SFSpeechURLRecognitionRequest Doesn't Output Results from Speech Recognition Until After Thread Calling It Completes - swift

Question

I am trying to use SFSpeechURLRecognitionRequest to transcribe audio files in a terminal application. While I have it working in a preliminary form, I'm running into an odd issue. It appears to only output the voice recognition results (both partial and complete) after the main thread terminates in my test applications. Note I am a Swift noob, so I might be missing something obvious.
Below I have a complete Xcode Playground application which demonstrates the issue. The output from the playground writes Playground Execution Complete and then I begin receiving partial outputs followed by the final output. Note that if I add a sleep(5) prior to the print it will wait 5 seconds and then output the print, and only then after the main thread has concluded begin processing the text. I have seen similar behavior in a GUI test application, where it only begins processing the text after the method call kicking off the request completes.
I have tried repeatedly checking the state of the task that is returned, sleeping between each check with no luck.
I have also tried calling the recognition task inside a DispatchQueue, which appears to run successfully in the background based on CPU usage, but the Partial and Final prints never appear until the application completes, at which point the console fills up with Partials followed by the Final.
Does anyone know of a way to have the speech recognition begin processing without the application thread completing? Ideally I would like to be able to kick it off and sleep for brief periods repeatedly, checking if the recognition task has completed in between each.
Edited below to match version immediately prior to figuring out the solution.
import Speech
var complete = false
SFSpeechRecognizer.requestAuthorization {
authStatus in
DispatchQueue.main.async {
if authStatus == .authorized {
print("Good to go!")
} else {
print("Transcription permission was declined.")
exit(1)
}
}
}
guard let myRecognizer = SFSpeechRecognizer() else {
print("Recognizer not supported for current locale!")
exit(1)
}
if !myRecognizer.isAvailable {
// The recognizer is not available right now
print("Recognizer not available right now")
exit(1)
}
if !myRecognizer.supportsOnDeviceRecognition {
print("On device recognition not possible!")
exit(1)
}
let path_to_wav = NSURL.fileURL(withPath: "/tmp/harvard.wav", isDirectory: false)
let request = SFSpeechURLRecognitionRequest(url: path_to_wav)
request.requiresOnDeviceRecognition = true
print("About to create recognition task...")
myRecognizer.recognitionTask(with: request) {
(result, error) in
guard let result = result else {
// Recognition failed, so check error for details and handle it
print("Recognition failed!!!")
print(error!)
exit(1)
}
if result.isFinal {
print("Final: \(result.bestTranscription.formattedString)")
complete = true
} else {
print("Partial: \(result.bestTranscription.formattedString)")
}
}
print("Playground execution complete.")

Answer 1

I figured it out! sleep doesn't actually let background tasks execute. Instead by adding the following:
let runLoop = RunLoop.current
let distantFuture = NSDate.distantFuture as NSDate
while complete == false && runLoop.run(mode: RunLoop.Mode.default, before: distantFuture as Date) {}
to the end just before the last print works (results begin appearing immediately, and the final print prints right after the final results).