Call to get_associated_ws_objects can be slow - ~ 4 sec #14
Description
Now that genome features are linked to ontology terms, the number of links can be very large.
It just so happens that one the terms I've been testing with, GO:0008152 (metabolic process), has a great many features linked. And clearly when we are linking features there can be many features across many genomes.
In any case, the call to get_associated_ws_objects for this term currently takes about 4 seconds, even with a small range of 10 items (offset=0, limit=10). Although perhaps acceptable for the initial request, this is quite painful for implementing paging. It could also be argued that paging or scrolling through thousands of items is not an optimal strategy. I'll have another ticket which addresses this, at least partly.
Depending on whether this can be fixed, and how fast it can get, we may need to develop an alternate strategy for clients. (E.g. use larger page sizes larger, with a background process to pre-fetch additional pages as need be.)