refactor so that the default dataset fetch is done in the top-level xena app

[nathandunn]

• Can use memoization (history of N) or hashing (what I'm currently doing) to not reprocess.
• Create a service class for calling (each level could be cached): fetch -> filter -> sort
• Calls service from top-level and then drops all of the states back down, which then filters down to individual component properties.
• this would allow us to get rid of the caching then.

[nathandunn]

[github-actions[bot]]

Stale issue message

[nathandunn]

This is done