import { rm, writeFile } from "node:fs/promises";
import { test, expect } from "@playwright/test";
import {
  buildIndex,
  readJson,
  resetData,
  resolvePath,
  writeSite,
} from "./helpers";

const CHANNEL = "test-youtube";
const VIDEO_DIR = "20240101_test1234567";
// The index keys a record by metadata.info.json's `id`, which the fixture now
// keeps equal to its directory name (see that file — a mismatch made snapshot
// generation rename the directory mid-spec).
const VIDEO_ID = VIDEO_DIR;
const DATA = `test-transcripts/channels/${CHANNEL}/data/${VIDEO_DIR}`;
const TRANSCRIPTS_PAGE = `test-transcripts/.export-index/shared/transcripts/${CHANNEL}/page-0000.json`;
// parseVtt only emits cues from lines carrying YouTube's inline word-timing
// tags, so the body needs <hh:mm:ss.mmm> markers to produce a real cue.
const VTT =
  "WEBVTT\n\n00:00:00.000 --> 00:00:05.000\n" +
  "<00:00:00.000><c> hello</c><00:00:02.000><c> regional</c>\n";

type Detail = { id: string; cues?: Array<unknown> };

// The site has to exist before the index build runs; buildIndex() from helpers
// is the shared "open the Pool, press the button, wait for Done" (the local name
// would shadow the import, hence seedAndBuild).
async function seedAndBuild(page: import("@playwright/test").Page) {
  await writeSite("testsite", {
    channels: [{ slug: CHANNEL, groupId: "default" }],
  });
  await buildIndex(page);
}

// YouTube sometimes serves a video's English captions only under a regional or
// auto code (transcript.en-US.vtt) with no plain transcript.en.vtt. buildIndex
// should still parse it as the primary transcript via the English-VTT fallback,
// so the video's shared transcript page carries cues.
test("build index parses a video with only transcript.en-US.vtt", async ({
  page,
}) => {
  await resetData("one-youtube-channel-with-data");
  // Replace the canonical transcript.en.vtt with a regional-only English track.
  await rm(resolvePath(`${DATA}/transcript.en.vtt`), { force: true });
  await writeFile(resolvePath(`${DATA}/transcript.en-US.vtt`), VTT);

  await seedAndBuild(page);

  const detailsPage = await readJson<Detail[]>(TRANSCRIPTS_PAGE);
  const detail = detailsPage.find((d) => d.id === VIDEO_ID);
  expect(detail).toBeDefined();
  // cues are written only when a primary transcript was recognized.
  expect(detail?.cues?.length ?? 0).toBeGreaterThan(0);
});

// Regression guard: a translation track (transcript.es-en-US.vtt) is NOT English
// and must not be mistaken for the primary transcript — the video lands in the
// index (it has metadata) but carries no cues.
test("build index ignores a non-English translation track", async ({ page }) => {
  await resetData("one-youtube-channel-with-data");
  await rm(resolvePath(`${DATA}/transcript.en.vtt`), { force: true });
  await writeFile(resolvePath(`${DATA}/transcript.es-en-US.vtt`), VTT);

  await seedAndBuild(page);

  const detailsPage = await readJson<Detail[]>(TRANSCRIPTS_PAGE);
  const detail = detailsPage.find((d) => d.id === VIDEO_ID);
  expect(detail).toBeDefined();
  expect(detail?.cues?.length ?? 0).toBe(0);
});