README.md

metadata

title: DRL Trading Bot
emoji: 🤖
colorFrom: blue
colorTo: green
sdk: streamlit
sdk_version: 1.42.0
app_file: src/ui/app.py
pinned: false

DRL Trading System

An autonomous Deep Reinforcement Learning trading system using PPO-LSTM for Binance Testnet trading with real-time Streamlit monitoring.

Features

• 🧠 PPO-LSTM Agent: Captures time-series dependencies for smarter trading decisions
• 🔄 Self-Improvement Loop: Automatically fine-tunes on successful trades every 24 hours
• 📊 Real-time Dashboard: TradingView-style charts with live Buy/Sell signals
• 🛡️ Risk Management: Circuit breaker stops trading at 5% daily loss
• 📈 Backtesting: Validated on 2024-2025 historical data

Quick Start

# 1. Create virtual environment
python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate

# 2. Install dependencies
pip install -r requirements.txt

# 3. Configure API keys
cp .env.example .env
# Edit .env with your Binance Testnet credentials

# 4. Run backtest first
python -m src.backtest.engine

# 5. Launch UI
streamlit run src/ui/app.py

Project Structure

drl-trading-system/
├── config/
│   └── config.yaml          # All configuration
├── src/
│   ├── env/                  # Gymnasium environment
│   ├── brain/                # PPO-LSTM agent
│   ├── api/                  # Binance connector
│   ├── backtest/             # Backtesting engine
│   └── ui/                   # Streamlit dashboard
├── data/
│   ├── historical/           # Cached OHLCV data
│   └── models/               # Saved checkpoints
└── tests/

Architecture

The system uses a Sharpe/Sortino ratio reward function to prioritize risk-adjusted returns over raw profit.

License

MIT