Skip to content

UU114/AI-QTRD

BranchesTags

Repository files navigation

AI-QTRD

AI-Powered Quantitative Trading & Research Development

AI驱动的量化交易与研究开发平台

AI-QTRD is an advanced quantitative trading platform that combines VeighNa framework with cutting-edge LLM technologies (DeepSeek) to enable natural language-driven strategy development and AI-enhanced trading decisions.

AI-QTRD 是一个先进的量化交易平台，结合了 VeighNa框架 与前沿的 大语言模型技术（DeepSeek），实现自然语言驱动的策略开发和AI增强的交易决策。

🚀 Core Features | 核心特性

1. AI-Enhanced Trading | AI增强交易

  • Natural Language Strategy | 自然语言策略编写

    • Describe trading strategies in plain language | 使用自然语言描述交易策略
    • DeepSeek LLM auto-converts to executable code | DeepSeek自动转换为可执行代码

  • Hybrid Decision System | 混合决策系统

    • Rule-based signals for deterministic logic | 规则信号处理确定性逻辑
    • Real-time LLM analysis for fuzzy judgments | 实时LLM分析处理模糊判断

  • News & Sentiment Analysis | 新闻与情绪分析

    • AI-powered market news interpretation | AI驱动的市场新闻解读
    • Real-time sentiment scoring | 实时情绪评分

2. Machine Learning Alpha | 机器学习Alpha

Built on VeighNa 4.0's vnpy.alpha module for professional ML-based quantitative research:

基于VeighNa 4.0的 vnpy.alpha 模块，提供专业的机器学习量化研究能力：

  • Factor Engineering | 因子工程

    • Polars-based high-performance feature computation | 基于Polars的高性能特征计算
    • Alpha101, Alpha158 factor sets | Alpha101、Alpha158因子集

  • ML Models | 机器学习模型

    • Lasso, LightGBM, MLP neural networks | Lasso、LightGBM、MLP神经网络
    • Unified API for model training and backtesting | 统一的模型训练和回测API

  • Complete Workflow | 完整工作流

    • Data management → Feature engineering → Model training → Signal generation → Backtesting
    • 数据管理 → 特征工程 → 模型训练 → 信号生成 → 策略回测

3. Event-Driven Architecture | 事件驱动架构

  • MainEngine orchestrates gateways, apps, and strategies | MainEngine协调网关、应用和策略
  • EventEngine enables asynchronous component communication | EventEngine实现异步组件通信
  • High-performance tick/bar data processing | 高性能tick/bar数据处理

4. Multi-Market Support | 多市场支持

Domestic Markets | 国内市场

  • Futures & Options | 期货期权: CTP, Mini CTP, SOPT, UFT, Femas, Esunny
  • A-Share | A股: XTP, TORA, OST, EMT
  • Gold TD | 黄金TD: KSGold, SGIT

International Markets | 海外市场

  • Interactive Brokers: Global stocks, futures, options | 全球股票、期货、期权
  • Overseas Futures | 海外期货: TAP, Direct Access

5. Comprehensive Trading Apps | 完善的交易应用

  • CTA Strategy | CTA策略: High-frequency CTA with fine-grained order control | 支持高频CTA策略
  • Portfolio Strategy | 组合策略: Multi-asset Alpha strategies | 多资产Alpha策略
  • Spread Trading | 价差交易: Custom spreads with algo execution | 自定义价差与算法执行
  • Option Master | 期权大师: Pricing models, Greeks tracking | 期权定价、希腊值跟踪
  • Algo Trading | 算法交易: TWAP, Iceberg, BestLimit algorithms | TWAP、冰山、最优限价算法
  • Risk Manager | 风险管理: Flow control, position limits | 交易流控、持仓限制

📦 Installation | 安装

Prerequisites | 环境要求

  • Python: 3.10+ (3.13 recommended) | Python 3.10以上（推荐3.13）
  • OS: Windows 11+ / Ubuntu 22.04+ / macOS | Windows 11以上 / Ubuntu 22.04以上 / macOS
  • DeepSeek API Key (for AI features) | DeepSeek API密钥（用于AI功能）

Quick Install | 快速安装

# Windows
install.bat

# Ubuntu / macOS
bash install.sh

# Development mode | 开发模式
pip install -e .

# With ML features | 安装ML功能依赖
pip install -e ".[alpha]"

🎯 Quick Start | 快速开始

1. AI-Enhanced Trading | AI增强交易

Natural Language Strategy Backtesting | 自然语言策略回测

cd vnpy_ai_trader

# Set DeepSeek API key | 配置DeepSeek密钥
export DEEPSEEK_API_KEY="your_api_key"

# Run backtest with natural language | 使用自然语言运行回测
python scripts/run_backtest.py \
    --strategy "沪深300成分股，RSI低于30时买入，盈利10%或亏损5%卖出" \
    --symbols "000001,000002,000063" \
    --start "2023-01-01" \
    --end "2024-01-01" \
    --capital 1000000

Web-Based Trading Interface | Web交易界面

cd vnpy_ai_trader

# Start web server | 启动Web服务器
python scripts/run_web_server.py

# Access UI | 访问界面
# http://localhost:8000
# API Docs: http://localhost:8000/docs

2. ML Alpha Research | ML Alpha研究

from vnpy.alpha.lab import AlphaLab
from vnpy.alpha.dataset.datasets.alpha_158 import Alpha158Dataset
from vnpy.alpha.model.models.lgb_model import LGBModel

# Initialize research lab | 初始化研究环境
lab = AlphaLab("my_alpha_project")

# Load Alpha158 features | 加载Alpha158因子
dataset = Alpha158Dataset()
lab.load_dataset(dataset)

# Train LightGBM model | 训练LightGBM模型
model = LGBModel()
lab.train_model(model)

# Generate signals and backtest | 生成信号并回测
lab.generate_signals()
lab.run_backtest()

3. Traditional VeighNa Trading | 传统VeighNa交易

from vnpy.event import EventEngine
from vnpy.trader.engine import MainEngine
from vnpy.trader.ui import MainWindow, create_qapp

from vnpy_ctp import CtpGateway
from vnpy_ctastrategy import CtaStrategyApp

def main():
    qapp = create_qapp()

    event_engine = EventEngine()
    main_engine = MainEngine(event_engine)

    # Add gateways and apps | 添加网关和应用
    main_engine.add_gateway(CtpGateway)
    main_engine.add_app(CtaStrategyApp)

    main_window = MainWindow(main_engine, event_engine)
    main_window.showMaximized()

    qapp.exec()

if __name__ == "__main__":
    main()

📁 Project Structure | 项目结构

AI-QTRD/
├── vnpy/                      # VeighNa core framework | 核心框架
│   ├── trader/                # Trading engine | 交易引擎
│   ├── alpha/                 # ML Alpha module | ML Alpha模块
│   ├── event/                 # Event engine | 事件引擎
│   ├── chart/                 # Chart visualization | 图表可视化
│   └── rpc/                   # RPC communication | RPC通信
│
├── vnpy_ai_trader/            # AI-enhanced subsystem | AI增强子系统
│   ├── src/
│   │   ├── ai_core/           # DeepSeek integration | DeepSeek集成
│   │   ├── strategy/          # AI strategies | AI策略
│   │   ├── datafeed/          # Free data sources | 免费数据源
│   │   ├── gateway/           # Trading gateways | 交易网关
│   │   └── web/               # Web interface | Web界面
│   ├── scripts/               # Execution scripts | 执行脚本
│   └── config/                # Configuration files | 配置文件
│
└── examples/
    └── alpha_research/        # ML research notebooks | ML研究示例

🔧 Code Quality | 代码质量

# Lint with ruff | 使用ruff检查代码
ruff check .

# Type check with mypy | 使用mypy类型检查
mypy vnpy

# Run tests | 运行测试
pytest tests/

📚 Documentation | 文档

  • VeighNa Documentation | VeighNa官方文档: www.vnpy.com/docs
  • AI Trader Guide | AI交易指南: See vnpy_ai_trader/doc/ | 见 vnpy_ai_trader/doc/
  • API Reference | API参考: Auto-generated at /docs endpoint | Web服务的 /docs 端点自动生成

🎓 Key Dependencies | 核心依赖

  • PySide6: GUI framework | GUI框架
  • Polars: High-performance DataFrames | 高性能数据框架
  • NumPy / Pandas: Data processing | 数据处理
  • TA-Lib: Technical indicators | 技术指标
  • LightGBM / scikit-learn: ML models | 机器学习模型
  • FastAPI: Web API framework | Web API框架
  • OpenAI SDK: DeepSeek client | DeepSeek客户端

🤝 Contributing | 贡献

We welcome contributions! Please follow the standard GitHub workflow:

欢迎贡献代码！请遵循标准的GitHub工作流程：

  1. Fork the repository | Fork本仓库
  2. Create a feature branch | 创建功能分支
  3. Commit changes with clear messages | 提交明确的更改说明
  4. Ensure code passes ruff check . and mypy vnpy | 确保代码通过代码检查
  5. Submit a Pull Request | 提交Pull Request

📖 Credits | 致谢

AI-QTRD is built upon the excellent work of:

AI-QTRD 基于以下优秀项目构建：

  • VeighNa: Core quantitative trading framework | 核心量化交易框架
  • Qlib: Inspiration for vnpy.alpha design | vnpy.alpha设计灵感来源
  • DeepSeek: LLM inference engine | 大语言模型推理引擎
  • AData: Free A-share data source | 免费A股数据源

📄 License | 许可证

MIT License

🔗 Links | 链接

AI-QTRD: Where Traditional Quantitative Trading Meets Modern AI
AI-QTRD: 传统量化交易与现代AI的完美结合

About

No description, website, or topics provided.

Resources

Readme

License

MIT license

Code of conduct

Code of conduct
Activity

Stars

2 stars

Watchers

0 watching

Forks

0 forks
Report repository

Releases

No releases published

Packages

 
 
 

Contributors

Languages