"""
data = pd.read_csv(file_name, sep=sep, header=0)
return [rdflib.URIRef(x) for x in data[col_entity]], list(data[col_label])
test_entities, test_labels = load_data(DATASET["test"][0], DATASET["test"][1],
DATASET["test"][2])
train_entities, train_labels = load_data(DATASET["train"][0],
DATASET["train"][1],
DATASET["train"][2])
entities = train_entities + test_entities
labels = train_labels + test_labels
kg = rdflib_to_kg(OUTPUT, label_predicates=LABEL_PREDICATES)
train_embeddings, test_embeddings = create_embeddings(kg, entities,
len(train_entities),
WALKERS)
# Fit a support vector machine on train embeddings and evaluate on test
clf = SVC(random_state=42)
clf.fit(train_embeddings, train_labels)
print("Support Vector Machine:")
print(
f"Accuracy = {accuracy_score(test_labels, clf.predict(test_embeddings))}")
print(confusion_matrix(test_labels, clf.predict(test_embeddings)))
# Create t-SNE embeddings from RDF2Vec embeddings (dimensionality reduction)
walk_tsne = TSNE(random_state=42)
X_walk_tsne = walk_tsne.fit_transform(train_embeddings + test_embeddings)
def test_rdflib_to_kg():
kg = rdflib_to_kg("samples/mutag.owl", label_predicates=label_predicates)
assert type(kg) == KnowledgeGraph
import random
import rdflib
from pyrdf2vec.converters import rdflib_to_kg
from pyrdf2vec.graph import Vertex
from pyrdf2vec.walkers import RandomWalker
LABEL_PREDICATE = "http://dl-learner.org/carcinogenesis#isMutagenic"
KG = rdflib_to_kg("samples/mutag.owl", label_predicates=[LABEL_PREDICATE])
def generate_entities():
return [
rdflib.URIRef(
f"{LABEL_PREDICATE.split('#')[0] + '#'}{random.randint(0, 335)}")
for _ in range(random.randint(0, 200))
]
class TestRandomWalker:
def test_extract_random_walks(self):
walks = RandomWalker(4, float("inf")).extract_random_walks(
KG, Vertex(str(generate_entities())))
assert type(walks) == list
def test_extract(self):
canonical_walks = RandomWalker(4, float("inf")).extract(
KG, generate_entities())
assert type(canonical_walks) == set