Start using a custom singleuser image

wildtreetech / ohjh

ohjh - The OpenHumans JupyterHub deployment

MIT License

8 stars 7 forks source link

Start using a custom singleuser image #15

Closed betatim closed 6 years ago

betatim commented 6 years ago

This adds a custom image with basic "data-science" libraries (R, Julia and Python + libraries). This starts closing #4.

In addition it adds a memory limit per user pod. We guarantee each user 1GB and limit them to 4GB.

betatim commented 6 years ago

Questions to check:

[ ] does this include devtools for R so people can install new packages? If yes, does it work and install stuff into the proper location?
[ ] does pip install do the right thing for this repo? What about conda?

gedankenstuecke commented 6 years ago

ip <- as.data.frame(installed.packages()[,c(1,3:4)])
rownames(ip) <- NULL
ip <- ip[is.na(ip$Priority),1:2,drop=FALSE]
print(ip, row.names=FALSE)

gives you a list of all installed packages. I just played around with an R notebook and it does include devtool as well as tidyverse. I noticed that tidyverse is version 1.0.0, while the current version is 1.2.1, so an update here would be nice (though people should be able to use devtools to upgrade themselves but it takes forever in my try)

betatim commented 6 years ago

Let's see if we can get the base image modified https://github.com/jupyter/docker-stacks/pull/562

betatim commented 6 years ago

Let's merge this, as notebooks.oh.org is currently deployed with this setup. We can fine tune image/installed packages in a new PR.