Objectives

Develop a new importer that accepts rdf serialized in RDF/XML and/or in n3 in place of the old "legacy format".

Specs/Ideas

Declaring RDF Resources as Talia Sources

The importer should let the user declare in some way which subset of the RDF Resources (things with an URI) should be treated as Talia Sources by the application.

Handling Dataobjects

There should be a RDF property (or a little model, I guess we will need more than just a property) to link data objects to Sources. The imported should of course understand these and properly link DOs to Sources.

Declaring timeline properties

The RDF property and the RDF object (Literal, datatype?) to be used for building the timeline should be configurable at import time.

Declaring map properties

Same as below for the map.

Import from SPARQL endpoints

todo

Test datasets

BAMI dataset

This is the data used to build http://bami.cilea.it. Resources of type foaf:person, foaf:group, bami:Manif*, bami:Item, bami:Work*, bami:publicationWork, wgs84_pos:SpatialThing (and subclasses), should be treated as Talia Sources.

DBPedia Persondata

little (stupid and "flat") slice of dbpedia with person data in ntriples is attached. instances of foaf:person (rdf:resources with type foaf:person) should be treated as Talia sources.

Islamic Art demo dataset

Small (but well linked) dataset used for the Talia Islamic Art Demo. Enrich it with some data from DBpedia and/or geonames.

Serialization formats

It is straightforward to transform between one serialization format (most common ones are RDF/XML, n-triples, n3 and Turtle), most RDF parsers and libraries can do it. There are also many online tools that can parse and serialize in many formats (e.g. any23, rdfconvert, google for more).

This is a simple tutorial on how to do transformations using only unix command line tools (and the raptor rdf parser): tutorial on N3 and N-Triples

Sesame and Jena RDF APIs also read, write and convert into most serialization formats. In Ruby there are plugins of RDF.rb that make it capable of dealing with a number of different formats (it natively supports only ntriples).