UHD_Media_Tool/apps/server/test/scrape/scrape-layout-mp.test.ts

import { mkdir, readdir, readFile, stat, writeFile } from "node:fs/promises";
import { dirname, join, relative } from "node:path";
import type { TmdbRichDetail, TmdbRichSeasonDetail } from "@umt/adapters";
import type { ScrapeMatch } from "@umt/contracts";
import { afterAll, beforeAll, describe, expect, it } from "vitest";
import { queueScrape, runScrapeRecord } from "../../src/modules/acquisition-organize/services/scrape.ts";
import { createFmHarness, type FmHarness, pgAvailable } from "../file-manager/harness.ts";

// SCR-2 刮削文件布局差分:同一组目录结构,MP ScrapingChain.scrape_metadata(gen-scrape-layout-mp.py 生成
// scrape-layout-mp.json)写哪些文件 ⇔ 我们就地刮削写哪些文件。偏离只允许下面 DEVIATIONS 里登记的几类
// (理由见 docs/runbook/scrape.md「与 MP 的差分」)。

const skip = !(await pgAvailable());
const MP = JSON.parse(await readFile(new URL("./scrape-layout-mp.json", import.meta.url), "utf8")) as Record<
  string,
  { target: string; ok: boolean; written: string[] }
>;

const SHOW = "人生切割术 (2022)";
const EP1 = `${SHOW}/Season 1/人生切割术 - S01E01 - 有关地狱的好消息.mkv`;
const EP2 = `${SHOW}/Season 1/人生切割术 - S01E02 - 半循环.mkv`;
const MOVIE_FILE = "星际穿越 (2014)/星际穿越 (2014) - 1080p.mkv";
const LOOSE = "Interstellar.2014.1080p.BluRay.mkv";
const TREE = [EP1, EP2, MOVIE_FILE, LOOSE, "Heat (1995)/BDMV/index.bdmv"];

/** 登记的偏离:把 MP 的输出换算成我们应有的输出 */
const DEVIATIONS: { id: string; apply: (files: string[], scenario: string) => string[] }[] = [
  // 1. 不写 backdrop.jpg 副本,只写 fanart.jpg(Emby / Jellyfin / Kodi 都认 fanart)
  { id: "no_backdrop_copy", apply: (fs) => fs.filter((f) => !/(^|\/)backdrop\.jpg$/.test(f)) },
  // 2. 剧照 <视频名>-thumb.jpg(Kodi 约定;MP 写 <视频名>.jpg)
  {
    id: "episode_thumb_suffix",
    apply: (fs) => fs.map((f) => (/ - S\d+E\d+ - [^/]+\.jpg$/.test(f) ? f.replace(/\.jpg$/, "-thumb.jpg") : f)),
  },
  // 3. 视频直接在登记目录根:图片写 <视频名>-poster.jpg / -fanart.jpg,不往根里放 poster.jpg(MP 会)
  {
    id: "root_movie_images",
    apply: (fs, s) =>
      s !== "movie_file_in_root"
        ? fs
        : fs.map((f) => (f === "poster.jpg" || f === "fanart.jpg" ? `${LOOSE.replace(/\.mkv$/, "")}-${f}` : f)),
  },
];

const detail = (entityType: "movie" | "tv", externalId: string): TmdbRichDetail =>
  ({
    entityType,
    externalId,
    title: "x",
    titleLanguage: "zh-CN",
    titleZh: "x",
    titleZhLanguage: "zh-CN",
    originalTitle: null,
    originalLanguage: "en",
    overview: null,
    overviewLanguage: null,
    tagline: null,
    year: 2022,
    releaseDate: null,
    lastAirDate: null,
    genres: [],
    voteAverage: null,
    voteCount: null,
    status: null,
    runtime: null,
    seasonCount: null,
    episodeCount: null,
    networks: [],
    originCountry: [],
    posterPath: "/poster.jpg",
    backdropPath: "/backdrop.jpg",
    cast: [],
    crew: [],
    externalIds: { imdb: null, tvdb: null, wikidata: null },
    translations: [],
    alternativeTitles: [],
    seasons: [],
    languages: ["zh-CN"],
  }) satisfies TmdbRichDetail;

const season1: TmdbRichSeasonDetail = {
  tvId: "95396",
  seasonNumber: 1,
  name: null,
  nameLanguage: null,
  overview: null,
  overviewLanguage: null,
  airDate: null,
  posterPath: "/season.jpg",
  fetchedLanguages: ["zh-CN"],
  episodes: [1, 2].map((n) => ({
    seasonNumber: 1,
    episodeNumber: n,
    name: null,
    nameLanguage: null,
    overview: null,
    overviewLanguage: null,
    airDate: null,
    runtime: null,
    stillPath: "/still.jpg",
    voteAverage: null,
    voteCount: null,
  })),
};

async function files(root: string): Promise<Set<string>> {
  const out = new Set<string>();
  const walk = async (d: string) => {
    for (const n of await readdir(d).catch(() => [])) {
      if (n.startsWith(".")) continue;
      const p = join(d, n);
      if ((await stat(p)).isDirectory()) await walk(p);
      else out.add(relative(root, p));
    }
  };
  await walk(root);
  return out;
}

describe.skipIf(skip)("刮削文件布局与 MP 差分", () => {
  let h: FmHarness;
  beforeAll(async () => {
    h = await createFmHarness(
      "scrlayout",
      {
        ao: {
          scrapeImage: async () => ({ content: new Uint8Array([0xff, 0xd8, 0xff]), mime: "image/jpeg" }),
          scrapeTmdbDetail: async (type, id) => detail(type, id),
          scrapeTmdbSeason: async (_tv, n) => (n === 1 ? season1 : null),
        },
      },
      { runTransfers: false },
    );
  });
  afterAll(async () => h?.close());

  for (const [scenario, mp] of Object.entries(MP)) {
    it(`${scenario}:${mp.target}`, async () => {
      // 每个场景一份独立目录树;"视频直接在根"的场景用下载根(整个根就是这棵树)
      const inRoot = scenario === "movie_file_in_root";
      const base = inRoot ? h.dl : join(h.lib, scenario);
      for (const f of TREE) {
        await mkdir(dirname(join(base, f)), { recursive: true });
        await writeFile(join(base, f), "video");
      }
      const before = await files(base);
      const tv = mp.target.startsWith(SHOW);
      const match: ScrapeMatch = {
        mediaEntityId: null,
        provider: "tmdb",
        externalId: tv ? "95396" : "157336",
        kind: tv ? "tv" : "movie",
        title: "x",
        year: null,
        source: "chosen",
      };
      const q = await queueScrape(h.module.deps().ao, {
        ...(inRoot
          ? { rootId: h.dlRoot, rootPath: h.dl, libraryRootId: null }
          : { rootId: h.libRoot, rootPath: h.lib, libraryRootId: h.libraryRootId }),
        path: inRoot ? mp.target : `${scenario}/${mp.target}`,
        match,
        mode: "overwrite",
        trigger: "manual",
      });
      const rec = await runScrapeRecord(h.module.deps().ao, q.recordId);
      expect(rec.status).toBe("succeeded");
      const after = await files(base);
      const ours = [...after].filter((f) => !before.has(f)).sort();
      const expected = DEVIATIONS.reduce((fs, d) => d.apply(fs, scenario), [...mp.written]).sort();
      expect(ours).toEqual(expected);
    });
  }
});