feat: add deterministic cleared project media

scripts/lib/privacy-scan.mjs

@@ -0,0 +1,118 @@
+import { access, readFile } from 'node:fs/promises';
+import { join } from 'node:path';
+import sharp from 'sharp';
+import { fileTypeFromFile } from 'file-type';
+import { createWorker, OEM } from 'tesseract.js';
+import {
+  BinaryBitmap,
+  HybridBinarizer,
+  MultiFormatReader,
+  NotFoundException,
+  RGBLuminanceSource,
+} from '@zxing/library';
+
+const normalize = (value) => value.normalize('NFKC').toLocaleLowerCase('en-US');
+const fail = (rule, location = '') => { throw new Error(location ? `${rule}:${location}` : rule); };
+
+export const loadDenylist = async (path) => {
+  let text;
+  try {
+    text = await readFile(path, 'utf8');
+  } catch {
+    fail('PRIVACY-DENYLIST-001');
+  }
+  const terms = [...new Set(text.split(/\r?\n/u).map((line) => line.trim()).filter(Boolean).map(normalize))];
+  if (terms.length === 0 || terms.some((term) => term.length < 3 || /[\u0000-\u001f\u007f]/u.test(term))) {
+    fail('PRIVACY-DENYLIST-001');
+  }
+  return terms;
+};
+
+export const scanText = (bytesOrText, denylist, rule = 'PRIVACY-BYTES-001') => {
+  const text = normalize(Buffer.isBuffer(bytesOrText) || bytesOrText instanceof Uint8Array
+    ? Buffer.from(bytesOrText).toString('utf8')
+    : String(bytesOrText));
+  if (denylist.some((term) => text.includes(normalize(term)))) fail(rule);
+};
+
+const scanBarcode = async (path) => {
+  let decoded;
+  try {
+    decoded = await sharp(path).removeAlpha().greyscale().raw().toBuffer({ resolveWithObject: true });
+  } catch {
+    fail('ASSET-DECODE-001');
+  }
+  const { data, info } = decoded;
+  const luminance = new RGBLuminanceSource(new Uint8ClampedArray(data), info.width, info.height);
+  const bitmap = new BinaryBitmap(new HybridBinarizer(luminance));
+  try {
+    new MultiFormatReader().decode(bitmap);
+    fail('PRIVACY-BARCODE-001');
+  } catch (error) {
+    if (error instanceof NotFoundException || error?.name === 'NotFoundException') return;
+    if (error instanceof Error && error.message === 'PRIVACY-BARCODE-001') throw error;
+    fail('PRIVACY-BARCODE-002');
+  }
+};
+
+const scanOcr = async (path, denylist) => {
+  const candidates = ['/usr/share/tessdata', '/usr/share/tesseract-ocr/5/tessdata'];
+  let langPath;
+  for (const candidate of candidates) {
+    try {
+      await access(join(candidate, 'eng.traineddata'));
+      langPath = candidate;
+      break;
+    } catch {}
+  }
+  if (!langPath) fail('PRIVACY-OCR-002');
+  let worker;
+  try {
+    worker = await createWorker('eng', OEM.LSTM_ONLY, { langPath, gzip: false, cacheMethod: 'none' });
+  } catch {
+    fail('PRIVACY-OCR-002');
+  }
+  let timer;
+  try {
+    const result = await Promise.race([
+      worker.recognize(path),
+      new Promise((_, reject) => { timer = setTimeout(() => reject(new Error('timeout')), 25_000); }),
+    ]);
+    scanText(result.data.text, denylist, 'PRIVACY-OCR-001');
+  } catch (error) {
+    if (error instanceof Error && error.message === 'PRIVACY-OCR-001') throw error;
+    fail('PRIVACY-OCR-002');
+  } finally {
+    if (timer !== undefined) clearTimeout(timer);
+    await worker.terminate().catch(() => {});
+  }
+};
+
+export const scanImage = async (path, denylist, options = {}) => {
+  let type;
+  try { type = await fileTypeFromFile(path); } catch { fail('PRIVACY-TYPE-001'); }
+  if (!type || !['image/png', 'image/webp', 'image/jpeg'].includes(type.mime)) fail('PRIVACY-TYPE-001');
+  let metadata;
+  try {
+    metadata = await sharp(path, { failOn: 'error', limitInputPixels: 50_000_000 }).metadata();
+  } catch {
+    fail('ASSET-DECODE-001');
+  }
+  if (metadata.exif || metadata.icc || metadata.iptc || metadata.xmp || metadata.orientation !== undefined) {
+    fail('PRIVACY-METADATA-001');
+  }
+  scanText(await readFile(path), denylist);
+  await scanBarcode(path);
+  if (options.runOcr !== false) await scanOcr(path, denylist);
+  return { width: metadata.width, height: metadata.height, mediaType: type.mime };
+};
+
+export const scanBinary = async (path, denylist) => {
+  let bytes;
+  try { bytes = await readFile(path); } catch { fail('PRIVACY-TYPE-001'); }
+  scanText(bytes, denylist);
+  let type;
+  try { type = await fileTypeFromFile(path); } catch { fail('PRIVACY-TYPE-001'); }
+  if (!type || !['video/webm', 'video/mp4'].includes(type.mime)) fail('PRIVACY-TYPE-001');
+  return { mediaType: type.mime };
+};