Installation via Source

Note

Below are some of the main instructions to build goSPL from the git source tree. This approach is mainly for experienced users working on a Linux environment. It is highly recommended to use conda, for quick installation and for packages and dependencies updates.

Update System

apt-get update -qq
apt-get install -yq --no-install-recommends bash-completion build-essential
apt-get install -yq --no-install-recommends python3-minimal python3-dev python3-pip
apt-get install -yq --no-install-recommends python3-tk python3-dbg cmake
apt-get install -yq --no-install-recommends python3-setuptools wget gfortran
apt-get install -yq --no-install-recommends proj-bin

MPICH

mkdir /tmp/mpich-build
wget http://www.mpich.org/static/downloads/${MPICH_VERSION}/mpich-3.3.tar.gz
tar xvzf mpich-3.3.tar.gz
cd mpich-3.3
./configure --enable-fast=all,O3 --prefix=/opt/mpich
make -j4
make install
ldconfig
cd /tmp
rm -fr *

export MPI_DIR=/opt/mpich
export PATH=/opt/mpich/bin:$PATH

PETSc

mkdir /tmp/petsc-build
wget http://ftp.mcs.anl.gov/pub/petsc/release-snapshots/petsc-lite-3.13.5.tar.gz
tar zxf petsc-lite-3.13.5.tar.gz
cd petsc-3.13.5
./configure --with-debugging=0 --prefix=/opt/petsc
            --COPTFLAGS="-g -O3" --CXXOPTFLAGS="-g -O3" --FOPTFLAGS="-g -O3"
            --with-zlib=1
            --download-fblaslapack=1
            --download-ctetgen=1
            --download-triangle=1
            --download-hdf5=1
            --download-mumps=1
            --download-parmetis=1
            --download-eigen=1
            --download-metis=1
            --download-hypre=1
            --download-scalapack=1
            --download-pragmatic=1
            --useThreads=1
            --with-shared-libraries
            --with-cxx-dialect=C++11
make PETSC_DIR=/tmp/petsc-build/petsc-3.13.5 PETSC_ARCH=arch-linux-c-opt all
make PETSC_DIR=/tmp/petsc-build/petsc-3.13.5 PETSC_ARCH=arch-linux-c-opt install
make PETSC_DIR=/opt/petsc PETSC_ARCH="" check
cd /tmp
rm -fr *
export PETSC_DIR=/opt/petsc
export PATH=/opt/petsc/bin:$PATH

Dependencies

goSPL has many required dependencies. If a dependency is not installed, goSPL will raise an ImportError when the method/class requiring that dependency is called.

A dependency XXXX is installed via the following command in a terminal:

pip install XXXX

Dependency	Minimum Version	Notes
NumPy	1.22.2	Numerical computing tools.
SciPy	1.5.2	Optimization, linear algebra, integration, interpolation
Cython	0.29.21	Superset of the Python programming language
mpi4py	3.1.1	Bindings for the Message Passing Interface standard
petsc4py	3.13.0	Interface to PETSc libraries
h5py	2.10.0	Interface to the HDF5 binary data format
pandas	1.1.2	Data analysis and manipulation tool
ruamel.yaml	0.16.12	Parsing YAML to Python objects
vtk	9.0.3	Toolkit for 3D computer graphics and image processing
numpy-indexed	0.3.5	Functionality for indexed operations on numpy ndarrays
gflex	1.1.0	Compute flexural isostasy
pyshtools	4.10	Spherical-harmonic backend used by the flexure module
meson-python	0.15.0	Build backend

Note

meshio and pyproj are no longer required by goSPL. The flexure module switched from the older isoFlex wrapper to a pyshtools spherical-harmonic implementation; if your environment still lists isoFlex, it can be removed.

goSPL installation

Once all the listed dependencies above have been installed, goSPL source files are available through GitHub:

git clone https://github.com/Geodels/gospl

It can then be installed locally on your system using:

pip install --no-build-isolation -e .

The --no-build-isolation flag is required so that pip uses the dependencies already installed in your environment (notably meson-python, numpy, cython, mpi4py and petsc4py) rather than fetching a fresh build environment that may not see your local PETSc/MPI installation.