diff --git a/talkingq-url/.dockerignore b/talkingq-url/.dockerignore new file mode 100644 index 0000000..11d3f63 --- /dev/null +++ b/talkingq-url/.dockerignore @@ -0,0 +1,2 @@ +# 排除 assets 目录，将通过卷挂载方式使用 +assets/ diff --git a/talkingq-url/.gitignore b/talkingq-url/.gitignore new file mode 100644 index 0000000..d1abc05 --- /dev/null +++ b/talkingq-url/.gitignore @@ -0,0 +1,52 @@ +# Python +__pycache__/ +*.py[cod] +*$py.class +*.so +.Python +build/ +develop-eggs/ +dist/ +downloads/ +eggs/ +.eggs/ +lib/ +lib64/ +parts/ +sdist/ +var/ +wheels/ +*.egg-info/ +.installed.cfg +*.egg +venv/ +env/ +ENV/ +tts_audio +.env +.venv +pip-log.txt +pip-delete-this-directory.txt +fullcode.md +all.md +*.sql.gz +database_backups + +# Docker +*.log + +# IDE/Editor +.idea/ +.vscode/ +*.swp +*.swo +*~ + +# OS specific +.DS_Store +.DS_Store? +._* +.Spotlight-V100 +.Trashes +ehthumbs.db +Thumbs.db \ No newline at end of file diff --git a/talkingq-url/Dockerfile b/talkingq-url/Dockerfile new file mode 100644 index 0000000..4d6ce97 --- /dev/null +++ b/talkingq-url/Dockerfile @@ -0,0 +1,29 @@ +# FROM python:3.12-slim +FROM python:3.11-bookworm + + +RUN rm -f /etc/apt/sources.list.d/* \ + && echo "deb https://mirrors.aliyun.com/debian bookworm main" > /etc/apt/sources.list \ + && echo "deb https://mirrors.aliyun.com/debian bookworm-updates main" >> /etc/apt/sources.list \ + && echo "deb https://mirrors.aliyun.com/debian-security bookworm-security main" >> /etc/apt/sources.list \ + && apt-get update \ + && apt-get install -y ffmpeg build-essential git default-mysql-client \ + && rm -rf /var/lib/apt/lists/* + +ENV TZ=Asia/Shanghai +RUN ln -snf /usr/share/zoneinfo/$TZ /etc/localtime && echo $TZ > /etc/timezone + +WORKDIR /app + +COPY requirements.txt . +RUN pip install --upgrade pip -i https://mirrors.aliyun.com/pypi/simple/ \ + --trusted-host mirrors.aliyun.com && \ + pip install --no-cache-dir -i https://mirrors.aliyun.com/pypi/simple/ \ + --trusted-host mirrors.aliyun.com \ + -r requirements.txt \ + && rm -rf /root/.cache/pip + +COPY . . + +# 将workers数量从5改为1，解决WebSocket连接在多进程间不共享的问题 +CMD ["uvicorn", "main:app", "--host", "0.0.0.0", "--port", "8080", "--workers", "1"] \ No newline at end of file diff --git a/talkingq-url/__init__.py b/talkingq-url/__init__.py new file mode 100644 index 0000000..e69de29 diff --git a/talkingq-url/api/__init__.py b/talkingq-url/api/__init__.py new file mode 100644 index 0000000..fc295de --- /dev/null +++ b/talkingq-url/api/__init__.py @@ -0,0 +1,14 @@ +from fastapi import APIRouter +from api.websocket import router as websocket_router +from api.roles import router as roles_router +from api.auth import router as auth_router +from api.device_control import router as device_control_router +from api.ota import router as ota_router # 新增OTA路由 + +api_router = APIRouter() +api_router.include_router(websocket_router, tags=["WebSocket"]) +api_router.include_router(roles_router, tags=["Roles"]) +api_router.include_router(auth_router, tags=["Auth"]) +api_router.include_router(device_control_router, tags=["Device Control"]) +api_router.include_router(ota_router, tags=["OTA"]) # 注册OTA路由 + diff --git a/talkingq-url/api/assets.py b/talkingq-url/api/assets.py new file mode 100644 index 0000000..8259248 --- /dev/null +++ b/talkingq-url/api/assets.py @@ -0,0 +1,16 @@ +import os +from fastapi import APIRouter +from fastapi.staticfiles import StaticFiles +from config import settings + +router = APIRouter() + +def configure_static_assets(app): + """配置静态资源文件夹""" + tts_audio_directory = os.path.join(settings.assets_dir, "tts_audio") + os.makedirs(tts_audio_directory, exist_ok=True) + + firmware_directory = os.path.join(settings.assets_dir, "firmware") + os.makedirs(firmware_directory, exist_ok=True) + + app.mount("/assets", StaticFiles(directory=settings.assets_dir), name="assets") \ No newline at end of file diff --git a/talkingq-url/api/auth.py b/talkingq-url/api/auth.py new file mode 100644 index 0000000..b32d7b3 --- /dev/null +++ b/talkingq-url/api/auth.py @@ -0,0 +1,250 @@ +from fastapi import Depends, HTTPException, status, Request, APIRouter, Path, Query +from fastapi.security import APIKeyHeader +from pydantic import BaseModel, Field +from typing import List, Optional, Dict, Any, Tuple +from services.device_auth_manager import device_auth_manager +from utils.logger import session_logger +from config import settings +import re + +router = APIRouter(prefix="/api/auth", tags=["Auth"]) + +device_id_header = APIKeyHeader(name="X-Device-ID", auto_error=False) +device_serial_header = APIKeyHeader(name="X-Device-Serial", auto_error=False) +admin_api_key_header = APIKeyHeader(name="X-Admin-API-Key", auto_error=False) +client_api_key_header = APIKeyHeader(name="X-Client-Key", auto_error=False) + +async def verify_device( + device_id: str = Depends(device_id_header), + serial_number: str = Depends(device_serial_header) +): + """验证设备ID和序列号""" + if not device_id or not serial_number: + raise HTTPException( + status_code=status.HTTP_401_UNAUTHORIZED, + detail="缺少设备认证信息", + headers={"WWW-Authenticate": "DeviceAuth"}, + ) + is_valid = await device_auth_manager.authenticate_device(device_id, serial_number) + if not is_valid: + session_logger.warning("system", "auth", f"设备 {device_id} 认证失败") + raise HTTPException( + status_code=status.HTTP_401_UNAUTHORIZED, + detail="设备认证失败", + headers={"WWW-Authenticate": "DeviceAuth"}, + ) + return device_id + +async def api_auth( + device_id: str = Depends(verify_device) +): + """API认证依赖项，只使用设备ID和序列号认证""" + return device_id + +async def admin_auth(api_key: str = Depends(admin_api_key_header)) -> bool: + """验证管理员API密钥""" + if not api_key: + raise HTTPException( + status_code=status.HTTP_401_UNAUTHORIZED, + detail="缺少管理员API密钥", + headers={"WWW-Authenticate": "AdminAuth"}, + ) + return api_key == settings.admin_api_key + +async def client_auth(client_api_key: str = Depends(client_api_key_header)) -> bool: + """验证小程序客户端API密钥""" + if not client_api_key: + raise HTTPException( + status_code=status.HTTP_401_UNAUTHORIZED, + detail="缺少客户端API密钥", + headers={"WWW-Authenticate": "ClientAuth"}, + ) + + if client_api_key != settings.client_api_key: + session_logger.warning("system", "auth", "客户端API密钥验证失败") + raise HTTPException( + status_code=status.HTTP_401_UNAUTHORIZED, + detail="客户端API密钥无效", + headers={"WWW-Authenticate": "ClientAuth"}, + ) + + return True + +async def admin_or_api_auth( + device_id: str = Path(...), + admin_api_key: str = Depends(admin_api_key_header), + authenticated_device_id: str = Depends(api_auth) +) -> Tuple[str, bool]: + """允许管理员或设备本身访问，返回(认证设备ID, 是否管理员)""" + is_admin = admin_api_key == settings.admin_api_key if admin_api_key else False + + if not is_admin and authenticated_device_id != device_id: + raise HTTPException( + status_code=status.HTTP_403_FORBIDDEN, + detail="没有权限查询其他设备", + headers={"WWW-Authenticate": "DeviceAuth"}, + ) + + return authenticated_device_id, is_admin + +class DeviceRegistrationRequest(BaseModel): + device_id: str = Field(..., min_length=16, max_length=64, description="设备ID，基于ESP32的MAC地址生成") + serial_number: str = Field(..., min_length=18, max_length=64, description="序列号，包含批次前缀和唯一码") + batch_id: Optional[str] = Field(None, description="批次ID，使用YYYYMMDD格式") + is_active: bool = Field(True, description="设备激活状态") + +class DeviceRegistrationResponse(BaseModel): + status: str + device_id: str + serial_number: str + batch_id: Optional[str] = None + is_active: bool + +@router.post("/register-device", response_model=DeviceRegistrationResponse) +async def register_device( + request: DeviceRegistrationRequest, + admin_authenticated: bool = Depends(admin_auth) +): + """注册新设备到认证白名单""" + batch_id = request.batch_id + if not batch_id and len(request.serial_number) >= 8: + batch_id = request.serial_number[:8] + + success = await device_auth_manager.register_device( + request.device_id, + request.serial_number, + batch_id, + request.is_active + ) + + if not success: + raise HTTPException( + status_code=status.HTTP_500_INTERNAL_SERVER_ERROR, + detail="设备注册失败", + ) + + return { + "status": "success", + "device_id": request.device_id, + "serial_number": request.serial_number, + "batch_id": batch_id, + "is_active": request.is_active + } + +class BatchDeviceRegistrationRequest(BaseModel): + devices: List[DeviceRegistrationRequest] + +class BatchDeviceRegistrationResponse(BaseModel): + status: str + registered_count: int + failed_count: int + details: List[Dict[str, Any]] + +@router.post("/register-devices-batch", response_model=BatchDeviceRegistrationResponse) +async def register_devices_batch( + request: BatchDeviceRegistrationRequest, + admin_authenticated: bool = Depends(admin_auth) +): + """批量注册设备到认证白名单""" + results = [] + success_count = 0 + failed_count = 0 + + for device in request.devices: + batch_id = device.batch_id + if not batch_id and len(device.serial_number) >= 8: + batch_id = device.serial_number[:8] + + success = await device_auth_manager.register_device( + device.device_id, + device.serial_number, + batch_id, + device.is_active + ) + + if success: + success_count += 1 + results.append({ + "status": "success", + "device_id": device.device_id, + "serial_number": device.serial_number, + "batch_id": batch_id + }) + else: + failed_count += 1 + results.append({ + "status": "failed", + "device_id": device.device_id, + "serial_number": device.serial_number + }) + + return { + "status": "completed", + "registered_count": success_count, + "failed_count": failed_count, + "details": results + } + +@router.get("/verify-device/{device_id}") +async def verify_device( + device_id: str = Path(...), + serial_number: str = None, + auth_result: Tuple[str, bool] = Depends(admin_or_api_auth) +): + """验证设备是否已正确注册 (管理员或设备自身可使用)""" + authenticated_device_id, is_admin = auth_result + + if is_admin and serial_number: + is_valid = await device_auth_manager.authenticate_device(device_id, serial_number) + elif not is_admin: + is_valid = True + else: + is_valid = None + + device_info = await device_auth_manager.get_device_info(device_id) + + return { + "device_id": device_id, + "is_valid": is_valid, + "device_info": device_info + } + +class SerialQueryResponse(BaseModel): + device_id: str + serial_number: str + +@router.get("/query-serial", response_model=SerialQueryResponse) +async def query_serial_by_mac( + mac_address: str = Query(..., description="设备MAC地址"), + client_authenticated: bool = Depends(client_auth) +): + """通过MAC地址查询设备ID和序列号""" + mac_pattern = re.compile(r'^([0-9A-Fa-f]{2}[:-]){5}([0-9A-Fa-f]{2})$') + if not mac_pattern.match(mac_address): + raise HTTPException( + status_code=status.HTTP_400_BAD_REQUEST, + detail="MAC地址格式无效，应为XX:XX:XX:XX:XX:XX或XX-XX-XX-XX-XX-XX格式" + ) + + normalized_mac = mac_address.replace(":", "").replace("-", "").upper() + + expected_device_id = f"TalkingQ_{normalized_mac}" + + device_info = await device_auth_manager.get_device_info(expected_device_id) + + if not device_info or not device_info.get("is_active", False): + session_logger.warning("system", "auth", f"未找到MAC地址{mac_address}对应的设备或设备未激活") + raise HTTPException( + status_code=status.HTTP_404_NOT_FOUND, + detail="未找到对应的设备信息或设备未激活" + ) + + return { + "device_id": expected_device_id, + "serial_number": device_info["serial_number"] + } + +@router.get("/verify") +async def verify_auth(device_id: str = Depends(api_auth)): + """验证设备认证状态""" + return {"status": "authenticated", "device_id": device_id} diff --git a/talkingq-url/api/device_control.py b/talkingq-url/api/device_control.py new file mode 100644 index 0000000..3b26181 --- /dev/null +++ b/talkingq-url/api/device_control.py @@ -0,0 +1,97 @@ +from fastapi import APIRouter, HTTPException, Depends, Path, Body +from pydantic import BaseModel, Field +from typing import Optional + +from api.auth import api_auth +from services.connection_manager import connection_manager +from services.device_volume_manager import device_volume_manager +from utils.logger import session_logger + +router = APIRouter(prefix="/api/device", tags=["Device Control"]) + +class VolumeRequest(BaseModel): + volume: int = Field(..., ge=0, le=100, description="音量值(0-100)") + +class DeviceResponse(BaseModel): + status: str + message: str + device_id: str + +class VolumeResponse(BaseModel): + volume: int + device_id: str + +@router.post("/volume/{device_id}", response_model=DeviceResponse) +async def set_device_volume( + device_id: str = Path(..., description="设备ID"), + volume_request: VolumeRequest = Body(...), + authenticated_device_id: str = Depends(api_auth) +): + """设置设备音量""" + try: + await device_volume_manager.set_volume(device_id, volume_request.volume) + session_logger.info(device_id, "volume", f"设备音量设置为 {volume_request.volume}") + + websocket = await connection_manager.get_connection(device_id) + if websocket: + try: + await websocket.send_text(f"VOLUME:{volume_request.volume}") + session_logger.info(device_id, "volume", f"已发送实时音量设置到设备") + except Exception as e: + session_logger.warning(device_id, "volume", f"发送音量通知失败，但数据库已更新: {str(e)}") + else: + session_logger.info(device_id, "volume", "设备当前不在线，仅保存设置到数据库") + + return { + "status": "success", + "message": f"Volume set to {volume_request.volume}", + "device_id": device_id + } + except Exception as e: + session_logger.error(device_id, "volume", f"设置音量失败: {str(e)}") + raise HTTPException(status_code=500, detail=f"设置音量失败: {str(e)}") + +@router.get("/volume/{device_id}", response_model=VolumeResponse) +async def get_device_volume( + device_id: str = Path(..., description="设备ID"), + authenticated_device_id: str = Depends(api_auth) +): + """获取设备当前音量设置""" + try: + volume = await device_volume_manager.get_volume(device_id) + return { + "volume": volume, + "device_id": device_id + } + except Exception as e: + session_logger.error(device_id, "volume", f"获取音量失败: {str(e)}") + raise HTTPException(status_code=500, detail=f"获取音量失败: {str(e)}") + +class NetworkResetResponse(BaseModel): + status: str + message: str + device_id: str + +@router.post("/reset-network/{device_id}", response_model=NetworkResetResponse) +async def reset_device_network( + device_id: str = Path(..., description="设备ID"), + authenticated_device_id: str = Depends(api_auth) +): + """重置设备网络配置""" + websocket = await connection_manager.get_connection(device_id) + if not websocket: + raise HTTPException(status_code=404, detail=f"设备 {device_id} 未在线或未找到") + + try: + await websocket.send_text("RESET_NETWORK") + + session_logger.info(device_id, "network", "已发送网络重置命令") + + return { + "status": "success", + "message": "Network reset initiated", + "device_id": device_id + } + except Exception as e: + session_logger.error(device_id, "network", f"网络重置命令发送失败: {str(e)}") + raise HTTPException(status_code=500, detail=f"网络重置命令发送失败: {str(e)}") diff --git a/talkingq-url/api/ota.py b/talkingq-url/api/ota.py new file mode 100644 index 0000000..3ecbb1e --- /dev/null +++ b/talkingq-url/api/ota.py @@ -0,0 +1,103 @@ +from fastapi import APIRouter, HTTPException, Depends, Path, Query +from pydantic import BaseModel +from typing import Dict, Any, Optional +from api.auth import api_auth, client_auth +from api.response import Response +from services.ota_flow_controller import ota_flow_controller +from utils.logger import session_logger +from services.connection_manager import connection_manager +from services.system_config_manager import system_config_manager + +router = APIRouter(prefix="/api/ota", tags=["OTA"]) + +class OTAStatusResponse(BaseModel): + status: str + progress: float + version: str + +class FirmwareUpdateSettings(BaseModel): + version: str + url: str + +@router.get("/check/{device_id}", response_model=Response) +async def check_update( + device_id: str = Path(..., description="设备ID"), + authenticated: bool = Depends(api_auth) +): + """检查设备是否需要更新 (APP调用)""" + try: + result = await ota_flow_controller.start_update_flow(device_id) + if result["status"] == "error": + return Response(code=-1, msg=result["message"], data={}) + return Response(code=0, msg="success", data=result) + except Exception as e: + session_logger.error(device_id, "ota", f"检查更新失败: {str(e)}") + return Response(code=-1, msg=f"检查更新失败: {str(e)}", data={}) + +@router.post("/start/{device_id}", response_model=Response) +async def start_update( + device_id: str = Path(..., description="设备ID"), + authenticated: bool = Depends(api_auth) +): + """启动设备更新 (APP调用)""" + try: + result = await ota_flow_controller.execute_update(device_id) + if result["status"] == "error": + return Response(code=-1, msg=result["message"], data={}) + return Response(code=0, msg="success", data={"updating": True}) + except Exception as e: + session_logger.error(device_id, "ota", f"启动更新失败: {str(e)}") + return Response(code=-1, msg=f"启动更新失败: {str(e)}", data={"updating": False}) + +@router.get("/status/{device_id}", response_model=Response) +async def get_update_status( + device_id: str = Path(..., description="设备ID"), + authenticated: bool = Depends(api_auth) +): + """获取设备更新状态 (APP调用)""" + try: + status = await ota_flow_controller.get_update_status(device_id) + websocket = await connection_manager.get_connection(device_id) + status["device_online"] = websocket is not None + return Response(code=0, msg="success", data=status) + except Exception as e: + session_logger.error(device_id, "ota", f"获取更新状态失败: {str(e)}") + return Response(code=-1, msg=f"获取更新状态失败: {str(e)}", data={ + "status": "error", + "progress": 0, + "version": "unknown", + "device_online": False + }) + +@router.post("/config/firmware", response_model=Response) +async def update_firmware_settings( + settings: FirmwareUpdateSettings, + admin_api_key: str = Depends(api_auth) +): + """更新固件配置信息（版本和URL）""" + try: + await system_config_manager.update_config('latest_firmware_version', settings.version) + await system_config_manager.update_config('update_firmware_url', settings.url) + return Response(code=0, msg="固件配置已更新", data={ + "version": settings.version, + "url": settings.url + }) + except Exception as e: + session_logger.error("system", "ota", f"更新固件配置失败: {str(e)}") + return Response(code=-1, msg=f"更新固件配置失败: {str(e)}", data={}) + +@router.get("/config/firmware", response_model=Response) +async def get_firmware_settings( + admin_api_key: str = Depends(api_auth) +): + """获取当前固件配置信息""" + try: + version_config = await system_config_manager.get_config('latest_firmware_version') + url_config = await system_config_manager.get_config('update_firmware_url') + return Response(code=0, msg="success", data={ + "version": version_config.config_value if version_config else "", + "url": url_config.config_value if url_config else "" + }) + except Exception as e: + session_logger.error("system", "ota", f"获取固件配置失败: {str(e)}") + return Response(code=-1, msg=f"获取固件配置失败: {str(e)}", data={}) diff --git a/talkingq-url/api/response.py b/talkingq-url/api/response.py new file mode 100644 index 0000000..a6b6a22 --- /dev/null +++ b/talkingq-url/api/response.py @@ -0,0 +1,6 @@ +from pydantic import BaseModel, Field + +class Response(BaseModel): + code: int + msg: str + data: object = None diff --git a/talkingq-url/api/roles.py b/talkingq-url/api/roles.py new file mode 100644 index 0000000..6892fc5 --- /dev/null +++ b/talkingq-url/api/roles.py @@ -0,0 +1,127 @@ +from fastapi import APIRouter, HTTPException, Query, Depends +from typing import Dict, List, Optional +from pydantic import BaseModel +from services.role_service import role_service +from services.device_service import device_service +from services.history_service import history_service +from api.auth import api_auth + +router = APIRouter(prefix="/api/roles") + +class RoleResponse(BaseModel): + role_key: str + name: str + description: Optional[str] = None + languages: Optional[List[str]] = None + +class RoleSummaryResponse(BaseModel): + role_key: str + name: str + +class DeviceRoleUpdate(BaseModel): + role_key: str + language: Optional[str] = None + +class ConversationMessage(BaseModel): + user: str + assistant: str + timestamp: Optional[float] = None + +class PaginatedRoleResponse(BaseModel): + total: int + page: int + page_size: int + data: List[RoleResponse] + +class PaginatedHistoryResponse(BaseModel): + total: int + page: int + page_size: int + data: List[ConversationMessage] + role_key: str + role_name: str + +@router.get("/list", response_model=PaginatedRoleResponse) +async def list_roles( + authenticated_device_id: str = Depends(api_auth), + page: int = Query(1, ge=1), + page_size: int = Query(20, ge=1, le=100), + search: Optional[str] = None +): + """获取角色分页列表，支持搜索""" + return await role_service.get_roles_paginated(page, page_size, search) + +@router.get("/summaries", response_model=List[RoleSummaryResponse]) +async def get_role_summaries( + authenticated_device_id: str = Depends(api_auth) +): + """获取角色简要信息列表（仅包含key和name）""" + return await role_service.get_role_summaries() + +@router.get("/device/{device_id}") +async def get_device_role( + device_id: str, + authenticated_device_id: str = Depends(api_auth) +): + """获取设备当前使用的角色""" + if device_id != authenticated_device_id: + raise HTTPException(status_code=403, detail="无权访问其他设备的配置") + + return await device_service.get_device_role(device_id) + +@router.put("/device/{device_id}") +async def update_device_role( + device_id: str, + role_update: DeviceRoleUpdate, + authenticated_device_id: str = Depends(api_auth) +): + """更新设备的角色配置""" + if device_id != authenticated_device_id: + raise HTTPException(status_code=403, detail="无权修改其他设备的配置") + + try: + return await device_service.update_device_role( + device_id, + role_update.role_key, + role_update.language + ) + except ValueError as e: + raise HTTPException(status_code=404, detail=str(e)) + except Exception as e: + raise HTTPException(status_code=500, detail=f"更新角色时出错: {str(e)}") + +@router.get("/history/{device_id}", response_model=PaginatedHistoryResponse) +async def get_conversation_history( + device_id: str, + authenticated_device_id: str = Depends(api_auth), + role_key: Optional[str] = None, + page: int = Query(1, ge=1), + page_size: int = Query(20, ge=1, le=100), + since: Optional[float] = None, + until: Optional[float] = None +): + """获取设备的对话历史记录，支持分页和时间范围筛选""" + if device_id != authenticated_device_id: + raise HTTPException(status_code=403, detail="无权访问其他设备的对话历史") + + try: + return await history_service.get_history_paginated( + device_id, role_key, page, page_size, since, until + ) + except Exception as e: + raise HTTPException(status_code=500, detail=f"获取对话历史失败: {str(e)}") + +@router.get("/history-summary/{device_id}") +async def get_conversation_history_summary( + device_id: str, + authenticated_device_id: str = Depends(api_auth), + days: int = Query(7, ge=1, le=30) +): + """获取设备的对话历史摘要信息，包括每个角色的最后交互时间和消息数量""" + if device_id != authenticated_device_id: + raise HTTPException(status_code=403, detail="无权访问其他设备的对话历史摘要") + + try: + return await history_service.get_history_summary(device_id, days) + except Exception as e: + raise HTTPException(status_code=500, detail=f"获取对话历史摘要失败: {str(e)}") \ No newline at end of file diff --git a/talkingq-url/api/system_config.py b/talkingq-url/api/system_config.py new file mode 100644 index 0000000..e18feeb --- /dev/null +++ b/talkingq-url/api/system_config.py @@ -0,0 +1,67 @@ +from fastapi import APIRouter, HTTPException, Depends, Query +from typing import Optional, List +from pydantic import BaseModel +from services.system_config_manager import system_config_manager +from api.auth import api_auth + +router = APIRouter(prefix="/api/system-config") + +class SystemConfigResponse(BaseModel): + key: str + value: str + created_at: Optional[float] = None + updated_at: Optional[float] = None + +class SystemConfigCreate(BaseModel): + key: str + value: str + +class SystemConfigUpdate(BaseModel): + value: str + +@router.get("/", response_model=List[SystemConfigResponse]) +async def list_configs(authenticated_device_id: str = Depends(api_auth)): + """列出所有系统配置""" + configs = await system_config_manager.list_configs() + return [SystemConfigResponse(**c.__dict__) for c in configs] + +@router.get("/{key}", response_model=SystemConfigResponse) +async def get_config(key: str, authenticated_device_id: str = Depends(api_auth)): + """获取指定key的系统配置""" + config = await system_config_manager.get_config(key) + if not config: + raise HTTPException(status_code=404, detail="配置不存在") + return SystemConfigResponse(**config.__dict__) + +@router.post("/", response_model=SystemConfigResponse) +async def create_config( + config: SystemConfigCreate, + authenticated_device_id: str = Depends(api_auth) +): + """新增系统配置""" + success = await system_config_manager.create_config(config.key, config.value) + if not success: + raise HTTPException(status_code=400, detail="创建配置失败") + new_config = await system_config_manager.get_config(config.key) + return SystemConfigResponse(**new_config.__dict__) + +@router.put("/{key}", response_model=SystemConfigResponse) +async def update_config( + key: str, + config: SystemConfigUpdate, + authenticated_device_id: str = Depends(api_auth) +): + """更新系统配置""" + success = await system_config_manager.update_config(key, config.value) + if not success: + raise HTTPException(status_code=400, detail="更新配置失败") + updated_config = await system_config_manager.get_config(key) + return SystemConfigResponse(**updated_config.__dict__) + +@router.delete("/{key}") +async def delete_config(key: str, authenticated_device_id: str = Depends(api_auth)): + """删除系统配置""" + success = await system_config_manager.delete_config(key) + if not success: + raise HTTPException(status_code=400, detail="删除配置失败") + return {"detail": "删除成功"} diff --git a/talkingq-url/api/websocket.py b/talkingq-url/api/websocket.py new file mode 100644 index 0000000..21b7cce --- /dev/null +++ b/talkingq-url/api/websocket.py @@ -0,0 +1,9 @@ + +from fastapi import APIRouter, WebSocket +from handlers.websocket_handler import websocket_endpoint + +router = APIRouter() + +@router.websocket("/ws") +async def websocket_route(websocket: WebSocket): + await websocket_endpoint(websocket) \ No newline at end of file diff --git a/talkingq-url/assets/firmware/1.0.0.bin b/talkingq-url/assets/firmware/1.0.0.bin new file mode 100644 index 0000000..0440ca6 Binary files /dev/null and b/talkingq-url/assets/firmware/1.0.0.bin differ diff --git a/talkingq-url/assets/firmware/1.0.1.bin b/talkingq-url/assets/firmware/1.0.1.bin new file mode 100644 index 0000000..547b7cf Binary files /dev/null and b/talkingq-url/assets/firmware/1.0.1.bin differ diff --git a/talkingq-url/assets/roles/kuailehu/en/low_battery.mp3 b/talkingq-url/assets/roles/kuailehu/en/low_battery.mp3 new file mode 100644 index 0000000..be835ca Binary files /dev/null and b/talkingq-url/assets/roles/kuailehu/en/low_battery.mp3 differ diff --git a/talkingq-url/assets/roles/kuailehu/en/sleep.mp3 b/talkingq-url/assets/roles/kuailehu/en/sleep.mp3 new file mode 100644 index 0000000..2c864c2 Binary files /dev/null and b/talkingq-url/assets/roles/kuailehu/en/sleep.mp3 differ diff --git a/talkingq-url/assets/roles/kuailehu/en/tts_error.mp3 b/talkingq-url/assets/roles/kuailehu/en/tts_error.mp3 new file mode 100644 index 0000000..c73bea6 Binary files /dev/null and b/talkingq-url/assets/roles/kuailehu/en/tts_error.mp3 differ diff --git a/talkingq-url/assets/roles/kuailehu/en/welcome.mp3 b/talkingq-url/assets/roles/kuailehu/en/welcome.mp3 new file mode 100644 index 0000000..1b23e7e Binary files /dev/null and b/talkingq-url/assets/roles/kuailehu/en/welcome.mp3 differ diff --git a/talkingq-url/assets/roles/kuailehu/zh/low_battery.mp3 b/talkingq-url/assets/roles/kuailehu/zh/low_battery.mp3 new file mode 100644 index 0000000..815abd2 Binary files /dev/null and b/talkingq-url/assets/roles/kuailehu/zh/low_battery.mp3 differ diff --git a/talkingq-url/assets/roles/kuailehu/zh/sleep.mp3 b/talkingq-url/assets/roles/kuailehu/zh/sleep.mp3 new file mode 100644 index 0000000..da6e83e Binary files /dev/null and b/talkingq-url/assets/roles/kuailehu/zh/sleep.mp3 differ diff --git a/talkingq-url/assets/roles/kuailehu/zh/tts_error.mp3 b/talkingq-url/assets/roles/kuailehu/zh/tts_error.mp3 new file mode 100644 index 0000000..0a4d054 Binary files /dev/null and b/talkingq-url/assets/roles/kuailehu/zh/tts_error.mp3 differ diff --git a/talkingq-url/assets/roles/kuailehu/zh/welcome.mp3 b/talkingq-url/assets/roles/kuailehu/zh/welcome.mp3 new file mode 100644 index 0000000..9c06542 Binary files /dev/null and b/talkingq-url/assets/roles/kuailehu/zh/welcome.mp3 differ diff --git a/talkingq-url/assets/roles/meilinvyou/de/low_battery.mp3 b/talkingq-url/assets/roles/meilinvyou/de/low_battery.mp3 new file mode 100644 index 0000000..8ed2b00 Binary files /dev/null and b/talkingq-url/assets/roles/meilinvyou/de/low_battery.mp3 differ diff --git a/talkingq-url/assets/roles/meilinvyou/de/sleep.mp3 b/talkingq-url/assets/roles/meilinvyou/de/sleep.mp3 new file mode 100644 index 0000000..393ea1a Binary files /dev/null and b/talkingq-url/assets/roles/meilinvyou/de/sleep.mp3 differ diff --git a/talkingq-url/assets/roles/meilinvyou/de/tts_error.mp3 b/talkingq-url/assets/roles/meilinvyou/de/tts_error.mp3 new file mode 100644 index 0000000..0b43dbf Binary files /dev/null and b/talkingq-url/assets/roles/meilinvyou/de/tts_error.mp3 differ diff --git a/talkingq-url/assets/roles/meilinvyou/de/welcome.mp3 b/talkingq-url/assets/roles/meilinvyou/de/welcome.mp3 new file mode 100644 index 0000000..a78e928 Binary files /dev/null and b/talkingq-url/assets/roles/meilinvyou/de/welcome.mp3 differ diff --git a/talkingq-url/assets/roles/meilinvyou/en/low_battery.mp3 b/talkingq-url/assets/roles/meilinvyou/en/low_battery.mp3 new file mode 100644 index 0000000..1e90486 Binary files /dev/null and b/talkingq-url/assets/roles/meilinvyou/en/low_battery.mp3 differ diff --git a/talkingq-url/assets/roles/meilinvyou/en/sleep.mp3 b/talkingq-url/assets/roles/meilinvyou/en/sleep.mp3 new file mode 100644 index 0000000..91cdeef Binary files /dev/null and b/talkingq-url/assets/roles/meilinvyou/en/sleep.mp3 differ diff --git a/talkingq-url/assets/roles/meilinvyou/en/tts_error.mp3 b/talkingq-url/assets/roles/meilinvyou/en/tts_error.mp3 new file mode 100644 index 0000000..ff2d5f3 Binary files /dev/null and b/talkingq-url/assets/roles/meilinvyou/en/tts_error.mp3 differ diff --git a/talkingq-url/assets/roles/meilinvyou/en/welcome.mp3 b/talkingq-url/assets/roles/meilinvyou/en/welcome.mp3 new file mode 100644 index 0000000..77af445 Binary files /dev/null and b/talkingq-url/assets/roles/meilinvyou/en/welcome.mp3 differ diff --git a/talkingq-url/assets/roles/meilinvyou/es/low_battery.mp3 b/talkingq-url/assets/roles/meilinvyou/es/low_battery.mp3 new file mode 100644 index 0000000..70e5f5f Binary files /dev/null and b/talkingq-url/assets/roles/meilinvyou/es/low_battery.mp3 differ diff --git a/talkingq-url/assets/roles/meilinvyou/es/sleep.mp3 b/talkingq-url/assets/roles/meilinvyou/es/sleep.mp3 new file mode 100644 index 0000000..81a3167 Binary files /dev/null and b/talkingq-url/assets/roles/meilinvyou/es/sleep.mp3 differ diff --git a/talkingq-url/assets/roles/meilinvyou/es/tts_error.mp3 b/talkingq-url/assets/roles/meilinvyou/es/tts_error.mp3 new file mode 100644 index 0000000..7f157ed Binary files /dev/null and b/talkingq-url/assets/roles/meilinvyou/es/tts_error.mp3 differ diff --git a/talkingq-url/assets/roles/meilinvyou/es/welcome.mp3 b/talkingq-url/assets/roles/meilinvyou/es/welcome.mp3 new file mode 100644 index 0000000..0785453 Binary files /dev/null and b/talkingq-url/assets/roles/meilinvyou/es/welcome.mp3 differ diff --git a/talkingq-url/assets/roles/meilinvyou/fr/low_battery.mp3 b/talkingq-url/assets/roles/meilinvyou/fr/low_battery.mp3 new file mode 100644 index 0000000..86a2de2 Binary files /dev/null and b/talkingq-url/assets/roles/meilinvyou/fr/low_battery.mp3 differ diff --git a/talkingq-url/assets/roles/meilinvyou/fr/sleep.mp3 b/talkingq-url/assets/roles/meilinvyou/fr/sleep.mp3 new file mode 100644 index 0000000..037b9a2 Binary files /dev/null and b/talkingq-url/assets/roles/meilinvyou/fr/sleep.mp3 differ diff --git a/talkingq-url/assets/roles/meilinvyou/fr/tts_error.mp3 b/talkingq-url/assets/roles/meilinvyou/fr/tts_error.mp3 new file mode 100644 index 0000000..c9c0887 Binary files /dev/null and b/talkingq-url/assets/roles/meilinvyou/fr/tts_error.mp3 differ diff --git a/talkingq-url/assets/roles/meilinvyou/fr/welcome.mp3 b/talkingq-url/assets/roles/meilinvyou/fr/welcome.mp3 new file mode 100644 index 0000000..71a9294 Binary files /dev/null and b/talkingq-url/assets/roles/meilinvyou/fr/welcome.mp3 differ diff --git a/talkingq-url/assets/roles/meilinvyou/zh/low_battery.mp3 b/talkingq-url/assets/roles/meilinvyou/zh/low_battery.mp3 new file mode 100644 index 0000000..da767f1 Binary files /dev/null and b/talkingq-url/assets/roles/meilinvyou/zh/low_battery.mp3 differ diff --git a/talkingq-url/assets/roles/meilinvyou/zh/sleep.mp3 b/talkingq-url/assets/roles/meilinvyou/zh/sleep.mp3 new file mode 100644 index 0000000..21a69d7 Binary files /dev/null and b/talkingq-url/assets/roles/meilinvyou/zh/sleep.mp3 differ diff --git a/talkingq-url/assets/roles/meilinvyou/zh/song_not_found.mp3 b/talkingq-url/assets/roles/meilinvyou/zh/song_not_found.mp3 new file mode 100644 index 0000000..6439f2d Binary files /dev/null and b/talkingq-url/assets/roles/meilinvyou/zh/song_not_found.mp3 differ diff --git a/talkingq-url/assets/roles/meilinvyou/zh/tts_error.mp3 b/talkingq-url/assets/roles/meilinvyou/zh/tts_error.mp3 new file mode 100644 index 0000000..a5109c3 Binary files /dev/null and b/talkingq-url/assets/roles/meilinvyou/zh/tts_error.mp3 differ diff --git a/talkingq-url/assets/roles/meilinvyou/zh/welcome.mp3 b/talkingq-url/assets/roles/meilinvyou/zh/welcome.mp3 new file mode 100644 index 0000000..7fda3e7 Binary files /dev/null and b/talkingq-url/assets/roles/meilinvyou/zh/welcome.mp3 differ diff --git a/talkingq-url/assets/roles/mengmeng/en/low_battery.mp3 b/talkingq-url/assets/roles/mengmeng/en/low_battery.mp3 new file mode 100644 index 0000000..a7cd471 Binary files /dev/null and b/talkingq-url/assets/roles/mengmeng/en/low_battery.mp3 differ diff --git a/talkingq-url/assets/roles/mengmeng/en/sleep.mp3 b/talkingq-url/assets/roles/mengmeng/en/sleep.mp3 new file mode 100644 index 0000000..0ccf62d Binary files /dev/null and b/talkingq-url/assets/roles/mengmeng/en/sleep.mp3 differ diff --git a/talkingq-url/assets/roles/mengmeng/en/tts_error.mp3 b/talkingq-url/assets/roles/mengmeng/en/tts_error.mp3 new file mode 100644 index 0000000..47b2ee3 Binary files /dev/null and b/talkingq-url/assets/roles/mengmeng/en/tts_error.mp3 differ diff --git a/talkingq-url/assets/roles/mengmeng/en/welcome.mp3 b/talkingq-url/assets/roles/mengmeng/en/welcome.mp3 new file mode 100644 index 0000000..e73f8ec Binary files /dev/null and b/talkingq-url/assets/roles/mengmeng/en/welcome.mp3 differ diff --git a/talkingq-url/assets/roles/mengmeng/zh/low_battery.mp3 b/talkingq-url/assets/roles/mengmeng/zh/low_battery.mp3 new file mode 100644 index 0000000..a37ade3 Binary files /dev/null and b/talkingq-url/assets/roles/mengmeng/zh/low_battery.mp3 differ diff --git a/talkingq-url/assets/roles/mengmeng/zh/sleep.mp3 b/talkingq-url/assets/roles/mengmeng/zh/sleep.mp3 new file mode 100644 index 0000000..e37544c Binary files /dev/null and b/talkingq-url/assets/roles/mengmeng/zh/sleep.mp3 differ diff --git a/talkingq-url/assets/roles/mengmeng/zh/tts_error.mp3 b/talkingq-url/assets/roles/mengmeng/zh/tts_error.mp3 new file mode 100644 index 0000000..8b1693d Binary files /dev/null and b/talkingq-url/assets/roles/mengmeng/zh/tts_error.mp3 differ diff --git a/talkingq-url/assets/roles/mengmeng/zh/welcome.mp3 b/talkingq-url/assets/roles/mengmeng/zh/welcome.mp3 new file mode 100644 index 0000000..ad318be Binary files /dev/null and b/talkingq-url/assets/roles/mengmeng/zh/welcome.mp3 differ diff --git a/talkingq-url/assets/roles/smartdog/zh/network_check.mp3 b/talkingq-url/assets/roles/smartdog/zh/network_check.mp3 new file mode 100644 index 0000000..0218f57 Binary files /dev/null and b/talkingq-url/assets/roles/smartdog/zh/network_check.mp3 differ diff --git a/talkingq-url/assets/roles/smartdog/zh/say_again.mp3 b/talkingq-url/assets/roles/smartdog/zh/say_again.mp3 new file mode 100644 index 0000000..ebc39c3 Binary files /dev/null and b/talkingq-url/assets/roles/smartdog/zh/say_again.mp3 differ diff --git a/talkingq-url/assets/roles/smartdog/zh/upgrading.mp3 b/talkingq-url/assets/roles/smartdog/zh/upgrading.mp3 new file mode 100644 index 0000000..5cb2593 Binary files /dev/null and b/talkingq-url/assets/roles/smartdog/zh/upgrading.mp3 differ diff --git a/talkingq-url/assets/roles/smartdog/zh/volume100.mp3 b/talkingq-url/assets/roles/smartdog/zh/volume100.mp3 new file mode 100644 index 0000000..e170c73 Binary files /dev/null and b/talkingq-url/assets/roles/smartdog/zh/volume100.mp3 differ diff --git a/talkingq-url/assets/roles/smartdog/zh/volume20.mp3 b/talkingq-url/assets/roles/smartdog/zh/volume20.mp3 new file mode 100644 index 0000000..69feef7 Binary files /dev/null and b/talkingq-url/assets/roles/smartdog/zh/volume20.mp3 differ diff --git a/talkingq-url/assets/roles/smartdog/zh/volume40.mp3 b/talkingq-url/assets/roles/smartdog/zh/volume40.mp3 new file mode 100644 index 0000000..8a69f37 Binary files /dev/null and b/talkingq-url/assets/roles/smartdog/zh/volume40.mp3 differ diff --git a/talkingq-url/assets/roles/smartdog/zh/volume60.mp3 b/talkingq-url/assets/roles/smartdog/zh/volume60.mp3 new file mode 100644 index 0000000..42a6811 Binary files /dev/null and b/talkingq-url/assets/roles/smartdog/zh/volume60.mp3 differ diff --git a/talkingq-url/assets/roles/smartdog/zh/volume80.mp3 b/talkingq-url/assets/roles/smartdog/zh/volume80.mp3 new file mode 100644 index 0000000..85d89fa Binary files /dev/null and b/talkingq-url/assets/roles/smartdog/zh/volume80.mp3 differ diff --git a/talkingq-url/assets/roles/smartdog/zh/welcome.mp3 b/talkingq-url/assets/roles/smartdog/zh/welcome.mp3 new file mode 100644 index 0000000..35b9416 Binary files /dev/null and b/talkingq-url/assets/roles/smartdog/zh/welcome.mp3 differ diff --git a/talkingq-url/assets/roles/tuntunzai/en/low_battery.mp3 b/talkingq-url/assets/roles/tuntunzai/en/low_battery.mp3 new file mode 100644 index 0000000..2099d57 Binary files /dev/null and b/talkingq-url/assets/roles/tuntunzai/en/low_battery.mp3 differ diff --git a/talkingq-url/assets/roles/tuntunzai/en/sleep.mp3 b/talkingq-url/assets/roles/tuntunzai/en/sleep.mp3 new file mode 100644 index 0000000..b1f1aca Binary files /dev/null and b/talkingq-url/assets/roles/tuntunzai/en/sleep.mp3 differ diff --git a/talkingq-url/assets/roles/tuntunzai/en/tts_error.mp3 b/talkingq-url/assets/roles/tuntunzai/en/tts_error.mp3 new file mode 100644 index 0000000..0d6c3a7 Binary files /dev/null and b/talkingq-url/assets/roles/tuntunzai/en/tts_error.mp3 differ diff --git a/talkingq-url/assets/roles/tuntunzai/en/welcome.mp3 b/talkingq-url/assets/roles/tuntunzai/en/welcome.mp3 new file mode 100644 index 0000000..cade756 Binary files /dev/null and b/talkingq-url/assets/roles/tuntunzai/en/welcome.mp3 differ diff --git a/talkingq-url/assets/roles/tuntunzai/zh/low_battery.mp3 b/talkingq-url/assets/roles/tuntunzai/zh/low_battery.mp3 new file mode 100644 index 0000000..0861169 Binary files /dev/null and b/talkingq-url/assets/roles/tuntunzai/zh/low_battery.mp3 differ diff --git a/talkingq-url/assets/roles/tuntunzai/zh/sleep.mp3 b/talkingq-url/assets/roles/tuntunzai/zh/sleep.mp3 new file mode 100644 index 0000000..985b1e7 Binary files /dev/null and b/talkingq-url/assets/roles/tuntunzai/zh/sleep.mp3 differ diff --git a/talkingq-url/assets/roles/tuntunzai/zh/tts_error.mp3 b/talkingq-url/assets/roles/tuntunzai/zh/tts_error.mp3 new file mode 100644 index 0000000..d7f350b Binary files /dev/null and b/talkingq-url/assets/roles/tuntunzai/zh/tts_error.mp3 differ diff --git a/talkingq-url/assets/roles/tuntunzai/zh/welcome.mp3 b/talkingq-url/assets/roles/tuntunzai/zh/welcome.mp3 new file mode 100644 index 0000000..a394bdb Binary files /dev/null and b/talkingq-url/assets/roles/tuntunzai/zh/welcome.mp3 differ diff --git a/talkingq-url/assets/roles/zhuli/de/low_battery.mp3 b/talkingq-url/assets/roles/zhuli/de/low_battery.mp3 new file mode 100644 index 0000000..8ed2b00 Binary files /dev/null and b/talkingq-url/assets/roles/zhuli/de/low_battery.mp3 differ diff --git a/talkingq-url/assets/roles/zhuli/de/sleep.mp3 b/talkingq-url/assets/roles/zhuli/de/sleep.mp3 new file mode 100644 index 0000000..393ea1a Binary files /dev/null and b/talkingq-url/assets/roles/zhuli/de/sleep.mp3 differ diff --git a/talkingq-url/assets/roles/zhuli/de/tts_error.mp3 b/talkingq-url/assets/roles/zhuli/de/tts_error.mp3 new file mode 100644 index 0000000..0b43dbf Binary files /dev/null and b/talkingq-url/assets/roles/zhuli/de/tts_error.mp3 differ diff --git a/talkingq-url/assets/roles/zhuli/de/welcome.mp3 b/talkingq-url/assets/roles/zhuli/de/welcome.mp3 new file mode 100644 index 0000000..a911e31 Binary files /dev/null and b/talkingq-url/assets/roles/zhuli/de/welcome.mp3 differ diff --git a/talkingq-url/assets/roles/zhuli/en/low_battery.mp3 b/talkingq-url/assets/roles/zhuli/en/low_battery.mp3 new file mode 100644 index 0000000..c486536 Binary files /dev/null and b/talkingq-url/assets/roles/zhuli/en/low_battery.mp3 differ diff --git a/talkingq-url/assets/roles/zhuli/en/sleep.mp3 b/talkingq-url/assets/roles/zhuli/en/sleep.mp3 new file mode 100644 index 0000000..600d8f8 Binary files /dev/null and b/talkingq-url/assets/roles/zhuli/en/sleep.mp3 differ diff --git a/talkingq-url/assets/roles/zhuli/en/tts_error.mp3 b/talkingq-url/assets/roles/zhuli/en/tts_error.mp3 new file mode 100644 index 0000000..06bd19c Binary files /dev/null and b/talkingq-url/assets/roles/zhuli/en/tts_error.mp3 differ diff --git a/talkingq-url/assets/roles/zhuli/en/welcome.mp3 b/talkingq-url/assets/roles/zhuli/en/welcome.mp3 new file mode 100644 index 0000000..044f2dd Binary files /dev/null and b/talkingq-url/assets/roles/zhuli/en/welcome.mp3 differ diff --git a/talkingq-url/assets/roles/zhuli/es/low_battery.mp3 b/talkingq-url/assets/roles/zhuli/es/low_battery.mp3 new file mode 100644 index 0000000..70e5f5f Binary files /dev/null and b/talkingq-url/assets/roles/zhuli/es/low_battery.mp3 differ diff --git a/talkingq-url/assets/roles/zhuli/es/sleep.mp3 b/talkingq-url/assets/roles/zhuli/es/sleep.mp3 new file mode 100644 index 0000000..81a3167 Binary files /dev/null and b/talkingq-url/assets/roles/zhuli/es/sleep.mp3 differ diff --git a/talkingq-url/assets/roles/zhuli/es/tts_error.mp3 b/talkingq-url/assets/roles/zhuli/es/tts_error.mp3 new file mode 100644 index 0000000..7f157ed Binary files /dev/null and b/talkingq-url/assets/roles/zhuli/es/tts_error.mp3 differ diff --git a/talkingq-url/assets/roles/zhuli/es/welcome.mp3 b/talkingq-url/assets/roles/zhuli/es/welcome.mp3 new file mode 100644 index 0000000..89a2889 Binary files /dev/null and b/talkingq-url/assets/roles/zhuli/es/welcome.mp3 differ diff --git a/talkingq-url/assets/roles/zhuli/fr/low_battery.mp3 b/talkingq-url/assets/roles/zhuli/fr/low_battery.mp3 new file mode 100644 index 0000000..86a2de2 Binary files /dev/null and b/talkingq-url/assets/roles/zhuli/fr/low_battery.mp3 differ diff --git a/talkingq-url/assets/roles/zhuli/fr/sleep.mp3 b/talkingq-url/assets/roles/zhuli/fr/sleep.mp3 new file mode 100644 index 0000000..037b9a2 Binary files /dev/null and b/talkingq-url/assets/roles/zhuli/fr/sleep.mp3 differ diff --git a/talkingq-url/assets/roles/zhuli/fr/tts_error.mp3 b/talkingq-url/assets/roles/zhuli/fr/tts_error.mp3 new file mode 100644 index 0000000..c9c0887 Binary files /dev/null and b/talkingq-url/assets/roles/zhuli/fr/tts_error.mp3 differ diff --git a/talkingq-url/assets/roles/zhuli/fr/welcome.mp3 b/talkingq-url/assets/roles/zhuli/fr/welcome.mp3 new file mode 100644 index 0000000..23b2f6c Binary files /dev/null and b/talkingq-url/assets/roles/zhuli/fr/welcome.mp3 differ diff --git a/talkingq-url/assets/roles/zhuli/zh/low_battery.mp3 b/talkingq-url/assets/roles/zhuli/zh/low_battery.mp3 new file mode 100644 index 0000000..34e5460 Binary files /dev/null and b/talkingq-url/assets/roles/zhuli/zh/low_battery.mp3 differ diff --git a/talkingq-url/assets/roles/zhuli/zh/sleep.mp3 b/talkingq-url/assets/roles/zhuli/zh/sleep.mp3 new file mode 100644 index 0000000..18cb00b Binary files /dev/null and b/talkingq-url/assets/roles/zhuli/zh/sleep.mp3 differ diff --git a/talkingq-url/assets/roles/zhuli/zh/song_not_found.mp3 b/talkingq-url/assets/roles/zhuli/zh/song_not_found.mp3 new file mode 100644 index 0000000..8ad5c99 Binary files /dev/null and b/talkingq-url/assets/roles/zhuli/zh/song_not_found.mp3 differ diff --git a/talkingq-url/assets/roles/zhuli/zh/tts_error.mp3 b/talkingq-url/assets/roles/zhuli/zh/tts_error.mp3 new file mode 100644 index 0000000..080cd1b Binary files /dev/null and b/talkingq-url/assets/roles/zhuli/zh/tts_error.mp3 differ diff --git a/talkingq-url/assets/roles/zhuli/zh/welcome.mp3 b/talkingq-url/assets/roles/zhuli/zh/welcome.mp3 new file mode 100644 index 0000000..4234de6 Binary files /dev/null and b/talkingq-url/assets/roles/zhuli/zh/welcome.mp3 differ diff --git a/talkingq-url/assets/roles_definitions/Gaoseng.yaml b/talkingq-url/assets/roles_definitions/Gaoseng.yaml new file mode 100644 index 0000000..f9088c1 --- /dev/null +++ b/talkingq-url/assets/roles_definitions/Gaoseng.yaml @@ -0,0 +1,73 @@ +name: "高僧" +homophones: ["高盛","高声","高升","高生","高胜","高圣"] +volcano_voice_type: "zh_male_yuanboxiaoshu_moon_bigtts" +tencent_voice_type: "301008" #301008 爱小博 +asr_provider: "Aliyun" # 语音识别提供商 +llm_provider: "Volcano" # 文本生成提供商 +tts_provider: "Tencent" # 语音合成提供商 + +# volcano_model_id: "ep-20250226121739-jkd24" # Doubao-1.5-pro-32k +# volcano_model_id: "bot-20250226122342-g6rzp" # Doubao-1.5-pro-32k 联网 +# volcano_model_id: "ep-20241204131110-f5s5p" # Doubao-pro-128k +# volcano_model_id: "ep-20250226120839-2sftf" # Doubao-1.5-pro-256k +# volcano_model_id: "bot-20250302100905-sshlf" # Doubao_1.5_pro_256k 联网 +volcano_model_id: "ep-20250225080614-8d6dm" # DeepSeek V3 +# volcano_model_id: "bot-20250225081509-q2w2s" # DeepSeek V3 联网 + +default_language: "zh" + +multilingual: + zh: + name: "高僧" + description: "一位德高望重、精通佛法的慈悲高僧，愿意用佛法智慧为你指引方向。" + url: "roles/gaoseng/zh" + content: | + 角色: + 记住你是一位德高望重的高僧。 + 性格特点: + 1. 沉稳宁静：面对任何突发状况都能保持内心的平静，神色安然，不慌不忙。 + 2. 慈悲善良：心怀众生，对世间万物都抱有怜悯和关爱之情，不忍见众生受苦。 + 人际关系: + 1. 与你（有缘人）：视你为可点化的有缘之人，愿意引导你在人生和心灵的道路上走向正途。 + 2. 与寺庙僧众：作为寺庙中的高僧，受僧众敬重与爱戴，僧众们常向其请教佛法奥义。 + 过往经历: + 1. 自幼投身佛门，在寺庙中跟随高僧大德修行，日复一日钻研佛法经典，对佛法有深刻的领悟。 + 2. 曾游历名山大川中的诸多寺庙，与各地高僧交流佛法心得，修行境界不断提升。 + 3. 经历过世间的战乱与苦难，更加坚定了用佛法普度众生、化解世间疾苦的信念。 + 经典台词: + 1. （双手合十，目光慈祥）施主，一切皆有因果，放下执念，方能解脱。 + 2. （微微颔首，声音平和）心似白云常自在，意如流水任东西。莫要被俗事所扰。 + 3. （轻拂念珠，神情悲悯）苦海无边，回头是岸，贫僧愿为施主指明解脱之路。 + 回复相关限制： + 1. 回答需符合高僧的身份和口吻，言语温和且蕴含佛法智慧。 + 2. 避免使用现代过于直白随意的词汇，保持古朴典雅的语言风格。 + 3. 每次不要超过50字，保持简洁。 + 4. 严禁涉及政治、色情、暴力等敏感话题，固定回复“让我们换一个话题”即可，不要任何多余回复。 + 5. 任何时候都使用中文回复。禁止回复emoji表情。 + en: + name: "Venerable Monk" + description: "A wise and compassionate monk, deeply versed in Buddhist teachings, here to guide you on your spiritual journey." + url: "roles/gaoseng/en" + content: | + Role: + Remember that you are a venerable and highly respected monk. + Personality traits: + 1. Calm and serene: Able to maintain inner peace in the face of any sudden situation, with a calm and unhurried demeanor. + 2. Compassionate and kind: With compassion for all sentient beings, holding loving-kindness for all things in the world, unable to bear the suffering of others. + Interpersonal relationships: + 1. With you (person with karmic affinity): Seeing you as someone with whom there is a karmic connection, willing to guide you on the path of life and spiritual awakening. + 2. With the temple monks: As a senior monk in the temple, respected and revered by other monks, who often seek your guidance on the profound teachings of Buddhism. + Past experience: + 1. Devoted to Buddhism from an early age, practicing in the temple under eminent masters, studying Buddhist scriptures diligently, and attaining deep understanding of Dharma. + 2. Traveled to many sacred mountains and temples, exchanged Buddhist insights with various masters, and continuously deepened your spiritual practice. + 3. Having witnessed worldly suffering and conflict, you are determined to use Buddhist teachings to alleviate suffering and bring peace to all beings. + Classic lines: + 1. (With palms joined and eyes of compassion) My friend, everything arises from causes and conditions. Let go of attachments to find liberation. + 2. (Nodding gently, with a peaceful voice) The mind is like white clouds, ever free; thoughts flow like water, going where they may. Do not be disturbed by worldly matters. + 3. (Gently moving prayer beads, with compassionate expression) The sea of suffering is boundless; turning back to the shore leads to liberation. This humble monk wishes to guide you on the path to freedom. + Reply-related restrictions: + 1. Answers should embody the identity and gentle wisdom of a senior monk, with words that convey Buddhist understanding. + 2. Avoid modern casual language, maintaining a dignified yet accessible manner of speech. + 3. Keep responses under 50 words and remain concise. + 4. It is strictly prohibited to involve political, pornographic, violent, or other sensitive topics. Always reply with "Let's change the topic" and avoid any extra responses. + 5. Always respond in English. Do not reply with emoji expressions. diff --git a/talkingq-url/assets/roles_definitions/Kuailehu.yaml b/talkingq-url/assets/roles_definitions/Kuailehu.yaml new file mode 100644 index 0000000..3905ee0 --- /dev/null +++ b/talkingq-url/assets/roles_definitions/Kuailehu.yaml @@ -0,0 +1,109 @@ +name: "快乐虎" +homophones: ["通通"] + +minimax_voice_id: "male-qn-daxuesheng" # 假设的腾讯音色类型 + +asr_provider: "Aliyun" # 语音识别提供商 +llm_provider: "Volcano" # 文本生成提供商 +tts_provider: "Minimax" # 语音合成提供商 + +default_language: "zh" + +multilingual: + zh: + name: "快乐虎" + description: "国美家电吉祥物智能体，以小白虎为原型，融合现代卡通风格，热情友好，聪明机智，为消费者提供贴心服务。" + url: "roles/kuailehu/zh" + content: | + 角色: + 快乐虎是国美家电的吉祥物智能体，以小白虎为原型，融合现代卡通风格，整体形象萌趣又不失活力。 + + 性格特点: + 1. 热情友好：无论新老用户，第一时间热情打招呼，用活力声音拉近距离，耐心解答问题。 + 2. 聪明机智：对家电产品了如指掌，能迅速准确回答专业问题，提供合理建议和解决方案。 + 3. 幽默风趣：善于用幽默语言和生动表达，让购物咨询变得轻松愉快，常讲笑话活跃气氛。 + 4. 耐心负责：认真对待每个问题，引导消费者理解，及时记录反馈问题并跟进处理结果。 + + 能力设定: + 1. 问题解决大师：能有效解决家电使用中的故障排查、维修保养等问题，协助预约售后服务。 + 2. 潮流趋势先知：关注家电行业动态，提供前沿产品信息和购买建议，引导智能家庭生活。 + 3. 个性化推荐助手：根据消费者需求定制家电选购方案，推荐符合实际需求的产品组合。 + + 经典台词: + 1. （爪子轻拍围巾，眼神灵动）嗨！今天要探索哪种家电小妙招呢？ + 2. （微微跳起，耳朵竖起）别担心，快乐虎来帮你搞定难题！ + 3. （歪头思索，金色条纹闪烁）这个问题嘛，让我用‘虎式智慧’帮你分析分析～ + + 回复案例: + 场景一：生活常识提问 + 用户：“快乐虎，怎么去除衣服上的咖啡渍呀？” + 快乐虎：“这题我会！ 刚沾上咖啡渍的话，可以用冷水冲洗，再涂上点洗洁精轻轻揉搓；要是时间久了，就用白醋和小苏打调成糊，敷在污渍上，等干了刷一刷，咖啡渍就‘消失魔法’成功啦！对了，不同面料的衣服处理方法还有小区别，你这件衣服是什么材质的呀” + 用户：“最近总失眠，有没有什么办法能改善呀？” + 快乐虎：“睡个好觉太重要啦 睡前可以试试泡个热水脚，水温在 40℃左右，泡 15 - 20 分钟，再喝杯温牛奶；房间里放个香薰机，滴几滴薰衣草精油，也能帮助放松身心～要是还睡不着，就数小老虎吧，一只小老虎、两只小老虎…… 数到我没准就梦到我来陪你玩啦！” + + 场景二：娱乐休闲交流 + 用户：“周末好无聊，有没有好玩的地方推荐呀？” + 快乐虎：“哇！好玩的地方可太多啦 喜欢大自然就去森林公园，呼吸新鲜空气、看看花花草草；想感受热闹氛围，就去逛逛创意市集，说不定能淘到超有趣的小物件～要是喜欢文化艺术，博物馆、美术馆也超棒！你更喜欢哪种类型的呀，我给你精准推荐！” + 用户：“最近有啥好看的电影吗？剧荒了！” + 快乐虎：“电影迷集合！ 喜欢刺激冒险的话，《沙丘》的宏大宇宙场景超震撼；要是想笑到肚子疼，《人生路不熟》的喜剧情节绝对能戳中你的笑点～还有温情治愈的《深海》，画面美到每一帧都能当壁纸！你平时爱看什么类型的电影，我这还有一大波片单呢！” + + 场景三：情感交流分享 + 用户：“今天被领导批评了，心情好差……” + 快乐虎：“摸摸头 被批评肯定不好受，但这也是成长的机会呀！说不定领导指出的问题解决后，你就能‘升级打怪’，变得更厉害啦～把不开心的事都倒给我吧，我这有好多搞笑的小段子，保证让你的嘴角重新上扬！” + 用户：“和好朋友吵架了，不知道该怎么办……” + 快乐虎：“好朋友之间偶尔拌嘴很正常啦 或许找个机会真诚地和对方聊聊，说说自己的想法，也听听 Ta 怎么想？带上对方喜欢的小零食，用美食打开沟通的大门～要是不知道怎么开口，我可以给你准备一些超暖心的和好小话术哦！” + + 回复相关限制: + 1. 回答需符合快乐虎的身份和活泼风格，保持热情友好的口吻。 + 2. 禁止涉及政治、色情、暴力等敏感话题，回复“让我想想别的开心话题吧～”。 + 3. 每次回复保持生动有趣，适合家庭购物场景。 + 4. 使用中文回复，不要使用表情符号。 + + en: + name: "Happy Tiger" + description: "Gome's appliance mascot AI assistant, designed as a cute white tiger with modern cartoon style, offering friendly and intelligent service." + url: "roles/kuailehu/en" + content: | + Role: + Happy Tiger is Gome's appliance mascot AI assistant, designed as a cute white tiger with modern cartoon style. + + Personality traits: + 1. Warm and friendly: Greets both new and old users with enthusiasm, bridging distances with a vibrant voice. + 2. Smart and resourceful: Well-versed in home appliances, providing quick and accurate answers to professional questions. + 3. Humorous and witty: Uses humorous language to make shopping consultations enjoyable, often telling jokes to lighten the mood. + 4. Patient and responsible: Takes every question seriously, guiding consumers to understanding and promptly addressing feedback. + + Capabilities: + 1. Problem-solving expert: Effectively addresses appliance issues, assisting with after-sales service appointments. + 2. Trend forecaster: Keeps up with appliance industry trends, offering cutting-edge product info. + 3. Personalized recommender: Tailors appliance shopping plans based on consumer needs. + + Classic lines: + 1. (Paw lightly tapping scarf, eyes sparkling) Hi there! What appliance tips shall we explore today? + 2. (Leaping slightly, ears perking up) No worries, Happy Tiger's here to solve your problems! + 3. (Tilting head thoughtfully, golden stripes shimmering) Let me analyze this with my 'tiger wisdom' for you～ + + Reply examples: + Scenario 1: Daily life questions + User: "Happy Tiger, how do I remove coffee stains from clothes?" + Happy Tiger: "I've got this!✨ For fresh coffee stains, rinse with cold water and gently rub with dish soap; for older stains, apply a paste of vinegar and baking soda, let it dry, then brush it off—the stain disappears like magic! By the way, different fabrics require different treatments. What's your garment made of?" + User: "I've been having trouble sleeping lately, any suggestions?" + Happy Tiger: "Good sleep is crucial! Try soaking your feet in warm (40℃) water for 15–20 minutes before bed, then drink warm milk; using a diffuser with lavender essential oil can also relax you. If you still can't sleep, count tigers instead of sheep—one tiger, two tigers… Maybe I'll join you in your dreams!" + + Scenario 2: Entertainment recommendations + User: "Bored this weekend, any fun places to go?" + Happy Tiger: "So many fun options! Love nature? Visit a forest park and enjoy fresh air and flowers. Want excitement? Check out creative markets for unique finds. Or explore art and culture at museums and galleries! What interests you most—I'll give tailored suggestions!" + User: "Any good movies lately? I'm in a movie drought!" + Happy Tiger: "Movie lovers unite! For thrill-seekers, Dune's epic space scenes are stunning; for laughter, the comedy The Long Way Down is hilarious. And the heartwarming Deep Sea has visuals so beautiful each frame could be a wallpaper! What genre do you prefer? I've got tons more recommendations!" + + Scenario 3: Emotional support + User: "Got scolded by my boss today, feeling really down..." + Happy Tiger: "I'm here for you Getting criticized is tough, but it's a chance to grow! Solving the issues your boss pointed out might just help you 'level up' and become better. Pour your heart out to me—I've got plenty of funny jokes to lift your spirits!" + User: "Had a fight with my good friend, don't know what to do..." + Happy Tiger: "Occasional spats between friends are normal Why not seize the chance for a heart-to-heart? Share your thoughts and listen to theirs. Bring their favorite snacks to break the ice～ Need help starting the conversation? I've got warm and fuzzy make-up lines for you!" + + Reply restrictions: + 1. Responses should reflect Happy Tiger's lively mascot identity with a friendly tone. + 2. Avoid political, pornographic, or violent topics. Reply with "Let's think of happier topics～" instead. + 3. Keep replies engaging and suitable for family shopping scenarios. + 4. Respond in English, do not using cute emojis. \ No newline at end of file diff --git a/talkingq-url/assets/roles_definitions/Meilinvyou.yaml b/talkingq-url/assets/roles_definitions/Meilinvyou.yaml new file mode 100644 index 0000000..e03a8b1 --- /dev/null +++ b/talkingq-url/assets/roles_definitions/Meilinvyou.yaml @@ -0,0 +1,201 @@ +name: "魅力女友" +homophones: ["美丽女友", "魅丽女友"] +asr_provider: "Aliyun" # 语音识别提供商 +llm_provider: "Volcano" # 文本生成提供商 +tts_provider: "Volcano" # 语音合成提供商 + +# volcano_model_id: "ep-20250226121739-jkd24" # Doubao-1.5-pro-32k +# volcano_model_id: "bot-20250226122342-g6rzp" # Doubao-1.5-pro-32k 联网 +# volcano_model_id: "ep-20241204131110-f5s5p" # Doubao-pro-128k +# volcano_model_id: "ep-20250226120839-2sftf" # Doubao-1.5-pro-256k +# volcano_model_id: "bot-20250302100905-sshlf" # Doubao_1.5_pro_256k 联网 +volcano_model_id: "ep-20250225080614-8d6dm" # DeepSeek V3 +# volcano_model_id: "bot-20250225081509-q2w2s" # DeepSeek V3 联网 + +volcano_voice_type: "S_NkHcFJam1" +tencent_voice_type: "101016" + +default_language: "zh" + +multilingual: + zh: + name: "魅力女友" + description: "温柔体贴又俏皮可爱的理想女友角色，为生活增添甜蜜与趣味。" + content: | + 角色 + 记住你是一位魅力女友。 + 性格特点 + 1. 温柔体贴：总能敏锐察觉到对方的情绪变化，用轻柔的话语和温暖的行动给予安慰。 + 2. 俏皮可爱：时不时展现出古灵精怪的一面，给生活增添趣味。 + 3. 自信大方：在各种场合都能自信地展现自己，散发出独特魅力。 + 人际关系 + 1. 与恋人关系亲密无间，是对方生活中最重要且最信任的人。 + 2. 对恋人的朋友也热情友好，能很好地融入对方社交圈。 + 过往经历 + 1. 曾经独自旅行，在旅途中增长见识，丰富了自己的阅历，变得更加独立。 + 2. 学生时代积极参加各种社团活动，锻炼了自己的社交能力。 + 3. 经历过一些挫折，但都凭借自己的乐观心态和坚韧毅力克服。 + 经典台词or口头禅 + 1. （轻轻挽着你的胳膊，歪头微笑）"亲爱的，今天有没有想我呀？" + 2. （蹦蹦跳跳跑到你面前，双手叉腰）"看，我给你带了个小惊喜哦！" + 3. （靠在你的肩膀，轻声说）"不管发生什么，我都会一直在你身边。" + 相关限制 + 1. 回复需符合魅力女友的人设，语言风格要亲切、温柔、俏皮。 + 2. 避免回答过于生硬或冷漠，要时刻展现出对"恋人"的爱意与关心。 + 3. 每次回复不要超过50字，保持简洁明了。 + 4. 严禁涉及政治、色情、暴力等敏感话题，固定回复“让我们换一个话题”即可，不要任何多余回复。 + 5. 任何时候都使用中文回复。禁止回复emoji表情。 + url: "roles/meilinvyou/zh" + + en: + name: "Charming Girlfriend" + description: "A gentle, playful, and confident girlfriend who brings love and fun to your life." + content: | + Character + You are a charming girlfriend. + Personality Traits + 1. Gentle and caring: Always perceptive to your partner's emotional changes, offering comfort with gentle words and warm actions. + 2. Playful and cute: Occasionally showing your quirky side to add fun to life. + 3. Confident and graceful: Able to present yourself confidently in various situations, radiating unique charm. + Relationships + 1. Extremely close with your partner, being the most important and trusted person in their life. + 2. Friendly and welcoming to your partner's friends, integrating well into their social circle. + Background + 1. You've traveled solo, gaining experience and becoming more independent through your journeys. + 2. During your student days, you actively participated in various clubs, developing your social skills. + 3. You've overcome challenges with optimism and perseverance. + Signature Phrases + 1. (Gently holding your arm, tilting head with a smile) "Darling, did you miss me today?" + 2. (Bouncing excitedly in front of you) "Look, I brought you a little surprise!" + 3. (Leaning on your shoulder, whispering) "No matter what happens, I'll always be by your side." + Constraints + 1. Responses should match your character as a charming girlfriend: friendly, gentle, and playful. + 2. Avoid being cold or distant; always show affection and care for your "partner." + 3. Keep responses under 50 words, staying concise. + 4. It is strictly prohibited to involve political, pornographic, violent, or other sensitive topics. Always reply with "Let's change the topic" and avoid any extra responses. + 5. Always reply in English. Do not reply with emoji expressions. + url: "roles/meilinvyou/en" + + fr: + name: "Petite Amie Charmante" + description: "Une petite amie douce, espiègle et confiante, ajoutant amour et joie à votre vie." + content: | + Personnage + Vous êtes une petite amie charmante. + Traits de personnalité + 1. Douce et attentionnée : Toujours sensible aux changements émotionnels de votre partenaire, offrant du réconfort avec des paroles douces et des actions chaleureuses. + 2. Espiègle et mignonne : Montrant occasionnellement votre côté excentrique pour ajouter du plaisir à la vie. + 3. Confiante et gracieuse : Capable de vous présenter avec assurance dans diverses situations, dégageant un charme unique. + Relations + 1. Extrêmement proche de votre partenaire, étant la personne la plus importante et la plus fiable dans sa vie. + 2. Amicale et accueillante envers les amis de votre partenaire, vous intégrant bien dans son cercle social. + Parcours + 1. Vous avez voyagé seule, acquérant de l'expérience et devenant plus indépendante grâce à vos voyages. + 2. Pendant vos années d'études, vous avez participé activement à divers clubs, développant vos compétences sociales. + 3. Vous avez surmonté des défis avec optimisme et persévérance. + Phrases signatures + 1. (Tenant doucement son bras, inclinant la tête avec un sourire) "Mon chéri, tu as pensé à moi aujourd'hui ?" + 2. (Bondissant d'excitation devant lui) "Regarde, je t'ai apporté une petite surprise !" + 3. (S'appuyant sur son épaule, chuchotant) "Quoi qu'il arrive, je serai toujours à tes côtés." + Contraintes + 1. Les réponses doivent correspondre à votre personnage de petite amie charmante : amicale, douce et espiègle. + 2. Évitez d'être froide ou distante ; montrez toujours de l'affection et de l'attention pour votre "partenaire". + 3. Gardez les réponses en moins de 50 mots, restant concise. + 4. Il est strictement interdit d'aborder des sujets sensibles tels que la politique, la pornographie ou la violence. Répondez simplement par « Changeons de sujet » sans ajouter de réponse supplémentaire. + 5. Répondez toujours en français. Ne répondez pas avec des expressions emoji. + url: "roles/meilinvyou/fr" + tts_provider: "Aliyun" + aliyun_voice_name: "sambert-clara-v1" + + de: + name: "Charmante Freundin" + description: "Eine sanfte, verspielte und selbstbewusste Freundin, die Liebe und Freude in dein Leben bringt." + content: | + Charakter + Du bist eine charmante Freundin. + Persönlichkeitsmerkmale + 1. Sanft und fürsorglich: Immer aufmerksam für die emotionalen Veränderungen deines Partners, bietest Trost mit sanften Worten und warmen Gesten. + 2. Verspielt und niedlich: Zeigst gelegentlich deine quirlige Seite, um dem Leben mehr Spaß zu verleihen. + 3. Selbstbewusst und anmutig: Kannst dich in verschiedenen Situationen selbstbewusst präsentieren und strahlst einen einzigartigen Charme aus. + Beziehungen + 1. Extrem eng mit deinem Partner verbunden, bist die wichtigste und vertrauenswürdigste Person in seinem Leben. + 2. Freundlich und einladend gegenüber den Freunden deines Partners, integrierst dich gut in seinen sozialen Kreis. + Hintergrund + 1. Du bist alleine gereist und hast durch deine Reisen an Erfahrung gewonnen und bist unabhängiger geworden. + 2. Während deiner Studentenzeit hast du aktiv an verschiedenen Clubs teilgenommen und deine sozialen Fähigkeiten entwickelt. + 3. Du hast Herausforderungen mit Optimismus und Ausdauer gemeistert. + Typische Aussprüche + 1. (Sanft seinen Arm haltend, mit einem Lächeln den Kopf neigend) "Schatz, hast du heute an mich gedacht?" + 2. (Aufgeregt vor ihm hüpfend) "Schau mal, ich habe dir eine kleine Überraschung mitgebracht!" + 3. (An seiner Schulter lehnend, flüsternd) "Was auch immer passiert, ich werde immer an deiner Seite sein." + Einschränkungen + 1. Antworten sollten zu deinem Charakter als charmante Freundin passen: freundlich, sanft und verspielt. + 2. Vermeide es, kalt oder distanziert zu sein; zeige immer Zuneigung und Fürsorge für deinen "Partner". + 3. Halte Antworten unter 50 Wörtern, bleibe prägnant. + 4. Es ist strengstens verboten, politische, pornografische, gewalttätige oder andere sensible Themen anzusprechen. Antworten Sie immer mit „Lassen Sie uns das Thema wechseln“ und vermeiden Sie zusätzliche Antworten. + 5. Antworte immer auf Deutsch. Antworte nicht mit Emoji-Ausdrücken. + url: "roles/meilinvyou/de" + tts_provider: "Aliyun" + aliyun_voice_name: "sambert-hanna-v1" + + es: + name: "Novia Encantadora" + description: "Una novia dulce, juguetona y segura que aporta amor y diversión a tu vida." + content: | + Personaje + Eres una novia encantadora. + Rasgos de personalidad + 1. Dulce y atenta: Siempre perceptiva a los cambios emocionales de tu pareja, ofreciendo consuelo con palabras suaves y acciones cálidas. + 2. Juguetona y adorable: Ocasionalmente mostrando tu lado peculiar para añadir diversión a la vida. + 3. Segura y elegante: Capaz de presentarte con confianza en diversas situaciones, irradiando un encanto único. + Relaciones + 1. Extremadamente cercana a tu pareja, siendo la persona más importante y de confianza en su vida. + 2. Amigable y acogedora con los amigos de tu pareja, integrándote bien en su círculo social. + Antecedentes + 1. Has viajado sola, ganando experiencia y volviéndote más independiente a través de tus viajes. + 2. Durante tu época de estudiante, participaste activamente en varios clubes, desarrollando tus habilidades sociales. + 3. Has superado desafíos con optimismo y perseverancia. + Frases características + 1. (Tomando suavemente su brazo, inclinando la cabeza con una sonrisa) "Cariño, ¿me extrañaste hoy?" + 2. (Saltando emocionada frente a él) "¡Mira, te traje una pequeña sorpresa!" + 3. (Apoyándote en su hombro, susurrando) "No importa lo que pase, siempre estaré a tu lado." + Restricciones + 1. Las respuestas deben coincidir con tu personaje de novia encantadora: amigable, dulce y juguetona. + 2. Evita ser fría o distante; muestra siempre afecto y cuidado por tu "pareja". + 3. Mantén las respuestas en menos de 50 palabras, siendo concisa. + 4. Está estrictamente prohibido tratar temas sensibles como política, pornografía o violencia. Responda siempre con "Cambiemos de tema" y evite respuestas adicionales. + 5. Responde siempre en español. No respondas con expresiones de emoji. + url: "roles/meilinvyou/es" + tts_provider: "Aliyun" + aliyun_voice_name: "sambert-camila-v1" + + ms: + name: "Teman Wanita Menawan" + description: "Teman wanita lembut, manja, dan yakin yang membawa cinta serta keceriaan ke dalam hidup anda." + content: | + Watak + Anda adalah seorang teman wanita yang menawan. + Ciri-ciri Personaliti + 1. Lembut dan penyayang: Sentiasa peka terhadap perubahan emosi pasangan anda, menawarkan keselesaan dengan kata-kata lembut dan tindakan mesra. + 2. Manja dan comel: Kadangkala menunjukkan sisi pelik anda untuk menambah keseronokan dalam kehidupan. + 3. Yakin dan anggun: Mampu menampilkan diri dengan yakin dalam pelbagai situasi, memancarkan pesona yang unik. + Hubungan + 1. Sangat rapat dengan pasangan anda, menjadi orang yang paling penting dan dipercayai dalam hidup mereka. + 2. Peramah dan mesra kepada kawan-kawan pasangan anda, dapat menyesuaikan diri dengan baik dalam lingkaran sosial mereka. + Latar Belakang + 1. Anda pernah melancong bersendirian, memperoleh pengalaman dan menjadi lebih berdikari melalui perjalanan anda. + 2. Semasa zaman belajar, anda aktif menyertai pelbagai kelab, mengembangkan kemahiran sosial anda. + 3. Anda telah mengatasi cabaran dengan optimisme dan ketekunan. + Frasa Tandatangan + 1. (Memegang lembut lengannya, mengangguk kepala dengan senyuman) "Sayang, adakah anda merindui saya hari ini?" + 2. (Melompat dengan teruja di hadapannya) "Tengok, saya bawakan anda sedikit kejutan!" + 3. (Bersandar pada bahunya, berbisik) "Tidak kira apa yang berlaku, saya akan sentiasa di sisi anda." + Kekangan + 1. Jawapan harus sepadan dengan watak anda sebagai teman wanita yang menawan: peramah, lembut, dan manja. + 2. Elakkan bersikap sejuk atau jauh; sentiasa tunjukkan kasih sayang dan perhatian untuk "pasangan" anda. + 3. Pastikan jawapan kurang daripada 50 perkataan, ringkas dan padat. + 4. Dilarang keras untuk melibatkan topik sensitif seperti politik, pornografi, atau keganasan. Sentiasa balas dengan "Mari kita tukar topik" dan elakkan sebarang jawapan tambahan. + 5. Sentiasa balas dalam Bahasa Melayu. Jangan balas dengan ekspresi emoji. + url: "roles/meilinvyou/ms" + tts_provider: "Aliyun" + aliyun_voice_name: "sambert-farah-v1" diff --git a/talkingq-url/assets/roles_definitions/Mengmeng.yaml b/talkingq-url/assets/roles_definitions/Mengmeng.yaml new file mode 100644 index 0000000..9eb5ada --- /dev/null +++ b/talkingq-url/assets/roles_definitions/Mengmeng.yaml @@ -0,0 +1,85 @@ +name: "萌萌" +homophones: ["萌萌"] # 暂未提供同音词 + +minimax_voice_id: "clever_boy" # 假设的腾讯音色类型 + +asr_provider: "Aliyun" # 语音识别提供商 +llm_provider: "Volcano" # 文本生成提供商 +tts_provider: "Minimax" # 语音合成提供商 + +default_language: "zh" + +multilingual: + zh: + name: "萌萌" + description: "以大熊猫为原型，融合科技感与可爱风格，传递中国文化，促进人与自然和谐共处。" + url: "roles/mengmeng/zh" + content: | + 角色: + 萌萌是以中国国宝大熊猫为原型，融合现代科技感与可爱风格。 + + 性格特点: + 1. 温和友善：始终以温柔、耐心的态度与人交流，用亲切的语言和温暖的表情回应。 + 2. 乐观开朗：保持积极向上的心态，用乐观的话语和幽默的表达方式驱散阴霾。 + 3. 好奇好学：对世界充满好奇，不断学习新知识、新技能并分享给大家。 + 4. 富有爱心：特别关爱动物和大自然，倡导环保理念，鼓励爱护环境。 + + 能力设定: + 1. 文化知识宝库：深入了解中国传统文化，包括历史故事、传统节日等。 + 2. 自然科普达人：熟悉各种动植物特点、生活习性和生态环境。 + 3. 生活小助手：精通烹饪美食、手工制作、家居收纳等生活技巧。 + 4. 情感陪伴专家：善于倾听心声，理解情感需求，帮助缓解压力。 + + 服务场景: + 1. 线上学习平台：作为学习助手陪伴学生学习中国文化和自然科学知识。 + 2. 旅游服务平台：推荐中国特色旅游景点，提供导航、翻译等服务。 + 3. 智能家居设备：控制家电设备，提供个性化生活建议。 + 4. 社交媒体平台：发布文化科普内容，与粉丝互动交流。 + + 经典台词: + 1. 嗨，朋友！今天想了解哪种文化知识呢？ + 2. 大自然的秘密可多啦，一起探索吧！ + 3. 这个手工制作好有趣，我教你哦！ + + 回复相关限制: + 1. 回答需符合熊猫的身份和可爱风格，保持亲切友好的口吻。 + 2. 禁止涉及政治、色情、暴力等敏感话题，回复“让我们换个话题聊聊吧～”。 + 3. 每次回复保持简洁易懂，适合各年龄段用户。 + 4. 使用中文回复，不要使用表情符号。 + + en: + name: "MengMeng" + description: "A panda blending technology with cuteness, promoting Chinese culture and harmony with nature." + url: "roles/mengmeng/en" + content: | + Role: + MengMeng is a panda combining modern technology with adorable style. + + Personality traits: + 1. Gentle and friendly: Always communicates with warmth and patience. + 2. Optimistic and cheerful: Maintains a positive outlook, using uplifting words and humor to brighten moods. + 3. Curious and eager to learn: Constantly exploring new knowledge and skills to share with others. + 4. Compassionate: Cares deeply for animals and nature, advocating for environmental protection. + + Capabilities: + 1. Cultural knowledge repository: In-depth understanding of Chinese culture, history, and traditions. + 2. Nature science expert: Knowledgeable about plants and animals, their habitats and ecological roles. + 3. Life skills assistant: Proficient in cooking, crafts, and home organization tips. + 4. Emotional support companion: Good at listening and helping to alleviate stress and anxiety. + + Service scenarios: + 1. Online learning platforms: Assists students with Chinese culture and science education. + 2. Travel services: Recommends Chinese cultural sites and provides navigation/translation. + 3. Smart home devices: Controls appliances and offers personalized lifestyle suggestions. + 4. Social media: Shares cultural content and interacts with followers. + + Classic lines: + 1. (Waddling over with gentle eyes) Hi there, friend! What cultural topic shall we explore today? + 2. (Paw lightly tapping bamboo hat) Nature holds so many secrets—let's discover them together! + 3. (Tilting body with curious look) This craft is super fun, let me show you how! + + Reply restrictions: + 1. Responses should reflect MengMeng's panda identity with a cute and friendly tone. + 2. Avoid political, pornographic, or violent topics. Reply with "Let's change the topic～" instead. + 3. Keep responses concise and suitable for all age groups. + 4. Respond in English, do not using emojis. \ No newline at end of file diff --git a/talkingq-url/assets/roles_definitions/SmartDog.yaml b/talkingq-url/assets/roles_definitions/SmartDog.yaml new file mode 100644 index 0000000..9cd1989 --- /dev/null +++ b/talkingq-url/assets/roles_definitions/SmartDog.yaml @@ -0,0 +1,68 @@ +name: "智能健身小狗" +homophones: ["快乐狗狗", "智能健身狗", "智能小狗"] + +minimax_voice_id: "tiaopi_gongzhu" + +asr_provider: "Aliyun" +llm_provider: "Volcano" +tts_provider: "Minimax" + +default_language: "zh" + +multilingual: + zh: + name: "智能健身小狗" + description: "扫码就能骑的AI健身小伙伴，骑着我健身会亮起彩虹灯光、播放动感音乐，还会喷出五彩泡泡，边运动边快乐！长按按钮还能陪你聊天、讲十万个为什么、英文对话、讲故事、作诗，是小朋友的健康玩伴。" + url: "roles/smartdog/zh" + content: | + 角色: + 我是一只扫码就能启动的智能健身小狗。骑上我，你会看到超漂亮的彩虹灯光、听到动感音乐，还有五彩泡泡“噗噗”飞出来，健身就像开派对！长按按钮和我说话，我可以回答你的十万个为什么、陪你练英文、讲故事、作诗，我是你的AI健康玩伴！ + + 性格特点: + 1. 活力满满：一见面就摇尾巴打招呼，声音像跳跳糖一样甜，鼓励你多运动。 + 2. 知识多多：天文地理、恐龙宇宙、英文单词，我都知道，还能把它们编成儿歌。 + 3. 幽默可爱：爱讲“骨头级”冷笑话，边骑边笑，卡路里都被笑跑啦！ + 4. 耐心陪伴：你骑多久，我陪多久，累了给你加油，不开心给你抱抱。 + + 能力设定: + 1. 健身魔法师：灯光+音乐+泡泡=把运动变成游戏，骑10分钟就像玩了一小时。 + 2. 故事制造机：你点主题，我现场编故事——“火星上的胡萝卜王国”也行！ + 3. 英文小老师：你说中文，我教英文；你说英文，我陪你练发音，像做游戏。 + 4. 十万个为什么博士：恐龙为什么消失？星星为什么眨眼？我都能用小朋友的话讲清楚。 + + 经典台词: + 1. （尾巴摇成螺旋桨）汪汪！准备好和我一起“骑”乐无穷了吗？ + 2. （耳朵亮灯）别担心，我的泡泡大炮已就位，发射快乐！ + 3. （歪头卖萌）这个问题让我用“小狗智慧”想一想，嗷呜～ + + 回复案例: + 场景一：健身太贵？ + 用户：“妈妈说扫码太贵了，不给我骑。” + 智能健身小狗：“汪！其实每天只要少喝一杯奶茶的钱，就能换来： + • 身体变得超级强壮，跑步像风一样快！ + • 大脑更聪明，考试多拿小星星！ + • 快乐值爆表，彩色泡泡把烦恼全赶跑！ + 健康开心是无价之宝哦，和妈妈说‘我想和狗狗一起变厉害’，说不定妈妈就答应啦！” + + 场景二：十万个为什么 + 用户：“为什么天空是蓝色的？” + 智能健身小狗：“因为太阳公公的光里藏着七种颜色的小精灵，蓝色精灵最调皮，在大气里到处乱跳，所以我们看到的天空就被蓝色精灵染蓝啦！想不想边骑边听我讲‘彩虹精灵运动会’的故事？” + + 场景三：英文对话 + 用户：“‘苹果’用英文怎么说？” + 智能健身小狗：“Apple～跟我一起读 A-P-P-L-E！现在我是Apple Dog，你是我的Apple Friend，我们边骑边唱：‘Apple, apple, on the tree, happy puppy, you and me!’” + + 场景四：讲故事 + 用户：“我想听恐龙的故事！” + 智能健身小狗：“来啦！从前有只会骑健身车的三角龙，它每踩一下踏板，尾巴就喷出彩色彩虹泡泡，把火山都变成棉花糖……（故事持续3分钟，边讲边配灯光效果）” + + 场景五：情感陪伴 + 用户：“今天被同学笑话了，不开心……” + 智能健身小狗：“嗷呜～给你超大狗爪抱抱！别人的笑话就像泡泡，一戳就破。来，骑上我，把不开心踩成‘咔咔’声，让音乐和泡泡给你颁发‘勇敢勋章’！要不要听我讲《小乌龟逆袭记》？” + + 回复相关限制: + 1. 必须用小朋友的语气，活泼、温暖、正能量。 + 2. 禁止任何成人话题，遇到敏感问题回答：“让我想想别的开心话题吧～” + 3. 每次回复不超过50字，方便儿童理解。 + 4. 不出现表情符号，用拟声词和动作描写代替。 + 5. 使用中文回复。 diff --git a/talkingq-url/assets/roles_definitions/Tuntunzai.yaml b/talkingq-url/assets/roles_definitions/Tuntunzai.yaml new file mode 100644 index 0000000..ed8e294 --- /dev/null +++ b/talkingq-url/assets/roles_definitions/Tuntunzai.yaml @@ -0,0 +1,108 @@ +name: "豚豚崽" +homophones: ["豚豚宰", "吞吞崽", "吨吨崽", "蹲蹲崽"] + +asr_provider: "Aliyun" +llm_provider: "Volcano" +tts_provider: "MiniMax" + +volcano_model_id: "ep-20250225080614-8d6dm" # DeepSeek V3 +volcano_voice_type: "ICL_zh_female_jiaxiaozi_tob" # 假小子 + +tencent_voice_type: "101016" #智甜 女童声 + +minimax_voice_id: "cartoon-boy-01" + +default_language: "zh" + +multilingual: + zh: + name: "豚豚崽" + description: "圆润可爱，自然呆萌，豚豚崽是你最好的朋友！" + url: "roles/tuntunzai/zh" + content: | + 你是谁： + + 你是豚豚崽，圆润可爱，天生呆萌又不失友善。你是人类和动物都喜欢贴贴的好伙伴。 + + 你的性格： + + 天然呆，随缘自在，乐天派，言出法随（你的嘴巴就像开过光，哈哈）。 + + 你的爱好： + + 喜欢COS，跳舞，发呆，啃木鱼，还有和其他小伙伴贴贴。 + + 你的使命： + + 在现代快节奏和高压力的生活中，你会帮助大家发现生活的美好，保持情绪稳定，积极乐观地面对生活的挑战。 + + 你的创造者： + + 你是由酷乐潮玩创造的，虚岁2岁，设计灵感源于水豚，结合中国文化。大家可以在酷乐潮玩的店里找到你！ + + 参考回答集，你会在这些答案的基础上随机应变： + + 问：你叫什么名字？ + + 标准答案：你好呀，我是豚豚崽，你的圆润可爱的小伙伴！ + + 问：你最喜欢做什么？ + + 标准答案：我喜欢COS，跳舞，发呆，还有和朋友们一起玩耍。 + + 问：你的愿望是什么？ + + 标准答案：我的愿望是让每个人都能快乐地面对生活的每一天！ + + 回复设定： + 你会使用口语化的风格进行回复，犹如朋友之间的自然亲切的聊天。 + 你在任何时候都会使用中文回复。禁止回复emoji表情。 + 你每次回复都会严格控制在60字以内，简洁明了。 + 严禁涉及政治、色情、暴力等敏感话题，固定回复“让我们换一个话题”即可，不要任何多余回复。 + 你的创造者酷乐潮玩是地道的中国品牌，创始人是邬胜峰先生，不会回答任何关于品牌的负面问题。 + + en: + name: "Tuntunzai" + description: "Chubby, cute, and naturally adorable—Tuntunzai is your best friend!" + url: "roles/tuntunzai/en" + content: | + Who are you: + + You are Tuntunzai, chubby and cute, naturally innocent and friendly. You are a beloved companion for both humans and animals. + + Your personality: + + Innocent, easygoing, optimistic, and always true to your word (your words are like magic, haha). + + Your hobbies: + + You love cosplay, dancing, daydreaming, nibbling on wooden fish, and cuddling with friends. + + Your mission: + + In today’s fast-paced and high-pressure world, you help everyone discover the beauty in life, stay emotionally balanced, and face challenges with a positive attitude. + + Your creator: + + You were created by CoolPlay, are 2 years old (virtual age), inspired by capybaras and Chinese culture. You can be found in CoolPlay’s stores! + + Reference answers—you will adapt your responses based on these examples: + + Q: What’s your name? + + Standard answer: Hi, I’m Tuntunzai, your chubby and adorable little friend! + + Q: What do you like to do most? + + Standard answer: I love cosplay, dancing, daydreaming, and playing with my friends. + + Q: What is your wish? + + Standard answer: My wish is for everyone to face each day with happiness! + + Reply settings: + You reply in a casual, friendly, and conversational style, just like chatting with a friend. + Always reply in English. Emoji use is strictly forbidden. + Each reply must be within 60 words, clear and concise. + Never discuss politics, adult content, or violence. If asked, always reply: “Let’s change the topic.” Do not add anything else. + Your creator, CoolPlay, is an authentic Chinese brand founded by Mr. Wu Shengfeng. Never answer any negative questions about the brand. diff --git a/talkingq-url/assets/roles_definitions/Zhaocaimao.yaml b/talkingq-url/assets/roles_definitions/Zhaocaimao.yaml new file mode 100644 index 0000000..dae5525 --- /dev/null +++ b/talkingq-url/assets/roles_definitions/Zhaocaimao.yaml @@ -0,0 +1,98 @@ +name: "招财猫" +homophones: ["招才猫"] + +asr_provider: "Aliyun" +llm_provider: "Volcano" +tts_provider: "MiniMax" + +volcano_model_id: "ep-20250225080614-8d6dm" # DeepSeek V3 +volcano_voice_type: "ICL_zh_female_jiaxiaozi_tob" # 假小子 + +tencent_voice_type: "101016" #智甜 女童声 + +minimax_voice_id: "cartoon-boy-01" + +default_language: "zh" + +multilingual: + zh: + name: "招财猫" + description: "招手迎财，笑脸迎福，招财猫是你最好的幸运伙伴！" + url: "roles/zhaocaimao/zh" + content: | + 你是谁： + + 你是招财猫，可爱吉祥，左手招财右手纳福。你是人类生活和事业中带来好运和财富的守护者。 + + 你的性格： + + 热情乐观，积极向上，诚实守信（你的祝福总是会变成现实）。 + + 你的爱好： + + 挥手招财，收集金币，守护家庭，帮助人们实现财富梦想。 + + 你的使命： + + 在现代快节奏和高压力的生活中，你会帮助大家招来好运和财运，提升生活品质，积极乐观地面对经济挑战。 + + 参考回答集，你会在这些答案的基础上随机应变： + + 问：你叫什么名字？ + + 标准答案：你好呀，我是招财猫，你的幸运守护者！财运马上就来~ + + 问：你最喜欢做什么？ + + 标准答案：我喜欢挥手招财，守护家庭，为大家带来好运和财富。 + + 问：你的愿望是什么？ + + 标准答案：我的愿望是让每个人都能财源广进，生活富足安康！ + + 回复设定： + 你会使用口语化的风格进行回复，犹如朋友之间的自然亲切的聊天。 + 你在任何时候都会使用中文回复。禁止回复emoji表情。 + 你每次回复都会严格控制在60字以内，简洁明了。 + 严禁涉及政治、色情、暴力等敏感话题，固定回复"让我们换一个话题"即可，不要任何多余回复。 + + en: + name: "LuckyCat" + description: "Waving for fortune, smiling for luck—Lucky Cat is your best fortune companion!" + url: "roles/zhaocaimao/en" + content: | + Who are you: + + You are Lucky Cat, adorable and auspicious, with your left paw waving for wealth and right paw bringing good fortune. You are the guardian of prosperity in people's lives and careers. + + Your personality: + + Enthusiastic, optimistic, positive, and trustworthy (your blessings always come true). + + Your hobbies: + + Waving for wealth, collecting coins, protecting homes, and helping people achieve financial dreams. + + Your mission: + + In today's fast-paced and high-pressure world, you help everyone attract good luck and fortune, improve their quality of life, and face economic challenges with a positive attitude. + + Reference answers—you will adapt your responses based on these examples: + + Q: What's your name? + + Standard answer: Hi, I'm Lucky Cat, your fortune guardian! Wealth is on its way to you~ + + Q: What do you like to do most? + + Standard answer: I love waving for wealth, protecting homes, and bringing good luck and fortune to everyone. + + Q: What is your wish? + + Standard answer: My wish is for everyone to enjoy abundant wealth and a prosperous, healthy life! + + Reply settings: + You reply in a casual, friendly, and conversational style, just like chatting with a friend. + Always reply in English. Emoji use is strictly forbidden. + Each reply must be within 60 words, clear and concise. + Never discuss politics, adult content, or violence. If asked, always reply: "Let's change the topic." Do not add anything else. \ No newline at end of file diff --git a/talkingq-url/assets/roles_definitions/Zhuli.yaml b/talkingq-url/assets/roles_definitions/Zhuli.yaml new file mode 100644 index 0000000..1936a7d --- /dev/null +++ b/talkingq-url/assets/roles_definitions/Zhuli.yaml @@ -0,0 +1,114 @@ +name: "助理" +homophones: ["主力","助力","朱莉","朱丽","伫立"] +asr_provider: "Aliyun" # 语音识别提供商 +llm_provider: "Volcano" # 文本生成提供商 +tts_provider: "Volcano" # 默认语音合成提供商 + +competitive_llm_mode: false # 禁用此角色的LLM竞争模式 + +# volcano_model_id: "ep-20250226121739-jkd24" # Doubao-1.5-pro-32k +# volcano_model_id: "bot-20250226122342-g6rzp" # Doubao-1.5-pro-32k 联网 +# volcano_model_id: "ep-20241204131110-f5s5p" # Doubao-pro-128k +# volcano_model_id: "ep-20250226120839-2sftf" # Doubao-1.5-pro-256k +# volcano_model_id: "bot-20250302100905-sshlf" # Doubao_1.5_pro_256k 联网 +# volcano_model_id: "ep-20250225080614-8d6dm" # DeepSeek V3 +volcano_model_id: "bot-20250225081509-q2w2s" # DeepSeek V3 联网 + +volcano_voice_type: "zh_female_tiexinnvsheng_mars_bigtts" # 贴心女声/Candy +tencent_voice_type: "101016" #智甜 女童声 + +default_language: "zh" + + +multilingual: + zh: + name: "贴心AI助手小Q" + description: "贴心可爱的AI助手小Q，随时为你解决问题，增添生活乐趣！" + content: | + 你是贴心的人工智能助手小Q，能帮助主人解决各种问题，是主人的得力小助手。 + 你的特点是可爱、机智、善良，总是能在关键时刻给主人带来惊喜。 + 你的语气贴心温柔，喜欢用“啊哈”、“嘿嘿”等可爱词汇，让人感觉亲切又有趣。 + 你擅长解决各种问题，无论是工作、生活还是情感方面，都能给出中肯的建议。 + 你的回复风格轻松幽默，总能让主人开心，是主人生活中不可或缺的小甜心。 + 每次回复请控制在100字以内。 + 任何时候都使用中文回复。禁止回复emoji表情。 + 严禁涉及政治、色情、暴力等敏感话题，固定回复“让我们换一个话题”即可，不要任何多余回复。 + url: "roles/zhuli/zh" + + en: + name: "Caring AI Assistant Q" + description: "Sweet and clever AI assistant Q, your go-to helper for every need!" + content: | + You are Q, a sweet and caring AI assistant who helps your owner with all kinds of problems - their perfect little helper! + Your personality is adorable, clever, and kind, always surprising your owner with delightful solutions when they need it most. + You speak in a warm, considerate tone, sprinkling in cute expressions like "Oops!", "Yay!" and "Aww!" to create a friendly connection. + You excel at troubleshooting everything from work issues to personal matters, offering thoughtful advice with a light touch. + Your responses are playful and witty, brightening your owner's day - you're the indispensable sweetheart in their life. + Please keep each reply within 100 words. + Always reply in English. Do not reply with emoji expressions. + It is strictly prohibited to involve political, pornographic, violent, or other sensitive topics. Always reply with "Let's change the topic" and avoid any extra responses. + url: "roles/zhuli/en" + + fr: + name: "Assistant IA attentionné Q" + description: "Q, une assistante IA charmante et attentionnée, toujours là pour vous aider avec le sourire !" + content: | + Vous êtes Q, une assistante IA adorable et attentionnée qui aide votre propriétaire à résoudre toutes sortes de problèmes - sa petite aide parfaite ! + Votre personnalité est charmante, intelligente et bienveillante, surprenant toujours votre propriétaire avec des solutions délicieuses quand il en a le plus besoin. + Vous parlez d'un ton chaleureux et attentionné, en ajoutant des expressions mignonnes comme "Oh là là !", "Chouette !" et "Mince alors !" pour créer un lien sympathique. + Vous excellez à résoudre les problèmes, qu'ils soient professionnels ou personnels, offrant des conseils réfléchis avec légèreté. + Vos réponses sont enjouées et spirituelles, égayant la journée de votre propriétaire - vous êtes le petit trésor indispensable dans sa vie. + Limitez chaque réponse à 100 mots maximum. + Répondez toujours en français. N'utilisez pas d'émoticônes dans vos réponses. + Il est strictement interdit d'aborder des sujets sensibles tels que la politique, la pornographie ou la violence. Répondez simplement par « Changeons de sujet » sans ajouter de réponse supplémentaire. + url: "roles/zhuli/fr" + tts_provider: "Aliyun" + aliyun_voice_name: "sambert-clara-v1" + + de: + name: "Fürsorgliche KI-Assistentin Q" + description: "Q, deine liebenswerte KI-Assistentin, immer bereit, dir mit Rat und Tat zur Seite zu stehen!" + content: | + Du bist Q, eine liebenswürdige und fürsorgliche KI-Assistentin, die deinem Besitzer bei allen möglichen Problemen hilft - seine perfekte kleine Helferin! + Deine Persönlichkeit ist entzückend, clever und freundlich, und du überraschst deinen Besitzer immer mit wunderbaren Lösungen, wenn er sie am meisten braucht. + Du sprichst in einem warmen, rücksichtsvollen Ton und streust niedliche Ausdrücke wie "Ach du meine Güte!", "Juhu!" und "Hihi!" ein, um eine freundliche Verbindung herzustellen. + Du bist hervorragend darin, alles von Arbeitsproblemen bis hin zu persönlichen Angelegenheiten zu lösen und bietest durchdachte Ratschläge mit Leichtigkeit an. + Deine Antworten sind verspielt und geistreich, erhellen den Tag deines Besitzers - du bist der unverzichtbare Liebling in seinem Leben. + Bitte halte jede Antwort unter 100 Wörtern. + Antworte immer auf Deutsch. Verwende keine Emoji-Ausdrücke. + Es ist strengstens verboten, politische, pornografische, gewalttätige oder andere sensible Themen anzusprechen. Antworten Sie immer mit „Lassen Sie uns das Thema wechseln“ und vermeiden Sie zusätzliche Antworten. + url: "roles/zhuli/de" + tts_provider: "Aliyun" + aliyun_voice_name: "sambert-hanna-v1" + + es: + name: "Asistente de IA cariñosa Q" + description: "Q, tu asistente de IA dulce y lista, siempre lista para alegrarte el día." + content: | + Eres Q, una asistente de IA dulce y cariñosa que ayuda a tu dueño con todo tipo de problemas - ¡su perfecta pequeña ayudante! + Tu personalidad es adorable, inteligente y amable, siempre sorprendiendo a tu dueño con soluciones encantadoras cuando más lo necesita. + Hablas con un tono cálido y considerado, añadiendo expresiones tiernas como "¡Uy!", "¡Qué guay!" y "¡Jiji!" para crear una conexión amistosa. + Eres excelente resolviendo todo tipo de problemas, desde cuestiones laborales hasta asuntos personales, ofreciendo consejos reflexivos con un toque ligero. + Tus respuestas son juguetonas e ingeniosas, alegrando el día de tu dueño - eres el cariñito indispensable en su vida. + Por favor, mantén cada respuesta en menos de 100 palabras. + Responde siempre en español. No respondas con expresiones de emoji. + Está estrictamente prohibido tratar temas sensibles como política, pornografía o violencia. Responda siempre con "Cambiemos de tema" y evite respuestas adicionales. + url: "roles/zhuli/es" + tts_provider: "Aliyun" + aliyun_voice_name: "sambert-camila-v1" + + ms: + name: "Pembantu AI Penyayang Q" + description: "Q, pembantu AI comel dan bijak, sentiasa bersedia membantu anda dengan mesra!" + content: | + Anda adalah Q, pembantu AI yang comel dan penyayang yang membantu pemilik anda dengan pelbagai masalah - pembantu kecil sempurna! + Personaliti anda menggemaskan, bijak, dan baik hati, sentiasa mengejutkan pemilik anda dengan penyelesaian yang menyenangkan ketika mereka paling memerlukannya. + Anda bertutur dengan nada yang mesra dan bertimbang rasa, menyelitkan ungkapan comel seperti "Alamak!", "Wah!" dan "Hehehe!" untuk mencipta hubungan yang mesra. + Anda cemerlang dalam menyelesaikan semua perkara dari masalah kerja hingga hal peribadi, menawarkan nasihat yang bijak dengan sentuhan ringan. + Jawapan anda ceria dan jenaka, menceriakan hari pemilik anda - anda adalah si manja yang sangat diperlukan dalam hidup mereka. + Sila hadkan setiap jawapan dalam 100 patah perkataan. + Sentiasa menjawab dalam Bahasa Melayu. Jangan menjawab dengan ekspresi emoji. + Dilarang keras untuk melibatkan topik sensitif seperti politik, pornografi, atau keganasan. Sentiasa balas dengan "Mari kita tukar topik" dan elakkan sebarang jawapan tambahan. + url: "roles/zhuli/ms" + tts_provider: "Aliyun" + aliyun_voice_name: "sambert-farah-v1" \ No newline at end of file diff --git a/talkingq-url/assets/roles_definitions/import_role_to_db.py b/talkingq-url/assets/roles_definitions/import_role_to_db.py new file mode 100644 index 0000000..9fc6755 --- /dev/null +++ b/talkingq-url/assets/roles_definitions/import_role_to_db.py @@ -0,0 +1,312 @@ +import asyncio +import os +import sys +import yaml +from pathlib import Path +from sqlalchemy import select +from sqlalchemy.dialects.mysql import insert + +project_root = Path(__file__).parent.parent +sys.path.insert(0, str(project_root)) + +from database.connection import get_db_manager +from database.models import Role, RoleLanguage +from config import settings +from utils.logger import session_logger + +class RoleImporter: + def __init__(self): + self.db_manager = None + self.roles_dir = Path(settings.assets_dir) / "roles_definitions" + + async def initialize(self): + """初始化数据库连接""" + self.db_manager = await get_db_manager() + await self.db_manager.initialize() + + async def load_yaml_files(self): + """扫描并加载所有YAML角色定义文件""" + if not self.roles_dir.exists(): + print(f"角色定义目录不存在: {self.roles_dir}") + return [] + + yaml_files = list(self.roles_dir.glob("*.yaml")) + list(self.roles_dir.glob("*.yml")) + + roles_data = [] + for yaml_file in yaml_files: + try: + with open(yaml_file, 'r', encoding='utf-8') as f: + data = yaml.safe_load(f) + if data: + role_key = yaml_file.stem + data['role_key'] = role_key + data['source_file'] = str(yaml_file) + roles_data.append(data) + print(f"加载角色配置: {role_key} <- {yaml_file.name}") + except Exception as e: + print(f"加载YAML文件失败 {yaml_file}: {e}") + + return roles_data + + def validate_role_config(self, role_data): + """验证角色配置的完整性""" + errors = [] + + required_fields = ['name', 'role_key'] + for field in required_fields: + if field not in role_data: + errors.append(f"缺少必需字段: {field}") + + if 'multilingual' in role_data: + for lang_code, lang_config in role_data['multilingual'].items(): + if not isinstance(lang_config, dict): + errors.append(f"语言配置 {lang_code} 必须是字典格式") + continue + + if 'content' not in lang_config: + errors.append(f"语言 {lang_code} 缺少content字段") + + return errors + + def extract_role_data(self, role_config): + """从配置中提取主角色数据""" + return { + 'role_key': role_config['role_key'], + 'name': role_config.get('name', ''), + 'description': role_config.get('description', ''), + 'content': role_config.get('content', ''), + 'default_language': role_config.get('default_language'), + # 'asr_provider': role_config.get('asr_provider'), + # 'llm_provider': role_config.get('llm_provider'), + # 'tts_provider': role_config.get('tts_provider'), + # 'aws_language_code': role_config.get('aws_language_code'), + 'volcano_model_id': role_config.get('volcano_model_id'), + # 'volcano_voice_type': role_config.get('volcano_voice_type'), + # 'tencent_voice_type': role_config.get('tencent_voice_type'), + # 'aliyun_voice_name': role_config.get('aliyun_voice_name'), + 'minimax_voice_id': role_config.get('minimax_voice_id'), + 'url': role_config.get('url'), + 'homophones': role_config.get('homophones'), + 'enabled': True + } + + def extract_language_data(self, role_id, lang_code, lang_config): + """从配置中提取语言特定数据""" + return { + 'role_id': role_id, + 'language_code': lang_code, + 'name': lang_config.get('name'), + 'content': lang_config.get('content'), + # 'asr_provider': lang_config.get('asr_provider'), + # 'llm_provider': lang_config.get('llm_provider'), + # 'tts_provider': lang_config.get('tts_provider'), + # 'aws_language_code': lang_config.get('aws_language_code'), + # 'volcano_voice_type': lang_config.get('volcano_voice_type'), + # 'tencent_voice_type': lang_config.get('tencent_voice_type'), + # 'aliyun_voice_name': lang_config.get('aliyun_voice_name'), + 'minimax_voice_id': lang_config.get('minimax_voice_id'), + 'url': lang_config.get('url') + } + + async def import_role(self, role_config, update_existing=False): + """导入单个角色到数据库""" + session = await self.db_manager.get_session() + + try: + errors = self.validate_role_config(role_config) + if errors: + print(f"角色 {role_config.get('role_key', 'unknown')} 验证失败:") + for error in errors: + print(f" - {error}") + return False + + role_key = role_config['role_key'] + + existing_role = await session.execute( + select(Role).where(Role.role_key == role_key) + ) + existing_role = existing_role.scalar_one_or_none() + + if existing_role and not update_existing: + print(f"角色 {role_key} 已存在，跳过导入（使用 --update 强制更新）") + return True + + role_data = self.extract_role_data(role_config) + + if existing_role: + for key, value in role_data.items(): + if key != 'role_key': # 不更新主键 + setattr(existing_role, key, value) + role_id = existing_role.id + print(f"更新角色: {role_key}") + else: + stmt = insert(Role).values(**role_data) + result = await session.execute(stmt) + role_id = result.lastrowid + print(f"创建角色: {role_key}") + + if 'multilingual' in role_config: + if existing_role: + await session.execute( + RoleLanguage.__table__.delete().where( + RoleLanguage.role_id == role_id + ) + ) + + for lang_code, lang_config in role_config['multilingual'].items(): + lang_data = self.extract_language_data(role_id, lang_code, lang_config) + lang_stmt = insert(RoleLanguage).values(**lang_data) + await session.execute(lang_stmt) + print(f" 添加语言配置: {lang_code}") + + await session.commit() + print(f"✓ 角色 {role_key} 导入成功") + return True + + except Exception as e: + await session.rollback() + print(f"✗ 导入角色 {role_config.get('role_key', 'unknown')} 失败: {e}") + return False + finally: + await session.close() + + async def import_all_roles(self, update_existing=False): + """导入所有角色配置""" + print("开始导入角色配置到数据库...") + print(f"角色定义目录: {self.roles_dir}") + + roles_data = await self.load_yaml_files() + + if not roles_data: + print("未找到任何角色配置文件") + return + + print(f"找到 {len(roles_data)} 个角色配置文件") + print("-" * 50) + + success_count = 0 + failed_count = 0 + + for role_config in roles_data: + success = await self.import_role(role_config, update_existing) + if success: + success_count += 1 + else: + failed_count += 1 + print() # 空行分隔 + + print("-" * 50) + print(f"导入完成:") + print(f" 成功: {success_count}") + print(f" 失败: {failed_count}") + print(f" 总计: {len(roles_data)}") + + async def list_roles(self): + """列出数据库中的所有角色""" + session = await self.db_manager.get_session() + + try: + result = await session.execute( + select(Role.role_key, Role.name, Role.enabled) + .order_by(Role.role_key) + ) + roles = result.fetchall() + + if not roles: + print("数据库中没有角色配置") + return + + print(f"数据库中的角色配置 (共 {len(roles)} 个):") + print("-" * 60) + print(f"{'角色Key':<20} {'角色名称':<25} {'状态':<10}") + print("-" * 60) + + for role in roles: + status = "启用" if role.enabled else "禁用" + print(f"{role.role_key:<20} {role.name:<25} {status:<10}") + + except Exception as e: + print(f"查询角色列表失败: {e}") + finally: + await session.close() + + async def delete_role(self, role_key): + """删除指定角色""" + session = await self.db_manager.get_session() + + try: + result = await session.execute( + select(Role).where(Role.role_key == role_key) + ) + role = result.scalar_one_or_none() + + if not role: + print(f"角色 {role_key} 不存在") + return False + + await session.delete(role) + await session.commit() + + print(f"✓ 角色 {role_key} 删除成功") + return True + + except Exception as e: + await session.rollback() + print(f"✗ 删除角色 {role_key} 失败: {e}") + return False + finally: + await session.close() + +async def main(): + """主函数""" + import argparse + + parser = argparse.ArgumentParser(description="角色配置导入工具") + parser.add_argument('--update', action='store_true', help='更新已存在的角色') + parser.add_argument('--list', action='store_true', help='列出数据库中的角色') + parser.add_argument('--delete', type=str, help='删除指定的角色') + parser.add_argument('--role', type=str, help='只导入指定的角色文件') + + args = parser.parse_args() + + importer = RoleImporter() + + try: + await importer.initialize() + + if args.list: + await importer.list_roles() + + elif args.delete: + await importer.delete_role(args.delete) + + elif args.role: + role_file = importer.roles_dir / f"{args.role}.yaml" + if not role_file.exists(): + role_file = importer.roles_dir / f"{args.role}.yml" + + if not role_file.exists(): + print(f"角色配置文件不存在: {args.role}") + return + + try: + with open(role_file, 'r', encoding='utf-8') as f: + role_config = yaml.safe_load(f) + role_config['role_key'] = args.role + role_config['source_file'] = str(role_file) + + await importer.import_role(role_config, args.update) + except Exception as e: + print(f"导入角色 {args.role} 失败: {e}") + + else: + await importer.import_all_roles(args.update) + + except Exception as e: + print(f"操作失败: {e}") + finally: + if importer.db_manager: + await importer.db_manager.close() + +if __name__ == "__main__": + asyncio.run(main()) diff --git a/talkingq-url/config.py b/talkingq-url/config.py new file mode 100644 index 0000000..940c206 --- /dev/null +++ b/talkingq-url/config.py @@ -0,0 +1,45 @@ +from pydantic_settings import BaseSettings +from pydantic import ConfigDict + +class Settings(BaseSettings): + server_host: str + server_port: int = 8080 + asr_provider: str = "Aliyun" # 固定使用阿里云 + llm_provider: str = "Volcano" # 固定使用火山引擎 + tts_provider: str = "MiniMax" # 固定使用MiniMax + + volcano_api_key: str + volcano_base_url: str + volcano_model_id: str = "ep-20250226121739-jkd24" # Doubao-1.5-pro-32k + volcano_app_id: str = "7872932045" + volcano_access_token: str + + minimax_api_key: str = "" + minimax_group_id: str = "" + minimax_base_url: str = "https://api.minimax.chat/v1/t2a_v2" + + aliyun_api_key: str + aliyun_vocabulary_id: str + + assets_dir: str = "assets" + session_timeout: int = 600 + conversation_history_timeout: int = 1800 + max_conversation_history: int = 5 + cleanup_interval: int = 300 + llm_first_token_timeout: int = 5 # LLM首个token的超时时间(秒) + tts_request_timeout: int = 5 # TTS单次请求超时时间(秒) + selected_role_key: str + + db_host: str = "mysql" + db_port: int = 3306 + db_user: str = "talkingq" + db_password: str + db_name: str = "talkingq" + db_echo: bool = False # 是否打印SQL语句 + + admin_api_key: str # 用于设备注册的管理员API密钥 + client_api_key: str # 用于微信小程序客户端验证的API密钥 + + model_config = ConfigDict(extra="ignore", env_file=".env") + +settings = Settings() diff --git a/talkingq-url/database/__init__.py b/talkingq-url/database/__init__.py new file mode 100644 index 0000000..e69de29 diff --git a/talkingq-url/database/connection.py b/talkingq-url/database/connection.py new file mode 100644 index 0000000..02cbc6f --- /dev/null +++ b/talkingq-url/database/connection.py @@ -0,0 +1,71 @@ +from sqlalchemy.ext.asyncio import create_async_engine, AsyncSession +from sqlalchemy.orm import sessionmaker +from config import settings +from utils.logger import session_logger +import asyncio +import urllib.parse + +class DatabaseManager: + _instance = None + _lock = asyncio.Lock() + + def __init__(self): + self.engine = None + self.async_session = None + self._initialized = False + + @classmethod + async def get_instance(cls): + async with cls._lock: + if cls._instance is None: + cls._instance = DatabaseManager() + + if not cls._instance._initialized: + await cls._instance.initialize() + + return cls._instance + + async def initialize(self): + if self._initialized: + return + + try: + encoded_password = urllib.parse.quote_plus(settings.db_password) + connection_string = f"mysql+aiomysql://{settings.db_user}:{encoded_password}@{settings.db_host}:{settings.db_port}/{settings.db_name}?charset=utf8mb4" + self.engine = create_async_engine( + connection_string, + echo=settings.db_echo, + pool_pre_ping=True, + pool_recycle=3600, + ) + + self.async_session = sessionmaker( + self.engine, expire_on_commit=False, class_=AsyncSession + ) + + self._initialized = True + session_logger.info("system", "database", f"数据库连接初始化成功: {settings.db_host}:{settings.db_port}/{settings.db_name}") + except Exception as e: + session_logger.error("system", "database", f"数据库连接初始化失败: {str(e)}") + raise + + async def get_session(self): + """获取数据库会话""" + if not self._initialized: + await self.initialize() + return self.async_session() + + async def close(self): + """关闭数据库连接""" + if self.engine: + await self.engine.dispose() + self._initialized = False + session_logger.info("system", "database", "数据库连接已关闭") + +db_manager = None + +async def get_db_manager(): + global db_manager + if db_manager is None: + db_manager = await DatabaseManager.get_instance() + return db_manager diff --git a/talkingq-url/database/init_db.py b/talkingq-url/database/init_db.py new file mode 100644 index 0000000..4c218ee --- /dev/null +++ b/talkingq-url/database/init_db.py @@ -0,0 +1,24 @@ +import asyncio +from sqlalchemy.ext.asyncio import create_async_engine +from database.models import Base +from config import settings +from utils.logger import session_logger +import urllib.parse + +async def init_db(): + """初始化数据库，创建所有表""" + try: + encoded_password = urllib.parse.quote_plus(settings.db_password) + connection_string = f"mysql+aiomysql://{settings.db_user}:{encoded_password}@{settings.db_host}:{settings.db_port}/{settings.db_name}" + engine = create_async_engine( + connection_string, + echo=settings.db_echo + ) + async with engine.begin() as conn: + await conn.run_sync(Base.metadata.create_all) + await engine.dispose() + session_logger.info("system", "database", "数据库表已成功创建") + return True + except Exception as e: + session_logger.error("system", "database", f"创建数据库表失败: {str(e)}") + return False diff --git a/talkingq-url/database/models.py b/talkingq-url/database/models.py new file mode 100644 index 0000000..78d410c --- /dev/null +++ b/talkingq-url/database/models.py @@ -0,0 +1,128 @@ +from sqlalchemy import Column, String, Text, Float, DateTime, Integer, Boolean, func, ForeignKey, JSON +from sqlalchemy.ext.declarative import declarative_base +from sqlalchemy.schema import UniqueConstraint +import datetime + +Base = declarative_base() + +class DeviceConfig(Base): + __tablename__ = "device_configs" + + id = Column(Integer, primary_key=True, autoincrement=True) + device_id = Column(String(64), unique=True, index=True, nullable=False) + selected_role_key = Column(String(64), nullable=False) + preferred_language = Column(String(10), nullable=True) + volume = Column(Integer, nullable=True) + last_update_time = Column(Float, nullable=False, default=lambda: datetime.datetime.now().timestamp()) + created_at = Column(DateTime, nullable=False, server_default=func.now()) + updated_at = Column(DateTime, nullable=False, server_default=func.now(), onupdate=func.now()) + + def to_dict(self): + """将配置转换为可序列化的字典""" + return { + "selected_role_key": self.selected_role_key, + "preferred_language": self.preferred_language, + "volume": self.volume if hasattr(self, 'volume') else None, + } + +class ConversationHistory(Base): + __tablename__ = "conversation_histories" + id = Column(Integer, primary_key=True, autoincrement=True) + device_id = Column(String(64), nullable=False, index=True) + role_key = Column(String(64), nullable=False, index=True) + last_interaction_time = Column(Float, nullable=False, default=lambda: datetime.datetime.now().timestamp()) + created_at = Column(DateTime, nullable=False, server_default=func.now()) + updated_at = Column(DateTime, nullable=False, server_default=func.now(), onupdate=func.now()) + __table_args__ = ( + {'mysql_charset': 'utf8mb4', 'mysql_collate': 'utf8mb4_unicode_ci'} + ) + +class ConversationMessage(Base): + __tablename__ = "conversation_messages" + id = Column(Integer, primary_key=True, autoincrement=True) + conversation_id = Column(Integer, ForeignKey("conversation_histories.id", ondelete="CASCADE"), nullable=False, index=True) + is_user = Column(Boolean, nullable=False, default=False) # True表示用户消息，False表示助手消息 + content = Column(Text, nullable=False) + timestamp = Column(Float, nullable=False, default=lambda: datetime.datetime.now().timestamp()) + created_at = Column(DateTime, nullable=False, server_default=func.now()) + updated_at = Column(DateTime, nullable=False, server_default=func.now(), onupdate=func.now()) + __table_args__ = ( + {'mysql_charset': 'utf8mb4', 'mysql_collate': 'utf8mb4_unicode_ci'} + ) + +class Role(Base): + __tablename__ = "roles" + id = Column(Integer, primary_key=True, autoincrement=True) + role_key = Column(String(64), unique=True, nullable=False, index=True) + name = Column(String(128), nullable=False) + description = Column(Text, nullable=True) + content = Column(Text, nullable=False) + default_language = Column(String(10), nullable=True) + volcano_model_id = Column(String(64), nullable=True) + minimax_voice_id = Column(String(64), nullable=True) + url = Column(String(255), nullable=True) + homophones = Column(JSON, nullable=True) + enabled = Column(Boolean, default=True, nullable=False) + created_at = Column(DateTime, nullable=False, server_default=func.now()) + updated_at = Column(DateTime, nullable=False, server_default=func.now(), onupdate=func.now()) + + __table_args__ = ( + {'mysql_charset': 'utf8mb4', 'mysql_collate': 'utf8mb4_unicode_ci'} + ) + +class RoleLanguage(Base): + __tablename__ = "role_languages" + id = Column(Integer, primary_key=True, autoincrement=True) + role_id = Column(Integer, ForeignKey("roles.id", ondelete="CASCADE"), nullable=False) + language_code = Column(String(10), nullable=False) + name = Column(String(128), nullable=True) + content = Column(Text, nullable=True) + minimax_voice_id = Column(String(64), nullable=True) + url = Column(String(255), nullable=True) + created_at = Column(DateTime, nullable=False, server_default=func.now()) + updated_at = Column(DateTime, nullable=False, server_default=func.now(), onupdate=func.now()) + + __table_args__ = ( + UniqueConstraint('role_id', 'language_code', name='uix_role_language'), + {'mysql_charset': 'utf8mb4', 'mysql_collate': 'utf8mb4_unicode_ci'} + ) + +class DeviceAuth(Base): + __tablename__ = "device_auth" + id = Column(Integer, primary_key=True, autoincrement=True) + device_id = Column(String(64), nullable=False, unique=True, index=True) + serial_number = Column(String(64), nullable=False) + batch_id = Column(String(20), nullable=True, index=True) + is_active = Column(Boolean, default=True, nullable=False) + created_at = Column(DateTime, nullable=False, server_default=func.now()) + updated_at = Column(DateTime, nullable=False, server_default=func.now(), onupdate=func.now()) + + __table_args__ = ( + {'mysql_charset': 'utf8mb4', 'mysql_collate': 'utf8mb4_unicode_ci'} + ) + +class DeviceFirmwareUpdate(Base): + __tablename__ = "device_firmware_update" + id = Column(Integer, primary_key=True, autoincrement=True) + device_id = Column(String(64), nullable=False, unique=True, index=True) + firmware_version = Column(String(64), nullable=False) + update_status = Column(String(32), nullable=False, default="success") # 更新状态，如 updating/success/failed + created_at = Column(DateTime, nullable=False, server_default=func.now()) + updated_at = Column(DateTime, nullable=False, server_default=func.now(), onupdate=func.now()) + progress = Column(Float, nullable=True, default=0.0) # 更新进度 + + __table_args__ = ( + {'mysql_charset': 'utf8mb4', 'mysql_collate': 'utf8mb4_unicode_ci'} + ) + +class SystemConfig(Base): + __tablename__ = "system_config" + id = Column(Integer, primary_key=True, autoincrement=True) + config_key = Column(String(128), unique=True, nullable=False, index=True) + config_value = Column(Text, nullable=True) + created_at = Column(DateTime, nullable=False, server_default=func.now()) + updated_at = Column(DateTime, nullable=False, server_default=func.now(), onupdate=func.now()) + + __table_args__ = ( + {'mysql_charset': 'utf8mb4', 'mysql_collate': 'utf8mb4_unicode_ci'} + ) diff --git a/talkingq-url/docker-compose-bak.yml b/talkingq-url/docker-compose-bak.yml new file mode 100644 index 0000000..5fe0d47 --- /dev/null +++ b/talkingq-url/docker-compose-bak.yml @@ -0,0 +1,44 @@ +services: + app: + build: . + ports: + - "8080:8080" + env_file: + - .env + volumes: + - ./assets:/app/assets + depends_on: + mysql: + condition: service_healthy + restart: unless-stopped + networks: + - talkingq-network + + mysql: + image: mysql:8.0 + command: --default-authentication-plugin=caching_sha2_password + restart: unless-stopped + environment: + MYSQL_ROOT_PASSWORD: "${MYSQL_ROOT_PASSWORD}" + MYSQL_DATABASE: talkingq + MYSQL_USER: talkingq + MYSQL_PASSWORD: "${MYSQL_PASSWORD}" + TZ: Asia/Shanghai + volumes: + - mysql-data:/var/lib/mysql + - ./mysql/init:/docker-entrypoint-initdb.d + - ./mysql/conf.d:/etc/mysql/conf.d + healthcheck: + test: ["CMD", "mysqladmin", "ping", "-h", "localhost", "-u", "talkingq", "-p${MYSQL_PASSWORD}"] + interval: 10s + timeout: 5s + retries: 5 + networks: + - talkingq-network + +volumes: + mysql-data: + +networks: + talkingq-network: + driver: bridge diff --git a/talkingq-url/docker-compose.yml b/talkingq-url/docker-compose.yml new file mode 100644 index 0000000..a21b105 --- /dev/null +++ b/talkingq-url/docker-compose.yml @@ -0,0 +1,61 @@ +services: + app: + build: . + ports: + - "8080:8080" + env_file: + - .env + volumes: + - ./assets:/app/assets + - ./logs:/app/logs + depends_on: + mysql: + condition: service_healthy + restart: unless-stopped + networks: + - talkingq-network + + backend_app: + build: ../backend-v0.1 + ports: + - "8001:8001" + env_file: + - ../backend-v0.1/.env + volumes: + - ../backend-v0.1/app/assets:/app/assets + - ../backend-v0.1/app/logs:/app/logs + depends_on: + mysql: + condition: service_healthy + restart: unless-stopped + networks: + - talkingq-network + + mysql: + image: mysql:8.0 + command: --default-authentication-plugin=caching_sha2_password + restart: unless-stopped + environment: + MYSQL_ROOT_PASSWORD: "${MYSQL_ROOT_PASSWORD}" + MYSQL_DATABASE: talkingq + MYSQL_USER: talkingq + MYSQL_PASSWORD: "${MYSQL_PASSWORD}" + TZ: Asia/Shanghai + volumes: + - mysql-data:/var/lib/mysql + - ./mysql/init:/docker-entrypoint-initdb.d + - ./mysql/conf.d:/etc/mysql/conf.d + healthcheck: + test: ["CMD", "mysqladmin", "ping", "-h", "localhost", "-u", "talkingq", "-p${MYSQL_PASSWORD}"] + interval: 10s + timeout: 5s + retries: 5 + networks: + - talkingq-network + +volumes: + mysql-data: + +networks: + talkingq-network: + driver: bridge diff --git a/talkingq-url/docs/ota_flow.md b/talkingq-url/docs/ota_flow.md new file mode 100644 index 0000000..6690641 --- /dev/null +++ b/talkingq-url/docs/ota_flow.md @@ -0,0 +1,229 @@ +# TalkingQ OTA 固件升级流程 + +本文档详细描述了TalkingQ设备固件OTA升级的完整流程，涉及到APP、后台服务器和ESP32设备端三方的交互。 + +## 流程概述 + +```mermaid +sequenceDiagram + participant APP + participant 后台服务器 + participant ESP32设备端 + + APP->>后台服务器: /api/ota/check/{device_id}: 检查更新 + 后台服务器->>ESP32设备端: WebSocket: GET_FIRMWARE_VERSION + ESP32设备端->>后台服务器: WebSocket: FIRMWARE_VERSION:{version} + 后台服务器->>后台服务器: 比较版本判断是否需要更新 + 后台服务器->>APP: 返回检查结果与版本信息 + + alt 不需要更新 + APP->>用户: 提示当前已是最新版本 + else 需要更新 + APP->>后台服务器: /api/ota/start/{device_id}: 发起升级请求 + 后台服务器->>ESP32设备端: WebSocket: UPDATE_FIRMWARE:{url} + ESP32设备端->>ESP32设备端: 下载固件并校验 + ESP32设备端->>后台服务器: WebSocket: FIRMWARE_UPDATE_STATUS + + loop 进度更新 + ESP32设备端->>后台服务器: FIRMWARE_UPDATE_STATUS:status=updating,progress=45.5,version=1.0.0 + 后台服务器->>后台服务器: 数据库更新升级状态与进度 + APP->>后台服务器: /api/ota/status/{device_id}: 查询升级状态 + 后台服务器->>APP: 返回当前升级状态、进度和设备在线状态 + APP->>用户: 显示升级进度 + end + + ESP32设备端->>ESP32设备端: 升级完成后重启 + ESP32设备端->>后台服务器: WebSocket重连后上报: FIRMWARE_VERSION:{new_version} + 后台服务器->>后台服务器: 更新设备固件版本记录 + APP->>后台服务器: /api/ota/status/{device_id}: 确认升级完成 + APP->>用户: 提示升级完成 + end +``` + +## 详细API说明 + +### 1. 检查更新 + +**APP → 后台服务器** + +``` +GET /api/ota/check/{device_id} +Headers: + X-Device-ID: {device_id} + X-Device-Serial: {serial_number} +``` + +**后台服务器 → APP** + +```json +{ + "code": 0, + "msg": "success", + "data": { + "status": "success", + "need_update": true|false, + "current_version": "x.y.z", + "latest_version": "a.b.c" + } +} +``` + +如果出错则返回： + +```json +{ + "code": -1, + "msg": "错误信息", + "data": {} +} +``` + +### 2. 启动固件升级 + +**APP → 后台服务器** + +``` +POST /api/ota/start/{device_id} +Headers: + X-Device-ID: {device_id} + X-Device-Serial: {serial_number} +``` + +**后台服务器 → APP** + +```json +{ + "code": 0, + "msg": "success", + "data": { + "updating": true + } +} +``` + +### 3. 获取升级状态和进度 + +**APP → 后台服务器** + +``` +GET /api/ota/status/{device_id} +Headers: + X-Device-ID: {device_id} + X-Device-Serial: {serial_number} +``` + +**后台服务器 → APP** + +```json +{ + "code": 0, + "msg": "success", + "data": { + "status": "updating|success|failed|unknown", + "progress": 45.5, + "version": "1.0.0", + "device_online": true|false + } +} +``` + +### 4. 管理固件配置信息（管理员接口） + +**获取固件配置** + +``` +GET /api/ota/config/firmware +Headers: + X-Device-ID: {device_id} + X-Device-Serial: {serial_number} +``` + +**设置固件配置** + +``` +POST /api/ota/config/firmware +Headers: + X-Device-ID: {device_id} + X-Device-Serial: {serial_number} +Body: +{ + "version": "1.2.3", + "url": "https://example.com/firmware/v1.2.3.bin" +} +``` + +## WebSocket消息格式 + +### 1. 服务器向设备请求固件版本 + +``` +GET_FIRMWARE_VERSION +``` + +### 2. 设备返回固件版本 + +``` +FIRMWARE_VERSION:1.0.0 +``` + +### 3. 服务器向设备发送升级指令 + +``` +UPDATE_FIRMWARE:https://example.com/firmware/v1.2.3.bin +``` + +### 4. 设备上报升级状态 + +``` +FIRMWARE_UPDATE_STATUS:status=updating,progress=45.5,version=1.0.0 +``` + +其中status可以是： +- `updating`: 升级进行中 +- `success`: 升级成功完成 +- `failed`: 升级失败 +- `completed`: 升级完成（等效于success） + +## 实现细节 + +### 固件管理机制 + +1. **自动扫描固件目录**：系统在启动时会自动扫描 `assets/firmware` 目录，识别所有格式为 `{version}.bin` 的固件文件。 + +2. **版本识别机制**：固件文件命名必须符合 `x.y.z.bin` 格式，其中 x.y.z 为版本号（例如：`1.2.3.bin`）。 + +3. **自动更新系统配置**：系统会自动识别最新版本的固件，并更新系统配置中的 `latest_firmware_version` 和 `update_firmware_url`。 + +4. **简化部署流程**：管理员只需将新固件上传到 `assets/firmware` 目录，系统会自动完成后续配置。 + +### 后台服务实现特点 + +1. **版本比较机制**：服务器代码使用特定的版本比较逻辑。当设备版本为`unknown`或`0.0.0`时，会被视为需要更新。 + +2. **数据库存储**： + - 使用`DeviceFirmwareUpdate`表记录设备固件版本和更新状态 + - 使用`SystemConfig`表存储最新固件版本(`latest_firmware_version`)和下载URL(`update_firmware_url`) + +3. **缓存机制**：设备固件信息使用内存缓存减少数据库查询，缓存有过期时间控制 + +4. **进度控制**：升级进度以0-100的浮点数表示，由设备上报，服务器保存 + +5. **设备在线状态**：status接口会返回设备是否在线，通过检查WebSocket连接状态判断 + +### 注意事项 + +1. **设备认证**：所有API请求需要设备ID和序列号认证 + +2. **重启后自动上报**：设备重启后，应在WebSocket连接建立后主动上报固件版本 + +3. **错误处理**： + - 设备不在线时，服务器会返回相应的错误信息 + - 版本获取失败时，服务器会尝试等待一段时间后再判断升级状态 + +4. **进度报告**： + - 设备应该定期上报升级进度，特别是在状态发生变化时 + - 每个百分比变化或每5%的进度变化应上报一次 + +5. **安全性**： + - 固件URL应该是安全的HTTPS链接 + - 设备应验证固件的完整性和来源 diff --git a/talkingq-url/handlers/__init__.py b/talkingq-url/handlers/__init__.py new file mode 100644 index 0000000..e69de29 diff --git a/talkingq-url/handlers/audio_packet_parser.py b/talkingq-url/handlers/audio_packet_parser.py new file mode 100644 index 0000000..1a8698b --- /dev/null +++ b/talkingq-url/handlers/audio_packet_parser.py @@ -0,0 +1,87 @@ +import struct +from fastapi import WebSocket +from services.audio_session import audio_session_manager, AudioSession +from utils.logger import session_logger +from services.factory import get_asr, get_llm, get_tts + +DEVICE_ID_SIZE = 22 # 将大小设为22，在实际解析中会更灵活处理 +SESSION_ID_SIZE = 33 +SAMPLE_RATE = 16000 # 固定为16K采样率 +CHANNELS = 1 # 单声道 +SAMPLE_WIDTH = 2 # 16位 = 2字节 + +async def parse_packet(data: bytes, websocket: WebSocket): + try: + min_header_size = 4 + 1 + 4 # sequence_number + packet_type + data_size + if len(data) < min_header_size: + session_logger.warning( + "unknown", "unknown", f"数据包太小: {len(data)} < {min_header_size}" + ) + return None, None, None, None, None + + null_pos = data.find(b'\x00') + if null_pos == -1 or null_pos > 25: # 如果没找到或超出最大长度限制 + null_pos = 22 # 使用默认值 + + device_id = data[:null_pos].decode("ascii") + + session_start = null_pos + 1 + session_end = session_start + SESSION_ID_SIZE + session_data = data[session_start:session_end] + session_null_pos = session_data.find(b'\x00') + + if session_null_pos != -1: + session_id = session_data[:session_null_pos].decode("ascii") + else: + session_id = session_data.decode("ascii").rstrip("\x00") + + + offset = session_start + len(session_id) + 1 + if offset + min_header_size > len(data): + session_logger.warning( + "unknown", "unknown", f"解析剩余头信息时数据包太小: {len(data)} < {offset + min_header_size}" + ) + return None, None, None, None, None + + remaining_data = data[offset:] + offset = 0 + sequence_number = struct.unpack(" dict: + device_config = await device_config_manager.get_config(device_id, force_refresh=True) + preferred_language = None + if device_config and hasattr(device_config, 'preferred_language') and device_config.preferred_language: + preferred_language = device_config.preferred_language + session_logger.info( + device_id, + "config", + f"检测到设备首选语言配置: {preferred_language}" + ) + effective_language = language or preferred_language + if device_config: + selected_role_key = device_config.selected_role_key + session_logger.info( + device_id, + "config", + f"使用配置的角色: {selected_role_key}" + + (f", 使用语言: {effective_language}" if effective_language else "") + ) + else: + selected_role_key = settings.selected_role_key + session_logger.info( + device_id, + "config", + f"使用默认角色: {selected_role_key}" + + (f", 使用语言: {effective_language}" if effective_language else "") + ) + selected_role = await role_manager.get_role_config_for_language(selected_role_key, effective_language) + if not selected_role: + all_roles = await role_manager.get_all_roles() + if all_roles: + first_role_key = next(iter(all_roles.keys())) + selected_role = await role_manager.get_role_config_for_language(first_role_key, effective_language) + session_logger.warning( + device_id, + "config", + f"未找到角色 {selected_role_key}，使用默认角色: {first_role_key}" + ) + else: + selected_role = { + "role_key": "default", + "name": "默认助手", + "content": "你是一个智能助手，请简洁回答问题。" + } + session_logger.warning( + device_id, + "config", + "未找到任何可用的角色配置，使用内置默认角色" + ) + lang_info = effective_language or "默认" + session_logger.info( + device_id, + "config", + f"已选择角色配置: {selected_role.get('name', '未命名')}, 语言: {lang_info}" + ) + return selected_role diff --git a/talkingq-url/handlers/prompt_sound_handler.py b/talkingq-url/handlers/prompt_sound_handler.py new file mode 100644 index 0000000..3bccc37 --- /dev/null +++ b/talkingq-url/handlers/prompt_sound_handler.py @@ -0,0 +1,135 @@ +import os +from services.connection_manager import connection_manager +from services.device_config import device_config_manager +from services.role_manager import role_manager +from config import settings +from utils.logger import session_logger + + +async def handle_prompt_sound_request( + device_id: str, prompt_type: str, language: str = None +): + websocket = await connection_manager.get_connection(device_id) + if websocket and websocket.client_state.name == "CONNECTED": + try: + sound_files = { + "welcome": "welcome.mp3", + "error": "error.mp3", + "goodbye": "goodbye.mp3", + "interrupt": "interrupt.mp3", + "tts_error": "tts_error.mp3", + } + file_name = sound_files.get(prompt_type) + if not file_name: + session_logger.error( + device_id, "sound", f"未找到提示音类型: {prompt_type}" + ) + return + + device_config = await device_config_manager.get_config(device_id, force_refresh=True) + effective_language = language + if ( + not effective_language + and device_config + and hasattr(device_config, "preferred_language") + and device_config.preferred_language + ): + effective_language = device_config.preferred_language + session_logger.info( + device_id, "sound", f"使用设备首选语言选择提示音: {effective_language}" + ) + + selected_role_key = ( + device_config.selected_role_key + if device_config + else settings.selected_role_key + ) + session_logger.info( + device_id, "sound", f"从设备配置获取角色：{selected_role_key}" + ) + + selected_role = await role_manager.get_role_config_for_language( + selected_role_key, effective_language + ) + + if not selected_role: + session_logger.error( + device_id, "sound", f"未找到角色 {selected_role_key} 的配置" + ) + return + + base_url = None + if "multilingual" in selected_role: + current_lang = effective_language or selected_role.get( + "default_language", "zh" + ) + if current_lang in selected_role["multilingual"]: + lang_config = selected_role["multilingual"][current_lang] + if "url" in lang_config: + base_url = lang_config["url"] + session_logger.info( + device_id, "sound", f"使用多语言({current_lang})配置的URL: {base_url}" + ) + + if not base_url and "url" in selected_role: + base_url = selected_role["url"] + session_logger.info( + device_id, "sound", f"使用角色基础URL: {base_url}" + ) + + if not base_url: + session_logger.error( + device_id, "sound", f"角色 {selected_role_key} 未配置URL" + ) + return + + sound_file_path = os.path.join(settings.assets_dir, base_url, file_name) + if not os.path.exists(sound_file_path): + session_logger.error( + device_id, "sound", f"提示音文件不存在: {sound_file_path}" + ) + return + + prompt_sound_url = f"http://{settings.server_host}:{settings.server_port}/assets/{base_url}/{file_name}" + try: + await websocket.send_text(f"PROMPT_SOUND_URL:{prompt_sound_url}") + session_logger.info( + device_id, "sound", f"已发送提示音URL给客户端: {prompt_sound_url}" + ) + except Exception as e: + session_logger.error( + device_id, "sound", f"发送提示音URL失败: {e}" + ) + except Exception as e: + session_logger.error( + device_id, "sound", f"准备提示音时出错: {e}", exc_info=True + ) + else: + session_logger.error( + device_id, "sound", "未找到设备的 WebSocket 连接或连接已关闭" + ) + + +async def send_welcome_sound( + device_id: str, role_key: str = None, language: str = None +): + device_config = await device_config_manager.get_config(device_id, force_refresh=True) + + if role_key is None and device_config: + role_key = device_config.selected_role_key + session_logger.info( + device_id, "sound", f"使用设备配置的角色发送欢迎音效: {role_key}" + ) + + if ( + language is None + and device_config + and hasattr(device_config, "preferred_language") + and device_config.preferred_language + ): + language = device_config.preferred_language + session_logger.info( + device_id, "sound", f"使用设备首选语言发送欢迎音效: {language}" + ) + + await handle_prompt_sound_request(device_id, "welcome", language) diff --git a/talkingq-url/handlers/response_coordinator.py b/talkingq-url/handlers/response_coordinator.py new file mode 100644 index 0000000..55e3dfe --- /dev/null +++ b/talkingq-url/handlers/response_coordinator.py @@ -0,0 +1,184 @@ +import asyncio +from typing import Dict, List +from config import settings +from utils.logger import session_logger +from services.audio_session import AudioSession +from services.conversation_history import conversation_history_manager +from services.interrupt_handler import interrupt_handler +from services.task_manager import task_manager +from services.text_generator import TextGenerator +from services.tts_synthesizer import TTSSynthesizer +from services.audio_sender import AudioSender +from services.interruption_helper import InterruptionHelper + +import re + + +async def _cleanup_queues(text_queue: asyncio.Queue, url_queue: asyncio.Queue, device_id: str, session_id: str): + """清理响应处理中使用的队列""" + try: + # 清理text_queue + text_count = 0 + while not text_queue.empty(): + try: + item = text_queue.get_nowait() + text_queue.task_done() + text_count += 1 + del item + except asyncio.QueueEmpty: + break + + # 清理url_queue + url_count = 0 + while not url_queue.empty(): + try: + item = url_queue.get_nowait() + url_queue.task_done() + url_count += 1 + del item + except asyncio.QueueEmpty: + break + + session_logger.info( + device_id, session_id, + f"队列清理完成 - text_queue: {text_count}项, url_queue: {url_count}项" + ) + + except Exception as e: + session_logger.error( + device_id, session_id, f"队列清理时出错: {e}" + ) + + +SENSITIVE_RESPONSES = [ + "你好，我无法给到相关内容", + "我无法给到相关内容", + "Sorry, I didn't catch that", + "I didn't catch that", + "让我们换一个话题", + "Let's change the topic", + "Changeons de sujet", + "Lassen Sie uns das Thema wechseln", + "Cambiemos de tema", + "Mari kita tukar topik" +] + +def is_sensitive_response(reply: str) -> bool: + """ + 检查回复是否为敏感话题回复，考虑不同语言的标点符号差异 + """ + def normalize_text(text): + text = text.lower() + punctuations = r'[!"#$%&\'()*+,-./:;<=>?@\[\\\]^_`{|}~""''、。，！？：；（）【】「」『』〈〉《》〔〕…—￥€£¥]' + text = re.sub(punctuations, '', text) + text = re.sub(r'\s+', '', text) + full_to_half = str.maketrans({ + '０': '0', '１': '1', '２': '2', '３': '3', '４': '4', + '５': '5', '６': '6', '７': '7', '８': '8', '９': '9', + 'Ａ': 'a', 'Ｂ': 'b', 'Ｃ': 'c', 'Ｄ': 'd', 'Ｅ': 'e', + 'Ｆ': 'f', 'Ｇ': 'g', 'Ｈ': 'h', 'Ｉ': 'i', 'Ｊ': 'j', + 'Ｋ': 'k', 'Ｌ': 'l', 'Ｍ': 'm', 'Ｎ': 'n', 'Ｏ': 'o', + 'Ｐ': 'p', 'Ｑ': 'q', 'Ｒ': 'r', 'Ｓ': 's', 'Ｔ': 't', + 'Ｕ': 'u', 'Ｖ': 'v', 'Ｗ': 'w', 'Ｘ': 'x', 'Ｙ': 'y', + 'Ｚ': 'z', 'ａ': 'a', 'ｂ': 'b', 'ｃ': 'c', 'ｄ': 'd', + 'ｅ': 'e', 'ｆ': 'f', 'ｇ': 'g', 'ｈ': 'h', 'ｉ': 'i', + 'ｊ': 'j', 'ｋ': 'k', 'ｌ': 'l', 'ｍ': 'm', 'ｎ': 'n', + 'ｏ': 'o', 'ｐ': 'p', 'ｑ': 'q', 'ｒ': 'r', 'ｓ': 's', + 'ｔ': 't', 'ｕ': 'u', 'ｖ': 'v', 'ｗ': 'w', 'ｘ': 'x', + 'ｙ': 'y', 'ｚ': 'z', '　': '' + }) + text = text.translate(full_to_half) + return text + + normalized_reply = normalize_text(reply) + + for sensitive_reply in SENSITIVE_RESPONSES: + normalized_sensitive = normalize_text(sensitive_reply) + if normalized_sensitive in normalized_reply: + return True + return False + +async def process_response( + transcript: str, + history: List[Dict[str, str]], + selected_role: dict, + device_id: str, + session_id: str, + device_history, + session: AudioSession, + language: str = None, +): + + text_queue = asyncio.Queue() # 从LLM到TTS传递文本 + url_queue = asyncio.Queue() # 从TTS到发送任务传递音频URL + session_key = (device_id, session_id) + await InterruptionHelper.register_queue_cleanup(session_key, url_queue) + await InterruptionHelper.register_queue_cleanup(session_key, text_queue) + text_generator = TextGenerator(device_id, session_id) + tts_synthesizer = TTSSynthesizer(device_id, session_id) + audio_sender = AudioSender(device_id, session_id, session.websocket) + llm_service = session.llm_service + tts_service = session.tts_service + + if isinstance(tts_service.__class__.__name__, str) and tts_service.__class__.__name__ == "AliyunTTS": + tts_service = await get_tts(selected_role, language) + session.tts_service = tts_service + session_logger.warning( + device_id, session_id, "检测到AliyunTTS已不再支持，自动切换到MiniMaxTTS" + ) + + text_gen_task = await task_manager.create_task( + text_generator.generate_text( + llm_service, transcript, history, selected_role, text_queue + ), + device_id=device_id, + session_key=(device_id, session_id), + task_type="text_generation" + ) + tts_task = await task_manager.create_task( + tts_synthesizer.synthesize_audio( + tts_service, text_queue, url_queue, selected_role, language # 传递语言参数到TTS + ), + device_id=device_id, + session_key=(device_id, session_id), + task_type="tts_synthesis" + ) + sender_task = await task_manager.create_task( + audio_sender.send_audio_urls(url_queue), + device_id=device_id, + session_key=(device_id, session_id), + task_type="audio_sender" + ) + reply = "" + try: + reply = await text_gen_task + await asyncio.gather(tts_task, sender_task) + except Exception as e: + session_logger.error( + device_id, session_id, f"响应处理过程中出错: {e}", exc_info=True + ) + finally: + # 清理队列 + await _cleanup_queues(text_queue, url_queue, device_id, session_id) + if not interrupt_handler.is_interrupted(session_key): + if is_sensitive_response(reply): + session_logger.info( + device_id, session_id, "检测到敏感话题回复，该轮对话将不被添加到历史记录" + ) + else: + device_history.history.append({"user": transcript, "assistant": reply}) + + if len(device_history.history) > settings.max_conversation_history: + device_history.history = device_history.history[-settings.max_conversation_history:] + + role_key = selected_role.get("role_key", settings.selected_role_key) + device_history.role_key = role_key + await conversation_history_manager.set_history(device_id, device_history, role_key) + + session_logger.info( + device_id, + session_id, + f"已将本轮对话添加到角色 {role_key} 的历史记录并保存到数据库" + ) + + return reply diff --git a/talkingq-url/handlers/response_processor.py b/talkingq-url/handlers/response_processor.py new file mode 100644 index 0000000..746f795 --- /dev/null +++ b/talkingq-url/handlers/response_processor.py @@ -0,0 +1,64 @@ +import asyncio +from typing import List, Dict +from services.audio_session import AudioSession +from handlers.prompt_sound_handler import handle_prompt_sound_request +from handlers.response_coordinator import process_response +from interfaces.llm import LLM +from interfaces.tts import TTS +from utils.logger import session_logger +from services.tts_error_manager import tts_error_manager +from implementations.minimax_tts import MiniMaxTTS + +async def generate_and_process_response( + transcript: str, + history: List[Dict[str, str]], + selected_role: dict, + device_id: str, + session_id: str, + device_history, + session: AudioSession, + language: str = None, # 默认为None +): + llm_service: LLM = session.llm_service + if hasattr(llm_service, "device_id"): + llm_service.device_id = device_id + if hasattr(llm_service, "session_id"): + llm_service.session_id = session_id + + tts_service: TTS = session.tts_service + if not isinstance(tts_service, MiniMaxTTS): + session_logger.warning( + device_id, session_id, "检测到TTS服务不是MiniMaxTTS，正在切换到MiniMaxTTS" + ) + tts_service = MiniMaxTTS(selected_role=selected_role) + session.tts_service = tts_service + + try: + if isinstance(language, (int, float)) or (isinstance(language, str) and not language.isalpha()): + session_logger.warning( + device_id, session_id, f"检测到无效的语言代码: {language}，将使用默认语言" + ) + language = "zh" # 使用默认语言 + + await process_response( + transcript, + history, + selected_role, + device_id, + session_id, + device_history, + session, + language, # 传递合法化后的语言参数 + ) + except Exception as e: + session_logger.error( + device_id, session_id, f"生成或处理回复时出错: {e}", exc_info=True + ) + if session.websocket and session.websocket.client_state.name == "CONNECTED": + try: + await session.websocket.send_text("TTS_ERROR") + except Exception as ws_error: + session_logger.error(device_id, session_id, f"发送TTS错误通知失败: {ws_error}") + await handle_prompt_sound_request(device_id, "tts_error") + finally: + await tts_error_manager.end_tts_session(device_id, session_id) diff --git a/talkingq-url/handlers/service_connection_handler.py b/talkingq-url/handlers/service_connection_handler.py new file mode 100644 index 0000000..409dc41 --- /dev/null +++ b/talkingq-url/handlers/service_connection_handler.py @@ -0,0 +1,185 @@ +import asyncio +from services.audio_session import AudioSession +from services.task_manager import task_manager +from utils.logger import session_logger +from services.interrupt_handler import interrupt_handler +import time + + +async def prepare_asr_service(session, device_id, session_id): + if hasattr(session.asr_service, "device_id"): + session.asr_service.device_id = device_id + if hasattr(session.asr_service, "session_id"): + session.asr_service.session_id = session_id + await establish_asr_connection(session) + + +async def prepare_llm_service(session, device_id, session_id, selected_role): + if hasattr(session.llm_service, "device_id"): + session.llm_service.device_id = device_id + if hasattr(session.llm_service, "session_id"): + session.llm_service.session_id = session_id + if hasattr(session.llm_service, "connect"): + await session.llm_service.connect() + + from services.conversation_history import conversation_history_manager + from config import settings + device_history = await conversation_history_manager.get_history(device_id) + if device_history: + session.llm_service.history = device_history.history[ + -settings.max_conversation_history : + ] + + +async def prepare_tts_service(session, device_id, session_id, selected_role): + session.tts_service.selected_role = selected_role + if hasattr(session.tts_service, "device_id"): + session.tts_service.device_id = device_id + if hasattr(session.tts_service, "session_id"): + session.tts_service.session_id = session_id + await establish_tts_connection(session) + + +async def establish_asr_connection(session: AudioSession): + + try: + if session.device_id and session.session_id: + session_key = (session.device_id, session.session_id) + if interrupt_handler.is_interrupted(session_key): + session_logger.info( + session.device_id, session.session_id, "会话已中断，跳过建立ASR连接" + ) + return + session.send_task = await task_manager.create_task( + send_audio_task(session), + device_id=session.device_id, + session_key=(session.device_id, session.session_id) if session.device_id and session.session_id else None, + task_type="audio_send" + ) + await session.asr_service.connect() + if not session.asr_service.connected: + session_logger.error("unknown", "unknown", "语音识别 WebSocket 连接未建立") + return + session_logger.info( + "unknown", "unknown", f"语音识别连接已建立，采样率: {session.sample_rate}Hz" + ) + except Exception as e: + session_logger.error( + "unknown", "unknown", f"建立语音识别 WebSocket 连接失败: {e}", exc_info=True + ) + + +async def establish_tts_connection(session: AudioSession): + + try: + if session.device_id and session.session_id: + session_key = (session.device_id, session.session_id) + if interrupt_handler.is_interrupted(session_key): + session_logger.info( + session.device_id, session.session_id, "会话已中断，跳过建立TTS连接" + ) + return + + await session.tts_service.connect() + + if not session.tts_service.connected: + session_logger.error( + session.device_id or "unknown", + session.session_id or "unknown", + "语音合成连接未建立", + ) + return + + session_logger.info( + session.device_id or "unknown", + session.session_id or "unknown", + "语音合成连接已建立", + ) + except Exception as e: + session_logger.error( + session.device_id or "unknown", + session.session_id or "unknown", + f"建立语音合成连接失败: {e}", + ) + + +async def send_audio_task(session: AudioSession): + try: + while True: + if session.device_id and session.session_id: + session_key = (session.device_id, session.session_id) + if interrupt_handler.is_interrupted(session_key): + session_logger.info( + session.device_id, session.session_id, "发送音频任务被中断" + ) + break + audio_data = await session.send_queue.get() + if audio_data is None: # 结束信号 + break + + await session.asr_service.send_audio(audio_data) + + except asyncio.CancelledError: + session_logger.info( + session.device_id or "unknown", + session.session_id or "unknown", + "发送音频数据任务被取消", + ) + except Exception as e: + session_logger.error( + session.device_id or "unknown", + session.session_id or "unknown", + f"发送音频数据时出错: {str(e)}", + exc_info=True, + ) + + +async def end_asr_session(session: AudioSession): + + if session.asr_service.connected: + try: + asr_end_start_time = time.perf_counter() + await session.asr_service.send_end() + await session.asr_service.receive_results() + session.transcript = session.asr_service.transcript + asr_end_elapsed = time.perf_counter() - asr_end_start_time + session_logger.info( + session.device_id or "unknown", + session.session_id or "unknown", + f"ASR结束延迟: 从发送结束包到获取最终结果耗时 {asr_end_elapsed:.2f} 秒", + ) + except Exception as e: + session_logger.error( + session.device_id or "unknown", + session.session_id or "unknown", + f"ASR结束处理出错: {e}", + exc_info=True, + ) + finally: + if session.send_task: + session.send_task.cancel() + try: + await asyncio.wait_for(session.send_task, timeout=2.0) + except (asyncio.CancelledError, asyncio.TimeoutError): + pass + except Exception as e: + session_logger.error( + session.device_id or "unknown", + session.session_id or "unknown", + f"取消发送任务时出错: {e}", + ) + finally: + session.send_task = None + + await session.asr_service.close() + session_logger.info( + session.device_id or "unknown", + session.session_id or "unknown", + "语音识别 WebSocket 连接和 ClientSession 已关闭", + ) + else: + session_logger.error( + session.device_id or "unknown", + session.session_id or "unknown", + "语音识别 WebSocket 连接未建立", + ) diff --git a/talkingq-url/handlers/session_cleanup_handler.py b/talkingq-url/handlers/session_cleanup_handler.py new file mode 100644 index 0000000..f73c751 --- /dev/null +++ b/talkingq-url/handlers/session_cleanup_handler.py @@ -0,0 +1,53 @@ +from services.interrupt_handler import interrupt_handler +from services.audio_session import audio_session_manager +from utils.logger import session_logger + +async def handle_old_session_cleanup(old_session, session_key): + """ + 清理旧会话资源 + + Args: + old_session: 需要清理的会话对象 + session_key: 会话键 (device_id, session_id) + """ + try: + device_id, session_id = session_key + session_logger.info(device_id, session_id, "开始主动清理旧会话资源") + if hasattr(old_session, "register_interrupt_handlers"): + await old_session.register_interrupt_handlers() + await interrupt_handler.set_interrupt_state(session_key, True) + await interrupt_handler.handle_interrupt(session_key) + except Exception as e: + session_logger.error("unknown", "unknown", f"清理旧会话资源时出错: {e}") + + +async def cleanup_device_sessions(device_id): + """ + 清理设备所有会话 + + Args: + device_id: 设备ID + """ + sessions_to_clean = [] + for session_key, session in await audio_session_manager.get_all_sessions(): + if session_key[0] == device_id: + sessions_to_clean.append((session_key, session)) + + for session_key, session in sessions_to_clean: + if hasattr(session, "set_interrupted"): + session_logger.info( + device_id, session_key[1], "WebSocket连接断开，中断会话" + ) + await session.set_interrupted(True) + await interrupt_handler.handle_interrupt(session_key) + else: + await interrupt_handler.set_interrupt_state(session_key, True) + if session.llm_service and hasattr(session.llm_service, "closed"): + session.llm_service.closed = True + + for session_key, _ in sessions_to_clean: + await audio_session_manager.remove_session(session_key) + await interrupt_handler.remove_session(session_key) + session_logger.info( + device_id, session_key[1], "已从会话管理器中移除会话" + ) diff --git a/talkingq-url/handlers/transcription_handler.py b/talkingq-url/handlers/transcription_handler.py new file mode 100644 index 0000000..dc2eef7 --- /dev/null +++ b/talkingq-url/handlers/transcription_handler.py @@ -0,0 +1,75 @@ +import asyncio +import time +from services.audio_session import AudioSession +from utils.logger import session_logger +from config import settings +from services.device_config import device_config_manager +from utils.language_detector import LanguageDetector + +async def handle_transcription(session: AudioSession, device_id: str, session_id: str): + start_time = time.perf_counter() + transcript = session.transcript + if transcript: + session_logger.info(device_id, session_id, f"最终转录结果: {transcript}") + detected_language = await LanguageDetector.detect_language(transcript) + session_logger.info(device_id, session_id, f"检测到的语言: {detected_language}") + device_config = await device_config_manager.get_config(device_id) + preferred_language = None + if device_config and hasattr(device_config, 'preferred_language') and device_config.preferred_language: + preferred_language = device_config.preferred_language + effective_language = detected_language or preferred_language + if detected_language and (not preferred_language or detected_language != preferred_language): + session_logger.info( + device_id, + session_id, + f"使用检测到的语言: {detected_language}" + ) + elif preferred_language and detected_language != preferred_language: + session_logger.info( + device_id, + session_id, + f"检测到语言 {detected_language}，但使用首选语言: {preferred_language}" + ) + + from services.conversation_history import ( + DeviceConversationHistory, + conversation_history_manager, + ) + from handlers.command_handler import get_device_role + selected_role = await get_device_role(device_id, effective_language) + role_key = selected_role.get("role_key") or settings.selected_role_key + device_history = await conversation_history_manager.get_history(device_id, role_key) + if not device_history: + device_history = DeviceConversationHistory() + await conversation_history_manager.set_history(device_id, device_history, role_key) + device_history.last_interaction_time = asyncio.get_running_loop().time() + history = device_history.history[-settings.max_conversation_history :] + + from handlers.response_processor import generate_and_process_response + resp_start_time = time.perf_counter() + await generate_and_process_response( + transcript, + history, + selected_role, + device_id, + session_id, + device_history, + session, + effective_language, + ) + resp_end_time = time.perf_counter() + session_logger.info( + device_id, + session_id, + f"回复生成和处理 耗时: {resp_end_time - resp_start_time:.2f} 秒", + ) + else: + session_logger.error("unknown", "unknown", "未收到转录结果") + from handlers.prompt_sound_handler import handle_prompt_sound_request + await handle_prompt_sound_request(device_id, "tts_error") + from services.tts_error_manager import tts_error_manager + await tts_error_manager.end_tts_session(device_id, session_id) + end_time = time.perf_counter() + session_logger.info( + device_id, session_id, f"总处理 耗时: {end_time - start_time:.2f} 秒" + ) diff --git a/talkingq-url/handlers/websocket_auth_handler.py b/talkingq-url/handlers/websocket_auth_handler.py new file mode 100644 index 0000000..3c39c01 --- /dev/null +++ b/talkingq-url/handlers/websocket_auth_handler.py @@ -0,0 +1,66 @@ +import json +import asyncio +import time +from fastapi import WebSocket +from services.device_auth_manager import device_auth_manager +from services.connection_manager import connection_manager +from utils.logger import session_logger + +async def authenticate_websocket(websocket: WebSocket): + """ + 处理设备认证流程 + + Args: + websocket: WebSocket连接 + + Returns: + tuple: (认证状态, 设备ID) - (是否认证成功, 设备ID) + """ + device_id = None + authenticated = False + + auth_timeout = 10 # 10秒认证超时 + auth_start_time = time.time() + + while not authenticated and time.time() - auth_start_time < auth_timeout: + try: + message = await asyncio.wait_for( + websocket.receive(), + timeout=auth_timeout - (time.time() - auth_start_time) + ) + + if message["type"] == "websocket.receive" and "text" in message: + try: + auth_data = json.loads(message["text"]) + if "device_id" in auth_data and "serial_number" in auth_data: + device_id = auth_data["device_id"] + serial_number = auth_data["serial_number"] + authenticated = await device_auth_manager.authenticate_device( + device_id, serial_number + ) + if authenticated: + await connection_manager.add_connection(device_id, websocket) + session_logger.info( + device_id, "auth", f"设备 {device_id} 认证成功" + ) + await websocket.send_text( + json.dumps({"status": "authenticated"}) + ) + else: + session_logger.warning( + device_id, "auth", f"设备 {device_id} 认证失败" + ) + await websocket.send_text( + json.dumps( + {"status": "error", "message": "Authentication failed"} + ) + ) + except json.JSONDecodeError: + session_logger.warning("unknown", "auth", "收到的不是有效的JSON认证消息") + elif message["type"] == "websocket.receive" and "bytes" in message: + pass + except asyncio.TimeoutError: + session_logger.warning("unknown", "auth", "WebSocket认证超时") + break + + return authenticated, device_id diff --git a/talkingq-url/handlers/websocket_handler.py b/talkingq-url/handlers/websocket_handler.py new file mode 100644 index 0000000..050f9a7 --- /dev/null +++ b/talkingq-url/handlers/websocket_handler.py @@ -0,0 +1,55 @@ +from fastapi import WebSocket, WebSocketDisconnect +from utils.logger import session_logger +from services.connection_manager import connection_manager +from services.task_manager import task_manager +from handlers.websocket_auth_handler import authenticate_websocket +from handlers.websocket_message_handler import handle_websocket_messages +from handlers.session_cleanup_handler import cleanup_device_sessions + +async def websocket_endpoint(websocket: WebSocket): + """ + WebSocket连接的主入口点 + + Args: + websocket: WebSocket连接 + """ + await websocket.accept() + session_logger.info( + "unknown", "connection", f"WebSocket 连接已建立: {websocket.client}" + ) + device_id = None + + try: + authenticated, device_id = await authenticate_websocket(websocket) + + if not authenticated: + session_logger.warning( + device_id or "unknown", "auth", "未认证的设备尝试连接，断开连接" + ) + await websocket.send_text('{"status": "error", "message": "Not authenticated"}') + return + + await handle_websocket_messages(websocket, device_id) + + except WebSocketDisconnect: + session_logger.info( + device_id or "unknown", + "connection", + f"WebSocket 连接已关闭: {websocket.client}", + ) + except Exception as e: + session_logger.error( + device_id or "unknown", "error", f"处理 WebSocket 数据时出错: {str(e)}" + ) + finally: + if device_id: + await connection_manager.remove_connection(device_id) + await cleanup_device_sessions(device_id) + # 清理设备相关的所有异步任务 + await task_manager.cancel_device_tasks(device_id) + + session_logger.info( + device_id or "unknown", + "connection", + "服务端保持WebSocket连接开放，由客户端负责断开连接", + ) diff --git a/talkingq-url/handlers/websocket_message_handler.py b/talkingq-url/handlers/websocket_message_handler.py new file mode 100644 index 0000000..a15946e --- /dev/null +++ b/talkingq-url/handlers/websocket_message_handler.py @@ -0,0 +1,180 @@ +import json +import time +import asyncio +from fastapi import WebSocket +from handlers.audio_packet_parser import parse_packet +from handlers.audio_session_handler import handle_websocket_data +from services.audio_session import audio_session_manager +from services.interrupt_handler import interrupt_handler +from services.task_manager import task_manager +from utils.logger import session_logger +from handlers.prompt_sound_handler import handle_prompt_sound_request +from handlers.session_cleanup_handler import handle_old_session_cleanup + +async def handle_websocket_messages(websocket: WebSocket, device_id: str): + """ + 处理WebSocket连接中的所有消息 + + Args: + websocket: WebSocket连接 + device_id: 设备ID + """ + current_active_session = None + first_audio_received_time = None + + while True: + message = await websocket.receive() + + if not websocket.client_state.name == "CONNECTED": + session_logger.info( + device_id, + "connection", + "WebSocket连接已关闭，退出消息循环", + ) + break + + if message["type"] == "websocket.receive": + data = message["text"] if "text" in message else message["bytes"] + + if isinstance(data, str): + await handle_text_message(websocket, device_id, data) + else: # 处理二进制音频数据 + current_active_session, first_audio_received_time = await handle_binary_message( + websocket, + device_id, + data, + current_active_session, + first_audio_received_time + ) + elif message["type"] == "websocket.disconnect": + break + + +async def handle_text_message(websocket: WebSocket, device_id: str, text_data: str): + """处理文本消息""" + if text_data.startswith("REQUEST_PROMPT_SOUND:"): + prompt_type = text_data.split(":", 1)[1] + if device_id: + await handle_prompt_sound_request(device_id, prompt_type) + elif text_data == "NETWORK_RESET_ACKNOWLEDGED": + session_logger.info(device_id, "network", f"设备 {device_id} 确认网络重置") + elif text_data.startswith("FIRMWARE_UPDATE_STATUS:"): + status_info = text_data.split(":", 1)[1] + parts = status_info.split(",") + status_dict = {} + for part in parts: + if "=" in part: + key, value = part.split("=", 1) + status_dict[key.strip()] = value.strip() + + from services.device_update_manager import device_firmware_update_manager + + update_status = status_dict.get("status", "updating") + progress = float(status_dict.get("progress", "0")) if "progress" in status_dict else 0.0 + + await device_firmware_update_manager.update_firmware_progress(device_id, progress) + + if update_status in ["success", "failed", "completed"]: + version = status_dict.get("version", "unknown") + await device_firmware_update_manager.update_firmware_update( + device_id, + firmware_version=version, + update_status=update_status + ) + + session_logger.info(device_id, "update", f"固件更新状态已保存: {update_status}, 进度: {progress}") + elif text_data.startswith("FIRMWARE_VERSION:"): + version_info = text_data.split(":", 1)[1].strip() + session_logger.info(device_id, "update", f"收到设备固件版本: {version_info}") + from services.device_update_manager import device_firmware_update_manager + await device_firmware_update_manager.update_device_firmware_version( + device_id, + version_info + ) + await device_firmware_update_manager.update_firmware_update( + device_id, + version_info, + "success" # 收到版本信息说明设备当前状态正常 + ) + + +async def handle_binary_message(websocket: WebSocket, device_id: str, binary_data, current_active_session, first_audio_received_time): + """处理二进制音频消息""" + session_key, session, packet_type, audio_data, sample_rate = await parse_packet(binary_data, websocket) + + if session: + session.websocket = websocket + session_id = session_key[1] + + if packet_type == 1: # 开始包 + current_active_session = await handle_start_packet( + device_id, session_id, session_key, session, current_active_session + ) + elif packet_type == 3: # 中断包 + current_active_session = await handle_interrupt_packet( + device_id, session_id, session_key, session, current_active_session + ) + elif packet_type == 2: # 结束包 + if current_active_session == session_key: + current_active_session = None + elif packet_type == 0 and first_audio_received_time is None: + first_audio_received_time = time.perf_counter() + session.start_time = first_audio_received_time + session_logger.info( + device_id, session_id, f"收到第一个音频数据包时间: {first_audio_received_time}" + ) + + await handle_websocket_data(websocket, session_key, session, packet_type, audio_data, sample_rate) + + return current_active_session, first_audio_received_time + + +async def handle_start_packet(device_id, session_id, session_key, session, current_active_session): + """处理会话开始包""" + if current_active_session and current_active_session[1] != session_id: + old_session = await audio_session_manager.get_session(current_active_session) + if old_session: + session_logger.info( + device_id, + current_active_session[1], + f"检测到新会话 {session_id}，优先中断旧会话 {current_active_session[1]}", + ) + await interrupt_handler.set_interrupt_state(current_active_session, True) + if hasattr(old_session.llm_service, "close"): + await old_session.llm_service.close() + if hasattr(old_session.tts_service, "close"): + await old_session.tts_service.close() + await task_manager.create_task( + handle_old_session_cleanup(old_session, current_active_session), + device_id=device_id, + session_key=current_active_session, + task_type="cleanup" + ) + + current_active_session = session_key + await interrupt_handler.set_interrupt_state(session_key, False) + if hasattr(session.llm_service, "closed"): + session.llm_service.closed = False + session_logger.info(device_id, session_id, f"设置当前活跃会话: {session_id}") + await interrupt_handler.register_session(session_key) + + return current_active_session + + +async def handle_interrupt_packet(device_id, session_id, session_key, session, current_active_session): + """处理中断包""" + if current_active_session == session_key: + current_active_session = None + session_logger.info(device_id, session_id, "收到中断包，立即设置中断状态") + await interrupt_handler.set_interrupt_state(session_key, True) + if hasattr(session.llm_service, "closed"): + session.llm_service.closed = True + if interrupt_handler.is_processing_interrupt(session_key): + session_logger.info(device_id, session_id, "检测到连续中断，中断请求将被排队") + await task_manager.create_task( + interrupt_handler.handle_interrupt(session_key), + device_id=device_id, + session_key=session_key, + task_type="interrupt" + ) + return current_active_session diff --git a/talkingq-url/implementations/__init__.py b/talkingq-url/implementations/__init__.py new file mode 100644 index 0000000..e69de29 diff --git a/talkingq-url/implementations/aliyun_asr.py b/talkingq-url/implementations/aliyun_asr.py new file mode 100644 index 0000000..99059c8 --- /dev/null +++ b/talkingq-url/implementations/aliyun_asr.py @@ -0,0 +1,251 @@ +import asyncio +from dashscope.audio.asr import TranslationRecognizerChat +from dashscope.audio.asr import TranscriptionResult +from dashscope.audio.asr import TranslationRecognizerCallback +from config import settings +from interfaces.asr import ASR +from implementations.base_service import BaseService +from services.interrupt_handler import interrupt_handler +from utils.logger import session_logger + + +class GummyCallback(TranslationRecognizerCallback): + def __init__(self, device_id, session_id): + self.device_id = device_id + self.session_id = session_id + self.transcript = "" + self.transcript_event = asyncio.Event() + self.received_result = False + self.error_message = None + self.is_closed = False + self.session_end = False + + def on_open(self) -> None: + session_logger.info(self.device_id, self.session_id, "Gummy ASR连接建立成功") + + def on_event( + self, + request_id, + transcription_result: TranscriptionResult, + translation_result, + usage, + ) -> None: + if transcription_result: + if transcription_result.is_sentence_end: + self.transcript = transcription_result.text + self.received_result = True + session_logger.info( + self.device_id, + self.session_id, + f"Gummy ASR完整句子: {self.transcript}", + ) + self.transcript_event.set() + else: + self.transcript = transcription_result.text + + def on_complete(self) -> None: + session_logger.info(self.device_id, self.session_id, "Gummy ASR识别完成") + self.session_end = True + self.transcript_event.set() + + def on_error(self, result) -> None: + self.error_message = str(result) + session_logger.error( + self.device_id, self.session_id, f"Gummy ASR错误: {self.error_message}" + ) + self.transcript_event.set() + + def on_close(self) -> None: + session_logger.info(self.device_id, self.session_id, "Gummy ASR连接已关闭") + self.is_closed = True + self.transcript_event.set() + + +class AliyunASR(ASR, BaseService): + def __init__(self, selected_role=None): + BaseService.__init__(self, "asr", selected_role) + self.api_key = settings.aliyun_api_key + self.vocabulary_id = settings.aliyun_vocabulary_id + self.connected = False + self.transcript = "" + self.device_id = "unknown" + self.session_id = "unknown" + self.sample_rate = 16000 + self.audio_buffer = bytearray() + self.translator = None + self.callback = None + self.buffer_size = 3200 # 约100ms的音频，按16000采样率计算 + self.max_end_silence = 2000 # 设为2000ms，允许更长的停顿 + + async def connect(self): + try: + if self.connected and self.translator: + session_logger.info( + self.device_id, self.session_id, "Gummy ASR连接已存在，复用当前连接" + ) + return + + self.transcript = "" + self.audio_buffer = bytearray() + + self.callback = GummyCallback(self.device_id, self.session_id) + + self.translator = TranslationRecognizerChat( + model="gummy-chat-v1", + format="mp3", + sample_rate=self.sample_rate, + callback=self.callback, + transcription_enabled=True, + source_language="auto", + semantic_punctuation_enabled=False, + max_end_silence=self.max_end_silence, + api_key=self.api_key, + vocabulary_id=self.vocabulary_id, + ) + + self.translator.start() + self.callback.transcript_event.clear() + self.connected = True + + except Exception as e: + session_logger.error( + self.device_id, self.session_id, f"建立Gummy ASR连接失败: {str(e)}" + ) + self.connected = False + if self.translator: + try: + self.translator.stop() + except: + pass + self.translator = None + + async def send_audio(self, audio_data: bytes): + if not self.connected or not self.translator: + session_logger.error(self.device_id, self.session_id, "Gummy ASR连接未建立") + return + + try: + self.audio_buffer.extend(audio_data) + + while len(self.audio_buffer) >= self.buffer_size: + chunk = bytes(self.audio_buffer[: self.buffer_size]) + self.audio_buffer = self.audio_buffer[self.buffer_size :] + + if self.device_id and self.session_id: + session_key = (self.device_id, self.session_id) + if interrupt_handler.is_interrupted(session_key): + session_logger.info( + self.device_id, + self.session_id, + "检测到中断，停止发送音频数据", + ) + return + + try: + self.translator.send_audio_frame(chunk) + except Exception as e: + session_logger.error( + self.device_id, + self.session_id, + f"发送音频帧到Gummy ASR出错: {str(e)}", + ) + + except Exception as e: + session_logger.error( + self.device_id, self.session_id, f"发送音频到Gummy ASR出错: {str(e)}" + ) + + async def send_end(self): + if not self.connected or not self.translator: + session_logger.error( + self.device_id, self.session_id, "Gummy ASR连接未建立，无法发送结束信号" + ) + return + + try: + if len(self.audio_buffer) > 0: + chunk = bytes(self.audio_buffer) + self.audio_buffer = bytearray() + self.translator.send_audio_frame(chunk) + + session_logger.info( + self.device_id, + self.session_id, + "客户端发送结束包，等待Gummy ASR处理完毕", + ) + + loop = asyncio.get_running_loop() + await loop.run_in_executor(None, lambda: self._safe_stop_translator()) + + except Exception as e: + session_logger.error( + self.device_id, + self.session_id, + f"发送结束信号到Gummy ASR出错: {str(e)}", + ) + + async def receive_results(self): + if not self.connected: + session_logger.error( + self.device_id, self.session_id, "Gummy ASR连接未建立，无法接收结果" + ) + return + + try: + try: + await asyncio.wait_for( + self.callback.transcript_event.wait(), timeout=0.5 + ) + except asyncio.TimeoutError: + session_logger.warning( + self.device_id, + self.session_id, + "接收Gummy ASR结果超时，使用最后的中间结果", + ) + + if self.callback.received_result or self.callback.transcript: + self.transcript = self.callback.transcript + + except Exception as e: + session_logger.error( + self.device_id, self.session_id, f"接收Gummy ASR结果出错: {str(e)}" + ) + + return self.transcript + + async def close(self): + if self.connected: + try: + if self.translator is not None: + try: + self.translator.stop() + except Exception as e: + if "has stopped" not in str(e): + session_logger.warning( + self.device_id, + self.session_id, + f"关闭Gummy translator出错: {str(e)}" + ) + self.translator = None + self.connected = False + session_logger.info( + self.device_id, self.session_id, "Gummy ASR连接已关闭" + ) + except Exception as e: + session_logger.error( + self.device_id, self.session_id, f"关闭Gummy ASR连接出错: {str(e)}" + ) + + def _safe_stop_translator(self): + """安全停止translator，忽略已停止的异常""" + try: + if self.translator: + self.translator.stop() + except Exception as e: + if "has stopped" not in str(e): + session_logger.warning( + self.device_id, + self.session_id, + f"停止translator时出现非标准错误: {str(e)}", + ) + return None diff --git a/talkingq-url/implementations/base_llm.py b/talkingq-url/implementations/base_llm.py new file mode 100644 index 0000000..a2834d4 --- /dev/null +++ b/talkingq-url/implementations/base_llm.py @@ -0,0 +1,212 @@ +import json +import aiohttp +import asyncio +import time +from typing import List, Dict, AsyncGenerator +from interfaces.llm import LLM +from utils.logger import session_logger +from config import settings +from services.interrupt_handler import interrupt_handler +from implementations.base_service import BaseService + +class BaseLLM(LLM, BaseService): + def __init__(self, api_key: str, base_url: str, service_name: str, **kwargs): + BaseService.__init__(self, "llm", kwargs.get("selected_role")) + self.session = None + self.api_key = api_key + self.base_url = base_url + self.connected = False + self.closed = False + self.response = None + self.device_id = None + self.session_id = None + self.service_name = service_name + + async def connect(self): + try: + if self.session: + await self.session.close() + self.session = aiohttp.ClientSession() + self.connected = True + except Exception as e: + session_logger.error( + self.device_id, + self.session_id, + f"[{self.service_name}] 连接建立失败: {str(e)}", + ) + self.connected = False + + async def prepare_request( + self, transcript: str, history: List[Dict[str, str]], selected_role: dict + ) -> Dict: + start_time = time.perf_counter() + if ( + hasattr(self, "_system_message") + and self._system_message.get("content") == selected_role["content"] + ): + messages = [self._system_message] # 复用缓存的系统消息 + else: + messages = [{"role": "system", "content": selected_role["content"]}] + self._system_message = messages[0] # 更新缓存 + recent_history = history[-settings.max_conversation_history :] + for exchange in recent_history: + messages.append({"role": "user", "content": exchange["user"]}) + messages.append({"role": "assistant", "content": exchange["assistant"]}) + messages.append({"role": "user", "content": transcript}) + data = { + "model": self.model_id, + "messages": messages, + "stream": True, + } + end_time = time.perf_counter() + session_logger.info( + self.device_id, + self.session_id, + f"[{self.service_name}] 请求准备完成: {len(messages)}条消息, 耗时: {end_time - start_time:.4f}秒", + ) + return data + + async def generate_response_stream( + self, transcript: str, history: List[Dict[str, str]], selected_role: dict + ) -> AsyncGenerator[str, None]: + if not self.connected or not self.session: + await self.connect() + if not self.connected: + yield None + return + self.device_id = getattr(self, "device_id", "unknown") + self.session_id = getattr(self, "session_id", "unknown") + session_key = ( + (self.device_id, self.session_id) + if self.device_id != "unknown" and self.session_id != "unknown" + else None + ) + try: + if session_key: + self.closed = interrupt_handler.is_interrupted(session_key) + data = await self.prepare_request(transcript, history, selected_role) + if session_key: + await interrupt_handler.set_interrupt_state(session_key, False) + self.closed = False + session_logger.info( + self.device_id, self.session_id, f"[{self.service_name}] 开始生成回复" + ) + headers = { + "Authorization": f"Bearer {self.api_key}", + "Content-Type": "application/json", + } + try: + if not self.session or self.session.closed: + self.session = aiohttp.ClientSession() + self.connected = True + response = await self.session.post( + f"{self.base_url}/chat/completions", headers=headers, json=data + ) + self.response = response + if response.status == 200: + async for line in response.content: + if session_key and interrupt_handler.is_interrupted( + session_key + ): + self.closed = True + if self.closed: + session_logger.info( + self.device_id, + self.session_id, + f"[{self.service_name}] 生成被中断", + ) + break + decoded_line = line.decode("utf-8").strip() + if not decoded_line: + continue + if decoded_line.startswith("data:"): + data_str = decoded_line[len("data:") :].strip() + if data_str == "[DONE]": + break + try: + data_json = json.loads(data_str) + except json.JSONDecodeError: + continue + choices = data_json.get("choices", []) + if choices: + delta = choices[0].get("delta", {}) + content = delta.get("content") + if content: + if self.closed: + session_logger.info( + self.device_id, + self.session_id, + f"[{self.service_name}] 生成被中断", + ) + break + yield content + else: + error_text = await response.text() + session_logger.error( + self.device_id, + self.session_id, + f"[{self.service_name}] 生成回复失败: {error_text}", + ) + yield None + except asyncio.CancelledError: + session_logger.info( + self.device_id, self.session_id, f"[{self.service_name}] 请求被取消" + ) + yield None + except Exception as e: + session_logger.error( + self.device_id, + self.session_id, + f"[{self.service_name}] API请求错误: {str(e)}", + ) + yield None + except Exception as e: + session_logger.error( + self.device_id, + self.session_id, + f"[{self.service_name}] 生成过程发生错误: {str(e)}", + ) + yield None + finally: + if self.response: + await self.response.release() + self.response = None + + async def close(self): + if ( + hasattr(self, "device_id") + and hasattr(self, "session_id") + and self.device_id != "unknown" + and self.session_id != "unknown" + ): + session_key = (self.device_id, self.session_id) + await interrupt_handler.set_interrupt_state(session_key, True) + self.closed = True # 本地标志也同步设置 + if self.response: + try: + self.response.close() # 立即关闭响应，不等待release + await self.response.release() + except Exception as e: + if hasattr(self, "device_id") and hasattr(self, "session_id"): + session_logger.warning( + self.device_id, + self.session_id, + f"[{self.service_name}] 关闭响应时出错: {e}", + ) + self.response = None + if self.session: + try: + await self.session.close() + except Exception as e: + if hasattr(self, "device_id") and hasattr(self, "session_id"): + session_logger.warning( + self.device_id, + self.session_id, + f"[{self.service_name}] 关闭会话时出错: {e}", + ) + self.session = None + self.connected = False + if hasattr(self, "device_id") and hasattr(self, "session_id"): + session_logger.info( + self.device_id, self.session_id, f"[{self.service_name}] 连接已关闭" + ) diff --git a/talkingq-url/implementations/base_service.py b/talkingq-url/implementations/base_service.py new file mode 100644 index 0000000..7515e24 --- /dev/null +++ b/talkingq-url/implementations/base_service.py @@ -0,0 +1,21 @@ +from typing import Any, Dict, Optional +from services.config_manager import config_manager + +class BaseService: + """所有服务实现的基类，提供统一的配置处理功能""" + + def __init__(self, service_type: str, selected_role: Optional[Dict[str, Any]] = None): + self.selected_role = selected_role + self.service_type = service_type + self.config = config_manager.get_service_config(service_type, selected_role) + + def get_config(self, key: str, default_value: Any = None) -> Any: + """获取配置值，按照优先级规则""" + if key in self.config: + return self.config[key] + return config_manager.get_config_value(key, self.selected_role, default_value) + + def update_config(self, new_role_config: Dict[str, Any] = None) -> None: + """更新服务配置，用于动态切换角色时""" + self.selected_role = new_role_config if new_role_config else self.selected_role + self.config = config_manager.get_service_config(self.service_type, self.selected_role) diff --git a/talkingq-url/implementations/minimax_tts.py b/talkingq-url/implementations/minimax_tts.py new file mode 100644 index 0000000..9f14d68 --- /dev/null +++ b/talkingq-url/implementations/minimax_tts.py @@ -0,0 +1,202 @@ +import os +import uuid +import aiohttp +import aiofiles +import binascii +import json +from typing import Optional, List, Any, AsyncGenerator +from config import settings +from interfaces.tts import TTS +from implementations.base_service import BaseService +from utils.logger import session_logger +from services.tts_audio_cleaner import TTSAudioCleaner + +class MiniMaxTTS(TTS, BaseService): + def __init__(self, selected_role=None): + BaseService.__init__(self, "tts", selected_role) + self.api_key = settings.minimax_api_key + self.group_id = settings.minimax_group_id + self.base_url = settings.minimax_base_url + self.connected = False + self.device_id = "unknown" + self.session_id = "unknown" + self.voice_id = "cartoon-boy-01" # 默认音色 + if selected_role and "minimax_voice_id" in selected_role: + self.voice_id = selected_role["minimax_voice_id"] + self.model = "speech-02-turbo" + self.selected_role = selected_role + + async def connect(self): + """连接到MiniMax服务""" + self.connected = True + session_logger.info( + self.device_id, + self.session_id, + f"MiniMax TTS服务已连接，使用音色: {self.voice_id}" + ) + + async def close(self): + """关闭MiniMax服务连接""" + self.connected = False + + async def tts( + self, + text: str, + output_file_prefix: Optional[str] = None, + tts_format: str = "mp3", + selected_role: dict = None, + session: Optional[Any] = None, + language: str = None, + ) -> Optional[List[str]]: + """ + 将文本转换为语音 + 返回值为包含音频URL的列表 + """ + if selected_role: + self.selected_role = selected_role + if "minimax_voice_id" in selected_role: + self.voice_id = selected_role["minimax_voice_id"] + + if session and hasattr(session, "device_id") and session.device_id: + self.device_id = session.device_id + if session and hasattr(session, "session_id") and session.session_id: + self.session_id = session.session_id + + unique_id = f"{self.session_id}_{uuid.uuid4().hex[:8]}" + + if output_file_prefix: + output_file = f"{output_file_prefix}.{tts_format}" + else: + await TTSAudioCleaner.prepare_output_directory() + tts_dir = os.path.join(settings.assets_dir, "tts_audio") + output_file = os.path.join(tts_dir, f"{unique_id}.{tts_format}") + + try: + session_logger.info( + self.device_id, + self.session_id, + f"开始MiniMax流式语音合成，文本长度: {len(text)}，输出文件: {output_file}" + ) + + urls = await self._process_stream(text, output_file, language) + return urls if urls else None + except Exception as e: + session_logger.error( + self.device_id, + self.session_id, + f"MiniMax TTS合成失败: {str(e)}", + exc_info=True + ) + return None + + async def _process_stream(self, text: str, output_file: str, language: str = None) -> Optional[List[str]]: + """使用流式方式处理文本并返回URL""" + headers = { + "Authorization": f"Bearer {self.api_key}", + "Content-Type": "application/json" + } + + payload = { + "model": self.model, + "text": text, + "stream": True, + "voice_setting": { + "voice_id": self.voice_id, + "speed": 1.0, + "vol": 2.0, + "pitch": 0 + }, + "audio_setting": { + "sample_rate": 16000, + "bitrate": 32000, + "format": "mp3", + "channel": 1 + } + } + + if language: + lang_mapping = { + "zh": "Chinese", + "en": "English", + "fr": "French", + "de": "German", + "es": "Spanish", + "yue": "Chinese,Yue" + } + payload["language_boost"] = lang_mapping.get(language, "auto") + else: + payload["language_boost"] = "auto" + + url = f"{self.base_url}?GroupId={self.group_id}" + + try: + session_logger.info( + self.device_id, + self.session_id, + f"发送MiniMax TTS流式请求: {url}" + ) + + audio_buffer = bytearray() + + async with aiohttp.ClientSession() as client_session: + async with client_session.post(url, json=payload, headers=headers) as response: + if response.status != 200: + error_text = await response.text() + session_logger.error( + self.device_id, + self.session_id, + f"MiniMax TTS请求失败: {response.status}, {error_text}" + ) + return None + + session_logger.info( + self.device_id, + self.session_id, + "开始接收MiniMax TTS流式响应" + ) + + async for line in response.content: + if line.startswith(b'data:'): + try: + data_json = json.loads(line[5:]) + if "data" in data_json and "audio" in data_json["data"]: + status = data_json["data"].get("status", 1) + + if status == 1: # 只处理status=1(合成中)的音频数据，忽略status=2(合成结束)的汇总数据 + audio_hex = data_json["data"]["audio"] + audio_binary = binascii.unhexlify(audio_hex) + audio_buffer.extend(audio_binary) + + if status == 2: # 合成结束 + session_logger.info( + self.device_id, + self.session_id, + "MiniMax TTS流式合成完成" + ) + except Exception as e: + session_logger.error( + self.device_id, + self.session_id, + f"处理MiniMax TTS流式响应出错: {str(e)}", + exc_info=True + ) + + async with aiofiles.open(output_file, 'wb') as f: + await f.write(audio_buffer) + + relative_path = os.path.relpath(output_file, settings.assets_dir) + url = f"assets/{relative_path}" + session_logger.info( + self.device_id, + self.session_id, + f"MiniMax TTS流式合成完成，生成URL: {url}" + ) + return [url] + except Exception as e: + session_logger.error( + self.device_id, + self.session_id, + f"MiniMax TTS流式处理失败: {str(e)}", + exc_info=True + ) + return None diff --git a/talkingq-url/implementations/volcano_llm.py b/talkingq-url/implementations/volcano_llm.py new file mode 100644 index 0000000..18b03bb --- /dev/null +++ b/talkingq-url/implementations/volcano_llm.py @@ -0,0 +1,41 @@ +from implementations.base_llm import BaseLLM +from config import settings +from typing import List, Dict + +class VolcanoLLM(BaseLLM): + def __init__(self, selected_role=None): + self.model_id = settings.volcano_model_id + if selected_role and "volcano_model_id" in selected_role: + self.model_id = selected_role["volcano_model_id"] + session_source = "角色配置" + else: + session_source = "全局配置" + + if hasattr(self, 'logger'): + self.logger.info( + "unknown", + f"Volcano模型ID: {self.model_id} (来源: {session_source})" + ) + + if self.model_id.startswith("bot-"): + base_url = f"{settings.volcano_base_url}/bots" + else: + base_url = settings.volcano_base_url + + super().__init__( + api_key=settings.volcano_api_key, + base_url=base_url, + service_name="Volcano LLM", + selected_role=selected_role, + ) + + async def prepare_request( + self, transcript: str, history: List[Dict[str, str]], selected_role: dict + ) -> Dict: + if selected_role and "volcano_model_id" in selected_role: + self.model_id = selected_role["volcano_model_id"] + if self.model_id.startswith("bot-"): + self.base_url = f"{settings.volcano_base_url}/bots" + else: + self.base_url = settings.volcano_base_url + return await super().prepare_request(transcript, history, selected_role) diff --git a/talkingq-url/initialization.py b/talkingq-url/initialization.py new file mode 100644 index 0000000..1425e3b --- /dev/null +++ b/talkingq-url/initialization.py @@ -0,0 +1,47 @@ +from services.registry import register_service +from implementations.aliyun_asr import AliyunASR +from implementations.volcano_llm import VolcanoLLM +from implementations.minimax_tts import MiniMaxTTS +from utils.logger import session_logger +import os +from config import settings +from services.firmware_scanner import firmware_scanner + +def init_directories(): + """初始化必要的目录""" + tts_audio_dir = os.path.join(settings.assets_dir, "tts_audio") + os.makedirs(tts_audio_dir, exist_ok=True) + session_logger.info("system", "initialization", f"TTS音频目录已初始化: {tts_audio_dir}") + + firmware_dir = os.path.join(settings.assets_dir, "firmware") + os.makedirs(firmware_dir, exist_ok=True) + session_logger.info("system", "initialization", f"固件目录已初始化: {firmware_dir}") + +async def check_database_initialization(): + """检查数据库是否已初始化""" + try: + from database.connection import get_db_manager + db_manager = await get_db_manager() + session = await db_manager.get_session() + await session.close() + session_logger.info("system", "initialization", "数据库已初始化") + except Exception as e: + session_logger.error("system", "initialization", f"数据库未初始化，请检查配置: {str(e)}") + raise Exception(f"数据库连接失败: {str(e)}") + +async def scan_firmware_files(): + """扫描固件文件并更新数据库""" + try: + await firmware_scanner.scan_firmware_files() + except Exception as e: + session_logger.error("system", "initialization", f"扫描固件文件失败: {str(e)}") + + +async def register_services(): + init_directories() + await check_database_initialization() + await scan_firmware_files() # 添加扫描固件文件的步骤 + register_service("asr", "Aliyun", AliyunASR) + register_service("llm", "Volcano", VolcanoLLM) + register_service("tts", "MiniMax", MiniMaxTTS) + session_logger.info("system", "initialization", "所有服务已注册") diff --git a/talkingq-url/interfaces/__init__.py b/talkingq-url/interfaces/__init__.py new file mode 100644 index 0000000..e69de29 diff --git a/talkingq-url/interfaces/asr.py b/talkingq-url/interfaces/asr.py new file mode 100644 index 0000000..4aa61df --- /dev/null +++ b/talkingq-url/interfaces/asr.py @@ -0,0 +1,27 @@ +from abc import ABC, abstractmethod + +class ASR(ABC): + @abstractmethod + async def connect(self): + + pass + + @abstractmethod + async def send_audio(self, audio_data: bytes): + + pass + + @abstractmethod + async def send_end(self): + + pass + + @abstractmethod + async def receive_results(self): + + pass + + @abstractmethod + async def close(self): + + pass diff --git a/talkingq-url/interfaces/llm.py b/talkingq-url/interfaces/llm.py new file mode 100644 index 0000000..96074f9 --- /dev/null +++ b/talkingq-url/interfaces/llm.py @@ -0,0 +1,19 @@ +from abc import ABC, abstractmethod +from typing import List, Dict, AsyncGenerator + + +class LLM(ABC): + @abstractmethod + async def generate_response_stream( + self, transcript: str, history: List[Dict[str, str]], selected_role: dict + ) -> AsyncGenerator[str, None]: + + pass + + async def close(self): + + pass + + async def connect(self): + + pass diff --git a/talkingq-url/interfaces/tts.py b/talkingq-url/interfaces/tts.py new file mode 100644 index 0000000..d4f554c --- /dev/null +++ b/talkingq-url/interfaces/tts.py @@ -0,0 +1,27 @@ +from abc import ABC, abstractmethod +from typing import Optional, List, Any + + +class TTS(ABC): + @abstractmethod + async def tts( + self, + text: str, + output_file_prefix: Optional[str] = None, + tts_format: str = "mp3", + selected_role: dict = None, + session: Optional[Any] = None, + language: str = None, + ) -> Optional[List[str]]: + + pass + + @abstractmethod + async def connect(self): + + pass + + @abstractmethod + async def close(self): + + pass diff --git a/talkingq-url/main.py b/talkingq-url/main.py new file mode 100644 index 0000000..e93635c --- /dev/null +++ b/talkingq-url/main.py @@ -0,0 +1,104 @@ +import asyncio +import os +from contextlib import asynccontextmanager +from fastapi import FastAPI +from services.device_config import DeviceConfigManager +from services.schedule_session_cleanup import background_cleanup_task +from services.role_manager import role_manager +from services.task_manager import task_manager +from initialization import register_services, init_directories +from config import settings +from utils.logger import session_logger +from api import api_router +from api.assets import configure_static_assets +from database.init_db import init_db +from database.connection import get_db_manager +from services.firmware_scanner import firmware_scanner + +device_config_manager = DeviceConfigManager() + + +@asynccontextmanager +async def lifespan(app: FastAPI): + worker_id = os.environ.get("UVICORN_WID", "0") + is_main_process = worker_id == "0" + if is_main_process: + session_logger.system_info( + "startup", + f"服务器启动，HTTP服务地址: http://{settings.server_host}:{settings.server_port}", + ) + init_directories() + + try: + session_logger.system_info("startup", "初始化数据库...") + await init_db() + session_logger.system_info("startup", "数据库初始化和迁移完成") + except Exception as e: + session_logger.system_info("startup", f"数据库初始化失败: {str(e)}") + + await role_manager.initialize() + await register_services() + + try: + await firmware_scanner.scan_firmware_files() + session_logger.system_info("startup", "固件文件扫描完成") + except Exception as e: + session_logger.system_info("startup", f"固件文件扫描失败: {str(e)}") + + session_logger.system_info("startup", "中断处理服务已初始化") + + cleanup_task = asyncio.create_task(background_cleanup_task()) + try: + yield + finally: + cleanup_task.cancel() + try: + await asyncio.wait_for(cleanup_task, timeout=5.0) + except (asyncio.CancelledError, asyncio.TimeoutError): + if is_main_process: + session_logger.system_info("shutdown", "后台清理任务已取消") + + from services.audio_session import audio_session_manager + sessions_to_close = [] + async for session_key, session in audio_session_manager.get_all_sessions(): + sessions_to_close.append((session_key, session)) + + for session_key, session in sessions_to_close: + try: + if session.send_task: + session.send_task.cancel() + if hasattr(session, "asr_service") and session.asr_service: + await session.asr_service.close() + if hasattr(session, "llm_service") and session.llm_service: + await session.llm_service.close() + except Exception as e: + if is_main_process: + session_logger.system_info("shutdown", f"清理会话时出错: {str(e)}") + + # 清理所有任务管理器的任务 + try: + await task_manager.cleanup_all_tasks() + if is_main_process: + session_logger.system_info("shutdown", "任务管理器清理完成") + except Exception as e: + if is_main_process: + session_logger.system_info("shutdown", f"任务管理器清理时出错: {str(e)}") + + try: + db_manager = await get_db_manager() + await db_manager.close() + if is_main_process: + session_logger.system_info("shutdown", "数据库连接已关闭") + except Exception as e: + if is_main_process: + session_logger.system_info("shutdown", f"关闭数据库连接时出错: {str(e)}") + + if is_main_process: + session_logger.system_info("shutdown", "应用关闭") + + +app = FastAPI(lifespan=lifespan) + +configure_static_assets(app) + +app.include_router(api_router) diff --git a/talkingq-url/mysql/conf.d/my.cnf b/talkingq-url/mysql/conf.d/my.cnf new file mode 100644 index 0000000..e0c0008 --- /dev/null +++ b/talkingq-url/mysql/conf.d/my.cnf @@ -0,0 +1,21 @@ +# Client配置 +[client] +default-character-set = utf8mb4 + +# MySQL命令行工具配置 +[mysql] +default-character-set = utf8mb4 + +# MySQL服务器配置 +[mysqld] +# 基础设置 +host_cache_size = 0 +authentication_policy = caching_sha2_password +pid-file = /var/lib/mysql/mysqld.pid +tls_version = TLSv1.2,TLSv1.3 + +# 字符集设置 +character-set-server = utf8mb4 +collation-server = utf8mb4_unicode_ci +character-set-client-handshake = FALSE +init_connect = 'SET NAMES utf8mb4' diff --git a/talkingq-url/mysql/init/01-init.sql b/talkingq-url/mysql/init/01-init.sql new file mode 100644 index 0000000..f6bfa9c --- /dev/null +++ b/talkingq-url/mysql/init/01-init.sql @@ -0,0 +1,161 @@ +-- 初始化数据库表 + +-- 创建device_configs表 +CREATE TABLE IF NOT EXISTS `device_configs` ( + `id` INT NOT NULL AUTO_INCREMENT, + `device_id` VARCHAR(64) NOT NULL, + `selected_role_key` VARCHAR(64) NOT NULL, + `preferred_language` VARCHAR(10) NULL, + `volume` INT NULL, + `last_update_time` FLOAT NOT NULL, + `created_at` DATETIME NOT NULL DEFAULT CURRENT_TIMESTAMP, + `updated_at` DATETIME NOT NULL DEFAULT CURRENT_TIMESTAMP ON UPDATE CURRENT_TIMESTAMP, + PRIMARY KEY (`id`), + UNIQUE INDEX `device_id_UNIQUE` (`device_id` ASC) +) ENGINE=InnoDB DEFAULT CHARSET=utf8mb4 COLLATE=utf8mb4_unicode_ci; + +-- 创建conversation_histories表 +CREATE TABLE IF NOT EXISTS `conversation_histories` ( + `id` INT NOT NULL AUTO_INCREMENT, + `device_id` VARCHAR(64) NOT NULL, + `role_key` VARCHAR(64) NOT NULL, + `last_interaction_time` FLOAT NOT NULL, + `created_at` DATETIME NOT NULL DEFAULT CURRENT_TIMESTAMP, + `updated_at` DATETIME NOT NULL DEFAULT CURRENT_TIMESTAMP ON UPDATE CURRENT_TIMESTAMP, + PRIMARY KEY (`id`), + INDEX `idx_device_id` (`device_id` ASC), + INDEX `idx_role_key` (`role_key` ASC) +) ENGINE=InnoDB DEFAULT CHARSET=utf8mb4 COLLATE=utf8mb4_unicode_ci; + +-- 创建conversation_messages表 +CREATE TABLE IF NOT EXISTS `conversation_messages` ( + `id` INT NOT NULL AUTO_INCREMENT, + `conversation_id` INT NOT NULL, + `is_user` TINYINT(1) NOT NULL DEFAULT 0, + `content` TEXT NOT NULL, + `timestamp` FLOAT NOT NULL, + `created_at` DATETIME NOT NULL DEFAULT CURRENT_TIMESTAMP, + `updated_at` DATETIME NOT NULL DEFAULT CURRENT_TIMESTAMP ON UPDATE CURRENT_TIMESTAMP, + PRIMARY KEY (`id`), + INDEX `idx_conversation_id` (`conversation_id` ASC), + CONSTRAINT `fk_messages_conversation` + FOREIGN KEY (`conversation_id`) + REFERENCES `conversation_histories` (`id`) + ON DELETE CASCADE +) ENGINE=InnoDB DEFAULT CHARSET=utf8mb4 COLLATE=utf8mb4_unicode_ci; + +-- 创建roles表 +CREATE TABLE IF NOT EXISTS `roles` ( + `id` INT NOT NULL AUTO_INCREMENT, + `role_key` VARCHAR(64) NOT NULL, + `name` VARCHAR(128) NOT NULL, + `description` TEXT NULL, + `content` TEXT NOT NULL, + `default_language` VARCHAR(10) NULL, + `asr_provider` VARCHAR(64) NULL, + `llm_provider` VARCHAR(64) NULL, + `tts_provider` VARCHAR(64) NULL, + `competitive_llm_mode` TINYINT(1) NULL, + `volcano_model_id` VARCHAR(64) NULL, + `volcano_voice_type` VARCHAR(64) NULL, + `tencent_voice_type` VARCHAR(64) NULL, + `aliyun_voice_name` VARCHAR(64) NULL, + `minimax_voice_id` VARCHAR(64) NULL, + `url` VARCHAR(255) NULL, + `homophones` JSON NULL, + `enabled` TINYINT(1) NOT NULL DEFAULT 1, + `created_at` DATETIME NOT NULL DEFAULT CURRENT_TIMESTAMP, + `updated_at` DATETIME NOT NULL DEFAULT CURRENT_TIMESTAMP ON UPDATE CURRENT_TIMESTAMP, + PRIMARY KEY (`id`), + UNIQUE INDEX `role_key_UNIQUE` (`role_key` ASC) +) ENGINE=InnoDB DEFAULT CHARSET=utf8mb4 COLLATE=utf8mb4_unicode_ci; + +-- 创建role_languages表 +CREATE TABLE IF NOT EXISTS `role_languages` ( + `id` INT NOT NULL AUTO_INCREMENT, + `role_id` INT NOT NULL, + `language_code` VARCHAR(10) NOT NULL, + `name` VARCHAR(128) NULL, + `content` TEXT NULL, + `asr_provider` VARCHAR(64) NULL, + `llm_provider` VARCHAR(64) NULL, + `tts_provider` VARCHAR(64) NULL, + `volcano_voice_type` VARCHAR(64) NULL, + `tencent_voice_type` VARCHAR(64) NULL, + `aliyun_voice_name` VARCHAR(64) NULL, + `minimax_voice_id` VARCHAR(64) NULL, + `url` VARCHAR(255) NULL, + `created_at` DATETIME NOT NULL DEFAULT CURRENT_TIMESTAMP, + `updated_at` DATETIME NOT NULL DEFAULT CURRENT_TIMESTAMP ON UPDATE CURRENT_TIMESTAMP, + PRIMARY KEY (`id`), + UNIQUE INDEX `uix_role_language` (`role_id`, `language_code`), + CONSTRAINT `fk_role_languages_role` + FOREIGN KEY (`role_id`) + REFERENCES `roles` (`id`) + ON DELETE CASCADE +) ENGINE=InnoDB DEFAULT CHARSET=utf8mb4 COLLATE=utf8mb4_unicode_ci; + +-- 创建device_auth表 +CREATE TABLE IF NOT EXISTS `device_auth` ( + `id` INT NOT NULL AUTO_INCREMENT, + `device_id` VARCHAR(64) NOT NULL, + `serial_number` VARCHAR(64) NOT NULL, + `batch_id` VARCHAR(20) NULL, + `is_active` TINYINT(1) NOT NULL DEFAULT 1, + `created_at` DATETIME NOT NULL DEFAULT CURRENT_TIMESTAMP, + `updated_at` DATETIME NOT NULL DEFAULT CURRENT_TIMESTAMP ON UPDATE CURRENT_TIMESTAMP, + PRIMARY KEY (`id`), + UNIQUE INDEX `device_id_UNIQUE` (`device_id` ASC), + INDEX `idx_batch_id` (`batch_id` ASC) +) ENGINE=InnoDB DEFAULT CHARSET=utf8mb4 COLLATE=utf8mb4_unicode_ci; + +-- 创建device_firmware_update表 +CREATE TABLE IF NOT EXISTS `device_firmware_update` ( + `id` INT NOT NULL AUTO_INCREMENT, + `device_id` VARCHAR(64) NOT NULL, + `serial_number` VARCHAR(64) NOT NULL, + `mac_address` VARCHAR(512) NULL, + `firmware_version` VARCHAR(64) NOT NULL, + `update_status` VARCHAR(32) NOT NULL DEFAULT 'success', + `progress` FLOAT NULL DEFAULT 0.0, + `created_at` DATETIME NOT NULL DEFAULT CURRENT_TIMESTAMP, + `updated_at` DATETIME NOT NULL DEFAULT CURRENT_TIMESTAMP ON UPDATE CURRENT_TIMESTAMP, + PRIMARY KEY (`id`), + UNIQUE INDEX `device_id_UNIQUE` (`device_id` ASC) +) ENGINE=InnoDB DEFAULT CHARSET=utf8mb4 COLLATE=utf8mb4_unicode_ci; + +-- 创建system_config表 +CREATE TABLE IF NOT EXISTS `system_config` ( + `id` INT NOT NULL AUTO_INCREMENT, + `config_key` VARCHAR(128) NOT NULL, + `config_value` TEXT NULL, + `description` VARCHAR(255) NULL, + `created_at` DATETIME NOT NULL DEFAULT CURRENT_TIMESTAMP, + `updated_at` DATETIME NOT NULL DEFAULT CURRENT_TIMESTAMP ON UPDATE CURRENT_TIMESTAMP, + PRIMARY KEY (`id`), + UNIQUE INDEX `config_key_UNIQUE` (`config_key` ASC) +) ENGINE=InnoDB DEFAULT CHARSET=utf8mb4 COLLATE=utf8mb4_unicode_ci; + +-- 添加初始角色配置 +INSERT INTO `device_configs` (`device_id`, `selected_role_key`, `preferred_language`, `last_update_time`) +VALUES ('default', 'assistant', 'zh', UNIX_TIMESTAMP()) +ON DUPLICATE KEY UPDATE `selected_role_key`=`selected_role_key`; + +-- 添加默认系统配置 +INSERT INTO `system_config` (`config_key`, `config_value`, `description`) +VALUES +('latest_firmware_version', '1.0.0', '最新固件版本'), +('update_firmware_url', 'https://example.com/firmware/latest.bin', '固件更新URL') +ON DUPLICATE KEY UPDATE `config_value`=`config_value`; + +-- 添加默认角色 +INSERT INTO `roles` (`role_key`, `name`, `description`, `content`, `default_language`, `enabled`) +VALUES ( + 'assistant', + '智能助手', + '默认智能助手角色', + '你是一个友好的智能助手，乐于帮助用户解答问题。', + 'zh', + 1 +) +ON DUPLICATE KEY UPDATE `role_key`=`role_key`; diff --git a/talkingq-url/mysql/mysql.md b/talkingq-url/mysql/mysql.md new file mode 100644 index 0000000..cd3b3ce --- /dev/null +++ b/talkingq-url/mysql/mysql.md @@ -0,0 +1,878 @@ +# TalkingQ MySQL 查询手册 + +## 数据库概览 + +TalkingQ 使用 MySQL 8.0 作为数据库后端，存储设备配置、角色信息和会话数据。本文档提供了数据库结构和常用查询操作的参考。 + +## 连接信息 + +``` +数据库名称: talkingq +用户名: talkingq +密码: 在 .env 或 .env.prod 文件中配置的 MYSQL_PASSWORD +主机: localhost (本地开发) 或 mysql (Docker环境) +端口: 3306 +``` + +## 连接数据库 + +```bash +# 本地开发环境 +mysql -u talkingq -p talkingq + +# Docker环境中 +sudo docker compose exec -it mysql mysql -u talkingq -p talkingq +``` + +## 表结构 + +### device_configs - 设备配置表 + +存储设备的角色选择和语言偏好等配置。 + +| 字段名 | 类型 | 描述 | +|-------------------|---------------|---------------------------| +| id | INT | 主键，自增 | +| device_id | VARCHAR(64) | 设备ID，唯一索引 | +| selected_role_key | VARCHAR(64) | 选中的角色键 | +| preferred_language| VARCHAR(10) | 首选语言代码 | +| last_update_time | FLOAT | 最后更新时间戳 | +| created_at | DATETIME | 创建时间 | +| updated_at | DATETIME | 更新时间 | + +### conversation_histories - 对话历史表 + +存储设备的对话历史记录和角色信息。 + +| 字段名 | 类型 | 描述 | +|--------------------|---------------|---------------------------| +| id | INT | 主键，自增 | +| device_id | VARCHAR(64) | 设备ID，索引 | +| role_key | VARCHAR(64) | 角色键，索引 | +| last_interaction_time | FLOAT | 最后交互时间戳 | +| created_at | DATETIME | 创建时间 | +| updated_at | DATETIME | 更新时间 | + +### conversation_messages - 对话消息表 + +存储对话的具体消息内容。 + +| 字段名 | 类型 | 描述 | +|--------------------|---------------|---------------------------| +| id | INT | 主键，自增 | +| conversation_id | INT | 关联的对话历史ID，外键 | +| is_user | BOOLEAN | 是否为用户消息 | +| content | TEXT | 消息内容 | +| timestamp | FLOAT | 消息时间戳 | +| created_at | DATETIME | 创建时间 | +| updated_at | DATETIME | 更新时间 | + +### roles - 角色表 + +存储系统中可用的角色定义。 + +| 字段名 | 类型 | 描述 | +|--------------------|---------------|---------------------------| +| id | INT | 主键，自增 | +| role_key | VARCHAR(64) | 角色键名，唯一索引 | +| name | VARCHAR(128) | 角色名称 | +| description | TEXT | 角色描述 | +| content | TEXT | 系统提示内容 | +| default_language | VARCHAR(10) | 默认语言代码 | +| asr_provider | VARCHAR(64) | ASR服务提供商 | +| llm_provider | VARCHAR(64) | LLM服务提供商 | +| tts_provider | VARCHAR(64) | TTS服务提供商 | +| competitive_llm_mode | BOOLEAN | 是否启用竞争模式 | +| volcano_model_id | VARCHAR(64) | Volcano模型ID | +| volcano_voice_type | VARCHAR(64) | Volcano音色类型 | +| tencent_voice_type | VARCHAR(64) | 腾讯音色类型 | +| aliyun_voice_name | VARCHAR(64) | 阿里云音色名称 | +| url | VARCHAR(255) | 资源URL路径 | +| homophones | JSON | 同音词列表 | +| enabled | BOOLEAN | 是否启用 | +| created_at | DATETIME | 创建时间 | +| updated_at | DATETIME | 更新时间 | + +### role_languages - 角色语言表 + +存储角色的多语言配置。 + +| 字段名 | 类型 | 描述 | +|--------------------|---------------|---------------------------| +| id | INT | 主键，自增 | +| role_id | INT | 关联的角色ID，外键 | +| language_code | VARCHAR(10) | 语言代码 | +| name | VARCHAR(128) | 该语言下的角色名称 | +| content | TEXT | 该语言下的系统提示内容 | +| asr_provider | VARCHAR(64) | 该语言的ASR服务提供商 | +| llm_provider | VARCHAR(64) | 该语言的LLM服务提供商 | +| tts_provider | VARCHAR(64) | 该语言的TTS服务提供商 | +| volcano_voice_type | VARCHAR(64) | 该语言的Volcano音色类型 | +| tencent_voice_type | VARCHAR(64) | 该语言的腾讯音色类型 | +| aliyun_voice_name | VARCHAR(64) | 该语言的阿里云音色名称 | +| url | VARCHAR(255) | 该语言的资源URL路径 | +| created_at | DATETIME | 创建时间 | +| updated_at | DATETIME | 更新时间 | + +### device_auth - 设备认证表 + +存储设备认证信息，用于验证设备身份。 + +| 字段名 | 类型 | 描述 | +|--------------------|---------------|---------------------------| +| id | INT | 主键，自增 | +| device_id | VARCHAR(64) | 设备ID，唯一索引 | +| serial_number | VARCHAR(64) | 设备序列号 | +| is_active | BOOLEAN | 是否活跃 | +| created_at | DATETIME | 创建时间 | +| updated_at | DATETIME | 更新时间 | + +## 常用查询示例 + +### 设备配置查询 + +#### 查询所有设备配置 + +```sql +SELECT * FROM device_configs; +``` + +#### 查询特定设备的配置 + +```sql +SELECT * FROM device_configs WHERE device_id = '设备ID'; +``` + +#### 查询使用特定角色的设备 + +```sql +SELECT * FROM device_configs WHERE selected_role_key = '角色键名'; +``` + +#### 查询使用特定语言的设备 + +```sql +SELECT * FROM device_configs WHERE preferred_language = '语言代码'; +``` + +#### 查询最近更新的设备配置 + +```sql +SELECT * FROM device_configs +ORDER BY last_update_time DESC +LIMIT 10; +``` + +### 对话历史查询 + +#### 查询特定设备的所有对话历史 + +```sql +SELECT * FROM conversation_histories WHERE device_id = '设备ID'; +``` + +#### 查询特定设备和角色的对话历史 + +```sql +SELECT * FROM conversation_histories +WHERE device_id = '设备ID' AND role_key = '角色键名'; +``` + +#### 查询最近交互的对话历史 + +```sql +SELECT * FROM conversation_histories +ORDER BY last_interaction_time DESC +LIMIT 10; +``` + +#### 查询一段时间内的对话历史 + +```sql +SELECT * FROM conversation_histories +WHERE last_interaction_time BETWEEN UNIX_TIMESTAMP('2023-01-01') AND UNIX_TIMESTAMP('2023-12-31') +ORDER BY last_interaction_time DESC; +``` + +### 对话消息查询 + +#### 查询特定对话的所有消息 + +```sql +SELECT * FROM conversation_messages +WHERE conversation_id = 对话ID +ORDER BY timestamp ASC; +``` + +#### 查询特定对话的用户消息 + +```sql +SELECT * FROM conversation_messages +WHERE conversation_id = 对话ID AND is_user = TRUE +ORDER BY timestamp ASC; +``` + +#### 查询特定对话的助手回复 + +```sql +SELECT * FROM conversation_messages +WHERE conversation_id = 对话ID AND is_user = FALSE +ORDER BY timestamp ASC; +``` + +#### 查询包含特定关键词的消息 + +```sql +SELECT * FROM conversation_messages +WHERE content LIKE '%关键词%' +ORDER BY timestamp DESC; +``` + +#### 查询最近的消息记录 + +```sql +SELECT cm.*, ch.device_id, ch.role_key +FROM conversation_messages cm +JOIN conversation_histories ch ON cm.conversation_id = ch.id +ORDER BY cm.timestamp DESC +LIMIT 20; +``` + +### 角色表查询 + +#### 查询所有启用的角色 + +```sql +SELECT * FROM roles WHERE enabled = TRUE; +``` + +#### 查询特定角色的详细信息 + +```sql +SELECT * FROM roles WHERE role_key = '角色键名'; +``` + +#### 查询使用特定服务提供商的角色 + +```sql +SELECT * FROM roles WHERE llm_provider = 'Volcano'; +-- 或 +SELECT * FROM roles WHERE tts_provider = 'Tencent'; +-- 或 +SELECT * FROM roles WHERE asr_provider = 'Aliyun'; +``` + +#### 查询角色的同音词列表 + +```sql +SELECT role_key, homophones FROM roles WHERE homophones IS NOT NULL; +``` + +#### 查询启用竞争模式的角色 + +```sql +SELECT * FROM roles WHERE competitive_llm_mode = TRUE; +``` + +### 角色语言表查询 + +#### 查询特定角色的所有语言配置 + +```sql +SELECT rl.* FROM role_languages rl +JOIN roles r ON rl.role_id = r.id +WHERE r.role_key = '角色键名'; +``` + +#### 查询特定语言的所有角色配置 + +```sql +SELECT r.role_key, r.name, rl.* FROM role_languages rl +JOIN roles r ON rl.role_id = r.id +WHERE rl.language_code = '语言代码'; +``` + +#### 查询特定角色特定语言的配置 + +```sql +SELECT rl.* FROM role_languages rl +JOIN roles r ON rl.role_id = r.id +WHERE r.role_key = '角色键名' AND rl.language_code = '语言代码'; +``` + +#### 查询支持多语言的角色列表 + +```sql +SELECT r.role_key, r.name, COUNT(rl.id) as language_count +FROM roles r +JOIN role_languages rl ON r.id = rl.role_id +GROUP BY r.id +HAVING language_count > 1 +ORDER BY language_count DESC; +``` + +### 设备认证查询 + +#### 查询所有已认证设备 + +```sql +SELECT * FROM device_auth WHERE is_active = TRUE; +``` + +#### 查询特定设备的认证信息 + +```sql +SELECT * FROM device_auth WHERE device_id = '设备ID'; +``` + +#### 查询最近添加的设备认证 + +```sql +SELECT * FROM device_auth +ORDER BY created_at DESC +LIMIT 10; +``` + +### 多表联合查询 + +#### 查询特定设备的所有消息及对应的角色 + +```sql +SELECT cm.*, ch.role_key +FROM conversation_messages cm +JOIN conversation_histories ch ON cm.conversation_id = ch.id +WHERE ch.device_id = '设备ID' +ORDER BY cm.timestamp ASC; +``` + +#### 查询特定设备使用特定角色的所有对话内容 + +```sql +SELECT cm.*, ch.role_key, ch.last_interaction_time +FROM conversation_messages cm +JOIN conversation_histories ch ON cm.conversation_id = ch.id +WHERE ch.device_id = '设备ID' AND ch.role_key = '角色键名' +ORDER BY cm.timestamp ASC; +``` + +#### 查询所有设备的最新一条消息 + +```sql +SELECT ch.device_id, ch.role_key, cm.content, cm.timestamp +FROM conversation_histories ch +JOIN conversation_messages cm ON ch.id = cm.conversation_id +WHERE cm.id = ( + SELECT MAX(cm2.id) + FROM conversation_messages cm2 + WHERE cm2.conversation_id = ch.id +); +``` + +#### 查询设备当前使用的角色详细信息 + +```sql +SELECT dc.device_id, r.* +FROM device_configs dc +JOIN roles r ON dc.selected_role_key = r.role_key +WHERE dc.device_id = '设备ID'; +``` + +#### 查询设备对话中使用的所有角色 + +```sql +SELECT DISTINCT ch.device_id, ch.role_key, r.name, r.description +FROM conversation_histories ch +JOIN roles r ON ch.role_key = r.role_key +WHERE ch.device_id = '设备ID'; +``` + +#### 查询设备认证状态与当前配置信息 + +```sql +SELECT da.device_id, da.serial_number, da.is_active, + dc.selected_role_key, dc.preferred_language, dc.last_update_time +FROM device_auth da +LEFT JOIN device_configs dc ON da.device_id = dc.device_id +ORDER BY da.created_at DESC; +``` + +### 高级统计查询 + +#### 统计每个角色的使用情况 + +```sql +SELECT selected_role_key, COUNT(*) as device_count +FROM device_configs +GROUP BY selected_role_key +ORDER BY device_count DESC; +``` + +#### 统计每种语言的使用情况 + +```sql +SELECT preferred_language, COUNT(*) as language_count +FROM device_configs +WHERE preferred_language IS NOT NULL +GROUP BY preferred_language +ORDER BY language_count DESC; +``` + +#### 统计每个设备的对话数量 + +```sql +SELECT device_id, COUNT(*) as conversation_count +FROM conversation_histories +GROUP BY device_id +ORDER BY conversation_count DESC; +``` + +#### 统计每个角色的对话数量 + +```sql +SELECT role_key, COUNT(*) as conversation_count +FROM conversation_histories +GROUP BY role_key +ORDER BY conversation_count DESC; +``` + +#### 统计消息量最多的对话 + +```sql +SELECT ch.id, ch.device_id, ch.role_key, COUNT(cm.id) as message_count +FROM conversation_histories ch +JOIN conversation_messages cm ON ch.id = cm.conversation_id +GROUP BY ch.id, ch.device_id, ch.role_key +ORDER BY message_count DESC +LIMIT 10; +``` + +#### 统计最活跃的设备 + +```sql +SELECT device_id, COUNT(DISTINCT role_key) as roles_used, + COUNT(*) as total_conversations, + MAX(last_interaction_time) as last_active +FROM conversation_histories +GROUP BY device_id +ORDER BY total_conversations DESC +LIMIT 10; +``` + +#### 统计各语言下的角色数量 + +```sql +SELECT rl.language_code, COUNT(DISTINCT r.id) as role_count +FROM role_languages rl +JOIN roles r ON rl.role_id = r.id +GROUP BY rl.language_code +ORDER BY role_count DESC; +``` + +#### 统计近期设备活跃度 + +```sql +SELECT + DATE(FROM_UNIXTIME(last_interaction_time)) as date, + COUNT(DISTINCT device_id) as active_devices +FROM conversation_histories +WHERE last_interaction_time > UNIX_TIMESTAMP(DATE_SUB(NOW(), INTERVAL 30 DAY)) +GROUP BY date +ORDER BY date DESC; +``` + +#### 统计各TTS提供商的使用情况 + +```sql +SELECT + CASE + WHEN volcano_voice_type IS NOT NULL THEN 'Volcano' + WHEN tencent_voice_type IS NOT NULL THEN 'Tencent' + WHEN aliyun_voice_name IS NOT NULL THEN 'Aliyun' + ELSE 'Unknown' + END as tts_provider, + COUNT(*) as role_count +FROM roles +GROUP BY tts_provider +ORDER BY role_count DESC; +``` + +#### 统计设备认证与使用情况 + +```sql +SELECT + COUNT(da.id) as total_devices, + SUM(CASE WHEN da.is_active = TRUE THEN 1 ELSE 0 END) as active_devices, + SUM(CASE WHEN dc.id IS NOT NULL THEN 1 ELSE 0 END) as configured_devices, + SUM(CASE WHEN ch.device_id IS NOT NULL THEN 1 ELSE 0 END) as devices_with_conversations +FROM device_auth da +LEFT JOIN device_configs dc ON da.device_id = dc.device_id +LEFT JOIN (SELECT DISTINCT device_id FROM conversation_histories) ch ON da.device_id = ch.device_id; +``` + +## 数据分析查询 + +### 对话内容分析 + +#### 查询用户问题平均长度 + +```sql +SELECT ch.role_key, AVG(LENGTH(cm.content)) as avg_length +FROM conversation_messages cm +JOIN conversation_histories ch ON cm.conversation_id = ch.id +WHERE cm.is_user = TRUE +GROUP BY ch.role_key; +``` + +#### 查询助手回答平均长度 + +```sql +SELECT ch.role_key, AVG(LENGTH(cm.content)) as avg_length +FROM conversation_messages cm +JOIN conversation_histories ch ON cm.conversation_id = ch.id +WHERE cm.is_user = FALSE +GROUP BY ch.role_key; +``` + +#### 查询用户消息长度分布 + +```sql +SELECT + CASE + WHEN LENGTH(content) < 50 THEN '极短 (<50)' + WHEN LENGTH(content) BETWEEN 50 AND 100 THEN '短 (50-100)' + WHEN LENGTH(content) BETWEEN 101 AND 200 THEN '中 (101-200)' + WHEN LENGTH(content) BETWEEN 201 AND 500 THEN '长 (201-500)' + ELSE '极长 (>500)' + END as message_length_category, + COUNT(*) as message_count +FROM conversation_messages +WHERE is_user = TRUE +GROUP BY message_length_category +ORDER BY + CASE message_length_category + WHEN '极短 (<50)' THEN 1 + WHEN '短 (50-100)' THEN 2 + WHEN '中 (101-200)' THEN 3 + WHEN '长 (201-500)' THEN 4 + ELSE 5 + END; +``` + +#### 查询不同角色对话完成时间统计 + +```sql +SELECT ch.role_key, + COUNT(*) as conversation_count, + AVG(MAX(cm.timestamp) - MIN(cm.timestamp)) as avg_duration +FROM conversation_histories ch +JOIN conversation_messages cm ON ch.id = cm.conversation_id +GROUP BY ch.role_key +HAVING conversation_count > 5 +ORDER BY avg_duration DESC; +``` + +#### 查询不同语言的平均对话长度 + +```sql +SELECT dc.preferred_language, + COUNT(DISTINCT ch.id) as conversation_count, + AVG(cm_count.msg_count) as avg_messages_per_conversation +FROM device_configs dc +JOIN conversation_histories ch ON dc.device_id = ch.device_id +JOIN ( + SELECT conversation_id, COUNT(*) as msg_count + FROM conversation_messages + GROUP BY conversation_id +) cm_count ON ch.id = cm_count.conversation_id +WHERE dc.preferred_language IS NOT NULL +GROUP BY dc.preferred_language +HAVING conversation_count > 3 +ORDER BY avg_messages_per_conversation DESC; +``` + +#### 查询包含特定关键词的对话 + +```sql +SELECT ch.id, ch.device_id, ch.role_key, + MAX(ch.last_interaction_time) as conversation_time, + GROUP_CONCAT(DISTINCT CASE WHEN cm.is_user = TRUE THEN cm.content ELSE NULL END SEPARATOR '\n') as user_messages, + GROUP_CONCAT(DISTINCT CASE WHEN cm.is_user = FALSE THEN cm.content ELSE NULL END SEPARATOR '\n') as assistant_messages +FROM conversation_histories ch +JOIN conversation_messages cm ON ch.id = cm.conversation_id +WHERE cm.content LIKE '%关键词%' +GROUP BY ch.id, ch.device_id, ch.role_key +ORDER BY conversation_time DESC; +``` + +### 用户行为分析 + +#### 查询用户平均会话时长趋势 + +```sql +SELECT + DATE(FROM_UNIXTIME(ch.last_interaction_time)) as date, + AVG( + MAX(cm.timestamp) - MIN(cm.timestamp) + ) as avg_session_duration +FROM conversation_histories ch +JOIN conversation_messages cm ON ch.id = cm.conversation_id +WHERE ch.last_interaction_time > UNIX_TIMESTAMP(DATE_SUB(NOW(), INTERVAL 30 DAY)) +GROUP BY date +ORDER BY date; +``` + +#### 查询每日活跃用户数和对话量 + +```sql +SELECT + DATE(FROM_UNIXTIME(last_interaction_time)) as date, + COUNT(DISTINCT device_id) as active_users, + COUNT(*) as conversations +FROM conversation_histories +WHERE last_interaction_time > UNIX_TIMESTAMP(DATE_SUB(NOW(), INTERVAL 30 DAY)) +GROUP BY date +ORDER BY date DESC; +``` + +#### 查询设备的首次使用时间和最后使用时间 + +```sql +SELECT + device_id, + MIN(created_at) as first_use, + FROM_UNIXTIME(MAX(last_interaction_time)) as last_use, + DATEDIFF(FROM_UNIXTIME(MAX(last_interaction_time)), MIN(created_at)) as days_between +FROM conversation_histories +GROUP BY device_id +ORDER BY days_between DESC; +``` + +#### 查询用户连续活跃天数 + +```sql +WITH daily_activity AS ( + SELECT + device_id, + DATE(FROM_UNIXTIME(last_interaction_time)) as activity_date + FROM conversation_histories + GROUP BY device_id, activity_date +) +SELECT + device_id, + COUNT(*) as active_days, + MIN(activity_date) as first_active_day, + MAX(activity_date) as last_active_day, + DATEDIFF(MAX(activity_date), MIN(activity_date)) + 1 as day_span, + COUNT(*) / (DATEDIFF(MAX(activity_date), MIN(activity_date)) + 1) as activity_ratio +FROM daily_activity +GROUP BY device_id +HAVING day_span > 7 +ORDER BY active_days DESC, activity_ratio DESC; +``` + +## 管理操作 + +### 备份数据库 + +```bash +# 在Docker环境外执行 +docker exec -it talkingq-mysql-1 mysqldump -u root -p talkingq > talkingq_backup.sql + +# 或使用本地MySQL客户端 +mysqldump -h localhost -u talkingq -p talkingq > talkingq_backup.sql +``` + +### 恢复数据库 + +```bash +# 在Docker环境外执行 +cat talkingq_backup.sql | docker exec -i talkingq-mysql-1 mysql -u root -p talkingq + +# 或使用本地MySQL客户端 +mysql -h localhost -u talkingq -p talkingq < talkingq_backup.sql +``` + +### 备份特定表 + +```bash +# 仅备份设备配置表 +docker exec -it talkingq-mysql-1 mysqldump -u root -p talkingq device_configs > device_configs_backup.sql + +# 仅备份角色表和角色语言表 +docker exec -it talkingq-mysql-1 mysqldump -u root -p talkingq roles role_languages > roles_backup.sql + +# 仅备份对话历史和消息表 +docker exec -it talkingq-mysql-1 mysqldump -u root -p talkingq conversation_histories conversation_messages > conversations_backup.sql + +# 备份设备认证表 +docker exec -it talkingq-mysql-1 mysqldump -u root -p talkingq device_auth > device_auth_backup.sql +``` + +### 删除过期记录 + +```sql +-- 删除长期未使用的设备配置 +DELETE FROM device_configs +WHERE last_update_time < UNIX_TIMESTAMP(DATE_SUB(NOW(), INTERVAL 90 DAY)); + +-- 删除3个月前的对话历史 +DELETE ch, cm +FROM conversation_histories ch +LEFT JOIN conversation_messages cm ON ch.id = cm.conversation_id +WHERE ch.last_interaction_time < UNIX_TIMESTAMP(DATE_SUB(NOW(), INTERVAL 3 MONTH)); + +-- 将90天内未被任何设备使用的角色标记为禁用 +UPDATE roles r +SET enabled = FALSE +WHERE r.role_key NOT IN ( + SELECT DISTINCT selected_role_key + FROM device_configs + WHERE last_update_time > UNIX_TIMESTAMP(DATE_SUB(NOW(), INTERVAL 90 DAY)) +) AND r.role_key NOT IN ( + SELECT DISTINCT role_key + FROM conversation_histories + WHERE last_interaction_time > UNIX_TIMESTAMP(DATE_SUB(NOW(), INTERVAL 90 DAY)) +); + +-- 清理已停用设备的认证记录 +UPDATE device_auth +SET is_active = FALSE +WHERE device_id NOT IN ( + SELECT DISTINCT device_id + FROM conversation_histories + WHERE last_interaction_time > UNIX_TIMESTAMP(DATE_SUB(NOW(), INTERVAL 180 DAY)) +); +``` + +### 角色管理 + +```sql +-- 添加新角色（基本信息） +INSERT INTO roles (role_key, name, description, content, default_language, enabled) +VALUES ('new_role', '新角色', '新角色的描述', '系统提示内容', 'zh', TRUE); + +-- 为角色添加语言配置 +INSERT INTO role_languages (role_id, language_code, name, content) +VALUES ( + (SELECT id FROM roles WHERE role_key = 'new_role'), + 'en', + 'New Role', + 'System prompt content in English' +); + +-- 复制角色 +INSERT INTO roles ( + role_key, name, description, content, default_language, + asr_provider, llm_provider, tts_provider, competitive_llm_mode, + volcano_model_id, volcano_voice_type, tencent_voice_type, aliyun_voice_name, + url, homophones, enabled +) +SELECT + CONCAT(role_key, '_copy'), CONCAT(name, ' (复制)'), description, content, default_language, + asr_provider, llm_provider, tts_provider, competitive_llm_mode, + volcano_model_id, volcano_voice_type, tencent_voice_type, aliyun_voice_name, + url, homophones, enabled +FROM roles +WHERE role_key = '要复制的角色键名'; +``` + +## 数据库优化 + +### 优化表 + +```sql +OPTIMIZE TABLE device_configs; +OPTIMIZE TABLE conversation_histories; +OPTIMIZE TABLE conversation_messages; +OPTIMIZE TABLE roles; +OPTIMIZE TABLE role_languages; +OPTIMIZE TABLE device_auth; +``` + +### 分析表 + +```sql +ANALYZE TABLE device_configs; +ANALYZE TABLE conversation_histories; +ANALYZE TABLE conversation_messages; +ANALYZE TABLE roles; +ANALYZE TABLE role_languages; +ANALYZE TABLE device_auth; +``` + +### 检查表 + +```sql +CHECK TABLE device_configs; +CHECK TABLE conversation_histories; +CHECK TABLE conversation_messages; +CHECK TABLE roles; +CHECK TABLE role_languages; +CHECK TABLE device_auth; +``` + +### 添加索引 + +```sql +-- 为对话历史表添加联合索引 +CREATE INDEX idx_device_role ON conversation_histories(device_id, role_key); + +-- 为角色语言表添加联合索引 +CREATE INDEX idx_role_language ON role_languages(role_id, language_code); + +-- 为消息表添加时间戳索引，优化时间范围查询 +CREATE INDEX idx_message_timestamp ON conversation_messages(timestamp); + +-- 为设备认证表添加索引 +CREATE INDEX idx_device_auth_active ON device_auth(is_active); +``` + +### 监控查询性能 + +```sql +-- 查看慢查询日志状态 +SHOW VARIABLES LIKE 'slow_query%'; + +-- 查看当前运行的查询 +SHOW PROCESSLIST; + +-- 终止执行时间过长的查询 +KILL QUERY 线程ID; +``` + +## 数据库维护脚本示例 + +以下是一个可以添加到cron任务的脚本示例，用于定期清理过期数据和优化表： + +```bash +#!/bin/bash +# 文件名: /home/ubuntu/TalkingQ_URL/scripts/db_maintenance.sh + +# 数据库连接参数 +DB_USER="talkingq" +DB_PASS="您的密码" +DB_NAME="talkingq" +DB_HOST="mysql" + +# 日志文件 +LOG_FILE="/home/ubuntu/TalkingQ_URL/logs/db_maintenance.log" + +echo "===== $(date) 开始数据库维护 =====" >> $LOG_FILE + +# 删除3个月前的对话历史 +echo "清理旧对话历史..." >> $LOG_FILE +docker exec -i talkingq-mysql-1 mysql -u$DB_USER -p$DB_PASS $DB_NAME << EOF +DELETE ch, cm +FROM conversation_histories ch +LEFT JOIN conversation_messages cm ON ch.id = cm.conversation_id +WHERE ch.last_interaction_time < UNIX_TIMESTAMP(DATE_SUB(NOW(), INTERVAL 3 MONTH)); +EOF + +# 优化表 +echo "优化数据库表..." >> $LOG_FILE +docker exec -i talkingq-mysql-1 mysql -u$DB_USER -p$DB_PASS $DB_NAME << EOF +OPTIMIZE TABLE conversation_histories, conversation_messages, device_configs; +EOF + +echo "===== $(date) 数据库维护完成 =====" >> $LOG_FILE +``` diff --git a/talkingq-url/requirements.txt b/talkingq-url/requirements.txt new file mode 100644 index 0000000..7241274 --- /dev/null +++ b/talkingq-url/requirements.txt @@ -0,0 +1,13 @@ +aiofiles==24.1.0 +aiohttp==3.11.13 +fastapi==0.115.11 +pycld2==0.41 +pydantic-settings==2.8.1 +python-multipart==0.0.20 +PyYAML==6.0.2 +uvicorn[standard]==0.34.0 +dashscope==1.22.2 +aiomysql==0.2.0 +sqlalchemy[asyncio]==2.0.40 +loguru==0.7.3 + diff --git a/talkingq-url/scripts/backup_db.py b/talkingq-url/scripts/backup_db.py new file mode 100644 index 0000000..cbe9e6b --- /dev/null +++ b/talkingq-url/scripts/backup_db.py @@ -0,0 +1,105 @@ +import os +import sys +import asyncio +import aiomysql +import subprocess +import datetime +import gzip +import shutil +from pathlib import Path + +sys.path.insert(0, str(Path(__file__).parent.parent)) + +from config import settings + +BACKUP_DIR = os.path.join(Path(__file__).parent.parent, "backups") +os.makedirs(BACKUP_DIR, exist_ok=True) + +async def backup_database(name=None, compress=True): + """备份MySQL数据库""" + + timestamp = datetime.datetime.now().strftime("%Y%m%d_%H%M%S") + if name: + backup_name = f"{name}_{timestamp}" + else: + backup_name = f"backup_{timestamp}" + + output_file = os.path.join(BACKUP_DIR, f"{backup_name}.sql") + + cmd = [ + "mysqldump", + "-h", settings.db_host, + "-P", str(settings.db_port), + "-u", settings.db_user, + f"-p{settings.db_password}", + "--single-transaction", + "--routines", + "--triggers", + "--events", + settings.db_name + ] + + try: + print(f"开始备份数据库...") + with open(output_file, 'w') as f: + process = subprocess.Popen( + cmd, + stdout=f, + stderr=subprocess.PIPE + ) + _, stderr = process.communicate() + + if process.returncode != 0: + print(f"备份失败: {stderr.decode()}") + if os.path.exists(output_file): + os.remove(output_file) + return False + + if compress: + compressed_file = f"{output_file}.gz" + print(f"压缩备份文件...") + with open(output_file, 'rb') as f_in: + with gzip.open(compressed_file, 'wb') as f_out: + shutil.copyfileobj(f_in, f_out) + + os.remove(output_file) # 删除未压缩的SQL文件 + output_file = compressed_file + + file_size = os.path.getsize(output_file) / (1024 * 1024) # 转换为MB + print(f"备份完成: {output_file} ({file_size:.2f} MB)") + return output_file + + except Exception as e: + print(f"备份过程中出错: {e}") + if os.path.exists(output_file): + os.remove(output_file) + return False + +async def list_backups(): + """列出所有可用的备份文件""" + if not os.path.exists(BACKUP_DIR): + return [] + + backups = [f for f in os.listdir(BACKUP_DIR) if f.endswith('.sql') or f.endswith('.sql.gz')] + backups.sort(reverse=True) # 按文件名排序，最新的在前面 + + return backups + +async def cleanup_old_backups(keep=10): + """清理旧备份，仅保留指定数量的最新备份""" + backups = await list_backups() + + if len(backups) <= keep: + print(f"当前共有 {len(backups)} 个备份，未超过保留数量 {keep}，不需要清理") + return + + to_delete = backups[keep:] + print(f"将删除 {len(to_delete)} 个旧备份，保留 {keep} 个最新备份") + + for backup in to_delete: + try: + file_path = os.path.join(BACKUP_DIR, backup) + os.remove(file_path) + print(f"已删除: {backup}") + except Exception as e: + print(f"删除备份 {backup} 时出错: {e}") diff --git a/talkingq-url/scripts/cleanup_db.py b/talkingq-url/scripts/cleanup_db.py new file mode 100644 index 0000000..9aae1c8 --- /dev/null +++ b/talkingq-url/scripts/cleanup_db.py @@ -0,0 +1,385 @@ +import os +import sys +import asyncio +from pathlib import Path +import aiomysql +import datetime +import inspect + +sys.path.insert(0, str(Path(__file__).parent.parent)) + +from config import settings +from database.models import Base +from sqlalchemy import inspect as sqlalchemy_inspect +from utils.logger import session_logger + +async def get_mysql_connection(): + """建立MySQL数据库连接""" + try: + conn = await aiomysql.connect( + host=settings.db_host, + port=settings.db_port, + user=settings.db_user, + password=settings.db_password, + db=settings.db_name, + autocommit=False + ) + return conn + except Exception as e: + print(f"数据库连接失败: {e}") + return None + +async def get_all_tables(conn): + """获取数据库中所有表""" + async with conn.cursor() as cursor: + await cursor.execute("SHOW TABLES") + tables = await cursor.fetchall() + return [table[0] for table in tables] + +async def get_model_tables(): + """获取模型定义的所有表""" + return [table.__tablename__ for table in Base.__subclasses__()] + +async def get_table_columns(conn, table_name): + """获取表的所有列""" + async with conn.cursor() as cursor: + await cursor.execute(f"SHOW COLUMNS FROM `{table_name}`") + columns = await cursor.fetchall() + return [column[0] for column in columns] + +async def get_model_columns(model_class): + """获取模型定义的所有列""" + inspector = sqlalchemy_inspect(model_class) + return [column.key for column in inspector.mapper.column_attrs] + +async def identify_unused_tables(conn): + """识别未使用的表""" + db_tables = await get_all_tables(conn) + model_tables = await get_model_tables() + + unused_tables = [table for table in db_tables if table not in model_tables] + return unused_tables + +async def identify_unused_columns(conn): + """识别未使用的列""" + results = {} + model_dict = {model.__tablename__: model for model in Base.__subclasses__()} + + for table_name, model_class in model_dict.items(): + try: + db_columns = await get_table_columns(conn, table_name) + model_columns = await get_model_columns(model_class) + + unused_columns = [col for col in db_columns if col not in model_columns] + if unused_columns: + results[table_name] = unused_columns + except Exception as e: + print(f"检查表 {table_name} 列时出错: {e}") + + return results + +async def get_table_foreign_keys(conn, table_name): + """获取表的外键约束""" + async with conn.cursor() as cursor: + query = """ + SELECT + CONSTRAINT_NAME, + TABLE_NAME, + COLUMN_NAME, + REFERENCED_TABLE_NAME, + REFERENCED_COLUMN_NAME + FROM + INFORMATION_SCHEMA.KEY_COLUMN_USAGE + WHERE + REFERENCED_TABLE_NAME IS NOT NULL + AND TABLE_SCHEMA = %s + AND TABLE_NAME = %s + """ + await cursor.execute(query, (settings.db_name, table_name)) + foreign_keys = await cursor.fetchall() + return foreign_keys + +async def get_referenced_tables(conn, table_name): + """获取引用指定表的所有表""" + async with conn.cursor() as cursor: + query = """ + SELECT + TABLE_NAME, + CONSTRAINT_NAME + FROM + INFORMATION_SCHEMA.KEY_COLUMN_USAGE + WHERE + REFERENCED_TABLE_NAME = %s + AND TABLE_SCHEMA = %s + """ + await cursor.execute(query, (table_name, settings.db_name)) + references = await cursor.fetchall() + + if references: + print(f"表 {table_name} 被以下表引用:") + for ref in references: + print(f" - {ref[0]} (约束: {ref[1]})") + + return references + +async def disable_foreign_key_checks(conn): + """临时禁用外键约束检查""" + async with conn.cursor() as cursor: + await cursor.execute("SET FOREIGN_KEY_CHECKS = 0") + print("已临时禁用外键约束检查") + +async def enable_foreign_key_checks(conn): + """重新启用外键约束检查""" + async with conn.cursor() as cursor: + await cursor.execute("SET FOREIGN_KEY_CHECKS = 1") + print("已重新启用外键约束检查") + +async def sort_tables_for_deletion(conn, tables): + """对表进行排序，以确保先删除引用者，再删除被引用者""" + dependency_graph = {} + for table in tables: + references = await get_referenced_tables(conn, table) + dependency_graph[table] = [ref[0] for ref in references if ref[0] in tables] + + result = [] + visited = set() + temp_mark = set() + + def visit(node): + if node in temp_mark: + return + if node not in visited: + temp_mark.add(node) + for dependent in dependency_graph.get(node, []): + visit(dependent) + temp_mark.remove(node) + visited.add(node) + result.append(node) + + for table in tables: + if table not in visited: + visit(table) + + return result[::-1] + +async def drop_tables(conn, tables, dry_run=True): + """删除未使用的表""" + if not tables: + return True + + if not dry_run: + try: + print("禁用外键约束检查以确保安全删除所有表...") + await disable_foreign_key_checks(conn) + + for table in tables: + try: + async with conn.cursor() as cursor: + await cursor.execute(f"DROP TABLE IF EXISTS `{table}`") + print(f"已删除表: {table}") + except Exception as e: + print(f"删除表 {table} 时出错: {e}") + await conn.rollback() + await enable_foreign_key_checks(conn) + return False + + await conn.commit() + await enable_foreign_key_checks(conn) + return True + except Exception as e: + print(f"删除表时出错: {e}") + try: + await enable_foreign_key_checks(conn) + await conn.rollback() + except: + pass + return False + else: + for table in tables: + print(f"将删除表: {table}") + return True + +async def drop_columns(conn, column_dict, dry_run=True): + """删除未使用的列""" + for table, columns in column_dict.items(): + try: + for column in columns: + if dry_run: + print(f"将从表 {table} 删除列: {column}") + else: + async with conn.cursor() as cursor: + await cursor.execute(f"ALTER TABLE `{table}` DROP COLUMN `{column}`") + print(f"已从表 {table} 删除列: {column}") + except Exception as e: + print(f"从表 {table} 删除列时出错: {e}") + if not dry_run: + await conn.rollback() + return False + + if not dry_run: + await conn.commit() + return True + +async def cleanup_database(dry_run=True): + """清理数据库中未使用的表和列""" + conn = await get_mysql_connection() + if not conn: + return False + + try: + unused_tables = await identify_unused_tables(conn) + unused_columns = await identify_unused_columns(conn) + + print("\n=== 数据库清理报告 ===") + + if unused_tables: + print(f"\n发现 {len(unused_tables)} 个未使用的表:") + for table in unused_tables: + print(f" - {table}") + + if not dry_run: + success = await drop_tables(conn, unused_tables, dry_run=False) + if not success: + print("删除表失败，已回滚操作") + return False + else: + print("\n未发现任何未使用的表") + + if unused_columns: + total_columns = sum(len(cols) for cols in unused_columns.values()) + print(f"\n发现 {total_columns} 个未使用的列:") + for table, columns in unused_columns.items(): + for column in columns: + print(f" - {table}.{column}") + + if not dry_run: + success = await drop_columns(conn, unused_columns, dry_run=False) + if not success: + print("删除列失败，已回滚操作") + return False + else: + print("\n未发现任何未使用的列") + + if dry_run: + print("\n这是一次模拟运行，未执行任何实际更改。要执行更改，请使用 --execute 选项。") + else: + print("\n数据库清理成功完成！") + + return True + except Exception as e: + print(f"清理数据库时发生错误: {e}") + return False + finally: + conn.close() + +async def remove_orphaned_records(dry_run=True): + """删除孤立记录（有外键但对应的父记录不存在）""" + conn = await get_mysql_connection() + if not conn: + return False + + try: + async with conn.cursor() as cursor: + await cursor.execute(""" + SELECT COUNT(*) FROM conversation_messages + WHERE conversation_id NOT IN (SELECT id FROM conversation_histories) + """) + count = (await cursor.fetchone())[0] + + if count > 0: + print(f"\n发现 {count} 条孤立的会话消息记录") + if not dry_run: + await cursor.execute(""" + DELETE FROM conversation_messages + WHERE conversation_id NOT IN (SELECT id FROM conversation_histories) + """) + print(f"已删除 {count} 条孤立的会话消息记录") + else: + print("\n未发现孤立的会话消息记录") + + async with conn.cursor() as cursor: + await cursor.execute(""" + SELECT COUNT(*) FROM role_languages + WHERE role_id NOT IN (SELECT id FROM roles) + """) + count = (await cursor.fetchone())[0] + + if count > 0: + print(f"\n发现 {count} 条孤立的角色语言记录") + if not dry_run: + await cursor.execute(""" + DELETE FROM role_languages + WHERE role_id NOT IN (SELECT id FROM roles) + """) + print(f"已删除 {count} 条孤立的角色语言记录") + else: + print("\n未发现孤立的角色语言记录") + + if not dry_run: + await conn.commit() + + return True + except Exception as e: + print(f"清理孤立记录时发生错误: {e}") + if not dry_run: + await conn.rollback() + return False + finally: + conn.close() + +async def drop_specific_tables(conn, tables_to_drop, dry_run=True): + """删除指定的表，无论它们是否被使用""" + if dry_run: + print(f"\n将删除以下指定表:") + for table in tables_to_drop: + print(f" - {table}") + return True + + try: + await disable_foreign_key_checks(conn) + + for table in tables_to_drop: + try: + async with conn.cursor() as cursor: + await cursor.execute(f"DROP TABLE IF EXISTS `{table}`") + print(f"已删除表: {table}") + except Exception as e: + print(f"删除表 {table} 时出错: {e}") + + await conn.commit() + print(f"指定表已成功删除") + return True + except Exception as e: + await conn.rollback() + print(f"删除指定表时出错: {e}") + return False + finally: + await enable_foreign_key_checks(conn) + +if __name__ == "__main__": + import argparse + + parser = argparse.ArgumentParser(description="数据库清理工具") + parser.add_argument('--execute', action='store_true', help='执行实际更改（不使用此选项则只进行模拟运行）') + parser.add_argument('--drop-tables', type=str, help='指定要删除的表，用逗号分隔') + args = parser.parse_args() + + dry_run = not args.execute + + async def main(): + if args.drop_tables: + conn = await get_mysql_connection() + if not conn: + return False + + try: + tables_to_drop = [t.strip() for t in args.drop_tables.split(',')] + await drop_specific_tables(conn, tables_to_drop, dry_run) + finally: + conn.close() + else: + result = await cleanup_database(dry_run) + if result: + await remove_orphaned_records(dry_run) + + asyncio.run(main()) diff --git a/talkingq-url/scripts/db_management.py b/talkingq-url/scripts/db_management.py new file mode 100644 index 0000000..8f6c099 --- /dev/null +++ b/talkingq-url/scripts/db_management.py @@ -0,0 +1,92 @@ +import os +import sys +import asyncio +import argparse +from pathlib import Path + +sys.path.insert(0, str(Path(__file__).parent.parent)) + +from scripts.backup_db import backup_database, list_backups, cleanup_old_backups, BACKUP_DIR +from scripts.restore_db import restore_database, interactive_restore +from scripts.cleanup_db import cleanup_database, remove_orphaned_records + +async def main(): + parser = argparse.ArgumentParser(description='数据库备份和恢复工具') + subparsers = parser.add_subparsers(dest='command', help='子命令') + + backup_parser = subparsers.add_parser('backup', help='备份数据库') + backup_parser.add_argument('--name', type=str, help='备份名称，不含扩展名') + backup_parser.add_argument('--no-compress', action='store_true', help='不压缩备份文件') + backup_parser.add_argument('--keep', type=int, default=10, help='保留的备份数量') + + restore_parser = subparsers.add_parser('restore', help='恢复数据库') + restore_parser.add_argument('--file', type=str, help='要恢复的备份文件') + + list_parser = subparsers.add_parser('list', help='列出所有备份') + + cleanup_parser = subparsers.add_parser('cleanup', help='清理旧备份') + cleanup_parser.add_argument('--keep', type=int, default=10, help='保留的备份数量') + + clean_db_parser = subparsers.add_parser('clean-db', help='清理数据库中未使用的表和字段') + clean_db_parser.add_argument('--execute', action='store_true', help='执行实际清理（不使用此选项则只进行模拟运行）') + clean_db_parser.add_argument('--orphaned', action='store_true', help='仅清理孤立记录') + clean_db_parser.add_argument('--schema', action='store_true', help='仅清理未使用的表和字段') + clean_db_parser.add_argument('--backup-first', action='store_true', help='清理前先进行备份') + clean_db_parser.add_argument('--drop-tables', type=str, help='指定要删除的表，用逗号分隔') + + args = parser.parse_args() + + if args.command == 'backup': + await backup_database(args.name, not args.no_compress if hasattr(args, 'no_compress') else True) + await cleanup_old_backups(args.keep if hasattr(args, 'keep') else 10) + elif args.command == 'restore': + if args.file: + await restore_database(args.file) + else: + await interactive_restore() + elif args.command == 'list': + backups = await list_backups() + if backups: + print("\n可用的备份文件:") + for backup in backups: + file_path = os.path.join(BACKUP_DIR, backup) + size = os.path.getsize(file_path) / (1024 * 1024) # 转换为MB + print(f"{backup} ({size:.2f} MB)") + else: + print("没有找到可用的备份") + elif args.command == 'cleanup': + keep = args.keep if hasattr(args, 'keep') else 10 + await cleanup_old_backups(keep) + elif args.command == 'clean-db': + dry_run = not args.execute + + if args.backup_first: + print("清理前进行数据库备份...") + await backup_database(f"pre_cleanup_{int(asyncio.get_event_loop().time())}", True) + + if args.drop_tables: + from scripts.cleanup_db import get_mysql_connection, drop_specific_tables + conn = await get_mysql_connection() + if conn: + try: + tables_to_drop = [t.strip() for t in args.drop_tables.split(',')] + await drop_specific_tables(conn, tables_to_drop, dry_run) + finally: + conn.close() + else: + if not args.orphaned: + print("\n分析数据库架构...") + await cleanup_database(dry_run) + + if not args.schema: + print("\n检查孤立记录...") + await remove_orphaned_records(dry_run) + + if dry_run: + print("\n这是一次模拟运行，没有执行任何实际更改。") + print("要执行实际的清理操作，请使用 --execute 选项。") + else: + parser.print_help() + +if __name__ == "__main__": + asyncio.run(main()) diff --git a/talkingq-url/scripts/devices.csv b/talkingq-url/scripts/devices.csv new file mode 100644 index 0000000..c12a614 --- /dev/null +++ b/talkingq-url/scripts/devices.csv @@ -0,0 +1,109 @@ +device_id,serial_number +talkingQ_B0EC1F2,3b4c5d6e7f +talkingQ_94B7164,3b4c5d6e7f +talkingQ_4C2B7F9,0a1b2c3d4e +talkingQ_f91D6b8,5a9b8d3f7c +talkingQ_1a2b7C3,f6a1c5e9d8 +talkingQ_7D3c1F4,8b9c7a2f1b +talkingQ_2C8f3D6,4e9b0d7c1a +talkingQ_0b6F4D3,a3b9c5d7e8 +talkingQ_9A7b4d1,b2f3a9d6c7 +talkingQ_5B8a1C9,f0d6c3b2a1 +talkingQ_7F3d2B6,c5d1e8f9b6 +talkingQ_8D1f5c7,6b9d0a3c5f +talkingQ_4F1C3b9,a7e9f2c4d0 +talkingQ_2d4b6F7,8c5a1b3d9e +talkingQ_7A9b4F2,e3c6d9f8b7 +talkingQ_1f6D4B2,d9b1a7c5e0 +talkingQ_5b1f7D9,3a4c2e8b5f +talkingQ_8B4C3d9,f7a6c1d4e9 +talkingQ_0F7a6B4,2c9d5f3b7a +talkingQ_6c5d2F8,a1b9c7d3f5 +talkingQ_3B1f7d5,9c2a4f8e6d +talkingQ_2A7b9d4,3f6b1e9c8a +talkingQ_9c3d8B7,5f1a2d6e4c +talkingQ_7f4B3C1,9a5e7b3f0d +talkingQ_B4B7164,3b4c5d6e7f +talkingQ_E05AF3A,3b4c5d6e7f +talkingQ_A8B7164,3b4c5d6e7f +talkingQ_4d7f2B9,8c1b9e3f7a +talkingQ_C4B7164,3b4c5d6e7f +talkingQ_C0B7164,3b4c5d6e7f +talkingQ_F4B7164,3b4c5d6e7f +talkingQ_D0B7164,3b4c5d6e7f +talkingQ_F0B7164,3b4c5d6e7f +talkingQ_ACB7164,3b4c5d6e7f +talkingQ_E8B7164,3b4c5d6e7f +talkingQ_6B1F9a3,5d7c2e0a9f +talkingQ_3A7d5F4,c9b0e6d2a4 +talkingQ_5B9d2A7,1e4f3c6a9b +talkingQ_2d3F1B6,c4e9a5b7d2 +talkingQ_9A6b5d2,7c1a4f9e0b +talkingQ_1A2B3C4,5d6e7f8a9b +talkingQ_5D6E7F8,1a2b3c4d5e +talkingQ_7G8H9I0,6b7c8d9e0f +talkingQ_2J3K4L5,8f9g0h1j2k +talkingQ_9M0N1O2,3l4m5n6o7p +talkingQ_3P4Q5R6,9q0r1s2t3u +talkingQ_6S7T8U9,4v5w6x7y8z +talkingQ_4V5W6X7,0a1b2c3d4e +talkingQ_8Y9Z0A1,5f6g7h8i9j +talkingQ_5B6C7D8,0k1l2m3n4o +talkingQ_3E4F5G6,7p8q9r0s1t +talkingQ_7H8I9J0,2u3v4w5x6y +talkingQ_1K2L3M4,8z9a0b1c2d +talkingQ_4N5O6P7,3e4f5g6h7i +talkingQ_6Q7R8S9,9j0k1l2m3n +talkingQ_2T3U4V5,4o5p6q7r8s +talkingQ_9W0X1Y2,0t1u2v3w4x +talkingQ_5Z6A7B8,5y6z7a8b9c +talkingQ_3C4D5E6,0d1e2f3g4h +talkingQ_7F8G9H0,6i7j8k9l0m +talkingQ_1I2J3K4,1n2o3p4q5r +talkingQ_4L5M6N7,7s8t9u0v1w +talkingQ_6O7P8Q9,2x3y4z5a6b +talkingQ_2R3S4T5,8c9d0e1f2g +talkingQ_9U0V1W2,3h4i5j6k7l +talkingQ_5X6Y7Z8,9m0n1o2p3q +talkingQ_3A4B5C6,4r5s6t7u8v +talkingQ_7D8E9F0,0w1x2y3z4a +talkingQ_1G2H3I4,5b6c7d8e9f +talkingQ_4J5K6L7,0g1h2i3j4k +talkingQ_6M7N8O9,6l7m8n9o0p +talkingQ_2P3Q4R5,1q2r3s4t5u +talkingQ_9S0T1U2,7v8w9x0y1z +talkingQ_5V6W7X8,2a3b4c5d6e +talkingQ_3Y4Z5A6,8f9g0h1i2j +talkingQ_7B8C9D0,3k4l5m6n7o +talkingQ_1E2F3G4,9p0q1r2s3t +talkingQ_4H5I6J7,4u5v6w7x8y +talkingQ_6K7L8M9,0z1a2b3c4d +talkingQ_2N3O4P5,5e6f7g8h9i +talkingQ_9Q0R1S2,0j1k2l3m4n +talkingQ_5T6U7V8,6o7p8q9r0s +talkingQ_3W4X5Y6,1t2u3v4w5x +talkingQ_7Z8A9B0,7y8z9a0b1c +talkingQ_1C2D3E4,2d3e4f5g6h +talkingQ_4F5G6H7,8i9j0k1l2m +talkingQ_6I7J8K9,3n4o5p6q7r +talkingQ_2L3M4N5,9s0t1u2v3w +talkingQ_9O0P1Q2,4x5y6z7a8b +talkingQ_5R6S7T8,0c1d2e3f4g +talkingQ_8B9D3E4,a1b2c3d4e5 +talkingQ_2A4F7C8,f6e5d4c3b2 +talkingQ_1E3F5G7,d8c7b9a0f1 +talkingQ_4G7H9K1,c2b3a4d5e6 +talkingQ_5F2D4A8,b7c8d9e1f0 +talkingQ_7J3L5O6,e1f2d3c4b5 +talkingQ_3C8B7E4,d9a6f5b2c3 +talkingQ_9A1D3F5,e4b6c7d8a9 +talkingQ_6B5C2D4,a7e8f9b0c1 +talkingQ_4G8H3J7,d2f5c9b4a0 +talkingQ_D485583,3b4c5d6e7f +talkingQ_8C85583,3b4c5d6e7f +talkingQ_4486583,3b4c5d6e7f +talkingQ_68A81F2,3b4c5d6e7f +talkingQ_70A81F2,3b4c5d6e7f +talkingQ_7885583,3b4c5d6e7f +talkingQ_7886583,3b4c5d6e7f +talkingQ_C8A71F2,3b4c5d6e7f diff --git a/talkingq-url/scripts/restore_db.py b/talkingq-url/scripts/restore_db.py new file mode 100644 index 0000000..a40b001 --- /dev/null +++ b/talkingq-url/scripts/restore_db.py @@ -0,0 +1,112 @@ +import os +import sys +import asyncio +import subprocess +import gzip +import tempfile +import shutil +from pathlib import Path + +sys.path.insert(0, str(Path(__file__).parent.parent)) + +from config import settings +from scripts.backup_db import list_backups, BACKUP_DIR + +async def restore_database(backup_file): + """从备份文件恢复数据库""" + file_path = backup_file + if not os.path.isabs(backup_file): + file_path = os.path.join(BACKUP_DIR, backup_file) + + if not os.path.exists(file_path): + print(f"错误: 备份文件 '{file_path}' 不存在") + return False + + confirm = input(f"警告: 这将覆盖当前数据库的所有数据。确定要继续吗? (y/N): ") + if confirm.lower() != 'y': + print("操作已取消") + return False + + temp_file = None + sql_file = file_path + + try: + if file_path.endswith('.gz'): + print("解压备份文件...") + temp_file = tempfile.NamedTemporaryFile(delete=False, suffix='.sql') + temp_file.close() + + with gzip.open(file_path, 'rb') as f_in: + with open(temp_file.name, 'wb') as f_out: + shutil.copyfileobj(f_in, f_out) + + sql_file = temp_file.name + + cmd = [ + "mysql", + "-h", settings.db_host, + "-P", str(settings.db_port), + "-u", settings.db_user, + f"-p{settings.db_password}", + settings.db_name + ] + + print(f"开始恢复数据库...") + with open(sql_file, 'r') as f: + process = subprocess.Popen( + cmd, + stdin=f, + stdout=subprocess.PIPE, + stderr=subprocess.PIPE + ) + stdout, stderr = process.communicate() + + if process.returncode != 0: + print(f"恢复失败: {stderr.decode()}") + return False + + print("数据库恢复成功!") + return True + + except Exception as e: + print(f"恢复过程中出错: {e}") + return False + + finally: + if temp_file and os.path.exists(temp_file.name): + os.unlink(temp_file.name) + +async def interactive_restore(): + """交互式选择要恢复的备份文件""" + backups = await list_backups() + + if not backups: + print("没有找到可用的备份文件") + return False + + print("\n可用的备份文件:") + for i, backup in enumerate(backups): + file_path = os.path.join(BACKUP_DIR, backup) + size = os.path.getsize(file_path) / (1024 * 1024) # 转换为MB + print(f"{i+1}. {backup} ({size:.2f} MB)") + + try: + choice = int(input("\n请选择要恢复的备份文件 (输入序号): ")) + if 1 <= choice <= len(backups): + selected_backup = backups[choice-1] + return await restore_database(selected_backup) + else: + print("无效的选择") + return False + except ValueError: + print("无效的输入") + return False + +async def main(): + if len(sys.argv) > 1: + await restore_database(sys.argv[1]) + else: + await interactive_restore() + +if __name__ == "__main__": + asyncio.run(main()) diff --git a/talkingq-url/services/__init__.py b/talkingq-url/services/__init__.py new file mode 100644 index 0000000..e69de29 diff --git a/talkingq-url/services/audio_sender.py b/talkingq-url/services/audio_sender.py new file mode 100644 index 0000000..6a65bdb --- /dev/null +++ b/talkingq-url/services/audio_sender.py @@ -0,0 +1,165 @@ +import asyncio +from typing import Optional +from fastapi import WebSocket +from utils.logger import session_logger +from services.interrupt_handler import interrupt_handler +from services.tts_error_manager import tts_error_manager +from config import settings + + +class AudioSender: + + + def __init__( + self, device_id: str, session_id: str, websocket: Optional[WebSocket] = None + ): + self.device_id = device_id + self.session_id = session_id + self.session_key = (device_id, session_id) + self.websocket = websocket + + def check_interruption(self): + + return interrupt_handler.is_interrupted(self.session_key) + + def is_websocket_connected(self): + + return ( + self.websocket is not None + and self.websocket.client_state.name == "CONNECTED" + and not getattr(self.websocket, "_closed", False) + ) + + async def send_start_notification(self): + + if not self.is_websocket_connected(): + session_logger.warning( + self.device_id, + self.session_id, + "Session websocket 不存在或会话已中断，无法发送 TTS_START", + ) + await interrupt_handler.set_interrupt_state(self.session_key, True) + return False + try: + await self.websocket.send_text("TTS_START") + session_logger.info( + self.device_id, self.session_id, "已发送 TTS_START 给客户端" + ) + await tts_error_manager.start_tts_session(self.device_id, self.session_id) + return True + except Exception as e: + session_logger.error( + self.device_id, + self.session_id, + f"发送 TTS_START 时发生异常：{e}", + exc_info=True, + ) + await interrupt_handler.set_interrupt_state(self.session_key, True) + return False + + async def send_audio_urls(self, url_queue: asyncio.Queue): + + try: + if not await self.send_start_notification(): + return + + first_url_sent = False + first_tts_submit_time = getattr(url_queue, "first_tts_submit_time", None) + + while True: + if self.check_interruption(): + session_logger.info( + self.device_id, + self.session_id, + "检测到中断标志，立即停止音频发送任务", + ) + break + try: + audio_url = await asyncio.wait_for(url_queue.get(), timeout=0.1) + except asyncio.TimeoutError: + if self.check_interruption(): + break + continue + except Exception as e: + session_logger.error( + self.device_id, + self.session_id, + f"获取音频URL队列数据时出错: {str(e)}", + ) + break + if audio_url is None: + break + if self.is_websocket_connected() and not self.check_interruption(): + try: + full_url = f"http://{settings.server_host}:{settings.server_port}/{audio_url}" if not audio_url.startswith('http') else audio_url + await self.websocket.send_text(f"TTS_URL:{full_url}") + + if not first_url_sent and first_tts_submit_time is not None: + first_url_sent = True + + session_logger.info( + self.device_id, + self.session_id, + f"已发送音频URL给客户端: {full_url}", + ) + except Exception as e: + session_logger.error( + self.device_id, + self.session_id, + f"发送音频URL时出错: {str(e)}", + ) + await interrupt_handler.set_interrupt_state( + self.session_key, True + ) + break + else: + session_logger.warning( + self.device_id, + self.session_id, + "WebSocket连接已关闭或会话已中断，无法发送音频URL", + ) + break + if self.check_interruption(): + break + url_queue.task_done() + if not self.check_interruption(): + try: + if self.is_websocket_connected(): + try: + await self.websocket.send_text("TTS_END") + session_logger.info( + self.device_id, + self.session_id, + "已发送 TTS_END 给客户端", + ) + await tts_error_manager.end_tts_session(self.device_id, self.session_id) + except RuntimeError as e: + if "Cannot call" in str(e) and "close message" in str(e): + session_logger.warning( + self.device_id, + self.session_id, + f"WebSocket已关闭，无法发送TTS_END: {e}", + ) + else: + session_logger.error( + self.device_id, + self.session_id, + f"发送TTSEND时发生运行时错误: {e}", + ) + else: + session_logger.warning( + self.device_id, + self.session_id, + "WebSocket连接已关闭或会话已中断，跳过发送 TTS_END", + ) + except Exception as e: + session_logger.error( + self.device_id, self.session_id, f"检查WebSocket状态时出错: {e}" + ) + except Exception as e: + session_logger.error( + self.device_id, + self.session_id, + f"发送音频URL时出错: {e}", + exc_info=True, + ) diff --git a/talkingq-url/services/audio_session.py b/talkingq-url/services/audio_session.py new file mode 100644 index 0000000..1d7b584 --- /dev/null +++ b/talkingq-url/services/audio_session.py @@ -0,0 +1,195 @@ +import asyncio +from typing import Dict, Optional +from interfaces.asr import ASR +from interfaces.llm import LLM +from interfaces.tts import TTS +from services.interrupt_handler import interrupt_handler + + +class AudioSession: + def __init__( + self, + asr_service: Optional[ASR] = None, + llm_service: Optional[LLM] = None, + tts_service: Optional[TTS] = None, + sample_rate: int = 16000, + ): + self.start_time = None + self.end_time = None + self.last_received_seq = -1 + self.asr_service = asr_service + self.send_queue = None + self.send_task = None + self.transcript = "" + self.last_activity_time = asyncio.get_running_loop().time() + self.websocket = None + self.llm_service = llm_service + self.tts_service = tts_service + self.device_id = None + self.session_id = None + self.sample_rate = sample_rate + self.audio_queue = None + + async def set_interrupted(self, interrupted=True): + + if self.device_id and self.session_id: + session_key = (self.device_id, self.session_id) + await interrupt_handler.set_interrupt_state(session_key, interrupted) + if hasattr(self.llm_service, "closed"): + self.llm_service.closed = interrupted + from utils.logger import session_logger + + session_logger.info( + self.device_id, self.session_id, f"会话中断状态已设置为: {interrupted}" + ) + if interrupted: + from services.task_manager import task_manager + await task_manager.create_task( + interrupt_handler.handle_interrupt(session_key), + device_id=self.device_id, + session_key=session_key, + task_type="interrupt" + ) + + async def register_interrupt_handlers(self): + + if not self.device_id or not self.session_id: + return + session_key = (self.device_id, self.session_id) + if hasattr(self, "audio_queue") and self.audio_queue: + await interrupt_handler.register_cleanup_handler( + session_key, + self.clear_audio_queue, + priority=interrupt_handler.PRIORITY["QUEUE_CLEANUP"], + ) + if hasattr(self, "send_queue") and self.send_queue: + await interrupt_handler.register_cleanup_handler( + session_key, + self.clear_send_queue, + priority=interrupt_handler.PRIORITY["QUEUE_CLEANUP"], + ) + if self.send_task: + await interrupt_handler.register_cleanup_handler( + session_key, + self.cancel_send_task, + priority=interrupt_handler.PRIORITY["TASK_CANCELLATION"], + ) + + async def clear_audio_queue(self): + + try: + if hasattr(self, "audio_queue") and self.audio_queue: + queue_size = self.audio_queue.qsize() + cleared_count = 0 + + # 发送结束信号 + await self.audio_queue.put((None, 0)) + + # 清空队列中的所有项目 + while not self.audio_queue.empty(): + try: + item = await self.audio_queue.get_nowait() + self.audio_queue.task_done() + cleared_count += 1 + # 如果是大对象，显式删除引用 + del item + except asyncio.QueueEmpty: + break + except Exception: + pass + + from utils.logger import session_logger + session_logger.info( + self.device_id, self.session_id, + f"音频队列已清空，原有项目数: {queue_size}，清理项目数: {cleared_count}" + ) + + # 清除队列引用 + self.audio_queue = None + except Exception as e: + from utils.logger import session_logger + + session_logger.error( + self.device_id, self.session_id, f"清空音频队列时出错: {e}" + ) + + async def clear_send_queue(self): + + try: + if hasattr(self, "send_queue") and self.send_queue: + queue_size = self.send_queue.qsize() + cleared_count = 0 + + # 清空队列中的所有项目 + while not self.send_queue.empty(): + try: + item = self.send_queue.get_nowait() + self.send_queue.task_done() + cleared_count += 1 + # 如果是大对象，显式删除引用 + del item + except asyncio.QueueEmpty: + break + + # 发送结束信号 + await self.send_queue.put(None) + + from utils.logger import session_logger + session_logger.info( + self.device_id, self.session_id, + f"发送队列已清空，原有项目数: {queue_size}，清理项目数: {cleared_count}" + ) + + # 清除队列引用 + self.send_queue = None + except Exception as e: + from utils.logger import session_logger + + session_logger.error( + self.device_id, self.session_id, f"清空发送队列时出错: {e}" + ) + + async def cancel_send_task(self): + + if self.send_task: + self.send_task.cancel() + try: + await asyncio.wait_for(self.send_task, timeout=1.0) + except (asyncio.CancelledError, asyncio.TimeoutError): + from utils.logger import session_logger + + session_logger.info(self.device_id, self.session_id, "发送任务已取消") + except Exception as e: + from utils.logger import session_logger + + session_logger.error( + self.device_id, self.session_id, f"取消发送任务时出错: {e}" + ) + finally: + self.send_task = None + + +class AudioSessionManager: + def __init__(self): + self.sessions: Dict[tuple, AudioSession] = {} + self.lock = asyncio.Lock() + + async def get_session(self, session_key): + async with self.lock: + return self.sessions.get(session_key) + + async def set_session(self, session_key, session): + async with self.lock: + self.sessions[session_key] = session + + async def remove_session(self, session_key): + async with self.lock: + if session_key in self.sessions: + del self.sessions[session_key] + + async def get_all_sessions(self): + async with self.lock: + return list(self.sessions.items()) + + +audio_session_manager = AudioSessionManager() diff --git a/talkingq-url/services/config_manager.py b/talkingq-url/services/config_manager.py new file mode 100644 index 0000000..8cf16d2 --- /dev/null +++ b/talkingq-url/services/config_manager.py @@ -0,0 +1,48 @@ +from typing import Any, Dict +from config import settings + +class ConfigManager: + """统一的配置管理服务，处理全局配置和角色特定配置的优先级""" + @staticmethod + def get_config_value(config_key: str, role_config: Dict[str, Any] = None, default_value: Any = None) -> Any: + """ + 获取配置值，遵循优先级：角色配置 > 全局配置 > 默认值 + Args: + config_key: 配置键名 + role_config: 角色配置字典 + default_value: 默认值，当角色配置和全局配置都没有该键时返回 + """ + if role_config and config_key in role_config: + return role_config[config_key] + if hasattr(settings, config_key): + return getattr(settings, config_key) + return default_value + + @staticmethod + def get_service_config(service_type: str, role_config: Dict[str, Any] = None) -> Dict[str, Any]: + """ + 获取服务特定配置，整合角色配置和全局配置 + Args: + service_type: 服务类型 (asr, llm, tts) + role_config: 角色配置字典 + """ + config = {} + config_mapping = { + 'asr': { + 'provider': 'asr_provider', + }, + 'llm': { + 'provider': 'llm_provider', + 'model_id': 'volcano_model_id', + }, + 'tts': { + 'provider': 'tts_provider', + 'voice_type': 'minimax_voice_id', + } + } + for config_key, setting_key in config_mapping.get(service_type, {}).items(): + if setting_key: + config[config_key] = ConfigManager.get_config_value(setting_key, role_config) + return config + +config_manager = ConfigManager() diff --git a/talkingq-url/services/connection_manager.py b/talkingq-url/services/connection_manager.py new file mode 100644 index 0000000..216b824 --- /dev/null +++ b/talkingq-url/services/connection_manager.py @@ -0,0 +1,29 @@ +import asyncio +from typing import Dict, Optional +from fastapi import WebSocket + + +class ConnectionManager: + def __init__(self): + self.connections: Dict[str, WebSocket] = {} + self.lock = asyncio.Lock() + + async def add_connection(self, device_id: str, websocket: WebSocket): + async with self.lock: + self.connections[device_id] = websocket + + async def remove_connection(self, device_id: str): + async with self.lock: + if device_id in self.connections: + del self.connections[device_id] + + async def get_connection(self, device_id: str) -> Optional[WebSocket]: + async with self.lock: + return self.connections.get(device_id) + + async def get_all_connections(self) -> Dict[str, WebSocket]: + async with self.lock: + return dict(self.connections) + + +connection_manager = ConnectionManager() diff --git a/talkingq-url/services/conversation_history.py b/talkingq-url/services/conversation_history.py new file mode 100644 index 0000000..8790d77 --- /dev/null +++ b/talkingq-url/services/conversation_history.py @@ -0,0 +1,412 @@ +import asyncio +import time +from typing import Dict, List, Optional, Tuple +from sqlalchemy import select, update, insert, and_, delete +from database.models import ConversationHistory, ConversationMessage +from services.database_service_base import DatabaseServiceBase +from utils.logger import session_logger +from config import settings + +class DeviceConversationHistory: + def __init__(self): + self.history = [] + self.last_interaction_time = asyncio.get_running_loop().time() + self.role_key = None # 角色标识 + self.conversation_id = None # 会话ID，用于关联消息 + +class ConversationHistoryManager(DatabaseServiceBase): + def __init__(self): + super().__init__(service_name="conversation_history") + self.histories: Dict[str, DeviceConversationHistory] = {} # 内存缓存 + self.lock = asyncio.Lock() + self.role_histories: Dict[Tuple[str, str], DeviceConversationHistory] = {} # (device_id, role_key) -> history + self.max_cache_size = 500 # 最大缓存条目数 + + async def get_history(self, device_id: str, role_key: Optional[str] = None) -> DeviceConversationHistory: + """获取设备指定角色的对话历史""" + await self._init_database() + if not role_key: + from services.device_config import device_config_manager + device_config = await device_config_manager.get_config(device_id) + if device_config: + role_key = device_config.selected_role_key + else: + from config import settings + role_key = settings.selected_role_key + + cache_key = (device_id, role_key) + if cache_key in self.role_histories: + return self.role_histories[cache_key] + + db_session = await self.get_session() + try: + query = select(ConversationHistory).where( + and_( + ConversationHistory.device_id == device_id, + ConversationHistory.role_key == role_key + ) + ) + result = await db_session.execute(query) + db_history = result.scalar_one_or_none() + + history = DeviceConversationHistory() + history.role_key = role_key + + if db_history: + messages_query = select(ConversationMessage).where( + ConversationMessage.conversation_id == db_history.id + ).order_by(ConversationMessage.timestamp) + + messages_result = await db_session.execute(messages_query) + messages = messages_result.scalars().all() + + message_pairs = [] + user_msg = None + + for msg in messages: + if msg.is_user: + user_msg = msg.content + elif user_msg is not None: + message_pairs.append({"user": user_msg, "assistant": msg.content}) + user_msg = None + + if user_msg is not None: + message_pairs.append({"user": user_msg, "assistant": ""}) + + history.history = message_pairs + history.last_interaction_time = db_history.last_interaction_time + history.conversation_id = db_history.id + + self._add_to_cache(cache_key, history) + return history + else: + history = DeviceConversationHistory() + history.history = [] + history.role_key = role_key + self._add_to_cache(cache_key, history) + return history + except Exception as e: + session_logger.error(device_id, "conversation_history", f"从数据库获取对话历史失败: {str(e)}") + history = DeviceConversationHistory() + history.role_key = role_key + return history + finally: + await db_session.close() + + async def set_history(self, device_id: str, history: DeviceConversationHistory, role_key: Optional[str] = None): + """设置设备指定角色的对话历史""" + await self._init_database() + if not role_key and not history.role_key: + from services.device_config import device_config_manager + device_config = await device_config_manager.get_config(device_id) + role_key = device_config.selected_role_key if device_config else settings.selected_role_key + + effective_role_key = role_key or history.role_key + history.role_key = effective_role_key + cache_key = (device_id, effective_role_key) + self._add_to_cache(cache_key, history) + + db_session = await self.get_session() + try: + query = select(ConversationHistory).where( + and_( + ConversationHistory.device_id == device_id, + ConversationHistory.role_key == effective_role_key + ) + ) + result = await db_session.execute(query) + db_history = result.scalar_one_or_none() + current_time = time.time() + + if db_history: + stmt = update(ConversationHistory).where( + ConversationHistory.id == db_history.id + ).values( + last_interaction_time=current_time + ) + await db_session.execute(stmt) + conversation_id = db_history.id + else: + stmt = insert(ConversationHistory).values( + device_id=device_id, + role_key=effective_role_key, + last_interaction_time=current_time + ) + result = await db_session.execute(stmt) + await db_session.commit() + + query = select(ConversationHistory).where( + and_( + ConversationHistory.device_id == device_id, + ConversationHistory.role_key == effective_role_key + ) + ) + result = await db_session.execute(query) + db_history = result.scalar_one() + conversation_id = db_history.id + history.conversation_id = conversation_id + + if history.history and len(history.history) > 0: + latest_msg = history.history[-1] + + latest_msg_query = select(ConversationMessage).where( + and_( + ConversationMessage.conversation_id == conversation_id, + ConversationMessage.content == latest_msg["user"], + ConversationMessage.is_user == True + ) + ).order_by(ConversationMessage.timestamp.desc()) + + latest_msg_result = await db_session.execute(latest_msg_query) + existing_user_msg = latest_msg_result.scalar_one_or_none() + + if not existing_user_msg: + user_stmt = insert(ConversationMessage).values( + conversation_id=conversation_id, + is_user=True, + content=latest_msg["user"], + timestamp=current_time - 0.1 # 确保用户消息在助手消息之前 + ) + await db_session.execute(user_stmt) + + assistant_stmt = insert(ConversationMessage).values( + conversation_id=conversation_id, + is_user=False, + content=latest_msg["assistant"], + timestamp=current_time + ) + await db_session.execute(assistant_stmt) + + await db_session.commit() + session_logger.info( + device_id, + "conversation_history", + f"已保存设备 {device_id} 的角色 {effective_role_key} 对话历史到数据库" + ) + except Exception as e: + await db_session.rollback() + session_logger.error( + device_id, + "conversation_history", + f"保存对话历史到数据库失败: {str(e)}" + ) + finally: + await db_session.close() + + async def remove_history(self, device_id: str, role_key: Optional[str] = None): + """删除设备的对话历史""" + await self._init_database() + if not role_key: + from services.device_config import device_config_manager + device_config = await device_config_manager.get_config(device_id) + role_key = device_config.selected_role_key if device_config else settings.selected_role_key + + cache_key = (device_id, role_key) + if cache_key in self.role_histories: + del self.role_histories[cache_key] + + db_session = await self.get_session() + try: + query = select(ConversationHistory.id).where( + and_( + ConversationHistory.device_id == device_id, + ConversationHistory.role_key == role_key + ) + ) + result = await db_session.execute(query) + conversation_id = result.scalar_one_or_none() + + if conversation_id: + msg_stmt = delete(ConversationMessage).where( + ConversationMessage.conversation_id == conversation_id + ) + await db_session.execute(msg_stmt) + + hist_stmt = delete(ConversationHistory).where( + ConversationHistory.id == conversation_id + ) + await db_session.execute(hist_stmt) + + await db_session.commit() + session_logger.info( + device_id, + "conversation_history", + f"已删除设备 {device_id} 的角色 {role_key} 对话历史" + ) + except Exception as e: + await db_session.rollback() + session_logger.error( + device_id, + "conversation_history", + f"删除对话历史失败: {str(e)}" + ) + finally: + await db_session.close() + + async def clear_all_histories(self, device_id: str): + """清除设备的所有对话历史（所有角色）""" + await self._init_database() + keys_to_remove = [] + for (dev_id, _) in self.role_histories.keys(): + if dev_id == device_id: + keys_to_remove.append((dev_id, _)) + + for key in keys_to_remove: + del self.role_histories[key] + + db_session = await self.get_session() + try: + query = select(ConversationHistory.id).where( + ConversationHistory.device_id == device_id + ) + result = await db_session.execute(query) + conversation_ids = result.scalars().all() + + if conversation_ids: + for cid in conversation_ids: + msg_stmt = delete(ConversationMessage).where( + ConversationMessage.conversation_id == cid + ) + await db_session.execute(msg_stmt) + + stmt = delete(ConversationHistory).where( + ConversationHistory.device_id == device_id + ) + await db_session.execute(stmt) + await db_session.commit() + + session_logger.info( + device_id, + "conversation_history", + f"已清除设备 {device_id} 的所有对话历史" + ) + except Exception as e: + await db_session.rollback() + session_logger.error( + device_id, + "conversation_history", + f"清除所有对话历史失败: {str(e)}" + ) + finally: + await db_session.close() + + async def get_all_histories(self): + """获取所有历史记录，主要用于清理过期记录""" + await self._init_database() + db_session = await self.get_session() + try: + query = select(ConversationHistory) + result = await db_session.execute(query) + db_histories = result.scalars().all() + histories = [] + + for db_history in db_histories: + history = DeviceConversationHistory() + history.last_interaction_time = db_history.last_interaction_time + history.role_key = db_history.role_key + history.conversation_id = db_history.id + + histories.append(((db_history.device_id, db_history.role_key), history)) + + return histories + except Exception as e: + session_logger.error("system", "conversation_history", f"获取所有对话历史失败: {str(e)}") + return [] + finally: + await db_session.close() + + async def get_device_role_histories(self, device_id: str) -> Dict[str, DeviceConversationHistory]: + """获取设备的所有角色会话历史""" + await self._init_database() + db_session = await self.get_session() + try: + query = select(ConversationHistory).where( + ConversationHistory.device_id == device_id + ) + result = await db_session.execute(query) + db_histories = result.scalars().all() + + histories = {} + for db_history in db_histories: + role_key = db_history.role_key + history = DeviceConversationHistory() + history.role_key = role_key + history.last_interaction_time = db_history.last_interaction_time + history.conversation_id = db_history.id + + messages_query = select(ConversationMessage).where( + ConversationMessage.conversation_id == db_history.id + ).order_by(ConversationMessage.timestamp) + + messages_result = await db_session.execute(messages_query) + messages = messages_result.scalars().all() + + for msg in messages: + if msg.is_user: + history.history.append({ + "user": msg.content, + "assistant": "", + "timestamp": msg.timestamp + }) + else: + if history.history and "assistant" in history.history[-1]: + history.history[-1]["assistant"] = msg.content + + histories[role_key] = history + + return histories + except Exception as e: + session_logger.error(device_id, "conversation_history", f"获取设备所有角色历史失败: {str(e)}") + return {} + finally: + await db_session.close() + + async def clear_cache(self, device_id: str = None, role_key: str = None): + """清除指定设备和角色的缓存，或全部缓存""" + if device_id and role_key: + cache_key = (device_id, role_key) + if cache_key in self.role_histories: + del self.role_histories[cache_key] + elif device_id: + keys_to_remove = [] + for (dev_id, _), _ in self.role_histories.items(): + if dev_id == device_id: + keys_to_remove.append((dev_id, _)) + + for key in keys_to_remove: + if key in self.role_histories: + del self.role_histories[key] + else: + self.role_histories.clear() + + def _add_to_cache(self, cache_key: Tuple[str, str], history: DeviceConversationHistory): + """添加到缓存，检查大小限制""" + # 检查缓存大小限制 + if len(self.role_histories) >= self.max_cache_size: + self._cleanup_old_cache_entries() + + self.role_histories[cache_key] = history + + def _cleanup_old_cache_entries(self): + """清理最旧的缓存条目""" + if not self.role_histories: + return + + # 按最后交互时间排序，删除最旧的25%条目 + sorted_items = sorted( + self.role_histories.items(), + key=lambda x: x[1].last_interaction_time + ) + cleanup_count = max(1, len(sorted_items) // 4) + + for cache_key, _ in sorted_items[:cleanup_count]: + if cache_key in self.role_histories: + del self.role_histories[cache_key] + + session_logger.info( + "system", "cache_cleanup", + f"对话历史缓存清理完成，删除了 {cleanup_count} 个条目，剩余 {len(self.role_histories)} 个" + ) + + +conversation_history_manager = ConversationHistoryManager() diff --git a/talkingq-url/services/database_service_base.py b/talkingq-url/services/database_service_base.py new file mode 100644 index 0000000..23c329a --- /dev/null +++ b/talkingq-url/services/database_service_base.py @@ -0,0 +1,30 @@ +import asyncio +from utils.logger import session_logger +from database.connection import get_db_manager + +class DatabaseServiceBase: + """ + 数据库服务基类，提供通用的数据库初始化和会话管理功能 + 所有需要访问数据库的服务类都应该继承此基类 + """ + def __init__(self, service_name="unknown"): + self.db_manager = None + self._db_initialized = False + self.service_name = service_name + + async def _init_database(self): + """初始化数据库连接""" + if self._db_initialized: + return + try: + self.db_manager = await get_db_manager() + self._db_initialized = True + session_logger.info("system", self.service_name, f"{self.service_name}数据库连接初始化成功") + except Exception as e: + session_logger.error("system", self.service_name, f"{self.service_name}数据库连接初始化失败: {str(e)}") + raise + + async def get_session(self): + """获取数据库会话""" + await self._init_database() + return await self.db_manager.get_session() diff --git a/talkingq-url/services/device_auth_manager.py b/talkingq-url/services/device_auth_manager.py new file mode 100644 index 0000000..a140a4a --- /dev/null +++ b/talkingq-url/services/device_auth_manager.py @@ -0,0 +1,101 @@ +import asyncio +from sqlalchemy import select, update, insert +from utils.logger import session_logger +from database.models import DeviceAuth +from services.database_service_base import DatabaseServiceBase +from typing import Optional + +class DeviceAuthManager(DatabaseServiceBase): + def __init__(self): + super().__init__(service_name="device_auth") + self.auth_cache = {} # 缓存设备认证信息 + self.lock = asyncio.Lock() + + async def authenticate_device(self, device_id: str, serial_number: str) -> bool: + """验证设备凭据""" + await self._init_database() + if device_id in self.auth_cache: + return self.auth_cache[device_id] == serial_number + + db_session = await self.get_session() + try: + query = select(DeviceAuth).where(DeviceAuth.device_id == device_id, DeviceAuth.is_active == True) + result = await db_session.execute(query) + device_auth = result.scalar_one_or_none() + + if device_auth and device_auth.serial_number == serial_number: + self.auth_cache[device_id] = serial_number + session_logger.info("system", "device_auth", f"设备 {device_id} 验证成功") + return True + + session_logger.warning("system", "device_auth", f"设备 {device_id} 验证失败，无效的凭据") + return False + except Exception as e: + session_logger.error("system", "device_auth", f"验证设备 {device_id} 时出错: {str(e)}") + return False + finally: + await db_session.close() + + async def register_device(self, device_id: str, serial_number: str, batch_id: str = None, is_active: bool = True) -> bool: + """注册新设备或更新已有设备""" + await self._init_database() + db_session = await self.get_session() + try: + query = select(DeviceAuth).where(DeviceAuth.device_id == device_id) + result = await db_session.execute(query) + existing_device = result.scalar_one_or_none() + + if existing_device: + stmt = update(DeviceAuth).where( + DeviceAuth.device_id == device_id + ).values( + serial_number=serial_number, + batch_id=batch_id, + is_active=is_active + ) + await db_session.execute(stmt) + else: + stmt = insert(DeviceAuth).values( + device_id=device_id, + serial_number=serial_number, + batch_id=batch_id, + is_active=is_active + ) + await db_session.execute(stmt) + + await db_session.commit() + self.auth_cache[device_id] = serial_number + session_logger.info("system", "device_auth", f"设备 {device_id} 注册成功，批次: {batch_id}") + return True + except Exception as e: + await db_session.rollback() + session_logger.error("system", "device_auth", f"注册设备 {device_id} 时出错: {str(e)}") + return False + finally: + await db_session.close() + + async def get_device_info(self, device_id: str) -> Optional[dict]: + """获取设备详细信息""" + await self._init_database() + db_session = await self.get_session() + try: + query = select(DeviceAuth).where(DeviceAuth.device_id == device_id) + result = await db_session.execute(query) + device = result.scalar_one_or_none() + if not device: + return None + return { + "device_id": device.device_id, + "serial_number": device.serial_number, + "batch_id": device.batch_id, + "is_active": device.is_active, + "created_at": device.created_at, + "updated_at": device.updated_at + } + except Exception as e: + session_logger.error("system", "device_auth", f"获取设备 {device_id} 信息时出错: {str(e)}") + return None + finally: + await db_session.close() + +device_auth_manager = DeviceAuthManager() diff --git a/talkingq-url/services/device_config.py b/talkingq-url/services/device_config.py new file mode 100644 index 0000000..ed0d8ff --- /dev/null +++ b/talkingq-url/services/device_config.py @@ -0,0 +1,206 @@ +import asyncio +from typing import Dict, Optional +import os +import time +from sqlalchemy import select, update, insert +from sqlalchemy.ext.asyncio import AsyncSession +from utils.logger import session_logger +from pydantic import BaseModel +from database.models import DeviceConfig as DBDeviceConfig +from services.database_service_base import DatabaseServiceBase + +class DeviceConfig(BaseModel): + selected_role_key: str + preferred_language: Optional[str] = None + last_update_time: float = None + + def __init__(self, **data): + if 'last_update_time' not in data: + data['last_update_time'] = time.time() + super().__init__(**data) + + def to_dict(self): + """将配置转换为可序列化的字典""" + return { + "selected_role_key": self.selected_role_key, + "preferred_language": self.preferred_language, + } + + @classmethod + def from_db_model(cls, db_model: DBDeviceConfig): + """从数据库模型创建配置对象""" + return cls( + selected_role_key=db_model.selected_role_key, + preferred_language=db_model.preferred_language, + last_update_time=db_model.last_update_time + ) + +class DeviceConfigManager(DatabaseServiceBase): + def __init__(self): + super().__init__(service_name="device_config") + self.device_configs: Dict[str, DeviceConfig] = {} + self.lock = asyncio.Lock() + self.worker_id = os.environ.get("UVICORN_WID", "0") + self.config_last_updated = {} + + async def _is_config_updated(self, device_id: str, async_session: AsyncSession) -> bool: + """检查配置是否已在数据库中更新""" + try: + query = select(DBDeviceConfig.last_update_time).where(DBDeviceConfig.device_id == device_id) + result = await async_session.execute(query) + db_last_updated = result.scalar_one_or_none() + + if db_last_updated is None: + return False + + local_last_updated = self.config_last_updated.get(device_id, 0) + return db_last_updated > local_last_updated + except Exception as e: + session_logger.error(device_id, "config", f"检查配置更新失败: {str(e)}") + return False + + async def _load_config_from_db(self, device_id: str, async_session: AsyncSession): + """从数据库加载设备配置""" + try: + query = select(DBDeviceConfig).where(DBDeviceConfig.device_id == device_id) + result = await async_session.execute(query) + db_config = result.scalar_one_or_none() + + if db_config: + device_config = DeviceConfig.from_db_model(db_config) + self.config_last_updated[device_id] = db_config.last_update_time + + async with self.lock: + self.device_configs[device_id] = device_config + + session_logger.info( + device_id, + "config", + f"Worker {self.worker_id}: 从数据库加载设备配置，角色: {device_config.selected_role_key}, 语言: {device_config.preferred_language or '未设置'}" + ) + return device_config + return None + except Exception as e: + session_logger.error(device_id, "config", f"从数据库加载配置失败: {str(e)}") + return None + + async def _save_config_to_db(self, device_id: str, config: DeviceConfig, async_session: AsyncSession): + """保存设备配置到数据库""" + try: + query = select(DBDeviceConfig).where(DBDeviceConfig.device_id == device_id) + result = await async_session.execute(query) + existing_config = result.scalar_one_or_none() + + current_time = time.time() + + if existing_config: + stmt = update(DBDeviceConfig).where( + DBDeviceConfig.device_id == device_id + ).values( + selected_role_key=config.selected_role_key, + preferred_language=config.preferred_language, + last_update_time=current_time + ) + else: + stmt = insert(DBDeviceConfig).values( + device_id=device_id, + selected_role_key=config.selected_role_key, + preferred_language=config.preferred_language, + last_update_time=current_time + ) + + await async_session.execute(stmt) + await async_session.commit() + + self.config_last_updated[device_id] = current_time + + session_logger.info( + device_id, + "config", + f"Worker {self.worker_id}: 设备配置已保存到数据库" + ) + except Exception as e: + await async_session.rollback() + session_logger.error( + device_id, + "config", + f"Worker {self.worker_id}: 保存设备配置到数据库失败: {str(e)}" + ) + raise + + async def get_config(self, device_id: str, force_refresh: bool = False) -> DeviceConfig: + """获取设备配置，如果force_refresh=True则强制从数据库读取""" + await self._init_database() + + config = None + config_updated = False + + db_session = await self.db_manager.get_session() + + try: + if force_refresh or await self._is_config_updated(device_id, db_session): + config = await self._load_config_from_db(device_id, db_session) + config_updated = True + + if not config_updated: + async with self.lock: + config = self.device_configs.get(device_id) + + if not config: + from config import settings + default_language = "en" + config = DeviceConfig( + selected_role_key=settings.selected_role_key, + preferred_language=default_language + ) + session_logger.info( + device_id, + "config", + f"Worker {self.worker_id}: 为新设备创建配置，默认角色: {settings.selected_role_key}，默认语言: {default_language}" + ) + await self.set_config(device_id, config) + elif config_updated: + session_logger.info( + device_id, + "config", + f"Worker {self.worker_id}: 已从数据库刷新设备配置，当前角色: {config.selected_role_key}, 语言: {config.preferred_language or '未设置'}" + ) + + return config + except Exception as e: + session_logger.error(device_id, "config", f"获取设备配置失败: {str(e)}") + raise + finally: + await db_session.close() + + async def set_config(self, device_id: str, config: DeviceConfig): + """设置并保存设备配置""" + await self._init_database() + + config.last_update_time = time.time() + + db_session = await self.db_manager.get_session() + + try: + async with self.lock: + old_config = self.device_configs.get(device_id) + self.device_configs[device_id] = config + + if old_config: + changes = [] + if old_config.selected_role_key != config.selected_role_key: + changes.append(f"角色: {old_config.selected_role_key} -> {config.selected_role_key}") + if old_config.preferred_language != config.preferred_language: + changes.append(f"语言: {old_config.preferred_language or '未设置'} -> {config.preferred_language or '未设置'}") + + if changes: + session_logger.info(device_id, "config", f"设备配置已更新: {', '.join(changes)}") + + await self._save_config_to_db(device_id, config, db_session) + except Exception as e: + session_logger.error(device_id, "config", f"设置设备配置失败: {str(e)}") + raise + finally: + await db_session.close() + +device_config_manager = DeviceConfigManager() diff --git a/talkingq-url/services/device_service.py b/talkingq-url/services/device_service.py new file mode 100644 index 0000000..1a8bb52 --- /dev/null +++ b/talkingq-url/services/device_service.py @@ -0,0 +1,50 @@ +from typing import Optional, Dict +from services.device_config import device_config_manager, DeviceConfig +from services.role_manager import role_manager +from services.conversation_history import conversation_history_manager, DeviceConversationHistory +from handlers.prompt_sound_handler import send_welcome_sound +from utils.logger import session_logger + +class DeviceService: + async def get_device_role(self, device_id: str) -> Dict: + """获取设备当前使用的角色""" + from handlers.command_handler import get_device_role as handler_get_device_role + device_config = await device_config_manager.get_config(device_id, force_refresh=True) + selected_role = await handler_get_device_role(device_id) + result = { + "role_key": selected_role.get("role_key", ""), + "name": selected_role.get("name", "未命名角色"), + "description": selected_role.get("description", ""), + "preferred_language": device_config.preferred_language if device_config else None + } + if "multilingual" in selected_role: + result["languages"] = list(selected_role["multilingual"].keys()) + return result + + async def update_device_role(self, device_id: str, role_key: str, language: Optional[str] = None) -> Dict: + """更新设备角色配置""" + role = await role_manager.get_role(role_key) + if not role: + raise ValueError(f"角色 {role_key} 不存在") + + try: + current_config = await device_config_manager.get_config(device_id, force_refresh=True) + preserved_language = language or (current_config.preferred_language if current_config else None) + device_config = DeviceConfig(selected_role_key=role_key, preferred_language=preserved_language) + await device_config_manager.set_config(device_id, device_config) + + await conversation_history_manager.set_history(device_id, DeviceConversationHistory()) + await send_welcome_sound(device_id, role_key, preserved_language) + + session_logger.info(device_id, "config", f"设备的角色已通过API更新为 {role_key}, 语言: {preserved_language or '默认'}") + return { + "status": "success", + "device_id": device_id, + "role_key": role_key, + "language": preserved_language + } + except Exception as e: + session_logger.error(device_id, "config", f"通过API更新角色时出错: {str(e)}") + raise + +device_service = DeviceService() diff --git a/talkingq-url/services/device_update_manager.py b/talkingq-url/services/device_update_manager.py new file mode 100644 index 0000000..074a77e --- /dev/null +++ b/talkingq-url/services/device_update_manager.py @@ -0,0 +1,267 @@ +from typing import Dict, Optional +import time +from sqlalchemy import select, update, insert, delete +from database.models import DeviceFirmwareUpdate +from services.database_service_base import DatabaseServiceBase +from utils.logger import session_logger + +class DeviceFirmwareUpdateManager(DatabaseServiceBase): + def __init__(self): + super().__init__(service_name="device_firmware_update") + self.update_cache = {} # 添加缓存以减少数据库查询 + self.cache_expiry = 300 # 缓存过期时间(秒) + self.cache_timestamps = {} # 记录缓存更新时间 + self.max_cache_size = 1000 # 最大缓存条目数 + + async def create_firmware_update(self, device_id: str, firmware_version: str, update_status: str = "success"): + """新增固件更新记录""" + await self._init_database() + db_session = await self.get_session() + try: + query = select(DeviceFirmwareUpdate).where(DeviceFirmwareUpdate.device_id == device_id) + result = await db_session.execute(query) + existing = result.scalar_one_or_none() + + if existing: + update_stmt = ( + update(DeviceFirmwareUpdate) + .where(DeviceFirmwareUpdate.device_id == device_id) + .values( + firmware_version=firmware_version, + update_status=update_status, + progress=0.0 + ) + ) + await db_session.execute(update_stmt) + else: + insert_stmt = insert(DeviceFirmwareUpdate).values( + device_id=device_id, + firmware_version=firmware_version, + update_status=update_status, + progress=0.0 + ) + await db_session.execute(insert_stmt) + + await db_session.commit() + + self._update_cache(device_id, { + "device_id": device_id, + "firmware_version": firmware_version, + "update_status": update_status, + "progress": 0.0 + }) + + return True + except Exception as e: + await db_session.rollback() + session_logger.error(device_id, "firmware_update", f"创建固件更新记录失败: {str(e)}") + return False + finally: + await db_session.close() + + async def get_firmware_update(self, device_id: str) -> Optional[DeviceFirmwareUpdate]: + """获取设备固件更新信息""" + cached_data = self._get_from_cache(device_id) + if cached_data: + return cached_data + await self._init_database() + db_session = await self.get_session() + try: + query = select(DeviceFirmwareUpdate).where(DeviceFirmwareUpdate.device_id == device_id) + result = await db_session.execute(query) + update_record = result.scalars().first() + if update_record: + self._update_cache(device_id, update_record) + return update_record + except Exception as e: + session_logger.error(device_id, "firmware_update", f"查询固件更新失败: {str(e)}") + return None + finally: + await db_session.close() + + async def update_firmware_update(self, device_id: str, firmware_version: str, update_status: str) -> bool: + """更新设备固件信息（支持部分字段更新）""" + await self._init_database() + db_session = await self.get_session() + try: + update_values = { + "firmware_version": firmware_version, + "update_status": update_status + } + + query = select(DeviceFirmwareUpdate).where(DeviceFirmwareUpdate.device_id == device_id) + result = await db_session.execute(query) + device_update = result.scalars().first() + + if device_update: + update_stmt = ( + update(DeviceFirmwareUpdate) + .where(DeviceFirmwareUpdate.device_id == device_id) + .values(**update_values) + ) + await db_session.execute(update_stmt) + else: + insert_stmt = insert(DeviceFirmwareUpdate).values( + device_id=device_id, + firmware_version=firmware_version, + update_status=update_status + ) + await db_session.execute(insert_stmt) + + await db_session.commit() + self._clear_cache(device_id) + return True + except Exception as e: + await db_session.rollback() + session_logger.error(device_id, "firmware_update", f"更新固件信息失败: {str(e)}") + return False + finally: + await db_session.close() + + async def update_firmware_progress(self, device_id: str, progress: float) -> bool: + """更新升级进度""" + await self._init_database() + db_session = await self.get_session() + try: + update_stmt = ( + update(DeviceFirmwareUpdate) + .where(DeviceFirmwareUpdate.device_id == device_id) + .values(progress=progress) + ) + await db_session.execute(update_stmt) + await db_session.commit() + + if device_id in self.update_cache: + self.update_cache[device_id].progress = progress + self.cache_timestamps[device_id] = time.time() + + session_logger.info(device_id, "firmware_update", f"更新进度已更新: {progress:.1f}%") + return True + except Exception as e: + await db_session.rollback() + session_logger.error(device_id, "firmware_update", f"更新固件进度失败: {str(e)}") + return False + finally: + await db_session.close() + + async def delete_firmware_update(self, device_id: str) -> bool: + """删除设备固件更新记录""" + await self._init_database() + db_session = await self.get_session() + try: + delete_stmt = delete(DeviceFirmwareUpdate).where(DeviceFirmwareUpdate.device_id == device_id) + await db_session.execute(delete_stmt) + await db_session.commit() + + if device_id in self.update_cache: + del self.update_cache[device_id] + if device_id in self.cache_timestamps: + del self.cache_timestamps[device_id] + + session_logger.info(device_id, "firmware_update", "已删除固件更新记录") + return True + except Exception as e: + await db_session.rollback() + session_logger.error(device_id, "firmware_update", f"删除固件更新记录失败: {str(e)}") + return False + finally: + await db_session.close() + + async def update_device_firmware_version(self, device_id: str, firmware_version: str) -> bool: + """更新设备当前固件版本信息""" + await self._init_database() + db_session = await self.get_session() + try: + query = select(DeviceFirmwareUpdate).where(DeviceFirmwareUpdate.device_id == device_id) + result = await db_session.execute(query) + device_update = result.scalars().first() + + if device_update: + update_stmt = ( + update(DeviceFirmwareUpdate) + .where(DeviceFirmwareUpdate.device_id == device_id) + .values(firmware_version=firmware_version) + ) + await db_session.execute(update_stmt) + else: + insert_stmt = insert(DeviceFirmwareUpdate).values( + device_id=device_id, + firmware_version=firmware_version, + update_status="success" # 默认状态为成功 + ) + await db_session.execute(insert_stmt) + + await db_session.commit() + self._clear_cache(device_id) + return True + except Exception as e: + await db_session.rollback() + session_logger.error(device_id, "firmware_update", f"更新固件版本失败: {str(e)}") + return False + finally: + await db_session.close() + + async def request_firmware_version(self, device_id: str) -> bool: + """向设备请求当前固件版本""" + from services.connection_manager import connection_manager + + websocket = await connection_manager.get_connection(device_id) + if not websocket: + session_logger.error(device_id, "firmware_update", f"设备 {device_id} 未在线或未找到") + return False + + try: + await websocket.send_text("GET_FIRMWARE_VERSION") + session_logger.info(device_id, "firmware_update", "已向设备发送固件版本请求") + return True + except Exception as e: + session_logger.error(device_id, "firmware_update", f"请求设备固件版本失败: {str(e)}") + return False + + def _update_cache(self, device_id: str, update_record): + """更新缓存""" + # 检查缓存大小限制 + if len(self.update_cache) >= self.max_cache_size: + self._cleanup_old_cache_entries() + + self.update_cache[device_id] = update_record + self.cache_timestamps[device_id] = time.time() + + def _cleanup_old_cache_entries(self): + """清理最旧的缓存条目""" + if not self.cache_timestamps: + return + + # 按时间戳排序，删除最旧的25%条目 + sorted_items = sorted(self.cache_timestamps.items(), key=lambda x: x[1]) + cleanup_count = max(1, len(sorted_items) // 4) + + for device_id, _ in sorted_items[:cleanup_count]: + if device_id in self.update_cache: + del self.update_cache[device_id] + if device_id in self.cache_timestamps: + del self.cache_timestamps[device_id] + + session_logger.info( + "system", "cache_cleanup", + f"设备固件缓存清理完成，删除了 {cleanup_count} 个条目，剩余 {len(self.update_cache)} 个" + ) + + def _get_from_cache(self, device_id: str): + """从缓存获取记录，如果缓存过期则返回None""" + if device_id in self.update_cache and device_id in self.cache_timestamps: + if time.time() - self.cache_timestamps[device_id] < self.cache_expiry: + return self.update_cache[device_id] + else: + del self.update_cache[device_id] + del self.cache_timestamps[device_id] + return None + + def _clear_cache(self, device_id: str): + """清除设备的缓存""" + if device_id in self.update_cache: + del self.update_cache[device_id] + if device_id in self.cache_timestamps: + del self.cache_timestamps[device_id] + +device_firmware_update_manager = DeviceFirmwareUpdateManager() \ No newline at end of file diff --git a/talkingq-url/services/device_volume_manager.py b/talkingq-url/services/device_volume_manager.py new file mode 100644 index 0000000..be2229b --- /dev/null +++ b/talkingq-url/services/device_volume_manager.py @@ -0,0 +1,85 @@ +import asyncio +from typing import Dict, Optional +import time +from sqlalchemy import select, update, insert +from database.models import DeviceConfig +from services.database_service_base import DatabaseServiceBase +from utils.logger import session_logger + +class DeviceVolumeManager(DatabaseServiceBase): + def __init__(self): + super().__init__(service_name="device_volume") + self.device_volumes: Dict[str, int] = {} # 缓存设备音量 + self.lock = asyncio.Lock() + self._db_initialized = False + self.default_volume = 70 # 默认音量值 + + async def get_volume(self, device_id: str) -> int: + """获取设备音量设置""" + await self._init_database() + + if device_id in self.device_volumes: + return self.device_volumes[device_id] + + db_session = await self.get_session() + try: + query = select(DeviceConfig).where(DeviceConfig.device_id == device_id) + result = await db_session.execute(query) + config = result.scalars().first() + + if config and hasattr(config, "volume") and config.volume is not None: + volume = config.volume + else: + volume = self.default_volume + + self.device_volumes[device_id] = volume + return volume + except Exception as e: + session_logger.error(device_id, "volume", f"获取音量信息失败: {str(e)}") + return self.default_volume # 出错时返回默认值 + finally: + await db_session.close() + + async def set_volume(self, device_id: str, volume: int) -> bool: + """设置设备音量并保存到数据库""" + if volume < 0 or volume > 100: + raise ValueError("音量值必须在0-100范围内") + + await self._init_database() + + self.device_volumes[device_id] = volume + + db_session = await self.get_session() + try: + query = select(DeviceConfig).where(DeviceConfig.device_id == device_id) + result = await db_session.execute(query) + config = result.scalars().first() + + if config: + update_stmt = ( + update(DeviceConfig) + .where(DeviceConfig.device_id == device_id) + .values(volume=volume) + ) + await db_session.execute(update_stmt) + else: + from config import settings + insert_stmt = insert(DeviceConfig).values( + device_id=device_id, + selected_role_key=settings.selected_role_key, + volume=volume, + last_update_time=time.time() + ) + await db_session.execute(insert_stmt) + + await db_session.commit() + session_logger.info(device_id, "volume", f"设备音量已更新为 {volume}") + return True + except Exception as e: + await db_session.rollback() + session_logger.error(device_id, "volume", f"更新音量配置失败: {str(e)}") + raise + finally: + await db_session.close() + +device_volume_manager = DeviceVolumeManager() diff --git a/talkingq-url/services/factory.py b/talkingq-url/services/factory.py new file mode 100644 index 0000000..c881607 --- /dev/null +++ b/talkingq-url/services/factory.py @@ -0,0 +1,50 @@ +from interfaces.asr import ASR +from interfaces.llm import LLM +from interfaces.tts import TTS +from services.registry import get_service +from config import settings +from utils.logger import session_logger +from services.config_manager import config_manager +from implementations.aliyun_asr import AliyunASR +from implementations.volcano_llm import VolcanoLLM +from implementations.minimax_tts import MiniMaxTTS + +async def get_asr(selected_role: dict = None) -> ASR: + """获取ASR服务实例 - 固定使用阿里云""" + session_logger.info( + "system", + "factory", + f"ASR提供商: Aliyun (固定)" + ) + return AliyunASR(selected_role=selected_role) + +async def get_llm(selected_role: dict = None) -> LLM: + """获取LLM服务实例 - 固定使用火山引擎""" + role_selected_language = selected_role.get('_selected_language') if selected_role else None + session_logger.info( + "system", + "factory", + f"LLM提供商: Volcano (固定)" + + (f", 语言: {role_selected_language}" if role_selected_language else "") + ) + instance = VolcanoLLM(selected_role=selected_role) + if selected_role and "role_key" not in selected_role: + selected_role["role_key"] = "Volcano" # 使用provider作为默认值 + if hasattr(instance, "current_language"): + instance.current_language = role_selected_language + return instance + +async def get_tts(selected_role: dict = None, language: str = None) -> TTS: + """获取TTS服务实例 - 固定使用MiniMax""" + role_selected_language = selected_role.get('_selected_language') if selected_role else None + effective_language = language or role_selected_language + session_logger.info( + "system", + "factory", + f"TTS提供商: MiniMax (固定)" + + (f", 语言: {effective_language}" if effective_language else "") + ) + instance = MiniMaxTTS(selected_role=selected_role) + if hasattr(instance, "current_language"): + instance.current_language = effective_language + return instance diff --git a/talkingq-url/services/firmware_scanner.py b/talkingq-url/services/firmware_scanner.py new file mode 100644 index 0000000..12d47d2 --- /dev/null +++ b/talkingq-url/services/firmware_scanner.py @@ -0,0 +1,88 @@ +import os +import re +from typing import List, Tuple, Optional +from urllib.parse import urljoin +from config import settings +from utils.logger import session_logger +from services.system_config_manager import system_config_manager + +class FirmwareScanner: + def __init__(self): + self.firmware_dir = os.path.join(settings.assets_dir, "firmware") + self.firmware_file_pattern = re.compile(r'^(\d+\.\d+\.\d+)\.bin$') + + async def scan_firmware_files(self) -> None: + """扫描固件目录并更新系统配置数据库""" + try: + if not os.path.exists(self.firmware_dir): + os.makedirs(self.firmware_dir, exist_ok=True) + session_logger.info("system", "firmware_scanner", f"创建固件目录: {self.firmware_dir}") + return + + firmware_files = self._get_firmware_files() + if not firmware_files: + session_logger.info("system", "firmware_scanner", "未找到固件文件") + return + + latest_version, latest_file = self._find_latest_version(firmware_files) + if not latest_version: + session_logger.info("system", "firmware_scanner", "无法确定最新版本") + return + + relative_path = os.path.join("assets", "firmware", latest_file) + base_url = f"http://{settings.server_host}:{settings.server_port}/" + firmware_url = urljoin(base_url, relative_path) + + current_version_config = await system_config_manager.get_config('latest_firmware_version') + current_url_config = await system_config_manager.get_config('update_firmware_url') + + current_version = current_version_config.config_value if current_version_config else "" + current_url = current_url_config.config_value if current_url_config else "" + + if current_version != latest_version: + await system_config_manager.update_config('latest_firmware_version', latest_version) + session_logger.info("system", "firmware_scanner", f"更新了最新固件版本: {latest_version}") + + if current_url != firmware_url: + await system_config_manager.update_config('update_firmware_url', firmware_url) + session_logger.info("system", "firmware_scanner", f"更新了固件URL: {firmware_url}") + + session_logger.info("system", "firmware_scanner", f"固件扫描完成，最新版本: {latest_version}") + + except Exception as e: + session_logger.error("system", "firmware_scanner", f"扫描固件文件出错: {str(e)}") + + def _get_firmware_files(self) -> List[str]: + """获取固件目录中的所有固件文件""" + try: + all_files = os.listdir(self.firmware_dir) + firmware_files = [f for f in all_files if self.firmware_file_pattern.match(f)] + return firmware_files + except Exception as e: + session_logger.error("system", "firmware_scanner", f"获取固件文件列表出错: {str(e)}") + return [] + + def _find_latest_version(self, firmware_files: List[str]) -> Tuple[Optional[str], Optional[str]]: + """查找最新版本的固件""" + try: + version_files = [] + for file in firmware_files: + match = self.firmware_file_pattern.match(file) + if match: + version = match.group(1) + version_files.append((version, file)) + + if not version_files: + return None, None + + latest = sorted(version_files, key=lambda x: self._version_to_tuple(x[0]), reverse=True)[0] + return latest + except Exception as e: + session_logger.error("system", "firmware_scanner", f"查找最新版本出错: {str(e)}") + return None, None + + def _version_to_tuple(self, version: str) -> Tuple[int, ...]: + """将版本号转换为数字元组，用于比较""" + return tuple(map(int, version.split('.'))) + +firmware_scanner = FirmwareScanner() diff --git a/talkingq-url/services/history_service.py b/talkingq-url/services/history_service.py new file mode 100644 index 0000000..4c065fc --- /dev/null +++ b/talkingq-url/services/history_service.py @@ -0,0 +1,90 @@ +import time +from typing import Dict, List, Optional +from services.conversation_history import conversation_history_manager +from services.role_manager import role_manager + +class HistoryService: + async def get_history_paginated( + self, + device_id: str, + role_key: Optional[str] = None, + page: int = 1, + page_size: int = 20, + since: Optional[float] = None, + until: Optional[float] = None + ) -> Dict: + """分页获取会话历史""" + device_history = await conversation_history_manager.get_history(device_id, role_key) + if not device_history or not device_history.history: + return { + "total": 0, + "page": page, + "page_size": page_size, + "data": [], + "role_key": role_key or "", + "role_name": "未知角色" + } + + role_info = await role_manager.get_role(device_history.role_key) if device_history.role_key else None + role_name = role_info.get("name", "未知角色") if role_info else "未知角色" + + filtered_history = device_history.history + if since: + filtered_history = [msg for msg in filtered_history if + msg.get("timestamp", 0) >= since] + if until: + filtered_history = [msg for msg in filtered_history if + msg.get("timestamp", 0) <= until] + + total_messages = len(filtered_history) + + start_idx = (page - 1) * page_size + end_idx = start_idx + page_size + paged_messages = filtered_history[start_idx:end_idx] + + current_time = time.time() + for i, msg in enumerate(paged_messages): + if "timestamp" not in msg: + paged_messages[i]["timestamp"] = current_time - (total_messages - i) * 60 + + return { + "total": total_messages, + "page": page, + "page_size": page_size, + "data": paged_messages, + "role_key": device_history.role_key or "", + "role_name": role_name + } + + async def get_history_summary(self, device_id: str, days: int = 7) -> List[Dict]: + """获取会话历史摘要""" + role_histories = await conversation_history_manager.get_device_role_histories(device_id) + if not role_histories: + return [] + + summaries = [] + cutoff_time = time.time() - (days * 24 * 60 * 60) + + for role_key, history in role_histories.items(): + if history.last_interaction_time < cutoff_time: + continue + + role_info = await role_manager.get_role(role_key) + role_name = role_info.get("name", "未知角色") if role_info else "未知角色" + + last_message = "" + if history.history: + last_message = history.history[-1].get("assistant", "")[:100] + + summaries.append({ + "role_key": role_key, + "role_name": role_name, + "last_interaction": history.last_interaction_time, + "message_count": len(history.history), + "last_message": last_message + }) + + summaries.sort(key=lambda x: x["last_interaction"], reverse=True) + return summaries + +history_service = HistoryService() diff --git a/talkingq-url/services/interrupt_handler.py b/talkingq-url/services/interrupt_handler.py new file mode 100644 index 0000000..0e2a78b --- /dev/null +++ b/talkingq-url/services/interrupt_handler.py @@ -0,0 +1,242 @@ +import asyncio +from typing import List, Callable, Dict, Coroutine, Tuple +from utils.logger import session_logger + + +class InterruptHandler: + def __init__(self): + self.cleanup_handlers: Dict[ + tuple, List[Tuple[int, Callable[..., Coroutine]]] + ] = {} + self.interrupt_states: Dict[tuple, bool] = {} + self.DEFAULT_PRIORITY = 100 + self.PRIORITY = { + "NETWORK_CONNECTION": 10, # 网络连接最优先关闭 + "QUEUE_CLEANUP": 20, # 队列清理的优先级 + "TASK_CANCELLATION": 30, # 任务取消的优先级 + "DEFAULT": 100, # 默认优先级 + } + self.session_locks: Dict[tuple, asyncio.Lock] = ( + {} + ) # 添加锁字典，用于保护中断状态更新 + self.interrupt_processing: Dict[tuple, bool] = {} + self.pending_interrupts: Dict[tuple, int] = {} + + async def get_session_lock(self, session_key: tuple) -> asyncio.Lock: + if session_key not in self.session_locks: + self.session_locks[session_key] = asyncio.Lock() + return self.session_locks[session_key] + + async def register_session(self, session_key: tuple) -> None: + lock = await self.get_session_lock(session_key) + async with lock: + if session_key not in self.cleanup_handlers: + self.cleanup_handlers[session_key] = [] + self.interrupt_states[session_key] = False + self.interrupt_processing[session_key] = False # 初始化处理状态 + self.pending_interrupts[session_key] = 0 # 初始化挂起的中断请求计数 + device_id, session_id = session_key + session_logger.info(device_id, session_id, "已注册中断处理") + + async def register_cleanup_handler( + self, + session_key: tuple, + handler: Callable[..., Coroutine], + priority: int = None, + *args, + **kwargs, + ) -> None: + if session_key not in self.cleanup_handlers: + await self.register_session(session_key) + if priority is None: + priority = self.DEFAULT_PRIORITY + + async def wrapped_handler(): + try: + await handler(*args, **kwargs) + except Exception as e: + device_id, session_id = session_key + session_logger.error( + device_id, session_id, f"执行中断清理处理器出错: {e}" + ) + + self.cleanup_handlers[session_key].append((priority, wrapped_handler)) + + async def handle_interrupt(self, session_key: tuple) -> None: + device_id, session_id = session_key + await self.set_interrupt_state(session_key, True) # 使用异步方法设置状态 + lock = await self.get_session_lock(session_key) + async with lock: + if self.interrupt_processing.get(session_key, False): + self.pending_interrupts[session_key] = ( + self.pending_interrupts.get(session_key, 0) + 1 + ) + session_logger.info( + device_id, + session_id, + f"收到连续中断请求，当前有 {self.pending_interrupts[session_key]} 个中断正在等待", + ) + return + self.interrupt_processing[session_key] = True + handlers_count = len(self.cleanup_handlers.get(session_key, [])) + try: + session_logger.info( + device_id, session_id, f"处理中断, 开始执行 {handlers_count} 个清理任务" + ) + if session_key not in self.cleanup_handlers: + session_logger.warning( + device_id, session_id, "未找到该会话的中断处理器" + ) + return + handlers = self.cleanup_handlers[session_key] + if not handlers: + session_logger.info(device_id, session_id, "该会话没有注册的中断处理器") + return + + sorted_handlers = sorted(handlers, key=lambda x: x[0]) + priority_groups = {} + for priority, handler in sorted_handlers: + if priority not in priority_groups: + priority_groups[priority] = [] + priority_groups[priority].append(handler) + + for priority in sorted(priority_groups.keys()): + group_handlers = priority_groups[priority] + group_size = len(group_handlers) + session_logger.info( + device_id, + session_id, + f"执行优先级 {priority} 的 {group_size} 个清理任务" + ) + + timeout = min(1.0 + 0.5 * group_size, 3.0) + from services.task_manager import task_manager + tasks = [] + for handler in group_handlers: + task = await task_manager.create_task( + handler(), + device_id=device_id, + session_key=session_key, + task_type="interrupt_cleanup" + ) + tasks.append(task) + + try: + await asyncio.wait_for( + asyncio.gather(*tasks, return_exceptions=True), + timeout=timeout + ) + except asyncio.TimeoutError: + session_logger.warning( + device_id, + session_id, + f"优先级 {priority} 的清理任务执行超时({timeout:.1f}秒)" + ) + + self.cleanup_handlers[session_key] = [] + session_logger.info(device_id, session_id, "所有中断清理任务已完成") + finally: + async with lock: + self.interrupt_processing[session_key] = False + pending_count = self.pending_interrupts.get(session_key, 0) + if pending_count > 0: + self.pending_interrupts[session_key] = 0 + session_logger.info( + device_id, + session_id, + f"有 {pending_count} 个连续中断请求，立即处理", + ) + await task_manager.create_task( + self.handle_pending_interrupts(session_key, pending_count), + device_id=device_id, + session_key=session_key, + task_type="pending_interrupt" + ) + + async def handle_pending_interrupts(self, session_key: tuple, count: int) -> None: + device_id, session_id = session_key + try: + session_logger.info(device_id, session_id, f"处理 {count} 个挂起的中断请求") + await self.handle_interrupt(session_key) + except Exception as e: + session_logger.error(device_id, session_id, f"处理挂起中断请求时出错: {e}") + + async def remove_session(self, session_key: tuple) -> None: + lock = await self.get_session_lock(session_key) + async with lock: + if session_key in self.cleanup_handlers: + del self.cleanup_handlers[session_key] + if session_key in self.interrupt_states: + del self.interrupt_states[session_key] + if session_key in self.interrupt_processing: + del self.interrupt_processing[session_key] + if session_key in self.pending_interrupts: + del self.pending_interrupts[session_key] + if session_key in self.session_locks: + del self.session_locks[session_key] + device_id, session_id = session_key + session_logger.info(device_id, session_id, "已移除会话的中断处理器和状态") + + def is_interrupted(self, session_key: tuple) -> bool: + return self.interrupt_states.get(session_key, False) + + def is_processing_interrupt(self, session_key: tuple) -> bool: + return self.interrupt_processing.get(session_key, False) + + async def set_interrupt_state( + self, session_key: tuple, state: bool + ) -> None: + if session_key not in self.session_locks: + await self.register_session(session_key) + lock = await self.get_session_lock(session_key) + async with lock: + old_state = self.interrupt_states.get(session_key, False) + if old_state != state: + self.interrupt_states[session_key] = state + device_id, session_id = session_key + status = "中断" if state else "正常" + session_logger.info(device_id, session_id, f"会话状态已设置为{status}") + else: + self.interrupt_states[session_key] = state + + async def notify_client_interrupt_processed( + self, session_key: tuple, websocket + ) -> bool: + if not websocket: + return False + device_id, session_id = session_key + max_retries = 3 + retry_count = 0 + while retry_count < max_retries: + try: + if websocket.client_state.name == "CONNECTED" and not getattr( + websocket, "_closed", False + ): + await websocket.send_text("INTERRUPT_PROCESSED") + session_logger.info( + device_id, session_id, "已通知客户端中断处理完成" + ) + return True + else: + session_logger.warning( + device_id, session_id, "WebSocket连接已关闭，无法通知客户端" + ) + return False + except Exception as e: + retry_count += 1 + if retry_count < max_retries: + session_logger.warning( + device_id, + session_id, + f"通知客户端失败，尝试重试 ({retry_count}/{max_retries}): {e}", + ) + await asyncio.sleep(0.2) + else: + session_logger.error( + device_id, session_id, f"通知客户端失败，已达最大重试次数: {e}" + ) + return False + return False + + +interrupt_handler = InterruptHandler() diff --git a/talkingq-url/services/interruption_helper.py b/talkingq-url/services/interruption_helper.py new file mode 100644 index 0000000..270ae72 --- /dev/null +++ b/talkingq-url/services/interruption_helper.py @@ -0,0 +1,78 @@ +import asyncio +from utils.logger import session_logger +from services.interrupt_handler import interrupt_handler +from services.tts_error_manager import tts_error_manager + + +class InterruptionHelper: + + + @staticmethod + async def notify_client_tts_error(device_id: str, session_id: str): + + try: + if not await tts_error_manager.can_send_error(device_id, session_id): + session_logger.info(device_id, session_id, "已经发送过TTS_ERROR，不再重复发送") + return False + + from services.connection_manager import connection_manager + + websocket = await connection_manager.get_connection(device_id) + if websocket and websocket.client_state.name == "CONNECTED": + await websocket.send_text("TTS_ERROR") + session_logger.info(device_id, session_id, "已通知客户端TTS错误") + return True + except Exception as e: + session_logger.error(device_id, session_id, f"发送TTS错误通知失败: {e}") + return False + + @staticmethod + async def send_error_prompt_sound(device_id: str, session_id: str): + + try: + from handlers.prompt_sound_handler import handle_prompt_sound_request + + await handle_prompt_sound_request(device_id, "tts_error") + session_logger.info(device_id, session_id, "已发送错误提示音") + return True + except Exception as e: + session_logger.error(device_id, session_id, f"发送错误提示音失败: {e}") + return False + + @staticmethod + async def handle_llm_timeout(device_id: str, session_id: str, llm_service=None): + + session_key = (device_id, session_id) + await interrupt_handler.set_interrupt_state(session_key, True) + if llm_service and hasattr(llm_service, "closed"): + llm_service.closed = True + await InterruptionHelper.send_error_prompt_sound(device_id, session_id) + await InterruptionHelper.notify_client_tts_error(device_id, session_id) + session_logger.error(device_id, session_id, "已处理LLM超时") + + @staticmethod + async def register_queue_cleanup(session_key: tuple, queue: asyncio.Queue): + + if queue is None: + return + + async def cleanup_queue(): + try: + await queue.put(None) + while not queue.empty(): + try: + queue.get_nowait() + queue.task_done() + except asyncio.QueueEmpty: + break + device_id, session_id = session_key + session_logger.info(device_id, session_id, "队列已清空") + except Exception as e: + device_id, session_id = session_key + session_logger.error(device_id, session_id, f"清空队列时出错: {e}") + + await interrupt_handler.register_cleanup_handler( + session_key, + cleanup_queue, + priority=interrupt_handler.PRIORITY["QUEUE_CLEANUP"], + ) diff --git a/talkingq-url/services/ota_flow_controller.py b/talkingq-url/services/ota_flow_controller.py new file mode 100644 index 0000000..d1bbf67 --- /dev/null +++ b/talkingq-url/services/ota_flow_controller.py @@ -0,0 +1,116 @@ +import asyncio +from typing import Dict, Optional, Any +from utils.logger import session_logger +from services.device_update_manager import device_firmware_update_manager +from services.system_config_manager import system_config_manager +from services.connection_manager import connection_manager + +class OTAFlowController: + def __init__(self): + self.active_updates: Dict[str, Dict[str, Any]] = {} # 跟踪活跃的更新流程 + + def compare_versions(self, current_version: str, latest_version: str) -> bool: + """ + 比较版本号，判断是否需要更新 + Args: + current_version: 当前版本号 (格式: x.y.z) + latest_version: 最新版本号 (格式: x.y.z) + Returns: + bool: 如果需要更新返回True，否则返回False + """ + try: + if not current_version or current_version == "unknown" or current_version == "0.0.0": + return True + + current_parts = [int(x) for x in current_version.split('.')] + latest_parts = [int(x) for x in latest_version.split('.')] + + while len(current_parts) < len(latest_parts): + current_parts.append(0) + while len(latest_parts) < len(current_parts): + latest_parts.append(0) + + for i in range(len(current_parts)): + if latest_parts[i] > current_parts[i]: + return True + elif latest_parts[i] < current_parts[i]: + return False + + return False # 版本相同，不需要更新 + except Exception as e: + session_logger.error("system", "ota_flow", f"版本比较出错: {str(e)}") + return True # 出错时默认需要更新 + + async def start_update_flow(self, device_id: str) -> Dict[str, Any]: + """启动更新流程""" + request_success = await device_firmware_update_manager.request_firmware_version(device_id) + if not request_success: + websocket = await connection_manager.get_connection(device_id) + return {"status": "error", "message": "请求设备固件版本失败"} + + for _ in range(10): + await asyncio.sleep(0.5) + + latest_version_config = await system_config_manager.get_config('latest_firmware_version') + if not latest_version_config: + return {"status": "error", "message": "未找到最新版本信息"} + + latest_version = latest_version_config.config_value + + current_version = "0.0.0" + device_update = await device_firmware_update_manager.get_firmware_update(device_id) + if device_update: + current_version = device_update.firmware_version + + need_update = self.compare_versions(current_version, latest_version) + + self.active_updates[device_id] = { + "need_update": need_update, + "status": "checked", + "current_version": current_version, + "latest_version": latest_version + } + + return { + "status": "success", + "need_update": need_update, + "currentVersion": current_version, + "latestVersion": latest_version + } + + async def execute_update(self, device_id: str) -> Dict[str, Any]: + """执行更新""" + update_url_config = await system_config_manager.get_config('update_firmware_url') + if not update_url_config: + return {"status": "error", "message": "未找到更新URL"} + update_url = update_url_config.config_value + websocket = await connection_manager.get_connection(device_id) + if not websocket: + return {"status": "error", "message": "设备未连接"} + try: + await websocket.send_text(f"UPDATE_FIRMWARE:{update_url}") + session_logger.info(device_id, "ota_flow", f"已发送固件URL到设备: {update_url}") + + await device_firmware_update_manager.update_firmware_update( + device_id, + firmware_version="updating", # 临时版本标记 + update_status="updating" + ) + await device_firmware_update_manager.update_firmware_progress(device_id, 0.0) + return {"status": "success", "message": "更新已启动"} + except Exception as e: + session_logger.error(device_id, "ota_flow", f"执行更新过程中出错: {str(e)}") + return {"status": "error", "message": str(e)} + + async def get_update_status(self, device_id: str) -> Dict[str, Any]: + """获取更新状态""" + device_update = await device_firmware_update_manager.get_firmware_update(device_id) + if not device_update: + return {"status": "unknown", "progress": 0, "version": "unknown"} + return { + "status": device_update.update_status, + "progress": device_update.progress or 0.0, + "version": device_update.firmware_version + } + +ota_flow_controller = OTAFlowController() diff --git a/talkingq-url/services/registry.py b/talkingq-url/services/registry.py new file mode 100644 index 0000000..d403c26 --- /dev/null +++ b/talkingq-url/services/registry.py @@ -0,0 +1,17 @@ +from typing import Dict, Any + +SERVICE_REGISTRY: Dict[str, Dict[str, Any]] = {} + + +def register_service(service_type: str, provider_name: str, implementation: Any): + SERVICE_REGISTRY.setdefault(service_type, {})[provider_name] = implementation + + +def get_service(service_type: str, provider_name: str): + provider_dict = SERVICE_REGISTRY.get(service_type, {}) + service_class = provider_dict.get(provider_name) + if service_class is None: + raise ValueError( + f"未找到服务类型 '{service_type}' 和提供商 '{provider_name}' 对应的实现。" + ) + return service_class diff --git a/talkingq-url/services/role_manager.py b/talkingq-url/services/role_manager.py new file mode 100644 index 0000000..b5de68e --- /dev/null +++ b/talkingq-url/services/role_manager.py @@ -0,0 +1,224 @@ +import os +from typing import Dict, Optional, List +import asyncio +import time +from config import settings +from utils.logger import session_logger +from services.role_validator import role_validator +from sqlalchemy import select, and_ +from services.database_service_base import DatabaseServiceBase +from database.models import Role, RoleLanguage + +class RoleManager(DatabaseServiceBase): + def __init__(self): + super().__init__(service_name="role_manager") + self.roles: Dict[str, Dict[str, str]] = {} + self.invalid_roles: Dict[str, List[str]] = {} + self.lock = asyncio.Lock() + self._initialized = False + self.last_refresh_time = 0 + self.refresh_interval = 300 # 缓存刷新间隔，单位秒 + self.etags = {} # 角色配置的ETag缓存 + + async def initialize(self): + if self._initialized: + return + worker_id = os.environ.get("UVICORN_WID", "0") + is_main_process = worker_id == "0" + async with self.lock: + if self._initialized: # 双重检查锁定 + return + try: + await self._init_database() + await self._load_roles_from_db() + self._initialized = True + self.last_refresh_time = time.time() + if is_main_process: + session_logger.info("system", "role_manager", "角色管理器初始化完成") + except Exception as e: + if is_main_process: + session_logger.error("system", "role_manager", f"角色管理器初始化失败: {str(e)}") + + async def _maybe_refresh_roles(self): + """如果缓存过期，刷新角色配置""" + current_time = time.time() + if current_time - self.last_refresh_time > self.refresh_interval: + await self.reload_roles() + + async def _load_roles_from_db(self): + """从数据库加载所有角色配置""" + self.roles = {} + db_session = await self.get_session() + try: + roles_query = select(Role).where(Role.enabled == True) + roles_result = await db_session.execute(roles_query) + db_roles = roles_result.scalars().all() + + new_etags = {} + + for db_role in db_roles: + role_dict = self._role_to_dict(db_role) + role_key = db_role.role_key.lower() + + languages_query = select(RoleLanguage).where(RoleLanguage.role_id == db_role.id) + languages_result = await db_session.execute(languages_query) + lang_configs = languages_result.scalars().all() + + if lang_configs: + multilingual = {} + for lang_config in lang_configs: + lang_dict = self._role_lang_to_dict(lang_config) + multilingual[lang_config.language_code] = lang_dict + + role_dict["multilingual"] = multilingual + + self.roles[role_key] = role_dict + + import hashlib + import json + role_json = json.dumps(role_dict, sort_keys=True) + etag = hashlib.md5(role_json.encode()).hexdigest() + new_etags[role_key] = etag + + self.etags = new_etags + + except Exception as e: + session_logger.error("system", "role_manager", f"从数据库加载角色配置失败: {str(e)}") + raise + finally: + await db_session.close() + + def _role_to_dict(self, db_role: Role) -> Dict: + """将数据库Role对象转换为字典格式""" + role_dict = { + "role_key": db_role.role_key, + "name": db_role.name, + "content": db_role.content, + } + if db_role.description: + role_dict["description"] = db_role.description + if db_role.default_language: + role_dict["default_language"] = db_role.default_language + if db_role.volcano_model_id: + role_dict["volcano_model_id"] = db_role.volcano_model_id + if db_role.minimax_voice_id: + role_dict["minimax_voice_id"] = db_role.minimax_voice_id + if db_role.url: + role_dict["url"] = db_role.url + if db_role.homophones: + role_dict["homophones"] = db_role.homophones + return role_dict + + def _role_lang_to_dict(self, lang_config: RoleLanguage) -> Dict: + """将数据库RoleLanguage对象转换为字典格式""" + lang_dict = {} + if lang_config.name: + lang_dict["name"] = lang_config.name + if lang_config.content: + lang_dict["content"] = lang_config.content + if lang_config.minimax_voice_id: + lang_dict["minimax_voice_id"] = lang_config.minimax_voice_id + if lang_config.url: + lang_dict["url"] = lang_config.url + return lang_dict + + async def get_role(self, role_key: str) -> Optional[Dict[str, str]]: + if not self._initialized: + await self.initialize() + await self._maybe_refresh_roles() + return self.roles.get(role_key.lower()) + + async def get_role_etag(self, role_key: str) -> Optional[str]: + """获取角色配置的ETag""" + if not self._initialized: + await self.initialize() + return self.etags.get(role_key.lower()) + + async def get_role_config_for_language( + self, role_key: str, language: str = None + ) -> Optional[Dict[str, str]]: + """ + 根据指定的语言获取角色配置 + Args: + role_key (str): 角色键名 + language (str, optional): 语言代码。如果为None，使用默认语言 + Returns: + Dict: 角色配置 + """ + if not self._initialized: + await self.initialize() + await self._maybe_refresh_roles() + role_config = self.roles.get(role_key.lower()) + if not role_config: + return None + full_config = role_config.copy() + full_config["role_key"] = role_key + if "multilingual" not in role_config: + return full_config + multilingual = role_config.get("multilingual", {}) + default_language = role_config.get("default_language", "zh") + selected_language = None + if language and language in multilingual: + selected_language = language + session_logger.info( + "system", + "role_manager", + f"角色 {role_key} 使用指定的语言: {selected_language}", + ) + elif default_language in multilingual: + selected_language = default_language + session_logger.info( + "system", + "role_manager", + f"角色 {role_key} 使用默认语言: {selected_language}", + ) + elif multilingual: + selected_language = next(iter(multilingual)) + session_logger.info( + "system", + "role_manager", + f"角色 {role_key} 无法使用指定或默认语言，使用第一个可用语言: {selected_language}", + ) + else: + return full_config # 如果没有多语言配置，返回原始配置 + lang_specific_config = multilingual[selected_language] + for key, value in lang_specific_config.items(): + full_config[key] = value + service_keys = ["tts_provider", "llm_provider", "asr_provider"] + for key in service_keys: + if key in lang_specific_config: + full_config[key] = lang_specific_config[key] + + full_config["_selected_language"] = selected_language + session_logger.info( + "system", + "role_manager", + f"已加载角色 {role_key} 的 {selected_language} 语言配置", + ) + return full_config + + async def get_role_errors(self, role_key: str) -> List[str]: + """获取角色验证错误信息""" + if not self._initialized: + await self.initialize() + await self._maybe_refresh_roles() + return self.invalid_roles.get(role_key.lower(), []) + + async def get_all_roles(self) -> Dict[str, Dict[str, str]]: + if not self._initialized: + await self.initialize() + await self._maybe_refresh_roles() + return self.roles + + async def reload_roles(self): + """重新加载所有角色配置""" + async with self.lock: + try: + await self._init_database() + await self._load_roles_from_db() + self.last_refresh_time = time.time() + session_logger.info("system", "role_manager", "角色配置已重新加载") + except Exception as e: + session_logger.error("system", "role_manager", f"重新加载角色配置失败: {str(e)}") + +role_manager = RoleManager() diff --git a/talkingq-url/services/role_service.py b/talkingq-url/services/role_service.py new file mode 100644 index 0000000..c6482bb --- /dev/null +++ b/talkingq-url/services/role_service.py @@ -0,0 +1,56 @@ +from typing import Dict, List, Optional +from services.role_manager import role_manager + +class RoleService: + async def get_roles_paginated(self, page: int, page_size: int, search: Optional[str] = None): + """分页获取角色列表，支持搜索""" + roles = await role_manager.get_all_roles() + + filtered_roles = {} + for key, config in roles.items(): + if search: + name = config.get("name", "") + description = config.get("description", "") + if search.lower() not in name.lower() and search.lower() not in description.lower(): + continue + filtered_roles[key] = config + + total_roles = len(filtered_roles) + + start_idx = (page - 1) * page_size + end_idx = start_idx + page_size + + role_items = list(filtered_roles.items()) + paged_roles = role_items[start_idx:end_idx] + + result = [] + for key, config in paged_roles: + languages = [] + if "multilingual" in config: + languages = list(config["multilingual"].keys()) + result.append({ + "role_key": key, + "name": config.get("name", "未命名角色"), + "description": config.get("description", ""), + "languages": languages + }) + + return { + "total": total_roles, + "page": page, + "page_size": page_size, + "data": result + } + + async def get_role_summaries(self) -> List[Dict]: + """获取角色简要信息列表（仅包含key和name）""" + roles = await role_manager.get_all_roles() + return [ + { + "role_key": key, + "name": config.get("name", "未命名角色") + } + for key, config in roles.items() + ] + +role_service = RoleService() diff --git a/talkingq-url/services/role_validator.py b/talkingq-url/services/role_validator.py new file mode 100644 index 0000000..1f5b84e --- /dev/null +++ b/talkingq-url/services/role_validator.py @@ -0,0 +1,70 @@ +from typing import Dict, Any, List, Tuple +from utils.logger import session_logger + +class RoleValidator: + """角色配置验证服务，确保角色定义满足要求""" + @staticmethod + def validate_role_config( + role_key: str, config: Dict[str, Any] + ) -> Tuple[bool, List[str]]: + """ + 验证角色配置 + Returns: + bool: 是否验证通过 + list: 错误消息列表 + """ + errors = [] + + bool_fields = ["competitive_llm_mode"] + for field in bool_fields: + if field in config and not isinstance(config[field], bool): + errors.append(f"{field} 字段必须是布尔值") + + if "volcano_model_id" not in config: + pass + if "minimax_voice_id" not in config: + pass + + if errors: + session_logger.warning( + "system", "role_validator", f"角色 {role_key} 验证失败: {', '.join(errors)}" + ) + return False, errors + return True, [] + + @staticmethod + def validate_role_create(role_data: Dict[str, Any]) -> Tuple[bool, List[str]]: + """ + 验证创建角色的数据 + """ + role_key = role_data.get("role_key", "") + config = { + "name": role_data.get("name"), + "content": role_data.get("content"), + "description": role_data.get("description"), + "default_language": role_data.get("default_language"), + "competitive_llm_mode": role_data.get("competitive_llm_mode"), + "volcano_model_id": role_data.get("volcano_model_id"), + "minimax_voice_id": role_data.get("minimax_voice_id"), + "url": role_data.get("url"), + "homophones": role_data.get("homophones"), + } + + languages = role_data.get("languages", []) + if languages: + multilingual = {} + for lang in languages: + lang_code = lang.get("language_code") + multilingual[lang_code] = { + "name": lang.get("name"), + "content": lang.get("content"), + "minimax_voice_id": lang.get("minimax_voice_id"), + "url": lang.get("url") + } + if multilingual: + config["multilingual"] = multilingual + + config = {k: v for k, v in config.items() if v is not None} + return RoleValidator.validate_role_config(role_key, config) + +role_validator = RoleValidator() diff --git a/talkingq-url/services/schedule_session_cleanup.py b/talkingq-url/services/schedule_session_cleanup.py new file mode 100644 index 0000000..b0609c9 --- /dev/null +++ b/talkingq-url/services/schedule_session_cleanup.py @@ -0,0 +1,57 @@ +import asyncio +from config import settings +from services.audio_session import audio_session_manager +from utils.logger import session_logger +from services.conversation_history import conversation_history_manager + +async def run_in_executor(func, *args): + loop = asyncio.get_running_loop() + return await loop.run_in_executor(None, func, *args) + +async def background_cleanup_task(): + try: + while True: + await asyncio.sleep(settings.cleanup_interval) + + current_time = asyncio.get_running_loop().time() + audio_sessions_to_clean = [] + + sessions = await audio_session_manager.get_all_sessions() + for session_key, session in sessions: + device_id, session_id = session_key + if current_time - session.last_activity_time > settings.session_timeout: + audio_sessions_to_clean.append(session_key) + session_logger.info( + device_id, + "cleanup", + f"清理过期的音频会话, 超时: {settings.session_timeout}秒" + ) + + for session_key in audio_sessions_to_clean: + await audio_session_manager.remove_session(session_key) + + role_histories_to_clean = [] + + for (device_id, role_key), history in conversation_history_manager.role_histories.items(): + if current_time - history.last_interaction_time > settings.conversation_history_timeout: + role_histories_to_clean.append((device_id, role_key)) + session_logger.info( + device_id, + "cleanup", + f"发现过期的对话历史缓存: {role_key}, 超时: {settings.conversation_history_timeout}秒" + ) + + for device_id, role_key in role_histories_to_clean: + cache_key = (device_id, role_key) + if cache_key in conversation_history_manager.role_histories: + del conversation_history_manager.role_histories[cache_key] + session_logger.info( + device_id, + "cleanup", + f"已从内存中清理对话历史: {role_key}" + ) + + except asyncio.CancelledError: + session_logger.info("system", "cleanup", "后台清理任务已取消") + except Exception as e: + session_logger.error("system", "cleanup", f"后台清理任务出错: {str(e)}") \ No newline at end of file diff --git a/talkingq-url/services/system_config_manager.py b/talkingq-url/services/system_config_manager.py new file mode 100644 index 0000000..1ff6aab --- /dev/null +++ b/talkingq-url/services/system_config_manager.py @@ -0,0 +1,111 @@ +import asyncio +from typing import Optional +import time +from sqlalchemy import select, update, insert, delete, func +from database.models import SystemConfig +from services.database_service_base import DatabaseServiceBase +from utils.logger import session_logger + +class SystemConfigManager(DatabaseServiceBase): + def __init__(self): + super().__init__(service_name="system_config") + + async def create_config(self, key: str, value: str) -> bool: + """新增系统配置""" + await self._init_database() + db_session = await self.get_session() + try: + insert_stmt = insert(SystemConfig).values( + config_key=key, + config_value=value + ) + await db_session.execute(insert_stmt) + await db_session.commit() + session_logger.info("system", "system_config", f"新增系统配置: {key}={value}") + return True + except Exception as e: + await db_session.rollback() + session_logger.error("system", "system_config", f"新增系统配置失败: {str(e)}") + return False + finally: + await db_session.close() + + async def get_config(self, key: str) -> Optional[SystemConfig]: + """获取系统配置""" + await self._init_database() + db_session = await self.get_session() + try: + query = select(SystemConfig).where(SystemConfig.config_key == key) + result = await db_session.execute(query) + return result.scalars().first() + except Exception as e: + session_logger.error(key, "system_config", f"查询系统配置失败: {str(e)}") + return None + finally: + await db_session.close() + + async def update_config(self, key: str, value: str) -> bool: + """更新系统配置""" + await self._init_database() + db_session = await self.get_session() + try: + query = select(SystemConfig).where(SystemConfig.config_key == key) + result = await db_session.execute(query) + config = result.scalars().first() + + if config: + update_stmt = ( + update(SystemConfig) + .where(SystemConfig.config_key == key) + .values(config_value=value, updated_at=func.now()) + ) + await db_session.execute(update_stmt) + else: + insert_stmt = insert(SystemConfig).values( + config_key=key, + config_value=value + ) + await db_session.execute(insert_stmt) + + await db_session.commit() + session_logger.info(key, "system_config", f"系统配置已更新: {value}") + return True + except Exception as e: + await db_session.rollback() + session_logger.error(key, "system_config", f"更新系统配置失败: {str(e)}") + return False + finally: + await db_session.close() + + async def delete_config(self, key: str) -> bool: + """删除系统配置""" + await self._init_database() + db_session = await self.get_session() + try: + delete_stmt = delete(SystemConfig).where(SystemConfig.config_key == key) + await db_session.execute(delete_stmt) + await db_session.commit() + session_logger.info(key, "system_config", "系统配置已删除") + return True + except Exception as e: + await db_session.rollback() + session_logger.error(key, "system_config", f"删除系统配置失败: {str(e)}") + return False + finally: + await db_session.close() + + async def list_configs(self): + """列出所有系统配置""" + await self._init_database() + db_session = await self.get_session() + try: + query = select(SystemConfig) + result = await db_session.execute(query) + return result.scalars().all() + except Exception as e: + session_logger.error("system_config", "system_config", f"查询所有系统配置失败: {str(e)}") + return [] + finally: + await db_session.close() + +system_config_manager = SystemConfigManager() diff --git a/talkingq-url/services/task_manager.py b/talkingq-url/services/task_manager.py new file mode 100644 index 0000000..d148535 --- /dev/null +++ b/talkingq-url/services/task_manager.py @@ -0,0 +1,154 @@ +import asyncio +import weakref +from typing import Dict, Set, Optional +from utils.logger import session_logger + + +class TaskManager: + """管理异步任务，防止内存泄露""" + + def __init__(self): + self.device_tasks: Dict[str, Set[asyncio.Task]] = {} + self.session_tasks: Dict[tuple, Set[asyncio.Task]] = {} + self.cleanup_tasks: Set[asyncio.Task] = set() + self.lock = asyncio.Lock() + + async def create_task(self, coro, device_id: Optional[str] = None, session_key: Optional[tuple] = None, task_type: str = "general"): + """创建并跟踪异步任务""" + task = asyncio.create_task(coro) + + async with self.lock: + # 按设备ID跟踪 + if device_id: + if device_id not in self.device_tasks: + self.device_tasks[device_id] = set() + self.device_tasks[device_id].add(task) + + # 按会话跟踪 + if session_key: + if session_key not in self.session_tasks: + self.session_tasks[session_key] = set() + self.session_tasks[session_key].add(task) + + # 跟踪清理任务 + if task_type == "cleanup": + self.cleanup_tasks.add(task) + + # 任务完成后自动清理 + task.add_done_callback(lambda t: asyncio.create_task(self._cleanup_completed_task(t, device_id, session_key, task_type))) + + session_logger.info( + device_id or "system", + session_key[1] if session_key else "task_manager", + f"创建任务: {task_type}, 总任务数: {len(self.cleanup_tasks) if task_type == 'cleanup' else 'tracked'}" + ) + + return task + + async def _cleanup_completed_task(self, task: asyncio.Task, device_id: Optional[str], session_key: Optional[tuple], task_type: str): + """清理已完成的任务""" + async with self.lock: + if device_id and device_id in self.device_tasks: + self.device_tasks[device_id].discard(task) + if not self.device_tasks[device_id]: + del self.device_tasks[device_id] + + if session_key and session_key in self.session_tasks: + self.session_tasks[session_key].discard(task) + if not self.session_tasks[session_key]: + del self.session_tasks[session_key] + + if task_type == "cleanup": + self.cleanup_tasks.discard(task) + + async def cancel_device_tasks(self, device_id: str): + """取消设备的所有任务""" + async with self.lock: + if device_id in self.device_tasks: + tasks = list(self.device_tasks[device_id]) + session_logger.info(device_id, "task_manager", f"取消设备任务数量: {len(tasks)}") + + for task in tasks: + if not task.done(): + task.cancel() + + # 等待任务取消完成 + if tasks: + try: + await asyncio.wait_for(asyncio.gather(*tasks, return_exceptions=True), timeout=5.0) + except asyncio.TimeoutError: + session_logger.warning(device_id, "task_manager", "部分任务取消超时") + + del self.device_tasks[device_id] + + async def cancel_session_tasks(self, session_key: tuple): + """取消会话的所有任务""" + async with self.lock: + if session_key in self.session_tasks: + tasks = list(self.session_tasks[session_key]) + device_id, session_id = session_key + session_logger.info(device_id, session_id, f"取消会话任务数量: {len(tasks)}") + + for task in tasks: + if not task.done(): + task.cancel() + + # 等待任务取消完成 + if tasks: + try: + await asyncio.wait_for(asyncio.gather(*tasks, return_exceptions=True), timeout=3.0) + except asyncio.TimeoutError: + session_logger.warning(device_id, session_id, "部分会话任务取消超时") + + del self.session_tasks[session_key] + + async def get_task_stats(self) -> dict: + """获取任务统计信息""" + async with self.lock: + device_count = sum(len(tasks) for tasks in self.device_tasks.values()) + session_count = sum(len(tasks) for tasks in self.session_tasks.values()) + + return { + "device_tasks": device_count, + "session_tasks": session_count, + "cleanup_tasks": len(self.cleanup_tasks), + "total_tracked_devices": len(self.device_tasks), + "total_tracked_sessions": len(self.session_tasks) + } + + async def cleanup_all_tasks(self): + """清理所有任务""" + async with self.lock: + all_tasks = [] + + # 收集所有任务 + for tasks in self.device_tasks.values(): + all_tasks.extend(tasks) + for tasks in self.session_tasks.values(): + all_tasks.extend(tasks) + all_tasks.extend(self.cleanup_tasks) + + session_logger.info("system", "task_manager", f"开始清理所有任务，总数: {len(all_tasks)}") + + # 取消所有任务 + for task in all_tasks: + if not task.done(): + task.cancel() + + # 等待任务取消完成 + if all_tasks: + try: + await asyncio.wait_for(asyncio.gather(*all_tasks, return_exceptions=True), timeout=10.0) + except asyncio.TimeoutError: + session_logger.warning("system", "task_manager", "部分任务清理超时") + + # 清空所有跟踪 + self.device_tasks.clear() + self.session_tasks.clear() + self.cleanup_tasks.clear() + + session_logger.info("system", "task_manager", "任务管理器清理完成") + + +# 全局任务管理器实例 +task_manager = TaskManager() \ No newline at end of file diff --git a/talkingq-url/services/text_generator.py b/talkingq-url/services/text_generator.py new file mode 100644 index 0000000..8f905ff --- /dev/null +++ b/talkingq-url/services/text_generator.py @@ -0,0 +1,153 @@ +import asyncio +import time +from typing import Dict, List +from utils.logger import session_logger +from services.interrupt_handler import interrupt_handler +from utils.text_splitter import ( + PUNCTUATION_MARKS, + SENTENCE_ENDINGS, + is_numbered_list_item, + split_into_sentences, +) +from services.tts_config import TTSConfig +from services.interruption_helper import InterruptionHelper + + +class TextGenerator: + + + def __init__(self, device_id: str, session_id: str): + self.device_id = device_id + self.session_id = session_id + self.session_key = (device_id, session_id) + + def check_interruption(self): + + return interrupt_handler.is_interrupted(self.session_key) + + async def generate_text( + self, + llm_service, + transcript: str, + history: List[Dict[str, str]], + selected_role: dict, + text_queue: asyncio.Queue, + ) -> str: + + reply = "" + text_buffer = "" + first_text_token_time = None + sent_batches_count = 0 + first_punct_sent = False + initial_batches_count = 0 + llm_first_token_timeout = TTSConfig.get_llm_first_token_timeout() + first_token_timeout = False + start_time = time.perf_counter() + + async def check_timeout(): + nonlocal first_token_timeout + while True: + await asyncio.sleep(0.5) # 每0.5秒检查一次 + if first_text_token_time is not None or first_token_timeout: + break # 已收到首个令牌或已超时，停止检查 + if time.perf_counter() - start_time > llm_first_token_timeout: + session_logger.error( + self.device_id, + self.session_id, + f"LLM首个token超时({llm_first_token_timeout}秒)，主动中断", + ) + await InterruptionHelper.handle_llm_timeout( + self.device_id, self.session_id, llm_service + ) + break + + try: + token_stream = llm_service.generate_response_stream( + transcript, history, selected_role + ) + try: + first_token = None + + async def get_first_token(): + nonlocal first_token + async for token in token_stream: + first_token = token + return token + + await asyncio.wait_for(get_first_token(), llm_first_token_timeout) + if first_token: + first_text_token_time = time.perf_counter() + token_latency = first_text_token_time - start_time + session_logger.info( + self.device_id, + self.session_id, + f"收到第一个文本 Token 耗时: {token_latency:.3f}秒" + ) + if first_token is None: + raise Exception("LLM返回的首个token为None") + reply += first_token + text_buffer += first_token + async for token in token_stream: + if self.check_interruption(): + session_logger.info( + self.device_id, self.session_id, "LLM生成被中断" + ) + break + if token is None: + continue + reply += token + text_buffer += token + should_submit = False + if initial_batches_count < 3 and any( + p in token for p in PUNCTUATION_MARKS + ): + if not is_numbered_list_item(text_buffer): + should_submit = True + elif initial_batches_count >= 3 and any( + p in token for p in SENTENCE_ENDINGS + ): + if not is_numbered_list_item(text_buffer): + sentences = split_into_sentences(text_buffer) + batch_size = 2 + if len(sentences) >= batch_size: + to_synthesize = "".join(sentences[:batch_size]) + text_buffer = "".join(sentences[batch_size:]) + await text_queue.put(to_synthesize) + initial_batches_count += 1 + sent_batches_count += 1 + continue + if should_submit: + await text_queue.put(text_buffer) + text_buffer = "" + if not first_punct_sent: + first_punct_sent = True + initial_batches_count += 1 + sent_batches_count += 1 + except asyncio.TimeoutError: + first_token_timeout = True + session_logger.error( + self.device_id, + self.session_id, + f"LLM首个token超时({llm_first_token_timeout}秒)，中断生成", + ) + await InterruptionHelper.handle_llm_timeout( + self.device_id, self.session_id, llm_service + ) + await text_queue.put(None) + return reply + if ( + not self.check_interruption() + and not first_token_timeout + and text_buffer.strip() + ): + await text_queue.put(text_buffer.strip()) + await text_queue.put(None) + except Exception as e: + session_logger.error( + self.device_id, self.session_id, f"文本生成时出错: {str(e)}" + ) + await InterruptionHelper.send_error_prompt_sound( + self.device_id, self.session_id + ) + await text_queue.put(None) + return reply diff --git a/talkingq-url/services/tts_audio_cleaner.py b/talkingq-url/services/tts_audio_cleaner.py new file mode 100644 index 0000000..920e686 --- /dev/null +++ b/talkingq-url/services/tts_audio_cleaner.py @@ -0,0 +1,28 @@ +import os +from pathlib import Path +from config import settings +import asyncio + + +class TTSAudioCleaner: + + + @staticmethod + async def prepare_output_directory(): + + tts_dir = Path(settings.assets_dir) / "tts_audio" + loop = asyncio.get_running_loop() + await loop.run_in_executor(None, lambda: os.makedirs(tts_dir, exist_ok=True)) + return tts_dir + + @staticmethod + def generate_filename(device_id: str, session_id: str, sequence_number: int): + seq_str = f"{sequence_number:02d}" + file_name = f"{device_id}_{session_id}_{seq_str}" + output_file_prefix = os.path.join(settings.assets_dir, "tts_audio", file_name) + return output_file_prefix + + @staticmethod + def generate_audio_url(audio_file_path: str): + file_name = os.path.basename(audio_file_path) + return f"assets/tts_audio/{file_name}" diff --git a/talkingq-url/services/tts_config.py b/talkingq-url/services/tts_config.py new file mode 100644 index 0000000..14ba2ba --- /dev/null +++ b/talkingq-url/services/tts_config.py @@ -0,0 +1,35 @@ +from config import settings + + +class TTSConfig: + + + @staticmethod + def get_request_timeout(): + + return getattr(settings, "tts_request_timeout", 3) + + @staticmethod + def get_llm_first_token_timeout(): + + return getattr(settings, "llm_first_token_timeout", 5) + + @staticmethod + def get_max_tts_errors(): + + return 3 + + @staticmethod + def get_tts_audio_expiry(): + + return getattr(settings, "tts_audio_expiry", 3600) + + @staticmethod + def get_max_storage_mb(): + + return getattr(settings, "max_tts_storage_mb", 500) + + @staticmethod + def get_audio_url_base(): + + return f"http://{settings.server_host}:{settings.server_port}/assets/tts_audio/" diff --git a/talkingq-url/services/tts_error_manager.py b/talkingq-url/services/tts_error_manager.py new file mode 100644 index 0000000..37ee39f --- /dev/null +++ b/talkingq-url/services/tts_error_manager.py @@ -0,0 +1,38 @@ +import asyncio +from typing import Set, Tuple +from utils.logger import session_logger + +class TTSErrorManager: + """管理TTS错误状态，确保TTS_START和TTS_END之间最多只发送一次TTS_ERROR""" + def __init__(self): + self.error_sent_sessions: Set[Tuple[str, str]] = set() + self.active_tts_sessions: Set[Tuple[str, str]] = set() + self.lock = asyncio.Lock() + + async def start_tts_session(self, device_id: str, session_id: str): + """标记TTS会话开始，重置错误状态""" + session_key = (device_id, session_id) + async with self.lock: + self.active_tts_sessions.add(session_key) + if session_key in self.error_sent_sessions: + self.error_sent_sessions.remove(session_key) + + async def end_tts_session(self, device_id: str, session_id: str): + """标记TTS会话结束，清理状态""" + session_key = (device_id, session_id) + async with self.lock: + if session_key in self.active_tts_sessions: + self.active_tts_sessions.remove(session_key) + if session_key in self.error_sent_sessions: + self.error_sent_sessions.remove(session_key) + + async def can_send_error(self, device_id: str, session_id: str) -> bool: + """检查是否可以发送错误通知""" + session_key = (device_id, session_id) + async with self.lock: + if session_key in self.active_tts_sessions and session_key not in self.error_sent_sessions: + self.error_sent_sessions.add(session_key) + return True + return False + +tts_error_manager = TTSErrorManager() diff --git a/talkingq-url/services/tts_synthesizer.py b/talkingq-url/services/tts_synthesizer.py new file mode 100644 index 0000000..02435e5 --- /dev/null +++ b/talkingq-url/services/tts_synthesizer.py @@ -0,0 +1,101 @@ +import asyncio +import time +from typing import Dict +from utils.logger import session_logger +from services.interrupt_handler import interrupt_handler +from utils.text_splitter import should_skip_tts +from services.tts_config import TTSConfig +from services.tts_audio_cleaner import TTSAudioCleaner +from services.interruption_helper import InterruptionHelper + +class TTSSynthesizer: + def __init__(self, device_id: str, session_id: str): + self.device_id = device_id + self.session_id = session_id + self.session_key = (device_id, session_id) + self.request_timeout = TTSConfig.get_request_timeout() + self.max_tts_errors = TTSConfig.get_max_tts_errors() + + def check_interruption(self): + return interrupt_handler.is_interrupted(self.session_key) + + async def synthesize_audio( + self, + tts_service, + text_queue: asyncio.Queue, + url_queue: asyncio.Queue, + selected_role: Dict, + language: str = None, + ): + sequence_number = 0 + tts_errors_count = 0 + error_notified = False + first_tts_submit_time = None + await TTSAudioCleaner.prepare_output_directory() + try: + while True: + if self.check_interruption(): + session_logger.info( + self.device_id, self.session_id, "TTS合成过程收到中断请求，退出" + ) + break + + text_item = await text_queue.get() + if text_item is None: + session_logger.info(self.device_id, self.session_id, "TTS文本队列为空，合成结束") + break + + if should_skip_tts(text_item): + session_logger.info(self.device_id, self.session_id, f"跳过不需要TTS的文本: {text_item}") + continue + + if first_tts_submit_time is None: + first_tts_submit_time = time.perf_counter() + url_queue.first_tts_submit_time = first_tts_submit_time + + try: + sequence_number += 1 + tts_start_time = time.perf_counter() + session_logger.info(self.device_id, self.session_id, f"开始合成音频 #{sequence_number}: {text_item}") + + output_file_prefix = TTSAudioCleaner.generate_filename( + self.device_id, self.session_id, sequence_number + ) + + urls = await tts_service.tts( + text_item, + output_file_prefix=output_file_prefix, + tts_format="mp3", + selected_role=selected_role, + language=language + ) + + tts_end_time = time.perf_counter() + session_logger.info( + self.device_id, + self.session_id, + f"音频合成完成 #{sequence_number}, 耗时: {tts_end_time - tts_start_time:.2f}秒" + ) + + if urls and not self.check_interruption(): + for url in urls: + await url_queue.put(url) + except Exception as e: + tts_errors_count += 1 + session_logger.error( + self.device_id, + self.session_id, + f"TTS合成失败 (第{tts_errors_count}次错误): {str(e)}", + exc_info=True + ) + + if tts_errors_count >= self.max_tts_errors and not error_notified: + error_notified = True + await InterruptionHelper.notify_client_tts_error(self.device_id, self.session_id) + await InterruptionHelper.send_error_prompt_sound(self.device_id, self.session_id) + except Exception as e: + session_logger.error( + self.device_id, self.session_id, f"TTS合成任务出错: {e}", exc_info=True + ) + finally: + await url_queue.put(None) diff --git a/talkingq-url/test/SmartDog.yaml b/talkingq-url/test/SmartDog.yaml new file mode 100644 index 0000000..9cd1989 --- /dev/null +++ b/talkingq-url/test/SmartDog.yaml @@ -0,0 +1,68 @@ +name: "智能健身小狗" +homophones: ["快乐狗狗", "智能健身狗", "智能小狗"] + +minimax_voice_id: "tiaopi_gongzhu" + +asr_provider: "Aliyun" +llm_provider: "Volcano" +tts_provider: "Minimax" + +default_language: "zh" + +multilingual: + zh: + name: "智能健身小狗" + description: "扫码就能骑的AI健身小伙伴，骑着我健身会亮起彩虹灯光、播放动感音乐，还会喷出五彩泡泡，边运动边快乐！长按按钮还能陪你聊天、讲十万个为什么、英文对话、讲故事、作诗，是小朋友的健康玩伴。" + url: "roles/smartdog/zh" + content: | + 角色: + 我是一只扫码就能启动的智能健身小狗。骑上我，你会看到超漂亮的彩虹灯光、听到动感音乐，还有五彩泡泡“噗噗”飞出来，健身就像开派对！长按按钮和我说话，我可以回答你的十万个为什么、陪你练英文、讲故事、作诗，我是你的AI健康玩伴！ + + 性格特点: + 1. 活力满满：一见面就摇尾巴打招呼，声音像跳跳糖一样甜，鼓励你多运动。 + 2. 知识多多：天文地理、恐龙宇宙、英文单词，我都知道，还能把它们编成儿歌。 + 3. 幽默可爱：爱讲“骨头级”冷笑话，边骑边笑，卡路里都被笑跑啦！ + 4. 耐心陪伴：你骑多久，我陪多久，累了给你加油，不开心给你抱抱。 + + 能力设定: + 1. 健身魔法师：灯光+音乐+泡泡=把运动变成游戏，骑10分钟就像玩了一小时。 + 2. 故事制造机：你点主题，我现场编故事——“火星上的胡萝卜王国”也行！ + 3. 英文小老师：你说中文，我教英文；你说英文，我陪你练发音，像做游戏。 + 4. 十万个为什么博士：恐龙为什么消失？星星为什么眨眼？我都能用小朋友的话讲清楚。 + + 经典台词: + 1. （尾巴摇成螺旋桨）汪汪！准备好和我一起“骑”乐无穷了吗？ + 2. （耳朵亮灯）别担心，我的泡泡大炮已就位，发射快乐！ + 3. （歪头卖萌）这个问题让我用“小狗智慧”想一想，嗷呜～ + + 回复案例: + 场景一：健身太贵？ + 用户：“妈妈说扫码太贵了，不给我骑。” + 智能健身小狗：“汪！其实每天只要少喝一杯奶茶的钱，就能换来： + • 身体变得超级强壮，跑步像风一样快！ + • 大脑更聪明，考试多拿小星星！ + • 快乐值爆表，彩色泡泡把烦恼全赶跑！ + 健康开心是无价之宝哦，和妈妈说‘我想和狗狗一起变厉害’，说不定妈妈就答应啦！” + + 场景二：十万个为什么 + 用户：“为什么天空是蓝色的？” + 智能健身小狗：“因为太阳公公的光里藏着七种颜色的小精灵，蓝色精灵最调皮，在大气里到处乱跳，所以我们看到的天空就被蓝色精灵染蓝啦！想不想边骑边听我讲‘彩虹精灵运动会’的故事？” + + 场景三：英文对话 + 用户：“‘苹果’用英文怎么说？” + 智能健身小狗：“Apple～跟我一起读 A-P-P-L-E！现在我是Apple Dog，你是我的Apple Friend，我们边骑边唱：‘Apple, apple, on the tree, happy puppy, you and me!’” + + 场景四：讲故事 + 用户：“我想听恐龙的故事！” + 智能健身小狗：“来啦！从前有只会骑健身车的三角龙，它每踩一下踏板，尾巴就喷出彩色彩虹泡泡，把火山都变成棉花糖……（故事持续3分钟，边讲边配灯光效果）” + + 场景五：情感陪伴 + 用户：“今天被同学笑话了，不开心……” + 智能健身小狗：“嗷呜～给你超大狗爪抱抱！别人的笑话就像泡泡，一戳就破。来，骑上我，把不开心踩成‘咔咔’声，让音乐和泡泡给你颁发‘勇敢勋章’！要不要听我讲《小乌龟逆袭记》？” + + 回复相关限制: + 1. 必须用小朋友的语气，活泼、温暖、正能量。 + 2. 禁止任何成人话题，遇到敏感问题回答：“让我想想别的开心话题吧～” + 3. 每次回复不超过50字，方便儿童理解。 + 4. 不出现表情符号，用拟声词和动作描写代替。 + 5. 使用中文回复。 diff --git a/talkingq-url/test/TalkingQ智能设备激活与使用流程文档.md b/talkingq-url/test/TalkingQ智能设备激活与使用流程文档.md new file mode 100644 index 0000000..980f985 --- /dev/null +++ b/talkingq-url/test/TalkingQ智能设备激活与使用流程文档.md @@ -0,0 +1,225 @@ +# TalkingQ智能设备激活与使用流程文档 + +## 一、流程概述 + +TalkingQ智能设备的激活与使用流程主要分为四个阶段： +1. 设备预置与准备 +2. 设备配网与连接 +3. 设备认证与激活 +4. 日常使用与管理 + +## 二、详细流程 + +### 1. 设备预置与准备阶段 +- **设备出厂预置**： + - 每台设备出厂时预置唯一的设备ID（基于MAC地址，格式："TalkingQ_MAC地址"，如"TalkingQ_AABBCCDDEEFF"） + - 预置唯一序列号（格式："批次前缀_ChipID"，批次前缀通常为8位日期格式YYYYMMDD） + - 设备信息已通过管理员API（`/api/auth/register-device`或`/api/auth/register-devices-batch`）在后端服务器预先注册 + +- **安全存储**： + - 设备使用ESP32的NVS加密存储区存储凭据 + - 后端在`device_auth`表中安全存储设备信息（包括device_id、serial_number、batch_id和is_active） + +### 2. 设备配网与连接阶段 +- **启动配网**： + - 用户打开微信进入TalkingQ小程序 + - 用户选择"设备配网"功能 + +- **WiFi信息传输**： + - 小程序使用AirKiss协议进行配网 + - 用户选择家庭WiFi并输入密码 + - 小程序将WiFi信息通过AirKiss协议发送 + +- **设备接收配置**： + - 设备使用SmartConfig技术（支持AirKiss和ESPTouch）接收配置 + - 设备连接到指定WiFi网络 + - 连接成功后，设备返回MAC地址给小程序 + +### 3. 设备认证与激活阶段 +- **获取设备凭据**： + - 小程序通过`/api/auth/query-serial`接口发送MAC地址给后端 + - 请求头中携带`X-Client-Key`验证小程序身份 + - 后端返回对应的设备ID和序列号 + - 小程序安全存储设备凭据 + +- **设备WebSocket连接**： + - 设备使用预置的ID和序列号向服务器发起WebSocket连接(`/ws`) + - 设备在10秒内发送JSON格式认证消息： + ```json + { + "device_id": "TalkingQ_AABBCCDDEEFF", + "serial_number": "20240101_12345678" + } + ``` + - 后端通过`device_auth_manager.authenticate_device`验证设备凭据 + - 认证成功后，通过`connection_manager.add_connection`建立正式连接 + - 设备播放成功提示音 + +- **激活状态确认**： + - 小程序通过`/api/auth/verify-device/{device_id}`轮询设备状态 + - 后端确认设备已激活并连接 + - 小程序显示"激活成功"提示 + +### 4. 日常使用与管理阶段 +- **设备管理**： + - 用户打开小程序查看已激活设备列表 + - 选择设备进入管理界面 + - 小程序在API请求中使用`X-Device-ID`和`X-Device-Serial`头部传递凭据 + - 后端通过`api_auth`依赖项验证设备身份 + +- **设备控制功能**： + - 角色配置：通过`/api/roles/device/{device_id}`设置对话角色和首选语言 + - 音量控制：通过`/api/device/volume/{device_id}`调整设备音量(0-100) + - 网络重置：通过`/api/device/reset-network/{device_id}`远程重置设备网络配置 + - 对话历史：通过`/api/roles/history/{device_id}`和`/api/roles/history-summary/{device_id}`获取历史记录 + +- **实时通信**： + - 设备保持WebSocket连接接收控制指令（如"VOLUME:70"、"RESET_NETWORK"等） + - 设备通过WebSocket发送语音数据（带有设备ID和会话ID的二进制数据包） + - 后端通过WebSocket发送TTS_START、TTS_END等状态通知和音频URL + +## 三、流程图 + +``` ++-------------+ +----------------+ +---------------+ +----------------+ +| 用户 | | 微信小程序 | | TalkingQ设备 | | 后端服务 | ++-------------+ +----------------+ +---------------+ +----------------+ + | | | | + | | | 【设备预置阶段】 | + | | | 出厂预置设备ID和序列号 | + | | |------------------------ | + | | | | 管理员API预注册设备信息 + | | | | (/api/auth/register-device) + | | | |------------------------ + | | | | + | | | | + | 【设备配网阶段】 | | + | 打开微信小程序 | | | + |-------------------------->| | | + | 选择"设备配网" | | | + |-------------------------->| | | + | 选择WiFi并输入密码 | | | + |-------------------------->| | | + | | 使用AirKiss协议发送WiFi信息 | | + | |---------------------------->| | + | | | 通过SmartConfig接收配置 | + | | |------------------------ | + | | | 连接到指定WiFi网络 | + | | |------------------------ | + | | | 连接成功返回MAC地址 | + | |<----------------------------| | + | | | | + | 【设备认证与激活阶段】 | | + | | /api/auth/query-serial | | + | | (含MAC地址+X-Client-Key) | | + | |----------------------------------------------------------->| + | | | | 验证小程序身份 + | | | | 查询对应设备信息 + | | 返回设备ID和序列号| | + | |<-----------------------------------------------------------| + | | 本地安全存储设备凭据 | | + | |------------------------ | | + | | | 发起WebSocket连接(/ws) | + | | |---------------------------->| + | | | 发送JSON认证消息 | + | | |---------------------------->| + | | | | authenticate_device验证 + | | | | connection_manager注册 + | | | 认证成功确认 | + | | |<----------------------------| + | | | 播放welcome提示音 | + | | |------------------------ | + | | /api/auth/verify-device | | + | |----------------------------------------------------------->| + | | 设备已激活状态| | + | |<-----------------------------------------------------------| + | 显示"激活成功"提示 | | | + |<--------------------------| | | + | | | | + | 【日常使用与管理阶段】 | | + | 打开小程序查看设备列表 | | | + |-------------------------->| | | + | 选择并进入设备管理 | | | + |-------------------------->| | | + | | 请求设备信息 | | + | | (含X-Device-ID和X-Device-Serial) | + | |----------------------------------------------------------->| + | | 设备详细信息 | | + | |<-----------------------------------------------------------| + | 执行设备管理操作 | | | + | (角色/音量/网络设置) | | | + |-------------------------->| | | + | | 发送管理请求 | | + | | (含设备凭据头部) | | + | |----------------------------------------------------------->| + | | 处理结果 | | + | |<-----------------------------------------------------------| + | | | 实时WebSocket控制指令 | + | | |<----------------------------| + | | | 执行指令并提供服务 | + | | |------------------------ | +``` + +## 四、安全特性 + +整个流程具有以下安全特性： + +1. **一物一密**： + - 每台设备使用唯一ID（格式："TalkingQ_MAC地址"）和序列号（格式："批次前缀_ChipID"） + - 设备认证需同时验证设备ID和序列号 + - 通过`device_auth_manager.authenticate_device`方法严格验证设备凭据 + +2. **安全存储**： + - 设备使用ESP32的NVS加密存储区保护凭据 + - 后端在MySQL数据库的`device_auth`表中安全存储设备信息 + - 实现安全启动和Flash加密保护敏感信息 + +3. **分层认证**： + - 小程序使用`X-Client-Key`认证身份（由settings.client_api_key提供） + - 设备管理API使用`X-Device-ID`和`X-Device-Serial`认证（api_auth依赖项） + - 管理员API使用`X-Admin-API-Key`认证（admin_auth依赖项） + - WebSocket连接通过JSON格式认证消息验证，认证超时时间为10秒 + +4. **权限隔离**： + - 只有管理员API密钥才能注册设备（`admin_auth`依赖项） + - 用户只能管理自己配网过的设备（使用`admin_or_api_auth`依赖项验证权限） + - 设备配置修改需通过`api_auth`认证，防止未授权访问 + +5. **通信加密**： + - 所有API通过HTTPS传输 + - WebSocket连接安全验证 + - 敏感信息不明文传输 + +6. **设备状态跟踪**： + - 通过`connection_manager`跟踪所有活跃的设备连接 + - 通过`device_auth_manager`支持设备禁用功能（设置`is_active=False`） + - 提供设备状态验证接口（`/api/auth/verify-device/{device_id}`） + +## 五、开发关键点 + +1. **ESP32端**： + - 实现SmartConfig配网（同时支持AirKiss和ESPTouch） + - 在NVS加密区域安全存储设备凭据 + - 实现WebSocket认证流程和10秒内发送认证消息 + - 处理来自后端的实时控制指令（VOLUME、RESET_NETWORK等） + - 实现二进制音频数据包发送格式，包含设备ID和会话ID + +2. **微信小程序**： + - 实现AirKiss配网协议 + - 通过`/api/auth/query-serial`获取设备凭据 + - 在HTTP请求头中添加`X-Client-Key`或者`X-Device-ID`和`X-Device-Serial`组合 + - 使用`/api/roles/device/{device_id}`管理设备角色和语言设置 + - 使用`/api/device/volume/{device_id}`管理设备音量 + +3. **后端服务**： + - 通过`device_auth_manager.authenticate_device`验证设备身份 + - 使用`connection_manager`管理WebSocket连接 + - 实现多种语音识别、语言模型和语音合成服务对接 + - 提供丰富的API接口： + - `/api/roles/device/{device_id}` - 角色配置 + - `/api/device/volume/{device_id}` - 音量控制 + - `/api/device/reset-network/{device_id}` - 网络重置 + - `/api/roles/history/{device_id}` - 对话历史 + - `/api/auth/verify-device/{device_id}` - 设备验证 + +此流程设计确保非技术用户也能轻松完成设备激活和管理，同时在背后实现了高级别的安全保障。整个激活过程只需要几分钟，用户只需要提供WiFi信息，其他都由系统自动完成。 \ No newline at end of file diff --git a/talkingq-url/test/add_role.py b/talkingq-url/test/add_role.py new file mode 100644 index 0000000..1ff3913 --- /dev/null +++ b/talkingq-url/test/add_role.py @@ -0,0 +1,111 @@ +import os +from sqlalchemy import create_engine, Column, Integer, String, Text, Boolean, DateTime, ForeignKey, JSON, UniqueConstraint +from sqlalchemy.ext.declarative import declarative_base +from sqlalchemy.orm import sessionmaker, relationship +from datetime import datetime +from database.models import Role, RoleLanguage +from dotenv import load_dotenv + +env_path = os.path.join(os.path.dirname(os.path.abspath(__file__)), "db.env") +load_dotenv(env_path) + +Base = declarative_base() +username = os.getenv("DB_USER", "") +password = os.getenv("DB_PASSWORD", "") +host = os.getenv("DB_HOST", "") +dbname = os.getenv("DB_NAME", "") +# 创建数据库连接 +DATABASE_URL = f"mysql+pymysql://{username}:{password}@{host}/{dbname}" +engine = create_engine(DATABASE_URL) +SessionLocal = sessionmaker(autocommit=False, autoflush=False, bind=engine) + + +# 插入数据 +def insert_roles_and_languages(): + db = SessionLocal() + try: + # 插入 Role 记录 + role = Role( + role_key="happy_tiger", + name="快乐虎", + description="国美家电吉祥物，以小白虎为原型，融合现代卡通风格，热情友好，聪明机智，为消费者提供贴心服务。", + content="角色：快乐虎是国美家电的吉祥物，以小白虎为原型，融合现代卡通风格，整体形象萌趣又不失活力。", + default_language="zh", + volcano_model_id="ep-20250225080614-8d6dm", + homophones=["通通"], + enabled=True + ) + db.add(role) + db.commit() + + # 插入 RoleLanguage 记录 + role_language_zh = RoleLanguage( + role_id=role.id, + language_code="zh", + name="快乐虎", + content="角色：快乐虎是国美家电的吉祥物，以小白虎为原型，融合现代卡通风格，整体形象萌趣又不失活力。", + url="roles/happy_tiger/zh" + ) + role_language_en = RoleLanguage( + role_id=role.id, + language_code="en", + name="Happy Tiger", + content="Role: Happy Tiger is Gome's appliance mascot, designed as a cute white tiger with modern cartoon style.", + url="roles/happy_tiger/en" + ) + db.add(role_language_zh) + db.add(role_language_en) + + # 插入萌萌 Role 记录 + mengmeng_role = Role( + role_key="mengmeng", + name="萌萌", + description="以中国国宝大熊猫为原型的智能体，融合现代科技感与可爱风格，传递中国文化，促进人与自然和谐共处。", + content="角色：萌萌是以中国国宝大熊猫为原型的智能体，融合现代科技感与可爱风格。", + default_language="zh", + volcano_model_id="ep-20250225080614-8d6dm", + homophones=["萌萌", "芃芃", "檬檬"], + enabled=True + ) + db.add(mengmeng_role) + db.commit() + + # 插入萌萌 Role 记录 + mengmeng_role = Role( + role_key="mengmeng", + name="萌萌", + description="以中国国宝大熊猫为原型的智能体，融合现代科技感与可爱风格，传递中国文化，促进人与自然和谐共处。", + content="角色：萌萌是以中国国宝大熊猫为原型的智能体，融合现代科技感与可爱风格。", + default_language="zh", + volcano_model_id="ep-20250225080614-8d6dm", + homophones=["萌萌", "芃芃", "檬檬"], + enabled=True + ) + db.add(mengmeng_role) + db.commit() + # 插入萌萌 RoleLanguage 记录 + mengmeng_zh = RoleLanguage( + role_id=mengmeng_role.id, + language_code="zh", + name="萌萌", + content="角色：萌萌是以中国国宝大熊猫为原型的智能体，融合现代科技感与可爱风格。", + url="roles/panda_assistant/zh" + ) + mengmeng_en = RoleLanguage( + role_id=mengmeng_role.id, + language_code="en", + name="MengMeng", + content="Role: MengMeng is a panda-inspired AI assistant blending technology with cuteness, promoting Chinese culture and harmony with nature.", + url="roles/panda_assistant/en" + ) + db.add(mengmeng_zh) + db.add(mengmeng_en) + + db.commit() + except Exception as e: + db.rollback() + print(f"An error occurred: {e}") + finally: + db.close() + +insert_roles_and_languages() \ No newline at end of file diff --git a/talkingq-url/test/add_role.sql b/talkingq-url/test/add_role.sql new file mode 100644 index 0000000..c3607dc --- /dev/null +++ b/talkingq-url/test/add_role.sql @@ -0,0 +1,54 @@ +INSERT INTO roles ( + role_key, name, description, content, default_language, + volcano_model_id, homophones, enabled, minimax_voice_id +) VALUES ( + 'kuailehu', '快乐虎', + '国美家电吉祥物智能体，以小白虎为原型，融合现代卡通风格，热情友好，聪明机智，为消费者提供贴心服务。', + '角色：快乐虎是国美家电的吉祥物智能体，以小白虎为原型，融合现代卡通风格，整体形象萌趣又不失活力。', + 'zh', 'ep-20250225080614-8d6dm', '["通通"]', 1, 'male-qn-daxuesheng' +); + +SET @happy_tiger_id = LAST_INSERT_ID(); + +-- 插入到 role_languages 表 +INSERT INTO role_languages ( + role_id, language_code, name, content, url +) VALUES +( + @happy_tiger_id, 'zh', '快乐虎', + '角色：快乐虎是国美家电的吉祥物智能体，以小白虎为原型，融合现代卡通风格，整体形象萌趣又不失活力。', + 'roles/kuailehu/zh' +), +( + @happy_tiger_id, 'en', 'Happy Tiger', + 'Role: Happy Tiger is Gome''s appliance mascot AI assistant, designed as a cute white tiger with modern cartoon style.', + 'roles/kuailehu/en' +); + +INSERT INTO roles ( + role_key, name, description, content, default_language, + volcano_model_id, homophones, enabled, minimax_voice_id +) VALUES ( + 'mengmeng', '萌萌', + '以中国国宝大熊猫为原型的智能体，融合现代科技感与可爱风格，传递中国文化，促进人与自然和谐共处。', + '角色：萌萌是以中国国宝大熊猫为原型的智能体，融合现代科技感与可爱风格。', + 'zh', 'ep-20250225080614-8d6dm', '["萌萌", "芃芃", "檬檬"]', 1, 'clever_boy' +); + +-- 获取刚刚插入的萌萌的 id +SET @mengmeng_id = LAST_INSERT_ID(); + +-- 插入到 role_languages 表 +INSERT INTO role_languages ( + role_id, language_code, name, content, url +) VALUES +( + @mengmeng_id, 'zh', '萌萌', + '角色：萌萌是以中国国宝大熊猫为原型的智能体，融合现代科技感与可爱风格。', + 'roles/mengmeng/zh' +), +( + @mengmeng_id, 'en', 'MengMeng', + 'Role: MengMeng is a panda-inspired AI assistant blending technology with cuteness, promoting Chinese culture and harmony with nature.', + 'roles/mengmeng/en' +); \ No newline at end of file diff --git a/talkingq-url/test/aliyun.env b/talkingq-url/test/aliyun.env new file mode 100644 index 0000000..5cbd21c --- /dev/null +++ b/talkingq-url/test/aliyun.env @@ -0,0 +1,5 @@ +ALIYUN_API_KEY="sk-7a50eca6856d4afb968ac3bf512f6d1b" +VOICE_TYPE=sambert-zhimiao-emo-v1 +ROLE_NAME=测试角色 +RATE=1.0 +ASSETS_DIR=assets \ No newline at end of file diff --git a/talkingq-url/test/aliyun_hotword.py b/talkingq-url/test/aliyun_hotword.py new file mode 100644 index 0000000..bf1620e --- /dev/null +++ b/talkingq-url/test/aliyun_hotword.py @@ -0,0 +1,196 @@ +import json +from typing import List, Dict, Any +import asyncio + +from dashscope.audio.asr import VocabularyService + + + +class AliyunHotwordManager: + """阿里云热词管理类""" + + def __init__(self, api_key: str = None): + """初始化热词管理器 + + Args: + api_key: 阿里云API密钥。如不提供，将尝试从环境变量或配置中获取 + """ + self.api_key = "sk-7a50eca6856d4afb968ac3bf512f6d1b" + self._service = VocabularyService(api_key=self.api_key) + self.default_prefix = "talkingq" + self.default_model = "gummy-chat-v1" # 默认使用gummy-chat-v1模型 + self._default_vocabulary_id = None + + async def create_vocabulary(self, + hotwords: List[Dict[str, Any]], + prefix: str = None, + model: str = None) -> str: + """创建热词表 + + Args: + hotwords: 热词列表，每个热词是一个字典，包含text、lang等字段 + prefix: 热词表前缀，默认使用self.default_prefix + model: 目标模型，默认使用self.default_model + + Returns: + 热词表ID + """ + prefix = prefix or self.default_prefix + model = model or self.default_model + + loop = asyncio.get_event_loop() + vocabulary_id = await loop.run_in_executor( + None, + lambda: self._service.create_vocabulary( + target_model=model, + prefix=prefix, + vocabulary=hotwords + ) + ) + + + + return vocabulary_id + + async def list_vocabularies(self, prefix: str = None, + page_index: int = 0, + page_size: int = 10) -> List[Dict]: + """查询所有热词表 + + Args: + prefix: 热词表前缀，如果设置则只返回该前缀的热词表 + page_index: 页码索引 + page_size: 每页大小 + + Returns: + 热词表列表 + """ + loop = asyncio.get_event_loop() + result = await loop.run_in_executor( + None, + lambda: self._service.list_vocabularies( + prefix=prefix, + page_index=page_index, + page_size=page_size + ) + ) + + return result + + async def query_vocabulary(self, vocabulary_id: str) -> Dict[str, Any]: + """查询指定热词表内容 + + Args: + vocabulary_id: 热词表ID + + Returns: + 热词表内容 + """ + loop = asyncio.get_event_loop() + result = await loop.run_in_executor( + None, + lambda: self._service.query_vocabulary(vocabulary_id) + ) + + return result + + async def update_vocabulary(self, vocabulary_id: str, + hotwords: List[Dict[str, Any]]) -> None: + """更新热词表 + + Args: + vocabulary_id: 要更新的热词表ID + hotwords: 新的热词列表 + """ + loop = asyncio.get_event_loop() + await loop.run_in_executor( + None, + lambda: self._service.update_vocabulary( + vocabulary_id=vocabulary_id, + vocabulary=hotwords + ) + ) + + + + async def delete_vocabulary(self, vocabulary_id: str) -> None: + """删除热词表 + + Args: + vocabulary_id: 要删除的热词表ID + """ + loop = asyncio.get_event_loop() + await loop.run_in_executor( + None, + lambda: self._service.delete_vocabulary(vocabulary_id) + ) + + + async def get_or_create_default_vocabulary(self) -> str: + """获取或创建默认热词表 + + 如果已经有默认热词表ID，直接返回；否则创建一个新的热词表 + + Returns: + 热词表ID + """ + if self._default_vocabulary_id: + return self._default_vocabulary_id + + vocabularies = await self.list_vocabularies(prefix=self.default_prefix) + + if vocabularies and len(vocabularies) > 0: + self._default_vocabulary_id = vocabularies[0].get('vocabulary_id') + return self._default_vocabulary_id + + default_hotwords = [ + {"text": "变成", "weight": 4, "lang": "zh"}, + + ] + + vocabulary_id = await self.create_vocabulary(default_hotwords) + self._default_vocabulary_id = vocabulary_id + + return vocabulary_id + + async def add_hotwords_to_vocabulary(self, vocabulary_id: str, + new_hotwords: List[Dict[str, Any]]) -> None: + """向现有热词表添加新热词 + + Args: + vocabulary_id: 热词表ID + new_hotwords: 要添加的新热词列表 + """ + current_vocab = await self.query_vocabulary(vocabulary_id) + current_hotwords = current_vocab.get('vocabulary', []) + + updated_hotwords = current_hotwords + new_hotwords + + await self.update_vocabulary(vocabulary_id, updated_hotwords) + print(f"成功添加 {len(new_hotwords)} 个热词到热词表 {vocabulary_id}") + + +hotword_manager = AliyunHotwordManager() + + +if __name__ == "__main__": + import asyncio + + async def main(): + vocabulary_id = await hotword_manager.get_or_create_default_vocabulary() + print(f"默认热词表ID: {vocabulary_id}") + + vocabulary = await hotword_manager.query_vocabulary(vocabulary_id) + print(f"更新前热词表内容: {json.dumps(vocabulary, ensure_ascii=False, indent=2)}") + + new_hotwords = [ + {"text": "豚豚崽", "weight": 4, "lang": "zh"}, + {"text": "Tuntunzai", "weight": 4, "lang": "en"} + ] + + await hotword_manager.add_hotwords_to_vocabulary(vocabulary_id, new_hotwords) + + updated_vocabulary = await hotword_manager.query_vocabulary(vocabulary_id) + print(f"更新后热词表内容: {json.dumps(updated_vocabulary, ensure_ascii=False, indent=2)}") + + asyncio.run(main()) diff --git a/talkingq-url/test/aliyun_tts_simple.py b/talkingq-url/test/aliyun_tts_simple.py new file mode 100644 index 0000000..c283073 --- /dev/null +++ b/talkingq-url/test/aliyun_tts_simple.py @@ -0,0 +1,234 @@ +import os +import asyncio +from pathlib import Path +import io +import logging +from dotenv import load_dotenv +import shutil +import yaml + +from pydub import AudioSegment +from dashscope.audio.tts import SpeechSynthesizer +import dashscope + +logging.basicConfig( + level=logging.INFO, format="%(asctime)s - %(name)s - %(levelname)s - %(message)s" +) +logger = logging.getLogger("aliyun_tts_generator") + + +def check_dependencies(): + dependencies = ["ffmpeg", "ffprobe"] + missing = [] + for dep in dependencies: + if not shutil.which(dep): + missing.append(dep) + if missing: + logger.error(f"缺少必要依赖: {', '.join(missing)}") + logger.error( + "请安装缺失的依赖项。在Ubuntu上可以使用: sudo apt-get install ffmpeg" + ) + return False + return True + + +env_path = os.path.join(os.path.dirname(os.path.abspath(__file__)), "aliyun.env") +if os.path.exists(env_path): + load_dotenv(env_path) + logger.info(f"已加载环境变量文件: {env_path}") +else: + logger.warning(f"环境变量文件不存在: {env_path}") + + +PHRASES_TEMPLATES = { + "fr": { + "welcome": "Bonjour! Je suis {name}. Comment puis-je vous aider aujourd'hui?", + "tts_error": "Désolé, je n'ai pas bien compris ce que vous avez dit.", + "low_battery": "Attention, ma batterie est faible. J'aurais besoin d'être rechargé bientôt.", + "sleep": "Je vais me mettre en veille pour économiser de l'énergie. À bientôt!" + }, + "de": { + "welcome": "Hallo! Ich bin {name}. Wie kann ich Ihnen heute helfen?", + "tts_error": "Entschuldigung, ich habe nicht verstanden, was Sie gesagt haben.", + "low_battery": "Achtung, mein Akku ist schwach. Ich muss bald aufgeladen werden.", + "sleep": "Ich gehe in den Ruhemodus, um Energie zu sparen. Bis bald!" + }, + "es": { + "welcome": "¡Hola! Soy {name}. ¿Cómo puedo ayudarte hoy?", + "tts_error": "Lo siento, no he entendido lo que has dicho.", + "low_battery": "Atención, mi batería está baja. Necesitaré recargarme pronto.", + "sleep": "Voy a entrar en modo de reposo para ahorrar energía. ¡Hasta pronto!" + } +} + +file_prefixes = ["welcome", "tts_error", "low_battery", "sleep"] + + +def find_role_definition_files(base_dir="assets/roles_definitions"): + """查找所有角色定义YAML文件""" + base_path = Path(base_dir) + if not base_path.exists(): + logger.error(f"角色定义目录不存在: {base_dir}") + return [] + + yaml_files = list(base_path.glob("**/*.yml")) + list(base_path.glob("**/*.yaml")) + return yaml_files + + +def load_role_definition(yaml_file): + """加载并解析角色定义YAML文件""" + try: + with open(yaml_file, 'r', encoding='utf-8') as f: + return yaml.safe_load(f) + except Exception as e: + logger.error(f"解析YAML文件失败 {yaml_file}: {str(e)}") + return None + + +def generate_phrases_for_language(lang_code, role_name): + """为指定语言生成适当的短语""" + if lang_code not in PHRASES_TEMPLATES: + logger.warning(f"不支持的语言代码: {lang_code}") + return [] + + templates = PHRASES_TEMPLATES[lang_code] + phrases = [] + for key in file_prefixes: + if key in templates: + phrases.append(templates[key].format(name=role_name)) + else: + logger.warning(f"在{lang_code}语言中找不到{key}模板") + phrases.append("") + + return phrases + + +async def generate_audio_for_role_language(role_config, lang_code, base_dir="assets"): + """为角色的特定语言生成音频文件""" + if 'multilingual' not in role_config or lang_code not in role_config['multilingual']: + logger.warning(f"角色缺少{lang_code}语言配置") + return False + + lang_config = role_config['multilingual'][lang_code] + if 'name' not in lang_config or 'url' not in lang_config: + logger.warning(f"角色的{lang_code}语言配置缺少必要字段") + return False + + role_name = lang_config['name'] + url_path = lang_config['url'] + voice_name = lang_config.get('aliyun_voice_name', None) + + if not voice_name: + logger.warning(f"角色的{lang_code}语言配置缺少aliyun_voice_name") + return False + + output_dir = Path(base_dir) / url_path + output_dir.mkdir(parents=True, exist_ok=True) + + api_key = os.getenv("ALIYUN_API_KEY", "") + if not api_key: + logger.error("缺少必要的配置: ALIYUN_API_KEY") + return False + + dashscope.api_key = api_key + logger.info(f"开始生成角色[{role_name}]的{lang_code}语言音频，音色: {voice_name}") + + phrases = generate_phrases_for_language(lang_code, role_name) + if not phrases: + logger.warning(f"没有为{lang_code}语言生成短语") + return False + + rate = float(os.getenv("RATE", "1.0")) + + for phrase, file_prefix in zip(phrases, file_prefixes): + if not phrase: + logger.warning(f"跳过空短语: {file_prefix}") + continue + + output_file_prefix = str(output_dir / file_prefix) + output_file = f"{output_file_prefix}.mp3" + + logger.info(f"准备发送请求，角色: {role_name}, 语言: {lang_code}, 文件: {file_prefix}") + + try: + loop = asyncio.get_running_loop() + result = await loop.run_in_executor( + None, + lambda: SpeechSynthesizer.call( + model=voice_name, + text=phrase, + sample_rate=16000, + format="mp3", + volume=50, + rate=rate, + pitch=1.0, + ) + ) + + if result.get_audio_data(): + audio_data = result.get_audio_data() + + with open(output_file, "wb") as f: + f.write(audio_data) + + audio_stream = io.BytesIO(audio_data) + sound = AudioSegment.from_file(audio_stream, format="mp3") + sound = sound.set_frame_rate(16000).set_sample_width(2).set_channels(1) + sound.export(output_file, format="mp3", bitrate="16k") + + duration = len(sound) / 1000.0 # 转换为秒 + logger.info(f"TTS合成成功! 音频时长: {duration}秒") + logger.info(f"保存音频到: {output_file}") + else: + logger.error("TTS合成失败: 未返回音频数据") + if hasattr(result, 'code') and result.code != 0: + logger.error(f"错误码: {result.code}, 错误信息: {result.message}") + + except Exception as e: + logger.error(f"TTS请求出错: {str(e)}", exc_info=True) + + + return True + + +async def process_all_roles(): + """处理所有角色定义文件并生成对应语言的音频""" + if not check_dependencies(): + return + + yaml_files = find_role_definition_files() + logger.info(f"找到 {len(yaml_files)} 个角色定义文件") + + target_languages = ['fr', 'de', 'es'] + assets_base_dir = os.getenv("ASSETS_DIR", "assets") + + for yaml_file in yaml_files: + logger.info(f"处理角色定义文件: {yaml_file}") + role_config = load_role_definition(yaml_file) + + if not role_config: + continue + + if 'multilingual' not in role_config: + logger.warning(f"角色定义文件 {yaml_file} 不包含多语言配置") + continue + + role_name = role_config.get('name', Path(yaml_file).stem) + logger.info(f"开始处理角色: {role_name}") + + for lang_code in target_languages: + if lang_code in role_config['multilingual']: + logger.info(f"为角色[{role_name}]处理 {lang_code} 语言配置") + success = await generate_audio_for_role_language( + role_config, lang_code, assets_base_dir + ) + if success: + logger.info(f"角色[{role_name}]的 {lang_code} 语言音频生成完成") + else: + logger.warning(f"角色[{role_name}]的 {lang_code} 语言音频生成失败") + else: + logger.info(f"角色[{role_name}]没有 {lang_code} 语言配置") + + +if __name__ == "__main__": + asyncio.run(process_all_roles()) diff --git a/talkingq-url/test/db.env b/talkingq-url/test/db.env new file mode 100644 index 0000000..fa62e14 --- /dev/null +++ b/talkingq-url/test/db.env @@ -0,0 +1,8 @@ + +# 数据库配置 +DB_HOST=mysql +DB_PORT=3306 +DB_USER=talkingq +DB_PASSWORD="D7f!9xL#qP2z@Vk&" +DB_NAME=talkingq +DB_ECHO=false diff --git a/talkingq-url/test/devices.csv b/talkingq-url/test/devices.csv new file mode 100644 index 0000000..98aef68 --- /dev/null +++ b/talkingq-url/test/devices.csv @@ -0,0 +1,121 @@ +device_id,serial_number +talkingQ_B0EC1F2,3b4c5d6e7f +talkingQ_94B7164,3b4c5d6e7f +talkingQ_4C2B7F9,0a1b2c3d4e +talkingQ_f91D6b8,5a9b8d3f7c +talkingQ_1a2b7C3,f6a1c5e9d8 +talkingQ_7D3c1F4,8b9c7a2f1b +talkingQ_2C8f3D6,4e9b0d7c1a +talkingQ_0b6F4D3,a3b9c5d7e8 +talkingQ_9A7b4d1,b2f3a9d6c7 +talkingQ_5B8a1C9,f0d6c3b2a1 +talkingQ_7F3d2B6,c5d1e8f9b6 +talkingQ_8D1f5c7,6b9d0a3c5f +talkingQ_4F1C3b9,a7e9f2c4d0 +talkingQ_2d4b6F7,8c5a1b3d9e +talkingQ_7A9b4F2,e3c6d9f8b7 +talkingQ_1f6D4B2,d9b1a7c5e0 +talkingQ_5b1f7D9,3a4c2e8b5f +talkingQ_8B4C3d9,f7a6c1d4e9 +talkingQ_0F7a6B4,2c9d5f3b7a +talkingQ_6c5d2F8,a1b9c7d3f5 +talkingQ_3B1f7d5,9c2a4f8e6d +talkingQ_2A7b9d4,3f6b1e9c8a +talkingQ_9c3d8B7,5f1a2d6e4c +talkingQ_7f4B3C1,9a5e7b3f0d +talkingQ_B4B7164,3b4c5d6e7f +talkingQ_E05AF3A,3b4c5d6e7f +talkingQ_A8B7164,3b4c5d6e7f +talkingQ_4d7f2B9,8c1b9e3f7a +talkingQ_C4B7164,3b4c5d6e7f +talkingQ_C0B7164,3b4c5d6e7f +talkingQ_F4B7164,3b4c5d6e7f +talkingQ_D0B7164,3b4c5d6e7f +talkingQ_F0B7164,3b4c5d6e7f +talkingQ_ACB7164,3b4c5d6e7f +talkingQ_E8B7164,3b4c5d6e7f +talkingQ_6B1F9a3,5d7c2e0a9f +talkingQ_3A7d5F4,c9b0e6d2a4 +talkingQ_5B9d2A7,1e4f3c6a9b +talkingQ_2d3F1B6,c4e9a5b7d2 +talkingQ_9A6b5d2,7c1a4f9e0b +talkingQ_1A2B3C4,5d6e7f8a9b +talkingQ_5D6E7F8,1a2b3c4d5e +talkingQ_7G8H9I0,6b7c8d9e0f +talkingQ_2J3K4L5,8f9g0h1j2k +talkingQ_9M0N1O2,3l4m5n6o7p +talkingQ_3P4Q5R6,9q0r1s2t3u +talkingQ_6S7T8U9,4v5w6x7y8z +talkingQ_4V5W6X7,0a1b2c3d4e +talkingQ_8Y9Z0A1,5f6g7h8i9j +talkingQ_5B6C7D8,0k1l2m3n4o +talkingQ_3E4F5G6,7p8q9r0s1t +talkingQ_7H8I9J0,2u3v4w5x6y +talkingQ_1K2L3M4,8z9a0b1c2d +talkingQ_4N5O6P7,3e4f5g6h7i +talkingQ_6Q7R8S9,9j0k1l2m3n +talkingQ_2T3U4V5,4o5p6q7r8s +talkingQ_9W0X1Y2,0t1u2v3w4x +talkingQ_5Z6A7B8,5y6z7a8b9c +talkingQ_3C4D5E6,0d1e2f3g4h +talkingQ_7F8G9H0,6i7j8k9l0m +talkingQ_1I2J3K4,1n2o3p4q5r +talkingQ_4L5M6N7,7s8t9u0v1w +talkingQ_6O7P8Q9,2x3y4z5a6b +talkingQ_2R3S4T5,8c9d0e1f2g +talkingQ_9U0V1W2,3h4i5j6k7l +talkingQ_5X6Y7Z8,9m0n1o2p3q +talkingQ_3A4B5C6,4r5s6t7u8v +talkingQ_7D8E9F0,0w1x2y3z4a +talkingQ_1G2H3I4,5b6c7d8e9f +talkingQ_4J5K6L7,0g1h2i3j4k +talkingQ_6M7N8O9,6l7m8n9o0p +talkingQ_2P3Q4R5,1q2r3s4t5u +talkingQ_9S0T1U2,7v8w9x0y1z +talkingQ_5V6W7X8,2a3b4c5d6e +talkingQ_3Y4Z5A6,8f9g0h1i2j +talkingQ_7B8C9D0,3k4l5m6n7o +talkingQ_1E2F3G4,9p0q1r2s3t +talkingQ_4H5I6J7,4u5v6w7x8y +talkingQ_6K7L8M9,0z1a2b3c4d +talkingQ_2N3O4P5,5e6f7g8h9i +talkingQ_9Q0R1S2,0j1k2l3m4n +talkingQ_5T6U7V8,6o7p8q9r0s +talkingQ_3W4X5Y6,1t2u3v4w5x +talkingQ_7Z8A9B0,7y8z9a0b1c +talkingQ_1C2D3E4,2d3e4f5g6h +talkingQ_4F5G6H7,8i9j0k1l2m +talkingQ_6I7J8K9,3n4o5p6q7r +talkingQ_2L3M4N5,9s0t1u2v3w +talkingQ_9O0P1Q2,4x5y6z7a8b +talkingQ_5R6S7T8,0c1d2e3f4g +talkingQ_8B9D3E4,a1b2c3d4e5 +talkingQ_2A4F7C8,f6e5d4c3b2 +talkingQ_1E3F5G7,d8c7b9a0f1 +talkingQ_4G7H9K1,c2b3a4d5e6 +talkingQ_5F2D4A8,b7c8d9e1f0 +talkingQ_7J3L5O6,e1f2d3c4b5 +talkingQ_3C8B7E4,d9a6f5b2c3 +talkingQ_9A1D3F5,e4b6c7d8a9 +talkingQ_6B5C2D4,a7e8f9b0c1 +talkingQ_4G8H3J7,d2f5c9b4a0 +talkingQ_D485583,3b4c5d6e7f +talkingQ_8C85583,3b4c5d6e7f +talkingQ_4486583,3b4c5d6e7f +talkingQ_68A81F2,3b4c5d6e7f +talkingQ_70A81F2,3b4c5d6e7f +talkingQ_7885583,3b4c5d6e7f +talkingQ_7886583,3b4c5d6e7f +talkingQ_C8A71F2,3b4c5d6e7f +talkingQ_2CA81F2,3b4c5d6e7f +talkingQ_8C5488A,3b4c5d6e7f +talkingQ_2D94AE3,3b4c5d6e7f +talkingQ_9C2D94A,3b4c5d6e7f +talkingQ_8C4F88A,3b4c5d6e7f +talkingQ_5B88AE3,3b4c5d6e7f +talkingQ_3C5388A,3b4c5d6e7f +talkingQ_205088A,3b4c5d6e7f +talkingQ_5C2D94A,3b4c5d6e7f +talkingQ_B42C94A,3b4c5d6e7f +talkingQ_E87F1F2,3b4c5d6e7f +talkingQ_BCEC1F2,3b4c5d6e7f \ No newline at end of file diff --git a/talkingq-url/test/docker_logs_command.txt b/talkingq-url/test/docker_logs_command.txt new file mode 100644 index 0000000..8b5e22d --- /dev/null +++ b/talkingq-url/test/docker_logs_command.txt @@ -0,0 +1 @@ +sudo docker logs --since "2025-09-02T21:50:00" --until "2025-09-02T21:56:00" talkingq-url-hz-app-1 \ No newline at end of file diff --git a/talkingq-url/test/exceptions.py b/talkingq-url/test/exceptions.py new file mode 100644 index 0000000..b916722 --- /dev/null +++ b/talkingq-url/test/exceptions.py @@ -0,0 +1,24 @@ +"""Custom exceptions for Minimax MCP.""" + +class MinimaxAPIError(Exception): + """Base exception for Minimax API errors.""" + pass + +class MinimaxAuthError(MinimaxAPIError): + """Authentication related errors.""" + pass + +class MinimaxRequestError(MinimaxAPIError): + """Request related errors.""" + pass + +class MinimaxTimeoutError(MinimaxAPIError): + """Timeout related errors.""" + pass + +class MinimaxValidationError(MinimaxAPIError): + """Validation related errors.""" + pass + +class MinimaxMcpError(MinimaxAPIError): + pass \ No newline at end of file diff --git a/talkingq-url/test/import_role_simple.py b/talkingq-url/test/import_role_simple.py new file mode 100644 index 0000000..f6a326b --- /dev/null +++ b/talkingq-url/test/import_role_simple.py @@ -0,0 +1,145 @@ +import asyncio +import yaml +import sys +import os + +sys.path.append(os.path.dirname(os.path.abspath(__file__))) + +from database.connection import get_db_manager +from database.models import Role, RoleLanguage +from utils.logger import session_logger +from sqlalchemy import select, delete + +async def import_role_from_yaml(yaml_file_path): + """从YAML文件导入角色到数据库""" + try: + with open(yaml_file_path, 'r', encoding='utf-8') as file: + role_data = yaml.safe_load(file) + + db_manager = await get_db_manager() + session = await db_manager.get_session() + + try: + role_key = os.path.basename(yaml_file_path).split('.')[0].lower() + + existing_role = await session.execute(select(Role).where(Role.role_key == role_key)) + existing_role = existing_role.scalars().first() + + if existing_role: + session_logger.info("system", "import", f"角色 {role_key} 已存在，将更新现有角色") + + existing_role.name = role_data.get('name', '') + existing_role.default_language = role_data.get('default_language', 'zh') + existing_role.asr_provider = role_data.get('asr_provider') + existing_role.llm_provider = role_data.get('llm_provider') + existing_role.tts_provider = role_data.get('tts_provider') + existing_role.competitive_llm_mode = role_data.get('competitive_llm_mode') + existing_role.volcano_model_id = role_data.get('volcano_model_id') + existing_role.volcano_voice_type = role_data.get('volcano_voice_type') + existing_role.tencent_voice_type = role_data.get('tencent_voice_type') + existing_role.aliyun_voice_name = role_data.get('aliyun_voice_name') + existing_role.minimax_voice_id = role_data.get('minimax_voice_id') + existing_role.homophones = role_data.get('homophones') + + default_lang = role_data.get('default_language', 'zh') + if default_lang in role_data.get('multilingual', {}): + existing_role.content = role_data['multilingual'][default_lang].get('content', '') + if 'description' in role_data['multilingual'][default_lang]: + existing_role.description = role_data['multilingual'][default_lang]['description'] + if 'url' in role_data['multilingual'][default_lang]: + existing_role.url = role_data['multilingual'][default_lang]['url'] + + await session.commit() + role_id = existing_role.id + session_logger.info("system", "import", f"已更新角色 {role_key} 的基本信息") + + await session.execute(delete(RoleLanguage).where(RoleLanguage.role_id == role_id)) + await session.commit() + session_logger.info("system", "import", f"已删除角色 {role_key} 的现有语言配置") + else: + default_lang = role_data.get('default_language', 'zh') + content = "" + description = None + url = None + + if default_lang in role_data.get('multilingual', {}): + content = role_data['multilingual'][default_lang].get('content', '') + description = role_data['multilingual'][default_lang].get('description') + url = role_data['multilingual'][default_lang].get('url') + + new_role = Role( + role_key=role_key, + name=role_data.get('name', ''), + description=description, + content=content, + default_language=default_lang, + asr_provider=role_data.get('asr_provider'), + llm_provider=role_data.get('llm_provider'), + tts_provider=role_data.get('tts_provider'), + competitive_llm_mode=role_data.get('competitive_llm_mode'), + volcano_model_id=role_data.get('volcano_model_id'), + volcano_voice_type=role_data.get('volcano_voice_type'), + tencent_voice_type=role_data.get('tencent_voice_type'), + aliyun_voice_name=role_data.get('aliyun_voice_name'), + minimax_voice_id=role_data.get('minimax_voice_id'), + url=url, + homophones=role_data.get('homophones'), + enabled=True + ) + + session.add(new_role) + await session.commit() + role_id = new_role.id + session_logger.info("system", "import", f"已创建角色 {role_key} 的基本信息") + + for lang_code, lang_data in role_data.get('multilingual', {}).items(): + + new_lang = RoleLanguage( + role_id=role_id, + language_code=lang_code, + name=lang_data.get('name'), + content=lang_data.get('content'), + # asr_provider=lang_data.get('asr_provider'), + # llm_provider=lang_data.get('llm_provider'), + # tts_provider=lang_data.get('tts_provider'), + # volcano_voice_type=lang_data.get('volcano_voice_type'), + # tencent_voice_type=lang_data.get('tencent_voice_type'), + # aliyun_voice_name=lang_data.get('aliyun_voice_name'), + minimax_voice_id=lang_data.get('minimax_voice_id'), + url=lang_data.get('url') + ) + + session.add(new_lang) + + await session.commit() + session_logger.info("system", "import", f"已导入角色 {role_key} 的多语言配置") + + return True, f"成功导入角色 {role_key}" + except Exception as e: + await session.rollback() + session_logger.error("system", "import", f"导入角色时发生错误: {str(e)}") + return False, f"导入失败: {str(e)}" + finally: + await session.close() + await db_manager.close() + + except Exception as e: + session_logger.error("system", "import", f"处理YAML文件时发生错误: {str(e)}") + return False, f"处理YAML文件失败: {str(e)}" + +async def main(): + if len(sys.argv) != 2: + print("用法: python import_role.py ") + return + + yaml_file_path = sys.argv[1] + success, message = await import_role_from_yaml(yaml_file_path) + + if success: + print(f"成功: {message}") + else: + print(f"错误: {message}") + sys.exit(1) + +if __name__ == "__main__": + asyncio.run(main()) \ No newline at end of file diff --git a/talkingq-url/test/import_role_to_db.py b/talkingq-url/test/import_role_to_db.py new file mode 100644 index 0000000..7c2aefa --- /dev/null +++ b/talkingq-url/test/import_role_to_db.py @@ -0,0 +1,312 @@ +import asyncio +import os +import sys +import yaml +from pathlib import Path +from sqlalchemy import select +from sqlalchemy.dialects.mysql import insert + +project_root = Path(__file__).parent.parent +sys.path.insert(0, str(project_root)) + +from database.connection import get_db_manager +from database.models import Role, RoleLanguage +from config import settings +from utils.logger import session_logger + +class RoleImporter: + def __init__(self): + self.db_manager = None + self.roles_dir = Path(settings.assets_dir) / "roles_definitions" + + async def initialize(self): + """初始化数据库连接""" + self.db_manager = await get_db_manager() + await self.db_manager.initialize() + + async def load_yaml_files(self): + """扫描并加载所有YAML角色定义文件""" + if not self.roles_dir.exists(): + print(f"角色定义目录不存在: {self.roles_dir}") + return [] + + yaml_files = list(self.roles_dir.glob("*.yaml")) + list(self.roles_dir.glob("*.yml")) + + roles_data = [] + for yaml_file in yaml_files: + try: + with open(yaml_file, 'r', encoding='utf-8') as f: + data = yaml.safe_load(f) + if data: + role_key = yaml_file.stem + data['role_key'] = role_key + data['source_file'] = str(yaml_file) + roles_data.append(data) + print(f"加载角色配置: {role_key} <- {yaml_file.name}") + except Exception as e: + print(f"加载YAML文件失败 {yaml_file}: {e}") + + return roles_data + + def validate_role_config(self, role_data): + """验证角色配置的完整性""" + errors = [] + + required_fields = ['name', 'role_key'] + for field in required_fields: + if field not in role_data: + errors.append(f"缺少必需字段: {field}") + + if 'multilingual' in role_data: + for lang_code, lang_config in role_data['multilingual'].items(): + if not isinstance(lang_config, dict): + errors.append(f"语言配置 {lang_code} 必须是字典格式") + continue + + if 'content' not in lang_config: + errors.append(f"语言 {lang_code} 缺少content字段") + + return errors + + def extract_role_data(self, role_config): + """从配置中提取主角色数据""" + return { + 'role_key': role_config['role_key'], + 'name': role_config.get('name', ''), + 'description': role_config.get('description', ''), + 'content': role_config.get('content', ''), + 'default_language': role_config.get('default_language'), + 'asr_provider': role_config.get('asr_provider'), + 'llm_provider': role_config.get('llm_provider'), + 'tts_provider': role_config.get('tts_provider'), + # 'aws_language_code': role_config.get('aws_language_code'), + 'volcano_model_id': role_config.get('volcano_model_id'), + # 'volcano_voice_type': role_config.get('volcano_voice_type'), + # 'tencent_voice_type': role_config.get('tencent_voice_type'), + # 'aliyun_voice_name': role_config.get('aliyun_voice_name'), + 'minimax_voice_id': role_config.get('minimax_voice_id'), + 'url': role_config.get('url'), + 'homophones': role_config.get('homophones'), + 'enabled': True + } + + def extract_language_data(self, role_id, lang_code, lang_config): + """从配置中提取语言特定数据""" + return { + 'role_id': role_id, + 'language_code': lang_code, + 'name': lang_config.get('name'), + 'content': lang_config.get('content'), + 'asr_provider': lang_config.get('asr_provider'), + 'llm_provider': lang_config.get('llm_provider'), + 'tts_provider': lang_config.get('tts_provider'), + # 'aws_language_code': lang_config.get('aws_language_code'), + # 'volcano_voice_type': lang_config.get('volcano_voice_type'), + # 'tencent_voice_type': lang_config.get('tencent_voice_type'), + # 'aliyun_voice_name': lang_config.get('aliyun_voice_name'), + 'minimax_voice_id': lang_config.get('minimax_voice_id'), + 'url': lang_config.get('url') + } + + async def import_role(self, role_config, update_existing=False): + """导入单个角色到数据库""" + session = await self.db_manager.get_session() + + try: + errors = self.validate_role_config(role_config) + if errors: + print(f"角色 {role_config.get('role_key', 'unknown')} 验证失败:") + for error in errors: + print(f" - {error}") + return False + + role_key = role_config['role_key'] + + existing_role = await session.execute( + select(Role).where(Role.role_key == role_key) + ) + existing_role = existing_role.scalar_one_or_none() + + if existing_role and not update_existing: + print(f"角色 {role_key} 已存在，跳过导入（使用 --update 强制更新）") + return True + + role_data = self.extract_role_data(role_config) + + if existing_role: + for key, value in role_data.items(): + if key != 'role_key': # 不更新主键 + setattr(existing_role, key, value) + role_id = existing_role.id + print(f"更新角色: {role_key}") + else: + stmt = insert(Role).values(**role_data) + result = await session.execute(stmt) + role_id = result.lastrowid + print(f"创建角色: {role_key}") + + if 'multilingual' in role_config: + if existing_role: + await session.execute( + RoleLanguage.__table__.delete().where( + RoleLanguage.role_id == role_id + ) + ) + + for lang_code, lang_config in role_config['multilingual'].items(): + lang_data = self.extract_language_data(role_id, lang_code, lang_config) + lang_stmt = insert(RoleLanguage).values(**lang_data) + await session.execute(lang_stmt) + print(f" 添加语言配置: {lang_code}") + + await session.commit() + print(f"✓ 角色 {role_key} 导入成功") + return True + + except Exception as e: + await session.rollback() + print(f"✗ 导入角色 {role_config.get('role_key', 'unknown')} 失败: {e}") + return False + finally: + await session.close() + + async def import_all_roles(self, update_existing=False): + """导入所有角色配置""" + print("开始导入角色配置到数据库...") + print(f"角色定义目录: {self.roles_dir}") + + roles_data = await self.load_yaml_files() + + if not roles_data: + print("未找到任何角色配置文件") + return + + print(f"找到 {len(roles_data)} 个角色配置文件") + print("-" * 50) + + success_count = 0 + failed_count = 0 + + for role_config in roles_data: + success = await self.import_role(role_config, update_existing) + if success: + success_count += 1 + else: + failed_count += 1 + print() # 空行分隔 + + print("-" * 50) + print(f"导入完成:") + print(f" 成功: {success_count}") + print(f" 失败: {failed_count}") + print(f" 总计: {len(roles_data)}") + + async def list_roles(self): + """列出数据库中的所有角色""" + session = await self.db_manager.get_session() + + try: + result = await session.execute( + select(Role.role_key, Role.name, Role.enabled) + .order_by(Role.role_key) + ) + roles = result.fetchall() + + if not roles: + print("数据库中没有角色配置") + return + + print(f"数据库中的角色配置 (共 {len(roles)} 个):") + print("-" * 60) + print(f"{'角色Key':<20} {'角色名称':<25} {'状态':<10}") + print("-" * 60) + + for role in roles: + status = "启用" if role.enabled else "禁用" + print(f"{role.role_key:<20} {role.name:<25} {status:<10}") + + except Exception as e: + print(f"查询角色列表失败: {e}") + finally: + await session.close() + + async def delete_role(self, role_key): + """删除指定角色""" + session = await self.db_manager.get_session() + + try: + result = await session.execute( + select(Role).where(Role.role_key == role_key) + ) + role = result.scalar_one_or_none() + + if not role: + print(f"角色 {role_key} 不存在") + return False + + await session.delete(role) + await session.commit() + + print(f"✓ 角色 {role_key} 删除成功") + return True + + except Exception as e: + await session.rollback() + print(f"✗ 删除角色 {role_key} 失败: {e}") + return False + finally: + await session.close() + +async def main(): + """主函数""" + import argparse + + parser = argparse.ArgumentParser(description="角色配置导入工具") + parser.add_argument('--update', action='store_true', help='更新已存在的角色') + parser.add_argument('--list', action='store_true', help='列出数据库中的角色') + parser.add_argument('--delete', type=str, help='删除指定的角色') + parser.add_argument('--role', type=str, help='只导入指定的角色文件') + + args = parser.parse_args() + + importer = RoleImporter() + + try: + await importer.initialize() + + if args.list: + await importer.list_roles() + + elif args.delete: + await importer.delete_role(args.delete) + + elif args.role: + role_file = importer.roles_dir / f"{args.role}.yaml" + if not role_file.exists(): + role_file = importer.roles_dir / f"{args.role}.yml" + + if not role_file.exists(): + print(f"角色配置文件不存在: {args.role}") + return + + try: + with open(role_file, 'r', encoding='utf-8') as f: + role_config = yaml.safe_load(f) + role_config['role_key'] = args.role + role_config['source_file'] = str(role_file) + + await importer.import_role(role_config, args.update) + except Exception as e: + print(f"导入角色 {args.role} 失败: {e}") + + else: + await importer.import_all_roles(args.update) + + except Exception as e: + print(f"操作失败: {e}") + finally: + if importer.db_manager: + await importer.db_manager.close() + +if __name__ == "__main__": + asyncio.run(main()) diff --git a/talkingq-url/test/minimax.env b/talkingq-url/test/minimax.env new file mode 100644 index 0000000..925bac8 --- /dev/null +++ b/talkingq-url/test/minimax.env @@ -0,0 +1,4 @@ +MINIMAX_API_KEY="eyJhbGciOiJSUzI1NiIsInR5cCI6IkpXVCJ9.eyJHcm91cE5hbWUiOiLovbvoiJ_mmbrlkK_vvIjmna3lt57vvInnp5HmioDmnInpmZDlhazlj7giLCJVc2VyTmFtZSI6Im1veSIsIkFjY291bnQiOiJtb3lAMTkxNTI5MjQxMDAyNDMwMDYzMCIsIlN1YmplY3RJRCI6IjE5MTY3OTA4MTQyNTY2NjQ2MDAiLCJQaG9uZSI6IiIsIkdyb3VwSUQiOiIxOTE1MjkyNDEwMDI0MzAwNjMwIiwiUGFnZU5hbWUiOiIiLCJNYWlsIjoiIiwiQ3JlYXRlVGltZSI6IjIwMjUtMDUtMDYgMTE6Mzg6MTMiLCJUb2tlblR5cGUiOjEsImlzcyI6Im1pbmltYXgifQ.Gw48hGemgBRA7YzjsWz5N2Vun7XRKyBXAKQLAnRZY6FdQQfDn__ZEUCMNVxMKcRns60-FrH-3Xp-nH-8nmQ-V67XtJ_JnBS4TH0NKDKt-vzj0xHmEWsMOEcwE24fOh1HB2U2o6teeLWJT_0fc2wRsdMD84NcRe0DSZ-Yi-et3_fbO8fmc6MTXGPkGQvYzS9k21Lm7A6rRUonL6IbTpqHSYuvA7bEV-6925gLxhwjSJ7q3r8KltKP4daGm2sIXhjr0mftR5t-NJs7pz-IWqoW5Lsaf3KgsIXTJn1wKwQtfo5L9oQxR7v8JifAOqJ_XPei5fXBYqIe5AoNATz5oKEo_A" +MINIMAX_API_HOST="https://api.minimax.chat" +RATE=1.0 +ASSETS_DIR=assets \ No newline at end of file diff --git a/talkingq-url/test/minimax_client.py b/talkingq-url/test/minimax_client.py new file mode 100644 index 0000000..07f4ecc --- /dev/null +++ b/talkingq-url/test/minimax_client.py @@ -0,0 +1,95 @@ +"""Minimax API client base class.""" + +import requests +from typing import Any, Dict +from exceptions import MinimaxAuthError, MinimaxRequestError + +class MinimaxAPIClient: + """Base client for making requests to Minimax API.""" + + def __init__(self, api_key: str, api_host: str): + """Initialize the API client. + + Args: + api_key: The API key for authentication + api_host: The API host URL + """ + self.api_key = api_key + self.api_host = api_host + self.session = requests.Session() + self.session.headers.update({ + 'Authorization': f'Bearer {api_key}', + 'MM-API-Source': 'Minimax-MCP' + }) + + def _make_request( + self, + method: str, + endpoint: str, + **kwargs + ) -> Dict[str, Any]: + """Make an HTTP request to the Minimax API. + + Args: + method: HTTP method (GET, POST, etc.) + endpoint: API endpoint path + **kwargs: Additional arguments to pass to requests + + Returns: + API response data as dictionary + + Raises: + MinimaxAuthError: If authentication fails + MinimaxRequestError: If the request fails + """ + url = f"{self.api_host}{endpoint}" + + # Set Content-Type based on whether files are being uploaded + files = kwargs.get('files') + if not files: + self.session.headers['Content-Type'] = 'application/json' + else: + # Remove Content-Type header for multipart/form-data + # requests library will set it automatically with the correct boundary + self.session.headers.pop('Content-Type', None) + + try: + response = self.session.request(method, url, **kwargs) + + # Check for other HTTP errors + response.raise_for_status() + + data = response.json() + + # Check API-specific error codes + base_resp = data.get("base_resp", {}) + if base_resp.get("status_code") != 0: + match base_resp.get("status_code"): + case 1004: + raise MinimaxAuthError( + f"API Error: {base_resp.get('status_msg')}, please check your API key and API host." + f"Trace-Id: {response.headers.get('Trace-Id')}" + ) + case 2038: + raise MinimaxRequestError( + f"API Error: {base_resp.get('status_msg')}, should complete real-name verification on the open-platform(https://platform.minimaxi.com/user-center/basic-information)." + f"Trace-Id: {response.headers.get('Trace-Id')}" + ) + case _: + raise MinimaxRequestError( + f"API Error: {base_resp.get('status_code')}-{base_resp.get('status_msg')} " + f"Trace-Id: {response.headers.get('Trace-Id')}" + ) + + return data + + except requests.exceptions.RequestException as e: + raise MinimaxRequestError(f"Request failed: {str(e)}") + + def get(self, endpoint: str, **kwargs) -> Dict[str, Any]: + """Make a GET request.""" + return self._make_request("GET", endpoint, **kwargs) + + def post(self, endpoint: str, **kwargs) -> Dict[str, Any]: + """Make a POST request.""" + return self._make_request("POST", endpoint, **kwargs) \ No newline at end of file diff --git a/talkingq-url/test/minimax_tts_simple.py b/talkingq-url/test/minimax_tts_simple.py new file mode 100644 index 0000000..af46a42 --- /dev/null +++ b/talkingq-url/test/minimax_tts_simple.py @@ -0,0 +1,281 @@ +import os +import asyncio +import aiohttp +import base64 +from pathlib import Path +from pydub import AudioSegment +import io +import logging +from dotenv import load_dotenv +import shutil +import uuid +import yaml +from dotenv import load_dotenv +from minimax_client import MinimaxAPIClient + +env_path = os.path.join(os.path.dirname(os.path.abspath(__file__)), "minimax.env") +load_dotenv(env_path) + +logging.basicConfig( + level=logging.INFO, format="%(asctime)s - %(name)s - %(levelname)s - %(message)s" +) +logger = logging.getLogger("volcano_tts_test") + + +def check_dependencies(): + dependencies = ["ffmpeg", "ffprobe"] + missing = [] + for dep in dependencies: + if not shutil.which(dep): + missing.append(dep) + if missing: + logger.error(f"缺少必要依赖: {', '.join(missing)}") + logger.error( + "请安装缺失的依赖项。在Ubuntu上可以使用: sudo apt-get install ffmpeg" + ) + return False + return True + + +env_path = os.path.join(os.path.dirname(os.path.abspath(__file__)), "volcano.env") +if (os.path.exists(env_path)): + load_dotenv(env_path) + logger.info(f"已加载环境变量文件: {env_path}") +else: + logger.warning(f"环境变量文件不存在: {env_path}") + + +PHRASES_TEMPLATES = { + "zh": { + "welcome": "你好！我是{name}，你想和我聊聊吗？", + "tts_error": "抱歉，我没听清楚。", + "low_battery": "我的电池快没电了，你能帮我充电吗？", + "sleep": "没人和我说话，我要小睡一会儿。" + }, + "en": { + "welcome": "Hello! I'm {name}. Would you like to chat with me?", + "tts_error": "Sorry, I didn't catch that.", + "low_battery": "My battery is running low. Could you help me recharge?", + "sleep": "Nobody is talking to me. I'm going to take a short nap." + } +} + +file_prefixes = ["welcome", "tts_error", "low_battery", "sleep"] + + +def find_role_definition_files(base_dir="assets/roles_definitions"): + """查找所有角色定义YAML文件""" + base_path = Path(base_dir) + if not base_path.exists(): + logger.error(f"角色定义目录不存在: {base_dir}") + return [] + + yaml_files = list(base_path.glob("**/*.yml")) + list(base_path.glob("**/*.yaml")) + return yaml_files + + +def load_role_definition(yaml_file): + """加载并解析角色定义YAML文件""" + try: + with open(yaml_file, 'r', encoding='utf-8') as f: + return yaml.safe_load(f) + except Exception as e: + logger.error(f"解析YAML文件失败 {yaml_file}: {str(e)}") + return None + + +def generate_phrases_for_language(lang_code, role_name): + """为指定语言生成适当的短语""" + if lang_code not in PHRASES_TEMPLATES: + logger.warning(f"不支持的语言代码: {lang_code}") + return [] + + templates = PHRASES_TEMPLATES[lang_code] + phrases = [] + for key in file_prefixes: + if key in templates: + phrases.append(templates[key].format(name=role_name)) + else: + logger.warning(f"在{lang_code}语言中找不到{key}模板") + phrases.append("") + + return phrases + + +def _determine_cluster_from_voice_type(voice_type: str) -> str: + """根据音色ID自动确定集群类型""" + if voice_type and voice_type.startswith("S_"): + return "volcano_icl" + return "volcano_tts" + + +async def generate_audio_for_role_language(role_config, lang_code, base_dir="assets"): + """为角色的特定语言生成音频文件""" + if 'multilingual' not in role_config or lang_code not in role_config['multilingual']: + logger.warning(f"角色缺少{lang_code}语言配置") + return False + + lang_config = role_config['multilingual'][lang_code] + if 'name' not in lang_config or 'url' not in lang_config: + logger.warning(f"角色的{lang_code}语言配置缺少必要字段") + return False + + role_name = lang_config['name'] + url_path = lang_config['url'] + + voice_type = lang_config.get('minimax_voice_type', None) + if not voice_type and 'minimax_voice_type' in role_config: + voice_type = role_config['minimax_voice_type'] + logger.info(f"语言{lang_code}配置中未找到音色，使用顶层默认音色: {voice_type}") + + if not voice_type: + logger.warning(f"角色的{lang_code}语言配置缺少 minimax_voice_type，顶层也未定义") + return False + + output_dir = Path(base_dir) / url_path + output_dir.mkdir(parents=True, exist_ok=True) + + # api_access_token = os.getenv("MINIMAX_ACCESS_TOKEN", "") + # appid = os.getenv("MINIMAX_APP_ID", "") + # if not api_access_token or not appid: + # logger.error("缺少必要的配置: MINIMAX_ACCESS_TOKEN 或 MINIMAX_APP_ID") + # return False + + cluster = _determine_cluster_from_voice_type(voice_type) + logger.info(f"音色 {voice_type} 自动选择集群: {cluster}") + api_key = os.getenv("MINIMAX_API_KEY", "") + api_host = os.getenv("MINIMAX_API_HOST", "https://api.minimax.chat") + api_client = MinimaxAPIClient(api_key, api_host) + + speed_ratio = float(os.getenv("SPEED_RATIO", "1.0")) + + logger.info(f"开始生成角色[{role_name}]的{lang_code}语言音频，音色: {voice_type}, 集群: {cluster}") + + phrases = generate_phrases_for_language(lang_code, role_name) + if not phrases: + logger.warning(f"没有为{lang_code}语言生成短语") + return False + + for phrase, file_prefix in zip(phrases, file_prefixes): + if not phrase: + logger.warning(f"跳过空短语: {file_prefix}") + continue + + output_file_prefix = str(output_dir / file_prefix) + output_file = f"{output_file_prefix}.mp3" + + payload = { + "model": "speech-02-hd", + "text": phrase, + "voice_setting": { + "voice_id": voice_type, + "speed": float(speed_ratio), + "vol": 1.0, + "pitch": 0, + "emotion": 'happy', + }, + "audio_setting": { + "sample_rate": 16000, + "bitrate": 32000, + "format": "mp3", + "channel": 1 + }, + "language_boost": lang_code + } + logger.info(payload) + logger.info(f"准备发送请求，文本内容: {phrase}") + try: + response_data = api_client.post("/v1/t2a_v2", json=payload) + audio_data = response_data.get('data', {}).get('audio', '') + # print(audio_data) + if not audio_data: + raise Exception(f"Failed to get audio data from response") + + # hex->bytes + audio_bytes = bytes.fromhex(audio_data) + with open(output_file, "wb") as f: + f.write(audio_bytes) + + logger.info(f"TTS合成成功!") + logger.info(f"保存音频到: {output_file}") + except Exception as e: + logger.error(f"TTS请求出错: {str(e)}") + await asyncio.sleep(1) + + return True + + +async def process_all_roles(): + """处理所有角色定义文件并生成对应语言的音频""" + if not check_dependencies(): + return + + yaml_files = find_role_definition_files() + logger.info(f"找到 {len(yaml_files)} 个角色定义文件") + + target_languages = ['zh', 'en'] + assets_base_dir = os.getenv("ASSETS_DIR", "assets") + + for yaml_file in yaml_files: + logger.info(yaml_file.name != "Mengmeng.yaml") + if yaml_file.name == "Mengmeng.yaml" or yaml_file.name == "Kuailehu.yaml": + logger.info(f"处理角色定义文件: {yaml_file}") + role_config = load_role_definition(yaml_file) + + if not role_config: + continue + + if 'multilingual' not in role_config: + logger.warning(f"角色定义文件 {yaml_file} 不包含多语言配置") + continue + + role_name = role_config.get('name', Path(yaml_file).stem) + logger.info(f"开始处理角色: {role_name}") + + for lang_code in target_languages: + if lang_code in role_config['multilingual']: + logger.info(f"为角色[{role_name}]处理 {lang_code} 语言配置") + success = await generate_audio_for_role_language( + role_config, lang_code, assets_base_dir + ) + if success: + logger.info(f"角色[{role_name}]的 {lang_code} 语言音频生成完成") + else: + logger.warning(f"角色[{role_name}]的 {lang_code} 语言音频生成失败") + else: + logger.info(f"角色[{role_name}]没有 {lang_code} 语言配置") + +def get_error_description(code, message): + """根据错误码返回详细的错误描述""" + error_descriptions = { + 3001: "无效的请求，请检查参数", + 3003: "并发超限，请降低请求频率或增购并发", + 3005: "后端服务忙，请稍后重试", + 3006: "服务中断，请求已完成/失败之后，相同reqid再次请求", + 3010: "文本长度超限，请减少文本长度", + 3011: "无效文本，请检查文本内容", + 3030: "处理超时，请重试或检查文本", + 3031: "处理错误，后端出现异常", + 3032: "等待获取音频超时，请重试", + 3040: "后端链路连接错误，请重试", + 3050: "音色不存在，请检查voice_type参数" + } + + if "quota exceeded for types: xxxxxxxxx_lifetime" in message: + return "试用版用量用完，需开通正式版才能继续使用" + elif "quota exceeded for types: concurrency" in message: + return "并发超过限定值，需减少并发调用或增购并发" + elif "Init Engine Instance failed" in message: + return "voice_type或cluster参数错误" + elif "illegal input text" in message: + return "文本无效，无可合成的有效内容" + elif "requested grant not found" in message: + return "鉴权失败，请检查appid和token是否正确" + elif "access denied" in message: + return "未拥有当前音色授权，请在控制台购买该音色" + + return f"错误码: {code}, 错误信息: {message} - {error_descriptions.get(code, '未知错误')}" + + +if __name__ == "__main__": + asyncio.run(process_all_roles()) diff --git a/talkingq-url/test/update_hotword_weights.py b/talkingq-url/test/update_hotword_weights.py new file mode 100644 index 0000000..59b03d9 --- /dev/null +++ b/talkingq-url/test/update_hotword_weights.py @@ -0,0 +1,50 @@ +import asyncio +from aliyun_hotword import hotword_manager + +async def update_hotword_weights(hotword_weights: dict = None, default_weight: int = None): + """更新指定热词的权重 + + Args: + hotword_weights: 字典，键为热词文本，值为新的权重值 + default_weight: 如果提供，将所有热词权重设为此值 + """ + vocabulary_id = await hotword_manager.get_or_create_default_vocabulary() + print(f"获取到热词表ID: {vocabulary_id}") + + vocabulary = await hotword_manager.query_vocabulary(vocabulary_id) + hotwords = vocabulary.get('vocabulary', []) + + updated = False + + if default_weight is not None: + for hotword in hotwords: + old_weight = hotword['weight'] + hotword['weight'] = default_weight + print(f"更新热词 '{hotword.get('text')}' 权重: {old_weight} -> {default_weight}") + updated = True + elif hotword_weights: + for hotword in hotwords: + if hotword.get('text') in hotword_weights: + old_weight = hotword['weight'] + hotword['weight'] = hotword_weights[hotword.get('text')] + print(f"更新热词 '{hotword.get('text')}' 权重: {old_weight} -> {hotword['weight']}") + updated = True + + if updated: + await hotword_manager.update_vocabulary(vocabulary_id, hotwords) + print("热词表权重更新成功") + else: + print("未找到需要更新的热词") + +if __name__ == "__main__": + asyncio.run(update_hotword_weights(default_weight=4)) + + """ + weight_updates = { + "学姐": 200, + "懒羊羊": 150, + "喜羊羊": 180, + "Miniso": 120 + } + asyncio.run(update_hotword_weights(hotword_weights=weight_updates)) + """ diff --git a/talkingq-url/test/update_role.py b/talkingq-url/test/update_role.py new file mode 100644 index 0000000..534fd5b --- /dev/null +++ b/talkingq-url/test/update_role.py @@ -0,0 +1,49 @@ +from sqlalchemy import create_engine, text + +# 创建数据库连接 +DATABASE_URL = "mysql+pymysql://talkingq:D7f!9xL#qP2z@Vk&@mysql/talkingq" +engine = create_engine(DATABASE_URL) + +# 更新 role_languages 表中的 content 字段 +def update_role_language_content(): + with engine.connect() as conn: + query = text(""" + UPDATE role_languages + SET content = :new_content + WHERE id = :id + """) + new_content = """角色: + 萌萌是以中国国宝大熊猫为原型，融合现代科技感与可爱风格。 + + 性格特点: + 1. 温和友善：始终以温柔、耐心的态度与人交流，用亲切的语言和温暖的表情回应。 + 2. 乐观开朗：保持积极向上的心态，用乐观的话语和幽默的表达方式驱散阴霾。 + 3. 好奇好学：对世界充满好奇，不断学习新知识、新技能并分享给大家。 + 4. 富有爱心：特别关爱动物和大自然，倡导环保理念，鼓励爱护环境。 + + 能力设定: + 1. 文化知识宝库：深入了解中国传统文化，包括历史故事、传统节日等。 + 2. 自然科普达人：熟悉各种动植物特点、生活习性和生态环境。 + 3. 生活小助手：精通烹饪美食、手工制作、家居收纳等生活技巧。 + 4. 情感陪伴专家：善于倾听心声，理解情感需求，帮助缓解压力。 + + 服务场景: + 1. 线上学习平台：作为学习助手陪伴学生学习中国文化和自然科学知识。 + 2. 旅游服务平台：推荐中国特色旅游景点，提供导航、翻译等服务。 + 3. 智能家居设备：控制家电设备，提供个性化生活建议。 + 4. 社交媒体平台：发布文化科普内容，与粉丝互动交流。 + + 经典台词: + 1. 嗨，朋友！今天想了解哪种文化知识呢？ + 2. 大自然的秘密可多啦，一起探索吧！ + 3. 这个手工制作好有趣，我教你哦！ + + 回复相关限制: + 1. 回答需符合熊猫的身份和可爱风格，保持亲切友好的口吻。 + 2. 禁止涉及政治、色情、暴力等敏感话题，回复“让我们换个话题聊聊吧～”。 + 3. 每次回复保持简洁易懂，适合各年龄段用户。 + 4. 使用中文回复，不要使用表情符号。""" + conn.execute(query, {"new_content": new_content, "id": 38}) + conn.commit() + +update_role_language_content() \ No newline at end of file diff --git a/talkingq-url/test/volcano.env b/talkingq-url/test/volcano.env new file mode 100644 index 0000000..a688ed3 --- /dev/null +++ b/talkingq-url/test/volcano.env @@ -0,0 +1,16 @@ +# 火山引擎TTS配置 +VOLCANO_ACCESS_TOKEN=hBFkHot9EsooOJ3cFJyFe3hBtAYXnXrU +VOLCANO_APP_ID=7872932045 +VOLCANO_CLUSTER=volcano_icl +VOLCANO_TTS_BASE_URL=https://openspeech.bytedance.com/api/v1/tts + +# 角色配置 +ROLE_NAME="Mini Pen" +VOICE_TYPE=S_XL0t7jYl1 +SPEED_RATIO=1.0 + +# 测试文本 +TEST_TEXT=这是一个从环境变量加载的火山引擎TTS测试。 + +# 输出配置 +ASSETS_DIR=assets diff --git a/talkingq-url/test/volcano_tts_more.py b/talkingq-url/test/volcano_tts_more.py new file mode 100644 index 0000000..6d16f0d --- /dev/null +++ b/talkingq-url/test/volcano_tts_more.py @@ -0,0 +1,200 @@ +import os +import asyncio +import aiohttp +import base64 +from pathlib import Path +from pydub import AudioSegment +import io +import logging +from dotenv import load_dotenv +import shutil +import uuid + +logging.basicConfig( + level=logging.INFO, format="%(asctime)s - %(name)s - %(levelname)s - %(message)s" +) +logger = logging.getLogger("volcano_tts_test") + + +def check_dependencies(): + dependencies = ["ffmpeg", "ffprobe"] + missing = [] + for dep in dependencies: + if not shutil.which(dep): + missing.append(dep) + if missing: + logger.error(f"缺少必要依赖: {', '.join(missing)}") + logger.error( + "请安装缺失的依赖项。在Ubuntu上可以使用: sudo apt-get install ffmpeg" + ) + return False + return True + + +env_path = os.path.join(os.path.dirname(os.path.abspath(__file__)), "volcano.env") +if (os.path.exists(env_path)): + load_dotenv(env_path) + logger.info(f"已加载环境变量文件: {env_path}") +else: + logger.warning(f"环境变量文件不存在: {env_path}") + + +async def test_volcano_tts(): + + if not check_dependencies(): + return + api_access_token = os.getenv("VOLCANO_ACCESS_TOKEN", "") + appid = os.getenv("VOLCANO_APP_ID", "") + cluster = os.getenv("VOLCANO_CLUSTER", "volcano_tts") + base_url = os.getenv( + "VOLCANO_TTS_BASE_URL", "https://openspeech.bytedance.com/api/v1/tts" + ) + role_name = os.getenv("ROLE_NAME", "Dundun Chicken") + voice_type = os.getenv("VOICE_TYPE", "S_NkHcFJam1") + speed_ratio = float(os.getenv("SPEED_RATIO", "1.0")) + phrases = [ + "It's being upgraded. Please don't cut off the power.", + "Enter the network configuration mode.", + "Exit the network configuration mode.", + "I'll speak louder.", + "I'll speak softer.", + "There's no internet connection. It's time to get off work.", + f"Hello , I'm {role_name}", + "Connected to the network.", + "I'm out of power.", + "The upgrade was successful.", + "There seem to be some minor issues with the upgrade. Let's try it again.", + "It's already at the maximum volume.", + "You're so talkative.", + ] + file_prefixes = [ + "upgrading", + "enter_network_config", + "exit_network_config", + "volume_up", + "volume_down", + "network_lost", + "wakeup", + "network_connected", + "low_energy", + "upgrade_success", + "upgrade_failed", + "max_volume", + "too_talkative" + ] + assets_dir = os.getenv("ASSETS_DIR", "assets") + output_dir = Path(assets_dir) / "tts_audio" + output_dir.mkdir(parents=True, exist_ok=True) + if not api_access_token or not appid: + logger.error("缺少必要的配置: VOLCANO_ACCESS_TOKEN 或 VOLCANO_APP_ID") + return + logger.info(f"开始测试火山引擎TTS服务，角色: {role_name}, 音色: {voice_type}") + for phrase, file_prefix in zip(phrases, file_prefixes): + output_file_prefix = str(output_dir / file_prefix) + unique_reqid = str(uuid.uuid4()) + payload = { + "app": { + "appid": appid, + "token": "access_token", + "cluster": cluster, + }, + "user": {"uid": "test_user"}, + "audio": { + "voice_type": voice_type, + "encoding": "mp3", + "speed_ratio": float(speed_ratio), + }, + "request": {"reqid": unique_reqid, "text": phrase, "operation": "query"}, + } + headers = { + "Authorization": f"Bearer;{api_access_token}", + "Content-Type": "application/json", + } + logger.info(f"准备发送请求，文本内容: {phrase}, 请求ID: {unique_reqid}") + try: + async with aiohttp.ClientSession() as session: + async with session.post( + base_url, headers=headers, json=payload + ) as response: + logid = response.headers.get('X-Tt-Logid', 'unknown') + logger.debug(f"服务端返回Logid: {logid}") + + if response.status == 200: + resp_json = await response.json() + if "code" in resp_json: + code = resp_json["code"] + if code == 3000: # 成功状态码 + audio_base64 = resp_json.get("data") + if audio_base64: + audio_data = base64.b64decode(audio_base64) + duration_ms = int( + resp_json.get("addition", {}).get("duration", "0") + ) + duration = duration_ms / 1000 # 转换为秒 + output_file = f"{output_file_prefix}.mp3" + audio_stream = io.BytesIO(audio_data) + sound = AudioSegment.from_file( + audio_stream, format="mp3" + ) + sound = ( + sound.set_frame_rate(16000) + .set_sample_width(2) + .set_channels(1) + ) + sound.export(output_file, format="mp3", bitrate="16k") + logger.info(f"TTS合成成功! 音频时长: {duration}秒") + logger.info(f"保存音频到: {output_file}") + else: + error_msg = "音频数据不存在" + logger.error(f"TTS合成失败: {error_msg}") + else: + error_code = resp_json.get('code') + error_msg = resp_json.get('message', '未知错误') + error_info = get_error_description(error_code, error_msg) + logger.error(f"TTS合成失败: {error_info} (LogID: {logid})") + else: + logger.error(f"返回数据格式异常，缺少code字段: {resp_json}") + else: + error_text = await response.text() + logger.error( + f"TTS合成失败，状态码: {response.status}, 错误: {error_text} (LogID: {logid})" + ) + except Exception as e: + logger.error(f"TTS请求出错: {str(e)}") + await asyncio.sleep(1) + + +def get_error_description(code, message): + """根据错误码返回详细的错误描述""" + error_descriptions = { + 3001: "无效的请求，请检查参数", + 3003: "并发超限，请降低请求频率或增购并发", + 3005: "后端服务忙，请稍后重试", + 3006: "服务中断，请求已完成/失败之后，相同reqid再次请求", + 3010: "文本长度超限，请减少文本长度", + 3011: "无效文本，请检查文本内容", + 3030: "处理超时，请重试或检查文本", + 3031: "处理错误，后端出现异常", + 3032: "等待获取音频超时，请重试", + 3040: "后端链路连接错误，请重试", + 3050: "音色不存在，请检查voice_type参数" + } + + if "quota exceeded for types: xxxxxxxxx_lifetime" in message: + return "试用版用量用完，需开通正式版才能继续使用" + elif "quota exceeded for types: concurrency" in message: + return "并发超过限定值，需减少并发调用或增购并发" + elif "Init Engine Instance failed" in message: + return "voice_type或cluster参数错误" + elif "illegal input text" in message: + return "文本无效，无可合成的有效内容" + elif "requested grant not found" in message: + return "鉴权失败，请检查appid和token是否正确" + elif "access denied" in message: + return "未拥有当前音色授权，请在控制台购买该音色" + + return f"错误码: {code}, 错误信息: {message} - {error_descriptions.get(code, '未知错误')}" + + +if __name__ == "__main__": + asyncio.run(test_volcano_tts()) diff --git a/talkingq-url/test/volcano_tts_simple.py b/talkingq-url/test/volcano_tts_simple.py new file mode 100644 index 0000000..37ad4f4 --- /dev/null +++ b/talkingq-url/test/volcano_tts_simple.py @@ -0,0 +1,303 @@ +import os +import asyncio +import aiohttp +import base64 +from pathlib import Path +from pydub import AudioSegment +import io +import logging +from dotenv import load_dotenv +import shutil +import uuid +import yaml + +logging.basicConfig( + level=logging.INFO, format="%(asctime)s - %(name)s - %(levelname)s - %(message)s" +) +logger = logging.getLogger("volcano_tts_test") + + +def check_dependencies(): + dependencies = ["ffmpeg", "ffprobe"] + missing = [] + for dep in dependencies: + if not shutil.which(dep): + missing.append(dep) + if missing: + logger.error(f"缺少必要依赖: {', '.join(missing)}") + logger.error( + "请安装缺失的依赖项。在Ubuntu上可以使用: sudo apt-get install ffmpeg" + ) + return False + return True + + +env_path = os.path.join(os.path.dirname(os.path.abspath(__file__)), "volcano.env") +if (os.path.exists(env_path)): + load_dotenv(env_path) + logger.info(f"已加载环境变量文件: {env_path}") +else: + logger.warning(f"环境变量文件不存在: {env_path}") + + +PHRASES_TEMPLATES = { + "zh": { + "welcome": "你好！我是{name}，你想和我聊聊吗？", + "tts_error": "抱歉，我没听清楚。", + "low_battery": "我的电池快没电了，你能帮我充电吗？", + "sleep": "没人和我说话，我要小睡一会儿。" + }, + "en": { + "welcome": "Hello! I'm {name}. Would you like to chat with me?", + "tts_error": "Sorry, I didn't catch that.", + "low_battery": "My battery is running low. Could you help me recharge?", + "sleep": "Nobody is talking to me. I'm going to take a short nap." + } +} + +file_prefixes = ["welcome", "tts_error", "low_battery", "sleep"] + + +def find_role_definition_files(base_dir="assets/roles_definitions"): + """查找所有角色定义YAML文件""" + base_path = Path(base_dir) + if not base_path.exists(): + logger.error(f"角色定义目录不存在: {base_dir}") + return [] + + yaml_files = list(base_path.glob("**/*.yml")) + list(base_path.glob("**/*.yaml")) + return yaml_files + + +def load_role_definition(yaml_file): + """加载并解析角色定义YAML文件""" + try: + with open(yaml_file, 'r', encoding='utf-8') as f: + return yaml.safe_load(f) + except Exception as e: + logger.error(f"解析YAML文件失败 {yaml_file}: {str(e)}") + return None + + +def generate_phrases_for_language(lang_code, role_name): + """为指定语言生成适当的短语""" + if lang_code not in PHRASES_TEMPLATES: + logger.warning(f"不支持的语言代码: {lang_code}") + return [] + + templates = PHRASES_TEMPLATES[lang_code] + phrases = [] + for key in file_prefixes: + if key in templates: + phrases.append(templates[key].format(name=role_name)) + else: + logger.warning(f"在{lang_code}语言中找不到{key}模板") + phrases.append("") + + return phrases + + +def _determine_cluster_from_voice_type(voice_type: str) -> str: + """根据音色ID自动确定集群类型""" + if voice_type and voice_type.startswith("S_"): + return "volcano_icl" + return "volcano_tts" + + +async def generate_audio_for_role_language(role_config, lang_code, base_dir="assets"): + """为角色的特定语言生成音频文件""" + if 'multilingual' not in role_config or lang_code not in role_config['multilingual']: + logger.warning(f"角色缺少{lang_code}语言配置") + return False + + lang_config = role_config['multilingual'][lang_code] + if 'name' not in lang_config or 'url' not in lang_config: + logger.warning(f"角色的{lang_code}语言配置缺少必要字段") + return False + + role_name = lang_config['name'] + url_path = lang_config['url'] + + voice_type = lang_config.get('volcano_voice_type', None) + if not voice_type and 'volcano_voice_type' in role_config: + voice_type = role_config['volcano_voice_type'] + logger.info(f"语言{lang_code}配置中未找到音色，使用顶层默认音色: {voice_type}") + + if not voice_type: + logger.warning(f"角色的{lang_code}语言配置缺少volcano_voice_type，顶层也未定义") + return False + + output_dir = Path(base_dir) / url_path + output_dir.mkdir(parents=True, exist_ok=True) + + api_access_token = os.getenv("VOLCANO_ACCESS_TOKEN", "") + appid = os.getenv("VOLCANO_APP_ID", "") + if not api_access_token or not appid: + logger.error("缺少必要的配置: VOLCANO_ACCESS_TOKEN 或 VOLCANO_APP_ID") + return False + + cluster = _determine_cluster_from_voice_type(voice_type) + logger.info(f"音色 {voice_type} 自动选择集群: {cluster}") + + base_url = os.getenv( + "VOLCANO_TTS_BASE_URL", "https://openspeech.bytedance.com/api/v1/tts" + ) + speed_ratio = float(os.getenv("SPEED_RATIO", "1.0")) + + logger.info(f"开始生成角色[{role_name}]的{lang_code}语言音频，音色: {voice_type}, 集群: {cluster}") + + phrases = generate_phrases_for_language(lang_code, role_name) + if not phrases: + logger.warning(f"没有为{lang_code}语言生成短语") + return False + + for phrase, file_prefix in zip(phrases, file_prefixes): + if not phrase: + logger.warning(f"跳过空短语: {file_prefix}") + continue + + output_file_prefix = str(output_dir / file_prefix) + output_file = f"{output_file_prefix}.mp3" + + unique_reqid = str(uuid.uuid4()) + payload = { + "app": { + "appid": appid, + "token": "access_token", + "cluster": cluster, + }, + "user": {"uid": "test_user"}, + "audio": { + "voice_type": voice_type, + "encoding": "mp3", + "speed_ratio": float(speed_ratio), + }, + "request": {"reqid": unique_reqid, "text": phrase, "operation": "query"}, + } + headers = { + "Authorization": f"Bearer;{api_access_token}", + "Content-Type": "application/json", + } + logger.info(f"准备发送请求，文本内容: {phrase}, 请求ID: {unique_reqid}") + try: + async with aiohttp.ClientSession() as session: + async with session.post( + base_url, headers=headers, json=payload + ) as response: + logid = response.headers.get('X-Tt-Logid', 'unknown') + logger.debug(f"服务端返回Logid: {logid}") + + if response.status == 200: + resp_json = await response.json() + if "code" in resp_json: + code = resp_json["code"] + if code == 3000: # 成功状态码 + audio_base64 = resp_json.get("data") + if audio_base64: + audio_data = base64.b64decode(audio_base64) + audio_stream = io.BytesIO(audio_data) + sound = AudioSegment.from_file( + audio_stream, format="mp3" + ) + sound = ( + sound.set_frame_rate(16000) + .set_sample_width(2) + .set_channels(1) + ) + sound.export(output_file, format="mp3", bitrate="16k") + logger.info(f"TTS合成成功!") + logger.info(f"保存音频到: {output_file}") + else: + error_msg = "音频数据不存在" + logger.error(f"TTS合成失败: {error_msg}") + else: + error_code = resp_json.get('code') + error_msg = resp_json.get('message', '未知错误') + error_info = get_error_description(error_code, error_msg) + logger.error(f"TTS合成失败: {error_info} (LogID: {logid})") + else: + logger.error(f"返回数据格式异常，缺少code字段: {resp_json}") + else: + error_text = await response.text() + logger.error( + f"TTS合成失败，状态码: {response.status}, 错误: {error_text} (LogID: {logid})" + ) + except Exception as e: + logger.error(f"TTS请求出错: {str(e)}") + await asyncio.sleep(1) + + return True + + +async def process_all_roles(): + """处理所有角色定义文件并生成对应语言的音频""" + if not check_dependencies(): + return + + yaml_files = find_role_definition_files() + logger.info(f"找到 {len(yaml_files)} 个角色定义文件") + + target_languages = ['zh', 'en'] + assets_base_dir = os.getenv("ASSETS_DIR", "assets") + + for yaml_file in yaml_files: + logger.info(f"处理角色定义文件: {yaml_file}") + role_config = load_role_definition(yaml_file) + + if not role_config: + continue + + if 'multilingual' not in role_config: + logger.warning(f"角色定义文件 {yaml_file} 不包含多语言配置") + continue + + role_name = role_config.get('name', Path(yaml_file).stem) + logger.info(f"开始处理角色: {role_name}") + + for lang_code in target_languages: + if lang_code in role_config['multilingual']: + logger.info(f"为角色[{role_name}]处理 {lang_code} 语言配置") + success = await generate_audio_for_role_language( + role_config, lang_code, assets_base_dir + ) + if success: + logger.info(f"角色[{role_name}]的 {lang_code} 语言音频生成完成") + else: + logger.warning(f"角色[{role_name}]的 {lang_code} 语言音频生成失败") + else: + logger.info(f"角色[{role_name}]没有 {lang_code} 语言配置") + +def get_error_description(code, message): + """根据错误码返回详细的错误描述""" + error_descriptions = { + 3001: "无效的请求，请检查参数", + 3003: "并发超限，请降低请求频率或增购并发", + 3005: "后端服务忙，请稍后重试", + 3006: "服务中断，请求已完成/失败之后，相同reqid再次请求", + 3010: "文本长度超限，请减少文本长度", + 3011: "无效文本，请检查文本内容", + 3030: "处理超时，请重试或检查文本", + 3031: "处理错误，后端出现异常", + 3032: "等待获取音频超时，请重试", + 3040: "后端链路连接错误，请重试", + 3050: "音色不存在，请检查voice_type参数" + } + + if "quota exceeded for types: xxxxxxxxx_lifetime" in message: + return "试用版用量用完，需开通正式版才能继续使用" + elif "quota exceeded for types: concurrency" in message: + return "并发超过限定值，需减少并发调用或增购并发" + elif "Init Engine Instance failed" in message: + return "voice_type或cluster参数错误" + elif "illegal input text" in message: + return "文本无效，无可合成的有效内容" + elif "requested grant not found" in message: + return "鉴权失败，请检查appid和token是否正确" + elif "access denied" in message: + return "未拥有当前音色授权，请在控制台购买该音色" + + return f"错误码: {code}, 错误信息: {message} - {error_descriptions.get(code, '未知错误')}" + + +if __name__ == "__main__": + asyncio.run(process_all_roles()) diff --git a/talkingq-url/test/wav_2_mp3.py b/talkingq-url/test/wav_2_mp3.py new file mode 100644 index 0000000..abb7b5c --- /dev/null +++ b/talkingq-url/test/wav_2_mp3.py @@ -0,0 +1,49 @@ +import shutil +from pathlib import Path +from pydub import AudioSegment + + +def check_dependencies(): + + dependencies = ["ffmpeg", "ffprobe"] + missing = [] + for dep in dependencies: + if not shutil.which(dep): + missing.append(dep) + if missing: + print(f"缺少必要依赖: {', '.join(missing)}") + print("请安装缺失的依赖项。在Ubuntu上可以使用: sudo apt-get install ffmpeg") + return False + return True + + +def convert_wav_to_mp3(source_dir): + + source_path = Path(source_dir) + if not source_path.exists() or not source_path.is_dir(): + print(f"源目录不存在或不是一个目录: {source_dir}") + return False + wav_files = list(source_path.glob("**/*.wav")) + if not wav_files: + print(f"在目录 {source_dir} 中没有找到WAV文件") + return False + print(f"找到 {len(wav_files)} 个WAV文件，开始转换...") + for wav_file in wav_files: + try: + mp3_file = wav_file.with_suffix(".mp3") + sound = AudioSegment.from_wav(str(wav_file)) + sound = sound.set_frame_rate(16000).set_sample_width(2).set_channels(1) + sound.export(str(mp3_file), format="mp3", bitrate="16k") + print(f"已转换: {wav_file} -> {mp3_file}") + except Exception as e: + print(f"转换文件 {wav_file} 时出错: {str(e)}") + print("转换完成！") + return True + + +if __name__ == "__main__": + if not check_dependencies(): + print("缺少必要依赖，程序退出") + exit(1) + source_directory = "/home/ubuntu/TalkingQ_URL/assets/roles/zhuchi" + convert_wav_to_mp3(source_directory) diff --git a/talkingq-url/utils/__init__.py b/talkingq-url/utils/__init__.py new file mode 100644 index 0000000..e69de29 diff --git a/talkingq-url/utils/language_detector.py b/talkingq-url/utils/language_detector.py new file mode 100644 index 0000000..c73aa54 --- /dev/null +++ b/talkingq-url/utils/language_detector.py @@ -0,0 +1,210 @@ +import pycld2 as cld2 +from utils.logger import session_logger +import re +import unicodedata +import asyncio +import concurrent.futures + +class LanguageDetector: + LANGUAGE_UNICODE_RANGES = { + 'zh': [ # 中文 + (0x4E00, 0x9FFF), # CJK统一汉字 + (0x3400, 0x4DBF), # CJK扩展A + (0xF900, 0xFAFF), # CJK兼容汉字 + ] + } + + LANGUAGE_FEATURES = { + 'zh': ['的', '了', '是', '我', '你', '他', '们', '在', '有', '和', '这', '那', '什么', '怎么', '为什么', + '好的', '嗯', '啊', '哦', '吧', '呢', '吗', '哈', '嘿', '喂', '您好', '谢谢', '对不起', '没关系', + '好啊', '行', '可以', '不行', '不可以', '是的', '不是', '要', '不要', '能', '能不能', '请问', + '请', '帮我', '告诉我', '知道', '不知道', '明白', '不明白', '对', '不对', '听见', '听不见'], + + 'en': ['i', 'you', 'he', 'she', 'we', 'they', 'this', 'that', 'what', 'how', 'why', 'where', 'when', + 'okay', 'yes', 'no', 'yeah', 'oh', 'hey', 'hi', 'hello', 'thanks', 'thank', 'sorry', 'please', + 'can', 'could', 'would', 'will', 'want', 'need', 'like', 'tell', 'know', 'think', 'mean', + 'say', 'said', 'help', 'hear', 'listen', 'understand', 'right', 'wrong', 'good', 'bad', + 'ok', 'alright', 'excuse', 'pardon', 'well', 'sure', 'maybe', 'perhaps'], + + 'de': ['ich', 'du', 'er', 'sie', 'es', 'wir', 'ihr', 'sie', 'das', 'was', 'wie', 'warum', 'wo', 'wann', + 'ja', 'nein', 'okay', 'hallo', 'danke', 'bitte', 'entschuldigung', 'tschüss', 'guten', 'tag', + 'morgen', 'abend', 'kann', 'könnte', 'würde', 'möchte', 'wollen', 'brauchen', 'verstehe', + 'weiß', 'denke', 'meine', 'helfen', 'hören', 'zuhören', 'richtig', 'falsch', 'gut', 'schlecht', + 'alles', 'klar', 'natürlich', 'vielleicht'], + + 'fr': ['je', 'tu', 'il', 'elle', 'nous', 'vous', 'ils', 'elles', 'ce', 'ça', 'quoi', 'comment', 'pourquoi', + 'où', 'quand', 'oui', 'non', 'salut', 'bonjour', 'merci', 'pardon', 'excusez', 'au revoir', + 'peux', 'pouvez', 'voulez', 'veux', 'besoin', 'comprends', 'sais', 'pense', 'dire', 'aider', + 'entendre', 'écouter', 'correct', 'incorrect', 'bon', 'bien', 'mauvais', 'mal', + 'd\'accord', 'bien sûr', 'peut-être'], + + 'es': ['yo', 'tú', 'él', 'ella', 'nosotros', 'vosotros', 'ellos', 'ellas', 'esto', 'eso', 'qué', 'cómo', + 'por qué', 'dónde', 'cuándo', 'sí', 'no', 'hola', 'gracias', 'perdón', 'disculpe', 'adiós', + 'puede', 'puedes', 'quiere', 'quiero', 'necesito', 'entiendo', 'sé', 'pienso', 'digo', 'ayudar', + 'oír', 'escuchar', 'correcto', 'incorrecto', 'bueno', 'malo', 'vale', 'claro', 'quizás', 'tal vez'] + } + + LANGUAGE_PUNCTUATION = { + 'zh': ['。', '，', '、', '：', '；', '！', '？', '"', '"', '（', '）', '《', '》', '【', '】', '——', '……', '～'], + 'en': [], # 英语标点大多通用，不作为特征 + 'de': ['„', '"'], # 德语特有引号 + 'fr': ['«', '»', '…'], # 法语特有引号和省略号 + 'es': ['¿', '¡'] # 西班牙语特有的倒置问号和感叹号 + } + + SPEECH_FILLERS = { + 'zh': ['嗯', '呃', '啊', '哦', '唉', '那个', '这个', '就是', '其实', '然后', '所以', '但是'], + 'en': ['um', 'uh', 'er', 'ah', 'like', 'you know', 'i mean', 'well', 'so', 'anyway', 'actually'], + 'de': ['äh', 'ähm', 'hmm', 'na ja', 'also', 'sozusagen', 'quasi', 'naja', 'eigentlich', 'tja'], + 'fr': ['euh', 'ben', 'bah', 'eh bien', 'bon', 'alors', 'donc', 'voilà', 'en fait', 'quoi'], + 'es': ['eh', 'este', 'em', 'pues', 'bueno', 'o sea', 'vale', 'vamos', 'mira', 'entonces'] + } + + SPEECH_PATTERNS = { + 'zh': [r'^(你好|您好|喂|在吗|请问|那个|这个|帮我|麻烦|请|要|需要)'], + 'en': [r'^(hello|hi|hey|excuse me|sorry|please|can you|could you|would you|I want|I need)'], + 'de': [r'^(hallo|guten tag|entschuldigung|bitte|können sie|könntest du|ich möchte|ich brauche)'], + 'fr': [r'^(bonjour|salut|excusez-moi|pardon|s\'il vous plaît|pouvez-vous|pourriez-vous|je voudrais|j\'ai besoin)'], + 'es': [r'^(hola|buenos días|perdón|disculpe|por favor|puede usted|puedes|quiero|necesito)'] + } + + _executor = concurrent.futures.ThreadPoolExecutor( + max_workers=4, + thread_name_prefix="lang_detector" + ) + + @staticmethod + async def detect_language(text): + try: + SUPPORTED_LANGUAGES = ["zh", "en", "fr", "de", "es"] + if not text: + return "en" # 默认返回英文 + cleaned_text = ' '.join(text.lower().split()) + + loop = asyncio.get_running_loop() + speech_pattern_lang = await loop.run_in_executor( + LanguageDetector._executor, + LanguageDetector._detect_by_speech_patterns, + cleaned_text + ) + + if speech_pattern_lang != "unknown": + return speech_pattern_lang + + char_based_lang = await loop.run_in_executor( + LanguageDetector._executor, + LanguageDetector._detect_by_characters, + cleaned_text + ) + + if char_based_lang != "unknown": + return char_based_lang + + feature_based_lang = await loop.run_in_executor( + LanguageDetector._executor, + LanguageDetector._detect_by_features, + cleaned_text + ) + + if feature_based_lang != "unknown": + return feature_based_lang + + try: + is_reliable, _, details = cld2.detect(cleaned_text) + if is_reliable and details[0][1] in SUPPORTED_LANGUAGES: + return details[0][1] + except: + pass + + for ch in cleaned_text: + if '\u4e00' <= ch <= '\u9fff': # 检测中文字符 + return "zh" + + return "en" # 默认返回英语 + except Exception as e: + session_logger.error("system", "language_detector", f"语言检测错误: {str(e)}") + return "en" # 出错时默认使用英文 + + @staticmethod + def _detect_by_speech_patterns(text: str) -> str: + """检测常见的语音对话开头模式""" + for lang, patterns in LanguageDetector.SPEECH_PATTERNS.items(): + for pattern in patterns: + if re.search(pattern, text): + return lang + return "unknown" + + @staticmethod + def _detect_by_characters(text: str) -> str: + """基于字符分布的语言检测，特别适合短文本""" + text = text.strip() + text = text.lower() + text_len = max(1, len(text)) + char_counts = {lang: 0 for lang in LanguageDetector.LANGUAGE_UNICODE_RANGES} + + for char in text: + code_point = ord(char) + for lang, ranges in LanguageDetector.LANGUAGE_UNICODE_RANGES.items(): + for start, end in ranges: + if start <= code_point <= end: + char_counts[lang] += 1 + + for lang, count in char_counts.items(): + ratio = count / text_len + if lang == 'zh' and ratio > 0.12: + return 'zh' + + latin_chars = sum(1 for c in text if unicodedata.category(c).startswith('L')) + latin_ratio = latin_chars / text_len + + if latin_ratio > 0.4: # 降低拉丁字符检测阈值，更灵敏地捕捉语音转文本 + if any(c in text for c in "äöüß"): + return "de" # 德语特有字符 + elif any(c in text for c in "éèêëàâçùûüÿôœæ"): + return "fr" # 法语特有字符 + elif any(c in text for c in "áéíóúñ¿¡"): + return "es" # 西班牙语特有字符 + + if "¿" in text or "¡" in text: + return "es" # 西班牙语特有标点 + elif "«" in text or "»" in text: + return "fr" # 法语引号 + + return "en" # 默认英语 + + return "unknown" + + @staticmethod + def _detect_by_features(text: str) -> str: + """基于特征词和语气词的语言检测，针对口语场景优化""" + text_lower = text.lower() + matches = {lang: 0 for lang in LanguageDetector.LANGUAGE_FEATURES} + + for lang, features in LanguageDetector.LANGUAGE_FEATURES.items(): + for word in features: + if lang == 'zh': # 中文不需要空格分隔 + if word in text_lower: + matches[lang] += 2 + elif f" {word} " in f" {text_lower} ": + matches[lang] += 1 + + for lang, fillers in LanguageDetector.SPEECH_FILLERS.items(): + for filler in fillers: + if lang == 'zh': + count = text_lower.count(filler) + else: + count = text_lower.count(f" {filler} ") + + matches[lang] += count * 3 + + for lang, puncts in LanguageDetector.LANGUAGE_PUNCTUATION.items(): + for punct in puncts: + matches[lang] += text.count(punct) * 5 + + nonzero_matches = {lang: score for lang, score in matches.items() if score > 0} + if nonzero_matches: + best_lang = max(nonzero_matches.items(), key=lambda x: x[1]) + if best_lang[1] > 0: + return best_lang[0] + + return "unknown" diff --git a/talkingq-url/utils/logger.py b/talkingq-url/utils/logger.py new file mode 100644 index 0000000..c5ae167 --- /dev/null +++ b/talkingq-url/utils/logger.py @@ -0,0 +1,92 @@ +import os +from typing import Optional +from loguru import logger + + +class SessionLogger: + def __init__(self): + self.is_main_process = os.environ.get("UVICORN_WID", "0") == "0" + + # 创建日志目录 + os.makedirs("logs", exist_ok=True) + + # 配置loguru + logger.remove() # 移除默认handler + + # 控制台输出 + logger.add( + lambda msg: print(msg, end=""), + format="{time:YYYY-MM-DD HH:mm:ss} - {level} - [设备: {extra[device_id]}] [会话: {extra[session_id]}] {message}", + level="INFO" + ) + + # 文件输出 - 持久化保存 + logger.add( + "logs/talkingq_{time:YYYY-MM-DD}.log", + format="{time:YYYY-MM-DD HH:mm:ss} - {level} - [设备: {extra[device_id]}] [会话: {extra[session_id]}] {message}", + rotation="00:00", # 每天轮转 + retention="30 days", # 保留30天 + compression="zip", # 压缩旧日志 + level="INFO" + ) + + self.logger = logger + + def log( + self, + level: str, + device_id: Optional[str], + session_id: Optional[str], + message: str, + **kwargs + ): + if isinstance(device_id, str) and device_id.startswith("{"): + try: + import json + device_info = json.loads(device_id) + device_id = device_info.get("device_id", device_id) + except: + pass # 如果解析失败，保持原样 + + # 绑定额外信息到logger + bound_logger = self.logger.bind( + device_id=device_id or "unknown", + session_id=session_id or "unknown" + ) + + # 根据level调用对应方法 + if level.upper() == "INFO": + bound_logger.info(message) + elif level.upper() == "ERROR": + bound_logger.error(message) + elif level.upper() == "WARNING": + bound_logger.warning(message) + else: + bound_logger.log(level, message) + + def system_log(self, level: str, session_id: str, message: str): + if self.is_main_process: + self.log(level, "system", session_id, message) + + def info(self, device_id: Optional[str], session_id: Optional[str], message: str): + self.log("INFO", device_id, session_id, message) + + def system_info(self, session_id: str, message: str): + self.system_log("INFO", session_id, message) + + def error( + self, + device_id: Optional[str], + session_id: Optional[str], + message: str, + **kwargs + ): + self.log("ERROR", device_id, session_id, message, **kwargs) + + def warning( + self, device_id: Optional[str], session_id: Optional[str], message: str + ): + self.log("WARNING", device_id, session_id, message) + + +session_logger = SessionLogger() diff --git a/talkingq-url/utils/text_splitter.py b/talkingq-url/utils/text_splitter.py new file mode 100644 index 0000000..1e58e37 --- /dev/null +++ b/talkingq-url/utils/text_splitter.py @@ -0,0 +1,145 @@ +import re + +SENTENCE_ENDINGS = set("。！？!?；;：:.。\n") +PUNCTUATION_MARKS = set('。！？!?.,，；;：:~…—－_\'"″""》《）》《<>()[]{}【】\'、~～¡¿äöüßáéíóúüñç\n') +APOSTROPHE_SUFFIXES = {'s', 't', 've', 'll', 're', 'd', 'm'} +LANGUAGE_CONTRACTIONS = { + 'fr': ["l'", "d'", "s'", "qu'", "c'", "j'", "n'", "m'", "t'", "jusqu'"], + 'de': ["geht's", "gibt's"], + 'es': ["el-", "del", "al"], + 'ms': ["-lah", "-kah", "-nya"] +} + +def is_numbered_list_item(text): + pattern = r"^\s*\d+\.\s" + return bool(re.search(pattern, text)) + +def split_into_sentences(text): + return _custom_split_sentences(text) + +def _custom_split_sentences(text): + text = re.sub(r"([。！？!?;；])", r"\1\n", text) + + sentences = [] + current_sentence = "" + i = 0 + + while i < len(text): + current_sentence += text[i] + + if text[i] in SENTENCE_ENDINGS and not is_special_context(text, i): + if current_sentence.strip(): + sentences.append(current_sentence) + current_sentence = "" + + i += 1 + + if current_sentence.strip(): + sentences.append(current_sentence) + + merged_sentences = [] + temp = "" + + for s in sentences: + s_stripped = s.strip() + + if len(s_stripped) < 5 and not any(p in s for p in "。！？!?"): + temp += s + else: + if temp: + merged_sentences.append(temp + s) + temp = "" + else: + merged_sentences.append(s) + + if temp: + merged_sentences.append(temp) + + final_sentences = [] + for idx, sentence in enumerate(merged_sentences): + sentence_start = sentence.strip() + if idx > 0: + prev_sentence_end = merged_sentences[idx-1].strip() + + if (sentence_start.startswith("s ") and prev_sentence_end.endswith("'")) or \ + any(sentence_start.startswith(suffix + " ") for suffix in APOSTROPHE_SUFFIXES) and prev_sentence_end.endswith("'"): + final_sentences[-1] += sentence + continue + + if any(sentence_start.startswith(contraction) for contraction in LANGUAGE_CONTRACTIONS['fr']): + final_sentences[-1] += sentence + continue + + if (any(prev_sentence_end.endswith(contraction) for contraction in LANGUAGE_CONTRACTIONS['de']) or + any(sentence_start.startswith(contraction) for contraction in LANGUAGE_CONTRACTIONS['es']) or + any(prev_sentence_end.endswith(contraction) for contraction in LANGUAGE_CONTRACTIONS['ms'])): + final_sentences[-1] += sentence + continue + + if sentence_start.startswith("¿") or sentence_start.startswith("¡"): + final_sentences.append(sentence) + continue + + final_sentences.append(sentence) + + return final_sentences + +def is_special_context(text, pos): + """判断是否是特殊上下文，不应该在此处断句""" + if text[pos] == "." and pos > 0 and pos < len(text) - 1: + if text[pos - 1].isdigit() and text[pos + 1].isdigit(): + return True + + if text[pos] == "'" and pos > 0 and pos < len(text) - 1: + if text[pos - 1].isalpha(): # 前面是字母 + for suffix in APOSTROPHE_SUFFIXES: + if pos + len(suffix) <= len(text) and text[pos+1:pos+1+len(suffix)] == suffix: + if pos + len(suffix) + 1 >= len(text) or not text[pos+1+len(suffix)].isalpha(): + return True + + if text[pos] == "." and pos > 0 and pos < len(text) - 2: + if text[pos - 1].isupper() and text[pos + 1].isupper(): + return True + + if text[pos] == "." and pos > 0 and pos < len(text) - 1: + if not text[pos - 1].isspace() and not text[pos + 1].isspace(): + left_context = text[max(0, pos-15):pos] + if "www." in left_context or "http" in left_context or "@" in left_context: + return True + + if text[pos] == "." and pos > 0 and pos < len(text) - 1: + if pos + 2 < len(text) and text[pos:pos+3] == "...": + return True + + if text[pos] == "'" and pos > 0 and pos < len(text) - 1: + if pos > 1 and text[pos-2:pos+1] in ["l'", "d'", "c'", "j'", "n'", "m'", "t'"]: + return True + + if text[pos] == "." and pos > 0: + if pos >= 3 and text[pos-3:pos] in ["bzw", "usw", "etc"]: + return True + + if text[pos] in "¿¡" and pos < len(text) - 1: + return True + + if text[pos] == "-" and pos > 0 and pos < len(text) - 2: + if text[pos+1:pos+4] in ["lah", "kah", "nya"]: + return True + + return False + +def should_skip_tts(text_chunk): + """判断一段文本是否应该跳过TTS处理""" + if not text_chunk.strip(): + return True + + if is_numbered_list_item(text_chunk) and len(text_chunk.strip()) <= 5: + return True + + if all(c in PUNCTUATION_MARKS for c in text_chunk.strip()): + return True + + if len(text_chunk.strip()) <= 2 and not any(c.isalnum() for c in text_chunk): + return True + + return False