Feat: Add NewsArticle JSON-LD, large image previews and automatic modified dates

Articles get NewsArticle structured data (headline, image, dates, author,
publisher) plus article:published_time and article:modified_time, and the
homepage gets WebSite and Organization data. Every indexable page allows
max-image-preview:large. The modified date is the article folder's last
commit, so editing an article updates it, the sitemap lastmod included.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

scripts/generate-pages.mjs

@@ -1,4 +1,5 @@
 import fs from "node:fs/promises";
+import { execFileSync } from "node:child_process";
 import path from "node:path";
 import fg from "fast-glob";
 import matter from "gray-matter";
@@ -21,6 +22,17 @@ const getUnlistedSlugs = async () => {
   }
 };
 
+// When the article's folder was last committed, so editing an article (text or images) updates its modified date on
+// its own. CI checks out full history for this. "" for an uncommitted article, which then counts as unmodified.
+const lastCommit = (dir) => {
+  try {
+    const d = execFileSync("git", ["log", "-1", "--format=%aI", "--", dir], { encoding: "utf8" }).trim();
+    return d && new Date(d).toISOString();
+  } catch {
+    return "";
+  }
+};
+
 const copyArticleAssets = async (file, slug) => {
   const dir = path.dirname(file);
   const entries = await fs.readdir(dir, { withFileTypes: true });
@@ -54,6 +66,7 @@ const run = async () => {
       sourceFile: file,
       markdown: md,
       isUnlisted: unlisted.has(data.slug),
+      modified: lastCommit(path.dirname(file)),
       html: fixInternalLinks(marked.parse(content))
     });
   }

scripts/lib/renderers.mjs

@@ -1,6 +1,45 @@
 import { SITE_URL, TURNSTILE_SITE_KEY } from "./constants.mjs";
 import { footer, merch, nav, shellHead } from "./templates.mjs";
-import { escapeHtml, escapeXml, fmtDate, renderAuthorInline, toISODate } from "./utils.mjs";
+import { escapeHtml, escapeXml, fmtDate, renderAuthorInline, toISO, toISODate } from "./utils.mjs";
+
+// Structured data (JSON-LD) so Google reads the headline, image, dates and authors directly instead of guessing
+const absUrl = (u = "") => (/^https?:\/\//.test(u) ? u : `${SITE_URL}/${u.replace(/^\/+/, "")}`);
+const publisher = { "@type": "Organization", name: "apophenia.news", url: `${SITE_URL}/`, logo: `${SITE_URL}/favicon.svg` };
+
+// Linked names in the byline (people with an author page) become Person authors. Unlinked co-authors are AI models,
+// which schema.org has no type for, so they stay in the visible byline only.
+const authorsLd = (author = "") => {
+  const people = [...String(author).matchAll(/\[([^\]]+)\]\(([^)]+)\)/g)].map(([, name, href]) => ({ "@type": "Person", name: name.trim(), url: absUrl(href.trim()) }));
+  return people.length ? people : [publisher];
+};
+
+// Published is the frontmatter date. Modified is the article folder's last commit (see generate-pages.mjs), never
+// earlier than published.
+const articleDates = (a) => {
+  const published = toISO(a.date);
+  return { published, modified: a.modified > published ? a.modified : published };
+};
+
+const articleLd = (a, { published, modified }) => ({
+  "@context": "https://schema.org",
+  "@type": "NewsArticle",
+  headline: a.title,
+  description: a.description,
+  image: [absUrl(a.header_image)],
+  datePublished: published,
+  dateModified: modified,
+  author: authorsLd(a.author),
+  publisher,
+  mainEntityOfPage: `${SITE_URL}/${a.slug}/`
+});
+
+const homeLd = {
+  "@context": "https://schema.org",
+  "@graph": [
+    { "@type": "WebSite", name: "apophenia.news", url: `${SITE_URL}/` },
+    { ...publisher, sameAs: ["https://youtube.com/@ApopheniaLoop"] }
+  ]
+};
 
 export const renderHome = (articles) => `
 ${shellHead({
@@ -8,7 +47,8 @@ ${shellHead({
   desc: "Signals, anomalies, civilization trajectories, and deep pattern analysis.",
   image: "https://free.direct-img.link/celestial+map?i=1",
   url: `${SITE_URL}/`,
-  type: "website"
+  type: "website",
+  jsonLd: homeLd
 })}
 ${nav}
 <main class="shell py-10">
@@ -54,13 +94,15 @@ ${nav}
 ${footer}
 `;
 
-export const renderArticle = (article) => `
+export const renderArticle = (article, dates = articleDates(article)) => `
 ${shellHead({
   title: `${article.title} — apophenia.news`,
   desc: article.description,
   image: article.header_image,
   url: `${SITE_URL}/${article.slug}/`,
-  type: "article"
+  type: "article",
+  ...dates,
+  jsonLd: articleLd(article, dates)
 })}
 ${nav}
 <main class="shell py-10">
@@ -408,7 +450,7 @@ export const renderSitemap = (articles, authors = []) => {
     { loc: `${SITE_URL}/newsletter/`, lastmod: now },
     { loc: `${SITE_URL}/write/`, lastmod: now },
     ...authors.map((a) => ({ loc: `${SITE_URL}/author/${a.slug}/`, lastmod: now })),
-    ...articles.map((a) => ({ loc: `${SITE_URL}/${a.slug}/`, lastmod: toISODate(a.date || new Date()) }))
+    ...articles.map((a) => ({ loc: `${SITE_URL}/${a.slug}/`, lastmod: toISODate(articleDates(a).modified) }))
   ];
 
   return `<?xml version="1.0" encoding="UTF-8"?>

scripts/lib/templates.mjs

@@ -72,7 +72,7 @@ export const merch = `
 </section>
 `;
 
-export const shellHead = ({ title, desc, image, url, type = "website", noindex = false }) => `<!doctype html>
+export const shellHead = ({ title, desc, image, url, type = "website", noindex = false, published, modified, jsonLd }) => `<!doctype html>
 <html lang="en">
 <head>
   <meta charset="UTF-8" />
@@ -85,7 +85,10 @@ export const shellHead = ({ title, desc, image, url, type = "website", noindex =
   ${url ? `<meta property="og:url" content="${url}" />` : ""}
   ${image ? `<meta property="og:image" content="${image}" />` : ""}
   <meta name="theme-color" content="#fcfcfa" />
-  ${noindex ? `<meta name="robots" content="noindex" />` : ""}
+  <meta name="robots" content="${noindex ? "noindex" : "max-image-preview:large"}" />
+  ${published ? `<meta property="article:published_time" content="${published}" />` : ""}
+  ${modified ? `<meta property="article:modified_time" content="${modified}" />` : ""}
+  ${jsonLd ? `<script type="application/ld+json">${JSON.stringify(jsonLd).replace(/</g, "\\u003c")}</script>` : ""}
   ${url ? `<link rel="canonical" href="${url}" />` : ""}
   <link rel="icon" type="image/svg+xml" href="/favicon.svg" />
   <link rel="alternate" type="application/rss+xml" title="apophenia.news RSS" href="${SITE_URL}/rss.xml" />

scripts/lib/utils.mjs

@@ -3,6 +3,8 @@ export const fmtDate = (date) =>
 
 export const toISODate = (d) => new Date(d).toISOString().split("T")[0];
 
+export const toISO = (d) => new Date(d).toISOString();
+
 export const escapeHtml = (s = "") =>
   String(s ?? "").replace(/[&<>"']/g, (ch) => ({ "&": "&amp;", "<": "&lt;", ">": "&gt;", '"': "&quot;", "'": "&#39;" }[ch]));