This script packages a local dataset into a versioned Docker image and publishes it to a container registry.

package_dataset_image.sh

# Replace <DATASET_NAME> with the name of your dataset
DATASET="<DATASET_NAME>"  

# Start a temporary container to stage the data
docker run --detach --interactive --tty --name "${DATASET}_data_package" ubuntu:latest bash

# Copy local data into the container
docker cp data/. "${DATASET}_data_package":/data

# Snapshot container as an image
docker commit "${DATASET}_data_package" "islasgeci/${DATASET}_data:latest"

# Push image to registry
docker push "islasgeci/${DATASET}_data:latest"

# Remove temporary container
docker rm --force "${DATASET}_data_package"