Library does not produce same results as http://boilerpipe-web.appspot.com/ about boilerpipe HOT 5 OPEN

It looks like the issue is the KeepLargestBlockFilter which rejects every block 
except the largest. While taking out this filter in the library should return 
results closer to http://boilerpipe-web.appspot.com/, it looks like that will 
cause some non-article things that are rejected by 
http://boilerpipe-web.appspot.com/ to make it through.

Is there anyplace I can go to see what filters are being by the article 
extractor used by http://boilerpipe-web.appspot.com/?

I'm also unable to get the same results using the HTMLHighlighter in extraction 
mode.  The web API (http://boilerpipe-web.appspot.com) clearly states that:

"This Web Application probably uses a more recent version than the released 
versions in the Boilerpipe Google Code page. You might thus get slightly 
different (hopefully better) results."

In the web demo, are you using the HTMLHighligher to extract HTML, or have you 
added a different approach to boilerpipe-core?  If changes have been made to 
core, I'd be happy to contribute to the open source version.  Hints would be 
appreciated =)

I got the similar issue.
When trying the URL "http://www.hokkaido-np.co.jp/news/donai/424760.html"
With ArticleExtractor and "Plain Text" output

Library code did not produce same results as http://boilerpipe-web.appspot.com/

Building from source will fix this issue.

I have done a build from SVN. Still Japanese URL gives bad results. Same works 
gr8 on the Web App. Code snippet below:
    URL url = new URL("http://d.hatena.ne.jp/mkusunok/20130817/p1");
    String text = ArticleExtractor.INSTANCE.getText(url);