Overview
This repository serves as the central documentation hub for Jean Guillaume’s Trading (JGT) platform - a comprehensive algorithmic trading system designed for autonomous signal detection, multi-timeframe analysis, and intelligent trade execution.
The JGT platform consists of interconnected packages forming a complete trading ecosystem:
📊 Data & Analytics Layer
- jgtpy - Market data services and core indicator calculations
- jgtcore - Shared libraries and core functionality
🤖 Machine Learning & Signal Detection
- jgtml - ML models, signal detection, and pattern intelligence
- jgtapy - Technical analysis indicators
🔗 Trading Execution & Connectivity
- jgtfxcon - FX connectivity and order management
- jgtutils - Platform utilities and configuration
🧠 Autonomous Trading Intelligence
- jgtagentic - Agentic trading orchestration and campaign management
Key Capabilities
Signal Detection System
- FDB (Fractal Divergent Bar) analysis for precise entry signals
- Multi-degree Alligator analysis (Regular/Big/Tide timeframes)
- Five Dimensions Confluence validation across multiple indicators
- Pattern intelligence learning and adaptation
Multi-Timeframe Analysis
- Hierarchical timeframe analysis from monthly to minute charts
- Higher timeframe bias integration for trend alignment
- Confluence validation across timeframe dimensions
- Automated signal scanning across multiple instruments
Autonomous Execution
- Risk-managed order sizing based on signal quality
- Stop loss management using technical levels
- Campaign orchestration for multi-instrument strategies
- Performance tracking and pattern learning
Data Processing Pipeline
Market Data (PDS) → Technical Indicators (IDS) → Chart Analysis (CDS/ADS) →
Feature Engineering (TTF) → Matrix Analysis (MX) → Signal Detection → Order Execution
For Developers & LLMs
Documentation Access
- Central Hub: llms.txt - Complete platform overview for LLMs
- Package-Specific: Each package has dedicated llms.txt for focused documentation
- Pattern:
https://<package>.jgwill.com/llms.txt for published documentation
Development Philosophy
- Intent-Driven Development - Trading strategies specified in natural language
- Pattern Intelligence - Continuous learning from signal performance
- Multi-Timeframe Confluence - Systematic validation across timeframes
- Autonomous Execution - Minimal human intervention once configured
Current Development State
- Production: Core data services and signal detection
- Active Development: Pattern intelligence and signal quality prediction
- Refactoring: jgtutils → jgtcore migration for better separation of concerns
- Innovation: Intent-driven trading specification language (SpecLang)
Getting Started
For Traders
- Install core packages:
pip install jgtpy jgtml jgtfxcon
- Configure market data connection
- Run signal scanner:
jgtapp fdbscan
- Review generated trading signals
For Developers
- Review platform architecture in llms.txt
- Explore package-specific documentation
- Understand data pipeline and signal detection logic
- Contribute to pattern intelligence development
For LLMs
- Start with llms.txt for complete platform context
- Reference package-specific llms.txt files for detailed component information
- Use provided examples and CLI patterns for implementation guidance
Technical Specifications
Supported Markets
- Forex: Major and minor currency pairs
- Indices: SPX500, AUS200, and other major indices
- Commodities: XAU/USD and other precious metals
Timeframes
- Long-term: M1 (Monthly), W1 (Weekly)
- Medium-term: D1 (Daily), H4, H8 (Multi-hour)
- Short-term: H1, H2, H3 (Hourly)
- Intraday: m15, m5, m1 (Minutes)
Signal Types
- Entry Signals: FDB breakouts, Alligator confluences
- Exit Signals: Technical stop levels, pattern completions
- Risk Management: Dynamic position sizing, stop loss management
For Individual Traders
- Automated Signal Detection - No manual chart watching required
- Multi-Market Coverage - Simultaneous analysis across instruments
- Risk Management - Built-in position sizing and stop loss logic
- Performance Tracking - Continuous improvement through pattern learning
For Trading Teams
- Systematic Approach - Consistent signal detection methodology
- Scalable Analysis - Handle multiple strategies simultaneously
- Knowledge Retention - Pattern intelligence preserves team learning
- Campaign Management - Coordinate complex multi-instrument strategies
For Algorithm Developers
- Modular Architecture - Clear separation of concerns across packages
- Extensible Framework - Add new indicators and signal types
- Pattern Learning - Built-in machine learning for signal improvement
- Complete Testing - Backtesting and forward testing capabilities
Support & Resources
- Documentation: Package-specific llms.txt files
- Examples: Working implementations in each package
- Community: GitHub discussions and issue tracking
- Development: Continuous integration and testing
The JGT platform represents a comprehensive approach to algorithmic trading, combining traditional technical analysis with modern machine learning and autonomous execution capabilities.