def test_map(self):
result = StringMap(
labels={'a': 1.0},
inputCol='key_col',
outputCol='value_col',
).transform(self.input)
expected = StringMapTest.spark.createDataFrame([['a', 'b', 1.0]], OUTPUT_SCHEMA)
assert_df(expected, result)
def test_map_from_dataframe(self):
labels_df = StringMapTest.spark.createDataFrame([['a', 1.0]], 'key_col: string, value_col: double')
result = StringMap.from_dataframe(
labels_df=labels_df,
inputCol='key_col',
outputCol='value_col'
).transform(self.input)
expected = StringMapTest.spark.createDataFrame([['a', 'b', 1.0]], OUTPUT_SCHEMA)
assert_df(expected, result)
def test_serialize_to_bundle(self):
string_map = StringMap({'a': 1.0}, 'key_col', 'value_col')
pipeline = Pipeline(stages=[string_map]).fit(self.input)
pipeline_file = os.path.join(os.path.dirname(__file__), '..', '..',
'target', 'test_serialize_to_bundle-pipeline.zip')
_serialize_to_file(pipeline_file, self.input, pipeline)
deserialized_pipeline = _deserialize_from_file(pipeline_file)
result = deserialized_pipeline.transform(self.input)
expected = self.spark.createDataFrame([['a', 'b', 1.0]], OUTPUT_SCHEMA)
assert_df(expected, result)
def test_map_custom_default_value(self):
result = StringMap(
labels={'z': 1.0},
inputCol='key_col',
outputCol='value_col',
handleInvalid='keep',
defaultValue=-1.0
).transform(self.input)
expected = StringMapTest.spark.createDataFrame([['a', 'b', -1.0]], OUTPUT_SCHEMA)
assert_df(expected, result)
def test_serialize_to_bundle(self):
string_map = StringMap(
labels={'a': 1.0},
inputCol='key_col',
outputCol='value_col',
)
pipeline = Pipeline(stages=[string_map]).fit(self.input)
serialization_dataset = pipeline.transform(self.input)
jar_file_path = _serialize_to_file(pipeline, serialization_dataset)
deserialized_pipeline = _deserialize_from_file(jar_file_path)
result = deserialized_pipeline.transform(self.input)
expected = StringMapTest.spark.createDataFrame([['a', 'b', 1.0]], OUTPUT_SCHEMA)
assert_df(expected, result)
def test_map_default_value(self):
result = StringMap({'z': 1.0}, 'key_col', 'value_col', handleInvalid='keep').transform(self.input)
expected = self.spark.createDataFrame([['a', 'b', 0.0]], OUTPUT_SCHEMA)
assert_df(expected, result)
def test_map(self):
result = StringMap({'a': 1.0}, 'key_col', 'value_col').transform(self.input)
expected = self.spark.createDataFrame([['a', 'b', 1.0]], OUTPUT_SCHEMA)
assert_df(expected, result)
