OOP

These checks ensures that you use Python’s version of OOP correctly.

There are different gotchas in Python to write beautiful classes and using objects correctly. That’s the place we collect these kind of rules.

Summary

BuiltinSubclassViolation

WPS600 — Forbid subclassing lowercase builtins.

ShadowedClassAttributeViolation

WPS601 — Forbid shadowing class level attributes with instance level attributes.

StaticMethodViolation

WPS602 — Forbid @staticmethod decorator.

BadMagicMethodViolation

WPS603 — Forbid certain magic methods.

WrongClassBodyContentViolation

WPS604 — Forbid incorrect nodes inside class definitions.

MethodWithoutArgumentsViolation

WPS605 — Forbid methods without any arguments.

WrongBaseClassViolation

WPS606 — Forbid anything other than a class as a base class.

WrongSlotsViolation

WPS607 — Forbid incorrect __slots__ definition.

WrongSuperCallViolation

WPS608 — Forbid super() with parameters or outside of methods.

DirectMagicAttributeAccessViolation

WPS609 — Forbid directly calling certain magic attributes and methods.

AsyncMagicMethodViolation

WPS610 — Forbid certain async magic methods.

YieldMagicMethodViolation

WPS611 — Forbid yield inside of certain magic methods.

UselessOverwrittenMethodViolation

WPS612 — Forbid useless overwritten methods.

WrongSuperCallAccessViolation

WPS613 — Forbid super() with incorrect method or property access.

WrongDescriptorDecoratorViolation

WPS614 — Forbids descriptors in regular functions.

UnpythonicGetterSetterViolation

WPS615 — Forbids to use getters and setters in objects.

BuggySuperContextViolation

WPS616 — Calling super() in buggy context.

LambdaAttributeAssignedViolation

WPS617 — Forbid using lambda as an assigned attribute.

BuiltinSubclassViolation[source]

WPS600 — Forbid subclassing lowercase builtins.

We forbid to subclass builtins like int, str, bool, etc. We allow to subclass object and type, warnings, and exceptions.

See ALLOWED_BUILTIN_CLASSES for the whole list of whitelisted names.

Reasoning:

It is almost never a good idea (unless you do something sneaky) to subclass primitive builtins.

Solution:

Use custom objects around some wrapper. Use magic methods to emulate the desired behaviour.

Example:

# Correct:
class Some: ...
class MyValueException(ValueError): ...

# Wrong:
class MyInt(int): ...

Added in version 0.10.0.

Changed in version 0.11.0.

Changed in version 1.0.0: Allows subclassing builtins in enum.Enum definitions.

full_code: ClassVar[str] = 'WPS600'
summary: ClassVar[str] = 'Forbid subclassing lowercase builtins.'
ShadowedClassAttributeViolation[source]

WPS601 — Forbid shadowing class level attributes with instance level attributes.

Reasoning:

This way you will have two attributes inside your __mro__ chain: one from instance and one from class. It might cause errors. Needless to say, that this is just pointless to do so.

Also, if you ever want to optimise your code with a tool like mypyc, this rule is a requirement.

Solution:

Use either class attributes or instance attributes. Use ClassVar type on fields that are declared as class attributes.

Note, that we cannot find shadowed attributes that are defined in parent classes. That’s where ClassVar is required for mypy to check it for you.

Example:

# Correct:
from typing import ClassVar

class First:
    field: ClassVar[int] = 1

class Second:
    field: int

    def __init__(self) -> None:
        self.field = 1

# Wrong:
class Some:
    field = 1

    def __init__(self) -> None:
        self.field = 1

Added in version 0.10.0.

Changed in version 0.11.0.

Changed in version 0.14.0.

Changed in version 1.0.0: Allow to shadow class attribute names in @dataclass classes.

full_code: ClassVar[str] = 'WPS601'
summary: ClassVar[str] = 'Forbid shadowing class level attributes with instance level attributes.'
StaticMethodViolation[source]

WPS602 — Forbid @staticmethod decorator.

Reasoning:

Static methods are not required to be inside the class. Because they even do not have access to the current instance.

Solution:

Use instance methods, @classmethod, or functions instead.

Added in version 0.1.0.

Changed in version 0.11.0.

full_code: ClassVar[str] = 'WPS602'
summary: ClassVar[str] = 'Forbid ``@staticmethod`` decorator.'
BadMagicMethodViolation[source]

WPS603 — Forbid certain magic methods.

Reasoning:

We forbid to use magic methods related to the forbidden language parts. Likewise, we forbid to use del keyword, so we forbid to use all magic methods related to it.

Solution:

Refactor your code to use custom methods instead. It will give more context to your app.

See MAGIC_METHODS_BLACKLIST for the full blacklist of the magic methods.

Added in version 0.1.0.

Changed in version 0.11.0.

full_code: ClassVar[str] = 'WPS603'
summary: ClassVar[str] = 'Forbid certain magic methods.'
WrongClassBodyContentViolation[source]

WPS604 — Forbid incorrect nodes inside class definitions.

Reasoning:

Python allows us to have conditions, context managers, and even infinite loops inside class definitions. On the other hand, only methods, attributes, and docstrings make sense. So, we discourage using anything except these nodes in class bodies.

Solution:

If you have complex logic inside your class definition, most likely that you do something wrong. There are different options to refactor this mess. You can try metaclasses, decorators, builders, and other patterns.

Example:

# Wrong:
class Test:
    for _ in range(10):
        print('What?!')

We also allow some nested classes, check out NestedClassViolation for more information.

Added in version 0.7.0.

Changed in version 0.11.0.

full_code: ClassVar[str] = 'WPS604'
summary: ClassVar[str] = 'Forbid incorrect nodes inside ``class`` definitions.'
MethodWithoutArgumentsViolation[source]

WPS605 — Forbid methods without any arguments.

Reasoning:

Methods without arguments are allowed to be defined, but almost impossible to use, if they are not @staticmethods. Furthermore, they don’t have an access to self, so cannot access the inner state of the object. It might be an intentional design or just a typo.

Solution:

Move any methods with arguments to raw functions. Or just add an argument if it is actually required.

Example:

# Correct:
class Test:
    def method(self): ...

# Wrong:
class Test:
    def method(): ...

Added in version 0.7.0.

Changed in version 0.11.0.

Changed in version 1.1.0: Allows usage of @staticmethod with no arguments.

full_code: ClassVar[str] = 'WPS605'
summary: ClassVar[str] = 'Forbid methods without any arguments.'
WrongBaseClassViolation[source]

WPS606 — Forbid anything other than a class as a base class.

We only check base classes and not keywords. They can be anything you need.

Reasoning:

In Python you can specify anything in the base classes slot. In runtime this expression will be evaluated and executed. We need to prevent dirty hacks in this field.

Solution:

Use only attributes, names, and types to be your base classes. Use annotation future import in case you use strings in base classes.

Example:

# Correct:
class Test(module.ObjectName, MixinName, keyword=True): ...
class GenericClass(Generic[ValueType]): ...

# Wrong:
class Test((lambda: object)()): ...

Added in version 0.7.0.

Changed in version 0.7.1.

Changed in version 0.11.0.

Changed in version 0.12.0.

full_code: ClassVar[str] = 'WPS606'
summary: ClassVar[str] = 'Forbid anything other than a class as a base class.'
WrongSlotsViolation[source]

WPS607 — Forbid incorrect __slots__ definition.

Things that this rule checks:

  • That __slots__ is a tuple, name, attribute, star, or call

  • That __slots__ do not have duplicates

  • That __slots__ do not have empty strings or invalid python names

Reasoning:

__slots__ is a very special attribute. It completely changes your class. So, we need to be careful with it. We should not allow anything rather than tuples to define slots, we also need to check that fields defined in __slots__ are unique.

Solution:

Use tuples with unique elements to define __slots__ attribute. Use snake_case to define attributes in __slots__.

Example:

# Correct:
class Test:
    __slots__ = ('field1', 'field2')

class Other(Test):
    __slots__ = (*Test.__slots__, 'child')

# Wrong:
class Test:
    __slots__ = ['field1', 'field2', 'field2']

Note, that we do ignore all complex expressions for this field. So, we only check raw literals.

Added in version 0.7.0.

Changed in version 0.11.0.

Changed in version 0.12.0.

full_code: ClassVar[str] = 'WPS607'
summary: ClassVar[str] = 'Forbid incorrect ``__slots__`` definition.'
WrongSuperCallViolation[source]

WPS608 — Forbid super() with parameters or outside of methods.

Reasoning:

super() is a very special function. It implicitly relies on the context where it is used and parameters passed to it. So, we should be very careful with parameters and context.

Solution:

Use super() without arguments and only inside methods.

Example:

# Correct:
super().__init__()

# Wrong:
super(ClassName, self).__init__()

Added in version 0.7.0.

Changed in version 0.11.0.

full_code: ClassVar[str] = 'WPS608'
summary: ClassVar[str] = 'Forbid ``super()`` with parameters or outside of methods.'
DirectMagicAttributeAccessViolation[source]

WPS609 — Forbid directly calling certain magic attributes and methods.

Reasoning:

Certain magic methods are only meant to be called by particular functions or operators, not directly accessed.

Solution:

Use special syntax constructs that will call underlying magic methods.

Example:

# Correct:
super().__init__()
mymodule.__name__

# Wrong:
foo.__str__()  # use `str(foo)`
2..__truediv__(2)  # use `2 / 2`
d.__delitem__('a')  # use del d['a']

Note, that it is possible to directly use these magic attributes with self, cls, and super() as base names. We allow this because a lot of internal logic relies on these methods.

Added in version 0.8.0.

Changed in version 0.11.0.

Changed in version 0.16.0.

Changed in version 1.0.0: No longer produced, kept here for historic reasons. This is covered with ruff and pylint linters. See PLC2801.

disabled_since: ClassVar[str | None] = '1.0.0'
full_code: ClassVar[str] = 'WPS609'
summary: ClassVar[str] = 'Forbid directly calling certain magic attributes and methods.'
AsyncMagicMethodViolation[source]

WPS610 — Forbid certain async magic methods.

We allow to make __anext__, __aenter__, __aexit__ async. We allow to make __aiter__ async if it is a generator (contains yield). We also allow custom magic methods to be async.

See ASYNC_MAGIC_METHODS_BLACKLIST for the whole list of blacklisted async magic methods.

Reasoning:

Defining the magic methods as async which are not supposed to be async would not work as expected.

Solution:

Do not make this magic method async.

Example:

# Correct:
class Test:
    def __lt__(self, other): ...

# Wrong:
class Test:
    async def __lt__(self, other): ...

Added in version 0.12.0.

full_code: ClassVar[str] = 'WPS610'
summary: ClassVar[str] = 'Forbid certain async magic methods.'
YieldMagicMethodViolation[source]

WPS611 — Forbid yield inside of certain magic methods.

We allow to make __iter__ a generator. We allow to make __aiter__ an async generator. See YIELD_MAGIC_METHODS_BLACKLIST for the whole list of blacklisted generator magic methods.

Reasoning:

Python’s datamodel is strict. You cannot make generators from random magic methods. This rule enforces it.

Solution:

Remove yield from a magic method or rename it to be a custom method.

Example:

 # Correct:
class Example:
    def __init__(self):
        ...

# Wrong:
class Example:
    def __init__(self):
        yield 10

Added in version 0.3.0.

Changed in version 0.11.0.

Changed in version 0.12.0.

full_code: ClassVar[str] = 'WPS611'
summary: ClassVar[str] = 'Forbid ``yield`` inside of certain magic methods.'
UselessOverwrittenMethodViolation[source]

WPS612 — Forbid useless overwritten methods.

Reasoning:

Overwriting method without any changes does not have any positive impact.

Solution:

Do not overwrite method in case you do not want to do any changes inside it.

Example:

# Correct:
class Test(Base):
    def method(self, argument):
        super().method(argument)
        return argument  # or None, or anything!

# Wrong:
class Test:
    def method(self, argument):
        return super().method(argument)

Added in version 0.12.0.

Changed in version 1.0.0: Ignores cases when super().method is called when function parameters have defaults. Because defaults might be different.

full_code: ClassVar[str] = 'WPS612'
summary: ClassVar[str] = 'Forbid useless overwritten methods.'
WrongSuperCallAccessViolation[source]

WPS613 — Forbid super() with incorrect method or property access.

Reasoning:

Can only use super() method that matches the following context. super().some() and super().some in Child.some(), and super().prop and super().prop() in Child.prop

Solution:

Use super() methods and properties with the correct context.

Example:

# Correct:
class Child(Parent):
    def some_method(self):
        original = super().some_method()

# Wrong:
class Child(Parent):
    def some_method(self):
        other = super().other_method()

Added in version 0.13.0.

full_code: ClassVar[str] = 'WPS613'
summary: ClassVar[str] = 'Forbid ``super()`` with incorrect method or property access.'
WrongDescriptorDecoratorViolation[source]

WPS614 — Forbids descriptors in regular functions.

Forbids using @staticmethod, @classmethod and @property for functions not in class.

Reasoning:

Descriptors like @staticmethod, @classmethod and @property do magic only as methods. We would want to warn users if the descriptors are used on regular functions.

Solution:

Do not use @staticmethod, @classmethod and @property on regular functions or wrap the functions into a Class.

Example:

# Correct:
class TestClass:
    @property
    def my_method():
        ...

# Wrong:
@property
def my_function():
    ...

Added in version 0.15.0.

full_code: ClassVar[str] = 'WPS614'
summary: ClassVar[str] = 'Forbids descriptors in regular functions.'
UnpythonicGetterSetterViolation[source]

WPS615 — Forbids to use getters and setters in objects.

Reasoning:

Python does not need this abstraction.

Solution:

Either use @property or make the attribute public and change it directly.

Example:

# Correct:
class Example:
    def __init__(self):
        self.attribute = None

# Wrong:
class Example:
    def __init__(self):
        self._attribute = None

    def set_attribute(self, value):
        ...

    def get_attribute(self):
        ...

Added in version 0.15.0.

full_code: ClassVar[str] = 'WPS615'
summary: ClassVar[str] = 'Forbids to use getters and setters in objects.'
BuggySuperContextViolation[source]

WPS616 — Calling super() in buggy context.

Reasoning:

Call to super() without arguments will cause unexpected TypeError in a number of specific contexts, e.g. dict/set/list comprehensions and generator expressions.

Read more: https://bugs.python.org/issue46175

Solution:

Use super(cls, self) instead in those cases.

Example:

# Correct:
(super(cls, self).augment(it) for it in items)

# Wrong:
(super().augment(it) for it in items)

Added in version 0.18.0.

full_code: ClassVar[str] = 'WPS616'
summary: ClassVar[str] = 'Calling super() in buggy context.'
LambdaAttributeAssignedViolation[source]

WPS617 — Forbid using lambda as an assigned attribute.

Reasoning:

Assigning lambda as an attribute does not make much sense. And can lead to potentially incorrect code.

Solution:

Use def statements to create regular or class methods.

Example:

# Correct:
class Used:
    def login(self): ...

# Wrong:
class User:
    def __init__(self):
        self.login = lambda: ...

Added in version 1.0.0.

full_code: ClassVar[str] = 'WPS617'
summary: ClassVar[str] = 'Forbid using ``lambda`` as an assigned attribute.'