@adobe/spacecat-shared-utils 1.81.0 → 1.82.0

package/CHANGELOG.md

@@ -1,3 +1,17 @@
+# [@adobe/spacecat-shared-utils-v1.82.0](https://github.com/adobe/spacecat-shared/compare/@adobe/spacecat-shared-utils-v1.81.1...@adobe/spacecat-shared-utils-v1.82.0) (2025-12-03)
+
+
+### Features
+
+* move www url resolver in audit worker to shared utils ([#1207](https://github.com/adobe/spacecat-shared/issues/1207)) ([40dbf3a](https://github.com/adobe/spacecat-shared/commit/40dbf3a18681cd304c168602d74b51c0baf8ead6))
+
+# [@adobe/spacecat-shared-utils-v1.81.1](https://github.com/adobe/spacecat-shared/compare/@adobe/spacecat-shared-utils-v1.81.0...@adobe/spacecat-shared-utils-v1.81.1) (2025-11-28)
+
+
+### Bug Fixes
+
+* update to node 24 ([#1179](https://github.com/adobe/spacecat-shared/issues/1179)) ([0e60c0a](https://github.com/adobe/spacecat-shared/commit/0e60c0ab791b47662d07822f7c93009a8f7048fd))
+
 # [@adobe/spacecat-shared-utils-v1.81.0](https://github.com/adobe/spacecat-shared/compare/@adobe/spacecat-shared-utils-v1.80.0...@adobe/spacecat-shared-utils-v1.81.0) (2025-11-26)
 
 

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@adobe/spacecat-shared-utils",
-  "version": "1.81.0",
+  "version": "1.82.0",
   "description": "Shared modules of the Spacecat Services - utils",
   "type": "module",
   "exports": {
@@ -10,7 +10,7 @@
     }
   },
   "engines": {
-    "node": ">=22.0.0 <23.0.0",
+    "node": ">=22.0.0 <25.0.0",
     "npm": ">=10.9.0 <12.0.0"
   },
   "main": "src/index.js",
@@ -53,14 +53,15 @@
   },
   "dependencies": {
     "@adobe/fetch": "4.2.3",
-    "@aws-sdk/client-s3": "3.937.0",
-    "@aws-sdk/client-sqs": "3.936.0",
+    "@aws-sdk/client-s3": "3.940.0",
+    "@aws-sdk/client-sqs": "3.940.0",
     "@json2csv/plainjs": "7.0.6",
     "aws-xray-sdk": "3.12.0",
     "cheerio": "1.1.2",
     "date-fns": "4.1.0",
     "franc-min": "6.2.0",
     "iso-639-3": "3.0.1",
+    "urijs": "1.19.11",
     "validator": "^13.15.15",
     "world-countries": "5.1.0",
     "zod": "^4.1.11"

package/src/index.js

@@ -66,6 +66,9 @@ export {
   getSpacecatRequestHeaders,
   ensureHttps,
   urlMatchesFilter,
+  hasNonWWWSubdomain,
+  toggleWWWHostname,
+  wwwUrlResolver,
 } from './url-helpers.js';
 
 export {

package/src/url-helpers.js

@@ -11,6 +11,8 @@
  */
 
 import { context as h2, h1 } from '@adobe/fetch';
+import URI from 'urijs';
+import { hasText, isValidUrl } from './functions.js';
 import { SPACECAT_USER_AGENT } from './tracing-fetch.js';
 
 /* c8 ignore next 3 */
@@ -240,6 +242,77 @@ function urlMatchesFilter(url, filterUrls) {
   }
 }
 
+/**
+ * Checks if a URL has a subdomain other than 'www'.
+ * @param {string} baseUrl - The URL to check.
+ * @returns {boolean} - True if the URL has a non-www subdomain, false otherwise.
+ * @throws {Error} - If the baseURL cannot be parsed.
+ */
+function hasNonWWWSubdomain(baseUrl) {
+  try {
+    const uri = new URI(baseUrl);
+    return hasText(uri.domain()) && hasText(uri.subdomain()) && uri.subdomain() !== 'www';
+  } catch {
+    throw new Error(`Cannot parse baseURL: ${baseUrl}`);
+  }
+}
+
+/**
+ * Toggles the www subdomain in a given hostname.
+ * @param {string} hostname - The hostname to toggle the www subdomain in.
+ * @returns {string} - The hostname with the www subdomain toggled.
+ */
+function toggleWWWHostname(hostname) {
+  if (hasNonWWWSubdomain(`https://${hostname}`)) return hostname;
+  return hostname.startsWith('www.') ? hostname.replace('www.', '') : `www.${hostname}`;
+}
+
+/**
+ * Resolves the correct URL for a site by checking RUM data availability.
+ * Tries www-toggled version first, then falls back to original.
+ * @param {object} site - The site object with getBaseURL() and getConfig() methods.
+ * @param {object} rumApiClient - The RUM API client instance with retrieveDomainkey method.
+ * @param {object} log - Logger instance with debug() and error() methods.
+ * @returns {Promise<string>} - The resolved hostname without protocol.
+ */
+async function wwwUrlResolver(site, rumApiClient, log) {
+  const overrideBaseURL = site.getConfig()?.getFetchConfig()?.overrideBaseURL;
+  if (isValidUrl(overrideBaseURL)) {
+    return overrideBaseURL.replace(/^https?:\/\//, '');
+  }
+
+  const baseURL = site.getBaseURL();
+  const uri = new URI(baseURL);
+  const hostname = uri.hostname();
+  const subdomain = uri.subdomain();
+
+  if (hasText(subdomain) && subdomain !== 'www') {
+    log.debug(`Resolved URL ${hostname} since ${baseURL} contains subdomain`);
+    return hostname;
+  }
+
+  try {
+    const wwwToggledHostname = toggleWWWHostname(hostname);
+    await rumApiClient.retrieveDomainkey(wwwToggledHostname);
+    log.debug(`Resolved URL ${wwwToggledHostname} for ${baseURL} using RUM API Client`);
+    return wwwToggledHostname;
+  } catch (e) {
+    log.error(`Could not retrieved RUM domainkey for ${hostname}: ${e.message}`);
+  }
+
+  try {
+    await rumApiClient.retrieveDomainkey(hostname);
+    log.debug(`Resolved URL ${hostname} for ${baseURL} using RUM API Client`);
+    return hostname;
+  } catch (e) {
+    log.error(`Could not retrieved RUM domainkey for ${hostname}: ${e.message}`);
+  }
+
+  const fallback = hostname.startsWith('www.') ? hostname : `www.${hostname}`;
+  log.debug(`Fallback to ${fallback} for URL resolution for ${baseURL}`);
+  return fallback;
+}
+
 export {
   ensureHttps,
   getSpacecatRequestHeaders,
@@ -252,4 +325,7 @@ export {
   stripTrailingSlash,
   stripWWW,
   urlMatchesFilter,
+  hasNonWWWSubdomain,
+  toggleWWWHostname,
+  wwwUrlResolver,
 };