feat: refactor summarizer and PDF extraction pipeline

- Split summarizer into summary_generator and summary_persister modules
- Refactor pdf_image_extractor to two-phase pipeline with PicoDet layout detection
- Add layout_detector service for PicoDet-S_layout_3cls integration
- Add exceptions module with ConflictError and NotFoundError
- Improve admin dashboard with better statistics and task management
- Add design review document with system optimization suggestions
- Add new tests for crawler, pdf_downloader, pipeline, and summary_utils
- Update dependencies and configuration
- Clean up dead code and improve error handling

app/exceptions.py

@@ -0,0 +1,35 @@
+"""业务异常体系 — 统一错误类型，供路由层和 service 层使用。
+
+路由层通过 main.py 的 @app.exception_handler(AppError) 统一捕获，
+转为对应 HTTP 状态码 + JSON 响应。
+"""
+
+from __future__ import annotations
+
+
+class AppError(Exception):
+    """所有业务异常的基类。"""
+
+    def __init__(self, message: str = "", *, detail: str = ""):
+        self.message = message or detail or self.__class__.__name__
+        super().__init__(self.message)
+
+
+class NotFoundError(AppError):
+    """资源不存在（404）。"""
+
+
+class ValidationError(AppError):
+    """请求参数校验失败（400）。"""
+
+
+class ExternalAPIError(AppError):
+    """外部 API 调用失败（502）。"""
+
+
+class PdfProcessError(AppError):
+    """PDF 处理错误（500）。"""
+
+
+class ConflictError(AppError):
+    """资源冲突（409）— 如锁冲突、并发任务冲突。"""