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/services/cleaner.py

@@ -16,13 +16,10 @@ from app.models import (
     Paper,
     TaskLock,
 )
+from app.utils import PAPERS_DIR, TMP_DIR
 
 logger = logging.getLogger(__name__)
 
-_DATA_DIR = Path("data")
-_TMP_DIR = _DATA_DIR / "tmp"
-_PAPERS_DIR = _DATA_DIR / "papers"
-
 # 临时文件最大保留时间（小时）
 _MAX_TMP_AGE_HOURS = 24
 
@@ -39,7 +36,7 @@ def cleanup_tmp(max_age_hours: int = _MAX_TMP_AGE_HOURS) -> dict:
     Returns:
         清理统计 {"scanned": int, "removed": int, "errors": list[str]}
     """
-    if not _TMP_DIR.exists():
+    if not TMP_DIR.exists():
         return {"scanned": 0, "removed": 0, "errors": []}
 
     now = datetime.now(timezone.utc)
@@ -48,7 +45,7 @@ def cleanup_tmp(max_age_hours: int = _MAX_TMP_AGE_HOURS) -> dict:
     removed = 0
     errors: list[str] = []
 
-    for entry in _TMP_DIR.iterdir():
+    for entry in TMP_DIR.iterdir():
         if not entry.is_dir():
             continue
         scanned += 1
@@ -147,13 +144,13 @@ async def delete_papers_by_date_range(
                 logger.warning("Failed to delete %s from ChromaDB", arxiv_id, exc_info=True)
 
             # 2. 删除本地文件 data/papers/{arxiv_id}/
-            paper_dir = _PAPERS_DIR / arxiv_id
+            paper_dir = PAPERS_DIR / arxiv_id
             if paper_dir.exists():
                 shutil.rmtree(paper_dir)
                 logger.debug("Removed paper dir: %s", paper_dir)
 
             # 3. 删除临时文件 data/tmp/{arxiv_id}/
-            tmp_dir = _TMP_DIR / arxiv_id
+            tmp_dir = TMP_DIR / arxiv_id
             if tmp_dir.exists():
                 shutil.rmtree(tmp_dir)
                 logger.debug("Removed tmp dir: %s", tmp_dir)