retrieval.py

from langchain_community.retrievers import BM25Retriever
from langchain_classic.retrievers import EnsembleRetriever
from chunking import semantic_chunks
from vectordb import vector_store

dense_retriever = vector_store.as_retriever(search_type= "mmr", search_kwargs = {'k': 3, "lambda_mult": 0.7, "fetch_k": 10})

sparse_retriever = BM25Retriever.from_documents(semantic_chunks)
sparse_retriever.k = 3

retriever = EnsembleRetriever(
    retrievers=[dense_retriever, sparse_retriever],
    weights=[0.7,0.3]
)

if __name__=="__main__":
    query = "what is a covering index ?"
    results = retriever.invoke(query)
    
    print(f"\nFound {len(results)} results for: '{query}'\n")
    for i,doc in enumerate(results):
        print(f"------Reults{i+1}------")
        print(f"Source: {doc.metadata.get('source')}")
        print(doc.page_content)
        print()