
self.assertTrue(all([dvf.strip() == dvn.replace(' ', '_') for (dvf, dvn) in zip(dv_lines, l1.props)]))
def test_squeeze(self):
l1 = Library(Dimension('x', np.linspace(0, 1, 10)),
Dimension('y', np.linspace(-1, 1, 10)))
l1['f'] = np.exp(l1.x_grid) * np.cos(l1.y_grid)
l1.extra_attributes['name'] = 'my_library_name'
iy = 3
l3 = Library.squeeze(l1[:, iy])
self.assertTrue(
np.all(np.abs(l1['f'][:, iy] - l3['f']) < 10. * machine_epsilon))
self.assertTrue(
np.all(
np.abs(np.squeeze(l1.x_grid[:, iy]) - l3.x_grid) < 10. *
machine_epsilon))
self.assertTrue(
l1.extra_attributes['name'] == l3.extra_attributes['name'])
def test_copy(self):
l1 = Library(Dimension('x', np.linspace(0, 1, 16)))
l1['f'] = np.exp(l1.x_grid)
l1.extra_attributes['name'] = 'my_library_name'
l2 = copy(l1)
l3 = deepcopy(l1)
l4 = Library.deepcopy(l1)
self.assertTrue(
np.all(np.abs(l1.x_grid - l2.x_grid) < 10. * machine_epsilon))
self.assertTrue(
np.all(np.abs(l1['f'] - l2['f']) < 10. * machine_epsilon))
self.assertTrue(
np.all(np.abs(l1.x_grid - l3.x_grid) < 10. * machine_epsilon))
self.assertTrue(
np.all(np.abs(l1['f'] - l3['f']) < 10. * machine_epsilon))
self.assertTrue(
np.all(np.abs(l1.x_grid - l4.x_grid) < 10. * machine_epsilon))
self.assertTrue(
np.all(np.abs(l1['f'] - l4['f']) < 10. * machine_epsilon))
self.assertTrue(
l1.extra_attributes['name'] == l2.extra_attributes['name'])
def test_single(self):
l1 = Library(Dimension('x', np.linspace(0, 1, 10)),
Dimension('y', np.linspace(-1, 1, 10)))
l1['f'] = np.exp(l1.x_grid) * np.cos(l1.y_grid)
l1.extra_attributes['name'] = 'my_library_name'
n1x = 2
n2x = n1x + 1
n1y = 1
n2y = -1
l2 = l1[n1x:n2x, n1y:n2y]
self.assertTrue(
np.all(
np.abs(l1.x_grid[n1x:n2x, n1y:n2y] - l2.x_grid) < 10. *
machine_epsilon))
self.assertTrue(
np.all(
np.abs(l1['f'][n1x:n2x, n1y:n2y] - l2['f']) < 10. *
machine_epsilon))
self.assertTrue(l2.shape == l1.x_grid[n1x:n2x, n1y:n2y].shape)
self.assertTrue(l2.size == l1.x_grid[n1x:n2x, n1y:n2y].size)
self.assertTrue(
l1.extra_attributes['name'] == l2.extra_attributes['name'])
def test_raw_pickle_3d(self):
file_name = 'l1test.pkl'
if isfile(file_name):
remove(file_name)
l1 = Library(Dimension('x', np.linspace(0, 1, 16)),
Dimension('y', np.linspace(1, 2, 8)),
Dimension('z', np.linspace(2, 3, 4)))
l1['f'] = l1.x_grid + l1.y_grid + l1.z_grid
l1['g'] = np.exp(l1.x_grid) * np.cos(np.pi * 2. * l1.y_grid) * l1.z_grid
l1.extra_attributes['name'] = 'my_library_name'
with open(file_name, 'wb') as f:
pickle.dump(l1, f)
with open(file_name, 'rb') as f:
l2 = pickle.load(f)
remove(file_name)
self.assertTrue(np.all(np.abs(l1['f'] - l2['f']) < 10. * machine_epsilon))
self.assertTrue(np.all(np.abs(l1['g'] - l2['g']) < 10. * machine_epsilon))
self.assertTrue(l1.extra_attributes['name'] == l2.extra_attributes['name'])
def test_save_and_load_1d(self):
file_name = 'l1test.pkl'
if isfile(file_name):
remove(file_name)
l1 = Library(Dimension('x', np.linspace(0, 1, 16)))
l1['f'] = l1.x_grid
l1['g'] = np.exp(l1.x_grid)
l1.extra_attributes['name'] = 'my_library_name'
l1.save_to_file(file_name)
l2 = Library.load_from_file(file_name)
remove(file_name)
self.assertTrue(np.all(np.abs(l1['f'] - l2['f']) < 10. * machine_epsilon))
self.assertTrue(np.all(np.abs(l1['g'] - l2['g']) < 10. * machine_epsilon))
self.assertTrue(l1.extra_attributes['name'] == l2.extra_attributes['name'])
def test_view(self):
slices = (slice(0, None, None), slice(1, 3, None), slice(1, -3, None))
l1 = Library(Dimension('x', np.linspace(0, 1, 10)),
Dimension('y', np.linspace(-1, 1, 4)),
Dimension('z', np.logspace(-1, 1, 7)))
gold_float = 0.5
gold_array = np.exp(-2. * l1.x_grid[slices])
fvals = np.exp(l1.x_grid) * np.cos(l1.y_grid) * np.log(l1.z_grid)
# start with float argument
g = gold_float
# set slice of original array
fvals[:, :, :] = np.exp(l1.x_grid) * np.cos(l1.y_grid) * np.log(
l1.z_grid)
l1['f'] = fvals
l2 = l1[slices]
fvals[slices] = g
self.assertTrue(
np.all(np.abs(l1['f'][slices] - g) < 10. * machine_epsilon))
self.assertTrue(
np.all(np.abs(l1[slices]['f'] - g) < 10. * machine_epsilon))
self.assertTrue(
np.all(np.abs(fvals[slices] - g) < 10. * machine_epsilon))
self.assertTrue(np.all(np.abs(l2['f'] - g) < 10. * machine_epsilon))
# set slice of library property
fvals[:, :, :] = np.exp(l1.x_grid) * np.cos(l1.y_grid) * np.log(
l1.z_grid)
l1['f'] = fvals
l2 = l1[slices]
l1['f'][slices] = g
self.assertTrue(
np.all(np.abs(l1['f'][slices] - g) < 10. * machine_epsilon))
self.assertTrue(
np.all(np.abs(l1[slices]['f'] - g) < 10. * machine_epsilon))
self.assertTrue(
np.all(np.abs(fvals[slices] - g) < 10. * machine_epsilon))
self.assertTrue(np.all(np.abs(l2['f'] - g) < 10. * machine_epsilon))
# set property of library slice
fvals[:, :, :] = np.exp(l1.x_grid) * np.cos(l1.y_grid) * np.log(
l1.z_grid)
l1['f'] = fvals
l2 = l1[slices]
l1[slices]['f'] = g
self.assertTrue(
np.all(np.abs(l1['f'][slices] - g) < 10. * machine_epsilon))
self.assertTrue(
np.all(np.abs(l1[slices]['f'] - g) < 10. * machine_epsilon))
self.assertTrue(
np.all(np.abs(fvals[slices] - g) < 10. * machine_epsilon))
self.assertTrue(np.all(np.abs(l2['f'] - g) < 10. * machine_epsilon))
# set slice of library view property
fvals[:, :, :] = np.exp(l1.x_grid) * np.cos(l1.y_grid) * np.log(
l1.z_grid)
l1['f'] = fvals
l2 = l1[slices]
l2['f'][:, :, :] = g
self.assertTrue(
np.all(np.abs(l1['f'][slices] - g) < 10. * machine_epsilon))
self.assertTrue(
np.all(np.abs(l1[slices]['f'] - g) < 10. * machine_epsilon))
self.assertTrue(
np.all(np.abs(fvals[slices] - g) < 10. * machine_epsilon))
self.assertTrue(np.all(np.abs(l2['f'] - g) < 10. * machine_epsilon))
# set property of library view
fvals[:, :, :] = np.exp(l1.x_grid) * np.cos(l1.y_grid) * np.log(
l1.z_grid)
l1['f'] = fvals
l2 = l1[slices]
l2['f'] = g
self.assertTrue(
np.all(np.abs(l1['f'][slices] - g) < 10. * machine_epsilon))
self.assertTrue(
np.all(np.abs(l1[slices]['f'] - g) < 10. * machine_epsilon))
self.assertTrue(
np.all(np.abs(fvals[slices] - g) < 10. * machine_epsilon))
self.assertTrue(np.all(np.abs(l2['f'] - g) < 10. * machine_epsilon))
# repeat with numpy array argument
g = gold_array
# set slice of original array
fvals[:, :, :] = np.exp(l1.x_grid) * np.cos(l1.y_grid) * np.log(
l1.z_grid)
l1['f'] = fvals
l2 = l1[slices]
fvals[slices] = g
self.assertTrue(
np.all(np.abs(l1['f'][slices] - g) < 10. * machine_epsilon))
self.assertTrue(
np.all(np.abs(l1[slices]['f'] - g) < 10. * machine_epsilon))
self.assertTrue(
np.all(np.abs(fvals[slices] - g) < 10. * machine_epsilon))
self.assertTrue(np.all(np.abs(l2['f'] - g) < 10. * machine_epsilon))
# set slice of library property
fvals[:, :, :] = np.exp(l1.x_grid) * np.cos(l1.y_grid) * np.log(
l1.z_grid)
l1['f'] = fvals
l2 = l1[slices]
l1['f'][slices] = g
self.assertTrue(
np.all(np.abs(l1['f'][slices] - g) < 10. * machine_epsilon))
self.assertTrue(
np.all(np.abs(l1[slices]['f'] - g) < 10. * machine_epsilon))
self.assertTrue(
np.all(np.abs(fvals[slices] - g) < 10. * machine_epsilon))
self.assertTrue(np.all(np.abs(l2['f'] - g) < 10. * machine_epsilon))
# set property of library slice
fvals[:, :, :] = np.exp(l1.x_grid) * np.cos(l1.y_grid) * np.log(
l1.z_grid)
l1['f'] = fvals
l2 = l1[slices]
l1[slices]['f'] = g
self.assertTrue(
np.all(np.abs(l1['f'][slices] - g) < 10. * machine_epsilon))
self.assertTrue(
np.all(np.abs(l1[slices]['f'] - g) < 10. * machine_epsilon))
self.assertTrue(
np.all(np.abs(fvals[slices] - g) < 10. * machine_epsilon))
self.assertTrue(np.all(np.abs(l2['f'] - g) < 10. * machine_epsilon))
# set slice of library view property
fvals[:, :, :] = np.exp(l1.x_grid) * np.cos(l1.y_grid) * np.log(
l1.z_grid)
l1['f'] = fvals
l2 = l1[slices]
l2['f'][:, :, :] = g
self.assertTrue(
np.all(np.abs(l1['f'][slices] - g) < 10. * machine_epsilon))
self.assertTrue(
np.all(np.abs(l1[slices]['f'] - g) < 10. * machine_epsilon))
self.assertTrue(
np.all(np.abs(fvals[slices] - g) < 10. * machine_epsilon))
self.assertTrue(np.all(np.abs(l2['f'] - g) < 10. * machine_epsilon))
# set property of library view
fvals[:, :, :] = np.exp(l1.x_grid) * np.cos(l1.y_grid) * np.log(
l1.z_grid)
l1['f'] = fvals
l2 = l1[slices]
l2['f'] = g
self.assertTrue(
np.all(np.abs(l1['f'][slices] - g) < 10. * machine_epsilon))
self.assertTrue(
np.all(np.abs(l1[slices]['f'] - g) < 10. * machine_epsilon))
self.assertTrue(
np.all(np.abs(fvals[slices] - g) < 10. * machine_epsilon))
self.assertTrue(np.all(np.abs(l2['f'] - g) < 10. * machine_epsilon))
