Software Requirements
RPM-CRN is written in C++ for efficiency. The code has been written and tested extensively in a Linux/UNIX environment, has also been compiled and run on Windows using Code::Blocks and Linux Subsystem for Windows, but has not been extensively tested on Mac.
Running the model will require working at a unix/linux style command line interface.
There are a number of software requirements to run the model and visualise the results: - A C++ compiler, preferably GCC: the GNU Compiler Collection - The make utility to compile the model with our make files - A text editor or integrated development environment (e.g. Notepad++, gedit, vims, VSCode) - A Python distribution with Scipy, Numpy and Matplotlib libraries
Optional for Multiobjective Optimisation: - Dakota - QUESO
MAC
Reconsider your life choices. Apparently you can install VirtualBox on a MAC. So there's that.
Windows
This code has been developed both in Linux and Windows environments. When working in Windows, this usually involves setting up a Linux virtual environment, for which there are a couple of options.
- Linux on a Virtual Machine
If you do not already work in Linux or UNIX, then the easiest way to get started would be to use some virtualisation software such as VirtualBox, and installing a Linux debian-based distribution such as Ubuntu or Linux Mint as a virtual machine in VBox. There is a nice video explaining how to do this here.
- Windows Subsystem for Linux
Alternatively, you can use the Windows Subsystem for Linux (WSL) Instructions for installing Linux Subsystem for Windows can be found here.
We find that WSL integrates really nicely within Visual Studio Code, our preferred integrated development environment (IDE).
- Code::Blocks
Finally, if you prefer to continue using Windows and do not wish to dive into Linux, it is possible to get the model working using the Code::Blocks IDE with MinGW (Minimalist GNU for Windows) compilers. The pair are available to install together here. We have not tested RPM-CRN extensively in this environment but the examples here will all compile and run correctly from Code::Blocks.
Linux/UNIX
Once you have access to a Linux machine or Linux virtual machine (WSL or VBox as above), we will work in a terminal.
- Download code release
In a terminal, to get the code, you can download a copy:
HOME$ wget https://github.com/mdhurst1/Rocky-Profile-Model/archive/refs/heads/JOSS-Paper.zip
Unzip the repository:
HOME$ unzip JOSS-Paper.zip
- Clone the repository
Alternatively, you can clone the repository using git (particularly if you wish to do any development work). To install git (if you do not already have it: )
HOME$ sudo apt install git-all
The download a clone of the repository with:
HOME$ git clone https://github.com/mdhurst1/Rocky-Profile-Model.git
If your using VSCode, the git integration is fantastic and you can clone the repository this way too by going to View -> Comman Pallete and typing Git: Clone. It will ask for the repository URL as above.
Dakota with QUESO
The Dakota toolkit delivers flexible and extendable software for model optimisation and uncertainty quantification. The software package QUESO (Quantification of Uncertainty for Estimation, Simulation and Optimisation) for the solution of inverse statistical problems is wrapped within Dakota. RPM-CRN works with Dakota and QUESO to optimise model parameters to match observed shore platform topography and measured CRN concentrations, and quantify uncertainty, in order to derive a liely timeseries of rock coast development.
The following shell script contains the terminal commands to install Dakota, Queso and associated dependencies. Thanks to Geraldo Fiorini Neto for collating the shell script.
To run the script, place it in your $home directory, then to make it executable, run:
HOME$ chmod +x Dakota_Queso_Install.sh
Run the script with:
HOME$ ./Dakota_Queso_Install.sh
#!/bin/bash
# Dakota 6.11.0 with QUESO 0.57.1
# install lib dependencies
apt-get install -y gcc g++ gfortran cmake libboost-all-dev libblas-dev liblapack-dev libopenmpi-dev openmpi-bin gsl-bin libgsl-dev python perl cmake-curses-gui libboost-dev openmpi-doc xorg-dev libmotif-dev build-essential gfortran autotools-dev curl unzip git vim openssh-client openssh-server libgsl0-dev python-pip doxygen texlive-latex-extra
# Install python modules
pip install numpy scipy pandas mpmath==1.0.0 sympy==1.2 pyyaml
# Remove old version of Java and install newer
apt-get remove default-jre && apt-get install openjdk-8-jre default-jdk
# Download and compile Queso 0.57.1
mkdir -p /root/queso && cd /root/queso
wget https://github.com/libqueso/queso/releases/download/v0.57.1/queso-0.57.1.tar.gz
tar -zxvf queso-0.57.1.tar.gz
cd queso-0.57.1
./configure --prefix=/opt/queso/0.57.1/
make && make install
# Download and extract Dakota 6.11.0
mkdir -p /root/dakota && cd /root/dakota
wget https://dakota.sandia.gov/sites/default/files/distributions/public/dakota-6.11.0-release-public.src-UI.tar.gz
tar -zxvf dakota-6.11.0-release-public.src-UI.tar.gz
mkdir -p /root/dakota/build
# Set environment variables and run
set ENABLE_DAKOTA_DOCS:BOOL=TRUE
echo "export PATH=\${PATH}:/opt/dakota/6.11.0/bin" >> /etc/profile.d/dakota.sh
echo "export LD_LIBRARY_PATH=\${LD_LIBRARY_PATH}:/opt/dakota/6.11.0/lib" >> /etc/profile.d/dakota.sh
. /etc/profile.d/dakota.sh
echo "export PATH=\${PATH}:/opt/queso/0.57.1/bin" >> /etc/profile.d/queso.sh
echo "export LD_LIBRARY_PATH=\${LD_LIBRARY_PATH}:/opt/queso/0.57.1/lib" >> /etc/profile.d/queso.sh
. /etc/profile.d/queso.sh
# Configure Dakota
cd /root/dakota/build
cmake -DCMAKE_C_COMPILER=gcc -DCMAKE_CXX_COMPILER=g++ -DCMAKE_Fortran_COMPILER=gfortran -DCMAKE_C_FLAGS:STRING="-O2" -DCMAKE_CXX_FLAGS:STRING="-O2" -DCMAKE_Fortran_FLAGS:STRING="-O2" -DDAKOTA_HAVE_MPI:BOOL=TRUE -DDAKOTA_HAVE_GSL:BOOL=TRUE -DHAVE_QUESO:BOOL=TRUE -DBoost_NO_BOOST_CMAKE:BOOL=TRUE -DCMAKE_INSTALL_PREFIX=/opt/dakota/6.11.0/ -DENABLE_DAKOTA_DOCS:BOOL=TRUE ../dakota-6.11.0-release-public.src-UI/
# Compile and install Dakota
make -j 4 && make install