Re: Lucene Excerpt Highlighting

Andreas Hartmann wrote:

> Hi Michael, hi Lenya developers,
>
> there is a small problem with Lucene excerpt highlighting:
>
> When I use the lucene-tag-name and lucene-class-value META tags,
> the indexing works correctly (only the text within the tag is
> indexed). But if the search term appears in the excerpt before
> the indexed text, it is highlighted at this position.
>
> An example:
>
> On cocooncenter, I have a menu "Home Articles Contact Search"
> on every page. When I'm searching for "Contact", the word from
> the menu text is highlighted, because it appears before the
> actually found word in the body. This makes the excerpt virtually
> useless for these search terms.
>
> Can you imagine an easy solution?


Well, I guess a  simple solution would be to dump (during indexing) the 
text which is actually being indexed
into another directory, e.g.

Crawling (dump ...)

Original document:    ..../htdocs_dump/employees/andreas.html

Indexing (dump once more ...)

Indexed text for creating the excerpt: ..../excerpts/employees/andreas.html

and then  you have to specify this "excerpts" directory within your 
sitemap rather than the "htdocs_dump" directory

One remark:
The "problem" exists because on the Lucene mailing list they are 
recommending that
you should NOT add your indexed text to the index as a field, because 
the index is becoming to large
which is going to slow down the searching. Maybe it would help to do 
some numbers: how slow does it become
with regard to how many number of characters and documents being 
indexed, etc.

I guess it would be nice if the "Lenya indexer" could be configured such 
that it can do either or.
(And the searcher as well of course).

What do you think?

HTH

Michael



>
>
> Regards,
> Andreas
>
>
> _______________________________________________
> WyonaCMS-dev mailing list
> [email protected]
> http://wyona.org/cgi-bin/mailman/listinfo/wyonacms-dev
>