Implement use of "runs" in messages.py?

[chriscarrollsmith]

This documentation makes it seem like we should be using a "run" for every chat step, regardless of whether we expect tool calling. The purpose of runs is basically to encapsulate a process of assistant response generation that may or may not require multiple steps. https://platform.openai.com/docs/assistants/overview

Spec for the run object: https://platform.openai.com/docs/api-reference/runs/object
Spec for submitToolOutputs endpoint: https://platform.openai.com/docs/api-reference/runs/submitToolOutputs

Review the sample code here: https://platform.openai.com/docs/assistants/quickstart#step-4-create-a-run
Some useful info on run lifecycle management here: https://platform.openai.com/docs/assistants/deep-dive#run-lifecycle