Skip to main content

flake8 plugin which checks for code that can be simplified

Project description

PyPI version Code on Github Actions Status Code style: black

flake8-simplify

A flake8 plugin that helps you simplify your code.

Installation

Install with pip:

pip install flake8-simplify

Python 3.6 to 3.8 are supported.

Usage

Just call flake8 . in your package or flake your.py:

$ flake8 .
./foo/__init__.py:690:12: SIM101 Multiple isinstance-calls which can be merged into a single call for variable 'other'

Rules

Python-specific rules:

Simplifying Comparations:

  • SIM201: Use 'a != b' instead of 'not a == b' (example)
  • SIM202: Use 'a == b' instead of 'not a != b' (example)
  • SIM203: Use 'a not in b' instead of 'not (a in b)' (example)
  • SIM204: Use 'a >= b' instead of 'not (a < b)' (example)
  • SIM205: Use 'a > b' instead of 'not (a <= b)' (example)
  • SIM206: Use 'a <= b' instead of 'not (a > b)' (example)
  • SIM207: Use 'a < b' instead of 'not (a <= b)' (example)
  • SIM208: Use 'a' instead of 'not (not a)' (example)
  • SIM210: Use 'bool(a)' instead of 'True if a else False' (example)
  • SIM211: Use 'not a' instead of 'False if a else True' (example)
  • SIM212: Use 'a if a else b' instead of 'b if not a else a' (example)
  • SIM220: Use 'False' instead of 'a and not a' (example)
  • SIM221: Use 'True' instead of 'a or not a' (example)
  • SIM222: Use 'True' instead of '... or True' (example)
  • SIM223: Use 'False' instead of '... and False' (example)
  • SIM224: Reserved for SIM901 once it's stable
  • SIM300: Use 'age == 42' instead of '42 == age' (example)

Simplifying usage of dictionaries:

  • SIM401: Use 'a_dict.get(key, "default_value")' instead of an if-block (example)
  • SIM118: Use 'key in dict' instead of 'key in dict.keys()' (example)

General Code Style:

Experimental rules:

Getting rules right for every possible case is hard. I don't want to disturb peoples workflows. For this reason, flake8-simplify introduces new rules with the SIM9 prefix. Every new rule will start with SIM9 and stay there for at least 6 months. In this time people can give feedback. Once the rule is stable, the code will change to another number.

Current experimental rules:

  • SIM901: Use comparisons directly instead of wrapping them in a bool(...) call (example)
  • SIM904: Assign values to dictionary directly at initialization (example)
  • SIM905: Split string directly if only constants are used (example)

Disabling Rules

You might have good reasons to ignore some flake8 rules. To do that, use the standard Flake8 configuration. For example, within the setup.cfg file:

[flake8]
ignore = SIM106, SIM113, SIM119, SIM9

Examples

SIM101

# Bad
isinstance(a, int) or isinstance(a, float)

# Good
isinstance(a, (int, float))

SIM102

# Bad
if a:
    if b:
        c

# Good
if a and b:
    c

SIM105

# Bad
try:
    foo()
except ValueError:
    pass

# Good
from contextlib import suppress

with suppress(ValueError):
    foo()

Please note that contextlib.suppress is roughly 3x slower than try/except (source).

SIM107

# Bad: This example returns 3!
def foo():
    try:
        1 / 0
        return "1"
    except:
        return "2"
    finally:
        return "3"


# Good
def foo():
    return_value = None
    try:
        1 / 0
        return_value = "1"
    except:
        return_value = "2"
    finally:
        return_value = "3"  # you might also want to check if "except" happened
    return return_value

SIM108

# Bad
if a:
    b = c
else:
    b = d

# Good
b = c if a else d

SIM109

# Bad
if a == b or a == c:
    d

# Good
if a in (b, c):
    d

SIM110

# Bad
for x in iterable:
    if check(x):
        return True
return False

# Good
return any(check(x) for x in iterable)

SIM111

# Bad
for x in iterable:
    if check(x):
        return False
return True

# Good
return all(not check(x) for x in iterable)

SIM112

# Bad
os.environ["foo"]
os.environ.get("bar")

# Good
os.environ["FOO"]
os.environ.get("BAR")

SIM113

Using enumerate in simple cases like the loops below is a tiny bit easier to read as the reader does not have to figure out where / when the loop variable is incremented (or if it is incremented in multiple places).

# Bad
idx = 0
for el in iterable:
    ...
    idx += 1

# Good
for idx, el in enumerate(iterable):
    ...

SIM114

# Bad
if a:
    b
elif c:
    b

# Good
if a or c:
    b

SIM115

# Bad
f = open(...)
...  # (do something with f)
f.close()

# Good
with open(...) as f:
    ...  # (do something with f)

SIM116

# Bad
if a == "foo":
    return "bar"
elif a == "bar":
    return "baz"
elif a == "boo":
    return "ooh"
else:
    return 42

# Good
mapping = {"foo": "bar", "bar": "baz", "boo": "ooh"}
return mapping.get(a, 42)

SIM117

# Bad
with A() as a:
    with B() as b:
        print("hello")

# Good
with A() as a, B() as b:
    print("hello")

Thank you for pointing this one out, Aaron Gokaslan!

SIM118

# Bad
key in a_dict.keys()

# Good
key in a_dict

Thank you for pointing this one out, Aaron Gokaslan!

SIM119

Dataclasses were introduced with PEP 557 in Python 3.7. The main reason not to use dataclasses is to support legacy Python versions.

Dataclasses create a lot of the boilerplate code for you:

  • __init__
  • __eq__
  • __hash__
  • __str__
  • __repr__

A lot of projects use them:

SIM120

# Bad
class FooBar(object):
    ...

# Good
class FooBar:
    ...

Both notations are equivalent in Python 3, but the second one is shorter.

SIM201

# Bad
not a == b

# Good
a != b

SIM202

# Bad
not a != b

# Good
a == b

SIM203

# Bad
not a in b

# Good
a not in b

SIM204

# Bad
not a < b

# Good
a >= b

SIM205

# Bad
not a <= b

# Good
a > b

SIM206

# Bad
not a > b

# Good
a <= b

SIM207

# Bad
not a >= b

# Good
a < b

SIM208

# Bad
not (not a)

# Good
a

SIM210

# Bad
True if a else False

# Good
bool(a)

SIM211

# Bad
False if a else True

# Good
not a

SIM212

# Bad
b if not a else a

# Good
a if a else b

SIM220

# Bad
a and not a

# Good
False

SIM221

# Bad
a or not a

# Good
True

SIM222

# Bad
... or True

# Good
True

SIM223

# Bad
... and False

# Good
False

SIM300

# Bad; This is called a "Yoda-Condition"
42 == age

# Good
age == 42

SIM401

# Bad
if key in a_dict:
    value = a_dict[key]
else:
    value = "default_value"

# Good
thing = a_dict.get(key, "default_value")

SIM901

This rule will be renamed to SIM224 after its 6-month trial period is over. Please report any issues you encounter with this rule!

The trial period starts on 23rd of January and will end on 23rd of August 2022.

# Bad
bool(a == b)

# Good
a == b

SIM904

This rule will be renamed to SIM224 after its 6-month trial period is over. Please report any issues you encounter with this rule!

The trial period starts on 12th of February and will end on 12th of September 2022.

# Bad
a = {}
a["b"] = "c"

# Good
a = {"b": "c"}

SIM905

This rule will be renamed to SIM225 after its 6-month trial period is over. Please report any issues you encounter with this rule!

The trial period starts on 13th of February and will end on 13th of September 2022.

# Bad
domains = "de com net org".split()

# Good
domains = ["de", "com", "net", "org"]

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

flake8_simplify-0.17.1.tar.gz (19.5 kB view hashes)

Uploaded Source

Built Distribution

flake8_simplify-0.17.1-py3-none-any.whl (17.0 kB view hashes)

Uploaded Python 3

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page