Ensuring IMAS compatible geometry

[theo-brown]

Question mostly for @MateoBell: I have another code (freegsnke) that is producing an IMAS geometry file for use in TORAX.
However, it's missing some quantities that the TORAX IMAS loader is currently expecting.

I'm slightly unclear as to how strict the IMAS schema is meant to be: is my other code meant to produce a file that is guaranteed to have all of the entries in? Or is TORAX meant to be able to handle when any entry is missing?

Particular example in this case is the volume, which FreeGSNKE is not currently outputting, and so the following fails:

torax/torax/_src/geometry/imas.py

Lines 95 to 100 in b371a85

	if IMAS_data.profiles_1d.dvolume_dpsi:
	dvoldpsi = 1 * IMAS_data.profiles_1d.dvolume_dpsi # Sign changed ddv4.
	else:
	dvoldpsi = np.gradient(
	IMAS_data.profiles_1d.volume, IMAS_data.profiles_1d.psi
	)

Linked issue on FreeGSNKE: FusionComputingLab/freegsnke#35

[jcitrin]

I'm not sure if there's a definitive answer on how strict the IMAS schema is meant to be. As a rule of thumb, I'd say that unless a specific output is extremely niche and/or difficult to calculate + output, then the expectation is that the data will be there. Looking at the list of outputs in the linked issue, I'd say that all of those should be expected IMAS outputs of an equilibrium code.

[MateoBell]

Yes I agree. There is no minimum set of entries the IDS should contain in general case, but I think it is better to assume we get it filled will all the quantities listed when received by TORAX, as their calculation is more relevant to the equilibrium code.