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:org.lambda3.text.simplification.discourse.runner.discourse_tree.extraction.ExtractionRule.java

License:Open Source License

private static List<Word> appendWordsFromTree(List<Word> words, Tree tree) {
    List<Word> res = new ArrayList<Word>();
    res.addAll(words);//from ww w.  j  a  v a 2  s  .  c  o m

    TregexPattern p = TregexPattern.compile(tree.value() + " <<, NNP|NNPS");
    TregexMatcher matcher = p.matcher(tree);

    boolean isFirst = true;
    for (Word word : tree.yieldWords()) {
        if ((isFirst) && (!matcher.findAt(tree))) {
            res.add(WordsUtils.lowercaseWord(word));
        } else {
            res.add(word);
        }
        isFirst = false;
    }

    return res;
}

From source file:org.lambda3.text.simplification.discourse.runner.discourse_tree.extraction.ExtractionRule.java

License:Open Source License

protected static List<Word> getRephrasedParticipalS(Tree np, Tree vp, Tree s, Tree vbgn) {
    Number number = getNumber(np);
    Tense tense = getTense(vp);/*  w w w .ja  va 2s.  c om*/

    TregexPattern p = TregexPattern.compile(vbgn.value() + " <<: (having . (been . VBN=vbn))");
    TregexPattern p2 = TregexPattern.compile(vbgn.value() + " <<: (having . VBN=vbn)");
    TregexPattern p3 = TregexPattern.compile(vbgn.value() + " <<: (being . VBN=vbn)");

    TregexMatcher matcher = p.matcher(s);
    if (matcher.findAt(s)) {
        List<Word> res = new ArrayList<>();

        res.add(new Word((number.equals(Number.SINGULAR)) ? "has" : "have"));
        res.add(new Word("been"));
        List<Word> next = ParseTreeExtractionUtils.getFollowingWords(s, matcher.getNode("vbn"), true);
        if (next.size() > 0) {
            next.set(0, WordsUtils.lowercaseWord(next.get(0)));
        }
        res.addAll(next);

        return res;
    }

    matcher = p2.matcher(s);
    if (matcher.findAt(s)) {
        List<Word> res = new ArrayList<>();

        res.add(new Word((number.equals(Number.SINGULAR)) ? "has" : "have"));
        List<Word> next = ParseTreeExtractionUtils.getFollowingWords(s, matcher.getNode("vbn"), true);
        if (next.size() > 0) {
            next.set(0, WordsUtils.lowercaseWord(next.get(0)));
        }
        res.addAll(next);

        return res;
    }

    matcher = p3.matcher(s);
    if (matcher.findAt(s)) {
        List<Word> res = new ArrayList<>();
        if (tense.equals(Tense.PRESENT)) {
            res.add(new Word((number.equals(Number.SINGULAR)) ? "is" : "are"));
        } else {
            res.add(new Word((number.equals(Number.SINGULAR)) ? "was" : "were"));
        }
        List<Word> next = ParseTreeExtractionUtils.getFollowingWords(s, matcher.getNode("vbn"), true);
        if (next.size() > 0) {
            next.set(0, WordsUtils.lowercaseWord(next.get(0)));
        }
        res.addAll(next);

        return res;
    }

    // default
    List<Word> res = new ArrayList<>();
    if (tense.equals(Tense.PRESENT)) {
        res.add(new Word((number.equals(Number.SINGULAR)) ? "is" : "are"));
    } else {
        res.add(new Word((number.equals(Number.SINGULAR)) ? "was" : "were"));
    }
    List<Word> next = ParseTreeExtractionUtils.getFollowingWords(s, vbgn, true);
    if (next.size() > 0) {
        next.set(0, WordsUtils.lowercaseWord(next.get(0)));
    }
    res.addAll(next);

    return res;
}

From source file:org.lambda3.text.simplification.discourse.runner.discourse_tree.extraction.rules.CoordinationExtractor.java

License:Open Source License

private static boolean isNPVPClause(Tree s) {
    TregexPattern p = TregexPattern.compile(s.value() + " < (NP $.. VP)");
    TregexMatcher matcher = p.matcher(s);

    return (matcher.findAt(s));
}

From source file:org.lambda3.text.simplification.discourse.utils.SPOSplitter.java

License:Open Source License

public static Optional<Result> split(String sentence) {
    try {/*from  ww w.  j  a  v  a 2 s . c o m*/
        Tree parseTree = ParseTreeParser.parse(sentence);
        //            LOGGER.info(ParseTreeVisualizer.prettyPrint(parseTree));

        // pattern with object
        TregexPattern p = TregexPattern.compile(
                "ROOT <<: (S < (NP=np $.. (VP=vp [ <+(VP) (VP=lowestvp !< VP) | ==VP=lowestvp !< VP ])))");
        TregexMatcher matcher = p.matcher(parseTree);
        while (matcher.findAt(parseTree)) {
            Tree np = matcher.getNode("np");
            Tree vp = matcher.getNode("vp");
            Tree lowestvp = matcher.getNode("lowestvp");

            // has object?
            TregexPattern op = TregexPattern
                    .compile(lowestvp.value() + " < (PP|NP|S|SBAR=ob !$,, (PP|NP|S|SBAR))");
            TregexMatcher omatcher = op.matcher(lowestvp);
            if (omatcher.findAt(lowestvp)) {
                Tree ob = omatcher.getNode("ob");

                List<Word> subjectWords = ParseTreeExtractionUtils.getContainingWords(np);
                List<Word> predicateWords = ParseTreeExtractionUtils.getWordsInBetween(parseTree, vp, ob, true,
                        false);
                List<Word> objectWords = ParseTreeExtractionUtils.getFollowingWords(vp, ob, true);

                return Optional.of(new Result(WordsUtils.wordsToString(subjectWords),
                        WordsUtils.wordsToString(predicateWords), WordsUtils.wordsToString(objectWords)));
            } else {
                List<Word> subjectWords = ParseTreeExtractionUtils.getContainingWords(np);
                List<Word> predicateWords = ParseTreeExtractionUtils.getContainingWords(vp);
                List<Word> objectWords = new ArrayList<>();

                return Optional.of(new Result(WordsUtils.wordsToString(subjectWords),
                        WordsUtils.wordsToString(predicateWords), WordsUtils.wordsToString(objectWords)));
            }
        }

    } catch (ParseTreeException e) {
        LOGGER.error("Could not parse sentence '{}'", sentence);
    }

    return Optional.empty();
}

From source file:sg.edu.nus.comp.pdtb.parser.ArgExtComp.java

License:Open Source License

private static List<String> findUpwardPath(Tree root, Tree connNode, Tree lca) {
    List<String> path = new ArrayList<>();
    if (connNode == null || lca == null) {// || connNode.equals(lca)) {
        return path;
    }/*from w  w w.  ja  va2  s  .  c om*/

    Tree curr = connNode;
    while (curr != null && !curr.equals(lca)) {
        String val = curr.value();
        if (val != null) {
            int t = val.indexOf("-");
            if (t > 0) {
                val = val.substring(0, t);
            }
            path.add(val);
        }
        curr = curr.parent(root);
    }

    if (curr != null && curr.equals(lca)) {
        String val = curr.value();
        if (val != null) {
            int t = val.indexOf("-");
            if (t > 0) {
                val = val.substring(0, t);
            }
            path.add(val);
        }
    }

    if (curr == null && path.isEmpty()) {
        return new ArrayList<>();
    } else {
        return path;
    }
}

From source file:sg.edu.nus.comp.pdtb.parser.ArgExtComp.java

License:Open Source License

private Tree[] getBioArgNodes(Tree root, String[] cols, ArrayList<String> spanArray, String connCat,
        Tree connNode) {
    List<Tree> arg1Nodes = getTreeNodesFromSpan(cols[14], spanArray);

    if (connCat.equals("Coordinator") && connNode.value().equals("CC")) {
        Tree[] children = connNode.parent(root).children();
        for (Tree child : children) {
            int ind = arg1Nodes.indexOf(child);
            if (ind == -1 && isPuncTag(child.value())) {
                arg1Nodes.add(child);/*from ww  w . j  a  va2s.  c  o  m*/
            }
        }
    }

    Tree arg1Node = (arg1Nodes.size() == 1) ? arg1Nodes.get(0) : getLCA(root, arg1Nodes);

    List<Tree> arg2Nodes = getTreeNodesFromSpan(cols[20], spanArray);

    Tree arg2Node = (arg2Nodes.size() == 1) ? arg2Nodes.get(0) : getLCA(root, arg2Nodes);

    return new Tree[] { arg1Node, arg2Node };
}

From source file:sg.edu.nus.comp.pdtb.parser.ArgExtComp.java

License:Open Source License

private Tree[] getArgNodes(Tree root, String[] cols, ArrayList<String> spanArray, String connCat,
        Tree connNode) {
    List<Tree> arg1Nodes = getTreeNodesFromSpan(cols[22], spanArray);
    arg1Nodes.addAll(getTreeNodesFromSpan(cols[29], spanArray));

    if (connCat.equals("Coordinator") && connNode.value().equals("CC")) {
        Tree[] children = connNode.parent(root).children();
        for (Tree child : children) {
            int ind = arg1Nodes.indexOf(child);
            if (ind == -1 && isPuncTag(child.value())) {
                arg1Nodes.add(child);//from   w  w  w . j  a  va 2  s. c om
            }
        }
    }

    Tree arg1Node = (arg1Nodes.size() == 1) ? arg1Nodes.get(0) : getLCA(root, arg1Nodes);

    List<Tree> arg2Nodes = getTreeNodesFromSpan(cols[32], spanArray);
    arg2Nodes.addAll(getTreeNodesFromSpan(cols[39], spanArray));

    Tree arg2Node = (arg2Nodes.size() == 1) ? arg2Nodes.get(0) : getLCA(root, arg2Nodes);

    return new Tree[] { arg1Node, arg2Node };
}

From source file:sg.edu.nus.comp.pdtb.parser.ArgExtComp.java

License:Open Source License

private static List<Tree> getInternalNodes(Tree root, Tree node) {
    // @child_nodes.size == 1 and @child_nodes.first.class != Node

    List<Tree> result = new ArrayList<>();

    if (node != null
            && !(node.children().length == 1 && node.firstChild() != null && node.firstChild().isLeaf())) {
        Tree parent = node.parent(root);
        if (parent != null && !node.value().equals("-NONE-")) {
            result.add(node);//from ww w . ja v  a 2 s .c  om
        }
        Tree[] children = node.children();
        for (Tree child : children) {
            result.addAll(getInternalNodes(root, child));
        }
    }

    return result;
}

From source file:sg.edu.nus.comp.pdtb.parser.ArgPosComp.java

License:Open Source License

public List<String[]> generateFeatures(Corpus.Type corpus, File article, FeatureType featureType)
        throws IOException {

    List<String[]> features = new ArrayList<>();
    ArrayList<String> spanMap = null;
    Map<String, String> spanHashMap = null;
    List<Tree> trees = null;
    List<String> explicitSpans = null;

    if (corpus.equals(Type.PDTB)) {
        trees = Corpus.getTrees(article, featureType);
        spanHashMap = Corpus.getSpanMap(article, featureType);

        spanMap = Corpus.getSpanMapAsList(article, featureType);
        explicitSpans = Corpus.getExplicitSpans(article, featureType);

        if (featureType == FeatureType.ErrorPropagation || featureType == FeatureType.Auto) {
            explicitSpans = Corpus.filterErrorProp(explicitSpans, article, featureType);
        }/*from  w  w  w .j a  v a 2s  .  com*/

    } else if (corpus.equals(Type.BIO_DRB)) {
        trees = Corpus.getBioTrees(article, featureType);
        spanHashMap = Corpus.getBioSpanMap(article, featureType);

        spanMap = Corpus.getBioSpanMapAsList(article, featureType);
        explicitSpans = Corpus.getBioExplicitSpans(article, featureType);
        if (featureType == FeatureType.ErrorPropagation || featureType == FeatureType.Auto) {
            explicitSpans = Corpus.filterBioErrorProp(explicitSpans, article, featureType);
        }
    } else {
        log.error("Unimplemented corpus type: " + corpus);
    }

    int index = 0;
    int contIndex = 0;

    for (String rel : explicitSpans) {

        String[] cols = rel.split("\\|", -1);
        Set<Integer> done = new HashSet<>();
        String label = null;
        if (featureType == FeatureType.AnyText) {
            label = "NA";
        } else {
            if (corpus == Type.PDTB) {
                label = Corpus.getLabel(cols[23], cols[33]);
            } else {
                label = Corpus.getBioLabel(cols[14], cols[20], spanMap);
            }
        }

        label = label.endsWith("PS") ? "PS" : label;
        if (label.equals("FS")) {
            continue;
        }
        index = contIndex;

        List<Node> nodes = new ArrayList<>();
        Tree root = null;

        String[] spans = corpus.equals(Type.PDTB) ? cols[3].split(";") : cols[1].split(";");

        for (String spanTmp : spans) {
            String[] span = spanTmp.split("\\.\\.");

            for (; index < spanMap.size(); ++index) {
                // wsj_1371,0,6,9..21,Shareholders
                String line = spanMap.get(index);
                String[] spanCols = line.split(",");
                String[] canSpan = spanCols[3].split("\\.\\.");

                // Start matches
                if (span[0].equals(canSpan[0]) || (nodes.size() > 0 && spans.length == 1
                        && Integer.parseInt(canSpan[1]) <= Integer.parseInt(span[1]))) {
                    if (nodes.size() == 0) {
                        contIndex = index;
                    }
                    root = trees.get(Integer.parseInt(spanCols[1]));
                    List<Tree> leaves = root.getLeaves();
                    int start = Integer.parseInt(spanCols[2]);
                    Tree node = root.getNodeNumber(start);
                    int nodeNum = 0;

                    for (; nodeNum < leaves.size(); ++nodeNum) {
                        Tree potNode = leaves.get(nodeNum);
                        if (node.equals(potNode)) {
                            int tmp = potNode.nodeNumber(root);
                            String tmpSpan = spanHashMap.get(spanCols[1] + ":" + tmp);
                            if (tmpSpan.equals(spanCols[3])) {
                                break;
                            }
                        }
                    }

                    if (corpus == Type.PDTB) {
                        if (!done.contains(nodeNum) && cols[5].contains(node.value().trim())) {
                            done.add(nodeNum);
                            nodes.add(new Node(node, nodeNum));
                        }
                    } else {
                        if (!done.contains(nodeNum)) {
                            done.add(nodeNum);
                            nodes.add(new Node(node, nodeNum));
                        }
                    }
                    if (span[1].equals(canSpan[1])) {
                        ++index;
                        break;
                    }
                }
            }
        }

        if (!nodes.isEmpty()) {
            String feature = printFeature(root, nodes, label);
            features.add(new String[] { feature, rel });
        }
    }

    return features;
}

From source file:sg.edu.nus.comp.pdtb.parser.ArgPosComp.java

License:Open Source License

private String printFeature(Tree root, List<Node> nodes, String label) {
    StringBuilder tmp = new StringBuilder();
    StringBuilder tmp2 = new StringBuilder();
    for (Node node : nodes) {
        if (node.tree.parent(root) != null) {
            tmp.append(node.tree.parent(root).value() + " ");
            tmp2.append(node.tree.value() + " ");
        }//from   w w w.  j  a va2 s .co m
    }
    String POS = tmp.toString().trim().replace(' ', '_');
    String connStr = tmp2.toString().trim().replace(' ', '_');

    if (connStr.equalsIgnoreCase("if_then") || connStr.equalsIgnoreCase("either_or")
            || connStr.equalsIgnoreCase("neither_nor")) {
        connStr = connStr.replaceAll("_", "..");
    }
    List<Tree> leaves = root.getLeaves();

    int firstNodeNum = nodes.get(0).index;

    Tree prevNode = firstNodeNum > 0 ? leaves.get(--firstNodeNum) : null;

    StringBuilder feature = new StringBuilder();

    feature.append("conn:");
    feature.append(connStr);
    feature.append(' ');

    feature.append("conn_POS:");
    feature.append(POS);
    feature.append(' ');

    if (!connStr.contains("..")) {

        int pos = nodes.get(0).index;
        if (pos <= 2) {
            feature.append("sent_pos:");
            feature.append(pos);
            feature.append(' ');
        } else {
            pos = nodes.get(nodes.size() - 1).index;
            if (pos >= leaves.size() - 3) {
                int pos2 = pos - leaves.size();
                feature.append("sent_pos:");
                feature.append(pos2);
                feature.append(' ');
            }
        }

        if (prevNode != null) {
            while (prevNode.parent(root).value().equals("-NONE-") && firstNodeNum > 0) {
                prevNode = leaves.get(--firstNodeNum);
            }

            if (prevNode != null) {

                String prevPOS = prevNode.parent(root).value().replace(' ', '_');
                String prev = prevNode.value().replace(' ', '_');

                feature.append("prev1:");
                feature.append(prev);
                feature.append(' ');

                feature.append("prev1_POS:");
                feature.append(prevPOS);
                feature.append(' ');

                feature.append("with_prev1_full:");
                feature.append(prev);
                feature.append('_');
                feature.append(connStr);
                feature.append(' ');

                feature.append("with_prev1_POS_full:");
                feature.append(prevPOS);
                feature.append('_');
                feature.append(POS);
                feature.append(' ');
                if (firstNodeNum > 0) {
                    Tree prev2Node = leaves.get(--firstNodeNum);
                    if (prev2Node != null) {
                        while (prev2Node.parent(root).value().equals("-NONE-") && firstNodeNum > 0) {
                            prev2Node = leaves.get(--firstNodeNum);
                        }

                        if (prev2Node != null) {
                            String prev2POS = prev2Node.parent(root).value().replace(' ', '_');
                            String prev2 = prev2Node.value().replace(' ', '_');

                            feature.append("prev2:");
                            feature.append(prev2);
                            feature.append(' ');

                            feature.append("prev2_POS:");
                            feature.append(prev2POS);
                            feature.append(' ');

                            feature.append("with_prev2_full:");
                            feature.append(prev2);
                            feature.append('_');
                            feature.append(connStr);
                            feature.append(' ');

                            feature.append("with_prev2_POS_full:");
                            feature.append(prev2POS);
                            feature.append('_');
                            feature.append(POS);
                            feature.append(' ');
                        }
                    }
                }
            }
        }

    }

    feature.append(label.replace(' ', '_'));

    return feature.toString().replaceAll("/", "\\\\/");
}