Lab 1: R, RStudio, Quarto, Git, and SSH Keys

Biostat 203B

Author

Dr. Hua Zhou @ UCLA

Published

January 16, 2024

This tutorial goes through the process of installing the software environment (R, Quarto, RStudio, Git, SSH) for reproducing 203B course materials and working on 203B homework.

1 R, Quarto, and RStudio

Mac users be aware that R-x.x.x-arm64.pkg is for Apple Silicon CPUs (M1/M2/M3) and R-x.x.x-x86_64.pkg is for Intel CPUs.

Windows users also need to install the Rtools.

Enable GitHub Copilot in RStudio

You will thank me for this act. First enable GitHub Copilot in your GitHub account (Student Developer Pack). Then in RStudio Desktop, go to Tools -> Global Options... -> Copilot -> Enable GitHub Copilot -> Apply.

2 Terminal and Bash shell

  • Open the Terminal app. You can find Terminal from LaunchPad -> Other, or from Spotlight Search (type Terminal). For convenience, you can pin the Terminal program to the Dock.

  • Check that current shell is Bash

echo $SHELL
  • The default zsh shell on MacOS is mostly compatible with the Bash shell, and can be used interchangeably for this course. If necessary, you can change default shell to Bash by the following command in Terminal. Then close the Terminal window and reopen it.
chsh -s /bin/bash
  • Download and install Git for Windows v2.43 or later. Accept the default settings during the installation process.

  • Git Bash program is available as a component of Git for Windows. It provides a basic Bash shell and packages many commonly used Linux programs.

  • Instruct RStudio to use Git Bash as terminal: Tools -> Global Options... -> Terminal -> New terminals open with -> Git Bash.

  • Note that Git Bash is not a Linux system. It’s a Windows program that emulates a Bash shell, but lacks many Linux commands. Git Bash does not have a separate Linux file system. It piggybacks on the Windows file system, thus cannot do Linux-style file permission control. The user home of Git Bash is same as the user home on Windows, e.g., /c/Users/[USERNAME].

Our first lectures and homework depend heavily on Linux. It is recommended to install the WSL (Windows Subsystem for Linux) program, which is a complete Linux solution for Windows users.

  • Install WSL2 following instructions.

  • A standalone Ubuntu system (v22.04 Jammy) is available after WSL installation. Look for icon Ubuntu.

  • WSL is a full Linux system. It has its own Linux file system, separate from the Windows file system. The user home is at /home/[USERNAME] within the Linux file system.

    • Within WSL, it is easy to access the files in Windows file system, which is mounted at /mnt/c on WSL. For example, the file /c/Users/[USERNAME]/file is available within WSL at /mnt/c/Users/[USERNAME]/file.

    • Within Windows, it is easy to access the files in the WSL file system from the Linux tab in File Explorer.

  • For Homework 1 (extensive Linux exercises), it is recommended to use RStudio Server within WSL. Follow the instructions here (choose Ubuntu 22) to install R (v4.3.2 at time of writing) and RStudio Server on WSL2 Ubuntu v22.04.

  • Install Quarto CLI on WSL2 Ubuntu:

wget https://github.com/quarto-dev/quarto-cli/releases/download/v1.3.450/quarto-1.3.450-linux-amd64.deb
sudo gdebi quarto-1.3.450-linux-amd64.deb
  • In Windows, point your browser to http://localhost:8787 to access RStudio Server. The default username and the password is the same as your Ubuntu login credential.
Tip

Occasionally we need to lunch, restart, or stop RStudio Server on WSL2 Ubuntu. The Bash commands are

sudo rstudio-server start
sudo rstudio-server restart
sudo rstudio-server stop

3 Git and GitHub

  • Check whether git command is available in Terminal.

If git is not available, follow the instructions Installing on macOS section to install it.

git is included in either Git Bash or WSL Ubuntu.

4 SSH and keys

4.1 SSH client

  • SSH client should be available in Terminal (Mac Terminal, Windows Git Bash, or Windows WSL) by default.

  • In Terminal, the command to connect to a Linux machine is

ssh [USERNAME]@[IP_ADDRESS]

Replace [USERNAME] in the command by your actual user name on the Linux machine you are connecting to. If you cannot connect, you may not have an account. For example, to connect to my account on the Hoffman2 cluster at UCLA

ssh huazhou@hoffman2.idre.ucla.edu

4.2 SSH keys

  • First check whether you already have keys on your local machine. If you don’t have ~/.ssh folder, that means you have never used SSH before.
ls -al ~/.ssh
  • If no SSH keys yet, generate a pair of RSA keys
    • Method 1: Use RStudio to generate keys. Tools -> Global Options... -> Git/SVN -> Create SSH Key.
    • Method 2: Generate keys on Terminal, following the instructions in lecture notes.
    Using either method, make sure keys are in the default location ~/.ssh/
Tip

Make sure the permission for the key files are correct.

  • The permission for the ~/.ssh folder should be 700 (drwx------).

  • The permission for the private key ~/.ssh/id_rsa should be 600 (-rw-------).

  • The permission for the public key ~/.ssh/id_rsa.pub should be 644 (-rw-r--r--).

  • The permission for the ~/.ssh folder can be 755 (drwxr-xr-x).

  • The permission for the private key ~/.ssh/id_rsa can be 644 (-rw-r--r--).

  • The permission for the public key ~/.ssh/id_rsa.pub can be 644 (-rw-r--r--).

  • The permission for the ~/.ssh folder should be 700 (drwx------).

  • The permission for the private key ~/.ssh/id_rsa should be 600 (-rw-------).

  • The permission for the public key ~/.ssh/id_rsa.pub should be 644 (-rw-r--r--).

  • Upload the public SSH key to GitHub.com: Click the user Avatar on top right corner -> Settings -> SSH and GPG keys -> New SSH key.
Tip

In your local machine terminal, you may copy the public key to the clipboard by following command.

pbcopy ~/.ssh/id_rsa.pub
clip < ~/.ssh/id_rsa.pub

If the xclip and xsel are installed, then we can use

xclip -selection clipboard ~/.ssh/id_rsa.pub
  • On local machine, instruct RStudio to use the public key just generated: Tools -> Global Options... -> Git/SVN -> SSH key:.

  • After setting up SSH key, you can git push your local commits to GitHub repo seamlessly without passwords.

5 Obtain and reproduce course materials

  • In Terminal (Mac Terminal, Windows Git Bash, or Windows WSL), you can git clone a copy of course materials by
git clone https://github.com/ucla-biostat-203b/2024winter.git ~/203b-lecture

This command clones the GitHub repository to a folder called 203b-lecture under your home directory. You can clone multiple copies to different locations on the same machine if you want. Since this GitHub repository is frequently updated, you can sync the local copy on your computer with the repository on GitHub by running

git pull

within the 203b-lecture folder.

  • Navigate to the 203b-lecture folder, double click the 2024winter.Rproj file to open the project in RStudio. Within RStudio, you can navigate to the slides folder to open and render qmd files.
Tip

Alternatively, you can also use RStudio to git clone the biostat-203b-2024-winter repository to your local machine: File -> New Project... -> Version Control -> Git -> input Repository URL:, Project directory name: 203b-lecture.

  • In RStudio, navigate to the slides folder, open and render qmd files.

6 Docker and Hoffman2

There are more ways to run the course materials.

You can use Docker to run the course materials on your local machine. See Lab 3 for instructions.

You can also use Hoffman2 cluster at UCLA to run the course materials. See Hoffman2 documentation for instructions.

Please note this RStudio Server container image lacks many tools such as SSH and Quarto. You need to install them by yourself.