Closed nichtich closed 10 years ago
An option is to look at RDF dumps in the wild and try to import those and work out from these experiences to a more general method.
E.g. VIAF : http://viaf.org/viaf/data/ (one triple per line) ZDB : http://datahub.io/dataset/zdb/resource/7605c274-49c6-4e95-af79-9d75b4a11630 Europeana: http://datahub.io/dataset/europeana-lod
A triple importer could contain an internal triple store to accumulate all incomming triples. At the end of the stream you can then generate JSON from the internal cache.
moved to issue #5
Importing single triples would be very limited, but allow for Stream-based import (one-pass). Importing full graphs requires a clear encoding in JSON (e.g. JSON-LD, aREF, https://metacpan.org/pod/RDF::TrineX::Serializer::MockTurtleSoup ...). SPARQL could be used to get a list of items with selected fields, but SPARQL is very verbose (could be abbreviated).