2. Installation

While we aspire for the installation process for Aronnax to be as simple as pip install aronnax, it is not yet that easy.

2.1. Dependencies

Aronnax is tested with Python 2.7 and 3.6, and depends on several external libraries.

The Python components of Aronnax depend on

numpy
scipy
tools
future

The final two are required for Python 2/3 compatibility. We recommend installing these with your favourite package manager before installing Aronnax. While the previous dependencies will allow you to run the model, some of the functions for dealing with the output files also depend on

xarray
dask

Additionally, the Fortran core requires

make
gfortran >= 4.7.4
mpi

In particular, Aronnax will need the mpif90 command in order for it to compile its Fortran core. You will need to manually ensure that you have a working installation of each of these.

In addition to these dependencies, the automated tests also depend on pytest and matplotlib.

2.2. Installation instructions

  1. Clone the repository to a local directory
    • git clone --recursive https://github.com/edoddridge/aronnax.git
  2. Move into the base directory of the repository
    • cd aronnax
  3. Compile Hypre
    • move to the directory ‘lib/hypre/src’
      • cd lib/hypre/src
    • configure the Hypre installer
      • ./configure
    • compile Hypre. This will take a few minutes
      • make install
    • move back to root directory of the repository
      • cd ../../../
  4. install Aronnax
    • pip install -e ./

Aronnax is now installed and ready to use. To verify that everything is working, you may wish to run the test suite. Do this by executing pytest in the base directory of the repository. This requires that the pytest module is installed.

Note

Installing in HPC environments: If your cluster requires programs to be compiled on the compute cores, then you will need to perform step 3 on the compute cores.