Improve speech recognizer to handle empty transcriptions

src/voice_dialogue/services/speech/recognizer.py

@@ -39,8 +39,12 @@ class ASRWorker(BaseThread):
 
             voice_task.language = self.language
             voice_task.whisper_start_time = time.time()
+
             user_voice: np.array = voice_task.user_voice
             transcribed_text = self.client.transcribe(user_voice)
+            if not transcribed_text.strip():
+                continue
+
             voice_task.whisper_end_time = time.time()
 
             task_id = voice_task.id