This repository contains all ingredients to run a Python Flask Web application that starts multiple Docker containers which in turn run a Jupyter Notebook server using SSL encryption and password protection (cf. http://ipython.org).

You can see the application in action under http://docker.quant-platform.com.

Note, however, that this app runs on the smallest DigitalOcean droplet (cf. https://www.digitalocean.com/?refcode=fbe512dd3dac). When setting up such a droplet it is recommended to use the latest version of Ubuntu in combination with Docker (cf. http://docker.com).

In addition, when setting up an environment make sure to have Python, Flask, Flask-WTF and Flask-SQLAlchemy installed. Best you install Anaconda first (cf. http://continuum.io/downloads) and do then:

sudo apt-get install python-pip
pip install flask-wtf
pip install flask-sqlalchemy

I assume that you have cloned the repository on an appropriate infrastructure in the root directory /root:

git clone --depth=1 https://github.com/yhilpisch/ipynb-docker

In order to run the Python Flask Web application on a server, you have to first build two Docker containers.

The first one is a basic Ubuntu container with some Python in it. Assuming you are in the repo folder do:

cd pythoncontainer
docker build -t pythoncontainer .

The second, which relies on the first, adds the Notebook Server functionality. Built it via

cd ../jupserver
docker build -t jupserver .

First some configurations. You need to adjust the URL to the server in the template file found under /root/ipynb-docker/serverapp/templates/server.html to reflect the correct IP address or the domain used. In that file replace the following link

<a href="https://docker.quant-platform.com:{{ port }}" target="_blank">link</a>

by a link containing your IP address

<a href="https://your-ip-address:{{ port }}" target="_blank">link</a>

When using a domain, make sure that both http & https are configured for the web server (e.g. nginx). The configuration (in the file /etc/nginx/sites-enabled/default) used for the demo under http://docker.quant-platform.com is:

server {
        listen 80;
        listen 443 ssl;
        ssl_certificate     /root/ipynb-docker/jupserver/jup_cert.pem;
        ssl_certificate_key /root/ipynb-docker/jupserver/jup_cert.pem;
        server_name docker.quant-platform.com jupyter.quant-platform.com;
        location / {
            proxy_pass http://docker.quant-platform.com:8888;
            proxy_set_header Host $host;
            proxy_buffering off;
        }
}

If not installed yet, install nginx via sudo apt-get install nginx.

A self-signed certificate for SSL encryption can be generated as follows:

openssl req -x509 -nodes -days 365 -newkey rsa:1024 -keyout jup_cert.pem -out jup_cert.pem

Make sure that this certificate is stored in the folder /root/ipynb-docker/jupserver (best to generate it there).

Now go the the directory of the app and start the app:

cd /root/ipynb-docker/serverapp
python run_app.py &

Your app should now be reachable under http://your-ip-address:8888.

© Dr. Yves J. Hilpisch | The Python Quants GmbH

The code of this repository is BSD licensed (cf. http://opensource.org/licenses/BSD-3-Clause).

The code in this repository comes with no representations or warranties, to the extent permitted by applicable law.

http://tpq.io | [email protected] | http://twitter.com/dyjh

Python Quant Platform | http://quant-platform.com

Derivatives Analytics with Python (Wiley Finance) | http://derivatives-analytics-with-python.com

Python for Finance (O'Reilly) | http://python-for-finance.com