Julia @ IHI Code Club

MS Stream link to recorded presentation (24 June 2020)

Pre-requisites

If you can't (or don't want to ;-) install Julia on your machine, you will probably be able to try it during the live coding session on Binder. If you're happy to try it directly on your machine, follow the instructions below.

Installing Julia

Download the current stable version of Julia (v1.4.2 as of 2020-06-24) for your operating system from the official website. Read the platform-specific instructions should you need more guidance.

IDEs and text editors

Many popular IDEs and text editors have plugins for Julia. The two most popular ones are

• Julia extension for Visual Studio Code
• Juno, extension for Atom

If you don't have any preference, I slightly recommend Visual Studio Code over Atom, as it has a better support overall. If you already have your favourite development environment, go for it! There will likely be a plugin to add support for Julia.

Installing the required packages

During the live coding session we'll use some third-party packages. Julia comes with a built-in package manager, which you can use to install them. Start Julia, either in the terminal or in your favourite IDE, from the REPL you can enter the package manager mode with the ] key, then run the following command:

add DataFrames, PyCall
build PyCall

You can then exit the package manager mode by pressing backspace -- think about it like deleting the ] key you used to enter the package manager mode.

Alternatively, you can install the packages also with the following command directly in the REPL (this also works in Jupyter notebooks)

using Pkg
Pkg.add(["DataFrames", "PyCall"])
Pkg.build("PyCall")

After you have successfully installed and built the packages, make sure they work as expected with the following command in the REPL:

using DataFrames, PyCall

Brew a cup of coffee while you wait for the precompilation of the packages to finish :-)

Optional dependencies

If you want to run the Jupyter notebbok, you have to install the Julia kernel provided by the IJulia.jl package. To install and build it, either enter the package manager mode in the REPL with ] and run the commands

add IJulia
build IJulia

or run the coomands

using Pkg
Pkg.add("IJulia")
Pkg.build("IJulia")

How to follow the live coding session

To follow the live coding session on your computer, clone the git repository locally with the command

git clone https://github.com/IHI-Code-Club/Julia

You then have many options, depending on your preferred setup. You can follow what we'll do during the live coding session in either the MarkDown document julia.md, the Jupyter notebook julia.ipynb, or the Julia script src/julia.jl. The former two documents have been automatically generated from the latter one using a package for literate programming called Literate.jl.

Using the REPL

If you simply want to use Julia's REPL, start it and type the commands that we'll be running. You can also copy them from any of the three documents mentioned above.

Using an IDE

Juno

If you decided to use Juno, open src/julia.jl in Atom. You can evaluate a line of code or a selection of lines with Ctrl + Enter, the result of the evaluation will shown inline. For more information, read the Basic Usage instructions in the Juno documentation.

Visual Studio Code

The Julia extension for Visual Studio Code allows you to evaluate the code directly in the editor, similarly to what Juno does. Open the file src/julia.jl and hit Alt + Enter to evaluate the current code block and move to the next line, or use Ctrl + Enter to simply evaluate the current line. Refer to the Running Code section of the manual for more information.

Running the Jupyter notebook locally

If you enjoy using Jupyter notebooks, you may want to run julia.ipynb. Remember to install the IJulia.jl package as described above.

You can run the notebook as usual with

jupyter /PATH/TO/julia.ipynb

or in the Julia REPL with the commands

using IJulia
notebook(detached=true)

then browse to the directory where this repository is and open the notebook.

Running the Jupyter notebook on Binder

If you didn't have the possibility to install locally Julia and the package suggested for the live coding session, you may still have a chance: you can run the Jupyter notebook on Binder. This solution, however, depends on the availability of Binder resources at the time of live coding: many users connected at the same time may cause a slow down of the remote notebook.

Further resources

Many useful learning resources are listed on the official website. You may also be interested in

• Official Julia documentation
• Julia - Learn X in Y, a quick Julia cheatsheet
• Julia By Example, another cheatsheet