InterfaceValues for DG interface integration

[AbdAlazezAhmed]

Tasks:

• InterfaceValues
• syncing/transforming quadrature points
• docs
• tests

After merge

• open issue to discuss mixed grid support for InterfaceValues
• check subdomain support and possibly open issue
• open issue about the investigation of mortaring with the new infrastructure

[codecov-commenter]

Codecov Report

Attention: 11 lines in your changes are missing coverage. Please review.

Comparison is base (d0f0e07) 92.98% compared to head (8e665d6) 93.08%.

Files	Patch %	Lines
src/FEValues/interface_values.jl	93.05%	10 Missing ⚠️
src/Grid/grid.jl	92.30%	1 Missing ⚠️

❗ Your organization needs to install the Codecov GitHub app to enable full functionality.

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #743      +/-   ##
==========================================
+ Coverage   92.98%   93.08%   +0.09%     
==========================================
  Files          33       34       +1     
  Lines        4950     5161     +211     
==========================================
+ Hits         4603     4804     +201     
- Misses        347      357      +10     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[termi-official]

Some more comments. I think resolving these and the open ones and the PR should be good. We might want to leave mixed grids as future work.

[termi-official]

I guess it is only left to merge master and resolving #743 (comment)

[lijas]

I was able to use a lot of stuff in the PR for implementing "GhostPenalty" with my IGA elements, so thats really nice :)

Only thing I noticed was that I had an really old verision of Tensors that did not include the functions rotation_tensor(::Float64, Float64, Float64). I did not get any error/warning from the package manager either. Does this mean that we should require Tensors to be >= v1.14.0 in the project file?

[AbdAlazezAhmed]

There were a custom rotation tensor function in this PR previously but I removed it in favor of not repeating stuff since we already have it in Tensors.jl
The difference is that it used ´sinpi´ and ´cospi´ which seem to be more accurate and pass the tests, while using ´sin´ and ´cos´ required increasing tolerance a bit.

[termi-official]

Good catch @lijas . Can you bump the compat here and merge master @AbdAlazezAhmed ?

@fredrikekre do you have time for a final review from your side? I am happy with the state for now.:)

stale

[fredrikekre]

🚀

[fredrikekre]

Great job :) Perhaps you can add an entry in the CHANGELOG?