Decorator for iterator scoping / aclosing

[maxfischer2781]

While the pattern of scoped_iter and aclosing solves resource leaks (see PEP 533 as always) it feels a bit unwieldy to use. Many functions need the resource in the entire scope and only scope it as a safety measure, not for functionality. The explicit async with for scoping increases the indentation level and pulls in unrelated resource management; a practical shortcoming is that adding scoping leads to a diff of the entire function due to indent.

It would be interesting to have a decorator to automatically scope arguments to the entire function:

# explicit scoping
async def compress(data: AnyIterable[T], selectors: AnyIterable[bool]) -> AsyncIterator[T]:
    async with ScopedIter(data) as data_iter, ScopedIter(selectors) as selectors_iter:
        async for item, keep in zip(data_iter, selectors_iter):
            if keep:
                yield item

# implicit scoping
@autoscope
async def compress(data: AnyIterable[T], selectors: AnyIterable[bool]) -> AsyncIterator[T]:
    async for item, keep in zip(data_iter, selectors_iter):
        if keep:
            yield item

This isn't quite straightforward, since it needs a mechanism to identify which arguments to scope how. For example, many utilities of asyncstdlib itself take *iterables: AnyIterable[T] and would have to scope each iterable.