Hi reader,
In one of my previous questions I posted the following code.
Now I found out, that for some strange reason, the solution has type np.complex128 with the imaginary part being 0. This makes me wonder, because I put quite the efford into making sure each matrix or vector is set to the correct datatype. Can someone help me, what I missed?
Thanks in advance!
What installation of PETSc are you using?
How did you install DOLFINx?
I use the docker image: dolfinx/lab:stable.
! pip show fenics-dolfinx
returned:
Name: fenics-dolfinx
Version: 0.6.0
Summary: DOLFINx Python interface
Home-page: UNKNOWN
Author: FEniCS Project
Author-email:
License: UNKNOWN
Location: /usr/local/dolfinx-complex/lib/python3.10/dist-packages
Requires: cffi, fenics-ffcx, fenics-ufl, mpi4py, numpy, petsc4py
Required-by:
and
! pip show petsc4py
returned:
Name: petsc4py
Version: 3.17.5
Summary: PETSc for Python
Home-page: https://gitlab.com/petsc/petsc
Author: Lisandro Dalcin
Author-email: dalcinl@gmail.com
License: BSD
Location: /usr/local/lib/python3.10/dist-packages
Requires: numpy
Required-by: fenics-dolfinx, slepc4py
Call source dolfinx-real-mode in your terminal, as explained in GitHub - FEniCS/dolfinx: Next generation FEniCS problem solving environment
I restarting the kernel, run the command in the shell, rerun the notebook and everything is still the same. Did I miss something?
I mean:
uh.dtype
returns
dtype('complex128')
Mabye I should add that my assumption is , that the problem actually arises in:
import ufl
import dolfinx
import numpy as np
from petsc4py.PETSc import ScalarType
material_tags = np.unique(cell_markers.values)
T = dolfinx.fem.TensorFunctionSpace(mesh, ("DG", 0), shape=(6, 6))
stiffness = dolfinx.fem.Function(T, dtype=ScalarType)
def tensor(x, tens):
tens_reshaped = tens.reshape((6 * 6, 1))
return np.broadcast_to(tens_reshaped, (6 * 6, x.shape[1]))
for tag in material_tags:
domain = cell_markers.find(tag)
if tag < len(stack):
material = stack[tag]
stiffness.interpolate(lambda x: tensor(x, material["stiff_tens"].astype(ScalarType)), domain)
with:
stiffness.dtype
being complex128 already
replacing ScalarTypewith np.float32 in the code segment above changed this.
But later I get an error:
from dolfinx.fem.petsc import LinearProblem
problem = LinearProblem(a, L, bcs=[bc], petsc_options={"ksp_type": "preonly", "pc_type": "lu"})
uh = problem.solve()
throws:
---------------------------------------------------------------------------
TypeError Traceback (most recent call last)
Cell In[10], line 4
1 from dolfinx.fem.petsc import LinearProblem
----> 4 problem = LinearProblem(a, L, bcs=[bc], petsc_options={"ksp_type": "preonly", "pc_type": "lu"})
5 uh = problem.solve()
File /usr/local/dolfinx-complex/lib/python3.10/dist-packages/dolfinx/fem/petsc.py:566, in LinearProblem.__init__(self, a, L, bcs, u, petsc_options, form_compiler_options, jit_options)
538 def __init__(self, a: ufl.Form, L: ufl.Form, bcs: typing.List[DirichletBCMetaClass] = [],
539 u: typing.Optional[_Function] = None, petsc_options={}, form_compiler_options={}, jit_options={}):
540 """Initialize solver for a linear variational problem.
541
542 Args:
(...)
564 "pc_factor_mat_solver_type": "mumps"})
565 """
--> 566 self._a = _create_form(a, form_compiler_options=form_compiler_options, jit_options=jit_options)
567 self._A = create_matrix(self._a)
569 self._L = _create_form(L, form_compiler_options=form_compiler_options, jit_options=jit_options)
File /usr/local/dolfinx-complex/lib/python3.10/dist-packages/dolfinx/fem/forms.py:176, in form(form, dtype, form_compiler_options, jit_options)
173 return list(map(lambda sub_form: _create_form(sub_form), form))
174 return form
--> 176 return _create_form(form)
...
TypeError: __init__(): incompatible constructor arguments. The following argument types are supported:
1. dolfinx.cpp.fem.Form_complex128(spaces: List[dolfinx::fem::FunctionSpace], integrals: Dict[dolfinx::fem::IntegralType, Tuple[List[Tuple[int, object]], dolfinx.cpp.mesh.MeshTags_int32]], coefficients: List[dolfinx.cpp.fem.Function_complex128], constants: List[dolfinx.cpp.fem.Constant_complex128], need_permutation_data: bool, mesh: dolfinx.cpp.mesh.Mesh = None)
2. dolfinx.cpp.fem.Form_complex128(form: int, spaces: List[dolfinx::fem::FunctionSpace], coefficients: List[dolfinx.cpp.fem.Function_complex128], constants: List[dolfinx.cpp.fem.Constant_complex128], subdomains: Dict[dolfinx::fem::IntegralType, dolfinx.cpp.mesh.MeshTags_int32], mesh: dolfinx.cpp.mesh.Mesh)
Invoked with: , [, ], [], [], {: None, : None, : None, : None},
In a notebook, you would have to change the python kernel to the one that does not have complex in the name.
Ok, sorry. I see my mistake.
But is there even an option for the jupyter notebook?
I mean if I select a kernel I only see dolfinx with complex.
Additionally, I checked the fencisx and docker documentation and did not find any hint how to change this.
I mean there was this section but this does not work for the jupyter-notebook?
Yes, in the top menu of jupyter lab there is an option Kernel, which you then can press Change kernel
and change to Python 3 (ipykernel)
Thank you.
I did not even think that (ipykernel) would work, because the tutorial said:
Once connected, select the option Python 3 (DOLFINx complex) in the dropdown.
Thank you so much!