🎬 AI-Powered SQL Database Assistant

🚀 Project Overview

This application bridges the gap between non-technical business stakeholders and complex SQL databases. By leveraging Generative AI (OpenAI GPT-3.5) and LangChain, it allows users to ask questions in plain English (e.g., "Which movie category generated the highest revenue?") and receive accurate, data-backed answers immediately.

The system automatically translates natural language into SQL queries, executes them against a live PostgreSQL database (DVD Rental Schema), and presents the results in an interactive dashboard.

📸 Application Demo

(The AI Agent successfully navigating a complex Join across 5 tables: Payment → Rental → Inventory → Film → Category)

---

💼 Business Value & Problem Solved

The Problem: Business analysts and managers often need quick insights from databases but lack the SQL expertise to query complex schemas. This leads to a dependency on data teams for basic reports, creating bottlenecks.

The Solution: This Self-Service AI Tool empowers stakeholders to:

• Reduce Ad-Hoc Requests: Answer 80% of routine data questions instantly without engineering support.
• Accelerate Decision Making: Cut down "time-to-insight" from hours to seconds.
• Democratize Data: Make the 15-table DVD Rental dataset accessible to anyone on the team.

---

🛠️ Technical Architecture

The application uses a Retrieve-Then-Generate architecture:

1. User Input: Natural language query via Streamlit.
2. Schema Retrieval: LangChain inspects the PostgreSQL table definitions.
3. SQL Generation: OpenAI GPT-3.5 translates the prompt + schema into a valid SQL query.
4. Execution: The query is run against the local PostgreSQL database using psycopg2.
5. Response: The raw data is formatted into a natural language summary and displayed.

Key Features:

• Complex Query Handling: Capable of generating JOINs, GROUP BY, and aggregations.
• Error Correction: If the LLM generates invalid SQL, the agent catches the database error, feeds it back to the LLM, and retries automatically.
• Secure Connection: Uses environment variables to manage database credentials and API keys securely.

---

💻 How to Run Locally

1. Clone the Repository

git clone [https://github.com/devanshi4/sql-ai-assistant.git](https://github.com/devanshi4/sql-ai-assistant.git)
cd sql-ai-assistant

2 Create Virtual Environment

Bash
python -m venv venv
source venv/bin/activate   # On Windows: venv\Scripts\activate

3. Install Dependencies

Bash
pip install -r requirements.txt

4. Set Up Database Ensure you have a local PostgreSQL instance running with the DVD Rental sample database restored.

5. Run the Application

Bash
streamlit run app.py

📂 Project Structure

Plaintext

├── app.py                  # Main application logic (Streamlit + LangChain)
├── requirements.txt        # List of Python dependencies
├── .gitignore              # Files excluded from Git (passwords, venv)
├── README.md               # Project documentation
└── demo.png                # Screenshot for documentation

🔮 Future Improvements

* Visualization: Add automatic charting (Bar/Line graphs) for numerical results.

* Containerization: Dockerize the application for easy cloud deployment (AWS EC2).

* Open Source LLM: Experiment with Llama-2 or Mistral to run entirely offline.