The bot's architecture includes components like `trading_client.py` for executing trades and `ranking_client.py` for evaluating strategies. It requires setup with API keys and MongoDB credentials, and users are advised to run the ranking client for two weeks before live trading. The project is open for contributions under the MIT License, with a focus on maintaining code quality through a test branch.
Key takeaways:
- AmpyFin is an AI-powered trading bot designed for the NASDAQ-100, utilizing a ranked ensemble learning system to dynamically rank and prioritize high-performing trading strategies.
- The bot collects data from sources like Financial Modeling Prep API and Polygon API, storing it securely in MongoDB for fast access and analysis.
- AmpyFin employs various trading strategies, including mean reversion, momentum, arbitrage, and AI-driven custom strategies, to adapt to changing market conditions.
- To use AmpyFin, users must configure API keys and MongoDB credentials, and it is recommended to run the ranking client for at least two weeks before live trading to ensure strategies are well-ranked.