def test_pipeline_hierarchy_fit_correct(data_setup):
data = data_setup
train, _ = train_test_data_setup(data)
first = PrimaryNode(operation_type='logit')
second = SecondaryNode(operation_type='logit', nodes_from=[first])
third = SecondaryNode(operation_type='logit', nodes_from=[first])
final = SecondaryNode(operation_type='logit', nodes_from=[second, third])
pipeline = Pipeline()
for node in [first, second, third, final]:
pipeline.add_node(node)
pipeline.unfit()
train_predicted = pipeline.fit(input_data=train)
assert pipeline.root_node.descriptive_id == (
'((/n_logit_default_params;)/'
'n_logit_default_params;;(/'
'n_logit_default_params;)/'
'n_logit_default_params;)/'
'n_logit_default_params')
assert pipeline.length == 4
assert pipeline.depth == 3
assert train_predicted.predict.shape[0] == train.target.shape[0]
assert final.fitted_operation is not None
def create_pipeline_with_several_nested_atomized_model() -> Pipeline:
pipeline = Pipeline()
atomized_op = create_atomized_model_with_several_atomized_models()
node_atomized_model = PrimaryNode(operation_type=atomized_op)
node_atomized_model_secondary = SecondaryNode(
operation_type=create_atomized_model())
node_atomized_model_secondary.nodes_from = [node_atomized_model]
node_knn = SecondaryNode('knn')
node_knn.custom_params = {'n_neighbors': 9}
node_knn.nodes_from = [node_atomized_model]
node_knn_second = SecondaryNode('knn')
node_knn_second.custom_params = {'n_neighbors': 5}
node_knn_second.nodes_from = [
node_atomized_model, node_atomized_model_secondary, node_knn
]
node_atomized_model_secondary_second = \
SecondaryNode(operation_type=create_atomized_model_with_several_atomized_models())
node_atomized_model_secondary_second.nodes_from = [node_knn_second]
pipeline.add_node(node_atomized_model_secondary_second)
return pipeline
def pipeline_with_secondary_nodes_only():
first = SecondaryNode(operation_type='logit', nodes_from=[])
second = SecondaryNode(operation_type='logit', nodes_from=[first])
pipeline = Pipeline()
pipeline.add_node(first)
pipeline.add_node(second)
return pipeline
def pipeline_with_self_cycle():
first = PrimaryNode(operation_type='logit')
second = SecondaryNode(operation_type='logit', nodes_from=[first])
second.nodes_from.append(second)
pipeline = Pipeline()
pipeline.add_node(first)
pipeline.add_node(second)
return pipeline
def pipeline_with_multiple_roots():
first = PrimaryNode(operation_type='logit')
root_first = SecondaryNode(operation_type='logit', nodes_from=[first])
root_second = SecondaryNode(operation_type='logit', nodes_from=[first])
pipeline = Pipeline()
for node in [first, root_first, root_second]:
pipeline.add_node(node)
return pipeline
def pipeline_with_cycle():
first = PrimaryNode(operation_type='logit')
second = SecondaryNode(operation_type='logit', nodes_from=[first])
third = SecondaryNode(operation_type='logit', nodes_from=[second, first])
second.nodes_from.append(third)
pipeline = Pipeline()
for node in [first, second, third]:
pipeline.add_node(node)
return pipeline
def baseline_pipeline():
pipeline = Pipeline()
last_node = SecondaryNode(operation_type='xgboost', nodes_from=[])
for requirement_model in ['knn', 'logit']:
new_node = PrimaryNode(requirement_model)
pipeline.add_node(new_node)
last_node.nodes_from.append(new_node)
pipeline.add_node(last_node)
return pipeline
def pipeline_with_isolated_components():
first = PrimaryNode(operation_type='logit')
second = SecondaryNode(operation_type='logit', nodes_from=[first])
third = SecondaryNode(operation_type='logit', nodes_from=[])
fourth = SecondaryNode(operation_type='logit', nodes_from=[third])
pipeline = Pipeline()
for node in [first, second, third, fourth]:
pipeline.add_node(node)
return pipeline
def valid_pipeline():
first = PrimaryNode(operation_type='logit')
second = SecondaryNode(operation_type='logit', nodes_from=[first])
third = SecondaryNode(operation_type='logit', nodes_from=[second])
last = SecondaryNode(operation_type='logit', nodes_from=[third])
pipeline = Pipeline()
for node in [first, second, third, last]:
pipeline.add_node(node)
return pipeline
def pipeline_third():
# QDA
# | \
# RF RF
pipeline = Pipeline()
new_node = SecondaryNode('qda')
for model_type in ('rf', 'rf'):
new_node.nodes_from.append(PrimaryNode(model_type))
pipeline.add_node(new_node)
[pipeline.add_node(node_from) for node_from in new_node.nodes_from]
return pipeline
def test_pipeline_repr():
first = PrimaryNode(operation_type='logit')
second = PrimaryNode(operation_type='lda')
third = PrimaryNode(operation_type='knn')
final = SecondaryNode(operation_type='xgboost',
nodes_from=[first, second, third])
pipeline = Pipeline()
pipeline.add_node(final)
expected_pipeline_description = "{'depth': 2, 'length': 4, 'nodes': [xgboost, logit, lda, knn]}"
assert repr(pipeline) == expected_pipeline_description
def test_delete_node_with_redirection():
first = PrimaryNode(operation_type='logit')
second = PrimaryNode(operation_type='lda')
third = SecondaryNode(operation_type='knn', nodes_from=[first, second])
final = SecondaryNode(operation_type='xgboost',
nodes_from=[third])
pipeline = Pipeline()
pipeline.add_node(final)
pipeline.delete_node(third)
assert len(pipeline.nodes) == 3
assert first in pipeline.root_node.nodes_from
def create_pipeline() -> Pipeline:
pipeline = Pipeline()
node_logit = PrimaryNode('logit')
node_lda = PrimaryNode('lda')
node_lda.custom_params = {'n_components': 1}
node_xgboost = SecondaryNode('xgboost')
node_xgboost.custom_params = {'n_components': 1}
node_xgboost.nodes_from = [node_logit, node_lda]
pipeline.add_node(node_xgboost)
return pipeline
def pipeline_third():
# XG
# / | \
# KNN LDA KNN
root_of_tree = SecondaryNode('xgboost')
for model_type in ('knn', 'lda', 'knn'):
root_of_tree.nodes_from.append(PrimaryNode(model_type))
pipeline = Pipeline()
for node in root_of_tree.nodes_from:
pipeline.add_node(node)
pipeline.add_node(root_of_tree)
return pipeline
def test_update_node_in_pipeline_correct():
first = PrimaryNode(operation_type='logit')
final = SecondaryNode(operation_type='xgboost', nodes_from=[first])
pipeline = Pipeline()
pipeline.add_node(final)
new_node = PrimaryNode('svc')
replacing_node = SecondaryNode('logit', nodes_from=[new_node])
pipeline.update_node(old_node=first, new_node=replacing_node)
assert replacing_node in pipeline.nodes
assert new_node in pipeline.nodes
assert first not in pipeline.nodes
def test_secondary_nodes_is_invariant_to_inputs_order(data_setup):
data = data_setup
train, test = train_test_data_setup(data)
first = PrimaryNode(operation_type='logit')
second = PrimaryNode(operation_type='lda')
third = PrimaryNode(operation_type='knn')
final = SecondaryNode(operation_type='xgboost',
nodes_from=[first, second, third])
pipeline = Pipeline()
for node in [first, second, third, final]:
pipeline.add_node(node)
first = deepcopy(first)
second = deepcopy(second)
third = deepcopy(third)
final_shuffled = SecondaryNode(operation_type='xgboost',
nodes_from=[third, first, second])
pipeline_shuffled = Pipeline()
# change order of nodes in list
for node in [final_shuffled, third, first, second]:
pipeline_shuffled.add_node(node)
train_predicted = pipeline.fit(input_data=train)
train_predicted_shuffled = pipeline_shuffled.fit(input_data=train)
# train results should be invariant
assert pipeline.root_node.descriptive_id == pipeline_shuffled.root_node.descriptive_id
assert np.equal(train_predicted.predict, train_predicted_shuffled.predict).all()
test_predicted = pipeline.predict(input_data=test)
test_predicted_shuffled = pipeline_shuffled.predict(input_data=test)
# predict results should be invariant
assert np.equal(test_predicted.predict, test_predicted_shuffled.predict).all()
# change parents order for the nodes fitted pipeline
nodes_for_change = pipeline.nodes[3].nodes_from
pipeline.nodes[3].nodes_from = [nodes_for_change[2], nodes_for_change[0], nodes_for_change[1]]
pipeline.nodes[3].unfit()
pipeline.fit(train)
test_predicted_re_shuffled = pipeline.predict(input_data=test)
# predict results should be invariant
assert np.equal(test_predicted.predict, test_predicted_re_shuffled.predict).all()
def create_atomized_model_with_several_atomized_models() -> AtomizedModel:
pipeline = Pipeline()
node_atomized_model_primary = PrimaryNode(
operation_type=create_atomized_model())
node_atomized_model_secondary = SecondaryNode(
operation_type=create_atomized_model())
node_atomized_model_secondary_second = SecondaryNode(
operation_type=create_atomized_model())
node_atomized_model_secondary_third = SecondaryNode(
operation_type=create_atomized_model())
node_atomized_model_secondary.nodes_from = [node_atomized_model_primary]
node_atomized_model_secondary_second.nodes_from = [
node_atomized_model_primary
]
node_atomized_model_secondary_third.nodes_from = [
node_atomized_model_secondary, node_atomized_model_secondary_second
]
pipeline.add_node(node_atomized_model_secondary_third)
atomized_model = AtomizedModel(pipeline)
return atomized_model
def test_pipeline_with_custom_params_for_model(data_setup):
data = data_setup
custom_params = dict(n_neighbors=1,
weights='uniform',
p=1)
first = PrimaryNode(operation_type='logit')
second = PrimaryNode(operation_type='lda')
final = SecondaryNode(operation_type='knn', nodes_from=[first, second])
pipeline = Pipeline()
pipeline.add_node(final)
pipeline_default_params = deepcopy(pipeline)
pipeline.root_node.custom_params = custom_params
pipeline_default_params.fit(data)
pipeline.fit(data)
custom_params_prediction = pipeline.predict(data).predict
default_params_prediction = pipeline_default_params.predict(data).predict
assert not np.array_equal(custom_params_prediction, default_params_prediction)
def compose_pipeline(self, data: InputData,
initial_pipeline: Optional[Pipeline],
composer_requirements: ComposerRequirements,
metrics: Optional[Callable]) -> Pipeline:
# TODO: fix this later?
train_data = data
test_data = data
metric_function_for_nodes = partial(metric_for_nodes,
metric_function=metrics,
train_data=train_data,
test_data=test_data)
optimiser = RandomSearchOptimiser(self.__iter_num, PrimaryNode,
SecondaryNode)
best_nodes_set, history = optimiser.optimise(
metric_function_for_nodes, composer_requirements.primary,
composer_requirements.secondary)
best_pipeline = Pipeline()
[best_pipeline.add_node(nodes) for nodes in best_nodes_set]
return best_pipeline
def test_pipeline_with_datamodel_fit_correct(data_setup):
data = data_setup
train_data, test_data = train_test_data_setup(data)
pipeline = Pipeline()
node_data = PrimaryNode('logit')
node_first = PrimaryNode('bernb')
node_second = SecondaryNode('rf')
node_second.nodes_from = [node_first, node_data]
pipeline.add_node(node_data)
pipeline.add_node(node_first)
pipeline.add_node(node_second)
pipeline.fit(train_data)
results = np.asarray(probs_to_labels(pipeline.predict(test_data).predict))
assert results.shape == test_data.target.shape
def pipeline_first():
# XG
# | \
# XG KNN
# | \ | \
# LR LDA LR LDA
pipeline = Pipeline()
root_of_tree, root_child_first, root_child_second = \
[SecondaryNode(model) for model in ('xgboost', 'xgboost', 'knn')]
for root_node_child in (root_child_first, root_child_second):
for requirement_model in ('logit', 'lda'):
new_node = PrimaryNode(requirement_model)
root_node_child.nodes_from.append(new_node)
pipeline.add_node(new_node)
pipeline.add_node(root_node_child)
root_of_tree.nodes_from.append(root_node_child)
pipeline.add_node(root_of_tree)
return pipeline
def nodes_to_pipeline(nodes: List[Node]) -> Pipeline:
pipeline = Pipeline()
[pipeline.add_node(nodes) for nodes in nodes]
return pipeline
