MemHive: Experimental Python Subinterpreters Orchestrator

I love the readme:

"DO NOT USE IN PRODUCTION OR YOU WILL BE FIRED. UNLESS YOU WORK AT EDGEDB."

For context, this is a way to share data between the new subinterpretter feature in Python.

Given there is no doc, the tests give a glimpse of what it looks like:

https://github.com/edgedb/memhive/blob/main/tests/test_basic...