hex.genmodel.easy.error

Class CountingErrorConsumer

java.lang.Object

hex.genmodel.easy.EasyPredictModelWrapper.ErrorConsumer

hex.genmodel.easy.error.CountingErrorConsumer

All Implemented Interfaces:

java.io.Serializable

public class CountingErrorConsumer
extends EasyPredictModelWrapper.ErrorConsumer

An implementation of EasyPredictModelWrapper.ErrorConsumer counting number of each kind of error even received

See Also:

Serialized Form

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
static class	CountingErrorConsumer.Config

Constructor Summary

Constructors

Constructor and Description
CountingErrorConsumer(GenModel model)
CountingErrorConsumer(GenModel model, CountingErrorConsumer.Config config)

Method Summary

Modifier and Type	Method and Description
void	dataTransformError(java.lang.String columnName, java.lang.Object value, java.lang.String message) Observe transformation error for data from the predicted dataset.
long	getDataTransformationErrorsCount()
java.util.Map<java.lang.String,java.util.concurrent.atomic.AtomicLong>	getDataTransformationErrorsCountPerColumn() An unmodifiable, thread-safe map of all columns with counts of data transformation errors observed.
long	getTotalUnknownCategoricalLevelsSeen() Counts and returns all previously unseen categorical variables across all columns.
java.util.Map<java.lang.String,java.util.concurrent.atomic.AtomicLong>	getUnknownCategoricalsPerColumn() Returns a thread-safe Map with column names as keys and number of observed unknown categorical values associated with each column.
java.util.Map<java.lang.Object,java.util.concurrent.atomic.AtomicLong>	getUnseenCategoricals(java.lang.String column)
void	unseenCategorical(java.lang.String columnName, java.lang.Object value, java.lang.String message) Previously unseen categorical level has been detected

Methods inherited from class java.lang.Object

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

Constructor Detail

CountingErrorConsumer

public CountingErrorConsumer(GenModel model)

Parameters:

model - An instance of GenModel

CountingErrorConsumer

public CountingErrorConsumer(GenModel model,
                             CountingErrorConsumer.Config config)

Parameters:

model - An instance of GenModel

config - An instance of CountingErrorConsumer.Config

Method Detail

dataTransformError

public void dataTransformError(java.lang.String columnName,
                               java.lang.Object value,
                               java.lang.String message)

Description copied from class: EasyPredictModelWrapper.ErrorConsumer

Observe transformation error for data from the predicted dataset.

Specified by:

dataTransformError in class EasyPredictModelWrapper.ErrorConsumer

Parameters:

columnName - Name of the column for which the error is raised

value - Original value that could not be transformed properly

message - Transformation error message

unseenCategorical

public void unseenCategorical(java.lang.String columnName,
                              java.lang.Object value,
                              java.lang.String message)

Description copied from class: EasyPredictModelWrapper.ErrorConsumer

Previously unseen categorical level has been detected

Specified by:

unseenCategorical in class EasyPredictModelWrapper.ErrorConsumer

Parameters:

columnName - Name of the column to which the categorical value belongs

value - Original value

message - Reason and/or actions taken

getTotalUnknownCategoricalLevelsSeen

public long getTotalUnknownCategoricalLevelsSeen()

Counts and returns all previously unseen categorical variables across all columns. Results may vary when called during prediction phase.

Returns:

A sum of all previously unseen categoricals across all columns

getUnknownCategoricalsPerColumn

public java.util.Map<java.lang.String,java.util.concurrent.atomic.AtomicLong> getUnknownCategoricalsPerColumn()

Returns a thread-safe Map with column names as keys and number of observed unknown categorical values associated with each column. The map returned is a direct reference to the backing this CountingErrorConsumer. Iteration during prediction phase may end up with undefined results. All the columns are listed.

Returns:

A thread-safe map.

getUnseenCategoricals

public java.util.Map<java.lang.Object,java.util.concurrent.atomic.AtomicLong> getUnseenCategoricals(java.lang.String column)

getDataTransformationErrorsCountPerColumn

public java.util.Map<java.lang.String,java.util.concurrent.atomic.AtomicLong> getDataTransformationErrorsCountPerColumn()

An unmodifiable, thread-safe map of all columns with counts of data transformation errors observed. The map returned is a direct reference to the backing this CountingErrorConsumer. Iteration during prediction phase may end up with undefined results.

Returns:

A thread-safe instance of Map

getDataTransformationErrorsCount

public long getDataTransformationErrorsCount()

Returns:

Number of transformation errors found