I would like to develop a one search solution that enables me to search across domain object (using Hibernate Search) and documents (pdfs, rtfs, etc) using native Lucene. I would like to have a Search Manager that uses MultiSearcher to search across all data.

I am not sure if this is the right approach, I'm not sure what I should to return..do I return Document objects and/or domain objects.

I would be grateful if someone might be able to advise me on what approach I should take.


Again any help would be appreciated!

You are probably going to store the pdfs and all in a database anyway,
so if you map them as BLOB fields of some entity, you can create a hierarchy of entities and add all attributes you'll need in the presentation layer.

You can map the BLOB analyzers declaratively using Hibernate Search, and use the existing ones from the Lucene project's contributions.

You may want to think about what your presentations layer needs first: a generic interface is desirable IMHO.

Hi

Thanks for your quick reply! The documents are not stored in the database. These documents are attachments that users upload to a document repository.

I guess I could return either a document summary (link to document) and return domain object.

Not sure if I have provided enough information, I am preparing a proposal and I want to make sure I know what I;m talking about!


Cheers

Hi,
even if the documents are not stored in the database, you may want to have some metadata stored in the database, such as the link and/or path to the resource. This way you can still define your own Search Bridge to provide custom file format management.
I would still recommend to make sure all types returned by your search have something in common, can be an interface or a common supertype, to make sure you don't have to code different views to show the results.
You should read the book, it is plenty of examples helpful for this situation.

If you prefer you may access the indexes with Lucene's API, which is exposed through Hibernate Search, but you should not use it to make changes to the index. Of course you may want whatever you want with indexes not directly managed by H.S.

Hi

Thanks for your valuable advice. I have already taken some of them into account. I will create a Summary interface which will be returned from different search components. The SearchManager will return a SearchResponse which contains the list of summary objects.

We have a interesting xml requirement which is stored within a Domain object (not very happy with this) but I will create a Custom class brigde that would take the xml and values and index these.


Cheers