de.up.ling.irtg.corpus

Class Corpus

java.lang.Object

de.up.ling.irtg.corpus.Corpus

All Implemented Interfaces:

Iterable<Instance>

public class Corpus
extends Object
implements Iterable<Instance>

An (annotated or unannotated) corpus of input objects. You may attach a collection of parse charts for these input objects to the chart using the Charts class. See the examples to see the exact file format for corpora.

Blank lines in a corpus file are ignored. Furthermore, lines that start with the comment prefix are ignored as well. The comment prefix is taken from the non-blank line of the corpus, which needs to be

[ccc] IRTG unannotated corpus file, v1.0

or

[ccc] IRTG annotated corpus file, v1.0 respectively. Whatever you specify as [ccc] is used as the comment pattern, and all lines that start with the same pattern are ignored as comments. So if you use "# IRTG annotated ...", then all lines starting with "#" are comments, and if you use "// IRTG unannotated ...", then all lines starting with "//" are comments. You can freely choose your own comment prefix to suit the needs of your corpus.

Author:

koller

Constructor Summary

Constructors

Constructor and Description
Corpus() This creates a new empty corpus.

Method Summary

Modifier and Type	Method and Description
void	addInstance(Instance instance) Adds a new instance to the corpus.
void	attachCharts(ChartAttacher charts) This attaches parse charts to the instances of the corpus.
void	attachCharts(String filename) Reads charts from a file and attaches them to this corpus.
int	getNumberOfInstances() Returns the number of instances contained in this corpus.
String	getSource() Returns a string describing the source of the corpus, if this was passed to the corpus at some point.
boolean	hasCharts() Returns true if the instances in this corpus are associated with parse charts.
boolean	isAnnotated() Returns true if there are gold annotations for each instance.
Iterator<Instance>	iterator()
static Corpus	readCorpus(Reader reader, InterpretedTreeAutomaton irtg) Reads a corpus from a string format available via a reader.
static Corpus	readCorpusLenient(Reader reader, InterpretedTreeAutomaton irtg) A version of readCorpus that allows interpretations in the corpus file to not be declared in the grammar.
void	setSource(String source) This sets a value for the source of the corpus.
void	sort(Comparator<Instance> comparator) Re-orders the instances in this corpus according to the order induced by the comparator.

Methods inherited from class java.lang.Object

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

Methods inherited from interface java.lang.Iterable

forEach, spliterator

Constructor Detail

Corpus

public Corpus()

This creates a new empty corpus. In this corpus there will be no charts and no annotation.

Method Detail

isAnnotated

public boolean isAnnotated()

Returns true if there are gold annotations for each instance.

Returns:

hasCharts

public boolean hasCharts()

Returns true if the instances in this corpus are associated with parse charts.

Returns:

attachCharts

public void attachCharts(ChartAttacher charts)

This attaches parse charts to the instances of the corpus. Different ChartAttachers may have different strategies for this. A OnTheFlyCharts attacher will compute the parse charts when they are requested.

Parameters:

charts -

attachCharts

public void attachCharts(String filename)
                  throws IOException

Reads charts from a file and attaches them to this corpus.

Parameters:

filename -

Throws:

IOException

getNumberOfInstances

public int getNumberOfInstances()

Returns the number of instances contained in this corpus.

Returns:

iterator

public Iterator<Instance> iterator()

Specified by:

iterator in interface Iterable<Instance>

addInstance

public void addInstance(Instance instance)

Adds a new instance to the corpus. If the instance is not annotated, it will change set the whole corpus to being unAnnotated.

Parameters:

instance -

getSource

public String getSource()

Returns a string describing the source of the corpus, if this was passed to the corpus at some point.

Returns:

setSource

public void setSource(String source)

This sets a value for the source of the corpus.

Parameters:

source -

readCorpus

public static Corpus readCorpus(Reader reader,
                                InterpretedTreeAutomaton irtg)
                         throws IOException,
                                CorpusReadingException

Reads a corpus from a string format available via a reader.

Parameters:

reader -

irtg -

Returns:

Throws:

IOException

CorpusReadingException

readCorpusLenient

public static Corpus readCorpusLenient(Reader reader,
                                       InterpretedTreeAutomaton irtg)
                                throws IOException,
                                       CorpusReadingException

A version of readCorpus that allows interpretations in the corpus file to not be declared in the grammar. These interpretations are omitted in the returned corpus. Returns an error if none of the interpretations are declared in the grammar.

Parameters:

reader -

irtg -

Returns:

Throws:

IOException

CorpusReadingException

sort

public void sort(Comparator<Instance> comparator)

Re-orders the instances in this corpus according to the order induced by the comparator.

Parameters:

comparator -