2 Replies Latest reply on Sep 26, 2006 10:53 AM by r_mun

Full Text Search Within JBoss Portal

r_mun Sep 20, 2006 12:16 PM

Hello,

I'm new to JBoss portall, and portal application architecture.
However we are now working over the project, requiring such kind of architecture. Our portal would require full-text search over different kinds of resources (our database, HTML, PDF, etc).

I performed some investigation on the search engine and found out that JBoss Portal uses Jackrabbit (http://jackrabbit.apache.org/) as one of its components which in turn is built upon Lucene ( http://lucene.apache.org/ ) framework (a full-text search engine).

In our application we will need to maintain structured (properties defined in a database) and unstructured (.html, .pdf, .doc etc) information.

I have several questions about search capabilities and implementation of search.

[Q001] How is the search performed over the structured data in the database? Is it possible by means of JBoss Portal at all?

[Q002] How is the unstructured data put and maintained in the repository?

The least help I'm counting on is a list of links where I can find information concearning these questions. (I couldn't find any)

Regards,
Roman Muntyanu

1. Re: Full Text Search Within JBoss Portal

unibrew Sep 20, 2006 12:35 PM (in response to r_mun)

Hello

This topic should partially answer your questions: http://www.jboss.com/index.html?module=bb&op=viewtopic&t=86785

Regards
-------------------------
Ryszard Kozmik
JBoss Forums
JBoss Labs Team
Actions
2. Re: Full Text Search Within JBoss Portal

r_mun Sep 26, 2006 10:53 AM (in response to r_mun)

Hi, again

First of all, Ryszard, thanks for answering my questions.
I have made some investigations lately and I want to sare the results of my research.

[A001] If to use the search engines like Lucene the search is performed as described in
http://wiki.apache.org/jakarta-lucene/LuceneFAQ#head-109358021acbfc89456e446740dc2bbf9049950f

If you want some more details on this you can check the http://mail-archives.apache.org/mod_mbox/lucene-java-user.
Use mails starting from the most recent (for now its Sep 2006) and go back, month by month searching for "database" in the message topics.
Or google {+"index a database" +lucene}

If you dont have time to implement offered functionality you may use two of the following (when writting this I couldn't find more) http://www.dbsight.net/
or
http://www.opensymphony.com/compass/ (Chapters 16-21 of the reference manual). The latter one is an opensource.

[A002] I guess it's a case of study of a specific framework.

Hope this helps someone someday :)
If there are any thought about the above, or you know some other frameworks, or under JBoss Portal it's differen - I will gladly accept the remarks. Thanks.

Best regards,
Roman Muntyanu
Actions

Go to original post