Switch default to return pathlib.Path rather than py.path.local().

[dagewa]

Also update deprecation message to announce that the latter will be completely removed later.

[dagewa]

Seems sensible to stage this. So, this changes the default, then we can remove all pathlib=True in dials_data calls across DIALS and dxtbx. Finally we can remove the choice entirely.

[ndevenish]

I'd probably go full nuclear; outright remove; catch and raise an error for it's presence in kwargs, just always return Path.

Do a major version bump (which I think by the dials-data deployment we do here, e.g. update pyproject.toml) so that it's a semantic version break.

[dagewa]

Ok, fair enough - will go for that

[dagewa]

@ndevenish I rediscovered this. Are you happy with it, now it has full removal and version bump?

[codecov]

Codecov Report

❌ Patch coverage is 80.00000% with 1 line in your changes missing coverage. Please review.
✅ Project coverage is 49.69%. Comparing base (60ceb6e) to head (3cbee3c).
⚠️ Report is 1 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #509      +/-   ##
==========================================
- Coverage   56.25%   49.69%   -6.56%     
==========================================
  Files           6        6              
  Lines         368      330      -38     
==========================================
- Hits          207      164      -43     
- Misses        161      166       +5     

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[ndevenish]

Yes, good legacy cruft to remove

[ndevenish]

I'm going to revert this as this has killed all of dials (well, 737 tests). Having a redundant parameter should at best be a warning, not a ValueError.