Appion

When a CL2D run finishes, it uploads a set of clustering stacks (e.g. 2,4,8,16 ..., n class averages) and a reference stack (which is just a pointer to the last level of classification, i.e. level n). I've noticed that, occasionally, when a CL2D run finishes, one cannot see the particles associated with the reference stack, viewstack does not show any associations. the clustering stacks are always ok. This has been an issue for several years now, since I put CL2D into the pipeline. I finally decided to investigate this further, as it's rather annoying, and found that some particles do NOT get put into any class by CL2D. Thus, the SQL query that returns the resulting particles, for example in the case of 16 class averages, will return 16 arrays for ApClusteringReferenceData, but 17 arrays for ApAlignReferenceData. This is because, for the particles that don't get assigned to a class, the reference comes out as NULL (the particles are still saved to the db with align-particle-id and stack-particle-id).

I am committing a few fixes to this bug, including in viewstack.php and alignSubStack.py. These have been tested at Salk and seem to work ok with CL2D displays and substack inclusion / exclusion, but I have not tested with other methods commonly in use.