Support LDF endpoint?

[dpriskorn]

With this we could encourage people writing queries to offload to LDF when possible and do more computing locally now that WDQS expensive SPARQL endpoint receives >100.000 queries a second.

see https://github.com/pchampin/hydra-py for a Triple Patterns Fragments client example code:

"""
Use TPFStore and performs a SPARQL query on top of it.
Note that this is very (very!) slow as soon as the query becomes slightly complex... :-/
"""
import logging
from rdflib import Graph
import sys
logging.basicConfig(level=logging.INFO)

import hydra.tpf # required to register TPFStore plugin

URL = 'http://data.linkeddatafragments.org/dbpedia2014'
if len(sys.argv) > 1:
    URL = sys.argv[1]

g = Graph("TPFStore")
g.open(URL)

QUERY = """
    PREFIX rdfs: <http://www.w3.org/2000/01/rdf-schema#>
    PREFIX dbr: <http://dbpedia.org/resource/>
    PREFIX dbo: <http://dbpedia.org/ontology/>

    SELECT * {
        ?p a dbo:Person; dbo:birthPlace ?bp .
    }
    LIMIT 10
"""

print len(g)

results = g.query(QUERY)
for i in results:
    print i

[LeMyst]

I never used LDF or TPF. That seems interesting and, if I understand correctly, that can solve the timeout when querying the SPARQL server.

I'll not put this at the top of the list, for the moment, but that seems promising.

[dpriskorn]

Exactly, no timeouts but often a truck load of requests because we get only 100 results at a time. Complex queries take more time.