tingwu

backend/.env.example

@@ -0,0 +1,9 @@
+# Aliyun OSS Configuration
+ALIYUN_ACCESS_KEY_ID=LTAI5tE62GW8MKyoEaotzxXk
+ALIYUN_ACCESS_KEY_SECRET=Zdzqo1fgj57DxxioXOotNKhJdSfVQW
+ALIYUN_OSS_ENDPOINT=https://oss-cn-shanghai.aliyuncs.com
+ALIYUN_OSS_BUCKET_NAME=tangledup-ai-staging
+ALIYUN_OSS_INTERNAL_ENDPOINT=https://oss-cn-shanghai-internal.aliyuncs.com
+
+# Aliyun Tingwu Configuration
+ALIYUN_TINGWU_APP_KEY=

backend/ai_services/admin.py

@@ -0,0 +1,11 @@
+from django.contrib import admin
+from django.contrib.admin import ModelAdmin
+from unfold.admin import ModelAdmin as UnfoldModelAdmin
+from .models import TranscriptionTask
+
+@admin.register(TranscriptionTask)
+class TranscriptionTaskAdmin(UnfoldModelAdmin):
+    list_display = ['id', 'status', 'task_id', 'created_at']
+    list_filter = ['status', 'created_at']
+    search_fields = ['id', 'task_id', 'transcription', 'summary']
+    readonly_fields = ['id', 'created_at', 'updated_at', 'task_id']

backend/ai_services/apps.py

@@ -0,0 +1,5 @@
+from django.apps import AppConfig
+
+
+class AiServicesConfig(AppConfig):
+    name = 'ai_services'

backend/ai_services/migrations/0001_initial.py

@@ -0,0 +1,34 @@
+# Generated by Django 6.0.1 on 2026-03-11 05:11
+
+import uuid
+from django.db import migrations, models
+
+
+class Migration(migrations.Migration):
+
+    initial = True
+
+    dependencies = [
+    ]
+
+    operations = [
+        migrations.CreateModel(
+            name='TranscriptionTask',
+            fields=[
+                ('id', models.UUIDField(default=uuid.uuid4, editable=False, primary_key=True, serialize=False)),
+                ('file_url', models.URLField(max_length=1024, verbose_name='文件链接')),
+                ('task_id', models.CharField(blank=True, max_length=100, null=True, verbose_name='听悟任务ID')),
+                ('status', models.CharField(choices=[('PENDING', '等待中'), ('PROCESSING', '处理中'), ('SUCCEEDED', '成功'), ('FAILED', '失败')], default='PENDING', max_length=20, verbose_name='状态')),
+                ('transcription', models.TextField(blank=True, null=True, verbose_name='逐字稿')),
+                ('summary', models.TextField(blank=True, null=True, verbose_name='AI总结')),
+                ('error_message', models.TextField(blank=True, null=True, verbose_name='错误信息')),
+                ('created_at', models.DateTimeField(auto_now_add=True, verbose_name='创建时间')),
+                ('updated_at', models.DateTimeField(auto_now=True, verbose_name='更新时间')),
+            ],
+            options={
+                'verbose_name': '转写任务',
+                'verbose_name_plural': '转写任务',
+                'ordering': ['-created_at'],
+            },
+        ),
+    ]

backend/ai_services/migrations/0002_transcriptiontask_evaluation_transcriptiontask_score.py

@@ -0,0 +1,23 @@
+# Generated by Django 6.0.1 on 2026-03-11 05:12
+
+from django.db import migrations, models
+
+
+class Migration(migrations.Migration):
+
+    dependencies = [
+        ('ai_services', '0001_initial'),
+    ]
+
+    operations = [
+        migrations.AddField(
+            model_name='transcriptiontask',
+            name='evaluation',
+            field=models.TextField(blank=True, null=True, verbose_name='AI评语'),
+        ),
+        migrations.AddField(
+            model_name='transcriptiontask',
+            name='score',
+            field=models.IntegerField(blank=True, help_text='基于转写内容的评分', null=True, verbose_name='AI评分'),
+        ),
+    ]

backend/ai_services/models.py

@@ -0,0 +1,35 @@
+import uuid
+from django.db import models
+from django.utils.translation import gettext_lazy as _
+
+class TranscriptionTask(models.Model):
+    class Status(models.TextChoices):
+        PENDING = 'PENDING', _('等待中')
+        PROCESSING = 'PROCESSING', _('处理中')
+        SUCCEEDED = 'SUCCEEDED', _('成功')
+        FAILED = 'FAILED', _('失败')
+
+    id = models.UUIDField(primary_key=True, default=uuid.uuid4, editable=False)
+    file_url = models.URLField(verbose_name=_('文件链接'), max_length=1024)
+    task_id = models.CharField(verbose_name=_('听悟任务ID'), max_length=100, blank=True, null=True)
+    status = models.CharField(
+        verbose_name=_('状态'),
+        max_length=20,
+        choices=Status.choices,
+        default=Status.PENDING
+    )
+    transcription = models.TextField(verbose_name=_('逐字稿'), blank=True, null=True)
+    summary = models.TextField(verbose_name=_('AI总结'), blank=True, null=True)
+    score = models.IntegerField(verbose_name=_('AI评分'), blank=True, null=True, help_text=_('基于转写内容的评分'))
+    evaluation = models.TextField(verbose_name=_('AI评语'), blank=True, null=True)
+    error_message = models.TextField(verbose_name=_('错误信息'), blank=True, null=True)
+    created_at = models.DateTimeField(verbose_name=_('创建时间'), auto_now_add=True)
+    updated_at = models.DateTimeField(verbose_name=_('更新时间'), auto_now=True)
+
+    class Meta:
+        verbose_name = _('转写任务')
+        verbose_name_plural = _('转写任务')
+        ordering = ['-created_at']
+
+    def __str__(self):
+        return f"{self.id} - {self.get_status_display()}"

backend/ai_services/serializers.py

@@ -0,0 +1,8 @@
+from rest_framework import serializers
+from .models import TranscriptionTask
+
+class TranscriptionTaskSerializer(serializers.ModelSerializer):
+    class Meta:
+        model = TranscriptionTask
+        fields = ['id', 'file_url', 'task_id', 'status', 'transcription', 'summary', 'error_message', 'created_at', 'updated_at']
+        read_only_fields = ['id', 'file_url', 'task_id', 'status', 'transcription', 'summary', 'error_message', 'created_at', 'updated_at']

backend/ai_services/services.py

@@ -0,0 +1,115 @@
+import json
+import logging
+import time
+import oss2
+from aliyunsdkcore.client import AcsClient
+from aliyunsdkcore.acs_exception.exceptions import ClientException, ServerException
+# 尝试导入最新的 API 版本，如果有问题可能需要调整
+try:
+    from aliyunsdktingwu.request.v20230930 import CreateTaskRequest, GetTaskInfoRequest
+except ImportError:
+    # Fallback or error handling if version differs
+    pass
+    
+from django.conf import settings
+
+logger = logging.getLogger(__name__)
+
+class AliyunTingwuService:
+    def __init__(self):
+        self.access_key_id = settings.ALIYUN_ACCESS_KEY_ID
+        self.access_key_secret = settings.ALIYUN_ACCESS_KEY_SECRET
+        self.oss_bucket_name = settings.ALIYUN_OSS_BUCKET_NAME
+        self.oss_endpoint = settings.ALIYUN_OSS_ENDPOINT
+        self.tingwu_app_key = settings.ALIYUN_TINGWU_APP_KEY
+        self.region_id = "cn-beijing" # 听悟服务主要在北京
+
+        # 初始化 OSS Bucket
+        if self.access_key_id and self.access_key_secret and self.oss_endpoint:
+             auth = oss2.Auth(self.access_key_id, self.access_key_secret)
+             self.bucket = oss2.Bucket(auth, self.oss_endpoint, self.oss_bucket_name)
+        else:
+            self.bucket = None
+            logger.warning("Aliyun OSS configuration missing.")
+        
+        # 初始化听悟 Client
+        if self.access_key_id and self.access_key_secret:
+            self.client = AcsClient(
+                self.access_key_id,
+                self.access_key_secret,
+                self.region_id
+            )
+        else:
+            self.client = None
+            logger.warning("Aliyun AccessKey configuration missing.")
+
+    def upload_to_oss(self, file_obj, file_name):
+        """
+        上传文件到 OSS 并返回带签名的 URL (有效期 3 小时)
+        """
+        if not self.bucket:
+            raise Exception("OSS Client not initialized")
+            
+        try:
+            # 上传文件
+            # file_obj 应该是打开的文件对象或字节流
+            self.bucket.put_object(file_name, file_obj)
+            
+            # 生成签名 URL，有效期 3 小时 (3600 * 3)
+            url = self.bucket.sign_url('GET', file_name, 3600 * 3)
+            return url
+        except Exception as e:
+            logger.error(f"OSS Upload failed: {e}")
+            raise e
+
+    def create_transcription_task(self, file_url, language="cn"):
+        """
+        创建听悟转写任务
+        """
+        if not self.client:
+             raise Exception("Tingwu Client not initialized")
+
+        request = CreateTaskRequest.CreateTaskRequest()
+        request.set_AppKey(self.tingwu_app_key)
+        
+        # 配置 Input
+        input_param = {
+            "FileUrl": file_url,
+            "SourceLanguage": language,
+            "TaskKey": "transcription_task" 
+        }
+        request.set_Input(json.dumps(input_param))
+        
+        # 配置 Parameters (开启自动章节和摘要)
+        parameters = {
+            "Transcoding": {
+                "TargetAudioFormat": "mp3"
+            },
+            "AutoChaptersEnabled": True,
+            "SummarizationEnabled": True 
+        }
+        request.set_Parameters(json.dumps(parameters))
+
+        try:
+            response = self.client.do_action_with_exception(request)
+            return json.loads(response)
+        except (ClientException, ServerException) as e:
+            logger.error(f"Tingwu CreateTask failed: {e}")
+            raise e
+
+    def get_task_info(self, task_id):
+        """
+        查询任务状态和结果
+        """
+        if not self.client:
+             raise Exception("Tingwu Client not initialized")
+
+        request = GetTaskInfoRequest.GetTaskInfoRequest()
+        request.set_TaskId(task_id)
+        
+        try:
+            response = self.client.do_action_with_exception(request)
+            return json.loads(response)
+        except (ClientException, ServerException) as e:
+            logger.error(f"Tingwu GetTaskInfo failed: {e}")
+            raise e

backend/ai_services/tests.py

@@ -0,0 +1,3 @@
+from django.test import TestCase
+
+# Create your tests here.

backend/ai_services/urls.py

@@ -0,0 +1,11 @@
+from django.urls import path, include
+from rest_framework.routers import DefaultRouter
+from .views import TranscriptionTaskViewSet, tingwu_callback
+
+router = DefaultRouter()
+router.register(r'transcriptions', TranscriptionTaskViewSet)
+
+urlpatterns = [
+    path('callback/', tingwu_callback, name='tingwu-callback'),
+    path('', include(router.urls)),
+]

backend/ai_services/views.py

@@ -0,0 +1,181 @@
+import logging
+import uuid
+from rest_framework import viewsets, status
+from rest_framework.decorators import action, api_view, permission_classes, parser_classes
+from rest_framework.response import Response
+from rest_framework.parsers import MultiPartParser, FormParser, JSONParser
+from rest_framework.permissions import AllowAny
+from django.conf import settings
+from .models import TranscriptionTask
+from .serializers import TranscriptionTaskSerializer
+from .services import AliyunTingwuService
+
+logger = logging.getLogger(__name__)
+
+@api_view(['POST'])
+@permission_classes([AllowAny])
+def tingwu_callback(request):
+    """
+    处理阿里云听悟的回调消息
+    """
+    data = request.data
+    logger.info(f"收到听悟回调: {data}")
+    
+    # 1. 处理连通性测试消息
+    # 格式: {"Code": "0", "Data": {"Test": "..."}, "Message": "success", "RequestId": "..."}
+    if isinstance(data, dict) and 'Data' in data and 'Test' in data['Data']:
+        logger.info("收到听悟连通性测试请求")
+        return Response({'message': 'success'}, status=status.HTTP_200_OK)
+
+    # 2. 处理任务完成消息 (根据实际文档或后续调试完善)
+    # 通常会包含 TaskId 和 Status
+    # 注意：阿里云听悟回调的结构可能在 Header 或 Body 中不同，需根据实际情况调整
+    # 这里是一个通用的处理逻辑
+    task_id = data.get('TaskId')
+    task_status = data.get('Status')
+    
+    if task_id:
+        try:
+            task = TranscriptionTask.objects.filter(task_id=task_id).first()
+            if task:
+                if task_status == 'COMPLETE':
+                    logger.info(f"任务 {task_id} 完成，等待下一次查询刷新")
+                    # 可以在这里直接调用 get_task_info 刷新数据，但要注意超时
+                elif task_status == 'FAILED':
+                    task.status = TranscriptionTask.Status.FAILED
+                    task.error_message = data.get('StatusText', 'Callback reported failure')
+                    task.save()
+            else:
+                logger.warning(f"回调收到未知任务ID: {task_id}")
+        except Exception as e:
+            logger.error(f"处理回调异常: {e}")
+
+    return Response({'message': 'success'}, status=status.HTTP_200_OK)
+
+class TranscriptionTaskViewSet(viewsets.ModelViewSet):
+    queryset = TranscriptionTask.objects.all()
+    serializer_class = TranscriptionTaskSerializer
+    parser_classes = (MultiPartParser, FormParser)
+
+    def create(self, request, *args, **kwargs):
+        """
+        上传音频文件并创建听悟转写任务
+        """
+        file_obj = request.FILES.get('file')
+        if not file_obj:
+            return Response({'error': '未提供文件'}, status=status.HTTP_400_BAD_REQUEST)
+
+        service = AliyunTingwuService()
+        if not service.bucket or not service.client:
+            return Response({'error': '阿里云服务未配置'}, status=status.HTTP_503_SERVICE_UNAVAILABLE)
+
+        try:
+            # 1. 上传文件到 OSS
+            file_extension = file_obj.name.split('.')[-1]
+            file_name = f"transcription/{uuid.uuid4()}.{file_extension}"
+            
+            # 使用服务上传
+            oss_url = service.upload_to_oss(file_obj, file_name)
+            
+            # 2. 创建数据库记录
+            task_record = TranscriptionTask.objects.create(
+                file_url=oss_url,
+                status=TranscriptionTask.Status.PENDING
+            )
+
+            # 3. 调用听悟接口创建任务
+            try:
+                tingwu_response = service.create_transcription_task(oss_url)
+                task_id = tingwu_response.get('TaskId')
+                
+                if task_id:
+                    task_record.task_id = task_id
+                    task_record.status = TranscriptionTask.Status.PROCESSING
+                    task_record.save()
+                else:
+                    task_record.status = TranscriptionTask.Status.FAILED
+                    task_record.error_message = "未能获取 TaskId"
+                    task_record.save()
+                    return Response({'error': '未能获取 TaskId'}, status=status.HTTP_500_INTERNAL_SERVER_ERROR)
+
+            except Exception as e:
+                task_record.status = TranscriptionTask.Status.FAILED
+                task_record.error_message = str(e)
+                task_record.save()
+                logger.error(f"创建听悟任务失败: {e}")
+                return Response({'error': f"创建听悟任务失败: {str(e)}"}, status=status.HTTP_500_INTERNAL_SERVER_ERROR)
+
+            serializer = self.get_serializer(task_record)
+            return Response(serializer.data, status=status.HTTP_201_CREATED)
+
+        except Exception as e:
+            logger.error(f"处理上传请求失败: {e}")
+            return Response({'error': str(e)}, status=status.HTTP_500_INTERNAL_SERVER_ERROR)
+
+    @action(detail=True, methods=['get'])
+    def refresh_status(self, request, pk=None):
+        """
+        刷新任务状态并获取结果
+        """
+        task = self.get_object()
+        
+        # 如果任务已经完成或失败，直接返回当前状态
+        if task.status in [TranscriptionTask.Status.SUCCEEDED, TranscriptionTask.Status.FAILED]:
+            serializer = self.get_serializer(task)
+            return Response(serializer.data)
+
+        if not task.task_id:
+            return Response({'error': '任务ID不存在'}, status=status.HTTP_400_BAD_REQUEST)
+
+        service = AliyunTingwuService()
+        try:
+            result = service.get_task_info(task.task_id)
+            task_status = result.get('TaskStatus')
+            
+            if task_status == 'COMPLETE':
+                task.status = TranscriptionTask.Status.SUCCEEDED
+                
+                # 解析结果
+                task_result = result.get('Result', {})
+                
+                # 提取逐字稿
+                sentences = task_result.get('Transcription', {}).get('Sentences', [])
+                full_text = " ".join([s.get('Text', '') for s in sentences])
+                task.transcription = full_text
+                
+                # 提取总结
+                # 总结结果结构可能因配置不同而异，这里尝试获取摘要
+                summarization = task_result.get('Summarization', {})
+                # 听悟的总结通常在 Summarization.Text 或类似字段
+                # 如果是章节摘要，可能在 Chapters 中
+                # 假设是全文摘要
+                if 'Text' in summarization:
+                    task.summary = summarization['Text']
+                elif 'Headline' in summarization:
+                    task.summary = summarization['Headline']
+                else:
+                    # 尝试从章节摘要中提取
+                    chapters = task_result.get('Chapters', [])
+                    summary_parts = []
+                    for chapter in chapters:
+                        if 'Headline' in chapter:
+                            summary_parts.append(chapter['Headline'])
+                        if 'Summary' in chapter:
+                            summary_parts.append(chapter['Summary'])
+                    task.summary = "\n".join(summary_parts)
+
+                task.save()
+
+            elif task_status == 'FAILED':
+                task.status = TranscriptionTask.Status.FAILED
+                task.error_message = result.get('TaskStatusText', 'Unknown error')
+                task.save()
+            
+            # 其他状态 (PENDING, RUNNING) 不做更改
+            
+            serializer = self.get_serializer(task)
+            return Response(serializer.data)
+            
+        except Exception as e:
+            logger.error(f"刷新任务状态失败: {e}")
+            return Response({'error': str(e)}, status=status.HTTP_500_INTERNAL_SERVER_ERROR)

backend/config/settings.py

@@ -12,10 +12,14 @@ https://docs.djangoproject.com/en/6.0/ref/settings/
 
 import os
 from pathlib import Path
+from dotenv import load_dotenv
 
 # Build paths inside the project like this: BASE_DIR / 'subdir'.
 BASE_DIR = Path(__file__).resolve().parent.parent
 
+# Load .env file
+load_dotenv(BASE_DIR / '.env')
+
 
 # Quick-start development settings - unsuitable for production
 # See https://docs.djangoproject.com/en/6.0/howto/deployment/checklist/
@@ -48,6 +52,7 @@ INSTALLED_APPS = [
     'shop',
     'community',
     'competition',
+    'ai_services',
 ]
 
 MIDDLEWARE = [
@@ -348,6 +353,17 @@ UNFOLD = {
                     },
                 ],
             },
+            {
+                "title": "AI 听悟",
+                "separator": True,
+                "items": [
+                    {
+                        "title": "转写与总结任务",
+                        "icon": "record_voice_over",
+                        "link": reverse_lazy("admin:ai_services_transcriptiontask_changelist"),
+                    },
+                ],
+            },
             {
                 "title": "系统配置",
                 "separator": True,
@@ -390,3 +406,11 @@ LOGGING = {
         'level': 'INFO',
     },
 }
+
+# 阿里云配置
+ALIYUN_ACCESS_KEY_ID = os.environ.get('ALIYUN_ACCESS_KEY_ID', '')
+ALIYUN_ACCESS_KEY_SECRET = os.environ.get('ALIYUN_ACCESS_KEY_SECRET', '')
+ALIYUN_OSS_BUCKET_NAME = os.environ.get('ALIYUN_OSS_BUCKET_NAME', '')
+ALIYUN_OSS_ENDPOINT = os.environ.get('ALIYUN_OSS_ENDPOINT', 'oss-cn-shanghai.aliyuncs.com')
+ALIYUN_OSS_INTERNAL_ENDPOINT = os.environ.get('ALIYUN_OSS_INTERNAL_ENDPOINT', '')
+ALIYUN_TINGWU_APP_KEY = os.environ.get('ALIYUN_TINGWU_APP_KEY', '')  # 听悟AppKey

backend/config/urls.py

@@ -9,6 +9,7 @@ urlpatterns = [
     path('api/', include('shop.urls')),
     path('api/community/', include('community.urls')),
     path('api/competition/', include('competition.urls')),
+    path('api/ai/', include('ai_services.urls')),
     
     # Swagger文档路由
     path('api/schema/', SpectacularAPIView.as_view(), name='schema'),

backend/requirements.txt

@@ -24,3 +24,7 @@ django-filter
 django-admin-sortable2
 openpyxl
 
+aliyun-python-sdk-core==2.16.0
+aliyun-python-sdk-tingwu==1.0.7
+oss2==2.19.1
+python-dotenv