git_github_intro.md

Projects, git, github intro

A typical data science project

• multiple files with data analysis scripts
• development takes time - scripts evolve and have multiple versions
• growing - multiple directories/subdirectories after some time
• multiple files with data (but note: large files or files with sensitive content require different handling)
• self-shared (e.g. developed on a laptop and on a desktop machine)
• shared development with other contributors (programmers, supervisors, etc.)
• different states (snapshots) of the project exist on different computers

Git

• "Git is a free and open source distributed version control system designed to handle everything from small to very large projects with speed and efficiency."
• https://git-scm.com/
• Git is a tool for tracking changes in files and directories. It organizes a repository: a "database" which keeps history of changes in a project. Git detects which project files were changed/added/removed and it knows how to memorize these changes in the project repository. But Git is not the tool to edit/visualize/analyze the files - for this purpose we use other tools (e.g. text editors, Python, R, etc.).

GitHub

GitHub is a cloud-based web service where users can store, share and visualize their Git repositories.

Some example repositories:

• Materials of this course are hosted at GitHub, at the following location: https://github.com/LUMC/EfDS
• Another course which we developed with help of GitHub; the course has a longer development history:
  • The source code/materials are here: https://github.com/LUMC/rcourse
  • The web pages served by GitHub based on the sources: https://lumc.github.io/rcourse/
• Many python related repositories: https://github.com/python

First steps with Git

First directory and first repository:

• Let's assume that you have a project. All files of the project are located in a directory (possibly with subdirectories).
• First, a new Git local repository needs to be initialized for the project directory. In Git terminology the project directory is now called working directory. From now on, it will be possible to track changes in project files.
• Changes are not tracked automatically. For each file which you want to track, you need to execute Git add command. Later, when you execute Git commit command, a snapshot of all previously added files will be stored in the local Git repository.
• At any moment all committed files can be checked out (i.e. recreated in the working directory based on the local repository).

A cloned repository:

• Using data from "the first" repository created above, a new independent working directory for the project can be cloned.
• The clone is a new working directory with its own local repository.
• Now, the first repository and the cloned repository will keep independent histories of changes in the files of their own working directories.
• The changes of the histories can be synchronized (merged) later.

Tasks to practice

• Learn how to clone the course materials on your computer.

• Create your own account on GitHub.

• Create your first local repository:

  • First create a new directory for Python files developed during this course.
  • Initialize a git repository in this directory (git init console/terminal command)
  • Create a README.md markdown file with a trivial description of the directory content.
  • Copy a Python notebook file to the directory.
  • Use git add commands to declare that the new files should be put to the repository.
  • Use git commit command to create a new commit in the repository.

• Create a new repository on GitHub. You will see some instructions how to connect your local repository to the newly created GitHub one.

• Use git push to push your committed local repository to GitHub. After this operation you should see new content in GitHub.