It is time to add some configuration to the gdata server. The server is based on a Entry / Feed Object representation model using "Extension Profiles" to describe entry / feed elements. As atom feed allows foreign namespaces as elements users should be able to extend the BaseFeed / BaseEntry classes and create their own extension profil. The Feed / Entry type must be configured to associate the corresponding Extension Profile at runtime. All possible feed / entry types will be registered at servers startup. I also thought about a feed / entry type based indexing configuration which enables users to define custom search fiedls per feed / entry type. This could look like the following: <feed class="org.apache.BaseFeed " extensionProfil=" org.apache.ExtensionProfile"> <entry class="class="org.apache.BaseFeed" /> <field path="/feed/entry/title" analyzer="SimpleAnalyzer" store="YES" index="untokenized"/> <field path="/feed/entry/summary" analyzer="SimpleAnalyzer" store="no" index="tokenized"/> <field path="/feed/entry/content" analyzer="SimpleAnalyzer" store="no" index="tokenized"/> </feed>
The Path expression is a XPath expression following the attributes of the Entry / Feed Implementation so MyBean.getTitle() would be /Mybean/title (mights use JXPath for that) . I will also add the components configuration add to this file. All the configuration (maybe some JMX stuff later) can be located in this single xml file. I guess loading the file using commons-digester is a good idea. Some Solr guys out there having some exerience with configuring indexers out there (I guess sol *Yonik* :) regards simon
