add config

2025-09-26 13:11:14 +08:00 · 2025-09-26 13:11:14 +08:00 · a7876fb472
commit a7876fb472
parent 0f15b5060b
7 changed files with 389 additions and 11 deletions
--- a/SYSTEMD_SETUP.md
+++ b/SYSTEMD_SETUP.md
@ -0,0 +1,115 @@
+# Local Voice Systemd 服务配置指南
+
+## 问题解决方案
+
+### 核心问题
+- **直接运行**: 使用PulseAudio，自动转换16000Hz采样率 ✅
+- **systemd运行**: 直接访问USB设备，只支持48000/44100Hz ❌
+
+### 解决策略
+通过配置systemd服务访问用户会话的PulseAudio，利用其采样率转换能力。
+
+## 部署步骤
+
+### 1. 上传文件到树莓派
+```bash
+scp local-voice.service local-voice-user.service deploy-service.sh test-audio.py zhuchaowe@192.168.101.212:~/Local-Voice/
+```
+
+### 2. 运行部署脚本
+```bash
+cd ~/Local-Voice
+./deploy-service.sh
+```
+
+### 3. 测试音频访问
+```bash
+python3 test-audio.py
+```
+
+### 4. 启动服务 (推荐使用用户级服务)
+```bash
+# 用户级服务 (推荐)
+systemctl --user start local-voice-user.service
+systemctl --user enable local-voice-user.service
+
+# 查看日志
+journalctl --user -u local-voice-user.service -f
+
+# 如果用户级服务不工作，尝试系统级服务
+sudo systemctl start local-voice.service
+sudo systemctl enable local-voice.service
+
+# 查看系统服务日志
+sudo journalctl -u local-voice.service -f
+```
+
+## 服务配置说明
+
+### 系统级服务 (local-voice.service)
+- 适用于传统systemd环境
+- 添加了PulseAudio环境变量
+- 包含音频设备权限
+
+### 用户级服务 (local-voice-user.service)
+- 推荐方案，更好的音频访问
+- 自动继承用户会话环境
+- 无需额外环境变量配置
+
+## 故障排除
+
+### 音频设备不可用
+1. 检查pipewire-pulse服务:
+   ```bash
+   systemctl --user status pipewire-pulse
+   systemctl --user start pipewire-pulse
+   ```
+
+2. 检查音频设备权限:
+   ```bash
+   groups  # 确认在audio组中
+   ```
+
+3. 测试直接运行:
+   ```bash
+   python3 multiprocess_recorder.py --enable-nfc
+   ```
+
+### 权限问题
+如果遇到权限错误，确保用户在audio组中:
+```bash
+sudo usermod -a -G audio $USER
+# 重新登录或重启系统
+```
+
+### 环境变量问题
+如果用户级服务正常但系统级服务有问题，可能需要:
+1. 检查环境变量是否正确设置
+2. 确认pipewire-pulse服务在系统启动时运行
+3. 考虑使用用户级服务作为主要方案
+
+## 验证步骤
+
+1. **环境测试**: `python3 test-audio.py`
+2. **服务启动**: `systemctl --user start local-voice-user.service`
+3. **日志监控**: `journalctl --user -u local-voice-user.service -f`
+4. **功能测试**: 使用NFC卡片触发录音功能
+
+## 备用方案
+
+如果上述方案不工作，可以考虑:
+
+### 方案A: 使用screen会话
+```bash
+screen -dmS local-voice python3 ~/Local-Voice/multiprocess_recorder.py --enable-nfc
+screen -r local-voice  # 查看输出
+```
+
+### 方案B: 使用cron启动
+在crontab中添加:
+```
+@reboot cd ~/Local-Voice && python3 multiprocess_recorder.py --enable-nfc
+```
+
+### 方案C: 修改代码支持多种采样率
+如果需要修改代码支持48000Hz采样率，请告知。
--- a/config.json
+++ b/config.json
@ -22,7 +22,9 @@
    "enable_llm": true,
    "enable_tts": true,
    "character": "libai",
-    "max_tokens": 50
+    "max_tokens": 50,
+    "enable_chat_memory": true,
+    "max_history_length": 5
  },
  "detection": {
    "zcr_min": 2400,
--- a/control_system.py
+++ b/control_system.py
@ -91,6 +91,18 @@ class ControlSystem:
            'failed_processing': 0
        }
        
+        # 聊天历史记录
+        self.chat_history = []
+        # 从配置中读取聊天历史设置
+        self.enable_chat_memory = self.config.get('processing', {}).get('enable_chat_memory', True)
+        self.max_history_length = self.config.get('processing', {}).get('max_history_length', 5)
+        
+        # 显示聊天记忆状态
+        if self.enable_chat_memory:
+            print(f"💬 聊天记忆功能已启用，最多保存 {self.max_history_length} 轮对话")
+        else:
+            print("💬 聊天记忆功能已禁用")
+        
        # 运行状态
        self.running = True
        
@ -997,11 +1009,28 @@ class ControlSystem:
                "Authorization": f"Bearer {self.api_config['llm']['api_key']}"
            }
            
+            # 构建消息列表
            messages = [
-                {"role": "system", "content": system_prompt},
-                {"role": "user", "content": text}
+                {
+                    "role": "system",
+                    "content": system_prompt
+                }
            ]
            
+            # 如果启用聊天记忆，添加历史对话记录
+            if self.enable_chat_memory:
+                for history_item in self.chat_history:
+                    messages.extend([
+                        {"role": "user", "content": history_item["user"]},
+                        {"role": "assistant", "content": history_item["assistant"]}
+                    ])
+            
+            # 添加当前用户消息
+            messages.append({
+                "role": "user",
+                "content": text
+            })
+            
            data = {
                "model": self.api_config['llm']['model'],
                "messages": messages,
@ -1020,7 +1049,14 @@ class ControlSystem:
                result = response.json()
                if 'choices' in result and len(result['choices']) > 0:
                    content = result['choices'][0]['message']['content']
-                    return content.strip()
+                    filtered_content = self._filter_parentheses_content(content.strip())
+                    
+                    # 保存对话到历史记录
+                    if self.enable_chat_memory:
+                        self._add_to_chat_history(text, filtered_content)
+                        print(f"💬 对话已保存到历史记录 (当前历史长度: {len(self.chat_history)})")
+                    
+                    return filtered_content
            
            print(f"❌ LLM API调用失败: {response.status_code}")
            return None
@ -1059,13 +1095,23 @@ class ControlSystem:
                {
                    "role": "system",
                    "content": system_prompt
-                },
-                {
-                    "role": "user",
-                    "content": text
                }
            ]
            
+            # 如果启用聊天记忆，添加历史对话记录
+            if self.enable_chat_memory:
+                for history_item in self.chat_history:
+                    messages.extend([
+                        {"role": "user", "content": history_item["user"]},
+                        {"role": "assistant", "content": history_item["assistant"]}
+                    ])
+            
+            # 添加当前用户消息
+            messages.append({
+                "role": "user",
+                "content": text
+            })
+            
            data = {
                "model": self.api_config['llm']['model'],
                "messages": messages,
@ -1181,6 +1227,12 @@ class ControlSystem:
                        print(f"🎵 发送剩余句子: {filtered_sentence}")
                        self._send_streaming_text_to_output_process(filtered_sentence)
                
+                # 保存完整回复到历史记录
+                if accumulated_text and self.enable_chat_memory:
+                    filtered_response = self._filter_parentheses_content(accumulated_text.strip())
+                    self._add_to_chat_history(text, filtered_response)
+                    print(f"💬 对话已保存到历史记录 (当前历史长度: {len(self.chat_history)})")
+                
                # 通知输出进程LLM生成已完成
                self._notify_llm_complete()
                
@ -1651,6 +1703,10 @@ class ControlSystem:
        
        print(f"🔄 角色已切换: {old_character} -> {character_name}")
        
+        # 清空聊天历史
+        self.clear_chat_history()
+        print(f"🔄 角色切换，聊天历史已清空")
+        
        # 播放新角色打招呼
        print("🎭 播放新角色打招呼...")
        greeting_success = self.play_greeting()
@ -1756,6 +1812,33 @@ class ControlSystem:
            "current_character": self.nfc_manager.get_current_character(),
            "running": self.nfc_manager.running
        }
+    
+    def _add_to_chat_history(self, user_message, assistant_response):
+        """添加对话到历史记录"""
+        import time
+        
+        # 如果历史记录超过最大长度，移除最早的记录
+        if len(self.chat_history) >= self.max_history_length:
+            self.chat_history.pop(0)
+        
+        # 添加新的对话记录
+        self.chat_history.append({
+            "user": user_message,
+            "assistant": assistant_response,
+            "timestamp": time.time()
+        })
+    
+    def clear_chat_history(self):
+        """清空聊天历史"""
+        self.chat_history = []
+        print("💬 聊天历史已清空")
+    
+    def get_chat_history_summary(self):
+        """获取聊天历史摘要"""
+        if not self.chat_history:
+            return "暂无聊天历史"
+        
+        return f"当前有 {len(self.chat_history)} 轮对话记录"

 def main():
    """主函数"""
--- a/deploy-service.sh
+++ b/deploy-service.sh
@ -0,0 +1,78 @@
+#!/bin/bash
+
+# Local Voice 服务部署脚本
+# 用于解决systemd环境下的音频设备访问问题
+
+echo "🔧 Local Voice 服务部署脚本"
+echo "=================================="
+
+# 获取用户ID
+USER_ID=$(id -u)
+USERNAME=$(whoami)
+echo "👤 当前用户: $USERNAME (UID: $USER_ID)"
+
+# 备份原有服务文件
+echo "💾 备份原有服务文件..."
+if systemctl list-unit-files | grep -q "local-voice.service"; then
+    sudo systemctl stop local-voice.service 2>/dev/null
+    sudo systemctl disable local-voice.service 2>/dev/null
+    sudo cp /etc/systemd/system/local-voice.service /etc/systemd/system/local-voice.service.backup 2>/dev/null || true
+    echo "✅ 原有服务已备份"
+fi
+
+# 安装systemd服务
+echo "📦 安装systemd服务..."
+sudo cp local-voice.service /etc/systemd/system/
+sudo systemctl daemon-reload
+sudo systemctl enable local-voice.service
+echo "✅ 系统级服务已安装"
+
+# 安装用户级服务
+echo "🏠 安装用户级服务..."
+mkdir -p ~/.config/systemd/user/
+cp local-voice-user.service ~/.config/systemd/user/
+systemctl --user daemon-reload
+systemctl --user enable local-voice-user.service
+echo "✅ 用户级服务已安装"
+
+# 确保pipewire-pulse运行
+echo "🎵 检查音频服务..."
+if ! systemctl --user is-active --quiet pipewire-pulse; then
+    echo "🔧 启动pipewire-pulse服务..."
+    systemctl --user start pipewire-pulse
+    systemctl --user enable pipewire-pulse
+    echo "✅ pipewire-pulse已启动"
+else
+    echo "✅ pipewire-pulse正在运行"
+fi
+
+# 启用用户服务 lingering (允许用户服务在登出后继续运行)
+echo "🔄 启用用户服务lingering..."
+sudo loginctl enable-linger $USERNAME
+echo "✅ 用户服务lingering已启用"
+
+# 测试建议
+echo ""
+echo "🎯 测试建议:"
+echo "1. 先测试用户级服务: systemctl --user start local-voice-user.service"
+echo "2. 查看用户服务日志: journalctl --user -u local-voice-user.service -f"
+echo "3. 如果用户级服务正常，再考虑使用系统级服务"
+echo ""
+echo "🔍 故障排除:"
+echo "- 如果用户级服务正常: 建议使用 systemctl --user start local-voice-user.service"
+echo "- 如果系统级服务正常: 使用 sudo systemctl start local-voice.service"
+echo ""
+echo "📊 当前音频设备状态:"
+python3 -c "
+import pyaudio
+p = pyaudio.PyAudio()
+print('可用音频设备:')
+for i in range(p.get_device_count()):
+    info = p.get_device_info_by_index(i)
+    if info['maxInputChannels'] > 0:
+        print(f'  设备 {i}: {info[\"name\"]} (默认采样率: {info[\"defaultSampleRate\"]}Hz)')
+p.terminate()
+"
+
+echo ""
+echo "✅ 部署完成！"
--- a/local-voice-user.service
+++ b/local-voice-user.service
@ -0,0 +1,16 @@
+[Unit]
+Description=Local Voice Multiprocess Recorder with NFC (User Service)
+After=network.target sound.target pipewire-pulse.service
+Wants=network.target sound.target pipewire-pulse.service
+
+[Service]
+Type=simple
+WorkingDirectory=/home/zhuchaowe/Local-Voice
+ExecStart=/usr/bin/python3 /home/zhuchaowe/Local-Voice/multiprocess_recorder.py --enable-nfc
+Restart=always
+RestartSec=10
+StandardOutput=journal
+StandardError=journal
+
+[Install]
+WantedBy=default.target
--- a/local-voice.service
+++ b/local-voice.service
@ -1,17 +1,24 @@
 [Unit]
 Description=Local Voice Multiprocess Recorder with NFC
-After=network.target
-Wants=network.target
+After=network.target sound.target
+Wants=network.target sound.target

 [Service]
 Type=simple
 User=zhuchaowe
-WorkingDirectory=/home/zhuchaowe
+Group=audio
+WorkingDirectory=/home/zhuchaowe/Local-Voice
+Environment="PULSE_SERVER=unix:/run/user/%U/pulse/native"
+Environment="XDG_RUNTIME_DIR=/run/user/%U"
+Environment="DBUS_SESSION_BUS_ADDRESS=unix:path=/run/user/%U/bus"
 ExecStart=/usr/bin/python3 /home/zhuchaowe/Local-Voice/multiprocess_recorder.py --enable-nfc
 Restart=always
 RestartSec=10
 StandardOutput=journal
 StandardError=journal
+# 添加设备访问权限
+DeviceAllow=/dev/snd/*
+DevicePolicy=auto

 [Install]
 WantedBy=multi-user.target
--- a/test-audio.py
+++ b/test-audio.py
@ -0,0 +1,77 @@
+#!/usr/bin/env python3
+"""
+音频设备测试脚本
+用于测试不同环境下的音频设备访问
+"""
+
+import pyaudio
+import sys
+import os
+
+def test_audio_access():
+    print("🔍 音频设备访问测试")
+    print("=" * 50)
+    
+    try:
+        p = pyaudio.PyAudio()
+        print("✅ PyAudio 初始化成功")
+        
+        # 显示所有音频设备
+        print("\n📋 可用音频设备:")
+        for i in range(p.get_device_count()):
+            info = p.get_device_info_by_index(i)
+            if info['maxInputChannels'] > 0:
+                print(f"  设备 {i}: {info['name']}")
+                print(f"    输入通道: {info['maxInputChannels']}")
+                print(f"    默认采样率: {info['defaultSampleRate']}")
+        
+        # 测试16000Hz采样率
+        print(f"\n🎵 测试16000Hz采样率访问:")
+        
+        # 测试默认设备
+        try:
+            stream = p.open(
+                format=pyaudio.paInt16,
+                channels=1,
+                rate=16000,
+                input=True,
+                frames_per_buffer=1024
+            )
+            print("✅ 默认设备支持16000Hz")
+            stream.close()
+        except Exception as e:
+            print(f"❌ 默认设备不支持16000Hz: {e}")
+        
+        # 测试pulse设备
+        try:
+            stream = p.open(
+                format=pyaudio.paInt16,
+                channels=1,
+                rate=16000,
+                input=True,
+                frames_per_buffer=1024,
+                input_device_index=6  # pulse设备
+            )
+            print("✅ Pulse设备支持16000Hz")
+            stream.close()
+        except Exception as e:
+            print(f"❌ Pulse设备不支持16000Hz: {e}")
+        
+        p.terminate()
+        print("\n✅ 音频测试完成")
+        
+    except Exception as e:
+        print(f"❌ 音频测试失败: {e}")
+        sys.exit(1)
+
+def check_environment():
+    print("\n🌍 环境信息:")
+    print(f"用户: {os.getenv('USER', 'Unknown')}")
+    print(f"用户ID: {os.getuid()}")
+    print(f"PULSE_SERVER: {os.getenv('PULSE_SERVER', 'Not set')}")
+    print(f"XDG_RUNTIME_DIR: {os.getenv('XDG_RUNTIME_DIR', 'Not set')}")
+    print(f"DBUS_SESSION_BUS_ADDRESS: {os.getenv('DBUS_SESSION_BUS_ADDRESS', 'Not set')}")
+
+if __name__ == "__main__":
+    check_environment()
+    test_audio_access()