From 31aa8425de4f7656b52ce6f359ad91967d28887e Mon Sep 17 00:00:00 2001
From: Jacob Harris <jharris@nytimes.com>
Date: Sat, 28 Dec 2013 22:26:21 -0500
Subject: [PATCH] Don't tokenize empty tweets

---
 ebook.rb | 1 +
 1 file changed, 1 insertion(+)

diff --git a/ebook.rb b/ebook.rb
index 8c26b02..f3c3aff 100644
--- a/ebook.rb
+++ b/ebook.rb
@@ -82,6 +82,7 @@ else
   tokenizer = Punkt::SentenceTokenizer.new(source_tweets.join(" "))  # init with corpus of all sentences
 
   source_tweets.each do |twt|
+    next if twt.nil? || twt == ''
     sentences = tokenizer.sentences_from_text(twt, :output => :sentences_text)
 
     # sentences = text.split(/[.:;?!]/)