Renkulab Docker Images contain minimal dependencies for launching interactive environments like Jupyterlab and RStudio from the Renku platform. They also each contain a version of the renku cli.
The basic python (py3.7) and basic R (r) images are available via
renku project templates
that you select upon renku project creation on the renkulab platform, or locally
via renku init. If you would like to use an image built from this repo that is
not available via the renku project templates, follow these steps:
- create your project with the minimal template that matches the programming language you're using (there's also a minimal template for languages other than python or R)
- replace the image in the
FROMline in your project's Dockerfile with one of the images built from this repo. See "Naming Conventions" for how to choose which image to use
Images in https://hub.docker.com/r/renku/renkulab/tags follow the following naming conventions to help you know which image to choose for your renku project:
renku/renkulab:renku[renku-python version]-[image flavor][image flavor version]-[tag|hash]
For example:
renku/renkulab:renku0.10.3-py3.7-0.6.2
renku/renkulab: indicates this is an image you can use to spawn an environment from your project on Renkulab.renku[renku version](e.g.renku0.10.3): indicates the version of the renku CLI installed in the image. Note: if no version is present, it's the latest available development version.[image flavor][image flavor version](e.g.r3.6.3): image flavor refers either to the programming language installed in the environment, or the base image that extra dependencies are added to. See below for details about the available flavors.[tag|hash](e.g.0.6.2): the tag is a value given to a commit of the repository and indicates that the version is part of a release. If the version is not part of a release, this value is the first few chars of the git commit SHA from which the image is built.
Available via renku project templates
The basic Jupyter image with minimal dependencies. Based on https://hub.docker.com/r/jupyterhub/singleuser.
Currently with python 3.7.
Available via renku project templates
Based on the rocker docker image: https://github.com/rocker-org/binder,
chosen because rocker keeps a more up-to-date version of R than conda.
Includes the R Jupyter kernel as well as RStudio. To access RStudio,
simply replace /lab or /tree with /rstudio in the URL.
Several versions of R are available, including the latest 3.6.3.
Based on the bioconductor docker image: https://github.com/Bioconductor/bioconductor_docker.
Currently with bioconductor 3_10 & devel.
Based on the py3.7 image with CUDA 9.2 installed.
Based on the py3.7 with CUDA 10.0 and Tensorflow 1.14 installed.
The basic Jupyter image with Julia installed. Based on https://hub.docker.com/r/jupyterhub/singleuser.
Currently with julia 1.3.1.
Build with Docker by running docker build -t <name:tag> . in the directory
of the image you would like to build.
If you have any suggestions for different languages or base images you would like us to provide, feel free to submit an issue (or a pull request!) to this repo.