Question: material function with an unexpected keyword 'mode'

[Gavinlee960312]

I'm trying to use sfepy to build a mode solver for step_index fiber. But when I tried to define the permittivity in the materials, I cannot find any way to define a circle-shape parameter for materials. Here is my code:

from future import absolute_import
from sfepy import data_dir
import numpy as np
filename_mesh = data_dir + '/meshes/2d/test1.mesh'

lamda=1.55E-6
k0=2*np.pi/lamda
neigen=10
regions={
'Omega':'all',
'Outsider':('vertices of group 1','facet'),
}

fields={
'ElectricField':('complex',2,'Omega',1),
}

variables={
'Ef':('unknown field','ElectricField',0),
'et':('test field','ElectricField','Ef'),
}

ebcs={
'freespace':('Outsider',{'Ef.0':0.0}),
}

materials={
'epsl':'permitivity',
'one':({'one':1.0},),
}

equations={
'ModeEquation':
"""dw_laplace.2.Omega(one.one,et,Ef)
=%sdw_volume_dot.2.Omega(epsl.f,et,Ef)
-%sdw_volume_dot.2.Omega(one.one,et,Ef)"""
%(k02, neigen*k02)
}

options = {
'nls' : 'newton',
'ls' : 'ls',
}

solvers={
'ls' : ('ls.scipy_direct', {}),
'newton':('nls.newton',{
'i_max':1,
'eps_a':1E-10,
}),
}

def permitivity(coors, mode=None):
if mode=='qp':
r = np.sqrt(coors[:,0]**2.0 + coors[:,1]2.0)
val = 1.45(r<=60) + 1.4(r>60 and r <200)
val.shape=(coors.shape[0],1,1)
return {'f':val}
functions = {
'permitivity' : (permitivity,),
}

And the log is:

sfepy: left over: ['name', 'doc', 'package', 'loader', 'spec', 'file', 'cached', 'builtins', 'absolute_import', 'data_dir', 'np', 'lamda', 'k0', 'neigen', 'permitivity', 'verbose', '_filename']
sfepy: reading mesh (C:\Users\Gavin\OneDrive\Desktop\sfepy\meshes\2d\test1.mesh)...
sfepy: number of vertices: 147142
sfepy: number of cells:
sfepy: 2_3: 292260
sfepy: ...done in 1.33 s
sfepy: warning: bad element orientation, trying to correct...
sfepy: ...corrected
sfepy: creating regions...
sfepy: Omega
sfepy: Outsider
sfepy: ...done in 0.18 s
sfepy: equation "ModeEquation":
sfepy: dw_laplace.2.Omega(one.one,et,Ef)
=16432223768723.176dw_volume_dot.2.Omega(epsl.f,et,Ef)
-164322237687231.75dw_volume_dot.2.Omega(one.one,et,Ef)
sfepy: using solvers:
ts: no ts
nls: newton
ls: ls
sfepy: updating variables...
sfepy: ...done
sfepy: setting up dof connectivities...
sfepy: ...done in 0.03 s
sfepy: matrix shape: (148020, 148020)
sfepy: assembling matrix graph...
sfepy: ...done in 0.10 s
sfepy: matrix structural nonzeros: 1041232 (4.75e-05% fill)
sfepy: updating variables...
sfepy: ...done
sfepy: updating materials...
sfepy: one
sfepy: epsl
Traceback (most recent call last):
File "./simple.py", line 182, in
main()
File "./simple.py", line 179, in main
app()
File "C:\Users\Gavin\OneDrive\Desktop\sfepy\sfepy\applications\application.py", line 29, in call_basic
return self.call(**kwargs)
File "C:\Users\Gavin\OneDrive\Desktop\sfepy\sfepy\applications\pde_solver_app.py", line 228, in call
post_process_hook_final=self.post_process_hook_final)
File "C:\Users\Gavin\OneDrive\Desktop\sfepy\sfepy\discrete\problem.py", line 1435, in solve
status=status)
File "C:\Users\Gavin\OneDrive\Desktop\sfepy\sfepy\solvers\ts_solvers.py", line 34, in _standard_ts_call
status=status, **kwargs)
File "C:\Users\Gavin\OneDrive\Desktop\sfepy\sfepy\solvers\ts_solvers.py", line 71, in call
prestep_fun(ts, vec0)
File "C:\Users\Gavin\OneDrive\Desktop\sfepy\sfepy\discrete\problem.py", line 1260, in prestep_fun
self.update_materials(verbose=self.conf.get('verbose', True))
File "C:\Users\Gavin\OneDrive\Desktop\sfepy\sfepy\discrete\problem.py", line 591, in update_materials
problem=self, verbose=verbose)
File "C:\Users\Gavin\OneDrive\Desktop\sfepy\sfepy\discrete\equations.py", line 340, in time_update_materials
verbose=verbose)
File "C:\Users\Gavin\OneDrive\Desktop\sfepy\sfepy\discrete\materials.py", line 58, in time_update
mat.time_update(ts, equations, mode=mode, problem=problem)
File "C:\Users\Gavin\OneDrive\Desktop\sfepy\sfepy\discrete\materials.py", line 303, in time_update
self.update_data(key, ts, equations, term, problem=problem)
File "C:\Users\Gavin\OneDrive\Desktop\sfepy\sfepy\discrete\materials.py", line 220, in update_data
**self.extra_args)
File "C:\Users\Gavin\OneDrive\Desktop\sfepy\sfepy\discrete\functions.py", line 37, in call
return self.function(*args, **_kwargs)
TypeError: permitivity() got multiple values for argument 'mode'

Anyone can help me? I'll really appreciate it!

[rc]

The material definition functions need to have the following arguments (see https://sfepy.org/doc-devel/users_guide.html#defining-material-parameters-by-functions):

def get_pars(ts, coors, mode=None, **kwargs):
    if mode == 'qp':
        ...

so try using:

def permitivity(ts, coors, mode=None, **kwargs):
    if mode=='qp':
        r = np.sqrt(coors[:,0]**2.0 + coors[:,1]2.0)
        val = 1.45(r<=60) + 1.4(r>60 and r <200)
        val.shape=(coors.shape[0],1,1)
        return {'f':val}

[Gavinlee960312]

Thanks. I changed the code as you mentioned, but it pops another error:

ValueError: wrong arguments shapes for "+1.0 * dw_laplace.2.Omega(one.one, et, Ef)" term! (see above)

To avoid using dw_laplace, I used dw_div_grad instead, and I got the following error:

File "C:\Users\Gavin\OneDrive\Desktop\sfepy\sfepy\terms\terms_dot.py", line 50, in dw_dot
status = fun(out, mat, val_qp, vgeo, sgeo, fmode)
File "sfepy\terms\extmods\terms.pyx", line 1528, in sfepy.terms.extmods.terms.dw_volume_dot_vector
array2fmfield4(_coef, coef)
File "sfepy\discrete\common\extmods_fmfield.pyx", line 5, in sfepy.discrete.common.extmods._fmfield.array2fmfield4
cdef inline int array2fmfield4(FMField *out,
ValueError: Buffer has wrong number of dimensions (expected 4, got 3)

[rc]

You can use just dw_laplace.2.Omega(et,Ef), the same with dw_volume_dot. Or do you want to use other values in future?

[rc]

Ah, you variable is a vector. So yes, you should use dw_div_grad instead of dw_laplace.

[rc]

The question was answered, closing.