26 Pythonic Code Tips and Tricks

Pythonic tips from the video course

Michael Kennedy
@mkennedy

Write Pythonic Code Like a Seasoned Developer | Tips and Tricks from the Talk Python Training Course
Take the full video course at Talk Python

https://talkpython.fm/pythonic
 What is Pythonic code?

The idea of writing idiomatic code that is most

aligned with the language features and ideals is a
key concept in Python. We call this idiomatic code
Pythonic.

Write Pythonic Code Like a Seasoned Developer | Tips and Tricks from the Talk Python Training Course
 Truthiness

False # false is false

[] # empty lists / arrays are false
{} # empty dictionaries are false
"" # empty strings are false
0 # zero ints are false
0.0 # zero floats are false
None # None / null / nil pointers are false
Custom # Custom __bool__ / __nonzero__

# Everything else is true!

Write Pythonic Code Like a Seasoned Developer | Tips and Tricks from the Talk Python Training Course
 Testing for None
accounts = find_accounts(search_text)

if accounts is not None:

# list accounts...

Comparisons to singletons use is.

Negations with is are "is not" rather

than "not account is None"

Write Pythonic Code Like a Seasoned Developer | Tips and Tricks from the Talk Python Training Course
 Multiple tests against a single variable
class Moves(Enum): West=1; East=2; ...

if m in {Moves.North, Moves.South, Moves.West, Moves.East}:

print("That's a direct move.")
else:
print("That's a diagonal move.")

Use in set() to test single value against multiple conditions.

Write Pythonic Code Like a Seasoned Developer | Tips and Tricks from the Talk Python Training Course
 String formatting
name = 'Michael'
age = 43

# Pythonic
print("Hi, I'm {} and I'm {} years old.".format(name, age))
# Prints: Hi, I'm Michael and I'm 43 years old.

print("Hi, I'm {1} years old and my name is {0}, yeah {1} years.".format(name, age))
# Prints: Hi, I'm 43 years old and my name is Michael, yeah 43 years.

# If you have a dictionary then this is still better:

data = {'day': 'Saturday', 'office': 'Home office', 'other': 'UNUSED'}
print("On {day} I was working in my {office}!".format(**data))
# Prints: On Saturday I was working in my Home office!

# Finally, in Python 3.6, you can:

print(f"On {day} I was working in my {office}!")
# Prints: On Saturday I was working in my Home office!
 Send an exit code
confirm = input("Are you sure you want to format drive C: [yes, NO]? ")

if not confirm or confirm.lower() != 'yes':

print("Format cancelled!")
sys.exit(1)

format_drive()
print("Format completed successful.")

sys.exit(0)

External processes can now observe the success or failure states.

Write Pythonic Code Like a Seasoned Developer | Tips and Tricks from the Talk Python Training Course
 Flat is better than nested
def download():
print("Let's try to download a file")

if not s.check_download_url():
print("Bad url")
return
if not s.check_network():
print("No network")
return
if not s.check_dns():
print("No DNS")
return
if not s.check_access_allowed():
print("No access")
return

print("Sweet, we can download ...")

 PEP 8: Imports
Imports should be one line per package, almost always avoid wildcards

# BAD
import sys, os, multiprocessing
from statistics import *

# GOOD
import sys
import os
import multiprocessing

# ALSO GOOD
from os import path, chmod, chown

Multiple symbols from one module are fine however.

Write Pythonic Code Like a Seasoned Developer | Tips and Tricks from the Talk Python Training Course
 PEP 8: Naming Conventions
Modules have short, all-lowercase names.

# module data_access
Constants are UPPERCASE
# constant, not changed at runtime
TRANSCATION_MODE = 'serializable'

Classes have CapWord names. class CustomerRepository:

def __init__(self):
self.db = create_db()
Variables and args are lowercase
self.mode = 'dev'

Functions are lowercase def create_customer(self, name, email):

# ...

Exceptions should class RepositoryError(Exception): pass

end in Error
 Merging dictionaries
route = {'id': 271, 'title': 'Fast apps'}
query = {'id': 1, 'render_fast': True}
post = {'email': 'j@j.com', 'name': 'Jeff'}

m1 = {**query, **post, **route}

Py3

Write Pythonic Code Like a Seasoned Developer | Tips and Tricks from the Talk Python Training Course
Adding iteration to custom types
class ShoppingCart:
def __init__(self):
self.items = []

def add_item(self, it):

self.items.append(it)

def __iter__(self):
for it in sorted(self.items, key=lambda i: i.price):
yield it

cart = ShoppingCart()
cart.add_item(CartItem("guitar", 799))
cart.add_item(CartItem("cd", 19))
cart.add_item(CartItem("iPhone", 699))

for item in cart:

print('{} for ${}'.format(item.name, item.price))
 Testing for containment

nums_dict = {1: "one", 2: "two", 3:"three", 5: "five"}

word_version = nums_dict[2] # could KeyError

if 2 in nums_dict:
word_version = nums_dict[2] # safe!

Write Pythonic Code Like a Seasoned Developer | Tips and Tricks from the Talk Python Training Course
On demand computation with yield

def fibonacci_generator():
current, nxt = 0, 1
while True:
current, nxt = nxt, nxt + current
yield current

Write Pythonic Code Like a Seasoned Developer | Tips and Tricks from the Talk Python Training Course
 Recursive yields made easy
def get_files(folder):
for item in os.listdir(folder):
full_item = os.path.join(folder, item)

if os.path.isfile(full_item):
yield full_item
elif os.path.isdir(full_item):
yield from get_files(full_item)

Py3

Write Pythonic Code Like a Seasoned Developer | Tips and Tricks from the Talk Python Training Course
 Inline generators via expressions
# direct loop style
high_measurements = []
for m in measurements:
if m.value >= 70:
high_measurements.append(m.value)

# generator style
high_measurements = (
m.value
for m in measurements
if m.value >= 70
)

Write Pythonic Code Like a Seasoned Developer | Tips and Tricks from the Talk Python Training Course
 Lambda expressions
def find_special_numbers(special_selector, limit=10):
found = []
n = 0
while len(found) < limit:
if special_selector(n):
found.append(n)
n += 1
return found

find_special_numbers(lambda x: x % 6 == 0)

Lambda expressions are small, inline methods.

Avoid return values for error handling
EAFP ( easier to ask forgiveness than permission) style
def get_latest_file():
try:
data = s.download_file()
save_file('latest.png', data)
print('Successfully downloaded latest.png')
except ConnectionError as ce:
print("Problem with network: {}".format(ce))
except Exception as x:
print("Error: {}".format(x))

Exception handling catches all the cases and

allows for specific error handling.

Write Pythonic Code Like a Seasoned Developer | Tips and Tricks from the Talk Python Training Course
Passing variable number of arguments
def biggest(x, y): *NAME signals variable length arguments [0, …)
if x > y: args is the conventional name for this argument
return x
else:
return y

biggest(1, 7) # 7 def biggest(x, *args): # type(args) = tuple

b = x
for y in args:
if y > b:
b = y

return b

biggest_var(1, 7, 11, 99, 5, -2) # 99

Write Pythonic Code Like a Seasoned Developer | Tips and Tricks from the Talk Python Training Course
Mutable values as method defaults
Default value is created once per process.

def add_items_bad(name, times=1, lst=[]):

for _ in range(0, times):
lst.append(name)
return lst

list_1 = add_items_bad("a", 3) # list_1 = [a,a,a]

add_items_bad("b", 2, list_1) # list_1 = [a,a,a,b,b]

list_2 = add_items_bad("n", 3) # list_2 = [a,a,a,b,b,n,n,n]

id(list_1) == id(list_2) # True

Write Pythonic Code Like a Seasoned Developer | Tips and Tricks from the Talk Python Training Course
Mutable values as method defaults
List value is created as needed.

def add_items_bad(name, times=1, lst=None):

if lst is None:
lst = []

for _ in range(0, times):

lst.append(name)
return lst

list_1 = add_items_bad("a", 3) # list_1 = [a,a,a]

add_items_bad("b", 2, list_1) # list_1 = [a,a,a,b,b]

list_2 = add_items_bad("n", 3) # list_2 = [n,n,n]

id(list_1) == id(list_2) # False
 Using __main__
Only execute code (rather than method
definitions, etc.) when called directly

# program.py
import support

if __name__ == '__main__':
print("The variable is {}".format(support.var))
# prints: The variable is A variable

Write Pythonic Code Like a Seasoned Developer | Tips and Tricks from the Talk Python Training Course
 Virtual environments
> pip3 install virtualenv

> python3 -m virtualenv ./local_env

Using base prefix
'/Library/Frameworks/Python.framework/Versions/3.5'
New python executable in ./local_env/bin/python3
Also creating executable in ./local_env/bin/python
Installing setuptools, pip...done.

> cd local_env/bin/

> . activate

(local_env) > pip list

pip (8.0.2)
setuptools (19.6.2)
wheel (0.26.0)

(local_env) > which python

/Users/mkennedy/python/environments/local_env/bin/python
 Defining fields on classes
Define fields in __init__ only.

class NotSoPythonicPet:
def __init__(self, name, age):
self.age = age
self.name = name

def get_name(self):
self.name_checked = True
return self.name

def get_age(self):
Almost never create new return self.age
fields outside of __init__.

Write Pythonic Code Like a Seasoned Developer | Tips and Tricks from the Talk Python Training Course
 Encapsulation and data hiding
class PetSnake:
def __init__(self, name, age):
self.__age = age
self.__name = name Double _'s mean private.
self._protected = True Single _'s mean protected.

@property
def name(self):
return self.__name

@property
def age(self): py = PetSnake("Slide", 6)
a = py.age # a is 6
return self.__age
protected_v = py._protected # works with warning
direct_a = py.__age # crash:
# AttributeError: 'PetSnake' object has no attribute '__age'

Write Pythonic Code Like a Seasoned Developer | Tips and Tricks from the Talk Python Training Course
 Properties (Pythonic)
class PetSnake:
def __init__(self, name, age):
self.__age = age
self.__name = name

@property
def name(self):
return self.__name
py = PetSnake("Slide", 6)
@property
def age(self):
print("She is named {} and {} years old."
return self.__age
.format(py.name, py.age
))
# prints: She is named Slide and 6 years old.

Write Pythonic Code Like a Seasoned Developer | Tips and Tricks from the Talk Python Training Course
Wait, is there a numerical for loop? [2]
More often, you need the index and the value.
Use enumerate for this.

data = [1, 7, 11]

for idx, val in enumerate(data):

print(" {} --> {}".format(idx, val))

# Prints: 0 --> 1, 1 --> 7, 2 --> 11

Write Pythonic Code Like a Seasoned Developer | Tips and Tricks from the Talk Python Training Course
 Named Tuples
named tuples are clearer and explicitly named

Measurement = collections.namedtuple(
'Measurement',
'temp, lat, long, quality')

m = Measurement(22.5, 44.234, 19.02, 'strong')

temp = m[0]
temp = m.temp
quality = m.quality

print(m)
# Measurement(temp=22.5, lat=44.234,
long=19.02, quality='strong')
 Getting the source code

https://github.com/mikeckennedy/write-pythonic-code-demos