Integration in C++ projects

Table of Contents

• CMake
  • FetchContent
  • find_package
  • add_subdirectory

nda is a compiled library.

You can either

• build and install it beforehand (see Installation) and then link against it (see find_package) or
• let CMake fetch and build it directly as part of your project (see FetchContent).

To use nda in your own C++ code, you simply have to include the relevant header files. For example:

#include <nda/nda.hpp> // core library incl. linear algebra functionalities
#include <nda/mpi.hpp> // MPI support (optional)
#include <nda/h5.hpp>  // HDF5 support (optional)
#include <nda/sym_grp.hpp> // Symmetry support (optional)
 
// use nda

CMake

FetchContent

If you use CMake to build your source code, you can fetch the h5 directly from the Github repository using CMake's FetchContent module:

cmake_minimum_required(VERSION 3.20)
project(my_project CXX)
 
# fetch from github
include (FetchContent)
FetchContent_Declare(
  nda
  GIT_REPOSITORY https://github.com/TRIQS/nda.git
  GIT_TAG        1.3.x
)
FetchContent_MakeAvailable(nda)
 
# declare a target and link to nda
add_executable(my_executable main.cpp)
target_link_libraries(my_executable nda::nda_c)

This will link automatically to all of nda's dependencies, except for the HDF5 C library. If you need to use some of the HDF5 C library features, you can simply link to it via h5::hdf5.

Note that the above will also build goolgetest and the unit tests for nda. To disable this, you can put set(Build_Tests OFF CACHE BOOL "" FORCE) before fetching the content or by specifying -DBuild_Tests=OFF on the command line.

find_package

If you have already installed nda on your system by following the instructions from the Installation page, you can also make use of CMake's find_package command. This has the advantage that you don't need to download anything, i.e. no internet connection is required. Furthermore, you only need to build the library once and can use it in multiple independent projects.

Let's assume that nda has been installed to path_to_install_dir. Then linking your project to nda with CMake is as easy as

cmake_minimum_required(VERSION 3.20)
project(my_project CXX)
 
# find nda
find_package(nda REQUIRED CONFIG)
 
# declare a target and link to nda
add_executable(my_executable main.cpp)
target_link_libraries(my_executable nda::nda_c)

In case, CMake cannot find the package, you might have to tell it where to look for the nda-config.cmake file by setting the variable nda_DIR to path_to_install_dir/lib/cmake/nda or by sourcing the provided ndavars.sh before running CMake:

$ source path_to_install_dir/share/nda/ndavars.sh

add_subdirectory

You can also integrate nda into your CMake project by placing the entire source tree in a subdirectory and call add_subdirectory():

cmake_minimum_required(VERSION 3.20)
project(my_project CXX)
 
# add nda subdirectory
add_subdirectory(deps/nda)
 
# declare a target and link to nda
add_executable(my_executable main.cpp)
target_link_libraries(my_executable nda::nda_c)

Here, it is assumed that the nda source tree is in a subdirectory deps/nda relative to your CMakeLists.txt file.