Added Optional[] around key argument to itertools.groupby()

[achampion]

See #2788

The simplest case of:

import itertools
iterator = itertools.groupby(range(10), None)

gives the type error: error: No overload variant of "groupby" matches argument types "range", "None"

iterator = itertools.groupby(range(10))

has no error.
This example is a little contrived though legal, the case came up in the context of a Optional[] argument to function, e.g.:

T = TypeVar('T')
S = TypeVar('S')

def fn(iterable: Iterable, key: Optional[Callable[[T], S]] = None) -> Iterator[Tuple[S, Iterator[T]]]:
    iterator: Iterator[Tuple[S, Iterator[T]]] = itertools.groupby(iterable, key)
    return iterator

error: Argument 2 to "groupby" has incompatible type "Optional[Callable[[T], S]]"; expected "Callable[[T], S]"

[srittau]

Thanks!

[JelleZijlstra]

Shouldn't we add a separate overload instead? If key is None, there's nothing to constrain the value of _S.

[srittau]

@JelleZijlstra Good point, opened #2794.