Nathanael (Nate) Maytan

Make sqlalchemy use connection pooling for catalog db

4

This PR results from the investigation done in #663, as well as a group coding session to implement a new benchmark for the catalog db (especially with input from @genematx,...

In ref of #733. Both cases insist upon a dataframe-compatible shape. For `write_dataframe`: ``` d = {'a': [1,2,3], 'b': [4,5,6,7]} c.write_dataframe(d, key='n') [...] ArrowInvalid: Column 1 named b expected length...

This is a minor detail: if you generate a database migration script with alembic, the current mako template puts single quotes around the revision identifiers whereas the linter wants double...

If you call `update_metadata` from a client, it will accept a simple string argument (as it is serializable), but then causes server 500's when doing anything with that node. The...