Pybacktest: A Python Backtesting Framework

Pybacktest is a Python library for backtesting trading strategies. It allows you to define strategies using a simple JSON format, run simulations on historical stock data, and visualize the performance of your strategies.

Features

• Easy-to-use API: Simple and intuitive API for running backtests.
• Flexible Strategy Configuration: Define complex trading strategies using a JSON format.
• Historical Data: Fetches historical stock data from Yahoo Finance.
• Performance Visualization: Plot the performance of your portfolio over time.
• Streamlit UI: An interactive web interface for creating, editing, and backtesting strategies.

Installation

You can install Pybacktest and its core dependencies using pip:

1. Clone the repository:

   git clone https://github.com/sciencemj/pybacktest.git
   cd pybacktest

2. Install in editable mode (this will also install the core dependencies):

   pip install -e .

How to Use

Here's a simple example of how to use Pybacktest to backtest a trading strategy:

from pybacktest.backtest import Backtest
from pybacktest.strategy import StrategyWrapper, StrategyManager
from pybacktest.models import Stock

# 1. Define the backtest period and create Stock objects
start = '2023-01-01'
end = '2023-12-31'
apple = Stock("AAPL", start, end)
tqqq = Stock("TQQQ", start, end)

# 2. Define your trading strategy in JSON format
strategy_json = {
    "AAPL": {
        "buy": {
            "ticker": "AAPL",
            "by": ["current", "Change_Pct"],
            "period": False,
            "criteria": ["percent-change", 0.5],
            "quantity": ["count", 10]
        },
        "sell": {
            "ticker": "AAPL",
            "by": ["current", "Close"],
            "period": False,
            "criteria": ["profit-rate", 10],
            "quantity": ["percent", 100]
        }
    },
    "TQQQ": {
        "buy": {
            "ticker": "TQQQ",
            "by": ["average", "Close"],
            "period": 3,
            "criteria": ["point", 0.05],
            "quantity": ["value", 1000]
        },
        "sell": {
            "ticker": "AAPL",
            "by": ["current", "Change_Pct"],
            "period": False,
            "criteria": ["percent-change", -3],
            "quantity":["percent", 50]
        }
    }
}

# 3. Create a StrategyManager
strategy = StrategyManager("MyStrategy", StrategyWrapper(**strategy_json))

# 4. Create a Backtest instance and run the simulation
backtest = Backtest([apple, tqqq], [strategy], 100000)
backtest.run()

# 5. Plot the performance
backtest.plot_performance()

Strategy Configuration

Trading strategies are defined in a JSON format. Each key in the JSON object represents a stock ticker. The value is an object with buy and sell rules.

TradeAction

A TradeAction object has the following fields:

• ticker: The ticker of the stock to trade.
• by: A list specifying the value to use for comparison.
  • The first element can be "average" or "current".
  • The second element is the field to use from the stock data (e.g., "Close", "Change_Pct").
• period: The period for the moving average if by is "average". false if not used.
• criteria: The condition for the trade.
  • The first element is the type of criteria (e.g., "percent-change", "profit-rate", "point").
  • The second element is the value for the criteria.
• quantity: The quantity to trade.
  • The first element is the type of quantity (e.g., "count", "percent", "value").
  • The second element is the value for the quantity.
• trade_as: The price to use for the trade (e.g., "Close", "Open").

Streamlit UI

You can access the live Streamlit application here: pybacktest.streamlit.app

Pybacktest comes with an interactive Streamlit web interface to help you create, edit, and backtest your trading strategies.

To run the Streamlit app locally, execute the following command:

streamlit run streamlit_page.py

The UI allows you to:

• Create and edit strategies using a user-friendly form.
• View the JSON representation of your strategies.
• Run backtests and see the results, including performance plots and trade history.
• Download your strategies as a JSON file.

License

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