[rag evals][1/n] refactor base scoring fn & data schema check

[yanxi0830]

What does this PR do?

• Refactor BaseScoringFn to be just a minimal interface, add new RegistrableBaseScoring
• Refactor data schema check
  • To separately evaluate retrieval component in RAG, we will have scoring functions needing "context" column additionally.
• Refactor braintrust eval (more scoring fn added & tested in following PR)

Test Plan

pytest -v -s -m llm_as_judge_scoring_together_inference scoring/test_scoring.py --judge-model meta-llama/Llama-3.2-3B-Instruct
pytest -v -s -m basic_scoring_together_inference scoring/test_scoring.py
pytest -v -s -m braintrust_scoring_together_inference scoring/test_scoring.py

pytest -v -s -m meta_reference_eval_together_inference eval/test_eval.py
pytest -v -s -m meta_reference_eval_together_inference_huggingface_datasetio eval/test_eval.py

Before submitting

• This PR fixes a typo or improves the docs (you can dismiss the other checks if that's the case).
• Ran pre-commit to handle lint / formatting issues.
• Read the contributor guideline,
  Pull Request section?
• Updated relevant documentation.
• Wrote necessary unit or integration tests.

[yanxi0830]

We will likely need to revisit expected data schema when evaluating for tools.

[raghotham]

its better for scoring/evals/post_training to all define their own schema for what they want the dataset to obey. The actual checking of the schema can be done in datasetio.

[ashwinb]

I still don't understand why we have these base classes because we have already declared what our impls need to do in terms of datatypes in our APIs. so the datatype for a scoring function already exists. Let's say I am implementing a new scoring function -- why do I need another base class and inherit from there? If there is some utilities I need for implementing the functions, they would be just utils / free functions or in the worst case, some mixins.

Can you explain the need for base classes please? I am very allergic to inheritance as you and @raghotham knows :)

[yanxi0830]

This class BaseScoringFn(ABC): is mostly separated out based on feedback from Tejas for his use case without
registration (syncing with you offline).

For our llama-stack implementations, I agree we don't need this separate BaseScoringFn, and could just use RegisteredBaseScoringFn as mixins.

[raghotham]

Can datasetio abstract schema validation instead of creating mixins to be used by scoring and evals?