新增业务原子化；新增fireccrawl的baseurl修改

backend/services/data_service.py

@@ -93,7 +93,62 @@ class DataService:
         except:
             return url
 
-    # ... (保持 get_task_monitor_data, save_chunks, search 等方法不变) ...
+    def get_task_by_root_url(self, url: str):
+        """返回已存在任务的 id（如果没有则返回 None）"""
+        clean_url = normalize_url(url)
+        with self.db.engine.connect() as conn:
+            row = conn.execute(select(self.db.tasks.c.id).where(self.db.tasks.c.root_url == clean_url)).fetchone()
+            return row[0] if row else None
+
+    def create_task_with_urls(self, url: str, urls: list[str]):
+        """
+        原子化：在单个事务中创建任务并批量插入 URL（去重）。
+        如果任务已存在，则不会创建新任务，而是把新的 URL 去重后插入到该任务下。
+
+        返回：{"task_id": int, "is_new_task": bool, "added": int}
+        """
+        clean_root = normalize_url(url)
+        clean_urls = [normalize_url(u) for u in urls]
+        added_count = 0
+        with self.db.engine.begin() as conn:
+            # 1. 尝试获取已存在任务
+            existing = conn.execute(select(self.db.tasks.c.id).where(self.db.tasks.c.root_url == clean_root)).fetchone()
+            if existing:
+                task_id = existing[0]
+                is_new = False
+            else:
+                # 创建新任务并返回 id
+                stmt = insert(self.db.tasks).values(root_url=clean_root).returning(self.db.tasks.c.id)
+                task_id = conn.execute(stmt).fetchone()[0]
+                is_new = True
+
+            # 2. 批量插入 urls，跳过已存在项
+            # 使用临时表或单条插入均可，这里逐条检查以保证兼容性
+            for u in clean_urls:
+                try:
+                    exists_q = select(self.db.queue.c.id).where(
+                        and_(self.db.queue.c.task_id == task_id, self.db.queue.c.url == u)
+                    )
+                    if not conn.execute(exists_q).fetchone():
+                        conn.execute(insert(self.db.queue).values(task_id=task_id, url=u, status='pending'))
+                        added_count += 1
+                except Exception:
+                    # 单条插入失败时忽略，继续处理剩余 URL
+                    continue
+
+        return {"task_id": task_id, "is_new_task": is_new, "added": added_count}
+
+    def delete_task(self, task_id: int):
+        """删除任务与相关队列与分片（谨慎使用，主要用于回滚）"""
+        with self.db.engine.begin() as conn:
+            try:
+                conn.execute(text("DELETE FROM chunks WHERE task_id = :tid"), {"tid": task_id})
+                conn.execute(text("DELETE FROM queue WHERE task_id = :tid"), {"tid": task_id})
+                conn.execute(text("DELETE FROM tasks WHERE id = :tid"), {"tid": task_id})
+                return True
+            except Exception as e:
+                logger.error(f"Failed to delete task {task_id}: {e}")
+                return False
 
     def get_task_monitor_data(self, task_id: int):
         """[数据库层监控] 获取持久化的任务状态"""