cocoon-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Vadim Gritsenko" <vadim.gritse...@verizon.net>
Subject RE: Cocoon + Lucene + Xindice
Date Thu, 14 Feb 2002 13:37:21 GMT
> From: Yaser Al Masri [mailto:yalmasri1063@hotmail.com]
> 
> Hi,
> 
> I've been looking to the implementation of Lucene inside Cocoon, and
it's
> great what this combo is able to deliver.  Things will continue to
work
> perfectly until database shows.  The 'like' and 'contains' functions
in SQL
> and XPATH queries respectively cannot make much use of the indexing
supplied
> by the corresponding database because they're a sub-string search,
while
> search engines will help here because of the implementation of
full-text
> string searches, so I guess the justification for using search engines
> rather than query statements for this case is valid.
> 
> The problem here ASAIK is that to use lucene you have to supply it
with
> Directory object that points to the location of your data files, which
> doesn't exist here in case of database.

What do you mean here? Index is built independently of wherever your
data is stored, and I do not understand your conclusion.


> I thought I could mimic the
> implementation of Lucene inside cocoon but for the Xindice database
instead
> of the spreading files,

It already works on the content stored anywhere; search/indexing
mechanism does not care about data storage.


> so I started studying the crawler behavior and if it
> could be attained with the help of the pseudo protocol implementation
for
> Xindice, but I stuck there with the fact that I'm only mapping URLs to
> underlying database not to actual directories and files seen to the
indexer.

Indexer *does* *not* work on files.


> On more problem appeared to be there is the information retrieved from
the
> searcher.  For many cases that native XML databases help in, it would
be of
> no use to the use to get a URL for his search rather that a brief
> description to the document he searched (like in searching for items
in
> catalog where the description and the image matters more than the URL
that
> links to it).

I believe it is possible to have document abstract with Lucene. You need
to investigate Lucene capabilities and augment indexer with this
feature.


>  So I thought if Lucene could be adapted to retrieve documents
> inside collections rather than URL.

It is not a Lucene concern to retrieve documents; use components
specializing in this, like CIncludeTransformer.

 
> Can anybody give me an insight on this issue? Are there any future
plans to
> include search support for XML and other databases?

Again: Search engine does not care where your resources are stored.
Assuming this, search mechanism already have support for "XML and other
databases".


>  How do you think the
> support of XQuery language will help in the search process?

It's no help here. It's different domain.


Regards,
Vadim



---------------------------------------------------------------------
Please check that your question has not already been answered in the
FAQ before posting. <http://xml.apache.org/cocoon/faqs.html>

To unsubscribe, e-mail: <cocoon-users-unsubscribe@xml.apache.org>
For additional commands, e-mail: <cocoon-users-help@xml.apache.org>


Mime
View raw message