Minor improvement

examples/reachy2-remote/parse_pose.py

@@ -183,6 +183,7 @@ for event in node:
                                 metadata={"encoding": "xyzrpy", "duration": "1"},
                             )
                             event = wait_for_event(id="translate_base")
+                            arm_holding_object = "right"
                     else:
                         y += 0.03
                         node.send_output(
@@ -236,6 +237,7 @@ for event in node:
                                 metadata={"encoding": "xyzrpy", "duration": "1"},
                             )
                             event = wait_for_event(id="translate_base")
+                            arm_holding_object = "left"
                 case "release":
                     if len(values) == 0:
                         continue
@@ -254,6 +256,7 @@ for event in node:
                     if abs(y) > 0.5:
                         continue
                     if arm_holding_object is None:
+                        print("No arm holding object!!!")
                         continue
                     elif arm_holding_object == "right":
                         node.send_output(

examples/reachy2-remote/parse_whisper.py

@@ -58,9 +58,10 @@ for event in node:
                 node.send_output("points", pa.array([], type=pa.float64()))
             elif "follow" in text:
                 text = f"Given the prompt: {text}. Output the bounding boxes for the given followed object"
+                node.send_output("look_ahead", pa.array([1.0]))
+                time.sleep(0.5)  # Sync image
                 node.send_output("text", pa.array([text]), {"image_id": "image_left"})
                 node.send_output("follow_pose", pa.array([1.0]))
-                node.send_output("look_ahead", pa.array([1.0]))
             elif "raise your arms" in text:
                 node.send_output("raise_arm_pose", pa.array([1.0]))
             elif "grab " in text: