""" 新闻资讯模块迁移脚本 将老的新闻分类(news_categories)和新闻文章(news_posts)迁移到新的栏目(columns)和内容(articles)体系 执行方式: cd backend && .\venv-win\Scripts\python.exe migrate_news_to_columns.py """ import sys import time from sqlalchemy import create_engine, text from pypinyin import lazy_pinyin import re DB_PATH = "sqlite:///cms.db" engine = create_engine(DB_PATH) def generate_slug(text: str) -> str: """将中文标题转为拼音 slug""" if not text: return f"article-{int(time.time())}" py = lazy_pinyin(text, style=0) slug = "-".join(py) slug = re.sub(r"[^a-zA-Z0-9-]", "", slug).lower() slug = re.sub(r"-+", "-", slug).strip("-") return slug or f"article-{int(time.time())}" def migrate(): with engine.connect() as conn: trans = conn.begin() try: print("=" * 60) print("新闻资讯模块迁移开始") print("=" * 60) # ========== 阶段 0: 清空现有数据 ========== print("\n[阶段 0] 清空 columns 和 articles 表...") # 先删 articles(外键依赖) result = conn.execute(text("DELETE FROM articles")) print(f" 已清空 articles 表") # 再删 columns result = conn.execute(text("DELETE FROM columns")) print(f" 已清空 columns 表") # ========== 阶段 1: 创建"新闻资讯"顶级栏目 ========== print("\n[阶段 1] 创建'新闻资讯'顶级栏目...") result = conn.execute(text(""" INSERT INTO columns (name, slug, description, parent_id, sort_order, is_enabled, display_mode, layout_style) VALUES (:name, :slug, :description, NULL, 0, 1, 'list', 'left') """), { "name": "新闻资讯", "slug": "xin-wen-zi-xun", "description": "" }) # 获取刚插入的 ID row = conn.execute(text("SELECT id FROM columns WHERE slug = 'xin-wen-zi-xun'")).fetchone() news_top_id = row[0] print(f" 顶级栏目 ID: {news_top_id}, slug: xin-wen-zi-xun") # ========== 阶段 2: 迁移启用的 news_categories ========== print("\n[阶段 2] 迁移启用的 news_categories(作为二级子栏目)...") categories = conn.execute(text(""" SELECT id, name, slug, description, sort_order, is_enabled FROM news_categories WHERE is_enabled = 1 ORDER BY id """)).fetchall() cat_id_map = {} # old_id -> new_id for cat in categories: old_id, name, slug, description, sort_order, is_enabled = cat result = conn.execute(text(""" INSERT INTO columns (name, slug, description, parent_id, sort_order, is_enabled, display_mode, layout_style) VALUES (:name, :slug, :description, :parent_id, :sort_order, :is_enabled, 'list', 'left') """), { "name": name, "slug": slug, "description": description or "", "parent_id": news_top_id, "sort_order": sort_order or 0, "is_enabled": is_enabled }) new_id = conn.execute(text("SELECT last_insert_rowid()")).fetchone()[0] cat_id_map[old_id] = new_id print(f" [{old_id}] {name} -> new_id={new_id}") print(f" 共迁移 {len(cat_id_map)} 个子栏目") # ========== 阶段 3: 迁移 news_posts ========== print("\n[阶段 3] 迁移 news_posts(重新生成拼音 slug)...") posts = conn.execute(text(""" SELECT id, title, slug, content, content_type, summary, cover_image, category_id, is_published, is_draft, view_count, created_at, updated_at FROM news_posts ORDER BY id """)).fetchall() # 用于检测 slug 冲突 used_slugs = set() slug_conflicts = 0 migrated_count = 0 skipped_count = 0 for post in posts: (old_id, title, old_slug, content, content_type, summary, cover_image, category_id, is_published, is_draft, view_count, created_at, updated_at) = post # 只迁移已发布或有效的文章 # 状态映射: is_published=True -> published, 否则 draft status = "published" if is_published else "draft" # 重新生成拼音 slug base_slug = generate_slug(title) new_slug = base_slug suffix = 1 while new_slug in used_slugs: new_slug = f"{base_slug}-{suffix}" suffix += 1 slug_conflicts += 1 used_slugs.add(new_slug) # 栏目映射 new_column_id = cat_id_map.get(category_id) if not new_column_id: print(f" 跳过 post_id={old_id},category_id={category_id} 未找到映射") skipped_count += 1 continue conn.execute(text(""" INSERT INTO articles (title, slug, content, content_type, summary, cover_image, column_id, author, status, sort_order, view_count, created_at, updated_at) VALUES (:title, :slug, :content, :content_type, :summary, :cover_image, :column_id, '', :status, 0, :view_count, :created_at, :updated_at) """), { "title": title, "slug": new_slug, "content": content or "", "content_type": content_type or "markdown", "summary": summary or "", "cover_image": cover_image or "", "column_id": new_column_id, "status": status, "view_count": view_count or 0, "created_at": created_at, "updated_at": updated_at }) migrated_count += 1 if migrated_count % 50 == 0: print(f" 已迁移 {migrated_count} / {len(posts)} 篇...") print(f" 共迁移 {migrated_count} 篇,跳过 {skipped_count} 篇,slug 冲突解决 {slug_conflicts} 次") # ========== 阶段 4: 验证 ========== print("\n[阶段 4] 验证迁移结果...") col_count = conn.execute(text("SELECT COUNT(*) FROM columns")).fetchone()[0] art_count = conn.execute(text("SELECT COUNT(*) FROM articles")).fetchone()[0] top_col = conn.execute(text("SELECT id, name, slug FROM columns WHERE parent_id IS NULL")).fetchone() child_cols = conn.execute(text("SELECT COUNT(*) FROM columns WHERE parent_id IS NOT NULL")).fetchone()[0] print(f" columns 表总数: {col_count}") print(f" 顶级栏目: {top_col}") print(f" 子栏目数: {child_cols}") print(f" articles 表总数: {art_count}") # 每个子栏目下的文章数 print("\n 各子栏目文章分布:") for row in conn.execute(text(""" SELECT c.name, COUNT(a.id) as cnt FROM columns c LEFT JOIN articles a ON a.column_id = c.id WHERE c.parent_id IS NOT NULL GROUP BY c.id ORDER BY cnt DESC """)): print(f" {row[0]}: {row[1]} 篇") trans.commit() print("\n" + "=" * 60) print("迁移成功完成!") print("=" * 60) except Exception as e: trans.rollback() print(f"\n迁移失败,已回滚: {e}") import traceback traceback.print_exc() sys.exit(1) if __name__ == "__main__": migrate()