Skip to content

Latest commit

 

History

History
354 lines (251 loc) · 12.8 KB

CONTRIBUTING.rst

File metadata and controls

354 lines (251 loc) · 12.8 KB

Contributing

  1. Please sign one of the contributor license agreements below.
  2. Fork the repo, develop and test your code changes, add docs.
  3. Make sure that your commit messages clearly describe the changes.
  4. Send a pull request. (Please Read: Faster Pull Request Reviews)

In order to add a feature to google-cloud-python:

  • The feature must be documented in both the API and narrative documentation (in docs/).
  • The feature must work fully on the following CPython versions: 2.7, 3.4, 3.5, and 3.6 on both UNIX and Windows.
  • The feature must not add unnecessary dependencies (where "unnecessary" is of course subjective, but new dependencies should be discussed).

You'll have to create a development environment to hack on google-cloud-python, using a Git checkout:

  • While logged into your GitHub account, navigate to the google-cloud-python repo on GitHub.

  • Fork and clone the google-cloud-python repository to your GitHub account by clicking the "Fork" button.

  • Clone your fork of google-cloud-python from your GitHub account to your local computer, substituting your account username and specifying the destination as hack-on-google-cloud-python. E.g.:

    $ cd ${HOME}
    $ git clone git@github.com:USERNAME/google-cloud-python.git hack-on-google-cloud-python
    $ cd hack-on-google-cloud-python
    # Configure remotes such that you can pull changes from the google-cloud-python
    # repository into your local repository.
    $ git remote add upstream git@github.com:GoogleCloudPlatform/google-cloud-python.git
    # fetch and merge changes from upstream into master
    $ git fetch upstream
    $ git merge upstream/master
    

Now your local repo is set up such that you will push changes to your GitHub repo, from which you can submit a pull request.

To work on the codebase and run the tests, we recommend using nox, but you can also use a virtualenv of your own creation.

We use nox to instrument our tests.

  • To test your changes, run unit tests with nox:

    $ nox -f datastore/noxfile.py -s unit-2.7
    $ nox -f datastore/noxfile.py -s unit-3.6
    $ ...
    

    Note

    The unit tests and system tests are contained in the individual nox.py files in each directory; substitute datastore in the example above with the package of your choice.

    Alternatively, you can just navigate directly to the package you are currently developing and run tests there:

    $ export GIT_ROOT=$(pwd)
    $ cd ${GIT_ROOT}/datastore/
    $ nox -s "unit(py='3.6')"
    
  • As mentioned previously, using setuptools in develop mode or a pip editable install is not possible with this library. This is because this library uses namespace packages. For context see Issue #2316 and the relevant PyPA issue.

    Since editable / develop mode can't be used, packages need to be installed directly. Hence your changes to the source tree don't get incorporated into the already installed package.

If the error mentions Python.h not being found, install python-dev and try again. On Debian/Ubuntu:

$ sudo apt-get install python-dev
  • PEP8 compliance, with exceptions defined in the linter configuration. If you have nox installed, you can test that you have not introduced any non-compliant code via:

    $ nox -s lint
    
  • In order to make nox -s lint run faster, you can set some environment variables:

    export GOOGLE_CLOUD_TESTING_REMOTE="upstream"
    export GOOGLE_CLOUD_TESTING_BRANCH="master"
    

    By doing this, you are specifying the location of the most up-to-date version of google-cloud-python. The the suggested remote name upstream should point to the official GoogleCloudPlatform checkout and the the branch should be the main branch on that remote (master).

Exceptions to PEP8:

  • Many unit tests use a helper method, _call_fut ("FUT" is short for "Function-Under-Test"), which is PEP8-incompliant, but more readable. Some also use a local variable, MUT (short for "Module-Under-Test").
  • To run system tests for a given package, you can execute:

    $ nox -f datastore/noxfile.py -s system-3.6
    $ nox -f datastore/noxfile.py -s system-2.7
    

    Note

    System tests are only configured to run under Python 2.7 and Python 3.6. For expediency, we do not run them in older versions of Python 3.

    This alone will not run the tests. You'll need to change some local auth settings and change some configuration in your project to run all the tests.

  • System tests will be run against an actual project and so you'll need to provide some environment variables to facilitate authentication to your project:

    • GOOGLE_APPLICATION_CREDENTIALS: The path to a JSON key file; see system_tests/app_credentials.json.sample as an example. Such a file can be downloaded directly from the developer's console by clicking "Generate new JSON key". See private key docs for more details.
    • In order for Logging system tests to work, the Service Account will also have to be made a project Owner. This can be changed under "IAM & Admin". Additionally, cloud-logs@google.com must be given Editor permissions on the project.
  • Examples of these can be found in system_tests/local_test_setup.sample. We recommend copying this to system_tests/local_test_setup, editing the values and sourcing them into your environment:

    $ source system_tests/local_test_setup
    
  • For datastore tests, you'll need to create composite indexes with the gcloud command line tool:

    # Install the app (App Engine Command Line Interface) component.
    $ gcloud components install app-engine-python
    
    # Authenticate the gcloud tool with your account.
    $ GOOGLE_APPLICATION_CREDENTIALS="path/to/app_credentials.json"
    $ gcloud auth activate-service-account \
    > --key-file=${GOOGLE_APPLICATION_CREDENTIALS}
    
    # Create the indexes
    $ gcloud datastore create-indexes system_tests/data/index.yaml
    
  • For datastore query tests, you'll need stored data in your dataset. To populate this data, run:

    $ python datastore/tests/system/utils/populate_datastore.py
    
  • If you make a mistake during development (i.e. a failing test that prevents clean-up) you can clear all system test data from your datastore instance via:

    $ python datastore/tests/system/utils/clear_datastore.py
    
  • The codebase must have 100% test statement coverage after each commit. You can test coverage via nox -s cover.

If you fix a bug, and the bug requires an API or behavior modification, all documentation in this package which references that API or behavior must be changed to reflect the bug fix, ideally in the same commit that fixes the bug or adds the feature.

To build and review docs (where ${VENV} refers to the virtualenv you're using to develop google-cloud-python):

  1. After following the steps above in "Using a Development Checkout", install Sphinx and all development requirements in your virtualenv:

    $ cd ${HOME}/hack-on-google-cloud-python
    $ ${VENV}/bin/pip install Sphinx
    
  2. Change into the docs directory within your google-cloud-python checkout and execute the make command with some flags:

    $ cd ${HOME}/hack-on-google-cloud-python/google-cloud-python/docs
    $ make clean html SPHINXBUILD=${VENV}/bin/sphinx-build
    

    The SPHINXBUILD=... argument tells Sphinx to use the virtualenv Python, which will have both Sphinx and google-cloud-python (for API documentation generation) installed.

  3. Open the docs/_build/html/index.html file to see the resulting HTML rendering.

As an alternative to 1. and 2. above, if you have nox installed, you can build the docs via:

$ nox -s docs

The description on PyPI for the project comes directly from the README. Due to the reStructuredText (rst) parser used by PyPI, relative links which will work on GitHub (e.g. CONTRIBUTING.rst instead of https://github.com/GoogleCloudPlatform/google-cloud-python/blob/master/CONTRIBUTING.rst) may cause problems creating links or rendering the description.

All build scripts in the .circleci/config.yml configuration file which have Python dependencies are specified in the nox.py configuration. They are executed in the Travis build via nox -s ${ENV} where ${ENV} is the environment being tested.

We support:

Supported versions can be found in our noxfile.py config.

We explicitly decided not to support Python 2.5 due to decreased usage and lack of continuous integration support.

We have dropped 2.6 as a supported version as well since Python 2.6 is no longer supported by the core development team.

Python 2.7 support is deprecated. All code changes should maintain Python 2.7 compatibility until January 1, 2020.

We also explicitly decided to support Python 3 beginning with version 3.4. Reasons for this include:

This library follows Semantic Versioning.

Some packages are currently in major version zero (0.y.z), which means that anything may change at any time and the public API should not be considered stable.

Before we can accept your pull requests you'll need to sign a Contributor License Agreement (CLA):

  • If you are an individual writing original source code and you own the intellectual property, then you'll need to sign an individual CLA.
  • If you work for a company that wants to allow you to contribute your work, then you'll need to sign a corporate CLA.

You can sign these electronically (just scroll to the bottom). After that, we'll be able to accept your pull requests.