SnapOtter/tests/fixtures/gen-synthetic-content.mjs

#!/usr/bin/env node
// gen-synthetic-content.mjs -- Regenerates Category A & B content fixtures as CC0.
//
// Category A: Deterministically regenerable (QR codes, barcodes, OCR text, PDFs,
//             synthetic audio/video).
// Category B: Replaces copyrighted content (Simpsons GIF, ConvertICO SVG,
//             Shutterstock stock photo) with project-owned CC0 alternatives.
//
// All output is provably CC0 (generated by project scripts, no third-party IP).
// Usage: node tests/fixtures/gen-synthetic-content.mjs
//
// Dependencies (resolved from apps/api/node_modules):
//   sharp, qrcode
// System deps on PATH:
//   ffmpeg

import { existsSync, writeFileSync } from "node:fs";
import { dirname, join } from "node:path";
import { execSync } from "node:child_process";
import { fileURLToPath } from "node:url";
import { createRequire } from "node:module";

const __dirname = dirname(fileURLToPath(import.meta.url));
const IMAGE_VALID = join(__dirname, "image/valid");
const DOC_VALID = join(__dirname, "document/valid");
const require = createRequire(join(__dirname, "../../apps/api/package.json"));
const sharp = require("sharp");
const QRCode = require("qrcode");

const FORCE = process.argv.includes("--force");

// Write file only if it does not already exist (or --force is set).
// Committed synthetics must not be silently overwritten because
// encoder-version differences can change bytes and break manifest hashes.
function writeIfMissing(path, data, label) {
  if (!FORCE && existsSync(path)) {
    console.log(`  SKIP (exists): ${label || path}`);
    return false;
  }
  writeFileSync(path, data);
  console.log(`  ${label || path}`);
  return true;
}

// ── Code 128B barcode encoder ──────────────────────────────

// Standard Code 128 bar patterns (11 modules each, stop = 13 modules).
// Each string is a binary pattern: 1 = dark bar, 0 = light space.
const CODE128 = [
  "11011001100","11001101100","11001100110","10010011000","10010001100",
  "10001001100","10011001000","10011000100","10001100100","11001001000",
  "11001000100","11000100100","10110011100","10011011100","10011001110",
  "10111001100","10011101100","10011100110","11001110010","11001011100",
  "11001001110","11011100100","11001110100","11101101110","11101001100",
  "11100101100","11100100110","11101100100","11100110100","11100110010",
  "11011011000","11011000110","11000110110","10100011000","10001011000",
  "10001000110","10110001000","10001101000","10001100010","11010001000",
  "11000101000","11000100010","10110111000","10110001110","10001101110",
  "10111011000","10111000110","10001110110","11101110110","11010001110",
  "11000101110","11011101000","11011100010","11011101110","11101011000",
  "11101000110","11100010110","11101101000","11101100010","11100011010",
  "11101111010","11001000010","11110001010","10100110000","10100001100",
  "10010110000","10010000110","10000101100","10000100110","10110010000",
  "10110000100","10011010000","10011000010","10000110100","10000110010",
  "11000010010","11001010000","11110111010","11000010100","10001111010",
  "10100111100","10010111100","10010011110","10111100100","10011110100",
  "10011110010","11110100100","11110010100","11110010010","11011011110",
  "11011110110","11110110110","10101111000","10100011110","10001011110",
  "10111101000","10111100010","11110101000","11110100010","10111011110",
  "10111101110","11101011110","11110101110",
  "11010000100", // 103 Start A
  "11010010000", // 104 Start B
  "11010011100", // 105 Start C
];
const CODE128_STOP = "1100011101011";

function encodeCode128B(text) {
  const START_B = 104;
  const values = [START_B];
  let checksum = START_B;
  for (let i = 0; i < text.length; i++) {
    const v = text.charCodeAt(i) - 32;
    values.push(v);
    checksum += v * (i + 1);
  }
  values.push(checksum % 103);
  let bits = "";
  for (const v of values) bits += CODE128[v];
  bits += CODE128_STOP;
  return bits;
}

function barcodeSvg(text, width, height) {
  const bits = encodeCode128B(text);
  const barH = height - 40;
  const quietZone = 20;
  const barAreaW = width - 2 * quietZone;
  const unitW = barAreaW / bits.length;
  const bars = [];
  for (let i = 0; i < bits.length; i++) {
    if (bits[i] === "1") {
      bars.push(`<rect x="${(quietZone + i * unitW).toFixed(2)}" y="10" width="${unitW.toFixed(2)}" height="${barH}" fill="black"/>`);
    }
  }
  return `<svg width="${width}" height="${height}" xmlns="http://www.w3.org/2000/svg">
  <rect width="${width}" height="${height}" fill="white"/>
  ${bars.join("\n  ")}
  <text x="${width / 2}" y="${height - 8}" text-anchor="middle" font-family="monospace, Courier" font-size="16" fill="black">${text}</text>
</svg>`;
}

// ── Minimal PDF generator ──────────────────────────────────

function generatePdf(pageCount) {
  const objects = [];
  let nextObj = 1;
  const addObj = (content) => {
    const id = nextObj++;
    objects.push({ id, content });
    return id;
  };

  const catalogId = addObj(null);
  const pagesId = addObj(null);
  const fontId = addObj("<< /Type /Font /Subtype /Type1 /BaseFont /Helvetica >>");

  const pageIds = [];
  for (let i = 1; i <= pageCount; i++) {
    const stream = `BT /F1 24 Tf 72 700 Td (Page ${i}) Tj ET`;
    const sId = addObj(`<< /Length ${stream.length} >>\nstream\n${stream}\nendstream`);
    const pId = addObj(
      `<< /Type /Page /Parent ${pagesId} 0 R /MediaBox [0 0 612 792] ` +
      `/Contents ${sId} 0 R /Resources << /Font << /F1 ${fontId} 0 R >> >> >>`,
    );
    pageIds.push(pId);
  }

  objects[0].content = `<< /Type /Catalog /Pages ${pagesId} 0 R >>`;
  objects[1].content =
    `<< /Type /Pages /Kids [${pageIds.map((id) => `${id} 0 R`).join(" ")}] /Count ${pageCount} >>`;

  let pdf = "%PDF-1.7\n%\x81\x81\x81\x81\n\n";
  const offsets = [];
  for (const obj of objects) {
    offsets.push(pdf.length);
    pdf += `${obj.id} 0 obj\n${obj.content}\nendobj\n\n`;
  }

  const xrefOffset = pdf.length;
  pdf += `xref\n0 ${objects.length + 1}\n`;
  pdf += "0000000000 65535 f \n";
  for (const off of offsets) {
    pdf += `${String(off).padStart(10, "0")} 00000 n \n`;
  }
  pdf += `trailer\n<< /Size ${objects.length + 1} /Root ${catalogId} 0 R >>\n`;
  pdf += `startxref\n${xrefOffset}\n%%EOF\n`;
  return Buffer.from(pdf, "binary");
}

// ── SVG logo (project-owned, replaces ConvertICO brand) ───

function projectSvg() {
  return `<svg width="400" height="400" viewBox="0 0 400 400" fill="none" xmlns="http://www.w3.org/2000/svg">
  <title>SnapOtter Test Fixture</title>
  <desc>CC0 geometric test SVG generated by gen-synthetic-content.mjs</desc>
  <defs>
    <linearGradient id="bg" x1="0" y1="0" x2="400" y2="400">
      <stop stop-color="#E07832"/>
      <stop offset="1" stop-color="#C06520"/>
    </linearGradient>
    <filter id="glow" x="-20%" y="-20%" width="140%" height="140%">
      <feGaussianBlur in="SourceGraphic" stdDeviation="4" result="blur"/>
      <feMerge><feMergeNode in="blur"/><feMergeNode in="SourceGraphic"/></feMerge>
    </filter>
  </defs>
  <rect width="400" height="400" fill="#1a1210"/>
  <circle cx="200" cy="180" r="100" fill="url(#bg)" filter="url(#glow)"/>
  <ellipse cx="200" cy="180" rx="60" ry="40" fill="#1a1210" opacity="0.3"/>
  <circle cx="170" cy="160" r="12" fill="white"/>
  <circle cx="230" cy="160" r="12" fill="white"/>
  <circle cx="170" cy="160" r="5" fill="#1a1210"/>
  <circle cx="230" cy="160" r="5" fill="#1a1210"/>
  <ellipse cx="200" cy="195" rx="15" ry="8" fill="#1a1210" opacity="0.4"/>
  <text x="200" y="320" font-family="Helvetica, Arial, sans-serif" font-size="28" font-weight="bold" text-anchor="middle" fill="#F0EBE4">TEST FIXTURE</text>
  <text x="200" y="350" font-family="Helvetica, Arial, sans-serif" font-size="14" text-anchor="middle" fill="#6B6560">CC0 - Generated Asset</text>
</svg>`;
}

// ── Multi-face placeholder (replaces Shutterstock photo) ──

function multiFaceSvg(w, h) {
  const cols = 5;
  const rows = 2;
  const cw = w / cols;
  const ch = h / rows;
  const colors = [
    "#D4956A", "#C2885E", "#E8B88A", "#A67853", "#BF9570",
    "#D9A87C", "#CB9068", "#E5C49E", "#B08060", "#C89878",
  ];
  const shapes = [];
  for (let r = 0; r < rows; r++) {
    for (let c = 0; c < cols; c++) {
      const cx = cw * c + cw / 2;
      const cy = ch * r + ch / 2;
      const rx = cw * 0.3;
      const ry = ch * 0.38;
      const idx = r * cols + c;
      shapes.push(`<ellipse cx="${cx.toFixed(0)}" cy="${cy.toFixed(0)}" rx="${rx.toFixed(0)}" ry="${ry.toFixed(0)}" fill="${colors[idx]}" stroke="#888" stroke-width="0.5"/>`);
      shapes.push(`<circle cx="${(cx - rx * 0.3).toFixed(0)}" cy="${(cy - ry * 0.15).toFixed(0)}" r="3" fill="#333"/>`);
      shapes.push(`<circle cx="${(cx + rx * 0.3).toFixed(0)}" cy="${(cy - ry * 0.15).toFixed(0)}" r="3" fill="#333"/>`);
    }
  }
  return `<svg width="${w}" height="${h}" xmlns="http://www.w3.org/2000/svg">
  <rect width="${w}" height="${h}" fill="#d0d0d0"/>
  ${shapes.join("\n  ")}
</svg>`;
}

// ── ffmpeg helper (all args are hardcoded constants) ───────

function ff(args, label) {
  try {
    execSync(`ffmpeg ${args}`, { stdio: "pipe" });
    console.log(`  ${label}`);
  } catch (e) {
    console.error(`  FAILED: ${label}`, e.stderr?.toString().slice(-300));
    throw e;
  }
}

// Like ff(), but skip if the output file already exists (unless --force).
function ffIfMissing(outPath, args, label) {
  if (!FORCE && existsSync(outPath)) {
    console.log(`  SKIP (exists): ${label}`);
    return;
  }
  ff(args, label);
}

// ── Main ───────────────────────────────────────────────────
//
// Scoped outputs:
//   - image/valid/  : QR codes, barcodes, OCR text, SVG logo, multi-face, animated GIF
//   - document/valid/: alt-2page.pdf, multipage-6.pdf
//
// Deliberately EXCLUDED (committed real heroes, not synthetics):
//   - video/valid/media-30s.mp4  (Big Buck Bunny CC-BY clip)
//   - audio/valid/media-30s.wav  (committed TTS/synthetic, hash-locked)
//   - video/valid/speech-10s.mp4 (macOS TTS, committed)
//   - audio/valid/speech-*.{wav,flac,ogg,m4a,aac,opus} (macOS TTS, committed)
//   - video/valid/hero.{mov,webm,mkv,avi} (Big Buck Bunny CC-BY, committed)

async function main() {
  console.log("Regenerating Category A & B content fixtures...");
  console.log(`Mode: ${FORCE ? "FORCE (overwrite all)" : "skip existing files (pass --force to overwrite)"}\n`);

  // ── QR codes (A) ──
  console.log("QR codes:");
  const qrData = "https://snapotter.com";
  const qrPng = await QRCode.toBuffer(qrData, {
    width: 400, margin: 2, color: { dark: "#000000", light: "#ffffff" },
  });
  writeIfMissing(join(IMAGE_VALID, "qr-code.png"), qrPng, "qr-code.png");

  const qrSvg = await QRCode.toString(qrData, { type: "svg", width: 296, margin: 2 });
  writeIfMissing(join(IMAGE_VALID, "qr-code.svg"), qrSvg, "qr-code.svg");

  const qrAvif = await sharp(qrPng).resize(400, 400).avif({ quality: 80 }).toBuffer();
  writeIfMissing(join(IMAGE_VALID, "qr-code.avif"), qrAvif, "qr-code.avif");

  // ── Barcodes (A) ──
  console.log("Barcodes:");
  const bcText = "SNAPOTTER-TEST-123";
  const bcSvgStr = barcodeSvg(bcText, 699, 152);
  const bcPng = await sharp(Buffer.from(bcSvgStr)).png().toBuffer();
  writeIfMissing(join(IMAGE_VALID, "barcode.png"), bcPng, "barcode.png");

  const bcAvif = await sharp(Buffer.from(bcSvgStr)).avif({ quality: 80 }).toBuffer();
  writeIfMissing(join(IMAGE_VALID, "barcode.avif"), bcAvif, "barcode.avif");

  // ── OCR text images (A) ──
  console.log("OCR text images:");
  const ocrEngSvg = `<svg width="800" height="200" xmlns="http://www.w3.org/2000/svg">
    <rect width="800" height="200" fill="white"/>
    <text x="400" y="115" text-anchor="middle" font-family="Helvetica, Arial, sans-serif" font-size="42" font-weight="bold" fill="black">The quick brown fox 12345</text>
  </svg>`;
  const ocrEngPng = await sharp(Buffer.from(ocrEngSvg)).png().toBuffer();
  writeIfMissing(join(IMAGE_VALID, "ocr-clean.png"), ocrEngPng, "ocr-clean.png");

  const jpText =
    "日本語の表記においては，漢字や仮名だけで" +
    "なく，ローマ字やアラビア数字，さらに句読" +
    "点や括弧類などの記述記号を用いる。これら" +
    "を組み合わせて表す日本語の文書では，表記" +
    "上における種々の問題がある。";
  const chars = [...jpText];
  const lineLen = Math.ceil(chars.length / 6);
  const lines = [];
  for (let i = 0; i < chars.length; i += lineLen) {
    lines.push(chars.slice(i, i + lineLen).join(""));
  }
  const jpSvg = `<svg width="480" height="172" xmlns="http://www.w3.org/2000/svg">
    <rect width="480" height="172" fill="white"/>
    ${lines.map((line, i) => `<text x="16" y="${30 + i * 28}" font-family="Hiragino Sans, Hiragino Kaku Gothic Pro, sans-serif" font-size="20" fill="black">${line}</text>`).join("\n    ")}
  </svg>`;
  const jpPng = await sharp(Buffer.from(jpSvg)).png().toBuffer();
  writeIfMissing(join(IMAGE_VALID, "ocr-japanese.png"), jpPng, "ocr-japanese.png");

  // ── PDFs (A: deterministic generator, safe to overwrite) ──
  console.log("PDFs:");
  writeIfMissing(join(DOC_VALID, "alt-2page.pdf"), generatePdf(2), "alt-2page.pdf");
  writeIfMissing(join(DOC_VALID, "multipage-6.pdf"), generatePdf(6), "multipage-6.pdf");

  // ── SVG logo (B: replaces ConvertICO brand) ──
  console.log("SVG logo (replacing ConvertICO brand):");
  writeIfMissing(join(IMAGE_VALID, "svg-logo.svg"), projectSvg(), "svg-logo.svg");

  // ── Multi-face placeholder (B: replaces Shutterstock photo) ──
  console.log("Multi-face placeholder (replacing Shutterstock stock photo):");
  const mfSvg = multiFaceSvg(433, 280);
  const mfWebp = await sharp(Buffer.from(mfSvg)).webp({ quality: 80 }).toBuffer();
  writeIfMissing(join(IMAGE_VALID, "multi-face.webp"), mfWebp, "multi-face.webp");

  // ── Animated GIF (B: replaces copyrighted Simpsons clip) ──
  console.log("Animated GIF (replacing copyrighted Simpsons clip):");
  const gifOut = join(IMAGE_VALID, "animated-simpsons.gif");
  ffIfMissing(
    gifOut,
    `-f lavfi -i "testsrc=duration=3:size=320x320:rate=10" -pix_fmt rgb8 -loop 0 -y "${gifOut}"`,
    "animated-simpsons.gif",
  );

  // ── Synthetic audio/video (A) ──
  // NOTE: media-30s.mp4 and media-30s.wav are no longer generated here.
  // media-30s.mp4 is now a real Big Buck Bunny CC-BY hero clip (committed).
  // media-30s.wav is a committed synthetic whose hash is locked in manifest.json.
  // Regenerating either would clobber the committed content with different bytes.
  console.log("Synthetic audio/video (metadata-tagged synthetics only):");

  const audioTagsOut = join(__dirname, "audio/valid/audio-with-tags.mp3");
  ffIfMissing(
    audioTagsOut,
    `-f lavfi -i "sine=frequency=440:duration=1.2:sample_rate=8000" ` +
    `-c:a libmp3lame -b:a 64k -ar 8000 -ac 1 ` +
    `-metadata title="Test Song" -metadata artist="Test Artist" ` +
    `-metadata album="Test Album" -metadata date="2026" ` +
    `-metadata genre="Electronic" -metadata track="1/10" ` +
    `-y "${audioTagsOut}"`,
    "audio-with-tags.mp3",
  );

  const videoMetaOut = join(__dirname, "video/valid/video-with-meta.mp4");
  ffIfMissing(
    videoMetaOut,
    `-f lavfi -i "color=c=0xE07832:s=64x64:d=1:rate=16" ` +
    `-c:v libx264 -pix_fmt yuv420p -movflags +faststart ` +
    `-y "${videoMetaOut}"`,
    "video-with-meta.mp4",
  );

  console.log("\nDone. Run gen-manifest.mjs to re-stamp hashes if any files were regenerated.");
}

main().catch((e) => {
  console.error("FATAL:", e);
  process.exit(1);
});