Initial setpoint implementation

[ksunden]

Supercedes #41

Requires https://github.com/wright-group/WrightTools/pulls/883 and therefore will fail automated tests.

There are certainly some rough edges/tweaking of default parameters to be done.

Differences from wt2:

• splines in differential axis space, not absolute
• uses moments ranther than gaussian fitting
• fewer pre-cleaning operations performed on test
• pre-cleaning operations external to workup function, thus finding the "setpoint measured channel" has to be done prior to this call.
• relatively memory safe (though not perfect in that respect)
• units behavior may be different, have not tested extensively, fitting is done in wn space for now (see Handling of unit systems #38)
• Representation generated uses detuning in a signed color map, rather than plotting detuning vs absolute motor as wt2 did. This is more in line with how the intensity plot is generated.

[ksunden]

Example of the auto-generated plot, as well as comparison to wt2 differentials

[ksunden]

May wish to flip the cmap for energy data, such that "blue shift" and "red shift" are consistent, currently it follows the convention of "red positive", "blue negative" for a) simplicity, b) consistency with that colormap's usage in general, though red/blue shift is a good argument