Dither analysis

[julienguy]

This PR adds some documentation and a script to perform a dither analysis with desimeter. The most recent dither sequence from 2020/03/15 is used as an example.

On this particular sequence, the result is not great, the rms of residuals is of 0.5 arcsec (there is still work to be done, in particular validate the ADC rotation correction: signs and zero points).

[julienguy]

I had forgotten the option --zbfit to refit the Zhao-Burge polynomial coefficients from the fiducials. With this, the desimeter-dither rms reduces to 0.42 arcsec which is a bit better. I think this indicates we have to work more on the fvc2fp transform, and in particular as a function of the ADC angles.

[schlafly]

Just verifying---this is without the "cheat" refitting rotation + scale + offset parameters? The trends are clearly very low order.

[julienguy]

Yes, 'this is without the "cheat" refitting rotation + scale + offset parameters'.

[julienguy]

• expid is now a required argument.
• I added back the final 'cheat' transformation results because this is indeed useful information.
• this replaces the desicmx script because I think it is simpler to have all desimeter related codes in the same place, I will remove the desicmx script to avoid confusion.

number of valid targets from dither analysis = 1758
residual scale along HA       = 1-7.8e-05
residual scale along Dec      = 1-4.1e-06
residual pointing offset dHA  = -0.11 arcsec
residual pointing offset dDec = -0.15 arcsec
residual rotation             = -9.19 arcsec
   desimeter(guide+fvc)-target rms 2D = 0.44 arcsec
                 dither-target rms 2D = 0.23 arcsec
   desimeter(guide+fvc)-dither rms 2D = 0.42 arcsec
transformed(desimeter) -dither rms 2D = 0.26 arcsec