Streamline graph DB queries

[jsstevenson]

It's been true for a while now that some of our core API endpoints require multiple ordered requests to the graph DB to succeed. For example, searching for a study first retrieves that study and its relationships, but can then spur further calls for certain kinds of variations and therapies. This is totally reasonable for now but as requirements and the data models begin to settle, we should be aiming to condense those protocols into a single DB call eventually.

[jsstevenson]

It might also be worth looking at whether some queries should be managed transactions vs the new driver.execute_query API https://neo4j.com/docs/python-manual/current/query-simple/

[jsstevenson]

get_nested_study makes a LOT of queries. Ideally this would be one. At minimum, rather than iterating through relations, there should be a single structured result.

[jsstevenson]

I think we made the decision to table this one for a while until some of the data models are more solidified -- the changes that I have in mind would be quite brittle to schema updates.

[korikuzma]

@jsstevenson I plan on updating the normalizers/ga4gh.vrs and then pulling in cat-vrs/va-spec python model changes.