Insight Lane with Docker

Docker:

A basic Docker image has been created to run the project in a container, using the ContinuumIO miniconda3 base image (Python 3.6). The virtual environment 'crash-model' is installed and activated when the image is started via container, as well as an apache2 webserver via supervisord to serve the visualization.

You can download the latest stable image from D4D's Docker Hub repo by running the following command, from a machine with the Docker engine installed:

$ docker pull datafordemocracy/crash-model:latest

Automatic building of images from the project's GitHub project have been configured to run on every commit to a branch. To see all available tagged versions of the image and their date of creation, see https://hub.docker.com/r/datafordemocracy/crash-model/tags/

For testing purposes you can build the image yourself from the Dockerfile by running the following from within the project repo:

$ docker build --tag datafordemocracy/crash-model:[tag] .

Once you have the image, you can run it in a container. The project folder (/app) is intentionally empty within the image, so you'll also need the project repo from GitHub available on your local machine. To do this run:

$ docker run -d -p 8080:8080 --name bcm.local -v /local/path/to/project_repo:/app datafordemocracy/crash-model:[tag]

The arguments to this command perform the following:

1. -d detaches the container and runs it in the background (gives you your shell back)
2. -p 8080:8080 maps port 8080 from the container to 8080 on your local machine (required if you want to view the visualization via browser)
3. --name bcm.local names the container 'bcm.local' (or whatever value you specify)
4. -v /local/path/to/project_repo:/app mounts your local machine's copy of the project repo into /app in the container.

Once you have a running container, you can get a shell on it to run the pipeline, test scripts etc. by running:

$ docker exec -it bcm.local /bin/bash