Hints

Task 1 - The message

Hint #1

Click to reveal

Half of the nickname of Merry Brandybuck’s best friend will help you a lot!

Solution

Click to reveal

The nickname of the best friend of Merry Brandybuck is Pippin (sorry if you were not a Lord of the Rings fan).

So, the solution is to use the pip command to install the missing Python package. Two different ways to do this:

  • Create a new cell in your Notebook and use the !pip install stegano command to install the required Python packages.

  • Use the pip install stegano command in a terminal launched inside Jupyter.

Task 2 - The environment

Not really a hint, but before you lose too much time on digging this out, here is an info we got from a friend: quay.io/modh/odh-minimal-notebook-container@sha256:9592b9aed5248b77c7490f08004091174030035a76c3b198f9f01c8be0060074

You’re welcome!

Hint #1

Click to reveal

Think out of the box of the out-of-the-box Workbench images!

Hint #2

Click to reveal

If only there was something in OpenShift that can Build a new container image Config…​

Hint #3

Click to reveal

A basic Containerfile is enough to build a new container image. You just need to add what is missing…​

Solution

Click to reveal

You need to create a new container image that includes the missing Python package. To do this, a simple Containerfile is enough.

Here is an example you can use:

FROM quay.io/modh/odh-minimal-notebook-container@sha256:9592b9aed5248b77c7490f08004091174030035a76c3b198f9f01c8be0060074

RUN pip install stegano

Method 1

  • From the OpenShift Console, open an OpenShift command line terminal (click on the >_ logo at the top right of the console and start the Console).

  • Create the BuildConfig using the oc new-build command:

oc new-build --name=ai-mazing-race-stegano -D $'FROM quay.io/modh/odh-minimal-notebook-container@sha256:9592b9aed5248b77c7490f08004091174030035a76c3b198f9f01c8be0060074\nRUN pip install stegano'

  • You can watch the process of building the new container image by running the following command:

oc logs -f bc/ai-mazing-race-stegano

Method 2

  • From the OpenShift Console, create a new ImageStream called ai-mazing-race-stegano.

  • Create a new BuildConfig called ai-mazing-race-stegano and link it to the ImageStream. Here is the YAML configuration you can use:

kind: BuildConfig
apiVersion: build.openshift.io/v1
metadata:
  name: ai-mazing-race-stegano
  labels:
    build: ai-mazing-race-stegano
spec:
  output:
    to:
      kind: ImageStreamTag
      name: 'ai-mazing-race-stegano:latest'
  strategy:
    type: Docker
    dockerStrategy:
      from:
        kind: DockerImage
        name: 'odh-minimal-notebook-container@sha256:9592b9aed5248b77c7490f08004091174030035a76c3b198f9f01c8be0060074'
  source:
    type: Dockerfile
    dockerfile: |-
      FROM quay.io/modh/odh-minimal-notebook-container@sha256:9592b9aed5248b77c7490f08004091174030035a76c3b198f9f01c8be0060074
      RUN pip install stegano

  • Launch the build by clicking on Start Build in the Actions drop-down of the new BuildConfig.

Congratulations! You have created a new container image that includes the missing Python package. Take note of its address image-registry.openshift-image-registry.svc:5000/<your_project_name>/ai-mazing-race-stegano

Task 3 - Setting everything up

If you were an OpenShift AI admin, it would be easy to import the new container image into the OpenShift AI environment. But you are not. So you need to find a way to use this new container image for the government’s workbench…​

Hint #1

Click to reveal

Have you heard about the Notebook CR (Custom Resource) from OpenShift AI?

Hint #2

Click to reveal

You can create a new Notebook CR that uses the new container image you have created. Or just "hack" an existing one…​

Solution

Click to reveal

Behind the scene, when you create a Workbench, a Notebook CR is created in OpenShift. But you can also directly create a new Notebook CR that uses the new container image you have created. Or here, the easiest way would simply be to slightly modify the existing Notebook CR.

Recipe:

  • Stop your Workbench.

  • From the OpenShift Console, go to the Home→API Explorer.

  • Filter/find the Notebook object in the group kubeflow.org at version v1. Click on it.

  • Make sure you are on your project and click on the Instances tab.

  • Click on the Name of your Notebook (Workbench).

  • In the YAML tab, look for the field image: 'image-registry.openshift-image-registry.svc:5000/redhat-ods-applications…​

  • Replace the value of the image field by the ImageStream name you created. Normally it should be: image-registry.openshift-image-registry.svc:5000/<your_project_name>/ai-mazing-race-stegano.

  • Get back to OpenShift AI. Your Workbench will show that the Notebook image it is using is Deleted. That’s not exactly true, it’s just not available in the images that OpenShift AI knows about…​

  • Start your Workbench. It will use the new container image you have created. You can now directly run the cell! Without typing anything. Mission accomplished, the government sends its thanks!