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
2026-06-06 00:00:55 +08:00
parent ba9afa212c
commit 85c4cfb9e8
22 changed files with 843 additions and 780 deletions
@@ -0,0 +1,81 @@
+"""趋势统计服务 — 按日论文数量、热门标签、Upvotes 分布、总结完成率。"""
+
+from __future__ import annotations
+
+from datetime import date, timedelta
+
+from sqlalchemy import text
+from sqlalchemy.orm import Session
+
+
+def get_trends_data(db: Session) -> dict:
+    """从 DB 聚合趋势数据。"""
+    thirty_days_ago = (date.today() - timedelta(days=30)).isoformat()
+
+    # 1. 按日论文数量（近 30 天）
+    daily_rows = db.execute(text("""
+        SELECT paper_date, COUNT(*) as cnt
+        FROM papers
+        WHERE paper_date >= :start_date
+        GROUP BY paper_date
+        ORDER BY paper_date ASC
+    """), {"start_date": thirty_days_ago}).fetchall()
+    daily_counts = [
+        {"date": str(row[0]), "count": row[1]}
+        for row in daily_rows
+    ]
+
+    # 2. 热门标签 Top 20
+    tag_rows = db.execute(text("""
+        SELECT tag, COUNT(*) as cnt
+        FROM paper_tags
+        GROUP BY tag
+        ORDER BY cnt DESC
+        LIMIT 20
+    """)).fetchall()
+    top_tags = [
+        {"tag": row[0], "count": row[1]}
+        for row in tag_rows
+    ]
+
+    # 3. Upvotes 分布
+    upvote_rows = db.execute(text("""
+        SELECT
+            CASE
+                WHEN upvotes >= 100 THEN '100+'
+                WHEN upvotes >= 50 THEN '50-99'
+                WHEN upvotes >= 20 THEN '20-49'
+                WHEN upvotes >= 10 THEN '10-19'
+                WHEN upvotes >= 5 THEN '5-9'
+                ELSE '0-4'
+            END as bucket,
+            COUNT(*) as cnt
+        FROM papers
+        GROUP BY bucket
+        ORDER BY MIN(upvotes) DESC
+    """)).fetchall()
+    upvotes_dist = [
+        {"range": row[0], "count": row[1]}
+        for row in upvote_rows
+    ]
+
+    # 4. 总结完成率
+    summary_rows = db.execute(text("""
+        SELECT
+            COALESCE(ss.status, 'none') as status,
+            COUNT(*) as cnt
+        FROM papers p
+        LEFT JOIN summary_status ss ON ss.paper_id = p.id
+        GROUP BY status
+    """)).fetchall()
+    summary_completion = [
+        {"status": row[0], "count": row[1]}
+        for row in summary_rows
+    ]
+
+    return {
+        "daily_counts": daily_counts,
+        "top_tags": top_tags,
+        "upvotes_dist": upvotes_dist,
+        "summary_completion": summary_completion,
+    }