ai chat init

backend/ai/chat.py

@@ -0,0 +1,26 @@
+from channels.generic.websocket import AsyncWebsocketConsumer
+import json
+from ai.langchain_client import get_ai_reply_stream
+from ai.utils import get_first_available_ai_config
+
+
+class ChatConsumer(AsyncWebsocketConsumer):
+    async def connect(self):
+        await self.accept()
+
+    async def disconnect(self, close_code):
+        pass
+
+    async def receive(self, text_data):
+        data = json.loads(text_data)
+        user_message = data.get("message", "")
+
+        model, api_key, api_base = await get_first_available_ai_config()
+
+        async def send_chunk(chunk):
+            await self.send(text_data=json.dumps({"is_streaming": True, "message": chunk}))
+
+        await get_ai_reply_stream(user_message, send_chunk, model_name=model, api_key=api_key, api_base=api_base)
+
+        # 结束标记
+        await self.send(text_data=json.dumps({"done": True}))

backend/ai/langchain_client.py

@@ -0,0 +1,25 @@
+from langchain.schema import HumanMessage
+
+from langchain_core.callbacks import AsyncCallbackHandler
+from langchain_community.chat_models import ChatOpenAI
+
+
+class MyHandler(AsyncCallbackHandler):
+    def __init__(self, send_func):
+        super().__init__()
+        self.send_func = send_func
+
+    async def on_llm_new_token(self, token: str, **kwargs):
+        await self.send_func(token)
+
+async def get_ai_reply_stream(message: str, send_func, api_key, api_base, model_name):
+    # 实例化时就带回调
+    chat = ChatOpenAI(
+        openai_api_key=api_key,
+        openai_api_base=api_base,
+        model_name=model_name,
+        temperature=0.7,
+        streaming=True,
+        callbacks=[MyHandler(send_func)]
+    )
+    await chat.ainvoke([HumanMessage(content=message)])

backend/ai/models.py

@@ -218,14 +218,12 @@ class ChatRole(CoreModel):
         blank=True,
         related_name="roles",
         verbose_name="关联的知识库",
-        db_comment="关联的知识库"
     )
     tools = models.ManyToManyField(
         'Tool',
         blank=True,
         related_name="roles",
         verbose_name="关联的工具",
-        db_comment="关联的工具"
     )
 
     class Meta:

backend/ai/routing.py

@@ -0,0 +1,7 @@
+from django.urls import re_path
+
+from ai.chat import ChatConsumer
+
+websocket_urlpatterns = [
+    re_path(r'ws/chat/$', ChatConsumer.as_asgi()),
+]

backend/ai/utils.py

@@ -0,0 +1,11 @@
+from ai.models import AIModel
+from utils.models import CommonStatus
+from asgiref.sync import sync_to_async
+
+@sync_to_async
+def get_first_available_ai_config():
+    # 这里只取第一个可用的，可以根据实际业务加筛选条件
+    ai = AIModel.objects.filter(status=CommonStatus.ENABLED).prefetch_related('key').first()
+    if not ai:
+        raise Exception('没有可用的AI配置')
+    return ai.model, ai.key.api_key, ai.key.url

backend/backend/asgi.py

@@ -1,16 +1,17 @@
-"""
-ASGI config for backend project.
-
-It exposes the ASGI callable as a module-level variable named ``application``.
-
-For more information on this file, see
-https://docs.djangoproject.com/en/5.2/howto/deployment/asgi/
-"""
-
 import os
-
 from django.core.asgi import get_asgi_application
+from channels.routing import ProtocolTypeRouter, URLRouter
 
 os.environ.setdefault('DJANGO_SETTINGS_MODULE', 'backend.settings')
 
-application = get_asgi_application()
+# 延迟导入，避免 AppRegistryNotReady 错误
+def get_websocket_urlpatterns():
+    from ai.routing import websocket_urlpatterns
+    return websocket_urlpatterns
+
+application = ProtocolTypeRouter({
+    "http": get_asgi_application(),
+    "websocket": URLRouter(
+        get_websocket_urlpatterns()
+    ),
+})

backend/backend/settings.py

@@ -53,6 +53,7 @@ INSTALLED_APPS = [
     'django_filters',
     'corsheaders',
     'rest_framework.authtoken',
+    'channels',
     "system",
     "ai",
 ]
@@ -231,5 +232,15 @@ LOGGING = {
     }
 }
 
+ASGI_APPLICATION = 'backend.asgi.application'
+
+
+# 简单用内存通道层
+CHANNEL_LAYERS = {
+    'default': {
+        'BACKEND': 'channels.layers.InMemoryChannelLayer'
+    }
+}
+
 if os.path.exists(os.path.join(BASE_DIR, 'backend/local_settings.py')):
     from backend.local_settings import *

backend/requirements.txt

@@ -13,4 +13,9 @@ eventlet==0.40.0
 goofish_api==0.0.6
 flower==2.0.1
 gunicorn==23.0.0
-django_redis==6.0.0
+django_redis==6.0.0
+django-ninja==1.4.3
+openai==1.95
+daphne==4.2.1
+langchain==0.3.26
+langchain-community==0.3.27