Start Python asyncio courutines execution immediately

836 views Asked by At

I am trying to understand Python asyncio module. It seems to be much more complicated than similar functionalities in other languages.

Here is an excerpt from documentation:

Calling a coroutine does not start its code running – it is just a generator, and the coroutine object returned by the call is really a generator object, which doesn’t do anything until you iterate over it. In the case of a coroutine object, there are two basic ways to start it running: call yield from coroutine from another coroutine (assuming the other coroutine is already running!), or schedule its execution using the async() function or the BaseEventLoop.create_task() method.

Coroutines (and tasks) can only run when the event loop is running.

Does it mean that it is impossible to achieve similar execution flow like here with Hack?

function main() : void
{
    $info = Vector {};
    for ($i = 0; $i < 5; $i++) {
        $info[] = genInfo($i);
    }
 
    echo '[main] Calling Async Function'."\n";
    $asyncCall = GenVectorWaitHandle::create($info);
 
    echo '[main] Doing whatever...'."\n";
 
    echo '[main] Time To Request Async Return Information'."\n";
    $output = $asyncCall->join();
 
    // Output Vector Data
    foreach ($output as $key => $value)
    {
        echo '['.$key.'] => '.$value."\n";
    }
}
 
async function genInfo(int $id): Awaitable<String> {
    echo '[genInfo] Generating '.$id."\n";
 
    $tmp = [];
    $tmp['id'] = $id;
    $tmp['start'] = date('H:i:s');
 
    //do some heavy working.
    await SleepWaitHandle::create(mt_rand(1000000,15*1000000));
 
    $tmp['end'] = date('H:i:s');
 
    echo "[genInfo] Completed $id\n";
 
    return json_encode($tmp);
}
 
main();
 
/*
 * This is a modified example from http://kernelcurry.com/blog/asynchronous-hack/
 */
 
//Output
[genInfo] Generating 0
[genInfo] Generating 1
[genInfo] Generating 2
[genInfo] Generating 3
[genInfo] Generating 4
[main] Calling Async Function
[main] Doing whatever...
[main] Time To Request Async Return Information
[genInfo] Completed 0
[genInfo] Completed 3
[genInfo] Completed 2
[genInfo] Completed 1
[genInfo] Completed 4
[0] => {"id":0,"start":"19:46:00","end":"19:46:03"}
[1] => {"id":1,"start":"19:46:00","end":"19:46:08"}
[2] => {"id":2,"start":"19:46:00","end":"19:46:05"}
[3] => {"id":3,"start":"19:46:00","end":"19:46:04"}
[4] => {"id":4,"start":"19:46:00","end":"19:46:08"}

As you can see, genInfo() execution starts immediately after call and yields on the await statement. I think that such execution flow is much more natural then this one with Python.

Is there any way to achieve the same in Python?

I tried to iterate coroutines generators one time (in order to allow them to execute and approach first yield) before passing them to loop.run_until_complete(), but without success.

0

There are 0 answers