daukantas / beaker

A collaborative platform for rapid and reproducible research.

Beaker

Beaker is a collaborative platform for rapid and reproducible research.

Getting Started

1. Create an account at beaker.org and follow the instructions in your account settings. If you are an AI2 employee please use Google OAuth and your AI2 account.

   These instructions will guide you through installing and configuring the Beaker CLI. See below for more options.

   Request credentials to run an experiment from a Beaker admin. You can either ask on #beaker-users or email [email protected]. Please include the email address associated with your Beaker account.

2. Run your first experiment. The following example counts words in the text of Moby Dick.

   cat > count.yaml << EOF
   tasks:
   - spec:
       image: examples/wordcount
       resultPath: /output
       datasetMounts:
       - datasetId: examples/moby
         containerPath: /input
   EOF

   beaker experiment create -f count.yaml

3. For more in-depth documentation, see https://github.com/beaker/docs.

Installation

Prerequisites

Docker is the foundation for Beaker experiments and must be installed to take full advantage of Beaker.

Install Beaker CLI

The most direct way to install Beaker is to download a release and extract it to your path.

tar -xvzf beaker_*.tar.gz -C /usr/local/bin

OS X users can install Beaker through Homebrew with a custom tap.

brew tap allenai/homebrew-beaker https://github.com/allenai/homebrew-beaker.git
brew install beaker

Beaker can also be installed from source using standard Go tools.

go get -u github.com/allenai/beaker/...

Notices

Beaker dependencies and licenses