Mesh2D.save puts points in yz-plane

[gdmcbain]

While investigating the MeshTri returned by refinterp #59 , I was puzzled to find that save prepends rather than appends zero coordinates; i.e. it puts it into the yz-plane rather than the xy-plane.

See https://github.com/kinnala/scikit-fem/blame/master/skfem/mesh/mesh2d/mesh2d.py#L191

        # a row of zeros required in 2D
        p = np.vstack((np.zeros(self.p.shape[1]), self.p))

Is that intentional?

[gdmcbain]

Actually I'm not sure that any embedding into ℝ³ is needed.

Trying

meshio.write('M2.mesh', meshio.Mesh(self.p.T, cells))

produces a valid two-dimensional MEDIT mesh file. Gmsh opens and renders this O. K.

MeshVersionFormatted 2
Dimension 2

Vertices
11520
0.0 0.0 1
0.125 0.0 1
0.0625 0.0625 1
0.0625 0.0 1
0.03125 0.03125 1
⋮

[kinnala]

I think that embedding into R^3 was done due to vtk export. But yz-plane sounds strange indeed.

[gdmcbain]

Yes, something to do with VTK, but maybe not too serious.

meshio.write('M2.vtk', meshio.Mesh(M.p.T, {'triangle': M.t.T}))

WARNING:root:VTK requires 3D points, but 2D points given. Appending 0 third component.

[kinnala]

Originally scikit-fem used pyevtk so it might be a remnant. I think it can be freely removed.