dkavolis/polynomials

Python wrappers for boost polynomials

polynomials

[WIP] Polynomial python wrappers for boost polynomials.

Description

Exposed polynomials:

• Legendre
• Laguerre
• Hermite
• Chebyshev
• LegendreStieltjes

Every polynomials has additional wrappers for:

• <Polynomial>Sequence - polynomial value generator at a specified abscissa value x up to a specified order
• <Polynomial>Series - dense 1D polynomial series with mutable coefficients
• <Polynomial>Product - N-dimensional polynomial product
• <Polynomial>ProductSet - list of N-dimensional weighed polynomial products

Python only wrappers:

• <Polynomial>PCE - polynomial chaos expansion fitting and prediction, can be used with any sklearn linear model though sklearn.linear_model.Lasso works best
• AdaptivePCE - adaptive version of polynomial chaos expansion using difference of Sobol' indices or variances between successive fits as convergence metric

Building

Cmake, pybind11 and boost are required for building. Tested with pybind11 2.4.3 and boost 1.73.0 using MSVC 19.27.28826.0 but should compile with GCC and Clang.

To build:

python setup.py build

Additional cmake options can be passed as

python setup.py build --cmake-options="-DCMAKE_CXX_COMPILER=/usr/bin/g++"

or in setup.user.cfg:

[build_ext]
vcpkg_dir = ~/vcpkg
vcpkg_triplet =
vcpkg_manifest = true
cxx_compiler = /usr/bin/g++
cmake_generator = Ninja
cmake_options =

Installation

python setup.py install --user

Note

This project has been set up using PyScaffold 3.2.3. For details and usage information on PyScaffold see https://pyscaffold.org/.