def test_sink_raise(self):
with pytest.raises(ValueError, match="sink must be a Sink instance"):
FeatureSetPipeline(
spark_client=SparkClient(),
source=Mock(
spec=Source,
readers=[
TableReader(id="source_a", database="db", table="table",),
],
query="select * from source_a",
),
feature_set=Mock(
spec=FeatureSet,
name="feature_set",
entity="entity",
description="description",
features=[
Feature(
name="user_id",
description="The user's Main ID or device ID",
dtype=DataType.FLOAT,
),
Feature(
name="ts",
description="The timestamp feature",
dtype=DataType.TIMESTAMP,
),
],
key_columns=["user_id"],
timestamp_column="ts",
),
sink=Mock(writers=[HistoricalFeatureStoreWriter(db_config=None)],),
)
def test_construct(
self, feature_set_dataframe, fixed_windows_output_feature_set_dataframe
):
# given
spark_client = SparkClient()
# arrange
feature_set = FeatureSet(
name="feature_set",
entity="entity",
description="description",
features=[
Feature(
name="feature1",
description="test",
transformation=SparkFunctionTransform(
functions=[
Function(F.avg, DataType.FLOAT),
Function(F.stddev_pop, DataType.FLOAT),
]
).with_window(
partition_by="id",
order_by=TIMESTAMP_COLUMN,
mode="fixed_windows",
window_definition=["2 minutes", "15 minutes"],
),
),
Feature(
name="divided_feature",
description="unit test",
dtype=DataType.FLOAT,
transformation=CustomTransform(
transformer=divide, column1="feature1", column2="feature2",
),
),
],
keys=[
KeyFeature(
name="id",
description="The user's Main ID or device ID",
dtype=DataType.INTEGER,
)
],
timestamp=TimestampFeature(),
)
output_df = (
feature_set.construct(feature_set_dataframe, client=spark_client)
.orderBy(feature_set.timestamp_column)
.select(feature_set.columns)
)
target_df = fixed_windows_output_feature_set_dataframe.orderBy(
feature_set.timestamp_column
).select(feature_set.columns)
# assert
assert_dataframe_equality(output_df, target_df)
def test_construct_rolling_windows_with_end_date(
self,
feature_set_dataframe,
rolling_windows_output_feature_set_dataframe_base_date,
):
# given
spark_client = SparkClient()
# arrange
feature_set = AggregatedFeatureSet(
name="feature_set",
entity="entity",
description="description",
features=[
Feature(
name="feature1",
description="test",
transformation=AggregatedTransform(functions=[
Function(F.avg, DataType.DOUBLE),
Function(F.stddev_pop, DataType.DOUBLE),
], ),
),
Feature(
name="feature2",
description="test",
transformation=AggregatedTransform(functions=[
Function(F.avg, DataType.DOUBLE),
Function(F.stddev_pop, DataType.DOUBLE),
], ),
),
],
keys=[
KeyFeature(
name="id",
description="The user's Main ID or device ID",
dtype=DataType.INTEGER,
)
],
timestamp=TimestampFeature(),
).with_windows(definitions=["1 day", "1 week"])
# act
output_df = feature_set.construct(
feature_set_dataframe, client=spark_client,
end_date="2016-04-18").orderBy("timestamp")
target_df = rolling_windows_output_feature_set_dataframe_base_date.orderBy(
feature_set.timestamp_column).select(feature_set.columns)
# assert
assert_dataframe_equality(output_df, target_df)
def test_output_columns(self):
# arrange
h3_feature = Feature(
name="new_feature",
description="unit test",
dtype=DataType.STRING,
transformation=H3HashTransform(
h3_resolutions=[6, 7, 8, 9, 10, 11, 12],
lat_column="lat",
lng_column="lng",
),
)
target_columns = [
"lat_lng__h3_hash__6",
"lat_lng__h3_hash__7",
"lat_lng__h3_hash__8",
"lat_lng__h3_hash__9",
"lat_lng__h3_hash__10",
"lat_lng__h3_hash__11",
"lat_lng__h3_hash__12",
]
# act
output_columns = h3_feature.get_output_columns()
# assert
assert sorted(output_columns) == sorted(target_columns)
def test_feature_set_with_invalid_feature(self, key_id, timestamp_c,
dataframe):
spark_client = SparkClient()
with pytest.raises(ValueError):
AggregatedFeatureSet(
name="name",
entity="entity",
description="description",
features=[
Feature(
name="feature1",
description="test",
transformation=SparkFunctionTransform(functions=[
Function(functions.avg, DataType.FLOAT)
], ).with_window(
partition_by="id",
mode="row_windows",
window_definition=["2 events"],
),
),
],
keys=[key_id],
timestamp=timestamp_c,
).construct(dataframe, spark_client)
def test_feature_transform_with_distinct_empty_subset(
self, timestamp_c, feature_set_with_distinct_dataframe):
spark_client = SparkClient()
with pytest.raises(ValueError,
match="The distinct subset param can't be empty."):
AggregatedFeatureSet(
name="name",
entity="entity",
description="description",
features=[
Feature(
name="feature",
description="test",
transformation=AggregatedTransform(functions=[
Function(functions.sum, DataType.INTEGER)
]),
),
],
keys=[
KeyFeature(name="h3",
description="test",
dtype=DataType.STRING)
],
timestamp=timestamp_c,
).with_windows(["3 days"]).with_distinct(
subset=[],
keep="first").construct(feature_set_with_distinct_dataframe,
spark_client,
end_date="2020-01-10")
def test_feature_transform(self, feature_set_dataframe):
test_feature = Feature(
name="feature1_over_feature2",
description="unit test",
dtype=DataType.FLOAT,
transformation=SQLExpressionTransform(expression="feature1/feature2"),
)
df = test_feature.transform(feature_set_dataframe)
assert all(
[
a == b
for a, b in zip(
df.columns,
[
"feature1",
"feature2",
"id",
"timestamp",
"feature1_over_feature2",
],
)
]
)
def test_feature_transform_with_distinct(
self,
timestamp_c,
feature_set_with_distinct_dataframe,
target_with_distinct_dataframe,
):
spark_client = SparkClient()
fs = (AggregatedFeatureSet(
name="name",
entity="entity",
description="description",
features=[
Feature(
name="feature",
description="test",
transformation=AggregatedTransform(
functions=[Function(functions.sum, DataType.INTEGER)]),
),
],
keys=[
KeyFeature(name="h3",
description="test",
dtype=DataType.STRING)
],
timestamp=timestamp_c,
).with_windows(["3 days"]).with_distinct(subset=["id"], keep="last"))
# assert
output_df = fs.construct(feature_set_with_distinct_dataframe,
spark_client,
end_date="2020-01-10")
assert_dataframe_equality(output_df, target_with_distinct_dataframe)
def test_aggregations_with_filter_expression(self, spark_context):
# arrange
test_feature = Feature(
name="feature_with_filter",
description="unit test",
transformation=AggregatedTransform(
functions=[
Function(functions.avg, DataType.DOUBLE),
Function(functions.min, DataType.DOUBLE),
Function(functions.max, DataType.DOUBLE),
],
filter_expression="type = 'a'",
),
from_column="feature",
)
target_aggregations = [
agg(
functions.when(functions.expr("type = 'a'"),
functions.col("feature")))
for agg in [functions.avg, functions.min, functions.max]
]
# act
output_aggregations = [
agg.function for agg in test_feature.transformation.aggregations
]
# assert
# cast to string to compare the columns definitions because direct column
# comparison was not working
assert str(target_aggregations) == str(output_aggregations)
def test_h3_feature_set(self, h3_input_df, h3_target_df):
spark_client = SparkClient()
feature_set = AggregatedFeatureSet(
name="h3_test",
entity="h3geolocation",
description="Test",
keys=[
KeyFeature(
name="h3_id",
description="The h3 hash ID",
dtype=DataType.DOUBLE,
transformation=H3HashTransform(
h3_resolutions=[6, 7, 8, 9, 10, 11, 12],
lat_column="lat",
lng_column="lng",
).with_stack(),
)
],
timestamp=TimestampFeature(),
features=[
Feature(
name="house_id",
description="Count of house ids over a day.",
transformation=AggregatedTransform(
functions=[Function(F.count, DataType.BIGINT)]),
),
],
).with_windows(definitions=["1 day"])
output_df = feature_set.construct(h3_input_df,
client=spark_client,
end_date="2016-04-14")
assert_dataframe_equality(output_df, h3_target_df)
def test_construct_without_window(
self,
feature_set_dataframe,
target_df_without_window,
):
# given
spark_client = SparkClient()
# arrange
feature_set = AggregatedFeatureSet(
name="feature_set",
entity="entity",
description="description",
features=[
Feature(
name="feature1",
description="test",
dtype=DataType.DOUBLE,
transformation=AggregatedTransform(
functions=[Function(F.avg, DataType.DOUBLE)]),
),
Feature(
name="feature2",
description="test",
dtype=DataType.FLOAT,
transformation=AggregatedTransform(
functions=[Function(F.count, DataType.BIGINT)]),
),
],
keys=[
KeyFeature(
name="id",
description="The user's Main ID or device ID",
dtype=DataType.INTEGER,
)
],
timestamp=TimestampFeature(from_column="fixed_ts"),
)
# act
output_df = feature_set.construct(feature_set_dataframe,
client=spark_client)
# assert
assert_dataframe_equality(output_df, target_df_without_window)
def test_blank_aggregation(self, feature_set_dataframe):
with pytest.raises(ValueError):
Feature(
name="feature1",
description="unit test",
transformation=AggregatedTransform(
functions=[Function(func="", data_type="")]),
)
def test_unsupported_aggregation(self, feature_set_dataframe):
with pytest.raises(TypeError):
Feature(
name="feature1",
description="unit test",
transformation=AggregatedTransform(
functions=[Function("median", DataType.DOUBLE)]),
)
def test_blank_transformer(self, feature_set_dataframe):
with pytest.raises(ValueError):
Feature(
name="feature",
description="unit test",
dtype=DataType.BIGINT,
transformation=CustomTransform(transformer=None),
)
def test_feature_transform_invalid_output(self, feature_set_dataframe):
with pytest.raises(Exception):
test_feature = Feature(
name="feature1_plus_a",
description="unit test",
dtype=DataType.FLOAT,
transformation=SQLExpressionTransform(expression="feature2 + a"),
)
test_feature.transform(feature_set_dataframe).collect()
def test__get_features_columns(self):
# arrange
feature_1 = Feature("feature1", "description", DataType.FLOAT)
feature_1.get_output_columns = Mock(return_value=["col_a", "col_b"])
feature_2 = Feature("feature2", "description", DataType.FLOAT)
feature_2.get_output_columns = Mock(return_value=["col_c"])
feature_3 = Feature("feature3", "description", DataType.FLOAT)
feature_3.get_output_columns = Mock(return_value=["col_d"])
target_features_columns = ["col_a", "col_b", "col_c", "col_d"]
# act
result_features_columns = FeatureSet._get_features_columns(
feature_1, feature_2, feature_3)
# assert
assert target_features_columns == result_features_columns
def test_feature_transform_with_dtype(self, feature_set_dataframe):
test_feature = Feature(
name="feature",
description="unit test",
dtype=DataType.TIMESTAMP,
)
df = test_feature.transform(feature_set_dataframe)
assert dict(df.dtypes).get("feature") == "timestamp"
def test_feature_get_output_columns_without_transformations(self):
test_feature = Feature(
name="feature",
from_column="origin",
description="unit test",
dtype=DataType.BIGINT,
)
assert test_feature.get_output_columns() == [test_feature.name]
def test_output_columns(self):
test_feature = Feature(
name="feature1_over_feature2",
description="unit test",
dtype=DataType.FLOAT,
transformation=SQLExpressionTransform(expression="feature1/feature2"),
)
df_columns = test_feature.get_output_columns()
assert all([a == b for a, b in zip(df_columns, ["feature1_over_feature2"],)])
def test_run_with_repartition(self, spark_session):
test_pipeline = FeatureSetPipeline(
spark_client=SparkClient(),
source=Mock(
spec=Source,
readers=[TableReader(id="source_a", database="db", table="table",)],
query="select * from source_a",
),
feature_set=Mock(
spec=FeatureSet,
name="feature_set",
entity="entity",
description="description",
keys=[
KeyFeature(
name="user_id",
description="The user's Main ID or device ID",
dtype=DataType.INTEGER,
)
],
timestamp=TimestampFeature(from_column="ts"),
features=[
Feature(
name="listing_page_viewed__rent_per_month",
description="Average of something.",
transformation=SparkFunctionTransform(
functions=[
Function(functions.avg, DataType.FLOAT),
Function(functions.stddev_pop, DataType.FLOAT),
],
).with_window(
partition_by="user_id",
order_by=TIMESTAMP_COLUMN,
window_definition=["7 days", "2 weeks"],
mode="fixed_windows",
),
),
],
),
sink=Mock(
spec=Sink, writers=[HistoricalFeatureStoreWriter(db_config=None)],
),
)
# feature_set need to return a real df for streaming validation
sample_df = spark_session.createDataFrame([{"a": "x", "b": "y", "c": "3"}])
test_pipeline.feature_set.construct.return_value = sample_df
test_pipeline.run(partition_by=["id"])
test_pipeline.source.construct.assert_called_once()
test_pipeline.feature_set.construct.assert_called_once()
test_pipeline.sink.flush.assert_called_once()
test_pipeline.sink.validate.assert_called_once()
def test_feature_transform(self, feature_set_dataframe, target_df_spark):
test_feature = Feature(
name="feature",
description="unit test",
transformation=SparkFunctionTransform(
functions=[Function(functions.cos, DataType.DOUBLE)], ),
from_column="feature1",
)
output_df = test_feature.transform(feature_set_dataframe)
assert_dataframe_equality(output_df, target_df_spark)
def test_agg_feature_set_with_window(self, key_id, timestamp_c, dataframe,
rolling_windows_agg_dataframe):
spark_client = SparkClient()
fs = AggregatedFeatureSet(
name="name",
entity="entity",
description="description",
features=[
Feature(
name="feature1",
description="unit test",
transformation=AggregatedTransform(
functions=[Function(functions.avg, DataType.FLOAT)]),
),
Feature(
name="feature2",
description="unit test",
transformation=AggregatedTransform(
functions=[Function(functions.avg, DataType.FLOAT)]),
),
],
keys=[key_id],
timestamp=timestamp_c,
).with_windows(definitions=["1 week"])
# raises without end date
with pytest.raises(ValueError):
_ = fs.construct(dataframe, spark_client)
# filters with date smaller then mocked max
output_df = fs.construct(dataframe,
spark_client,
end_date="2016-04-17")
assert output_df.count() < rolling_windows_agg_dataframe.count()
output_df = fs.construct(dataframe,
spark_client,
end_date="2016-05-01")
assert_dataframe_equality(output_df, rolling_windows_agg_dataframe)
def test_negative_windows(self, feature_set_dataframe):
with pytest.raises(KeyError):
Feature(
name="feature1",
description="unit test",
transformation=SparkFunctionTransform(
functions=[Function(functions.avg,
DataType.DOUBLE)], ).with_window(
partition_by="id",
mode="fixed_windows",
window_definition=["-2 weeks"],
),
).transform(feature_set_dataframe)
def test_feature_transform_no_from_column(self, feature_set_dataframe):
test_feature = Feature(
name="feature",
description="unit test feature without transformation",
dtype=DataType.BIGINT,
)
df = test_feature.transform(feature_set_dataframe)
assert all([
a == b for a, b in zip(df.columns, feature_set_dataframe.columns)
])
def test_args_without_transformation(self):
test_feature = Feature(
name="feature",
from_column="origin",
description="unit test",
dtype=DataType.BIGINT,
)
assert test_feature.name == "feature"
assert test_feature.from_column == "origin"
assert test_feature.description == "unit test"
assert test_feature.dtype == DataType.BIGINT
def test_feature_transform(self, feature_set_dataframe, target_df_agg):
test_feature = Feature(
name="feature1",
description="unit test",
transformation=AggregatedTransform(functions=[
Function(functions.avg, DataType.DOUBLE),
Function(functions.stddev_pop, DataType.DOUBLE),
]),
)
# aggregated feature transform won't run transformations
# and depends on the feature set
with pytest.raises(NotImplementedError):
_ = test_feature.transform(feature_set_dataframe)
def test_feature_transform_output(self, feature_set_dataframe):
test_feature = Feature(
name="feature1_over_feature2",
description="unit test",
dtype=DataType.FLOAT,
transformation=SQLExpressionTransform(expression="feature1/feature2"),
)
df = test_feature.transform(feature_set_dataframe).collect()
assert df[0]["feature1_over_feature2"] == 1
assert df[1]["feature1_over_feature2"] == 1
assert df[2]["feature1_over_feature2"] == 1
assert df[3]["feature1_over_feature2"] == 1
def test_output_columns(self, feature_set_dataframe):
test_feature = Feature(
name="feature",
description="unit test",
dtype=DataType.BIGINT,
transformation=CustomTransform(
transformer=divide, column1="feature1", column2="feature2",
),
)
df_columns = test_feature.get_output_columns()
assert isinstance(df_columns, list)
assert df_columns == ["feature"]
def test_columns_not_in_dataframe(self, spark_context, spark_session):
# arrange
input_df = create_df_from_collection(self.input_data, spark_context,
spark_session)
feature = Feature(
name="id",
description="stack transformation",
dtype=DataType.STRING,
transformation=StackTransform("id_c", "id_d"),
)
# act and assert
with pytest.raises(ValueError,
match="Columns not found, columns in df: "):
feature.transform(input_df)
def test_custom_transform_output(self, feature_set_dataframe):
test_feature = Feature(
name="feature",
description="unit test",
dtype=DataType.BIGINT,
transformation=CustomTransform(
transformer=divide, column1="feature1", column2="feature2",
),
)
df = test_feature.transform(feature_set_dataframe).collect()
assert df[0]["feature"] == 1
assert df[1]["feature"] == 1
assert df[2]["feature"] == 1
assert df[3]["feature"] == 1