Update app.py

app.py

@@ -5,6 +5,18 @@ import os
 import numpy as np
 from sentence_transformers import SentenceTransformer
 from typing import List, Optional
+import logging
+
+# 配置日志
+logging.basicConfig(
+    level=logging.INFO,
+    format="%(asctime)s-%(name)s-%(levelname)s-%(message)s",
+    handlers=[
+        logging.FileHandler("embedding_service.log"),  # 日志写入文件
+        logging.StreamHandler()  # 同时输出到控制台
+    ]
+)
+logger = logging.getLogger("embedding_service")
 
 app = FastAPI()
 
@@ -27,21 +39,35 @@ MODEL_MAPPING = {
 models = {}
 
 def get_model(model_name: str):
+    logger.info(f"尝试获取模型: {model_name}")
     if model_name not in models:
         # 检查是否支持该模型
         if model_name not in MODEL_MAPPING:
-            raise HTTPException(status_code=400, detail=f"不支持的模型: {model_name}")
+            error_msg = f"不支持的模型: {model_name}"
+            logger.error(error_msg)
+            raise HTTPException(status_code=400, detail=error_msg)
         # 加载模型
-        models[model_name] = SentenceTransformer(MODEL_MAPPING[model_name])
+        logger.info(f"开始加载模型: {MODEL_MAPPING[model_name]}")
+        try:
+            models[model_name] = SentenceTransformer(MODEL_MAPPING[model_name])
+            logger.info(f"模型 {model_name} 加载成功")
+        except Exception as e:
+            error_msg = f"加载模型 {model_name} 失败: {str(e)}"
+            logger.error(error_msg)
+            raise HTTPException(status_code=500, detail=error_msg)
     return models[model_name]
 
 # 验证API密钥
 def verify_api_key(authorization: Optional[str] = None):
+    logger.info("验证API密钥")
     if not authorization or not authorization.startswith("Bearer "):
+        logger.warning("未提供有效的API密钥格式")
         raise HTTPException(status_code=401, detail="未提供有效的API密钥")
     api_key = authorization[len("Bearer "):]
     if api_key != os.getenv("API_KEY"):
+        logger.warning("无效的API密钥")
         raise HTTPException(status_code=401, detail="无效的API密钥")
+    logger.info("API密钥验证通过")
     return True
 
 # 请求体模型（对齐OpenAI格式）
@@ -67,15 +93,19 @@ async def create_embedding(
     request: EmbeddingRequest,
     _: bool = Depends(verify_api_key)
 ):
+    logger.info(f"收到嵌入请求，模型: {request.model}, 输入类型: {type(request.input)}")
     try:
         # 获取模型
         model = get_model(request.model)
         
         # 处理输入（支持单文本或文本列表）
         inputs = [request.input] if isinstance(request.input, str) else request.input
+        logger.info(f"处理输入，文本数量: {len(inputs)}")
         
         # 计算嵌入
+        logger.info("开始计算嵌入")
         embeddings = model.encode(inputs, normalize_embeddings=True)
+        logger.info(f"嵌入计算完成，嵌入形状: {embeddings.shape}")
         
         # 构建响应
         data = [
@@ -85,6 +115,7 @@ async def create_embedding(
         
         # 估算token数（简单近似：每个单词约1 token）
         prompt_tokens = sum(len(text.split()) for text in inputs)
+        logger.info(f"估算token数: {prompt_tokens}")
         
         return EmbeddingResponse(
             data=data,
@@ -92,14 +123,17 @@ async def create_embedding(
             usage={"prompt_tokens": prompt_tokens, "total_tokens": prompt_tokens}
         )
     except Exception as e:
-        raise HTTPException(status_code=500, detail=str(e))
+        error_msg = f"处理嵌入请求时发生错误: {str(e)}"
+        logger.error(error_msg)
+        raise HTTPException(status_code=500, detail=error_msg)
 
 # 健康检查接口
 @app.get("/health")
 async def health_check():
+    logger.info("健康检查请求")
     return {"status": "healthy", "models": list(MODEL_MAPPING.keys())}
 
 if __name__ == "__main__":
     import uvicorn
-    uvicorn.run(app, host="0.0.0.0", port=7860)
-    
+    logger.info("启动服务")
+    uvicorn.run(app, host="0.0.0.0", port=7860)