Topography resolution

[DavidAKopriva]

This pull request has the option of automatic curvature based refinement due to bottom topography. Include "sizing = ON" to the topography block to turn on. There is a simple test included, but more general testing is needed. It is also really expensive due to finite difference approximations to the topography computed from the elevation function. Speedup is needed.

[codecov[bot]]

Codecov Report

Merging #33 (e06a876) into main (6c43412) will increase coverage by 0.66%. The diff coverage is 95.32%.

@@            Coverage Diff             @@
##             main      #33      +/-   ##
==========================================
+ Coverage   66.85%   67.52%   +0.66%     
==========================================
  Files          66       68       +2     
  Lines        9948    10195     +247     
==========================================
+ Hits         6651     6884     +233     
- Misses       3297     3311      +14     

Flag	Coverage Δ
unittests	67.52% <95.32%> (+0.66%)	:arrow_up:

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
Source/IO/ControlFileReader.f90	83.12% <ø> (ø)
Source/Project/Model/SMModel.f90	75.45% <77.77%> (-0.01%)	:arrow_down:
Source/Surfaces/DataFileTopographyClass.f90	77.55% <84.61%> (-10.92%)	:arrow_down:
Source/Project/MeshProject.f90	72.82% <85.71%> (+0.46%)	:arrow_up:
Source/Surfaces/BiCubicClass.f90	98.57% <98.57%> (ø)
Source/Project/Sizer/Sizer.f90	77.05% <100.00%> (+1.08%)	:arrow_up:
Source/Surfaces/GaussianCurvature.f90	100.00% <100.00%> (ø)
Source/Surfaces/SMTopographyClass.f90	60.97% <100.00%> (+33.70%)	:arrow_up:
Source/Testing/CurveTests.f90	100.00% <100.00%> (ø)
Source/Testing/MeshingTests.f90	88.42% <100.00%> (+0.19%)	:arrow_up:
... and 1 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more Δ = absolute <relative> (impact), ø = not affected, ? = missing data Powered by Codecov. Last update 6c43412...e06a876. Read the comment docs.

[fluidnumerics-joe]

It took a little bit to catch up with the conversation here. The code looks good, and from what I can tell this feature is enabled in the topography block using sizing = ON. I'd say the only bit missing here is to update relevant sections of documentation that have reference to the TOPOGRAPHY block:

• appendix.md
• examples.md
• three-dimensional-hexahedral-meshes.md

I'll add what I think is relevant to these docs today.

[DavidAKopriva]

We also have a branch off of this one that is almost done that factors out the biCubic interpolation and uses it to compute the gaussian curvature. That branch is almost done. I'm getting the test example files prepared, and I have to edit the makefile. The branch is "WithBiCubics".

[DavidAKopriva]

My thought is to merge "WithBiCubics" into the topography branch, then make sure everything is OK before merging into main.

[DavidAKopriva]

This will be added to the documentation.

[andrewwinters5000]

I think this is ready to merge into main. But first, I cannot build the docs locally. Can someone check that the LaTeX displays properly? Also, I cannot workout where the new ApproximationOfGaussianCurvature.md is included in the docs

[DavidAKopriva]

I think the thing to do with the new doc is to add it to the list in the assumptions-and-algorithms.md with a link there to the document. I can add that today. As for rendering, all I can say is it does not render in two markdown editors that I have, but does render fine in visual studio with the plugin. I don't know how to test that on GitHub.

[DavidAKopriva]

I guess the thing I don't know is what the url is for the documents folder on GitHub. Can'tfind any examples so far in what we have...

[sloede]

As for rendering, all I can say is it does not render in two markdown editors that I have, but does render fine in visual studio with the plugin. I don't know how to test that on GitHub.

I often use grip, which opens a local web server and tries to mimic what a README would look like when rendered on GH. You can install it on macOS using brew install grip and then run it with grip myfile.md.

[DavidAKopriva]

Thanks. I'll give it a try. Also, I guess the URL is just (https://trixi-framework.github.io/HOHQMesh/NameOfFile/), which is what the other pages seem to have in the URL bar.

[DavidAKopriva]

Oh Well, grip didn't work for me. Just sat there doing nothing after printing a bunch of notices, the last of which was Running on... Press CTRL+C to quit, then nothing. That did quit something.

[andrewwinters5000]

Oh Well, grip didn't work for me. Just sat there doing nothing after printing a bunch of notices, the last of which was Running on... Press CTRL+C to quit, then nothing. That did quit something.

For me it opens the preview if I run, e.g., grip -b index.md (this will automatically open a browser tab)

[andrewwinters5000]

@DavidAKopriva @fluidnumerics-joe Now that I verified that the docs are okay and all built correctly, should this merge to main?

[DavidAKopriva]

Sure. This morning I added code to look at the radius of curvature along the gradient direction to give some refinement if the gaussian curvature is zero. It still passes the tests (the old ones, no new test for this explicitly, yet), but i want to play with it a little bit more. Does that sound OK? Or we can wait to include it in the next round.

[DavidAKopriva]

ShelfWithRefinementr.pdf

Here's what it does now. Without the "fix" there is no refinement due to topography.

[andrewwinters5000]

This morning I added code to look at the radius of curvature along the gradient direction to give some refinement if the gaussian curvature is zero. It still passes the tests (the old ones, no new test for this explicitly, yet), but i want to play with it a little bit more. Does that sound OK? Or we can wait to include it in the next round.

That sounds good, but I think we can open a new PR for this topography improvement. This PR already adds quite a bit, so I think we can declare it mergable and then make it better.

[DavidAKopriva]

OK, I just opened the new PR. I am OK with whichever way you want to go.

[andrewwinters5000]

Okay, I will merge this now