Hi Hibernator,
that's a nice idea and shouldn't be too hard to get working.
Quote:
I don't know if HS puts some metainformation in the indexes and solr can not read them...
Hibernate Search doesn't write directly to the index but uses Lucene, so everything written by Search can obviously be read (and searched for) by Solr too.
Basically it writes in the index what you have annotated on your entities, so you have good control of it. The only exception is the class name, it's a mandatory field which we need to store in the index, but Solr can ignore it, or make good use of it for some filters if you like.
Quote:
i also think that i have to create an schema in solr with the index fields...
That's right, but should be quite easy. You can inspect the index (assuming it's not empty) with Luke and verify how closely it matches your entity mapping.
It would be very nice if you could contribute a tool which generates such a schema directly from our mapping; it would need some changes in the DocumentBuilder in Hibernate Search to "export" it's metadata information.. feel free to ask more details as needed, or join the developer mailing list
http://hibernate.org/community/mailinglists