Allow for Pytree input for coil objective targets and bounds

[kianorr]

Input should allow for different bounds/targets for different coils. I don't think this is possible yet

[dpanici]

• Preferred: parse the inputs as a pytree like in FixSumCoilCurrents
• Or document the current method better

[dpanici]

1164 if I have a SumMagneticField with [ToroidalMagneticField, MixedCoilSet]

want to do CoilLength on this, where we target lengths in the mixed coilset differentlt

should target be like [ [], [{"length": 1}, {"length":2}]]

I think we should enforce it having the same pytree structure as the SumMagneticField, but just ignore the entries corresponding to non-coil/coilset leafs

[dpanici]

Alternatively, we could allow QuadraticFlux to take in a list of magnetic fields and it just automatically sums them, that way we can say things = (tf_field, coilset) and then have our ususal coil obj target the coilset without needing additional plumbing