Pyright has problem with combo of `@property` and `@cache`

[MapleCCC]

Describe the bug
The combo use of @property and @cache is backed by the document of functools.cached_property, so it should be considered a valid use case. However, pyright seems to have some bugs in handling such case.

To Reproduce

from functools import cache

class Rectangle:
    def __init__(self, length: int, width: int) -> None:
        self._length = length
        self._width = width

    @property
    @cache
    def area(self) -> int:
        return self._length * self._width

def is_large_rectangle(rec: Rectangle) -> bool:
    return rec.area >= 100

rec = Rectangle(10, 10)
print(is_large_rectangle(rec))

Pyright complains that Operator ">=" not supported for types "_lru_cache_wrapper[int]" and "Literal[100]".

Expected behavior
There should be no error for the above code.

VS Code extension or command-line
Pyright command-line tool 1.1.158

[erictraut]

This looks to be a limitation or a bug in the typeshed type definition for cache. I think pyright is applying the type information correctly here. It matches the behavior of mypy in this case.

    def cache(__user_function: Callable[..., _T]) -> _lru_cache_wrapper[_T]: ...

class _lru_cache_wrapper(Generic[_T]):
    __wrapped__: Callable[..., _T]
    def __call__(self, *args: Hashable, **kwargs: Hashable) -> _T: ...
    def cache_info(self) -> _CacheInfo: ...
    def cache_clear(self) -> None: ...

This definition implies that the resulting cache object is callable, and calling it will return the cached value. But that's not how it works at runtime.

I don't know why cache was defined as it currently is in functools.pyi. My advice is to file an issue in the typeshed repo and start a discussion there. We frequently pull changes from typeshed into pyright, so if a fix is made there, it will soon be fixed in pyright as well.

[MapleCCC]

@erictraut Thank you for your patient reply. To be honest, I am not familiar with the type annotation enough to understand your introspection. If it's convenient, could you consider open an issue in the typeshed library? I really don't know how to describe this complex type annotation thing 😵

[erictraut]

This will be fixed in the next release.

[erictraut]

This is now addressed in pyright 1.1.159, which I just published. It will also be included in the next release of pylance.