QuantForge

📖 Overview

QuantForge is a high-performance option pricing library implemented in Rust with Python bindings via PyO3. It provides Black-Scholes based pricing, Greeks calculation, and implied volatility computation with Rust's performance while maintaining Python's ease of use.

📋 Features and Implementation

Option Pricing Models

QuantForge supports multiple option pricing models optimized for various asset classes:

• Black-Scholes: European options on stocks
• American Options: Early exercise options with Bjerksund-Stensland (2002) approximation
• Merton: Options on dividend-paying assets
• Black76: Commodity and futures options
• Asian Options (coming soon): Path-dependent options
• Spread Options (coming soon): Multi-asset options
• Garman-Kohlhagen (coming soon): FX options

Core Features

• ⚡ High Performance: Up to 1x faster than NumPy+SciPy, 1x faster than Pure Python
• 🎯 Machine Precision: erf-based implementation achieving <1e-15 accuracy
• 📊 Complete Greeks: Delta, Gamma, Vega, Theta, Rho plus model-specific Greeks (Dividend Rho, Early Exercise Boundary)
• 🔥 Implied Volatility: Newton-Raphson solver up to 170x faster than Pure Python
• 🚀 Auto-Parallelization: Automatic Rayon parallelization for batches >30,000 elements
• 📦 Zero-Copy Design: Direct NumPy array access eliminating memory copy overhead
• ✅ Robustness: 250+ golden master tests with comprehensive coverage
• 🔧 Production Ready: Input validation, edge case handling, Put-Call parity verified

📊 Performance Benchmark Results

Environment: Linux - 6 cores - 29.3GB RAM - Python 3.12.5 - 2025-09-12 12:47:56

Latest Benchmark Results

Benchmark data not found

Performance varies by environment. Values shown are medians of 5 runs. See benchmarks for details.

🔥 Implied Volatility Performance

Fair comparison using Newton-Raphson method (same algorithm and parameters):

Maximum speedup: 170x vs Pure Python, 45x vs NumPy

📥 Installation

pip install quantforge

From Source (Development)

# Clone repository
git clone https://github.com/drillan/quantforge.git
cd quantforge

# Install Rust (if needed)
curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh

# Build and install
pip install maturin
maturin develop --release

Development Dependencies

# Using uv (recommended)
uv sync --group dev

# Or standard pip
pip install -e ".[dev]"

🚀 Quick Start

Basic Usage

import numpy as np
from quantforge.models import black_scholes

# Single option calculation
spot = 100.0   # Current price
strike = 110.0 # Strike price
time = 1.0     # Time to maturity (years)
rate = 0.05    # Risk-free rate
sigma = 0.2    # Volatility

# Call option price
call_price = black_scholes.call_price(spot, strike, time, rate, sigma)
print(f"Call Price: ${call_price:.4f}")

# Put option price
put_price = black_scholes.put_price(spot, strike, time, rate, sigma)
print(f"Put Price: ${put_price:.4f}")

# All Greeks calculation
greeks = black_scholes.greeks(spot, strike, time, rate, sigma, is_call=True)
print(f"Delta: {greeks.delta:.4f}")
print(f"Gamma: {greeks.gamma:.4f}")
print(f"Vega: {greeks.vega:.4f}")
print(f"Theta: {greeks.theta:.4f}")
print(f"Rho: {greeks.rho:.4f}")

Batch Processing (High Performance)

import numpy as np
from quantforge.models import black_scholes

# Generate 1 million random data points
n = 1_000_000
spots = np.random.uniform(80, 120, n)      # Uniform distribution 80-120
strikes = np.full(n, 100.0)                # Fixed strike
times = np.random.uniform(0.1, 2.0, n)     # 0.1-2 years
rates = np.full(n, 0.05)                   # Fixed rate
sigmas = np.random.uniform(0.1, 0.4, n)    # 10-40% volatility

# Batch processing (~56ms for 1M elements)
prices = black_scholes.call_price_batch(spots, strikes, times, rates, sigmas)

# Batch Greeks calculation
greeks = black_scholes.greeks_batch(spots, strikes, times, rates, sigmas, 
                                    is_call=np.full(n, True))

Implied Volatility Calculation

from quantforge.models import black_scholes

# Calculate implied volatility from market price
market_price = 12.50
iv = black_scholes.implied_volatility(
    price=market_price,
    s=100.0,  # Current price
    k=110.0,  # Strike price
    t=1.0,    # Time to maturity
    r=0.05,   # Risk-free rate
    is_call=True
)
print(f"Implied Volatility: {iv:.2%}")

🔄 Parallelization Optimization

QuantForge automatically balances computation and overhead by applying parallelization based on data size:

import numpy as np
from quantforge.models import black_scholes

# Large data: Automatic parallelization (all cores)
large_spots = np.random.uniform(90, 110, 1_000_000)
large_prices = black_scholes.call_price_batch(large_spots, 100, 1.0, 0.05, 0.2)

# Small data: Single-threaded (avoid overhead)
small_spots = np.array([100, 105, 110])
small_prices = black_scholes.call_price_batch(small_spots, 100, 1.0, 0.05, 0.2)

📊 Benchmarks

Practical Scenario: Volatility Surface Construction (10×10 Grid)

Practical Scenario: 10,000 Option Portfolio Risk Calculation

See performance benchmarks for detailed results.

🏗️ Architecture

quantforge/
├── src/                    # Rust core implementation
│   ├── models/            # Pricing models (Black-Scholes, Black76, Merton, etc.)
│   ├── math/              # Mathematical functions (erf, norm_cdf, etc.)
│   ├── validation.rs      # Input validation
│   └── traits.rs          # Batch processing traits
│
├── python/                 # Python bindings
│   └── quantforge/        # Python package
│       └── models/        # Model-specific modules
│
└── tests/                  # Test suite
    ├── unit/              # Unit tests
    ├── integration/       # Integration tests
    ├── golden_master/     # Golden master tests
    └── performance/       # Benchmark tests

Technology Stack

• Rust 1.88+: Core computation engine
• PyO3: Python-Rust bindings
• Rayon: Data parallel processing
• NumPy: Array interface
• maturin: Build and packaging

📚 Documentation

• Official Documentation (English)
• API Reference
• Performance Guide
• Developer Guide
• Detailed Benchmarks

🧪 Testing

# Run Python tests (450+ test cases)
pytest tests/

# Run Rust tests
cargo test --release

# Measure coverage
pytest tests/ --cov=quantforge --cov-report=html

# Run benchmarks
pytest tests/performance/ -m benchmark

🤝 Contributing

Pull requests are welcome! For major changes, please open an issue first to discuss what you would like to change.

1. Fork the repository
2. Create your feature branch (git checkout -b feature/amazing-feature)
3. Commit your changes (git commit -m 'Add amazing feature')
4. Push to the branch (git push origin feature/amazing-feature)
5. Open a Pull Request

See Contributing Guide for details.

📄 License

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

🙏 Acknowledgments

• QuantLib for implementation and validation data
• Rayon project for high-speed parallel processing
• PyO3 project for Python-Rust bindings

📮 Contact

For questions or suggestions, please open an issue or join our discussions.