Updated to Redland 1.0.16, Rasqal 0.9.30 and Raptor 2.0.9.
Updated to Rasqal 0.9.29 and Raptor 2.0.8.
Added the CSV, TSV and Table result formats for bindings that were added in more recent Rasqals.
Also fixed a long term issue with validating of external data that caused 500 Errors occasionally.
Announced Triplr RDF Query service which also operates as a SPARQL endpoint.
This does not yet run Rasqal Subversion which has a more functional query engine. At present Rasqal 0.9.16 fails to handle UNION, some OPTIONALs, all nested graph patterns and a lot of datatype details that are fixed or improved in subversion.
Updated the Triplr installation to use the latest released Redland (1.0.9), Rasqal (0.9.16) and Raptor (1.4.19).
I have removed RSS serializing due to abuse - used in automated feed readers that ignore robots.txt (most of them) or other automated processing.
I have updated Triplr to use the latest subversion Raptor RDF JSON serializers to restore the JSON callback functionality.
There are two JSON serializers in Raptor that
are invoked in Triplr using the name
the default being the resource-centric output
The alternate triples-dump output can be generated by adding parameter
triples such as with
The JSON output can have a callback defined which
was available with Triplr when it was using the perl JSON code, of
wrapping the output in a function
foo(...) when you call
Triplr with a parameter
callback=foo such as with a call like
The JSON output is also altered to restore the
footer keys that provide some Triplr acknowledgment
on the data output. This was also provided in the service with the
Here is an example of using both parameters:
Banned a bot for trying to GRDDL all of last.fm through triplr. Use raptor yourself if you want to do that.
Well, Danny revealed it already but I've changed the JSON serializer from being created by perl's JSON library, to generated by the subversion version of Raptor 1.4.17 itself using a resource-centric approach RDF/JSON by Ian Davis of Talis.
The fancy callback args are gone for now until I ummm, think about it some more.
RDFa support has been removed.
I got a complaint from a certain
that he couldn't condone Triplr's URI design since it included
http:// in the URI path. Since everyone
uses only http (right?) I've just switched the format to
to make Triplr read the URI
Of course, I've put a redirection in from the old form URIs so that should make everyone else not even notice this, if they are paying attention to 301s.
I took an item off the todo list, so that
you can now specify an input format as well as an output format.
This was needed to specify RDFa input since it cannot be easily detected, and the GRDDL parser is the normal parser for (X)HTML formats. Use like
Triplr is using a pre-release of Raptor 1.4.16 so there are few other new things such as parsing TRiG plus 100% support for the GRDDL proposed recommendation which was out today - congrats!
Update: RDFa support was removed 2007-09-16.
Updated Triplr to use a pre-release of Redland 1.0.6 that can return the namespaces used in parsing, so that the serializer now only declares namespaces seen. The result is that the Turtle, RDF/XML and other namespace-abbreviated results are more compact and readable.
Seems like wget(1)
strips out the second // if you try
wget http://triplr.org/rdf/http://example.org/, Triplr
receives the URL
so I've recognized that and added an HTTP 301 Permanent Redirect to the
right URL. It's my favourite type of Redirect. What's yours?
Wish I remembered to tell everyone especially Rev Dan before he wrote
Geonames, rdf, triplr, json, Yahoo! Pipes and the Semantic Web, oh my!
that the JSON writer has an
callback=name parameter. I guess
he still had fun! His first post using the JSON was
Geonames machine tags, Triplr.org and JSON, oh my!
Now using a real JSON writer, the CPAN module
JSON by Makamaka Hannyaharamitu
There can be optional Matrix URI parameters given to Triplr indicated by a ';' after the output name like
Here the foo parameter has the value bar and the
blah parameter has no explicit value, and defaults to '1'.
The JSON output uses a
pretty parameter. If set to non-0,
the JSON output is pretty/indented/long, otherwise it is ugly/compact.
HTML output as a table of triples
JSON output based on Serializing SPARQL Query Results in JSON