Citywide Data Science JupyterHub

This repository contains Docker images and Kubernetes configuration for the Citywide Data Science and Predictive Analytics JupyterHub deployment.

The Docker image can be found in image. It can be build using make build, and published to Dockerhub using make publish.

The Kubernetes configuration can be found in deploy. It is based on the Zero-to-JupyterHub guide. It can be deployed using make upgrade.

Setup

• To setup on WSL, you'll need the deploy/secrets.yaml file and the ~/.kube/config.

• Install kubectl using conda install -c conda-forge kubernetes

• You'll need to also be running the latest version of AWS CLI, or at least > 1.16.

• Then, create a profile for Kubernets inside your ~/.aws/credentials

• You'll need to set the profile key in the kubectl config under user:exec: as

env:
        # - name: AWS_PROFILE
        #   value: "<aws-profile>"

• For the secrets file, we use git crypt. Make sure you have the citywide-jhub.key file and then run git-crypt unlock ~/path/to/key in the base of the repo.

• From there, you should be able to run something like kubectl get svc.

• Finally, you'll need Helm, the package manager for Kubernetes. To install helm, follow the steps from Zero to JupyterHub.