Hi all

This provides an higher level overview about the work on STANBOL-353,
STANBOL-374, STANBOL-375, STANBOL-376, STANBOL-377 and STANBOL-380.
People that want/need to know the details please have a look at the
listed Issues and the commit statements of the linked commits.

Managed- and Referenced Solr Server
============================

Stanbol supports now two ways of running Solr servers (CoreContainer)
within the same JVM:

1. ManagedSolrServer: Allows to create/update/activate/deactivate and
remove Solr indexes (SolrCore).
2. ReferencedSolrServer: Allows to init a Solr server for a directory
on the local file system.

In both cases the CoreContainer and SolrCore instances are registered
as OSGI Services with a set of Metadata. Therefore other component can
use the Whiteboard Pattern to work with Solr servers and indexes. This
possibility is already internally used to publish SolrServers on the
StanbolRESTful interface. Both the Referenced- and the
ManagedSolrServer do use the same set of properties when registering
CoreContainer and SolrCores.

For the details how to use this new Services see the main
documentation file for the org.apache.stanbol.commons.solr artifacts
[1].

DBpedia bundle, Default data and upgrade to a full index.
=========================================

The dbpedia.defaultdata bundle was replaced by a dbpedia bundle.
Upgrading to a full dbpedia index does no longer require to first
remove the dbpedia.defaultdata bundle and than install an other
dbpedia bundle (e.g. the one created by the indexing tool).

Now to upgrade to an bigger dbpedia index only requires users to copy
the "dbpedia.solrindex.zip" file to the "/datafiles" folder of your
Stanbol installation. The file will be detected and the upgrade to the
bigger index will start immediately. During the upgrade queries will
use the old index. After the upgrade is complete the
"dbpedia.solrindex.zip" can be safely deleted from the "/datafiles"
directory.

To upgrade from one full index to an other one needs to

1. delete any existing "dbpedia.solrindex.zip" from the "/datafiles" folder.
2. stop the dbpedia bundle
3. copy the new "dbpedia.solrindex.zip" to the "/datafiles" folder
4. start the dbpedia bundle.

If step three is skipped the dbpedia index will be reinitialized based
on the default index including 43k entities.

Full DBPedia indexes can be downloaded from [2] or created by using
the "/entityhub/indexing/dbpedia" indexing utility.

This functionality uses the ability to specify multiple
"Index-Archives" in Solr-Index-Archive-Reference files. Details are
provided in [1]


best
Rupert Westenthaler

[1] 
http://svn.apache.org/repos/asf/incubator/stanbol/trunk/commons/solr/README.md
[2] http://dev.iks-project.eu/downloads/stanbol-indices/

-- 
| Rupert Westenthaler                            [email protected]
| Bodenlehenstraße 11                             ++43-699-11108907
| A-5500 Bischofshofen

Reply via email to