I want to develop a triple store for Jena. How can I integrate it to Jena?
Now I have these references:
1. Jena, TDB,SDB source code
2. The SPI portion of Jena api document
3. Jena document, specifically I think it is about assemblor 4. hbase-rdf project source code
5. JenaSesameModel source code
As my understanding, I think I should:
1.implement some of the interface in SPI package:
Is that right? Are there any other information? Thank you.
asked 11 Jan '13, 14:11
If you just want triples, you only need to implement Graph and there is GraphBase that provides defaults for most things.
If you want quads/named graph, implement DatasetGraph (many partial implementations - choose the one for a base class that fits your style, e.g. separate graphs vs quads).
Everything works at that point.
If you want performance for SPARQL, you also need to implement some of the operations in OpExecutor to exploit the internal abilities of your store.
An assemblers is a bonus - it would allow you to plug into Fuseki.
answered 11 Jan '13, 15:56