fix: 优化后端数据

crawler/realtime_conflict_service.py

@@ -333,7 +333,7 @@ def _extract_and_merge_panel_data(items: list) -> None:
         from datetime import timezone
         merged_any = False
         # 规则模式可多处理几条（无 Ollama）；AI 模式限制 5 条避免调用过多
-        limit = 10 if os.environ.get("CLEANER_AI_DISABLED", "0") == "1" else 5
+        limit = 25 if os.environ.get("CLEANER_AI_DISABLED", "0") == "1" else 10
         for it in items[:limit]:
             text = (it.get("title", "") or "") + " " + (it.get("summary", "") or "")
             if len(text.strip()) < 20:
@@ -383,6 +383,40 @@ async def _periodic_fetch() -> None:
 # ==========================
 # API 接口
 # ==========================
+@app.post("/crawler/backfill")
+def crawler_backfill():
+    """从 situation_update 重新解析并合并战损/报复等数据，用于修复历史数据未提取的情况"""
+    if not os.path.exists(DB_PATH):
+        return {"ok": False, "error": "db not found"}
+    try:
+        from db_merge import merge
+        if os.environ.get("CLEANER_AI_DISABLED", "0") == "1":
+            from extractor_rules import extract_from_news
+        else:
+            from extractor_ai import extract_from_news
+        conn = sqlite3.connect(DB_PATH, timeout=10)
+        rows = conn.execute(
+            "SELECT id, timestamp, category, summary FROM situation_update ORDER BY timestamp DESC LIMIT 50"
+        ).fetchall()
+        conn.close()
+        merged = 0
+        for r in rows:
+            uid, ts, cat, summary = r
+            text = ((cat or "") + " " + (summary or "")).strip()
+            if len(text) < 20:
+                continue
+            try:
+                extracted = extract_from_news(text, timestamp=ts)
+                if extracted and merge(extracted, db_path=DB_PATH):
+                    merged += 1
+            except Exception:
+                pass
+        _notify_node()
+        return {"ok": True, "processed": len(rows), "merged": merged}
+    except Exception as e:
+        return {"ok": False, "error": str(e)}
+
+
 @app.get("/crawler/status")
 def crawler_status():
     """爬虫状态：用于排查数据更新链路"""