Advanced git

Block 1.4: Filters and Hooks

Jan Simson

Automation & Control:
Git Filters & Hooks

Background: https://unsplash.com/photos/V1YEEItoOTE

What are Git Hooks? 🪝

• Git hooks are scripts that run automatically at specific points in your git workflow
• They allow you to automate tasks and enforce policies
• Can be used to:
  • Check code quality before commits
  • Validate commit messages
  • Run tests before pushing
  • …

Types of Git Hooks 🪝

• pre-commit: Before commit is created ⭐️
• prepare-commit-msg: Before commit editor opens
• commit-msg: After commit message is entered
• post-commit: After commit is created
• pre-push: Before pushing to remote

Where do hooks live? 🪝

• Hooks are stored in .git/hooks/ directory
• .git is an invisible directory that stores all the git information
  • If you delete it, you lose the local git history 🚨
• Sample hooks are provided (with .sample extension)

Warning

Hooks in .git/hooks/ are not tracked by git itself!

Example: A Simple pre-commit Hook

#!/bin/sh
# .git/hooks/pre-commit

# Check if there are any TODO comments
if git grep -q 'TODO' -- '*.py' '*.js'; then
    echo "Error: Found TODO comments in staged files"
    echo "Please resolve TODOs before committing"
    exit 1
fi

exit 0

Note

A non-zero exit code will abort the git operation

The pre-commit Framework 🐍

Python tool to manage and maintain (multi-language) pre-commit hooks.

https://pre-commit.com/

• Managing git hooks manually is tedious
• The pre-commit framework makes it easy:
  • Manage hooks via configuration file
  • Share hooks across team via git
  • Use hooks from the community
  • Run hooks automatically

Installing pre-commit

# Install pre-commit itself
pip install pre-commit

# Install the git hook scripts
pre-commit install

Note

After pre-commit install, hooks run automatically on git commit

Configuring pre-commit

Create a .pre-commit-config.yaml file in your repository:

repos:
  - repo: https://github.com/pre-commit/pre-commit-hooks
    rev: v4.5.0
    hooks:
      - id: trailing-whitespace
      - id: end-of-file-fixer
      - id: check-yaml
      - id: check-added-large-files

Popular pre-commit Hooks

• Code formatting
  • black (Python), styler (R), prettier (JS)
• Linting
  • ruff (Python), lintr (R)
• Security
  • detect-secrets, bandit
• General
  • trailing-whitespace, check-merge-conflict
  • mixed-line-ending

Running pre-commit Manually

# Run on all files
pre-commit run --all-files

# Run specific hook
pre-commit run black --all-files

# Run only on staged files
pre-commit run

Tip

Use --all-files when first setting up or updating hooks

Skipping Hooks (When Necessary)

Sometimes you need to bypass hooks:

# Skip all hooks
git commit --no-verify -m "Emergency fix"

# Skip specific hook with pre-commit
SKIP=black git commit -m "Skip formatting for now"

Hooks: More Info

• Tips
  1. Start small: Add hooks gradually
  2. Run fast: Keep hooks quick (< 10 seconds)
• 🔗 Discussion of Pros and Cons
• 🔗 hk: New alternative tool for managing git hooks

Git Filters 🔍

What are Git Filters?

• Filters process file content when moving between the working directory and git
• Two types:
  • Clean filters: Process files before staging (working dir → index (git))
  • Smudge filters: Process files after checkout (index (git) → working dir)
• Useful for:
  • Removing sensitive data before commits
  • Normalizing file formats
  • Handling generated content

Filter Example: Cleaning up Notebooks

# .git/config
[filter "nbstrip"]
    clean = jupyter nbconvert --ClearOutputPreprocessor.enabled=True --inplace
    smudge = Catch

# .gitattributes
*.ipynb filter=nbstrip

Warning

Filters can be complex, but often tools offer to set them up for you!

Questions?

More Info: https://pre-commit.com/ | https://git-scm.com/book/en/v2/Customizing-Git-Git-Hooks

Bonus Practical: Branching Chaos

1. Form teams of 2 people each
2. Open https://learngitbranching.js.org/?NODEMO (the last part is important)
3. Create a medium complex graph, by playing around with commits, branches, merging and rebasing
4. Reconstruct the graph created by your partner (and vice versa)

End of Section 🎉

Any Questions?

[🏡 Back to Overview]

[⏩️ Next Block]