Update app.py

app.py

@@ -1,540 +1,540 @@
-import json
-import os
-import random
-import time
-import uuid
-import asyncio
-from starlette.responses import Response, StreamingResponse
-from fastapi import FastAPI, Request
-from fastapi.middleware.cors import CORSMiddleware
-
-# 配置常量
-CONFIG = {
-    "API": {
-        "BASE_URL": "https://fragments.e2b.dev",
-        "API_KEY": os.getenv("API_KEY","sk-123456")
-    },
-    "RETRY": {
-        "MAX_ATTEMPTS": 1,
-        "DELAY_BASE": 1000
-    },
-    "MODEL_CONFIG": {
-        "o1": {
-            "id": "o1",
-            "provider": "OpenAI",
-            "providerId": "openai",
-            "name": "o1",
-            "multiModal": True,
-            "Systemprompt": "",
-            "opt_max": {
-                "temperatureMax": 2,
-                "max_tokensMax": 0,
-                "presence_penaltyMax": 2,
-                "frequency_penaltyMax": 2,
-                "top_pMax": 1,
-                "top_kMax": 500
-            }
-        },
-        "o1-mini": {
-            "id": "o1",
-            "provider": "OpenAI",
-            "providerId": "openai",
-            "name": "o1-mini",
-            "multiModal": False,
-            "Systemprompt": "",
-            "opt_max": {
-                "temperatureMax": 2,
-                "max_tokensMax": 0,
-                "presence_penaltyMax": 2,
-                "frequency_penaltyMax": 2,
-                "top_pMax": 1,
-                "top_kMax": 500
-            }
-        },
-        "o3-mini": {
-            "id": "o3-mini",
-            "provider": "OpenAI",
-            "providerId": "openai",
-            "name": "o3 mini",
-            "multiModal": False,
-            "Systemprompt": "",
-            "opt_max": {
-                "temperatureMax": 2,
-                "max_tokensMax": 0,
-                "presence_penaltyMax": 2,
-                "frequency_penaltyMax": 2,
-                "top_pMax": 1,
-                "top_kMax": 500
-            }
-        },
-        "gpt-4.5-preview": {
-            "id": "gpt-4.5-preview",
-            "provider": "OpenAI",
-            "providerId": "openai",
-            "name": "GPT-4.5",
-            "multiModal": True,
-            "Systemprompt": "",
-            "opt_max": {
-                "temperatureMax": 2,
-                "max_tokensMax": 0,
-                "presence_penaltyMax": 2,
-                "frequency_penaltyMax": 2,
-                "top_pMax": 1,
-                "top_kMax": 500
-            }
-        },
-        "gpt-4o": {
-            "id": "gpt-4o",
-            "provider": "OpenAI",
-            "providerId": "openai",
-            "name": "GPT-4o",
-            "multiModal": True,
-            "Systemprompt": "",
-            "opt_max": {
-                "temperatureMax": 2,
-                "max_tokensMax": 16380,
-                "presence_penaltyMax": 2,
-                "frequency_penaltyMax": 2,
-                "top_pMax": 1,
-                "top_kMax": 500
-            }
-        },
-        "gpt-4-turbo": {
-            "id": "gpt-4-turbo",
-            "provider": "OpenAI",
-            "providerId": "openai",
-            "name": "GPT-4 Turbo",
-            "multiModal": True,
-            "Systemprompt": "",
-            "opt_max": {
-                "temperatureMax": 2,
-                "max_tokensMax": 16380,
-                "presence_penaltyMax": 2,
-                "frequency_penaltyMax": 2,
-                "top_pMax": 1,
-                "top_kMax": 500
-            }
-        },
-        "gemini-1.5-pro": {
-            "id": "gemini-1.5-pro-002",
-            "provider": "Google Vertex AI",
-            "providerId": "vertex",
-            "name": "Gemini 1.5 Pro",
-            "multiModal": True,
-            "Systemprompt": "",
-            "opt_max": {
-                "temperatureMax": 2,
-                "max_tokensMax": 8192,
-                "presence_penaltyMax": 2,
-                "frequency_penaltyMax": 2,
-                "top_pMax": 1,
-                "top_kMax": 500
-            }
-        },
-        "gemini-exp-1121": {
-            "id": "gemini-exp-1121",
-            "provider": "Google Generative AI",
-            "providerId": "google",
-            "name": "Gemini Experimental 1121",
-            "multiModal": True,
-            "Systemprompt": "",
-            "opt_max": {
-                "temperatureMax": 2,
-                "max_tokensMax": 8192,
-                "presence_penaltyMax": 2,
-                "frequency_penaltyMax": 2,
-                "top_pMax": 1,
-                "top_kMax": 40
-            }
-        },
-        "gemini-2.0-flash-exp": {
-            "id": "models/gemini-2.0-flash-exp",
-            "provider": "Google Generative AI",
-            "providerId": "google",
-            "name": "Gemini 2.0 Flash",
-            "multiModal": True,
-            "Systemprompt": "",
-            "opt_max": {
-                "temperatureMax": 2,
-                "max_tokensMax": 8192,
-                "presence_penaltyMax": 2,
-                "frequency_penaltyMax": 2,
-                "top_pMax": 1,
-                "top_kMax": 40
-            }
-        },
-        "claude-3-7-sonnet-latest": {
-            "id": "claude-3-5-sonnet-latest",
-            "provider": "Anthropic",
-            "providerId": "anthropic",
-            "name": "Claude 3.7 Sonnet",
-            "multiModal": True,
-            "Systemprompt": "",
-            "opt_max": {
-                "temperatureMax": 1,
-                "max_tokensMax": 8192,
-                "presence_penaltyMax": 2,
-                "frequency_penaltyMax": 2,
-                "top_pMax": 1,
-                "top_kMax": 500
-            }
-        },
-        "claude-3-5-sonnet-latest": {
-            "id": "claude-3-5-sonnet-latest",
-            "provider": "Anthropic",
-            "providerId": "anthropic",
-            "name": "Claude 3.5 Sonnet",
-            "multiModal": True,
-            "Systemprompt": "",
-            "opt_max": {
-                "temperatureMax": 1,
-                "max_tokensMax": 8192,
-                "presence_penaltyMax": 2,
-                "frequency_penaltyMax": 2,
-                "top_pMax": 1,
-                "top_kMax": 500
-            }
-        },
-        "claude-3-5-haiku-latest": {
-            "id": "claude-3-5-haiku-latest",
-            "provider": "Anthropic",
-            "providerId": "anthropic",
-            "name": "Claude 3.5 Haiku",
-            "multiModal": False,
-            "Systemprompt": "",
-            "opt_max": {
-                "temperatureMax": 1,
-                "max_tokensMax": 8192,
-                "presence_penaltyMax": 2,
-                "frequency_penaltyMax": 2,
-                "top_pMax": 1,
-                "top_kMax": 500
-            }
-        }
-    },
-    "DEFAULT_HEADERS": {
-        'accept': '*/*',
-        'accept-language': 'zh-CN,zh;q=0.9,en;q=0.8,en-GB;q=0.7,en-US;q=0.6',
-        'cache-control': 'no-cache',
-        'content-type': 'application/json',
-        'origin': 'https://fragments.e2b.dev',
-        'pragma': 'no-cache',
-        'priority': 'u=1, i',
-        'referer': 'https://fragments.e2b.dev/',
-        'sec-ch-ua': '"Not(A:Brand";v="99", "Microsoft Edge";v="133", "Chromium";v="133"',
-        'sec-ch-ua-mobile': '?0',
-        'sec-ch-ua-platform': '"Windows"',
-        'sec-fetch-dest': 'empty',
-        'sec-fetch-mode': 'cors',
-        'sec-fetch-site': 'same-origin',
-        'user-agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/133.0.0.0 Safari/537.36 Edg/133.0.0.0'
-    },
-    "MODEL_PROMPT": "Chatting with users and starting role-playing, the most important thing is to pay attention to their latest messages, use only 'text' to output the chat text reply content generated for user messages, and finally output it in code"
-}
-
-
-# 工具类
-class Utils:
-    @staticmethod
-    def uuidv4():
-        return str(uuid.uuid4())
-
-    @staticmethod
-    async def config_opt(params, model_config):
-        if not model_config.get("opt_max"):
-            return None
-
-        options_map = {
-            "temperature": "temperatureMax",
-            "max_tokens": "max_tokensMax",
-            "presence_penalty": "presence_penaltyMax",
-            "frequency_penalty": "frequency_penaltyMax",
-            "top_p": "top_pMax",
-            "top_k": "top_kMax"
-        }
-
-        constrained_params = {}
-        for key, value in params.items():
-            max_key = options_map.get(key)
-            if (max_key and 
-                max_key in model_config["opt_max"] and 
-                value is not None):
-                constrained_params[key] = min(value, model_config["opt_max"][max_key])
-
-        return constrained_params
-
-
-# API客户端类
-class ApiClient:
-    def __init__(self, model_id):
-        if model_id not in CONFIG["MODEL_CONFIG"]:
-            raise ValueError(f"不支持的模型: {model_id}")
-        self.model_config = CONFIG["MODEL_CONFIG"][model_id]
-
-    def process_message_content(self, content):
-        if isinstance(content, str):
-            return content
-        if isinstance(content, list):
-            return "\n".join([item["text"] for item in content if item["type"] == "text"])
-        if isinstance(content, dict):
-            return content.get("text")
-        return None
-
-    async def prepare_chat_request(self, request, config=None):
-        opt_config = config or {"model": self.model_config["id"]}
-        return {
-            "userID": Utils.uuidv4(),
-            "messages": await self.transform_messages(request),
-            "template": {
-                "text": {
-                    "name": CONFIG["MODEL_PROMPT"],
-                    "lib": [""],
-                    "file": "pages/ChatWithUsers.txt",
-                    "instructions": self.model_config["Systemprompt"],
-                    "port": None
-                }
-            },
-            "model": {
-                "id": self.model_config["id"],
-                "provider": self.model_config["provider"],
-                "providerId": self.model_config["providerId"],
-                "name": self.model_config["name"],
-                "multiModal": self.model_config["multiModal"]
-            },
-            "config": {"model": self.model_config["id"]}
-        }
-
-    async def transform_messages(self, request):
-        merged_messages = []
-        for current in request["messages"]:
-            current_content = self.process_message_content(current["content"])
-            if current_content is None:
-                continue
-
-            if (merged_messages and 
-                current and 
-                merged_messages[-1]["role"] == current["role"]):
-                last_content = self.process_message_content(merged_messages[-1]["content"])
-                if last_content is not None:
-                    merged_messages[-1]["content"] = f"{last_content}\n{current_content}"
-                    continue
-            
-            merged_messages.append(current)
-
-        messages = []
-        for msg in merged_messages:
-            if msg["role"] in ["system", "user"]:
-                messages.append({
-                    "role": "user",
-                    "content": [{
-                        "type": "text",
-                        "text": msg["content"]
-                    }]
-                })
-            elif msg["role"] == "assistant":
-                messages.append({
-                    "role": "assistant",
-                    "content": [{
-                        "type": "text",
-                        "text": msg["content"]
-                    }]
-                })
-        
-        return messages
-
-
-# 响应处理类
-class ResponseHandler:
-    @staticmethod
-    async def handle_stream_response(chat_message, model):
-        async def stream_generator():
-            index = 0
-            while True:
-                # 如果已经发送完所有内容
-                if index >= len(chat_message):
-                    yield f"data: [DONE]\n\n"
-                    break
-
-                chunk_size = random.randint(15, 29)
-                chunk = chat_message[index:index + chunk_size]
-
-                event_data = {
-                    "id": Utils.uuidv4(),
-                    "object": "chat.completion.chunk",
-                    "created": int(time.time()),
-                    "model": model,
-                    "choices": [{
-                        "index": 0,
-                        "delta": {"content": chunk},
-                        "finish_reason": "stop" if index + chunk_size >= len(chat_message) else None
-                    }]
-                }
-                
-                try:
-                    payload = f"data: {json.dumps(event_data)}\n\n"
-                    yield payload
-                except Exception as error:
-                    raise Exception(f"json转换失败: {error}")
-
-                index += chunk_size
-                await asyncio.sleep(0.05)  # 50ms delay
-
-        return StreamingResponse(
-            stream_generator(),
-            media_type="text/event-stream",
-            headers={
-                "Cache-Control": "no-cache",
-                "Connection": "keep-alive",
-            }
-        )
-
-    @staticmethod
-    async def handle_normal_response(chat_message, model):
-        response_data = {
-            "id": Utils.uuidv4(),
-            "object": "chat.completion",
-            "created": int(time.time()),
-            "model": model,
-            "choices": [{
-                "index": 0,
-                "message": {
-                    "role": "assistant",
-                    "content": chat_message
-                },
-                "finish_reason": "stop"
-            }],
-            "usage": None
-        }
-        
-        return Response(
-            content=json.dumps(response_data),
-            media_type="application/json"
-        )
-
-
-# FastAPI 应用
-app = FastAPI()
-
-# 添加 CORS 中间件
-app.add_middleware(
-    CORSMiddleware,
-    allow_origins=["*"],
-    allow_credentials=True,
-    allow_methods=["*"],
-    allow_headers=["*"],
-)
-
-
-# 模型列表端点
-@app.get("/hf/v1/models")
-async def get_models():
-    return {
-        "object": "list",
-        "data": [
-            {
-                "id": model,
-                "object": "model",
-                "created": int(time.time()),
-                "owned_by": "e2b",
-            }
-            for model in CONFIG["MODEL_CONFIG"].keys()
-        ]
-    }
-
-
-# 聊天完成端点
-@app.post("/hf/v1/chat/completions")
-async def chat_completions(request: Request):
-    try:
-        # 验证授权
-        auth_header = request.headers.get("authorization", "")
-        auth_token = auth_header.replace("Bearer ", "") if auth_header else ""
-        
-        if auth_token != CONFIG["API"]["API_KEY"]:
-            return Response(
-                content=json.dumps({"error": "Unauthorized"}),
-                status_code=401,
-                media_type="application/json"
-            )
-        
-        # 解析请求体
-        request_body = await request.json()
-        model = request_body.get("model")
-        temperature = request_body.get("temperature")
-        max_tokens = request_body.get("max_tokens")
-        presence_penalty = request_body.get("presence_penalty")
-        frequency_penalty = request_body.get("frequency_penalty")
-        top_p = request_body.get("top_p")
-        top_k = request_body.get("top_k")
-        stream = request_body.get("stream", False)
-        
-        # 配置选项
-
-        config_opt = await Utils.config_opt(
-            {
-                "temperature": temperature,
-                "max_tokens": max_tokens,
-                "presence_penalty": presence_penalty,
-                "frequency_penalty": frequency_penalty,
-                "top_p": top_p,
-                "top_k": top_k
-            },
-            CONFIG["MODEL_CONFIG"][model]
-        )
-
-        #config_opt = {"model": model if CONFIG["MODEL_CONFIG"].get(model) else "gpt-4o"}
-        # 准备请求
-        api_client = ApiClient(model)
-        request_payload = await api_client.prepare_chat_request(request_body, config_opt)
-        
-        # 发送请求到上游服务
-        import httpx
-        async with httpx.AsyncClient() as client:
-            response = await client.post(
-                f"{CONFIG['API']['BASE_URL']}/api/chat",
-                headers=CONFIG["DEFAULT_HEADERS"],
-                json=request_payload
-            )
-            response_data = response.json()
-
-        # 处理响应
-        chat_message = (response_data.get('code', '') or response_data.get('text', '') or response_data or '').strip() or None
-
-        if not chat_message:
-            raise Exception("No response from upstream service")
-
-        # 返回流式或普通响应
-        if stream:
-            return await ResponseHandler.handle_stream_response(chat_message, model)
-        else:
-            return await ResponseHandler.handle_normal_response(chat_message, model)
-            
-    except Exception as error:
-        return Response(
-            content=json.dumps({
-                "error": {
-                    "message": f"{str(error)} 请求失败，可能是上下文超出限制或其他错误，请稍后重试。",
-                    "type": "server_error",
-                    "param": None,
-                    "code": getattr(error, "code", None)
-                }
-            }),
-            status_code=500,
-            media_type="application/json"
-        )
-
-
-# 404 处理
-@app.api_route("/{path:path}", methods=["GET", "POST", "PUT", "DELETE"])
-async def not_found(request: Request, path: str):
-    return Response(
-        content="服务运行成功，请使用正确请求路径",
-        status_code=404,
-        headers={"Access-Control-Allow-Origin": "*"}
-    )
-
-
-# 启动服务器
-if __name__ == "__main__":
-    import uvicorn
+import json
+import os
+import random
+import time
+import uuid
+import asyncio
+from starlette.responses import Response, StreamingResponse
+from fastapi import FastAPI, Request
+from fastapi.middleware.cors import CORSMiddleware
+
+# 配置常量
+CONFIG = {
+    "API": {
+        "BASE_URL": "https://fragments.e2b.dev",
+        "API_KEY": os.getenv("API_KEY","sk-123456")
+    },
+    "RETRY": {
+        "MAX_ATTEMPTS": 1,
+        "DELAY_BASE": 1000
+    },
+    "MODEL_CONFIG": {
+        "o1": {
+            "id": "o1",
+            "provider": "OpenAI",
+            "providerId": "openai",
+            "name": "o1",
+            "multiModal": True,
+            "Systemprompt": "",
+            "opt_max": {
+                "temperatureMax": 2,
+                "max_tokensMax": 0,
+                "presence_penaltyMax": 2,
+                "frequency_penaltyMax": 2,
+                "top_pMax": 1,
+                "top_kMax": 500
+            }
+        },
+        "o1-mini": {
+            "id": "o1",
+            "provider": "OpenAI",
+            "providerId": "openai",
+            "name": "o1-mini",
+            "multiModal": False,
+            "Systemprompt": "",
+            "opt_max": {
+                "temperatureMax": 2,
+                "max_tokensMax": 0,
+                "presence_penaltyMax": 2,
+                "frequency_penaltyMax": 2,
+                "top_pMax": 1,
+                "top_kMax": 500
+            }
+        },
+        "o3-mini": {
+            "id": "o3-mini",
+            "provider": "OpenAI",
+            "providerId": "openai",
+            "name": "o3 mini",
+            "multiModal": False,
+            "Systemprompt": "",
+            "opt_max": {
+                "temperatureMax": 2,
+                "max_tokensMax": 0,
+                "presence_penaltyMax": 2,
+                "frequency_penaltyMax": 2,
+                "top_pMax": 1,
+                "top_kMax": 500
+            }
+        },
+        "gpt-4.5-preview": {
+            "id": "gpt-4.5-preview",
+            "provider": "OpenAI",
+            "providerId": "openai",
+            "name": "GPT-4.5",
+            "multiModal": True,
+            "Systemprompt": "",
+            "opt_max": {
+                "temperatureMax": 2,
+                "max_tokensMax": 0,
+                "presence_penaltyMax": 2,
+                "frequency_penaltyMax": 2,
+                "top_pMax": 1,
+                "top_kMax": 500
+            }
+        },
+        "gpt-4o": {
+            "id": "gpt-4o",
+            "provider": "OpenAI",
+            "providerId": "openai",
+            "name": "GPT-4o",
+            "multiModal": True,
+            "Systemprompt": "",
+            "opt_max": {
+                "temperatureMax": 2,
+                "max_tokensMax": 16380,
+                "presence_penaltyMax": 2,
+                "frequency_penaltyMax": 2,
+                "top_pMax": 1,
+                "top_kMax": 500
+            }
+        },
+        "gpt-4-turbo": {
+            "id": "gpt-4-turbo",
+            "provider": "OpenAI",
+            "providerId": "openai",
+            "name": "GPT-4 Turbo",
+            "multiModal": True,
+            "Systemprompt": "",
+            "opt_max": {
+                "temperatureMax": 2,
+                "max_tokensMax": 16380,
+                "presence_penaltyMax": 2,
+                "frequency_penaltyMax": 2,
+                "top_pMax": 1,
+                "top_kMax": 500
+            }
+        },
+        "gemini-1.5-pro": {
+            "id": "gemini-1.5-pro-002",
+            "provider": "Google Vertex AI",
+            "providerId": "vertex",
+            "name": "Gemini 1.5 Pro",
+            "multiModal": True,
+            "Systemprompt": "",
+            "opt_max": {
+                "temperatureMax": 2,
+                "max_tokensMax": 8192,
+                "presence_penaltyMax": 2,
+                "frequency_penaltyMax": 2,
+                "top_pMax": 1,
+                "top_kMax": 500
+            }
+        },
+        "gemini-exp-1121": {
+            "id": "gemini-exp-1121",
+            "provider": "Google Generative AI",
+            "providerId": "google",
+            "name": "Gemini Experimental 1121",
+            "multiModal": True,
+            "Systemprompt": "",
+            "opt_max": {
+                "temperatureMax": 2,
+                "max_tokensMax": 8192,
+                "presence_penaltyMax": 2,
+                "frequency_penaltyMax": 2,
+                "top_pMax": 1,
+                "top_kMax": 40
+            }
+        },
+        "gemini-2.0-flash-exp": {
+            "id": "models/gemini-2.0-flash-exp",
+            "provider": "Google Generative AI",
+            "providerId": "google",
+            "name": "Gemini 2.0 Flash",
+            "multiModal": True,
+            "Systemprompt": "",
+            "opt_max": {
+                "temperatureMax": 2,
+                "max_tokensMax": 8192,
+                "presence_penaltyMax": 2,
+                "frequency_penaltyMax": 2,
+                "top_pMax": 1,
+                "top_kMax": 40
+            }
+        },
+        "claude-3-7-sonnet-latest": {
+            "id": "claude-3-5-sonnet-latest",
+            "provider": "Anthropic",
+            "providerId": "anthropic",
+            "name": "Claude 3.7 Sonnet",
+            "multiModal": True,
+            "Systemprompt": "",
+            "opt_max": {
+                "temperatureMax": 1,
+                "max_tokensMax": 8192,
+                "presence_penaltyMax": 2,
+                "frequency_penaltyMax": 2,
+                "top_pMax": 1,
+                "top_kMax": 500
+            }
+        },
+        "claude-3-5-sonnet-latest": {
+            "id": "claude-3-5-sonnet-latest",
+            "provider": "Anthropic",
+            "providerId": "anthropic",
+            "name": "Claude 3.5 Sonnet",
+            "multiModal": True,
+            "Systemprompt": "",
+            "opt_max": {
+                "temperatureMax": 1,
+                "max_tokensMax": 8192,
+                "presence_penaltyMax": 2,
+                "frequency_penaltyMax": 2,
+                "top_pMax": 1,
+                "top_kMax": 500
+            }
+        },
+        "claude-3-5-haiku-latest": {
+            "id": "claude-3-5-haiku-latest",
+            "provider": "Anthropic",
+            "providerId": "anthropic",
+            "name": "Claude 3.5 Haiku",
+            "multiModal": False,
+            "Systemprompt": "",
+            "opt_max": {
+                "temperatureMax": 1,
+                "max_tokensMax": 8192,
+                "presence_penaltyMax": 2,
+                "frequency_penaltyMax": 2,
+                "top_pMax": 1,
+                "top_kMax": 500
+            }
+        }
+    },
+    "DEFAULT_HEADERS": {
+        'accept': '*/*',
+        'accept-language': 'zh-CN,zh;q=0.9,en;q=0.8,en-GB;q=0.7,en-US;q=0.6',
+        'cache-control': 'no-cache',
+        'content-type': 'application/json',
+        'origin': 'https://fragments.e2b.dev',
+        'pragma': 'no-cache',
+        'priority': 'u=1, i',
+        'referer': 'https://fragments.e2b.dev/',
+        'sec-ch-ua': '"Not(A:Brand";v="99", "Microsoft Edge";v="133", "Chromium";v="133"',
+        'sec-ch-ua-mobile': '?0',
+        'sec-ch-ua-platform': '"Windows"',
+        'sec-fetch-dest': 'empty',
+        'sec-fetch-mode': 'cors',
+        'sec-fetch-site': 'same-origin',
+        'user-agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/133.0.0.0 Safari/537.36 Edg/133.0.0.0'
+    },
+    "MODEL_PROMPT": "Chatting with users and starting role-playing, the most important thing is to pay attention to their latest messages, use only 'text' to output the chat text reply content generated for user messages, and finally output it in code"
+}
+
+
+# 工具类
+class Utils:
+    @staticmethod
+    def uuidv4():
+        return str(uuid.uuid4())
+
+    @staticmethod
+    async def config_opt(params, model_config):
+        if not model_config.get("opt_max"):
+            return None
+
+        options_map = {
+            "temperature": "temperatureMax",
+            "max_tokens": "max_tokensMax",
+            "presence_penalty": "presence_penaltyMax",
+            "frequency_penalty": "frequency_penaltyMax",
+            "top_p": "top_pMax",
+            "top_k": "top_kMax"
+        }
+
+        constrained_params = {}
+        for key, value in params.items():
+            max_key = options_map.get(key)
+            if (max_key and 
+                max_key in model_config["opt_max"] and 
+                value is not None):
+                constrained_params[key] = min(value, model_config["opt_max"][max_key])
+
+        return constrained_params
+
+
+# API客户端类
+class ApiClient:
+    def __init__(self, model_id):
+        if model_id not in CONFIG["MODEL_CONFIG"]:
+            raise ValueError(f"不支持的模型: {model_id}")
+        self.model_config = CONFIG["MODEL_CONFIG"][model_id]
+
+    def process_message_content(self, content):
+        if isinstance(content, str):
+            return content
+        if isinstance(content, list):
+            return "\n".join([item["text"] for item in content if item["type"] == "text"])
+        if isinstance(content, dict):
+            return content.get("text")
+        return None
+
+    async def prepare_chat_request(self, request, config=None):
+        opt_config = config or {"model": self.model_config["id"]}
+        return {
+            "userID": Utils.uuidv4(),
+            "messages": await self.transform_messages(request),
+            "template": {
+                "text": {
+                    "name": CONFIG["MODEL_PROMPT"],
+                    "lib": [""],
+                    "file": "pages/ChatWithUsers.txt",
+                    "instructions": self.model_config["Systemprompt"],
+                    "port": None
+                }
+            },
+            "model": {
+                "id": self.model_config["id"],
+                "provider": self.model_config["provider"],
+                "providerId": self.model_config["providerId"],
+                "name": self.model_config["name"],
+                "multiModal": self.model_config["multiModal"]
+            },
+            "config": opt_config
+        }
+
+    async def transform_messages(self, request):
+        merged_messages = []
+        for current in request["messages"]:
+            current_content = self.process_message_content(current["content"])
+            if current_content is None:
+                continue
+
+            if (merged_messages and 
+                current and 
+                merged_messages[-1]["role"] == current["role"]):
+                last_content = self.process_message_content(merged_messages[-1]["content"])
+                if last_content is not None:
+                    merged_messages[-1]["content"] = f"{last_content}\n{current_content}"
+                    continue
+            
+            merged_messages.append(current)
+
+        messages = []
+        for msg in merged_messages:
+            if msg["role"] in ["system", "user"]:
+                messages.append({
+                    "role": "user",
+                    "content": [{
+                        "type": "text",
+                        "text": msg["content"]
+                    }]
+                })
+            elif msg["role"] == "assistant":
+                messages.append({
+                    "role": "assistant",
+                    "content": [{
+                        "type": "text",
+                        "text": msg["content"]
+                    }]
+                })
+        
+        return messages
+
+
+# 响应处理类
+class ResponseHandler:
+    @staticmethod
+    async def handle_stream_response(chat_message, model):
+        async def stream_generator():
+            index = 0
+            while True:
+                # 如果已经发送完所有内容
+                if index >= len(chat_message):
+                    yield f"data: [DONE]\n\n"
+                    break
+
+                chunk_size = random.randint(15, 29)
+                chunk = chat_message[index:index + chunk_size]
+
+                event_data = {
+                    "id": Utils.uuidv4(),
+                    "object": "chat.completion.chunk",
+                    "created": int(time.time()),
+                    "model": model,
+                    "choices": [{
+                        "index": 0,
+                        "delta": {"content": chunk},
+                        "finish_reason": "stop" if index + chunk_size >= len(chat_message) else None
+                    }]
+                }
+                
+                try:
+                    payload = f"data: {json.dumps(event_data)}\n\n"
+                    yield payload
+                except Exception as error:
+                    raise Exception(f"json转换失败: {error}")
+
+                index += chunk_size
+                await asyncio.sleep(0.05)  # 50ms delay
+
+        return StreamingResponse(
+            stream_generator(),
+            media_type="text/event-stream",
+            headers={
+                "Cache-Control": "no-cache",
+                "Connection": "keep-alive",
+            }
+        )
+
+    @staticmethod
+    async def handle_normal_response(chat_message, model):
+        response_data = {
+            "id": Utils.uuidv4(),
+            "object": "chat.completion",
+            "created": int(time.time()),
+            "model": model,
+            "choices": [{
+                "index": 0,
+                "message": {
+                    "role": "assistant",
+                    "content": chat_message
+                },
+                "finish_reason": "stop"
+            }],
+            "usage": None
+        }
+        
+        return Response(
+            content=json.dumps(response_data),
+            media_type="application/json"
+        )
+
+
+# FastAPI 应用
+app = FastAPI()
+
+# 添加 CORS 中间件
+app.add_middleware(
+    CORSMiddleware,
+    allow_origins=["*"],
+    allow_credentials=True,
+    allow_methods=["*"],
+    allow_headers=["*"],
+)
+
+
+# 模型列表端点
+@app.get("/hf/v1/models")
+async def get_models():
+    return {
+        "object": "list",
+        "data": [
+            {
+                "id": model,
+                "object": "model",
+                "created": int(time.time()),
+                "owned_by": "e2b",
+            }
+            for model in CONFIG["MODEL_CONFIG"].keys()
+        ]
+    }
+
+
+# 聊天完成端点
+@app.post("/hf/v1/chat/completions")
+async def chat_completions(request: Request):
+    try:
+        # 验证授权
+        auth_header = request.headers.get("authorization", "")
+        auth_token = auth_header.replace("Bearer ", "") if auth_header else ""
+        
+        if auth_token != CONFIG["API"]["API_KEY"]:
+            return Response(
+                content=json.dumps({"error": "Unauthorized"}),
+                status_code=401,
+                media_type="application/json"
+            )
+        
+        # 解析请求体
+        request_body = await request.json()
+        model = request_body.get("model")
+        temperature = request_body.get("temperature")
+        max_tokens = request_body.get("max_tokens")
+        presence_penalty = request_body.get("presence_penalty")
+        frequency_penalty = request_body.get("frequency_penalty")
+        top_p = request_body.get("top_p")
+        top_k = request_body.get("top_k")
+        stream = request_body.get("stream", False)
+        
+        # 配置选项
+
+        config_opt = await Utils.config_opt(
+            {
+                "temperature": temperature,
+                "max_tokens": max_tokens,
+                "presence_penalty": presence_penalty,
+                "frequency_penalty": frequency_penalty,
+                "top_p": top_p,
+                "top_k": top_k
+            },
+            CONFIG["MODEL_CONFIG"][model]
+        )
+
+        #config_opt = {"model": model if CONFIG["MODEL_CONFIG"].get(model) else "gpt-4o"}
+        # 准备请求
+        api_client = ApiClient(model)
+        request_payload = await api_client.prepare_chat_request(request_body, config_opt)
+        
+        # 发送请求到上游服务
+        import httpx
+        async with httpx.AsyncClient() as client:
+            response = await client.post(
+                f"{CONFIG['API']['BASE_URL']}/api/chat",
+                headers=CONFIG["DEFAULT_HEADERS"],
+                json=request_payload
+            )
+            response_data = response.json()
+
+        # 处理响应
+        chat_message = (response_data.get('code', '') or response_data.get('text', '') or response_data or '').strip() or None
+
+        if not chat_message:
+            raise Exception("No response from upstream service")
+
+        # 返回流式或普通响应
+        if stream:
+            return await ResponseHandler.handle_stream_response(chat_message, model)
+        else:
+            return await ResponseHandler.handle_normal_response(chat_message, model)
+            
+    except Exception as error:
+        return Response(
+            content=json.dumps({
+                "error": {
+                    "message": f"{str(error)} 请求失败，可能是上下文超出限制或其他错误，请稍后重试。",
+                    "type": "server_error",
+                    "param": None,
+                    "code": getattr(error, "code", None)
+                }
+            }),
+            status_code=500,
+            media_type="application/json"
+        )
+
+
+# 404 处理
+@app.api_route("/{path:path}", methods=["GET", "POST", "PUT", "DELETE"])
+async def not_found(request: Request, path: str):
+    return Response(
+        content="服务运行成功，请使用正确请求路径",
+        status_code=404,
+        headers={"Access-Control-Allow-Origin": "*"}
+    )
+
+
+# 启动服务器
+if __name__ == "__main__":
+    import uvicorn
     uvicorn.run(app, host="0.0.0.0", port=7860)