Sparql query JSON error from BNCF endpoint

Question

I'm trying to retrieve results from the BNCF at this endpoint.

My query (with "ab" as example) is:

PREFIX skos: <http://www.w3.org/2004/02/skos/core#>
SELECT DISTINCT ?source ?label ?content
                WHERE {
                    ?source a skos:Concept;
                        skos:prefLabel ?label; 
                        skos:scopeNote ?content.
                FILTER regex(str(?label), "ab", "i")
            }

The query is correct in fact if you try to run it works. But when I try to get the results from my python this is the error:

SyntaxError: JSON Parse error: Unexpected EOF

This is my python code:

__3store = "http://digitale.bncf.firenze.sbn.it/openrdf-workbench/repositories/NS_03_2014/query"
sparql = SPARQLUpdateStore(queryEndpoint=__3store)
sparql.setReturnFormat(JSON)
results = sparql.query(query_rdf).convert()
print json.dumps(result, separators=(',',':'))

I tried the code above according to this answer, before my code was like this:

__3store = "http://digitale.bncf.firenze.sbn.it/openrdf-workbench/repositories/NS_03_2014/query"
sparql = SPARQLWrapper(__3store,returnFormat="json")
sparql.setQuery(query_rdf)
result = sparql.query().convert() 
print json.dumps(result, separators=(',',':'))

but both throw the same error.

Does anyone know how to fix it? Thanks

EDIT:

This is python code, hope it is enough to understand

import sys
sys.path.append ('cgi/lib')
import rdflib
from rdflib.plugins.stores.sparqlstore import SPARQLUpdateStore, SPARQLStore
import json
from SPARQLWrapper import SPARQLWrapper, JSON

#MAIN
print "Content-type: application/json"
print
prefix_SKOS =       "prefix skos:      <http://www.w3.org/2004/02/skos/core#>"
crlf = "\n"
query_rdf = ""
query_rdf += prefix_SKOS + crlf
query_rdf += '''
            SELECT DISTINCT ?source ?title ?content
                WHERE {
                    ?source a skos:Concept;
                        skos:prefLabel ?title; 
                        skos:scopeNote ?content.
                FILTER regex(str(?title), "ab", "i")
            }

        '''
__3store = "http://digitale.bncf.firenze.sbn.it/openrdf-workbench/repositories/NS_03_2014/query"
sparql = SPARQLWrapper(__3store,returnFormat="json")
sparql.setQuery(query_rdf)
result = sparql.query().convert() 

print result

Running this in Python shell returns:

Content-type: application/json


Warning (from warnings module):
  File "/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/site-packages/SPARQLWrapper-1.6.4-py2.7.egg/SPARQLWrapper/Wrapper.py", line 689
RuntimeWarning: Format requested was JSON, but XML (application/sparql-results+xml;charset=UTF-8) has been returned by the endpoint
<xml.dom.minidom.Document instance at 0x105add710>

So I think the result is always an XML also if I specificied Json as a return format.

Answer 1

There are a couple of problems playing together here:

First, you should only use SPARQLUpdateStore from rdflib if you want to access a SPARQL store via rdflib's Graph interface (e.g., you can add triples, you can iterate over them, etc.). If you want to write a SPARQL query yourself you should use SPARQLWrapper.

Second, if you ask SPARQLWrapper to return JSON, what it does is actually ask the server for a couple of mime types that are most common and standardized for what we just call "json", as shown here and here:

_SPARQL_JSON = ["application/sparql-results+json", "text/javascript", "application/json"]

It seems as if your sever does understand application/sparql-results+json, but not a combined "give me any of these mime-types header" as rdflib compiles it for maximum interoperability (so your server essentially doesn't fully support HTTP Accept Headers):

curl -i -G -H 'Accept: application/sparql-results+json' --data-urlencode 'query=PREFIX skos: 
<http://www.w3.org/2004/02/skos/core#>
SELECT DISTINCT ?source ?label ?content
WHERE {
 ?source a skos:Concept;
 skos:prefLabel ?label;
 skos:scopeNote ?content.
 FILTER regex(str(?label), "ab", "i")
}' http://digitale.bncf.firenze.sbn.it/openrdf-workbench/repositories/NS_03_2014/query

will return:

HTTP/1.1 200 OK
Date: Mon, 18 May 2015 13:13:45 GMT
Server: Apache/2.2.17 (Unix) PHP/5.3.6 mod_jk/1.2.31
...
Content-Type: application/sparql-results+json;charset=UTF-8

{
  "head" : {
    "vars" : [ ],
    "vars" : [ "source", "label", "content" ],
    "link" : [ "info" ]
  },
  "results" : {
    "bindings" : [ {
      "content" : {
        "type" : "literal",
        "value" : "Il lasciare ingiustificatamente qualcuno o qualcosa di cui si è responsabili"
      },
      "source" : {
        "type" : "uri",
        "value" : "http://purl.org/bncf/tid/12445"
      },
      "label" : {
        "xml:lang" : "it",
        "type" : "literal",
        "value" : "Abbandono"
      }
    },
...

so everything is ok, but if we ask for the combined, more interoperable mime types:

curl -i -G -H 'Accept: application/sparql-results+json,text/javascript,application/json' --data-urlencode 'query=PREFIX skos: <http://www.w3.org/2004/02/skos/core#>
SELECT DISTINCT ?source ?label ?content
WHERE {
 ?source a skos:Concept;
 skos:prefLabel ?label;
 skos:scopeNote ?content.
 FILTER regex(str(?label), "ab", "i")
}' http://digitale.bncf.firenze.sbn.it/openrdf-workbench/repositories/NS_03_2014/query

we get an xml result:

HTTP/1.1 200 OK
Server: Apache/2.2.17 (Unix) PHP/5.3.6 mod_jk/1.2.31
...
Content-Type: application/sparql-results+xml;charset=UTF-8

<?xml version='1.0' encoding='UTF-8'?>
...

So long story short: it's a bug in the server you're using. The following is a nasty workaround (it seems SPARQLWrapper doesn't just allow us to manually set the headers, but unconditionally overrides them in _createRequest), but it works:

In [1]: import SPARQLWrapper as sw

In [2]: sparql = sw.SPARQLWrapper("http://digitale.bncf.firenze.sbn.it/openrdf-workbench/repositories/NS_03_2014/query")

In [3]: sparql.setReturnFormat(sw.JSON)

In [4]: sparql.setQuery('''                                                                                                     PREFIX skos: <http://www.w3.org/2004/02/skos/core#>
SELECT DISTINCT ?source ?label ?content
                WHERE {
                    ?source a skos:Concept;
                        skos:prefLabel ?label;
                        skos:scopeNote ?content.
                FILTER regex(str(?label), "ab", "i")
            }
''')

In [5]: request = sparql._createRequest()

In [6]: request.add_header('Accept', 'application/sparql-results+json')

In [7]: from urllib2 import urlopen

In [8]: response = urlopen(request)

In [9]: res = sw.Wrapper.QueryResult((response, sparql.returnFormat))

In [10]: result = res.convert()

In [11]: result
Out[11]:
{u'head': {u'link': [u'info'], u'vars': [u'source', u'label', u'content']},
 u'results': {u'bindings': [{u'content': {u'type': u'literal',
     u'value': u'Il lasciare ingiustificatamente qualcuno o qualcosa di cui si \xe8 responsabili'},
    u'label': {u'type': u'literal',
     u'value': u'Abbandono',
     u'xml:lang': u'it'},
    u'source': {u'type': u'uri', u'value': u'http://purl.org/bncf/tid/12445'}},
   ...