Hi Robert, Thanks for your detailed response. I've been successful in loading rdf into virtuoso and the sparql features seem to work great. They are publicly available at http://dbtune.org/cmn/sparql
In fact, with the help of Kingsley, Hugh, and the list, i've managed to get dereferencable uris straight out of the sparql store using virtuoso's url re-write tools. So my basic goal is to follow the dbpedia-style linked data method using the resource-data-page style. For example, http://dbtune.org/cmn/resource/WAGN derefs to http://dbtune.org/cmn/data/WAGN if you accept rdf.xml. then http://dbtune.org/cmn/data/WAGN makes a describe sparql query to output the appropriate rdf describing what is known about http://dbtune.org/resource/WAGN. All this works lovely. Now i want to server HTML versions for humans at http://dbtune.org/page/WAGN. I thought i understood using /about/html was an appropriate way to do this. however it would not work for me. the best i could get was info about xsd style sheets. if i understood Kingsley's email, this is because the rdf returned by http://dbtune.org/data/<> doesn't include foaf:topicOf properties. cheers, kurt j On Wed, Feb 11, 2009 at 4:38 PM, Robert Jones <rjo...@openlinksw.com> wrote: > Kurt J wrote: >> >> Ah I think I understand. So using /about/ is really not appropriate >> for the sort-of on-demand generation of rdf that i'm doing. in that >> http://dbtune.org/cmn/data/<composer> redirects to a sparql query for >> describing <composer> and there wouldn't be a foaf:primaryTopic >> property in that result. so to use /about/ i would want to generate >> static rdf files (ie http://dbtune.org/cmn/data/WAGN.rdf) and host >> them? is this what is done on dbpedia? is there a fancy programmatic >> way of doing this from the rdf store? >> > > Kurt, > > I'm not entirely sure about what you are trying to achieve so I will run > down the typical check-list for getting /about/ to work on a simple RDF > view. > > Assuming that you have installed a recent conductor vad, and rdf_mappers vad > you will automatically have an administration interface (conductor) that > will allow you create simple RDF views directly from DB tables in a few > clicks, and the /about/ virtual directory + required rewrite/redirect rules > courtesy of rdf_mappers. > > If the /about/ virtual directory has not been altered then it will > automatically retrieve and display RDF data from remote or local stores, > hence we now need to ensure that local Virtuoso returns suitable RDF. > > The simplest way to see how to create an RDF delivering endpoint is to allow > Virtuoso to do it for you from a database table. From the conductor go to > "Database" then "RDF Schema Object". Find a table that you would like from > RDF data from and click "Generate Single Mapping". You will be to a page > containing a SQL script to be run, note the following: > > 1. The necessary grants (as you have seen before) > 2. The IRI class, which will contain the URI for accessing the data > 3. The create graph statement which details the mapping from SQL table to > RDF. > 4. The 4 create rewrite rules which instruct Virtuoso how to handle certain > clients/methods for accessing the to-be-created endpoint. > 5. the virtual directory creation (vhost_define) using the rules from 4. > > Once you hit the execute button and wait for confirmation that x number of > RDF manipulations were done you should be able to point your web browser to > a URL that resembles the URI from 2. above. This should automatically > redirect you to the /about/ service for viewing the RDF in an HTML format. > > Please let me know how this compares with what you have already achieved and > are trying to achieve and we will see if we can overcome your current > hurdles. > > Regards, Rob. > -- > Software Development > OpenLink Software > >
