How to run a notebook as a stateless web app?

[Paul A]

The concept of collaboration while developing a notebook is great, but if one then wants to deploy the notebook in production it needs to be stateless, i.e. the actions of one users should not (by default) affect what the other users sees. 

Is there a workflow/nbextensions that help productionize jupyter notebooks as web apps?

[Peter Parente]

Check out the trio of dashboard projects in the jupyter incubator:

https://github.com/jupyter-incubator/dashboards/wiki

[Paul A]

Thank you. Tried installing locally. Got the dashboards and the widgets. Can't get the bindings to appear in jupyter. Anaconda 3.5.1 x64.

[Peter Parente]

> Can't get the bindings to appear in jupyter.

Can you explain what you mean by the bindings not appearing? If you mean none of the UI elements appear and the widget examples don't work, can you double-check that you followed the instructions for enabling the extensions after doing the pip install?

[Paul A]

The "Publish to local dashboards server" menu option does not appear after installing the bindings.

--
You received this message because you are subscribed to a topic in the Google Groups "Project Jupyter" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/jupyter/rnI9qlV9-V0/unsubscribe.
To unsubscribe from this group and all its topics, send an email to jupyter+u...@googlegroups.com.
To post to this group, send email to jup...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/jupyter/aeb70089-38d8-44c9-a968-75b8d0b39080%40googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

[Peter Parente]

Make sure you also installed the jupyter_cms extension *and enabled it*. I think we added that step to the README, but it's easy to miss. (Unfortunately, there's no way to have one extension enable another one today.)

Also, note that the https://github.com/jupyter-incubator/dashboards_setup/tree/master/docker_deploy repo has a recipe for bringing up all the necessary pieces easily with docker-compose. Even if you don't want to use docker, you can still refer to the Dockerfiles and docker-compose.yml there for the steps you have to take to install and configure everything properly.