Code Summary: Set up the Environment

The code below sets up the basic environment needed to begin building an AI Agent that uses MongoDB.

Link to code on GitHub

Import Packages

Update the imported packages in the main.py file with the following:

import key_param
from pymongo import MongoClient
from langchain_openai import ChatOpenAI

Initialize the MongoDB Client

Connect to MongoDB Atlas deployment using the following:

def init_mongodb():
    """
    Initialize MongoDB client and collections.

    Returns:
        tuple: MongoDB client, vector search collection, full documents collection.
    """
    mongodb_client = MongoClient(key_param.mongodb_uri)
    
    DB_NAME = "ai_agents"
    
    vs_collection = mongodb_client[DB_NAME]["chunked_docs"]
    
    full_collection = mongodb_client[DB_NAME]["full_docs"]
    
    return mongodb_client, vs_collection, full_collection

Create the Entry Point for the Application

Create a main function which will be the entry point for the AI agent, calling the `init_mongodb` function to initialize the MongoDB connection and initializing the `llm` (ChatOpenAI model).

def main():
    """
    Main function to initialize and execute the graph.
    """
    # Initialize MongoDB connections
    mongodb_client, vs_collection, full_collection = init_mongodb()
    
    # Initialize the ChatOpenAI model with API key
    llm = ChatOpenAI(openai_api_key=key_param.openai_api_key, temperature=0, model="gpt-4o")
    
    

# Execute main function when script is run directly
main()