Add EMPIAR 11618 experimental recon

[garrettwrong]

Adds gallery example for EMPIAR 11618 recon.

The data for this is clean enough that we can use the randomized class selection to avoid the expensive entire class average set generation. Raw data to recon volume with 1 A100 GPU is taking about 2 hours and about 20A at 129 pixels.

I did try using various preprocessing techniques. While most were slightly faster preprocessing, they all had decreased recon accuracy in my (limited) testing runs.

Most of my runs were at 50 neighbors per class average. I'm currently testing with 100 as written here and will decide based on the FSC and visuals what to initially recommend before proofreading and opening up for review.

[garrettwrong]

Using the non-legacy preprocessing seems to give up about 0.5A in this configuration so I went with that. Probably in the future it can improve.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 90.87%. Comparing base (50d79b3) to head (34006d1).
⚠️ Report is 4 commits behind head on develop.

Additional details and impacted files

@@           Coverage Diff            @@
##           develop    #1361   +/-   ##
========================================
  Coverage    90.87%   90.87%           
========================================
  Files          135      135           
  Lines        14678    14678           
========================================
  Hits         13339    13339           
  Misses        1339     1339           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

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

[j-c-c]

Looks good!

[garrettwrong]

Github having some issues, will return to restart these later. This code diff doesn't actually run in anything other than the doc build so the CI should go through normally ....

[garrettwrong]

Github having some issues, will return to restart these later. This code diff doesn't actually run in anything other than the doc build so the CI should go through normally ....

GH appears healthier today. Restarted.

[garrettwrong]

Going to merge this in just to keep the PR area tidy. We can tweak it before the next release if any concerns.