def compute(
self,
verbose: bool = False,
lineage: bool = False
) -> Union[float, np.array, Tuple[Union[float, np.array], str]]:
if self.result_var is None:
self.script = DMLScript()
self.script.build_code(self)
if lineage:
result_variables, ltrace = self.script.execute(lineage)
else:
result_variables = self.script.execute(lineage)
if self.output_type == OutputType.DOUBLE:
self.result_var = result_variables.getDouble(
self.script.out_var_name)
elif self.output_type == OutputType.MATRIX:
self.result_var = matrix_block_to_numpy(
get_gateway().jvm,
result_variables.getMatrixBlock(self.script.out_var_name))
if verbose:
print(self.script.dml_script)
# TODO further info
if lineage:
return self.result_var, ltrace
else:
return self.result_var
def execute(
self,
lineage: bool = False
) -> Union[JavaObject, Tuple[JavaObject, str]]:
"""If not already created, create a preparedScript from our DMLCode, pass python local data to our prepared
script, then execute our script and return the resultVariables
:return: resultVariables of our execution
"""
# we could use the gateway directly, non defined functions will be automatically
# sent to the entry_point, but this is safer
gateway = get_gateway()
entry_point = gateway.entry_point
if self.prepared_script is None:
input_names = self.inputs.keys()
connection = entry_point.getConnection()
self.prepared_script = connection.prepareScript(
self.dml_script, _list_to_java_array(input_names),
_list_to_java_array([self.out_var_name]))
for (name, input_node) in self.inputs.items():
input_node.pass_python_data_to_prepared_script(
gateway.jvm, name, self.prepared_script)
if lineage:
connection.setLineage(True)
ret = self.prepared_script.executeScript()
if lineage:
return ret, self.prepared_script.getLineageTrace(self.out_var_name)
return ret
def _list_to_java_array(
py_list: Union[Collection[str], KeysView[str]]) -> JavaArray:
"""Convert python collection to java array.
:param py_list: python collection
:return: java array
"""
gateway = get_gateway()
array = gateway.new_array(gateway.jvm.java.lang.String, len(py_list))
for (i, e) in enumerate(py_list):
array[i] = e
return array
def getlineage(self) -> str:
gateway = get_gateway()
entry_point = gateway.entry_point
if self.prepared_script is None:
input_names = self.inputs.keys()
connection = entry_point.getConnection()
self.prepared_script = connection.prepareScript(
self.dml_script, _list_to_java_array(input_names),
_list_to_java_array([self.out_var_name]))
for (name, input_node) in self.inputs.items():
input_node.pass_python_data_to_prepared_script(
gateway.jvm, name, self.prepared_script)
connection.setLineage(True)
ret = self.prepared_script.executeScript()
lineage = self.prepared_script.getLineageTrace(self.out_var_name)
return lineage