Fix for the preview and for dataset streaming

SemEval2015.py

@@ -143,7 +143,8 @@ class SemEvalXMLDataset():
         self.SentenceWithOpinions = []
         self.xml_path = file_name
 
-        self.sentenceXmlList = parse(self.xml_path).getElementsByTagName('sentence')
+        with open(self.xml_path) as f:
+            self.sentenceXmlList = parse(f).getElementsByTagName('sentence')
 
         for sentenceXml in self.sentenceXmlList:
             reviewId = sentenceXml.getAttribute("id").split(':')[0]