Integrate Coval for RAG Observability

Integrate Coval into your LlamaIndex RAG pipeline to monitor and evaluate its performance. Specifically, set up Coval to track query execution, retrieved chunks, and the final LLM response. Explain how you would use Coval's metrics (e.g., context precision, context recall, answer relevance) to identify areas for improvement in your RAG system. Provide pseudo-code or conceptual steps for integrating Coval's logging and evaluation hooks within your LlamaIndex `QueryEngine`.