Add subcell limiting support for StructuredMesh

[bennibolm]

This PR adds subcell limiting support for Structuredmesh. This includes the curved mesh version of calcflux_fhat and the computation of local bounds at interfaces and boundaries (calc_bounds_twosided_interface and calc_bounds_onesided_interface). Moreover, to get the inverse Jacobian for each node correctly, we use get_inverse_jacobian() for all mesh types.

The rest are just resorting and reformatting ugly lines of code.

For now, there is no support for non-conservative systems.

[github-actions[bot]]

Review checklist

This checklist is meant to assist creators of PRs (to let them know what reviewers will typically look for) and reviewers (to guide them in a structured review process). Items do not need to be checked explicitly for a PR to be eligible for merging.

Purpose and scope

• [ ] The PR has a single goal that is clear from the PR title and/or description.
• [ ] All code changes represent a single set of modifications that logically belong together.
• [ ] No more than 500 lines of code are changed or there is no obvious way to split the PR into multiple PRs.

Code quality

• [ ] The code can be understood easily.
• [ ] Newly introduced names for variables etc. are self-descriptive and consistent with existing naming conventions.
• [ ] There are no redundancies that can be removed by simple modularization/refactoring.
• [ ] There are no leftover debug statements or commented code sections.
• [ ] The code adheres to our conventions and style guide, and to the Julia guidelines.

Documentation

• [ ] New functions and types are documented with a docstring or top-level comment.
• [ ] Relevant publications are referenced in docstrings (see example for formatting).
• [ ] Inline comments are used to document longer or unusual code sections.
• [ ] Comments describe intent ("why?") and not just functionality ("what?").
• [ ] If the PR introduces a significant change or new feature, it is documented in NEWS.md with its PR number.

Testing

• [ ] The PR passes all tests.
• [ ] New or modified lines of code are covered by tests.
• [ ] New or modified tests run in less then 10 seconds.

Performance

• [ ] There are no type instabilities or memory allocations in performance-critical parts.
• [ ] If the PR intent is to improve performance, before/after time measurements are posted in the PR.

Verification

• [ ] The correctness of the code was verified using appropriate tests.
• [ ] If new equations/methods are added, a convergence test has been run and the results are posted in the PR.

Created with :heart: by the Trixi.jl community.

[codecov[bot]]

Codecov Report

Attention: Patch coverage is 99.05660% with 2 lines in your changes are missing coverage. Please review.

Project coverage is 96.11%. Comparing base (c2513e2) to head (83a8cef).

Files	Patch %	Lines
...rc/solvers/dgsem_structured/subcell_limiters_2d.jl	98.41%	2 Missing :warning:

Additional details and impacted files

```diff @@ Coverage Diff @@ ## main #1946 +/- ## ========================================== + Coverage 96.09% 96.11% +0.02% ========================================== Files 457 460 +3 Lines 36734 36926 +192 ========================================== + Hits 35298 35490 +192 Misses 1436 1436 ``` | [Flag](https://app.codecov.io/gh/trixi-framework/Trixi.jl/pull/1946/flags?src=pr&el=flags&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=trixi-framework) | Coverage Δ | | |---|---|---| | [unittests](https://app.codecov.io/gh/trixi-framework/Trixi.jl/pull/1946/flags?src=pr&el=flag&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=trixi-framework) | `96.11% <99.06%> (+0.02%)` | :arrow_up: | Flags with carried forward coverage won't be shown. [Click here](https://docs.codecov.io/docs/carryforward-flags?utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=trixi-framework#carryforward-flags-in-the-pull-request-comment) to find out more.

:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.

[bennibolm]

Benchmarking if using dispatched routines is faster than calling inverse_jacobian inside for node for loop. Running kelving helmholtz on TreeMesh with local limiting

using Trixi
equations = CompressibleEulerEquations2D(1.4)
basis = LobattoLegendreBasis(3)
limiter_idp = SubcellLimiterIDP(equations, basis;
                                local_twosided_variables_cons = ["rho"],
                                positivity_variables_cons = ["rho"],
                                positivity_variables_nonlinear = [pressure]) 
trixi_include("../examples/tree_2d_dgsem/elixir_euler_kelvin_helmholtz_instability_sc_subcell.jl", tspan = (0.0, 0.2), limiter_idp=limiter_idp)

Benchmarking:

limiter = semi.solver.volume_integral.limiter
alpha = limiter.cache.subcell_limiter_coefficients.alpha
u = Trixi.wrap_array(sol.u[end], mesh, equations, solver, semi.cache)
t=0.0
dt=0.1
variable = 1
using BenchmarkTools
@btime Trixi.idp_local_twosided!($alpha, $limiter, $u, $t, $dt, $semi, $mesh, $variable)

First version with

@threaded for element in eachelement(dg, semi.cache)
        for j in eachnode(dg), i in eachnode(dg)
            inverse_jacobian = cache.elements.inverse_jacobian[element]
            idp_local_twosided_inner!(alpha, inverse_jacobian, u, dt, dg, cache,
                                      variable, var_min, var_max, i, j, element)
        end
    end

and

@threaded for element in eachelement(dg, semi.cache)
        inverse_jacobian = cache.elements.inverse_jacobian[element]
        for j in eachnode(dg), i in eachnode(dg)
            idp_local_twosided_inner!(alpha, inverse_jacobian, u, dt, dg, cache,
                                      variable, var_min, var_max, i, j, element)
        end
    end

results in equivalent numbers: 338.607 μs (4 allocations: 136 bytes)

(The allocations are coming from creating the symbol in

var_min = variable_bounds[Symbol(variable_string, "_min")]
var_max = variable_bounds[Symbol(variable_string, "_max")]

Conclusion: Dispatching is not faster for the TreeMesh implementation. So, it's possible to use something like get_inverse_jacobian to get the jacobian for all mesh types within the inner loop. This saves a lot of lines in this PR.

[bennibolm]

Old description: Getting the inverse Jacobian makes this PR a bit longer. That's because for TreeMesh, inverse_jacobian is fixed within one element, while it is different for each node for StructuredMesh. To avoid something like

if mesh isa TreeMesh
    inverse_jacobian = cache.elements.inverse_jacobian[element]
end
for j in eachnode(dg), i in eachnode(dg)
    if mesh isa StructuredMesh
        inverse_jacobian = cache.elements.inverse_jacobian[i, j, element]
    end
    [...]
end

and also avoid unnecessary calls of inverse jacobian in TreeMesh simulations, I dispatched for the mesh type and extracted the code [...] to an inner function (This is also done for e.g. the routine apply_jacobian!).

UPDATE: Due to https://github.com/trixi-framework/Trixi.jl/pull/1946#issuecomment-2112619960 and the addition of get_inverse_jacobian, no dispatching is needed anymore.