Skip to content

Releases: TPTBusiness/Predix

v1.0.0 - Initial Release: EURUSD Trading Agent

02 Apr 18:44
@TPTBusiness TPTBusiness
v1.0.0
a3fa885

Choose a tag to compare

View all tags
v1.0.0 - Initial Release: EURUSD Trading Agent Latest
Latest

🎉 Predix v1.0.0 - Initial Release

📊 What is Predix?

Predix is an autonomous AI-powered quantitative trading agent for EUR/USD forex markets. It automates the full research and development cycle for trading strategies:

  • 📊 Data Analysis – Automatically analyzes market patterns and microstructure
  • 💡 Strategy Discovery – Proposes novel trading factors and signals
  • 🧠 Model Evolution – Iteratively improves predictive models
  • 📈 Backtesting – Validates strategies on historical 1-minute data

✨ Key Features

Autonomous Factor Generation

  • 110+ EURUSD factors generated autonomously using LLMs
  • Multi-agent debate system (Bull/Bear/Neutral analysts)
  • Stanley Druckenmiller-style macro analysis
  • Market regime detection using Hurst Exponent

Comprehensive Backtesting

  • Backtesting engine with IC, Sharpe Ratio, Max Drawdown metrics
  • SQLite database for tracking all backtest results
  • Correlation analysis and portfolio optimization
  • Risk management with position limits and leverage controls

Professional Testing

  • 97 unit tests with 98.77% code coverage
  • Edge case testing for all metrics
  • Integration tests for full workflows

Dashboards & UI

  • Web Dashboard (Flask + HTML) - Live trading progress
  • CLI Dashboard (Rich library) - Terminal-based UI
  • Real-time macro data (EURUSD, DXY, Volatility)
  • Session-aware analysis (Asian/London/NY sessions)

Developer Experience

  • All code comments in English
  • Comprehensive documentation (QWEN.md, README.md)
  • Clean git history with English commit messages
  • Proper attribution guidelines (ATTRIBUTION.md)

🏗️ Technical Stack

  • Python 3.10/3.11 - Primary language
  • PyTorch - Deep learning models
  • Qlib - Backtesting engine
  • LLM (Qwen3.5-35B) - Factor generation via local llama.cpp
  • Flask - Web dashboard API
  • SQLite - Results database
  • Rich/Typer - CLI interface
  • pytest - Testing framework (98.77% coverage)

📁 Project Structure

Predix/
├── rdagent/                    # Core agent framework
│   ├── components/
│   │   ├── backtesting/        # Backtest engine, metrics, database
│   │   └── coder/
│   │       └── factor_coder/   # Factor generation & EURUSD modules
│   └── scenarios/
│       └── qlib/               # Qlib integration for FX trading
├── test/                       # Unit tests (97 tests, 98.77% coverage)
├── web/                        # Dashboard frontend
├── results/                    # Backtest results (not in git)
├── docs/                       # Documentation
└── requirements.txt            # Dependencies

🚀 Quick Start

Installation

# Clone repository
git clone https://github.com/TPTBusiness/Predix.git
cd Predix

# Create conda environment
conda create -n predix python=3.10
conda activate predix

# Install in editable mode
pip install -e .[test,lint]

Configuration

# Create .env file
cat << EOF > .env
CHAT_MODEL=qwen3.5-35b
OPENAI_API_BASE=http://localhost:8081/v1
OPENAI_API_KEY=local
EMBEDDING_MODEL=nomic-embed-text
QLIB_DATA_DIR=~/.qlib/qlib_data/eurusd_1min_data
EOF

# Start LLM server (llama.cpp)
~/llama.cpp/build/bin/llama-server \
  --model ~/models/qwen3.5/Qwen3.5-35B-A3B-Q3_K_M.gguf \
  --n-gpu-layers 36 \
  --ctx-size 80000 \
  --port 8081

Run Trading Loop

# Start trading loop (24/7)
./start_loop.sh

# Or single run with dashboard
python predix.py fin_quant --with-dashboard

📊 Performance Metrics

Metric Target Status
Test Coverage >80% ✅ 98.77%
Tests Passed All ✅ 97/97
Factors Generated 100+ ✅ 110+
Code Quality Clean ✅ All English
Documentation Complete ✅ QWEN.md, README

📝 License & Attribution

License: MIT License

Attribution Requirements:

If you use this code or concepts in your project, you must:

  1. Include the MIT License text
  2. Keep the copyright notice: "Copyright (c) 2025 Predix Team"
  3. Provide attribution to the original project

See ATTRIBUTION.md for detailed guidelines and examples.

Legal Basis: MIT License requires:

"The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software."

🙏 Acknowledgments

This project draws inspiration from various open-source projects:

  • Microsoft RD-Agent (MIT License) - Foundation for our autonomous R&D agent framework
  • TradingAgents (Apache 2.0 License) - Inspiration for multi-agent debate system
  • ai-hedge-fund - Inspiration for macro analysis and risk management concepts

All code in Predix is originally written and implemented independently.

📦 Installation Options

Full Installation (with dev dependencies)

pip install -e .[test,lint,docs]

Minimal Installation

pip install -e .

Docker (Coming Soon)

docker pull predixai/predix:latest
docker run -it predixai/predix

🧪 Testing

# Run all tests
pytest test/ -v

# Run with coverage
pytest test/ --cov=rdagent --cov-report=html

# Run specific test suite
pytest test/backtesting/ -v

📚 Documentation

  • README.md - Installation and quick start
  • QWEN.md - Comprehensive development guide
  • ATTRIBUTION.md - Usage guidelines and attribution requirements
  • docs/ - Additional documentation

🔧 Development

# Linting
ruff check rdagent/

# Type checking
mypy rdagent/

# Format code
black rdagent/

# Run tests
pytest test/backtesting/ -v

📞 Support

⚠️ Disclaimer

This project is for educational and research purposes only.

  • Not intended for real trading or investment
  • No investment advice or guarantees provided
  • Creator assumes no liability for financial losses
  • Consult a financial advisor for investment decisions
  • Past performance does not indicate future results

By using this software, you agree to use it solely for learning purposes.

🎯 What's Next?

Planned for v1.1.0:

  • Live paper-trading integration
  • Additional factor types (ML-based)
  • Performance optimization
  • Docker containerization
  • CI/CD pipeline

📈 Statistics

  • Lines of Code: ~15,000+
  • Files: 100+
  • Commits: 20+
  • Contributors: 1
  • Stars: ⭐ (Be the first to star!)

Full changelog: CHANGELOG.md for v1.0.0

If you like this project, please ⭐ star this repository!

Assets 2
Loading