[WOUT2INDATA] add a tool that puts the plasma boundary geometry from a VMEC wout file into a new INDATA namelist

[jonathanschilling]

This PR adds a new tool WOUT2INDATA.

This tool:

• reads a VMEC indata namelist from a given VMEC input file,
• reads a given VMEC wout file and
• writes a new indata namelist to a new file, where the Fourier coefficients RBC, ... of the plasma boundary geometry have been replaced to represent the plasma boundary from the wout file.

The new output file is called input.<ext>_new and it is written with the same relative/absolute path as the given input file.

[lazersos]

So I was taking a look at this and I don't quite understand this. Wouldn't it be better to just have a utility which takes a wout file only and generates and input file from it?

[jons-pf]

The problem I see is that some information is not reconstructable from the wout file that was used to generate it, e.g., the intermediate entries in ns_array but also ncurr, nvacskip, pres_scale, tcon0, delt and lforbal. We could now patch up the wout writing, but that still would not bring that info back for all the existing VMEC runs, for which the input files should still exist on the other hand.

[jons-pf]

One could make the input file command line arg optional though and leave all parameters untouched wrt. their default values from the vmec_input module when generating a new input based on a wout file only.

[lazersos]

I would also suggest the name VMEC2INDATA as this is more consistent with our existing naming conventions.

[lazersos]

Also PRES_SCALE can be reconstructed from the wout file.

REAL(rprec), EXTERNAL :: pmass
IF (pmass(0.0) > 0.0) THEN
   PRES_SCALE = presf(1)/pmass(0)
END IF

[jons-pf]

Ok, I would then move forward as follows:

1. rename to VMEC2INDATA
2. make input file optional -> filename of generated new input file is then determined by <ext> from wout_<ext>.nc, since a wout filename always needs to be provided
3. if input file is not provided, reconstruct as much of INDATA as possible from wout (incl pres_scale as you suggested)
4. if input file is provided, use as much info from that as possible, i.e., overwrite only lasym, mpol, ntor and the boundary coefficients (as currently implemented)

@lazersos Does that address your comments to your satisfaction?

[lazersos]

@jons-pf has there been progress on this?

[jons-pf]

@lazersos Sorry for the delay, I was using the code as-is for now and did not get around yet to implementing your suggestions; will do ASAP.