Constituent (illinois-cogcomp-nlp 3.1.29 API)

java.lang.Object
- edu.illinois.cs.cogcomp.core.datastructures.textannotation.Constituent

All Implemented Interfaces:

HasAttributes, Serializable
```
public class Constituent
extends Object
implements Serializable, HasAttributes
```
A Constituent represents a unit of text (not necessarily contiguous) that participates in a view. In terms of the nodes and edges representation of views, Constituents are the nodes and Relations are edges. Each Constituent corresponds to a set of tokens in a TextAnnotation.

Author:

Vivek Srikumar

See Also:

Serialized Form

Field Summary

Fields
Modifier and Type	Field and Description
`protected Map<String,String>`	`attributes`
`protected double`	`constituentScore`
`protected int`	`endCharOffset`
`protected List<Relation>`	`incomingRelations`
`protected int`	`label`
`protected Map<String,Double>`	`labelsToScores`
`protected List<Relation>`	`outgoingRelations`
`protected IntPair`	`span`
`protected int`	`startCharOffset`
`protected TextAnnotation`	`textAnnotation`
`protected String`	`viewName` This indicates whether the element `Constituent#constituentTokens` is a two element list consisting of a start and an end tokenId, specifying a span, instead of explicitly listing the each element.

Constructor Summary

Constructors
Constructor and Description
`Constituent(Map<String,Double> labelsToScores, String viewName, TextAnnotation text, int start, int end)` instantiate a constituent with a set of labels and corresponding scores: the 'main' label (returned by #getLabel()) will be the label with the highest score, as decided by `ArgMax`.
`Constituent(String label, double score, String viewName, TextAnnotation text, int start, int end)` start, end offsets are token indexes, and use one-past-the-end indexing -- so a one-token constituent right at the beginning of a text has start/end (0,1) offsets are relative to the entire text span (i.e.
`Constituent(String label, String viewName, TextAnnotation text, int start, int end)` start, end offsets are token indexes, and use one-past-the-end indexing -- so a one-token constituent right at the beginning of a text has start/end (0,1) offsets are relative to the entire text span (i.e.

Method Summary

All Methods Instance Methods Concrete Methods Deprecated Methods
Modifier and Type	Method and Description
`void`	`addAttribute(String key, String value)`
`Constituent`	`cloneForNewView(String newViewName)`
`Constituent`	`cloneForNewViewWithDestinationLabel(String newViewName, String Dlabel)`
`boolean`	`doesConstituentCover(Constituent other)`
`boolean`	`doesConstituentCover(int tokenId)`
`boolean`	`doesConstituentCoverAll(Collection<Integer> tokenIds)`
`boolean`	`equals(Object obj)`
`boolean`	`equalsWithoutAttributeEqualityCheck(Constituent that)` This function can be used in scenarios where there is a need for Constituent equality, ignoring the values of the attributes.
`String`	`getAttribute(String key)`
`Set<String>`	`getAttributeKeys()`
`double`	`getConstituentScore()`
`TreeSet<Integer>`	`getConstituentTokens()` Deprecated.
`int`	`getEndCharOffset()`
`int`	`getEndSpan()`
`int`	`getInclusiveEndCharOffset()`
`int`	`getInclusiveStartCharOffset()`
`List<Relation>`	`getIncomingRelations()` Get a list of relations where the target is this constituent
`String`	`getLabel()`
`Map<String,Double>`	`getLabelsToScores()` Return map of labels to scores.
`int`	`getNumberOfTokens()`
`List<Relation>`	`getOutgoingRelations()` Get a list of relations where the source is this constituent.
`int`	`getSentenceId()` Return the identifier of the sentence that contains this constituent.
`IntPair`	`getSpan()`
`int`	`getStartCharOffset()`
`int`	`getStartSpan()`
`String`	`getSurfaceForm()` This method returns the original surface form of the constituent.
`TextAnnotation`	`getTextAnnotation()`
`String`	`getTokenizedSurfaceForm()` This method returns a tokenized representation of the surface form of the constituent.
`View`	`getView()`
`String`	`getViewName()`
`boolean`	`hasAttribute(String key)`
`int`	`hashCode()`
`boolean`	`isConsituentInRange(int start, int end)`
`int`	`length()`
`void`	`removeAllAttributes()` Removes all attributes from a Constituent.
`void`	`removeAllIncomingRelatons()`
`void`	`removeAllOutgoingRelaton()`
`void`	`removeIncomingRelaton(Relation r)`
`void`	`removeOutgoingRelation(Relation r)`
`int`	`size()`
`String`	`toSExpression()` Convert this constituent into an S-Expression, including relation labels.
`String`	`toSExpression(boolean includeEdgeLabels)`
`protected String`	`toSExpression(int spaces, boolean firstChild, String prefix, boolean includeEdgeLabels)`
`String`	`toString()`

Methods inherited from class java.lang.Object
clone, finalize, getClass, notify, notifyAll, wait, wait, wait

- Field Detail
  - constituentScore
```
protected final double constituentScore
```
  - textAnnotation
```
protected final TextAnnotation textAnnotation
```
  - span
```
protected final IntPair span
```
  - startCharOffset
```
protected final int startCharOffset
```
  - endCharOffset
```
protected final int endCharOffset
```
  - outgoingRelations
```
protected final List<Relation> outgoingRelations
```
  - incomingRelations
```
protected final List<Relation> incomingRelations
```
  - label
```
protected final int label
```
  - viewName
```
protected final String viewName
```
    This indicates whether the element Constituent#constituentTokens is a two element list consisting of a start and an end tokenId, specifying a span, instead of explicitly listing the each element. This would use less memory.
  - labelsToScores
```
protected final Map<String,Double> labelsToScores
```
  - attributes
```
protected Map<String,String> attributes
```
- Constructor Detail
  - Constituent
```
public Constituent(String label,
                   String viewName,
                   TextAnnotation text,
                   int start,
                   int end)
```
    start, end offsets are token indexes, and use one-past-the-end indexing -- so a one-token constituent right at the beginning of a text has start/end (0,1) offsets are relative to the entire text span (i.e. NOT sentence-relative) This constructor assigns default score to constituent
    
    Parameters:
    
    label - label of this Constituent
    
    viewName - name of View this Constituent belongs to
    
    text - TextAnnotation this Constituent belongs to
    
    start - start token offset
    
    end - end token offset (one-past-the-end)
  - Constituent
```
public Constituent(Map<String,Double> labelsToScores,
                   String viewName,
                   TextAnnotation text,
                   int start,
                   int end)
```
    instantiate a constituent with a set of labels and corresponding scores: the 'main' label (returned by #getLabel()) will be the label with the highest score, as decided by ArgMax. Start, end offsets are token indexes, and use one-past-the-end indexing -- so a one-token constituent right at the beginning of a text has start/end (0,1) offsets are relative to the entire text span (i.e. NOT sentence-relative) This constructor assigns default score to the constituent.
    
    Parameters:
    
    labelsToScores - set of possible labels and corresponding scores.
    
    viewName - name of View this Constituent belongs to
    
    text - TextAnnotation this Constituent belongs to
    
    start - start token offset
    
    end - end token offset (one-past-the-end)
  - Constituent
```
public Constituent(String label,
                   double score,
                   String viewName,
                   TextAnnotation text,
                   int start,
                   int end)
```
    start, end offsets are token indexes, and use one-past-the-end indexing -- so a one-token constituent right at the beginning of a text has start/end (0,1) offsets are relative to the entire text span (i.e. NOT sentence-relative)
    
    Parameters:
    
    label - label of this Constituent
    
    score - confidence in label
    
    viewName - name of View this Constituent belongs to
    
    text - TextAnnotation this Constituent belongs to
    
    start - start token offset
    
    end - end token offset (one-past-the-end)
- Method Detail
  - getLabelsToScores
```
public Map<String,Double> getLabelsToScores()
```
    Return map of labels to scores. If not explicitly created, returns null. The returned map is a copy, to avoid inadvertent changes to the label/score mapping.
    
    Returns:
    
    map of labels to scores
  - getStartCharOffset
```
public int getStartCharOffset()
```
  - getEndCharOffset
```
public int getEndCharOffset()
```
  - getInclusiveStartCharOffset
```
public int getInclusiveStartCharOffset()
```
  - getInclusiveEndCharOffset
```
public int getInclusiveEndCharOffset()
```
  - addAttribute
```
public void addAttribute(String key,
                         String value)
```
    Specified by:
    
    addAttribute in interface HasAttributes
  - doesConstituentCover
```
public boolean doesConstituentCover(int tokenId)
```
  - doesConstituentCover
```
public boolean doesConstituentCover(Constituent other)
```
  - doesConstituentCoverAll
```
public boolean doesConstituentCoverAll(Collection<Integer> tokenIds)
```
  - equalsWithoutAttributeEqualityCheck
```
public boolean equalsWithoutAttributeEqualityCheck(Constituent that)
```
    This function can be used in scenarios where there is a need for Constituent equality, ignoring the values of the attributes. Note that many equality functions in Java automatically call the `equals' function.
    
    Parameters:
    
    that - the input constituent you compare with
    
    Returns:
    
    whether the two constituents are the same or not.
  - equals
```
public boolean equals(Object obj)
```
    Overrides:
    
    equals in class Object
  - getAttribute
```
public String getAttribute(String key)
```
    Specified by:
    
    getAttribute in interface HasAttributes
  - getAttributeKeys
```
public Set<String> getAttributeKeys()
```
    Specified by:
    
    getAttributeKeys in interface HasAttributes
  - getConstituentScore
```
public double getConstituentScore()
```
    Returns:
    
    the constituentScore
  - getEndSpan
```
public int getEndSpan()
```
  - getLabel
```
public String getLabel()
```
  - getNumberOfTokens
```
public int getNumberOfTokens()
```
  - size
```
public int size()
```
  - length
```
public int length()
```
  - getStartSpan
```
public int getStartSpan()
```
  - getSpan
```
public IntPair getSpan()
```
  - getTokenizedSurfaceForm
```
public String getTokenizedSurfaceForm()
```
    This method returns a tokenized representation of the surface form of the constituent. This is not the original surface form of the constituent. To retrieve that, please use getSurfaceForm()
    
    Returns:
    
    A constructed form based on the tokens covered by the constituent and a single whitespace separating them.
  - getSurfaceForm
```
public String getSurfaceForm()
```
    This method returns the original surface form of the constituent. To retrieve a tokenized form please use getTokenizedSurfaceForm()
    
    Returns:
    
    The original (char-offset-based) form of the constituent.
  - getTextAnnotation
```
public TextAnnotation getTextAnnotation()
```
    Returns:
    
    the textAnnotation
  - getViewName
```
public String getViewName()
```
    Returns:
    
    the viewName
  - getView
```
public View getView()
```
  - hasAttribute
```
public boolean hasAttribute(String key)
```
    Specified by:
    
    hasAttribute in interface HasAttributes
  - hashCode
```
public int hashCode()
```
    Overrides:
    
    hashCode in class Object
    
    Returns:
  - isConsituentInRange
```
public boolean isConsituentInRange(int start,
                                   int end)
```
  - getOutgoingRelations
```
public List<Relation> getOutgoingRelations()
```
    Get a list of relations where the source is this constituent.
    
    Returns:
    
    A list of Relations
  - getIncomingRelations
```
public List<Relation> getIncomingRelations()
```
    Get a list of relations where the target is this constituent
    
    Returns:
    
    A list of Relations
  - toSExpression
```
public String toSExpression()
```
    Convert this constituent into an S-Expression, including relation labels.
    
    Returns:
    
    A string, where
  - toSExpression
```
public String toSExpression(boolean includeEdgeLabels)
```
  - toSExpression
```
protected String toSExpression(int spaces,
                               boolean firstChild,
                               String prefix,
                               boolean includeEdgeLabels)
```
  - toString
```
public String toString()
```
    Overrides:
    
    toString in class Object
  - getSentenceId
```
public int getSentenceId()
```
    Return the identifier of the sentence that contains this constituent. No sentence contains this constituent (that is, if this constituent is an implicit one), then return -1.
  - getConstituentTokens
```
@Deprecated
public TreeSet<Integer> getConstituentTokens()
```
    Deprecated.
    
    Do not use this function. Use getSpan() instead.
  - cloneForNewView
```
public Constituent cloneForNewView(String newViewName)
```
  - cloneForNewViewWithDestinationLabel
```
public Constituent cloneForNewViewWithDestinationLabel(String newViewName,
                                                       String Dlabel)
```
  - removeIncomingRelaton
```
public void removeIncomingRelaton(Relation r)
```
  - removeAllIncomingRelatons
```
public void removeAllIncomingRelatons()
```
  - removeOutgoingRelation
```
public void removeOutgoingRelation(Relation r)
```
  - removeAllOutgoingRelaton
```
public void removeAllOutgoingRelaton()
```
  - removeAllAttributes
```
public void removeAllAttributes()
```
    Removes all attributes from a Constituent.
    
    Specified by:
    
    removeAllAttributes in interface HasAttributes

Class Constituent

Field Summary

Constructor Summary

Method Summary

Methods inherited from class java.lang.Object

Field Detail

constituentScore

textAnnotation

span

startCharOffset

endCharOffset

outgoingRelations

incomingRelations

label

viewName

labelsToScores

attributes

Constructor Detail

Constituent

Constituent

Constituent

Method Detail

getLabelsToScores

getStartCharOffset

getEndCharOffset

getInclusiveStartCharOffset

getInclusiveEndCharOffset

addAttribute

doesConstituentCover

doesConstituentCover

doesConstituentCoverAll

equalsWithoutAttributeEqualityCheck

equals

getAttribute

getAttributeKeys

getConstituentScore

getEndSpan

getLabel

getNumberOfTokens

size

length

getStartSpan

getSpan

getTokenizedSurfaceForm

getSurfaceForm

getTextAnnotation

getViewName

getView

hasAttribute

hashCode

isConsituentInRange

getOutgoingRelations

getIncomingRelations

toSExpression

toSExpression

toSExpression

toString

getSentenceId

getConstituentTokens

cloneForNewView

cloneForNewViewWithDestinationLabel

removeIncomingRelaton

removeAllIncomingRelatons

removeOutgoingRelation

removeAllOutgoingRelaton

removeAllAttributes