@adobe/helix-html-pipeline 5.11.12 → 5.12.0

package/CHANGELOG.md

@@ -1,3 +1,17 @@
+# [5.12.0](https://github.com/adobe/helix-html-pipeline/compare/v5.11.13...v5.12.0) (2024-07-31)
+
+
+### Features
+
+* include json-ld from metadata ([#657](https://github.com/adobe/helix-html-pipeline/issues/657)) ([56c7d37](https://github.com/adobe/helix-html-pipeline/commit/56c7d376014583e8d6664886902ebc2400874bb2)), closes [#648](https://github.com/adobe/helix-html-pipeline/issues/648)
+
+## [5.11.13](https://github.com/adobe/helix-html-pipeline/compare/v5.11.12...v5.11.13) (2024-07-15)
+
+
+### Bug Fixes
+
+* .aem.* URLs should both be rewritten to keep existing content compatible ([b6b5b04](https://github.com/adobe/helix-html-pipeline/commit/b6b5b04d7d461471173bf33ac801e3b0f54e1a3c)), closes [#498](https://github.com/adobe/helix-html-pipeline/issues/498)
+
 ## [5.11.12](https://github.com/adobe/helix-html-pipeline/compare/v5.11.11...v5.11.12) (2024-07-07)
 
 

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@adobe/helix-html-pipeline",
-  "version": "5.11.12",
+  "version": "5.12.0",
   "description": "Helix HTML Pipeline",
   "main": "src/index.js",
   "types": "src/index.d.ts",
@@ -17,7 +17,7 @@
     "lint": "eslint .",
     "semantic-release": "semantic-release",
     "semantic-release-dry": "semantic-release --dry-run",
-    "prepare": "husky install"
+    "prepare": "husky"
   },
   "repository": {
     "type": "git",
@@ -82,13 +82,13 @@
     "eslint-import-resolver-exports": "1.0.0-beta.5",
     "eslint-plugin-header": "3.1.1",
     "eslint-plugin-import": "2.29.1",
-    "esmock": "2.6.6",
-    "husky": "9.0.11",
+    "esmock": "2.6.7",
+    "husky": "9.1.3",
     "js-yaml": "4.1.0",
-    "jsdom": "24.1.0",
-    "junit-report-builder": "3.2.1",
+    "jsdom": "24.1.1",
+    "junit-report-builder": "4.0.0",
     "lint-staged": "15.2.7",
-    "mocha": "10.6.0",
+    "mocha": "10.7.0",
     "mocha-multi-reporters": "1.5.1",
     "mocha-suppress-logs": "0.5.1",
     "semantic-release": "24.0.0"

package/src/steps/render.js

@@ -14,6 +14,7 @@
 import { h } from 'hastscript';
 import { unified } from 'unified';
 import rehypeParse from 'rehype-parse';
+import { cleanupHeaderValue } from '@adobe/helix-shared-utils';
 
 function appendElement($parent, $el) {
   if ($el) {
@@ -34,6 +35,13 @@ function createElement(name, ...attrs) {
   return h(name, properties);
 }
 
+function sanitizeJsonLd(jsonLd) {
+  if (jsonLd.toLowerCase().indexOf('</script>') >= 0) {
+    throw new Error('script tag not allowed');
+  }
+  return JSON.stringify(JSON.parse(jsonLd.trim()));
+}
+
 /**
  * @type PipelineStep
  * @param {PipelineState} state
@@ -59,7 +67,13 @@ export default async function render(state, req, res) {
     appendElement($head, createElement('link', 'rel', 'canonical', 'href', meta.canonical));
   }
 
+  let jsonLd;
   for (const [name, value] of Object.entries(meta.page)) {
+    if (name.toLowerCase() === 'json-ld') {
+      jsonLd = value;
+      // eslint-disable-next-line no-continue
+      continue;
+    }
     const attr = name.includes(':') && !name.startsWith('twitter:') ? 'property' : 'name';
     if (Array.isArray(value)) {
       for (const v of value) {
@@ -71,6 +85,19 @@ export default async function render(state, req, res) {
   }
   appendElement($head, createElement('link', 'rel', 'alternate', 'type', 'application/xml+atom', 'href', meta.feed, 'title', `${meta.title} feed`));
 
+  // inject json ld if valid
+  if (jsonLd) {
+    const props = { type: 'application/ld+json' };
+    try {
+      jsonLd = sanitizeJsonLd(jsonLd);
+    } catch (e) {
+      jsonLd = '';
+      props['data-error'] = `error in json-ld: ${cleanupHeaderValue(e.message)}`;
+    }
+    const script = h('script', props, jsonLd);
+    $head.children.push(script);
+  }
+
   // inject head.html
   const headHtml = state.helixConfig?.head?.data.html;
   if (headHtml) {

package/src/steps/utils.js

@@ -12,9 +12,9 @@
 
 const AZURE_BLOB_REGEXP = /^https:\/\/hlx\.blob\.core\.windows\.net\/external\//;
 
-const MEDIA_BLOB_REGEXP = /^https:\/\/.*\.hlx3?\.(live|page)\/media_.*/;
+const MEDIA_BLOB_REGEXP = /^https:\/\/.*\.(aem|hlx3?)\.(live|page)\/media_.*/;
 
-const HELIX_URL_REGEXP = /^https:\/\/(?!admin\.|www\.)[^.]+\.hlx3?\.(live|page)\/?.*/;
+const HELIX_URL_REGEXP = /^https:\/\/(?!admin\.|www\.)[^.]+\.(aem|hlx3?)\.(live|page)\/?.*/;
 
 /**
  * Returns the original host name from the request to the outer CDN.