seo-gravity-mcp 1.0.0 → 1.0.1

package/dist/utils/nlp.js

@@ -1,5 +1,40 @@
-import natural from 'natural';
-const tokenizer = new natural.WordTokenizer();
+// Native lightweight word tokenizer
+function tokenize(text) {
+    return text.match(/[a-zA-Z0-9'-]+/g) || [];
+}
+// Built-in TF-IDF implementation without heavy external dependencies
+class SimpleTfIdf {
+    docs = [];
+    addDocument(doc) {
+        const rawTokens = tokenize(doc.toLowerCase());
+        const tokens = rawTokens.filter(t => t.length >= 3 && !stopwords.has(t) && !/^\d+$/.test(t));
+        this.docs.push(tokens);
+    }
+    listTerms(docIndex) {
+        if (docIndex < 0 || docIndex >= this.docs.length)
+            return [];
+        const docTokens = this.docs[docIndex];
+        if (docTokens.length === 0)
+            return [];
+        const totalDocs = this.docs.length;
+        const termCounts = new Map();
+        for (const t of docTokens) {
+            termCounts.set(t, (termCounts.get(t) || 0) + 1);
+        }
+        const results = [];
+        for (const [term, count] of termCounts.entries()) {
+            const tf = count / docTokens.length;
+            let docFreq = 0;
+            for (const d of this.docs) {
+                if (d.includes(term))
+                    docFreq++;
+            }
+            const idf = Math.log((1 + totalDocs) / (1 + docFreq)) + 1;
+            results.push({ term, tfidf: tf * idf * 10 });
+        }
+        return results.sort((a, b) => b.tfidf - a.tfidf);
+    }
+}
 const stopwords = new Set([
     'a', 'about', 'above', 'after', 'again', 'against', 'all', 'am', 'an', 'and', 'any', 'are', 'aren\'t', 'as', 'at',
     'be', 'because', 'been', 'before', 'being', 'below', 'between', 'both', 'but', 'by', 'can', 'can\'t', 'cannot',
@@ -21,7 +56,7 @@ const stopwords = new Set([
  */
 export function extractKeyphrases(text, maxItems = 30) {
     const clean = text.toLowerCase().replace(/[^a-z0-9\s-]/g, ' ');
-    const rawTokens = tokenizer.tokenize(clean) || [];
+    const rawTokens = tokenize(clean);
     const tokens = rawTokens.filter(t => t.length > 2 && !stopwords.has(t) && !/^\d+$/.test(t));
     const totalTokens = tokens.length || 1;
     const phraseCounts = new Map();
@@ -54,7 +89,7 @@ export function extractKeyphrases(text, maxItems = 30) {
  * Computes TF-IDF between target text and a set of competitor texts to detect content gaps.
  */
 export function computeContentGapTfIdf(targetText, competitorTexts) {
-    const tfidf = new natural.TfIdf();
+    const tfidf = new SimpleTfIdf();
     // Doc 0 is target text
     tfidf.addDocument(targetText);
     // Docs 1..N are competitors
@@ -75,7 +110,7 @@ export function computeContentGapTfIdf(targetText, competitorTexts) {
         });
     }
     // Find target frequencies
-    const targetTokens = (tokenizer.tokenize(targetText.toLowerCase()) || []).filter(t => !stopwords.has(t));
+    const targetTokens = tokenize(targetText.toLowerCase()).filter(t => !stopwords.has(t));
     const targetTokenCounts = new Map();
     for (const t of targetTokens) {
         targetTokenCounts.set(t, (targetTokenCounts.get(t) || 0) + 1);
@@ -104,7 +139,7 @@ export function computeContentGapTfIdf(targetText, competitorTexts) {
 export function calculateReadability(text) {
     const clean = text.replace(/\s+/g, ' ').trim();
     const sentences = clean.split(/[.!?]+/).filter(s => s.trim().length > 0);
-    const words = (tokenizer.tokenize(clean) || []).filter(w => /[a-zA-Z]/.test(w));
+    const words = tokenize(clean).filter(w => /[a-zA-Z]/.test(w));
     const numSentences = Math.max(sentences.length, 1);
     const numWords = Math.max(words.length, 1);
     let numSyllables = 0;
@@ -129,7 +164,7 @@ export function calculateReadability(text) {
     // Long sentences (> 25 words)
     const longSentences = sentences
         .map(s => s.trim())
-        .filter(s => (tokenizer.tokenize(s) || []).length > 25);
+        .filter(s => tokenize(s).length > 25);
     return {
         fleschReadingEase: Math.max(0, Math.min(100, fleschReadingEase)),
         gradeLevel,

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "seo-gravity-mcp",
-  "version": "1.0.0",
+  "version": "1.0.1",
   "description": "Comprehensive Next-Gen SEO, GEO & Competitor Intelligence MCP Server for Antigravity",
   "main": "dist/index.js",
   "type": "module",
@@ -32,15 +32,12 @@
     "@modelcontextprotocol/sdk": "^1.6.1",
     "axios": "^1.7.9",
     "cheerio": "^1.0.0",
-    "dotenv": "^16.4.7",
     "fast-xml-parser": "^5.0.8",
     "jsdom": "^26.0.0",
-    "natural": "^8.0.1",
     "zod": "^3.24.2"
   },
   "devDependencies": {
     "@types/jsdom": "^21.1.7",
-    "@types/natural": "^5.1.5",
     "@types/node": "^22.13.5",
     "ts-node": "^10.9.2",
     "typescript": "^5.7.3"