@adobe/spacecat-shared-data-access 2.54.0 → 2.56.0

package/CHANGELOG.md

@@ -1,3 +1,17 @@
+# [@adobe/spacecat-shared-data-access-v2.56.0](https://github.com/adobe/spacecat-shared/compare/@adobe/spacecat-shared-data-access-v2.55.0...@adobe/spacecat-shared-data-access-v2.56.0) (2025-08-20)
+
+
+### Features
+
+* add cdn filters in llmo config ([#925](https://github.com/adobe/spacecat-shared/issues/925)) ([021110b](https://github.com/adobe/spacecat-shared/commit/021110bbf7da221d41a808af54ed43f5e5f513ea))
+
+# [@adobe/spacecat-shared-data-access-v2.55.0](https://github.com/adobe/spacecat-shared/compare/@adobe/spacecat-shared-data-access-v2.54.0...@adobe/spacecat-shared-data-access-v2.55.0) (2025-08-20)
+
+
+### Features
+
+* add scrape-client destination  ([#913](https://github.com/adobe/spacecat-shared/issues/913)) ([e208a87](https://github.com/adobe/spacecat-shared/commit/e208a87214874a2708ac2d7614fcfd4c0770fe17))
+
 # [@adobe/spacecat-shared-data-access-v2.54.0](https://github.com/adobe/spacecat-shared/compare/@adobe/spacecat-shared-data-access-v2.53.0...@adobe/spacecat-shared-data-access-v2.54.0) (2025-08-13)
 
 

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@adobe/spacecat-shared-data-access",
-  "version": "2.54.0",
+  "version": "2.56.0",
   "description": "Shared modules of the Spacecat Services - Data Access",
   "type": "module",
   "engines": {

package/src/models/audit/audit.model.js

@@ -73,6 +73,7 @@ class Audit extends BaseModel {
   static AUDIT_STEP_DESTINATIONS = {
     CONTENT_SCRAPER: 'content-scraper',
     IMPORT_WORKER: 'import-worker',
+    SCRAPE_CLIENT: 'scrape-client',
   };
 
   /**
@@ -86,7 +87,9 @@ class Audit extends BaseModel {
    *   [Audit.AUDIT_STEP_DESTINATIONS.IMPORT_WORKER]: {
    *     getQueueUrl: function,
    *     formatPayload: function
-   *   }
+   *   },
+   *   [Audit.AUDIT_STEP_DESTINATIONS.SCRAPE_CLIENT]: {
+   *   formatPayload: function
    * }}
    */
   static AUDIT_STEP_DESTINATION_CONFIGS = {
@@ -153,6 +156,26 @@ class Audit extends BaseModel {
         auditContext,
       }),
     },
+    [Audit.AUDIT_STEP_DESTINATIONS.SCRAPE_CLIENT]: {
+      /**
+       *
+       * @param stepResult - The result of the audit step.
+       * @param auditContext - The audit context.
+       * @param context - The context object.
+       * @returns {object} - The formatted payload for the scrape client.
+       */
+      formatPayload: (stepResult, auditContext, context) => ({
+        urls: stepResult.urls.map((urlObj) => urlObj.url),
+        processingType: stepResult.processingType || 'default',
+        options: stepResult.options || {},
+        maxScrapeAge: stepResult.maxScrapeAge || 24,
+        auditData: {
+          siteId: stepResult.siteId,
+          completionQueueUrl: stepResult.completionQueueUrl || context.env?.AUDIT_JOBS_QUEUE_URL,
+          auditContext,
+        },
+      }),
+    },
   };
 
   /**

package/src/models/site/config.js

@@ -269,6 +269,13 @@ export const configSchema = Joi.object({
       }),
     ).optional(),
     tags: Joi.array().items(Joi.string()).optional(),
+    cdnlogsFilter: Joi.array().items(
+      Joi.object({
+        key: Joi.string().required(),
+        value: Joi.array().items(Joi.string()).required(),
+        type: Joi.string().valid('include', 'exclude').optional(),
+      }),
+    ).optional(),
   }).optional(),
   cdnLogsConfig: Joi.object({
     bucketName: Joi.string().required(),
@@ -384,6 +391,7 @@ export const Config = (data = {}) => {
     const llmoConfig = self.getLlmoConfig();
     return llmoConfig?.customerIntent || [];
   };
+  self.getLlmoCdnlogsFilter = () => state?.llmo?.cdnlogsFilter;
 
   self.updateSlackConfig = (channel, workspace, invitedUserCount) => {
     state.slack = {
@@ -525,6 +533,11 @@ export const Config = (data = {}) => {
     );
   };
 
+  self.updateLlmoCdnlogsFilter = (cdnlogsFilter) => {
+    state.llmo = state.llmo || {};
+    state.llmo.cdnlogsFilter = cdnlogsFilter;
+  };
+
   self.updateImports = (imports) => {
     state.imports = imports;
   };