Run the hotword detected sound in another thread, or it's likely to mess up with the timing of the hotword-assistant interaction

platypush/backend/assistant/snowboy/__init__.py

@@ -6,6 +6,7 @@
 import json
 import os
 import subprocess
+import threading
 import time
 
 from platypush.backend import Backend
@@ -110,9 +111,11 @@ class AssistantSnowboyBackend(Backend):
                 raise AttributeError('No voice_model_file specified for model {}'.format(name))
 
             model_file = os.path.abspath(os.path.expanduser(model_file))
-            detect_sound = os.path.abspath(os.path.expanduser(detect_sound))
             assistant_plugin_name = conf.get('assistant_plugin')
 
+            if detect_sound:
+                detect_sound = os.path.abspath(os.path.expanduser(detect_sound))
+
             if not os.path.isfile(model_file):
                 raise FileNotFoundError('Voice model file {} does not exist or it not a regular file'.format(model_file))
 
@@ -128,13 +131,16 @@ class AssistantSnowboyBackend(Backend):
         """
         Callback called on hotword detection
         """
-
-        def callback():
         try:
             import snowboydecoder
         except ImportError:
             import snowboy.snowboydecoder as snowboydecoder
 
+
+        def sound_thread(sound):
+            snowboydecoder.play_audio_file(sound)
+
+        def callback():
             self.bus.post(HotwordDetectedEvent(hotword=hotword))
             model = self.models[hotword]
 
@@ -143,7 +149,7 @@ class AssistantSnowboyBackend(Backend):
             assistant_language = model.get('assistant_language')
 
             if detect_sound:
-                snowboydecoder.play_audio_file(detect_sound)
+                threading.Thread(target=sound_thread, args=(detect_sound,)).start()
 
             if assistant_plugin:
                 assistant_plugin.start_conversation(language=assistant_language)
@@ -158,7 +164,7 @@ class AssistantSnowboyBackend(Backend):
     def run(self):
         super().run()
         self.detector.start(detected_callback=[
-            lambda: self.hotword_detected(hotword)
+            self.hotword_detected(hotword)
             for hotword in self.models.keys()
         ])