src/control_backend/agents/communication/ri_communication_agent.py

@@ -55,6 +55,7 @@ class RICommunicationAgent(BaseAgent):
         self.connected = False
         self.gesture_agent: RobotGestureAgent | None = None
         self.speech_agent: RobotSpeechAgent | None = None
+        self.visual_emotion_recognition_agent: VisualEmotionRecognitionAgent | None = None
 
     async def setup(self):
         """
@@ -218,6 +219,7 @@ class RICommunicationAgent(BaseAgent):
                         socket_address=addr,
                         bind=bind,
                     )
+                    self.visual_emotion_recognition_agent = visual_emotion_agent
                     await visual_emotion_agent.start()
                 case _:
                     self.logger.warning("Unhandled negotiation id: %s", id)
@@ -323,6 +325,9 @@ class RICommunicationAgent(BaseAgent):
 
         if self.speech_agent is not None:
             await self.speech_agent.stop()
+        
+        if self.visual_emotion_recognition_agent is not None:
+            await self.visual_emotion_recognition_agent.stop()
 
         if self.pub_socket is not None:
             self.pub_socket.close()
@@ -332,6 +337,7 @@ class RICommunicationAgent(BaseAgent):
             self.connected = True
 
     async def handle_message(self, msg: InternalMessage):
+        return 
         try:
             pause_command = PauseCommand.model_validate_json(msg.body)
             await self._req_socket.send_json(pause_command.model_dump())

src/control_backend/agents/perception/visual_emotion_recognition_agent/visual_emotion_recognition_agent.py

@@ -7,6 +7,7 @@ import numpy as np
 import zmq
 import zmq.asyncio as azmq
 from pydantic_core import ValidationError
+import struct
 
 from control_backend.agents import BaseAgent
 from control_backend.agents.perception.visual_emotion_recognition_agent.visual_emotion_recognizer import (  # noqa
@@ -88,7 +89,7 @@ class VisualEmotionRecognitionAgent(BaseAgent):
         while self._running:
             try:
                 frame_bytes = await self.video_in_socket.recv()
-
+    
                 # Convert bytes to a numpy buffer
                 nparr = np.frombuffer(frame_bytes, np.uint8)
 
@@ -97,6 +98,7 @@ class VisualEmotionRecognitionAgent(BaseAgent):
 
                 if frame_image is None:
                     # Could not decode image, skip this frame
+                    self.logger.warning("Received invalid video frame, skipping.")
                     continue
 
                 # Get the dominant emotion from each face
@@ -124,6 +126,7 @@ class VisualEmotionRecognitionAgent(BaseAgent):
             except zmq.Again:
                 self.logger.warning("No video frame received within timeout.")
 
+    
     async def update_emotions(self, prev_emotions: set[str], emotions: set[str]):
         """
         Compare emotions from previous window and current emotions,