Update main.py

main.py

@@ -1,7 +1,6 @@
 import os
 import time
 import random
-import json
 import asyncio
 import requests
 from fastapi import FastAPI, HTTPException, Request
@@ -42,7 +41,10 @@ async def fetch_response(messages: List[ChatCompletionMessage], model: str):
         "user-agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/127.0.0.0 Safari/537.36"
     }
     
+    # 将消息列表转换为单个字符串，保留对话历史
     conversation = "\n".join([f"{msg.role}: {msg.content}" for msg in messages])
+    
+    # 添加指导语
     conversation += "\n请关注并回复user最近的消息并避免总结对话历史的回答"
     
     data = {
@@ -59,44 +61,17 @@ async def fetch_response(messages: List[ChatCompletionMessage], model: str):
     return response.json()
 
 async def stream_response(content: str):
-    chunk_size = len(content)  # 将整个内容作为一个块发送
-    chat_id = f"chatcmpl-{os.urandom(12).hex()}"
-    
-    # 发送开始的块
-    yield f"data: {json.dumps({
-        'id': chat_id,
-        'object': 'chat.completion.chunk',
-        'created': int(time.time()),
-        'model': 'gpt-3.5-turbo-0613',
-        'choices': [{
-            'index': 0,
-            'delta': {
-                'content': content
-            },
-            'finish_reason': None
-        }]
-    })}\n\n"
-
-    # 发送结束的块
-    yield f"data: {json.dumps({
-        'id': chat_id,
-        'object': 'chat.completion.chunk',
-        'created': int(time.time()),
-        'model': 'gpt-3.5-turbo-0613',
-        'choices': [{
-            'index': 0,
-            'delta': {},
-            'finish_reason': 'stop'
-        }]
-    })}\n\n"
-
+    # Send the entire content as a single chunk
+    yield f"data: {{'id': 'chatcmpl-{os.urandom(12).hex()}', 'object': 'chat.completion.chunk', 'created': 1677652288, 'model': 'gpt-3.5-turbo-0613', 'choices': [{'index': 0, 'delta': {{'content': '{content}'}}, 'finish_reason': None}]}}\n\n"
+    yield f"data: {{'id': 'chatcmpl-{os.urandom(12).hex()}', 'object': 'chat.completion.chunk', 'created': 1677652288, 'model': 'gpt-3.5-turbo-0613', 'choices': [{'index': 0, 'delta': {{}}, 'finish_reason': 'stop'}]}}\n\n"
     yield 'data: [DONE]\n\n'
 
-@app.post("/hf/v1/chat/completions")
+@app.post("/v1/chat/completions")
 async def chat_completions(request: Request):
     body = await request.json()
     chat_request = ChatCompletionRequest(**body)
 
+    # 传递整个消息历史到API
     api_response = await fetch_response(chat_request.messages, chat_request.model)
 
     content = api_response.get("response", "")