fix: updating content ai

src/headings/shared-utils.js

@@ -16,6 +16,10 @@ import { load as cheerioLoad } from 'cheerio';
 import SeoChecks from '../metatags/seo-checks.js';
 import { getTopPagesForSiteId } from '../utils/data-access.js';
 import { getObjectKeysUsingPrefix, getObjectFromKey } from '../utils/s3-utils.js';
+import {
+  extractBrandGuidelinesFromProfile,
+  formatBrandGuidelinesToMarkdown,
+} from '../utils/brand-profile.js';
 import {
   getHeadingLevel,
   getHeadingContext,
@@ -141,50 +145,14 @@ export async function loadScrapeJson(url, site, allKeys, s3Client, S3_SCRAPER_BU
   return getObjectFromKey(s3Client, S3_SCRAPER_BUCKET_NAME, s3Key, log);
 }
 
-/**
- * Extract brand guidelines from brand profile
- * @param {Object} brandProfile - Brand profile from site config
- * @returns {Object} Formatted brand guidelines
- */
-function extractBrandGuidelinesFromProfile(brandProfile) {
-  const mainProfile = brandProfile.main_profile || {};
-  // Extract brand persona (short description)
-  const brandPersona = mainProfile.brand_personality?.description || '';
-
-  // Extract tone
-  const toneAttributes = mainProfile.tone_attributes || {};
-  const primaryTones = toneAttributes.primary || [];
-  const tone = primaryTones.join(', ');
-
-  // Extract editorial guidelines
-  const editorialGuidelines = mainProfile.editorial_guidelines || {};
-  const dos = editorialGuidelines.dos || [];
-  const donts = editorialGuidelines.donts || [];
-
-  // Extract forbidden items
-  const languagePatterns = mainProfile.language_patterns || {};
-  const avoidPatterns = languagePatterns.avoid || [];
-  const avoidTones = toneAttributes.avoid || [];
-  const forbidden = [...avoidPatterns, ...avoidTones];
-
-  return {
-    brand_persona: brandPersona,
-    tone,
-    editorial_guidelines: {
-      do: dos,
-      dont: donts,
-    },
-    forbidden,
-  };
-}
-
 /**
  * Get brand guidelines from site config or generate from healthy tags using AI
+ * Returns formatted markdown string suitable for AI prompts
  * @param {Object} healthyTagsObject - Object with healthy title, description, h1
  * @param {Object} log - Logger instance
  * @param {Object} context - Audit context
  * @param {Object} site - Site object (optional, for accessing brand profile)
- * @returns {Promise<Object>} Brand guidelines
+ * @returns {Promise<string>} Brand guidelines formatted as markdown string
  */
 export async function getBrandGuidelines(healthyTagsObject, log, context, site = null) {
   // First, try to get brand profile from site config
@@ -195,8 +163,9 @@ export async function getBrandGuidelines(healthyTagsObject, log, context, site =
       if (brandProfile && typeof brandProfile === 'object' && Object.keys(brandProfile).length > 0) {
         log.info('[Brand Guidelines] Using brand profile from site config');
         const guidelines = extractBrandGuidelinesFromProfile(brandProfile);
-        log.debug(`[Brand Guidelines] Extracted guidelines: ${JSON.stringify(guidelines)}`);
-        return guidelines;
+        const formattedGuidelines = formatBrandGuidelinesToMarkdown(guidelines);
+        log.debug(`[Brand Guidelines] Extracted guidelines: ${formattedGuidelines}`);
+        return formattedGuidelines;
       }
     } catch (error) {
       log.warn(`[Brand Guidelines] Error accessing brand profile from site config: ${error.message}`);

src/utils/brand-profile.js

@@ -0,0 +1,185 @@
+/*
+ * Copyright 2025 Adobe. All rights reserved.
+ * This file is licensed to you under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License. You may obtain a copy
+ * of the License at http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software distributed under
+ * the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR REPRESENTATIONS
+ * OF ANY KIND, either express or implied. See the License for the specific language
+ * governing permissions and limitations under the License.
+ */
+
+/**
+ * Brand Profile Utilities
+ *
+ * Shared functions for extracting and formatting brand profile data.
+ * Matches the extraction used in Mystique BrandProfileTool.py for consistency
+ * across Readability, Summarization, Headings, and Content AI.
+ */
+
+/**
+ * Extract brand guidelines from brand profile
+ * Matches the extraction used in Mystique BrandProfileTool.py for Readability/Summarization
+ * @param {Object} brandProfile - Brand profile from site config
+ * @returns {Object} Formatted brand guidelines
+ */
+export function extractBrandGuidelinesFromProfile(brandProfile) {
+  const mainProfile = brandProfile.main_profile || {};
+
+  // Extract tone attributes (primary and avoid)
+  const toneAttributes = mainProfile.tone_attributes || {};
+  const tonePrimary = toneAttributes.primary || [];
+  const toneAvoid = toneAttributes.avoid || [];
+
+  // Extract vocabulary - signature phrases (top 5)
+  const vocabulary = mainProfile.vocabulary || {};
+  const signaturePhrases = (vocabulary.signature_phrases || []).slice(0, 5);
+
+  // Extract brand values - core values (top 5)
+  const brandValues = mainProfile.brand_values || {};
+  const coreValues = (brandValues.core_values || []).slice(0, 5).map((v) => ({
+    name: v.name || '',
+    evidence: v.evidence || '',
+  }));
+
+  // Extract language patterns (preferred and avoid) (top 5 each)
+  const languagePatterns = mainProfile.language_patterns || {};
+  const languagePreferred = (languagePatterns.preferred || []).slice(0, 5);
+  const languageAvoid = (languagePatterns.avoid || []).slice(0, 5);
+
+  // Extract communication style
+  const communicationStyle = mainProfile.communication_style || '';
+
+  // Extract editorial guidelines (dos and donts) (top 5 each)
+  const editorialGuidelines = mainProfile.editorial_guidelines || {};
+  const editorialDos = (editorialGuidelines.dos || []).slice(0, 5);
+  const editorialDonts = (editorialGuidelines.donts || []).slice(0, 5);
+
+  return {
+    tone_attributes: {
+      primary: tonePrimary,
+      avoid: toneAvoid,
+    },
+    signature_phrases: signaturePhrases,
+    brand_values: coreValues,
+    language_patterns: {
+      preferred: languagePreferred,
+      avoid: languageAvoid,
+    },
+    communication_style: communicationStyle,
+    editorial_guidelines: {
+      dos: editorialDos,
+      donts: editorialDonts,
+    },
+  };
+}
+
+/**
+ * Format brand guidelines to markdown string for AI prompts
+ * Matches the format used in Mystique BrandProfileTool.py
+ * @param {Object} guidelines - Extracted brand guidelines object
+ * @returns {string} Formatted markdown string
+ */
+export function formatBrandGuidelinesToMarkdown(guidelines) {
+  const parts = ['## Brand Guidelines (from Brand Profile)'];
+
+  // Tone attributes
+  if (guidelines.tone_attributes) {
+    parts.push('\n### TONE ATTRIBUTES');
+    if (guidelines.tone_attributes.primary?.length > 0) {
+      parts.push(`  ✓ MUST USE: ${guidelines.tone_attributes.primary.join(', ')}`);
+    }
+    if (guidelines.tone_attributes.avoid?.length > 0) {
+      parts.push(`  ✗ MUST AVOID: ${guidelines.tone_attributes.avoid.join(', ')}`);
+    }
+  }
+
+  // Signature phrases
+  if (guidelines.signature_phrases?.length > 0) {
+    parts.push('\n### SIGNATURE PHRASES');
+    parts.push('  ✓ USE these phrases when relevant:');
+    guidelines.signature_phrases.forEach((phrase) => {
+      parts.push(`    • "${phrase}"`);
+    });
+  }
+
+  // Brand values
+  if (guidelines.brand_values?.length > 0) {
+    parts.push('\n### BRAND VALUES');
+    guidelines.brand_values.forEach((value) => {
+      parts.push(`    • ${value.name}: ${value.evidence}`);
+    });
+  }
+
+  // Language patterns
+  if (guidelines.language_patterns) {
+    parts.push('\n### LANGUAGE PATTERNS');
+    if (guidelines.language_patterns.preferred?.length > 0) {
+      parts.push('  ✓ Preferred:');
+      guidelines.language_patterns.preferred.forEach((item) => {
+        parts.push(`    • ${item}`);
+      });
+    }
+    if (guidelines.language_patterns.avoid?.length > 0) {
+      parts.push('  ✗ Avoid:');
+      guidelines.language_patterns.avoid.forEach((item) => {
+        parts.push(`    • ${item}`);
+      });
+    }
+  }
+
+  // Communication style
+  if (guidelines.communication_style) {
+    parts.push('\n### COMMUNICATION STYLE');
+    parts.push(`  ${guidelines.communication_style}`);
+  }
+
+  // Editorial guidelines
+  if (guidelines.editorial_guidelines) {
+    parts.push('\n### EDITORIAL GUIDELINES');
+    if (guidelines.editorial_guidelines.dos?.length > 0) {
+      parts.push('  ✓ DO:');
+      guidelines.editorial_guidelines.dos.forEach((item) => {
+        parts.push(`    • ${item}`);
+      });
+    }
+    if (guidelines.editorial_guidelines.donts?.length > 0) {
+      parts.push("  ✗ DON'T:");
+      guidelines.editorial_guidelines.donts.forEach((item) => {
+        parts.push(`    • ${item}`);
+      });
+    }
+  }
+
+  return parts.join('\n');
+}
+
+/**
+ * Get formatted brand guidelines from a site's brand profile
+ * @param {Object} site - Site object with getConfig method
+ * @param {Object} log - Logger instance (optional)
+ * @returns {string} Formatted brand guidelines markdown string, or empty string if not available
+ */
+export function getBrandGuidelinesFromSite(site, log = null) {
+  if (!site) {
+    return '';
+  }
+
+  try {
+    const config = site.getConfig();
+    const brandProfile = config?.getBrandProfile?.();
+
+    if (brandProfile && typeof brandProfile === 'object' && Object.keys(brandProfile).length > 0) {
+      log?.info('[Brand Guidelines] Using brand profile from site config');
+      const guidelines = extractBrandGuidelinesFromProfile(brandProfile);
+      const formattedGuidelines = formatBrandGuidelinesToMarkdown(guidelines);
+      log?.debug(`[Brand Guidelines] Extracted guidelines: ${formattedGuidelines}`);
+      return formattedGuidelines;
+    }
+  } catch (error) {
+    log?.error(`[Brand Guidelines] Error accessing brand profile from site config: ${error.message}`);
+  }
+
+  return '';
+}

src/utils/content-ai.js

@@ -11,6 +11,7 @@
  */
 
 import { ImsClient } from '@adobe/spacecat-shared-ims-client';
+import { getBrandGuidelinesFromSite } from './brand-profile.js';
 
 /**
  * Calculates a weekly cron schedule set to run one hour from now.
@@ -208,6 +209,25 @@ export class ContentAIClient {
 
     this.log?.info(`Creating ContentAI configuration for site ${baseURL} with cron schedule ${cronSchedule} and name ${name}`);
 
+    // Get brand guidelines from site config
+    const brandGuidelines = getBrandGuidelinesFromSite(site, this.log);
+    const hasBrandGuidelines = brandGuidelines && brandGuidelines.length > 0;
+
+    if (hasBrandGuidelines) {
+      this.log?.info(`[ContentAI] Brand guidelines found for site ${baseURL}`);
+    } else {
+      this.log?.info(`[ContentAI] No brand guidelines found for site ${baseURL}`);
+    }
+
+    // Build system prompt with optional brand guidelines
+    let systemPrompt = 'You are a helpful AI Assistant powering the search experience.\nYou will answer questions using the provided context.\n';
+
+    if (hasBrandGuidelines) {
+      systemPrompt += `\n**Brand Guidelines**:\n${brandGuidelines}\n\nWhen generating responses, follow the brand guidelines to ensure answers match the brand's tone, vocabulary, and editorial standards.\n`;
+    }
+
+    systemPrompt += '\nContext: {context}\n';
+
     const contentAiData = {
       steps: [
         {
@@ -232,7 +252,7 @@ export class ContentAIClient {
           name: 'Comprehensive Q&A assitant',
           description: 'AI assistant for answering any user question about a topic in the indexed knowledge',
           prompts: {
-            system: 'You are a helpful AI Assistant powering the search experience.\nYou will answer questions using the provided context.\nContext: {context}\n',
+            system: systemPrompt,
             user: 'Please answer the following question: {question}\n',
           },
         },