Ion viscosity permissions

[malamast]

WIP

1. Fixed permissions. We need AA when bounce_frequency is enabled.
2. Apply parallel BC for eta. We probably don't need the values of eta at the parallel guard cells. It depends on how the parallel operators are implemented.
3. Applied a softFloor to collision frequency 1/nu as in head_conduction.

[codecov]

Codecov Report

❌ Patch coverage is 26.92308% with 19 lines in your changes missing coverage. Please review.
✅ Project coverage is 57.44%. Comparing base (a1c3ba3) to head (f45aea2).

Files with missing lines	Patch %	Lines
src/braginskii_ion_viscosity.cxx	26.92%	16 Missing and 3 partials ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #628      +/-   ##
==========================================
- Coverage   57.51%   57.44%   -0.08%     
==========================================
  Files          98       98              
  Lines       10239    10262      +23     
  Branches     1469     1478       +9     
==========================================
+ Hits         5889     5895       +6     
- Misses       3758     3772      +14     
- Partials      592      595       +3     

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

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[mikekryjak]

I don't understand parallel slices well, but Claude tells me that this if statement is always going to be true since eta.getMesh()->communicate(eta) always calculates the parallel slices.

A similar BC in braginskii_conduction has a simpler implementation:

hermes-3/src/braginskii_conduction.cxx

Lines 276 to 289 in a1c3ba3

	for (RangeIterator r = mesh->iterateBndryLowerY(); !r.isDone(); r++) {
	for (int jz = 0; jz < mesh->LocalNz; jz++) {
	auto i = indexAt(kappa_par, r.ind, mesh->ystart, jz);
	auto im = i.ym();
	kappa_par[im] = kappa_par[i];
	}
	}
	for (RangeIterator r = mesh->iterateBndryUpperY(); !r.isDone(); r++) {
	for (int jz = 0; jz < mesh->LocalNz; jz++) {
	auto i = indexAt(kappa_par, r.ind, mesh->yend, jz);
	auto ip = i.yp();
	kappa_par[ip] = kappa_par[i];
	}
	}

[mikekryjak]

Hi @malamast! Thanks for this. I'm not an expert on viscosity but found a possible issue and left a comment. I have two other suggestions:

• Can you change the name of the PR so that it mentions the fact this changes the ion viscosity parallel boundary condition? I think that is probably the most important change in this PR.
• Now that eta is neumann, it is non-zero at the boundary, which means viscosity will now flow through the parallel boundary. Is this intended? If so, could you mention the reason in the PR description?
• In braginskii_conduction we set kappa to Neumann but then disable conduction through the boundary by passing a bool to Div_par_k_Grad:

hermes-3/src/braginskii_conduction.cxx

Lines 294 to 295 in a1c3ba3

	add(species["energy_source"],
	Div_par_K_Grad_par_mod(kappa_par, T, flow_ylow_conduction, false));

[mikekryjak]

For consistency with conduction:

Suggested change

	{"long_name", std::string("Ion viscosity coefficient") + species_name},
	{"long_name", species_name + std::string(" ion viscosity coefficient")},