feat: requirement更新

.gitignore

@@ -114,3 +114,5 @@ node_modules/
 *.njsproj
 *.sln
 *.sw?
+
+*.mp3

backend/app/api/v1/endpoints/speaker.py

@@ -0,0 +1,10 @@
+from fastapi import APIRouter
+from app.constants.tts import SPEAKER_DATA
+from app.schemas import SpeakerResponse
+
+router = APIRouter()
+
+
+@router.get("/list", response_model=SpeakerResponse)
+async def get_model_vendors():
+    return SpeakerResponse(data=SPEAKER_DATA)

backend/app/api/v1/endpoints/tts.py

@@ -0,0 +1,496 @@
+import uuid
+import websockets
+import time
+import fastrand
+import json
+import asyncio
+import os
+import aiofiles
+from datetime import datetime
+from typing import Dict, Any, Optional as OptionalType
+
+from app.constants.tts import APP_ID, TOKEN, SPEAKER
+
+# 协议常量
+PROTOCOL_VERSION = 0b0001
+DEFAULT_HEADER_SIZE = 0b0001
+FULL_CLIENT_REQUEST = 0b0001
+AUDIO_ONLY_RESPONSE = 0b1011
+FULL_SERVER_RESPONSE = 0b1001
+ERROR_INFORMATION = 0b1111
+MsgTypeFlagWithEvent = 0b100
+NO_SERIALIZATION = 0b0000
+JSON = 0b0001
+COMPRESSION_NO = 0b0000
+
+# 事件类型
+EVENT_NONE = 0
+EVENT_Start_Connection = 1
+EVENT_FinishConnection = 2
+EVENT_ConnectionStarted = 50
+EVENT_StartSession = 100
+EVENT_FinishSession = 102
+EVENT_SessionStarted = 150
+EVENT_SessionFinished = 152
+EVENT_TaskRequest = 200
+EVENT_TTSSentenceEnd = 351
+EVENT_TTSResponse = 352
+
+# 音频文件保存目录
+TEMP_AUDIO_DIR = "./temp_audio"
+
+
+# 确保音频目录存在
+async def ensure_audio_dir():
+    """异步创建音频目录"""
+    if not os.path.exists(TEMP_AUDIO_DIR):
+        os.makedirs(TEMP_AUDIO_DIR, exist_ok=True)
+
+
+# 生成时间戳文件名
+def generate_audio_filename() -> str:
+    """生成基于时间戳的音频文件名"""
+    timestamp = datetime.now().strftime("%Y%m%d_%H%M%S_%f")[:-3]  # 精确到毫秒
+    return f"{timestamp}.mp3"
+
+
+# ... 保留所有原有的类定义和工具函数 ...
+
+class Header:
+    def __init__(self,
+                 protocol_version=PROTOCOL_VERSION,
+                 header_size=DEFAULT_HEADER_SIZE,
+                 message_type=0,
+                 message_type_specific_flags=0,
+                 serial_method=NO_SERIALIZATION,
+                 compression_type=COMPRESSION_NO,
+                 reserved_data=0):
+        self.header_size = header_size
+        self.protocol_version = protocol_version
+        self.message_type = message_type
+        self.message_type_specific_flags = message_type_specific_flags
+        self.serial_method = serial_method
+        self.compression_type = compression_type
+        self.reserved_data = reserved_data
+
+    def as_bytes(self) -> bytes:
+        return bytes([
+            (self.protocol_version << 4) | self.header_size,
+            (self.message_type << 4) | self.message_type_specific_flags,
+            (self.serial_method << 4) | self.compression_type,
+            self.reserved_data
+        ])
+
+
+class Optional:
+    def __init__(self, event=EVENT_NONE, sessionId=None, sequence=None):
+        self.event = event
+        self.sessionId = sessionId
+        self.errorCode = 0
+        self.connectionId = None
+        self.response_meta_json = None
+        self.sequence = sequence
+
+    def as_bytes(self) -> bytes:
+        option_bytes = bytearray()
+        if self.event != EVENT_NONE:
+            option_bytes.extend(self.event.to_bytes(4, "big", signed=True))
+        if self.sessionId is not None:
+            session_id_bytes = str.encode(self.sessionId)
+            size = len(session_id_bytes).to_bytes(4, "big", signed=True)
+            option_bytes.extend(size)
+            option_bytes.extend(session_id_bytes)
+        if self.sequence is not None:
+            option_bytes.extend(self.sequence.to_bytes(4, "big", signed=True))
+        return option_bytes
+
+
+class Response:
+    def __init__(self, header, optional):
+        self.optional = optional
+        self.header = header
+        self.payload = None
+        self.payload_json = None
+
+
+# 工具函数
+def gen_log_id():
+    """生成logID"""
+    ts = int(time.time() * 1000)
+    r = fastrand.pcg32bounded(1 << 24) + (1 << 20)
+    local_ip = "00000000000000000000000000000000"
+    return f"02{ts}{local_ip}{r:08x}"
+
+
+def get_payload_bytes(uid='1234', event=EVENT_NONE, text='', speaker='', audio_format='mp3',
+                      audio_sample_rate=24000):
+    return str.encode(json.dumps({
+        "user": {"uid": uid},
+        "event": event,
+        "namespace": "BidirectionalTTS",
+        "req_params": {
+            "text": text,
+            "speaker": speaker,
+            "audio_params": {
+                "format": audio_format,
+                "sample_rate": audio_sample_rate,
+                "enable_timestamp": True,
+            }
+        }
+    }))
+
+
+def read_res_content(res, offset):
+    content_size = int.from_bytes(res[offset: offset + 4], "big", signed=True)
+    offset += 4
+    content = str(res[offset: offset + content_size], encoding='utf8')
+    offset += content_size
+    return content, offset
+
+
+def read_res_payload(res, offset):
+    payload_size = int.from_bytes(res[offset: offset + 4], "big", signed=True)
+    offset += 4
+    payload = res[offset: offset + payload_size]
+    offset += payload_size
+    return payload, offset
+
+
+def parser_response(res) -> Response:
+    if isinstance(res, str):
+        raise RuntimeError(res)
+    response = Response(Header(), Optional())
+    # 解析结果
+    header = response.header
+    num = 0b00001111
+    header.protocol_version = res[0] >> 4 & num
+    header.header_size = res[0] & 0x0f
+    header.message_type = (res[1] >> 4) & num
+    header.message_type_specific_flags = res[1] & 0x0f
+    header.serial_method = res[2] >> num
+    header.message_compression = res[2] & 0x0f
+    header.reserved_data = res[3]
+
+    offset = 4
+    optional = response.optional
+    if header.message_type == FULL_SERVER_RESPONSE or AUDIO_ONLY_RESPONSE:
+        # read event
+        if header.message_type_specific_flags == MsgTypeFlagWithEvent:
+            optional.event = int.from_bytes(res[offset:offset + 4], "big", signed=True)
+            offset += 4
+            if optional.event == EVENT_NONE:
+                return response
+            # read connectionId
+            elif optional.event == EVENT_ConnectionStarted:
+                optional.connectionId, offset = read_res_content(res, offset)
+            elif optional.event == EVENT_SessionStarted or optional.event == EVENT_SessionFinished:
+                optional.sessionId, offset = read_res_content(res, offset)
+                optional.response_meta_json, offset = read_res_content(res, offset)
+            elif optional.event == EVENT_TTSResponse:
+                optional.sessionId, offset = read_res_content(res, offset)
+                response.payload, offset = read_res_payload(res, offset)
+            elif optional.event == EVENT_TTSSentenceEnd:
+                optional.sessionId, offset = read_res_content(res, offset)
+                response.payload_json, offset = read_res_content(res, offset)
+
+    elif header.message_type == ERROR_INFORMATION:
+        optional.errorCode = int.from_bytes(res[offset:offset + 4], "big", signed=True)
+        offset += 4
+        response.payload, offset = read_res_payload(res, offset)
+    return response
+
+
+async def send_event(ws, header, optional=None, payload=None):
+    full_client_request = bytearray(header)
+    if optional is not None:
+        full_client_request.extend(optional)
+    if payload is not None:
+        payload_size = len(payload).to_bytes(4, 'big', signed=True)
+        full_client_request.extend(payload_size)
+        full_client_request.extend(payload)
+    await ws.send(full_client_request)
+
+
+# TTS状态管理类，添加消息ID和任务追踪
+class TTSState:
+    def __init__(self, message_id: str):
+        self.message_id = message_id
+        self.volc_ws: OptionalType[websockets.WebSocketServerProtocol] = None
+        self.session_id: OptionalType[str] = None
+        self.task: OptionalType[asyncio.Task] = None  # 用于追踪异步任务
+        self.is_processing = False
+        self.audio_data = bytearray()  # 用于收集音频数据
+        self.audio_filename = None  # 保存的文件名
+
+
+# 全局状态管理
+class TTSManager:
+    def __init__(self):
+        # WebSocket -> 消息ID -> TTS状态
+        self.connections: Dict[any, Dict[str, TTSState]] = {}
+        # 会话ID -> 消息ID 的映射，用于路由响应
+        self.session_to_message: Dict[str, str] = {}
+        # 消息ID -> WebSocket 的映射
+        self.message_to_websocket: Dict[str, any] = {}
+
+    def get_connection_states(self, websocket) -> Dict[str, TTSState]:
+        """获取WebSocket连接的所有TTS状态"""
+        if websocket not in self.connections:
+            self.connections[websocket] = {}
+        return self.connections[websocket]
+
+    def add_tts_state(self, websocket, message_id: str) -> TTSState:
+        """添加新的TTS状态"""
+        states = self.get_connection_states(websocket)
+        if message_id in states:
+            # 如果已存在，先清理旧的
+            self.cleanup_message_state(websocket, message_id)
+
+        tts_state = TTSState(message_id)
+        states[message_id] = tts_state
+        self.message_to_websocket[message_id] = websocket
+        return tts_state
+
+    def get_tts_state(self, websocket, message_id: str) -> OptionalType[TTSState]:
+        """获取指定的TTS状态"""
+        states = self.get_connection_states(websocket)
+        return states.get(message_id)
+
+    def register_session(self, session_id: str, message_id: str):
+        """注册会话ID和消息ID的映射"""
+        self.session_to_message[session_id] = message_id
+
+    def get_message_by_session(self, session_id: str) -> OptionalType[str]:
+        """根据会话ID获取消息ID"""
+        return self.session_to_message.get(session_id)
+
+    def get_websocket_by_message(self, message_id: str):
+        """根据消息ID获取WebSocket"""
+        return self.message_to_websocket.get(message_id)
+
+    def cleanup_message_state(self, websocket, message_id: str):
+        """清理指定消息的状态"""
+        states = self.get_connection_states(websocket)
+        if message_id in states:
+            tts_state = states[message_id]
+            # 取消任务
+            if tts_state.task and not tts_state.task.done():
+                tts_state.task.cancel()
+            # 清理映射
+            if tts_state.session_id and tts_state.session_id in self.session_to_message:
+                del self.session_to_message[tts_state.session_id]
+            if message_id in self.message_to_websocket:
+                del self.message_to_websocket[message_id]
+            # 删除状态
+            del states[message_id]
+
+    def cleanup_connection(self, websocket):
+        """清理整个连接的状态"""
+        if websocket in self.connections:
+            states = self.connections[websocket]
+            for message_id in list(states.keys()):
+                self.cleanup_message_state(websocket, message_id)
+            del self.connections[websocket]
+
+
+# 全局TTS管理器实例
+tts_manager = TTSManager()
+
+
+# 初始化独立的TTS连接
+async def create_tts_connection() -> websockets.WebSocketServerProtocol:
+    """创建独立的TTS连接"""
+    log_id = gen_log_id()
+    ws_header = {
+        "X-Api-App-Key": APP_ID,
+        "X-Api-Access-Key": TOKEN,
+        "X-Api-Resource-Id": 'volc.service_type.10029',
+        "X-Api-Connect-Id": str(uuid.uuid4()),
+        "X-Tt-Logid": log_id,
+    }
+    url = 'wss://openspeech.bytedance.com/api/v3/tts/bidirection'
+    volc_ws = await websockets.connect(url, additional_headers=ws_header, max_size=1000000000)
+
+    # 启动连接
+    header = Header(message_type=FULL_CLIENT_REQUEST,
+                    message_type_specific_flags=MsgTypeFlagWithEvent).as_bytes()
+    optional = Optional(event=EVENT_Start_Connection).as_bytes()
+    payload = str.encode("{}")
+    await send_event(volc_ws, header, optional, payload)
+
+    # 等待连接确认
+    raw_data = await volc_ws.recv()
+    res = parser_response(raw_data)
+    if res.optional.event != EVENT_ConnectionStarted:
+        raise Exception("TTS连接失败")
+
+    return volc_ws
+
+
+# 异步保存音频文件
+async def save_audio_file(audio_data: bytes, filename: str) -> str:
+    """异步保存音频文件"""
+    await ensure_audio_dir()
+    file_path = os.path.join(TEMP_AUDIO_DIR, filename)
+
+    async with aiofiles.open(file_path, 'wb') as f:
+        await f.write(audio_data)
+
+    return file_path
+
+
+# 处理单个TTS任务
+async def process_tts_task(websocket, message_id: str, text: str, speaker: str = None):
+    """处理单个TTS任务（独立协程）"""
+    tts_state = None
+    # 使用传入的speaker，如果没有则使用默认的
+    selected_speaker = speaker if speaker else SPEAKER
+
+    try:
+        print(f"开始处理TTS任务 [{message_id}]: {text}, 使用说话人: {selected_speaker}")
+        # 获取TTS状态
+        tts_state = tts_manager.get_tts_state(websocket, message_id)
+        if not tts_state:
+            raise Exception(f"找不到TTS状态: {message_id}")
+        tts_state.is_processing = True
+        # 生成音频文件名
+        tts_state.audio_filename = generate_audio_filename()
+        # 创建独立的TTS连接
+        tts_state.volc_ws = await create_tts_connection()
+        # 创建会话
+        tts_state.session_id = uuid.uuid4().__str__().replace('-', '')
+        tts_manager.register_session(tts_state.session_id, message_id)
+        print(f"创建TTS会话 [{message_id}]: {tts_state.session_id}")
+        header = Header(message_type=FULL_CLIENT_REQUEST,
+                        message_type_specific_flags=MsgTypeFlagWithEvent,
+                        serial_method=JSON).as_bytes()
+        optional = Optional(event=EVENT_StartSession, sessionId=tts_state.session_id).as_bytes()
+        # 使用选择的speaker
+        payload = get_payload_bytes(event=EVENT_StartSession, speaker=selected_speaker)
+        await send_event(tts_state.volc_ws, header, optional, payload)
+        raw_data = await tts_state.volc_ws.recv()
+        res = parser_response(raw_data)
+        if res.optional.event != EVENT_SessionStarted:
+            raise Exception("TTS会话启动失败")
+        print(f"TTS会话创建成功 [{message_id}]: {tts_state.session_id}")
+        # 发送文本到TTS服务
+        print(f"发送文本到TTS服务 [{message_id}]...")
+        header = Header(message_type=FULL_CLIENT_REQUEST,
+                        message_type_specific_flags=MsgTypeFlagWithEvent,
+                        serial_method=JSON).as_bytes()
+        optional = Optional(event=EVENT_TaskRequest, sessionId=tts_state.session_id).as_bytes()
+        # 使用选择的speaker
+        payload = get_payload_bytes(event=EVENT_TaskRequest, text=text, speaker=selected_speaker)
+        await send_event(tts_state.volc_ws, header, optional, payload)
+        # 接收TTS响应并发送到前端
+        print(f"开始接收TTS响应 [{message_id}]...")
+        audio_count = 0
+        try:
+            while True:
+                raw_data = await asyncio.wait_for(
+                    tts_state.volc_ws.recv(),
+                    timeout=30
+                )
+                res = parser_response(raw_data)
+                print(f"收到TTS事件 [{message_id}]: {res.optional.event}")
+                if res.optional.event == EVENT_TTSSentenceEnd:
+                    print(f"句子结束事件 [{message_id}] - 直接完成")
+                    break
+                elif res.optional.event == EVENT_SessionFinished:
+                    print(f"收到会话结束事件 [{message_id}]")
+                    break
+                elif res.optional.event == EVENT_TTSResponse:
+                    audio_count += 1
+                    print(f"收到音频数据 [{message_id}] #{audio_count}，大小: {len(res.payload)}")
+
+                    # 收集音频数据
+                    tts_state.audio_data.extend(res.payload)
+
+                    # 发送音频数据到前端
+                    await websocket.send_json({
+                        "id": audio_count,
+                        "type": "tts_audio_data",
+                        "messageId": message_id,
+                        "audioData": res.payload.hex()  # 转为hex字符串
+                    })
+                else:
+                    print(f"未知TTS事件 [{message_id}]: {res.optional.event}")
+        except asyncio.TimeoutError:
+            print(f"TTS响应超时 [{message_id}]，强制结束")
+        # 异步保存音频文件
+        if tts_state.audio_data:
+            file_path = await save_audio_file(
+                bytes(tts_state.audio_data),
+                tts_state.audio_filename
+            )
+            print(f"音频文件已保存 [{message_id}]: {file_path}")
+
+        # 发送完成消息，包含文件路径和使用的speaker
+        await websocket.send_json({
+            "type": "tts_audio_complete",
+            "messageId": message_id,
+            "audioFile": tts_state.audio_filename,
+            "audioPath": os.path.join(TEMP_AUDIO_DIR, tts_state.audio_filename) if tts_state.audio_data else None,
+            "speaker": selected_speaker
+        })
+        print(f"TTS处理完成 [{message_id}]，共发送 {audio_count} 个音频包，使用说话人: {selected_speaker}")
+    except asyncio.CancelledError:
+        print(f"TTS任务被取消 [{message_id}]")
+        await websocket.send_json({
+            "type": "tts_error",
+            "messageId": message_id,
+            "message": "TTS任务被取消"
+        })
+    except Exception as e:
+        print(f"TTS处理异常 [{message_id}]: {e}")
+        import traceback
+        traceback.print_exc()
+        await websocket.send_json({
+            "type": "tts_error",
+            "messageId": message_id,
+            "message": f"TTS处理失败: {str(e)}"
+        })
+    finally:
+        # 清理资源
+        if tts_state:
+            tts_state.is_processing = False
+            if tts_state.volc_ws:
+                try:
+                    await tts_state.volc_ws.close()
+                except:
+                    pass
+        # 清理状态
+        tts_manager.cleanup_message_state(websocket, message_id)
+
+
+# 启动TTS文本转换
+async def handle_tts_text(websocket, message_id: str, text: str, speaker: str = None):
+    """启动TTS文本转换"""
+    # 创建新的TTS状态
+    print(speaker)
+    tts_state = tts_manager.add_tts_state(websocket, message_id)
+    # 启动异步任务，传入speaker参数
+    tts_state.task = asyncio.create_task(
+        process_tts_task(websocket, message_id, text, speaker)
+    )
+
+
+# 取消TTS任务
+async def handle_tts_cancel(websocket, message_id: str):
+    """取消TTS任务"""
+    tts_state = tts_manager.get_tts_state(websocket, message_id)
+    if tts_state and tts_state.task and not tts_state.task.done():
+        tts_state.task.cancel()
+        await websocket.send_json({
+            "type": "tts_complete",
+            "messageId": message_id
+        })
+    tts_manager.cleanup_message_state(websocket, message_id)
+
+
+# 清理连接的所有TTS资源
+async def cleanup_connection_tts(websocket):
+    """清理连接的所有TTS资源"""
+    print(f"清理连接的TTS资源...")
+    tts_manager.cleanup_connection(websocket)
+    print("TTS资源清理完成")

backend/app/api/v1/endpoints/voice_conversation.py

@@ -0,0 +1,105 @@
+import json
+import aiohttp
+import asyncio
+from fastapi.encoders import jsonable_encoder
+from starlette.websockets import WebSocket
+
+from . import tts
+from app.constants.model_data import tip_message, base_url, headers
+
+
+async def process_voice_conversation(websocket: WebSocket, asr_text: str, message_id: str, speaker: str):
+    try:
+        print(f"开始处理语音对话 [{message_id}]: {asr_text}")
+
+        # 1. 发送ASR识别结果到前端
+        await websocket.send_json({
+            "type": "asr_result",
+            "messageId": message_id,
+            "result": asr_text
+        })
+
+        # 2. 构建LLM请求
+        messages = [
+            tip_message,
+            {"role": "user", "content": asr_text}
+        ]
+        payload = {
+            "model": "gpt-4o",
+            "messages": messages,
+            "stream": True
+        }
+
+        print(f"发送LLM请求 [{message_id}]: {json.dumps(payload, ensure_ascii=False)}")
+
+        # 3. 流式处理LLM响应
+        full_response = ""
+        llm_completed = False
+
+        async with aiohttp.ClientSession() as session:
+            async with session.post(
+                    base_url,
+                    headers=headers,
+                    json=jsonable_encoder(payload)
+            ) as resp:
+                if resp.status != 200:
+                    error_text = await resp.text()
+                    raise Exception(f"LLM API请求失败: {resp.status} - {error_text}")
+
+                # 读取流式响应
+                async for line in resp.content:
+                    if line:
+                        line = line.decode('utf-8').strip()
+                        if line.startswith('data: '):
+                            data = line[6:].strip()
+                            if data == '[DONE]':
+                                llm_completed = True
+                                print(f"LLM响应完成 [{message_id}]")
+                                break
+
+                            try:
+                                result = json.loads(data)
+                                # 提取内容
+                                choices = result.get("choices", [])
+                                if not choices:
+                                    # 跳过空choices数据包
+                                    continue
+
+                                delta = choices[0].get("delta", {})
+                                content = delta.get("content")
+
+                                if content:
+                                    full_response += content
+
+                            except json.JSONDecodeError as e:
+                                print(f"JSON解析错误 [{message_id}]: {e}, 数据: {data}")
+                                continue
+                            except Exception as e:
+                                print(f"处理数据包异常 [{message_id}]: {e}, 数据: {data}")
+                                continue
+
+        # 4. LLM生成完成后，启动完整的TTS处理
+        if llm_completed and full_response:
+            print(f"LLM生成完成 [{message_id}], 总内容长度: {len(full_response)}")
+            print(f"完整内容: {full_response}")
+
+            # 发送完成消息
+            await websocket.send_json({
+                "type": "llm_complete_response",
+                "messageId": message_id,
+                "content": full_response
+            })
+
+            # 启动TTS处理完整内容
+            print(f"启动完整TTS处理 [{message_id}]: {full_response}")
+            await tts.handle_tts_text(websocket, message_id, full_response, speaker)
+
+    except Exception as e:
+        print(f"语音对话处理异常 [{message_id}]: {e}")
+        import traceback
+        traceback.print_exc()
+        await websocket.send_json({
+            "type": "voice_conversation_error",
+            "messageId": message_id,
+            "message": f"处理失败: {str(e)}"
+        })

backend/app/api/v1/endpoints/websocket_service.py

@@ -1,14 +1,21 @@
+# websocket_service.py
+import uuid
+
 from fastapi import APIRouter, WebSocket, WebSocketDisconnect
 from typing import Set
 from aip import AipSpeech
 from app.constants.asr import APP_ID, API_KEY, SECRET_KEY
 import json
 
+from . import tts
+from .voice_conversation import process_voice_conversation
+
 router = APIRouter()
 active_connections: Set[WebSocket] = set()
 
 asr_client = AipSpeech(APP_ID, API_KEY, SECRET_KEY)
 
+
 async def asr_buffer(buffer_data: bytes) -> str:
     result = asr_client.asr(buffer_data, 'pcm', 16000, {'dev_pid': 1537})
     if result.get('err_msg') == 'success.':
@@ -16,6 +23,7 @@ async def asr_buffer(buffer_data: bytes) -> str:
     else:
         return '语音转换失败'
 
+
 async def broadcast_online_count():
     data = {"online_count": len(active_connections), 'type': 'count'}
     to_remove = set()
@@ -27,12 +35,14 @@ async def broadcast_online_count():
     for ws in to_remove:
         active_connections.remove(ws)
 
+
 @router.websocket("/websocket")
 async def websocket_online_count(websocket: WebSocket):
     await websocket.accept()
     active_connections.add(websocket)
     await broadcast_online_count()
     temp_buffer = bytes()
+
     try:
         while True:
             message = await websocket.receive()
@@ -45,15 +55,66 @@ async def websocket_online_count(websocket: WebSocket):
                     except Exception:
                         continue
                     msg_type = data.get("type")
+
                     if msg_type == "ping":
                         await websocket.send_json({"online_count": len(active_connections), "type": "count"})
+
                     elif msg_type == "asr_end":
                         asr_text = await asr_buffer(temp_buffer)
-                        await websocket.send_json({"type": "asr_result", "result": asr_text})
+                        # 从data中获取messageId，如果不存在则生成一个新的ID
+                        message_id = data.get("messageId", "voice_" + str(uuid.uuid4()))
+                        if data.get("voiceConversation"):
+                            speaker = data.get("speaker")
+                            await process_voice_conversation(websocket, asr_text, message_id, speaker)
+                        else:
+                            await websocket.send_json({"type": "asr_result", "result": asr_text})
                         temp_buffer = bytes()
+
+                    # TTS处理
+                    elif msg_type == "tts_text":
+                        message_id = data.get("messageId")
+                        text = data.get("text", "")
+                        speaker = data.get("speaker")
+
+                        if not message_id:
+                            await websocket.send_json({
+                                "type": "tts_error",
+                                "message": "缺少messageId参数"
+                            })
+                            continue
+
+                        print(f"收到TTS文本请求 [{message_id}]: {text}")
+                        try:
+                            await tts.handle_tts_text(websocket, message_id, text, speaker)
+                        except Exception as e:
+                            print(f"TTS文本处理异常 [{message_id}]: {e}")
+                            await websocket.send_json({
+                                "type": "tts_error",
+                                "messageId": message_id,
+                                "message": f"TTS处理失败: {str(e)}"
+                            })
+
+                    elif msg_type == "tts_cancel":
+                        message_id = data.get("messageId")
+                        if message_id:
+                            print(f"收到TTS取消请求 [{message_id}]")
+                            try:
+                                await tts.handle_tts_cancel(websocket, message_id)
+                            except Exception as e:
+                                print(f"TTS取消处理异常 [{message_id}]: {e}")
+
     except WebSocketDisconnect:
-        active_connections.remove(websocket)
-        await broadcast_online_count()
-    except Exception:
-        active_connections.remove(websocket)
+        pass
+    except Exception as e:
+        print(f"WebSocket异常: {e}")
+    finally:
+        # 清理资源
+        active_connections.discard(websocket)
+
+        # 清理所有TTS资源
+        try:
+            await tts.cleanup_connection_tts(websocket)
+        except:
+            pass
+
         await broadcast_online_count()

backend/app/constants/model_data.py

@@ -19,7 +19,8 @@ MODEL_DATA = [
     {
         "vendor": "Anthropic",
         "models": [
-            {"model_id": "claude-sonnet-4-thinking", "model_name": "Claude Sonnet 4 thinking", "model_type": "reasoning"},
+            {"model_id": "claude-sonnet-4-thinking", "model_name": "Claude Sonnet 4 thinking",
+             "model_type": "reasoning"},
             {"model_id": "claude-sonnet-4", "model_name": "Claude Sonnet 4", "model_type": "text"},
         ]
     },
@@ -27,6 +28,7 @@ MODEL_DATA = [
         "vendor": "硅基流动",
         "models": [
             {"model_id": "deepseek-v3", "model_name": "DeepSeek V3", "model_type": "text"},
+            {"model_id": "deepseek-r1", "model_name": "DeepSeek R1", "model_type": "reasoning"},
         ]
     }
 ]

backend/app/constants/tts.py

@@ -0,0 +1,104 @@
+# APP_ID = '1142362958'
+# TOKEN = 'O-a4JkyLFrYkME9no11DxFkOY-UnAoFF'
+# SPEAKER = 'zh_male_beijingxiaoye_moon_bigtts'
+
+APP_ID = '2138450044'
+TOKEN = 'V04_QumeQZhJrQ_In1Z0VBQm7n0ttMNO'
+SPEAKER = 'zh_male_beijingxiaoye_moon_bigtts'
+
+SPEAKER_DATA = [
+    {
+        "category": "趣味口音",
+        "speakers": [
+            {
+                "speaker_id": "zh_male_jingqiangkanye_moon_bigtts",
+                "speaker_name": "京腔侃爷/Harmony",
+                "language": "中文-北京口音、英文",
+                "platforms": ["豆包", "Cici", "web demo"]
+            },
+            {
+                "speaker_id": "zh_female_wanwanxiaohe_moon_bigtts",
+                "speaker_name": "湾湾小何",
+                "language": "中文-台湾口音",
+                "platforms": ["豆包", "Cici"]
+            },
+            {
+                "speaker_id": "zh_female_wanqudashu_moon_bigtts",
+                "speaker_name": "湾区大叔",
+                "language": "中文-广东口音",
+                "platforms": ["豆包", "Cici"]
+            },
+            {
+                "speaker_id": "zh_female_daimengchuanmei_moon_bigtts",
+                "speaker_name": "呆萌川妹",
+                "language": "中文-四川口音",
+                "platforms": ["豆包", "Cici"]
+            },
+            {
+                "speaker_id": "zh_male_guozhoudege_moon_bigtts",
+                "speaker_name": "广州德哥",
+                "language": "中文-广东口音",
+                "platforms": ["豆包", "Cici"]
+            },
+            {
+                "speaker_id": "zh_male_beijingxiaoye_moon_bigtts",
+                "speaker_name": "北京小爷",
+                "language": "中文-北京口音",
+                "platforms": ["豆包"]
+            },
+            {
+                "speaker_id": "zh_male_haoyuxiaoge_moon_bigtts",
+                "speaker_name": "浩宇小哥",
+                "language": "中文-青岛口音",
+                "platforms": ["豆包"]
+            },
+            {
+                "speaker_id": "zh_male_guangxiyuanzhou_moon_bigtts",
+                "speaker_name": "广西远舟",
+                "language": "中文-广西口音",
+                "platforms": ["豆包"]
+            },
+            {
+                "speaker_id": "zh_female_meituojieer_moon_bigtts",
+                "speaker_name": "妹坨洁儿",
+                "language": "中文-长沙口音",
+                "platforms": ["豆包", "剪映"]
+            },
+            {
+                "speaker_id": "zh_male_yuzhouzixuan_moon_bigtts",
+                "speaker_name": "豫州子轩",
+                "language": "中文-河南口音",
+                "platforms": ["豆包"]
+            }
+        ]
+    },
+    {
+        "category": "角色扮演",
+        "speakers": [
+            {
+                "speaker_id": "zh_male_naiqimengwa_mars_bigtts",
+                "speaker_name": "奶气萌娃",
+                "language": "中文",
+                "platforms": ["剪映", "豆包"]
+            },
+            {
+                "speaker_id": "zh_female_popo_mars_bigtts",
+                "speaker_name": "婆婆",
+                "language": "中文",
+                "platforms": ["剪映C端", "抖音", "豆包"]
+            },
+            {
+                "speaker_id": "zh_female_gaolengyujie_moon_bigtts",
+                "speaker_name": "高冷御姐",
+                "language": "中文",
+                "platforms": ["豆包", "Cici"]
+            },
+            {
+                "speaker_id": "zh_male_aojiaobazong_moon_bigtts",
+                "speaker_name": "傲娇霸总",
+                "language": "中文",
+                "platforms": ["豆包"]
+            }
+        ]
+    }
+]

backend/app/main.py

@@ -1,5 +1,5 @@
 from fastapi import FastAPI
-from app.api.v1.endpoints import chat, model, websocket_service
+from app.api.v1.endpoints import chat, model, websocket_service,speaker
 
 app = FastAPI()
 
@@ -9,6 +9,7 @@ app.include_router(websocket_service.router, prefix="", tags=["websocket_service
 app.include_router(chat.router, prefix="/v1/chat", tags=["chat"])
 # 获取模型列表服务
 app.include_router(model.router, prefix="/v1/model", tags=["model_list"])
+app.include_router(speaker.router, prefix="/v1/speaker", tags=["speaker_list"])
 
 if __name__ == "__main__":
     import uvicorn

backend/app/schemas/__init__.py

@@ -5,4 +5,5 @@ from .chat import (
     ModelInfo,
     VendorModelList,
     VendorModelResponse,
+    SpeakerResponse
 )

backend/app/schemas/chat.py

@@ -33,3 +33,20 @@ class VendorModelList(BaseModel):
 
 class VendorModelResponse(BaseModel):
     data: List[VendorModelList]
+
+
+# Speaker相关模型
+class Speaker(BaseModel):
+    speaker_id: str
+    speaker_name: str
+    language: str
+    platforms: List[str]
+
+
+class CategorySpeakers(BaseModel):
+    category: str
+    speakers: List[Speaker]
+
+
+class SpeakerResponse(BaseModel):
+    data: List[CategorySpeakers]

web/.prettierignore

@@ -0,0 +1,3 @@
+# .prettierignore
+auto-imports.d.ts
+components.d.ts

web/.prettierrc.json

@@ -0,0 +1,8 @@
+{
+  "$schema": "https://json.schemastore.org/prettierrc",
+  "tabWidth": 2,
+  "singleQuote": false,
+  "printWidth": 80,
+  "trailingComma": "none",
+  "ignorePath": ".prettierignore"
+}

web/components.d.ts

@@ -7,16 +7,26 @@ export {}
 /* prettier-ignore */
 declare module 'vue' {
   export interface GlobalComponents {
+    Avatar: typeof import('./src/components/avatar.vue')['default']
     Markdown: typeof import('./src/components/markdown.vue')['default']
+    Message_tools: typeof import('./src/components/MessageTools.vue')['default']
+    MessageTools: typeof import('./src/components/MessageTools.vue')['default']
     NButton: typeof import('naive-ui')['NButton']
+    NCollapse: typeof import('naive-ui')['NCollapse']
+    NCollapseItem: typeof import('naive-ui')['NCollapseItem']
     NConfigProvider: typeof import('naive-ui')['NConfigProvider']
+    NDivider: typeof import('naive-ui')['NDivider']
+    NImage: typeof import('naive-ui')['NImage']
     NInput: typeof import('naive-ui')['NInput']
     NMessageProvider: typeof import('naive-ui')['NMessageProvider']
     NPopconfirm: typeof import('naive-ui')['NPopconfirm']
+    NPopover: typeof import('naive-ui')['NPopover']
     NScrollbar: typeof import('naive-ui')['NScrollbar']
     NSelect: typeof import('naive-ui')['NSelect']
+    NSpin: typeof import('naive-ui')['NSpin']
     NTag: typeof import('naive-ui')['NTag']
     RouterLink: typeof import('vue-router')['RouterLink']
     RouterView: typeof import('vue-router')['RouterView']
+    Tts: typeof import('./src/components/tts.vue')['default']
   }
 }

web/eslint.config.js

@@ -1,39 +1,38 @@
-import antfu from "@antfu/eslint-config"
+import antfu from "@antfu/eslint-config";
 
-export default antfu(
-  {
-    formatters: {
-      /**
-       * Format CSS, LESS, SCSS files, also the `<style>` blocks in Vue
-       * By default uses Prettier
-       */
-      css: true,
-      /**
-       * Format HTML files
-       * By default uses Prettier
-       */
-      html: true,
-      /**
-       * Format Markdown files
-       * Supports Prettier and dprint
-       * By default uses Prettier
-       */
-      markdown: "prettier",
-    },
-    stylistic: {
-      indent: 2,
-      quotes: "double",
-      semi: true,
-    },
-    vue: true,
-    ignores: ["node_modules", "dist"],
-    rules: {
-      "vue/html-self-closing": "off",
-      "antfu/top-level-function": "off",
-      "ts/no-unsafe-function-type": "off",
-      "no-console": "off",
-      "unused-imports/no-unused-vars": "warn",
-    },
+export default antfu({
+  formatters: {
+    /**
+     * Format CSS, LESS, SCSS files, also the `<style>` blocks in Vue
+     * By default uses Prettier
+     */
+    css: true,
+    /**
+     * Format HTML files
+     * By default uses Prettier
+     */
+    html: true,
+    /**
+     * Format Markdown files
+     * Supports Prettier and dprint
+     * By default uses Prettier
+     */
+    markdown: "prettier"
   },
-
-)
+  stylistic: {
+    indent: 2,
+    quotes: "double",
+    semi: true
+  },
+  vue: true,
+  ignores: ["node_modules", "dist"],
+  rules: {
+    "vue/html-self-closing": "off",
+    "antfu/top-level-function": "off",
+    "ts/no-unsafe-function-type": "off",
+    "no-console": "off",
+    "unused-imports/no-unused-vars": "warn",
+    "ts/no-use-before-define": "off",
+    "vue/operator-linebreak": "off",
+  }
+});

web/index.html

@@ -1,13 +1,16 @@
 <!doctype html>
-<html lang="en">
-  <head>
-    <meta charset="UTF-8" />
-    <link rel="icon" type="image/svg+xml" href="/vite.svg" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" />
-    <title>chat</title>
-  </head>
-  <body>
-    <div id="app"></div>
-    <script type="module" src="/src/main.ts"></script>
-  </body>
-</html>
+<html lang="zh-CN">
+
+<head>
+  <meta charset="UTF-8" />
+  <link rel="icon" type="image/svg+xml" href="/vite.svg" />
+  <meta name="viewport" content="width=device-width, initial-scale=1.0" />
+  <title>chat</title>
+</head>
+
+<body>
+  <div id="app"></div>
+  <script type="module" src="/src/main.ts"></script>
+</body>
+
+</html>

web/package.json

@@ -15,6 +15,7 @@
     "@vueuse/core": "^13.4.0",
     "axios": "^1.10.0",
     "highlight.js": "^11.11.1",
+    "lodash-es": "^4.17.21",
     "markdown-it": "^14.1.0",
     "pinia": "^3.0.3",
     "pinia-plugin-persistedstate": "^4.3.0",
@@ -27,13 +28,17 @@
   },
   "devDependencies": {
     "@antfu/eslint-config": "^4.16.1",
+    "@types/lodash-es": "^4.17.12",
     "@types/markdown-it": "^14.1.2",
     "@types/node": "^24.0.4",
     "@vitejs/plugin-vue": "^6.0.0",
     "@vue/tsconfig": "^0.7.0",
     "eslint": "^9.29.0",
+    "eslint-config-prettier": "^10.1.5",
     "eslint-plugin-format": "^1.0.1",
+    "eslint-plugin-prettier": "^5.5.1",
     "naive-ui": "^2.42.0",
+    "prettier": "^3.6.2",
     "typescript": "~5.8.3",
     "unplugin-vue-components": "^0.28.0",
     "vite": "^7.0.0",

web/pnpm-lock.yaml

@@ -20,6 +20,9 @@ importers:
       highlight.js:
         specifier: ^11.11.1
         version: 11.11.1
+      lodash-es:
+        specifier: ^4.17.21
+        version: 4.17.21
       markdown-it:
         specifier: ^14.1.0
         version: 14.1.0
@@ -51,6 +54,9 @@ importers:
       '@antfu/eslint-config':
         specifier: ^4.16.1
         version: 4.16.1(@vue/compiler-sfc@3.5.17)(eslint-plugin-format@1.0.1(eslint@9.29.0(jiti@2.4.2)))(eslint@9.29.0(jiti@2.4.2))(typescript@5.8.3)
+      '@types/lodash-es':
+        specifier: ^4.17.12
+        version: 4.17.12
       '@types/markdown-it':
         specifier: ^14.1.2
         version: 14.1.2
@@ -66,12 +72,21 @@ importers:
       eslint:
         specifier: ^9.29.0
         version: 9.29.0(jiti@2.4.2)
+      eslint-config-prettier:
+        specifier: ^10.1.5
+        version: 10.1.5(eslint@9.29.0(jiti@2.4.2))
       eslint-plugin-format:
         specifier: ^1.0.1
         version: 1.0.1(eslint@9.29.0(jiti@2.4.2))
+      eslint-plugin-prettier:
+        specifier: ^5.5.1
+        version: 5.5.1(eslint-config-prettier@10.1.5(eslint@9.29.0(jiti@2.4.2)))(eslint@9.29.0(jiti@2.4.2))(prettier@3.6.2)
       naive-ui:
         specifier: ^2.42.0
         version: 2.42.0(vue@3.5.17(typescript@5.8.3))
+      prettier:
+        specifier: ^3.6.2
+        version: 3.6.2
       typescript:
         specifier: ~5.8.3
         version: 5.8.3
@@ -1447,6 +1462,12 @@ packages:
     peerDependencies:
       eslint: ^9.5.0
 
+  eslint-config-prettier@10.1.5:
+    resolution: {integrity: sha512-zc1UmCpNltmVY34vuLRV61r1K27sWuX39E+uyUnY8xS2Bex88VV9cugG+UZbRSRGtGyFboj+D8JODyme1plMpw==}
+    hasBin: true
+    peerDependencies:
+      eslint: '>=7.0.0'
+
   eslint-flat-config-utils@2.1.0:
     resolution: {integrity: sha512-6fjOJ9tS0k28ketkUcQ+kKptB4dBZY2VijMZ9rGn8Cwnn1SH0cZBoPXT8AHBFHxmHcLFQK9zbELDinZ2Mr1rng==}
 
@@ -1538,6 +1559,20 @@ packages:
     peerDependencies:
       eslint: ^9.0.0
 
+  eslint-plugin-prettier@5.5.1:
+    resolution: {integrity: sha512-dobTkHT6XaEVOo8IO90Q4DOSxnm3Y151QxPJlM/vKC0bVy+d6cVWQZLlFiuZPP0wS6vZwSKeJgKkcS+KfMBlRw==}
+    engines: {node: ^14.18.0 || >=16.0.0}
+    peerDependencies:
+      '@types/eslint': '>=8.0.0'
+      eslint: '>=8.0.0'
+      eslint-config-prettier: '>= 7.0.0 <10.0.0 || >=10.1.0'
+      prettier: '>=3.0.0'
+    peerDependenciesMeta:
+      '@types/eslint':
+        optional: true
+      eslint-config-prettier:
+        optional: true
+
   eslint-plugin-regexp@2.9.0:
     resolution: {integrity: sha512-9WqJMnOq8VlE/cK+YAo9C9YHhkOtcEtEk9d12a+H7OSZFwlpI6stiHmYPGa2VE0QhTzodJyhlyprUaXDZLgHBw==}
     engines: {node: ^18 || >=20}
@@ -2427,8 +2462,8 @@ packages:
     resolution: {integrity: sha512-GbK2cP9nraSSUF9N2XwUwqfzlAFlMNYYl+ShE/V+H8a9uNl/oUqB1w2EL54Jh0OlyRSd8RfWYJ3coVS4TROP2w==}
     engines: {node: '>=6.0.0'}
 
-  prettier@3.6.1:
-    resolution: {integrity: sha512-5xGWRa90Sp2+x1dQtNpIpeOQpTDBs9cZDmA/qs2vDNN2i18PdapqY7CmBeyLlMuGqXJRIOPaCaVZTLNQRWUH/A==}
+  prettier@3.6.2:
+    resolution: {integrity: sha512-I7AIg5boAr5R0FFtJ6rCfD+LFsWHp81dolrFD8S79U9tb8Az2nGrJncnMSnys+bpQJfRUzqs9hnA81OAA3hCuQ==}
     engines: {node: '>=14'}
     hasBin: true
 
@@ -4276,6 +4311,10 @@ snapshots:
       '@eslint/compat': 1.3.1(eslint@9.29.0(jiti@2.4.2))
       eslint: 9.29.0(jiti@2.4.2)
 
+  eslint-config-prettier@10.1.5(eslint@9.29.0(jiti@2.4.2)):
+    dependencies:
+      eslint: 9.29.0(jiti@2.4.2)
+
   eslint-flat-config-utils@2.1.0:
     dependencies:
       pathe: 2.0.3
@@ -4321,7 +4360,7 @@ snapshots:
       eslint: 9.29.0(jiti@2.4.2)
       eslint-formatting-reporter: 0.0.0(eslint@9.29.0(jiti@2.4.2))
       eslint-parser-plain: 0.1.1
-      prettier: 3.6.1
+      prettier: 3.6.2
       synckit: 0.9.3
 
   eslint-plugin-import-lite@0.3.0(eslint@9.29.0(jiti@2.4.2))(typescript@5.8.3):
@@ -4401,6 +4440,15 @@ snapshots:
       tinyglobby: 0.2.14
       yaml-eslint-parser: 1.3.0
 
+  eslint-plugin-prettier@5.5.1(eslint-config-prettier@10.1.5(eslint@9.29.0(jiti@2.4.2)))(eslint@9.29.0(jiti@2.4.2))(prettier@3.6.2):
+    dependencies:
+      eslint: 9.29.0(jiti@2.4.2)
+      prettier: 3.6.2
+      prettier-linter-helpers: 1.0.0
+      synckit: 0.11.8
+    optionalDependencies:
+      eslint-config-prettier: 10.1.5(eslint@9.29.0(jiti@2.4.2))
+
   eslint-plugin-regexp@2.9.0(eslint@9.29.0(jiti@2.4.2)):
     dependencies:
       '@eslint-community/eslint-utils': 4.7.0(eslint@9.29.0(jiti@2.4.2))
@@ -5459,7 +5507,7 @@ snapshots:
     dependencies:
       fast-diff: 1.3.0
 
-  prettier@3.6.1: {}
+  prettier@3.6.2: {}
 
   pretty-ms@9.2.0:
     dependencies:

web/src/App.vue

@@ -1,13 +1,13 @@
 <script setup lang="ts">
-import type { GlobalThemeOverrides } from "naive-ui"
-import { zhCN } from "naive-ui"
-import { useWebSocketStore } from "@/services"
+import type { GlobalThemeOverrides } from "naive-ui";
+import { zhCN } from "naive-ui";
+import { useWebSocketStore } from "@/services";
 
-const websocketStore = useWebSocketStore()
+const websocketStore = useWebSocketStore();
 
 onMounted(() => {
-  websocketStore.connect()
-})
+  websocketStore.connect();
+});
 
 const themeOverrides: GlobalThemeOverrides = {
   common: {
@@ -17,12 +17,12 @@ const themeOverrides: GlobalThemeOverrides = {
     primaryColorSuppl: "#00bfff",
     fontWeightStrong: "600",
     borderRadius: "8px",
-    borderRadiusSmall: "5px",
+    borderRadiusSmall: "5px"
   },
   Button: {
-    textColor: "#0094c5",
-  },
-}
+    textColor: "#0094c5"
+  }
+};
 </script>
 
 <template>

web/src/assets/Icons/index.ts

@@ -1,4 +1,7 @@
-export { default as ExclamationTriangleIcon } from "./svg/heroicons/ExclamationTriangleIcon.svg?component"
-export { default as microphone } from "./svg/heroicons/MicrophoneIcon.svg?component"
-export { default as PaperAirplaneIcon } from "./svg/heroicons/PaperAirplaneIcon.svg?component"
-export { default as TrashIcon } from "./svg/heroicons/TrashIcon.svg?component"
+export { default as ChevronLeftIcon } from "./svg/heroicons/ChevronLeftIcon.svg?component";
+export { default as DocumentDuplicateIcon } from "./svg/heroicons/DocumentDuplicateIcon.svg?component";
+export { default as ExclamationTriangleIcon } from "./svg/heroicons/ExclamationTriangleIcon.svg?component";
+export { default as microphone } from "./svg/heroicons/MicrophoneIcon.svg?component";
+export { default as PaperAirplaneIcon } from "./svg/heroicons/PaperAirplaneIcon.svg?component";
+export { default as SpeakerWaveIcon } from "./svg/heroicons/SpeakerWaveIcon.svg?component";
+export { default as TrashIcon } from "./svg/heroicons/TrashIcon.svg?component";

web/src/assets/Icons/svg/heroicons/ChevronLeftIcon.svg

@@ -0,0 +1,3 @@
+<svg xmlns="http://www.w3.org/2000/svg" fill="none" viewBox="0 0 24 24" stroke-width="1.5" stroke="currentColor" class="size-6">
+  <path stroke-linecap="round" stroke-linejoin="round" d="M15.75 19.5 8.25 12l7.5-7.5" />
+</svg>

web/src/assets/Icons/svg/heroicons/DocumentDuplicateIcon.svg

@@ -0,0 +1,3 @@
+<svg xmlns="http://www.w3.org/2000/svg" fill="none" viewBox="0 0 24 24" stroke-width="1.5" stroke="currentColor" class="size-6">
+  <path stroke-linecap="round" stroke-linejoin="round" d="M15.75 17.25v3.375c0 .621-.504 1.125-1.125 1.125h-9.75a1.125 1.125 0 0 1-1.125-1.125V7.875c0-.621.504-1.125 1.125-1.125H6.75a9.06 9.06 0 0 1 1.5.124m7.5 10.376h3.375c.621 0 1.125-.504 1.125-1.125V11.25c0-4.46-3.243-8.161-7.5-8.876a9.06 9.06 0 0 0-1.5-.124H9.375c-.621 0-1.125.504-1.125 1.125v3.5m7.5 10.375H9.375a1.125 1.125 0 0 1-1.125-1.125v-9.25m12 6.625v-1.875a3.375 3.375 0 0 0-3.375-3.375h-1.5a1.125 1.125 0 0 1-1.125-1.125v-1.5a3.375 3.375 0 0 0-3.375-3.375H9.75" />
+</svg>

web/src/assets/Icons/svg/heroicons/SpeakerWaveIcon.svg

@@ -0,0 +1,3 @@
+<svg xmlns="http://www.w3.org/2000/svg" fill="none" viewBox="0 0 24 24" stroke-width="1.5" stroke="currentColor" class="size-6">
+  <path stroke-linecap="round" stroke-linejoin="round" d="M19.114 5.636a9 9 0 0 1 0 12.728M16.463 8.288a5.25 5.25 0 0 1 0 7.424M6.75 8.25l4.72-4.72a.75.75 0 0 1 1.28.53v15.88a.75.75 0 0 1-1.28.53l-4.72-4.72H4.51c-.88 0-1.704-.507-1.938-1.354A9.009 9.009 0 0 1 2.25 12c0-.83.112-1.633.322-2.396C2.806 8.756 3.63 8.25 4.51 8.25H6.75Z" />
+</svg>

web/src/components/MessageTools.vue

@@ -0,0 +1,26 @@
+<script setup lang="ts">
+import type { Message } from "@/interfaces";
+
+import { DocumentDuplicateIcon } from "@/assets/Icons";
+import { copy } from "@/utils";
+
+const { msg } = defineProps<{
+  msg: Message;
+}>();
+</script>
+
+<template>
+  <div class="flex items-center gap-2 justify-end mt-2">
+    <div v-if="msg.role !== 'user'">
+      <tts :text="msg.content || ''" :message-id="msg.id!" />
+    </div>
+    <NPopover trigger="hover">
+      <template #trigger>
+        <NButton quaternary circle @click="copy(msg.content || '')">
+          <DocumentDuplicateIcon class="!w-4 !h-4" />
+        </NButton>
+      </template>
+      <span>复制内容</span>
+    </NPopover>
+  </div>
+</template>

web/src/components/avatar.vue

@@ -0,0 +1,19 @@
+<script setup lang="ts">
+const { avatar } = defineProps<{
+  avatar: string;
+}>();
+</script>
+
+<template>
+  <NImage
+    :src="avatar"
+    width="64"
+    height="64"
+    :preview-disabled="true"
+    object-fit="cover"
+    class="!block !w-16 !min-w-16 !h-16"
+    :img-props="{
+      class: 'rounded-lg !block !w-16 !min-w-16 !h-16'
+    }"
+  />
+</template>

web/src/components/markdown.vue

@@ -1,10 +1,13 @@
 <script setup lang="ts">
-import hljs from "highlight.js"
-import markdownit from "markdown-it"
+import hljs from "highlight.js";
+import markdownit from "markdown-it";
+import { useWindowWidth } from "@/utils";
 
 const { content } = defineProps<{
-  content: string
-}>()
+  content: string;
+}>();
+
+const windowWidth = useWindowWidth();
 
 const md = markdownit({
   html: true,
@@ -13,24 +16,34 @@ const md = markdownit({
   highlight(str, lang) {
     if (lang && hljs.getLanguage(lang)) {
       try {
-        return hljs.highlight(str, { language: lang }).value
+        return hljs.highlight(str, { language: lang }).value;
+        // eslint-disable-next-line unused-imports/no-unused-vars
+      } catch (__) {
       }
-      // eslint-disable-next-line unused-imports/no-unused-vars
-      catch (__) { }
     }
 
-    return "" // use external default escaping
-  },
-})
+    return ""; // use external default escaping
+  }
+});
+
+// // 计算代码块宽度
+const codeWidth = computed(() => {
+  return windowWidth.value - 160 - 200; // 减去左右边距和其他元素的宽度
+});
 </script>
 
 <template>
-  <div class="markdown-body w-full text-base break-words whitespace-normal" v-html="md.render(content)">
-  </div>
+  <div
+    class="markdown-body w-full text-base break-words whitespace-normal"
+    :style="{ '--code-width': `${codeWidth}px` }"
+    v-html="md.render(content)"
+  ></div>
 </template>
 
 <style scoped>
 .markdown-body :deep(pre) {
+  width: var(--code-width, 100%);
+  max-width: 100%;
   background: #fafafacc;
   border-radius: 6px;
   padding: 16px;
@@ -38,4 +51,20 @@ const md = markdownit({
   overflow-x: auto;
   margin: 8px 0;
 }
+
+.markdown-body :deep(pre)::-webkit-scrollbar {
+  height: 8px;
+  background: #f0f0f0;
+  border-radius: 4px;
+}
+
+.markdown-body :deep(pre)::-webkit-scrollbar-thumb {
+  background: #bfbfbf;
+  border-radius: 4px;
+  transition: all 0.3s ease-in-out;
+}
+
+.markdown-body :deep(pre)::-webkit-scrollbar-thumb:hover {
+  background: #999;
+}
 </style>

web/src/components/tts.vue

@@ -0,0 +1,83 @@
+<script setup lang="ts">
+import { SpeakerWaveIcon } from "@/assets/Icons";
+import { useLayoutStore, useTtsStore } from "@/stores";
+
+const { text, messageId } = defineProps<{
+  text: string;
+  messageId: string;
+}>();
+
+const ttsStore = useTtsStore();
+const layoutStore = useLayoutStore();
+const { simpleMode } = storeToRefs(layoutStore);
+
+// 获取当前消息的状态
+const isPlaying = computed(() => ttsStore.isPlaying(messageId));
+const isLoading = computed(() => ttsStore.isLoading(messageId));
+const hasAudio = computed(() => ttsStore.hasAudio(messageId));
+
+// 处理按钮点击
+const handleClick = () => {
+  if (isLoading.value) {
+    return; // 合成中不响应点击
+  }
+
+  if (hasAudio.value) {
+    // 如果音频已准备好，切换播放/暂停
+    if (isPlaying.value) {
+      ttsStore.pause(messageId);
+    } else {
+      ttsStore.play(messageId);
+    }
+  } else {
+    // 如果没有音频，开始TTS转换
+    ttsStore.convertText(text, messageId);
+  }
+};
+
+// 文本改变清理之前的音频
+watch(
+  () => text,
+  () => {
+    ttsStore.clearAudio(messageId);
+  }
+);
+
+onUnmounted(() => {
+  ttsStore.clearAudio(messageId);
+});
+</script>
+
+<template>
+  <NPopover trigger="hover">
+    <template #trigger>
+      <NButton
+        :loading="isLoading"
+        @click="handleClick"
+        quaternary
+        circle
+        :disabled="!text.trim()"
+      >
+        <SpeakerWaveIcon
+          v-if="!isLoading"
+          class="!w-4 !h-4"
+          :class="{
+            '': !simpleMode,
+            'animate-pulse': isPlaying
+          }"
+        />
+      </NButton>
+    </template>
+    <span>
+      {{
+        isLoading
+          ? "合成中..."
+          : isPlaying
+            ? "点击暂停"
+            : hasAudio
+              ? "点击播放"
+              : "语音合成"
+      }}
+    </span>
+  </NPopover>
+</template>

web/src/interfaces/chat_service.ts

@@ -1,24 +1,128 @@
 export interface IChatWithLLMRequest {
-  messages: Message[]
+  messages: Message[];
   /**
    * 要使用的模型的 ID
    */
-  model: string
+  model: string;
 }
 
 export interface Message {
-  content?: string
-  role?: string
-  [property: string]: any
+  content?: string;
+  thinking?: string;
+  role?: string;
+  usage?: UsageInfo;
+  id?: string;
+  type?: "chat" | "voice";
+  [property: string]: any;
 }
 
 export interface ModelInfo {
-  model_id: string
-  model_name: string
-  model_type: string
+  model_id: string;
+  model_name: string;
+  model_type: string;
 }
 
 export interface ModelListInfo {
-  vendor: string
-  models: ModelInfo[]
+  vendor: string;
+  models: ModelInfo[];
+}
+
+export interface UsageInfo {
+  prompt_tokens: number;
+  completion_tokens: number;
+  total_tokens: number;
+}
+
+/**
+ * Speaker 语音合成器基本信息
+ */
+export interface Speaker {
+  /** speaker唯一标识ID */
+  speaker_id: string;
+  /** speaker显示名称 */
+  speaker_name: string;
+  /** 支持的语言/口音 */
+  language: string;
+  /** 支持的平台列表 */
+  platforms: string[];
+}
+
+/**
+ * Speaker分类信息
+ */
+export interface CategorySpeakers {
+  /** 分类名称 */
+  category: string;
+  /** 该分类下的speaker列表 */
+  speakers: Speaker[];
+}
+
+/**
+ * Speaker分类枚举
+ */
+export enum SpeakerCategory {
+  /** 趣味口音 */
+  ACCENT = "趣味口音",
+  /** 角色扮演 */
+  ROLE_PLAY = "角色扮演"
+}
+
+/**
+ * 常用平台枚举
+ */
+export enum SpeakerPlatform {
+  DOUYIN = "抖音",
+  DOUBAO = "豆包",
+  CICI = "Cici",
+  JIANYING = "剪映",
+  JIANYING_C = "剪映C端",
+  WEB_DEMO = "web demo",
+  STORY_AI = "StoryAi",
+  MAOXIANG = "猫箱"
+}
+
+/**
+ * Speaker选择器组件Props
+ */
+export interface SpeakerSelectorProps {
+  /** 当前选中的speaker */
+  selectedSpeaker?: Speaker;
+  /** speaker选择回调 */
+  onSpeakerChange: (speaker: Speaker) => void;
+  /** 是否禁用 */
+  disabled?: boolean;
+  /** 过滤特定分类 */
+  filterCategories?: SpeakerCategory[];
+  /** 过滤特定平台 */
+  filterPlatforms?: SpeakerPlatform[];
+}
+
+/**
+ * 语音合成参数
+ */
+export interface VoiceSynthesisParams {
+  /** 使用的speaker */
+  speaker: Speaker;
+  /** 要合成的文本 */
+  text: string;
+  /** 语速 (0.5-2.0) */
+  speed?: number;
+  /** 音调 (0.5-2.0) */
+  pitch?: number;
+  /** 音量 (0.0-1.0) */
+  volume?: number;
+}
+
+/**
+ * 语音合成响应
+ */
+export interface VoiceSynthesisResponse {
+  /** 音频文件URL */
+  audio_url: string;
+  /** 音频时长(秒) */
+  duration: number;
+  /** 合成状态 */
+  status: "success" | "error";
+  /** 错误信息 */
+  error_message?: string;
 }

web/src/interfaces/index.ts

@@ -1,9 +1,9 @@
 export interface ICommonResponse<T> {
-  code: number
-  msg: string
-  data: T
+  code: number;
+  msg: string;
+  data: T;
 }
 
-export type IMsgOnlyResponse = ICommonResponse<{ msg: string }>
+export type IMsgOnlyResponse = ICommonResponse<{ msg: string }>;
 
-export * from "./chat_service"
+export * from "./chat_service";

web/src/layouts/BasicLayout.vue

@@ -1,27 +1,64 @@
 <script setup lang="ts">
-import { NImage } from "naive-ui";
+import { ChevronLeftIcon } from "@/assets/Icons";
 import logo from "@/assets/logo.png";
-import { useChatStore } from "@/stores";
+import { useChatStore, useLayoutStore } from "@/stores";
 
 const chatStore = useChatStore();
 const { onlineCount } = storeToRefs(chatStore);
+const layoutStore = useLayoutStore();
+const { hiddenLeftSidebar, simpleMode } = storeToRefs(layoutStore);
 </script>
 
 <template>
-  <div class="h-screen flex overflow-hidden">
-    <div class="flex-none w-[200px] h-full flex flex-col">
+  <div class="relative h-screen flex overflow-hidden">
+    <div
+      class="absolute left-0 top-0 bottom-0 z-10 flex-none w-[200px] h-full flex flex-col bg-white transition-all ease-in-out"
+      :class="{
+        '-translate-x-[200px]': hiddenLeftSidebar
+      }"
+    >
+      <div
+        @click="hiddenLeftSidebar = !hiddenLeftSidebar"
+        class="absolute -right-3 translate-y-1/2 top-1/2 z-20 w-[24px] h-[24px] bg-[#0094c526] rounded-full flex items-center justify-center cursor-pointer"
+        :class="{
+          'rotate-180 -right-8': hiddenLeftSidebar
+        }"
+      >
+        <ChevronLeftIcon class="!w-4 !h-4 text-[#777]" />
+      </div>
       <router-link class="w-full my-6 cursor-pointer" to="/">
-        <NImage class="w-full object-cover" :src="logo" alt="logo" :preview-disabled="true" />
+        <NImage
+          class="w-full object-cover"
+          :src="logo"
+          alt="logo"
+          :preview-disabled="true"
+        />
       </router-link>
       <router-link
-        class="w-full h-[52px] px-8 flex items-center cursor-pointer" :class="$route.path === '/'
-          ? [
-            'bg-[rgba(37,99,235,0.04)] text-[#0094c5] border-r-2 border-[#0094c5]',
-          ]
-          : []
-        " to="/"
+        class="w-full h-[52px] px-8 flex items-center cursor-pointer"
+        :class="
+          $route.path === '/'
+            ? [
+                'bg-[rgba(37,99,235,0.04)] text-[#0094c5] border-r-2 border-[#0094c5]'
+              ]
+            : []
+        "
+        to="/"
       >
-        聊天
+        对话
+      </router-link>
+      <router-link
+        class="w-full h-[52px] px-8 flex items-center cursor-pointer"
+        :class="
+          $route.path === '/voice'
+            ? [
+                'bg-[rgba(37,99,235,0.04)] text-[#0094c5] border-r-2 border-[#0094c5]'
+              ]
+            : []
+        "
+        to="/voice"
+      >
+        语音聊天
       </router-link>
 
       <div class="w-full h-full flex flex-col items-center text-[#0094c5]">
@@ -32,7 +69,10 @@ const { onlineCount } = storeToRefs(chatStore);
         </div>
       </div>
     </div>
-    <div class="flex-1 relative">
+    <div
+      class="flex-1 relative"
+      :class="{ 'ml-[200px]': !hiddenLeftSidebar && !simpleMode }"
+    >
       <RouterView />
     </div>
   </div>

web/src/main.ts

@@ -1,14 +1,14 @@
-import { createPinia } from 'pinia'
-import { createApp } from 'vue'
-import App from './App.vue'
+import { createPinia } from "pinia";
+import { createApp } from "vue";
+import App from "./App.vue";
 
-import router from './router'
-import './style.css'
+import router from "./router";
+import "./style.css";
 
-const pinia = createPinia()
-const app = createApp(App)
+const pinia = createPinia();
+const app = createApp(App);
 
-app.use(pinia)
-app.use(router)
+app.use(pinia);
+app.use(router);
 
-app.mount('#app')
+app.mount("#app");

web/src/router/index.ts

@@ -1,28 +1,37 @@
-import { createRouter, createWebHistory } from 'vue-router'
+import { createRouter, createWebHistory } from "vue-router";
 
-import BasicLayout from '@/layouts/BasicLayout.vue'
-import { resetDescription, setTitle } from '@/utils'
-import community from '@/views/CommunityView.vue'
+import BasicLayout from "@/layouts/BasicLayout.vue";
+import { resetDescription, setTitle } from "@/utils";
+import ChatLLM from "@/views/ChatLLMView.vue";
+import VoiceView from "@/views/VoiceView.vue";
 
 const router = createRouter({
   history: createWebHistory(import.meta.env.BASE_URL),
   routes: [
     {
-      path: '/',
+      path: "/",
       component: BasicLayout,
       children: [
         {
-          path: '',
-          name: 'community',
-          component: community,
+          path: "",
+          name: "ChatLLM",
+          component: ChatLLM,
           meta: {
-            title: '社区',
-          },
+            title: "对话"
+          }
         },
-      ],
-    },
-  ],
-})
+        {
+          path: "/voice",
+          name: "Voice",
+          component: VoiceView,
+          meta: {
+            title: "语音对话"
+          }
+        }
+      ]
+    }
+  ]
+});
 
 // // 权限检查函数，检查并决定是否允许访问
 // const checkPermission: NavigationGuard = (to, from, next) => {
@@ -64,17 +73,17 @@ const router = createRouter({
 // // 添加导航守卫
 
 router.beforeEach((to, from, next) => {
-  setTitle(to.meta.title as string)
-  resetDescription()
+  setTitle(to.meta.title as string);
+  resetDescription();
 
   // context.loadingBar?.start();
   // 在每个路由导航前执行权限检查
   // checkPermission(to, from, next);
-  next()
-})
+  next();
+});
 
 // router.afterEach(() => {
 //   context.loadingBar?.finish();
 // });
 
-export default router
+export default router;

web/src/services/audio_websocket.ts

@@ -0,0 +1,30 @@
+import { useWebSocketStore } from "@/services";
+
+export const useAudioWebSocket = () => {
+  const webSocketStore = useWebSocketStore();
+
+  const sendMessage = (data: string | Uint8Array) => {
+    if (webSocketStore.connected) {
+      if (typeof data === "string") {
+        webSocketStore.send(data);
+      } else {
+        webSocketStore.websocket?.send(data);
+      }
+    }
+  };
+
+  const ensureConnection = async (): Promise<void> => {
+    if (!webSocketStore.connected) {
+      webSocketStore.connect();
+      await new Promise<void>((resolve) => {
+        const check = () => {
+          if (webSocketStore.connected) resolve();
+          else setTimeout(check, 100);
+        };
+        check();
+      });
+    }
+  };
+
+  return { sendMessage, ensureConnection };
+};

web/src/services/base_service.ts

@@ -5,27 +5,33 @@ import { context } from "@/utils";
 const BaseClientService = axios.create();
 
 // 添加请求拦截器
-BaseClientService.interceptors.request.use((config) => {
-  // 在发送请求之前做些什么
-  return config;
-}, (e) => {
-  // 对请求错误做些什么
-  return Promise.reject(e);
-});
-
-// 添加响应拦截器
-BaseClientService.interceptors.response.use((response) => {
-  // 2xx 范围内的状态码都会触发该函数。
-  // 对响应数据做点什么
-  return response;
-}, (e) => {
-  // “50”开头统一处理
-  if (e.response?.status >= 500 && e.response?.status < 600) {
-    context.message?.error("服务器开小差了，请稍后再试");
+BaseClientService.interceptors.request.use(
+  (config) => {
+    // 在发送请求之前做些什么
+    return config;
+  },
+  (e) => {
+    // 对请求错误做些什么
     return Promise.reject(e);
   }
-  return Promise.reject(e);
-});
+);
+
+// 添加响应拦截器
+BaseClientService.interceptors.response.use(
+  (response) => {
+    // 2xx 范围内的状态码都会触发该函数。
+    // 对响应数据做点什么
+    return response;
+  },
+  (e) => {
+    // “50”开头统一处理
+    if (e.response?.status >= 500 && e.response?.status < 600) {
+      context.message?.error("服务器开小差了，请稍后再试");
+      return Promise.reject(e);
+    }
+    return Promise.reject(e);
+  }
+);
 
 /** 基础URL */
 export const BaseUrl = "/v1";

web/src/services/chat_service.ts

@@ -1,91 +1,145 @@
-import type { AxiosRequestConfig } from "axios"
+import type { AxiosRequestConfig } from "axios";
 
-import type { IChatWithLLMRequest } from "@/interfaces"
-import BaseClientService, { BaseUrl } from "./base_service.ts"
+import type { IChatWithLLMRequest, UsageInfo } from "@/interfaces";
+import BaseClientService, { BaseUrl } from "./base_service.ts";
 
 export class ChatService {
-  public static basePath = BaseUrl
+  public static basePath = BaseUrl;
 
   /** Chat with LLM */
   public static async ChatWithLLM(
     accessToken: string,
     request: IChatWithLLMRequest,
     onProgress: (content: string) => void,
+    getUsageInfo: (object: UsageInfo) => void = () => {},
+    getThinking: (thinkingContent: string) => void
   ) {
-    let response
-    let buffer = ""
-    let accumulatedContent = ""
+    let response;
+    let buffer = "";
+    let accumulatedContent = "";
+    let thinking = false;
+    let thinkingContent = "";
     try {
       response = await fetch("/v1/chat/completions", {
         method: "POST",
         headers: {
-          "Authorization": `Bearer ${accessToken}`,
-          "Content-Type": "application/json",
+          Authorization: `Bearer ${accessToken}`,
+          "Content-Type": "application/json"
         },
-        body: JSON.stringify(request),
-      })
+        body: JSON.stringify(request)
+      });
 
       if (!response.ok) {
         // eslint-disable-next-line unicorn/error-message
-        throw new Error()
+        throw new Error();
       }
 
-      const reader = response.body?.getReader()
-      const decoder = new TextDecoder()
+      const reader = response.body?.getReader();
+      const decoder = new TextDecoder();
 
       while (true) {
-        const { done, value } = await reader!.read()
+        const { done, value } = await reader!.read();
 
-        if (done)
-          break
+        if (done) break;
 
         // 将二进制数据转为字符串并存入缓冲区
-        buffer += decoder.decode(value)
+        buffer += decoder.decode(value);
 
         // 查找换行符分割数据
-        const lines = buffer.split("\n")
+        const lines = buffer.split("\n");
 
         // 保留未处理完的部分
-        buffer = lines.pop() || ""
+        buffer = lines.pop() || "";
 
         // 处理每一行
         for (const line of lines) {
-          const trimmedLine = line.trim()
-          if (!trimmedLine)
-            continue
+          const trimmedLine = line.trim();
+          if (!trimmedLine) continue;
 
           if (trimmedLine.startsWith("data: ")) {
-            const jsonStr = trimmedLine.slice(6)
+            const jsonStr = trimmedLine.slice(6);
 
             // 处理结束标记
             if (jsonStr === "[DONE]") {
-              onProgress(accumulatedContent) // 最终更新
-              return
+              onProgress(accumulatedContent); // 最终更新
+              return;
             }
 
             try {
-              const data = JSON.parse(jsonStr)
+              const data = JSON.parse(jsonStr);
+
+              // 处理内容
               if (data.choices?.[0]?.delta?.content) {
-                // 累积内容
-                accumulatedContent += data.choices[0].delta.content
-                // 触发回调
-                onProgress(accumulatedContent)
+                const content = data.choices[0].delta.content;
+
+                if (content.includes("<think>")) {
+                  thinking = true;
+                  // 只处理<think>前的内容
+                  const [beforeThink, afterThink] = content.split("<think>");
+                  if (beforeThink) {
+                    accumulatedContent += beforeThink;
+                    onProgress(accumulatedContent);
+                  }
+                  thinkingContent = "";
+                  if (afterThink) {
+                    thinkingContent += afterThink;
+                    getThinking(thinkingContent);
+                  }
+                  continue;
+                }
+
+                if (content.includes("</think>")) {
+                  thinking = false;
+                  // 只处理</think>后的内容
+                  const [beforeEndThink, afterEndThink] =
+                    content.split("</think>");
+                  thinkingContent += beforeEndThink;
+                  getThinking(thinkingContent);
+                  if (afterEndThink) {
+                    accumulatedContent += afterEndThink;
+                    onProgress(accumulatedContent);
+                  }
+                  thinkingContent = "";
+                  continue;
+                }
+
+                if (thinking) {
+                  thinkingContent += content;
+                  getThinking(thinkingContent);
+                } else {
+                  accumulatedContent += content;
+                  onProgress(accumulatedContent);
+                }
               }
-            }
-            catch (err) {
-              console.error("JSON解析失败:", err)
+
+              // 处理使用信息
+              if (data.usage) {
+                const { prompt_tokens, completion_tokens, total_tokens } =
+                  data.usage;
+                getUsageInfo({
+                  prompt_tokens,
+                  completion_tokens,
+                  total_tokens
+                });
+              }
+            } catch (err) {
+              console.error("JSON解析失败:", err);
             }
           }
         }
       }
-    }
-    catch (err) {
-      console.error("Error:", err)
+    } catch (err) {
+      console.error("Error:", err);
     }
   }
 
   // 获取模型列表
   public static GetModelList(config?: AxiosRequestConfig<any>) {
-    return BaseClientService.get(`${this.basePath}/model/list`, config)
+    return BaseClientService.get(`${this.basePath}/model/list`, config);
+  }
+
+  // 获取音色列表
+  public static GetSpeakerList(config?: AxiosRequestConfig<any>) {
+    return BaseClientService.get(`${this.basePath}/speaker/list`, config);
   }
 }

web/src/services/index.ts

@@ -1,3 +1,4 @@
-export * from "./base_service"
-export * from "./chat_service"
-export * from "./websocket"
+export * from "./audio_websocket";
+export * from "./base_service";
+export * from "./chat_service";
+export * from "./websocket";

web/src/services/websocket.ts

@@ -1,28 +1,145 @@
-import { useChatStore } from "@/stores";
+import { useChatStore, useTtsStore } from "@/stores";
 
 // WebSocket
 export const useWebSocketStore = defineStore("websocket", () => {
   const websocket = ref<WebSocket>();
   const connected = ref(false);
   const chatStore = useChatStore();
+  const ttsStore = useTtsStore();
+  const router = useRouter();
 
   const { onlineCount } = storeToRefs(chatStore);
 
   const onmessage = (e: MessageEvent) => {
-    const data = JSON.parse(e.data);
-    switch (data.type) {
-      case "count":
-        onlineCount.value = data.online_count;
-        break;
-      case "asr_result":
-        chatStore.addMessageToHistory(data.result);
+    // 检查消息类型
+    if (e.data instanceof ArrayBuffer) {
+      // 处理二进制音频数据（兜底处理，新版本应该不会用到）
+      console.log("收到二进制音频数据，大小:", e.data.byteLength);
+      // 可以选择忽略或者作为兜底处理
+    } else if (e.data instanceof Blob) {
+      // 如果是Blob，转换为ArrayBuffer（兜底处理）
+      e.data.arrayBuffer().then((buffer: ArrayBuffer) => {
+        console.log("收到Blob音频数据，大小:", buffer.byteLength);
+      });
+    } else if (typeof e.data === "string") {
+      // 处理文本JSON消息
+      try {
+        const data = JSON.parse(e.data);
+        switch (data.type) {
+          case "count":
+            onlineCount.value = data.online_count;
+            break;
+          case "asr_result":
+            if (router.currentRoute.value.path === "/") {
+              chatStore.addMessageToHistory(data.result);
+            } else if (router.currentRoute.value.path === "/voice") {
+              // 在语音页面的处理
+              chatStore.addMessageToHistory(data.result, "user", "voice");
+            } else {
+              console.warn(data);
+            }
+            break;
+
+          // 新的TTS消息格式处理
+          case "tts_audio_data":
+            // 新的音频数据格式，包含messageId和hex格式的音频数据
+            if (data.messageId && data.audioData) {
+              console.log(
+                `收到TTS音频数据 [${data.messageId}]，hex长度:`,
+                data.audioData.length
+              );
+              try {
+                // 将hex字符串转换为ArrayBuffer
+                const bytes = data.audioData
+                  .match(/.{1,2}/g)
+                  ?.map((byte: string) => Number.parseInt(byte, 16));
+                if (bytes) {
+                  const buffer = new Uint8Array(bytes).buffer;
+                  console.log(
+                    `转换后的音频数据大小 [${data.messageId}]:`,
+                    buffer.byteLength
+                  );
+                  ttsStore.handleAudioData(buffer, data.messageId);
+                } else {
+                  console.error(`音频数据格式错误 [${data.messageId}]`);
+                }
+              } catch (error) {
+                console.error(`音频数据转换失败 [${data.messageId}]:`, error);
+                ttsStore.handleError(
+                  `音频数据转换失败: ${error}`,
+                  data.messageId
+                );
+              }
+            } else {
+              console.error("tts_audio_data消息格式错误:", data);
+            }
+            break;
+
+          case "tts_audio_complete":
+            // TTS音频传输完成
+            if (data.messageId) {
+              console.log(`TTS音频传输完成 [${data.messageId}]`);
+              ttsStore.finishConversion(data.messageId);
+            } else {
+              console.log("TTS音频传输完成（无messageId）");
+              ttsStore.finishConversion(data.messageId);
+            }
+            break;
+
+          case "tts_complete":
+            // TTS会话结束
+            if (data.messageId) {
+              console.log(`TTS会话结束 [${data.messageId}]`);
+            } else {
+              console.log("TTS会话结束");
+            }
+            break;
+
+          case "tts_error":
+            // TTS错误
+            if (data.messageId) {
+              console.error(`TTS错误 [${data.messageId}]:`, data.message);
+              ttsStore.handleError(data.message, data.messageId);
+            } else {
+              console.error("TTS错误:", data.message);
+              ttsStore.handleError(data.message, data.messageId || "unknown");
+            }
+            break;
+
+          case "llm_complete_response":
+            // LLM部分响应
+            if (router.currentRoute.value.path === "/voice") {
+              chatStore.addMessageToHistory(data.content, "assistant", "voice");
+            }
+            break;
+
+          default:
+            console.log("未知消息类型:", data.type, data);
+        }
+      } catch (error) {
+        console.error("JSON解析错误:", error, "原始数据:", e.data);
+      }
+    } else {
+      console.warn("收到未知格式的消息:", typeof e.data, e.data);
     }
   };
 
   const send = (data: string) => {
-    if (websocket.value && websocket.value.readyState === WebSocket.OPEN)
+    if (websocket.value && websocket.value.readyState === WebSocket.OPEN) {
       websocket.value?.send(data);
+    } else {
+      console.warn("WebSocket未连接，无法发送消息:", data);
+    }
   };
+
+  const sendBinary = (data: ArrayBuffer | Uint8Array) => {
+    if (websocket.value && websocket.value.readyState === WebSocket.OPEN) {
+      websocket.value?.send(data);
+    } else {
+      console.warn("WebSocket未连接，无法发送二进制数据");
+    }
+  };
+
   const close = () => {
     websocket.value?.close();
   };
@@ -33,12 +150,15 @@ export const useWebSocketStore = defineStore("websocket", () => {
 
     websocket.value.onopen = () => {
       connected.value = true;
+      console.log("WebSocket连接成功");
 
       let pingIntervalId: NodeJS.Timeout | undefined;
 
-      if (pingIntervalId)
-        clearInterval(pingIntervalId);
-      pingIntervalId = setInterval(() => send("ping"), 30 * 1000);
+      if (pingIntervalId) clearInterval(pingIntervalId);
+      pingIntervalId = setInterval(() => {
+        // 修改ping格式为JSON格式，与后端保持一致
+        send(JSON.stringify({ type: "ping" }));
+      }, 30 * 1000);
 
       if (websocket.value) {
         websocket.value.onmessage = onmessage;
@@ -46,21 +166,29 @@ export const useWebSocketStore = defineStore("websocket", () => {
         websocket.value.onerror = (e: Event) => {
           console.error(`WebSocket错误:${(e as ErrorEvent).message}`);
         };
-        websocket.value.onclose = () => {
+
+        websocket.value.onclose = (e: CloseEvent) => {
           connected.value = false;
+          console.log(`WebSocket连接关闭: ${e.code} ${e.reason}`);
           setTimeout(() => {
+            console.log("尝试重新连接WebSocket...");
             connect(); // 尝试重新连接
           }, 1000); // 1秒后重试连接
         };
       }
     };
+
+    websocket.value.onerror = (e: Event) => {
+      console.error("WebSocket连接错误:", e);
+    };
   };
 
   return {
     websocket,
     connected,
     send,
+    sendBinary,
     close,
-    connect,
+    connect
   };
 });

web/src/stores/asr_store.ts

@@ -1,5 +1,6 @@
 import { useWebSocketStore } from "@/services";
 import { convertToPCM16 } from "@/utils";
+import { useChatStore } from "./chat_store";
 
 export const useAsrStore = defineStore("asr", () => {
   // 是否正在录音
@@ -11,8 +12,8 @@ export const useAsrStore = defineStore("asr", () => {
   let mediaStreamSource: MediaStreamAudioSourceNode | null = null;
   let workletNode: AudioWorkletNode | null = null;
 
-  // 获取 WebSocket store 实例
   const webSocketStore = useWebSocketStore();
+  const router = useRouter();
 
   /**
    * 发送消息到 WebSocket
@@ -23,8 +24,7 @@ export const useAsrStore = defineStore("asr", () => {
     if (webSocketStore.connected) {
       if (typeof data === "string") {
         webSocketStore.send(data);
-      }
-      else {
+      } else {
         webSocketStore.websocket?.send(data);
       }
     }
@@ -53,8 +53,7 @@ export const useAsrStore = defineStore("asr", () => {
    * 开始录音
    */
   const startRecording = async () => {
-    if (isRecording.value)
-      return;
+    if (isRecording.value) return;
     messages.value = [];
     // 确保 WebSocket 已连接
     if (!webSocketStore.connected) {
@@ -62,8 +61,7 @@ export const useAsrStore = defineStore("asr", () => {
       // 等待连接建立
       await new Promise<void>((resolve) => {
         const check = () => {
-          if (webSocketStore.connected)
-            resolve();
+          if (webSocketStore.connected) resolve();
           else setTimeout(check, 100);
         };
         check();
@@ -73,24 +71,30 @@ export const useAsrStore = defineStore("asr", () => {
       // 获取麦克风音频流
       const stream = await navigator.mediaDevices.getUserMedia({ audio: true });
       // 创建音频上下文，采样率16kHz
-      audioContext = new (window.AudioContext || (window as any).webkitAudioContext)({
-        sampleRate: 16000,
+      audioContext = new (window.AudioContext ||
+        (window as any).webkitAudioContext)({
+        sampleRate: 16000
       });
       // 用Blob方式创建AudioWorklet模块的URL
-      const blob = new Blob([audioProcessorCode], { type: "application/javascript" });
+      const blob = new Blob([audioProcessorCode], {
+        type: "application/javascript"
+      });
       const processorUrl = URL.createObjectURL(blob);
       // 加载AudioWorklet模块
       await audioContext.audioWorklet.addModule(processorUrl);
-      // 释放URL对象（防止内存泄漏）
+      // 释放URL对象防止内存泄漏
       URL.revokeObjectURL(processorUrl);
+
       // 创建音频源节点
       mediaStreamSource = audioContext.createMediaStreamSource(stream);
+
       // 创建AudioWorkletNode
       workletNode = new AudioWorkletNode(audioContext, "audio-processor", {
         numberOfInputs: 1,
         numberOfOutputs: 1,
-        channelCount: 1,
+        channelCount: 1
       });
+
       // 监听来自AudioWorklet的音频数据
       workletNode.port.onmessage = (event) => {
         if (event.data.type === "audiodata") {
@@ -104,8 +108,7 @@ export const useAsrStore = defineStore("asr", () => {
       mediaStreamSource.connect(workletNode);
       workletNode.connect(audioContext.destination);
       isRecording.value = true;
-    }
-    catch (err) {
+    } catch (err) {
       // 麦克风权限失败或AudioWorklet加载失败
       console.error("需要麦克风权限才能录音", err);
     }
@@ -115,16 +118,22 @@ export const useAsrStore = defineStore("asr", () => {
    * 停止录音
    */
   const stopRecording = () => {
-    if (!isRecording.value)
-      return;
+    if (!isRecording.value) return;
 
+    const messageId = `voice_${Date.now()}`;
     // 通知后端录音结束
-    sendMessage(JSON.stringify({ type: "asr_end" }));
+    const msg: Record<string, any> = { type: "asr_end" };
+    if (router.currentRoute.value.path === "/voice") {
+      msg.messageId = messageId;
+      msg.voiceConversation = true;
+      msg.speaker = useChatStore().speakerInfo?.speaker_id;
+    }
+    sendMessage(JSON.stringify(msg));
 
     // 停止所有音轨
     if (mediaStreamSource?.mediaStream) {
       const tracks = mediaStreamSource.mediaStream.getTracks();
-      tracks.forEach(track => track.stop());
+      tracks.forEach((track) => track.stop());
     }
 
     // 断开音频节点
@@ -149,6 +158,6 @@ export const useAsrStore = defineStore("asr", () => {
     messages,
     startRecording,
     stopRecording,
-    sendMessage,
+    sendMessage
   };
 });

web/src/stores/chat_store.ts

@@ -1,94 +1,186 @@
-import type { IChatWithLLMRequest, ModelInfo, ModelListInfo } from "@/interfaces";
+import type {
+  CategorySpeakers,
+  IChatWithLLMRequest,
+  ModelInfo,
+  ModelListInfo,
+  Speaker,
+  UsageInfo
+} from "@/interfaces";
 import { ChatService } from "@/services";
 
 export const useChatStore = defineStore("chat", () => {
-  const token = ("sk-fkGVZBrAqvIxLjlF3b5f19EfBb63486c90Fa5a1fBd7076Ee");
+  const router = useRouter();
+  const token = "sk-fkGVZBrAqvIxLjlF3b5f19EfBb63486c90Fa5a1fBd7076Ee";
+
   // 默认模型
   const modelInfo = ref<ModelInfo | null>(null);
   // 历史消息
   const historyMessages = ref<IChatWithLLMRequest["messages"]>([]);
   // 是否正在响应
   const completing = ref<boolean>(false);
+  // 是否正在思考
+  const thinking = ref<boolean>(false);
+  // 模型列表
+  const modelList = ref<ModelListInfo[]>([]);
+  // 音色列表
+  const speakerList = ref<CategorySpeakers[]>([]);
+  // 当前音色信息
+  const speakerInfo = ref<Speaker | null>(null);
   // 在线人数
   const onlineCount = ref<number>(0);
 
+  // 生成消息ID方法
+  const generateMessageId = () => new Date().getTime().toString();
+
+  // 获取最后一条消息
+  const getLastMessage = () =>
+    historyMessages.value[historyMessages.value.length - 1];
+
   // 与 LLM 聊天
   const chatWithLLM = async (
     request: IChatWithLLMRequest,
-    onProgress: (content: string) => void, // 接收进度回调
+    onProgress: (content: string) => void,
+    getUsageInfo: (object: UsageInfo) => void = () => {},
+    getThinking: (thinkingContent: string) => void = () => {}
   ) => {
-    if (completing.value)
-      throw new Error("正在响应中");
+    if (completing.value) throw new Error("正在响应中");
 
-    completing.value = true; // 开始请求
+    completing.value = true;
     try {
-      await ChatService.ChatWithLLM(token, request, (content) => {
-        onProgress(content);
-      });
-    }
-    catch (error) {
+      await ChatService.ChatWithLLM(
+        token,
+        request,
+        onProgress,
+        getUsageInfo,
+        getThinking
+      );
+    } catch (error) {
       console.error("请求失败:", error);
-    }
-    finally {
+    } finally {
       completing.value = false;
     }
   };
 
   // 添加消息到历史记录
-  const addMessageToHistory = (message: string) => {
+  const addMessageToHistory = (
+    message: string,
+    role: "user" | "assistant" = "user",
+    type: "chat" | "voice" = "chat"
+  ) => {
     const content = message.trim();
-    if (!content)
-      return;
+    if (!content) return;
 
     historyMessages.value.push({
-      role: "user",
+      role,
       content,
+      type,
+      id: generateMessageId()
     });
   };
 
   // 清除历史消息
-  const clearHistoryMessages = () => {
-    historyMessages.value = [];
+  const clearHistoryMessages = (type?: "chat" | "voice") => {
+    historyMessages.value = type
+      ? historyMessages.value.filter((msg) => msg.type !== type)
+      : [];
   };
 
-  watch(historyMessages, (newVal) => {
-    // 当历史消息变化时，发送请求
-    if (newVal.length > 0) {
-      const lastMessage = newVal[newVal.length - 1];
-      if (lastMessage.role === "user" && modelInfo.value) {
-        chatWithLLM({
-          messages: newVal,
-          model: modelInfo.value?.model_id,
-        }, (content) => {
-          // 处理进度回调
-          if (
-            historyMessages.value.length === 0
-            || historyMessages.value[historyMessages.value.length - 1].role !== "assistant"
-          ) {
-            historyMessages.value.push({
-              role: "assistant",
-              content: "",
-            });
-          }
-          historyMessages.value[historyMessages.value.length - 1].content = content;
-        });
-      }
+  // 确保最后一条消息是助手消息
+  const ensureAssistantMessage = () => {
+    const lastMessage = getLastMessage();
+    if (!lastMessage || lastMessage.role !== "assistant") {
+      historyMessages.value.push({
+        role: "assistant",
+        content: ""
+      });
     }
-  }, { deep: true });
+  };
 
-  // 模型列表
-  const modelList = ref<ModelListInfo[]>([]);
+  // 处理聊天响应的逻辑
+  const handleChatResponse = async (
+    messages: IChatWithLLMRequest["messages"]
+  ) => {
+    if (!modelInfo.value) return;
+
+    // 过滤出type为chat的聊天消息
+    const filteredMessages = computed(() =>
+      messages.filter((msg) => msg.type === "chat" || !msg.type)
+    );
+
+    await chatWithLLM(
+      { messages: filteredMessages.value, model: modelInfo.value.model_id },
+      // 处理文本内容
+      (content) => {
+        ensureAssistantMessage();
+        thinking.value = false;
+        getLastMessage().content = content;
+      },
+      // 处理使用信息
+      (usageInfo: UsageInfo) => {
+        const lastMessage = getLastMessage();
+        if (lastMessage?.role === "assistant") {
+          lastMessage.usage = usageInfo;
+        }
+      },
+      // 处理思维链
+      (thinkingContent: string) => {
+        ensureAssistantMessage();
+        thinking.value = true;
+        getLastMessage().thinking = thinkingContent;
+      }
+    );
+
+    // 设置消息ID
+    getLastMessage().id = generateMessageId();
+  };
+
+  watch(
+    historyMessages,
+    (newVal) => {
+      if (newVal.length > 0 && router.currentRoute.value.path === "/") {
+        const lastMessage = newVal[newVal.length - 1];
+        if (lastMessage.role === "user") {
+          handleChatResponse(newVal);
+        }
+      }
+    },
+    { deep: true }
+  );
 
   // 获取模型列表
   const getModelList = async () => {
     try {
       const response = await ChatService.GetModelList();
       modelList.value = response.data.data;
-    }
-    catch (error) {
+    } catch (error) {
       console.error("获取模型列表失败:", error);
     }
   };
 
-  return { token, completing, chatWithLLM, historyMessages, addMessageToHistory, clearHistoryMessages, getModelList, modelList, modelInfo, onlineCount };
+  // 获取音色列表
+  const getSpeakerList = async () => {
+    try {
+      const response = await ChatService.GetSpeakerList();
+      speakerList.value = response.data.data;
+    } catch (error) {
+      console.error("获取音色·列表失败:", error);
+    }
+  };
+
+  return {
+    token,
+    completing,
+    thinking,
+    modelInfo,
+    modelList,
+    historyMessages,
+    chatWithLLM,
+    addMessageToHistory,
+    clearHistoryMessages,
+    getModelList,
+    onlineCount,
+    speakerList,
+    getSpeakerList,
+    speakerInfo
+  };
 });

web/src/stores/index.ts

@@ -1,2 +1,4 @@
-export * from "./asr_store"
-export * from "./chat_store"
+export * from "./asr_store";
+export * from "./chat_store";
+export * from "./layout_store";
+export * from "./tts_store";

web/src/stores/layout_store.ts

@@ -0,0 +1,44 @@
+import { matchMedia } from "@/utils";
+
+export const useLayoutStore = defineStore("layout", () => {
+  // 侧边栏状态
+  const hiddenLeftSidebar = ref(false);
+  // 简洁按钮
+  const simpleMode = ref(false);
+
+  const handleResize = () => {
+    matchMedia(
+      "sm",
+      () => {
+        hiddenLeftSidebar.value = true;
+      },
+      () => {
+        hiddenLeftSidebar.value = false;
+      }
+    );
+    matchMedia(
+      "536",
+      () => {
+        simpleMode.value = true;
+      },
+      () => {
+        simpleMode.value = false;
+      }
+    );
+  };
+
+  const toggleLeftSidebar = () => {
+    hiddenLeftSidebar.value = !hiddenLeftSidebar.value;
+  };
+
+  window.addEventListener("resize", handleResize);
+
+  onMounted(() => {
+    handleResize();
+  });
+
+  onUnmounted(() => {
+    window.removeEventListener("resize", handleResize);
+  });
+  return { hiddenLeftSidebar, toggleLeftSidebar, simpleMode };
+});

web/src/stores/tts_store.ts

@@ -0,0 +1,311 @@
+import { useAudioWebSocket } from "@/services";
+import { createAudioUrl, mergeAudioChunks } from "@/utils";
+import { useChatStore } from "./chat_store";
+
+interface AudioState {
+  isPlaying: boolean;
+  isLoading: boolean;
+  audioElement: HTMLAudioElement | null;
+  audioUrl: string | null;
+  audioChunks: ArrayBuffer[];
+  hasError: boolean;
+  errorMessage: string;
+}
+
+export const useTtsStore = defineStore("tts", () => {
+  const chatStore = useChatStore();
+  // 多音频状态管理 - 以消息ID为key
+  const audioStates = ref<Map<string, AudioState>>(new Map());
+
+  // 当前活跃的转换请求（保留用于兼容性）
+  const activeConversion = ref<string | null>(null);
+
+  // 会话状态
+  const hasActiveSession = ref(false);
+
+  // WebSocket连接
+  const { sendMessage, ensureConnection } = useAudioWebSocket();
+
+  /**
+   * 获取或创建音频状态
+   */
+  const getAudioState = (messageId: string): AudioState => {
+    if (!audioStates.value.has(messageId)) {
+      audioStates.value.set(messageId, {
+        isPlaying: false,
+        isLoading: false,
+        audioElement: null,
+        audioUrl: null,
+        audioChunks: [],
+        hasError: false,
+        errorMessage: ""
+      });
+    }
+    return audioStates.value.get(messageId)!;
+  };
+
+  /**
+   * 发送文本进行TTS转换
+   */
+  const convertText = async (text: string, messageId: string) => {
+    try {
+      await ensureConnection();
+
+      // 暂停其他正在播放的音频
+      pauseAll();
+
+      // 获取当前消息的状态
+      const state = getAudioState(messageId);
+
+      // 清理之前的音频和错误状态
+      clearAudioState(state);
+      state.isLoading = true;
+      state.audioChunks = [];
+
+      // 设置当前活跃转换
+      activeConversion.value = messageId;
+      hasActiveSession.value = true;
+
+      // 发送文本到TTS服务
+      sendMessage(
+        JSON.stringify({
+          type: "tts_text",
+          text,
+          messageId,
+          speaker: chatStore.speakerInfo?.speaker_id
+        })
+      );
+    } catch (error) {
+      handleError(`连接失败: ${error}`, messageId);
+    }
+  };
+
+  /**
+   * 处理接收到的音频数据 - 修改为支持messageId参数
+   */
+  const handleAudioData = (data: ArrayBuffer, messageId?: string) => {
+    // 如果传递了messageId就使用它，否则使用activeConversion
+    const targetMessageId = messageId || activeConversion.value;
+    if (!targetMessageId) {
+      console.warn("handleAudioData: 没有有效的messageId");
+      return;
+    }
+
+    console.log(`接收音频数据 [${targetMessageId}]，大小:`, data.byteLength);
+    const state = getAudioState(targetMessageId);
+    state.audioChunks.push(data);
+  };
+
+  /**
+   * 完成TTS转换，创建播放器并自动播放 - 修改为支持messageId参数
+   */
+  const finishConversion = async (messageId?: string) => {
+    // 如果传递了messageId就使用它，否则使用activeConversion
+    const targetMessageId = messageId || activeConversion.value;
+    if (!targetMessageId) {
+      console.warn("finishConversion: 没有有效的messageId");
+      return;
+    }
+
+    const state = getAudioState(targetMessageId);
+    console.log(
+      `完成TTS转换 [${targetMessageId}]，音频片段数量:`,
+      state.audioChunks.length
+    );
+
+    if (state.audioChunks.length === 0) {
+      handleError("没有接收到音频数据", targetMessageId);
+      return;
+    }
+
+    try {
+      // 合并音频片段
+      const mergedAudio = mergeAudioChunks(state.audioChunks);
+      console.log(
+        `合并后音频大小 [${targetMessageId}]:`,
+        mergedAudio.byteLength
+      );
+
+      // 创建音频URL和元素
+      state.audioUrl = createAudioUrl(mergedAudio);
+      state.audioElement = new Audio(state.audioUrl);
+
+      // 设置音频事件
+      setupAudioEvents(state, targetMessageId);
+
+      state.isLoading = false;
+
+      // 清除activeConversion（如果是当前活跃的）
+      if (activeConversion.value === targetMessageId) {
+        activeConversion.value = null;
+      }
+
+      console.log(`TTS音频准备完成 [${targetMessageId}]，开始自动播放`);
+
+      // 自动播放
+      await play(targetMessageId);
+    } catch (error) {
+      handleError(`音频处理失败: ${error}`, targetMessageId);
+    }
+  };
+
+  /**
+   * 设置音频事件监听
+   */
+  const setupAudioEvents = (state: AudioState, messageId: string) => {
+    if (!state.audioElement) return;
+
+    const audio = state.audioElement;
+
+    audio.addEventListener("ended", () => {
+      state.isPlaying = false;
+      console.log(`音频播放结束 [${messageId}]`);
+    });
+
+    audio.addEventListener("error", (e) => {
+      console.error(`音频播放错误 [${messageId}]:`, e);
+      handleError("音频播放失败", messageId);
+    });
+
+    audio.addEventListener("canplaythrough", () => {
+      console.log(`音频可以播放 [${messageId}]`);
+    });
+  };
+
+  /**
+   * 播放指定消息的音频
+   */
+  const play = async (messageId: string) => {
+    const state = getAudioState(messageId);
+
+    if (!state.audioElement) {
+      handleError("音频未准备好", messageId);
+      return;
+    }
+
+    try {
+      // 暂停其他正在播放的音频
+      pauseAll(messageId);
+
+      await state.audioElement.play();
+      state.isPlaying = true;
+      state.hasError = false;
+      state.errorMessage = "";
+      console.log(`开始播放音频 [${messageId}]`);
+    } catch (error) {
+      handleError(`播放失败: ${error}`, messageId);
+    }
+  };
+
+  /**
+   * 暂停指定消息的音频
+   */
+  const pause = (messageId: string) => {
+    const state = getAudioState(messageId);
+
+    if (!state.audioElement) return;
+
+    state.audioElement.pause();
+    state.isPlaying = false;
+    console.log(`暂停音频 [${messageId}]`);
+  };
+
+  /**
+   * 暂停所有音频
+   */
+  const pauseAll = (excludeMessageId?: string) => {
+    audioStates.value.forEach((state, messageId) => {
+      if (excludeMessageId && messageId === excludeMessageId) return;
+      if (state.isPlaying && state.audioElement) {
+        state.audioElement.pause();
+        state.isPlaying = false;
+      }
+    });
+  };
+
+  /**
+   * 处理TTS错误 - 修改为支持messageId参数
+   */
+  const handleError = (errorMsg: string, messageId?: string) => {
+    // 如果传递了messageId就使用它，否则使用activeConversion
+    const targetMessageId = messageId || activeConversion.value;
+    if (!targetMessageId) {
+      console.error(`TTS错误 (无messageId): ${errorMsg}`);
+      return;
+    }
+
+    console.error(`TTS错误 [${targetMessageId}]: ${errorMsg}`);
+    const state = getAudioState(targetMessageId);
+    state.hasError = true;
+    state.errorMessage = errorMsg;
+    state.isLoading = false;
+
+    if (activeConversion.value === targetMessageId) {
+      activeConversion.value = null;
+      hasActiveSession.value = false;
+    }
+  };
+
+  /**
+   * 清理指定消息的音频资源
+   */
+  const clearAudio = (messageId: string) => {
+    const state = getAudioState(messageId);
+    clearAudioState(state);
+    audioStates.value.delete(messageId);
+  };
+
+  /**
+   * 清理音频状态
+   */
+  const clearAudioState = (state: AudioState) => {
+    if (state.audioElement) {
+      state.audioElement.pause();
+      state.audioElement = null;
+    }
+
+    if (state.audioUrl) {
+      URL.revokeObjectURL(state.audioUrl);
+      state.audioUrl = null;
+    }
+
+    state.isPlaying = false;
+    state.audioChunks = [];
+    state.hasError = false;
+    state.errorMessage = "";
+  };
+
+  // 状态查询方法
+  const isPlaying = (messageId: string) => getAudioState(messageId).isPlaying;
+  const isLoading = (messageId: string) => getAudioState(messageId).isLoading;
+  const hasAudio = (messageId: string) =>
+    !!getAudioState(messageId).audioElement;
+  const hasError = (messageId: string) => getAudioState(messageId).hasError;
+  const getErrorMessage = (messageId: string) =>
+    getAudioState(messageId).errorMessage;
+
+  // 组件卸载时清理所有资源
+  onUnmounted(() => {
+    audioStates.value.forEach((state) => clearAudioState(state));
+    audioStates.value.clear();
+  });
+
+  return {
+    // 状态查询方法
+    isPlaying,
+    isLoading,
+    hasAudio,
+    hasError,
+    getErrorMessage,
+
+    // 核心方法
+    convertText,
+    handleAudioData,
+    finishConversion,
+    play,
+    pause,
+    pauseAll,
+    clearAudio,
+    handleError
+  };
+});

web/src/stores/user_store.ts

@@ -1,5 +1,3 @@
 export const useUserStore = defineStore("user", () => {
-  return {
-
-  };
+  return {};
 });

web/src/style.css

@@ -1,2 +1,2 @@
-@import 'tailwindcss';
-@import 'highlight.js/styles/github.css';
+@import "tailwindcss";
+@import "highlight.js/styles/github.css";

web/src/utils/audio.ts

@@ -0,0 +1,20 @@
+// 合并音频片段
+export const mergeAudioChunks = (chunks: ArrayBuffer[]): Uint8Array => {
+  const totalLength = chunks.reduce((acc, chunk) => acc + chunk.byteLength, 0);
+  const merged = new Uint8Array(totalLength);
+  let offset = 0;
+  chunks.forEach((chunk) => {
+    merged.set(new Uint8Array(chunk), offset);
+    offset += chunk.byteLength;
+  });
+  return merged;
+};
+
+// 创建音频播放URL
+export const createAudioUrl = (
+  audioData: Uint8Array,
+  mimeType = "audio/mp3"
+): string => {
+  const blob = new Blob([audioData as BlobPart], { type: mimeType });
+  return URL.createObjectURL(blob);
+};

web/src/utils/clipboard.ts

@@ -0,0 +1,23 @@
+const leagacyCopy = (text: string) => {
+  const input = document.createElement("input");
+  input.value = text;
+  document.body.appendChild(input);
+  input.select();
+  try {
+    document.execCommand("copy");
+  } catch (err) {
+    console.error(err);
+  }
+  document.body.removeChild(input);
+};
+
+export const copy = (text: string) => {
+  if (navigator.clipboard && navigator.clipboard.writeText) {
+    navigator.clipboard.writeText(text).catch((err) => {
+      console.error(err);
+      leagacyCopy(text); // 如果现代API失败，使用旧方法
+    });
+  } else {
+    leagacyCopy(text); // 如果没有现代API，使用旧方法
+  }
+};

web/src/utils/context.ts

@@ -1,9 +1,9 @@
-import type { LoadingBarApiInjection } from "naive-ui/es/loading-bar/src/LoadingBarProvider"
-import type { MessageApiInjection } from "naive-ui/es/message/src/MessageProvider"
-import type { NotificationApiInjection } from "naive-ui/es/notification/src/NotificationProvider"
+import type { LoadingBarApiInjection } from "naive-ui/es/loading-bar/src/LoadingBarProvider";
+import type { MessageApiInjection } from "naive-ui/es/message/src/MessageProvider";
+import type { NotificationApiInjection } from "naive-ui/es/notification/src/NotificationProvider";
 
 export const context: {
-  message?: MessageApiInjection
-  notification?: NotificationApiInjection
-  loadingBar?: LoadingBarApiInjection
-} = {}
+  message?: MessageApiInjection;
+  notification?: NotificationApiInjection;
+  loadingBar?: LoadingBarApiInjection;
+} = {};

web/src/utils/format.ts

@@ -0,0 +1,7 @@
+export const formatTime = (seconds: number): string => {
+  if (Number.isNaN(seconds) || !Number.isFinite(seconds)) return "00:00";
+
+  const minutes = Math.floor(seconds / 60);
+  const secs = Math.floor(seconds % 60);
+  return `${minutes.toString().padStart(2, "0")}:${secs.toString().padStart(2, "0")}`;
+};

web/src/utils/index.ts

@@ -1,6 +1,9 @@
-export * from "./context"
-export * from "./media"
-export * from "./pcm"
-export * from "./title"
-export * from "./title"
-export * from "./url"
+export * from "./audio";
+export * from "./clipboard";
+export * from "./context";
+export * from "./format";
+export * from "./media";
+export * from "./pcm";
+export * from "./title";
+export * from "./title";
+export * from "./url";

web/src/utils/media.ts

@@ -2,42 +2,54 @@
 export const matchMedia = (
   type: "sm" | "md" | "lg" | string,
   matchFunc?: Function,
-  mismatchFunc?: Function,
+  mismatchFunc?: Function
 ) => {
   if (type === "sm") {
     if (window.matchMedia("(max-width: 767.98px)").matches) {
       /* 窗口小于或等于 */
-      matchFunc?.()
+      matchFunc?.();
+    } else {
+      mismatchFunc?.();
     }
-    else {
-      mismatchFunc?.()
-    }
-  }
-  else if (type === "md") {
+  } else if (type === "md") {
     if (window.matchMedia("(max-width: 992px)").matches) {
       /* 窗口小于或等于 */
-      matchFunc?.()
+      matchFunc?.();
+    } else {
+      mismatchFunc?.();
     }
-    else {
-      mismatchFunc?.()
-    }
-  }
-  else if (type === "lg") {
+  } else if (type === "lg") {
     if (window.matchMedia("(max-width: 1200px)").matches) {
       /* 窗口小于或等于 */
-      matchFunc?.()
+      matchFunc?.();
+    } else {
+      mismatchFunc?.();
     }
-    else {
-      mismatchFunc?.()
-    }
-  }
-  else {
+  } else {
     if (window.matchMedia(`(max-width: ${type}px)`).matches) {
       /* 窗口小于或等于 */
-      matchFunc?.()
-    }
-    else {
-      mismatchFunc?.()
+      matchFunc?.();
+    } else {
+      mismatchFunc?.();
     }
   }
-}
+};
+
+/** 获取视窗宽度 */
+export const useWindowWidth = () => {
+  const width = ref(window.innerWidth);
+
+  const updateWidth = () => {
+    width.value = window.innerWidth;
+  };
+
+  onMounted(() => {
+    window.addEventListener("resize", updateWidth);
+  });
+
+  onUnmounted(() => {
+    window.removeEventListener("resize", updateWidth);
+  });
+
+  return width;
+};

web/src/utils/pcm.ts

@@ -1,12 +1,12 @@
 export const convertToPCM16 = (float32Array: Float32Array): Uint8Array => {
-  const int16Buffer = new Int16Array(float32Array.length)
+  const int16Buffer = new Int16Array(float32Array.length);
   for (let i = 0; i < float32Array.length; i++) {
-    int16Buffer[i] = Math.max(-1, Math.min(1, float32Array[i])) * 0x7FFF
+    int16Buffer[i] = Math.max(-1, Math.min(1, float32Array[i])) * 0x7FFF;
   }
-  const buffer = new ArrayBuffer(int16Buffer.length * 2)
-  const view = new DataView(buffer)
+  const buffer = new ArrayBuffer(int16Buffer.length * 2);
+  const view = new DataView(buffer);
   for (let i = 0; i < int16Buffer.length; i++) {
-    view.setInt16(i * 2, int16Buffer[i], true)
+    view.setInt16(i * 2, int16Buffer[i], true);
   }
-  return new Uint8Array(buffer)
-}
+  return new Uint8Array(buffer);
+};

web/src/utils/title.ts

@@ -1,25 +1,24 @@
-import { useTitle } from "@vueuse/core"
+import { useTitle } from "@vueuse/core";
 
-const DEFAULT_TITLE = "Agent"
+const DEFAULT_TITLE = "Agent";
 
 const DEFAULT_DESCRIPTION = document
   .querySelector("meta[name='description']")
-  ?.getAttribute("content")
+  ?.getAttribute("content");
 
 export function setTitle(title?: string) {
-  useTitle().value = (title ? `${title} | ` : "") + DEFAULT_TITLE
+  useTitle().value = (title ? `${title} | ` : "") + DEFAULT_TITLE;
 }
 
 export function resetDescription() {
   document
     .querySelector("meta[name='description']")
-    ?.setAttribute("content", DEFAULT_DESCRIPTION!)
+    ?.setAttribute("content", DEFAULT_DESCRIPTION!);
 }
 
 export function setDescription(description?: string) {
-  if (!description)
-    return
+  if (!description) return;
   document
     .querySelector("meta[name='description']")
-    ?.setAttribute("content", `${description} | ${DEFAULT_TITLE}`)
+    ?.setAttribute("content", `${description} | ${DEFAULT_TITLE}`);
 }

web/src/utils/url.ts

@@ -1,15 +1,15 @@
 /** 直接当前页面跳转到指定url */
 export const jump = (url: string) => {
-  window.location.href = url
-}
+  window.location.href = url;
+};
 
 /** 在新标签页中跳转到指定url */
 export const jumpBlank = (url: string) => {
-  window.open(url, "_blank")
-}
+  window.open(url, "_blank");
+};
 
 /** 将对象转换为url查询字符串 */
 export const queryFormat = (query: Record<string, any>) => {
-  const params = new URLSearchParams(query)
-  return params.toString() ? `?${params}` : ""
-}
+  const params = new URLSearchParams(query);
+  return params.toString() ? `?${params}` : "";
+};

web/src/views/ChatLLMView.vue

@@ -0,0 +1,289 @@
+<script setup lang="ts">
+import type { SelectGroupOption, SelectOption } from "naive-ui";
+import type { Message } from "@/interfaces";
+import { throttle } from "lodash-es";
+import AIAvatar from "@/assets/ai_avatar.png";
+import {
+  ExclamationTriangleIcon,
+  microphone,
+  PaperAirplaneIcon,
+  TrashIcon
+} from "@/assets/Icons";
+import UserAvatar from "@/assets/user_avatar.jpg";
+import markdown from "@/components/markdown.vue";
+import { useAsrStore, useChatStore, useLayoutStore } from "@/stores";
+
+const chatStore = useChatStore();
+const { historyMessages, completing, modelList, modelInfo, thinking } =
+  storeToRefs(chatStore);
+const asrStore = useAsrStore();
+const { isRecording } = storeToRefs(asrStore);
+const layoutStore = useLayoutStore();
+const { hiddenLeftSidebar, simpleMode } = storeToRefs(layoutStore);
+
+const inputData = ref("");
+const scrollbarRef = ref<HTMLElement | null>(null);
+const options = ref<Array<SelectGroupOption | SelectOption>>([]);
+// NCollapse 组件的折叠状态
+const collapseActive = ref<string[]>(
+  historyMessages.value.map((msg, idx) => String(msg.id ?? idx))
+);
+
+// 过滤出type为chat的聊天消息
+const filteredMessages = computed(() =>
+  historyMessages.value.filter((msg) => msg.type === "chat" || !msg.type)
+);
+
+const getName = (msg: Message, idx: number) => String(msg.id ?? idx);
+
+// TODO: bugfix: 未能正确展开
+watch(
+  historyMessages,
+  (newVal, oldVal) => {
+    // 取所有name
+    const newNames = newVal.map((msg, idx) => getName(msg, idx));
+    const oldNames = oldVal ? oldVal.map((msg, idx) => getName(msg, idx)) : [];
+    // 找出新增的name
+    const addedNames = newNames.filter((name) => !oldNames.includes(name));
+    // 保留原有已展开项
+    const currentActive = collapseActive.value.filter((name) =>
+      newNames.includes(name)
+    );
+    // 新增的默认展开
+    collapseActive.value = [...currentActive, ...addedNames];
+  },
+  { immediate: true, deep: true }
+);
+
+// 处理折叠项的点击事件，切换折叠状态
+const handleItemHeaderClick = (name: string) => {
+  if (collapseActive.value.includes(name)) {
+    collapseActive.value = collapseActive.value.filter((n) => n !== name);
+  } else {
+    collapseActive.value.push(name);
+  }
+};
+
+// 处理选中模型的 ID
+const selectedModelId = computed({
+  get: () => modelInfo.value?.model_id ?? null,
+  set: (id: string | null) => {
+    for (const vendor of modelList.value) {
+      const found = vendor.models.find((model) => model.model_id === id);
+      if (found) {
+        modelInfo.value = found;
+        return;
+      }
+    }
+    modelInfo.value = null;
+  }
+});
+
+// 监听模型列表变化，更新选项
+watch(
+  () => modelList.value,
+  (newVal) => {
+    if (newVal) {
+      options.value = newVal.map((vendor) => ({
+        type: "group",
+        label: vendor.vendor,
+        key: vendor.vendor,
+        children: vendor.models.map((model) => ({
+          label: model.model_name,
+          value: model.model_id,
+          type: model.model_type
+        }))
+      }));
+
+      if (newVal.length > 0 && newVal[0].models.length > 0) {
+        modelInfo.value = newVal[0].models[0];
+      }
+    }
+  },
+  { immediate: true, deep: true }
+);
+
+// 发送消息
+const handleSendMessage = () => {
+  if (inputData.value.trim() === "") return;
+  chatStore.addMessageToHistory(inputData.value);
+  inputData.value = "";
+};
+
+// 开关语音输入
+const toggleRecording = throttle(() => {
+  if (isRecording.value) {
+    asrStore.stopRecording();
+  } else {
+    asrStore.startRecording();
+  }
+}, 500);
+
+watch(completing, (newVal) => {
+  if (newVal) {
+    nextTick(() => {
+      scrollbarRef.value?.scrollTo({ top: 99999, behavior: "smooth" });
+    });
+  }
+});
+
+onMounted(() => {
+  chatStore.getModelList();
+});
+</script>
+
+<template>
+  <div
+    class="p-8 !pr-4 h-full w-full flex flex-col gap-4 border-l-[24px] border-l-[#FAFAFA] transition-all ease-in-out text-base"
+    :class="{ '!border-l-0': hiddenLeftSidebar || simpleMode }"
+  >
+    <!-- 历史消息区 -->
+    <NScrollbar ref="scrollbarRef" class="flex-1 pr-4 relative">
+      <div class="flex items-start mb-4">
+        <span class="rounded-lg overflow-hidden !w-16 !min-w-16 !h-16">
+          <avatar :avatar="AIAvatar" />
+        </span>
+        <div class="text-base w-full max-w-full ml-2 flex flex-col items-start">
+          <span class="text-base font-bold mb-4">助手：</span>
+          <span class="text-base"
+            >你好，我是你的智能助手，请问有什么可以帮助你的吗？</span
+          >
+          <NDivider />
+        </div>
+      </div>
+      <!-- 默认消息↑ 历史消息↓ -->
+      <div
+        v-for="(msg, idx) in filteredMessages"
+        :key="idx"
+        class="flex items-start mb-4"
+      >
+        <!-- 头像↓ -->
+        <span
+          v-if="msg.role === 'user'"
+          class="rounded-lg overflow-hidden !w-16 !min-w-16 !h-16"
+        >
+          <avatar :avatar="UserAvatar" />
+        </span>
+        <span v-else class="rounded-lg overflow-hidden">
+          <avatar :avatar="AIAvatar" />
+        </span>
+        <!-- 头像↑ 名称↓ -->
+        <div class="text-base w-full max-w-full ml-2 flex flex-col items-start">
+          <span class="text-base font-bold">{{
+            msg.role === "user" ? "你：" : "助手："
+          }}</span>
+          <!-- 使用信息 -->
+          <div
+            v-if="msg.role !== 'user'"
+            class="text-[12px] text-[#7A7A7A] mb-[2px]"
+          >
+            Tokens: <span class="mr-1">{{ msg.usage?.total_tokens }}</span>
+          </div>
+          <div class="w-full max-w-full">
+            <NCollapse
+              v-if="msg.thinking?.trim()"
+              :expanded-names="collapseActive[idx]"
+            >
+              <NCollapseItem
+                :title="
+                  thinking && idx === historyMessages.length - 1
+                    ? '思考中...'
+                    : '已深度思考'
+                "
+                :name="getName(msg, idx)"
+                @item-header-click="
+                  () => handleItemHeaderClick(getName(msg, idx))
+                "
+              >
+                <div
+                  class="text-[#7A7A7A] mb-4 border-l-2 border-[#E5E5E5] ml-2 pl-2"
+                >
+                  <markdown :content="msg.thinking || ''" />
+                </div>
+              </NCollapseItem>
+            </NCollapse>
+            <!-- 内容↓ 思维链↑ -->
+            <markdown :content="msg.content || ''" />
+            <MessageTools :msg="msg" />
+            <NDivider />
+          </div>
+        </div>
+      </div>
+      <div
+        v-if="isRecording"
+        class="absolute inset-0 pointer-events-none flex items-center justify-center text-[#7A7A7A] text-2xl bg-white/80"
+      >
+        正在语音输入...
+      </div>
+    </NScrollbar>
+    <!-- 输入框 -->
+    <NInput
+      v-model:value="inputData"
+      type="textarea"
+      placeholder="输入内容，Enter发送，Shift+Enter换行"
+      :autosize="{
+        minRows: 3,
+        maxRows: 15
+      }"
+      @keyup.enter="handleSendMessage"
+    />
+    <!-- 操作区 -->
+    <div class="flex justify-between items-center gap-2">
+      <div class="flex items-center gap-2">
+        <NSelect
+          v-model:value="selectedModelId"
+          label-field="label"
+          value-field="value"
+          children-field="children"
+          filterable
+          :options="options"
+        />
+      </div>
+      <div class="flex items-center gap-2">
+        <NPopconfirm
+          :positive-button-props="{ type: 'error' }"
+          positive-text="清除"
+          negative-text="取消"
+          @positive-click="chatStore.clearHistoryMessages('chat')"
+          @negative-click="() => {}"
+        >
+          <template #icon>
+            <ExclamationTriangleIcon class="!w-6 !h-6 text-[#d03050]" />
+          </template>
+          <template #trigger>
+            <NButton :disabled="isRecording || completing" type="warning">
+              <template v-if="!simpleMode"> 清除历史 </template>
+              <TrashIcon
+                class="!w-4 !h-4"
+                :class="{
+                  'ml-1': !simpleMode
+                }"
+              />
+            </NButton>
+          </template>
+          <span>确定要清除历史消息吗？</span>
+        </NPopconfirm>
+
+        <NButton :disabled="completing" @click="toggleRecording">
+          <template v-if="!simpleMode">
+            {{ isRecording ? "停止输入" : "语音输入" }}
+          </template>
+          <microphone
+            class="!w-4 !h-4"
+            :class="{
+              'ml-1': !simpleMode
+            }"
+          />
+        </NButton>
+        <NButton
+          :disabled="isRecording"
+          :loading="completing"
+          @click="handleSendMessage"
+        >
+          发送
+          <PaperAirplaneIcon class="!w-4 !h-4 ml-1" />
+        </NButton>
+      </div>
+    </div>
+  </div>
+</template>

web/src/views/CommunityView.vue

@@ -1,130 +0,0 @@
-<script setup lang="ts">
-import type { SelectGroupOption, SelectOption } from "naive-ui";
-import { ExclamationTriangleIcon, microphone, PaperAirplaneIcon, TrashIcon } from "@/assets/Icons";
-import markdown from "@/components/markdown.vue";
-import { useAsrStore, useChatStore } from "@/stores";
-
-const chatStore = useChatStore();
-const asrStore = useAsrStore();
-
-const { historyMessages, completing, modelList, modelInfo } = storeToRefs(chatStore);
-const { isRecording } = storeToRefs(asrStore);
-
-const inputData = ref("");
-
-const options = ref<Array<SelectGroupOption | SelectOption>>([]);
-
-// 处理选中模型的 ID
-const selectedModelId = computed({
-  get: () => modelInfo.value?.model_id ?? null,
-  set: (id: string | null) => {
-    for (const vendor of modelList.value) {
-      const found = vendor.models.find(model => model.model_id === id);
-      if (found) {
-        modelInfo.value = found;
-        return;
-      }
-    }
-    modelInfo.value = null;
-  },
-});
-
-// 监听模型列表变化，更新选项
-watch(() => modelList.value, (newVal) => {
-  if (newVal) {
-    options.value = newVal.map(vendor => ({
-      type: "group",
-      label: vendor.vendor,
-      key: vendor.vendor,
-      children: vendor.models.map(model => ({
-        label: model.model_name,
-        value: model.model_id,
-        type: model.model_type,
-      })),
-    }));
-
-    if (newVal.length > 0 && newVal[0].models.length > 0) {
-      modelInfo.value = newVal[0].models[0];
-    }
-    console.log("Options updated:", options.value);
-  }
-}, { immediate: true, deep: true });
-
-// 发送消息
-const handleSendMessage = () => {
-  if (inputData.value.trim() === "")
-    return;
-  chatStore.addMessageToHistory(inputData.value);
-  inputData.value = "";
-};
-
-// 开关语音输入
-const toggleRecording = () => {
-  if (isRecording.value) {
-    asrStore.stopRecording();
-  }
-  else {
-    asrStore.startRecording();
-  }
-};
-
-onMounted(() => {
-  chatStore.getModelList();
-});
-</script>
-
-<template>
-  <div class="p-8 !pr-4 h-full w-full flex flex-col gap-4 border-l-[24px] border-l-[#FAFAFA] text-base">
-    <NScrollbar class="flex-1 pr-4">
-      <div class="flex items-start mb-4">
-        <span class="text-base w-14 min-w-14">助手：</span>
-        <NTag type="success" class="text-base max-w-full !h-auto">
-          <span class="text-base">你好，我是你的智能助手，请问有什么可以帮助你的吗？</span>
-        </NTag>
-      </div>
-      <div v-for="(msg, idx) in historyMessages" :key="idx" class="flex items-start mb-4">
-        <span v-if="msg.role === 'user'" class="text-base w-14 min-w-14">你：</span>
-        <span v-else class="text-base w-14 min-w-14">助手：</span>
-        <NTag :type="msg.role === 'user' ? 'info' : 'success'" class="max-w-full !h-auto">
-          <markdown :content="msg.content || ''" />
-        </NTag>
-      </div>
-    </NScrollbar>
-    <NInput v-model:value="inputData" type="textarea" placeholder="在这里输入消息" @keyup.enter="handleSendMessage" />
-    <!-- 操作区 -->
-    <div class="flex justify-between items-center gap-2">
-      <div class="flex items-center gap-2">
-        <NSelect
-          v-model:value="selectedModelId" label-field="label" value-field="value" children-field="children"
-          filterable :options="options"
-        />
-      </div>
-      <div class="flex items-center gap-2">
-        <NPopconfirm
-          :positive-button-props="{ type: 'error' }" positive-text="清除" negative-text="取消"
-          @positive-click="chatStore.clearHistoryMessages" @negative-click="() => { }"
-        >
-          <template #icon>
-            <ExclamationTriangleIcon class="!w-6 !h-6 text-[#d03050]" />
-          </template>
-          <template #trigger>
-            <NButton :disabled="isRecording || completing" type="warning">
-              清除历史
-              <TrashIcon class="!w-4 !h-4 ml-1" />
-            </NButton>
-          </template>
-          <span>确定要清除历史消息吗？</span>
-        </NPopconfirm>
-
-        <NButton :disabled="completing" @click="toggleRecording">
-          {{ isRecording ? "停止输入" : "语音输入" }}
-          <microphone class="!w-4 !h-4 ml-1" />
-        </NButton>
-        <NButton :disabled="isRecording" :loading="completing" @click="handleSendMessage">
-          发送
-          <PaperAirplaneIcon class="!w-4 !h-4 ml-1" />
-        </NButton>
-      </div>
-    </div>
-  </div>
-</template>

web/src/views/VoiceView.vue

@@ -0,0 +1,261 @@
+<script setup lang="ts">
+import type { SelectGroupOption, SelectOption } from "naive-ui";
+import type { Message } from "@/interfaces";
+import { throttle } from "lodash-es";
+import AIAvatar from "@/assets/ai_avatar.png";
+import { ExclamationTriangleIcon, microphone, TrashIcon } from "@/assets/Icons";
+import UserAvatar from "@/assets/user_avatar.jpg";
+import markdown from "@/components/markdown.vue";
+import { useAsrStore, useChatStore, useLayoutStore } from "@/stores";
+
+const chatStore = useChatStore();
+const { historyMessages, completing, speakerList, speakerInfo, thinking } =
+  storeToRefs(chatStore);
+const asrStore = useAsrStore();
+const { isRecording } = storeToRefs(asrStore);
+const layoutStore = useLayoutStore();
+const { hiddenLeftSidebar, simpleMode } = storeToRefs(layoutStore);
+
+const scrollbarRef = ref<HTMLElement | null>(null);
+const options = ref<Array<SelectGroupOption | SelectOption>>([]);
+// NCollapse 组件的折叠状态
+const collapseActive = ref<string[]>(
+  historyMessages.value.map((msg, idx) => String(msg.id ?? idx))
+);
+
+// 过滤出type为voice的聊天消息
+const filteredMessages = computed(() =>
+  historyMessages.value.filter((msg) => msg.type === "voice")
+);
+
+const getName = (msg: Message, idx: number) => String(msg.id ?? idx);
+
+// TODO: bugfix: 未能正确展开
+watch(
+  historyMessages,
+  (newVal, oldVal) => {
+    // 取所有name
+    const newNames = newVal.map((msg, idx) => getName(msg, idx));
+    const oldNames = oldVal ? oldVal.map((msg, idx) => getName(msg, idx)) : [];
+    // 找出新增的name
+    const addedNames = newNames.filter((name) => !oldNames.includes(name));
+    // 保留原有已展开项
+    const currentActive = collapseActive.value.filter((name) =>
+      newNames.includes(name)
+    );
+    // 新增的默认展开
+    collapseActive.value = [...currentActive, ...addedNames];
+  },
+  { immediate: true, deep: true }
+);
+
+// 处理折叠项的点击事件，切换折叠状态
+const handleItemHeaderClick = (name: string) => {
+  if (collapseActive.value.includes(name)) {
+    collapseActive.value = collapseActive.value.filter((n) => n !== name);
+  } else {
+    collapseActive.value.push(name);
+  }
+};
+
+// 处理选中speaker的 ID
+const selectedSpeakerId = computed({
+  get: () => speakerInfo.value?.speaker_id ?? null,
+  set: (id: string | null) => {
+    for (const category of speakerList.value) {
+      const found = category.speakers.find(
+        (speaker) => speaker.speaker_id === id
+      );
+      if (found) {
+        speakerInfo.value = found;
+        return;
+      }
+    }
+    speakerInfo.value = null;
+  }
+});
+
+// 监听speaker列表变化，更新选项
+watch(
+  () => speakerList.value,
+  (newVal) => {
+    if (newVal) {
+      options.value = newVal.map((category) => ({
+        type: "group",
+        label: category.category,
+        key: category.category,
+        children: category.speakers.map((speaker) => ({
+          label: speaker.speaker_name,
+          value: speaker.speaker_id,
+          language: speaker.language,
+          platforms: speaker.platforms
+        }))
+      }));
+
+      // 默认选择第一个speaker
+      if (newVal.length > 0 && newVal[0].speakers.length > 0) {
+        speakerInfo.value = newVal[0].speakers[0];
+      }
+    }
+  },
+  { immediate: true, deep: true }
+);
+
+// 开关语音输入
+const toggleRecording = throttle(() => {
+  if (isRecording.value) {
+    asrStore.stopRecording();
+  } else {
+    asrStore.startRecording();
+  }
+}, 500);
+
+watch(completing, (newVal) => {
+  if (newVal) {
+    nextTick(() => {
+      scrollbarRef.value?.scrollTo({ top: 99999, behavior: "smooth" });
+    });
+  }
+});
+
+onMounted(() => {
+  chatStore.getSpeakerList();
+});
+</script>
+
+<template>
+  <div
+    class="p-8 !pr-4 h-full w-full flex flex-col gap-4 border-l-[24px] border-l-[#FAFAFA] transition-all ease-in-out text-base"
+    :class="{ '!border-l-0': hiddenLeftSidebar || simpleMode }"
+  >
+    <!-- 历史消息区 -->
+    <NScrollbar ref="scrollbarRef" class="flex-1 pr-4 relative">
+      <div class="flex items-start mb-4">
+        <span class="rounded-lg overflow-hidden !w-16 !min-w-16 !h-16">
+          <avatar :avatar="AIAvatar" />
+        </span>
+        <div class="text-base w-full max-w-full ml-2 flex flex-col items-start">
+          <span class="text-base font-bold mb-4">助手：</span>
+          <span class="text-base"
+            >你好，我是你的智能助手，请问有什么可以帮助你的吗？</span
+          >
+          <NDivider />
+        </div>
+      </div>
+      <!-- 默认消息↑ 历史消息↓ -->
+      <div
+        v-for="(msg, idx) in filteredMessages"
+        :key="idx"
+        class="flex items-start mb-4"
+      >
+        <!-- 头像↓ -->
+        <span
+          v-if="msg.role === 'user'"
+          class="rounded-lg overflow-hidden !w-16 !min-w-16 !h-16"
+        >
+          <avatar :avatar="UserAvatar" />
+        </span>
+        <span v-else class="rounded-lg overflow-hidden">
+          <avatar :avatar="AIAvatar" />
+        </span>
+        <!-- 头像↑ 名称↓ -->
+        <div class="text-base w-full max-w-full ml-2 flex flex-col items-start">
+          <span class="text-base font-bold">{{
+            msg.role === "user" ? "你：" : "助手："
+          }}</span>
+          <!-- 使用信息 -->
+          <div
+            v-if="msg.role !== 'user'"
+            class="text-[12px] text-[#7A7A7A] mb-[2px]"
+          >
+            Tokens: <span class="mr-1">{{ msg.usage?.total_tokens }}</span>
+          </div>
+          <div class="w-full max-w-full">
+            <NCollapse
+              v-if="msg.thinking?.trim()"
+              :expanded-names="collapseActive[idx]"
+            >
+              <NCollapseItem
+                :title="
+                  thinking && idx === historyMessages.length - 1
+                    ? '思考中...'
+                    : '已深度思考'
+                "
+                :name="getName(msg, idx)"
+                @item-header-click="
+                  () => handleItemHeaderClick(getName(msg, idx))
+                "
+              >
+                <div
+                  class="text-[#7A7A7A] mb-4 border-l-2 border-[#E5E5E5] ml-2 pl-2"
+                >
+                  <markdown :content="msg.thinking || ''" />
+                </div>
+              </NCollapseItem>
+            </NCollapse>
+            <!-- 内容↓ 思维链↑ -->
+            <markdown :content="msg.content || ''" />
+            <MessageTools :msg="msg" />
+            <NDivider />
+          </div>
+        </div>
+      </div>
+      <div
+        v-if="isRecording"
+        class="absolute inset-0 pointer-events-none flex items-center justify-center text-[#7A7A7A] text-2xl bg-white/80"
+      >
+        正在语音输入...
+      </div>
+    </NScrollbar>
+    <!-- 操作区 -->
+    <div class="flex justify-between items-center gap-2">
+      <div class="flex items-center gap-2">
+        <NSelect
+          v-model:value="selectedSpeakerId"
+          label-field="label"
+          value-field="value"
+          children-field="children"
+          filterable
+          :options="options"
+        />
+      </div>
+      <div class="flex items-center gap-2">
+        <NPopconfirm
+          :positive-button-props="{ type: 'error' }"
+          positive-text="清除"
+          negative-text="取消"
+          @positive-click="chatStore.clearHistoryMessages('voice')"
+          @negative-click="() => {}"
+        >
+          <template #icon>
+            <ExclamationTriangleIcon class="!w-6 !h-6 text-[#d03050]" />
+          </template>
+          <template #trigger>
+            <NButton :disabled="isRecording || completing" type="warning">
+              <template v-if="!simpleMode"> 清除历史 </template>
+              <TrashIcon
+                class="!w-4 !h-4"
+                :class="{
+                  'ml-1': !simpleMode
+                }"
+              />
+            </NButton>
+          </template>
+          <span>确定要清除历史消息吗？</span>
+        </NPopconfirm>
+
+        <NButton :disabled="completing" @click="toggleRecording">
+          <template v-if="!simpleMode">
+            {{ isRecording ? "停止输入" : "语音输入" }}
+          </template>
+          <microphone
+            class="!w-4 !h-4"
+            :class="{
+              'ml-1': !simpleMode
+            }"
+          />
+        </NButton>
+      </div>
+    </div>
+  </div>
+</template>

web/tsconfig.json

@@ -2,13 +2,22 @@
   "compilerOptions": {
     "baseUrl": ".",
     "paths": {
-      "@/*": ["./src/*"]
+      "@/*": [
+        "./src/*"
+      ]
     },
-    "types": ["vite-svg-loader"]
+    "types": [
+      "vite-svg-loader",
+      "lodash"
+    ]
   },
   "references": [
-    { "path": "./tsconfig.app.json" },
-    { "path": "./tsconfig.node.json" }
+    {
+      "path": "./tsconfig.app.json"
+    },
+    {
+      "path": "./tsconfig.node.json"
+    }
   ],
   "files": []
 }

web/vite.config.ts

@@ -48,4 +48,7 @@ export default defineConfig({
       "@": fileURLToPath(new URL("./src", import.meta.url)),
     },
   },
+  esbuild: {
+    treeShaking: true,
+  },
 });