feat: 添加mineru健康检查端点并优化类型注解

- 在 app.py 中新增健康检查端点以确认 API 正在运行。
- 优化了类型注解，使用更简洁的语法。
- 在 db_manager.py 中修正了日志信息的语言。
- 在 OCR 插件中添加了对 MinerU OCR 服务健康检查的逻辑。

docker-compose.yml

@@ -150,6 +150,8 @@ services:
       dockerfile: Dockerfile
     image: mineru-api:latest
     container_name: mineru-api
+    profiles:
+      - all
     deploy:
       resources:
         reservations:

scripts/mineru-api/app.py

@@ -4,7 +4,6 @@ import tempfile
 from base64 import b64encode
 from glob import glob
 from io import StringIO
-from typing import Tuple, Union
 
 import magic_pdf.model as model_config
 import uvicorn
@@ -54,9 +53,9 @@ def init_writers(
     file: UploadFile = None,
     output_path: str = None,
     output_image_path: str = None,
-) -> Tuple[
-    Union[S3DataWriter, FileBasedDataWriter],
-    Union[S3DataWriter, FileBasedDataWriter],
+) -> tuple[
+    S3DataWriter | FileBasedDataWriter,
+    S3DataWriter | FileBasedDataWriter,
     bytes,
 ]:
     """
@@ -113,8 +112,8 @@ def process_file(
     file_bytes: bytes,
     file_extension: str,
     parse_method: str,
-    image_writer: Union[S3DataWriter, FileBasedDataWriter],
-) -> Tuple[InferenceResult, PipeResult]:
+    image_writer: S3DataWriter | FileBasedDataWriter,
+) -> tuple[InferenceResult, PipeResult]:
     """
     Process PDF file content
 
@@ -128,7 +127,7 @@ def process_file(
         Tuple[InferenceResult, PipeResult]: Returns inference result and pipeline result
     """
 
-    ds: Union[PymuDocDataset, ImageDataset] = None
+    ds: PymuDocDataset | ImageDataset = None
     if file_extension in pdf_extensions:
         ds = PymuDocDataset(file_bytes)
     elif file_extension in office_extensions:
@@ -169,6 +168,13 @@ def encode_image(image_path: str) -> str:
         return b64encode(f.read()).decode()
 
 
+@app.get("/health", tags=["health"], summary="Health check endpoint")
+async def health_check():
+    """
+    Simple health check endpoint to confirm the API is running.
+    """
+    return JSONResponse(content={"status": "healthy"}, status_code=200)
+
 @app.post(
     "/file_parse",
     tags=["projects"],
@@ -206,9 +212,7 @@ async def file_parse(
         return_content_list: Whether to return parsed PDF content list. Default to False
     """
     try:
-        if (file is None and file_path is None) or (
-            file is not None and file_path is not None
-        ):
+        if (bool(file) == bool(file_path)):
             return JSONResponse(
                 content={"error": "Must provide either file or file_path"},
                 status_code=400,

server/db_manager.py

@@ -35,7 +35,7 @@ class DBManager:
         """创建数据库表"""
         # 确保所有表都会被创建
         Base.metadata.create_all(self.engine)
-        logger.info("数据库表创建/检查完成")
+        logger.info("Database tables created/checked")
 
     def get_session(self):
         """获取数据库会话"""
@@ -50,7 +50,7 @@ class DBManager:
             session.commit()
         except Exception as e:
             session.rollback()
-            logger.error(f"数据库操作失败: {e}")
+            logger.error(f"Database operation failed: {e}")
             raise
         finally:
             session.close()

src/core/graphbase.py

@@ -26,21 +26,22 @@ class GraphDatabase:
 
         # 尝试加载已保存的图数据库信息
         if not self.load_graph_info():
-            logger.debug(f"未找到已保存的图数据库信息，将创建新的配置")
+            logger.debug("创建新的图数据库配置")
 
         self.start()
 
     def start(self):
         if not config.enable_knowledge_graph or not config.enable_knowledge_base:
             return
+
         uri = os.environ.get("NEO4J_URI", "bolt://localhost:7687")
         username = os.environ.get("NEO4J_USERNAME", "neo4j")
         password = os.environ.get("NEO4J_PASSWORD", "0123456789")
-        logger.info(f"Connecting to Neo4j at {uri}/{self.kgdb_name}")
+        logger.info(f"Connecting to Neo4j: {uri}/{self.kgdb_name}")
         try:
             self.driver = GD.driver(f"{uri}/{self.kgdb_name}", auth=(username, password))
             self.status = "open"
-            logger.info(f"Connected to Neo4j at {uri}/{self.kgdb_name}, {self.get_graph_info(self.kgdb_name)}")
+            logger.info(f"Connected to Neo4j: {self.get_graph_info(self.kgdb_name)}")
             # 连接成功后保存图数据库信息
             self.save_graph_info(self.kgdb_name)
         except Exception as e:

src/core/knowledgebase.py

@@ -416,8 +416,10 @@ class KnowledgeBase:
                 logger.error(f"处理文件 {file_path} 时出错: {e}")
                 file_info_dict["status"] = "failed"
                 file_info_dict["error"] = str(e)
+                raise e
 
-            file_infos[file_id] = file_info_dict
+            finally:
+                file_infos[file_id] = file_info_dict
 
         return file_infos
 

src/plugins/_ocr.py

@@ -163,6 +163,11 @@ class OCRPlugin:
         import requests
         import json
 
+        health_check_response = requests.get(f"{mineru_ocr_uri}/health", timeout=5)
+        if health_check_response.status_code != 200 or health_check_response.json().get("status") != "healthy":
+            logger.error("Mineru OCR service health check failed.")
+            raise RuntimeError("Mineru OCR service health check failed. Please check the log use `docker logs mineru-api`")
+
         # 读取PDF文件
         with open(pdf_path, 'rb') as f:
             files = {'file': f}

web/src/apis/base.js

@@ -73,6 +73,8 @@ export async function apiRequest(url, options = {}, requiresAuth = false) {
         throw new Error('未授权，请先登录')
       } else if (response.status === 403) {
         throw new Error('没有权限执行此操作')
+      } else if (response.status === 500) {
+        throw new Error('Server 500 Error, please check the log use `docker logs api-dev`')
       }
 
       throw new Error(errorMessage)