Allow expressions with submeshes#4140
Conversation
…ordinate element hash
Co-authored-by: Jørgen Schartum Dokken <dokken92@gmail.com>
Co-authored-by: Jørgen Schartum Dokken <dokken92@gmail.com>
|
CI failing due to global Github issues: |
Co-authored-by: Jørgen Schartum Dokken <dokken92@gmail.com>
Co-authored-by: Paul T. Kühner <56360279+schnellerhase@users.noreply.github.com>
| def mark_left_facets(x): | ||
| return np.isclose(x[0], 1.0, atol=tol) | ||
|
|
||
| left_facets = dolfinx.mesh.locate_entities(mesh, mesh.topology.dim - 1, mark_left_facets) |
There was a problem hiding this comment.
This is an empty list. That is not desired for the test, right? Should the marker before be np.isclose(x[0], 0.0, atol=tol)?
|
@schnellerhase and I got further. Now we are missing a permutation to compare the output of the two approaches, as they produce the same output, but in a permuted order.. |
With submesh support, it is quite common that one want to evaluate an expression that uses quantities from both the parent and submesh.
This PR extends expression to be able to support:
Other structural changes
pack_coefficientsfordolfinx::fem::Expressioninto a multi-step approach:Requires: FEniCS/ffcx#828
Resolves: #3207