dials.scale: allow wildcard for experiment in exclude_images

[graeme-winter]

Just had to write a script to

dials.scale ../work/symmetrized.* d_min=1.5 \
exclude_images=0:21:100,1:21:100,2:21:100,3:21:100,4:21:100,5:21:100,6:21:100,7:21:100,8:21:100,9:21:100,10:21:100,11:21:100,12:21:100,13:21:100,14:21:100,15:21:100,16:21:100,17:21:100,18:21:100,19:21:100

but it feels like a legitimate use case for modest-large-N data sets to say exclude imges a:b for all runs.

Propose syntax: exclude_images=*:21:100 in above case (since this is a common sense expectation)

[graeme-winter]

use case: multi-crystal damaged data set where we hit it about 20x harder than intended

[graeme-winter]

Alt: by analogy with exclude_images_multiple -> exclude_images_all?

[dagewa]

Currently specifying exclude_images=start:stop with multiple experiments raises an error ("Exclude images must be in the form experimentnumber:start:stop for multiple experiments"). Perhaps this restriction is artificial? If exclude_images=start:stop was simply made to mean exclude these images from every experiment, then it would work for single experiments and multiple experiments alike. Might that be a neat solution?

[jbeilstenedmands]

I do like that suggestion @dagewa . As long as a message is printed out to say what is happening for the multiple experiments case, that would be a very neat solution.

[graeme-winter]

Certainly happy with that - essentially avoiding ${USER} from having to write a script to perform a task was my desire. I will update the PR accordingly.