Installation¶
There are two ways of installing pyPDAF.
Conda¶
The easiest approach is using conda
. Currently, pyPDAF
is available from conda
for Windows
, Linux
and MacOS
. The installation can be obtained via:
conda create -n pypdaf -c conda-forge yumengch::pypdaf
After installation, pyPDAF
can be used by activating the conda environment conda activate pypdaf
.
Source code¶
In some cases, it might be desirable to compile pyPDAF using your preferred compilers and environment. In this case, pyPDAF can be installed from source
git clone https://github.com/yumengch/pyPDAF.git
cd pyPDAF
git submodule update --init --recursive
pip install -v .
The installation congifuration is specified in setup.cfg
. Example setup configurations are provided in PDAFBuild directory.
Each entry of the configuration file is listed here:
pwd
: This is the path to the current directory of the repository, e.g./home/users/xxxx/pyPDAF
PDAF_dir
: This is the absolute, or relative (topwd
) path where PDAF directory is located. One do not need to change this option.cmake_config_path
is the path to the.cmake
file for PDAF compilation. This file contains CMake configurations for PDAF. There are example configurations in thePDAFBuild
directorycondaBuild
is a switch for building a conda package. If the purpose of compiling is not building a conda package, this option should always beFalse
.c_compiler
andfortran_compiler
specifies the compiler being used.c_compiler
can be:gcc
,msvc
,icc
,clang
andfortran_compiler
can be:gfortran
andifort
. Thefortran_compiler
should be consistent with the compiler used in cmake configuration file.useMKL
decides if you use Intel’s Math Kernel Library (MKL). IfTrue
is given,MKLROOT
must be specified which is the absolute path to the static MKL libraryLAPACK_PATH
andLAPACK_Flag
is the path to the BLAS and LAPACK directory and the linking flag respectively. They can be delimited by,
. For example, we can haveLAPACK_Flag=blas,lapack
. Do not give-lblas
assetuptools
deal with the format to the linker.MPI_LIB_PATH
is the path to the MPI library. In windows, this path is usuallyC:\Program Files (x86)\Microsoft SDKs\MPI\Lib\x64
.