MeshKernel is a library for creating and editing meshes. It supports 1D & 2D unstructured meshes as well as curvilinear meshes.
Wrappers for the following languages are available:
The library is separated in an API namespace (MeshKernelApi) used for communication with the client and a backend namespace (MeshKernel) where the algorithms are implemented. The API namespace contains several structures used as parameters for the API methods (see API usage section). These structures must be mirrored in the client application and filled with appropriate values.
- Creating a triangular mesh inside a polygon
In this example a mesh is created by discretizing the polygon perimeter with the desired edge length.
- Mesh orthogonalization
Finite volume staggered flow solvers require the mesh to be as much orthogonal as possible. MeshKernel provides an algorithm to adapt the mesh and achieve a good balance between mesh orthogonality and smoothness.
- Curvilinear mesh generation
Curvilinear meshes for rivers can be generated using splines.
- Mesh refinement
A mesh can be refined in areas based on samples or polygon selections.
- libgomp
The requirements are:
- Git
- CMake 3.23 or higher
- A C 20 compatible compiler. Currently the following compilers are supported:
- Windows: Microsoft Visual C
- Linux: GCC
- macOS: GCC. Clang is not yet supported. Care must be taken if gcc and g are symbolic links to clang and clang
- The Boost libraries
- NetCDF
- Doxygen (optional)
- Under Windows, precompiled boost binaries (with MSVC compiler) can be downloaded here. Alternatively, the source code is available here and the installation instructions can be found here. Once installed, set the boost environment variable
BOOST_INCLUDEDIR
to the appropriate path. For example, if version 1.74.0 in installed inC:\Apps
, setBOOST_INCLUDEDIR=C:\Apps\boost_1_74_0
. - Under Linux, Boost can be either obtained from the package repository of the used Linux distribution or built from source following these instructions.
- Under macOS, Boost can be installed by following these instructions.
The NetCDF static library is required for building MeshKernel. Optionally, a PowerShell script is made available for cloning and building NetCDF and its dependencies (HDF5, ZLIB, Curl, and m4). It can be used under Windows, Linux, and macOS. The script is compatible with Powershell version 7 and higher. If necessary, Windows users need to upgrade to from version 5 to 7. The procedure is described here.
To run the script in a PowerShell session, use
.\install_netcdf_static.ps1 -WorkDir '/path/to/work/directory' -InstallDir '/path/to/install/directory' -BuildType 'Release' -ParallelJobs 10 -GitTags @{zlib = 'v1.2.13'; curl = 'curl-7_88_1'; hdf5 = 'hdf5-1_14_0'; netcdf_c = 'v4.9.1'}
with /path/to/work/directory
and /path/to/install/directory
replaced with valid paths.
For more information regarding the script's options above, use
Get-Help .\install_netcdf_static.ps1 -Detailed
Upon successful installation, to build MeshKernel successfully, it is important to either
- add the path to the install directory to the system path, or
- configure the MeshKernel build with
-DCMAKE_PREFIX_PATH=/path/to/install/directory
.
Note: Additional dependencies may be required depending on the system configuration:
- Windows: Perl
- Linux and macOS:
- m4, OpenSSL, Curl (Provided by the used system)
- Powershell (Linux instructions, macOS instructions).
Follow the steps below to configure, build and install MeshKernel.
-
To configure under Windows with Visual Studio, a solution is generated using
cmake -S <path-to-source-dir> -B <path-to-build-dir> -G "Visual Studio 16 2019" --install-prefix <path-to-install-dir>
This example uses Visual Studio 16 2019. A different version can be specified. In the above
<path-to-source-dir>
is the path to the MeshKernel source tree (the top-level directory containing source files provided by the project).<path-to-build-dir>
is the path to the directory where MeshKernel is to be built.<path-to-install-dir>
is the path top the directory where MeshKernel is to be installed.
Under Linux and macOS , the generator option can be omitted or the following can be used:
cmake -S <path-to-source-dir> -B <path-to-build-dir> -G "Unix Makefiles" --install-prefix <path-to-install-dir>
-
To build the project's targets, use:
cmake --build <path-to-build-dir> --config <cfg> --parallel <jobs>
where
<cfg>
is the build type (Debug
,Release
,RelWithDebInfo
andMinSizeRel
), seeCMAKE_BUILD_TYPE
.<jobs>
is the maximum number of concurrent processes to use when building.
Note: this builds the documentation by default in
<path-to-build-dir>/docs/html
. -
To install, use:
cmake --install <path-to-build-dir>
or
cmake --install <path-to-build-dir> --prefix [<path-to-install-dir>]
to override the installation path specified during configuration (step 1).
MeshKernel can be configured with a set of options, which are summarized in the table below.
Option | Description | Type | Default | Notes |
---|---|---|---|---|
ENABLE_UNIT_TESTING | Enables building the unit test executables | Bool | ON | |
ENABLE_BENCHMARKING | Enables building the benchmark executable | Bool | OFF | |
ENABLE_BENCHMARKING_MEM_REPORT | Enables reporting memory usage statistics | Bool | OFF | Applicable only when ENABLE_BENCHMARKING is ON, ignored otherwise |
ENABLE_CODE_COVERAGE | Generates code coverage report | Bool | OFF | Valid only under Linux and macOS when a GNU compiler is used (requires gcov), ignored otherwise |
HAVE_SRC_LOC_IN_ERR_MSGS | Includes the source location information in custom exceptions | Bool | OFF |