- Motivation
- Features
- Installation
- Basic Usage
- Default Template
- Customizing Templates
- Makefile
- PDM
- pytest
The Data Science Cookiecutter πͺ is an opinionated, yet configurable, Python project that provides a template for organizing and setting up data science projects. It uses the Cookiecutter project structure to create a standardized and reproducible project layout.
Data science projects often require a well-structured project layout to ensure reproducibility and collaboration. The Data Science Cookiecutter aims to solve this problem by providing a project template. While it follows certain opinions about project organization, it also allows for easy customization to fit different project needs.
- Standardized project structure for data science projects
- Automatic generation of project files and folders
- Customizable templates for different project needs
- Customizable for multiple programming languages (currently, the default template currently only has Python)
- Easy project initialization with just a few command-line arguments
The Data Science Cookiecutter is on pypi and can be installed using pip
, poetry
, pdm
or conda
.
pdm add datascience-cookiecutter
To create a new data science project using the Data Science Cookiecutter, follow these steps:
- Open a terminal or command prompt.
cd
to the directory where you want to create the project.- Run the following command:
cookiecutter myproject
wheremyproject
is the name of your project. - profit π
.
βββ Makefile <- Makefile for project automation
βββ README.md <- Project documentation and instructions
βββ pyproject.toml <- Configuration file for dependencies and project metadata
βββ data <- Folder to store data
β βββ final <- Folder for final processed data
β βββ processed <- Folder for intermediate processed data
β βββ raw <- Folder for raw data
β βββ sim <- Folder for simulated data
βββ dev <- Folder for development-related files
β βββ notebooks <- Folder for Jupyter notebooks
β βββ scripts <- Folder for development scripts
βββ docs <- Folder for project documentation
βββ myproject <- Placeholder folder for the project itself (replaced with your project name)
β βββ __init__.py <- Python package initialization file
β βββ main.py <- Main Python script for the project
βββ references <- Folder for reference materials
βββ reports <- Folder for project reports
β βββ img <- Folder for images and visualizations used in reports
β βββ report.md <- Sample report file (Markdown format)
βββ tests <- Folder for project tests
If you want to customize the default template used by cookiecutter
, you can create a templates.py
file in your $HOME/.config/cookiecutter
directory. Follow these steps:
- Open a text editor and create a new file called
templates.py
. - Import the necessary classes
Folder
andFileTemplate
by adding the following lines totemplates.py
:
from datascience_cookiecutter import Folder, FileTemplate
Define your custom template using the Folder and FileTemplate classes. Here's a minimal example:
MYTEMPLATE = Folder(
name="{{name}}",
subfolders=[
Folder(name="src", files=[FileTemplate(filename="main.py", content="print('Hello, world!')")]),
Folder(name="data"),
Folder(name="docs"),
],
files=[
FileTemplate(filename="README.md", content="# My {{name}}"),
],
)
Occurences of {{name}}
will be replaced by the project name as provided
with cookiecutter myprojectname
. To use your custom template, simply run the
cookiecutter command with the --template
option followed by the name of
your custom template. For example:
$ cookiecutter myproject --template=MYTEMPLATE
Enjoy customizing your templates! β¨π§ββοΈ
A Makefile is a file containing a set of instructions, known as targets, used to automate tasks in software development. It provides a convenient way to define and organize common commands for building, testing, and managing a project.
In the provided Makefile, you have the following targets:
- install: Installs project dependencies using
pdm install
. - test: Runs project tests with
pytest
- format: Applies code formatting using
isort
andblack
. - lint: Performs linting and static type checking using
ruff
andmypy
To use the Makefile, open a terminal or command prompt, navigate to your project directory, and run the desired target using the make command followed by the target name. For example:
make install
PDM is a Python package manager and build tool that provides an alternative to other package managers like pip or Poetry. It aims to simplify and enhance the management of project dependencies, virtual environments, and building distributions. Follow the link to install it. If you dont want to use it, you can customize the template to create your own Makefile and pyproject.toml.
The template (and PDM) follow the PEP 621 standard for project metadata to use a pyproject.toml file instead of setup.py. This file contains the project metadata and dependencies. It also allows you to specify details like the Python version and the project entry point.
Pytest is a Python testing framework that allows you to write simple and scalable tests with a clean and expressive syntax. It provides powerful features like fixtures, test discovery, and test selection.
For more information, you can visit the official pytest website: pytest.org