fix: update rag_eval_ragas notebook for ragas>=0.4.3 API

[SyedShahmeerAli12]

Summary

Fixes #285 - updates the rag_eval_ragas cookbook notebook to work with ragas>=0.4.3, which introduced breaking API changes.

What changed:

• Replaced HaystackLLMWrapper (removed in ragas 0.4.x) with llm_factory and embedding_factory from the native ragas API
• Each metric now receives its LLM at construction time, e.g. AnswerRelevancy(llm=llm, embeddings=embeddings)
• Updated imports to use ragas.metrics.collections (non-deprecated path in 0.4.x)
• Replaced the standalone ragas.evaluate() section with RagasEvaluator.run() per sample
• Cleared all stale cell outputs

Tested:

• Ran end-to-end against real OpenAI after feat: Update Ragas integration to work with >=0.4.3 haystack-core-integrations#3207 merged
• RagasEvaluator in pipeline returns valid scores for answer_relevancy, context_precision, faithfulness
• All 31 unit tests in ragas-haystack pass

[review-notebook-app]

Check out this pull request on 

See visual diffs & provide feedback on Jupyter Notebooks.

---

Powered by ReviewNB

[kacperlukawski]

Thank you @SyedShahmeerAli12 I had a look at the suggested changes, and started wondering whether the integration itself shouldn't be extended to reflect the latest changes. The previous approach was just a bit more elegant. Let me have a look at ragas-haystack first and then I'll come back to this PR.