Share
Embed
Python Tricks That You Can't Live Without
My talk from PyCon Philippines 2012.
Audrey Roy
June 30, 2012
Transcript
45 slides
-
PYTHON TRICKS THAT YOU CAN’T LIVE WITHOUT Audrey Roy [email protected]
http://www.audreymroy.com @audreyr PYCON PHILIPPINES 2012 -
ABOUT ME • Principal at Cartwheel Web • Massachusetts Institute
of Technology EECS (winter 2005) • Filipina-American and very proud to be here flickr.com/photos/chrisjrn/6102009780/ -
Audrey Roy @audreyr I ♥ PYTHON • OpenComparison core dev,
and contributor to various open-source projects • Co-founded PyLadies • Helped organize #pyconph • Python Software Foundation member • I even met my fiancé Daniel Greenfeld at PyCon! • • I even met my fiancé Daniel Greenfeld at PyCon! http://www.flickr.com/photos/47628826@N05/4374285165/ -
Audrey Roy @audreyr OVERVIEW • Code readability • Linters and
code checkers • Where to find free reusable Python libraries • How to package your code for reuse -
CODE READABILITY The #1 trick to being a great Python
developer is writing clear, understandable code. -
Audrey Roy @audreyr CODE READABILITY • The best Python code
is compact, but not too compact • Write self-documenting code • And document it anyway :) -
Can this be made cleaner? CODE READABILITY def is_even(x): if
x % 2 == 0: return True else: return False -
Can this be made even cleaner? CODE READABILITY def is_even(x):
if x % 2 == 0: return True return False -
Don’t forget your docstrings CODE READABILITY def is_even(x): """ Returns
True if x is even, and False if x is odd. """ return x % 2 == 0 -
Keep in mind Python’s philosophy as you code. ZEN OF
PYTHON >>> import this The Zen of Python, by Tim Peters Beautiful is better than ugly. Explicit is better than implicit. Simple is better than complex. Complex is better than complicated. Flat is better than nested. Sparse is better than dense. Readability counts. Special cases aren't special enough to break the rules. Although practicality beats purity. Errors should never pass silently. Unless explicitly silenced. In the face of ambiguity, refuse the temptation to guess. ... -
Audrey Roy @audreyr PEP8 • Python style guide • 4
spaces. No tabs! • Blank lines between function & class defs • Much more... -
SublimeLinter Highlights lines of code that are not PEP8-compliant. Catches
potential style issues or errors. SUBLIME TEXT 2 + PLUGINS (also for CSS, JS, PHP, Ruby, etc.) -
By the way, Sublime Text 2 plugins are simple Python
files SUBLIME TEXT 2 + PLUGINS To write a plugin, you put a Python file in Sublime’s “Packages” directory -
A command-line PEP8 checker. PEP8.PY $ pep8 test2.py test2.py:13:1: E302
expected 2 blank lines, found 1 test2.py:20:1: W391 blank line at end of file http://pypi.python.org/pypi/pep8 -
Advanced Python source code analyzer. PYLINT $ pylint test2.py No
config file found, using default configuration ************* Module test2 C: 1,0: Missing docstring F: 1,0: Unable to import 'django.db.models' C: 3,0: Invalid name "compa2lookup" (should match (([A-Z_][A- Z0-9_]*)|(__.*__))$) C: 13,0:p_expression_ID: Invalid name "p_expression_ID" (should match [a-z_][a-z0-9_]{2,30}$) C: 13,0:p_expression_ID: Invalid name "p" (should match [a-z_][a- z0-9_]{2,30}$) C: 13,20:p_expression_ID: Invalid name "p" (should match [a-z_][a- z0-9_]{2,30}$) C: 18,4:p_expression_ID: Invalid name "d" (should match [a-z_][a- z0-9_]{2,30}$) W: 19,11:p_expression_ID: Used * or ** magic http://pypi.python.org/pypi/pylint -
Advanced Python source code analyzer. PYLINT Report ====== 8 statements
analysed. Messages by category -------------------- +-----------+-------+---------+-----------+ |type |number |previous |difference | +===========+=======+=========+===========+ |convention |6 |NC |NC | +-----------+-------+---------+-----------+ |refactor |0 |NC |NC | +-----------+-------+---------+-----------+ |warning |1 |NC |NC | +-----------+-------+---------+-----------+ http://pypi.python.org/pypi/pylint -
Audrey Roy @audreyr FINDING CODE TO REUSE Where to get
FREE reusable Python libraries: 1. Python Standard Library • Many great essentials, already on your system! • http://docs.python.org/library/index.html 2. Python Package Index • 21,000+ packages to download! • http://pypi.python.org/ -
Audrey Roy @audreyr WHY REUSE CODE? • Python helps you
avoid reinventing the wheel • “Not Invented Here” syndrome -
Audrey Roy @audreyr MORE ABOUT THE PYTHON STDLIB A collection
of highly useful modules • No need to install • Just import and start using them! -
STDLIB EXAMPLE: MATH >>> import math >>> math.ceil(2.03) 3.0 >>>
math.floor(2.99) 2.0 >>> math.log(32,2) 5.0 >>> math.erf(0.5) 0.5204998778130465 Mathematical functions defined by the C standard -
STDLIB EXAMPLE: RANDOM >>> import random >>> random.random() 0.12863367604888531 >>>
random.uniform(0,100) 25.374019279313988 >>> math.floor(random.uniform(0,100)) 77.0 >>> random.randrange(0,100) 69 -
Audrey Roy @audreyr MORE ABOUT PYPI • PyPI is “Python
Package Index” • 21,000+ packages • All created by community members like you • http://pypi.python.org -
Audrey Roy @audreyr PYPI EXAMPLES • You saw some great
examples already from PyPI (Python Package Index) • pep8: Simple PEP8 syntax checker • pylint: Advanced source code analyzer -
Audrey Roy @audreyr STDLIB VS. PYPI • The stdlib is
conservative • Few additions/changes/deprecations • On PyPI, anything goes! -
Audrey Roy @audreyr STDLIB VS. PYPI • Sometimes PyPI packages
are better than the equivalent stdlib ones • e.g. requests is better than urllib2 • If in doubt, ask around -
Audrey Roy @audreyr THE WRONG WAY • Systemwide installation of
Python libraries is generally bad • You can make a mess of your system -
Audrey Roy @audreyr THE RIGHT WAY You really should be
using these 2 tools: • pip - a good package installer • virtualenv - create isolated Python envs I strongly recommend virtualenvwrapper too. -
Create isolated virtualenvs for different projects. THE RIGHT WAY: VIRTUALENV
$ workon consumer_io (consumer_io) $ cd consumer_io/proj/ (consumer_io) $ python manage.py runserver (consumer_io) $ ... (consumer_io) $ deactivate $ cd ../../experiments $ workon experiments (experiments) $ python somethingelse.py (experiments) $ ... -
Use pip to install packages into virtualenvs. THE RIGHT WAY:
PIP (experiments) $ pip install Django==1.4 pip is like easy_install, but much better. -
THE RIGHT WAY: PIP+VIRTUALENV SCENARIO: You use Django 1.3 for
work, but you want to experiment with Django 1.4. With pip and virtualenv, you can switch between 1.3 and 1.4 on the same computer. -
You should pin your dependencies in requirements.txt! PIP REQUIREMENTS FILES
$ pip install -r requirements.txt # Your requirements.txt file Flask==0.8 glue==0.2.5 Pillow==1.7.7 Django==1.4 Use pip install PackageName==1.0.4 for experimentation only. -
Once installed, you can import Python code from modules: AFTER
INSTALLATION? from collections import deque Or from submodules: from os.path import abspath -
A module is a file containing Python definitions and statements.
Like this: MODULES # divisible.py def is_even(x): """ Returns True if x is even, and False if x is odd. """ return x % 2 == 0 -
A Python package is a collection of modules. PACKAGES sound/
__init__.py formats/ __init__.py wav.py aiff.py effects/ __init__.py echo.py surround.py -
A sample import from this package: PACKAGES sound/ __init__.py formats/
__init__.py wav.py aiff.py effects/ __init__.py echo.py surround.py from sound.formats.wav import read_wav -
Relative imports work between submodules of a package: INTRA-PACKAGE IMPORTS
from . import echo from .. import formats from ..filters import equalizer -
INTRA-PACKAGE IMPORTS Absolute imports work between submodules of a package:
# Use this from anywhere in the package from sound.effects import echo package root -
Audrey Roy @audreyr IMPORTING FROM OUTSIDE A PACKAGE • Can’t
use absolute/relative imports • What to do? One of these: • Good: Add the package to PYTHONPATH [edit env var or use sys.path.append()] • Better: Install the package into your active virtualenv. -
Audrey Roy @audreyr BETTER PACKAGING • Recommended reading: “The Hitchhiker’s
Guide To Packaging” • http://guide.python-distribute.org • Learn to make packages that are downloadable & installable from PyPI -
Audrey Roy @audreyr THANK YOU • Find me if you
have questions • Introduce yourself - I’d love to meet you! • Twitter: @audreyr • Email: [email protected]