Type hinting generator in Python 3.6

According to PEP-484, we should be able to type hinting a generator function as follows:

from typing import Generator

def generate() -> Generator[int, None, None]:
    for i in range(10):
        yield i

for i in generate():
    print(i)

However, the list comprehension gives the following error in PyCharm.

Expected 'collections.Iterable', got 'Generator[int, None, None]' instead less... (⌘F1)

Any idea why PyCharm is considering this as error?

A few clarification after reading some answers. I am using PyCharm Community Edition 2016.3.2 (the latest version) and have imported the typing.Generator (updated in the code). The above code runs just fine, but PyCharm considers this an error:

So, I'm wondering if this is actually an error or an unsupported feature in PyCharm.

Solution 1:

You need to import the typing module. As per docs:

The return type of generator functions can be annotated by the generic type Generator[yield_type, send_type, return_type] provided by typing.py module

Try this way instead:

from typing import Generator


def generate() -> Generator[int, None, None]:
    for i in range(10):
        yield i

The above will have the desired result:

l = [i for i in generate()]

Output:

[0, 1, 2, 3, 4, 5, 6, 7, 8, 9]

As pointed out in the comments, you might not use the last version of PyCharm. Try switching to version 2016.3.2 and you might be fine. Unfortunately this is a well-known bug, as per @AshwiniChaudhary comment.

More, the reported issue (for the last version of PyCharm) was submitted on December, last year. They probably fixed it and pushed the modifications into the same version.

Solution 2:

This isn't a direct answer to the question, but I think it is a better solution.

I'm using the typing specification below, using Iterator[int] instead of Generator. The validation is OK. I think it is a lot clearer. It better describes the code intention and is recommended by Python docs.

from typing import Iterator

def generate() -> Iterator[int]:
    for i in range(10):
        yield i

It would also allow future refactorings if you change your Generator for a list or other iterable.

I'm using Visual Studio Code with PyLance for typing validation. PyCharm mypy should have the same behavior.