clean up test suite errors

[xzrderek]

Note

Updates Langfuse integration to be compatible with 2.x.

• Replace from langfuse import get_client with from langfuse import Langfuse and add a get_client() shim returning Langfuse() in eval_protocol/adapters/langfuse.py
• Update Langfuse-related tests to instantiate Langfuse() directly and keep observe usage
• Add pytest.skip markers for HuggingFace datasets no longer available (e.g., gsm8k, squad) and one integration test

^{Written by Cursor Bugbot for commit 3f174a7. This will update automatically on new commits. Configure here.}