feat: add dynamic sitemap generation and API endpoint

apps/trustlab/next.config.mjs

@@ -80,6 +80,10 @@ const nextConfig = {
         source: "/robots.txt",
         destination: "/api/v1/robots",
       },
+      {
+        source: "/sitemap.xml",
+        destination: "/api/v1/sitemap",
+      },
     ];
   },
 };

apps/trustlab/src/lib/data/common/sitemap.js

@@ -0,0 +1,134 @@
+import { formatPagePath } from "@commons-ui/payload";
+
+import { site } from "@/trustlab/utils";
+
+function normalizePathname(pathname) {
+  if (!pathname || typeof pathname !== "string") {
+    return null;
+  }
+
+  if (pathname === "/") {
+    return pathname;
+  }
+
+  const trimmed = pathname.trim();
+  if (!trimmed) {
+    return null;
+  }
+
+  const withLeadingSlash = trimmed.startsWith("/") ? trimmed : `/${trimmed}`;
+  return withLeadingSlash.replace(/\/+$/, "");
+}
+
+function getAbsoluteUrl(pathname) {
+  const normalizedPathname = normalizePathname(pathname);
+  if (!normalizedPathname) {
+    return null;
+  }
+
+  const siteUrl = site.url.replace(/\/+$/, "");
+  return `${siteUrl}${normalizedPathname}`;
+}
+
+function getLastModified(doc) {
+  const rawDate = doc?.updatedAt || doc?.createdAt;
+  if (!rawDate) {
+    return null;
+  }
+
+  const parsedDate = new Date(rawDate);
+  if (Number.isNaN(parsedDate.getTime())) {
+    return null;
+  }
+
+  return parsedDate.toISOString();
+}
+
+function toSitemapEntry(doc, pathname) {
+  const url = getAbsoluteUrl(pathname);
+  if (!url) {
+    return null;
+  }
+
+  return {
+    url,
+    lastModified: getLastModified(doc),
+  };
+}
+
+function getPagePathname(doc) {
+  if (!doc) {
+    return null;
+  }
+
+  const pathname = formatPagePath("pages", doc);
+  return normalizePathname(pathname);
+}
+
+async function getPagesEntries(api) {
+  const { docs } = await api.getCollection("pages", {
+    pagination: false,
+    select: {
+      slug: true,
+      parent: true,
+      breadcrumbs: true,
+      updatedAt: true,
+      createdAt: true,
+    },
+    where: {
+      slug: {
+        not_in: ["404", "500"],
+      },
+    },
+  });
+
+  return docs
+    .map((doc) => toSitemapEntry(doc, getPagePathname(doc)))
+    .filter(Boolean);
+}
+
+function dedupeEntries(entries) {
+  const seen = new Set();
+
+  return entries.filter((entry) => {
+    if (seen.has(entry.url)) {
+      return false;
+    }
+
+    seen.add(entry.url);
+    return true;
+  });
+}
+
+function escapeXml(value) {
+  return String(value)
+    .replaceAll("&", "&")
+    .replaceAll("<", "&lt;")
+    .replaceAll(">", "&gt;")
+    .replaceAll('"', "&quot;")
+    .replaceAll("'", "&apos;");
+}
+
+async function getSitemapEntries(api) {
+  const pages = await getPagesEntries(api);
+  return dedupeEntries(pages).sort((left, right) =>
+    left.url.localeCompare(right.url),
+  );
+}
+
+async function buildSitemapXml(api) {
+  const entries = await getSitemapEntries(api);
+  const xmlEntries = entries
+    .map(({ url, lastModified }) => {
+      const lastModifiedNode = lastModified
+        ? `\n    <lastmod>${lastModified}</lastmod>`
+        : "";
+
+      return `  <url>\n    <loc>${escapeXml(url)}</loc>${lastModifiedNode}\n  </url>`;
+    })
+    .join("\n");
+
+  return `<?xml version="1.0" encoding="UTF-8"?>\n<urlset xmlns="http://www.sitemaps.org/schemas/sitemap/0.9">\n${xmlEntries}\n</urlset>\n`;
+}
+
+export default buildSitemapXml;

apps/trustlab/src/lib/data/index.js

@@ -1,3 +1,8 @@
-export { getPageStaticPaths, getPageStaticProps, getRobotsTxt } from "./local";
+export {
+  getPageStaticPaths,
+  getPageStaticProps,
+  getRobotsTxt,
+  getSitemapXml,
+} from "./local";
 
 export default undefined;

apps/trustlab/src/lib/data/local/index.js

@@ -1,4 +1,5 @@
 import { getPageProps, getPagePaths } from "@/trustlab/lib/data/common";
+import buildSitemapXml from "@/trustlab/lib/data/common/sitemap";
 import api from "@/trustlab/lib/payload";
 
 export async function getPageStaticPaths() {
@@ -21,4 +22,8 @@ export async function getRobotsTxt() {
   return siteSettings?.robotsTxt;
 }
 
+export async function getSitemapXml() {
+  return buildSitemapXml(api);
+}
+
 export default undefined;

apps/trustlab/src/pages/api/v1/sitemap.page.js

@@ -0,0 +1,24 @@
+import * as Sentry from "@sentry/nextjs";
+
+import { getSitemapXml } from "@/trustlab/lib/data";
+
+export default async function handler(req, res) {
+  if (req.method !== "GET") {
+    res.setHeader("Allow", "GET");
+    res.status(405).end();
+    return;
+  }
+
+  try {
+    const sitemapXml = await getSitemapXml();
+    res.setHeader(
+      "Cache-Control",
+      "public, max-age=3600, stale-while-revalidate=86400",
+    );
+    res.setHeader("Content-Type", "application/xml; charset=utf-8");
+    res.send(sitemapXml);
+  } catch (error) {
+    Sentry.captureException(error);
+    res.status(500).end();
+  }
+}