Add TGLFBasedTransportModel

[theo-brown]

Pulling out the good bits of #477 and bringing them up to date.

Aim for this PR is to focus on implementing the convertor between TORAX outputs and TGLF inputs. Other parts of #477 are being divided up, as more people are getting involved in working on it.

[theo-brown]

Input conversion validated against JETTO TGLF runs for a spherical tokamak test case, using the EQDSK and IMAS equilibrium IDS from JETTO as the input to TORAX.

[theo-brown]

Thanks for taking a look at this @Nush395, I moved a bunch of stuff to the docstring as suggested.

[jcitrin]

FYI @gsnoep , @aaronkho

[theo-brown]

@jcitrin sorry I noticed some things that still needed doing. Is it ok for me to push to this branch now that you've marked it as copybara:import-manual, or is there a better option?

[jcitrin]

@theo-brown , sure go ahead and add. We hadn't actually started the internal review yet, no worries. Will hold off on it.

[theo-brown]

Thanks! I'm working on #1436 and this in parallel, having validated that the TGLFNNukaea implemented in fusion_surrogates is correctly reproducing the fluxes from the tglfnn-ukaea repo. Some issues I find are to do with how I'm talking to the NN, and some are with how I'm generating inputs/outputs, so there's a bit of bouncing between the two issues as I iterate. Hopefully should be able to get this in this week?

[theo-brown]

Summary of changes since last review:

• Rebased onto main, removed DynamicRuntime objects
• Updated links to GACODE docs

[theo-brown]

@jcitrin this is ready for a re-review / for the copybara-import-manual label to be added back!