Contributing
Contributions are welcome, and they are greatly appreciated! Every little bit helps, and credit will always be given.
Types of Contributions
You can contribute in many ways:
Report Bugs
Report bugs at https://github.com/cookiecutter/cookiecutter/issues.
If you are reporting a bug, please include:
Your operating system name and version.
Any details about your local setup that might be helpful in troubleshooting.
If you can, provide detailed steps to reproduce the bug.
If you don’t have steps to reproduce the bug, just note your observations in as much detail as you can. Questions to start a discussion about the issue are welcome.
Fix Bugs
Look through the GitHub issues for bugs. Anything tagged with “bug” is open to whoever wants to implement it.
Implement Features
Look through the GitHub issues for features. Anything tagged with “enhancement” and “please-help” is open to whoever wants to implement it.
Please do not combine multiple feature enhancements into a single pull request.
Note: this project is very conservative, so new features that aren’t tagged with “please-help” might not get into core. We’re trying to keep the code base small, extensible, and streamlined. Whenever possible, it’s best to try and implement feature ideas as separate projects outside of the core codebase.
Write Documentation
Cookiecutter could always use more documentation, whether as part of the official Cookiecutter docs, in docstrings, or even on the web in blog posts, articles, and such.
If you want to review your changes on the documentation locally, you can do:
pip install -r docs/requirements.txt
make servedocs
This will compile the documentation, open it in your browser and start watching the files for changes, recompiling as you save.
Submit Feedback
The best way to send feedback is to file an issue at https://github.com/cookiecutter/cookiecutter/issues.
If you are proposing a feature:
Explain in detail how it would work.
Keep the scope as narrow as possible, to make it easier to implement.
Remember that this is a volunteer-driven project, and that contributions are welcome :)
Setting Up the Code for Local Development
Here’s how to set up cookiecutter
for local development.
Fork the
cookiecutter
repo on GitHub.Clone your fork locally:
git clone git@github.com:your_name_here/cookiecutter.git
Install your local copy into a virtualenv. Assuming you have virtualenvwrapper installed, this is how you set up your fork for local development:
mkvirtualenv cookiecutter
cd cookiecutter/
python setup.py develop
Create a branch for local development:
git checkout -b name-of-your-bugfix-or-feature
Now you can make your changes locally.
When you’re done making changes, check that your changes pass the tests and lint check:
pip install tox
tox
Please note that tox runs lint check automatically, since we have a test environment for it.
If you feel like running only the lint environment, please use the following command:
tox -e lint
Ensure that your feature or commit is fully covered by tests. Check report after regular tox run. You can also run coverage only report and get html report with statement by statement highlighting:
tox -e cov-report
You report will be placed to htmlcov
directory. Please do not include this directory to your commits. By default this directory in our .gitignore
file.
Commit your changes and push your branch to GitHub:
git add .
git commit -m "Your detailed description of your changes."
git push origin name-of-your-bugfix-or-feature
Submit a pull request through the GitHub website.
Contributor Guidelines
Pull Request Guidelines
Before you submit a pull request, check that it meets these guidelines:
The pull request should include tests.
If the pull request adds functionality, the docs should be updated. Put your new functionality into a function with a docstring, and add the feature to the list in README.md.
The pull request must pass all CI/CD jobs before being ready for review.
If one CI/CD job is failing for unrelated reasons you may want to create another PR to fix that first.
Coding Standards
PEP8
Functions over classes except in tests
Quotes via http://stackoverflow.com/a/56190/5549
Use double quotes around strings that are used for interpolation or that are natural language messages
Use single quotes for small symbol-like strings (but break the rules if the strings contain quotes)
Use triple double quotes for docstrings and raw string literals for regular expressions even if they aren’t needed.
Example:
LIGHT_MESSAGES = {
'English': "There are %(number_of_lights)s lights.",
'Pirate': "Arr! Thar be %(number_of_lights)s lights."
}
def lights_message(language, number_of_lights):
"""Return a language-appropriate string reporting the light count."""
return LIGHT_MESSAGES[language] % locals()
def is_pirate(message):
"""Return True if the given message sounds piratical."""
return re.search(r"(?i)(arr|avast|yohoho)!", message) is not None
Write new code in Python 3.
Testing with tox
Tox uses py.test under the hood, hence it supports the same syntax for selecting tests.
For further information please consult the pytest usage docs
.
To run a particular test class with tox:
tox -e py '-k TestFindHooks'
To run some tests with names matching a string expression:
tox -e py '-k generate'
Will run all tests matching “generate”, test_generate_files for example.
To run just one method:
tox -e py '-k "TestFindHooks and test_find_hook"'
To run all tests using various versions of python in virtualenvs defined in tox.ini, just run tox:
tox
This configuration file setup the pytest-cov plugin and it is an additional dependency. It generate a coverage report after the tests.
It is possible to tests with some versions of python, to do this the command is:
tox -e py27,py35,pypy
Will run py.test with the python2.7, python3.5 and pypy interpreters, for example.
Core Committer Guide
Vision and Scope
Core committers, use this section to:
Guide your instinct and decisions as a core committer
Limit the codebase from growing infinitely
Command-Line Accessible
Provides a command-line utility that creates projects from cookiecutters
Extremely easy to use without having to think too hard
Flexible for more complex use via optional arguments
API Accessible
Entirely function-based and stateless (Class-free by intentional design)
Usable in pieces for developers of template generation tools
Being Jinja2-specific
Sets a standard baseline for project template creators, facilitating reuse
Minimizes the learning curve for those who already use Flask or Django
Minimizes scope of Cookiecutter codebase
Extensible
Being extendable by people with different ideas for Jinja2-based project template tools.
Entirely function-based
Aim for statelessness
Lets anyone write more opinionated tools
Freedom for Cookiecutter users to build and extend.
No officially-maintained cookiecutter templates, only ones by individuals
Commercial project-friendly licensing, allowing for private cookiecutters and private Cookiecutter-based tools
Fast and Focused
Cookiecutter is designed to do one thing, and do that one thing very well.
Cover the use cases that the core committers need, and as little as possible beyond that :)
Generates project templates from the command-line or API, nothing more
Minimize internal line of code (LOC) count
Ultra-fast project generation for high performance downstream tools
Inclusive
Cross-platform and cross-version support are more important than features/functionality
Fixing Windows bugs even if it’s a pain, to allow for use by more beginner coders
Stable
Aim for 100% test coverage and covering corner cases
No pull requests will be accepted that drop test coverage on any platform, including Windows
Conservative decisions patterned after CPython’s conservative decisions with stability in mind
Stable APIs that tool builders can rely on
New features require a +1 from 3 core committers
VCS-Hosted Templates
Cookiecutter project templates are intentionally hosted VCS repos as-is.
They are easily forkable
It’s easy for users to browse forks and files
They are searchable via standard Github/Bitbucket/other search interface
Minimizes the need for packaging-related cruft files
Easy to create a public project template and host it for free
Easy to collaborate
Process: Pull Requests
If a pull request is untriaged:
Look at the roadmap
Set it for the milestone where it makes the most sense
Add it to the roadmap
How to prioritize pull requests, from most to least important:
Fixes for broken tests. Broken means broken on any supported platform or Python version.
Extra tests to cover corner cases.
Minor edits to docs.
Bug fixes.
Major edits to docs.
Features.
Ensure that each pull request meets all requirements in checklist.
Process: Issues
If an issue is a bug that needs an urgent fix, mark it for the next patch release.
Then either fix it or mark as please-help.
For other issues: encourage friendly discussion, moderate debate, offer your thoughts.
New features require a +1 from 2 other core committers (besides yourself).
Process: Roadmap
The roadmap located here
Due dates are flexible. Core committers can change them as needed. Note that GitHub sort on them is buggy.
How to number milestones:
Follow semantic versioning. Look at: http://semver.org
Milestone size:
If a milestone contains too much, move some to the next milestone.
Err on the side of more frequent patch releases.
Process: Pull Request merging and HISTORY.md maintenance
If you merge a pull request, you’re responsible for updating AUTHORS.md
and HISTORY.md
When you’re processing the first change after a release, create boilerplate following the existing pattern:
## x.y.z (Development)
The goals of this release are TODO: release summary of features
Features:
* Feature description, thanks to [@contributor](https://github.com/contributor) (#PR).
Bug Fixes:
* Bug fix description, thanks to [@contributor](https://github.com/contributor) (#PR).
Other changes:
* Description of the change, thanks to [@contributor](https://github.com/contributor) (#PR).
Process: Accepting Template Pull Requests
Run the template to generate the project.
Attempt to start/use the rendered project.
Merge the template in.
Update the history file.
note: Adding a template doesn’t give authors credit.
Process: Your own code changes
All code changes, regardless of who does them, need to be reviewed and merged by someone else.
This rule applies to all the core committers.
Exceptions:
Minor corrections and fixes to pull requests submitted by others.
While making a formal release, the release manager can make necessary, appropriate changes.
Small documentation changes that reinforce existing subject matter. Most commonly being, but not limited to spelling and grammar corrections.
Responsibilities
Ensure cross-platform compatibility for every change that’s accepted. Windows, Mac, Debian & Ubuntu Linux.
Ensure that code that goes into core meets all requirements in this checklist: https://gist.github.com/audreyr/4feef90445b9680475f2
Create issues for any major changes and enhancements that you wish to make. Discuss things transparently and get community feedback.
Don’t add any classes to the codebase unless absolutely needed. Err on the side of using functions.
Keep feature versions as small as possible, preferably one new feature per version.
Be welcoming to newcomers and encourage diverse new contributors from all backgrounds. Look at Code of Conduct.
Becoming a Core Committer
Contributors may be given core commit privileges. Preference will be given to those with:
Past contributions to Cookiecutter and other open-source projects. Contributions to Cookiecutter include both code (both accepted and pending) and friendly participation in the issue tracker. Quantity and quality are considered.
A coding style that the other core committers find simple, minimal, and clean.
Access to resources for cross-platform development and testing.
Time to devote to the project regularly.