Add TTS events

Author: kPatch

coffee_ws/src/coffee_voice_agent/coffee_voice_agent/voice_agent_bridge.py

@@ -194,6 +194,21 @@ async def _handle_websocket_message(self, message: str):
                 # Handle startup/ready events from voice agent
                 self.get_logger().info(f"Voice agent startup: {data.get('message', 'Ready')} (version: {data.get('version', 'unknown')})")
 
+            elif message_type == 'TTS_EVENT':
+                # Handle TTS started/finished events - parse nested data structure
+                event_data = data.get('data', {})
+                event = event_data.get('event', 'unknown')
+                emotion = event_data.get('emotion', 'unknown')
+                source = event_data.get('source', 'unknown')
+                text = event_data.get('text', '')
+                text_preview = text[:50] + "..." if len(text) > 50 else text
+                
+                self.get_logger().info(f"TTS {event}: emotion={emotion}, source={source}, text='{text_preview}'")
+                
+                # TODO: Publish to ROS2 topics when robot coordination is ready
+                # self.tts_started_pub.publish(...) for event == "started"
+                # self.tts_finished_pub.publish(...) for event == "finished"
+                
             else:
                 self.get_logger().warn(f"Unknown message type from voice agent: {message_type}")
 

coffee_ws/src/coffee_voice_agent/scripts/state/state_manager.py

@@ -524,8 +524,17 @@ async def say_with_emotion(self, text: str, emotion: str = None):
         logger.info(f"🔍 DEBUG: say_with_emotion emotion: {emotion}")
 
         if self.session:
+            # Send TTS_STARTED event
+            await self._send_tts_event("started", text, emotion or self.current_emotion, "manual")
+            
             logger.info("🔍 DEBUG: Calling session.say() directly (bypasses llm_node)")
-            await self.session.say(text)
+            handle = await self.session.say(text)
+            
+            # Wait for TTS completion
+            await handle.wait_for_playout()
+            
+            # Send TTS_FINISHED event
+            await self._send_tts_event("finished", text, emotion or self.current_emotion, "manual")
 
             if emotion:
                 logger.info(f"🎭 Speaking with emotion: {emotion}")
@@ -537,4 +546,18 @@ def get_random_greeting(self) -> str:
         """Get a random greeting from the greeting pool"""
         from utils.greeting_data import get_random_greeting
 
-        return get_random_greeting() 
+        return get_random_greeting()
+
+    async def _send_tts_event(self, event: str, text: str, emotion: str, source: str):
+        """Send TTS event through agent's WebSocket connection"""
+        if self.agent and hasattr(self.agent, '_send_websocket_event'):
+            event_data = {
+                "event": event,
+                "text": text[:100] + "..." if len(text) > 100 else text,  # Truncate long text
+                "emotion": emotion,
+                "source": source,
+                "timestamp": datetime.now().isoformat()
+            }
+            await self.agent._send_websocket_event("TTS_EVENT", event_data)
+        else:
+            logger.debug(f"Cannot send TTS {event} event - no agent WebSocket connection")