Install and Set Up Git

Table of contents

Install and Set Up Git

Setting up git and authenticating yourself to GitHub is an important first step in managing and tracking your code and various projects. Follow the steps for your platform below

GUI or Command-Line?

Developers should always feel empowered to select and use the tools they feel most comfortable with. There is no superiority or shame in selecting a GUI over the command-line git tool. Generally, most every feature is available in either.

While most data scientists tend to use git from the command-line, many use a GUI every day.

Advantanges of using a GUI for git:

Simple to use.
Handles dragging and dropping of files.
Visually displays changes, errors, and status.
Makes resolving merge conflicts much simpler.
GitHub Desktop is an excellent, free choice. There are many other free and paid git clients available.

Here is what GitHub Desktop looks like:

GitHub Desktop

Advantages of using the CLI version of git:

Fastest way to access all commands.
Displays only what the developer wants to see, allowing for focus.
Might be the only way to use git in a remote system (HPC cluster, cloud instances, etc.)

In order to focus on specific operations and workflow steps using git this documentation references CLI commands only. We recommend learning those commands to get the concepts, and then learn how they are done in a GUI.

Command-Line `git`

Data scientists and software developers who write code regularly greatly tend to use the terminal/command-line to interact with git. This primer assumes this and does not address various GUI tools for working with git and GitHub.

Installation

Download git

Install Git using the instructions at the following link. A git GUI version is optional, but the command line version is recommended: https://git-scm.com/book/en/v2/Getting-Started-Installing-Git
Check that the `git` CLI is installed in your environment by issuing the git command.

First-time Configuration

Confirm your installation is working by issuing the git command from within your terminal. You should see a number of subcommands avaialable to you.

The first time you use git you will encounter a couple of setup issues:

The first time you clone a repository from GitHub using a new ssh key you will get a “key fingerprint” approval request. You can safely say “Yes” to this request, which you will only be asked once.
The first time you try to commit code on your laptop using git you will be asked to configure two global settings - your name and email. These settings identify you in the log of commits and changes. Here are the two settings; replace the values in quotes with your own info:
```
git config --global user.name "Your Name"
git config --global user.email "mst3k@virginia.edu
```

Signing In

GitHub offers two ways to authenticate your local computer to GitHub. Usernames and passwords are no longer an option, as they present security risks. Instead you can use SSH keys or Tokens. We recommend you use a token.

Personal Access Tokens (PATs) - PATs are long, randomized tokens that can be scoped with specific levels of permissions. More on how to authenticate using PATs can be found here. More details about how to configure your local system are also available. Or watch the video below:
SSH Keys - you can also use SSH keypairs to authenticate your computer to GitHub. These allow your pushes and pulls, etc. to authenticate seamlessly to GitHub as a full owner of the repository. More about SSH authentication to GitHub is available here.

Troubleshooting Authentication

SSH key warnings are verbose and should be somewhat easy to interpret.
When a PAT is invalid, the user will be given a username and password prompt. This is likely because the PAT has expired or is malformed.
Remember that if you inject your PAT token value into the GitHub repository URL in the moment of cloning, that token is hard-coded into the .git/config file of that repository. You should update that URL with a fresh PAT as needed.