Updated utils

Now takes a string instead of a list of words and tokenize that string
of words using NLTK's word_tokenizer

Author: mac389

.DS_Store

@@ -1,11 +1,6 @@
 import couchdb	
 
 from sys import argv,path
-
-
-
-
-#initialization
 path.append('/Volumes/My Book/Dropbox/ToxTweet/Software/APIs/ytpy/lib')
 path.append('/Volumes/My Book/Dropbox/ToxTweet/Software/APIs/ytpy/examples')
 
@@ -54,7 +49,8 @@ def save(video,format='txt'):
 
 data = YouTubeSearch(drug)
 
-def update_database
+def update_database():
+	pass
 
 server  = couchdb.Server()
 

examples/ytpy.py

@@ -160,8 +160,9 @@ def clean(tweet,keepTags = False):
 
 def remove_stopwords(utterance, languages=['english','spanish','french']): 
 	#languages is a list
+	utterance = word_tokenize(utterance)
 	from nltk.corpus import stopwords
-	my_stopwords = map(lambda word: word.rstrip('\n'),open('stopwords','rb').readlines())
+	my_stopwords = map(lambda word: word.rstrip('\n'),open('/Volumes/My Book/Dropbox/ToxTweet/Software/APIs/ytpy/constants/stopwords','rb').readlines())
 	allowed_languages = {'english','spanish','french'} #This is a set
 	#Serial processing is least obfuscated
 	for language in set(languages).intersection(allowed_languages):