2,2 KiB
2,2 KiB
OKX Crypto Trading Strategy
This project is a quantitative trading strategy designed for OKX cryptocurrency exchange, based on an AI-driven architecture using DeepSeek and Qwen models.
Project Structure
crypto/trading_bot.py: Main entry point for the trading bot. Orchestrates data fetching, AI analysis, and trade execution.crypto/okx_data_processor.py: Handles interaction with OKX API viaccxt. Fetches OHLCV data and calculates technical indicators (EMA, RSI, ATR).crypto/ai_client_factory.py: Factory for creating AI clients.crypto/deepseek_client.py: Client for DeepSeek API (Market Structure Analysis).crypto/qwen_client.py: Client for Qwen API (Strategy Optimization and Decision Making).crypto/database_manager.py: Manages the SQLite database for trade logging.crypto/.env: Configuration file for API keys.crypto/requirements.txt: Python dependencies.
Setup
-
Install Dependencies:
pip install -r crypto/requirements.txt -
Configuration: The
crypto/.envfile has been pre-configured with your provided OKX API keys. Note: You need to add yourSILICONFLOW_API_KEYto the.envfile for the AI models to work. -
Run the Bot: To run a single pass of the strategy:
python crypto/trading_bot.pyOr use the watchdog script to keep it running:
./crypto/run_bot_watchdog.sh
Strategy Logic
- Data Acquisition: Fetches historical candle data from OKX.
- Feature Engineering: Calculates technical indicators (EMA, RSI, ATR, Volatility).
- Market Analysis (DeepSeek): Analyzes market structure, support/resistance, and trends.
- Decision Making (Qwen): Combines market analysis with technical signals to generate trading decisions (Buy/Sell/Hold) and manage risk (Stop Loss/Take Profit).
- Execution: Executes trades on OKX (currently in logging mode, uncomment execution lines in
trading_bot.pyto enable live trading).
Notes
- The bot currently logs actions instead of executing real trades for safety. Review
trading_bot.py'sexecute_trademethod to enable real orders. - Ensure your API keys have the necessary permissions (Trade, Read).