Fix sign of phase shift in DataSet.apply_symop

[JBGreisman]

Based on visual inspection of maps, it appeared that DataSet.apply_symop() was applying phase shifts in the opposite direction of the specified symmetry operation. This PR changes the sign for how phase shifts are applied, and updates the relevant test.

[JBGreisman]

To do:

• It would be useful to make a "gold standard" pair of F/phi with a known real-space translation between the corresponding maps so that we can write a cleaner test.

[codecov-commenter]

Codecov Report

Merging #63 (15c41a7) into main (cb5b764) will decrease coverage by 0.05%. The diff coverage is 95.23%.

:exclamation: Current head 15c41a7 differs from pull request most recent head e75519d. Consider uploading reports for the commit e75519d to get more accurate results

@@            Coverage Diff             @@
##             main      #63      +/-   ##
==========================================
- Coverage   98.88%   98.83%   -0.06%     
==========================================
  Files          38       38              
  Lines        1441     1453      +12     
==========================================
+ Hits         1425     1436      +11     
- Misses         16       17       +1     

Flag	Coverage Δ
unittests	98.83% <95.23%> (-0.06%)	:arrow_down:

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
reciprocalspaceship/dataset.py	98.82% <95.23%> (-0.21%)	:arrow_down:

---

Continue to review full report at Codecov.

Legend - Click here to learn more Δ = absolute <relative> (impact), ø = not affected, ? = missing data Powered by Codecov. Last update 9dc62ba...e75519d. Read the comment docs.

[JBGreisman]

My most recent push added a test that uses fmodel data to compute a real-space map, shifts the map based on a desired translation using np.roll, and then back computes corresponding complex structure factors. These are compared to the results of DataSet.apply_symop().

[JBGreisman]

I am still working on this for the complex structure factors. This test may not be working as well as I had hoped, but I am not yet ready to come to that conclusion

[JBGreisman]

This PR now fixes DataSet.apply_symop() for use with both |F|+phi and complex SFs. I updated the test to use a P1 dataset avoid an interdependency caused by the use of apply_symop() in DataSet.to_reciprocalgrid().

P.S. -- I updated my emacs recently to auto-apply black formatting, which is why there are additional changes to the test_dataset_symops.py file