edu.northwestern.at.morphadorner.corpuslinguistics.sentencemelder

Class SentenceMelder

java.lang.Object

edu.northwestern.at.morphadorner.corpuslinguistics.sentencemelder.SentenceMelder

Direct Known Subclasses:

AdornedXMLReader.MySentenceMelder, XMLSentenceMelder

public class SentenceMelder
extends java.lang.Object

Sentence melder

Constructs a formatted sentence by melding a list of sentence tokens containing words and punctuation.

Example: Printing reconstituted sentence text.

Assume "sentences" is a java.util.List of sentences. Each entry in "sentences" is a sentence -- a list of words expressed as a java.util.list of word objects. The word objects may be simple strings or any object implementing a "toString()" method that yields the text associated with that word. In the following example, we use String as the type for the word objects.

  SentenceMelder melder   = new SentenceMelder();

  for ( int i = 0 ; i < sentences.size() ; i++ )
  {
      List sentence   = sentences.get( i );

      String sentenceText =
          melder.reconstituteSentence( sentences.get( i ) );

      System.out.println( sentenceText );
  }
  

Field Summary

Fields

Modifier and Type	Field and Description
protected static java.lang.String	LDQUOTESTR Left double curly quote.
protected static java.lang.String	LSQUOTESTR Left single curly quote.
protected static java.lang.String	RDQUOTESTR Right double curly quote.
protected static java.lang.String	RSQUOTESTR Right single curly quote.
protected SentenceMelderState	state Sentence melder state.

Constructor Summary

Constructors

Constructor and Description
SentenceMelder() Create sentence melder.

Method Summary

Methods

Modifier and Type	Method and Description
java.lang.String	endSentence() Finish sentence.
SentenceMelderState	getState() Get state.
void	outputBlank() Add blank to sentence.
protected void	outputWord(java.lang.String word) Add word to sentence.
void	processWord(java.lang.String word) Process word.
java.lang.String	reconstituteSentence(java.util.List<?> sentence) Reconstitute a list of words to a sentence.
void	reset() Reset the melder.
void	setState(SentenceMelderState state) Set state.
boolean	shouldOutputBlank(java.lang.String word, boolean isFirstWord) Should blank be output.
void	startSentence() Start sentence.

Methods inherited from class java.lang.Object

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

Field Detail

LSQUOTESTR

protected static final java.lang.String LSQUOTESTR

Left single curly quote.

See Also:

Constant Field Values

RSQUOTESTR

protected static final java.lang.String RSQUOTESTR

Right single curly quote.

See Also:

Constant Field Values

LDQUOTESTR

protected static final java.lang.String LDQUOTESTR

Left double curly quote.

See Also:

Constant Field Values

RDQUOTESTR

protected static final java.lang.String RDQUOTESTR

Right double curly quote.

See Also:

Constant Field Values

state

protected SentenceMelderState state

Sentence melder state.

Constructor Detail

SentenceMelder

public SentenceMelder()

Create sentence melder.

Method Detail

reconstituteSentence

public java.lang.String reconstituteSentence(java.util.List<?> sentence)

Reconstitute a list of words to a sentence.

Parameters:

sentence - List of words and punctuation in a sentence. Normally each word is a plain string, but the "toString()" operation is performed on each, so it may be any object type whose toString() yields a proper word (e.g., an AdornedWord).

Returns:

The reconstituted sentence as a string.

startSentence

public void startSentence()

Start sentence.

outputBlank

public void outputBlank()

Add blank to sentence.

outputWord

protected void outputWord(java.lang.String word)

Add word to sentence.

Parameters:

word - The word to add.

In general this should not be called directly.

processWord

public void processWord(java.lang.String word)

Process word.

Parameters:

word - The word to add.

endSentence

public java.lang.String endSentence()

Finish sentence.

Returns:

Returned sentence.

shouldOutputBlank

public boolean shouldOutputBlank(java.lang.String word,
                        boolean isFirstWord)

Should blank be output.

Parameters:

word - The word.

isFirstWord - First word in sentence.

Returns:

True to output blank before word.

getState

public SentenceMelderState getState()

Get state.

Returns:

Sentence melder state.

setState

public void setState(SentenceMelderState state)

Set state.

Parameters:

state - The sentence melder state.

reset

public void reset()

Reset the melder.

Clears the internal state which tracks quote mark pairing, word count, and the previous word encountered. Usually called between text divisions such as paragraphs.