https://github.com/vpavlin/jupyter-publish-extension
https://github.com/vpavlin/jupyter-publish-extension
Last synced: 2 months ago
JSON representation
- Host: GitHub
- URL: https://github.com/vpavlin/jupyter-publish-extension
- Owner: vpavlin
- License: gpl-3.0
- Created: 2018-05-31T09:02:26.000Z (almost 7 years ago)
- Default Branch: master
- Last Pushed: 2018-11-15T16:00:12.000Z (over 6 years ago)
- Last Synced: 2025-01-23T03:45:24.604Z (4 months ago)
- Language: JavaScript
- Size: 21.5 KB
- Stars: 1
- Watchers: 2
- Forks: 0
- Open Issues: 1
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# JupyterHub Publish Extension
This extension and service allow you to publish and share your notebooks in JupyterHub
Requirements:
* Using KubeSpawner (i.e. running on top of Kubernetes/OpenShift)
* Installing this repo through pip
* Installing & enabling `publish.js` as nbextension
* Edit your `jupyterhub_config.py`
All bellow is done automatically if you use Open Data Hub deployment of JupyterHub (see https://gitlab.com/opendatahub/jupyterhub-ansible)
## Install and run the publish service
Add git URL to your `requirements.txt` for JupyterHub
```
git+https://github.com/vpavlin/jupyter-publish-extension.git
```
And add the following to your `jupyterhub_config.py` to enable the service to run
```
import uuid
c.ConfigurableHTTPProxy.auth_token = str(uuid.uuid4())
public_service_dict = {
'PROXY_TOKEN': c.ConfigurableHTTPProxy.auth_token,
'PROXY_API_URL': 'http://%s:%d/' % ("127.0.0.1", 8082)
}
public_service_dict.update(os.environ)
c.JupyterHub.services = [
{
'name': 'public',
'command': ['bash', '-c', 'jupyter_publish_service'],
'environment': public_service_dict
}
]
```
You also need to extend the Jupyter single user server to run a nbviewer sidecar container
```
c.KubeSpawner.singleuser_extra_containers = [
{
"name": "nbviewer",
"image": "nbviewer:latest",
"ports": [
{
"containerPort": 9090,
"protocol": "TCP"
}
],
"env" : [
{
"name": "NBVIEWER_LOCALFILES",
"value": "/opt/app-root/src/public_notebooks"
},
{
"name": "NBVIEWER_TEMPLATES",
"value": "/opt/app-root/src"
},
{
"name": "NBVIEWER_PORT",
"value": "9090"
},
{
"name": "JUPYTERHUB_SERVICE_PREFIX",
"value": "/user/{username}/public/"
},
{
"name": "CACHE_EXPIRY_MIN",
"value": "30"
},
{
"name": "CACHE_EXPIRY_MAX",
"value": "60"
},
{
"name": "NO_CACHE",
"value": "true"
}
],
"volumeMounts": [
{
"mountPath": "/opt/app-root/src",
"name": "data"
}
]
}
]
```
The `nbviewer` image is based on https://github.com/vpavlin/nbviewer-quickstart which is designed to be built and run on top of OpenShift. To get the image available, run
```
oc apply -f https://raw.githubusercontent.com/vpavlin/nbviewer-quickstart/master/images.json
```
## Install the extension
```
[ -z "${PUBLISH_EXTENSION_VERSION}" ] && PUBLISH_EXTENSION_VERSION=master
jupyter nbextension install --sys-prefix https://raw.githubusercontent.com/vpavlin/jupyter-publish-extension/${PUBLISH_EXTENSION_VERSION}/publish.js
jupyter nbextension enable --sys-prefix publish
```