[Evals API][6/n] meta-reference llm as judge, registration for ScoringFnDefs (#330)

* wip scoring refactor

* llm as judge, move folders

* test full generation + eval

* extract score regex to llm context

* remove prints, cleanup braintrust in this branch

* change json -> class

* remove initialize

* address nits

* check identifier prefix

* udpate MANIFEST

llama_stack/providers/registry/scoring.py

@@ -20,6 +20,7 @@ def available_providers() -> List[ProviderSpec]:
             api_dependencies=[
                 Api.datasetio,
                 Api.datasets,
+                Api.inference,
             ],
         ),
     ]