The regression test fails because of errors that are likely in OMFIT:

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
TestOmasExamples.test_plot_omas_omfit
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Possible problems found with gacode profiles file:
 * pow_e_line is negative (should always be positive)
make: *** [Makefile:26: test] Error 1
Error: Process completed with exit code 2.

What OMFIT version does the test environment use?

This has been tested as part of #257 and seems to work for that case.

Stale pull request message

I tried following a series of PRs here on omas that all depend on each other. Can @AreWeDreaming open an issue here on omas with the correct order of operations for merging? Then we can tackle each one sequentially.

This is the PR that needs to be merged next.

I don't see any reviewers assigned. Is this still waiting on someone before the merge?

Congratulations @bechtt you have fallen to one of my classic blunders.

Need to create OMFIT PR with density_thermal change to match this PR.

It sounds like the plan is to hold off on the electron density vs density_thermal change until we hear some clarification from IMAS devs about the possible use cases for this (through JIRA)

@AreWeDreaming I would think that density_thermal is the proper entry to be filled for the sample (or both?). Perhaps the failing regression could be fixed with a ods.physics_core_profiles_consistent()?

@jmcclena do you understand why density_thermal would be useful for electrons? It seems like it only makes sense for ions

While not as well understood as fast ions, I believe suprathermal instabilities like fishbones have been observed with ECH heating (https://iopscience.iop.org/article/10.1088/0029-5515/47/11/022/pdf). You could also have a fast runaway electron population near a disruption.

Additionally, while there is no fast electron temperature, there is a fast parallel and perp pressure (just like how the the ions are set up)

@jmcclena I did my thesis on non-thermal electron distributions functions generated by ECH. You have to try very hard (core density 1.e19 + 0.8 MW ECH) to generate non-thermal electron distributions, and even then they make up less than 1 % of the electron density.
However, the real issue is that I cannot think of a situation where a two-fluid description of the electron distribution would be a good choice, maybe with the exception of run-away electrons, but even there you will have a hard time fitting that kind of distribution with two Maxwellians.
The other thing is that all our means to measure or model n_e are using the assumption that the electrons are thermally distributed. Splitting the electron density into thermal and fast implies that this distinction has been made and something has been done to determine that the electrons are thermally distributed.
I'll create a JIRA tracker at ITER and ask them for guidance. Let's see what they think.

Shouldn't fast (and total) electron density also be set according to ITER's recommendation? https://jira.iter.org/browse/IMAS-5050

Probably but that is out of scope for this PR

Probably but that is out of scope for this PR

How so? It looks like you are adding the mappings to omas from MDS+ for electron density. What would be a more appropriate time to set up the proper mapping?

Well to do it properly every time electron density_thermal is set we would need to set density_fast as well. I could do that but we have 4 more PRs to work through.

I tried my best to test this (and the original branch) with omas viewer, but couldn't get any plots to work. That probably isn't an ideal test here (since the development has stalled), but I don't have anything better.

It could be worth running the OMFIT regression tests with this omas branch to ensure nothing breaks there. If that works I'd be comfortable with merging this

Stale pull request message

This passed 73 OMFIT regression tests here https://github.com/gafusion/OMFIT-source/pull/6967 and all omas tests so I don't see any reason not to merge