Fix for the preview and for dataset streaming
Browse filesSince the dataset viewer uses dataset streaming to show the first rows of your dataset, you have to make the dataset compatible with streaming for the preview to work.
However `xml.dom.minidom.parse` does not support streaming out-of-the-box (it can't read a remote file). To fix it, you need to `open` the file first, before passing it to `xml.dom.minidom.parse`.
This works because `open` in the dataset script is extended to support reading remote files over HTTP, especially to enable streaming.
- SemEval2015.py +2 -1
SemEval2015.py
CHANGED
@@ -143,7 +143,8 @@ class SemEvalXMLDataset():
|
|
143 |
self.SentenceWithOpinions = []
|
144 |
self.xml_path = file_name
|
145 |
|
146 |
-
|
|
|
147 |
|
148 |
for sentenceXml in self.sentenceXmlList:
|
149 |
reviewId = sentenceXml.getAttribute("id").split(':')[0]
|
|
|
143 |
self.SentenceWithOpinions = []
|
144 |
self.xml_path = file_name
|
145 |
|
146 |
+
with open(self.xml_path) as f:
|
147 |
+
self.sentenceXmlList = parse(f).getElementsByTagName('sentence')
|
148 |
|
149 |
for sentenceXml in self.sentenceXmlList:
|
150 |
reviewId = sentenceXml.getAttribute("id").split(':')[0]
|