Yesterday morning I had thought that the new thing would just be a composition of datasets... but if it actually embeds a viewconf, then it's reasonable for it to be handled here: It won't take much code, and this will let us:
set the has_visualization facet at index time. (I think that's the name?)
catch some errors at index time?
use vis_preview.py consistently. (I think the pipeline developers have used it to preview the dataset before publication.)
FWIW: I'd go for a longer vitessce hint string than just uia: Bytes are cheap! Comprehension is expensive!
Filing as draft: Prereqs need to be in place before this can be used.
Yesterday morning I had thought that the new thing would just be a composition of datasets... but if it actually embeds a viewconf, then it's reasonable for it to be handled here: It won't take much code, and this will let us:
has_visualizationfacet at index time. (I think that's the name?)vis_preview.pyconsistently. (I think the pipeline developers have used it to preview the dataset before publication.)FWIW: I'd go for a longer vitessce hint string than just
uia: Bytes are cheap! Comprehension is expensive!Filing as draft: Prereqs need to be in place before this can be used.