tfidf.rb in company-mapping-0.2.0

- old
+ new

@@ -1,9 +1,10 @@
 module Company
   module Mapping
 
-#TFIDF class implements Term Frequency Inverse Document Frequency statistic.
+#TFIDF class implements Term Frequency Inverse Document Frequency statistic. Term frequency–inverse document frequency,
+# is a numerical statistic that is intended to reflect how important a word is to a document in a collection or corpus.
     class TFIDF
       attr_accessor :tf, :idf
 
       def initialize(corpus)
         @corpus = corpus
@@ -11,70 +12,41 @@
 
       #Calculates the tf-idf weights in the given corpus
       def calculate
         @tfidf = Hash.new
 
-        if (@idf==nil)
-          @idf = InverseDocumentFrequency.new(@corpus)
-        end
-
-        if (@tf==nil)
-          _tokenizer = BasicTokenizer.new
-          @tf = NormalizedTermFrequency.new(_tokenizer)
-        end
-
+        @idf ||= InverseDocumentFrequency.new(@corpus)
+        @tf ||= NormalizedTermFrequency.new(BasicTokenizer.new)
         @idf_weights = @idf.calculate
 
-        @corpus.each {
-            |doc|
+        @corpus.each do |doc|
+          termfreq = @tf.calculate(doc.contents)
 
-          _termfreq = @tf.calculate(doc.contents)
-
-          _tfidf_weights = Hash.new
-
-          _termfreq.each do |term, tf|
-            _weight = tf * @idf_weights[term]
-            _tfidf_weights[term] = _weight
-          end
-
-          @tfidf[doc.id] = _tfidf_weights
-        }
-        return @tfidf
+          @tfidf[doc.id] =
+              termfreq.each_with_object({}) do |(term, tf), tfidf_weights|
+                weight = tf * @idf_weights[term]
+                tfidf_weights[term] = weight
+              end
+        end
+        @tfidf
       end
 
       #Calculates tfidf weights of new incoming document without importing the document in the corpus and re-calculating the tf-idf weights for the entire corpus
       def calculate_tfidf_weights_of_new_document(new_doc)
-        _termfreq = @tf.calculate(new_doc.contents)
+        termfreq = @tf.calculate(new_doc.contents)
 
-        _tfidf_weights = Hash.new
-
-        _termfreq.each do |term, tf|
-          if (@idf_weights.has_key? term)
-            _weight = tf * @idf_weights[term]
-          else
-            _weight = tf * @idf.maxIDF
-          end
-          _tfidf_weights[term] = _weight
+        @tfidf[new_doc.id] = termfreq.each_with_object({}) do |(term, tf), tfidf_weights|
+          weight = tf * (@idf_weights[term] || @idf.maxIDF)
+          tfidf_weights[term] = weight
         end
-        @tfidf[new_doc.id] = _tfidf_weights
-        return @tfidf
+        @tfidf
       end
 
       #Calculates tf-idf similarity between two given documents. It is actually
       #the calculated Cosine Similarity by using tf*idf weights.
       def similarity(doc1_id, doc2_id)
-        if (@tfidf==nil)
-          calculate
-        end
-
-        _cosine_similarity = CosineSimilarity.new
-        return _cosine_similarity.calculate(@tfidf[doc1_id], @tfidf[doc2_id])
-      end
-
-      def info
-        " term frequency–inverse document frequency, is a numerical "
-        +"statistic that is intended to reflect how important a word "
-        +"is to a document in a collection or corpus"
+        @tfidf ||= calculate
+        CosineSimilarity.new.calculate(@tfidf[doc1_id], @tfidf[doc2_id])
       end
     end
 
   end
 end
\ No newline at end of file