Rust

Rust is a programming language meant to replace traditionally “difficult” compiled languages like C and C++. Rust can be used for almost any application, data science included.

Getting started on the cluster

The traditional way to install Rust is via its package manager: rustup. Now, rustup is the “toolchain manager”, but you can think of it as a package manager, but only for different versions of Rust. This should already be installed in the cluster, you can simply type rustup to get started.

Now, rustup has lots of capabilities, there is even a book about it, but the only rustup command you should ever have to run is the following: 

rustup default stable

This will install the stable version of Rust and set it as the default, which is fine in most cases.

Hello, world

Once, you’ve installed Rust, you can create a sample project. Make a directory you would like to have your rust project in, then run the following: 

mkdir myproject # create your project directory
cd myproject # go inside the directory
cargo init # sets up a rust project

Cargo is rust’s package manager for packages, known as “crates”. It also works as a build system, like Make or Scons. By default, cargo init sets up the project as a “hello, world” program. To compile and run the project, run the following: 

cargo run # this should print "hello, world"

The entry point of a Rust program is main.rs, so feel free to experiment in there. Rust is designed to work with editors, so you’re really missing out if you’re just using the plain vim or nano. Consider trying out VS code with the Rust extension.

Resources

One of the great things about Rust is its ecosystem. The Rust organization provides a tool called mdbook, which makes it easy to create static website “books” in and outside the rust ecosystem.

Rust (the organization) provides two books:

  • “The Book”: great if you want theory or extended explanations about Rust code
  • “Rust by Example”: Great for quick and dirty learning, and a good no-nonsense reference when you just need it to work

Additionally, Rust provides an interactive course meant to be run locally:

Videos

Here is a playlist of videos that may sell you on rust. They are short and easy to digest.

Alternative ways of running rust

Rust is traditionally a compiled language, though the editor support sidesteps many of the problems associated with compiled languages (runtime uncertainty). That being said, there are ways of interpreting Rust, rather than compiling it.

The central project creating interpreter support for Rust is evcxr, which provides an interactive repl as well as a jupyter kernel for use with jupyter notebooks.