Open GoogleCodeExporter opened 8 years ago
Issue 45 has been merged into this issue.
Original comment by ckkohl79
on 25 Mar 2012 at 2:12
Adding FORM as an ignorable element at the highlighter (but not at the
Extractor itself) has two disadvantages:
1. The highlighted HTML will not be consistent with the TextDocument's content
information (i.e., it should rather go into DefaultTagActionMap)
2. FORM may span over many otherwise relevant text blocks. Adding FORM as
TA_IGNORABLE_ELEMENT significantly reduces extraction accuracy (with L3S-GN1,
avg. token-level F1 goes down by about 5% to 7%).
Could you please give examples (URLs) where boilerpipe fails currently?
Please try the very latest version from trunk, or -- preferably -- the version
at http://boilerpipe-web.appspot.com/
Original comment by ckkohl79
on 25 Mar 2012 at 2:19
Adding FORM as an ignorable element at the highlighter (but not at the
Extractor itself) has two disadvantages:
1. The highlighted HTML will not be consistent with the TextDocument's content
information (i.e., it should rather go into DefaultTagActionMap)
2. FORM may span over many otherwise relevant text blocks. Adding FORM as
TA_IGNORABLE_ELEMENT significantly reduces extraction accuracy (with L3S-GN1,
avg. token-level F1 goes down by about 5% to 7%).
Could you please give examples (URLs) where boilerpipe fails currently?
Please try the very latest version from trunk, or -- preferably -- the version
at http://boilerpipe-web.appspot.com/
Original comment by ckkohl79
on 25 Mar 2012 at 2:19
Original issue reported on code.google.com by
xavi.beu...@gmail.com
on 24 Mar 2012 at 6:40