QTrade

QTrade is a small, modular Python library for backtesting trading strategies and training reinforcement learning agents on financial data.

Features

• Single-asset and portfolio backtesting: pass one DataFrame for single-asset, or a dict[str, DataFrame] for multi-asset / portfolio strategies sharing one cash pool.
• Walk-forward optimization: rolling-window training + immediate out-of-sample test, the standard antidote to in-sample overfitting.
• RL trading environment: a Gymnasium TradingEnv with pluggable Action / Observation / Reward schemes, ready to drop into stable-baselines3 and friends.
• Bokeh reports: equity, drawdown, per-asset OHLC panels, trade scatter — saved as a single self-contained HTML.
• Curated metrics: Sharpe, Sortino, Calmar, Omega, max drawdown duration, win rate, profit factor, etc.

See COMPARISON.md for an honest take on where QTrade fits next to backtrader / vectorbt / Backtesting.py.

Installation

pip install qtrade-lib

Optional extras:

pip install "qtrade-lib[rl]"   # adds stable-baselines3 for RL evaluation utils

To work from source:

git clone https://github.com/gguan/qtrade.git
cd qtrade
pip install -e ".[dev]"
pre-commit install
pytest

Quickstart

import yfinance as yf
from qtrade.backtest import Strategy, Backtest

class SmaCrossover(Strategy):
    n1 = 5
    n2 = 20

    def prepare(self):
        for df in self._data.values():
            df['sma1'] = df['Close'].rolling(self.n1).mean()
            df['sma2'] = df['Close'].rolling(self.n2).mean()

    def on_bar_close(self):
        s1, s2 = self.data['sma1'], self.data['sma2']
        if s1.iloc[-2] < s2.iloc[-2] and s1.iloc[-1] > s2.iloc[-1]:
            self.buy()
        elif s1.iloc[-2] > s2.iloc[-2] and s1.iloc[-1] < s2.iloc[-1]:
            self.close()

data = yf.download("GC=F", start="2023-01-01", end="2024-01-01",
                   interval="1d", multi_level_index=False)

bt = Backtest(data, SmaCrossover, cash=10_000, trade_on_close=True)
bt.run()
bt.show_stats()
bt.plot()

For a multi-asset / walk-forward example, see examples/portfolio_strategy.py.

Documentation

• User Guide
• Multi-asset / portfolio backtests
• RL trading environment
• API reference
• Releasing
• Changelog

Requirements

• Python ≥ 3.10
• Runtime dependencies declared in pyproject.toml (numpy, pandas, scipy, matplotlib, bokeh, gymnasium, mplfinance, tqdm).

References

This project is inspired by:

• backtesting.py — single-asset backtest API design
• tensortrade — RL trading framework

License

MIT — see LICENSE.