Example usage for edu.stanford.nlp.trees Tree value

List of usage examples for edu.stanford.nlp.trees Tree value

  1. HOME
  2. Java
  3. edu.stanford.nlp
  4. edu.stanford.nlp.trees.*
  5. Tree
  6. value

Introduction

In this page you can find the example usage for edu.stanford.nlp.trees Tree value.

Prototype

@Override
    public String value()

Source Link

Usage

From source file:edu.jhu.hlt.concrete.stanford.PreNERCoreMapWrapper.java

License:Open Source License

/**
*
* @param root//w  w  w  .  ja  v a  2 s  .com
* @param left
* @param right
* @param n
*          is the length of the sentence is tokens.
* @param p
* @param tokenizationUUID
* @return The constituent ID
* @throws AnalyticException
*/
private static int constructConstituent(Tree root, int left, int right, int n, Parse p, UUID tokenizationUUID,
        HeadFinder hf) throws AnalyticException {

    Constituent constituent = new Constituent();
    constituent.setId(p.getConstituentListSize());
    constituent.setTag(root.value());
    constituent.setStart(left);
    constituent.setEnding(right);
    p.addToConstituentList(constituent);
    Tree headTree = null;
    if (!root.isLeaf()) {
        try {
            headTree = hf.determineHead(root);
        } catch (java.lang.IllegalArgumentException iae) {
            LOGGER.warn("Failed to find head, falling back on rightmost constituent.");
            headTree = root.children()[root.numChildren() - 1];
        }
    }
    int i = 0, headTreeIdx = -1;

    int leftPtr = left;
    for (Tree child : root.getChildrenAsList()) {
        int width = child.getLeaves().size();
        int childId = constructConstituent(child, leftPtr, leftPtr + width, n, p, tokenizationUUID, hf);
        constituent.addToChildList(childId);

        leftPtr += width;
        if (headTree != null && child == headTree) {
            assert (headTreeIdx < 0);
            headTreeIdx = i;
        }
        i++;
    }

    if (headTreeIdx >= 0)
        constituent.setHeadChildIndex(headTreeIdx);

    if (!constituent.isSetChildList())
        constituent.setChildList(new ArrayList<Integer>());
    return constituent.getId();
}

From source file:edu.ucla.cs.scai.qa.questionclassifier.SyntacticTreeNode.java

public SyntacticTreeNode(Tree t, ArrayList<CoreLabel> tokens, SyntacticTreeNode parent) throws Exception {
    this.parent = parent;
    value = t.value();
    if (t.isLeaf()) {
        CoreLabel c = tokens.remove(0);/*from  w  w  w  .jav  a2 s  .  c  om*/
        begin = c.beginPosition();
        end = c.endPosition();
        if (c == null) {
            throw new Exception("Mapping between TreeNode and CoreLabel not found");
        } else {
            lemma = c.lemma();
            ner = c.ner();
            //System.out.println(value + " -> " + c.value());
            if (!value.equals(c.value())) {
                throw new Exception("Different words have been matched!");
            }
        }
    } else {
        boolean hasNPchildren = false;
        boolean hasWHNPchildren = false;
        boolean hasQPchildren = false;
        begin = Integer.MAX_VALUE;
        end = Integer.MIN_VALUE;
        for (Tree c : t.children()) {
            SyntacticTreeNode child = new SyntacticTreeNode(c, tokens, this);
            children.add(child);
            if (child.value.equals("NP")) {
                hasNPchildren = true;
            } else if (child.value.equals("QP")) {
                hasQPchildren = true;
            } else if (child.value.equals("WHNP")) {
                hasWHNPchildren = true;
            }
            begin = Math.min(begin, child.begin);
            end = Math.max(end, child.end);
        }
        if (value.equals("NP")) {
            if (hasNPchildren) {
                npCompound = true;
            } else if (hasQPchildren) {
                npQp = true;
            } else {
                npSimple = true;
            }
        } else if (value.equals("WHNP")) { //can a WHNP node have QP children?
            if (hasNPchildren || hasWHNPchildren) {
                whnpCompound = true;
            } else if (!hasQPchildren) {
                whnpSimple = true;
            }
        }
    }
}

From source file:elkfed.coref.features.pairs.FE_CCommand.java

License:Apache License

public static Boolean getCCommand(PairInstance inst) {

    // should be in the same sentence
    if (inst.getAnaphor().getSentId() != inst.getAntecedent().getSentId())
        return false;

    //Ana should not be reflexive or reciprocal pronoun
    if (inst.getAnaphor().getReflPronoun())
        return false;

    // should have not-null  maxnp-trees (otherwise -- problematic mentions)

    Tree sentenceTree = inst.getAnaphor().getSentenceTree();
    Tree AnaTree = inst.getAnaphor().getMaxNPParseTree();
    Tree AnteTree = inst.getAntecedent().getMaxNPParseTree();
    if (sentenceTree == null)
        return false;
    if (AnaTree == null)
        return false;
    if (AnteTree == null)
        return false;

    // should not dominate each other
    if (AnaTree.dominates(AnteTree))
        return false;
    if (AnteTree.dominates(AnaTree))
        return false;

    //the first branching node for ante should dominate ana (but not via S-node)
    AnteTree = AnteTree.parent(sentenceTree);
    while (AnteTree != null) {

        if (AnteTree.children().length > 1) {
            if (!AnteTree.dominates(AnaTree))
                return false;
            while (AnaTree != null && AnaTree != AnteTree) {
                if (AnaTree.value().toLowerCase().startsWith("s"))
                    return false;
                AnaTree = AnaTree.parent(sentenceTree);
            }//w w w.  j ava 2 s  .c om
            return true;
        }
        AnteTree = AnteTree.parent(sentenceTree);
    }

    return false;
}

From source file:elkfed.coref.features.pairs.FE_Copula.java

License:Apache License

public static Boolean getCopula(PairInstance inst) {

    // should be in the same sentence
    if (inst.getAnaphor().getSentId() != inst.getAntecedent().getSentId()) {
        return false;
    }//w  ww. j  ava  2 s  . c om

    // should have not-null  maxnp-trees (otherwise -- problematic mentions)

    Tree sentenceTree = inst.getAnaphor().getSentenceTree();
    Tree AnaTree = inst.getAnaphor().getMaxNPParseTree();
    Tree AnteTree = inst.getAntecedent().getMaxNPParseTree();
    if (sentenceTree == null) {
        return false;
    }
    if (AnaTree == null) {
        return false;
    }
    if (AnteTree == null) {
        return false;
    }

    // exclude "there is .." (ToDo: exclude other expletives!)

    if (inst.getAntecedent().getMarkableString().toLowerCase().matches(NONREF_NP)) {
        return false;
    }

    //exclude date and time

    if (inst.getAnaphor().getSemanticClass() == SemanticClass.TIME) {
        return false;
    }
    if (inst.getAnaphor().getSemanticClass() == SemanticClass.DATE) {
        return false;
    }
    if (inst.getAntecedent().getSemanticClass() == SemanticClass.TIME) {
        return false;
    }
    if (inst.getAntecedent().getSemanticClass() == SemanticClass.DATE) {
        return false;
    }

    // should be subj-obj of the same verb

    Tree vp = AnaTree.parent(sentenceTree);
    if (vp == null) {
        return false;
    }
    if (!vp.value().equalsIgnoreCase("vp")) {
        return false;
    }
    while (vp.parent(sentenceTree) != null && vp.parent(sentenceTree).value().equalsIgnoreCase("vp")) {
        vp = vp.parent(sentenceTree);
    }
    if (vp.parent(sentenceTree) == null) {
        return false;
    }

    Boolean foundante = false;
    Tree[] chldsup = vp.parent(sentenceTree).children();
    for (int i = 0; i < chldsup.length; i++) {
        if (chldsup[i] == AnteTree) {
            foundante = true;
        }
        if (chldsup[i] == vp && foundante == false) {
            return false;
        }
    }

    vp = AnaTree.parent(sentenceTree); // we do not want to go higher here -- "S is *ing O" fires otherwise

    // should not contain a modal verb
    Tree[] chlds = vp.children();
    for (int i = 0; i < chlds.length; i++) {
        if (chlds[i].value().equalsIgnoreCase("rb")) {
            return false;
        }
        if (chlds[i].value().equalsIgnoreCase("md")
                && chlds[i].getLeaves().get(0).value().toLowerCase().matches(MODAL_VERB)) {
            return false;
        }
    }

    // the verb should be one of the copula verbs

    for (int i = 0; i < chlds.length; i++) {

        if (chlds[i].value().equalsIgnoreCase("vbd") || chlds[i].value().equalsIgnoreCase("aux")
                || chlds[i].value().equalsIgnoreCase("vbn") || chlds[i].value().equalsIgnoreCase("vb")
                || chlds[i].value().equalsIgnoreCase("vbd") || chlds[i].value().equalsIgnoreCase("vbp")
                || chlds[i].value().equalsIgnoreCase("vbz") || chlds[i].value().equalsIgnoreCase("vbg")) {
            if (chlds[i].getLeaves().get(0).value().toLowerCase().matches(COPULA_VERB)) {
                /*
                System.out.println("Found positive copula verb (" 
                 +chlds[i].getLeaves().get(0).value() +
                ") for ("+
                inst.getAnaphor().getMarkableString()+
                "),("+
                inst.getAntecedent().getMarkableString()+
                ") ");
                */
                return true;
            }
        }
    }
    return false;
}

From source file:elkfed.coref.features.pairs.FE_Span.java

License:Apache License

public static Boolean getSpanEmbed(PairInstance inst) {

    // should be in the same sentence
    if (inst.getAnaphor().getSentId() != inst.getAntecedent().getSentId())
        return false;

    // should not be in apposition
    if (FE_AppositiveParse.getAppositivePrs(inst))
        return false;

    // should not be adjacent (should take from appo_icab, but doesn't work for some reason)

    Markable m1 = inst.getAntecedent().getMarkable();
    Markable m2 = inst.getAnaphor().getMarkable();

    int sana = m2.getLeftmostDiscoursePosition();
    int sante = m1.getLeftmostDiscoursePosition();
    int eana = m2.getRightmostDiscoursePosition();
    int eante = m1.getRightmostDiscoursePosition();

    MiniDiscourse doc = m1.getMarkableLevel().getDocument();

    if (m1.getAttributeValue("min_ids") != null) {
        String[] spans = MarkableHelper.parseRanges(m1.getAttributeValue("min_ids"));
        sante = doc.DiscoursePositionFromDiscourseElementID(spans[0]);
        eante = doc.DiscoursePositionFromDiscourseElementID(spans[spans.length - 1]);
    }/*  w w  w.j a  v a  2s . com*/
    if (m2.getAttributeValue("min_ids") != null) {
        String[] spans = MarkableHelper.parseRanges(m2.getAttributeValue("min_ids"));
        sana = doc.DiscoursePositionFromDiscourseElementID(spans[0]);
        eana = doc.DiscoursePositionFromDiscourseElementID(spans[spans.length - 1]);
    }

    if (eana == sante - 1)
        return false;
    if (eante == sana - 1)
        return false;

    // check for trivial embedding, if maxnps are missing
    if (inst.getAntecedent().getMaxNPParseTree() == null && sana >= sante && eana <= eante)
        return true;
    if (inst.getAnaphor().getMaxNPParseTree() == null && sana <= sante && eana >= eante)
        return true;

    // check for maximal np embedding

    sana -= inst.getAnaphor().getSentenceStart();
    sante -= inst.getAntecedent().getSentenceStart();
    eana -= inst.getAnaphor().getSentenceStart();
    eante -= inst.getAntecedent().getSentenceStart();

    // if anaphor does have MaxNPParseTree -- check that it does not span over the antecedent
    // if it does -- check whether there is an s-node in between

    if (inst.getAnaphor().getMaxNPParseTree() != null) {
        Tree sentTree = inst.getAntecedent().getSentenceTree();
        List<Tree> Leaves = sentTree.getLeaves();
        Tree startNode = Leaves.get(sante);
        if (eante < Leaves.size()) {
            // check <leaves.size for markables spanning over sentence boundaries
            Tree endNode = Leaves.get(eante);
            if (inst.getAnaphor().getMaxNPParseTree().dominates(endNode)
                    && inst.getAnaphor().getMaxNPParseTree().dominates(startNode)) {
                Boolean sfound = false;
                Tree t = startNode;
                while (t != null && t != inst.getAnaphor().getMaxNPParseTree() && !sfound) {
                    if (t.value().toLowerCase().startsWith("s"))
                        sfound = true;
                    t = t.parent(sentTree);
                }
                if (!sfound)
                    return true;
            }
        }

    }

    // same for antecedent
    if (inst.getAntecedent().getMaxNPParseTree() != null) {
        Tree sentTree = inst.getAntecedent().getSentenceTree();
        List<Tree> Leaves = sentTree.getLeaves();
        Tree startNode = Leaves.get(sana);
        if (eana < Leaves.size()) {
            Tree endNode = Leaves.get(eana);

            if (inst.getAntecedent().getMaxNPParseTree().dominates(endNode)
                    && inst.getAntecedent().getMaxNPParseTree().dominates(startNode)) {
                Boolean sfound = false;
                Tree t = startNode;
                while (t != null && t != inst.getAntecedent().getMaxNPParseTree() && !sfound) {
                    if (t.value().toLowerCase().startsWith("s"))
                        sfound = true;
                    t = t.parent(sentTree);
                }
                if (!sfound)
                    return true;

            }
        }
    }

    return false;

}

From source file:elkfed.coref.mentions.Mention.java

License:Apache License

public String getRootPath() {
    Tree top = getSentenceTree();//from w w  w.  j  a  v a  2s.  c  o  m
    Tree here = getHighestProjection();
    StringBuffer sb = new StringBuffer();
    String lastValue = null;
    while (here != top) {
        here = here.parent(top);
        String val = here.value();
        if (!val.equals(lastValue))
            sb.append(here.value()).append(".");
        lastValue = val;
    }
    return sb.toString();
}

From source file:elkfed.coref.mentions.Mention.java

License:Apache License

/**
 * @author samuel//from   w  w w. j av  a  2s. c om
 * @param markableSubtree
 *            A subtree of the sentenceTreeWithDiscIds
 * @param postag
 *            Parts of speech tag
 * @return Array of disc ids representing the highest projecting phrase
 *         inside the markables subtree with the given postag
 */
public ArrayList<String> getHighestProjectingPhraseWithPOS(Tree markableSubtree, String postag) {
    LinkedList<Tree> stack = new LinkedList<Tree>();
    stack.add(markableSubtree);
    ArrayList<String> result = new ArrayList<String>();
    Tree head = null;
    fifo: while (stack.size() > 0) {
        head = stack.removeFirst();
        for (Tree child : head.children()) {
            if (child.value().toString().equalsIgnoreCase(postag)) {
                break fifo;
            } else if (child.value().toString().equalsIgnoreCase("nx")) {
                stack.add(child);
            }
        }
    }
    for (Tree child : head.children()) {
        if (child.value().toString().equalsIgnoreCase(postag)) {
            result.add(child.children()[0].value().toString());
        }
    }
    if (result.size() > 0) {
        return result;
    } else {
        return null;
    }
}

From source file:elkfed.expletives.EF_Tree.java

License:Apache License

public static Tree tree_pred(Tree node) {
    LabelFactory lf = new StringLabelFactory();
    Tree result = new LabeledScoredTreeNode();
    result.setLabel(lf.newLabel(node.value() + "-PRD"));
    if (node.value().equals("PP") && node.children().length == 2 && node.children()[0].value().equals("IN")) {
        Tree[] dtrs = new Tree[2];
        dtrs[0] = node.children()[0];/*from  w ww.ja  v a  2s.c  o m*/
        dtrs[1] = node.children()[1].headPreTerminal(new ModCollinsHeadFinder());
        result.setChildren(dtrs);
    }
    Tree[] dtrs = new Tree[1];
    dtrs[0] = node.headPreTerminal(new ModCollinsHeadFinder());
    result.setChildren(dtrs);
    return result;
}

From source file:elkfed.expletives.EF_Tree.java

License:Apache License

/** constructs a marked subtree for a subclause
 * outside the path to the pronoun/*from  w w w  .  ja va  2 s . c  o  m*/
 * @param node the starting point
 * @return a marked subtree for the tree starting with node
 */
public static Tree tree_pruned(Tree node) {
    LabelFactory lf = new StringLabelFactory();
    Tree result = new LabeledScoredTreeNode();
    result.setLabel(lf.newLabel(node.value() + "-X"));
    List<Tree> dtrs = new ArrayList<Tree>();
    boolean cpl_seen = false;
    if (node.value().matches("S|SBAR|VP")) {
        for (Tree t : node.children()) {
            // modals are copied verbatim
            String cat = t.value();
            if (cat.matches("TO|MD|IN")) {
                dtrs.add(t);
                cpl_seen = true;
            } else if (cat.startsWith("WH")) {
                Tree dtr = tagged_word(cat, "WH");
                cpl_seen = true;
            } else if (t.value().startsWith("VB")) {
                break;
            } else if (t.value().matches("S|SBAR|VP")) {
                if (cpl_seen) {
                    //ignore
                } else {
                    dtrs.add(tree_pruned(t));
                }
            }
        }
    }
    result.setChildren(dtrs);
    return result;
}

From source file:elkfed.expletives.EF_Tree.java

License:Apache License

private static void dtrs_inside(Tree node, List<Tree> dtrs) {
    boolean after_be = false;
    for (Tree t : node.children()) {
        String cat = t.value();
        if (cat.matches("''|``")) {
            // ignore
        } else if ((cat.startsWith("VB") || cat.equals("AUX"))
                && t.children()[0].value().matches("'s|is|was|be|being|been")) {
            //dtrs.add(t);
            dtrs.add(tagged_word(cat, "BE"));
            after_be = true;//  w  w w. j  a  v a  2s  .com
        } else if (after_be && cat.matches("NP|ADJP|VP|PP")) {
            if (cat.equals("VP") && t.children()[0].value().matches("VB[GN]")) {
                dtrs_inside(t, dtrs);
                //dtrs.add(tree_outside(t));
            } else {
                dtrs.add(tree_pred(t));
            }
            after_be = false;
        } else {
            if (cat.matches("S|SBAR")) {
                dtrs.add(tree_pruned(t));
            } else if (cat.equals("VP")) {
                dtrs_inside(t, dtrs);
                //dtrs.add(tree_outside(t));
            } else {
                dtrs.add(tree_outside(t));
            }
        }
    }
}