Add ability to mask out specific regions of a CCD

[martinjohndyer]

We have the (m)ask function in setaper, but you can only create a mask if it's linked to an aperture. It came up in a student viva in the case where you have a bright Compo star you actually want to mask out a that specific region of the CCD, notably have it stay static on a given pixel region not move with the aperture.

@vikdhillon @SGParsons and I talked through a few options of how you could block out a section, I think a capital (M)ask pixels might have you somehow click on the plot to black out a region you want to be excluded.

How that would work in practice might be interesting. Ideally it would be part of setaper, or maybe a separate setmask? You'd either have to extend the .ape file definition or have a separate mask file that has to be included in the reduction. Maybe the second would be more reusable, define a compo.msk file? I'm sure there could be other good uses for that. But this is just throwing out ideas, the first thing to do would be to look at how the aperture masks are implemented in the code.

[StuartLittlefair]

I'm in favour of a mask file. Thinking down the line mask.msk files can be re-used with shiftadd etc.

One thing we'd need to decide is what we do with masked pixels. There are multiple scenarios:

1. Include masked pixels in calculation, but flag data as masked-sky or masked-object in the log file
2. Reject masked pixels silently
3. Reject masked pixels from calculations, and also flag data as in 1.

Personally I feel option 3 is the correct way to handle this