Diffusion bits #311
Diffusion bits #311
Conversation
solver updates -Add a diffusion solver -symbolic ddt -Eulerian ddt CM updates -Anisotropic diffusion CM -Flux CM, allowing flux terms to be passed directly into the solver
Functions now appear correctly in the view() function
Folded into solvers.py
There was a problem hiding this comment.
Pull Request Overview
This PR introduces new diffusion functionality, including a diffusion solver, a generic flux constitutive model, and an anisotropic diffusion constitutive model, along with Eulerian and symbolic strategies for history management.
- Adds SNES_Diffusion class and integrates Eulerian and Symbolic history methods in solvers.
- Updates module initializers to expose the new diffusion solver.
- Extends maths functions with an L2_norm function and introduces new constitutive models for anisotropic diffusion and generic flux.
Reviewed Changes
Copilot reviewed 6 out of 6 changed files in this pull request and generated no comments.
Show a summary per file
| File | Description |
|---|---|
| src/underworld3/systems/solvers.py | Adds SNES_Diffusion with new history management and solver setup |
| src/underworld3/systems/init.py | Exposes the new diffusion-only solver and relevant DDt imports |
| src/underworld3/maths/functions.py | Introduces a new L2_norm function for norm computations |
| src/underworld3/maths/init.py | Imports the new L2_norm function |
| src/underworld3/constitutive_models.py | Adds AnisotropicDiffusionModel and GenericFluxModel to support diffusion properties |
Comments suppressed due to low confidence (1)
src/underworld3/systems/init.py:37
- The module imports 'SemiLagragian_DDt' but elsewhere the union types and similar imports use the name 'SemiLagrangian_DDt'. Consider aligning the naming for consistency.
from .ddt import SemiLagragian_DDt
|
@bknight1 - 2 tests are failing. I think the first one is a @lmoresi issue. The second one I'm not sure about. Do you know anything about it? Also the testing script isn't failing on test errors - this needs fixing! |
|
@julesghub the advDiff test is the old one, which was failing for the reviewers of the joss paper. I updated the test there but don't think it's been merged. I'll update it for this branch now. |
Aligns with the already updated test in the joss paper branch
…into BK-diffusion
|
I'm happy to look at the bugs I introduced ! |
Add in: