ocr

/*
 * file to text ocr
 * Base: https://imagetotext.my
 * Sumber: https://whatsapp.com/channel/0029VbB4Kw8EFeXfeExaXc3Q
 * Note: extract teks dari file
 *
 * Usage:
 *   node ocr.js 
 *             ^ untuk info cara pakai
 */
const fs   = require('fs');
const path = require('path');

// ── Config ───────────────────────────────────────────────────────────────────

const BASE = 'https://imagetotext.my';
const API  = `${BASE}/index.php`;
const MIME = {
  '.jpg': 'image/jpeg', '.jpeg': 'image/jpeg', '.png': 'image/png',
  '.gif': 'image/gif',  '.webp': 'image/webp',  '.bmp': 'image/bmp',
  '.tif': 'image/tiff', '.tiff': 'image/tiff',  '.pdf': 'application/pdf',
};

const BASE_HEADERS = {
  'User-Agent':      'Mozilla/5.0 (X11; Linux x86_64; rv:140.0) Gecko/20100101 Firefox/140.0',
  'Accept':          '*/*',
  'Accept-Language': 'en-US,en;q=0.5',
  'Referer':         `${BASE}/`,
  'Origin':          BASE,
};

const sleep = ms => new Promise(r => setTimeout(r, ms));

// ── Cookie auto-fetch ─────────────────────────────────────────────────────────

let _cookieCache = null;

async function fetchCookie() {
  if (_cookieCache) return _cookieCache;

  const res = await fetch(BASE, { method: 'GET', headers: BASE_HEADERS, redirect: 'follow' });

  // Ambil semua Set-Cookie header
  // Node 18+ fetch: res.headers.getSetCookie() → string[]
  // Node <18 / undici lama: res.headers.get('set-cookie') → comma-joined string
  let cookies = [];
  if (typeof res.headers.getSetCookie === 'function') {
    cookies = res.headers.getSetCookie();
  } else {
    const raw = res.headers.get('set-cookie');
    if (raw) cookies = raw.split(/,(?=[^ ])/).map(s => s.trim());
  }

  // Ambil nilai cookie (sebelum ;) yang namanya 'login'
  const loginCookie = cookies
    .map(c => c.split(';')[0].trim())
    .find(c => c.startsWith('login='));

  if (!loginCookie) throw new Error('Gagal ambil session cookie dari homepage');

  _cookieCache = loginCookie;
  return loginCookie;
}

function clearCookieCache() {
  _cookieCache = null;
}

// ── Core ─────────────────────────────────────────────────────────────────────

async function upload(input, name) {
  const cookie   = await fetchCookie();
  const buf      = typeof input === 'string' ? fs.readFileSync(input) : input;
  const filename = name || (typeof input === 'string' ? path.basename(input) : 'image.png');
  const mime     = MIME[path.extname(filename).toLowerCase()] || 'application/octet-stream';
  const boundary = `----OCR${Date.now()}`;

  const head = Buffer.from(
    `--${boundary}\r\nContent-Disposition: form-data; name="op"\r\n\r\nupload_direct\r\n` +
    `--${boundary}\r\nContent-Disposition: form-data; name="file"; filename="${filename}"\r\nContent-Type: ${mime}\r\n\r\n`
  );
  const tail = Buffer.from(`\r\n--${boundary}--\r\n`);
  const body = Buffer.concat([head, buf, tail]);

  const headers = {
    ...BASE_HEADERS,
    'Cookie':         cookie,
    'Content-Type':  `multipart/form-data; boundary=${boundary}`,
    'Content-Length': String(body.length),
  };

  const res  = await fetch(API, { method: 'POST', headers, body });
  const json = await res.json();

  if (!json?.success) {
    // Cookie mungkin expired walau baru di-fetch, clear cache & lempar error
    clearCookieCache();
    throw new Error(
      json?.error === 'file_too_large'
        ? 'File terlalu besar (max 10 MB)'
        : json?.message || 'Upload gagal'
    );
  }

  const { key, file_id, url } = json.data;
  if (!key || !file_id) throw new Error('Response upload tidak valid');
  return { fileId: String(file_id), key, url };
}

async function poll(fileId, key, timeout = 5 * 60 * 1000) {
  const cookie = await fetchCookie();
  const params = new URLSearchParams({
    op: 'status', action: 'check_task_status',
    file_id: fileId, filename: key,
  });
  const headers = { ...BASE_HEADERS, 'Cookie': cookie };
  const start   = Date.now();

  while (true) {
    if (Date.now() - start > timeout)
      throw new Error(`Timeout setelah ${Math.round((Date.now() - start) / 1000)}s`);

    try {
      const res    = await fetch(`${API}?${params}`, { headers });
      const json   = await res.json();
      const data   = json?.data || json || {};
      const status = String(data.status || '').toLowerCase();

      if (['completed', 'complete', 'done', 'success'].includes(status)) {
        const r = data.result || data;
        return { text: r.ocr_text || r.text || '', confidence: r.confidence ?? null, language: r.language ?? null };
      }
      if (status.includes('fail') || status.includes('error'))
        throw new Error(`OCR error: ${status}`);
    } catch (e) {
      if (e.message.startsWith('OCR error') || e.message.startsWith('Timeout')) throw e;
    }

    await sleep(2000);
  }
}

// ── Public API ────────────────────────────────────────────────────────────────

async function extract(input, opts = {}) {
  const name            = opts.filename || (typeof input === 'string' ? path.basename(input) : 'image.png');
  const { fileId, key } = await upload(input, name);
  const result          = await poll(fileId, key, opts.timeout);
  return { text: result.text, confidence: result.confidence, language: result.language, fileId };
}

async function extractBatch(inputs, opts = {}) {
  // Pre-fetch cookie sekali sebelum batch dimulai
  await fetchCookie();

  const limit   = opts.concurrency ?? 5;
  const results = new Array(inputs.length);
  let idx = 0;

  const worker = async () => {
    while (idx < inputs.length) {
      const i = idx++;
      try {
        results[i] = { file: inputs[i], ...(await extract(inputs[i], opts)) };
      } catch (e) {
        results[i] = { file: inputs[i], text: null, error: e.message };
      }
    }
  };

  await Promise.all(Array.from({ length: Math.min(limit, inputs.length) }, worker));
  return results;
}

module.exports = { extract, extractBatch, upload, poll, fetchCookie, clearCookieCache };

// ── CLI ───────────────────────────────────────────────────────────────────────

if (require.main === module) {
  const argv = process.argv.slice(2);

  if (!argv.length) {
    console.log(`
imagetotext.my OCR — Node.js wrapper

Usage:
  node ocr.js <file> [options]
  node ocr.js <file1> <file2> ...  (batch)

Options:
  -o <output>   Simpan ke file (contoh: -o hasil.json / hasil.txt / hasil.md)
                Tanpa -o → langsung print JSON ke terminal

Examples:
  node ocr.js foto.png
  node ocr.js scan.pdf -o hasil.txt
  node ocr.js a.png b.jpg c.png -o batch.json

Note: Cookie session di-fetch otomatis dari homepage, tidak perlu hardcode.
`);
    process.exit(0);
  }

  const files = [];
  let outFile = null;
  for (let i = 0; i < argv.length; i++) {
    if (argv[i] === '-o' && argv[i + 1]) { outFile = argv[++i]; }
    else files.push(argv[i]);
  }

  const ext   = outFile ? path.extname(outFile).toLowerCase() : '.json';
  const isTxt = ext === '.txt' || ext === '.md';

  (async () => {
    process.stderr.write('Fetching session cookie...\n');
    await fetchCookie();
    process.stderr.write('Processing...\n');

    try {
      const raw = files.length === 1
        ? await extract(files[0])
        : await extractBatch(files);

      let output;
      if (isTxt && files.length === 1) {
        output = raw.text;
      } else if (isTxt) {
        output = raw.map(r => `=== ${r.file} ===\n${r.text || '[error] ' + r.error}`).join('\n\n');
      } else {
        output = JSON.stringify(raw, null, 2);
      }

      if (outFile) {
        fs.writeFileSync(outFile, output, 'utf8');
        process.stderr.write(`Saved: ${outFile}\n`);
      } else {
        console.log(output);
      }
    } catch (err) {
      console.error(JSON.stringify({ error: err.message }));
      process.exit(1);
    }
  })();
}