def diagnosis_example(num_graphs=200,
num_processing_steps_tr=5,
num_processing_steps_ge=5,
num_training_iterations=1000,
keyspace=KEYSPACE,
uri=URI):
"""
Run the diagnosis example from start to finish, including traceably ingesting predictions back into Grakn
Args:
num_graphs: Number of graphs to use for training and testing combined
num_processing_steps_tr: The number of message-passing steps for training
num_processing_steps_ge: The number of message-passing steps for testing
num_training_iterations: The number of training epochs
keyspace: The name of the keyspace to retrieve example subgraphs from
uri: The uri of the running Grakn instance
Returns:
Final accuracies for training and for testing
"""
tr_ge_split = int(num_graphs * 0.5)
generate_example_graphs(num_graphs, keyspace=keyspace, uri=uri)
client = GraknClient(uri=uri)
session = client.session(keyspace=keyspace)
graphs = create_concept_graphs(list(range(num_graphs)), session)
with session.transaction().read() as tx:
# Change the terminology here onwards from thing -> node and role -> edge
node_types = get_thing_types(tx)
[node_types.remove(el) for el in TYPES_TO_IGNORE]
edge_types = get_role_types(tx)
[edge_types.remove(el) for el in ROLES_TO_IGNORE]
print(f'Found node types: {node_types}')
print(f'Found edge types: {edge_types}')
ge_graphs, solveds_tr, solveds_ge = pipeline(
graphs,
tr_ge_split,
node_types,
edge_types,
num_processing_steps_tr=num_processing_steps_tr,
num_processing_steps_ge=num_processing_steps_ge,
num_training_iterations=num_training_iterations,
continuous_attributes=CONTINUOUS_ATTRIBUTES,
categorical_attributes=CATEGORICAL_ATTRIBUTES,
output_dir=f"./events/{time.time()}/")
with session.transaction().write() as tx:
write_predictions_to_grakn(ge_graphs, tx)
session.close()
client.close()
return solveds_tr, solveds_ge
def diagnosis_example(num_graphs=200,
num_processing_steps_tr=10,
num_processing_steps_ge=10,
num_training_iterations=1000,
keyspace="diagnosis",
uri="localhost:48555"):
tr_ge_split = int(num_graphs * 0.5)
generate_example_graphs(num_graphs, keyspace=keyspace, uri=uri)
client = GraknClient(uri=uri)
session = client.session(keyspace=keyspace)
graphs = create_concept_graphs(list(range(num_graphs)), session)
with session.transaction().read() as tx:
# Change the terminology here onwards from thing -> node and role -> edge
node_types = get_thing_types(tx)
edge_types = get_role_types(tx)
print(f'Found node types: {node_types}')
print(f'Found edge types: {edge_types}')
ge_graphs, solveds_tr, solveds_ge = pipeline(
graphs,
tr_ge_split,
node_types,
edge_types,
num_processing_steps_tr=num_processing_steps_tr,
num_processing_steps_ge=num_processing_steps_ge,
num_training_iterations=num_training_iterations,
continuous_attributes=CONTINUOUS_ATTRIBUTES,
categorical_attributes=CATEGORICAL_ATTRIBUTES,
output_dir=f"./events/{time.time()}/")
with session.transaction().write() as tx:
write_predictions_to_grakn(ge_graphs, tx)
session.close()
client.close()
return solveds_tr, solveds_ge