- Step 1: Read http://www.python.org/dev/peps/pep-0008/
- Step 2: Read http://www.python.org/dev/peps/pep-0008/ again
- Step 3: Read on
- Put two newlines between top-level code (funcs, classes, etc)
- Put one newline between methods in classes and anywhere else
- Do not write "except:", use "except Exception:" at the very least
- Include your name with TODOs as in "#TODO(termie)"
- Do not name anything the same name as a built-in or reserved word
- Do not make relative imports
- Order your imports by the full module path
- Organize your imports according to the following template
Example:
# vim: tabstop=4 shiftwidth=4 softtabstop=4 {{stdlib imports in human alphabetical order}} \n {{third-party lib imports in human alphabetical order}} \n {{glance imports in human alphabetical order}} \n \n {{begin your code}}
Example:
import httplib import logging import random import StringIO import time import unittest import eventlet import webob.exc import glance.api.middleware from glance.api import images from glance.auth import users import glance.common from glance.endpoint import cloud from glance import test
Docstrings are required for all functions and methods.
Docstrings should ONLY use triple-double-quotes ("""
)
Single-line docstrings should NEVER have extraneous whitespace between enclosing triple-double-quotes.
INCORRECT
""" There is some whitespace between the enclosing quotes :( """
CORRECT
"""There is no whitespace between the enclosing quotes :)"""
Docstrings that span more than one line should look like this:
Example:
""" Start the docstring on the line following the opening triple-double-quote If you are going to describe parameters and return values, use Sphinx, the appropriate syntax is as follows. :param foo: the foo parameter :param bar: the bar parameter :returns: return_type -- description of the return value :returns: description of the return value :raises: AttributeError, KeyError """
DO NOT leave an extra newline before the closing triple-double-quote.
If a dictionary (dict) or list object is longer than 80 characters, its items should be split with newlines. Embedded iterables should have their items indented. Additionally, the last item in the dictionary should have a trailing comma. This increases readability and simplifies future diffs.
Example:
my_dictionary = { "image": { "name": "Just a Snapshot", "size": 2749573, "properties": { "user_id": 12, "arch": "x86_64", }, "things": [ "thing_one", "thing_two", ], "status": "ACTIVE", }, }
Calls to methods 80 characters or longer should format each argument with newlines. This is not a requirement, but a guideline:
unnecessarily_long_function_name('string one', 'string two', kwarg1=constants.ACTIVE, kwarg2=['a', 'b', 'c'])
Rather than constructing parameters inline, it is better to break things up:
list_of_strings = [ 'what_a_long_string', 'not as long', ] dict_of_numbers = { 'one': 1, 'two': 2, 'twenty four': 24, } object_one.call_a_method('string three', 'string four', kwarg1=list_of_strings, kwarg2=dict_of_numbers)
In order to support multiple languages, we have a mechanism to support automatic translations of exception and log strings.
Example:
msg = _("An error occurred") raise HTTPBadRequest(explanation=msg)
If you have a variable to place within the string, first internationalize the template string then do the replacement.
Example:
msg = _("Missing parameter: %s") % ("flavor",) LOG.error(msg)
If you have multiple variables to place in the string, use keyword parameters. This helps our translators reorder parameters when needed.
Example:
msg = _("The server with id %(s_id)s has no key %(m_key)s") LOG.error(msg % {"s_id": "1234", "m_key": "imageId"})
For every new feature, unit tests should be created that both test and (implicitly) document the usage of said feature. If submitting a patch for a bug that had no unit test, a new passing unit test should be added. If a submitted bug fix does have a unit test, be sure to add a new one that fails without the patch and passes with the patch.
A number of modules from openstack-common are imported into the project.
These modules are "incubating" in openstack-common and are kept in sync with the help of openstack-common's update.py script. See:
http://wiki.openstack.org/CommonLibrary#Incubation
The copy of the code should never be directly modified here. Please always update openstack-common first and then run the script to copy the changes across.
Use __name__ as the name of your logger and name your module-level logger objects 'LOG':
LOG = logging.getLogger(__name__)