// 预生成解说语音（DashScope qwen-tts），输出 e2e/voice/*.wav 与 clips.json（文本→{file,dur}）。
// run.mjs 据此让每步停留足够时长；add-voice.mjs 据此把语音放到字幕出现的时间点。
import { execSync } from "child_process";
import crypto from "crypto";
import fs from "fs";
import path from "path";
import { NARRATION, clean } from "./narration.mjs";

const OUT = path.resolve("e2e/voice");
const QWEN_VOICE = process.env.QWEN_TTS_VOICE ?? "Cherry";
const FALLBACK_VOICE = process.env.TTS_VOICE ?? "Tingting";

function readApiKey() {
  const env = fs.readFileSync(path.resolve(".env"), "utf-8");
  const m = env.match(/^AI_API_KEY=(.*)$/m);
  return m ? m[1].trim() : "";
}
const KEY = readApiKey();
const sh = (cmd) => execSync(cmd, { stdio: ["ignore", "pipe", "pipe"] }).toString();
const durationOf = (f) =>
  parseFloat(sh(`ffprobe -v error -show_entries format=duration -of csv=p=0 "${f}"`).trim());

async function qwenTts(text, outFile) {
  const r = await fetch(
    "https://dashscope.aliyuncs.com/api/v1/services/aigc/multimodal-generation/generation",
    {
      method: "POST",
      headers: { Authorization: `Bearer ${KEY}`, "Content-Type": "application/json" },
      body: JSON.stringify({ model: "qwen-tts", input: { text, voice: QWEN_VOICE } }),
      signal: AbortSignal.timeout(40000),
    }
  );
  const j = await r.json();
  const url = j?.output?.audio?.url;
  if (!url) throw new Error(JSON.stringify(j).slice(0, 160));
  const a = await fetch(url, { signal: AbortSignal.timeout(40000) });
  if (!a.ok) throw new Error(`下载音频失败 ${a.status}`);
  fs.writeFileSync(outFile, Buffer.from(await a.arrayBuffer()));
}

const main = async () => {
  fs.rmSync(OUT, { recursive: true, force: true });
  fs.mkdirSync(OUT, { recursive: true });
  const texts = Array.from(new Set(Object.values(NARRATION).map(clean))).filter(Boolean);
  const clips = {};
  for (const text of texts) {
    const id = crypto.createHash("md5").update(text).digest("hex").slice(0, 10);
    const wav = path.join(OUT, `${id}.wav`);
    let engine = "qwen-tts";
    let ok = false;
    for (let attempt = 0; attempt < 5 && !ok; attempt++) {
      try {
        await qwenTts(text, wav);
        ok = true;
      } catch (e) {
        if (attempt < 4) await new Promise((r) => setTimeout(r, 2500 * (attempt + 1)));
      }
    }
    if (!ok) {
      engine = "say";
      const aiff = path.join(OUT, `${id}.aiff`);
      sh(`say -v ${FALLBACK_VOICE} -o "${aiff}" "${text.replace(/"/g, "")}"`);
      sh(`ffmpeg -y -i "${aiff}" "${wav}"`);
      fs.rmSync(aiff, { force: true });
    }
    const dur = durationOf(wav);
    clips[text] = { file: path.relative(process.cwd(), wav), dur, engine };
    console.log(`✓ ${dur.toFixed(1)}s [${engine}] ${text.slice(0, 26)}…`);
    await new Promise((r) => setTimeout(r, 1200)); // 限速，避免触发节流
  }
  fs.writeFileSync(path.join(OUT, "clips.json"), JSON.stringify(clips, null, 2));
  console.log(`完成：${texts.length} 条语音 → e2e/voice/clips.json`);
};

main().catch((e) => {
  console.error(e);
  process.exit(1);
});