search

HTTP-APIs / hydra-python-agent

Python Hydra smart client and test console
MIT License
22 stars 42 forks source link

More interoperability with pyRDFlib #56

Open Mec-iS opened 6 years ago

Mec-iS commented 6 years ago

What we want is probably to use some parts from RDFlib with our Redis backend. Investigate.

shravandoda commented 5 years ago

RDFlib uses it's own implementation of graph and uses SPARQL for querying the graph. I don't know how would we integrate the two. Redisgraph is a general purpose graph database which uses open-cypher as the querying language. RDFlib has the advantage that it's designed to work with Linked-Data rather than being a general purpose graph database. But still the final decision will depend upon how the user wants to query the data. Moreover, implementing TPF is easier with RDFlib as it makes it easier to deal with triples.

Mec-iS commented 5 years ago

My first idea was to use RDFlib as intermediate representation for then storing the result into Redisgraph. We should first see if there are pros in doing something like that. As Hydra is a Linked Data technology, we probably want to make Redisgraph "a little less general" and more "specialized" in Linked Data by using some functionalities from RDFlib. In general:

server-side data --> client fetches -->
relevant sub-graph is temporary parsed in RDFlib --> data is updated in Redis

In this case the nodes/properties stored in Redisgraph may contain some additional flags or "headers" carried on from the RDFlib representation. For how to use RDFlib with Hydra see the original hydra-py client.