What is the primary function of the `retriever` class in a hybrid search pipeline?	To wrap multiple search implementations (like vector index and BM25) and merge their results into a single output.
What prerequisite must the different search implementations (e.g., Vector Index and BM25 Index) have to be easily combined into a retriever?	They must share an almost identical public API, including methods like `add document` and `search`.
What specific technique is used to combine and merge results from different search methodologies?	Reciprocal Rank Fusion (RRF).
How does Reciprocal Rank Fusion (RRF) determine the relevance of a single text chunk?	It calculates a score for the chunk based on its rank across all input search lists.
What is the general formula used in Reciprocal Rank Fusion for a given rank?	The term is calculated as 1 / (1 + rank value).
What is the final step after calculating scores using Reciprocal Rank Fusion?	The combined table of results is sorted based on the calculated score from greatest to least.
What is the main benefit of designing individual search indexes with a standardized API?	It allows new, completely different search functionalities to be easily added to the retriever class without modifying existing code.
What is the overall goal of implementing a hybrid search approach (combining vector and lexical search)?	To improve the accuracy of the Retrieval-Augmented Generation (RAG) pipeline.