feat: merge main into migration branch

Author: alinarublea

.gitignore

@@ -10,3 +10,6 @@ test-results.xml
 .env
 .idea/
 .vscode/
+.aws-sam/
+env.sh
+output.txt

.nycrc.json

@@ -10,5 +10,8 @@
   "all": true,
   "include": [
     "src/**/*.js"
+  ],
+  "exclude": [
+    "src/index-local.js"
   ]
 }

CHANGELOG.md

@@ -545,4 +545,21 @@ Here's how messages flow between workers in a step-based audit:
 }
 ```
 
-Each message preserves the `auditContext` to maintain the step chain. The `next` field determines which step runs next, while `auditId` and `fullAuditRef` track the audit state across workers.
+Each message preserves the `auditContext` to maintain the step chain. The `next` field determines which step runs next, while `auditId` and `fullAuditRef` track the audit state across workers.
+
+## Run Audits locally
+
+You can run the audit locally using AWS SAM and Docker.
+
+1. Ensure you have [Docker](https://docs.docker.com/desktop/setup/install/mac-install/), [AWS SAM](https://docs.aws.amazon.com/serverless-application-model/latest/developerguide/install-sam-cli.html) and [jq](https://jqlang.org/) installed.
+2. Login to AWS using KLAM and login with your AWS CLI.
+3. To provide secrets to the audit, please run `./populate-env.sh` once. It will fetch all secrets from the AWS Secret Manager.
+4. To run the audit locally, execute the following commands:
+    ```bash
+    source env.sh
+    npm run local-build
+    npm run local-run
+    ```
+5. Starting point of the execution is `src/index-local.js`. Output of the audit can be found in `output.txt`.
+
+If you need to add additional secrets, make sure to adjust the Lambda `template.yml` accordingly.

README.md

@@ -1,6 +1,6 @@
 {
   "name": "@adobe/spacecat-audit-worker",
-  "version": "1.55.0",
+  "version": "1.68.16",
   "description": "SpaceCat Audit Worker",
   "main": "src/index.js",
   "type": "module",
@@ -20,7 +20,9 @@
     "deploy": "hedy -v --deploy --aws-deploy-bucket=spacecat-prod-deploy --pkgVersion=latest",
     "deploy-ci": "hedy -v --deploy --pkgVersion=ci$CI_BUILD_NUM -l latest --aws-deploy-bucket=spacecat-dev-deploy --cleanup-ci=24h",
     "deploy-secrets": "hedy --aws-update-secrets --params-file=secrets/secrets.env",
-    "prepare": "husky"
+    "prepare": "husky",
+    "local-build": "sam build",
+    "local-run": "sam local invoke -l output.txt"
   },
   "wsk": {
     "target": "aws",
@@ -37,7 +39,9 @@
     "nodeVersion": 22,
     "static": [
       "static/prompts/broken-backlinks.prompt",
-      "static/prompts/broken-backlinks-followup.prompt"
+      "static/prompts/broken-backlinks-followup.prompt",
+      "static/prompts/structured-data-suggest.prompt",
+      "static/prompts/image-alt-text.prompt"
     ]
   },
   "repository": {
@@ -57,43 +61,44 @@
     "reporter-options": "configFile=.mocha-multi.json"
   },
   "dependencies": {
-    "@adobe/fetch": "4.1.11",
+    "@adobe/fetch": "4.2.0",
     "@adobe/helix-shared-secrets": "2.2.10",
     "@adobe/helix-shared-wrap": "2.0.2",
     "@adobe/helix-status": "10.1.5",
-    "@adobe/helix-universal": "5.0.8",
-    "@adobe/helix-universal-logger": "3.0.23",
-    "@adobe/spacecat-shared-ahrefs-client": "1.6.7",
-    "@adobe/spacecat-shared-data-access": "2.7.1",
-    "@adobe/spacecat-shared-google-client": "1.4.10",
-    "@adobe/spacecat-shared-gpt-client": "1.5.0",
-    "@adobe/spacecat-shared-http-utils": "1.9.8",
-    "@adobe/spacecat-shared-rum-api-client": "2.21.1",
+    "@adobe/helix-universal": "5.0.9",
+    "@adobe/helix-universal-logger": "3.0.24",
+    "@adobe/spacecat-shared-ahrefs-client": "1.6.12",
+    "@adobe/spacecat-shared-data-access": "2.13.1",
+    "@adobe/spacecat-shared-google-client": "1.4.15",
+    "@adobe/spacecat-shared-gpt-client": "1.5.4",
+    "@adobe/spacecat-shared-http-utils": "1.9.12",
+    "@adobe/spacecat-shared-rum-api-client": "2.22.0",
     "@adobe/spacecat-shared-rum-api-client-v1": "npm:@adobe/[email protected]",
-    "@adobe/spacecat-shared-utils": "1.32.0",
-    "@aws-sdk/client-s3": "3.749.0",
-    "@aws-sdk/client-lambda": "3.749.0",
-    "@aws-sdk/credential-provider-node": "3.749.0",
-    "@aws-sdk/client-sqs": "3.749.0",
+    "@adobe/spacecat-shared-utils": "1.35.0",
+    "@aws-sdk/client-s3": "3.758.0",
+    "@aws-sdk/client-lambda": "3.758.0",
+    "@aws-sdk/credential-provider-node": "3.758.0",
+    "@aws-sdk/client-sqs": "3.758.0",
     "aws-xray-sdk": "3.10.3",
-    "@aws-sdk/s3-request-presigner": "3.749.0",
+    "@aws-sdk/s3-request-presigner": "3.758.0",
+    "cheerio": "1.0.0",
     "diff": "7.0.0",
     "jsdom": "26.0.0",
     "urijs": "1.19.11"
   },
   "devDependencies": {
     "@adobe/eslint-config-helix": "2.0.9",
     "@adobe/helix-deploy": "https://gitpkg.now.sh/alinarublea/helix-deploy?test-deploy",
-    "@adobe/helix-universal": "5.0.8",
-    "@adobe/semantic-release-coralogix": "1.1.34",
+    "@adobe/helix-universal": "5.0.9",
+    "@adobe/semantic-release-coralogix": "1.1.35",
     "@adobe/semantic-release-skms-cmr": "1.1.5",
-    "@redocly/cli": "1.29.0",
+    "@redocly/cli": "1.33.0",
     "@semantic-release/changelog": "6.0.3",
     "@semantic-release/exec": "7.0.3",
     "@semantic-release/git": "10.0.1",
     "@semantic-release/npm": "12.0.1",
-    "@typescript-eslint/eslint-plugin": "8.24.0",
-    "@typescript-eslint/parser": "8.24.0",
+    "@typescript-eslint/eslint-plugin": "8.26.0",
+    "@typescript-eslint/parser": "8.26.0",
     "c8": "10.1.3",
     "chai": "5.2.0",
     "chai-as-promised": "8.0.1",
@@ -111,7 +116,7 @@
     "semantic-release": "24.2.3",
     "sinon": "19.0.2",
     "sinon-chai": "4.0.0",
-    "typescript": "5.7.3",
+    "typescript": "5.8.2",
     "yaml": "2.7.0"
   },
   "lint-staged": {

package-lock.json

@@ -0,0 +1,15 @@
+#!/bin/bash
+
+# Define the secret name and region
+SECRET_NAME="/helix-deploy/spacecat-services/audit-worker/ci"
+REGION="us-east-1"
+
+# Retrieve the secret from AWS Secrets Manager
+SECRET_JSON=$(aws secretsmanager get-secret-value --secret-id $SECRET_NAME --region $REGION --query SecretString --output text)
+
+touch env.sh
+
+# Parse the JSON and write to env.sh
+echo "$SECRET_JSON" | jq -r 'to_entries | .[] | "export \(.key)=\(.value | @sh)"' >> env.sh
+
+echo "env.sh file has been populated with secrets from AWS Secrets Manager."

package.json

@@ -1,5 +1,6 @@
 {
   "$schema": "https://docs.renovatebot.com/renovate-schema.json",
+  "enabled": false,
   "extends": [
     "local>adobe-rnd/spacecat-audit-worker"
   ],