def predict(
self,
parameter_values_pandas_frame,
t=None,
context_values_pandas_frame=None,
objective_name=None) -> Prediction: # pylint: disable=unused-argument
# TODO: make this streaming and/or using arrow.
#
if context_values_pandas_frame is not None:
raise NotImplementedError(
"Context not currently supported on remote optimizers")
feature_values_dict = parameter_values_pandas_frame.to_dict(
orient='list')
prediction_request = OptimizerMonitoringService_pb2.PredictRequest(
OptimizerHandle=self.
optimizer_handle_for_optimizer_monitoring_service,
Features=OptimizerMonitoringService_pb2.Features(
FeaturesJsonString=json.dumps(feature_values_dict)))
prediction_response = self._optimizer_monitoring_stub.Predict(
prediction_request)
# To be compliant with the OptimizerBase, we need to recover a single Prediction object and return it.
#
objective_predictions_pb2 = prediction_response.ObjectivePredictions
assert len(objective_predictions_pb2) == 1
only_prediction_pb2 = objective_predictions_pb2[0]
objective_name = only_prediction_pb2.ObjectiveName
valid_predictions_df = Prediction.dataframe_from_json(
only_prediction_pb2.PredictionDataFrameJsonString)
prediction = Prediction.create_prediction_from_dataframe(
objective_name=objective_name, dataframe=valid_predictions_df)
return prediction
def encode_optimization_problem(
optimization_problem: OptimizationProblem
) -> OptimizerMonitoringService_pb2.OptimizationProblem:
return OptimizerMonitoringService_pb2.OptimizationProblem(
ParameterSpace=OptimizerMonitoringServiceEncoder.encode_hypergrid(
optimization_problem.parameter_space),
ObjectiveSpace=OptimizerMonitoringServiceEncoder.encode_hypergrid(
optimization_problem.objective_space),
Objectives=[
OptimizerMonitoringService_pb2.Objective(
Name=objective.name, Minimize=objective.minimize)
for objective in optimization_problem.objectives
],
ContextSpace=None if optimization_problem.context_space is None
else OptimizerMonitoringServiceEncoder.encode_hypergrid(
optimization_problem.context_space))
def encode_empty_dimension(
dimension: EmptyDimension
) -> OptimizerMonitoringService_pb2.EmptyDimension:
assert isinstance(dimension, EmptyDimension)
return OptimizerMonitoringService_pb2.EmptyDimension(
Name=dimension.name,
DimensionType=OptimizerMonitoringServiceEncoder.
dimension_types_to_pb2_types[dimension.type])
def encode_subgrid(
subgrid: SimpleHypergrid.JoinedSubgrid
) -> OptimizerMonitoringService_pb2.GuestSubgrid:
assert isinstance(subgrid, SimpleHypergrid.JoinedSubgrid)
return OptimizerMonitoringService_pb2.GuestSubgrid(
Subgrid=OptimizerMonitoringServiceEncoder.encode_hypergrid(
subgrid.subgrid),
ExternalPivotDimension=OptimizerMonitoringServiceEncoder.
encode_dimension(subgrid.join_dimension))
def encode_categorical_dimension(
dimension: CategoricalDimension
) -> OptimizerMonitoringService_pb2.CategoricalDimension:
assert isinstance(dimension, CategoricalDimension)
return OptimizerMonitoringService_pb2.CategoricalDimension(
Name=dimension.name,
Values=[
OptimizerMonitoringServiceEncoder.encode_primitive_value(value)
for value in dimension.values
])
def encode_continuous_dimension(
dimension: ContinuousDimension
) -> OptimizerMonitoringService_pb2.ContinuousDimension:
assert isinstance(dimension, ContinuousDimension)
return OptimizerMonitoringService_pb2.ContinuousDimension(
Name=dimension.name,
Min=dimension.min,
Max=dimension.max,
IncludeMin=dimension.include_min,
IncludeMax=dimension.include_max)
def encode_primitive_value(
value: Union[int, float, bool, str]
) -> OptimizerMonitoringService_pb2.PrimitiveValue:
assert isinstance(value, (int, float, bool, str))
if isinstance(value, bool):
return OptimizerMonitoringService_pb2.PrimitiveValue(
BoolValue=value)
if isinstance(value, int):
return OptimizerMonitoringService_pb2.PrimitiveValue(
IntValue=value)
if isinstance(value, float):
return OptimizerMonitoringService_pb2.PrimitiveValue(
DoubleValue=value)
if isinstance(value, str):
return OptimizerMonitoringService_pb2.PrimitiveValue(
StringValue=value)
raise TypeError(
f"{value} is of type: {type(value)} but must be one of (int, float, bool, str)"
)
def encode_ordinal_dimension(
dimension: OrdinalDimension
) -> OptimizerMonitoringService_pb2.OrdinalDimension:
assert isinstance(dimension, OrdinalDimension)
return OptimizerMonitoringService_pb2.OrdinalDimension(
Name=dimension.name,
Ascending=dimension.ascending,
OrderedValues=[
OptimizerMonitoringServiceEncoder.encode_primitive_value(value)
for value in dimension.values
])
def encode_dimension(
dimension: Dimension) -> OptimizerMonitoringService_pb2.Dimension:
if isinstance(dimension, EmptyDimension):
return OptimizerMonitoringService_pb2.Dimension(
EmptyDimension=OptimizerMonitoringServiceEncoder.
encode_empty_dimension(dimension))
if isinstance(dimension, ContinuousDimension):
return OptimizerMonitoringService_pb2.Dimension(
ContinuousDimension=OptimizerMonitoringServiceEncoder.
encode_continuous_dimension(dimension))
if isinstance(dimension, DiscreteDimension):
return OptimizerMonitoringService_pb2.Dimension(
DiscreteDimension=OptimizerMonitoringServiceEncoder.
encode_discrete_dimension(dimension))
if isinstance(dimension, OrdinalDimension):
return OptimizerMonitoringService_pb2.Dimension(
OrdinalDimension=OptimizerMonitoringServiceEncoder.
encode_ordinal_dimension(dimension))
if isinstance(dimension, CategoricalDimension):
return OptimizerMonitoringService_pb2.Dimension(
CategoricalDimension=OptimizerMonitoringServiceEncoder.
encode_categorical_dimension(dimension))
if isinstance(dimension, CompositeDimension):
return OptimizerMonitoringService_pb2.Dimension(
CompositeDimension=OptimizerMonitoringServiceEncoder.
encode_composite_dimension(dimension))
raise TypeError(f"Unsupported dimension type: {type(dimension)}")
def encode_composite_dimension(
dimension: CompositeDimension
) -> OptimizerMonitoringService_pb2.CompositeDimension:
assert isinstance(dimension, CompositeDimension)
encoded_chunks = []
for chunk in dimension.enumerate_chunks():
if dimension.chunks_type is ContinuousDimension:
encoded_chunks.append(
OptimizerMonitoringService_pb2.Dimension(
ContinuousDimension=OptimizerMonitoringServiceEncoder.
encode_continuous_dimension(chunk)))
elif dimension.chunks_type is DiscreteDimension:
encoded_chunks.append(
OptimizerMonitoringService_pb2.Dimension(
DiscreteDimension=OptimizerMonitoringServiceEncoder.
encode_discrete_dimension(chunk)))
elif dimension.chunks_type is OrdinalDimension:
encoded_chunks.append(
OptimizerMonitoringService_pb2.Dimension(
OrdinalDimension=OptimizerMonitoringServiceEncoder.
encode_ordinal_dimension(chunk)))
elif dimension.chunks_type is CategoricalDimension:
encoded_chunks.append(
OptimizerMonitoringService_pb2.Dimension(
CategoricalDimension=OptimizerMonitoringServiceEncoder.
encode_categorical_dimension(chunk)))
else:
raise TypeError(
f"Unsupported chunk type: {dimension.chunks_type.__name__}"
)
return OptimizerMonitoringService_pb2.CompositeDimension(
Name=dimension.name,
ChunkType=OptimizerMonitoringServiceEncoder.
dimension_types_to_pb2_types[dimension.chunks_type],
Chunks=encoded_chunks)
def encode_hypergrid(
hypergrid: SimpleHypergrid
) -> OptimizerMonitoringService_pb2.SimpleHypergrid:
assert isinstance(hypergrid, SimpleHypergrid)
encoded_subgrids = []
for _, subgrids in hypergrid.joined_subgrids_by_pivot_dimension.items(
):
for subgrid in subgrids:
encoded_subgrid = OptimizerMonitoringServiceEncoder.encode_subgrid(
subgrid)
encoded_subgrids.append(encoded_subgrid)
return OptimizerMonitoringService_pb2.SimpleHypergrid(
Name=hypergrid.name,
Dimensions=[
OptimizerMonitoringServiceEncoder.encode_dimension(dimension)
for dimension in hypergrid.root_dimensions
],
GuestSubgrids=encoded_subgrids)
def encode_discrete_dimension(
dimension: DiscreteDimension
) -> OptimizerMonitoringService_pb2.DiscreteDimension:
assert isinstance(dimension, DiscreteDimension)
return OptimizerMonitoringService_pb2.DiscreteDimension(
Name=dimension.name, Min=dimension.min, Max=dimension.max)
def optimizer_handle_for_optimizer_monitoring_service(self):
return OptimizerMonitoringService_pb2.OptimizerHandle(Id=self.id)