Mask from chopper acceptance diagram

[bingli621]

It could be useful to have a mask generated from the chopper acceptance diagram. This mask can be applied to the neutron source, to get neutrons that will go through all chopper.
The masked source can help with the following two things:

1. Get an estimate of the flux on sample when summed within the subframes.

An alternative way to estimate the flux is to read the monitor counts from the monitor placed right before the sample. But this reading comes from a sampling of all netruons the source provide, and won't be accurate if undersampled. I wonder if a direct summation of regions allowed by the mask will be more accurate?

2. Speed up the simulation by ignoring neutrons that will be blocked by any chopper.

This will speed up direct geometer instrument significantly since the majority of neutrons are blocked by choppers. See attached chopper acceptance diafgram and frame propagation diagram for reference.

This idea came from brief discussions with @nvaytet @YooSunYoung and @SimonHeybrock

[nvaytet]

I think this should simply be implemented as a new source where the mask has been applied manually.
The source data is just a Scipp data array with dimensions wavelength and time.

[bingli621]

For T-REX and CSPEC, the acceptance diagram will change, depending on the central wavelength user requests, what is the repetition rate multiplication (rrm) number, and whether to use the high flux (wide) chopper slits or high resolution (narrow) slits. So we need to have a way to update the mask according to the operational parameters.