Integrate the ParSocial code Task Queues into the eqsql package

The ParSocial paper code implemented a TaskQueue API with 3 implementations, all of which used the core eqsql.eq code.

RESTful web API used flask to call eq code on a remote machine where the DB was located.
Globus Compute (i.e., funcx) based API used funcx to call the same functions that web-based API used.
Local API that just forwards directly to the eqsql.eq code.

These only implement a subset of the eqsql.eq functionality. Determine what more needs to be added, and harden the web and funcx ones appropriately.

[x] protocol for task queue
[x] local task queue implementation
[x] update swift-t extension and tests of that (tests/api_tests) to use new local queue
[x] remote task queue globus compute implementation
[x] remote task queue flask implementation (wrapper around globus compute)
[x] R-API updates to match new task queues
[ ] Performance Tests -- local vs. remote with 1000s of tasks
[ ] Non-persistent ME Use Case - ME submits tasks, stops (saves future ids?), restarts (reconnects, futures recreated etc.)
[x] Update mews-project creator to match new API

emews / EQ-SQL