A Python backtesting engine with machine learning capabilities for testing trading strategies. Built with FastAPI, scikit-learn, and pandas. All data stored in memory.

• Three trading strategies: Moving Average Crossover, RSI, Bollinger Bands
• Real market data via Yahoo Finance
• Performance metrics: return, Sharpe ratio, max drawdown, win rate
• Random Forest ML models for price prediction and signal generation
• FastAPI REST API with HTML dashboard
• In-memory storage (no database)

git clone <repository-url>
cd backtesting-engine
uv sync

Start the API server:

uv run python main.py api

API available at http://localhost:8000

Start the frontend:

uv run python server.py

Dashboard available at http://localhost:8080

Test the engine:

uv run python main.py test

• POST /backtest/run - Start a backtest
• GET /backtest/status/{run_id} - Check status
• GET /results/metrics - Get performance metrics
• POST /ml/train/{symbol} - Train ML models
• GET /ml/signals/{symbol} - Get trading signals
• GET /data/prices/{symbol} - Get price data

Full API documentation at http://localhost:8000/docs

Moving Average Crossover: Buy/sell on MA crossovers. Parameters: short_period, long_period, lookback

RSI Strategy: Buy when oversold, sell when overbought. Parameters: period, oversold, overbought, lookback

Bollinger Bands: Buy at lower band, sell at upper band. Parameters: period, std_dev, lookback

import requests

# Run a backtest
response = requests.post('http://localhost:8000/backtest/run', json={
    "strategy": "moving_average_crossover",
    "symbol": "AAPL",
    "start_date": "2024-01-01",
    "end_date": "2024-12-31",
    "initial_capital": 100000,
    "parameters": {"short_period": 10, "long_period": 20}
})

# Train ML models
requests.post('http://localhost:8000/ml/train/AAPL')

# Get signals
signals = requests.get('http://localhost:8000/ml/signals/AAPL')