WIP: Add JSON serialization and additional frame testing utilities

[chadbailey59]

I'm working on adding some custom FrameProcessor testing capability to the pipecat CLI. This PR adds a few things to help with that.

The goal is to be able to define input frames and optional expected output frames in JSON, like this:

{
  "input_frames": [
    {"type": "TextFrame", "text": "hello world"},
    {"type": "TextFrame", "text": "testing 123"}
  ],
  "expected_output": [
    {"type": "TextFrame", "text": "HELLO WORLD"},
    {"type": "TextFrame", "text": "TESTING 123"}
  ]
}

The existing run_test function requires you to define expected output frames and has test assertions built in. This PR effectively re-implements run test as run_test_from_file to move the assertions part to a separate function. This lets us make expected_output optional, so we can run the FrameProcessor without knowing what output we expect when we're building our FrameProcessor.

The custom validator also adds a bit more utility, such as validating field contents instead of just frame types. It also starts to build out a bit of a testing DSL, like adding _contains to a string field to test for contains instead of equals, like: {"type": "TextFrame", "text_contains": "TESTING"} from the example above.

There's some mild duplication in creating a separate run_test_from_file function. It's just extremely simple pipeline creation, and I think it's worth it not to disturb the existing run_test function. But I could be convinced otherwise.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.