Deprecate datablock.*Diff pseudoclasses

[Anthchirp]

BeamDiff, ScanDiff, GoniometerDiff, DetectorDiff, SequenceDiff in dxtbx.datablock are replaced by functions in dxtbx.model.compare (beam_diff, etc.)

They don't really have anything to do with datablocks or experiment lists.

[Anthchirp]

Question here: There are a bunch more of these comparator "classes" (BeamDiff, ScanDiff, GoniometerDiff, DetectorDiff) in dxtbx.datablock. Should these go into dxtbx.model.experiment_list, or elsewhere collectively, eg. dxtbx.model.compare?

They don't really have anything to do with datablocks or experiment lists.

[codecov]

Codecov Report

Merging #302 (ebf6549) into main (8cdb7d3) will increase coverage by 0.00%.
The diff coverage is 10.00%.

@@           Coverage Diff           @@
##             main     #302   +/-   ##
=======================================
  Coverage   46.73%   46.74%           
=======================================
  Files         232      233    +1     
  Lines       19207    19221   +14     
  Branches     2763     2764    +1     
=======================================
+ Hits         8977     8985    +8     
- Misses       9733     9739    +6     
  Partials      497      497           

[phyy-nx]

Looks like no impact on my side if these classes are moved.

[Anthchirp]

@graeme-winter You self-requested a review on this, I remember it had something to do with tolerance phil thingies. Do you want to add your 2p here or is this a separate PR?

Otherwise this here is done - tests pass and everything.

[graeme-winter]

@graeme-winter You self-requested a review on this, I remember it had something to do with tolerance phil thingies. Do you want to add your 2p here or is this a separate PR?

Otherwise this here is done - tests pass and everything.

Warning: Ambiguous parameter definition: wavelength = 0.979499
Best matches:
  input.tolerance.beam.wavelength
  geometry.beam.wavelength
Assuming geometry.beam.wavelength was intended.

I think that this did not used to be the case? 🤔

[graeme-winter]

@graeme-winter You self-requested a review on this, I remember it had something to do with tolerance phil thingies. Do you want to add your 2p here or is this a separate PR?
Otherwise this here is done - tests pass and everything.

Warning: Ambiguous parameter definition: wavelength = 0.979499
Best matches:
  input.tolerance.beam.wavelength
  geometry.beam.wavelength
Assuming geometry.beam.wavelength was intended.

I think that this did not used to be the case? 🤔

Grey-Area override :) $ dials.import ../insu_7_testbeamline_2.nxs origin=-82.7366,77.3027,-332.998
Sorry: Ambiguous parameter definition: origin = -82.7366,77.3027,-332.998
Best matches:
  input.tolerance.detector.origin
  geometry.detector.panel.origin
  geometry.detector.hierarchy.origin
  geometry.detector.hierarchy.group.origin

There we go... I think this was what I was meaning - I think this is somewhat related but I am minded to make that a different pull request.

[graeme-winter]

Scanned change set, did not see anything which would make future debuging harder and moving away from pseudoclasses is always a good thing, thank you for the work here.

Am slightly concerned by use of functools.partial as an indicator of poor overall design, but that is out of context for this change set. I will raise an issue with my other concerns as raised in the conversation.