🤖 Natural Language to SQL Query Generator

A Streamlit app that converts English questions into SQL queries using AI models. Simply ask questions about your data and get instant results!

🚀 Quick Setup

1. Clone and navigate to the project:

git clone [email protected]:mdarshad1000/Text-2-SQL.git
cd [email protected]:mdarshad1000/Text-2-SQL.git

2. Create .env file with your API keys:

OPENAI_API_KEY=your_key_here
GEMINI_API_KEY=your_key_here

3. Start the app:

docker compose up -d

Visit http://0.0.0.0:8501 in your browser.

💡 How to Use

1. Select an AI model (GPT-4, Llama3.2:latest, or Gemini) {{ If you want to use the Llama3.2 model, ensure the Ollama server is running. }}
2. Type your question (e.g., "Show me all sales from last month")
3. Click "Generate Results"
4. View your data in tables or charts
5. Download results as CSV if needed

🏗️ Code Structure & Documentation

Folder Structure

`Chinook_PostgreSql.sql`: SQL script to create database schema
`Dockerfile`: Docker file to build Docker image
`README.md`: This README file
`app.py`: Main application file
`config.py`: Configuration file for LLM models
`docker-compose.yml`: Docker Compose file to run Docker image
`llm_integration.py`: Abstract base class for LLM integrations
`query_executor.py`: Database executor class
`requirements.txt`: Python dependencies

Classes

LLMQueryGenerator (Abstract Base Class)

Base class for all AI model integrations.

• generate_sql_from_nl(db_schema, nl_query): Abstract method to convert natural language to SQL

OpenAIQueryGenerator

Handles GPT-4 integration.

• __init__(model, sys_prompt, user_prompt): Initializes with model settings
• generate_sql_from_nl(db_schema, nl_query): Generates SQL using OpenAI API

OllamaQueryGenerator

Manages Llama3.2:latest integration.

• __init__(model, sys_prompt, user_prompt): Sets up Ollama configuration
• generate_sql_from_nl(db_schema, nl_query): Generates SQL using Ollama

GeminiQueryGenerator

Handles Google's Gemini integration.

• __init__(model, sys_prompt, user_prompt): Configures Gemini settings
• generate_sql_from_nl(db_schema, nl_query): Generates SQL using Gemini API

DatabaseExecutor

Manages database operations.

• connect(): Establishes database connection
• disconnect(): Closes database connection
• execute_query(query): Runs SQL queries and returns results as DataFrame
• get_schema_info(): Returns database structure information

Workflow

1. Initialization

   • App starts and creates DatabaseExecutor instance
   • Initializes AI models (GPT-4, Llama3.2, Gemini)

2. User Interaction

   • User selects AI model
   • Views database schema
   • Enters natural language question

3. Query Processing

   User Question → AI Model → SQL Query → Database → Results → Visualization
   

4. Data Display

   • Results shown in table format
   • Optional visualization in charts
   • CSV export available

🛠️ Troubleshooting

If something's not working:

1. Check container status: docker compose ps
2. View logs: docker compose logs app
3. Restart if needed: docker compose restart

📝 Notes

• Uses Chinook sample database (digital media store data)
• Supports basic data visualization
• Downloads available in CSV format