@adobe/spacecat-shared-html-analyzer 1.0.3 → 1.0.4

package/CHANGELOG.md

@@ -1,3 +1,10 @@
+# [@adobe/spacecat-shared-html-analyzer-v1.0.4](https://github.com/adobe/spacecat-shared/compare/@adobe/spacecat-shared-html-analyzer-v1.0.3...@adobe/spacecat-shared-html-analyzer-v1.0.4) (2025-11-03)
+
+
+### Bug Fixes
+
+* adding wordCountBefore & wordCountAfter ([#1084](https://github.com/adobe/spacecat-shared/issues/1084)) ([d3ae176](https://github.com/adobe/spacecat-shared/commit/d3ae1769e9a8548da3396027897e55d479afec2f))
+
 # [@adobe/spacecat-shared-html-analyzer-v1.0.3](https://github.com/adobe/spacecat-shared/compare/@adobe/spacecat-shared-html-analyzer-v1.0.2...@adobe/spacecat-shared-html-analyzer-v1.0.3) (2025-10-29)
 
 

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@adobe/spacecat-shared-html-analyzer",
-  "version": "1.0.3",
+  "version": "1.0.4",
   "description": "Analyze HTML content visibility for AI crawlers and citations - compare static HTML vs fully rendered content",
   "type": "module",
   "engines": {

package/src/analyzer.js

@@ -74,7 +74,9 @@ export async function calculateStats(originalHTML, currentHTML, ignoreNavFooter
   // Calculate word counts using consistent tokenization
   const originalTokens = tokenize(originalText, 'word');
   const currentTokens = tokenize(currentText, 'word');
-  const wordDiff = Math.abs(currentTokens.length - originalTokens.length);
+  const wordCountBefore = originalTokens.length;
+  const wordCountAfter = currentTokens.length;
+  const wordDiff = Math.abs(wordCountAfter - wordCountBefore);
 
   // Calculate content increase ratio (how many times content increased)
   let contentIncreaseRatio;
@@ -89,6 +91,8 @@ export async function calculateStats(originalHTML, currentHTML, ignoreNavFooter
     ? Math.min(100, (originalTokens.length / currentTokens.length) * 100) : 100;
 
   return {
+    wordCountBefore,
+    wordCountAfter,
     wordDiff,
     contentIncreaseRatio: Math.round(contentIncreaseRatio * 100) / 100, // Round to 1 decimal place
     citationReadability: Math.round(citationReadability),
@@ -109,6 +113,8 @@ export async function calculateBothScenarioStats(originalHTML, currentHTML) {
   const statsNotIgnored = await calculateStats(originalHTML, currentHTML, false);
   return {
     withNavFooterIgnored: {
+      wordCountBefore: statsIgnored.wordCountBefore,
+      wordCountAfter: statsIgnored.wordCountAfter,
       wordDiff: statsIgnored.wordDiff,
       contentIncreaseRatio: statsIgnored.contentIncreaseRatio,
       citationReadability: statsIgnored.citationReadability,
@@ -116,6 +122,8 @@ export async function calculateBothScenarioStats(originalHTML, currentHTML) {
       missingWords: statsIgnored.wordDiff,
     },
     withoutNavFooterIgnored: {
+      wordCountBefore: statsNotIgnored.wordCountBefore,
+      wordCountAfter: statsNotIgnored.wordCountAfter,
       wordDiff: statsNotIgnored.wordDiff,
       contentIncreaseRatio: statsNotIgnored.contentIncreaseRatio,
       citationReadability: statsNotIgnored.citationReadability,

package/src/index.d.ts

@@ -125,12 +125,16 @@ interface TextComparison {
 }
 
 interface BasicStats {
+  wordCountBefore: number;
+  wordCountAfter: number;
   wordDiff: number;
   contentIncreaseRatio: number;
   citationReadability: number;
 }
 
 interface ScenarioStats {
+  wordCountBefore: number;
+  wordCountAfter: number;
   wordDiff: number;
   contentIncreaseRatio: number;
   citationReadability: number;

package/test/index.test.js

@@ -52,10 +52,14 @@ describe('HTML Visibility Analyzer', () => {
     it('should provide basic comparison statistics', async () => {
       const result = await calculateStats(simpleHtml, richHtml);
 
+      expect(result).to.have.property('wordCountBefore');
+      expect(result).to.have.property('wordCountAfter');
       expect(result).to.have.property('wordDiff');
       expect(result).to.have.property('contentIncreaseRatio');
       expect(result).to.have.property('citationReadability');
 
+      expect(result.wordCountBefore).to.be.a('number');
+      expect(result.wordCountAfter).to.be.a('number');
       expect(result.wordDiff).to.be.a('number');
       expect(result.contentIncreaseRatio).to.be.a('number');
       expect(result.citationReadability).to.be.a('number');
@@ -68,7 +72,16 @@ describe('HTML Visibility Analyzer', () => {
 
       expect(result).to.have.property('withNavFooterIgnored');
       expect(result).to.have.property('withoutNavFooterIgnored');
+
+      // Verify withNavFooterIgnored has all required properties
+      expect(result.withNavFooterIgnored).to.have.property('wordCountBefore');
+      expect(result.withNavFooterIgnored).to.have.property('wordCountAfter');
       expect(result.withNavFooterIgnored).to.have.property('contentGain');
+      expect(result.withNavFooterIgnored).to.have.property('missingWords');
+
+      // Verify withoutNavFooterIgnored has all required properties
+      expect(result.withoutNavFooterIgnored).to.have.property('wordCountBefore');
+      expect(result.withoutNavFooterIgnored).to.have.property('wordCountAfter');
       expect(result.withoutNavFooterIgnored).to.have.property('missingWords');
     });
   });