Fix RRF algorithm to match AI Search algo

[pamelafox]

This algorithm matches what AI Search describes. I stepped through the results with the debugger and they seemed reasonable.

[Copilot]

Pull Request Overview

This PR updates the reciprocal rank fusion implementation to align with the AI Search algorithm. The changes include:

• Replacing the previous weighted fusion (using alpha) with an RRF scoring approach using a k offset.
• Updating the hybrid_search function to use the new fusion mechanism.
• Modifying sample user query strings in both the Spanish and English versions.

Reviewed Changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated no comments.

File	Description
spanish/rag_documents_hybrid.py	Updates the reciprocal_rank_fusion function and sample query string.
rag_documents_hybrid.py	Updates the reciprocal_rank_fusion function and sample query string.

Comments suppressed due to low confidence (2)

spanish/rag_documents_hybrid.py:85

• [nitpick] Consider using (i+1+k) instead of (i+k) to properly reflect 1-indexed ranks for RRF scoring, ensuring no off-by-one error.

scores[doc["id"]] += 1 / (i + k)

rag_documents_hybrid.py:85

• [nitpick] Consider using (i+1+k) instead of (i+k) so that the ranking starts at 1 as expected by standard RRF implementations, which may avoid off-by-one issues.

scores[doc["id"]] += 1 / (i + k)