From b943471d2a51326ebca422b9bf1dfa34f1810523 Mon Sep 17 00:00:00 2001
From: "Ossama W. Obeid" <owo@owobeid.com>
Date: Thu, 1 Jan 2015 17:57:02 +0300
Subject: [PATCH] Added a function to measure the similarity of two words.

---
 glove/glove.py | 31 +++++++++++++++++++++++++++++++
 1 file changed, 31 insertions(+)

diff --git a/glove/glove.py b/glove/glove.py
index d2340b4..bfbd6ea 100644
--- a/glove/glove.py
+++ b/glove/glove.py
@@ -227,6 +227,37 @@ def most_similar(self, word, number=5):
 
         return self._similarity_query(self.word_vectors[word_idx], number)[1:]
 
+    def _similarity(self, word1_vec, word2_vec):
+        dst = (np.dot(word1_vec, word2_vec)
+               / np.linalg.norm(word1_vec)
+               / np.linalg.norm(word2_vec))
+
+        return dst
+
+    def similarity(self, word1, word2):
+        """
+        Return the similarity measure between word1 and word2.
+        """
+
+        if self.word_vectors is None:
+            raise Exception('Model must be fit before querying')
+
+        if self.dictionary is None:
+            raise Exception('No word dictionary supplied')
+        
+        try:
+            word1_idx = self.dictionary[word1]
+        except KeyError:
+            raise Exception('Word not in dictionary')
+
+        try:
+            word2_idx = self.dictionary[word2]
+        except KeyError:
+            raise Exception('Word not in dictionary')
+
+        return self._distance(self.word_vectors[word1_idx],
+                              self.word_vectors[word2_idx])
+        
     def most_similar_paragraph(self, paragraph, number=5, **kwargs):
         """
         Return words most similar to a given paragraph (iterable of tokens).