Docker Cheat Sheet

comments

Docker is an increasingly popular tool designed to make it easier to create, deploy and run applications within a container. We recently published an article – Data Scientist guide for getting started with Docker – which hopefully laid out some of the basics. As we have done in the past with SQL, Python Regular Expressions and many others, we thought it would be useful to have a centralized cheat sheet for Docker commands, which we’ve based on Docker’s official cheat sheet:

Install

First off, you’ll need to head over to the Docker site to install a version of the software.

To ensure it’s been installed correctly, open the command line and type docker version. This should display something like the below:

Build

docker build [OPTIONS] PATH | URL | -

Builds an image from a Dockerfile

docker images

Lists all of the images that are locally stored in the docker engine

docker rmi IMAGE_NAME

Removes one or more images

Ship

docker pull IMAGE_NAME[:TAG]

Pulls an image or a repository from a registry

docker push IMAGE_NAME[:TAG]

Pushes an image or a repository to a registry

docker tag SOURCE_IMAGE[:TAG] TARGET_IMAGE[:TAG]

Creates a tag TARGET_IMAGE that refers to SOURCE_IMAGE

docker login [OPTIONS] [SERVER]

Logs into a docker registry

docker logout [SERVER]

Logs out of a docker registry

Run

docker create [OPTIONS] IMAGE_NAME

Creates a new container. This is the same as docker run, but the container is never started. See docker run for options

docker run [OPTIONS] IMAGE_NAME

Run a command in a new container

docker start CONTAINER_NAME

Start one or more containers

docker stop CONTAINER_NAME

Stop one or more running containers

docker kill CONTAINER_NAME

Kill one or more running containers

docker ps

List all containers

docker rm CONTAINER_NAME

Remove one or more containers

Data Science Examples

FROM ubuntu

RUN apt-get install python3

This Dockerfile would install python3 on top of Ubuntu layer. Dockerfiles are text files that define the environment inside the container

 
FROM ubuntu:latest 
RUN apt-get update && apt-get install -y python3 \ python3-pip 
RUN pip3 install jupyter
WORKDIR /home/jupyter
COPY  /src/jupyter ./
EXPOSE 8000
ENTRYPOINT ["jupyter", "notebook", "--ip=*"]

This Dockerfile would:
- install python3 on top of Ubuntu layer
- create a /home/jupyter directory on the container
- copy in contents from the /src/jupyter folder on the user's machine
- Expose port 8000
- Run Jupyter notebook

docker pull tensorflow/tensorflow

Pull a latest TensorFlow image down

docker run -it -p 8000:8000 tensorflow/tensorflow

Run TensorFlow image in a container on port 8000

Related:

• Data Scientist guide for getting started with Docker
  
• Torus for Docker-First Data Science
  
• Setting up your AI Dev Environment in 5 Minutes