This guide starts out with our general Python coding style guidelines and ends with a section on how we write & run (Python) tests.
Our starting point is PEP8, the standard "Style Guide for Python Code." Many Python IDEs will check your code against PEP8. (Note that PEP8 isn't frozen; it actually changes over time, but slowly.)
BigchainDB uses Python 3.5+, so you can ignore all PEP8 guidelines specific to Python 2.
We use pre-commit to check some of the rules below before every commit but not everything is realized yet. The hooks we use can be found in the .pre-commit-config.yaml file.
PEP8 says some things about docstrings, but not what to put in them or how to structure them. PEP257 was one proposal for docstring conventions, but we prefer Google-style docstrings instead: they're easier to read and the napoleon extension for Sphinx lets us turn them into nice-looking documentation. Here are some references on Google-style docstrings:
- Google's docs on Google-style docstrings
- napoleon's docs include an overview of Google-style docstrings
- Example Google-style docstrings (from napoleon's docs)
PEP8 has some maximum line length guidelines, starting with "Limit all lines to a maximum of 79 characters" but "for flowing long blocks of text with fewer structural restrictions (docstrings or comments), the line length should be limited to 72 characters."
We discussed this at length, and it seems that the consensus is: try to keep line lengths less than 79/72 characters, unless you have a special situation where longer lines would improve readability. (The basic reason is that 79/72 works for everyone, and BigchainDB is an open source project.) As a hard limit, keep all lines less than 119 characters (which is the width of GitHub code review).
Python lets you use single or double quotes. PEP8 says you can use either, as long as you're consistent. We try to stick to using single quotes, except in cases where using double quotes is more readable. For example:
print('This doesn\'t look so nice.')
print("Doesn't this look nicer?")Should we use parentheses or slashes (\) to break strings across multiple lines, i.e.
my_string = ('This is a very long string, so long that it will not fit into just one line '
'so it must be split across multiple lines.')
# or
my_string = 'This is a very long string, so long that it will not fit into just one line ' \
'so it must be split across multiple lines.'It seems the preference is for slashes, but using parentheses is okay too. (There are good arguments either way. Arguing about it seems like a waste of time.)
If you need to import lots of names from a module or package, and they won't all fit in one line (without making the line too long), then use parentheses to spread the names across multiple lines, like so:
from Tkinter import (
Tk, Frame, Button, Entry, Canvas, Text,
LEFT, DISABLED, NORMAL, RIDGE, END,
)
# Or
from Tkinter import (Tk, Frame, Button, Entry, Canvas, Text,
LEFT, DISABLED, NORMAL, RIDGE, END)For the rationale, see PEP 328.
Given the choice:
x = 'name: %s; score: %d' % (name, n)
# or
x = 'name: {}; score: {}'.format(name, n)we use the format() version. The official Python documentation says, "This method of string formatting is the new standard in Python 3, and should be preferred to the % formatting described in String Formatting Operations in new code."
We use Flake8 to check our Python code style. Once you have it installed, you can run it using:
flake8 --max-line-length 119 bigchaindb/
The content of this section was moved to bigchaindb/tests/README.md.
Note: We automatically run all tests on all pull requests (using Travis CI), so you should definitely run all tests locally before you submit a pull request. See the above-linked README file for instructions.