Adapt a function on a refined mesh

mesh
1

I have a mesh, and I refine it using the adapt function

mesh1 = adapt(mesh,markedCells)

In Fenics 2017.2, I was able to adapt a function from the original mesh onto the adapted one using the adapt function

u1 = adapt(u,mesh1)

Nevertheless, I am not able to do it in Fenics 2019.1. I get the following error:

TypeError: adapt(): incompatible function arguments. The following argument types are supported:
    1. (arg0: dolfin.cpp.mesh.Mesh) -> dolfin.cpp.mesh.Mesh
    2. (arg0: dolfin.cpp.mesh.MeshFunctionSizet, arg1: dolfin.cpp.mesh.Mesh) -> dolfin.cpp.mesh.MeshFunctionSizet

Is there this functionality on Fenics 2019.1?

Thank you in advance!

2

A function with the signature that you’re looking for is still in

https://bitbucket.org/fenics-project/dolfin/src/master/dolfin/adaptivity/adapt.cpp

But, looking at

https://bitbucket.org/fenics-project/dolfin/src/master/python/src/fem.cpp

it doesn’t seem to have made it into the Python interface after the switch to pybind11. You can implement a quick workaround as follows:

from dolfin import *

cpp_code = """
#include <pybind11/pybind11.h>
#include <dolfin/adaptivity/adapt.h>
#include <dolfin/function/Function.h>
#include <dolfin/mesh/Mesh.h>
namespace py = pybind11;
PYBIND11_MODULE(SIGNATURE, m)
{
m.def("adapt", [](const dolfin::Function &function,
		  std::shared_ptr<const dolfin::Mesh> adapted_mesh,
                  bool interpolate){
	         return dolfin::adapt(function, adapted_mesh, interpolate);});
}
"""
m = compile_cpp_code(cpp_code)
def adaptFunction(f,mesh,interp=True):
    return m.adapt(f.cpp_object(),mesh,interp)

# Test:
mesh = UnitSquareMesh(2,2)
V = FunctionSpace(mesh,"CG",1)
f = interpolate(Expression("x[0]",degree=1),V)

rmesh = refine(mesh)
rf = adaptFunction(f,rmesh)

# Plot:
from matplotlib import pyplot as plt
plot(rf)
plot(rf.function_space().mesh())
plt.show()
3

Thank you!

I would like to add that rf is of a pybind_11 type. If, for some reason, you need a pure python type, you can use the assign function as

python_rf.assign(rf)

where python_rf is a function of a function space.

Thank you again!