This project implements an AI-powered query agent that generates, executes, and interprets result using LangGraph, MongoDB and Docker. It is designed to answer natural language questions about a MongoDB database, process results, and provide concise answer.

• Automatically generates MongoDB pipelines from plain English questions.
• Executes queries on MongoDB and retrieves results.
• Formats query results into concise, human-readable answers.
• Fully modular and extensible with LangGraph.

├── agent_with_mongodb.ipynb # AI agent logic and architecture
├── insert_data_mongo.py # Data entry in MongoDB
├── docker-compose-infra.yml # Docker setup for MongoDB
├── README.md # Project documentation

1. State-based Agent:

   • generate_pipeline: Generates MongoDB aggregation pipeline from natural language.
   • execute_query: Executes the pipeline against MongoDB and fetches results.
   • format_answer: Formats the results into a concise, human-readable answer.

2. Pipeline Parsing:

   • Handles nested dictionaries and lists in aggregation pipelines.

3. LLM Integration:

   • Uses language model by Groq LLaMA3-70B and LangGraph.
   • Generate valid MongoDB pipelines.
   • Format query results into readable answers.

1. Configure Docker.
2. Configure MongoDB.
3. Insert data in MongoDB database.
4. Run agent code to retrieve data from database.