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/services/searcher.py

@@ -213,11 +213,7 @@ def _search_semantic(
     arxiv_ids = [c["arxiv_id"] for c in candidates]
     distance_map = {c["arxiv_id"]: c["distance"] for c in candidates}
 
-    stmt = (
-        select(Paper)
-        .where(Paper.arxiv_id.in_(arxiv_ids))
-        .options(*PAPER_FULL_LOAD)
-    )
+    stmt = select(Paper).where(Paper.arxiv_id.in_(arxiv_ids)).options(*PAPER_FULL_LOAD)
     if tag:
         stmt = stmt.where(Paper.tags.any(tag=tag))
 
@@ -298,9 +294,7 @@ def _load_papers_by_ids(
 
     papers = (
         db.execute(
-            select(Paper)
-            .where(Paper.id.in_(paper_ids))
-            .options(*PAPER_FULL_LOAD)
+            select(Paper).where(Paper.id.in_(paper_ids)).options(*PAPER_FULL_LOAD)
         )
         .unique()
         .scalars()