I admit I don't really know how the current system operates, but it seems from the "Dataset comparisons" section that the producer of the plots has to specify all the comparisons that they want to be possible up-front? I'm imagining a scheme where all comparisons are possible and one just ticks some boxes to decide what datasets to overlay. I think you'd have that on the same page where the list of dataset comparisons already exists, and then within the page itself also have tick boxes to turn different components on and off for different plots.
#1 Updated by Jeremy Wolcott about 3 years ago
- Status changed from New to Rejected
That said, "old" comparison directories (from previous runs of the Python stuff) aren't removed, and will show up on the index page. So one can run it several times with different datasets enabled to get different comparisons. Example on my test page: http://nusoft.fnal.gov/nova/users/jwolcott/validation/new_test (note how there are two "dataset comparisons" links).
#3 Updated by Jeremy Wolcott about 3 years ago
Well, because the JSROOT component of this was added on top of a system that already existed, that's how they're delivered here. (Originally they were just static images, remember.) The logic for combining them all lives in the server-side Python.