Elsai Retrievers

The Elsai Retrievers package provides an interface for performing hybrid document retrieval by combining semantic and sparse (BM25) search strategies. This enables more accurate and flexible information retrieval across multiple sources.

Prerequisites

• Python >= 3.9

• .env file with appropriate API keys and configuration variables

Installation

To install the elsai-retrievers package:

pip install --index-url https://elsai-core-package.optisolbusiness.com/root/elsai-retrievers/ elsai-retrievers==0.1.0

Component

1. HybridRetriever

HybridRetriever is a class that enables combining results from multiple retrievers — including dense vector-based (semantic) and sparse (BM25) — into a single ranked list.

from elsai_retrievers.hybrid_retriever import HybridRetriever

# Create the hybrid retriever instance
hybrid_retriever = HybridRetriever()

# Perform hybrid retrieval
results = hybrid_retriever.hybrid_retrieve(
    chunks=chunks,  # Optional: for BM25-based retrieval
    retrievers=["retriever1", "retriever2"],  # Semantic retrievers that support `.retrieve()`
    question="How does hybrid retrieval work?"
)