def test_literal_multiline_quotes(self):
s, p, o = split_nt_line(self.literal_multiline_quotes)
self.assertEqual(s, "http://example.org/#spiderman")
self.assertEqual(p, "http://example.org/text")
self.assertEqual(
o,
"\"This is a multi-line\\nliteral with many quotes (\\\"\\\"\\\"\\\"\\\")"
"\\nand two apostrophes ('').\"")
def test_literal_xmls(self):
s, p, o = split_nt_line(self.literal_xmls)
self.assertEqual(s, "http://example.org/show/218")
self.assertEqual(p, "http://www.w3.org/2000/01/rdf-schema#label")
self.assertEqual(
o,
""""That Seventies Show"^^<http://www.w3.org/2001/XMLSchema#string>"""
)
def lines_to_triples(self,
triples: List[str]) -> List[Tuple[str, str, str]]:
errors = 0
results = []
for triple_line in tqdm(triples,
desc='Converting triples',
unit=' triples'):
# split it and check if it is a triple
try:
triple = split_nt_line(triple_line)
results.append(triple)
except ValueError:
errors += 1
continue
print(f'{errors} occurred during triple conversion')
return results
def test_literal_double(self):
s, p, o = split_nt_line(self.literal_double)
self.assertEqual(s, "http://en.wikipedia.org/wiki/Helium")
self.assertEqual(p, "http://example.org/elements/specificGravity")
self.assertEqual(
o, "\"1.663E-4\"^^<http://www.w3.org/2001/XMLSchema#double>")
def test_literal_integer(self):
s, p, o = split_nt_line(self.literal_integer)
self.assertEqual(s, "http://en.wikipedia.org/wiki/Helium")
self.assertEqual(p, "http://example.org/elements/atomicNumber")
self.assertEqual(o,
"\"2\"^^<http://www.w3.org/2001/XMLSchema#integer>")
def test_literal_region(self):
s, p, o = split_nt_line(self.literal_region)
self.assertEqual(s, "http://example.org/show/218")
self.assertEqual(p, "http://example.org/show/localName")
self.assertEqual(o, "\"Cette Série des Années Septante\"@fr-be")
def test_literal_language(self):
s, p, o = split_nt_line(self.literal_language)
self.assertEqual(s, "http://example.org/show/218")
self.assertEqual(p, "http://example.org/show/localName")
self.assertEqual(o, "\"That Seventies Show\"@en")
def test_literal_untyped(self):
s, p, o = split_nt_line(self.literal_untyped)
self.assertEqual(s, "http://example.org/show/218")
self.assertEqual(p, "http://www.w3.org/2000/01/rdf-schema#label")
self.assertEqual(o, "\"That Seventies Show\"")
def test_split_simple_tabs(self):
s, p, o = split_nt_line(self.simple_line_tabs)
self.assertEqual(s, "http://example.org/#spiderman")
self.assertEqual(
p, "http://www.perceive.net/schemas/relationship/enemyOf")
self.assertEqual(o, "http://example.org/#green-goblin")
def map_triple_lines(self, triple_lines: List[str]):
"""
Assigns each entity and each relation an id and creates a list of triples consisting of the ids.
:param triple_lines: List of Triples
:return: Tuple with the mapping of entity -> ID, relation -> ID and a list of the triples with the integer ids.
"""
# prepare for mapping
dict_ent = dict()
dict_rel = dict()
list_triples = []
num_ent = 0
num_rel = 0
# iterate through every line containing a triple
sys.stdout.write("Processing Triple lines ... ")
sys.stdout.flush()
skipped = 0
for triple_line in tqdm(triple_lines,
desc='Processing',
unit=' triples'):
# split it and check if it is a triple
try:
triple = split_nt_line(triple_line)
except ValueError as e:
if self.fail_silently:
skipped += 1
continue
raise e
# check if subject is in entities, add if not
if triple[0] not in dict_ent:
idx_sub = num_ent
dict_ent.update({triple[0]: idx_sub})
num_ent += 1
else:
idx_sub = dict_ent[triple[0]]
# check if predicate is in relations, add if not
if triple[1] not in dict_rel:
idx_rel = num_rel
dict_rel.update({triple[1]: idx_rel})
num_rel += 1
else:
idx_rel = dict_rel[triple[1]]
# check if object is in entities, add if not
if triple[2] not in dict_ent:
idx_obj = num_ent
dict_ent.update({triple[2]: idx_obj})
num_ent += 1
else:
idx_obj = dict_ent[triple[2]]
# check triple
if idx_sub < 0 or idx_rel < 0 or idx_obj < 0:
sys.exit("Failure: Mapped Triple has invalid Indeces")
# add to the mapped triples list
# careful: OpenKE format is "subject object relation"
mapped_triple = [idx_sub, idx_obj, idx_rel]
list_triples.append(mapped_triple)
# # output progess (avoid spamming)
# progess += 1
# percentage = int((progess * 100) / finish)
# if percentage > last_percentage:
# sys.stdout.write("\rProcessing Triple lines ... " + str(percentage) + "%")
# sys.stdout.flush()
# last_percentage = percentage
# output results
print("")
print(str(len(dict_ent)) + " Distinct Entities")
print(str(len(dict_rel)) + " Distinct Relations")
print(str(len(list_triples)) + " Distinct Triples")
print("Skipped " + str(skipped) + " lines")
return dict_ent, dict_rel, list_triples