fix issue

app/api/v1/images.py

@@ -21,7 +21,7 @@ from sqlalchemy import select, func, text, delete
 from sqlalchemy.ext.asyncio import AsyncSession
 
 from app.config import settings
-from app.database import get_db
+from app.database import get_db, async_session_factory
 from app.models.base import OCRImage
 from app.services.ocr_service import OCRService
 from app.services.llm_service import LLMService
@@ -299,15 +299,11 @@ async def search_images(
     keyword: str,
     limit: int = Query(5, ge=1, le=100),
     sort: str = Query("time_desc"),
-    db: AsyncSession = Depends(get_db),
 ):
     """
     AI 搜索图片（SSE 流式返回）。
-    1. 从 DB 取所有图片，按指定排序
-    2. 每批 5 条发给 Qwen3-8B 判断
-    3. 并发池最多 10 个请求
-    4. 匹配结果实时 SSE 推送
-    5. 找够 limit 个或遍历完 DB 后结束
+    注意：不使用 Depends(get_db)，因为 StreamingResponse 的生命周期
+    超出 request scope，需要在生成器内部自行管理 session。
     """
 
     async def event_stream():
@@ -320,6 +316,8 @@ async def search_images(
         # 发送开始事件
         yield f"data: {json.dumps({'type': 'start', 'keyword': keyword, 'limit': limit}, ensure_ascii=False)}\n\n"
 
+        # 在生成器内部自行管理 session，避免 get_db 提前关闭
+        async with async_session_factory() as db:
             # 先获取总记录数
             count_result = await db.execute(
                 select(func.count(OCRImage.id)).where(OCRImage.status == "completed")

app/main.py

@@ -8,6 +8,7 @@ from __future__ import annotations
 import asyncio
 import json
 import logging
+import os
 import threading
 from contextlib import asynccontextmanager
 from pathlib import Path
@@ -193,7 +194,11 @@ async def health_check():
 @app.get("/data/images/{image_name}", tags=["前端"])
 async def serve_image(image_name: str):
     """访问 images 目录下的图片文件（用于前端预览）"""
-    img_path = settings.images_dir / image_name
+    # 防止路径遍历攻击
+    safe_name = os.path.basename(image_name)
+    if safe_name != image_name or ".." in image_name:
+        return JSONResponse(status_code=400, content={"detail": "非法文件名"})
+    img_path = settings.images_dir / safe_name
     if not img_path.exists():
         return JSONResponse(status_code=404, content={"detail": "图片不存在"})
     return FileResponse(str(img_path))

app/models/base.py

@@ -81,6 +81,7 @@ class OCRImage(Base, TimestampMixin):
         String(20), default="pending", comment="状态: pending/processing/completed/failed"
     )
     blocks: Mapped[str | None] = mapped_column(Text, nullable=True, comment="OCR 文本块（JSON）")
+    error_message: Mapped[str | None] = mapped_column(Text, nullable=True, comment="错误信息")
 
     @property
     def blocks_list(self) -> list[dict]:

app/services/llm_service.py

@@ -286,10 +286,12 @@ class LLMService:
 
         try:
             from openai import AsyncOpenAI
+            http_client = httpx.AsyncClient(timeout=60.0)
+            try:
                 client = AsyncOpenAI(
                     api_key=settings.MINIMAX_API_KEY,
                     base_url=settings.MINIMAX_BASE_URL,
-                http_client=httpx.AsyncClient(timeout=60.0),
+                    http_client=http_client,
                 )
                 response = await client.chat.completions.create(
                     model=settings.MINIMAX_CHAT_MODEL,
@@ -317,6 +319,8 @@ class LLMService:
                         matched_ids.append(articles[num - 1]["id"])
 
                 return matched_ids
+            finally:
+                await http_client.aclose()
         except Exception as exc:
             logger.warning("LLM 批量判断失败: %s", exc)
             return []

app/services/ocr_service.py

@@ -102,7 +102,7 @@ class PaddleOCRProvider(OCRProviderBase):
     async def recognize(self, image_path: str) -> OCRResult:
         """使用 PaddleOCR 识别图片"""
         self._init_ocr()
-        loop = asyncio.get_event_loop()
+        loop = asyncio.get_running_loop()
 
         def _run_ocr():
             result = self._ocr.ocr(image_path, cls=True)
@@ -164,7 +164,7 @@ class AliyunOCRProvider(OCRProviderBase):
 
         import asyncio
 
-        loop = asyncio.get_event_loop()
+        loop = asyncio.get_running_loop()
 
         def _call_api():
             from alibabacloud_tea_openapi.models import Config
@@ -249,7 +249,7 @@ class TencentOCRProvider(OCRProviderBase):
         import base64
         import json
 
-        loop = asyncio.get_event_loop()
+        loop = asyncio.get_running_loop()
 
         def _call_api():
             from tencentcloud.common import credential