what's this?

this is the simplest example of sharing notebooks with docker.

lets get started

get it

clone this repo with git clone https://github.com/mynameisvinn/Docker-for-Data-Scientists.

start docker daemon

do systemctl start docker.service

building da container

do docker build -t mynameisvinn/ds . which tells docker to build the container.

it does so by executing a sequence of commands specified by the dockerfile including (a) fetching a base python 3.6 layer; (b) copying requirements.txt and the notebooks folder into the container; (c) executing shell commands such as pip install, setting the working directory to home, exposing port 8888, and finally launching a notebook.

FROM python:3.6-slim

COPY requirements.txt requirements.txt
RUN pip install -r requirements.txt

COPY notebooks /home
WORKDIR /home

# Exposing ports
EXPOSE 8888

# Running jupyter notebook
# --NotebookApp.token ='demo' is the password
CMD ["jupyter", "notebook", "--no-browser", "--ip 0.0.0.0", "--allow-root", "--NotebookApp.token='demo'"]

running da container

now that we've built the container, we can run it.

we run with docker run -d -p 8887:8888 mynameisvinn/ds. this container launches a notebook at localhost:8887 (password: demo).

running it on ec2 - why not?

assuming youve enabled traffic to port 8887, you could do everything above in an ec2 instance, and then access the notebook from your local machine by pointing your browser to $EC2_IP:8887.

syncing files between local and container

we could also do docker run -d -p 8887:8888 -v /Users/vincent1/dropbox/temp/Docker-for-Data-Scientists/shared_data:/home mynameisvinn/ds, which maps local's shared_data folder with the container's home folder.

other helpful commands:

• ssh into container do docker run -it mynameisvinn/ds
• stop all containers. do docker stop $(docker ps -a -q).
• remove images. do docker rmi $(docker images -a -q). this will delete all images from local disk.
• removing dangling images do docker system prune will purge intermediate unnamed images. Depending on the sizes of containers you're using, this can save a considerable amount of disk space.

docker compose

alternatively, assuming the image can be pulled from dockerhub, you can run docker-compose up.