Skip to content

Latest commit

 

History

History
183 lines (113 loc) · 7.33 KB

README.md

File metadata and controls

183 lines (113 loc) · 7.33 KB

🌟 AmpyFin Trading Bot

🚀 Introduction

Welcome to AmpyFin, an advanced AI-powered trading bot designed for the NASDAQ-100. Imagine having expert traders working for you 24/7—AmpyFin makes this a reality.

📊 AmpyFin’s Data Collection Power

🔍 Data Sources

  • Financial Modeling Prep API: Retrieves NASDAQ-100 tickers to gain crucial market insights.
  • Polygon API: Monitors real-time market conditions, ensuring that the bot acts based on the most current data.

💾 Data Storage

All data and trading logs are securely stored in MongoDB, allowing fast access to historical trading information and supporting in-depth analysis.

🤖 Machine Learning at Work

At the core of AmpyFin are diverse algorithms optimized for different market conditions. Rather than relying on a single strategy or multiple strategies, AmpyFin relies on a ranked ensemble learning system that dynamically ranks each strategy and gives more influence in the final decision to strategies with better performance.

📈 Trading Strategies

Some of the strategies AmpyFin employs include:

  • 📊 Mean Reversion: Predicts asset prices will return to their historical average.
  • 📈 Momentum: Capitalizes on prevailing market trends.
  • 💱 Arbitrage: Identifies and exploits price discrepancies between related assets.
  • 🧠 AI-Driven Custom Strategies: Continuously refined through machine learning for enhanced performance.

These strategies work collaboratively, ensuring AmpyFin is always prepared for changing market dynamics.

🔗 How Dynamic Ranking Works

Managing multiple algorithms is simplified with AmpyFin’s dynamic ranking system, which ranks each algorithm based on performance.

🏆 Ranking System

Each strategy starts with a base score of 0 and a mock balance of $50,000. The system evaluates their performance and assigns a weight based on the following function:

$$ \left( \frac{e^e}{e^2} - 1 \right)^{2i} $$

Where (i) is the strategy's rank. Please keep in mind that the strategy's rank

⏳ Time Delta Coefficient

This ensures that strategies with better recent performance have a greater influence on decision-making while maintaining balance by also accounting for old performance as well.

💡 Benefits of Dynamic Ranking

  • 📉 Quickly adapts to changing market conditions.
  • 📊 Prioritizes high-performing algorithms.
  • ⚖️ Balances risk while maximizing potential returns.

📂 File Structure and Objectives

🤝 trading_client.py

Objective: Executes trades based on algorithmic decisions.

Features:

  • Executes trades every 60 seconds by default (adjustable based on user).
  • Ensures a minimum spending balance of $15,000 (adjustable based on user) and maintains 30% liquidity (adjustable based on user).
  • Logs trades with details like timestamp, stock, and reasoning.

🏆 ranking_client.py

Objective: Runs the ranking system to evaluate trading strategies.

Features:

  • Downloads NASDAQ-100 tickers and stores them in MongoDB.
  • Updates algorithm scores and rankings every 30 seconds (adjustable based on user).

📜 strategies/*

Objective: Defines various trading strategies. Houses strategies like mean reversion, momentum, and arbitrage.

Features:

  • trading_strategies_v1.py: Archived first iteration of AmpyFin used 5 strategies. This file is not supported anymore but is a great reference material
  • trading_strategies_v2.py: Archived second gen older strategies being used in the ranking system. Contains 50 strategies with a lot leaning towards momentum.
  • trading_strategies_v2_1.py: Archived second gen older strategies that complements the older strategies in trading_strategies_v2.py. Houses 10 more strategies. This is where newer strategies will be implemented until it caps at 50 strategies as well.
  • talib_indicators.py: Contains all the technical indicators used in the strategies. To visit the documentation for each technical indicator, please visit the following link: Link to TA. These indicators were not developed by me, but I have modified their use to fit the needs of AmpyFin.

🔧 helper_files/*

Objective: Helper Files to help with both trading client and ranking client. Houses functions for retrieving a Mongo Client, getting latest prices, current strategies implemented etc.

Features:

  • client_helper.py: Contains common functions for client operations in both ranking and trading.

⚙️ Installation

1️⃣ Clone the Repository

git clone https://github.com/yeonholee50/AmpyFin.git
cd AmpyFin

2️⃣ Install Dependencies

  • Run the following command to install the required Python packages:
pip install -r requirements.txt
  • We have recently migrated to using Ta-Lib for trading. Please follow the installation instructions here:

👉 Ta-Lib Python Original

👉 Ta-Lib Python Easy Installation

3️⃣ Configuration

  1. Create config.py:
    • Copy config_template.py to config.py and enter your API keys and MongoDB credentials.
    POLYGON_API_KEY = "your_polygon_api_key"
FINANCIAL_PREP_API_KEY = "your_fmp_api_key"
MONGO_DB_USER = "your_mongo_user"
MONGO_DB_PASS = "your_mongo_password"
API_KEY = "your_alpaca_api_key"
API_SECRET = "your_alpaca_secret_key"
BASE_URL = "https://paper-api.alpaca.markets"
mongo_url = "your mongo connection string"

4️⃣ API Setup

  • Polygon API
  1. Sign up at Polygon.io and get an API key.
  2. Add it to config.py as POLYGON_API_KEY.
  • Financial Modeling Prep API
  1. Sign up at Financial Modeling Prep and get an API key.
  2. Add it to config.py as FINANCIAL_PREP_API_KEY.
  • Alpaca API
  1. Sign up at Alpaca and get API keys.
  2. Add them to config.py as API_KEY and API_SECRET.

5️⃣ Set Up MongoDB

  • Sign up for a MongoDB cluster (e.g., via MongoDB Atlas).
  • Create a database for stock data storage and replace the mongo_url in 'config.py' with your connection string. Make sure to give yourself Network Access.
  • Run the setup script setup.py:
  • After running the mongo setup script, the MongoDB setup for the rest will be completed on the first minute in trading for both ranking and trading.

⚡ Usage

To run the bot, execute on two separate terminals:

python ranking_client.py
python trading_client.py

⚠️ IMPORTANT

For people looking to do live trading, I suggest training via running ranking_client.py for at least two weeks before running the trading bot altogether. This way, you're running with a client that has been trained to a certain extent (with strategies ranked) and is ready to go. Otherwise, you will most likely be buying random stocks.

📑 Logging

  • system.log: Tracks major events like API errors and MongoDB operations.
  • rank_system.log: Logs all ranking-related events and updates.

🛠️ Contributing

Contributions are welcome! 🎉 Feel free to submit pull requests or report issues. All contributions should be made on the test branch. Please avoid committing directly to the main branch.

📜 License

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