[FIXED] How to perform unit tests of asynchronous functions?

Issue

I'm using Bleak to discover & connect to the nearest Bluetooth Low Energy (BLE) device, and I'm currently writing unit tests (using pytest).

I am new to Python tests and I don't know what to do with these patch/mock to make it work on async functions.

I do not know if I should use the actual function, or apply patches to the default functions to make the test executable without the BLE dongle.

Here is a sample of code (improvement of discover.py) :

def list(op_sys: str) -> list:
    """list BLE devices

    Returns:
        list: status & list or error message
    """
    import asyncio, platform
    from bleak import discover

    async def run() -> list:
        """discover BLE devices

        Returns:
            list: status & list or error message
        """
        BLElist = []
        try:
            devices = await discover()
            for d in devices:
                print("'%s'" % d.name) # list devices
                BLElist.append(d.name)
            return 'success', BLElist
        except:
            return 'error', 'You don\'t have any BLE dongle.'

    # linux = 3.6, windows = 3.7, need a new loop to work
    if op_sys == "Windows":
        asyncio.set_event_loop(asyncio.new_event_loop())

    loop = asyncio.get_event_loop()
    return loop.run_until_complete(run())

I'm wondering if I should rewrite the function to move the run() part outside, and mock it.

Solution

The outer function list(op_sys) -> list is not async because it does a call to loop.run_until_complete.

So that one can be unit tested like any synchronous python function.

If you want to unit test async functions like the inner function run() -> list, take a look over here: https://pypi.org/project/asynctest/.

Answered By - Freek Wiekmeijer