Greenlens/server/lib/tiktok-assets.js

const sharp = require('sharp');

const MAX_TIKTOK_PHOTO_BYTES = 20 * 1024 * 1024;

const createInputError = (message) => {
  const error = new Error(message);
  error.status = 400;
  error.code = 'BAD_REQUEST';
  return error;
};

const normalizePublicBaseUrl = (value) => String(value || '').trim().replace(/\/$/, '');

const isAllowedR2MediaUrl = (value) => {
  try {
    const candidate = new URL(value);
    const configuredHosts = String(process.env.TIKTOK_MEDIA_SOURCE_HOSTS || '')
      .split(',')
      .map((host) => host.trim().toLowerCase())
      .filter(Boolean);
    const hostname = candidate.hostname.toLowerCase();
    return candidate.protocol === 'https:'
      && !candidate.username
      && !candidate.password
      && !candidate.port
      && (hostname.endsWith('.r2.dev') || configuredHosts.includes(hostname));
  } catch {
    return false;
  }
};

const isOwnedTiktokMediaUrl = (value, publicBaseUrl) => {
  try {
    const candidate = new URL(value);
    const base = new URL(normalizePublicBaseUrl(publicBaseUrl));
    const basePath = base.pathname.replace(/\/$/, '');
    return candidate.protocol === 'https:'
      && !candidate.username
      && !candidate.password
      && candidate.origin === base.origin
      && (candidate.pathname === basePath || candidate.pathname.startsWith(`${basePath}/`));
  } catch {
    return false;
  }
};

const readBase64Photo = (photo) => {
  if (Buffer.isBuffer(photo)) return photo;
  if (Buffer.isBuffer(photo?.buffer)) return photo.buffer;

  let encoded = photo?.imageBase64 || photo?.dataBase64 || photo?.base64;
  if (typeof encoded !== 'string' || !encoded.trim()) return null;
  encoded = encoded.trim();

  const dataUrlMatch = encoded.match(/^data:image\/[a-z0-9.+-]+;base64,(.+)$/is);
  if (dataUrlMatch) encoded = dataUrlMatch[1];

  if (!/^[a-z0-9+/\s]+={0,2}$/i.test(encoded)) {
    throw createInputError('Each photo must contain valid base64 image data.');
  }

  const buffer = Buffer.from(encoded.replace(/\s/g, ''), 'base64');
  if (!buffer.length) throw createInputError('Each photo must contain image data.');
  return buffer;
};

const normalizeTiktokPhoto = async (buffer) => {
  if (buffer.length > MAX_TIKTOK_PHOTO_BYTES) {
    throw createInputError('Each TikTok photo must be 20 MB or smaller.');
  }

  try {
    return await sharp(buffer)
      .rotate()
      .resize({ width: 1080, height: 1080, fit: 'inside', withoutEnlargement: true })
      .jpeg({ quality: 92, mozjpeg: true })
      .toBuffer();
  } catch {
    throw createInputError('Each photo must be a valid image.');
  }
};

const downloadRemotePhoto = async (sourceUrl) => {
  if (!isAllowedR2MediaUrl(sourceUrl)) {
    throw createInputError('External photo URL host is not approved for GreenLens imports.');
  }

  const controller = new AbortController();
  const timeout = setTimeout(() => controller.abort(), 15_000);
  try {
    const response = await fetch(sourceUrl, { redirect: 'error', signal: controller.signal });
    if (!response.ok) throw createInputError('GreenLens could not download an external photo.');

    const contentType = String(response.headers.get('content-type') || '').split(';')[0].trim().toLowerCase();
    if (!['image/jpeg', 'image/png', 'image/webp'].includes(contentType)) {
      throw createInputError('External photo must be JPEG, PNG, or WebP.');
    }

    const declaredSize = Number(response.headers.get('content-length') || 0);
    if (declaredSize > MAX_TIKTOK_PHOTO_BYTES) {
      throw createInputError('Each TikTok photo must be 20 MB or smaller.');
    }

    const chunks = [];
    let totalBytes = 0;
    for await (const chunk of response.body) {
      totalBytes += chunk.length;
      if (totalBytes > MAX_TIKTOK_PHOTO_BYTES) {
        controller.abort();
        throw createInputError('Each TikTok photo must be 20 MB or smaller.');
      }
      chunks.push(Buffer.from(chunk));
    }
    return Buffer.concat(chunks, totalBytes);
  } catch (error) {
    if (error?.status) throw error;
    throw createInputError('GreenLens could not securely download the external photo.');
  } finally {
    clearTimeout(timeout);
  }
};

const sleepMs = (ms) => new Promise((resolve) => setTimeout(resolve, ms));

// Preflight before TikTok ever sees a URL: it must answer 200 without a
// redirect, be JPEG or WebP, and stay within TikTok's 20 MiB limit. Also
// serves as the read-after-write check right after a MinIO upload. Retries a
// couple of times to ride out brief 404/5xx windows after a fresh write.
const verifyPublicPhotoUrl = async (url, { attempts = 3, fetchImpl = fetch } = {}) => {
  let lastReason = 'unreachable';
  for (let attempt = 0; attempt < attempts; attempt++) {
    if (attempt > 0) await sleepMs(1000 * attempt);
    const controller = new AbortController();
    const timeout = setTimeout(() => controller.abort(), 10_000);
    try {
      const response = await fetchImpl(url, { method: 'HEAD', redirect: 'error', signal: controller.signal });
      if (response.status !== 200) {
        lastReason = `HTTP ${response.status}`;
        continue;
      }
      const contentType = String(response.headers.get('content-type') || '').split(';')[0].trim().toLowerCase();
      if (!['image/jpeg', 'image/webp'].includes(contentType)) {
        lastReason = `unsupported content type ${contentType || '(none)'}`;
        break;
      }
      const size = Number(response.headers.get('content-length') || 0);
      if (size > MAX_TIKTOK_PHOTO_BYTES) {
        lastReason = 'larger than 20 MB';
        break;
      }
      return;
    } catch {
      lastReason = 'unreachable or redirected';
    } finally {
      clearTimeout(timeout);
    }
  }
  throw createInputError(`Photo URL is not usable for TikTok (${lastReason}): ${url}`);
};

const prepareTiktokPhotoUrls = async (photos, options) => {
  const publicBaseUrl = normalizePublicBaseUrl(options.publicBaseUrl);
  if (!publicBaseUrl) throw new Error('MINIO_PUBLIC_URL is not configured.');
  const verifyUrl = options.verifyPhotoUrl || verifyPublicPhotoUrl;

  const urls = [];
  for (const photo of photos) {
    const sourceUrl = typeof photo === 'string' ? photo.trim() : String(photo?.url || '').trim();
    if (sourceUrl && isOwnedTiktokMediaUrl(sourceUrl, publicBaseUrl)) {
      await verifyUrl(sourceUrl);
      urls.push(sourceUrl);
      continue;
    }

    let inputBuffer = readBase64Photo(photo);
    if (!inputBuffer && sourceUrl) {
      inputBuffer = await (options.downloadPhoto || downloadRemotePhoto)(sourceUrl);
    }
    if (!inputBuffer) {
      throw createInputError('Each photo must include a GreenLens storage URL or imageBase64.');
    }

    const jpegBuffer = await (options.normalizePhoto || normalizeTiktokPhoto)(inputBuffer);
    const uploaded = await options.uploadImage(jpegBuffer.toString('base64'), 'image/jpeg');
    if (!isOwnedTiktokMediaUrl(uploaded?.url, publicBaseUrl)) {
      throw new Error('GreenLens image storage returned an unverified public URL.');
    }
    await verifyUrl(uploaded.url);
    urls.push(uploaded.url);
  }

  return urls;
};

module.exports = {
  MAX_TIKTOK_PHOTO_BYTES,
  isAllowedR2MediaUrl,
  isOwnedTiktokMediaUrl,
  verifyPublicPhotoUrl,
  prepareTiktokPhotoUrls,
};