This is a definition file to illustrate the use of an Ubuntu22.04 container to create a virtual python environment and install with pip specific packages listed in a requirements.txt file located in the same folder as the Dockerfile

A container image built with cftime==1.6.4, netCDF4==1.7.1.post1 and xarray==2024.6.0 was automatically created on https://quay.io/repository/jeani/pipenv and can be pulled and used straight away (see below).

Those who want to install different packages have to rebuild the container as follows:

• clone this repository git clone https://quay.io/repository/jeani/pipenv
• modify the content of the requirements.txt file to suit their needs
• build the container with docker build --progress=plain -t pipenv -f Dockerfile .
• if needed save the container image docker save pipenv -o pipenv.tar
• and convert it to .sif: apptainer build pipenv.sif docker-archive://pipenv.tar

Assuming that your data and scripts are in the directory called my_data_and_scripts_on_the_host you have to make it visible from inside the container using -v (for Docker) or --bind (for Singularity/Apptainer) and provide a location not already in use (here I made up something nammed /opt/uio/my_data_and_scripts_in_the_container).

Pull the container image with:

docker pull quay.io/jeani/pipenv:latest

execute it with:

docker run -it --rm -v my_data_and_scripts_on_the_host:/opt/uip/my_data_and_scripts_in_the_container pipenv:latest

Pull the container image with:

apptainer pull docker://quay.io/jeani/pipenv:latest

execute it with:

apptainer shell --bind my_data_and_scripts_on_the_host:/opt/uio/my_data_and_scripts_in_the_container pipenv_latest.sif

then, once inside the container activate the environment and perform your data processing as usual:

• source /opt/uio/my_new_pythonenv/bin/activate
• change directory to /opt/uio/my_data_and_scripts_in_the_container and run for instance python run_my_script.py