send2streamer/server.js

require('dotenv').config();

const path = require('path');
const express = require('express');
const session = require('express-session');
const { createServer } = require('http');
const { Server } = require('socket.io');
const { OAuth2Client } = require('google-auth-library');
const { execFile } = require('child_process');

const PORT = process.env.PORT || 3000;
const GOOGLE_CLIENT_ID = process.env.GOOGLE_CLIENT_ID;
const SESSION_SECRET = process.env.SESSION_SECRET;
const TTS_VOICE = process.env.TTS_VOICE || 'ru-RU-DmitryNeural';

if (!GOOGLE_CLIENT_ID) {
  console.error('Missing GOOGLE_CLIENT_ID in .env (see .env.example)');
  process.exit(1);
}
if (!SESSION_SECRET) {
  console.error('Missing SESSION_SECRET in .env (see .env.example)');
  process.exit(1);
}

const oauthClient = new OAuth2Client(GOOGLE_CLIENT_ID);

const app = express();
const httpServer = createServer(app);
const io = new Server(httpServer);

const sessionMiddleware = session({
  secret: SESSION_SECRET,
  resave: false,
  saveUninitialized: false,
  cookie: {
    maxAge: 1000 * 60 * 60 * 24 * 7, // 7 days
    // set to true once served over HTTPS in production
    secure: false,
    sameSite: 'lax',
  },
});

app.use(sessionMiddleware);
app.use(express.json());
app.use(express.static(path.join(__dirname, 'public')));

// Share the express session with socket.io connections.
const wrap = (middleware) => (socket, next) => middleware(socket.request, {}, next);
io.use(wrap(sessionMiddleware));

app.get('/config', (req, res) => {
  res.json({ googleClientId: GOOGLE_CLIENT_ID });
});

app.get('/me', (req, res) => {
  res.json({ user: req.session.user || null });
});

app.post('/auth/google', async (req, res) => {
  const { credential } = req.body;
  if (!credential) {
    return res.status(400).json({ error: 'missing credential' });
  }

  try {
    const ticket = await oauthClient.verifyIdToken({
      idToken: credential,
      audience: GOOGLE_CLIENT_ID,
    });
    const payload = ticket.getPayload();

    req.session.user = {
      email: payload.email,
      name: payload.name,
      picture: payload.picture,
    };

    console.log('[auth] logged in:', payload.email);
    res.json({ user: req.session.user });
  } catch (err) {
    console.error('Google token verification failed:', err.message);
    res.status(401).json({ error: 'invalid token' });
  }
});

// Shells out to the `edge-tts` Python CLI (same engine as t2sTelegramBot uses)
// instead of the edge-tts-node npm package, whose request-signing algorithm
// currently gets rejected by Microsoft's servers with a 403.
function synthesizeSpeech(text) {
  return new Promise((resolve, reject) => {
    execFile(
      'edge-tts',
      ['-t', text, '-v', TTS_VOICE],
      { encoding: 'buffer', maxBuffer: 10 * 1024 * 1024 },
      (err, stdout) => {
        if (err) return reject(err);
        resolve(stdout);
      }
    );
  });
}

app.post('/auth/logout', (req, res) => {
  req.session.destroy(() => res.json({ ok: true }));
});

io.on('connection', (socket) => {
  const connectedUser = socket.request.session.user;
  console.log(`[socket] connected id=${socket.id} user=${connectedUser ? connectedUser.email : '(none)'}`);

  socket.on('disconnect', (reason) => {
    console.log(`[socket] disconnected id=${socket.id} reason=${reason}`);
  });

  socket.on('send_message', (text) => {
    // The session snapshot on socket.request is captured once, at connect time.
    // If the socket connected before login (page load opens it immediately),
    // it never sees a session updated later by the separate /auth/google request
    // unless we explicitly reload it from the store here.
    socket.request.session.reload((err) => {
      const user = !err && socket.request.session.user;
      console.log(`[send_message] id=${socket.id} user=${user ? user.email : '(none)'} text=${JSON.stringify(text)}`);

      if (!user) {
        socket.emit('send_error', 'Not authenticated');
        return;
      }
      if (typeof text !== 'string' || !text.trim()) {
        return;
      }

      const message = text.trim().slice(0, 500);
      const clientCount = io.engine.clientsCount;
      console.log(`[broadcast] "${message}" -> ${clientCount} connected client(s)`);
      io.emit('display_message', {
        text: message,
        from: user.name,
        at: Date.now(),
      });

      // Voice is generated after the fact so the text shows up immediately
      // instead of waiting on the round trip to the TTS service.
      synthesizeSpeech(message)
        .then((audioBuffer) => {
          io.emit('display_audio', {
            audio: audioBuffer.toString('base64'),
            mimeType: 'audio/mpeg',
          });
        })
        .catch((err) => {
          console.error('[tts] synthesis failed:', err);
        });
    });
  });
});

httpServer.listen(PORT, () => {
  console.log(`Listening on http://localhost:${PORT}`);
});