Users have trouble finding what they really want because they get confused between the parent and child datasets. Proposal:
At index time, for parents, pull in the assets list for child, under a new attribute. (Want to keep parent and child files separate in UI: probably will not list child files.)
for children, add a flag so they can be hidden in search results.
children still accessible by URL, but we will avoid making direct links in UI.
Questions:
Are we confident there will only be one child, and no grandchildren?
This only deals with datasets, or are we interested in pulling dataset assets into samples?
cc @ilan-gold
Users have trouble finding what they really want because they get confused between the parent and child datasets. Proposal:
Questions: