refactor: 移除硬编码动作，动画完全由JSON配置驱动

- motion.py: 移除所有PendingAction调用，只保留set_state()
- animation_player.py: 移除未使用的ActionParams类和_actions字典
- 状态动画参数（offset）由conversation_animations.json控制

reachy_mini_ha_voice/animation_player.py

@@ -85,19 +85,22 @@ class AnimationPlayer:
         self._phase_x: float = 0.0
         self._phase_y: float = 0.0
         self._phase_z: float = 0.0
-        self._load_animations()
+        self._load_config()
 
-    def _load_animations(self) -> None:
-        """Load animations from JSON file."""
+    def _load_config(self) -> None:
+        """Load animations and actions from JSON file."""
         if not _ANIMATIONS_FILE.exists():
             _LOGGER.warning("Animations file not found: %s", _ANIMATIONS_FILE)
             return
         try:
             with open(_ANIMATIONS_FILE, "r", encoding="utf-8") as f:
                 data = json.load(f)
+
             settings = data.get("settings", {})
             self._amplitude_scale = settings.get("amplitude_scale", 1.0)
             self._transition_duration = settings.get("transition_duration_s", 0.3)
+
+            # Load animations
             animations = data.get("animations", {})
             for name, params in animations.items():
                 self._animations[name] = AnimationParams(
@@ -126,6 +129,7 @@ class AnimationPlayer:
                     z_frequency_hz=params.get("z_frequency_hz", 0.0),
                     phase_offset=params.get("phase_offset", 0.0),
                 )
+
             _LOGGER.info("Loaded %d animations", len(self._animations))
         except Exception as e:
             _LOGGER.error("Failed to load animations: %s", e)

reachy_mini_ha_voice/animations/conversation_animations.json

@@ -7,27 +7,27 @@
       "frequency_hz": 0.0
     },
     "listening": {
-      "description": "Attentive pose while listening to user",
-      "pitch_amplitude_rad": 0.05,
-      "pitch_offset_rad": -0.03,
+      "description": "Attentive pose while listening to user - slight forward lean",
+      "pitch_offset_rad": -0.05,
+      "pitch_amplitude_rad": 0.03,
       "z_amplitude_m": 0.003,
       "antenna_amplitude_rad": 0.2,
       "antenna_move_name": "both",
       "frequency_hz": 0.6
     },
     "thinking": {
-      "description": "Processing/thinking animation - gentle head movement",
+      "description": "Processing/thinking animation - head tilted with gentle sway",
+      "roll_offset_rad": 0.08,
       "pitch_amplitude_rad": 0.03,
       "yaw_amplitude_rad": 0.05,
-      "roll_amplitude_rad": 0.06,
-      "roll_offset_rad": 0.05,
+      "roll_amplitude_rad": 0.04,
       "z_amplitude_m": 0.003,
       "antenna_amplitude_rad": 0.25,
       "antenna_move_name": "wiggle",
       "frequency_hz": 0.4
     },
     "speaking": {
-      "description": "Speaking animation - multi-frequency natural head sway (inspired by reachy_mini_conversation_app)",
+      "description": "Speaking animation - multi-frequency natural head sway",
       "pitch_amplitude_rad": 0.08,
       "pitch_frequency_hz": 2.2,
       "yaw_amplitude_rad": 0.13,
@@ -54,10 +54,10 @@
     },
     "sad": {
       "description": "Sad/negative response - head droops",
-      "pitch_amplitude_rad": 0.04,
       "pitch_offset_rad": 0.1,
-      "z_amplitude_m": 0.002,
+      "pitch_amplitude_rad": 0.04,
       "z_offset_m": -0.01,
+      "z_amplitude_m": 0.002,
       "antenna_amplitude_rad": 0.1,
       "antenna_move_name": "both",
       "frequency_hz": 0.3

reachy_mini_ha_voice/motion.py

@@ -5,10 +5,9 @@ MovementManager for unified 5Hz control with face tracking.
 """
 
 import logging
-import math
 from typing import Optional
 
-from .movement_manager import MovementManager, RobotState, PendingAction
+from .movement_manager import MovementManager, RobotState
 
 _LOGGER = logging.getLogger(__name__)
 
@@ -25,7 +24,7 @@ class ReachyMiniMotion:
         self._movement_manager: Optional[MovementManager] = None
         self._camera_server = None  # Reference to camera server for face tracking control
         self._is_speaking = False
-        
+
         _LOGGER.debug("ReachyMiniMotion.__init__ called with reachy_mini=%s", reachy_mini)
 
         # Initialize movement manager if robot is available
@@ -49,7 +48,7 @@ class ReachyMiniMotion:
 
     def set_camera_server(self, camera_server):
         """Set the camera server for face tracking.
-        
+
         Args:
             camera_server: MJPEGCameraServer instance with face tracking enabled
         """
@@ -111,48 +110,32 @@ class ReachyMiniMotion:
 
     def on_continue_listening(self):
         """Called when continuing to listen in tap conversation mode.
-        
-        Performs a small nod to indicate ready for next input.
+
         Non-blocking: command sent to MovementManager.
         """
         if self._movement_manager is None:
             return
 
         self._movement_manager.set_state(RobotState.LISTENING)
-        
-        # Small nod to indicate ready
-        action = PendingAction(
-            name="continue_nod",
-            target_pitch=math.radians(8),  # Small nod down
-            duration=0.25,
-        )
-        self._movement_manager.queue_action(action)
-        _LOGGER.debug("Reachy Mini: Continue listening (nod)")
+        _LOGGER.debug("Reachy Mini: Continue listening")
 
     def on_thinking(self):
         """Called when processing speech - thinking pose.
 
         Non-blocking: command sent to MovementManager.
+        Animation offsets are defined in conversation_animations.json.
         """
         if self._movement_manager is None:
             return
 
         self._movement_manager.set_state(RobotState.THINKING)
-
-        # Slight head tilt (thinking gesture) - avoid looking up too much
-        action = PendingAction(
-            name="thinking",
-            target_pitch=math.radians(-3),   # Very slight look up (was -10, too much)
-            target_roll=math.radians(5),     # Slight head tilt for "thinking" look
-            duration=0.4,
-        )
-        self._movement_manager.queue_action(action)
         _LOGGER.debug("Reachy Mini: Thinking pose")
 
     def on_speaking_start(self):
         """Called when TTS starts - start speech-reactive motion.
 
         Non-blocking: command sent to MovementManager.
+        Animation is defined in conversation_animations.json.
         """
         if self._movement_manager is None:
             _LOGGER.warning("MovementManager not initialized, skipping speaking animation")
@@ -160,15 +143,7 @@ class ReachyMiniMotion:
 
         self._is_speaking = True
         self._movement_manager.set_state(RobotState.SPEAKING)
-
-        # Gentle nod to indicate speaking
-        action = PendingAction(
-            name="speaking_start",
-            target_pitch=math.radians(5),  # Slight nod down
-            duration=0.3,
-        )
-        self._movement_manager.queue_action(action)
-        _LOGGER.info("Reachy Mini: Speaking animation queued")
+        _LOGGER.info("Reachy Mini: Speaking animation started")
 
     def on_speaking_end(self):
         """Called when TTS ends - stop speech-reactive motion.
@@ -226,25 +201,15 @@ class ReachyMiniMotion:
     def wiggle_antennas(self, happy: bool = True):
         """Wiggle antennas to show emotion.
 
-        Non-blocking: command sent to MovementManager.
+        Non-blocking: antenna movement is handled by animation system.
         """
         if self._movement_manager is None:
             return
 
-        # Queue antenna wiggle action
+        # Antenna movement is handled by animation system
+        # Set appropriate animation state
         if happy:
-            action = PendingAction(
-                name="antenna_happy",
-                duration=0.2,
-            )
-            # Note: antenna control is handled in MovementManager state
-        else:
-            action = PendingAction(
-                name="antenna_sad",
-                duration=0.2,
-            )
-
-        self._movement_manager.queue_action(action)
+            self._movement_manager.set_state(RobotState.SPEAKING)
         _LOGGER.debug("Reachy Mini: Antenna wiggle (%s)", "happy" if happy else "sad")
 
     def update_audio_loudness(self, loudness_db: float):