Numba error while exporting to paraview (trying to reproduce an example from an article)

[Mikhail Kondratyev]

I'm trying to reproduce an example from an article "PorePy: an open-source software for simulation of multiphysicsprocesses" https://doi.org/10.1007/s10596-020-10002-5.

Here is the code given in the Fig 7.

import porepy as pp
import numpy as np
from scipy.sparse.linalg import spsolve

f1 = pp.Fracture(np.array([[0,1,1,0], [0,0,1,1], [0,0,0,0]]))
f2 = pp.Fracture(np.array([[0,0,0,0], [0,1,1,0], [0,0,1,1]]))
network = pp.FractureNetwork3d([f1, f2])

gb = network.mesh({'mesh_size_frac': 0.1,
                   'mesh_size_bond': 1,
                   'mesh_size_min': 0.01,
                   })

diffusion_discr = pp.Mpfa('flow')
accumulation_discr = pp.MassMatrix('flow')
for g, d in gb:
    pp.initialize_default_data(g, d, 'flow')
    d[pp.PRIMARY_VARIABLES] = {'p': {'cells': 1}}
    d[pp.DISCRETIZATION] = {'p': {'accumulation': accumulation_discr,
                                  'diffusion': diffusion_discr}}

for e, d in gb.edges():
    g_l, g_h = gb.nodes_of_edge(e)
    mg = d['mortar_grid']
    data = {'normal_diffusivity': 1}
    pp.initialize_data(mg, d, 'flow', data)
    d[pp.PRIMARY_VARIABLES] = {'mortar_flux': {'cells': 1}}
    interface_discr = pp.RobinCoupling(
            'flow',
            diffusion_discr,
            diffusion_discr)
    d[pp.COUPLING_DISCRETIZATION] = {
            'interface_flux': {
                g_h: ('p', 'diffusion'),
                g_l: ('p', 'diffusion'),
                e: ('mortar_flux', interface_discr)
                }
            }

assembler = pp.Assembler(gb)
assembler.discretize()
A, b = assembler.assemble_matrix_rhs()

p = spsolve(A, b)
assembler.distribute_variable(p)

paraview_exporter = pp.Exporter(gb, file_name= 'foo')
paraview_exporter.write_vtk('p')

On one of the last line when exporting to paraview I get the following error:

  File "/home/user/.local/lib/python3.8/site-packages/numba/core/byteflow.py", line 772, in op_CALL_FUNCTION_EX
    raise UnsupportedError(errmsg)
numba.core.errors.UnsupportedError: Failed in nopython mode pipeline (step: analyzing bytecode)
CALL_FUNCTION_EX with **kwargs not supported

What could it be?

[Eirik Keilegavlen]

Sounds like a compatibility issue with numba. Which version of numba, numpy and python do you have, and how updated is your PorePy install?

[Mikhail Kondratyev]

It's Ubuntu 19.10.
python 3.8.3
numba 0.50.1
numpy 1.19.5

Numba and Numpy were installed via pip

[Eirik Keilegavlen]

Thanks; I can reproduce the error. No idea what happened, but I will investigate, and give updates here.

[Eirik Keilegavlen]

@mkondratyev85 can you please pull latest version of PorePy (after #514 was merged) and then try?

It seems the error was introduced when we added some logging functionality a few weeks ago.

[Mikhail Kondratyev]

Thank you for help. It works now. I just had do change write_vtk to write_vtu on the last line.

[Eirik Keilegavlen]

Oh yead, that's another recent minor change to the code.

Thanks a lot for reporting this!

[Eirik Keilegavlen]

Resolved by #514