refactor: restructure services and add image/pdf extraction utilities

- Add image_extractor, pdf_downloader, pi_client, trends services
- Add shared utils module
- Refactor summarizer, embedder, routes for cleaner separation
- Update tests to match new service structure

app/main.py

@@ -2,14 +2,13 @@
 
 import logging
 import os
+from contextlib import asynccontextmanager
 
 from fastapi import FastAPI
 from fastapi.staticfiles import StaticFiles
-from starlette.staticfiles import StaticFiles as StarletteStaticFiles
 
 from app.config import settings
-from app.database import engine
-from app.models import init_db
+from app.database import engine, init_db
 from app.routes.admin import router as admin_router
 from app.routes.compare import router as compare_router
 from app.routes.pages import router as pages_router
@@ -24,11 +23,30 @@ logging.basicConfig(
 logger = logging.getLogger(__name__)
 
 
+@asynccontextmanager
+async def lifespan(app: FastAPI):
+    """应用生命周期管理：启动与关闭。"""
+    # ── startup ──
+    from app.services.scheduler import start_scheduler
+    from app.services.embedder import init_chroma
+
+    start_scheduler()
+    init_chroma()
+
+    yield
+
+    # ── shutdown ──
+    from app.services.scheduler import stop_scheduler
+
+    stop_scheduler()
+
+
 def create_app() -> FastAPI:
     app = FastAPI(
         title="HF Daily Papers",
         description="HuggingFace Daily Papers — 中文论文导览站",
         version="0.1.0",
+        lifespan=lifespan,
     )
 
     # 确保数据目录存在
@@ -65,23 +83,6 @@ def create_app() -> FastAPI:
     app.include_router(trends_router)
     app.include_router(compare_router)
 
-    # 调度器（Phase 4）
-    @app.on_event("startup")
-    async def _start_scheduler():
-        from app.services.scheduler import start_scheduler
-        start_scheduler()
-
-    # Phase 5: 初始化 ChromaDB
-    @app.on_event("startup")
-    async def _init_chroma():
-        from app.services.embedder import init_chroma
-        init_chroma()
-
-    @app.on_event("shutdown")
-    async def _stop_scheduler():
-        from app.services.scheduler import stop_scheduler
-        stop_scheduler()
-
     return app