class GraphTester(object):
"""Compares MLR graphs obtained from converting LOM to expected graphs.
The graph tester is stateful.
1. Receives a LOM fragment and stores it.
2. Receives a required N3 MLR graph and LOM conversion parameters
A. Converts the LOM fragment to a converted graph (with parameters.)
B. Checks that all triples in the N3 graph are found in the converted
graph. Stores the required graph.
3. Receives a forbidden N3 MLR graph:
Checks that all triples in the forbidden graph are either present in
the required graph or absent from the converted graph.
"""
MISSING = 1
UNEXPECTED = 2
def __init__(self):
self.converter = Converter()
self.parser = self.converter.populate_argparser()
self.reset()
def normalizeTerm(self, graph, term):
if isinstance(term, URIRef):
return graph.normalizeUri(term)
return term
def reset(self):
"Clears the state"
self.last_lom = None
self.last_graph = None
self.last_comparator = None
def set_lom(self, lom):
"Sets the LOM fragment."
self.reset()
self.last_lom = etree.fromstring(LOM_TEMPLATE % (lom,))
def process_line(self, format, code, args=None):
"Process a fragment."
format = format.lower()
if format == 'xml':
self.set_lom(code)
return None, []
elif format == 'n3':
return self.test_n3(code, args)
elif format == 'rdf-xml':
return self.test_graph(Graph().parse(data=code, format="xml"), args)
assert False, 'format should be xml or n3'
def parse_n3(self, n3):
"Parse a N3 graph"
return Graph().parse(data=N3_PREFIXES + n3, format="n3")
def test_n3(self, n3, args=None):
"Test that the provided n3 fragment is conformant"
return self.test_graph(self.parse_n3(n3), args)
def find_missing(self, expected_graph, obtained_graph):
"List triples in the expected graph missing from the obtained graph."
errors = []
nsm = obtained_graph.namespace_manager
comparator_eo = GraphCorrespondence(expected_graph, obtained_graph)
self.last_comparator = comparator_eo
comparator_eo.identify()
triples = expected_graph.triples((None, None, None))
triples = comparator_eo.translate_triple_list(triples)
for triple in triples:
if not list(obtained_graph.triples(triple)):
errors.append(tuple([self.normalizeTerm(nsm, x) for x in triple]))
return errors
def find_forbidden(self, forbidden_graph, obtained_graph):
"""List triples in the forbidden graph found in the obtained graph
and absent from the previous (stored) required graph.
"""
assert self.last_comparator
expected_graph = self.last_comparator.source
errors = []
nsm = obtained_graph.namespace_manager
comparator_fo = GraphCorrespondence(forbidden_graph, obtained_graph)
comparator_fo.identify()
map_oe = dict(
(o, e) for (e, o) in
self.last_comparator.blank_map.iteritems())
map_fe = dict(
(f, map_oe.get(o, None)) for f, o in
comparator_fo.blank_map.iteritems())
triples = forbidden_graph.triples((None, None, None))
for triple in triples:
if list(obtained_graph.triples(comparator_fo.translate_triple(triple))) \
and not list(expected_graph.triples(translate_triple(triple, map_fe))):
errors.append(tuple([self.normalizeTerm(nsm, x) for x in triple]))
return errors
def test_graph(self, graph, args=None):
"Test that the provided graph is conformant"
assert self.last_lom is not None
errors = []
if args and args.lower() == 'forbidden':
assert(self.last_graph)
errors = [(self.UNEXPECTED, e) for e in
self.find_forbidden(graph, self.last_graph)]
else:
if args:
options = self.parser.parse_args(args.split())
self.converter.set_options_from_dict(vars(options))
else:
self.converter.set_options_from_dict()
obtained_graph = self.converter.lomxml2graph(self.last_lom)
self.last_graph = obtained_graph
errors = [(self.MISSING, e) for e in self.find_missing(graph, obtained_graph)]
return self.last_graph, errors
def __init__(self):
self.converter = Converter()
self.parser = self.converter.populate_argparser()
self.reset()
class testMlr(unittest.TestCase):
@classmethod
def setupClass(self):
self.converter = Converter()
self.graph = self.converter.lomfile2graph(EXAMPLE)
def triple_from_n3(self, n3):
g = Graph().parse(data=n3, format="n3")
return next(g.triples((None, None, None)))
def test_not_empty(self):
assert (len(self.graph) > 0)
def test_has_lang(self):
expected = '%s <%s> mlr3:DES0500 "fra-CA".' % (prologue, TEST_ID)
assert (len(list(self.graph.triples(self.triple_from_n3(expected)))))
def test_has_all_mlr_values(self):
triples = list(self.graph.triples((term.URIRef(TEST_ID), None, None)))
#sys.stderr.write(`triples`)
predicates = set(p for (s, p, o) in triples)
missing_predicates = []
for (p, n) in Element_names.iteritems():
if p in Known_Missing:
continue
if p not in predicates:
missing_predicates.append((p, n))
assert not missing_predicates, missing_predicates
def test_has_only_mlr_values(self):
triples = list(self.graph.triples((term.URIRef(TEST_ID), None, None)))
predicates = set(p for (s, p, o) in triples)
extra_predicates = []
for p in predicates:
if p == term.URIRef('http://www.inria.fr/acacia/corese#graph'):
continue
if not p in Element_names:
extra_predicates.append(p)
assert not extra_predicates, extra_predicates
def test_codomain(self):
wrong_codomain_type = []
for predicate in MLR_codomain.iterkeys():
for s, p, o in self.graph.triples(
(None, term.URIRef(predicate), None)):
for s2, p2, o2 in self.graph.triples((o, RDF.type, None)):
if o2 not in MLR_codomain[predicate]:
wrong_codomain_type.append((predicate, o2))
assert not wrong_codomain_type, wrong_codomain_type
def test_subobjects_has_only_mlr_values(self):
extra_predicates = []
for s, p, o in self.graph.triples((None, RDF.type, None)):
if o == MLR1.RC0002:
continue
assert o in MLR_Subclass_attributes, "Missing type:" + o
attributes = MLR_Subclass_attributes[o]
for s2, p2, o2 in self.graph.triples((s, None, None)):
if RDF.type == p2:
continue
if not p2 in attributes:
extra_predicates.append((o, p2))
assert not extra_predicates, extra_predicates
def test_subobjects_has_all_mlr_values(self):
missing_types = []
missing_predicates = []
res_by_type = defaultdict(set)
pred_by_type = defaultdict(set)
for s, p, o in self.graph.triples((None, RDF.type, None)):
res_by_type[o].add(s)
for type_uri, ressources in res_by_type.iteritems():
for r in ressources:
for s, p, o in self.graph.triples((r, None, None)):
pred_by_type[type_uri].add(p)
for type_uri, predicates in pred_by_type.iteritems():
if type_uri == MLR1.RC0002:
continue
if type_uri not in MLR_Subclass_attributes:
missing_types.append(type_uri)
continue
attributes = MLR_Subclass_attributes[type_uri]
for p, name in attributes.iteritems():
if p in Known_Missing:
continue
if not p in predicates:
missing_predicates.append((type_uri, p, name))
assert not missing_types, missing_types
assert not missing_predicates, missing_predicates
def test_value_types(self):
triples = list(self.graph.triples((None, None, None)))
for s, p, o in triples:
if p in Type_constraints:
assert Type_constraints[p](str(o))
def test_lom_files(self):
ex_dir = os.path.join(this_dir, 'data')
examples = os.listdir(ex_dir)
all_clear = True
for example in examples:
print(example)
graph = self.converter.lomfile2graph(os.path.join(ex_dir, example))
def setupClass(self):
self.converter = Converter()
self.graph = self.converter.lomfile2graph(EXAMPLE)
def __init__(self):
self.converter = Converter()
parser = argparse.ArgumentParser(description="Graph Comparison")
self.parser = self.converter.populate_argparser(parser)
self.reset()
def setupClass(self):
self.converter = Converter()
self.graph = self.converter.lomfile2graph(EXAMPLE)
class testMlr(unittest.TestCase):
@classmethod
def setupClass(self):
self.converter = Converter()
self.graph = self.converter.lomfile2graph(EXAMPLE)
def triple_from_n3(self, n3):
g = Graph().parse(data=n3, format="n3")
return next(g.triples((None,None,None)))
def test_not_empty(self):
assert(len(self.graph)>0)
def test_has_lang(self):
expected = '%s <%s> mlr3:DES0500 "fra-CA".' % (prologue, TEST_ID)
assert(len(list(self.graph.triples(self.triple_from_n3(expected)))))
def test_has_all_mlr_values(self):
triples = list(self.graph.triples((term.URIRef(TEST_ID), None, None)))
#sys.stderr.write(`triples`)
predicates = set(p for (s, p, o) in triples)
missing_predicates = []
for (p, n) in Element_names.iteritems():
if p in Known_Missing:
continue
if p not in predicates:
missing_predicates.append((p, n))
assert not missing_predicates, missing_predicates
def test_has_only_mlr_values(self):
triples = list(self.graph.triples((term.URIRef(TEST_ID), None, None)))
predicates = set(p for (s, p, o) in triples)
extra_predicates = []
for p in predicates:
if p == term.URIRef('http://www.inria.fr/acacia/corese#graph'):
continue
if not p in Element_names:
extra_predicates.append(p)
assert not extra_predicates, extra_predicates
def test_codomain(self):
wrong_codomain_type = []
for predicate in MLR_codomain.iterkeys():
for s, p, o in self.graph.triples((None, term.URIRef(predicate), None)):
for s2, p2, o2 in self.graph.triples((o, RDF.type, None)):
if o2 not in MLR_codomain[predicate]:
wrong_codomain_type.append((predicate, o2))
assert not wrong_codomain_type, wrong_codomain_type
def test_subobjects_has_only_mlr_values(self):
extra_predicates = []
for s, p, o in self.graph.triples((None, RDF.type, None)):
if o == MLR1.RC0002:
continue
assert o in MLR_Subclass_attributes, "Missing type:" + o
attributes = MLR_Subclass_attributes[o]
for s2, p2, o2 in self.graph.triples((s, None, None)):
if RDF.type == p2:
continue
if not p2 in attributes:
extra_predicates.append((o, p2))
assert not extra_predicates, extra_predicates
def test_subobjects_has_all_mlr_values(self):
missing_types = []
missing_predicates = []
res_by_type = defaultdict(set)
pred_by_type = defaultdict(set)
for s, p, o in self.graph.triples((None, RDF.type, None)):
res_by_type[o].add(s)
for type_uri, ressources in res_by_type.iteritems():
for r in ressources:
for s, p, o in self.graph.triples((r, None, None)):
pred_by_type[type_uri].add(p)
for type_uri, predicates in pred_by_type.iteritems():
if type_uri == MLR1.RC0002:
continue
if type_uri not in MLR_Subclass_attributes:
missing_types.append(type_uri)
continue
attributes = MLR_Subclass_attributes[type_uri]
for p, name in attributes.iteritems():
if p in Known_Missing:
continue
if not p in predicates:
missing_predicates.append((type_uri, p, name))
assert not missing_types, missing_types
assert not missing_predicates, missing_predicates
def test_value_types(self):
triples = list(self.graph.triples((None, None, None)))
for s, p, o in triples:
if p in Type_constraints:
assert Type_constraints[p](str(o))
def test_lom_files(self):
ex_dir = os.path.join(this_dir, 'data')
examples = os.listdir(ex_dir)
all_clear = True
for example in examples:
print(example)
graph = self.converter.lomfile2graph(os.path.join(ex_dir, example))
def __init__(self):
self.converter = Converter()
self.parser = self.converter.populate_argparser()
self.reset()
class GraphTester(object):
"""Compares MLR graphs obtained from converting LOM to expected graphs.
The graph tester is stateful.
1. Receives a LOM fragment and stores it.
2. Receives a required N3 MLR graph and LOM conversion parameters
A. Converts the LOM fragment to a converted graph (with parameters.)
B. Checks that all triples in the N3 graph are found in the converted
graph. Stores the required graph.
3. Receives a forbidden N3 MLR graph:
Checks that all triples in the forbidden graph are either present in
the required graph or absent from the converted graph.
"""
MISSING = 1
UNEXPECTED = 2
def __init__(self):
self.converter = Converter()
self.parser = self.converter.populate_argparser()
self.reset()
def normalizeTerm(self, graph, term):
if isinstance(term, URIRef):
return graph.normalizeUri(term)
return term
def reset(self):
"Clears the state"
self.last_lom = None
self.last_graph = None
self.last_comparator = None
def set_lom(self, lom):
"Sets the LOM fragment."
self.reset()
self.last_lom = etree.fromstring(LOM_TEMPLATE % (lom, ))
def process_line(self, format, code, args=None):
"Process a fragment."
format = format.lower()
if format == 'xml':
self.set_lom(code)
return None, []
elif format == 'n3':
return self.test_n3(code, args)
elif format == 'rdf-xml':
return self.test_graph(Graph().parse(data=code, format="xml"),
args)
assert False, 'format should be xml or n3'
def parse_n3(self, n3):
"Parse a N3 graph"
return Graph().parse(data=N3_PREFIXES + n3, format="n3")
def test_n3(self, n3, args=None):
"Test that the provided n3 fragment is conformant"
return self.test_graph(self.parse_n3(n3), args)
def find_missing(self, expected_graph, obtained_graph):
"List triples in the expected graph missing from the obtained graph."
errors = []
nsm = obtained_graph.namespace_manager
comparator_eo = GraphCorrespondence(expected_graph, obtained_graph)
self.last_comparator = comparator_eo
comparator_eo.identify()
triples = expected_graph.triples((None, None, None))
triples = comparator_eo.translate_triple_list(triples)
for triple in triples:
if not list(obtained_graph.triples(triple)):
errors.append(
tuple([self.normalizeTerm(nsm, x) for x in triple]))
return errors
def find_forbidden(self, forbidden_graph, obtained_graph):
"""List triples in the forbidden graph found in the obtained graph
and absent from the previous (stored) required graph.
"""
assert self.last_comparator
expected_graph = self.last_comparator.source
errors = []
nsm = obtained_graph.namespace_manager
comparator_fo = GraphCorrespondence(forbidden_graph, obtained_graph)
comparator_fo.identify()
map_oe = dict(
(o, e) for (e, o) in self.last_comparator.blank_map.iteritems())
map_fe = dict((f, map_oe.get(o, None))
for f, o in comparator_fo.blank_map.iteritems())
triples = forbidden_graph.triples((None, None, None))
for triple in triples:
if list(obtained_graph.triples(comparator_fo.translate_triple(triple))) \
and not list(expected_graph.triples(translate_triple(triple, map_fe))):
errors.append(
tuple([self.normalizeTerm(nsm, x) for x in triple]))
return errors
def test_graph(self, graph, args=None):
"Test that the provided graph is conformant"
assert self.last_lom is not None
errors = []
if args and args.lower() == 'forbidden':
assert (self.last_graph)
errors = [(self.UNEXPECTED, e)
for e in self.find_forbidden(graph, self.last_graph)]
else:
if args:
options = self.parser.parse_args(args.split())
self.converter.set_options_from_dict(vars(options))
else:
self.converter.set_options_from_dict()
obtained_graph = self.converter.lomxml2graph(self.last_lom)
self.last_graph = obtained_graph
errors = [(self.MISSING, e)
for e in self.find_missing(graph, obtained_graph)]
return self.last_graph, errors