Jupyter All-in-One

This is an Docker image of jupyter lab with multiple common language kernels included. Some use cases:

• Remote workspace for coding
• Playground of language features
• Quick validation of programming ideas
• One-off scripting

Features

• Multiple languages are baked in:
  • Python3
  • C++ 11, 14, 17
  • Golang
  • Haskell
  • Java
  • Scala
  • Rust
  • Racket
  • TypeScript & JavaScript
• Both code-server(VS Code) and Jupyter Lab are exposed

Usage

docker run -d \
  -p 8080:8080 -p 8081:8081 \
  -v ~/data/lab:/data \
  -u $(id -u):$(id -g) \
  -e TOKEN=your_token \
  sineliu/jupyterlab-all-in-one:latest

code-server is listening on 8080, jupyter lab is listening on 8081.

Note: Rust is having issue running in code-server. link

Default user and password is sine. With sudo access. Build your own image for a different username.

Miscellaneous

Why not jupyterwith?

Nix has some limitations on flexibility:

• Installing python packages outside of nixpkgs can be troublesome
• Requires hack to make kernels compatible with code-server jupyter extension
• /nix is immutable, but sometimes it's convenient to edit installed packages for debugging

Problems on uid

As docker image delievers the filesystem, the user id and group id of installed kernels are determined at build time. There is no way to change it at runtime except for a tedious chown. This image is built with user id 1000 and group id 1000. Other runtime overrides are untested and may cause errors.