def test_valid(self, simple_tuple_node_list):
nodes = [node(*tup) for tup in simple_tuple_node_list]
assert len(nodes) == len(simple_tuple_node_list)
def test_bad_node(func, expected):
with pytest.raises(ValueError, match=expected):
node(*func())
def test_tag_nodes(self):
tagged_node = node(identity, "input", "output",
tags=["hello"]).tag(["world"])
assert "hello" in tagged_node.tags
assert "world" in tagged_node.tags
assert len(tagged_node.tags) == 2
def test_node_less_than(self):
first = node(identity, "input1", "output1", name="A")
second = node(identity, "input1", "output1", name="B")
assert first < second
assert first is not second
def test_different_input_list_order_not_equal(self):
first = node(biconcat, ["input1", "input2"], "output1", name="A")
second = node(biconcat, ["input2", "input1"], "output1", name="A")
assert first != second
def test_inputs_none(self):
dummy_node = node(constant_output, None, "output")
assert dummy_node.inputs == []
def test_outputs_str(self):
dummy_node = node(identity, "input1", "output1")
assert dummy_node.outputs == ["output1"]
def saving_none_pipeline():
return Pipeline([
node(random, None, "A"),
node(null, "A", "B"),
node(identity, "B", "C")
])
def branchless_pipeline():
return Pipeline([
node(identity, "ds1", "ds2", name="node1"),
node(identity, "ds2", "ds3", name="node2"),
])
def create_pipeline(**kwargs):
# conn start
start_conn = [
node(hktvmall_conn_node,
inputs="params:hktvmall_home_url",
outputs="HktvmallHeader",
tags="Preparation")
]
# get HKTV mall categories
get_category = [
node(request_hktvmall_catagory_code,
inputs=["HktvmallHeader", "params:hktvmall_category_diction_url"],
outputs="category_raw_req",
tags="Preparation"),
node(categories_df_etl,
inputs="category_raw_req",
outputs="category_df",
tags="Preparation")
]
# generate urls by type for requests
gen_url_list = [
node(gen_hktvmall_product_by_method_and_cat_links,
inputs=[
'params:hktvmall_catagory_code',
'params:hktvmall_browse_method',
"params:product_by_method_catcode_url"
],
outputs=dict(method1="promotiondiff_url_list",
method2="hotpickorder_url_list"),
tags="Preparation"),
node(gen_hktvmall_full_site_links,
inputs=["category_df", 'params:hktvmall_cat_product_url'],
outputs="fullsite_url_list",
tags="Preparation")
]
# multi threading requires for raw data
req_raw_df = [
node(multi_threading_req,
inputs=["HktvmallHeader", "promotiondiff_url_list"],
outputs="promotiondiff_raw_list",
tags="Requests"),
node(multi_threading_req,
inputs=["HktvmallHeader", "hotpickorder_url_list"],
outputs="hotpickorder_raw_list",
tags="Requests"),
node(multi_threading_req,
inputs=["HktvmallHeader", 'fullsite_url_list'],
outputs="fullsite_raw_list",
tags="Requests")
]
# ETL on df columns for proper columns
etl_on_df = [
node(raw_etl,
inputs="promotiondiff_raw_list",
outputs="promotiondiff_raw_df",
tags="ETL"),
node(raw_etl,
inputs="hotpickorder_raw_list",
outputs="hotpickorder_raw_df",
tags="ETL"),
node(raw_etl,
inputs="fullsite_raw_list",
outputs="fullsite_raw_df",
tags="ETL"),
]
# turn df to CSVDataSet
df_to_csv = [
node(df_to_kedro_csvdataset,
inputs=["category_df", "params:category_path"],
outputs="category_raw",
tags="Saving Data"),
node(df_to_kedro_csvdataset,
inputs=["promotiondiff_raw_df", "params:promotiondiff_path"],
outputs="promotiondiff_raw",
tags="Saving Data"),
node(df_to_kedro_csvdataset,
inputs=["hotpickorder_raw_df", "params:hotpickorder_path"],
outputs="hotpickorder_raw",
tags="Saving Data"),
node(df_to_kedro_csvdataset,
inputs=["fullsite_raw_df", "params:fullsite_path"],
outputs="fullsite_raw",
tags="Saving Data")
]
pipe = start_conn + get_category + gen_url_list + req_raw_df + etl_on_df + df_to_csv
# pipe = start_conn + get_category
# pipe.append(
# node(
# gen_hktvmall_full_site_links,
# inputs=["category_df", 'params:hktvmall_cat_product_url'],
# outputs="fullsite_url_list",
# tags="Preparation"
# )
# )
# pipe.append(
# node(
# multi_threading_req,
# inputs=["HktvmallHeader", 'fullsite_url_list'],
# outputs="fullsite_raw_list",
# tags="Requests"
# )
# )
# pipe.append(
# node(
# raw_etl,
# inputs="fullsite_raw_list",
# outputs="fullsite_raw_df",
# tags="ETL"
# ),
# )
# pipe.append(
# node(
# df_to_kedro_csvdataset,
# inputs=["fullsite_raw_df", "params:fullsite_path"],
# outputs="fullsite_raw",
# tags="Saving Data"
# )
# )
return Pipeline(pipe)
def create_pipeline(**kwargs):
return Pipeline([
node(make_prediction, ["test_x", "rf_model"], "predict"),
node(rmse_cv, ["test_x", "test_y", "rf_model"], "quality")
])
def _create_pipelines():
bad_pipeline_middle = Pipeline(
[
node(identity, "cars", "boats", name="node1", tags=["tag1"]),
node(identity, "boats", "trains", name="node2"),
node(bad_node, "trains", "ships", name="nodes3"),
node(identity, "ships", "planes", name="node4"),
],
tags="bad_pipeline",
)
bad_pipeline_head = Pipeline(
[
node(bad_node, "cars", "boats", name="node1", tags=["tag1"]),
node(identity, "boats", "trains", name="node2"),
node(identity, "trains", "ships", name="nodes3"),
node(identity, "ships", "planes", name="node4"),
],
tags="bad_pipeline",
)
default_pipeline = Pipeline(
[
node(identity, "cars", "boats", name="node1", tags=["tag1"]),
node(identity, "boats", "trains", name="node2"),
node(identity, "trains", "ships", name="node3"),
node(identity, "ships", "planes", name="node4"),
],
tags="pipeline",
)
return {
"__default__": default_pipeline,
"empty": Pipeline([]),
"simple": Pipeline([node(identity, "cars", "boats")]),
"bad_pipeline_middle": bad_pipeline_middle,
"bad_pipeline_head": bad_pipeline_head,
}
def test_task_exception(self, fan_out_fan_in, catalog):
catalog.add_feed_dict(feed_dict=dict(A=42))
pipeline = Pipeline([fan_out_fan_in, node(exception_fn, "Z", "X")])
with pytest.raises(Exception, match="test exception"):
ThreadRunner().run(pipeline, catalog)
def test_labelled(self):
assert "labeled_node: <lambda>([input1]) -> [output1]" in str(
node(lambda x: None, "input1", "output1", name="labeled_node"))
def test_no_input(self):
assert "constant_output(None) -> [output1]" in str(
node(constant_output, None, "output1"))
def saving_result_pipeline():
return Pipeline([node(identity, "ds", "dsX")])
def test_no_output(self):
assert "<lambda>([input1]) -> None" in str(
node(lambda x: None, "input1", None))
@pytest.fixture
def dummy_dataframe():
return pd.DataFrame({"col1": [1, 2], "col2": [4, 5], "col3": [5, 6]})
def identity(input1: str):
return input1 # pragma: no cover
def bad_node(x):
raise ValueError("Oh no!")
bad_pipeline_middle = Pipeline(
[
node(identity, "cars", "boats", name="node1", tags=["tag1"]),
node(identity, "boats", "trains", name="node2"),
node(bad_node, "trains", "ships", name="nodes3"),
node(identity, "ships", "planes", name="node4"),
],
tags="bad_pipeline",
)
expected_message_middle = (
"There are 2 nodes that have not run.\n"
"You can resume the pipeline run by adding the following "
"argument to your previous command:\n"
' --from-nodes "nodes3"')
bad_pipeline_head = Pipeline(
[
def test_outputs_none(self):
dummy_node = node(identity, "input", None)
assert dummy_node.outputs == []
def _get_pipelines(self) -> Dict[str, Pipeline]:
return {
"__default__": Pipeline([node(identity, "cars", "boats")])
}
def test_node_equals(self):
first = node(identity, "input1", "output1", name="a node")
second = node(identity, "input1", "output1", name="a node")
assert first == second
assert first is not second
def test_spark_pickle(self, is_async, data_catalog):
"""SparkDataSet(load) -> node -> PickleDataSet (save)"""
pipeline = Pipeline([node(identity, "spark_in", "pickle_ds")])
pattern = ".* was not serialized due to.*"
with pytest.raises(DataSetError, match=pattern):
SequentialRunner(is_async=is_async).run(pipeline, data_catalog)
def test_node_invalid_equals(self):
n = node(identity, "input1", "output1", name="a node")
assert n != "hello"
def create_pipeline(**kwargs):
return Pipeline([
node(
func=tidy_data,
inputs="filt_list",
outputs="data",
),
node(
func=buscar_numero_legajo,
inputs="data",
outputs="numero_legajo",
),
node(
func=buscar_violencia_genero,
inputs="data",
outputs="violencia_de_genero",
),
node(
func=buscar_violencia_fisica,
inputs="data",
outputs="violencia_fisica",
),
node(
func=buscar_info_denunciante,
inputs="data",
outputs="info_denunciante",
),
node(
func=buscar_info_acusado,
inputs="data",
outputs="info_acusado",
),
node(
func=buscar_genero_den,
inputs="info_denunciante",
outputs="genero_denunciante",
),
node(
func=buscar_nacionalidad_denunciante,
inputs="info_denunciante",
outputs="nacionalidad_denunciante",
),
node(
func=buscar_est_civil_denunciente,
inputs="info_denunciante",
outputs="est_civil_denunciante",
),
node(
func=buscar_edad_denunciante,
inputs="info_denunciante",
outputs="edad_denunciante",
),
node(
func=buscar_est_denunciante,
inputs="info_denunciante",
outputs="est_denunciante",
),
node(
func=buscar_domic_denunciante,
inputs="info_denunciante",
outputs="domic_denunciante",
),
node(
func=buscar_villa_denunciante,
inputs="info_denunciante",
outputs="villa_denunciante",
),
node(
func=buscar_ocupac_denunciante,
inputs="info_denunciante",
outputs="ocupac_denunciante",
),
node(
func=buscar_genero_acusado,
inputs="info_acusado",
outputs="genero_acusado",
),
node(
func=buscar_nacionalidad_acusado,
inputs="info_acusado",
outputs="nacionalida_acusado",
),
node(
func=buscar_est_civil_acusado,
inputs="info_acusado",
outputs="est_civil_acusado",
),
node(
func=buscar_edad_acusado,
inputs="info_acusado",
outputs="edad_acusado",
),
node(
func=buscar_instruccion_acusado,
inputs="info_acusado",
outputs="instruccion_acusado",
),
node(
func=buscar_domicilio_acusado,
inputs="info_acusado",
outputs="domicilio_acusado",
),
node(
func=buscar_ocupacion_acusado,
inputs="info_acusado",
outputs="ocupacion_acusado",
),
node(
func=buscar_relacion,
inputs="info_acusado",
outputs="relacion",
),
node(
func=chequear_conv,
inputs="info_acusado",
outputs="convivencia",
),
node(
func=buscar_info_episodio,
inputs="data",
outputs="info_episodio",
),
node(
func=buscar_denuncia_anterior,
inputs="info_episodio",
outputs="denuncia_anterior",
),
node(
func=buscar_medidas_prot,
inputs="info_episodio",
outputs="medidas_prot",
),
node(
func=buscar_dia_hecho,
inputs="data",
outputs="dia_hecho",
),
node(
func=buscar_conclusiones,
inputs="data",
outputs="conclusiones",
),
node(
func=buscar_riesgo,
inputs="conclusiones",
outputs="riesgo",
),
node(
func=buscar_informe_final,
inputs="data",
outputs="informe_final",
),
node(
func=buscar_violencia_amb,
inputs="informe_final",
outputs="violencia_amb",
),
node(
func=buscar_violencia_econ,
inputs="informe_final",
outputs="violencia_econ",
),
node(
func=buscar_violencia_genero,
inputs="informe_final",
outputs="violencia_genero",
),
node(
func=buscar_violencia_psico,
inputs="informe_final",
outputs="violencia_psico",
),
node(
func=buscar_violencia_sex,
inputs="informe_final",
outputs="violencia_sex",
),
node(
func=buscar_violencia_simb,
inputs="informe_final",
outputs="violencia_simb",
),
node(
func=buscar_violencia_soc,
inputs="informe_final",
outputs="violencia_soc",
),
node(
func=buscar_hijos,
inputs="informe_final",
outputs="hijos",
),
node(
func=buscar_hijos_en_comun,
inputs="hijos",
outputs="hijos_en_comun",
),
node(
func=buscar_frecuencia,
inputs="informe_final",
outputs="frecuencia",
),
node(
func=buscar_dijo,
inputs="data",
outputs="frases_agresion",
),
node(
func=buscar_dijo_sin_comillas,
inputs="data",
outputs="frases_sin_comillas",
),
node(
func=buscar_comillas,
inputs="data",
outputs="frases_comillas",
),
node(
func=buscar_fecha_del_hecho,
inputs=["info_episodio", "fecha_denuncia"],
outputs="fecha_del_hecho",
),
node(
func=buscar_horario_hecho,
inputs="info_episodio",
outputs="horario_hecho",
),
node(
func=buscar_fecha_denuncia,
inputs="data",
outputs="fecha_denuncia",
),
node(
func=to_excel,
inputs=[
"violencia_de_genero",
"violencia_fisica",
"genero_denunciante",
"nacionalidad_denunciante",
"est_civil_denunciante",
"edad_denunciante",
"est_denunciante",
"domic_denunciante",
"villa_denunciante",
"ocupac_denunciante",
"genero_acusado",
"nacionalida_acusado",
"est_civil_acusado",
"edad_acusado",
"instruccion_acusado",
"domicilio_acusado",
"ocupacion_acusado",
"relacion",
"convivencia",
"denuncia_anterior",
"medidas_prot",
"dia_hecho",
"riesgo",
"violencia_psico",
"violencia_econ",
"violencia_sex",
"violencia_soc",
"violencia_amb",
"violencia_simb",
"hijos_en_comun",
"frecuencia",
"frases_sin_comillas",
"frases_agresion",
"frases_comillas",
"fecha_del_hecho",
"fecha_denuncia",
"horario_hecho",
"numero_legajo",
],
outputs="excel_ovd_data",
),
])
def test_different_output_list_order_not_equal(self):
first = node(identity, "input1", ["output1", "output2"], name="A")
second = node(identity, "input1", ["output2", "output1"], name="A")
assert first != second
"""Contents of hello_kedro.py"""
from kedro.io import DataCatalog, MemoryDataSet
from kedro.pipeline import node, Pipeline
from kedro.runner import SequentialRunner
# Prepare a data catalog
data_catalog = DataCatalog({"example_data": MemoryDataSet()})
# Prepare second node
def join_statements(greeting):
return f"{greeting} Kedro!"
join_statements_node = node(join_statements,
inputs="my_salutation",
outputs="my_message")
# Prepare first node
def return_greeting():
return "Bonjourno"
return_greeting_node = node(return_greeting,
inputs=None,
outputs="my_salutation")
# Assemble nodes into a pipeline
pipeline = Pipeline([join_statements_node, return_greeting_node])
def test_bad_input(func, expected):
with pytest.raises(TypeError, match=expected):
node(*func())
def create_pipeline(**kwargs):
return Pipeline(
[
#node(
# func=split_train_pool,
# inputs=dict(
# y_train_full="y_train_full",
# n_init="params:N_INIT"
# ),
# outputs=["full_id", "train_id", "pool_id"],
# tags=["sampling"]
#),
node(
func=truncate_dataset,
inputs=dict(
X_train_full="X_train_full",
y_train_full="y_train_full",
size="params:SIZE_ANALYSIS"
),
outputs=["X_train_trunc", "y_train_trunc"],
tags=["pre_sampling"]
),
node(
func=compute_gaussian_kernel,
inputs=dict(
X="X_train_trunc"
),
outputs="K_FIXE",
tags=["pre_sampling"]
),
node(
func=al_performances,
inputs=dict(
bs="params:BATCH_SEQ",
budget="params:BUDGET",
n_simu="params:N_SIMULATIONS",
X_train_full="X_train_trunc",
y_train_full="y_train_trunc",
X_test="X_test",
y_test="y_test",
K_FIXE="K_FIXE",
n_init="params:N_INIT"
),
outputs="al_perfs",
tags=["sampling", "active_sampling"]
),
node(
func=lambda_analysis,
inputs=dict(
b="params:BATCH_SIZE",
budget="params:BUDGET",
n_simu="params:N_SIMULATIONS",
X_train_full="X_train_trunc",
y_train_full="y_train_trunc",
X_test="X_test",
y_test="y_test",
n_init="params:N_INIT",
K_FIXE="K_FIXE"
),
outputs="al_lam_perfs",
tags=["sampling", "active_lambda_sampling"]
),
node(
func=pl_performances,
inputs=dict(
bs="params:BATCH_SEQ",
budget="params:BUDGET",
n_simu="params:N_SIMULATIONS",
X_train_full="X_train_trunc",
y_train_full="y_train_trunc",
X_test="X_test",
y_test="y_test",
n_init="params:N_INIT"
),
outputs="pl_perfs",
tags=["sampling", "passive_sampling"]
),
node(
func=b_descent_analysis,
inputs=dict(
budget="params:BUDGET",
n_simu="params:N_SIMULATIONS",
X_train_full="X_train_trunc",
y_train_full="y_train_trunc",
X_test="X_test",
y_test="y_test",
n_init="params:N_INIT",
b_descent_size="params:BATCH_DESCENT_SIZE",
K_FIXE="K_FIXE"
),
outputs="b_descent_perfs",
tags=["sampling", "active_descent_sampling"]
),
node(
func=b_ascent_analysis,
inputs=dict(
budget="params:BUDGET",
n_simu="params:N_SIMULATIONS",
X_train_full="X_train_trunc",
y_train_full="y_train_trunc",
X_test="X_test",
y_test="y_test",
n_init="params:N_INIT",
b_ascent_size="params:BATCH_DESCENT_SIZE",
K_FIXE="K_FIXE"
),
outputs="b_ascent_perfs",
tags=["sampling", "active_descent_sampling"]
),
]
)
def test_tag_nodes_single_tag(self):
tagged_node = node(identity, "input", "output",
tags="hello").tag("world")
assert "hello" in tagged_node.tags
assert "world" in tagged_node.tags
assert len(tagged_node.tags) == 2
def test_updated_partial(self):
n = node(update_wrapper(partial(identity), identity), ["in"], ["out"])
assert str(n) == "identity([in]) -> [out]"
assert n.name == "identity([in]) -> [out]"
assert n.short_name == "Identity"
