Required Packages

Install and load the latest versions of the following packages, preferably through Spack. Read more about how to install Spack, and how to install and load packages using Spack here.

Packages
boost	cmake	py-pytz
py-numpy	py-setuptools-scm	py-kiwisolver
py-python-dateutil	pkgconf	py-numexpr
py-setuptools	py-et-xmlfile	py-matplotlib
py-bottleneck	py-argparse	py-jdcal
py-pyparsing	py-cython	py-pandas
py-subprocess32	py-cycler	py-openpyxl
py-six	mpich	python

Note: You may skip MPI installation if you are running on an HPC cluster system with built-in MPI distribution (e.g. CrayPE, ibrun, Intel MPI etc.) and simply load that into your environment. Similarly, you can use internal Python distribution and packages if already available and skip installation.

Example Package Install

Install the py-numpy and py-pandas packages dependencies using Spack like this:

# Spack automatically installs the dependecies

# install py-numpy with gcc@<version> compiler
$ spack install py-numpy%gcc@<version>
# load the package along with all its dependencies
$ spack load -r py-numpy

# install py-pandas with gcc@<version> compiler
$ spack install py-pandas%gcc@<version>
# load the package along with all its dependencies
$ spack load -r py-pandas

# install cmake with gcc@<version> compiler
$ spack install cmake%gcc@<version>
# load the package along with all its dependencies
$ spack load -r cmake