Text to SQL with Deepseek

A web-based application that converts natural language questions into SQL queries and executes them against SQLite databases.

Features

• Natural language to SQL query conversion
• SQLite database connection and schema analysis
• Query execution with result display
• Conversation history tracking
• Sample database creation for testing
• Direct SQL query execution interface

Requirements

Core Dependencies:

• 
• 
• 
• 
• 

Installation

1. Clone the repository
2. Install dependencies:

pip install pandas langchain-core langchain-ollama gradio

3. Ensure Ollama is installed and running with the deepseek-coder model:

ollama pull deepseek-coder:6.7b

Usage

Starting the Application

python app.py

The application will launch in your web browser at http://localhost:7860

Database Connection

1. Upload a SQLite database file (.db, .sqlite, .sqlite3)
2. Or create a sample e-commerce database for testing

Querying Data

Submit questions in natural language through the Query Interface:

• "Show all customers from USA"
• "List top 5 products by sales"
• "Find orders over $500"

Direct SQL Execution

Use the SQL Executor tab to run custom SQL queries directly against the connected database.

Configuration

The application uses the deepseek-coder:6.7b model by default. To change the model:

app = TextToSQLApp(model_name="your-preferred-model")

Sample Database Schema

The application can generate a sample e-commerce database with the following tables:

• customers - Customer information
• products - Product catalog
• orders - Order records
• order_items - Order line items

Limitations

• Supports SQLite databases only
• Requires Ollama runtime with compatible language models
• Query complexity depends on model capabilities
• No authentication or user management

Architecture

• Backend: with for LLM integration
• Database: with for data handling
• Frontend: web interface
• LLM: compatible models for SQL generation

Error Handling

The application includes error handling for:

• Database connection failures
• Invalid SQL syntax
• Missing database schema
• Model communication errors