migrate.ts

import { readFileSync, writeFileSync, mkdirSync, existsSync } from "fs";
import { basename, extname } from "path";

const SITE_DIR = "./site";
const ASSETS_DIR = `${SITE_DIR}/assets`;
const CDN_BASE =
  "https://cdn.prod.website-files.com/640f4a7e6dfcea721c5ea20f/";

const pages = ["index.html", "design.html", "development.html", "pricing.html"];

// 1. Read all HTML files and extract CDN URLs
const allUrls = new Set<string>();

for (const page of pages) {
  const html = readFileSync(`${SITE_DIR}/${page}`, "utf-8");
  // Match CDN URLs - handle parentheses in filenames
  const regex = new RegExp(
    `https://cdn\\.prod\\.website-files\\.com/640f4a7e6dfcea721c5ea20f/[^"'\\s>]+`,
    "g"
  );
  let match;
  while ((match = regex.exec(html)) !== null) {
    allUrls.add(match[0]);
  }
}

console.log(`Found ${allUrls.size} unique asset URLs`);

// 2. Create a mapping from CDN URL to local path
const urlMap = new Map<string, string>();

for (const url of allUrls) {
  const relativePath = url.replace(CDN_BASE, "");
  const decoded = decodeURIComponent(relativePath);

  if (decoded.includes("css/")) {
    urlMap.set(url, `assets/css/${basename(decoded)}`);
  } else if (decoded.includes("js/")) {
    // Skip Webflow JS
    continue;
  } else {
    // Clean up filename
    const clean = decoded.replace(/[() ]/g, (c) =>
      c === " " ? "-" : c === "(" ? "" : ""
    );
    urlMap.set(url, `assets/images/${clean}`);
  }
}

// 3. Download all assets
mkdirSync(`${ASSETS_DIR}/images`, { recursive: true });
mkdirSync(`${ASSETS_DIR}/css`, { recursive: true });

async function downloadAll() {
  const entries = [...urlMap.entries()];
  console.log(`Downloading ${entries.length} assets...`);

  for (const [url, localPath] of entries) {
    const fullPath = `${SITE_DIR}/${localPath}`;
    if (existsSync(fullPath)) {
      continue;
    }
    try {
      const resp = await fetch(url);
      if (!resp.ok) {
        console.error(`Failed to download ${url}: ${resp.status}`);
        continue;
      }
      const buffer = Buffer.from(await resp.arrayBuffer());
      writeFileSync(fullPath, buffer);
      console.log(`  Downloaded: ${localPath}`);
    } catch (e: any) {
      console.error(`  Error downloading ${url}: ${e.message}`);
    }
  }

  // 4. Rewrite HTML files
  for (const page of pages) {
    let html = readFileSync(`${SITE_DIR}/${page}`, "utf-8");

    // Sort URLs by length descending to avoid partial replacements
    const sortedEntries = [...urlMap.entries()].sort(
      (a, b) => b[0].length - a[0].length
    );

    for (const [url, localPath] of sortedEntries) {
      html = html.replaceAll(url, localPath);
    }

    // Remove Webflow JS script tags (but keep CSS)
    html = html.replace(
      /<script\s+src="[^"]*webflow[^"]*\.js"[^>]*><\/script>/g,
      ""
    );
    // Remove WebFont loader script tag
    html = html.replace(
      /<script\s+src="[^"]*webfont\.js"[^>]*><\/script>/g,
      ""
    );
    // Remove Webflow class manipulation inline script
    html = html.replace(
      /<script\s+type="text\/javascript">!function\(o,c\).*?<\/script>/gs,
      ""
    );
    // Remove WebFont.load inline script
    html = html.replace(
      /<script\s+type="text\/javascript">WebFont\.load\(.*?\);<\/script>/gs,
      ""
    );
    // Replace Google Fonts preconnect with proper CSS link
    html = html.replace(
      '<link href="https://fonts.gstatic.com" rel="preconnect" crossorigin="anonymous"/>',
      '<link href="https://fonts.gstatic.com" rel="preconnect" crossorigin="anonymous"/>\n<link href="https://fonts.googleapis.com/css2?family=Epilogue:wght@400;600&display=swap" rel="stylesheet"/>'
    );
    // Remove jQuery (loaded for Webflow)
    html = html.replace(
      /<script\s+src="[^"]*jquery[^"]*\.js[^"]*"[^>]*><\/script>/g,
      ""
    );
    // Add our slider JS before </body>
    html = html.replace(
      "</body>",
      '<script src="assets/js/slider.js"></script></body>'
    );

    // Remove Webflow badge / data attributes
    html = html.replace(/\s*data-wf-[a-z-]+="[^"]*"/g, "");
    html = html.replace(/\s*data-w-id="[^"]*"/g, "");

    // Fix internal links: /design -> /design.html etc. (for local serving)
    // Actually for Cloudflare Pages, clean URLs work. Let's keep them as-is.
    // But we need /design to serve design.html. Cloudflare Pages does this automatically.

    // Update www.bento-ds.com links to relative
    html = html.replace(/https:\/\/www\.bento-ds\.com\//g, "/");
    html = html.replace(/https:\/\/www\.bento-ds\.com#/g, "/#");
    html = html.replace(/https:\/\/www\.bento-ds\.com/g, "/");

    writeFileSync(`${SITE_DIR}/${page}`, html);
    console.log(`Rewrote ${page}`);
  }

  console.log("\nDone! Static site is in ./site/");
}

downloadAll();