calummackervoy
commented
3 years ago Hi! For info DjangoLDP now lives here: https://git.startinblox.com/djangoldp-packages/djangoldp
Linked Data on top of native Django objects
I think that it's useful that we build on top of Django's model class to provide an RDF view onto SQL data, overall I think our approach of using Django models is quite beginner-friendly. We realise that we need some mapping for fields (https://git.startinblox.com/djangoldp-packages/djangoldp/issues/352)
We're also looking at how to extend our support so that the ORM can be used with a Triplestore backend. Currently I'm considering if we could cut off the Django tree at the SQL level, and hit a wrapper to RDFLib instead, because Django's database backends are tightly coupled to SQL. It would be incredible to combine the power of Django and especially the ORM with the open world assumption provided by graphs
could be extended to support the mapping and namespace handling RDF_IO provides
Obviously this issue is a couple of years old.. are you still interested in this direction ?
RDF_IO handles the ORM -> graph configuration and RDF utilities like namespace handling and RDF ingest mechanisms.
pushing a Linked Data API to the user using these mapping is currently pushed off to other services - such as RDF4J or Marmotta where scalability and UI can be addressed,
A "native" capability exploiting Django's templating similar to https://github.com/RDFLib/pyLDAPI
would be useful
or an alternative approach for Linked Data on top of native Django objects (e.g. https://git.happy-dev.fr/startinblox/djangoldp-packages/djangoldp) could be extended to support the mapping and namespace handling RDF_IO provides.
any approach should be extensible to support multiple profiles - see https://www.w3.org/TR/2018/WD-dx-prof-conneg-20181218/