Hi,
ad 1) the highlighter branch is there, did you pull the newest version?
ad 2) the basic QueryScorer calls extractTerms() method on the supplied query,
so you can use it only with queries that implement this method.
All primitive queries do implement it, Wildcard query is not a primitive query
and so you have to call its rewrite method and use the rewritten query.
ad 3) the new implementation – SpanHighlightScorer – is able to rewrite the
queries itself because it makes (if necessary) its own memoryindex
containing just the highlighted document. The queries are rewritten using this
small index. For a wildcard query this normally leads to a much smaller
number of matching terms and so you avoid the max clause problem. (But of
course, if your Wildcard query matches more than maxTerms different terms
in the highlighted text, than you have to increase this number) – How big are
your documents?
Jiri
From: muhammad ismael [mailto:m.ismae...@gmail.com]
Sent: Wednesday, January 05, 2011 7:00 PM
To: clucene-developers@lists.sourceforge.net
Subject: Re: [CLucene-dev] CLucene-developers Digest, Vol 57, Issue 4
please could you write how you instantiate the highlighter? Which
scorer do you use?
m_highlightQuery = m_pMainQuery->clone();
Query * q = m_highlightQuery->rewrite(m_pReader);
m_queryScorer = new QueryScorer(q);
m_highlighter = new Highlighter(m_queryScorer);
Are you using the highlighter version from the highlighter branch?
No i am on master branch and i can not find highlighter branch , i merged the
wildcardQuery_fix and memory_leaks branches to master.
I would try to find the place, where the exception is thrown.
But still, there is a problem along with the ConstantScoreQuery. Both
ways to highlight a document
that are available now will not work because:
1) If you use the "basic" highlighting - means you use the
CL_NS2(search,highlight)::QueryScorer - than this scorer calls the
extractTerms() method which is not
implemented in ConstantScoreQuery as this query uses filters to match
documents.
2 the "exact" highlighting - means you use the
CL_NS2(search,highlight)::SpanHighlightScorer - also does not highlight
ConstantScoreQueries for the same reason.
is this means that i can not use WildcardQuery with highlighter without setting
the maximum clauses?
But, you could use this scorer with the original WildcardQuery and
allow rewriting queries in place. Then if your documents are not too long and
the single documents does not contain more the the max clause limit
different matching terms, then this would work.
If you want to use the new highlighter scorer then I could send you
piece of code, how to call it.
Note :-
I can not use setMaxClauses because my documents are too long and if i use
setMaxClauses i always get too many clauses exception
Mohammad Ismael
------------------------------------------------------------------------------
Learn how Oracle Real Application Clusters (RAC) One Node allows customers
to consolidate database storage, standardize their database environment, and,
should the need arise, upgrade to a full multi-node Oracle RAC database
without downtime or disruption
http://p.sf.net/sfu/oracle-sfdevnl
_______________________________________________
CLucene-developers mailing list
CLucene-developers@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/clucene-developers
