Home

Welcome to the skimage-docker wiki!

The following text should help you setup skimage-docker on your machine.

About

In short, this setup listens for code execution requests, processes and returns the output. To do this it spins up Docker containers in the back-end, all execution takes place within it and the output is returned in a legible format to the front-end which requested it in the first place.

It was born as a result of a SoC project to setup an interactive gallery for examples @ scikit-image, the Image Processing library for Python.

This is live here @ interactive gallery for scikit-image examples

Setting it up

Get the latest copy on your server - git clone https://github.com/scikit-image/skimage-docker/

Prerequisites

• Docker
• Flask - the webserver using which we expose the service
• docker-py - Python library for interacting with the Docker daemon

Docker works on the concept of images and containers and a Dockerfile is what is usually used to build an image

The Dockerfile provided with the repository helps one build an image which caters to running code which uses the scikit-image library. Basically installs all the dependencies on top of the base image. This started as an attempt to setup an interactive gallery for examples @ scikit-image, remember?

Building the image - docker build -t docker-skimage:1.0 .

This has to be run when present in the root directory of the repository, since the Dockerfile is present there and that is what this command uses

After a successful build, - docker images should list 'docker-skimage'

Start the server - python start.py

This also prints the current URL at which it is running. All left to do now is set up the front-end which interacts with this setup.

Sample Demo

Here is a very minimalist demo.

I'll try to explain some of the interactions taking place here. Obviously a lot more can be achieved as desired. See the interactive gallery ;)