def get_collection(self, c_id=None, filter=[]):
# todo: ASK and check if collection exists
if c_id is not None:
result = self.sparql.select(self.marmotta.ldp(encoder.encode(c_id)))
graph = result.toDataset().graph(self.marmotta.ldp(encoder.encode(c_id)))
contents = self.RDA.graph_to_object(graph)
if len(contents) is 0:
raise NotFoundError()
else:
binds = [Bind(Variable('s'), self.marmotta.ldp()), Bind(Variable('p'), LDP.ns.contains)]
filters = [
Filter(s=Variable('o'),p=URIRef(v.get('label')) if k.type is RDA.Collection else LDP.contains/LDP.contains/URIRef(v.get('label')),
o=v.get('rdf', Literal)(k.value)) for v,k in [
(access(self.RDA.dictionary.get(f.type).inverted, f.path), f) for f in filter
]
]
result = self.sparql.list(binds, filters)
collections = [dct[Variable('o')] for dct in result.bindings]
contents = []
if len(collections):
result = self.sparql.select(collections)
ds = result.toDataset()
graphs = [ds.graph(collection) for collection in collections]
for graph in graphs:
contents += self.RDA.graph_to_object(graph)
return contents
def _enrich(g, database, schema, graph):
ns_tbox = _generate_tbox_namespace(graph)
# update namespaces
ns_abox = default_namespace_of(graph)[0]
_update_namespaces(g.namespace_manager,\
('rws.' + database.lower(), ns_abox),
('geo', Namespace("http://www.opengis.net/ont/geosparql#")))
for entry in schema.from_database(database):
source_def = entry['source']
target_def = entry['target']
source_class = _classname_from_def(database, source_def[0])
attributes = [definition['property'] for definition in source_def]
properties = [ property_from_mapping(ns_tbox, source_class, attr)\
for attr in attributes ]
# select target references
q = _generate_query(URIRef(ns_tbox + source_class), attributes,
properties)
for binding in graph.query(q).bindings:
source_uri = binding[Variable('source_id')]
if source_uri is None or source_uri == "":
continue
values = { '['+attr+']': binding[Variable(attr)].toPython() for attr in attributes\
if Variable(attr) in binding.keys() }
g += _generate_branch(schema, ns_abox, target_def, source_uri,
values)
def rdf_cksum(self, gr):
'''
Generate a checksum for a graph.
What this method does is ordering the graph by subject, predicate,
object, then creating a pickle string and a checksum of it.
N.B. The context of the triples is ignored, so isomorphic graphs would
have the same checksum regardless of the context(s) they are found in.
@TODO This can be later reworked to use a custom hashing algorithm.
:param rdflib.Graph: gr The graph to be hashed.
:rtype: str
:return: SHA1 checksum.
'''
# Remove the messageDigest property, which very likely reflects the
# previous state of the resource.
gr.remove((Variable('s'), nsc['premis'].messageDigest, Variable('o')))
ord_gr = sorted(list(gr), key=lambda x: (x[0], x[1], x[2]))
hash = sha1(pickle.dumps(ord_gr)).hexdigest()
return hash
def test_24_result_bindings(self):
b = list(self.graph.query('SELECT (1 as ?x) {}'))
assert b[0][0] == Literal(1)
assert b[0]['x'] == Literal(1)
assert b[0][Variable('x')] == Literal(1)
b = self.graph.query('SELECT (1 as ?x) {}').bindings
assert b[0][Variable('x')] == Literal(1)
assert b[0]['x'] == Literal(1)
def _construct(self, sparql: str, **kwargs) -> Graph:
result = self._select(sparql, **kwargs)
S = Variable("s")
P = Variable("p")
O = Variable("o")
neo = Graph()
for fact in result.bindings:
neo.add((fact[S], fact[P], fact[O]))
return neo
def set_member(self, c_id, m_obj):
if isinstance(m_obj, Model):
m_obj = [m_obj]
elif not isinstance(m_obj, list):
raise ParseError()
c_ldp_id = self.marmotta.ldp(encoder.encode(c_id))
collection = self.get_collection(c_id).pop() # 404 if collection not found
if len(set([m.id for m in m_obj])) is not len(m_obj):
raise ForbiddenError()
if not collection.capabilities.membershipIsMutable:
raise ForbiddenError()
if collection.capabilities.restrictedToType:
for m in m_obj:
if not(hasattr(m,"datatype") and m.datatype in collection.capabilities.restrictedToType):
raise ForbiddenError()
if collection.capabilities.maxLength >= 0:
size = self.sparql.size(c_ldp_id).bindings.pop().get(Variable('size'))
if int(size) > collection.capabilities.maxLength-len(m_obj):
raise ForbiddenError()#"Operation forbidden. Collection of maximum size {} is full.".format(collection.capabilities.maxLength))
ds = Dataset()
ldp = ds.graph(identifier=LDP.ns)
for m in m_obj:
m_id = self.marmotta.ldp(encoder.encode(c_id)+"/member/"+encoder.encode(m.id))
member = ds.graph(identifier=m_id)
member += self.RDA.object_to_graph(member.identifier,m)
ldp += LDP.add_contains(c_ldp_id+"/member",m_id,False)
res = self.sparql.insert(ds)
if res.status_code is not 200:
raise DBError()
return m_obj
def __init__(self,graphNodeList=None):
self.propVals = []
if graphNodeList:
self._list = graphNodeList
self.identifier = Variable(BNode())
else:
self._list = graphNodeList and graphNodeList or []
self.identifier = URIRef('http://www.w3.org/1999/02/22-rdf-syntax-ns#nil')
def _query_bindings(triple, g=None, to_n3=True):
(s, p, o) = triple
if isinstance(g, Graph):
g = g.identifier
if s is None: s = Variable("S")
if p is None: p = Variable("P")
if o is None: o = Variable("O")
if g is None: g = Variable("G")
if isinstance(s, BNode):
s = _bnode_to_nodeid(s)
if isinstance(p, BNode):
p = _bnode_to_nodeid(p)
if isinstance(o, BNode):
o = _bnode_to_nodeid(o)
if isinstance(g, BNode):
g = _bnode_to_nodeid(g)
if to_n3:
return dict(zip("SPOG", [x.n3() for x in (s, p, o, g)]))
else:
return dict(zip("SPOG", [x for x in (s, p, o, g)]))
def frame2triple(self,
root: dict,
parent: Variable,
triples: list,
extra: list,
pref=None) -> None:
"""
convert a json frame to s-p-o triples recursively
:param root: json frame
:param parent: subject for now
:param triples: a list to store the results
:param extra: indeces of extra triples
"""
for k, v in root.items():
if k == '@context':
continue
p = self.to_node(k)
if isinstance(v, str):
o = self.to_node(v)
else:
if (parent.toPython(), k) in self.exist_triples:
o_str = self.exist_triples[(parent.n3(), k)]
if o_str[0] in ['$', '?']:
o = Variable(o_str)
elif o_str[0] in ['"', "'"]:
o = Literal(o_str.strip('"').strip("'"))
elif o_str[0] in ['<']:
o = URIRef(o_str)
else:
o = Literal(o_str.strip('"').strip("'"))
else:
o = Variable('var%d' % self.current_naming)
self.current_naming += 1
extra.append(len(triples))
if not isinstance(p, Literal):
triples.append(pref + [[parent, p, o]])
else:
print('failed to find context of %s.' % p)
if isinstance(v, dict) and len(v):
self.frame2triple(v, o, triples, extra,
pref + [[parent, p, o]])
def test_14_initBindings(self):
TST = Namespace('http://example.com/ns/')
self.graph.add((TST.a, TST.b, TST.c))
self.graph.add((TST.d, TST.e, TST.f))
result = self.graph.query(
"SELECT * { ?s ?p ?o }",
initBindings={
"p": TST.b,
Variable("o"): TST.c,
},
)
assert result.type == "SELECT", result.type
assert len(result) == 1
def f(bindings):
if unBound:
# @@note, param must be reassigned to avoid tricky issues of scope
# see: http://docs.python.org/ref/naming.html
_param = isinstance(param, Variable) and param or Variable(
param[1:])
val = bindings[_param]
if isinstance(val, Literal):
return getLiteralValue(val)
else:
return val
else:
return value
def _sparql_select(self, q, cursor):
log.debug("_sparql_select")
results = cursor.execute(q.encode("utf-8"))
vars = [Variable(col[0]) for col in results.description]
var_dict = VirtuosoResultRow.prepare_var_dict(vars)
def f():
for r in results:
try:
yield VirtuosoResultRow([resolve(cursor, x) for x in r],
var_dict)
except Exception, e:
log.debug("skip row, because of %s", e)
pass
def get_member(self, c_id, m_id=None, filter=[]):
# todo: ASK and check if member exists
if m_id is not None:
if not isinstance(m_id, list):
m_id = [m_id]
members = [self.marmotta.ldp(encoder.encode(c_id)+"/member/"+encoder.encode(id)) for id in m_id]
# ds = self.sparql.select(ids).toDataset()
# contents = [self.RDA.graph_to_member(ds.graph(id)) for id in ids]
else:
id = self.marmotta.ldp(encoder.encode(c_id))
if not self.sparql.ask(id, self.RDA.ns.Collection).askAnswer:
raise NotFoundError
lst = self.sparql.list([Bind(Variable('s'), self.marmotta.ldp(encoder.encode(c_id)+"/member")), Bind(Variable('p'), LDP.ns.contains)])
members = [dct[Variable('o')] for dct in lst.bindings]
contents=[]
if len(members):
dataset = self.sparql.select(members).toDataset()
graphs = [dataset.graph(member) for member in members]
for graph in graphs:
contents += self.RDA.graph_to_object(graph)
if m_id is not None and len(contents) is 0:
raise NotFoundError()
return contents
def testToldBNode(self):
for s, p, o in self.graph.triples((None, RDF.type, None)):
pass
query = """SELECT ?obj WHERE { %s ?prop ?obj }""" % s.n3()
print query
rt = self.graph.query(query, DEBUG=debug)
self.failUnless(
len(rt) == 1,
"BGP should only match the 'told' BNode by name (result set size: %s)"
% len(rt))
bindings = {Variable('subj'): s}
query = """SELECT ?obj WHERE { ?subj ?prop ?obj }"""
print query
rt = self.graph.query(query, initBindings=bindings, DEBUG=debug)
self.failUnless(
len(rt) == 1,
"BGP should only match the 'told' BNode by name (result set size: %s, BNode: %s)"
% (len(rt), s.n3()))
def _select(self, sparql: str, **kwargs) -> dict:
that = self._wrapper()
that.setQuery(sparql)
that.setReturnFormat(JSON)
json_result = that.queryAndConvert()
res = {}
res["type_"] = "SELECT"
res["vars_"] = [Variable(v) for v in json_result["head"]["vars"]]
column = OrderedDict()
bindings = []
for json_row in json_result["results"]["bindings"]:
rdf_row = {}
for variable in res["vars_"]:
if str(variable) in json_row:
rdf_row[variable] = self.jsonToNode(
json_row[str(variable)])
else:
rdf_row[variable] = None
bindings.append(rdf_row)
res["bindings"] = bindings
return SPARQLResult(res)
def _sparql_select(self, q, cursor, must_close):
log.debug("_sparql_select")
results = cursor.execute(q)
vars = [Variable(col[0]) for col in results.description]
var_dict = VirtuosoResultRow.prepare_var_dict(vars)
def f():
try:
for r in results:
try:
yield VirtuosoResultRow(
[resolve(cursor, x) for x in r], var_dict)
except Exception as e:
log.debug("skip row, because of %s", e)
pass
finally:
if must_close:
cursor.close()
e = EagerIterator(f())
e.vars = vars
e.selectionF = e.vars
return e
def test_rdflib_mysql_test(self):
"""
test taken from rdflib/test/test_mysql.py
"""
implies = URIRef("http://www.w3.org/2000/10/swap/log#implies")
testN3="""
@prefix rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#> .
@prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> .
@prefix : <http://test/> .
{:a :b :c;a :foo} => {:a :d :c,?y}.
_:foo a rdfs:Class.
:a :d :c."""
#Thorough test suite for formula-aware store
g = self.rdflib_graph
g.parse(data=testN3, format="n3")
#print g.store
for s,p,o in g.triples((None,implies,None)):
formulaA = s
formulaB = o
self.assertTrue(type(formulaA)==QuotedGraph and type(formulaB)==QuotedGraph)
a = URIRef('http://test/a')
b = URIRef('http://test/b')
c = URIRef('http://test/c')
d = URIRef('http://test/d')
v = Variable('y')
universe = ConjunctiveGraph(g.store)
#test formula as terms
self.assertTrue(len(list(universe.triples((formulaA,implies,formulaB))))==1)
#test variable as term and variable roundtrip
self.assertTrue(len(list(formulaB.triples((None,None,v))))==1)
for s,p,o in formulaB.triples((None,d,None)):
if o != c:
self.assertTrue(isinstance(o,Variable))
self.assertTrue(o == v)
s = list(universe.subjects(RDF.type, RDFS.Class))[0]
self.assertTrue(isinstance(s,BNode))
self.assertTrue( len(list(universe.triples((None,implies,None)))) == 1)
self.assertTrue( len(list(universe.triples((None,RDF.type,None)))) ==1)
self.assertTrue( len(list(formulaA.triples((None,RDF.type,None))))==1)
self.assertTrue( len(list(formulaA.triples((None,None,None))))==2)
self.assertTrue( len(list(formulaB.triples((None,None,None))))==2)
self.assertTrue( len(list(universe.triples((None,None,None))))==3)
self.assertTrue( len(list(formulaB.triples((None,URIRef('http://test/d'),None))))==2)
self.assertTrue( len(list(universe.triples((None,URIRef('http://test/d'),None))))==1)
#context tests
#test contexts with triple argument
self.assertTrue( len(list(universe.contexts((a,d,c))))==1)
#Remove test cases
universe.remove((None,implies,None))
self.assertTrue( len(list(universe.triples((None,implies,None))))==0)
self.assertTrue( len(list(formulaA.triples((None,None,None))))==2)
self.assertTrue( len(list(formulaB.triples((None,None,None))))==2)
formulaA.remove((None,b,None))
self.assertTrue( len(list(formulaA.triples((None,None,None))))==1)
formulaA.remove((None,RDF.type,None))
self.assertTrue( len(list(formulaA.triples((None,None,None))))==0)
universe.remove((None,RDF.type,RDFS.Class))
#remove_context tests
universe.remove_context(formulaB)
self.assertTrue( len(list(universe.triples((None,RDF.type,None))))==0)
self.assertTrue( len(universe)==1)
self.assertTrue( len(list(formulaB.triples((None,None,None))))==0)
universe.remove((None,None,None))
self.assertTrue( len(universe)==0)
"illegal argument, pattern must be a tuple of 3 or 4 element, got %s"
% len(tupl))
if len(tupl) == 3:
(s, p, o) = tupl
f = None
else:
(s, p, o, f) = tupl
final = []
for c in (s, p, o):
if isinstance(c, Variable):
if not c in self.unbounds:
self.unbounds.append(c)
final.append(c)
elif isinstance(c, BNode):
#Do nothing - BNode name management is handled by SPARQL parser
final.append(c)
else:
final.append(_createResource(c))
final.append(f)
return tuple(final)
def fetchTerminalExpression(self):
yield self
if __name__ == '__main__':
v1 = Variable("a")
g = BasicGraphPattern([("a", "?b", 24), ("?r", "?c", 12345),
(v1, "?c", 3333)])
print g
def ask_member(self, c_id, m_id):
if not isinstance(m_id, list):
m_id = [m_id]
ids = [self.marmotta.ldp(encoder.encode(c_id)+"/member/"+encoder.encode(m)) for m in m_id]
result = self.sparql.find(ids,self.RDA.ns.Member)
return float(result.bindings.pop().get(Variable('size')))/len(m_id)
import pyparsing
match_definitions = [
(parser.BaseDecl,
[
('BASE <s:ex>', URIRef('s:ex')),
('BASE <>', URIRef('')),
]),
(parser.PrefixDecl,
[
('PREFIX s: <s:ex>', components.PrefixDeclaration(
's:', 's:ex')),
]),
(parser.Var,
[
('?foo', Variable('foo')),
('$bar', Variable('bar')),
]),
(parser.IRIref,
[
('pre:local', components.QName('pre:local')),
('<s:ex>', URIRef('s:ex')),
]),
(parser.RDFLiteral,
[
('"foo"', Literal("foo")),
('"foo"@en-US', Literal("foo", lang='en-US')),
("'bar'^^pre:type", components.ParsedDatatypedLiteral(
"bar", components.QName('pre:type'))),
("'bar'^^<s:type>", components.ParsedDatatypedLiteral(
"bar", URIRef('s:type'))),
def testN3Store(store="default", configString=None):
g = ConjunctiveGraph(store=store)
if configString:
g.destroy(configString)
g.open(configString)
g.parse(data=testN3, format="n3")
print g.store
try:
for s, p, o in g.triples((None, implies, None)):
formulaA = s
formulaB = o
assert type(formulaA) == QuotedGraph and type(formulaB) == QuotedGraph
a = URIRef('http://test/a')
b = URIRef('http://test/b')
c = URIRef('http://test/c')
d = URIRef('http://test/d')
v = Variable('y')
universe = ConjunctiveGraph(g.store)
#test formula as terms
assert len(list(universe.triples((formulaA, implies, formulaB)))) == 1
#test variable as term and variable roundtrip
assert len(list(formulaB.triples((None, None, v)))) == 1
for s, p, o in formulaB.triples((None, d, None)):
if o != c:
assert isinstance(o, Variable)
assert o == v
s = list(universe.subjects(RDF.type, RDFS.Class))[0]
assert isinstance(s, BNode)
assert len(list(universe.triples((None, implies, None)))) == 1
assert len(list(universe.triples((None, RDF.type, None)))) == 1
assert len(list(formulaA.triples((None, RDF.type, None)))) == 1
assert len(list(formulaA.triples((None, None, None)))) == 2
assert len(list(formulaB.triples((None, None, None)))) == 2
assert len(list(universe.triples((None, None, None)))) == 3
assert len(
list(formulaB.triples((None, URIRef('http://test/d'), None)))) == 2
assert len(
list(universe.triples((None, URIRef('http://test/d'), None)))) == 1
#context tests
#test contexts with triple argument
assert len(list(universe.contexts((a, d, c)))) == 1
#Remove test cases
universe.remove((None, implies, None))
assert len(list(universe.triples((None, implies, None)))) == 0
assert len(list(formulaA.triples((None, None, None)))) == 2
assert len(list(formulaB.triples((None, None, None)))) == 2
formulaA.remove((None, b, None))
assert len(list(formulaA.triples((None, None, None)))) == 1
formulaA.remove((None, RDF.type, None))
assert len(list(formulaA.triples((None, None, None)))) == 0
universe.remove((None, RDF.type, RDFS.Class))
#remove_context tests
universe.remove_context(formulaB)
assert len(list(universe.triples((None, RDF.type, None)))) == 0
assert len(universe) == 1
assert len(formulaB) == 0
universe.remove((None, None, None))
assert len(universe) == 0
g.store.destroy(configString)
except:
g.store.destroy(configString)
raise
def ask_collection(self, c_id):
if not isinstance(c_id, list):
c_id = [c_id]
ids = [self.marmotta.ldp(encoder.encode(id)) for id in c_id]
result = self.sparql.find(ids,self.RDA.ns.Collection)
return float(result.bindings.pop().get(Variable('size')))/len(c_id)
def univar(self, label, sic=False):
if not sic:
self.counter += 1
label += str(self.counter)
return Variable(label)