116 lines
5.9 KiB
TypeScript
116 lines
5.9 KiB
TypeScript
/**
|
||
* # 脚本 增量导入原站新增内容（dump 之后发布/更新的 post 与 page）
|
||
*
|
||
* 背景：MariaDB dump 快照截至 2026-10-29（最大 post id 161733），之后原站陆续发布新文章
|
||
* （161855/161748 等出现在内嵌模板文章列表、/blog/ 归档）。本脚本走原站 REST API
|
||
* （wp-json/wp/v2/posts|pages，_embed 特色图与术语）增量补齐；Elementor 原始数据
|
||
* （_elementor_data）不经 REST 暴露，导入的 html 走博客详情 content-html 分支渲染。
|
||
*
|
||
* 用法：bun scripts/import-live-posts.ts [--dry] [after=ISO日期]
|
||
* after 缺省 = 本地已发布内容最大 updated_at/published_at（dump 快照点）
|
||
* 幂等：slug 已存在则跳过（与 dump 增量合并，不覆盖既有行）。
|
||
*/
|
||
import { SQL } from "bun";
|
||
|
||
const LIVE = "https://www.dusuniot.com";
|
||
const sql = new SQL(process.env.DATABASE_URL || "postgres://dusuniot:dusuniot@127.0.0.1:25432/dusuniot", { max: 1 });
|
||
const dry = process.argv.includes("--dry");
|
||
const afterArg = process.argv.find((a) => a.startsWith("after="))?.slice(6);
|
||
|
||
/** # 函数 取本地内容快照水位（已有内容的最大发布时间，向后取 1 天重叠窗口） */
|
||
async function watermark(): Promise<string> {
|
||
if (afterArg) return afterArg;
|
||
const r = await sql`select max(published_at)::text as t from contents`;
|
||
return r[0]?.t ? new Date(new Date(r[0].t).getTime() - 86400_000).toISOString() : "2026-10-29T00:00:00";
|
||
}
|
||
|
||
/** # 函数 剥 HTML 标签并解码常见实体（摘要/纯文本用） */
|
||
function stripHtml(html: string): string {
|
||
return html
|
||
.replace(/<[^>]*>/g, " ")
|
||
.replace(/&[a-z#0-9]+;/gi, " ")
|
||
.replace(/\s+/g, " ")
|
||
.trim();
|
||
}
|
||
|
||
/** # 函数 原站媒体 URL → 本地 /uploads 路径（不在 uploads 的返回 null） */
|
||
function localUploadPath(url: string): string | null {
|
||
const i = url.indexOf("/uploads/");
|
||
return i === -1 ? null : url.slice(i + "/uploads/".length);
|
||
}
|
||
|
||
/** # 函数 确保媒体行存在（按 legacy_id 查，缺则插入），返回 media id */
|
||
async function ensureMedia(m: { id: number; source_url: string; alt_text?: string; title?: { rendered?: string }; mime_type?: string; media_details?: { width?: number; height?: number } } | null | undefined): Promise<number | null> {
|
||
if (!m?.source_url) return null;
|
||
const found = await sql`select id from media where legacy_id = ${m.id} limit 1`;
|
||
if (found[0]) return found[0].id as number;
|
||
const path = localUploadPath(m.source_url);
|
||
if (!path) return null;
|
||
const [res]: any = await sql`insert into media (path, url, mime_type, width, height, alt_text, title, ref_count, legacy_id, created_at)
|
||
values (${path}, ${"/uploads/" + path}, ${m.mime_type || "image/jpeg"}, ${m.media_details?.width ?? null}, ${m.media_details?.height ?? null},
|
||
${m.alt_text || ""}, ${stripHtml(m.title?.rendered || "") || null}, 0, ${m.id}, now()) returning id`;
|
||
return res.id;
|
||
}
|
||
|
||
/** # 函数 WP 分类 slug → 本地 categories.id（按 slug 匹配；无则忽略该术语） */
|
||
async function categoryId(wpSlug: string): Promise<number | null> {
|
||
const r = await sql`select id from categories where slug = ${wpSlug} limit 1`;
|
||
return r[0]?.id ?? null;
|
||
}
|
||
|
||
/** # 函数 导入一批 REST 内容（post 或 page），返回新增条数 */
|
||
async function importBatch(kind: "posts" | "pages", after: string): Promise<number> {
|
||
const url = `${LIVE}/wp-json/wp/v2/${kind}?after=${encodeURIComponent(after)}&per_page=100&orderby=date&order=asc&_embed=wp:featuredmedia,wp:term`;
|
||
const res = await fetch(url, { headers: { "User-Agent": "Mozilla/5.0" } });
|
||
if (!res.ok) {
|
||
console.log(`SKIP(REST ${res.status}):`, url);
|
||
return 0;
|
||
}
|
||
const items = (await res.json()) as any[];
|
||
let imported = 0;
|
||
for (const p of items) {
|
||
const slug = p.slug;
|
||
const exists = await sql`select id from contents where slug = ${slug} limit 1`;
|
||
if (exists[0]) {
|
||
console.log("SKIP(已存在):", slug);
|
||
continue;
|
||
}
|
||
// WooCommerce 桩页面跳过（shop/cart/checkout/my-account 由本地路由实现）
|
||
if (["shop", "cart", "checkout", "my-account"].includes(slug)) continue;
|
||
const html: string = p.content?.rendered || "";
|
||
const excerpt: string = stripHtml(p.excerpt?.rendered || "") || stripHtml(html).slice(0, 160);
|
||
const featured = await ensureMedia(p._embedded?.["wp:featuredmedia"]?.[0]);
|
||
const type = kind === "posts" ? "post" : "page";
|
||
const published = new Date(p.date_gmt ? p.date_gmt + "Z" : p.date);
|
||
const updated = new Date(p.modified_gmt ? p.modified_gmt + "Z" : p.modified);
|
||
if (dry) {
|
||
console.log("DRY:", type, slug, published.toISOString().slice(0, 10));
|
||
continue;
|
||
}
|
||
const [row]: any = await sql`insert into contents (type, slug, title, excerpt, html, status, featured_media_id, template,
|
||
menu_order, seo_title, seo_description, meta, legacy_id, published_at, created_at, updated_at)
|
||
values (${type}, ${slug}, ${stripHtml(p.title?.rendered || "")}, ${excerpt}, ${html}, ${"published"},
|
||
${featured}, ${"default"}, 0, ${stripHtml(p.title?.rendered || "")}, ${excerpt.slice(0, 160)},
|
||
${{ elementor: null }}, ${p.id}, ${published}, ${published}, ${updated}) returning id`;
|
||
// 术语关联（REST _embedded wp:term 第一组 = categories）
|
||
for (const group of p._embedded?.["wp:term"] || []) {
|
||
for (const t of group) {
|
||
if (t.taxonomy !== "category") continue;
|
||
const cid = await categoryId(t.slug);
|
||
if (cid) await sql`insert into content_terms (content_id, category_id) values (${row.id}, ${cid}) on conflict do nothing`;
|
||
else console.log(" 未匹配分类(忽略):", t.slug, "→", slug);
|
||
}
|
||
}
|
||
imported++;
|
||
console.log("导入:", type, p.id, slug);
|
||
}
|
||
return imported;
|
||
}
|
||
|
||
const after = await watermark();
|
||
console.log("水位:", after);
|
||
const n1 = await importBatch("posts", after);
|
||
const n2 = await importBatch("pages", after);
|
||
console.log("完成 posts:", n1, "pages:", n2);
|
||
await sql.end();
|