import { test } from "node:test";
import assert from "node:assert/strict";
import {
  archiveOrgDetailsUrl,
  archiveOrgDownloadUrl,
  archiveOrgTorrentUrl,
  archiveOrgVideoId,
  archiveOrgVideoIdFromNativeId,
  parseArchiveOrgUrl,
  dateFromMirrorFileName,
  titleFromMirrorFileName,
  youtubeIdFromFileName,
  youtubeIdFromIdentifier,
} from "./archiveOrgId";
import { extractVideoId } from "./videoId";
import { defaultWebpageUrl, detectPlatform, queueKeyForUrl } from "./platform";
import { dataDirIdForUrl } from "../ytdlp/runYtdlp";

// Run with:
//   pnpm --filter yt-dlp-transcript-common exec tsx --test lib/archiveOrgId.test.ts
//
// Every identifier, file and YouTube id here is invented.

const ITEM = "example-item";
const FILE = "Example Talk (Part 1)-AbC123xyz_9.mp4";

test("archive.org item hosts are detected; the Wayback Machine is not", () => {
  assert.equal(detectPlatform(`https://archive.org/details/${ITEM}`), "archiveorg");
  assert.equal(detectPlatform(`https://www.archive.org/details/${ITEM}`), "archiveorg");
  assert.equal(detectPlatform("https://web.archive.org/web/2020/https://example.com/"), null);
  assert.equal(queueKeyForUrl(`https://archive.org/details/${ITEM}`), "platform:archiveorg");
});

test("a whole item's id is its identifier, from details, embed and download URLs", () => {
  for (const url of [
    `https://archive.org/details/${ITEM}`,
    `https://archive.org/details/${ITEM}/`,
    `https://archive.org/details/${ITEM}?autoplay=1`,
    `https://archive.org/embed/${ITEM}`,
    `https://archive.org/download/${ITEM}`,
  ]) {
    assert.equal(extractVideoId(url), ITEM, url);
  }
});

test("a file inside an item gets a stable, filesystem-safe, unique id", () => {
  const url = archiveOrgDetailsUrl({ identifier: ITEM, file: FILE });
  assert.equal(url, `https://archive.org/details/${ITEM}/Example%20Talk%20(Part%201)-AbC123xyz_9.mp4`);
  const id = extractVideoId(url)!;
  assert.match(id, /^example-item__Example-Talk-Part-1-AbC123xyz_9-[0-9a-f]{8}$/);
  // The same file by every URL form, encoded or not (yt-dlp unquotes `+`).
  assert.equal(extractVideoId(`https://archive.org/embed/${ITEM}/${encodeURIComponent(FILE)}`), id);
  assert.equal(extractVideoId(archiveOrgDownloadUrl(ITEM, FILE)), id);
  assert.equal(
    extractVideoId(`https://archive.org/details/${ITEM}/Example+Talk+(Part+1)-AbC123xyz_9.mp4`),
    id,
  );
  // yt-dlp's native id for the entry maps to the same id.
  assert.equal(archiveOrgVideoIdFromNativeId(`${ITEM}/${FILE}`), id);
  assert.equal(archiveOrgVideoIdFromNativeId(ITEM), ITEM);
  // The output path pins to it (a safe directory name).
  assert.equal(dataDirIdForUrl(url), id);
  // Paths that slug alike still differ.
  const a = archiveOrgVideoId({ identifier: ITEM, file: "a b.mp4" });
  const b = archiveOrgVideoId({ identifier: ITEM, file: "a_b.mp4" });
  const c = archiveOrgVideoId({ identifier: ITEM, file: "a b.mkv" });
  assert.equal(new Set([a, b, c]).size, 3);
  // Sub-directories are part of the path.
  assert.equal(
    parseArchiveOrgUrl(`https://archive.org/details/${ITEM}/disc1/01%20Intro.mp3`)?.file,
    "disc1/01 Intro.mp3",
  );
});

test("URLs that name no item have no id", () => {
  assert.equal(parseArchiveOrgUrl("https://archive.org/search?query=x"), null);
  assert.equal(parseArchiveOrgUrl("https://archive.org/details/"), null);
  assert.equal(extractVideoId("https://archive.org/search?query=x"), null);
});

test("the torrent and page URLs", () => {
  assert.equal(archiveOrgTorrentUrl(ITEM), `https://archive.org/download/${ITEM}/${ITEM}_archive.torrent`);
  assert.equal(defaultWebpageUrl("archiveorg", ITEM), `https://archive.org/details/${ITEM}`);
  const fileId = archiveOrgVideoId({ identifier: ITEM, file: FILE });
  assert.equal(defaultWebpageUrl("archiveorg", fileId), `https://archive.org/details/${ITEM}`);
});

test("YouTube ids read from mirror names", () => {
  assert.equal(youtubeIdFromIdentifier("youtube-AbC123xyz_9"), "AbC123xyz_9");
  assert.equal(youtubeIdFromIdentifier(ITEM), null);
  assert.equal(youtubeIdFromFileName(FILE), "AbC123xyz_9");
  assert.equal(youtubeIdFromFileName("Some Title [Zz9-Qq8_Ww7].webm"), "Zz9-Qq8_Ww7");
  assert.equal(youtubeIdFromFileName("dir/Some Title [Zz9-Qq8_Ww7].info.json"), "Zz9-Qq8_Ww7");
  // An eleven-letter lowercase word is not an id.
  assert.equal(youtubeIdFromFileName("interview-performance.mp4"), null);
  assert.equal(youtubeIdFromFileName("plain.mp4"), null);
});

test("a mirrored file's name gives its title: date, view count, id and extension off", () => {
  const cases: [string, string | null][] = [
    // `YYYYMMDD <title> [<n> views]-<id>.<ext>`
    ["20210102 Example Talk at the Hall [1234 views]-AbC123xyz_9.mp4", "Example Talk at the Hall"],
    // A `<word>_` before the date, ` _ ` for ` | `.
    ["extras_20200304 A Long Chat _ Guest Name _ TOPIC _ Some Show [9876543 views]-Def456uvw_8.mp4",
      "A Long Chat | Guest Name | TOPIC | Some Show"],
    // No date; a separator inside the title; a dashed id that starts with `-`.
    ["Speaker - 'A Quoted Line!' & Why! _ Daily Clip--bC123xyz_9.mp4", "Speaker - 'A Quoted Line!' & Why! | Daily Clip"],
    // The bracketed id, a date then ` - `, commas in the count.
    ["20191231 - New Year Stream [1,234 views] [Ghi789rst_7].mkv", "New Year Stream"],
    // A sanitised colon: an underscore glued to a word, a space after it.
    ["20220505 Part One_ The Beginning-Jkl012mno_6.webm", "Part One: The Beginning"],
    // An underscore anywhere else stays.
    ["snake_case and_more [12 views]-Mno345pqr_5.mp4", "snake_case and_more"],
    // Not a calendar day (no 31st of April): kept.
    ["20210431 Not a Day-Abc234def_5.mp4", "20210431 Not a Day"],
    // Not a calendar date: kept. Brackets that are not a count: kept.
    ["20221340 Numbers First [NEW] Thing-Pqr678stu_4.mp4", "20221340 Numbers First [NEW] Thing"],
    // A dot inside the title survives; only the one extension goes.
    ["20180101 Talk with A.B. Someone-Stu901vwx_3.mp4", "Talk with A.B. Someone"],
    // A path inside the item: only its last part.
    ["sub/dir/20180101 Inside a Folder-Vwx234yza_2.mp4", "Inside a Folder"],
    // A name with no id at all is cleaned the same way.
    ["20180101 Plain Recording.mp3", "Plain Recording"],
    // Nothing sensible left.
    ["20180101 [55 views]-Yza567bcd_1.mp4", null],
    ["___.mp4", null],
  ];
  for (const [name, want] of cases) assert.equal(titleFromMirrorFileName(name), want, name);
});

test("a mirrored file's name gives its upload day: the leading [word_]YYYYMMDD, a real calendar day only", () => {
  const cases: [string, string | null][] = [
    ["20210102 Example Talk [1234 views]-AbC123xyz_9.mp4", "20210102"],
    ["extras_20200304 A Long Chat _ Some Show-Def456uvw_8.mp4", "20200304"],
    ["20191231 - New Year Stream [Ghi789rst_7].mkv", "20191231"],
    ["sub/dir/20180101 Inside a Folder-Vwx234yza_2.mp4", "20180101"],
    ["20180101.mp3", "20180101"],
    // Leap days: only in a leap year.
    ["20200229 Leap Day-Jkl012mno_6.mp4", "20200229"],
    ["20190229 No Leap Day-Jkl012mno_6.mp4", null],
    ["20210431 Not a Day-Abc234def_5.mp4", null],
    ["20221340 Numbers First-Pqr678stu_4.mp4", null],
    // Not at the start, or glued to the title: no date.
    ["Talk from 20210102-Stu901vwx_3.mp4", null],
    ["20210102Talk-Stu901vwx_3.mp4", null],
    ["two_words_20210102 Talk-Stu901vwx_3.mp4", null],
    ["Plain Recording.mp3", null],
  ];
  for (const [name, want] of cases) assert.equal(dateFromMirrorFileName(name), want, name);
});

test("dateFromMirrorFileName: a file with no date of its own takes its folder's", () => {
  assert.equal(dateFromMirrorFileName("Show/20190303_The Example Show - A Guest/The Example Show - A Guest.mkv"), "20190303");
  assert.equal(dateFromMirrorFileName("Show/20190303_Folder/20180101 Own date.mkv"), "20180101", "the file's own date wins");
  assert.equal(dateFromMirrorFileName("Show/Folder without date/Plain name.mkv"), null);
  assert.equal(dateFromMirrorFileName("Show/20191340_Bad date/Plain.mkv"), null, "not a calendar day");
});