Investigate DataclassBridge caching and deallocation performance

[tangkong]

Current Behavior

The QDataclassBridge structure was vendored from atef nearly verbatim, along with some comments suggesting we revisit the dataclass caching mechanisms

Expected Behavior

Framework works as intended, with no noticeable issues

Context / environment

Noticed in #38

Steps to Reproduce (for bugs)

N/A

Suggested Solution

Look for issues in caching, make sure garbage collection is occurring as planned, etc

[tangkong]

To be clear, we (ECS) haven't run into any problems with this in our time using the pattern. We created this issue preemptively, but we've not run into a bug yet