A training repository for learning GitHub Copilot best practices through hands-on ML and MCP server exercises.

IMPORTANT: Before the training, you must complete the pre-training setup.

See PRE-TRAINING-SETUP.md and finish all steps at least 48 hours in advance.

📚 What's Inside

This repository contains:

• Training exercises for GitHub Copilot workflows and Model Context Protocol (MCP) servers
• Minimal ML pipeline for airline discount prediction using synthetic data
• Complete setup automation with cross-platform support (Mac/Linux/Windows)

🎯 Learning Path

Exercise 01: Setup Copilot Instructions

Learn to configure repo-wide and path-scoped Copilot instructions for consistent, high-quality code generation.

📁 exercises/01-setup/

Exercise 02: MCP Server for GitHub

Explore GitHub's MCP server for repository management, issue tracking, and code search through Copilot.

📁 exercises/02-mcp-server-for-github/

Exercise 03: Build a Custom MCP Server

Create your own MCP server that wraps the Synth CLI for synthetic data generation, then connect it to VS Code.

📁 exercises/03-custom-mcp-server/

• Setup guide: Install Synth, generate data, load into SQLite
• Main exercise: Build FastAPI MCP server, add tools, connect to Copilot
• Homework: Deploy with Docker, add security, share with team

Exercise 04: Unit Tests with pytest

Master test-driven development with GitHub Copilot by writing comprehensive unit tests.

📁 exercises/04-unit-tests/

🚀 Quick Start

For the ML Pipeline

cd airline-discount-ml
./setup.sh  # Mac/Linux
# OR
setup.bat   # Windows

# Verify setup
make test

See airline-discount-ml/README.md for detailed instructions.

For Exercises

Each exercise folder contains:

• Detailed step-by-step instructions
• Code samples and templates
• Troubleshooting guides
• Expected outcomes

Start with Exercise 01 to configure your Copilot environment.

📁 Repository Structure

airlst-github-copilot-training/
├── README.md                    # This file
├── exercises/                   # Training exercises
│   ├── 01-setup/               # Copilot instructions setup
│   ├── 02-mcp-server-for-github/  # GitHub MCP server
│   ├── 03-custom-mcp-server/   # Build custom MCP server
│   └── 04-unit-tests/          # pytest unit testing with copilot
└── airline-discount-ml/        # ML project
    ├── src/                    # Source code
    ├── tests/                  # Test suite
    ├── data/                   # SQLite database and schemas
    ├── notebooks/              # Jupyter notebooks
    ├── setup.sh / setup.bat    # Automated setup
    └── README.md               # Detailed project docs

🎓 Who Is This For?

• Developers learning GitHub Copilot best practices
• Teams adopting AI-assisted development workflows
• Anyone interested in building custom MCP servers
• Data scientists exploring synthetic data generation

📝 Prerequisites

• VS Code with GitHub Copilot extensions
• Python 3.8+
• Git
• (Optional) Docker for custom MCP deployment

🔗 Key Technologies

• GitHub Copilot - AI pair programming
• Model Context Protocol (MCP) - Extend Copilot with custom tools
• Synth CLI - Generate realistic synthetic data
• FastAPI - Build web APIs for MCP servers
• SQLite - Lightweight database for training
• pytest - Python testing framework
• scikit-learn - Machine learning models

🤝 Contributing

This is a training repository. Feel free to fork and adapt for your own training needs.

📄 License

Training materials for educational purposes.

---

Ready to start? Head to exercises/01-setup/ and follow the instructions to configure your Copilot environment.