DependencyInstance.java

package mstparser;

import java.io.IOException;
import java.io.ObjectInputStream;
import java.io.ObjectOutputStream;
import java.io.Serializable;
import java.util.Arrays;
import java.util.List;

public class DependencyInstance implements Serializable {

    public FeatureVector fv;
    public String actParseTree;
    
    // The various data types. Here's an example from Portuguese:
    //
    // 3  eles ele   pron       pron-pers M|3P|NOM 4    SUBJ   _     _
    // ID FORM LEMMA COURSE-POS FINE-POS  FEATURES HEAD DEPREL PHEAD PDEPREL
    //
    // We ignore PHEAD and PDEPREL for now. 
    
    // FORM: the forms - usually words, like "thought"
    public String[] forms;
    
    // LEMMA: the lemmas, or stems, e.g. "think"
    public String[] lemmas;
    
    // COURSE-POS: the course part-of-speech tags, e.g."V"
    public String[] cpostags;
    
    // FINE-POS: the fine-grained part-of-speech tags, e.g."VBD"
    public String[] postags;
    
    // FEATURES: some features associated with the elements separated by "|", e.g. "PAST|3P"
    public String[][] feats;
    
    // HEAD: the IDs of the heads for each element
    public int[] heads;
    
    // DEPREL: the dependency relations, e.g. "SUBJ"
    public String[] deprels;
    
    // afm 03-07-08 --- heads predicted by the base classifier in a stacked learning framework
    public int[] heads_pred;

    // afm 03-07-08 --- dependency relations predicted by the base classifier in a stacked learning framework
    public String[] deprels_pred;
    
    // afm 03-07-08 --- true if there are stacked learning features
    public boolean stacked=false;
    
    // RELATIONAL FEATURE: relational features that hold between items
    public RelationalFeature[] relFeats;
    
    // Confidence scores per edge
    public double[] confidenceScores;
    
    public List<String> numbers;

    public DependencyInstance() {
    }

    public DependencyInstance(DependencyInstance source) {
        this.fv = source.fv;
        this.actParseTree = source.actParseTree;
    }

    public DependencyInstance(String[] forms, FeatureVector fv) {
        this.forms = forms;
        this.fv = fv;
    }

    public DependencyInstance(String[] forms, String[] postags, FeatureVector fv) {
        this(forms, fv);
        this.cpostags = this.cpostags = postags;
    }

    public DependencyInstance(String[] forms, String[] postags,
            String[] labs, FeatureVector fv) {
        this(forms, postags, fv);
        this.deprels = labs;
    }

    public DependencyInstance(String[] forms, String[] postags,
            String[] labs, int[] heads) {
        this.stacked = false; // afm 03-07-2008
        this.forms = forms;
        this.cpostags = this.postags = postags;
        this.deprels = labs;
        this.heads = heads;
    }

    public DependencyInstance(String[] forms, String[] postags,
            String[] labs, int[] heads, double[] confidenceScores) {
        this(forms, postags, labs, heads);
        this.confidenceScores = confidenceScores;
    }

    public DependencyInstance(String[] forms, String[] lemmas, String[] cpostags,
            String[] postags, String[][] feats, String[] labs, int[] heads, List<String> numbers) {
        this(forms, postags, labs, heads);
        this.lemmas = lemmas;
        this.cpostags = cpostags;
        this.feats = feats;
        this.numbers = numbers;
    }
    
    public DependencyInstance(String[] forms, String[] lemmas, String[] cpostags, 
            String[] postags, String[][] feats, String[] labs, int[] heads,
            RelationalFeature[] relFeats, String[] deprels_pred, int[] heads_pred, 
            boolean stacked, List<String> numbers) {
    	this(forms, lemmas, cpostags, postags, feats, labs, heads, numbers);
        this.relFeats = relFeats;
    	this.stacked = stacked;
    	this.deprels_pred = deprels_pred;
    	this.heads_pred = heads_pred;
    }

    public DependencyInstance(String[] forms, String[] lemmas, String[] cpostags,
            String[] postags, String[][] feats, String[] labs, int[] heads,
            RelationalFeature[] relFeats, double[] confidenceScores, List<String> numbers) {
        this(forms, lemmas, cpostags, postags, feats, labs, heads, numbers);
        this.relFeats = relFeats;
        this.confidenceScores = confidenceScores;
    }

    public void setFeatureVector(FeatureVector fv) {
        this.fv = fv;
    }

    public int length() {
        return forms.length;
    }

    @Override
    public String toString() {
        StringBuilder sb = new StringBuilder();
        sb.append(Arrays.toString(forms)).append("\n");
        return sb.toString();
    }

    // used in "MSTParserStacked"
    public void setPreviousLevelsInfo(String[] depRelPred, int[] headPred)
    {
    	this.deprels_pred = depRelPred;
    	this.heads_pred = headPred;
    }
    
    private void writeObject(ObjectOutputStream out) throws IOException {
        out.writeObject(forms);
        out.writeObject(lemmas);
        out.writeObject(cpostags);
        out.writeObject(postags);
        out.writeObject(heads);
        out.writeObject(deprels);
        out.writeObject(actParseTree);
        out.writeObject(feats);
        out.writeObject(relFeats);
	out.writeObject(deprels_pred);
	out.writeObject(heads_pred);
        out.writeObject(stacked);
    }

    private void readObject(ObjectInputStream in) throws IOException, ClassNotFoundException {
        forms = (String[]) in.readObject();
        lemmas = (String[]) in.readObject();
        cpostags = (String[]) in.readObject();
        postags = (String[]) in.readObject();
        heads = (int[]) in.readObject();
        deprels = (String[]) in.readObject();
        actParseTree = (String) in.readObject();
        feats = (String[][]) in.readObject();
        relFeats = (RelationalFeature[]) in.readObject();
        
        // stacked
	deprels_pred = (String[])in.readObject();
	heads_pred = (int[])in.readObject();
	stacked = (Boolean)in.readObject();
    }
}