PyMaple 2.1

Maple Container Utility

Introduction

Maple is a Python API and CLI that acts as a wrapper around docker/singularity to implement containerization of HPC applications and their dependencies.

Tutorial

Installation

mkdir -p $HOME/.local/bin
export PATH="$PATH:$HOME/.local/bin"
./setup develop && ./setup install

Writing a Maplefile

Maplefile is used to define environment variables required by maple. Following is a list of variables:

base: Name of the base image

container: Name of the local container

target: Name of the target dir to mount source dir

backend: Backend (docker/singularity)

maple passes these variables to its internal Dockerfile to build the images and containers.

CLI use:

• Building an image

  maple image build <image>: Build local image from remote image

• Getting shell access:

  maple container pour --image=<image>: Pour local image into a container

  maple container shell: Provides shell access to the container

  maple container commit --image=<image>: Save changes from local container to local image (only available with docker backend)

  maple container rinse: This commands stops and deletes the local container (only available with docker backend)

  maple image squash --image=<image>: Prune redundant layers from a local image (do this to reduce size of an image after maple container commit, only available with docker backend)

• Launch an ipython notebook inside the

  maple container notebook --image=<image> --port=<port>: launches the notebook server

• Run commands inside the container

  maple container run --image=<image> "echo Hello World!": example to launch specific command inside the container, use –comit to save changes to the image

• Cleanup

  maple container rinse <container1> <container2> <container3>: deletes containers

  maple image delete <image1> <image2> <image3>: deletes images

• Remote interface maple pull <image> and maple push <image>

Examples

See examples/create examples/bubblebox examples/flashsim