import { readFile, writeFile } from "node:fs/promises";
import { test, expect } from "@playwright/test";
import {
  channelStage,
  generateReport,
  pathExists,
  resetData,
  resolvePath,
  writeSettings,
} from "./helpers";

const CHANNEL = "test-transcribe";
const DATA = `test-transcripts/channels/${CHANNEL}/data`;

// Minimal metadata so transcribeOne's normalize pass runs (it needs
// metadata.info.json) and writes transcript.cues.json.
const META = JSON.stringify({
  id: "vidA",
  title: "Synthetic parakeet video",
  upload_date: "20240101",
  duration: 10,
});

async function selectParakeet() {
  await writeSettings({
    adminTitle: "Test Admin",
    maxTranscriptPageBytes: 8388608,
    sleepBetweenDownloadsSeconds: 0,
    transcriptionApp: "parakeet",
    transcriptionApps: { parakeet: {} },
  });
}

test("parakeet app transcribes audio and writes a transcript.json", async ({
  page,
}) => {
  await resetData("one-transcribe-channel-with-audio");
  await selectParakeet();
  await generateReport(page, CHANNEL);
  await page.goto(channelStage(CHANNEL, "transcribe"));
  await page.getByRole("button", { name: "Transcribe missing" }).click();
  await expect(page.getByLabel("Transcribe missing output")).toContainText(
    "3 succeeded",
    { timeout: 30_000 },
  );

  for (const id of ["vidA", "vidB", "vidC"]) {
    expect(await pathExists(`${DATA}/${id}/transcript.json`)).toBe(true);
  }

  // The wrapper emits chough-native JSON (chunk_data in seconds). Reaching this
  // shape proves the parakeet app's argv ran AND that transcribeOne correctly
  // renamed the wrapper's exact `--output` path (no ".json" appended) to
  // transcript.json.
  const raw = JSON.parse(
    await readFile(resolvePath(`${DATA}/vidA/transcript.json`), "utf8"),
  );
  expect(Array.isArray(raw.chunk_data)).toBe(true);
  expect(raw.transcription).toBeUndefined();
});

test("parakeet output normalizes to chough-tagged cues", async ({ page }) => {
  await resetData("one-transcribe-channel-with-audio");
  await selectParakeet();
  // Give vidA metadata so the normalize pass produces transcript.cues.json.
  await writeFile(resolvePath(`${DATA}/vidA/metadata.info.json`), META);

  await generateReport(page, CHANNEL);
  await page.goto(channelStage(CHANNEL, "transcribe"));
  await page.getByRole("button", { name: "Transcribe missing" }).click();
  await expect(page.getByLabel("Transcribe missing output")).toContainText(
    "3 succeeded",
    { timeout: 30_000 },
  );

  await expect
    .poll(() => pathExists(`${DATA}/vidA/transcript.cues.json`), {
      timeout: 15_000,
    })
    .toBe(true);

  const cues = JSON.parse(
    await readFile(resolvePath(`${DATA}/vidA/transcript.cues.json`), "utf8"),
  );
  // The stitched output is chough-shaped, so it carries the chough-json tag…
  expect(cues.transcriptFormat).toBe("chough-json");
  // …and the cues come from seconds-based chunk_data (start 0, end 5), not
  // misread as whisper milliseconds.
  expect(cues.cues.length).toBeGreaterThan(0);
  expect(cues.cues[0].start).toBe(0);
  expect(cues.cues[0].end).toBe(5);
});