This is a Jupyter extension that shows a list of notebooks that users can easily preview and copy for their own use.

The targeted application is a JupyterHub deployment, where it is useful to distribute a collection of curated examples or templates and make it possible for hub uses to quickly share notebooks.

A new "Shared Notebooks" page lists notebooks from some configured storage. For each notebook there are two buttons, "preview" and "use".

Clicking "preview" shows a static HTML version of the notebook.

Clicking "use" opens a dialog box to prompt user for a filename or filepath (relative to their home dir).

On the notebook toolbar, a new "Share this notebook" button (the "paper airplane" icon at right) submits the notebook to the list of "Shared Notebooks".

• /tree#shared is the "Shared Notebooks" tab on the user's home page
• /shared returns JSON that populates the contents of that tab
• /shared/preview?notebook_id=xpcs.ipynb shows a static HTML preview (similar to nbviewer)
• /shared/fetch?notebook_id=xpcs.ipynb&dest=my-xpcs.ipynb makes a copy of the notebook in the user's home directory, stripping out the example output
• /shared/submit?notebook_id=my-new-example.ipynb copies a clean notebook into a shared
• /shared/delete?notebook_id=my-new-example.ipynb delete a notebook to the storage

• nbconvert
• nbformat
• hdfs for the hdfs storage

python setup.py install

In addition to installing the nshared packages, the installation adds a server extension to the jupyter notebook config file:

c.NotebookApp.server_extensions.append('nbshared.handlers')

Mutilple configuration possible:

• Local
• hdfs

Set the location of the shared notebooks to be distributed by adding this line to the jupyter notebook config file:

c.SharedManager.storage_class = "nbshared.LocalStorage"
c.SharedManager.storage_options = {
    "path": "/path/to/directory/to/share"
}

The intention is that path is a globally-writable directory.

Set the location of the shared notebooks to be distributed by adding this line to the jupyter notebook config file:

c.SharedManager.storage_class = "nbshared.HdfsStorage"
c.SharedManager.storage_options = {
    "url": "http://NAMENODE_ADDRESS:PORT",
    "hdfsUser": None,
    "path": "/path/to/directory/to/share"
}

If hdfsUser is set to None, the user who launch the jupyter instance will be use.

This project is indebted to the nbexamples project and the nbgrader project, a related (and much more complex!) application.