feat: add deterministic cleared project media

scripts/lib/asset-recipes.mjs

@@ -0,0 +1,91 @@
+import { readFile, realpath, stat } from 'node:fs/promises';
+import { basename, extname, isAbsolute, relative, resolve, sep } from 'node:path';
+
+const idPattern = /^[a-z][a-z0-9-]{0,63}$/u;
+const outputPattern = /^project-[a-z0-9][a-z0-9-]*\.(webp|png|webm)$/u;
+
+const fail = (rule) => { throw new Error(rule); };
+
+const exactKeys = (record, allowed, rule) => {
+  if (Object.keys(record).some((key) => !allowed.includes(key))) fail(rule);
+};
+
+export const readJson = async (path, rule = 'ASSET-JSON-001') => {
+  try {
+    return JSON.parse(await readFile(path, 'utf8'));
+  } catch {
+    fail(rule);
+  }
+};
+
+export const loadRecipes = async (path) => {
+  const input = await readJson(path, 'ASSET-RECIPE-001');
+  if (!input || typeof input !== 'object' || input.schema !== 'portfolio-asset-recipes/v1' || !Array.isArray(input.assets)) {
+    fail('ASSET-RECIPE-001');
+  }
+  exactKeys(input, ['schema', 'assets'], 'ASSET-RECIPE-001');
+  const seenIds = new Set();
+  const seenOutputs = new Set();
+  const assets = input.assets.map((asset) => {
+    if (!asset || typeof asset !== 'object') fail('ASSET-RECIPE-001');
+    exactKeys(asset, [
+      'id', 'kind', 'output', 'maxWidth', 'startSeconds', 'durationSeconds',
+      'rightsHolder', 'license', 'attribution',
+    ], 'ASSET-RECIPE-001');
+    if (!idPattern.test(asset.id) || !['image', 'video'].includes(asset.kind) || !outputPattern.test(asset.output)) {
+      fail('ASSET-RECIPE-001');
+    }
+    if (asset.kind === 'image' && !/\.(webp|png)$/u.test(asset.output)) fail('ASSET-RECIPE-001');
+    if (asset.kind === 'video' && !asset.output.endsWith('.webm')) fail('ASSET-RECIPE-001');
+    if (!Number.isInteger(asset.maxWidth) || asset.maxWidth < 64 || asset.maxWidth > 4096) fail('ASSET-RECIPE-001');
+    if (asset.kind === 'video' && (
+      typeof asset.startSeconds !== 'number' || asset.startSeconds < 0 ||
+      typeof asset.durationSeconds !== 'number' || asset.durationSeconds <= 0 || asset.durationSeconds > 30
+    )) fail('ASSET-RECIPE-001');
+    for (const field of ['rightsHolder', 'license', 'attribution']) {
+      if (typeof asset[field] !== 'string' || asset[field].trim().length === 0 || /[\u0000-\u001f\u007f]/u.test(asset[field])) {
+        fail('ASSET-RECIPE-001');
+      }
+    }
+    if (seenIds.has(asset.id) || seenOutputs.has(asset.output.toLowerCase())) fail('ASSET-RECIPE-002');
+    seenIds.add(asset.id);
+    seenOutputs.add(asset.output.toLowerCase());
+    return Object.freeze({ ...asset });
+  });
+  return Object.freeze(assets);
+};
+
+export const loadPrivateMapping = async (path, recipes) => {
+  const input = await readJson(path, 'ASSET-MAPPING-001');
+  if (!input || typeof input !== 'object' || input.schema !== 'portfolio-private-source-map/v1' ||
+    !input.sources || typeof input.sources !== 'object' || Array.isArray(input.sources)) fail('ASSET-MAPPING-001');
+  exactKeys(input, ['schema', 'sources'], 'ASSET-MAPPING-001');
+  const ids = recipes.map(({ id }) => id).sort();
+  if (JSON.stringify(Object.keys(input.sources).sort()) !== JSON.stringify(ids)) fail('ASSET-MAPPING-002');
+  for (const value of Object.values(input.sources)) {
+    if (typeof value !== 'string' || value.length === 0 || isAbsolute(value) || value.includes('\\') ||
+      value.split('/').some((part) => part === '' || part === '.' || part === '..')) fail('ASSET-MAPPING-001');
+  }
+  return input.sources;
+};
+
+export const resolvePrivateSource = async (sourceRoot, mappedPath) => {
+  const root = await realpath(sourceRoot);
+  const candidate = resolve(root, mappedPath);
+  const rel = relative(root, candidate);
+  if (!rel || rel.startsWith(`..${sep}`) || rel === '..' || isAbsolute(rel)) fail('ASSET-SOURCE-001');
+  let resolved;
+  try {
+    resolved = await realpath(candidate);
+  } catch {
+    fail('ASSET-SOURCE-001');
+  }
+  const resolvedRel = relative(root, resolved);
+  if (resolvedRel.startsWith(`..${sep}`) || resolvedRel === '..' || isAbsolute(resolvedRel)) fail('ASSET-SOURCE-001');
+  const info = await stat(resolved);
+  if (!info.isFile() || info.size <= 0 || info.size > 100 * 1024 * 1024) fail('ASSET-SOURCE-002');
+  return resolved;
+};
+
+export const publicFileNameFor = (recipe) => basename(recipe.output);
+export const publicExtensionFor = (recipe) => extname(recipe.output).slice(1);

scripts/lib/hash-file.mjs

@@ -0,0 +1,10 @@
+import { createHash } from 'node:crypto';
+import { createReadStream } from 'node:fs';
+
+export const hashFile = async (path) => new Promise((resolve, reject) => {
+  const digest = createHash('sha256');
+  const stream = createReadStream(path);
+  stream.on('error', reject);
+  stream.on('data', (chunk) => digest.update(chunk));
+  stream.on('end', () => resolve(digest.digest('hex')));
+});

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 };
+};

scripts/lib/video-frames.mjs

@@ -0,0 +1,74 @@
+import { execFile } from 'node:child_process';
+import { mkdir, readFile } from 'node:fs/promises';
+import { join } from 'node:path';
+import { promisify } from 'node:util';
+
+const execFileAsync = promisify(execFile);
+const fail = (rule) => { throw new Error(rule); };
+
+export const runMediaCommand = async (command, args, rule, timeout = 60_000) => {
+  try {
+    return await execFileAsync(command, args, {
+      encoding: 'utf8', timeout, maxBuffer: 8 * 1024 * 1024, windowsHide: true,
+    });
+  } catch {
+    fail(rule);
+  }
+};
+
+export const probeVideo = async (path, options = {}) => {
+  const { stdout } = await runMediaCommand('ffprobe', [
+    '-v', 'error', '-show_entries', 'format=duration:stream=index,codec_type,width,height,duration',
+    '-of', 'json', path,
+  ], 'ASSET-VIDEO-PROBE-001', 20_000);
+  let probe;
+  try { probe = JSON.parse(stdout); } catch { fail('ASSET-VIDEO-PROBE-001'); }
+  const videoStreams = probe.streams?.filter((stream) => stream.codec_type === 'video') ?? [];
+  const audioStreams = probe.streams?.filter((stream) => stream.codec_type === 'audio') ?? [];
+  const unexpected = probe.streams?.filter((stream) => !['video', 'audio'].includes(stream.codec_type)) ?? [];
+  const duration = Number(probe.format?.duration ?? videoStreams[0]?.duration);
+  if (videoStreams.length !== 1 || unexpected.length > 0 || (options.allowAudio === false && audioStreams.length > 0) ||
+    !Number.isFinite(duration) || duration <= 0 || duration > 120) {
+    fail('ASSET-VIDEO-PROBE-001');
+  }
+  return { duration, width: Number(videoStreams[0].width), height: Number(videoStreams[0].height) };
+};
+
+export const encodeVideoDerivatives = async ({ source, webmPath, mp4Path, posterPath, startSeconds, durationSeconds, maxWidth }) => {
+  const common = [
+    '-hide_banner', '-loglevel', 'error', '-nostdin', '-y', '-threads', '1',
+    '-ss', String(startSeconds), '-i', source, '-t', String(durationSeconds), '-map', '0:v:0',
+    '-an', '-sn', '-dn', '-vf', `scale=w='min(${maxWidth},iw)':h=-2:flags=lanczos`,
+    '-map_metadata', '-1', '-map_chapters', '-1', '-pix_fmt', 'yuv420p',
+  ];
+  await runMediaCommand('ffmpeg', [
+    ...common, '-fflags', '+bitexact', '-flags:v', '+bitexact',
+    '-c:v', 'libvpx-vp9', '-deadline', 'good', '-cpu-used', '4', '-row-mt', '0',
+    '-b:v', '0', '-crf', '34', '-threads', '1', webmPath,
+  ], 'ASSET-VIDEO-ENCODE-001', 120_000);
+  await runMediaCommand('ffmpeg', [
+    ...common, '-fflags', '+bitexact', '-flags:v', '+bitexact',
+    '-c:v', 'libx264', '-preset', 'slow', '-crf', '25', '-threads', '1',
+    '-movflags', '+faststart', mp4Path,
+  ], 'ASSET-VIDEO-ENCODE-001', 120_000);
+  await runMediaCommand('ffmpeg', [
+    '-hide_banner', '-loglevel', 'error', '-nostdin', '-y', '-threads', '1',
+    '-ss', '0.05', '-i', webmPath, '-frames:v', '1', '-an', '-sn', '-dn', '-map_metadata', '-1', posterPath,
+  ], 'ASSET-VIDEO-POSTER-001', 30_000);
+};
+
+export const extractVideoFrames = async ({ source, durationSeconds, outputDir, prefix = 'frame' }) => {
+  await mkdir(outputDir, { recursive: true });
+  const points = [...new Set([0.05, durationSeconds / 2, Math.max(0.05, durationSeconds - 0.1)].map((value) => Math.min(durationSeconds, Math.max(0, value)).toFixed(3)))];
+  const paths = [];
+  for (const [index, point] of points.entries()) {
+    const path = join(outputDir, `${prefix}-${String(index + 1).padStart(2, '0')}.png`);
+    await runMediaCommand('ffmpeg', [
+      '-hide_banner', '-loglevel', 'error', '-nostdin', '-y', '-threads', '1',
+      '-ss', point, '-i', source, '-frames:v', '1', '-an', '-sn', '-dn', '-map_metadata', '-1', path,
+    ], 'ASSET-VIDEO-FRAME-001', 30_000);
+    if ((await readFile(path)).length === 0) fail('ASSET-VIDEO-FRAME-001');
+    paths.push(path);
+  }
+  return paths;
+};