lifelib-pyql 0.0.1

QuantLib Cython wrappers

lifelib-pyql

Cython-based Python bindings for the QuantLib C++ library.

This project is a fork of PyQL (originally developed by Enthought) that replaces SWIG wrappers with a more Pythonic API built on Cython. It provides comprehensive coverage of QuantLib's functionality, including term structures, instruments, pricing engines, short-rate and equity models, and more.

Portfolio Module

Beyond the core QuantLib bindings, lifelib-pyql is developing a portfolio submodule for processing a large number of instruments of the same type efficiently.

When working with thousands or millions of instruments, creating a separate Python object for each one is too slow. The portfolio submodule provides a set of new APIs that let you create a single Python object representing a large set of instruments of the same type, whose methods operate on all instruments in the set at once rather than one at a time.

Currently, the portfolio module includes:

• Schedules -- A vectorized collection of payment schedules. It bulk-generates QuantLib Schedule objects from arrays of parameters and stores the resulting dates in a compact 2D NumPy datetime64[D] array.

from lifelib_pyql.portfolio.api import Schedules
import numpy as np

schedules = Schedules(
    effective_dates=np.array(['2024-01-15', '2024-03-01'], dtype='datetime64[D]'),
    termination_dates=np.array(['2029-01-15', '2029-03-01'], dtype='datetime64[D]'),
    tenors=np.array([6, 3]),       # months
    max_size=20,
)

schedules.dates    # 2D datetime64[D] array, NaT-padded
schedules.size     # number of dates per schedule
schedules[0]       # returns a single Schedule object
schedules[0:1]     # returns a new Schedules slice

• FixedRateBonds -- A vectorized collection of fixed-rate bonds stored as parallel NumPy arrays. It bulk-stores bond parameters and reconstructs individual QuantLib FixedRateBond objects on demand via indexing.

from lifelib_pyql.portfolio.api import Schedules, FixedRateBonds
from lifelib_pyql.time.daycounters.actual_actual import ActualActual
from lifelib_pyql.time.businessdayconvention import Following
import numpy as np

bonds = FixedRateBonds(
    settlement_days=np.array([3, 3, 3]),
    face_amounts=np.array([100.0, 100.0, 100.0]),
    schedules=schedules,                        # a Schedules object
    coupons=np.array([0.05, 0.04, 0.06]),       # one coupon rate per bond
    accrual_day_counter=ActualActual(ActualActual.ISMA),
    payment_convention=Following,
    redemptions=100.0,                          # scalar broadcasts to all
    issue_dates=np.array(['2024-01-15', '2024-03-01', '2024-06-01'],
                         dtype='datetime64[D]'),
)

bonds[0]           # returns a single FixedRateBond object
bonds[0:2]         # returns a new FixedRateBonds slice
len(bonds)         # number of bonds in the collection
bonds.coupons      # 1D or 2D array of coupon rates

Prerequisites

• QuantLib 1.41 or higher (with Boost 1.78.0+)
• Cython 3.0 or higher
• Python 3.10+
• NumPy

Building

# Build Cython extensions in-place
make build

# Run the test suite
make tests

# Build documentation (requires Sphinx, nbsphinx)
make docs

See the Getting started guide for full details.

License

BSD 3-Clause. See LICENSE.txt. The original PyQL license is preserved in LICENSE_ORIGINAL_PYQL.txt.