edu.illinois.cs.cogcomp.lbjava.nlp

Class DataReader

java.lang.Object

edu.illinois.cs.cogcomp.lbjava.nlp.DataReader

All Implemented Interfaces:

edu.illinois.cs.cogcomp.lbjava.parse.Parser

Direct Known Subclasses:

PrepSRLDataReader

public abstract class DataReader
extends Object
implements edu.illinois.cs.cogcomp.lbjava.parse.Parser

A generic data reader class that can be used to interface with the `LBJava` Parser code. It also provides a simple caching mechanism based on TextAnnotationMapDBHandler. The main assumption is that an LBJava-based classifier is going to be using a collection of Constituents as the set of training/testing examples. Subclasses need to implement specific methods for candidate generation and input format reading.

Field Summary

Fields

Modifier and Type	Field and Description
static String	CANDIDATE
protected String	corpusName
protected IResetableIterator<TextAnnotation>	dataset
protected String	file
protected static org.slf4j.Logger	logger
protected String	viewName

Constructor Summary

Constructors

Constructor and Description
DataReader(String file, String corpusName, String viewName)

Method Summary

Modifier and Type	Method and Description
abstract List<Constituent>	candidateGenerator(TextAnnotation ta)
void	close()
protected List<Constituent>	getFinalCandidates(View goldView, List<Constituent> candidates)
Object	next() Fetches the next available data instance for training/testing.
abstract void	preprocess(TextAnnotation ta)
abstract List<TextAnnotation>	readData()
void	reset()

Methods inherited from class java.lang.Object

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

Field Detail

logger

protected static org.slf4j.Logger logger

CANDIDATE

public static final String CANDIDATE

See Also:

Constant Field Values

dataset

protected IResetableIterator<TextAnnotation> dataset

viewName

protected String viewName

corpusName

protected String corpusName

file

protected final String file

Constructor Detail

DataReader

public DataReader(String file,
                  String corpusName,
                  String viewName)

Method Detail

preprocess

public abstract void preprocess(TextAnnotation ta)
                         throws AnnotatorException

Throws:

AnnotatorException

readData

public abstract List<TextAnnotation> readData()

candidateGenerator

public abstract List<Constituent> candidateGenerator(TextAnnotation ta)

getFinalCandidates

protected List<Constituent> getFinalCandidates(View goldView,
                                               List<Constituent> candidates)

next

public Object next()

Fetches the next available data instance for training/testing. Also, pre-processes each new TextAnnotation object before accessing its members.

Specified by:

next in interface edu.illinois.cs.cogcomp.lbjava.parse.Parser

Returns:

A Constituent (which might be a part of a Relation, depending on the type of View )

reset

public void reset()

Specified by:

reset in interface edu.illinois.cs.cogcomp.lbjava.parse.Parser

close

public void close()

Specified by:

close in interface edu.illinois.cs.cogcomp.lbjava.parse.Parser