feat(sites): on-demand site detail page content (INT-2)

- SiteReleaseDetail.page: description, MediaInfo/BDInfo, screenshots, IMDb / Douban / TMDB ids, file list, tags, uploader, plus a parse probe (labels found) for layout-change diagnosis. Generic NexusPHP rules (field label -> next sibling) starting from the MovieClaw detail selectors; site overrides under umt.detail.page (HHClub: div labels). M-Team builds the same structure from descr BBCode + mediainfo. - Fetched only on demand (drawer open without cache, "复核详情", release.inspect); content in kv_cache region site.detail for the site detail cache period (>= 24 h); within the window the site is not requested again. External ids persist in source_release.raw._detail, which list snapshots do not overwrite. - Screenshots only as proxy paths /api/v1/img/site-detail/{rid}/{i}: public-address check on every hop, no site credentials, magic-byte image check, 8 MB cap. Poster proxy shares the same helper. - Identity: site ids count as explicit-id evidence; when TMDB is queried, a TMDB id goes straight to details and an IMDb id to /find (no title search on a hit). Douban only for local recall. - Release drawer "站点详情页" section; fixtures and tests for HHClub div layout, rowhead layout, BBCode, cache/proxy/identity end to end. Co-Authored-By: Claude Opus 5.5 <[email protected]>
2026-09-29 16:21:17 +08:00
1 parent 6e9e8efec2
commit 04cffcd1f1
@@ -258,6 +258,9 @@
        "imdb": "a[href*='imdb.com']::attr(href)",
        "douban": "a[href*='douban.com']::attr(href)",
        "fileList": "td.rowfollow > ul > li::text"
+      },
+      "page": {
+        "labelSelectors": ["div.font-bold.leading-6", "span.font-bold"]
      }
    },
    "vipLevels": ["贵宾", "VIP"],
@@ -141,6 +141,16 @@ const UmtNexusSchema = z.strictObject({
          fileList: z.string().optional(),
        })
        .optional(),
+      page: z
+        .strictObject({
+          labelSelectors: z.array(z.string()).optional(),
+          mediaInfo: z.string().optional(),
+          screenshots: z.string().optional(),
+          tags: z.string().optional(),
+          uploader: z.string().optional(),
+          files: z.string().optional(),
+        })
+        .optional(),
    })
    .optional(),
  downloadPath: z.string().min(1),
@@ -0,0 +1,108 @@
+/**
+ * 详情页扩展内容的通用文本规则(NexusPHP HTML 与 M-Team BBCode 共用):MediaInfo 识别、截图过滤、TMDB 链接。
+ */
+import type { SiteDetailPage } from "./types.ts";
+import { extractDoubanId, extractImdbId } from "./util.ts";
+
+/** 表情 / 图标 / 旗帜 / 头像等小图,不算截图 */
+const NOT_SCREENSHOT =
+  /(smilies|smiley|emoji|\/icons?\/|pic\/(?:trans|cattrans|flag)|\/flags?\/|avatar|\.svg(?:$|\?))/i;
+
+export const MAX_SCREENSHOTS = 20;
+export const MAX_FILES = 500;
+
+/** MediaInfo / BDInfo 文本:有概览段落 + 视频段落,且至少 5 行「键 : 值」 */
+export function looksLikeMediaInfo(text: string): boolean {
+  if (!/(^|\n)\s*(General|概览|常规|DISC INFO|Disc Title|Complete name)\b/i.test(text)) return false;
+  if (!/(^|\n)\s*(Video|视频)\b/i.test(text)) return false;
+  return (text.match(/^[^\n:：]{2,40}[ \t]*[:：][ \t]*\S/gm) ?? []).length >= 5;
+}
+
+export function extractTmdb(text: string | null | undefined): SiteDetailPage["tmdb"] {
+  const m = text ? /themoviedb\.org\/(movie|tv)\/(\d+)/.exec(text) : null;
+  return m ? { type: m[1] as "movie" | "tv", id: m[2]! } : null;
+}
+
+/** 图片地址转绝对 http(s) 地址并过滤非截图;无效返回 null */
+export function screenshotUrl(src: string | null | undefined, base: string): string | null {
+  if (!src || src.startsWith("data:")) return null;
+  let u: URL;
+  try {
+    u = new URL(src.trim(), base);
+  } catch {
+    return null;
+  }
+  if (u.protocol !== "http:" && u.protocol !== "https:") return null;
+  if (NOT_SCREENSHOT.test(u.pathname)) return null;
+  return u.toString();
+}
+
+/** 多行文本规整:去 \r、行尾空白,连续空行压成一行 */
+export function tidyText(text: string): string {
+  return text
+    .replace(/\r/g, "")
+    .replace(/ /g, " ")
+    .split("\n")
+    .map((l) => l.replace(/\s+$/, ""))
+    .join("\n")
+    .replace(/\n{3,}/g, "\n\n")
+    .trim();
+}
+
+export function emptyDetailPage(): SiteDetailPage {
+  return {
+    descriptionText: null,
+    mediaInfo: null,
+    screenshots: [],
+    imdbId: null,
+    doubanId: null,
+    tmdb: null,
+    files: [],
+    tags: [],
+    uploader: null,
+    probe: { labels: [], hasDescription: false, imageCount: 0 },
+  };
+}
+
+/**
+ * M-Team 详情(BBCode 简介 + 独立 mediainfo 字段)→ 扩展内容。
+ * 截图取 [img] 标签;简介去掉 BBCode 标记与 MediaInfo 引用块。
+ */
+export function pageFromBbcode(
+  descr: string | null | undefined,
+  base: string,
+  extra: { mediaInfo?: string | null; imdbId?: string | null; doubanId?: string | null; tags?: string[] } = {},
+): SiteDetailPage {
+  const page = emptyDetailPage();
+  const src = descr ?? "";
+  const imgs: string[] = [];
+  for (const m of src.matchAll(/\[img(?:=([^\]]+))?\]([^[]*)\[\/img\]|\[img=([^\]]+)\]/gi)) {
+    const u = screenshotUrl((m[2] || m[1] || m[3] || "").trim(), base);
+    if (u && !imgs.includes(u)) imgs.push(u);
+  }
+  page.probe.imageCount = imgs.length;
+  page.screenshots = imgs.slice(0, MAX_SCREENSHOTS);
+  let mediaInfo = extra.mediaInfo?.trim() || null;
+  let body = src.replace(/\[img(?:=[^\]]+)?\][^[]*\[\/img\]|\[img=[^\]]+\]/gi, "");
+  // [quote] / [code] / [mediainfo] 块里的 MediaInfo
+  body = body.replace(
+    /\[(quote|code|mediainfo|box|hide|spoiler)(=[^\]]*)?\]([\s\S]*?)\[\/\1\]/gi,
+    (all, _t, _a, inner) => {
+      const text = tidyText(String(inner).replace(/\[\/?[a-z*]+(?:=[^\]]*)?\]/gi, ""));
+      if (looksLikeMediaInfo(text)) {
+        mediaInfo ??= text;
+        return "";
+      }
+      return all;
+    },
+  );
+  const text = tidyText(body.replace(/\[\/?[a-z*]+(?:=[^\]]*)?\]/gi, ""));
+  page.probe.hasDescription = !!src.trim();
+  page.descriptionText = text ? text.slice(0, 20_000) : null;
+  page.mediaInfo = mediaInfo ? mediaInfo.slice(0, 50_000) : null;
+  page.imdbId = extra.imdbId ?? extractImdbId(src) ?? null;
+  page.doubanId = extra.doubanId ?? extractDoubanId(src) ?? null;
+  page.tmdb = extractTmdb(src);
+  page.tags = [...new Set(extra.tags ?? [])].slice(0, 20);
+  return page;
+}
@@ -90,6 +90,8 @@ export {
  type SiteCapabilities,
  type SiteCategory,
  type SiteCredentialKind,
+  type SiteDetailFile,
+  type SiteDetailPage,
  SiteError,
  type SiteErrorKind,
  type SiteHealth,
@@ -11,6 +11,7 @@
 */
 import { redactUrl } from "../../rss/redact.ts";
 import { buildHttpContext, type SiteAdapterDeps } from "../deps.ts";
+import { pageFromBbcode } from "../detail-text.ts";
 import { readBody, siteFetch } from "../http.ts";
 import { healthFromError } from "../spider/adapter.ts";
 import { promotionFromFactors } from "../spider/map.ts";
@@ -154,6 +155,12 @@ export function createMteamAdapter(input: MteamDefinition, deps: SiteAdapterDeps
      const [row] = parseMteamResult([data], def.domain);
      const r: SiteReleaseDetail = mteamTorrentToRelease(row!, data, def, http.now().toISOString());
      if (data.descr) r.descriptionText = String(data.descr).slice(0, 20_000);
+      r.page = pageFromBbcode(data.descr, def.domain, {
+        mediaInfo: typeof data.mediainfo === "string" ? data.mediainfo : null,
+        imdbId: r.imdbId ?? null,
+        doubanId: r.doubanId ?? null,
+        tags: r.tags,
+      });
      return r;
    },

@@ -46,6 +46,7 @@ export interface MteamTorrentJson {
  labels?: string | number | null;
  labelsNew?: string[] | null;
  descr?: string | null;
+  mediainfo?: string | null;
  status?: {
    seeders?: string | number | null;
    leechers?: string | number | null;
@@ -0,0 +1,107 @@
+import { HHANCLUB_DETAIL_PAGE_FULL } from "@umt/testkit";
+import { describe, expect, it } from "vitest";
+import { hhanclubDefinition } from "../definitions/index.ts";
+import { looksLikeMediaInfo, pageFromBbcode } from "../detail-text.ts";
+import { getSiteDefinition } from "../registry.ts";
+import { parseDetailPage } from "./detail.ts";
+import type { IndexerDefinition } from "./schema.ts";
+
+const OBSERVED = "2026-09-29T00:00:00.000Z";
+
+describe("详情页扩展内容(HHClub div 模板)", () => {
+  const { kind, detail } = parseDetailPage(HHANCLUB_DETAIL_PAGE_FULL, hhanclubDefinition, "91002", OBSERVED);
+  const page = detail!.page!;
+
+  it("外部 ID、上传者、标签", () => {
+    expect(kind).toBe("ok");
+    expect(page).toMatchObject({
+      imdbId: "tt14452776",
+      doubanId: "36000001",
+      tmdb: { type: "tv", id: "136315" },
+      uploader: "uploader01",
+      tags: ["官方", "中字", "完结"],
+    });
+    expect(detail!.imdbId).toBe("tt14452776");
+    expect(detail!.tags).toEqual(["官方", "中字", "完结"]);
+  });
+
+  it("MediaInfo 从简介的代码块取出,简介里不再重复", () => {
+    expect(page.mediaInfo).toMatch(/^General\nUnique ID : 1234\n/);
+    expect(page.mediaInfo).toContain("Format : HEVC");
+    expect(page.descriptionText).toContain("◎译　　名　示例剧集");
+    expect(page.descriptionText).not.toContain("Unique ID");
+  });
+
+  it("截图:相对地址补全、优先 data-src、过滤表情与占位图", () => {
+    expect(page.screenshots).toEqual([
+      "https://hhanclub.net/attachments/202609/poster.jpg",
+      "https://img.example-host.com/i/shot1.png",
+      "https://img.example-host.com/i/shot2.jpg",
+    ]);
+  });
+
+  it("文件清单:「名称 (大小)」", () => {
+    expect(page.files).toEqual([
+      { name: "Example.Show.S01E01.mkv", sizeBytes: Math.round(1.04 * 1024 ** 3) },
+      { name: "Example.Show.S01E02.mkv", sizeBytes: Math.round(1.02 * 1024 ** 3) },
+    ]);
+  });
+
+  it("诊断:记录页面上的字段标签", () => {
+    expect(page.probe.labels).toEqual(expect.arrayContaining(["标题", "副标题", "标签", "上传者", "文件列表"]));
+    expect(page.probe.hasDescription).toBe(true);
+  });
+});
+
+describe("详情页扩展内容(标准 NexusPHP rowhead 表格)", () => {
+  const def = getSiteDefinition("audiences") as IndexerDefinition;
+  const html = `<html><body><a href="logout.php">退出</a><h1 id="top">Some.Movie.2024.1080p.BluRay.x264-GRP</h1>
+<table>
+<tr><td class="rowhead">副标题</td><td class="rowfollow">某电影</td></tr>
+<tr><td class="rowhead">基本信息</td><td class="rowfollow"><b>大小：</b>8.00 GB &nbsp; <b>类型:</b> 电影</td></tr>
+<tr><td class="rowhead">标签</td><td class="rowfollow"><span>中字</span><span>DIY</span></td></tr>
+<tr><td class="rowhead">上传者</td><td class="rowfollow"><i>匿名</i></td></tr>
+<tr><td class="rowhead">简介</td><td class="rowfollow"><div id="kdescr">一段简介<br>第二行
+<div class="codemain">DISC INFO:<br>Disc Title: X<br>Size: 1<br>Protection: AACS<br>Playlist: 00800.MPLS<br>Length: 2:00:00<br>Video: MPEG-4 AVC<br>Total Bitrate: 30 Mbps</div>
+<img src="https://img.example-host.com/a.png"></div></td></tr>
+</table>
+<table id="filelist"><tr><td>Some.Movie.mkv</td><td>8.00 GB</td></tr></table></body></html>`;
+
+  it("rowhead 标签、BDInfo、#filelist 表格", () => {
+    const { detail } = parseDetailPage(html, def, "1", OBSERVED);
+    const page = detail!.page!;
+    expect(page.tags).toEqual(["中字", "DIY"]);
+    expect(page.uploader).toBe("匿名");
+    expect(page.mediaInfo).toContain("Disc Title: X");
+    expect(page.descriptionText).toBe("一段简介\n第二行");
+    expect(page.screenshots).toEqual(["https://img.example-host.com/a.png"]);
+    expect(page.files).toEqual([{ name: "Some.Movie.mkv", sizeBytes: 8 * 1024 ** 3 }]);
+    expect(page.imdbId).toBeNull();
+  });
+});
+
+describe("M-Team BBCode 简介", () => {
+  it("截图取 [img],MediaInfo 取引用块,外部 ID 从正文取", () => {
+    const descr = `[img]https://img.example-host.com/p.jpg[/img]
+◎简介 这是简介 https://www.imdb.com/title/tt15239678/
+[quote]General
+Unique ID : 1
+Complete name : a.mkv
+Format : Matroska
+File size : 1 GiB
+
+Video
+Format : AVC[/quote]
+[img=https://img.example-host.com/s.png]`;
+    const page = pageFromBbcode(descr, "https://kp.m-team.cc/", { tags: ["中字"] });
+    expect(page.screenshots).toEqual(["https://img.example-host.com/p.jpg", "https://img.example-host.com/s.png"]);
+    expect(page.mediaInfo).toMatch(/^General\n/);
+    expect(page.descriptionText).toBe("◎简介 这是简介 https://www.imdb.com/title/tt15239678/");
+    expect(page.imdbId).toBe("tt15239678");
+    expect(page.tags).toEqual(["中字"]);
+  });
+
+  it("MediaInfo 识别不误判普通简介", () => {
+    expect(looksLikeMediaInfo("General\n一段文字\nVideo")).toBe(false);
+  });
+});
@@ -0,0 +1,208 @@
+/**
+ * 详情页扩展内容(简介 / MediaInfo / 截图 / IMDb·豆瓣·TMDB / 文件清单 / 标签 / 上传者)—— 本项目扩展。
+ * 以 MovieClaw detail_*_css(description / imdb / douban / file_list)为起点,其余按 NexusPHP details.php 公开结构
+ * 做通用规则:字段标签(td.rowhead 或站点配置的 div 标签)的下一个兄弟元素是值。站点特化写在定义 `umt.detail.page`。
+ * 文件清单只取页面上已有的内容,不为它额外请求 viewfilelist.php(请求克制)。
+ */
+import type { CheerioAPI } from "cheerio";
+import {
+  emptyDetailPage,
+  extractTmdb,
+  looksLikeMediaInfo,
+  MAX_FILES,
+  MAX_SCREENSHOTS,
+  screenshotUrl,
+  tidyText,
+} from "../detail-text.ts";
+import type { SiteDetailFile, SiteDetailPage } from "../types.ts";
+import { extractDoubanId, extractImdbId } from "../util.ts";
+import { mcSelect } from "./mc-selector.ts";
+import type { IndexerDefinition } from "./schema.ts";
+import { parseSizeMp } from "./spider.ts";
+
+type DetailConf = NonNullable<IndexerDefinition["umt"]["detail"]>;
+type El = ReturnType<CheerioAPI>;
+
+const LABELS = {
+  description: ["简介", "Description"],
+  mediaInfo: ["MediaInfo", "媒体信息", "BDInfo"],
+  tags: ["标签", "Tags"],
+  uploader: ["上传者", "发布者", "发布人", "上传人", "Uploader"],
+  files: ["文件列表", "文件清单", "File list", "Files"],
+};
+
+/** 简介里可能装 MediaInfo 的块 */
+const MI_BLOCKS = "fieldset, div.codemain, pre, code, blockquote, .spoiler-content, .hide, div.nexus-media-info-raw";
+
+const normLabel = (t: string) =>
+  t
+    .replace(/\s+/g, "")
+    .replace(/[:：]+$/, "")
+    .toLowerCase();
+
+/** 保留换行的文本:<br> 与块级元素换行 */
+function blockText($: CheerioAPI, el: El): string {
+  const c = el.clone();
+  c.find("script, style").remove();
+  c.find("br").replaceWith("\n");
+  c.find("div, p, li, tr, pre, fieldset, blockquote, h1, h2, h3, h4, h5, h6").each((_, n) => {
+    $(n).append("\n");
+  });
+  return tidyText(c.text());
+}
+
+const SIZE_TAIL = /^(.*?)\s*[(（[]\s*([\d.,]+\s*[KMGTP]?i?B)\s*[)）\]]\s*$/i;
+const HAS_EXT = /\.[A-Za-z0-9]{2,5}$/;
+
+function fileFromText(text: string): SiteDetailFile | null {
+  const t = text.replace(/\s+/g, " ").trim();
+  if (!t) return null;
+  const m = SIZE_TAIL.exec(t);
+  if (m?.[1]) return { name: m[1], sizeBytes: parseSizeMp(m[2]!) || null };
+  return HAS_EXT.test(t) ? { name: t, sizeBytes: null } : null;
+}
+
+function filesFrom($: CheerioAPI, rows: El): SiteDetailFile[] {
+  const out: SiteDetailFile[] = [];
+  rows.each((_, n) => {
+    if (out.length >= MAX_FILES) return;
+    const row = $(n);
+    const cells = row.children("td");
+    if (cells.length >= 2) {
+      const name = cells.first().text().replace(/\s+/g, " ").trim();
+      const size = cells.eq(1).text().trim();
+      if (name && HAS_EXT.test(name) && /\d/.test(size)) out.push({ name, sizeBytes: parseSizeMp(size) || null });
+      return;
+    }
+    const f = fileFromText(row.text());
+    if (f) out.push(f);
+  });
+  return out;
+}
+
+export function extractDetailPage($: CheerioAPI, conf: DetailConf, base: string): SiteDetailPage {
+  const page = emptyDetailPage();
+  const p = conf.page ?? {};
+
+  // 字段标签 → 值元素
+  const pairs: { label: string; value: El }[] = [];
+  const addLabels = (sel: string, valueFor: (e: El) => El) => {
+    $(sel).each((_, n) => {
+      const e = $(n);
+      const label = normLabel(e.text());
+      if (!label || label.length > 16) return;
+      const value = valueFor(e);
+      if (value.length) pairs.push({ label, value });
+    });
+  };
+  addLabels(conf.rowLabel, (e) => e.nextAll(conf.rowValue).first());
+  for (const sel of p.labelSelectors ?? []) addLabels(sel, (e) => e.next());
+  const byLabel = (names: string[]): El | null =>
+    pairs.find((x) => names.some((n) => x.label.startsWith(normLabel(n))))?.value ?? null;
+  page.probe.labels = [...new Set(pairs.map((x) => x.label))].slice(0, 40);
+
+  // 简介容器
+  const descSel = conf.fields?.description ?? conf.description;
+  let desc: El | null = descSel ? $(descSel).first() : null;
+  if (!desc?.length) desc = byLabel(LABELS.description);
+  page.probe.hasDescription = !!desc?.length;
+
+  // MediaInfo:站点选择器 → 字段标签 → 简介里的引用 / 代码块(取最内层文本,整块从简介里去掉)
+  let mediaInfoEl: El | null = null;
+  let miBlockIndex = -1;
+  if (p.mediaInfo) {
+    const el = $(p.mediaInfo).first();
+    if (el.length) mediaInfoEl = el;
+  }
+  mediaInfoEl ??= byLabel(LABELS.mediaInfo);
+  const blocks = desc?.length ? desc.find(MI_BLOCKS) : null;
+  if (!mediaInfoEl && blocks) {
+    blocks.each((i, n) => {
+      if (miBlockIndex >= 0) return;
+      if (looksLikeMediaInfo(blockText($, $(n)))) miBlockIndex = i;
+    });
+    if (miBlockIndex >= 0) {
+      const outer = blocks.eq(miBlockIndex);
+      const inner = outer
+        .find(MI_BLOCKS)
+        .filter((_, n) => looksLikeMediaInfo(blockText($, $(n))))
+        .last();
+      mediaInfoEl = inner.length ? inner : outer;
+    }
+  }
+  if (mediaInfoEl) {
+    const t = blockText($, mediaInfoEl);
+    if (t) page.mediaInfo = t.slice(0, 50_000);
+  }
+
+  // 截图
+  const imgs: string[] = [];
+  const imgEls = p.screenshots ? $(p.screenshots) : desc?.length ? desc.find("img") : null;
+  imgEls?.each((_, n) => {
+    const e = $(n);
+    const u = screenshotUrl(e.attr("data-src") ?? e.attr("data-original") ?? e.attr("src"), base);
+    if (u && !imgs.includes(u)) imgs.push(u);
+  });
+  page.probe.imageCount = imgs.length;
+  page.screenshots = imgs.slice(0, MAX_SCREENSHOTS);
+
+  // 简介纯文本:去掉 MediaInfo 块与图片
+  if (desc?.length) {
+    const c = desc.clone();
+    if (miBlockIndex >= 0) c.find(MI_BLOCKS).eq(miBlockIndex).remove();
+    c.find("img").remove();
+    const t = blockText($, c);
+    if (t) page.descriptionText = t.slice(0, 20_000);
+  }
+
+  // 外部 ID:页面上的链接优先,其次简介正文
+  const descHtml = desc?.html() ?? "";
+  const f = conf.fields ?? {};
+  page.imdbId =
+    extractImdbId(mcSelect($, null, f.imdb ?? "a[href*='imdb.com']::attr(href)")) ?? extractImdbId(descHtml) ?? null;
+  page.doubanId =
+    extractDoubanId(mcSelect($, null, f.douban ?? "a[href*='douban.com']::attr(href)")) ??
+    extractDoubanId(descHtml) ??
+    null;
+  page.tmdb = extractTmdb($("a[href*='themoviedb.org']").first().attr("href")) ?? extractTmdb(descHtml);
+
+  // 文件清单:站点选择器 → MovieClaw file_list(去伪元素)→ 字段标签 → #filelist 表格
+  const fileSel = p.files ?? f.fileList?.replace(/::text$/, "");
+  let files = fileSel ? filesFrom($, $(fileSel)) : [];
+  if (!files.length) {
+    const v = byLabel(LABELS.files);
+    if (v) files = filesFrom($, v.find("li, tr"));
+  }
+  if (!files.length) files = filesFrom($, $("#filelist tr"));
+  page.files = files;
+
+  // 标签
+  const tagEl = p.tags ? $(p.tags) : byLabel(LABELS.tags);
+  if (tagEl?.length) {
+    const parts = tagEl.find("span, a").length
+      ? tagEl
+          .find("span, a")
+          .map((_, n) => $(n).text().trim())
+          .get()
+      : tagEl.text().split(/[\s,，|/]+/);
+    page.tags = [...new Set(parts.map((t) => t.trim()).filter((t) => t && t.length <= 20))].slice(0, 20);
+  }
+
+  // 上传者
+  const upEl = p.uploader ? $(p.uploader).first() : byLabel(LABELS.uploader);
+  if (upEl?.length) {
+    const a = upEl.is("a") ? upEl : upEl.find("a[href*='userdetails.php']").first();
+    const name = (a.length ? a.text() : upEl.text()).replace(/\s+/g, " ").trim();
+    if (name) page.uploader = name.slice(0, 64);
+  }
+  if (!page.uploader) {
+    for (const x of pairs) {
+      const m = /(?:上传者|发布者|发布人|Uploader)\s*[:：]\s*(\S+)/i.exec(x.value.text());
+      if (m?.[1]) {
+        page.uploader = m[1].slice(0, 64);
+        break;
+      }
+    }
+  }
+  return page;
+}
@@ -6,6 +6,7 @@
 import type { CheerioAPI } from "cheerio";
 import type { SiteReleaseDetail } from "../types.ts";
 import { extractDoubanId, extractImdbId, parseCount, parseSiteTime } from "../util.ts";
+import { extractDetailPage } from "./detail-page.ts";
 import { categoryName, promotionFromFactors } from "./map.ts";
 import { mcSelect } from "./mc-selector.ts";
 import type { IndexerDefinition } from "./schema.ts";
@@ -115,5 +116,10 @@ export function parseDetailPage(
  if (imdb) detail.imdbId = imdb;
  const douban = extractDoubanId(mcSelect($, null, f.douban)) ?? extractDoubanId(descHtml);
  if (douban) detail.doubanId = douban;
+  const page = extractDetailPage($, conf, def.domain);
+  page.imdbId ??= detail.imdbId ?? null;
+  page.doubanId ??= detail.doubanId ?? null;
+  if (page.tags.length) detail.tags = page.tags;
+  detail.page = page;
  return { kind: "ok", detail };
 }
@@ -198,6 +198,20 @@ export interface UmtSiteExtension {
    fields?: Partial<
      Record<"title" | "subtitle" | "size" | "description" | "download" | "imdb" | "douban" | "fileList", string>
    >;
+    /**
+     * 详情页扩展内容(简介 / MediaInfo / 截图 / 文件 / 标签 / 上传者)的站点特化,均可省略(走通用规则,见 detail-page.ts)。
+     * 选择器为普通 CSS。
+     */
+    page?: {
+      /** 除 rowLabel 外的字段标签元素(div 版详情页:标签元素的下一个兄弟元素是值) */
+      labelSelectors?: string[];
+      mediaInfo?: string;
+      screenshots?: string;
+      tags?: string;
+      uploader?: string;
+      /** 文件清单行(每行一个文件,文本形如「名称 (大小)」或两列表格) */
+      files?: string;
+    };
  };
  /** 下载路径(按资源 ID 拼接) */
  downloadPath: string;
@@ -60,8 +60,34 @@ export interface SiteRelease {
  observedAt: string;
 }

+/** 详情页文件清单一项;大小未知为 null */
+export interface SiteDetailFile {
+  name: string;
+  sizeBytes: number | null;
+}
+
+/**
+ * 详情页扩展内容(只在按需抓详情时得到,服务端放 kv_cache,不进 source_release 列)。
+ * 截图是站点页面上的原始图片地址,展示时经服务端图片代理,不让浏览器直连外站。
+ */
+export interface SiteDetailPage {
+  /** 简介纯文本(已去掉 MediaInfo 块) */
+  descriptionText: string | null;
+  mediaInfo: string | null;
+  screenshots: string[];
+  imdbId: string | null;
+  doubanId: string | null;
+  tmdb: { type: "movie" | "tv"; id: string } | null;
+  files: SiteDetailFile[];
+  tags: string[];
+  uploader: string | null;
+  /** 解析诊断:页面上找到的字段标签、是否有简介容器、图片数(用于站点改版排查,不展示) */
+  probe: { labels: string[]; hasDescription: boolean; imageCount: number };
+}
+
 export interface SiteReleaseDetail extends SiteRelease {
  descriptionText?: string;
+  page?: SiteDetailPage;
 }

 export interface SiteListFilters {
@@ -348,6 +348,16 @@ export class TmdbClient {
    return [...m, ...t].sort((a, b) => (b.popularity ?? 0) - (a.popularity ?? 0));
  }

+  /** /find/{imdb_id}?external_source=imdb_id:IMDb ID → 电影 / 剧集候选(站点详情页给出的 IMDb 链接走这里) */
+  async findByImdb(imdbId: string): Promise<MetadataCandidate[]> {
+    if (!/^tt\d{5,10}$/.test(imdbId)) return [];
+    const lang = this.langs[0]!;
+    const p = await this.get<Json>(`/find/${imdbId}`, { external_source: "imdb_id", language: lang });
+    const movies = (Array.isArray(p.movie_results) ? p.movie_results : []) as Json[];
+    const tvs = (Array.isArray(p.tv_results) ? p.tv_results : []) as Json[];
+    return [...movies.map((r) => this.fromBrief("movie", r, lang)), ...tvs.map((r) => this.fromBrief("tv", r, lang))];
+  }
+
  /** 详情:多语言标题、别名(alternative_titles)、外部 ID、季列表(剧集)。不存在抛 notfound。 */
  async details(type: MediaEntityType, id: string | number): Promise<MetadataCandidate> {
    const [first, ...rest] = await Promise.all(
@@ -636,6 +636,24 @@ export const WorkbenchReleasePage = z.object({
 });
 export type WorkbenchReleasePage = z.infer<typeof WorkbenchReleasePage>;

+/**
+ * 站点详情页扩展内容(按需抓详情时得到,服务端缓存 24 小时;过期后需再次复核详情)。
+ * 截图地址是服务端图片代理路径(/api/v1/img/site-detail/{releaseId}/{index}),不直连外站。
+ */
+export const SiteReleasePage = z.object({
+  fetchedAt: Timestamp,
+  descriptionText: z.string().nullable(),
+  mediaInfo: z.string().nullable(),
+  screenshots: z.array(z.object({ url: z.string(), host: z.string() })),
+  imdbId: z.string().nullable(),
+  doubanId: z.string().nullable(),
+  tmdb: z.object({ type: z.enum(["movie", "tv"]), id: z.string() }).nullable(),
+  files: z.array(z.object({ name: z.string(), sizeBytes: z.number().int().nullable() })),
+  tags: z.array(z.string()),
+  uploader: z.string().nullable(),
+});
+export type SiteReleasePage = z.infer<typeof SiteReleasePage>;
+
 export const WorkbenchReleaseDetail = z.object({
  release: WorkbenchRelease,
  observations: z.array(
@@ -657,5 +675,9 @@ export const WorkbenchReleaseDetail = z.object({
  }),
  /** 刷新详情失败时保留既有数据并给出原因 */
  refreshError: z.string().nullable(),
+  /** 详情页扩展内容;没抓过、缓存已过期或站点不支持详情页时为 null */
+  page: SiteReleasePage.nullable(),
+  /** 站点能否抓详情页(false 时前端不自动抓) */
+  detailSupported: z.boolean(),
 });
 export type WorkbenchReleaseDetail = z.infer<typeof WorkbenchReleaseDetail>;
@@ -80,3 +80,31 @@ export const HHANCLUB_LIST_PAGE_2026 = `<!DOCTYPE html><html><head><title>HHCLUB
 </div>
 <div class="pagination"><a href="?page=0">1-100</a><a href="?page=1">101-200</a></div>
 </body></html>`;
+
+/**
+ * 详情页扩展内容(简介 / MediaInfo / 截图 / 外部链接 / 标签 / 上传者 / 文件清单)fixture:
+ * 结构按 HHClub div 模板(「标签」div + 相邻值 div)与 NexusPHP #kdescr 简介编写,数值均为示例。
+ */
+export const HHANCLUB_DETAIL_PAGE_FULL = `<!DOCTYPE html><html><head><title>HHCLUB :: 种子详情</title></head><body>
+<div id="user-info-panel"><a href="userdetails.php?id=7" class="VIP_Name">bob</a> <a href="logout.php">退出</a></div>
+<div class="grid">
+  <div class="font-bold leading-6">标题</div><div class="font-bold leading-6">Example.Show.S01.2026.2160p.WEB-DL.H265.AAC-HHWEB</div>
+  <div class="font-bold leading-6">副标题</div><div class="font-bold leading-6">示例剧集 | 全12集 | 4K</div>
+  <div><span class="font-bold">大小：</span><span>12.50 GB</span></div>
+  <div class="font-bold leading-6">标签</div><div class="font-bold leading-6"><span class="tag">官方</span><span class="tag">中字</span><span class="tag">完结</span></div>
+  <div class="font-bold leading-6">上传者</div><div class="font-bold leading-6"><a href="userdetails.php?id=321">uploader01</a></div>
+  <div><a href="download.php?id=91002">下载种子</a></div>
+  <div class="font-bold leading-6">简介</div>
+  <div id="kdescr">
+    <img src="attachments/202609/poster.jpg" alt="poster"><br>
+    ◎译　　名　示例剧集<br>◎年　　代　2026<br>
+    ◎IMDb链接 <a href="https://www.imdb.com/title/tt14452776/">https://www.imdb.com/title/tt14452776/</a><br>
+    ◎豆瓣链接 <a href="https://movie.douban.com/subject/36000001/">https://movie.douban.com/subject/36000001/</a><br>
+    ◎TMDB <a href="https://www.themoviedb.org/tv/136315">themoviedb</a><br>
+    <img src="pic/smilies/1.gif">
+    <fieldset><legend>MediaInfo</legend><div class="codemain">General<br>Unique ID : 1234<br>Complete name : Example.Show.S01E01.mkv<br>Format : Matroska<br>File size : 1.04 GiB<br>Duration : 45 min<br><br>Video<br>ID : 1<br>Format : HEVC<br>Width : 3 840 pixels</div></fieldset>
+    <img src="https://img.example-host.com/i/shot1.png"><img data-src="https://img.example-host.com/i/shot2.jpg" src="pic/trans.gif">
+  </div>
+  <div class="font-bold leading-6">文件列表</div>
+  <div class="font-bold leading-6"><ul><li>Example.Show.S01E01.mkv (1.04 GB)</li><li>Example.Show.S01E02.mkv (1.02 GB)</li></ul></div>
+</div></body></html>`;
@@ -144,7 +144,7 @@ function detailBody(r: MockNexusRelease): string {
 <tr><td class="rowhead">副标题</td><td class="rowfollow" align="left">${esc(r.subtitle ?? "")}</td></tr>
 <tr><td class="rowhead">基本信息</td><td class="rowfollow" align="left"><b>大小：</b>${esc(r.sizeText)}&nbsp;&nbsp;&nbsp;<b>类型:</b> ${esc(r.catName)}&nbsp;&nbsp;发布于 <span title="${r.uploadedAt}">${esc(r.relativeTime ?? r.uploadedAt)}</span></td></tr>
 <tr><td class="rowhead">同伴</td><td class="rowfollow" align="left">${r.seeders.replace(/,/g, "")}个做种者 | ${r.leechers.replace(/,/g, "")}个下载者</td></tr>
-<tr><td class="rowhead">简介</td><td class="rowfollow" align="left"><div id="kdescr">${esc(r.description ?? "暂无简介")}</div></td></tr>
+<tr><td class="rowhead">简介</td><td class="rowfollow" align="left"><div id="kdescr">${r.descriptionHtml ?? esc(r.description ?? "暂无简介")}</div></td></tr>
 </table>`;
 }

@@ -25,6 +25,8 @@ export interface MockNexusRelease {
  hr?: boolean;
  uploader?: string;
  description?: string;
+  /** 简介原样 HTML(优先于 description;用于详情页扩展内容测试) */
+  descriptionHtml?: string;
  /** 下载返回磁力而不是 .torrent */
  magnet?: string;
 }
@@ -19,7 +19,7 @@ const notFound: MockReply = {
  json: { success: false, status_code: 34, status_message: "The resource you requested could not be found." },
 };

-/** mock TMDB v3:/3/search/{movie,tv}、/3/{movie,tv}/:id(append_to_response)、/3/tv/:id/season/:n */
+/** mock TMDB v3:/3/search/{movie,tv}、/3/find/:imdbId、/3/{movie,tv}/:id(append_to_response)、/3/tv/:id/season/:n */
 export async function createMockTmdb(opts: MockTmdbOptions = {}): Promise<MockTmdb> {
  const fx = opts.fixtures ?? TMDB_FIXTURES;
  const apiKey = opts.apiKey ?? "test-tmdb-key";
@@ -84,6 +84,23 @@ export async function createMockTmdb(opts: MockTmdbOptions = {}): Promise<MockTm
          return { json: { page: 1, results, total_pages: 1, total_results: results.length } };
        },
      ],
+      [
+        /^\/3\/find\/(tt\d+)$/,
+        (m, q) => {
+          if (q.external_source !== "imdb_id") return { json: { movie_results: [], tv_results: [] } };
+          const hit = fx.filter((f) => f.external_ids.imdb_id === m[1]);
+          return {
+            json: {
+              movie_results: hit
+                .filter((f) => f.type === "movie")
+                .map((f) => ({ ...brief(f, q.language), media_type: "movie" })),
+              tv_results: hit
+                .filter((f) => f.type === "tv")
+                .map((f) => ({ ...brief(f, q.language), media_type: "tv" })),
+            },
+          };
+        },
+      ],
      [
        /^\/3\/tv\/(\d+)\/season\/(\d+)$/,
        (m, q) => {