Example: Contextual AI Reranker Integration with DSPy

[NitishGourishetty]

I'm was interested in adding Contextual AI as a retriever integration within DSPy.

This PR includes an example notebook (docs/docs/tutorials/contextual-reranker/contextual_reranker_tutorial.ipynb) that integrates Contextual AI's instruction-tuned reranker (ctxl-rerank-v2-instruct-multilingual) into DSPy as a post-retrieval reranking step.

The notebook shows:

• How to subclass dspy.Retrieve for external API rerankers
• An end-to-end RAG pipeline on the RAG-QA Tech dataset
• Evaluation using DSPy's SemanticF1 metric

This is just a demo / POC and I would love any suggestions on how to align this more closely with DSPy's retriever conventions and how I can get it into the /retrievers or maybe a new rerankers folder.

Any guidance on next steps (like whether to keep it as an example or formalize it under dspy/retrievers/) would be much appreciated! Thanks!