As an atlas builder, I would like to export Nunaliit documents in the RDF(s) model. Two formats would be useful: I) Turtle because it is mature, broadly implemented, is easy for a trained human to read and understand, and a W3C standard (https://www.w3.org/TR/turtle/); ii) JSON-LD because it is now widely used by web developers and other users and has also recently become a standard (https://www.w3.org/TR/json-ld11/). There are many tools to convert from Turtle to JSON-LD and other RDF serializations (formats), so the support of JSON-LD is initially not critical.
The atlas builder should be able to qualify attributes in the schema with a namespace URI and a short form that identifies the vocabulary(ies) being used for a particular data value. For example, short form = rdf: namespace URI = http://www.w3.org/1999/02/22-rdf-syntax-ns#. These prefixes and namespaces should be used in the export files. If no namespace is asserted, then a unique, default prefix:namespace should be generated and assigned using some form of the atlas URL, for example (mpde: HTTP://mpde.gcrc.carleton.ca/ontology/). This will help to avoid term collisions when the data is used in broader contexts.
The atlas builder should be able to customize the schema elements that are exported. Ideally, the atlas developer could use the Nunaliit export tool and the export functionality in the data modification tool to execute the export. Additionally, an API that supports export would have very, very high value.
A specific use case for this functionality is to export the graph in the MPDE atlas for import into a triple store such as Fuseki, or visualization in a tool such as http://vowl.visualdataweb.org/webvowl.html.
[x] Support more methods other than just schema (layer? add the filtering stuff back?)
[x] Modify the export page to also support this endpoint?
[ ] Handle how the user defined semantic annotations should be created, parsed, and written out as (presently, it defaults to the atlas-specific ontology/namespace and uses the properties from export.json to choose the values to be written out)
As an atlas builder, I would like to export Nunaliit documents in the RDF(s) model. Two formats would be useful: I) Turtle because it is mature, broadly implemented, is easy for a trained human to read and understand, and a W3C standard (https://www.w3.org/TR/turtle/); ii) JSON-LD because it is now widely used by web developers and other users and has also recently become a standard (https://www.w3.org/TR/json-ld11/). There are many tools to convert from Turtle to JSON-LD and other RDF serializations (formats), so the support of JSON-LD is initially not critical.
The atlas builder should be able to qualify attributes in the schema with a namespace URI and a short form that identifies the vocabulary(ies) being used for a particular data value. For example, short form = rdf: namespace URI = http://www.w3.org/1999/02/22-rdf-syntax-ns#. These prefixes and namespaces should be used in the export files. If no namespace is asserted, then a unique, default prefix:namespace should be generated and assigned using some form of the atlas URL, for example (mpde: HTTP://mpde.gcrc.carleton.ca/ontology/). This will help to avoid term collisions when the data is used in broader contexts.
The atlas builder should be able to customize the schema elements that are exported. Ideally, the atlas developer could use the Nunaliit export tool and the export functionality in the data modification tool to execute the export. Additionally, an API that supports export would have very, very high value.
A specific use case for this functionality is to export the graph in the MPDE atlas for import into a triple store such as Fuseki, or visualization in a tool such as http://vowl.visualdataweb.org/webvowl.html.