Draft: Segmentation-based contact using at-points operators [WIP]

Zach Atkinsrequested to merge
zach/at-points-segmentation into main

Summary

This is very rough. I don't really like what I'm doing with MatComposite to make this work, but I think there's something promising there. Right now, only Jacobi preconditioning is supported and the Jacobian definitely still has a couple bugs. Nitsche isn't right and I'm not sure how/if I'm going to make it work.

Oh, and it doesn't work with gen right now, since I'm abusing how composite operators work. I want a sequential operator, needed to handle the fact that I have two different element restrictions with different elements but the same points. That's probably the main issue with the Jacobian. Open to other ideas of how to do that.

Checklist

Closes #292

  • Atomic commit history (It's quite long, but actually pretty atomic)
  • All new features documented
  • Updated CHANGELOG.md

Merge request reports