pyquaver 0.1.0

Standalone PyPI library for quantitative trading signal generation and backtesting

quaver

Standalone Python library for quantitative trading signal generation and walk-forward backtesting.

Installation

pip install pyquaver

With optional extras:

# Interactive notebooks with yfinance data
pip install pyquaver[notebooks]

# Development tools (pytest, ruff, mypy)
pip install pyquaver[dev]

# Sphinx documentation
pip install pyquaver[docs]

Quick Start

Single-asset backtest

import pandas as pd
from quaver.backtest import run_backtest

# Load your OHLCV data (must have columns: ts, open, high, low, close, volume)
candles = pd.read_csv("data.csv")

result = run_backtest(
    engine_name="mean_reversion",
    parameters={"fast_period": 20, "slow_period": 50, "threshold": 0.02},
    candles=candles,
    instrument_id="AAPL",
    initial_capital=10_000.0,
)

print(result.summary())

Multi-asset pairs backtest

from quaver.backtest import run_multi_asset_backtest

results = run_multi_asset_backtest(
    engine_name="pairs_mean_reversion",
    parameters={
        "instrument_a": "AAPL",
        "instrument_b": "MSFT",
        "spread_window": 60,
        "entry_z": 2.0,
        "exit_z": 0.5,
    },
    candles_map={"AAPL": candles_aapl, "MSFT": candles_msft},
    initial_capital=10_000.0,
    allow_shorting=True,
)

for iid, r in results.items():
    print(f"{iid}: {r.summary()}")

Discover strategies

from quaver.strategies.registry import StrategyRegistry
import quaver.strategies  # auto-registers all built-in engines

print(StrategyRegistry.list_engines())
# ['mean_reversion', 'pairs_mean_reversion', 'regime_mean_reversion', 'vsa_stopping_volume']

Built-in Strategies

Strategy	Type	Description
mean_reversion	single	Dual moving-average mean reversion
regime_mean_reversion	single	Regime-based probabilistic mean reversion
vsa_stopping_volume	single	VSA stopping-volume reversal pattern
pairs_mean_reversion	multi	Statistical arbitrage pairs trading

Requirements

• Python >= 3.12
• numpy >= 2.4.2
• pandas >= 3.0.1

License

MIT