Merge pull request nltk#1011 from longdt219/fix_embedding

Fix bug import nltk#1002

Author: stevenbird

nltk/corpus/__init__.py

@@ -64,7 +64,6 @@
 from nltk.tokenize import RegexpTokenizer
 from nltk.corpus.util import LazyCorpusLoader
 from nltk.corpus.reader import *
-from nltk.data import find
 
 abc = LazyCorpusLoader(
     'abc', PlaintextCorpusReader, r'(?!\.).*\.txt', encoding=[
@@ -253,7 +252,6 @@
     'semcor', SemcorCorpusReader, r'brown./tagfiles/br-.*\.xml',
     wordnet) # Must be defined *after* wordnet corpus.
 
-word2vec_sample = str(find('models/word2vec_sample/pruned.word2vec.bin'))
 
 def demo():
     # This is out-of-date:

nltk/test/gensim.doctest

@@ -38,8 +38,8 @@ Using the pre-trained model
 ~~~~~~~~~~~~~~~~~~~
 NLTK also include a pre-trained model which is part of a model that is trained on 100 billion words from Google News Dataset. 
 The full model is from https://code.google.com/p/word2vec/  which is about 3 Gb.    
- 
-	>>> from nltk.corpus import word2vec_sample
+ 	>>> from nltk.data import find
+	>>> word2vec_sample = str(find('models/word2vec_sample/pruned.word2vec.bin'))
 	>>> model = gensim.models.Word2Vec.load(word2vec_sample)
 
 We pruned the model to only include the most common words (~44k words).