HTMLStripStandardTokenizerFactory doesn't interpret word boundaries on html
tags correctly.
-------------------------------------------------------------------------------------------
Key: SOLR-678
URL: https://issues.apache.org/jira/browse/SOLR-678
Project: Solr
Issue Type: Bug
Components: search
Affects Versions: 1.2
Environment: Mac OS X 10.5.4, java version "1.5.0_13"
Reporter: Matt Connolly
The HTMLStripStandardTokenizerFactory filter does not place word boundaries on
HTML tags like it should.
For example, indexing the text "<h2>title</h2><p>some comment</p>" results in
two words being indexed: "titlesome" and "comment" when there should be three
words: "title" "some" and "comment".
Not all tags need this, for example, it may be perfectly reasonable to write
"<b>sub</b>script" to be indexed as "subscript" since the <b> is interpretted
as inline, not block.
I would suggest all block or paragraph tags be translated into spaces so that
text on either side of the tag is considered separate tokens. eg: p div h1 h2
h3 h4 h5 h6 br hr pre (etc)
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
