retriever.py

from langchain_community.retrievers import BM25Retriever
from langchain.docstore.document import Document
from langchain.tools import StructuredTool
from typing import List
import datasets

# Internal variable to cache retriever
_guest_bm25_retriever = None

def load_guest_dataset():
    global _guest_bm25_retriever

    if _guest_bm25_retriever is not None:
        return _guest_bm25_retriever

    # Load the dataset
    guest_dataset = datasets.load_dataset("agents-course/unit3-invitees", split="train")

    # Convert dataset entries into Document objects
    docs = [
        Document(
            page_content="\n".join([
                f"Name: {guest['name']}",
                f"Relation: {guest['relation']}",
                f"Description: {guest['description']}",
                f"Email: {guest['email']}"
            ]),
            metadata={"name": guest["name"]}
        )
        for guest in guest_dataset
    ]

    # Create and cache the retriever
    _guest_bm25_retriever = BM25Retriever.from_documents(docs)
    return _guest_bm25_retriever

def retrieve_guest_info(query: str) -> str:
    retriever = load_guest_dataset()
    results: List[Document] = retriever.get_relevant_documents(query)

    if results:
        return "\n\n".join([doc.page_content for doc in results[:3]])
    else:
        return "No matching guest information found."

GuestInfoTool = StructuredTool.from_function(
    name="guest_info",
    description="Retrieves detailed information about gala guests based on their name or relation.",
    func=retrieve_guest_info
)