# Using pre-commit hooks with uv for Data Science

## 1. What is a *pre-commit hook*?

A **hook** is a small script that runs automatically at certain stages of using Git.
A **pre-commit hook** runs right before you save your changes with `git commit`.

The idea: automate checks or fixes **before your code enters Git history**.

Examples:
- check Python code style (PEP8),
- automatically format code,
- prevent committing large data files,
- clean up Jupyter notebooks before committing.



## 2. Why should data scientists care?

When working in a team, clean code and reproducibility matter.
Hooks help you:
- **avoid silly mistakes** (typos, forgotten formatting),
- **keep the repo clean** (notebooks without huge outputs),
- **save time** (problems are fixed before review),
- **collaborate smoothly** with engineers.



## 3. Install and initialize `uv`

If you don’t have `uv` yet, install it:

```bash
curl -LsSf https://astral.sh/uv/install.sh | sh
````

Then create and enter a project:

```bash
uv init my-data-project
cd my-data-project
```

This creates a virtual environment and a `pyproject.toml`.



## 4. Add `pre-commit` with `uv`

Instead of `pip install pre-commit`, use:

```bash
uv add --dev pre-commit
```

This installs `pre-commit` as a **development dependency** in your `pyproject.toml`.



## 5. Initialize hooks in Git

Inside your project, run:

```bash
uv run pre-commit install
```

From now on, every `git commit` will trigger the hooks you configure.



## 6. Configure `.pre-commit-config.yaml`

At the root of your repo, create a `.pre-commit-config.yaml`.
Here’s a setup useful for data scientists:

```yaml
repos:
  # Basic checks
  - repo: https://github.com/pre-commit/pre-commit-hooks
    rev: v4.6.0
    hooks:
      - id: trailing-whitespace
      - id: end-of-file-fixer
      - id: check-merge-conflict
      - id: check-added-large-files

  # Python code formatter (Black)
  - repo: https://github.com/psf/black
    rev: 23.9.1
    hooks:
      - id: black
        language_version: python3

  # Organize imports (isort)
  - repo: https://github.com/pycqa/isort
    rev: 5.12.0
    hooks:
      - id: isort

  # Clean Jupyter notebooks
  - repo: https://github.com/kynan/nbstripout
    rev: 0.6.1
    hooks:
      - id: nbstripout
```

What these do:

* `trailing-whitespace`, `end-of-file-fixer`: keep files tidy,
* `black`: auto-format Python code,
* `isort`: clean and order imports,
* `nbstripout`: remove execution outputs from Jupyter notebooks.



## 7. Run hooks manually

To check all files at once:

```bash
uv run pre-commit run --all-files
```



## 8. Best practices for data scientists

* Keep `pre-commit` in **dev dependencies** only.
* Document in your README:

  ```bash
  uv run pre-commit install
  ```

  so teammates also enable hooks.
* Regularly update hooks:

  ```bash
  uv run pre-commit autoupdate
  ```



## Conclusion

With `uv` and *pre-commit hooks*, you get:

* faster installations than pip,
* reproducible environments,
* automatic code quality checks before commits.