Ignore URLs by ignoring strings with .

1 files changed, 1 insertions, 1 deletions

diff --git a/data.py b/data.py
index 0a1fbc8..87dbc91 100644
--- a/data.py
+++ b/data.py
@@ -24,7 +24,7 @@ text = [unescape(sub(r'<[^>]*>', ' ', status[2])) for status in statuses]
 words = [word for message in text for word in message.split()]
 # Remove URLs and special characters and convert to lowercase
 words = [sub(r'[^a-z0-9]', '', word.lower())
-         for word in words if word.find('://') == -1]
+         for word in words if word.find('.') == -1]
 # Remove empty strings
 words = [word for word in words if word != '']