Update app.py

app.py

@@ -1,13 +1,26 @@
-from fastapi import FastAPI, HTTPException, Depends
+
 from fastapi.security import APIKeyQuery, APIKeyHeader
-from pydantic import BaseModel
-from transformers import AutoTokenizer, AutoModelForSequenceClassification
-import torch
-import os
 from typing import List, Optional
 from datetime import datetime  # 需在文件开头导入
-from fastapi.responses import MarkdownResponse
 import os
+import uuid
+from datetime import datetime
+from fastapi import FastAPI, HTTPException, Depends
+from fastapi.responses import MarkdownResponse  # 确保导入这行
+from pydantic import BaseModel
+from transformers import AutoTokenizer, AutoModelForSequenceClassification
+import torch
+
+# 【双重保障】设置 Hugging Face 缓存目录（与 Dockerfile 一致）
+os.environ["TRANSFORMERS_CACHE"] = "/tmp/huggingface_cache"
+os.environ["HUGGINGFACE_HUB_CACHE"] = "/tmp/huggingface_cache"
+
+# 初始化 FastAPI 应用
+app = FastAPI(title="Cross-Encoder 重排序 API（兼容 GPT 格式）")
+
+# 后续代码不变（API 认证、模型加载、根路径首页、/v1/chat/completions 接口等）
+# ...
+
 os.environ["TRANSFORMERS_CACHE"] = "/tmp/huggingface_cache"
 class CrossEncoderWrapper:
     def __init__(self):