Python etc
6.34K subscribers
17 photos
156 links
Regular tips about Python and programming in general

Owner — @pushtaev
The current season is run by @orsinium

Tips are appreciated: https://ko-fi.com/pythonetc / https://sobe.ru/na/pythonetc

© CC BY-SA 4.0 — mention if repost
Download Telegram
to view and join the conversation
Let's talk a bit more about scopes.

Any class and function can implicitly use variables from the global scope:

v = 'global'
def f():
print(f'{v=}')
f()
# v='global'


Or from any other enclosing scope, even if it is defined after the fucntion definition:

def f():
v1 = 'local1'
def f2():
def f3():
print(f'{v1=}')
print(f'{v2=}')
v2 = 'local2'
f3()
f2()
f()
# v1='local1'
# v2='local2'


Class body is a tricky case. It is not considered an enclosing scope for functions defined inside of it:

v = 'global'
class A:
v = 'local'
print(f'A {v=}')
def f():
print(f'f {v=}')
# A v='local'

A.f()
# f v='global'
Any enclosing variable can be shadowed in the local scope without affecting the global one:

v = 'global'
def f():
v = 'local'
print(f'f {v=}')
f()
# f v='local'

print(f'{v=}')
# v='global'


And if you try to use a variable and then shadow it, the code will fail at runtime:

v = 'global'
def f():
print(v)
v = 'local'
f()
# UnboundLocalError: local variable 'v' referenced before assignment


If you want to re-define the global variable instead of locally shadowing it, it can be achieved using global and nonlocal statements:

v = 'global'
def f():
global v
v = 'local'
print(f'f {v=}')
f()
# f v='local'
print(f'g {v=}')
# g v='local'

def f1():
v = 'non-local'
def f2():
nonlocal v
v = 'local'
print(f'f2 {v=}')
f2()
print(f'f1 {v=}')
f1()
# f2 v='local'
# f1 v='local'


Also, global can be used to skip non-local definitions:

v = 'global'
def f1():
v = 'non-local'
def f2():
global v
print(f'f2 {v=}')
f2()
f1()
# f2 v='global'


To be said, using global and nonlocal is considered a bad practice that complicates the code testing and usage. If you want a global state, think if it can be achieved in another way. If you desperately need a global state, consider using singleton pattern which is a little bit better.
Let's learn a bit more about strings performance. What if instead of an unknown amount of strings we have only a few known variables?

s1 = 'hello, '
s2 = '@pythonetc'

%timeit s1+s2
# 56.7 ns ± 6.17 ns per loop

%timeit ''.join([s1, s2])
# 110 ns ± 6.09 ns per loop

%timeit '{}{}'.format(s1, s2)
# 63.3 ns ± 6.69 ns per loop

%timeit f'{s1}{s2}'
# 57 ns ± 5.43 ns per loop


No surprises here, + and f-strings are equally good, str.format is quite close. But what if we have numbers instead?

n1 = 123
n2 = 456
%timeit str(n1)+str(n2)
# 374 ns ± 7.09 ns per loop

%timeit '{}{}'.format(n1, n2)
# 249 ns ± 4.73 ns per loop

%timeit f'{n1}{n2}'
# 208 ns ± 3.49 ns per loop


In this case, formatting is faster because it doesn't create intermediate strings. However, there is something else about f-strings. Let's measure how long it takes just to convert an int into an str:

%timeit str(n1)
# 138 ns ± 4.86 ns per loop

%timeit '{}'.format(n1)
# 148 ns ± 3.49 ns per loop

%timeit format(n1, '')
# 91.8 ns ± 6.12 ns per loop

%timeit f'{n1}'
# 63.8 ns ± 6.13 ns per loop


Wow, f-strings are twice faster than just str! This is because f-strings are part of the grammar but str is just a function that requires function-lookup machinery:

import dis
dis.dis("f'{n1}'")
1 0 LOAD_NAME 0 (n1)
2 FORMAT_VALUE 0
4 RETURN_VALUE

dis.dis("str(n1)")
1 0 LOAD_NAME 0 (str)
2 LOAD_NAME 1 (n1)
4 CALL_FUNCTION 1
6 RETURN_VALUE


And once more, disclaimer: readability is more important than performance until proven otherwise. Use your knowledge with caution :)
Types str and bytes are immutable. As we learned in previous posts, + is optimized for str but sometimes you need a fairly mutable type. For such cases, there is bytearray type. It is a "hybrid" of bytes and list:

b = bytearray(b'hello, ')
b.extend(b'@pythonetc')
b
# bytearray(b'hello, @pythonetc')

b.upper()
# bytearray(b'HELLO, @PYTHONETC')


The type bytearray has all methods of both bytes and list except sort:

set(dir(bytearray)) ^ (set(dir(bytes)) | set(dir(list)))
# {'__alloc__', '__class_getitem__', '__getnewargs__', '__reversed__', 'sort'}


If you're looking for reasons why there is no bytearray.sort, there is the only answer we found: stackoverflow.com/a/22783330/8704691.
Suppose, you have 10 lists:

lists = [list(range(10_000)) for _ in range(10)]


What's the fastest way to join them into one? To have a baseline, let's just + everything together:

s = lists
%timeit s[0] + s[1] + s[2] + s[3] + s[4] + s[5] + s[6] + s[7] + s[8] + s[9]
# 1.65 ms ± 25.1 µs per loop


Now, let's try to use functools.reduce. It should be about the same but cleaner and doesn't require to know in advance how many lists we have:

from functools import reduce
from operator import add
%timeit reduce(add, lists)
# 1.65 ms ± 27.2 µs per loop


Good, about the same speed. However, reduce is not "pythonic" anymore, this is why it was moved from built-ins into functools. The more beautiful way to do it is using sum:

%timeit sum(lists, start=[])
# 1.64 ms ± 83.8 µs per loop


Short and simple. Now, can we make it faster? What if we itertools.chain everything together?

from itertools import chain
%timeit list(chain(*lists))
# 599 µs ± 20.4 µs per loop


Wow, this is about 3 times faster. Can we do better? Let's try something more straightforward:

%%timeit
r = []
for lst in lists:
r.extend(lst)
# 250 µs ± 5.96 µs per loop


Turned out, the most straightforward and simple solution is the fastest one.
Starting Python 3.8, the interpreter warns about is comparison of literals.

Python 3.7:

>>> 0 is 0
True


Python 3.8:

>>> 0 is 0
<stdin>:1: SyntaxWarning: "is" with a literal. Did you mean "=="?
True


The reason is that it is an infamous Python gotcha. While == does values comparison (which is implemented by calling __eq__ magic method, in a nutshell), is compares memory addresses of objects. It's true for ints from -5 to 256 but it won't work for ints out of this range or for objects of other types:

a = -5
a is -5 # True
a = -6
a is -6 # False
a = 256
a is 256 # True
a = 257
a is 257 # False
Floating point numbers in Python and most of the modern languages are implemented according to IEEE 754. The most interesting and hardcore part is "arithmetic formats" which defines a few special values:

+ inf and -inf representing infinity.
+ nan representing a special "Not a Number" value.
+ -0.0 representing "negative zero"

Negative zero is the easiest case, for all operations it considered to be the same as the positive zero:

-.0 == .0  # True
-.0 < .0 # False


Nan returns False for all comparison operations (except !=) including comparison with inf:

import math

math.nan < 10 # False
math.nan > 10 # False
math.nan < math.inf # False
math.nan > math.inf # False
math.nan == math.nan # False
math.nan != 10 # True


And all binary operations on nan return nan:

math.nan + 10  # nan
1 / math.nan # nan


You can read more about nan in previous posts:

+ https://t.me/pythonetc/561
+ https://t.me/pythonetc/597

Infinity is bigger than anything else (except nan). However, unlike in pure math, infinity is equal to infinity:

10 < math.inf         # True
math.inf == math.inf # True


The sum of positive and negative infinity is nan:

-math.inf + math.inf  # nan
Infinity has an interesting behavior on division operations. Some of them are expected, some of them are surprising. Without further talking, there is a table:

truediv (/)
| -8 | 8 | -inf | inf
-8 | 1.0 | -1.0 | 0.0 | -0.0
8 | -1.0 | 1.0 | -0.0 | 0.0
-inf | inf | -inf | nan | nan
inf | -inf | inf | nan | nan

floordiv (//)
| -8 | 8 | -inf | inf
-8 | 1 | -1 | 0.0 | -1.0
8 | -1 | 1 | -1.0 | 0.0
-inf | nan | nan | nan | nan
inf | nan | nan | nan | nan

mod (%)
| -8 | 8 | -inf | inf
-8 | 0 | 0 | -8.0 | inf
8 | 0 | 0 | -inf | 8.0
-inf | nan | nan | nan | nan
inf | nan | nan | nan | nan


The code used to generate the table:

import operator
cases = (-8, 8, float('-inf'), float('inf'))
ops = (operator.truediv, operator.floordiv, operator.mod)
for op in ops:
print(op.__name__)
row = ['{:4}'.format(x) for x in cases]
print(' ' * 6, ' | '.join(row))
for x in cases:
row = ['{:4}'.format(x)]
for y in cases:
row.append('{:4}'.format(op(x, y)))
print(' | '.join(row))
PEP-589 (landed in Python 3.8) introduced typing.TypedDict as a way to annotate dicts:

from typing import TypedDict

class Movie(TypedDict):
name: str
year: int

movie: Movie = {
'name': 'Blade Runner',
'year': 1982,
}


It cannot have keys that aren't explicitly specified in the type:

movie: Movie = {
'name': 'Blade Runner',
'year': 1982,
'director': 'Ridley Scott', # fails type checking
}


Also, all specified keys are required by default but it can be changed by passing total=False:

movie: Movie = {} # fails type checking

class Movie2(TypedDict, total=False):
name: str
year: int

movie2: Movie2 = {} # ok
PEP-526, introducing syntax for variable annotations (laded in Python 3.6), allows annotating any valid assignment target:

c.x: int = 0
c.y: int

d = {}
d['a']: int = 0
d['b']: int


The last line is the most interesting one. Adding annotations to an expression suppresses its execution:

d = {}

# fails
d[1]
# KeyError: 1

# nothing happens
d[1]: 1


Despite being a part of the PEP, it's not supported by mypy:

$ cat tmp.py
d = {}
d['a']: int
d['b']: str
reveal_type(d['a'])
reveal_type(d['b'])

$ mypy tmp.py
tmp.py:2: error: Unexpected type declaration
tmp.py:3: error: Unexpected type declaration
tmp.py:4: note: Revealed type is 'Any'
tmp.py:5: note: Revealed type is 'Any'
In most of the programming languages (like C, PHP, Go, Rust) values can be passed into a function either as value or as reference (pointer):

+ Call by value means that the value of the variable is copied, so all modification with the argument value inside the function won't affect the original value. This is an example of how it works in Go:

package main

func f(v2 int) {
v2 = 2
println("f v2:", v2)
// Output: f v2: 2
}

func main() {
v1 := 1
f(v1)
println("main v1:", v1)
// Output: main v1: 1
}


+ Call by reference means that all modifications that are done by the function, including reassignment, will modify the original value:

package main

func f(v2 *int) {
*v2 = 2
println("f v2:", *v2)
// Output: f v2: 2
}

func main() {
v1 := 1
f(&v1)
println("main v1:", v1)
// Output: main v1: 2
}


So, which one is used in Python? Well, neither.

In Python, the caller and the function share the same value:

def f(v2: list):
v2.append(2)
print('f v2:', v2)
# f v2: [1, 2]

v1 = [1]
f(v1)
print('v1:', v1)
# v1: [1, 2]


However, the function can't replace the value (reassign the variable):

def f(v2: int):
v2 = 2
print('f v2:', v2)
# f v2: 2

v1 = 1
f(v1)
print('v1:', v1)
# v1: 1


This approach is called Call by sharing. That means the argument is always passed into a function as a copy of the pointer. So, both variables point to the same boxed object in memory but if the pointer itself is modified inside the function, it doesn't affect the caller code.
What if we want to modify a collection inside a function but don't want these modifications to affect the caller code? Then we should explicitly copy the value.

For this purpose, all mutable built-in collections provide method .copy:

def f(v2):
v2 = v2.copy()
v2.append(2)
print(f'{v2=}')
# v2=[1, 2]
v1 = [1]
f(v1)
print(f'{v1=}')
# v1=[1]


Custom objects (and built-in collections too) can be copied using copy.copy:

import copy

class C:
pass

def f(v2: C):
v2 = copy.copy(v2)
v2.p = 2
print(f'{v2.p=}')
# v2.p=2

v1 = C()
v1.p = 1
f(v1)
print(f'{v1.p=}')
# v1.p=1


However, copy.copy copies only the object itself but not underlying objects:

v1 = [[1]]
v2 = copy.copy(v1)
v2.append(2)
v2[0].append(3)
print(f'{v1=}, {v2=}')
# v1=[[1, 3]], v2=[[1, 3], 2]


So, if you need to copy all subobjects recursively, use, copy.deepcopy:

v1 = [[1]]
v2 = copy.deepcopy(v1)
v2[0].append(2)
print(f'{v1=}, {v2=}')
# v1=[[1]], v2=[[1, 2]]
Python uses eager evaluation. When a function is called, all its arguments are evaluated from left to right and only then their results are passed into the function:

print(print(1) or 2, print(3) or 4)
# 1
# 3
# 2 4


Operators and and or are lazy, the right value is evaluated only if needed (for or if the left value is falsy, and for and if the left value is truthy):

print(1) or print(2) and print(3)
# 1
# 2


For mathematical operators, the precedence is how it is in math:

1 + 2 * 3
# 7


The most interesting case is operator ** (power) which is (supposedly, the only thing in Python which is) evaluated from right to left:

2 ** 3 ** 4 == 2 ** (3 ** 4)
# True
Most of the exceptions raised from the standard library or built-ins have a quite descriptive self-contained message:

try:
[][0]
except IndexError as e:
exc = e

exc.args
# ('list index out of range',)


However, KeyError is different: instead of a user-friendly error message it contains the key which is missed:

try:
{}[0]
except KeyError as e:
exc = e

exc.args
# (0,)


So, if you log an exception as a string, make sure you save the class name (and the traceback) as well, or at least use repr instead of str:

repr(exc)
# 'KeyError(0)'
When something fails, usually you want to log it. Let's have a look at a small toy example:

from logging import getLogger

logger = getLogger(__name__)
channels = {}

def update_channel(slug, name):
try:
old_name = channels[slug]
except KeyError as exc:
logger.error(repr(exc))
...

update_channel('pythonetc', 'Python etc')
# Logged: KeyError('pythonetc')


This example has a few issues:

+ There is no explicit log message. So, when it fails, you can't search in the project where this log record comes from.
+ There is no traceback. When the try block execution is more complicated, we want to be able to track where exactly in the call stack the exception occurred. To achieve it, logger methods provide exc_info argument. When it is set to True, the current exception with traceback will be added to the log message.

So, this is how we can do it better:

def update_channel(slug, name):
try:
old_name = channels[slug]
except KeyError as exc:
logger.error('channel not found', exc_info=True)
...

update_channel('pythonetc', 'Python etc')
# channel not found
# Traceback (most recent call last):
# File "...", line 3, in update_channel
# old_name = channels[slug]
# KeyError: 'pythonetc'


Also, the logger provides a convenient method exception which is the same as error with exc_info=True:

logger.exception('channel not found')
Let's have a look at the following log message:

import logging
logger = logging.getLogger(__name__)
logger.warning('user not found')
# user not found


When this message is logged, it can be hard based on it alone to reproduce the given situation, to understand what went wrong. So, it's good to provide some additional context. For example:

user_id = 13
logger.warning(f'user #{user_id} not found')


That's better, now we know what user it was. However, it's hard to work with such kinds of messages. For example, we want to get a notification when the same type of error messages occurred too many times in a minute. Before, it was one error message, "user not found". Now, for every user, we get a different message. Or another example, if we want to get all messages related to the same user. If we just search for "13", we will get many false positives where "13" means something else, not user_id.

The solution is to use structured logging. The idea of structured logging is to store all additional values as separate fields instead of mixing everything in one text message. In Python, it can be achieved by passing the variables as the extra argument. Most of the logging libraries will recognize and store everything passed into extra. For example, how it looks like in python-json-logger:

from pythonjsonlogger import jsonlogger

logger = logging.getLogger()

handler = logging.StreamHandler()
formatter = jsonlogger.JsonFormatter()
handler.setFormatter(formatter)
logger.addHandler(handler)

logger.warning('user not found', extra=dict(user_id=13))
# {"message": "user not found", "user_id": 13}


However, the default formatter doesn't show extra:

logger = logging.getLogger()
logger.warning('user not found', extra=dict(user_id=13))
# user not found


So, if you use extra, stick to the third-party formatter you use or write your own.
Multiline string literal preserves every symbol between opening and closing quotes, including indentation:

def f():
return """
hello
world
"""
f()
# '\n hello\n world\n '


A possible solution is to remove indentation, Python will still correctly parse the code:

def f():
return """
hello
world
"""
f()
# '\nhello\n world\n'


However, it's difficult to read because it looks like the literal is outside of the function body but it's not. So, a much better solution is not to break the indentation but instead remove it from the string content using textwrap.dedent:

from textwrap import dedent

def f():
return dedent("""
hello
world
""")
f()
# '\nhello\n world\n'
If any function can modify any passed argument, how to prevent a value from modification? Make it immutable! That means the object doesn't have methods to modify it in place, only methods returning a new value. This is how numbers and str are immutable. While list has append method that modifies the object in place, str just doesn't have anything like this, all modifications return a new str:

a = b = 'ab'
a is b # True
b += 'cd'
a is b # False


This is why every built-in collection has an immutable version:

+ Immutable list is tuple.
+ Immutable set is frozenset.
+ Immutable bytearray is bytes.
+ dict doesn't have an immutable version but since Python 3.3 it has types.MappingProxyType wrapper that makes it immutable:

from types import MappingProxyType

orig = {1: 2}
immut = MappingProxyType(orig)

immut[3] = 4
# TypeError: 'mappingproxy' object does not support item assignment


And since it is just a proxy, not a new type, it reflects all the changes in the original mapping:

orig[3] = 4
immut[3]
# 4
Python has a built-in module sqlite3 to work with SQLite database.

import sqlite3
conn = sqlite3.connect(':memory:')
cur = conn.cursor()
cur.execute('SELECT UPPER(?)', ('hello, @pythonetc!',))
cur.fetchone()
# ('HELLO, @PYTHONETC!',)


Fun fact: for explanation what is SQL Injection the documentation links xkcd about Bobby tables instead of some smart article or Wikipedia page.
Since Python doesn't have a char type, an element of str is always str:

'@pythonetc'[0][0][0][0][0]
# '@'


This is an infinite type and you can't construct in a strictly typed language (and why would you?) because it's unclear how to construct the first instance (thing-in-itself?). For example, in Haskell:

Prelude> str = str str

<interactive>:1:7: error:
• Occurs check: cannot construct the infinite type: t1 ~ t -> t1