Rescaling (suggested feature)

[mishapadidar]

This is a feature suggestion, not a bug.
It is common to "rescale" configurations by either enlarging the surface by a constant multiple or multiplying the field strength by a constant.
In these cases, the equilibrium can be rescaled instead of recomputed.

For instance, if B -> cB then J -> cJ and gradp -> c^2 gradp.
It could be convenenient to have this utility in VMEC++.
See DESC's rescale for example.

[jons-pf]

Good idea, I agree that this could be useful. Re-scaling the equilibrium itself is probably not too big of an issue.

Re-scaling all the post-processed output individually could be challenging, at least because of the sheer number of output quantities of a VMEC run. However, that could be tackled by splitting off the output quantity computation and running it on the re-scaled equilibrium...

Are we really only talking about the two transformations

1. B --> c * B
2. R --> c * R

or are there also other transformations that we would need to consider?

[mishapadidar]

Splitting the output computation sounds just right!

One thing to keep in mind is that users often scale slightly different variants of the same quantity, i.e.
users may want to scale B via B0, <B>, B_max or the spatial scales via a, R, V.
It also may be more common to specify the target value, such as a_target, instead of the scale factor, c.

[jurasic-pf]

For scaling the magnetic field we have to clearly communicate what happens to the pressure (c^2}, so there's no surprises for users.
Doing this on the C++ sode would probably be easiest for now

[mishapadidar]

perhaps a scale_pressure=True argument, to optionally scale the pressure to keep force balance

[jons-pf]

@mishapadidar like in here? https://github.com/landreman/vmec_equilibria/blob/master/scale_vmec/scale_vmec.f90

[mishapadidar]

Yes that works. Scaling pres_scale seems like it should work.