New reducers for refinement criteria

[lkotipal]

Adds new datareducers for AMR:

• vg_amr_alpha1_target and vg_amr_alpha2_target for calculating target refinement levels based on alpha and J/B. These need to be shifted based the refine thresholds.

[markusbattarbee]

Now that we're changing the output alphas in vlsv files to already include the dx scaling, this isn't needed. (And for old stuff like FID, it would need to be reverted to take in alpha and jperb, let's not go there)

[lkotipal]

Now that we're changing the output alphas in vlsv files to already include the dx scaling, this isn't needed. (And for old stuff like FID, it would need to be reverted to take in alpha and jperb, let's not go there)

I don't think this ever happened, nor is there a PR to Vlasiator for this. I think this should be re-opened

[markusbattarbee]

https://github.com/fmihpc/vlasiator/blob/9e7be0f8ad1480c198ce14fdfb077f1a687c6e4a/fieldsolver/derivatives.cpp#L810

In here

https://github.com/fmihpc/vlasiator/pull/920/files#diff-66e9971f9feb14198d5119745329c6e5b7e4dc813a428c1a13ae36b2aa79f2e0