Add field to things for boundary error objectives

[f0uriest]

• Adds the field parameters as optional inputs to {Vacuum}BoundaryError.compute
• Adds field_fixed option to objectives - currently this doesn't do anything since we usually can't precompute field qtys for a changing boundary, but could add special logic for certain field types later. Mainly this flag is for use in regular free boundary solve so you don't need to include explicit constraints targeting the field.

Resolves #855

Should ideally merge #726 and #853 first so we have some sensible coil constraints before doing single stage stuff

[review-notebook-app]

Check out this pull request on 

See visual diffs & provide feedback on Jupyter Notebooks.

---

Powered by ReviewNB

[codecov]

Codecov Report

Attention: Patch coverage is 96.66667% with 1 lines in your changes are missing coverage. Please review.

Project coverage is 94.63%. Comparing base (073efec) to head (6797100).

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #935      +/-   ##
==========================================
- Coverage   94.64%   94.63%   -0.02%     
==========================================
  Files          86       86              
  Lines       21580    21586       +6     
==========================================
+ Hits        20424    20427       +3     
- Misses       1156     1159       +3     

Files	Coverage Δ
desc/magnetic_fields/_core.py	82.35% <100.00%> (-0.31%)	⬇️
desc/objectives/_free_boundary.py	96.87% <100.00%> (+0.06%)	⬆️
desc/io/equilibrium_io.py	86.22% <92.30%> (-0.96%)	⬇️

[ddudt]

Why were these changes necessary for this PR?

[f0uriest]

SplineMagneticField requires some custom behavior to work correctly as a pytree (required for it to be part of things), and the previous method of defining a custom flattening method didn't consider all the extra attributes that get added with the Optimizable base class. This refactors things a bit so that each class can use the default flattening method but also define specific attributes as static or dynamic without having to re-define the entire flattening/unflattening method (and then having to account for every attribute individually)

[f0uriest]

Note benchmarks are expected to fail due to updated argument name (renamed ext_field -> field)