An early holiday gift from @ProjectJupyter and @mybinderteam: Introducing repo2docker, a lightweight command-line tool for building Docker images from Git repositories blog.jupyter.org/introducing-re…
tl;dr: repo2docker is a lightweight tool built in Python for:
1. inferring the environment needed to run a repository by utilizing common config files (e.g. `requirements.txt`)
2. Building a docker image that runs that environment
3. Running a jupyter server to let you explore
1. inferring the environment needed to run a repository by utilizing common config files (e.g. `requirements.txt`)
2. Building a docker image that runs that environment
3. Running a jupyter server to let you explore
It is workflow-, language-, and infrastructure-agnostic. It can be used to run lots of different user interfaces and can be used locally on your own computer, or remotely as a part of cloud deployments
The @mybinderteam has been using repo2docker within Kubernetes to handle all of its environment generation steps, then passing the images to a JupyterHub to serve user sessions. However you can use repo2docker *only* for building images suitable for a JupyterHub
You can also use it purely locally (provided you have access to Docker somewhere) in order to help you build reproducible images from your data science environments
If you'd like to get involved in repo2docker, @mybinderteam, or @ProjectJupyter please reach out! We're an open community and we'd love your input! Links in the article, but here are some good ones:
* r2d repo: github.com/jupyter/repo2d…
* Jupyter discourse: discourse.jupyter.org
* r2d repo: github.com/jupyter/repo2d…
* Jupyter discourse: discourse.jupyter.org
