Add a 'scaled' flag to the reflection table.

[jbeilstenedmands]

Introduce a 'scaled' flag, which is set at the end of scaling to indicate which are the good scaled reflections (not excluded, not outliers etc). This is the inverse of the current 'bad_for_scaling' flag, which is unclear and prone to misuse, as indicated by the confusion in the discussion in #1268.

Ideally, the scaled flag will replace all uses of 'bad_for_scaling' outside scaling code - this has not been implemented yet, as it will require updating some of the regression test files (e.g. l_cysteine_4_sweeps_scaled). If we're happy with the proposed changes here, then once merged I can update the test files and implement further changes of bad_for_scaling to scaled in places such as dials.util.filter_reflections

[codecov]

Codecov Report

Merging #1377 into master will increase coverage by 0.01%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##           master    #1377      +/-   ##
==========================================
+ Coverage   64.70%   64.71%   +0.01%     
==========================================
  Files         616      616              
  Lines       69735    69762      +27     
  Branches     9546     9549       +3     
==========================================
+ Hits        45120    45145      +25     
- Misses      22835    22837       +2     
  Partials     1780     1780              

[ndevenish]

I think we're ready to merge this when the conflicts are resolved

[ndevenish]

What's the relation between scaled and bad_for_scaling? It seems to set scaled as the exact opposite quite often, but I can't tell if there are scenarios where they diverge.

[ndevenish]

Also, needs a newsfragment (probably .api?)

[jbeilstenedmands]

@ndevenish you are correct, as in the PR description, I'm moving over to the scaled flag in general outside of internal scaling code, although keeping bad_for_scaling for convenience within scaling (at least for now). Newsfragment incoming.

[ndevenish]

D'oh, that's what I get for reading the code and not the description.

Is the long term idea to properly deprecate the bad_for_scaling?