AlgoTradeX: Intelligent Automated Trading System

This is an automated trading system designed for live trading and strategy testing using the Alpaca API. It retrieves real-time market data, calculates indicators, generates trade signals and executes orders based on predefined strategies.

It supports:

• Live trading using Alpaca's API on AWS EC2, allowing continuous execution in a cloud environment.
• Local execution using Docker for development and strategy testing.
• Automated deployments to EC2 via a GitHub Actions CD pipeline, ensuring seamless updates when changes are pushed.

⚠️ Note: A legacy TCP-based simulated trading approach is included in this repo but is no longer the primary method.

📂 Project Structure

📂 trading-system/
│── 📜 README.md             # Project documentation & setup instructions
│── 📜 requirements.txt      # Python dependencies
│── 📜 Dockerfile            # 🐳 Docker setup for containerized deployment
│── 📜 main_day_trader.py    # 🚀 Entry point for Alpaca paper trading (Alpaca API)

├── 📂 data/                 # 📊 Historical market data and trade logs

├── 📂 models/               # 🧠 Trained ML models and feature scalers  

├── 📂 notebooks/            # 📓 Strategy research, backtesting and model training  

├── 📂 src/                  # 🏗️ Core trading system  
│   ├── 📂 alpaca_utils/     # 📡 Alpaca API modules
│   ├── 📂 local_trading/    # 🔌 Simulated trading & backtesting (TCP-based)
│   └── 📂 scripts/          # 🛠️ Utility scripts (data fetching, API checks)

├── 📂 .github/workflows/    # 🔄 CI/CD pipeline  
│   └── 📜 deploy.yml        # 🚀 GitHub Actions workflow for EC2 auto-deploy

🚀 Getting Started

1️⃣ Clone the Repository

Using SSH (Recommended):

git clone git@github.com:sohammandal/trading-system.git
cd trading-system

2️⃣ Create a Virtual Environment & Install Dependencies

python -m venv venv
source venv/bin/activate  # Mac/Linux
venv\Scripts\activate     # Windows
pip install -r requirements.txt

3️⃣ Configure API Keys

Obtain API Keys from Alpaca, create a .env file as seen below in the project root and ensure it is included in .gitignore:

ALPACA_API_KEY=your_api_key
ALPACA_SECRET_KEY=your_secret_key

4️⃣ Verify Setup by Checking Account Balance

Run the following command to confirm your API keys are working correctly:

python src/scripts/check_balance.py # Checks API Connectivity

🛠 Running the Trading Bot with Docker

Easily run the trading bot locally using Alpaca's API in a containerized environment:

1️⃣ Build the Docker Image

docker build -t trading-bot .

2️⃣ Run the Trading Bot

docker run -d --restart unless-stopped --env-file .env --name trading-bot trading-bot

This will start the bot in a detached mode, automatically restarting if it stops.

3️⃣ Check Logs

docker logs -f trading-bot

4️⃣ Stop & Remove the Container

docker stop trading-bot && docker rm trading-bot

👉 Make sure your .env file contains valid API credentials before running the container.

🌍 Deploying to AWS EC2

The trading bot can run continuously on AWS EC2, leveraging GitHub Actions for automated deployments whenever updates are pushed to the repository. This ensures that the bot is always running the latest version without manual intervention.

1️⃣ Set Up an AWS EC2 Instance

• Launch a t2.micro (free-tier) EC2 instance running Ubuntu.
• Install Docker & Git on the instance.
• Configure security groups to allow SSH (port 22).
• Generate an SSH key pair in the instance for GitHub authentication.
• Clone the repository and configure environment variables (create a .env file in the project root).

2️⃣ Set Up GitHub Actions for Automated Deployment

• Add the EC2 SSH private key as a GitHub Secret (EC2_SSH_PRIVATE_KEY).
• Store EC2 public IP, username (ubuntu), and other details as repo secrets (EC2_HOST, EC2_USER).
• Configure the GitHub Actions workflow (.github/workflows/deploy.yml) to:
  • SSH into EC2.
  • Pull the latest code from GitHub.
  • Rebuild and restart the Docker container running the bot.

Once set up, any push to the main branch will automatically deploy the latest version to EC2.

3️⃣ Monitor & Manage the Bot

• Check live logs:

  docker logs -f trading-bot

• Restart the bot if needed:

  docker restart trading-bot

• Manually update & restart:

  git pull origin main
  docker stop trading-bot && docker rm trading-bot
  docker build -t trading-bot .
  docker run -d --restart unless-stopped --env-file .env --name trading-bot trading-bot

This setup ensures a reliable, automated, and cloud-based live trading environment with minimal maintenance.

⚡ Legacy: Simulated Trading (TCP-based, No API Keys Required)

This older testing method streams historical data from CSV files to simulate real-time trading.

1️⃣ Start the TCP Server

Streams market data to connected clients:

python src/local_trading/tcp_server.py -p 9999 -f data/historical_stock_data_5min_6months.csv -t 0.1

2️⃣ Run the Trading Client

Executes trades based on the simulated market data:

python src/local_trading/four_factor_day_trader.py

3️⃣ Stop the Client

Use Ctrl + C to stop execution.

All trade logs are saved in the data/ directory. 🚀