Smarter M1 gain

[tepickering]

We want to be able to set gains for each Zernike mode based on how closely the mode matches a mirror bending mode. Some of the high order modes we fit by default are very stiff and difficult to correct. Since the mirror is unlikely to bend in such a way, it is probably best to ignore them or apply a very small gain to them when calculating forces.

[tepickering]

Still worth considering, but this has been somewhat mitigated by changing the default number of modes to 22 which encompasses all of the major mirror bending modes.

[tepickering]

which modes are used to calculate forces is now hard-coded. however, the same gain is applied to all of them equally. it would probably be better to have a tapered gain for the stiffer modes. need to go back through the old documentation for more insight on this.