mirror of
https://github.com/triqbit/mt5-ai-xauusd-trader.git
synced 2026-11-02 06:56:15 +00:00
andonly1348GitHubgoogle-labs-jules[bot] <161369871+google-labs-jules[bot]@users.noreply.github.com>andonly1348This change defines the formal operational data retention policy and provides an automated cleanup script to enforce it. Key improvements: - Created docs/DATA_RETENTION_POLICY.md defining retention windows: - Trade records, Linked Signals, Audit Logs: 7 years (Compliance) - Risk Events, Performance Metrics: 2 years (Audit) - Unlinked Signals, App Logs: 90 days (Operational) - Backtest Results: 1 year (Research) - Enhanced scripts/data_cleanup.py with: - Support for audit_log purging. - Multi-database URL support (handling separate audit DBs). - Safe deletion logic for linked records. - Mandatory dry-run and safety warnings for production DBs. - Improved tests/test_data_cleanup.py with SQLAlchemy StaticPool for reliable in-memory SQLite testing. Signed-off-by: Jules03 (andonly1348) Co-authored-by: google-labs-jules[bot] <161369871+google-labs-jules[bot]@users.noreply.github.com> Co-authored-by: andonly1348 <279141437+andonly1348@users.noreply.github.com>
🤖 MT5 AI/ML Trading Bot - Enterprise Edition
Institutional-Grade Algorithmic Trading System for MetaTrader 5
🏛️ Executive Summary
The MT5 AI/ML Trading Bot is a production-ready, enterprise-grade automated trading engine specifically optimized for the XAUUSD (Gold) market. It leverages cutting-edge Deep Reinforcement Learning (DRL) and Ensemble Machine Learning to deliver consistent, risk-adjusted returns within the MetaTrader 5 ecosystem.
Built on an architectural foundation that integrates 25+ top-tier quantitative finance repositories, this system provides a unified interface for model training, backtesting, and live execution.
🏛️ Technical Credibility & Trust
The MT5 AI/ML Trading Bot is built for institutional-grade reliability. We prioritize transparency, evidence-based development, and clear system boundaries.
- Architecture Quick-Start: 5-minute overview of system components, data flow, and maturity levels.
- System Maturity Map: Transparent status of production vs. experimental subsystems.
- Project Health Dashboard: Real-time visibility into CI status, stability KPIs, and strategic risks.
- Audit Evidence: Direct routing to ADRs, performance reports, and security scorecards.
🚀 Core Features
🧠 Advanced Intelligence
- DRL Architectures: PPO (Proximal Policy Optimization), Dreamer V3, and LSTM-based actors.
- Ensemble Engine: Real-time signal consensus from multiple neural networks to minimize variance.
- Explainability System: Structured attribution breakdowns for every trade signal, providing institutional-grade transparency into model and risk decisions.
- Dynamic Feature Engineering: 140+ market indicators including multi-timeframe TA-Lib features and macro-sentiment integration.
🛡️ Institutional Risk Management
- Ray Dalio All-Weather Allocation: Scenario-based risk parity across multi-currency pairs.
- 6-Layer Execution Filter: Cascade validation using ATR, Trend Angle, Momentum, and EMA sequencing.
- Circuit Breakers: Automated drawdown protection, per-session loss limits, and daily profit targets.
⚡ Production Infrastructure
- CI/CD Pipeline: Fully automated GitHub Actions for linting, security audits (
pip-audit), and unit testing. - Dockerized Deployment: Multi-stage builds for lightweight, cross-platform cloud deployment.
- Hybrid Connector: Native MT5 SDK support with MetaAPI cloud failover.
📊 Performance Benchmark
| Metric | Target Value | Verification Method |
|---|---|---|
| Annualized Return | 60% - 90% | Walk-forward Backtest (10Y) |
| Sharpe Ratio | 2.8 - 3.5 | Risk-Adjusted Return Analysis |
| Max Drawdown | < 12% | Dynamic Equity Protection |
| Profit Factor | 2.5+ | Gross Profit / Gross Loss |
🛠️ Technology Stack
- Frameworks: PyTorch, Stable-Baselines3, Gymnasium
- Data: Pandas, NumPy, TA-Lib
- DevOps: Docker, GitHub Actions, Ruff
- Settings: Pydantic Settings V2
📦 Project Structure
mt5-ai-xauusd-trader/
├── .github/workflows/ # Automated CI/CD (Quality, Security, Tests)
├── src/ # Core Package Content
│ ├── core/ # Environment-driven Configuration (Pydantic)
│ ├── models/ # AI/ML Architectures (Ensemble, LSTM, DRL)
│ └── trading/ # MT5 SDK Connectors & Risk Engines
├── tests/ # Comprehensive Unit & Integration Suite
├── main.py # Unified CLI Entrypoint
├── Dockerfile # Multi-stage Production Build
└── requirements-ci.txt # Pinned, CVE-free Dependencies
🏁 Quick Start
1. Installation
git clone https://github.com/triqbit/mt5-ai-xauusd-trader.git
cd mt5-ai-xauusd-trader
pip install -r requirements.txt
2. Configuration
Create a .env file based on src/core/config.py defaults:
MT5_LOGIN=your_account
MT5_PASSWORD=your_password
MT5_SERVER=your_broker_server
MODE=demo
3. Execution
# Run validation and start in demo mode
python main.py --mode demo --symbol XAUUSD --verbose
📜 Documentation Index
| Guide | Description |
|---|---|
| Architecture Quick-Start | Primary technical overview and system maturity map. |
| DEVELOPMENT_PLAN.md | Technical roadmap and implementation milestones. |
| ENTERPRISE_STANDARDS.md | Coding standards, CI/CD requirements, and security policies. |
| DEPLOYMENT_GUIDE.md | Step-by-step instructions for Docker and Cloud deployment. |
| DATABASE_STANDARDS.md | Schemas for trade logging and performance tracking. |
| Contributing Guide | How to contribute safely and effectively. |
| Contribution Map | Safe vs. Sensitive zone navigation. |
⚖️ License
Distributed under the MIT License. See LICENSE for more information.
Disclaimer: Trading involves significant risk. This software is for educational purposes only. The developers assume no liability for financial losses.