A-Zero

A-Zero is a modular, data-driven, and safety-oriented framework written in Java for building, backtesting, and deploying automated crypto trading strategies.

What is A-Zero?

This project provides the core infrastructure needed to apply a systematic and rigorous approach to algorithmic trading. It is built with an "Open Core" philosophy: the framework itself is open-source, allowing you to build your own proprietary trading strategies on a robust and well-tested foundation.

Our core guiding principles are:

• Safety First: Emphasizing risk management, robust error handling, and a phased approach to deployment.
• Modularity: A clean, microservice-ready architecture that separates concerns like data ingestion, signal generation, and execution.
• Testability: A powerful backtesting engine is at the heart of the project, enabling data-driven validation of all strategies before they ever touch a live market.

Core Features

• Command-Line Backtester: A powerful and fast CLI tool to run your strategies against historical data and generate performance reports.
• Declarative Strategy Definition: Define complex trading strategies in simple, human-readable YAML files using a rules-based approach.
• Realistic Simulation Engine: A backtesting library with a portfolio margin model, trading costs (fees, slippage), and forced liquidation for margin calls.
• Data Ingestion Utility: A simple CLI tool for downloading historical K-line data from exchanges like Binance.
• Pluggable Strategy Interface: Define your trading logic by implementing a simple Strategy interface.
• Clean, Modern Java: Built with modern Java and a minimal set of dependencies, ready for integration with frameworks like Spring.

Build Status & CI

This project maintains a high standard of code quality through a Continuous Integration (CI) pipeline powered by GitHub Actions. Every push and pull request to our main development branches is automatically built and verified by running a comprehensive test suite.

For detailed logs and a history of all recent builds, please visit the Actions tab.

Project Roadmap

This project is in its early stages. Our planned development path is:

• v0.1: Foundational libraries and data ingestion utility.
  • a0-core: Shared data models and interfaces.
  • a0-data-ingestor: CLI tool for downloading historical data.
  • a0-backtester: Core library for strategy simulation.
• v0.2: Command-Line Backtester and Strategy Integration.
  • a0-backtester: Implement strategy loading and execution within the BacktestEngine.
  • a0-strategy-rules-engine: Build the engine to parse YAML-based strategies.
  • a0-backtester-cli: Build the command-line application to run backtests.
• v0.3: Initial hooks and interfaces for sentiment analysis modules.
• v0.4: Integration with exchange Testnet APIs for paper trading.
• v1.0: A stable, production-ready framework for live spot trading.

Getting Started

The A-Zero framework is under active development. It currently provides two primary command-line tools: a data ingestor and a backtester.

Downloading Historical Data

1. Clone the repository: git clone https://github.com/impatient0/a-zero.git
2. Navigate to the project directory: cd a-zero
3. Build the project using Maven: mvn clean package
4. Run the data ingestor tool:

   java -jar a0-data-ingestor/target/a0-data-ingestor-0.1.0-SNAPSHOT.jar --symbol=BTCUSDT --timeframe=1h --start-date=2023-01-01

This will create a BTCUSDT-1h.csv file in your project root, ready for use with the backtesting engine.

2. Running a Backtest

Once you have data, you can run a simulation using the backtester-cli.

1. Define a Strategy: Create a strategy configuration file (e.g., my-strategy.yaml). You can use the example in the a0-strategy-rules-engine README as a template.

2. Run the Backtester: Execute the backtester JAR, pointing it to your strategy and data files.

   java -jar a0-backtester-cli/target/a0-backtester-cli-*.jar \
     --strategy-file=my-strategy.yaml \
     --data-file=data/BTCUSDT-1h.csv \
     --symbol=BTCUSDT \
     --account-mode=MARGIN \
     --leverage=10

   This will run the simulation and print a performance summary to your console.

License

This project is licensed under the Apache License 2.0. See the LICENSE file for details.