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/models.py

@@ -12,6 +12,7 @@ from sqlalchemy import (
     String,
     Text,
     UniqueConstraint,
+    select,
 )
 from sqlalchemy.orm import joinedload, relationship
 
@@ -93,7 +94,7 @@ class PaperAuthor(Base):
 
     id = Column(Integer, primary_key=True, autoincrement=True)
     paper_id = Column(
-        Integer, ForeignKey("papers.id", ondelete="CASCADE"), nullable=False
+        Integer, ForeignKey("papers.id", ondelete="CASCADE"), nullable=False, index=True
     )
     name = Column(String, nullable=False)
     position = Column(Integer, default=0)
@@ -108,7 +109,7 @@ class PaperTag(Base):
 
     id = Column(Integer, primary_key=True, autoincrement=True)
     paper_id = Column(
-        Integer, ForeignKey("papers.id", ondelete="CASCADE"), nullable=False
+        Integer, ForeignKey("papers.id", ondelete="CASCADE"), nullable=False, index=True
     )
     tag = Column(String, nullable=False)
     source = Column(String, default="hf")
@@ -155,7 +156,7 @@ class SummaryStatus(Base):
     paper_id = Column(
         Integer, ForeignKey("papers.id", ondelete="CASCADE"), nullable=False
     )
-    status = Column(String, nullable=False, default="pending")
+    status = Column(String, nullable=False, default="pending", index=True)
     quality = Column(String)
     error_type = Column(String)
     error = Column(Text)
@@ -219,7 +220,7 @@ class UserReadingStatus(Base):
     paper_id = Column(
         Integer, ForeignKey("papers.id", ondelete="CASCADE"), nullable=False
     )
-    status = Column(String, nullable=False, default="unread")
+    status = Column(String, nullable=False, default="unread", index=True)
     updated_at = Column(DateTime, nullable=False)
 
     paper = relationship("Paper", back_populates="reading_status")
@@ -271,3 +272,15 @@ PAPER_FULL_LOAD = (
     joinedload(Paper.bookmark),
     joinedload(Paper.reading_status),
 )
+
+
+def get_paper_by_arxiv_id(db, arxiv_id: str, *, load=PAPER_DEFAULT_LOAD):
+    """按 arxiv_id 查询论文（带关联加载），未找到返回 None。"""
+    stmt = select(Paper).where(Paper.arxiv_id == arxiv_id).options(*load)
+    return db.execute(stmt).unique().scalar_one_or_none()
+
+
+def get_paper_by_id(db, paper_id: int, *, load=PAPER_DEFAULT_LOAD):
+    """按主键查询论文（带关联加载），未找到返回 None。"""
+    stmt = select(Paper).where(Paper.id == paper_id).options(*load)
+    return db.execute(stmt).unique().scalar_one_or_none()