Python Extensions¶

The pyquantlib.extensions module contains pure Python implementations that extend QuantLib’s functionality. These demonstrate how to prototype new features without C++ compilation.

Smile Sections¶

SviSmileSection¶

class pyquantlib.extensions.SviSmileSection(time_to_expiry, forward, svi_params, validate=True)¶

Pure Python implementation of SVI smile section.

This class demonstrates extending QuantLib in Python without C++ compilation. It can be validated against ql.SviSmileSection from QuantLib’s experimental folder.

Parameters:

time_to_expiry (float) – Time to expiry in years
forward (float) – Forward price
svi_params (list[float]) – SVI parameters [a, b, sigma, rho, m]
validate (bool) – Whether to validate parameters (default True)

from pyquantlib.extensions import SviSmileSection

# Pure Python - same results as C++ ql.SviSmileSection
py_smile = SviSmileSection(1.0, 100.0, [0.04, 0.1, 0.3, -0.4, 0.0])
print(py_smile.volatility(100.0))

# Access parameters
print(py_smile.a, py_smile.b, py_smile.sigma, py_smile.rho, py_smile.m)

Helper Functions¶

pyquantlib.extensions.svi_total_variance(a, b, sigma, rho, m, k)¶: Compute SVI total variance: a + b * (rho * (k - m) + sqrt((k - m)^2 + sigma^2))

pyquantlib.extensions.check_svi_parameters(a, b, sigma, rho, m)¶: Validate SVI parameters for no-arbitrage conditions. Raises ValueError if invalid.

Pricing Engines¶

ModifiedKirkEngine¶

class pyquantlib.extensions.ModifiedKirkEngine(process1, process2, correlation)¶

Modified Kirk engine for spread option pricing.

This engine implements the Modified Kirk approximation which adds a volatility skew correction to the standard Kirk formula. The correction is derived using Malliavin calculus and significantly improves accuracy when the correlation between the two underlying assets is high (rho > 0.9).

The spread option payoff is: max(S1 - S2 - K, 0) for a call.

Parameters:

process1 (GeneralizedBlackScholesProcess) – Black-Scholes process for the first asset (S1)
process2 (GeneralizedBlackScholesProcess) – Black-Scholes process for the second asset (S2)
correlation (float) – Correlation between the two assets, in [-1, 1]

References:

Alos, E., & Leon, J.A. (2015). Quantitative Finance, 16(1), 31-42.
Kirk, E. (1995). “Correlation in the energy markets.” Managing Energy Price Risk.

from pyquantlib.extensions import ModifiedKirkEngine

engine = ModifiedKirkEngine(process1, process2, correlation=0.95)
option.setPricingEngine(engine)
print(f"NPV: {option.NPV():.4f}")

Static Methods¶

static ModifiedKirkEngine.kirk_volatility(F1, F2, K, sigma1, sigma2, rho)¶: Calculate Kirk’s approximation volatility (without skew correction).

static ModifiedKirkEngine.skew_slope(F1, F2, K, sigma1, sigma2, rho)¶: Calculate the skew slope correction term from Alos & Leon (2015).