Setting up Docker for data science: rstudio and jupyter

References
https://towardsdatascience.com/a-short-guide-to-using-docker-for-your-data-science-environment-912617b3603e

https://jupyter-docker-stacks.readthedocs.io/en/latest/

https://hub.docker.com/r/rocker/rstudio/


Installation:

Tricky installation on Windows

• download the desktop version
• click nexts
• run run docker hello-world  -> throws an error
• Fixed by IT, issue with Hyper V
• On boot took a long time to start -> Docker right click and reset

RStudio

docker run --rm -p 8787:8787 -e PASSWORD=yourpasswordhere rocker/rstudio

Use any password except rstudio

open the URL 127.0.0.1:8787

Enter rstudio and password


Jupyter

docker run -p 8888:8888 jupyter/scipy-notebook:17aba6048f44

Then open link which is printed out


Mounting a local folder

docker run -P -v "path_to_shared_local_directory":/home/jovyan/work/ -t cvbi/datascience-environment

Change "path_to_shared_local_directory" to a folder on the laptop

Got an error

docker: Error response from daemon: Drive sharing seems blocked by a firewall.
See 'docker run --help'.


Stop and remove a container

docker stop image_name
docker rm image_name



Save changes to container

After doing the changes

docker commit CONTAINER_ID new_name

docker stop CONTAINER_ID

docker rm CONTAINER_ID

Script to automate saving a container

https://github.com/zoobab/docker-save-running-containers

Restart

docker run -p xxxx:xxxx new_name