Charles Reid f73f2f7719 adding --link to run_jupyter
note that --link just adds a container to /etc/hosts
that means you can connect to mongodb on another container
named "inspiring_boyd" by having mongo connect to the database
at "inspiring_boyd" instead of "localhost".
This automatically resolves to the IP address of the mongo container.
2017-04-20 03:13:08 +00:00
2017-04-19 05:49:49 +00:00
2017-04-20 03:13:08 +00:00

Jupyter Notebook Docker Container

This repo contains scripts for building Docker containers that run Jupyter notebooks.

Jupyter + Mongo

Dockerfile

The Dockerfile uses aptitude to install the following components:

  • Python
  • Pip
  • Numpy
  • Scipy
  • iPython
  • Pandas
  • Sympy

Note that we need to install the python-scipy version of Scipy because running pip install scipy causes this machine to run out of memory.

The script then uses pip to install the following:

  • Pymongo

Building

To build the container, use the build_jupyter.sh script. This script does not require any configuration.

$ ./build_jupyter.sh

Running

To run the container, use the run_jupyter.sh script. This script links the Jupyter container to the Mongo container, and must be configured to point to the correct name of the MongoDB container.

The container exposes port 8888 to the outside world. Because it is linked to the MongoDB container, port 27017 (MongoDB) is also accessible from the Jupyter notebook.

$ ./run_jupyter.sh
Description
Files for creating a Jupyter notebook docker container.
Readme 30 KiB
Languages
Dockerfile 56.7%
Shell 43.3%