natural 6.7.2 → 6.8.1

package/lib/natural/brill_pos_tagger/lib/Lexicon.js

@@ -24,18 +24,20 @@ const dutchLexicon = require('../data/Dutch/brill_Lexicon.json')
 const DEBUG = false
 
 // Constructor creates a Lexicon for language
-function Lexicon (language, defaultCategory, defaultCategoryCapitalised) {
-  switch (language) {
-    case 'EN':
-      this.lexicon = englishLexicon
-      break
-    case 'DU':
-      this.lexicon = dutchLexicon
-      break
-    default:
-      this.lexicon = dutchLexicon
-      break
-  }
+function Lexicon (language, defaultCategory, defaultCategoryCapitalised, extendedLexicon) {
+  const lexicon = (() => {
+    switch (language) {
+      case 'EN':
+        return englishLexicon
+      case 'DU':
+        return dutchLexicon
+      default:
+        return dutchLexicon
+    }
+  })()
+
+  this.lexicon = Object.assign(lexicon, extendedLexicon || {})
+
   if (defaultCategory) {
     this.defaultCategory = defaultCategory
     if (defaultCategoryCapitalised) {

package/lib/natural/sentiment/SentimentAnalyzer.js

@@ -22,6 +22,8 @@
 
 'use strict'
 
+const DEBUG = true
+
 // Afinn
 const englishAfinnVoca = require('afinn-165')
 const spanishAfinnVoca = require('./Spanish/afinnShortSortedSpanish.json')
@@ -87,6 +89,17 @@ class SentimentAnalyzer {
     // this.vocabulary must be a copy of the languageFiles object
     // or in subsequent execution the polarity will be undefined
     // shallow copy - requires ES6
+    if (languageFiles[type]) {
+      if (languageFiles[type][language]) {
+        if (languageFiles[type][language][0]) {
+          this.vocabulary = Object.create(languageFiles[type][language][0])
+        }
+      } else {
+        throw new Error('Type ' + type + ' for Language ' + language + ' not supported')
+      }
+    } else {
+      throw new Error('Type Language ' + type + ' not supported')
+    }
     this.vocabulary = Object.assign({}, languageFiles[type][language][0])
     Object.setPrototypeOf(this.vocabulary, null)
     if (type === 'senticon') {
@@ -98,7 +111,6 @@ class SentimentAnalyzer {
         Object.keys(this.vocabulary).forEach(word => {
           this.vocabulary[word] = this.vocabulary[word].polarity
         })
-        // console.log(JSON.stringify(this.vocabulary, null, 2));
       }
     }
 
@@ -120,24 +132,24 @@ class SentimentAnalyzer {
   getSentiment (words) {
     let score = 0
     let negator = 1
-    // let nrHits = 0
+    let nrHits = 0
 
     words.forEach((token) => {
       const lowerCased = token.toLowerCase()
       if (this.negations.indexOf(lowerCased) > -1) {
         negator = -1
-        // nrHits++
+        DEBUG && nrHits++
       } else {
         // First try without stemming
         if (this.vocabulary[lowerCased] !== undefined) {
           score += negator * this.vocabulary[lowerCased]
-          // nrHits++
+          DEBUG && nrHits++
         } else {
           if (this.stemmer) {
             const stemmedWord = this.stemmer.stem(lowerCased)
             if (this.vocabulary[stemmedWord] !== undefined) {
               score += negator * this.vocabulary[stemmedWord]
-              // nrHits++
+              DEBUG && nrHits++
             }
           }
         }
@@ -145,7 +157,7 @@ class SentimentAnalyzer {
     })
 
     score = score / words.length
-    // console.log("Number of hits: " + nrHits);
+    DEBUG && console.log('Number of hits: ' + nrHits)
 
     return score
   }

package/package.json

@@ -1,7 +1,7 @@
 {
   "name": "natural",
   "description": "General natural language (tokenizing, stemming (English, Russian, Spanish), part-of-speech tagging, sentiment analysis, classification, inflection, phonetics, tfidf, WordNet, jaro-winkler, Levenshtein distance, Dice's Coefficient) facilities for node.",
-  "version": "6.7.2",
+  "version": "6.8.1",
   "homepage": "https://github.com/NaturalNode/natural",
   "repository": {
     "type": "git",