com.pervasive.datarush.operators.model

Class GetModel<T>

java.lang.Object

com.pervasive.datarush.operators.AbstractLogicalOperator

com.pervasive.datarush.operators.StreamingOperator

com.pervasive.datarush.operators.ExecutableOperator

com.pervasive.datarush.operators.model.GetModel<T>

Type Parameters:

T - the type of the model object

All Implemented Interfaces:

LogicalOperator, SinkOperator<AbstractModelPort<T>>

Direct Known Subclasses:

GetPMML

public class GetModel<T>
extends ExecutableOperator
implements SinkOperator<AbstractModelPort<T>>

Provides a way to update an in-memory reference to a model object. This is typically used when a graph is embedded within a larger context where we need to obtain a handle to a model object. This operator is always implicitly non-parallel; if the source is parallel, a merge must be performed and thus a mergeHandler must be specified.

NOTE: this operator is non-parallel

Constructor Summary

Constructors

Constructor and Description
GetModel(LogicalPortFactory<? extends AbstractModelPort<T>> factory) Gets a model from a graph.

Method Summary

Modifier and Type	Method and Description
protected ExecutableOperator	cloneForExecution() Performs a deep copy of the operator for execution.
protected void	computeMetadata(StreamingMetadataContext ctx) Implementations must adhere to the following contracts
protected void	execute(ExecutionContext ctx) Executes the operator.
AbstractModelPort<T>	getInput() The input port
ModelMergeHandler<T>	getMergeHandler() Returns the merge handler to be used to merge model fragments into a single model.
T	getModel() Following graph execution, will return an in-memory model object.
Reference<T>	getModelReference() For advanced use only; returns the reference location where the model is to be set.
void	setMergeHandler(ModelMergeHandler<T> mergeHandler) Sets the merge handler to be used to merge model fragments into a single model.
void	setModelReference(Reference<T> reference) For advanced use only; specifies a reference location where the model is to be set.

Methods inherited from class com.pervasive.datarush.operators.ExecutableOperator

getNumInputCopies, getPortSettings, handleInactiveOutput

Methods inherited from class com.pervasive.datarush.operators.AbstractLogicalOperator

disableParallelism, getInputPorts, getOutputPorts, newInput, newInput, newOutput, newRecordInput, newRecordInput, newRecordOutput, notifyError

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Methods inherited from interface com.pervasive.datarush.operators.LogicalOperator

disableParallelism, getInputPorts, getOutputPorts

Constructor Detail

GetModel

public GetModel(LogicalPortFactory<? extends AbstractModelPort<T>> factory)

Gets a model from a graph.

Parameters:

factory - the factory for the model port.

Method Detail

getInput

public final AbstractModelPort<T> getInput()

The input port

Specified by:

getInput in interface SinkOperator<AbstractModelPort<T>>

Returns:

the input port

getModel

public final T getModel()

Following graph execution, will return an in-memory model object. Behavior is undefined if this method is invoked prior to successful graph completion.

Returns:

the output model

getModelReference

public final Reference<T> getModelReference()

For advanced use only; returns the reference location where the model is to be set.

Returns:

the reference location

setModelReference

public final void setModelReference(Reference<T> reference)

For advanced use only; specifies a reference location where the model is to be set.

Parameters:

reference - the reference location

getMergeHandler

public final ModelMergeHandler<T> getMergeHandler()

Returns the merge handler to be used to merge model fragments into a single model. This value must be specified if the upstream operator is parallel.

Returns:

the merge handler to use

setMergeHandler

public final void setMergeHandler(ModelMergeHandler<T> mergeHandler)

Sets the merge handler to be used to merge model fragments into a single model. This value must be specified if the upstream operator is parallel.

Parameters:

mergeHandler - the merge handler to use

computeMetadata

protected final void computeMetadata(StreamingMetadataContext ctx)

Description copied from class: StreamingOperator

Implementations must adhere to the following contracts

General

Regardless of input ports/output port types, all implementations must do the following:

1. Validation. Validation of configuration should always be performed first.
2. Declare parallelizability.. Implementations must declare parallelizability by calling StreamingMetadataContext.parallelize(ParallelismStrategy).

Input record ports

Implementations with input record ports must declare the following:

1. Required data ordering:
Implementations that have data ordering requirements must declare them by calling RecordPort#setRequiredDataOrdering, otherwise data may arrive in any order.
2. Required data distribution (only applies to parallelizable operators):
Implementations that have data distribution requirements must declare them by calling RecordPort#setRequiredDataDistribution, otherwise data will arrive in an unspecified partial distribution.

Note that if the upstream operator's output distribution/ordering is compatible with those required, we avoid a re-sort/re-distribution which is generally a very large savings from a performance standpoint. In addition, some operators may chose to query the upstream output distribution/ordering by calling RecordPort#getSourceDataDistribution and RecordPort#getSourceDataOrdering. These should be viewed as a hints to help chose a more efficient algorithm. In such cases, though, operators must still declare data ordering and data distribution requirements; otherwise there is no guarantee that data will arrive sorted/distributed as required.

Output record ports

Implementations with output record ports must declare the following:

1. Type: Implementations must declare their output type by calling RecordPort#setType.

Implementations with output record ports may declare the following:

1. Output data ordering: Implementations that can make guarantees as to their output ordering may do so by calling RecordPort#setOutputDataOrdering
2. Output data distribution (only applies to parallelizable operators): Implementations that can make guarantees as to their output distribution may do so by calling RecordPort#setOutputDataDistribution

Note that both of these properties are optional; if unspecified, performance may suffer since the framework may unnecessarily re-sort/re-distributed the data.

Input model ports

In general, there is nothing special to declare for input model ports. Models are implicitly duplicated to all partitions when going from non-parallel to parallel operators. The case of a model going from a parallel to a non-parallel node is a special case of a "model reducer" operator. In the case of a model reducer, the downstream operator, must declare the following:

1. Merge handler: Model reducers must declare a merge handler by calling AbstractModelPort#setMergeHandler.

Note that MergeModel is a convenient, re-usable model reducer, parameterized with a merge-handler.

Output model ports

SimpleModelPort's have no associated metadata and therefore there is never any output metadata to declare. PMMLPort's, on the other hand, do have associated metadata. For all PMMLPorts, implementations must declare the following:

1. pmmlModelSpec: Implementations must declare the PMML model spec by calling PMMLPort.setPMMLModelSpec.

Specified by:

computeMetadata in class StreamingOperator

Parameters:

ctx - the context

cloneForExecution

protected final ExecutableOperator cloneForExecution()

Description copied from class: ExecutableOperator

Performs a deep copy of the operator for execution. The default implementation is implemented in terms of JSON serialization: we perform a JSON serialization followed by a JSON deserialization. As a best-practice, operator implementations should not override this method. If they must override, though, then they must guarantee that cloneForExecution copies any instance variables that are modified by execute.

Overrides:

cloneForExecution in class ExecutableOperator

Returns:

a deep copy of this operator

execute

protected final void execute(ExecutionContext ctx)

Description copied from class: ExecutableOperator

Executes the operator. Implementations should adhere to the following contracts:

1. Following execution, all input ports must be at end-of-data.
2. Following execution, all output ports must be at end-of-data.

Specified by:

execute in class ExecutableOperator

Parameters:

ctx - context in which to lookup physical ports bound to logical ports