Fix failure to ignore darks/flats when getting the other from separate file

[yousefmoazzam]

Fixes IMGDA-650

An alternative to #656, as a precursor to the refactoring of the darks/flats config class.

Checklist

• I have added tests that prove my fix is effective or that my feature works
• New and existing unit tests pass locally with my changes
• I have made corresponding changes to the documentation

[dkazanc]

thanks @yousefmoazzam , looks good. So in this instance _empty_aux_array from dataset.py is not called as you create a dummy array of zeroes or ones directly here? I'm trying to understand in what situation _empty_aux_array is triggered anyway?

[yousefmoazzam]

I've not looked into _empty_aux_array(), I actually accidentally stumbled upon this way to fix the issue during refactoring when I was explicitly enumerating each possible case in terms of code paths.

From a cursory look at where the function is mentioned https://github.com/search?q=repo%3ADiamondLightSource%2Fhttomo%20_empty_aux_array(&type=code, it looks like it gets called only when the darks/flats property is accessed but their value is None.

So I'd guess that, somehow, in the other PR to fix this, the darks/flats were None for some reason - and so changing _empty_aux_data() was a viable way to fix it.

(As I was already doing refactoring that was on top of this, I decided to split this out to make it clear where the bug fix was, rather than having the bug fix be buried within the subsequent refactoring).

[dkazanc]

Thanks! Yes, I was curious about it just because there was this bug which might need fixing if any other case refers to _empty_aux_array() and that array is getting build. For that change all looks fine, thanks for the separate PR on this.