Add unit cell angle distribution plots to cluster plotting code

[jbeilstenedmands]

This PR add plots of unit cell angle distributions to the unit cell cluster plotting code, e.g.

Distributions are only plotted for non-symmetry constrained angles, e.g. won't appear for cubic crystals, only beta distribution plotted for P2.

I have also implemented a heatmap version of the unit cell scatter plot to the cluster plotting code. For now I have implemented this as an option and kept the default as before for dials.cosym, xia2.multiplex etc, but I have added the new version to the dials.merge report :
New option:

Default:

As you can see, the default is probably more appropriate for up to several hundred crystals (the plots here are 34 crystals), but when you get into the 1000s, I think the heatmap is a more useful plot due to the point crowding.

[codecov]

Codecov Report

Merging #2197 (1a4f6a4) into main (6a7ada4) will increase coverage by 0.01%.
The diff coverage is 92.59%.

@@            Coverage Diff             @@
##             main    #2197      +/-   ##
==========================================
+ Coverage   80.65%   80.66%   +0.01%     
==========================================
  Files         588      588              
  Lines       66601    66920     +319     
  Branches     9393     9476      +83     
==========================================
+ Hits        53718    53982     +264     
- Misses      10804    10850      +46     
- Partials     2079     2088       +9     

[rjgildea]

Thanks for adding this, it would be good to see some examples for SSX data, i.e. with 1000s of unit cells. For the scatter plot version improvements it might be worth exploring adding alpha levels to help reduce overcrowding of the plot (although probably outside of the scope of this PR).