import crypto from 'node:crypto';
import fs from 'node:fs';
import path from 'node:path';
import cors from 'cors';
import express from 'express';
import { MsEdgeTTS, OUTPUT_FORMAT } from 'msedge-tts';

const app = express();
const port = Number(process.env.TTS_PORT ?? 3333);
const host = process.env.TTS_HOST ?? '0.0.0.0';
const cacheDir = path.resolve(process.cwd(), '.tts-cache');

fs.mkdirSync(cacheDir, { recursive: true });

app.use(cors());
app.use(express.json({ limit: '64kb' }));
app.use('/audio', express.static(cacheDir, { maxAge: '1h' }));

app.get('/health', (_request, response) => {
  response.json({ ok: true, voice: getVoiceFromEnv() });
});

app.get('/voices', async (_request, response) => {
  const tts = new MsEdgeTTS({});
  const voices = await tts.getVoices();
  response.json(
    voices
      .filter((voice) => voice.Locale.startsWith('en-'))
      .map((voice) => ({
        name: voice.ShortName,
        friendlyName: voice.FriendlyName,
        gender: voice.Gender,
        locale: voice.Locale,
      })),
  );
});

app.post('/tts', async (request, response) => {
  const text = sanitizeText(String(request.body?.text ?? ''));

  if (!text) {
    response.status(400).json({ error: 'Missing text' });
    return;
  }

  const voice = String(request.body?.voice ?? getVoiceFromEnv());
  const rate = String(request.body?.rate ?? process.env.TTS_RATE ?? '-8%');
  const pitch = String(request.body?.pitch ?? process.env.TTS_PITCH ?? '+0Hz');
  const hash = crypto.createHash('sha256').update(`${voice}|${rate}|${pitch}|${text}`).digest('hex');
  const filename = `${hash}.mp3`;
  const filePath = path.join(cacheDir, filename);

  if (!fs.existsSync(filePath)) {
    const tts = new MsEdgeTTS({});
    await tts.setMetadata(voice, OUTPUT_FORMAT.AUDIO_24KHZ_96KBITRATE_MONO_MP3);
    const { audioFilePath, metadataFilePath } = await tts.toFile(cacheDir, escapeXml(text), { rate, pitch });
    fs.renameSync(audioFilePath, filePath);
    if (metadataFilePath && fs.existsSync(metadataFilePath)) {
      fs.unlinkSync(metadataFilePath);
    }
    tts.close();
  }

  response.json({ audioUrl: `${request.protocol}://${request.get('host')}/audio/${filename}` });
});

app.listen(port, host, () => {
  console.log(`TTS server listening on http://${host}:${port}`);
  console.log(`Voice: ${getVoiceFromEnv()}`);
});

function getVoiceFromEnv() {
  return process.env.TTS_VOICE ?? 'en-US-JennyNeural';
}

function sanitizeText(value) {
  return value
    .replace(/<think>[\s\S]*?<\/think>/gi, '')
    .replace(/```[\s\S]*?```/g, ' ')
    .replace(/`([^`]+)`/g, '$1')
    .replace(/\*\*([^*]+)\*\*/g, '$1')
    .replace(/[*_#>~]/g, '')
    .replace(/https?:\/\/\S+/g, '')
    .replace(/\s+/g, ' ')
    .trim()
    .slice(0, 1800);
}

function escapeXml(value) {
  return value
    .replace(/&/g, '&amp;')
    .replace(/</g, '&lt;')
    .replace(/>/g, '&gt;')
    .replace(/"/g, '&quot;')
    .replace(/'/g, '&apos;');
}