GG Quant - Comprehensive Quantitative Trading Framework

GG Quant is a powerful, modular framework for quantitative trading that combines traditional financial analysis with cutting-edge AI technologies including Large Language Models (LLMs) and Reinforcement Learning (RL).

🚀 Features

📊 Data Management

• Multi-source Data: Support for Yahoo Finance, Tushare, and custom data sources
• SQLite Database: Efficient local storage and retrieval
• Technical Indicators: Built-in technical analysis (RSI, MACD, Bollinger Bands, etc.)
• Data Processing: Automated cleaning, validation, and feature engineering

🤖 AI-Powered Analysis

• Multiple LLM Providers: Ollama (local), OpenAI, HuggingFace
• Free Model Support: Cost-effective analysis with local models
• Sentiment Analysis: Market sentiment extraction from text
• AI-Generated Insights: Automated market analysis and recommendations

🧠 Reinforcement Learning

• Trading Environment: Gymnasium-compatible trading environments
• Multiple Agents: Q-Learning, Deep Q-Networks, and more
• Custom Strategies: Train RL agents for specific trading scenarios
• Performance Evaluation: Comprehensive agent testing and comparison

📈 Backtesting System

• Strategy Library: Pre-built trading strategies (MA crossover, RSI, etc.)
• Performance Metrics: Comprehensive risk and return analysis
• Strategy Comparison: Multi-strategy performance comparison
• Custom Strategies: Easy strategy development framework

🧪 Testing Framework

• Unit Tests: Comprehensive module testing
• Integration Tests: End-to-end workflow validation
• Performance Tests: System performance validation

📦 Installation

Quick Install

pip install gg-quant

Development Install

git clone https://github.com/ggquant/gg-quant.git
cd gg-quant
pip install -e ".[dev]"

Dependencies

GG Quant requires Python 3.8+ and the following main dependencies:

• pandas, numpy - Data analysis
• yfinance, tushare - Market data
• langchain, ollama - LLM integration
• gymnasium, stable-baselines3 - Reinforcement learning
• matplotlib, seaborn - Visualization

🎯 Quick Start

1. Data Management

from gg_quant import DataManager

# Initialize data manager
data_manager = DataManager("data/my_portfolio.db")

# Fetch and store data
symbols = ["AAPL", "GOOGL", "MSFT"]
results = data_manager.fetch_and_store(symbols, period="1y")

# Retrieve processed data
aapl_data = data_manager.get_data("AAPL", start_date="2023-01-01")
print(aapl_data.head())

2. AI-Powered Analysis

from gg_quant import LLMAnalyzer

# Initialize with local Ollama model
analyzer = LLMAnalyzer("ollama")

# Analyze sentiment
sentiment = analyzer.analyze_sentiment("Apple stock is performing well today")
print(f"Sentiment: {sentiment}")

# Generate market insights
market_analysis = analyzer.analyze_market("AAPL", aapl_data, {"rsi": 45.2})
print(market_analysis["analysis"])

# Get trading signal
signal = analyzer.generate_signal("AAPL", aapl_data, strategy="technical")
print(f"Signal: {signal['signal']} (Confidence: {signal['confidence']})")

3. Backtesting

from gg_quant import BacktestEngine
from gg_quant.backtest.strategies import MovingAverageStrategy

# Create strategy
strategy = MovingAverageStrategy(short_window=10, long_window=50)

# Run backtest
engine = BacktestEngine(initial_capital=10000)
result = engine.run_backtest(aapl_data, strategy)

# View results
print(f"Total Return: {result['metrics']['total_return']:.2%}")
print(f"Sharpe Ratio: {result['metrics']['sharpe_ratio']:.2f}")
print(f"Max Drawdown: {result['metrics']['max_drawdown']:.2%}")

4. Reinforcement Learning

from gg_quant import RLAgent

# Create RL agent
agent = RLAgent()

# Create trading environment
env = agent.create_environment(aapl_data, initial_balance=10000)

# Train agent
train_result = agent.train(total_timesteps=5000)
print(f"Training completed: {train_result}")

# Evaluate agent
eval_result = agent.evaluate(n_episodes=10)
print(f"Average reward: {eval_result['avg_reward']:.2f}")

📁 Project Structure

gg_quant/
├── data/                   # Data management
│   ├── database/           # SQLite database management
│   ├── fetchers/           # Data fetching from various sources
│   ├── processors/         # Data processing and technical analysis
│   └── models/            # Data models and structures
├── llm/                   # LLM integration
│   ├── clients/           # Various LLM client implementations
│   ├── analyzers/         # AI-powered analysis tools
│   └── models/            # LLM configurations and templates
├── rl/                    # Reinforcement learning
│   ├── environments/      # Trading environments
│   ├── agents/            # RL agent implementations
│   └── trainers/          # Training utilities
├── backtest/              # Backtesting system
│   ├── engine/            # Core backtesting engine
│   ├── metrics/           # Performance metrics
│   └── strategies/        # Trading strategies
├── test/                  # Testing framework
│   ├── unit/              # Unit tests
│   └── integration/       # Integration tests
└── utils/                 # Utilities and helpers

🔧 Configuration

Environment Variables

Create a .env file in your project root:

# Tushare API (optional)
TUSHARE_TOKEN=your_tushare_token

# OpenAI API (optional)
OPENAI_API_KEY=your_openai_key

# HuggingFace API (optional)
HUGGINGFACE_API_KEY=your_huggingface_key

# Ollama configuration
OLLAMA_BASE_URL=http://localhost:11434
OLLAMA_MODEL=gemma:2b

Custom Configuration

from gg_quant.llm.models import LLMConfig, ModelRegistry

# Add custom model configuration
custom_config = LLMConfig(
    name="custom_gpt",
    provider="openai",
    model_id="gpt-4",
    api_key="your_key",
    max_tokens=2000
)

registry = ModelRegistry()
registry.add_model(custom_config)

📊 Available Strategies

Technical Analysis Strategies

• Moving Average Crossover: Classic MA strategy
• RSI Mean Reversion: RSI-based overbought/oversold strategy
• Bollinger Bands: Volatility-based breakout strategy
• MACD Strategy: Trend following using MACD crossovers
• Buy and Hold: Simple benchmark strategy

AI-Enhanced Strategies

• Sentiment-Based: Trade based on news sentiment
• AI Signals: Use LLM-generated trading signals
• Hybrid: Combine technical and AI signals

🧪 Testing

Run the test suite:

# Run all tests
python -m pytest

# Run unit tests only
python -m pytest gg_quant/test/unit/

# Run integration tests only
python -m pytest gg_quant/test/integration/

# Run with coverage
python -m pytest --cov=gg_quant

📈 Examples

The examples/ directory contains comprehensive examples:

• basic_backtest.py - Simple backtesting example
• rl_training.py - Train an RL trading agent
• ai_analysis.py - Use LLM for market analysis
• data_pipeline.py - Complete data processing pipeline

🤝 Contributing

We welcome contributions! Please see our Contributing Guide for details.

Development Setup

git clone https://github.com/ggquant/gg-quant.git
cd gg-quant
pip install -e ".[dev]"
pre-commit install

Code Style

We use:

• Black for code formatting
• Flake8 for linting
• MyPy for type checking

# Format code
black gg_quant/ tests/

# Run linting
flake8 gg_quant/ tests/

# Type checking
mypy gg_quant/

📚 Documentation

Full documentation is available at gg-quant.readthedocs.io

API Documentation

• Data Module
• LLM Module
• RL Module
• Backtest Module

Tutorials

• Getting Started Guide
• Advanced RL Training
• Custom Strategy Development

🗺️ Roadmap

Version 1.1 (Planned)

• Web-based dashboard
• Real-time data streaming
• Options and derivatives support
• Portfolio optimization

Version 1.2 (Planned)

• Advanced neural network architectures
• Multi-asset portfolio strategies
• Risk management tools
• Cloud deployment options

📄 License

This project is licensed under the MIT License - see the LICENSE file for details.

🙏 Acknowledgments

• Gymnasium - RL environments
• LangChain - LLM integration
• Pandas - Data manipulation
• Stable-Baselines3 - RL algorithms
• yfinance - Market data

📞 Support

• Documentation: gg-quant.readthedocs.io
• Issues: GitHub Issues
• Discussions: GitHub Discussions

⭐ Star History

---

Disclaimer: This software is for educational and research purposes only. Past performance is not indicative of future results. Always conduct your own research before making investment decisions.