LOSC C++ library

Introduction

The LOSC library provides a C++ interface to perform calculations of LOSC, including the construction of LOSC curvature matrix, LOSC localization, local occupation number and LOSC corrections.

The provided C++ interface (C++ header files) includes

Header	Description
<losc/curvature.hpp>	for the LOSC curvature matrix.
<losc/localization.hpp>	for the LOSC localization.
<losc/local_occupation.hpp>	for the LOSC local occupation.
<losc/correction.hpp>	for the LOSC corrections.
<losc/exception.hpp>	for the exceptions that may be thrown in LOSC library.
<losc/eigen_def.hpp>	for the declarations of matrix used in LOSC library.
<losc/losc.hpp>	A convenient header that includes everything you need.

In order to use the C++ interface of the LOSC library, following the steps of installation. Then all you need to do is to include the corresponding headers in your project to use the LOSC C++ library.

Data Structure

The main data structure in the LOSC C++ library is the representation of matrix/vector objects. In the LOSC C++ library, we use the popular Eigen library, which lets us achieve the manipulation of matrices and vectors easily. Particularly, we use Eigen::Matrix to represent matrices, and Eigen::Vector to represent vectors.

Being different to the default behavior in Eigen that the storage of matrix is column-wise (Fortran-style), the storage of matrix in the LOSC C++ library is row-wise (C-style). To avoid declaring the row-wise Eigen::Matrix in a cumbersome way being like

Eigen::Matrix<double, Eigen::Dynamic, Eigen::Dynamic, Eigen::RowMajor>;

an alias LOSCMatrix is defined in <losc/eigen_def.hpp> as

using LOSCMatrix = Eigen::Matrix<double, Eigen::Dynamic, Eigen::Dynamic, Eigen::RowMajor>;

There is a similar alias LOSCVector for the vector objects. For other alias of Eigen types that are used the LOSC C++ library, please refer to <losc/eigen_def.hpp>.

Detailed References for the API

Please refer this section for the full documentation of C++ interface of the LOSC library.