Install pipelines SDK in Jupyter images

[jlewi]

We should install the pipelines SDK in our Jupyter images so that pipelines can be used.

/assign @qimingj

[qimingj]

Basically we need to install KFP SDK on all images (probably https://github.com/kubeflow/kubeflow/tree/76107ff3aa7f9fde6fd7f15f2ec0a69a35874efa/components/tensorflow-notebook-image).

We are finalizing a few things in SDK and I will cut a new release later this week or early next week. Then we can start the integration.

What's the release cadence of these images?

[IronPan]

@jlewi could you confirm this the right place to pip install the DSL?
https://github.com/kubeflow/kubeflow/blob/76107ff3aa7f9fde6fd7f15f2ec0a69a35874efa/components/tensorflow-notebook-image/Dockerfile

The SDK release is currently part of pipeline deployment. So the release process would be

1. cut a pipeline release 2. update the jupyter dockerfile and ks registry 3. cut a kubeflow release

does that sounds right?

[abhi-g]

@qimingj we can cherry pick into the v0.3.4 patch release

[kunmingg]

@IronPan
here defines available images when start new notebook.
https://github.com/kubeflow/kubeflow/blob/master/kubeflow/jupyter/ui/default/config.yaml
Can edit it once post-submit of #2008 is done

[IronPan]

/assign @IronPan

[IronPan]

I saw that the new notebook images are created with tag v-base-b5bfe47-920
Just wait for the new release. Feel free to close this issue

[jlewi]

https://console.cloud.google.com/gcr/images/kubeflow-images-public/GLOBAL/tensorflow-1.11.0-notebook-cpu@sha256:d570b9351103056c2ea6b9509f27577c8a5f7a43803018852b28fb17f929f66a/details?tab=info