stills_indexer: Add a deterministic option for experiment identifiers

[dwpaley]

No description provided.

[codecov]

Codecov Report

Merging #1864 (e90423d) into main (fc216c8) will increase coverage by 0.03%.
The diff coverage is 13.33%.

❗ Current head e90423d differs from pull request most recent head b49dc97. Consider uploading reports for the commit b49dc97 to get more accurate results

@@            Coverage Diff             @@
##             main    #1864      +/-   ##
==========================================
+ Coverage   67.10%   67.14%   +0.03%     
==========================================
  Files         619      617       -2     
  Lines       70002    69532     -470     
  Branches     9732     9671      -61     
==========================================
- Hits        46977    46686     -291     
+ Misses      21059    20906     -153     
+ Partials     1966     1940      -26     

[Anthchirp]

Reference to previous identifier discussions: PRs #902 and #1086

[dwpaley]

Thanks, @Anthchirp . According to the previous discussion and the unimplemented "timestamp" identifier option, it looks like a feature like this had been intended. I think this PR meets the goal of the original timestamp idea, but also (1) handles multiple lattices in a shot with unique identifiers for each, and (2) assumes nothing about the image format.

[dwpaley]

After cctbx/dxtbx#438 we think this is no longer needed because we don't have to delete imagesets during cctbx.xfel.merge jobs. Thus we can always look at the imageset if we want to connect processed data to the original images, and no longer need to duplicate that information in the identifier.