Update app.py

app.py

@@ -9,26 +9,6 @@ app = Flask(__name__)
 def generate_random_ip():
     return f"{random.randint(1,255)}.{random.randint(0,255)}.{random.randint(0,255)}.{random.randint(0,255)}"
 
-def generate_user_agent():
-    os_list = ['Windows NT 10.0', 'Windows NT 6.1', 'Mac OS X 10_15_7', 'Ubuntu', 'Linux x86_64']
-    browser_list = ['Chrome', 'Firefox', 'Safari', 'Edge']
-    chrome_version = f"{random.randint(70, 126)}.0.{random.randint(1000, 9999)}.{random.randint(100, 999)}"
-    firefox_version = f"{random.randint(70, 100)}.0"
-    safari_version = f"{random.randint(600, 615)}.{random.randint(1, 9)}.{random.randint(1, 9)}"
-    edge_version = f"{random.randint(80, 100)}.0.{random.randint(1000, 9999)}.{random.randint(100, 999)}"
-
-    os = random.choice(os_list)
-    browser = random.choice(browser_list)
-
-    if browser == 'Chrome':
-        return f"Mozilla/5.0 ({os}) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/{chrome_version} Safari/537.36"
-    elif browser == 'Firefox':
-        return f"Mozilla/5.0 ({os}; rv:{firefox_version}) Gecko/20100101 Firefox/{firefox_version}"
-    elif browser == 'Safari':
-        return f"Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/{safari_version} (KHTML, like Gecko) Version/{safari_version.split('.')[0]}.1.2 Safari/{safari_version}"
-    elif browser == 'Edge':
-        return f"Mozilla/5.0 ({os}) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/{edge_version} Safari/537.36 Edg/{edge_version}"
-        
 def format_openai_response(content, finish_reason=None):
     return {
         "id": "chatcmpl-123",
@@ -43,7 +23,6 @@ def format_openai_response(content, finish_reason=None):
     }
 
 @app.route('/hf/v1/chat/completions', methods=['POST'])
-
 def chat_completions():
     data = request.json
     messages = data.get('messages', [])
@@ -52,15 +31,8 @@ def chat_completions():
     if not messages:
         return {"error": "No messages provided"}, 400
     
-    # 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"
-
     model = data.get('model', 'gpt-4o')
-    
-    # 根据模型名称设置endpoint和URL
+
     if model.startswith('gpt'):
         endpoint = "openAI"
         original_api_url = 'https://chatpro.ai-pro.org/api/ask/openAI'
@@ -69,43 +41,57 @@ def chat_completions():
         original_api_url = 'https://chatpro.ai-pro.org/api/ask/claude'
     else:
         return {"error": "Unsupported model"}, 400
-    
+
     headers = {
         'content-type': 'application/json',
         'X-Forwarded-For': generate_random_ip(),
         'origin': 'https://chatpro.ai-pro.org',
         'user-agent': generate_user_agent()
     }
-    
-    payload = {
-        "text": conversation,
-        "endpoint": endpoint,
-        "model": model
-    }
-    
+
     def generate():
-        last_content = ""
-        response = requests.post(original_api_url, headers=headers, json=payload, stream=True)
-        client = sseclient.SSEClient(response)
-        
-        for event in client.events():
-            if event.data.startswith('{"text":'):
-                data = json.loads(event.data)
-                new_content = data['text'][len(last_content):]
-                last_content = data['text']
-                
-                if new_content:
-                    yield f"data: {json.dumps(format_openai_response(new_content))}\n\n"
+        nonlocal messages
+        full_response = ""
+        while True:
+            conversation = "\n".join([f"{msg['role']}: {msg['content']}" for msg in messages])
+            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"
             
-            elif '"final":true' in event.data:
-                yield f"data: {json.dumps(format_openai_response('', 'stop'))}\n\n"
-                yield "data: [DONE]\n\n"
-                break
+            payload = {
+                "text": conversation,
+                "endpoint": endpoint,
+                "model": model
+            }
+            
+            response = requests.post(original_api_url, headers=headers, json=payload, stream=True)
+            client = sseclient.SSEClient(response)
+            
+            for event in client.events():
+                if event.data.startswith('{"text":'):
+                    data = json.loads(event.data)
+                    new_content = data['text'][len(full_response):]
+                    full_response = data['text']
+                    
+                    if new_content:
+                        yield f"data: {json.dumps(format_openai_response(new_content))}\n\n"
+                
+                elif '"final":true' in event.data:
+                    final_data = json.loads(event.data)
+                    finish_reason = final_data.get('responseMessage', {}).get('finish_reason', 'stop')
+                    if finish_reason == 'length':
+                        # ��果因为长度被截断，添加已生成的回复到消息列表，并继续生成
+                        messages.append({"role": "assistant", "content": full_response})
+                        messages.append({"role": "user", "content": "请继续你的输出"})
+                        break
+                    else:
+                        yield f"data: {json.dumps(format_openai_response('', finish_reason))}\n\n"
+                        yield "data: [DONE]\n\n"
+                        return
 
     if stream:
         return Response(stream_with_context(generate()), content_type='text/event-stream')
     else:
         full_response = ""
+        finish_reason = "stop"
         for chunk in generate():
             if chunk.startswith("data: ") and not chunk.strip() == "data: [DONE]":
                 response_data = json.loads(chunk[6:])
@@ -113,6 +99,8 @@ def chat_completions():
                     delta = response_data['choices'][0].get('delta', {})
                     if 'content' in delta:
                         full_response += delta['content']
+                    if 'finish_reason' in delta:
+                        finish_reason = delta['finish_reason']
 
         return {
             "id": "chatcmpl-123",
@@ -125,7 +113,7 @@ def chat_completions():
                     "role": "assistant",
                     "content": full_response
                 },
-                "finish_reason": "stop"
+                "finish_reason": finish_reason
             }],
             "usage": {
                 "prompt_tokens": 0,