Summary

This PR adds a practical troubleshooting page for users who build retrieval, reranking, and RAG pipelines with Rankify.

The new page focuses on common failure patterns (for example “good retrieval, bad reranking”) and gives concrete checklists that help users quickly localize the most likely source of error.

Changes

• Added docs/docs/troubleshooting_rag_and_reranking.md
• Linked the new page in docs/mkdocs.yml as “Troubleshooting RAG & Reranking”
• No changes to the core library or API

Motivation

From my own experience and from issues I often see in RAG projects, many problems are not caused by a single bug in the code, but by small mismatches between:

• retriever configuration
• reranker configuration
• evaluation setup
• data preparation

This page tries to give Rankify users a structured way to:

1. Match their symptoms to a failure pattern
2. Walk through a checklist for that pattern
3. Use the results to open a more precise GitHub issue or discussion if needed

I am very happy to adjust the wording, structure, or title if you prefer a different style. Feedback is welcome.