fix run issues

CMakeLists.txt

@@ -7,7 +7,7 @@ endif()
 
 find_package(ament_cmake REQUIRED)
 find_package(ament_cmake_python REQUIRED)
-find_package(rosidl_default_generators REQUIRED)
+find_package(interfaces REQUIRED)
 
 # 确保使用系统 Python（而不是 conda/miniconda 的 Python）
 find_program(PYTHON3_CMD python3 PATHS /usr/bin /usr/local/bin NO_DEFAULT_PATH)
@@ -19,13 +19,6 @@ if(PYTHON3_CMD)
   set(PYTHON_EXECUTABLE ${PYTHON3_CMD} CACHE FILEPATH "Python executable" FORCE)
 endif()
 
-rosidl_generate_interfaces(${PROJECT_NAME}
-  "srv/ASRRecognize.srv"
-  "srv/TTSSynthesize.srv"
-  "srv/VADEvent.srv"
-  "srv/AudioData.srv"
-)
-
 install(CODE "
   execute_process(
     COMMAND ${PYTHON3_CMD} -m pip install --prefix=${CMAKE_INSTALL_PREFIX} --no-deps ${CMAKE_CURRENT_SOURCE_DIR}
@@ -77,25 +70,6 @@ if robot_speaker_src:
         print(f'Copied robot_speaker from {robot_speaker_src} to {ros2_site_packages}')
     else:
         print(f'robot_speaker already in correct location')
-    
-    # 复制 ROS2 生成的 srv 模块（rosidl_generate_interfaces 生成的）
-    rosidl_py_src = os.path.join(build_dir, 'rosidl_generator_py', 'robot_speaker')
-    if os.path.exists(rosidl_py_src):
-        # 复制 srv 目录
-        srv_src = os.path.join(rosidl_py_src, 'srv')
-        srv_dest = os.path.join(robot_speaker_dest, 'srv')
-        if os.path.exists(srv_src):
-            if os.path.exists(srv_dest):
-                shutil.rmtree(srv_dest)
-            shutil.copytree(srv_src, srv_dest)
-            print(f'Copied srv module to {srv_dest}')
-        
-        # 复制生成的接口文件（.so 和 .c 文件）
-        for pattern in ['robot_speaker_s__rosidl_typesupport*.so', '_robot_speaker_s*.c']:
-            for file in glob.glob(os.path.join(rosidl_py_src, pattern)):
-                dest_file = os.path.join(robot_speaker_dest, os.path.basename(file))
-                shutil.copy2(file, dest_file)
-                print(f'Copied {os.path.basename(file)} to {robot_speaker_dest}')
 
 # 处理 entry_points 脚本
 lib_dir = os.path.join(install_prefix, 'lib', 'robot_speaker')

config/voice.yaml

@@ -55,12 +55,12 @@ system:
   session_timeout: 3.0  # 会话超时时间（秒）
   shutup_keywords: "bi zui"  # 闭嘴指令关键词（拼音，逗号分隔）
   interrupt_command_queue_depth: 10  # 中断命令订阅的队列深度（QoS）
-  sv_enabled: true  # 是否启用声纹识别
-  # sv_model_path: "~/hivecore_robot_os1/voice_model" # 声纹模型路径
-  sv_model_path: "~/ros_learn/speech_campplus_sv_zh-cn_16k-common" # 声纹模型路径
+  sv_enabled: false  # 是否启用声纹识别
+  sv_model_path: "~/hivecore_robot_os1/voice_model" # 声纹模型路径
+  # sv_model_path: "~/ros_learn/speech_campplus_sv_zh-cn_16k-common" # 声纹模型路径
   sv_threshold: 0.65  # 声纹识别阈值（0.0-1.0，值越小越宽松，值越大越严格）
-  # sv_speaker_db_path: "~/hivecore_robot_os1/config/speakers.json"  # 声纹数据库保存路径（JSON格式，相对于ROS2包share目录）
-  sv_speaker_db_path: "~/ros_learn/hivecore_robot_voice/config/speakers.json"  # 声纹数据库保存路径（JSON格式，相对于ROS2包share目录）
+  sv_speaker_db_path: "~/hivecore_robot_os1/config/speakers.json"  # 声纹数据库保存路径（JSON格式，相对于ROS2包share目录）
+  # sv_speaker_db_path: "~/ros_learn/hivecore_robot_voice/config/speakers.json"  # 声纹数据库保存路径（JSON格式，相对于ROS2包share目录）
   sv_buffer_size: 96000  # 声纹验证录音缓冲区大小（样本数，48kHz下2秒=96000）
   continue_without_image: true  # 多模态意图（skill_sequence/chat_camera）未获取到图片时是否继续推理
   
@@ -69,5 +69,5 @@ camera:
     jpeg_quality: 85  # JPEG压缩质量（0-100，85是质量和大小平衡点）
 
 interfaces:
-  # root_path: "~/hivecore_robot_os1/hivecore_robot_interfaces/src"  # 接口文件根目录，支持 ~ 展开和相对路径
-  root_path: "~/ros_learn/hivecore_robot_interfaces/src"  # 接口文件根目录，支持 ~ 展开和相对路径
+  root_path: "~/hivecore_robot_os1/hivecore_robot_interfaces/src"  # 接口文件根目录，支持 ~ 展开和相对路径
+  # root_path: "~/ros_learn/hivecore_robot_interfaces/src"  # 接口文件根目录，支持 ~ 展开和相对路径

package.xml

@@ -15,9 +15,6 @@
   <depend>interfaces</depend>
   <buildtool_depend>ament_cmake</buildtool_depend>
   <buildtool_depend>ament_cmake_python</buildtool_depend>
-  <buildtool_depend>rosidl_default_generators</buildtool_depend>
-  <exec_depend>rosidl_default_runtime</exec_depend>
-  <member_of_group>rosidl_interface_packages</member_of_group>
 
   <exec_depend>python3-pyaudio</exec_depend>
   <exec_depend>python3-requests</exec_depend>

robot_speaker/core/register_speaker_node.py

@@ -9,7 +9,7 @@ import queue
 import rclpy
 from rclpy.node import Node
 from ament_index_python.packages import get_package_share_directory
-from robot_speaker.srv import ASRRecognize, AudioData, VADEvent
+from interfaces.srv import ASRRecognize, AudioData, VADEvent
 from robot_speaker.core.speaker_verifier import SpeakerVerificationClient
 from pypinyin import pinyin, Style
 

robot_speaker/core/robot_speaker_node.py

@@ -15,7 +15,7 @@ import os
 import yaml
 from typing import Optional
 from ament_index_python.packages import get_package_share_directory
-from robot_speaker.srv import VADEvent, ASRRecognize, TTSSynthesize, AudioData
+from interfaces.srv import VADEvent, ASRRecognize, TTSSynthesize, AudioData
 from openai import OpenAI
 from robot_speaker.core.context_manager import ConversationHistory
 from robot_speaker.core.speaker_verifier import SpeakerVerificationClient, SpeakerState

robot_speaker/core/speaker_verifier.py

@@ -29,8 +29,8 @@ class SpeakerVerificationClient:
         self._lock = threading.Lock()
         
         # # 优化CPU性能：限制Torch使用的线程数，防止多线程竞争导致性能骤降
-        # import torch
-        # torch.set_num_threads(1)
+        import torch
+        torch.set_num_threads(1)
 
         from funasr import AutoModel
         model_path = os.path.expanduser(self.model_path)

robot_speaker/services/asr_audio_node.py

@@ -1,6 +1,6 @@
 import rclpy
 from rclpy.node import Node
-from robot_speaker.srv import ASRRecognize, AudioData, VADEvent
+from interfaces.srv import ASRRecognize, AudioData, VADEvent
 import threading
 import queue
 import time

robot_speaker/services/tts_audio_node.py

@@ -1,7 +1,7 @@
 import rclpy
 from rclpy.node import Node
 from rclpy.callback_groups import ReentrantCallbackGroup
-from robot_speaker.srv import TTSSynthesize
+from interfaces.srv import TTSSynthesize
 import threading
 import yaml
 import os