Example usage for edu.stanford.nlp.util Pair Pair

List of usage examples for edu.stanford.nlp.util Pair Pair

  1. HOME
  2. Java
  3. edu.stanford.nlp
  4. edu.stanford.nlp.util.*
  5. Pair
  6. Pair

Introduction

In this page you can find the example usage for edu.stanford.nlp.util Pair Pair.

Prototype

public Pair(T1 first, T2 second)

Source Link

Usage

From source file:BuildBinarizedDataset.java

public static void extractLabels(Map<Pair<Integer, Integer>, String> spanToLabels, List<HasWord> tokens,
        String line) {//from  w w w  .ja  v  a2s.c  om
    String[] pieces = line.trim().split("\\s+");
    if (pieces.length == 0) {
        return;
    }
    if (pieces.length == 1) {
        String error = "Found line with label " + line + " but no tokens to associate with that line";
        throw new RuntimeException(error);
    }

    //TODO: BUG: The pieces are tokenized differently than the splitting, e.g., on possessive markers as in "actors' expenses"
    for (int i = 0; i < tokens.size() - pieces.length + 2; ++i) {
        boolean found = true;
        for (int j = 1; j < pieces.length; ++j) {
            if (!tokens.get(i + j - 1).word().equals(pieces[j])) {
                found = false;
                break;
            }
        }
        if (found) {
            spanToLabels.put(new Pair<>(i, i + pieces.length - 1), pieces[0]);
        }
    }
}

From source file:edu.cmu.ark.AnalysisUtilities.java

License:Open Source License

public static void addPeriodIfNeeded(Tree input) {
    String tregexOpStr = "ROOT < (S=mainclause !< /\\./)";
    TregexPattern matchPattern = TregexPatternFactory.getPattern(tregexOpStr);
    TregexMatcher matcher = matchPattern.matcher(input);

    if (matcher.find()) {
        TsurgeonPattern p;//from ww  w . ja  v a2s  . c o  m
        List<TsurgeonPattern> ps = new ArrayList<TsurgeonPattern>();
        List<Pair<TregexPattern, TsurgeonPattern>> ops = new ArrayList<Pair<TregexPattern, TsurgeonPattern>>();

        ps.add(Tsurgeon.parseOperation("insert (. .) >-1 mainclause"));
        p = Tsurgeon.collectOperations(ps);
        ops.add(new Pair<TregexPattern, TsurgeonPattern>(matchPattern, p));
        Tsurgeon.processPatternsOnTree(ops, input);
    }
}

From source file:edu.cmu.ark.AnalysisUtilities.java

License:Open Source License

/**
 * Remove traces and non-terminal decorations (e.g., "-SUBJ" in "NP-SUBJ") from a Penn Treebank-style tree.
 *
 * @param inputTree//w  w  w.  ja  v a 2 s  .c  o  m
 */
public void normalizeTree(Tree inputTree) {
    inputTree.label().setFromString("ROOT");

    List<Pair<TregexPattern, TsurgeonPattern>> ops = new ArrayList<Pair<TregexPattern, TsurgeonPattern>>();
    List<TsurgeonPattern> ps = new ArrayList<TsurgeonPattern>();
    String tregexOpStr;
    TregexPattern matchPattern;
    TsurgeonPattern p;
    TregexMatcher matcher;

    tregexOpStr = "/\\-NONE\\-/=emptynode";
    matchPattern = TregexPatternFactory.getPattern(tregexOpStr);
    matcher = matchPattern.matcher(inputTree);
    ps.add(Tsurgeon.parseOperation("prune emptynode"));
    matchPattern = TregexPatternFactory.getPattern(tregexOpStr);
    p = Tsurgeon.collectOperations(ps);
    ops.add(new Pair<TregexPattern, TsurgeonPattern>(matchPattern, p));
    Tsurgeon.processPatternsOnTree(ops, inputTree);

    Label nonterminalLabel;

    tregexOpStr = "/.+\\-.+/=nonterminal < __";
    matchPattern = TregexPatternFactory.getPattern(tregexOpStr);
    matcher = matchPattern.matcher(inputTree);
    while (matcher.find()) {
        nonterminalLabel = matcher.getNode("nonterminal");
        if (nonterminalLabel == null)
            continue;
        nonterminalLabel.setFromString(tlp.basicCategory(nonterminalLabel.value()));
    }

}

From source file:edu.cmu.ark.AnalysisUtilities.java

License:Open Source License

/**
 * remove extra quotation marks//from   w w w  . j a v  a2  s. com
 * (a hack due to annoying PTB conventions by which quote marks aren't in the same consituent)
 *
 * @param input
 */
public static void removeExtraQuotes(Tree input) {
    List<Pair<TregexPattern, TsurgeonPattern>> ops = new ArrayList<Pair<TregexPattern, TsurgeonPattern>>();
    String tregexOpStr;
    TregexPattern matchPattern;
    TsurgeonPattern p;
    List<TsurgeonPattern> ps;

    ps = new ArrayList<TsurgeonPattern>();
    tregexOpStr = "ROOT [ << (``=quote < `` !.. ('' < '')) | << (''=quote < '' !,, (`` < ``)) ] ";
    matchPattern = TregexPatternFactory.getPattern(tregexOpStr);
    ps.add(Tsurgeon.parseOperation("prune quote"));
    p = Tsurgeon.collectOperations(ps);
    ops.add(new Pair<TregexPattern, TsurgeonPattern>(matchPattern, p));
    Tsurgeon.processPatternsOnTree(ops, input);

}

From source file:edu.cmu.ark.nlp.sent.SentenceSimplifier.java

License:Open Source License

/**
 * /*from w  w w . j  a va2  s. c om*/
 * e.g., However, John did not study. -> John did not study.
 * 
 * @param q
 * @return
 */
private boolean removeClauseLevelModifiers(Question q) {
    List<Pair<TregexPattern, TsurgeonPattern>> ops;
    String tregexOpStr;
    TregexPattern matchPattern;
    TsurgeonPattern p;
    List<TsurgeonPattern> ps;

    boolean modified = false;

    //remove subordinate clauses and various phrases
    //leave conditional antecedents (i.e., with "if" or "unless" as complementizers.  punt on "even if")
    tregexOpStr = "ROOT=root < (S=mainclause < (/SBAR|ADVP|ADJP|CC|PP|S|NP/=fronted !< (IN < if|unless) !$ ``  $++ NP=subject))";
    matchPattern = TregexPatternFactory.getPattern(tregexOpStr);
    TregexMatcher matcher = matchPattern.matcher(q.getIntermediateTree());
    if (matcher.find()) {
        ops = new ArrayList<Pair<TregexPattern, TsurgeonPattern>>();
        ps = new ArrayList<TsurgeonPattern>();
        tregexOpStr = "ROOT=root < (S=mainclause < (/[,:]/=comma $ (/SBAR|ADVP|ADJP|CC|PP|S|NP/=fronted !< (IN < if|unless) $++ NP=subject)))";
        matchPattern = TregexPatternFactory.getPattern(tregexOpStr);
        ps.add(Tsurgeon.parseOperation("prune comma"));
        p = Tsurgeon.collectOperations(ps);
        ops.add(new Pair<TregexPattern, TsurgeonPattern>(matchPattern, p));
        Tsurgeon.processPatternsOnTree(ops, q.getIntermediateTree());

        ops = new ArrayList<Pair<TregexPattern, TsurgeonPattern>>();
        ps = new ArrayList<TsurgeonPattern>();
        tregexOpStr = "ROOT=root < (S=mainclause < (/SBAR|ADVP|ADJP|CC|PP|S|NP/=fronted !< (IN < if|unless) $++ NP=subject))";
        matchPattern = TregexPatternFactory.getPattern(tregexOpStr);
        ps.add(Tsurgeon.parseOperation("prune fronted"));
        p = Tsurgeon.collectOperations(ps);
        ops.add(new Pair<TregexPattern, TsurgeonPattern>(matchPattern, p));
        Tsurgeon.processPatternsOnTree(ops, q.getIntermediateTree());

        addQuotationMarksIfNeeded(q.getIntermediateTree());
        if (this.getComputeFeatures)
            q.setFeatureValue("removedClauseLevelModifiers", 1.0);
        modified = true;
    }

    return modified;
}

From source file:edu.cmu.ark.nlp.sent.SentenceSimplifier.java

License:Open Source License

/**
 * /* w  ww.ja  va2s  . co  m*/
 * e.g., John studied, hoping to get a good grade. -> John studied.
 * 
 * @param input
 * @return whether or not a change was made
 */
private boolean removeVerbalModifiersAfterCommas(Question q) {
    List<Pair<TregexPattern, TsurgeonPattern>> ops;
    String tregexOpStr;
    TregexPattern matchPattern;
    TsurgeonPattern p;
    List<TsurgeonPattern> ps;

    ops = new ArrayList<Pair<TregexPattern, TsurgeonPattern>>();

    tregexOpStr = "ROOT=root << (VP !< VP < (/,/=comma $+ /[^`].*/=modifier))";
    matchPattern = TregexPatternFactory.getPattern(tregexOpStr);

    //remove modifiers
    ps = new ArrayList<TsurgeonPattern>();
    if (matchPattern.matcher(q.getIntermediateTree()).find()) {
        ps.add(Tsurgeon.parseOperation("prune modifier"));
        p = Tsurgeon.collectOperations(ps);
        ops.add(new Pair<TregexPattern, TsurgeonPattern>(matchPattern, p));
        Tsurgeon.processPatternsOnTree(ops, q.getIntermediateTree());

        //now remove the comma
        ops.clear();
        ps.clear();
        tregexOpStr = "ROOT=root << (VP !< VP < /,/=comma)";
        matchPattern = TregexPatternFactory.getPattern(tregexOpStr);
        ps.add(Tsurgeon.parseOperation("prune comma"));
        p = Tsurgeon.collectOperations(ps);
        ops.add(new Pair<TregexPattern, TsurgeonPattern>(matchPattern, p));
        Tsurgeon.processPatternsOnTree(ops, q.getIntermediateTree());
        addQuotationMarksIfNeeded(q.getIntermediateTree());

        if (this.getComputeFeatures)
            q.setFeatureValue("removedVerbalModifiersAfterCommas", 1.0);
        return true;
    } else {
        return false;
    }
}

From source file:edu.cmu.ark.nlp.sent.SentenceSimplifier.java

License:Open Source License

private void addQuotationMarksIfNeeded(Tree input) {
    String tregexOpStr;/*from www  . j a va  2  s.c  om*/
    TregexPattern matchPattern;
    TregexMatcher matcher;

    tregexOpStr = "__=parent < (/`/ !.. /'/)";

    matchPattern = TregexPatternFactory.getPattern(tregexOpStr);
    matcher = matchPattern.matcher(input);

    if (matcher.find()) {
        TsurgeonPattern p;
        List<TsurgeonPattern> ps = new ArrayList<TsurgeonPattern>();
        List<Pair<TregexPattern, TsurgeonPattern>> ops = new ArrayList<Pair<TregexPattern, TsurgeonPattern>>();

        ps.add(Tsurgeon.parseOperation("insert ('' '') >-1 parent"));
        p = Tsurgeon.collectOperations(ps);
        ops.add(new Pair<TregexPattern, TsurgeonPattern>(matchPattern, p));
        Tsurgeon.processPatternsOnTree(ops, input);
    }
}

From source file:edu.cmu.ark.nlp.sent.SentenceSimplifier.java

License:Open Source License

/**
 * Convert a non-definite determiner to "the".
 * Used when extracting from noun modifiers such as relative clauses.
 * E.g., "A tall man, who was named Bob, entered the store." 
 * -> "A tall man was named Bob."/*from  w w  w . j av a  2 s. c o m*/
 * -> "THE tall man was named Bob."
 *  
 * @param np
 */
private void makeDeterminerDefinite(Tree np) {
    String tregexOpStr = "NP !> __ <+(NP) (DT=det !< the)";
    TregexPattern matchPattern = TregexPatternFactory.getPattern(tregexOpStr);

    TsurgeonPattern p;
    List<TsurgeonPattern> ps = new ArrayList<TsurgeonPattern>();
    List<Pair<TregexPattern, TsurgeonPattern>> ops = new ArrayList<Pair<TregexPattern, TsurgeonPattern>>();

    ps.add(Tsurgeon.parseOperation("replace det (DT the)"));
    p = Tsurgeon.collectOperations(ps);
    ops.add(new Pair<TregexPattern, TsurgeonPattern>(matchPattern, p));
    Tsurgeon.processPatternsOnTree(ops, np);
}

From source file:edu.cmu.ark.nlp.sent.SentenceSimplifier.java

License:Open Source License

/**
 * e.g., John, who hoped to get a good grade, studied. -> John studied.
 *   /*  www . ja va2 s  .c om*/
 */
private boolean removeNonRestrRelClausesAndParticipials(Question q) {
    List<Pair<TregexPattern, TsurgeonPattern>> ops = new ArrayList<Pair<TregexPattern, TsurgeonPattern>>();
    String tregexOpStr;
    TregexPattern matchPattern;
    TsurgeonPattern p;
    List<TsurgeonPattern> ps;

    ps = new ArrayList<TsurgeonPattern>();
    tregexOpStr = "NP < (VP|SBAR=mod $- /,/=punc !$+ /,/ !$ CC|CONJP)";
    matchPattern = TregexPatternFactory.getPattern(tregexOpStr);
    boolean modified = false;
    if (matchPattern.matcher(q.getIntermediateTree()).find()) {
        ps.add(Tsurgeon.parseOperation("prune punc"));
        ps.add(Tsurgeon.parseOperation("prune mod"));
        p = Tsurgeon.collectOperations(ps);
        ops.add(new Pair<TregexPattern, TsurgeonPattern>(matchPattern, p));
        Tsurgeon.processPatternsOnTree(ops, q.getIntermediateTree());
        if (this.getComputeFeatures)
            q.setFeatureValue("removedNonRestrRelClausesAndParticipials", 1.0);
        modified = true;
    }

    ps = new ArrayList<TsurgeonPattern>();
    tregexOpStr = "NP < (VP|SBAR=mod $- /,/=punc $+ /,/=punc2 !$ CC|CONJP)";
    matchPattern = TregexPatternFactory.getPattern(tregexOpStr);
    if (matchPattern.matcher(q.getIntermediateTree()).find()) {
        ps.add(Tsurgeon.parseOperation("prune punc"));
        ps.add(Tsurgeon.parseOperation("prune mod"));
        ps.add(Tsurgeon.parseOperation("prune punc2"));
        p = Tsurgeon.collectOperations(ps);
        ops.add(new Pair<TregexPattern, TsurgeonPattern>(matchPattern, p));
        Tsurgeon.processPatternsOnTree(ops, q.getIntermediateTree());
        if (this.getComputeFeatures)
            q.setFeatureValue("removedNonRestrRelClausesAndParticipials", 1.0);
        modified = true;
    }

    return modified;
}

From source file:edu.cmu.ark.nlp.sent.SentenceSimplifier.java

License:Open Source License

/**
 * /* w  w  w.j a  v a  2 s .co  m*/
 * e.g., John Smith (1931-1992) was a fireman. -> John Smith was a Fireman.
 * 
 * @return whether or not a change was made
 */
private boolean removeParentheticals(Question q) {
    List<Pair<TregexPattern, TsurgeonPattern>> ops = new ArrayList<Pair<TregexPattern, TsurgeonPattern>>();
    String tregexOpStr;
    TregexPattern matchPattern;
    TsurgeonPattern p;
    List<TsurgeonPattern> ps;
    boolean res = false;

    ps = new ArrayList<TsurgeonPattern>();
    tregexOpStr = "__=parenthetical [ $- /-LRB-/=leadingpunc $+ /-RRB-/=trailingpunc "
            + " | $+ /,/=leadingpunc $- /,/=trailingpunc !$ CC|CONJP "
            + " | $+ (/:/=leadingpunc < --) $- (/:/=trailingpunc < /--/) ]";
    matchPattern = TregexPatternFactory.getPattern(tregexOpStr);
    if (matchPattern.matcher(q.getIntermediateTree()).find()) {
        ps.add(Tsurgeon.parseOperation("prune leadingpunc"));
        ps.add(Tsurgeon.parseOperation("prune parenthetical"));
        ps.add(Tsurgeon.parseOperation("prune trailingpunc"));
        p = Tsurgeon.collectOperations(ps);
        ops.add(new Pair<TregexPattern, TsurgeonPattern>(matchPattern, p));
        Tsurgeon.processPatternsOnTree(ops, q.getIntermediateTree());

        if (res)
            addQuotationMarksIfNeeded(q.getIntermediateTree());
        if (this.getComputeFeatures)
            q.setFeatureValue("removedParentheticals", 1.0);
        res = true;
    }

    ps = new ArrayList<TsurgeonPattern>();
    tregexOpStr = "PRN=parenthetical";
    matchPattern = TregexPatternFactory.getPattern(tregexOpStr);
    if (matchPattern.matcher(q.getIntermediateTree()).find()) {
        ps.add(Tsurgeon.parseOperation("prune parenthetical"));
        p = Tsurgeon.collectOperations(ps);
        ops.add(new Pair<TregexPattern, TsurgeonPattern>(matchPattern, p));
        Tsurgeon.processPatternsOnTree(ops, q.getIntermediateTree());
        if (this.getComputeFeatures)
            q.setFeatureValue("removedParentheticals", 1.0);
        res = true;
    }

    return res;
}