Given I want to properly using type annotations for named tuples from the typing module:
from typing import NamedTuple, List
class Foo(NamedTuple):
my_list: List[int] = []
foo1 = Foo()
foo1.my_list.append(42)
foo2 = Foo()
print(foo2.my_list) # prints [42]
What is the best or cleanest ways to avoid the mutable default value misery in Python? I have a few ideas, but nothing really seems to be good
Using None
as default
class Foo(NamedTuple):
my_list: Optional[List[int]] = None
foo1 = Foo()
if foo1.my_list is None
foo1 = foo1._replace(my_list=[]) # super ugly
foo1.my_list.append(42)
Overwriting __new__
or __init__
won't work:
AttributeError: Cannot overwrite NamedTuple attribute __init__
AttributeError: Cannot overwrite NamedTuple attribute __new__
Special @classmethod
class Foo(NamedTuple):
my_list: List[int] = []
@classmethod
def use_me_instead(cls, my_list=None):
if not my_list:
my_list = []
return cls(my_list)
foo1 = Foo.use_me_instead()
foo1.my_list.append(42) # works!
Maybe using frozenset
and avoid mutable attributes altogether? But that won't work with Dict
s as there are no frozendict
s.
Does anyone have a good answer?
EDIT:
Blending my approach with Sebastian Wagner's idea of using a decorator, we can achieve something like this:
from typing import NamedTuple, List, Callable, TypeVar, Type, Any, cast
from functools import wraps
T = TypeVar('T')
def default_factory(**factory_kw: Callable[[], Any]) -> Callable[[Type[T]], Type[T]]:
def wrapper(wcls: Type[T], /) -> Type[T]:
@wraps(wcls.__new__)
def __new__(cls: Type[T], *args: Any, **kwargs: Any) -> T:
for key, factory in factory_kw.items():
kwargs.setdefault(key, factory())
new = super(cls, cls).__new__(cls, *args, **kwargs) # type: ignore[misc]
# This call to cast() is necessary if you run MyPy with the --strict argument
return cast(T, new)
cls_name = wcls.__name__
wcls.__name__ = wcls.__qualname__ = f'_{cls_name}'
return type(cls_name, (wcls, ), {'__new__': __new__, '__slots__': ()})
return wrapper
@default_factory(my_list=list)
class Foo(NamedTuple):
# You do not *need* to have the default value in the class body,
# but it makes MyPy a lot happier
my_list: List[int] = []
foo1 = Foo()
foo1.my_list.append(42)
foo2 = Foo()
print(f'foo1 list: {foo1.my_list}') # prints [42]
print(f'foo2 list: {foo2.my_list}') # prints []
print(Foo) # prints <class '__main__.Foo'>
print(Foo.__mro__) # prints (<class '__main__.Foo'>, <class '__main__._Foo'>, <class 'tuple'>, <class 'object'>)
from inspect import signature
print(signature(Foo.__new__)) # prints (_cls, my_list: List[int] = [])
Run it through MyPy, and MyPy informs us that the revealed type of foo1
and foo2
is still "Tuple[builtins.list[builtins.int], fallback=__main__.Foo]"
Original answer below.
How about this? (Inspired by this answer here):
from typing import NamedTuple, List, Optional, TypeVar, Type
class _Foo(NamedTuple):
my_list: List[int]
T = TypeVar('T', bound="Foo")
class Foo(_Foo):
"A namedtuple defined as `_Foo(mylist)`, with a default value of `[]`"
__slots__ = ()
def __new__(cls: Type[T], mylist: Optional[List[int]] = None) -> T:
mylist = [] if mylist is None else mylist
return super().__new__(cls, mylist) # type: ignore
f, g = Foo(), Foo()
print(isinstance(f, Foo)) # prints "True"
print(isinstance(f, _Foo)) # prints "True"
print(f.mylist is g.mylist) # prints "False"
Run it through MyPy and the revealed type of f
and g
will be: "Tuple[builtins.list[builtins.int], fallback=__main__.Foo]"
.
I'm not sure why I had to add the # type: ignore
to get MyPy to stop complaining — if anybody can enlighten me on that, I'd be interested. Seems to work fine at runtime.