jupyter-nbgallery

Overview

This package provides several Jupyter extensions to enable integration with nbgallery. More information is available here in the nbgallery github project.

Installation

This package contains UI extensions (nbextensions) for NBClassic only. The server extension was migrated to jupyterlab_nbgallery To install with pip:

pip install jupyter_nbgallery

As of version 2.0, the minimal set of extensions required for nbgallery integration will be automatically installed and enabled. This includes the post_hack server extension, the nbgallery environment registration UI extension, and the Gallery menu UI extension.

To enable the optional extensions:

# Enable all extensions
jupyter nbclassic-extension enable --sys-prefix --py jupyter_nbgallery

# Enable individual extensions
jupyter nbclassic-extension enable --sys-prefix --section tree jupyter_nbgallery/autodownload/autodownload
jupyter nbclassic-extension enable --sys-prefix --section notebook jupyter_nbgallery/easy_buttons/easy_buttons
jupyter nbclassic-extension enable --sys-prefix --section notebook jupyter_nbgallery/instrumentation/instrumentation

Jupyter configuration

Once the extension is installed, you will also need to modify your Jupyter configuration to interact with your nbgallery server.

First, your Jupyter instance needs to know about your nbgallery server in order to save notebooks, download preferences, etc. This is configured in nbconfig/common.json underneath one of Jupyter's configuration directories. You can use the included jupyter nbgallery script to set the location of your nbgallery, as well as the Jupyter client name that will appear in nbgallery's "Run in Jupyter" Environments page.

# Set nbgallery server location:
# jupyter nbgallery [--user|--system|--sys-prefix] configure url <nbgallery server address>
# Example:
jupyter nbgallery --sys-prefix configure url http://localhost:3000

# Set nbgallery environment client name
# Example:
jupyter nbgallery --sys-prefix configure client.name my-jupyter-instance

After configuration, nbconfig/common.json should look something like this (potentially with other unrelated sections):

{
  "nbgallery": {
    "url": "http://localhost:3000",
    "client": {
      "name": "my-jupyter-instance"
    }
  }
}

Second, in order for nbgallery's "Run in Jupyter" button to work, Jupyter must be configured to allow cross-site AJAX. Note these are security-relevant configuration settings. These can be set in jupyter_notebook_config.py:

c.NotebookApp.allow_origin = <URL of your nbgallery instance>`
c.NotebookApp.allow_credentials = True
c.NotebookApp.disable_check_xsrf = True

These can also be set on the command line when launching Jupyter; for example:

jupyter notebook --NotebookApp.allow_origin='http://localhost:3000' --NotebookApp.allow_credentials=True --NotebookApp.disable_check_xsrf=True

If you are launching jupyter lab instead of jupyter notebook, be aware that as of Lab 3.0, these are now ServerApp settings that can be set in jupyter_server_config.py (see the migration guide).

List of Extensions

UI Extensions (nbextensions)

Minimal Set

• environment registers this Jupyter instance with nbgallery so notebooks can be launched with the "Run in Jupyter" button. It also downloads preferences stored in nbgallery.
• gallery_menu adds the Gallery and Preferences menus to the Notebook interface. These are necessary for saving notebooks and preferences back to nbgallery.

Optional

• autodownload will download Starred and Recently Executed notebooks from nbgallery when you first visit the Jupyter tree page.
• easy_buttons adds buttons to the active cell for commonly-used actions.
• instrumentation reports cell execution status back to nbgallery for notebook health assessment.