eus.ixa.ixa.pipe.ml.sequence

Class SequenceLabelerEventStream

java.lang.Object

opennlp.tools.util.AbstractEventStream<SequenceLabelSample>

eus.ixa.ixa.pipe.ml.sequence.SequenceLabelerEventStream

All Implemented Interfaces:

AutoCloseable, opennlp.tools.util.ObjectStream<opennlp.tools.ml.model.Event>

public class SequenceLabelerEventStream
extends opennlp.tools.util.AbstractEventStream<SequenceLabelSample>

Class for creating an event stream out of data files for training an name finder.

Constructor Summary

Constructors

Constructor and Description
SequenceLabelerEventStream(opennlp.tools.util.ObjectStream<SequenceLabelSample> dataStream)
SequenceLabelerEventStream(opennlp.tools.util.ObjectStream<SequenceLabelSample> dataStream, String type, SequenceLabelerContextGenerator contextGenerator, SequenceLabelerCodec codec) Creates a new name finder event stream using the specified data stream and context generator.

Method Summary

Modifier and Type	Method and Description
static String[][]	additionalContext(String[] tokens, Map<String,String> prevMap) Generated previous decision features for each token based on contents of the specified map.
protected Iterator<opennlp.tools.ml.model.Event>	createEvents(SequenceLabelSample sample)
static List<opennlp.tools.ml.model.Event>	generateEvents(String[] sentence, String[] outcomes, SequenceLabelerContextGenerator cg)
static String[]	generateOutcomes(Span[] names, String type, int length) Deprecated. use the BioCodec implementation of the SequenceValidator instead!

Methods inherited from class opennlp.tools.util.AbstractEventStream

close, read, reset

Methods inherited from class java.lang.Object

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

Constructor Detail

SequenceLabelerEventStream

public SequenceLabelerEventStream(opennlp.tools.util.ObjectStream<SequenceLabelSample> dataStream,
                                  String type,
                                  SequenceLabelerContextGenerator contextGenerator,
                                  SequenceLabelerCodec codec)

Creates a new name finder event stream using the specified data stream and context generator.

Parameters:

dataStream - The data stream of events.

type - null or overrides the type parameter in the provided samples

contextGenerator - The context generator used to generate features for the event stream.

codec - the encoding

SequenceLabelerEventStream

public SequenceLabelerEventStream(opennlp.tools.util.ObjectStream<SequenceLabelSample> dataStream)

Method Detail

generateOutcomes

@Deprecated
public static String[] generateOutcomes(Span[] names,
                                                    String type,
                                                    int length)

Deprecated. use the BioCodec implementation of the SequenceValidator instead!

Generates the name tag outcomes (start, continue, other) for each token in a sentence with the specified length using the specified name spans.

Parameters:

names - Token spans for each of the names.

type - null or overrides the type parameter in the provided samples

length - The length of the sentence.

Returns:

An array of start, continue, other outcomes based on the specified names and sentence length.

generateEvents

public static List<opennlp.tools.ml.model.Event> generateEvents(String[] sentence,
                                                                String[] outcomes,
                                                                SequenceLabelerContextGenerator cg)

createEvents

protected Iterator<opennlp.tools.ml.model.Event> createEvents(SequenceLabelSample sample)

Specified by:

createEvents in class opennlp.tools.util.AbstractEventStream<SequenceLabelSample>

additionalContext

public static String[][] additionalContext(String[] tokens,
                                           Map<String,String> prevMap)

Generated previous decision features for each token based on contents of the specified map.

Parameters:

tokens - The token for which the context is generated.

prevMap - A mapping of tokens to their previous decisions.

Returns:

An additional context array with features for each token.