enode-backtester 0.1.5

Professional event-driven backtesting framework for quantitative trading strategies

enode-backtester

Lightweight event-driven backtesting framework that wires together data handlers, strategies, sizing logic, a portfolio, execution simulation, and analysis helpers. The library is designed as a foundation you can extend with richer metrics, visualisations, and strategy research.

---

Repository Layout

• backtester/__init__.py – Public entry point exporting run_backtest, utility helpers, and default sizing/execution models.
• backtester/data.py – Data-handling abstractions. DataFrameDataHandler streams pandas data frames into the event queue as StockEvents.
• backtester/event.py – Event hierarchy that flows through the engine (market data, signals, orders, fills).
• backtester/strategy.py – Base strategy contract; provides the signal helper for emitting SignalEvents.
• backtester/sizer.py – Base sizing abstraction plus a simple FixedSizeSizer.
• backtester/portfolio.py – Tracks cash, positions, trade log, equity curve, and reacts to incoming events.
• backtester/execution.py – Simulated execution handler with configurable slippage and commission models.
• backtester/engine.py – Core event loop that coordinates the handlers and processes the event queue.
• backtester/analysis.py – Post-run utilities, including an optional PyFolio tear sheet and trade log export.
• backtester/models.py – Pydantic models shared across modules (e.g., StockCandle).

---

Installation

# using uv (recommended)
uv sync

# or using pip / virtualenv
python -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt

pyfolio is optional; install it (pip install pyfolio) if you need tear-sheet generation.

---

Creating a Strategy

1. Subclass backtester.strategy.BaseStrategy.
2. Implement on_stock_event(self, event) to react to incoming candles (event.payload is a StockCandle).
3. Use self.signal(symbol, timestamp, action) to enqueue SignalEvents (action is typically "buy" or "sell").

from collections import deque

from backtester.strategy import BaseStrategy
from backtester.event import StockEvent


class MySMAStrategy(BaseStrategy):
    def __init__(self, event_queue, data_handler, window: int = 20):
        super().__init__(event_queue, data_handler)
        self.window = window
        self.prices: dict[str, deque[float]] = {}

    def on_stock_event(self, event: StockEvent) -> None:
        symbol = event.payload.symbol
        price = event.payload.close
        history = self.prices.setdefault(symbol, deque(maxlen=self.window))
        history.append(price)

        if len(history) < history.maxlen:
            return

        avg_price = sum(history) / len(history)
        if price > avg_price:
            self.signal(symbol, event.payload.timestamp, "buy")
        elif price < avg_price:
            self.signal(symbol, event.payload.timestamp, "sell")

---

Running a Backtest

import pandas as pd

from backtester import FixedSizeSizer, run_backtest, get_trade_log_df
from backtester.data import DataFrameDataHandler
from my_strategies import MySMAStrategy

# 1. Prepare your market data as pandas DataFrames keyed by symbol
data_dict: dict[str, pd.DataFrame] = {
    "AAPL": aapl_df,
    "GOOGL": googl_df,
}

# 2. Choose a sizing model (optional)
sizer = FixedSizeSizer(default_size=10)

# 3. Execute the backtest
portfolio = run_backtest(
    data_dict=data_dict,
    strategy_class=MySMAStrategy,
    initial_cash=100_000.0,
    sizer=sizer,
)

# 4. Inspect results
trade_log = get_trade_log_df(portfolio)
print(trade_log.tail())
print(f"Final equity: {portfolio.equity_curve[-1][1]:,.2f}")

run_backtest instantiates all core components for you:

• DataFrameDataHandler feeds candles into the event queue.
• Your strategy consumes StockEvents and emits SignalEvents.
• Portfolio generates OrderEvents using the provided Sizer.
• SimulatedExecutionHandler fills orders with configurable commission and slippage.

After the engine finishes processing all data, you receive the final Portfolio object for downstream analysis.

---

Extending the Framework

• Metrics & Visuals: Add new functions under backtester.analysis or integrate notebooks that consume the trade log and equity curve.
• Strategies: Drop additional strategy classes alongside your own module; they only need to inherit from BaseStrategy.
• Execution/Sizing: Create custom classes that follow the abstract base class signatures in execution.py and sizer.py.
• Data: Implement new handlers by subclassing BaseDataHandler if you need live feeds, streaming data, or alternative storage.

---

Contributing

Pull requests and issues welcome—especially around additional strategy templates, validation rules, and analytical tooling.