familyfedie-website/scripts/extract-content.mjs

import fs from "node:fs";
import path from "node:path";

const repoRoot = process.cwd();
const legacyHtmlRoot = path.join(repoRoot, "archive-source");
const contentRoot = path.join(repoRoot, "content");
const contentReadme = `# Content

This directory is the future source of truth for editable site content.

These Markdown files were extracted from the exported HTML posts in \`archive-source/\`.
The body content intentionally keeps its original HTML markup so the migration
does not lose formatting, links, headings, or embedded images.

- \`content/speeches/<year>/\` - Rev. Sun Myung Moon speech posts grouped by speech year
- \`content/speeches/mrs-hak-ja-han-moon/\` - Mrs. Hak Ja Han Moon speech posts
- \`content/blog/<year>/\` - non-speech blog posts grouped by published year

Regenerate this content from \`archive-source/blog/\` with:

\`\`\`bash
npm run extract:content
\`\`\`
`;

function walk(dir) {
  return fs.readdirSync(dir, { withFileTypes: true }).flatMap((entry) => {
    const fullPath = path.join(dir, entry.name);
    if (entry.isDirectory()) {
      return walk(fullPath);
    }
    return fullPath.endsWith(".html") ? [fullPath] : [];
  });
}

function decodeEntities(value = "") {
  return value
    .replace(/&#8211;/g, "-")
    .replace(/&#8212;/g, "-")
    .replace(/&#8216;/g, "'")
    .replace(/&#8217;/g, "'")
    .replace(/&#8220;/g, '"')
    .replace(/&#8221;/g, '"')
    .replace(/&raquo;/g, "»")
    .replace(/&laquo;/g, "«")
    .replace(/&amp;/g, "&")
    .replace(/&quot;/g, '"')
    .replace(/&#039;/g, "'")
    .replace(/&nbsp;/g, " ")
    .replace(/&hellip;/g, "...")
    .replace(/<[^>]*>/g, "")
    .replace(/\s+/g, " ")
    .trim();
}

function yamlString(value) {
  return JSON.stringify(value ?? "");
}

function yamlArray(key, values) {
  if (!values.length) {
    return `${key}: []`;
  }
  return [`${key}:`, ...values.map((value) => `  - ${yamlString(value)}`)].join("\n");
}

function matchFirst(html, pattern) {
  return html.match(pattern)?.[1] ?? "";
}

function extractLinks(html, containerClass) {
  const startMatch = html.match(new RegExp(`<[^>]*class=["'][^"']*${containerClass}[^"']*["'][^>]*>`));
  if (!startMatch?.index) {
    return [];
  }

  const start = startMatch.index + startMatch[0].length;
  const afterStart = html.slice(start);
  const endIndex = afterStart.search(containerClass === "footer-tags" ? /<\/div>/ : /<\/span>/);
  const body = endIndex === -1 ? afterStart : afterStart.slice(0, endIndex);

  return Array.from(body.matchAll(/<a\s+[^>]*href=["']([^"']+)["'][^>]*>([\s\S]*?)<\/a>/g)).map((match) => ({
    href: match[1],
    label: decodeEntities(match[2]),
  }));
}

function extractEntryContent(html) {
  const content = matchFirst(html, /<div class="entry-content">\s*([\s\S]*?)\s*<\/div><!-- \.entry-content -->/);
  return content
    .replace(/<div class="pdfprnt-buttons[\s\S]*?<\/div>/g, "")
    .replace(/\n[ \t]+/g, "\n")
    .trim();
}

function slugFromPath(filePath) {
  return path.basename(filePath, ".html");
}

function relativeUrl(filePath) {
  return `/${path.relative(legacyHtmlRoot, filePath).split(path.sep).join("/")}`;
}

function destinationFor(post) {
  const speechCategory = post.categories.find((category) => category.href.startsWith("/speeches/categories/"));
  if (speechCategory) {
    const year =
      speechCategory.href.match(/(?:^|-)(1\d{3}|20\d{2})(?:-|\.html$)/)?.[1] ??
      speechCategory.label.match(/\b(1\d{3}|20\d{2})\b/)?.[1];
    const collection =
      year ??
      path.basename(speechCategory.href, ".html").replace(/^rev-sun-myung-moon$/, "rev-sun-myung-moon-general");
    return path.join(contentRoot, "speeches", collection, `${post.slug}.md`);
  }

  const publishYear = post.published.match(/^\d{4}/)?.[0] ?? "undated";
  return path.join(contentRoot, "blog", publishYear, `${post.slug}.md`);
}

function extractPost(filePath) {
  const html = fs.readFileSync(filePath, "utf8");
  const entryClass = matchFirst(html, /<div id="post-[^"]*" class="([^"]*post type-post[^"]*)">/);
  if (!entryClass) {
    return null;
  }

  const title = decodeEntities(matchFirst(html, /<h1 class="entry-title">([\s\S]*?)<\/h1>/));
  const published = matchFirst(html, /<time class="onDate date published" datetime="([^"]+)"/);
  const updated = matchFirst(html, /<time class="updated"\s+datetime="([^"]+)"/);
  const categories = extractLinks(html, "bl_categ");
  const tags = extractLinks(html, "footer-tags");
  const content = extractEntryContent(html);
  const slug = slugFromPath(filePath);
  const source = relativeUrl(filePath);

  return {
    title,
    published,
    updated,
    source,
    slug,
    categories,
    tags,
    content,
  };
}

function renderMarkdown(post) {
  const speechCategory = post.categories.find((category) => category.href.startsWith("/speeches/categories/"));
  const type = speechCategory ? "speech" : "blog";
  const speechYear =
    speechCategory?.href.match(/(?:^|-)(1\d{3}|20\d{2})(?:-|\.html$)/)?.[1] ??
    speechCategory?.label.match(/\b(1\d{3}|20\d{2})\b/)?.[1];
  const speechCollection =
    speechCategory && !speechYear
      ? path.basename(speechCategory.href, ".html").replace(/^rev-sun-myung-moon$/, "rev-sun-myung-moon-general")
      : null;
  const categoryLabels = post.categories.map((category) => category.label);
  const tagLabels = post.tags.map((tag) => tag.label);

  const frontmatter = [
    "---",
    `title: ${yamlString(post.title)}`,
    `type: ${yamlString(type)}`,
    `published: ${yamlString(post.published)}`,
    `updated: ${yamlString(post.updated)}`,
    `source: ${yamlString(post.source)}`,
    speechYear ? `speechYear: ${yamlString(speechYear)}` : null,
    speechCollection ? `speechCollection: ${yamlString(speechCollection)}` : null,
    yamlArray("categories", categoryLabels),
    yamlArray("tags", tagLabels),
    "---",
  ]
    .filter(Boolean)
    .join("\n");

  return `${frontmatter}\n\n${post.content}\n`;
}

fs.rmSync(contentRoot, { recursive: true, force: true });

const postFiles = walk(path.join(legacyHtmlRoot, "blog")).filter((filePath) => /\/\d{4}\/\d{2}\/\d{2}\//.test(filePath));
const posts = postFiles.map(extractPost).filter(Boolean);

for (const post of posts) {
  const destination = destinationFor(post);
  fs.mkdirSync(path.dirname(destination), { recursive: true });
  fs.writeFileSync(destination, renderMarkdown(post));
}

fs.writeFileSync(path.join(contentRoot, "README.md"), contentReadme);

const speechCount = posts.filter((post) => destinationFor(post).includes(`${path.sep}speeches${path.sep}`)).length;
const blogCount = posts.length - speechCount;

console.log(`Extracted ${posts.length} posts to content/.`);
console.log(`- speeches: ${speechCount}`);
console.log(`- blog: ${blogCount}`);