triplestore with revisions

Question

I need to store RDF triples, while keeping track of modifications. Is there anything out there for this service? Alternatively, how would you keep track of revision while using a context-aware triplestore ?

Edit: please note that I don't need only the changeset. I also need to see the triples at any time in the past.

Halford

Do you need branches like git? – Hemianopsia

Copernic does exactly that github.com/amirouche/copernic it is based on #55143985 – Hemianopsia

Answer 1

Take a look at this Semantic Overflow question on the same subject - semanticoverflow.com: How to implement semantic data versioning? - there's a variety of relevant links there

Generally versioning in Triple stores relies on either using named graphs and storing meta-data about each named graph in a separate graph or the Triple Store having some in-built mechanism for versioning such as the Talis platform (though that's not a context aware store so not suitable for you)

Meraree

Answer 2

I tried to address this sort of thing by storing additional triples describing who said what, when. But you need reification, named graphs, or quads which can ruin decidability.

Rosemari

Answer 3

in addition to the other answers, also take a look at the Lensfield system (research software at the moment) developed by Jim Downing et al.. Lensfield will not only keep track of different versions of your data, but also of any scripts/transformation instructions you use to process your data and change your graph. In short, it's a mixture of version control, maven and make for semantic data. Maybe that helps.

Bigname

Yes, that was my plan, if PMR granted me the help I asked. – Halford

Answer 4

For the problem of versioning multiple RDF graphs, we have developed the Quit Store (Quads in Git). It provides:

• Context aware versioning of a whole RDF data set (multiple named graphs in one repositroy)
• An access interface to any time in the past of the repository by providing a SPARQL endpoint for each individual version of the data set.

We have also published some papers about various aspects of this topic. You can find them all here: http://aksw.org/Projects/Quit

Wysocki