Skip to content

Latest commit

 

History

History
136 lines (93 loc) · 4.53 KB

CONTRIBUTING.md

File metadata and controls

136 lines (93 loc) · 4.53 KB

Contributing

Issues

A good issue includes a short, self contained, correct example of the problem, something like:

assert github.Github().get_user("jacquev6").name == "Vincent Jacques"

It is even better if you provide the debug logs associated with your issue. Enable them with github.enable_console_debug_logging and copy them in the body of the issue. Warning: you may want to remove some private information (authentication information is removed, but there may be private stuff in the messages)

If for any reason you are not able to do that, open your issue anyway and a maintainer will see what is needed to solve your problem.

Pull Requests

Pull Requests should clearly describe two things:

  1. The problem they attempt to solve
  2. How the author went about solving the problem

Ideally, changes should be made in logical commits and tests added to improve the project's coverage of the GitHub API.

Coding style

PyGithub adopts the black coding style.

To manually format the code:

tox -e lint

Pre-commit plugin

To forget about coding style and let pre-commit fix your flake8/isort/black issue.

pre-commit install

That's it!

Adding missing attributes for a GithubObject

$ python scripts/add_attribute.py [class_name] [attribute_name] [attribute_type]

# For example, if you want to add a `url` attribute of string type to the Commit class
# Note: adding multiple attributes you have to run the script multiple times

$ python scripts/add_attribute.py Commit url string

Deprecation warning

Before removing attributes/methods, consider adding deprecation warnings instead. The Deprecated packages provides a handy decorator to add deprecation warnings with an optional reason.

from deprecated import deprecated

@property
@deprecated
def rate(self):
   pass

@deprecated(reason="Deprecated in favor of the new branch protection")
def get_protected_branch(self):
   pass

Automated tests

First you need to install the test dependencies:

pip install -r requirements/test.txt

Then you can run the tests through pytest. Run a specific test with pytest tests/tests_filename.py or pytest tests/tests_filename.py -k testMethod or pytest -k TestClass.testMethod.

If you add a new test, for example Issue139.testCompletion, you have to run pytest -k Issue139.testCompletion --record to create the tests/ReplayData/*.txt files needed for your new test. Check them and commit them as well. You will need a GithubCredentials.py file at the root of the project with the following contents:

login = "my_login"
password = "my_password"                # Can be left empty if not used
oauth_token = "my_token"                # Can be left empty if not used
jwt = "my_json_web_token"               # Can be left empty if not used
app_id = "my_app_id"                    # Can be left empty if not used
app_private_key = "my_app_private_key"  # Can be left empty if not used

If you use 2 factor authentication on your Github account, tests that require a login/password authentication will fail. You can use pytest Issue139.testCompletion --record --auth_with_token to use the oauth_token field specified in GithubCredentials.py when recording a unit test interaction. Note that the password = "" (empty string is ok) must still be present in GithubCredentials.py to run the tests even when the --auth_with_token arg is used.

Also note that if you record your test data with --auth_with_token then you also need to be in token authentication mode when running the test. You can do this by setting tokenAuthMode to be true like so:

    def setUp(self):
        self.tokenAuthMode = True
        super().setUp()
        ...

A simple alternative is to replace token private_token_removed with Basic login_and_password_removed in all your newly generated ReplayData files.

Similarly, you can use pytest Issue139.testCompletion --record --auth_with_jwt to use the jwt field specified in GithubCredentials.py to access endpoints that require JWT.

To run manual tests with external scripts that use the PyGithub package, you can install your development version with:

pip install --editable path/to/project

You may also want to investigate tox to run tests:

pip install tox
tox -epy310

Build documentation locally

pip install -r requirements/docs.txt
sphinx-build doc build

If you use tox:

tox -edocs