Generator and context manager at the same time

Question

Imagine I have some code that I want it to run:

with F() as o:
    while True:
        a = o.send(2)
        print(a)

It means that the F class should return an generator and also it is context manager, generally I want a context manager to be generator too.

I tried this:

class F:

    def __enter__(self):
        return self

    def __exit__(self, *exc):
        print('exit')

    def __next__(self):
        return 5

    def __iter__(self):
        return self

As expected this will return AttributeError: 'F' object has no attribute 'send', I handled this error by adding:

def send(self, param):
    self.__next__()

but I think it is not a good way to do this, I look around and find this, but they are not using send as I want, I need that instance to be a generator.

Answer 1

You can use collections.abc and subclass your class F from Generator (manual pages). If you implement enter and exit, your instance will be generator and have context manager support as well:

from collections.abc import Generator

class F(Generator):
    def __init__(self):
        self.__my_generator = self._my_generator()
        next(self.__my_generator)   # prime the generator

    def _my_generator(self):
        while True:
            v = yield 42
            print('generator received ', v)

    # context manager interace:
    def __enter__(self):
        return self

    def __exit__(self, *exc):
        print('exit')

    # Generator interface:
    def send(self, value):
        return self.__my_generator.send(value)

    def throw(self, typ, value=None, traceback=None):
        return self.__my_generator.throw(typ, value, traceback)


with F() as o:
    while True:
        a = o.send(2)
        print('I received ', a)

Prints:

generator received  2
I received  42
...etc.