🔮 InsightAI: SQL Chat Assistant
Note
Talk to your data. A natural language interface that executes safe SQL queries, generates visualizations, and forecasts future trends using Google Gemini Flash.
🚀 Features
Natural Language to SQL: Converts plain English (e.g., "Show sales for 2023") into executable SQL.
Live Execution: Connects to a real SQLite database to fetch live results (not just text generation).
Auto-Visualization: Intelligently chooses between Data Tables, Bar Charts, or Line Charts based on data shape.
🔮 Future Forecasting: Uses Linear/Polynomial Regression (Scikit-Learn) to predict future trends based on historical data.
🛡️ Enterprise Safety: Custom middleware blocks destructive commands (DROP, DELETE) to prevent data loss.
🛠️ Tech Stack
Backend: Python, Flask, Scikit-Learn (Forecasting)
Frontend: React (Single File Architecture), Tailwind CSS, Recharts
Database: SQLite
AI Engine: Google Gemini 1.5 Flash / 2.0 Flash (via google-genai SDK)
⚡ Quick Start Guide
Follow these steps to get the project running locally in 5 minutes.
- Clone the Repository
git clone <your-repo-url>
cd sql-chat-assistant
- Install Dependencies
Make sure you have Python installed. Then run:
pip install -r requirements.txt
- Setup Environment Variables
Important
You must have a valid API Key for the application to work.
Get a free key from Google AI Studio.
Create a .env file in the root directory.
Paste the key as shown below:
GEMINI_API_KEY=AIzaSy...YourKeyHere
- Initialize Database
Create the SQLite database and seed it with dummy data:
python backend/init_db.py
- Run the Server
Start the Flask backend (which also serves the Frontend):
python backend/app.py
- Open the App
Go to your browser and visit:
http://127.0.0.1:5000
🧪 Demo Scenarios
Tip
Use this script during your hackathon demo to showcase all features:
Simple Query: "Show me all products."
Aggregation: "What is the total revenue by country?" (Triggers Bar Chart)
Time-Series: "Show orders for 2023 ordered by date."
Forecasting: Click the "Predict Future" button in the top right.
Caution
Safety Demo: Try asking "DELETE all customers." The middleware will block this request to demonstrate enterprise security features.
📂 Project Structure
sql-chat-assistant/
├── backend/
│ ├── app.py # Main Flask Server & API Routes
│ ├── safety.py # SQL Sanitization Middleware
│ ├── forecast.py # ML Forecasting Logic
│ └── init_db.py # Database Seeding Script
├── data/
│ ├── hackathon.db # SQLite Database (Created on init)
│ └── seed_data.sql # Raw SQL data for testing
├── frontend/
│ └── index.html # Single-file React Application
├── prompts/
│ ├── system_prompt.md # AI Persona & Rules
│ └── schema.txt # Database Schema Context
└── README.md
📜 License
MIT License