lucenenet-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Simon Svensson <>
Subject Re: FastVectorHilighter
Date Tue, 28 Aug 2012 09:22:32 GMT

It sounds like you've hit LUCENE-3332 
(FastVectorHighlighter ignores MultiPhraseQuery (and more)). It's not 
yet fixed in Lucene, and this makes me wonder; do we "fix" things that 
have yet to be fixed in Lucene? Wouldn't that mean that we're ahead of 
the Lucene project for some files, and behind for others? Is this a problem?

Anyway, creating a Jira issue and attaching any patches is always welcome.

// Simon

On 2012-08-28 10:53, Gian Maria Ricci wrote:
> Hi to everyone,
> I'm using the fastVectorHilighter but I found little problems, first 
> of all it does not highlight if you use PrefixQuery or payload query, 
> the problem is quite simple, is just because in the code there are 
> checks like this one
> if(query isTermQuery)
>                 {
>                     AddTerm(((TermQuery)query).GetTerm(), 
> query.GetBoost());
> }
> That prevent from using other types of query. Another issue is that if 
> I want a piece of text of X character length, the match word is at the 
> very beginning, but I really wish to have them in the middle, because 
> I'm interested in text around the word, and I think that this should 
> be a reasonable default.
> I've modified the Contrib.FastVectoryHilighter project to fix those 
> two problem (I'm still testing everything) so I'm asking if there is 
> plan to already fix those stuff in the next version so I can do 
> minimal change now waiting for the next version, or if I can invest 
> some of my time continue to fix those issue in the actual version, (if 
> you want to review my patch to include in the code you are welcome).
> Thanks for your attention and thanks for the exceptional work you are 
> doing with Lucene.NET.
> --
> Ricci Gian Maria
> MVP Visual Studio ALM
> <>
> Twitter:
> Msn: <>
> Description: cid:image001.png@01CC0BF5.2E57C9A0

  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message