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
@@ -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;