Author: suat
Date: Thu Oct 6 08:50:17 2011
New Revision: 1179535
URL: http://svn.apache.org/viewvc?rev=1179535&view=rev
Log:
STANBOL-340:
-Added the README.md file (provided by Anil) to Contenthub
-Deleted the outdated README from search
Added:
incubator/stanbol/trunk/contenthub/README.md
Removed:
incubator/stanbol/trunk/contenthub/search/README.md
Added: incubator/stanbol/trunk/contenthub/README.md
URL:
http://svn.apache.org/viewvc/incubator/stanbol/trunk/contenthub/README.md?rev=1179535&view=auto
==============================================================================
--- incubator/stanbol/trunk/contenthub/README.md (added)
+++ incubator/stanbol/trunk/contenthub/README.md Thu Oct 6 08:50:17 2011
@@ -0,0 +1,67 @@
+# Apache Stanbol Contenthub
+
+Apache Stanbol Contenthub is a persistent document store which enables text
based document submission
+and semantic search together with faceted search capability on submitted
documents.
+
+## ContentHub Store
+
+It is the subcomponent that actually stores the document and its metadata
persistently. In current implementation only text/plain documents are allowed.
+
+Storage part of Contenthub provides basic methods such as create, put, get and
delete. When a document is
+submitted, it delegates the textual content to Stanbol Enhancer to get its
enhancements. While submitting the document, it is also possible to specify
external metadata as field value pairs to the document.
+
+The document itself and all specified external metadata are indexed through an
embedded Apache Solr core which is created specifically for Contenthub.
+Since documents are given unique IDs while indexing, using its unique ID, a
document can be retrieved or deleted from Contenthub.
+ContentHub provides an HTML interface for its functionalities under the
following endpoint, which is available after running the Full Launcher of
Apache Stanbol:
+
+ http://localhost:8080/contenthub
+
+## ContentHub Search
+
+ContentHub has a semantic search subcomponent that allows to make search over
submitted documents. HTML interface for search functionality can be reached
under:
+
+ http://localhost:8080/contenthub/search
+
+To start a search, one should enter a keyword and choose the search engines
that will execute. After having the first search results, all facets and values
of these facets will also arrive. Later on, when a facet constraint is chosen,
documents and facets will be dynamically updated according to chosen
constraint(s).
+
+Contenthub Search API also provides specifying an ontology which carries
semantic information to make the search more semantic. How this external
ontology is exploited will be explained within the search engine documentations
below. Furthermore, Search API enables specifying constraints for the search
operation. The aim is to provide faceted search functionality through Java
interface based on the specified constraints.
+
+Search part of component is formed by several search engines that work
sequentially and contribute to the search results. Each search engine works
with a given search context. The initialization of the search context is
performed before the execution of any search engine. Each search engine makes
use of the information embedded in the search context and populates the context
with new results, such as resulting documents, related ontological resources,
new keywords etc...
+
+Currently, three search engines are active in search subcomponent:
+
+### Ontology Resource Search Engine
+
+This engine works when an additional ontology is specified at the beginning of
the search. A SPARQL query based on a LARQ index is executed on specified
ontology to find individuals and classes related with the keyword. When a class
is found, it is added to search context as a related class resource and then,
subclasses, superclasses and instances
+of all these classes are found and added to the search context.
+
+When an individual about keyword is found it is added as a related individual
resource to search context and it's classes are found. These classes are added
to the search context using the same methodology explained in the previos
paragraph.
+
+### Enhancement Search Engine
+
+This engine designed to work on enhancement graph which contains all
enhancements of content items submitted to the Contenthub.
+
+When a document is submitted to ContentHub, its content is enhanced
automatically by Enhancer component.
+In a single Clerezza graph, all the enhancements are kept together and this
graph is indexed with LARQ. The LARQ index is automatically updated when a new
enhancement is added.
+
+Enhancement Search Engine, executes a SPARQL query on enhancement graph to
find enhancements about the given keyword.
+When an enhancement is found, the document from which the enhancement was
obtained is added to search context as a related document resource.
+
+### Solr Search Engine
+
+Solr Search Engine is the engine that gives full-text and faceted search
capabilities to the Contenthub.
+
+Since every document is indexed to Solr (to the core created for Contenthub),
it is possible to do full-text
+search over documents' content and metadata. After the first search, all the
facet constraints of resulting documents will be available for faceted search.
When a facet constraint is chosen, resulting documents and facet constraints
are updated dynamically.
+
+Later on, related class and individual resources about the keyword, which are
found by Ontology Resource Search Engine, are searched over Solr using their
resource name.
+
+After all, document resources founded by Enhancement Search Engine is
examined. If there is a document whose field values does not match with facet
constraints, these document resources are removed from the search results.
+
+## Building and Launching ContentHub
+
+Since ContentHub is included in Full Launcher of Apache Stanbol
+it is built with Apache Stanbol and can be launched under
+Apache Stanbol Full Launcher. For detailed instructions to build and launch
Apache Stanbol see the README file through the following link:
+
+ http://svn.apache.org/repos/asf/incubator/stanbol/trunk/README.md
