A cookiecutter template for Stanford's Programming in Journalism course (Comm 177P/277P) in winter quarter 2020.

The template is designed to introduce standardized development workflows and code hygiene to students.

We encourage you to fork and modify our code and send us feedback!

This template is intended for use with datakit, a command-line tool for automating data analysis worfklows. It provides the following features:

• Generates a standard file/directory structure
• Simplifies sharing of source data by saving it in git
• Encourages use of virtual environments via Pipenv
• Auto-generates GitHub projects
• Provides command-line tools to easily save code in git and push to GitHub
• Automatically adds instructors as repo collaborators

• Python >= 3.5
• Follow the datakit-github install docs. Skip the section called "Install DataKit libraries" (these are handled in the step below).
• Download and install requirements.txt in your system Python.

curl -s https://raw.githubusercontent.com/stanfordjournalism/cookiecutter-stanford-progj/master/requirements.txt | xargs pip install 

This cookiecutter template will be installed locally after creating your first project:

datakit project create --template gh:stanfordjournalism/cookiecutter-stanford-progj

If it's your only cookiecutter template (or the default), on subsequent runs you can simply use:

datakit project create

For more details on using cookiecutter templates with datakit, see the datakit-project docs.

Projects generated using this template include a README.md with important reminders on the standard workflow, such as using pipenv shell when running code or tasks for the project.

The template builds out the below project structure:

├── Pipfile
├── README.md
├── data
│   ├── processed (Raw data that has been transformed)
│   └── raw  (Copy of original source data)
├── lib (Re-usable Python code in .py files)
│   ├── __init__.py
│   └── utils.py
├── notebooks (Jupyter notebooks)
├── scripts (Number-prefixed data processing scripts)
│   └── 1-etl.py
└── tasks (invoke task definitions)
    ├── __init__.py
    └── code.py
        

Projects generated with this template include invoke command-line tasks designed to simplify the use of version control for new programmers.

# activate the project environment
cd your-project
pipenv shell

# save code changes in local git
invoke code.save

# push commits to GitHub
invoke code.push

These commands are limited by design. Most notably, they require students to be on the master branch of the repo, and do not provide a way to pull upstream changes.

More advanced students who want to branch, pull code from GitHub, etc. should use the standard git command-line tools or a Desktop client of their choice.

Note, we do not recommend GitHub's official desktop client for this template since it is designed for https-based interactions rather than ssh.

Students who use invoke commands to save and push work to GitHub should generally be following a unidirectional workflow (save on master and push). If an instructor needs to suggest code changes or alternative implementations as part of the feedback process, we recommend one of the below strategies:

• Add comments on files on GitHub
• Use GH issues
• Create a branch with changes and push to GitHub for discussion
• Push changes to master branch and pair program with the student to guide them through use of lower-level git commands to pull down changes

Environment settings can be used to control template behavior. These can be useful when testing the template to avoid spamming collaborators.

• PROGJ_GH_INTEGRATE (default: true) - set to "false" to disable GitHub project generation
• PROGJ_GH_ADD_COLLABS (default: true) - set to "false" to disable adding collaborators. This is most useful when instructors are developing on the repo and want to avoid spamming other instructors.
• PROGJ_GH_COLLABORATORS - colon-separated list of GitHub usernames (e.g. user1:user2)