Examples with DSubTree edu.rice.cs.caper.bayou.core.dsl.DSubTree used on opensource projects

Example 16 with DSubTree

use of edu.rice.cs.caper.bayou.core.dsl.DSubTree in project bayou by capergroup.

the class JaccardAPICallsMetric method compute.

/**
 * Computes the minimum Jaccard distance on the set of API calls
 * between the original and the predicted ASTs.
 */
@Override
public float compute(DSubTree originalAST, List<DSubTree> predictedASTs) {
    List<Float> jaccard = new ArrayList<>();
    jaccard.add((float) 1);
    Set<DAPICall> A = originalAST.bagOfAPICalls();
    for (DSubTree predictedAST : predictedASTs) {
        Set<DAPICall> B = predictedAST.bagOfAPICalls();
        // A union B
        Set<DAPICall> AunionB = new HashSet<>();
        AunionB.addAll(A);
        AunionB.addAll(B);
        // A intersect B
        Set<DAPICall> AinterB = new HashSet<>();
        AinterB.addAll(A);
        AinterB.retainAll(B);
        jaccard.add(1 - ((float) AinterB.size()) / AunionB.size());
    }
    return Metric.min(jaccard);
}

Example 17 with DSubTree

use of edu.rice.cs.caper.bayou.core.dsl.DSubTree in project bayou by capergroup.

the class JaccardSequencesMetric method compute.

/**
 * Computes the minimum Jaccard distance on the set of sequences of API calls
 * between the original and the predicted ASTs.
 */
@Override
public float compute(DSubTree originalAST, List<DSubTree> predictedASTs) {
    List<Float> jaccard = new ArrayList<>();
    jaccard.add((float) 1);
    Set<Sequence> A;
    try {
        List<Sequence> _A = new ArrayList<>();
        _A.add(new Sequence());
        originalAST.updateSequences(_A, 999, 999);
        A = new HashSet<>(_A);
    } catch (DASTNode.TooManySequencesException | DASTNode.TooLongSequenceException e) {
        return (float) 1;
    }
    for (DSubTree predictedAST : predictedASTs) {
        Set<Sequence> B;
        try {
            List<Sequence> _B = new ArrayList<>();
            _B.add(new Sequence());
            predictedAST.updateSequences(_B, 999, 999);
            B = new HashSet<>(_B);
        } catch (DASTNode.TooManySequencesException | DASTNode.TooLongSequenceException e) {
            jaccard.add((float) 1);
            continue;
        }
        // A union B
        Set<Sequence> AunionB = new HashSet<>();
        AunionB.addAll(A);
        AunionB.addAll(B);
        // A intersect B
        Set<Sequence> AinterB = new HashSet<>();
        AinterB.addAll(A);
        AinterB.retainAll(B);
        jaccard.add(1 - ((float) AinterB.size()) / AunionB.size());
    }
    return Metric.min(jaccard);
}

Example 18 with DSubTree

use of edu.rice.cs.caper.bayou.core.dsl.DSubTree in project bayou by capergroup.

the class MetricCalculator method execute.

public void execute() throws IOException {
    if (cmdLine == null)
        return;
    int topk = cmdLine.hasOption("t") ? Integer.parseInt(cmdLine.getOptionValue("t")) : 3;
    Metric metric;
    String m = cmdLine.getOptionValue("m");
    switch(m) {
        case "equality-ast":
            metric = new EqualityASTMetric();
            break;
        case "jaccard-sequences":
            metric = new JaccardSequencesMetric();
            break;
        case "jaccard-api-calls":
            metric = new JaccardAPICallsMetric();
            break;
        case "num-control-structures":
            metric = new NumControlStructuresMetric();
            break;
        case "num-statements":
            metric = new NumStatementsMetric();
            break;
        case "latency":
            metric = null;
            break;
        default:
            System.err.println("invalid metric: " + cmdLine.getOptionValue("m"));
            return;
    }
    List<JSONInputFormat.DataPoint> data = JSONInputFormat.readData(cmdLine.getOptionValue("f"));
    List<Float> values = new ArrayList<>();
    for (JSONInputFormat.DataPoint datapoint : data) {
        DSubTree originalAST = datapoint.ast;
        List<DSubTree> predictedASTs = datapoint.out_asts.subList(0, Math.min(topk, datapoint.out_asts.size()));
        if (m.equals("latency"))
            values.add(datapoint.latency);
        else
            values.add(metric.compute(originalAST, predictedASTs));
    }
    float average = Metric.mean(values);
    float stdv = Metric.standardDeviation(values);
    System.out.println(String.format("%.2f,%.2f", average, stdv));
}

Example 19 with DSubTree

use of edu.rice.cs.caper.bayou.core.dsl.DSubTree in project bayou by capergroup.

the class JaccardSequencesMetric method compute.

/**
 * Computes the minimum Jaccard distance on the set of sequences of API calls
 * between the original and the predicted ASTs.
 */
@Override
public float compute(DSubTree originalAST, List<DSubTree> predictedASTs, String aggregate) {
    List<Float> jaccard = new ArrayList<>();
    jaccard.add((float) 1);
    Set<Sequence> A;
    try {
        List<Sequence> _A = new ArrayList<>();
        _A.add(new Sequence());
        originalAST.updateSequences(_A, 999, 999);
        A = new HashSet<>(_A);
    } catch (DASTNode.TooManySequencesException | DASTNode.TooLongSequenceException e) {
        return (float) 1;
    }
    for (DSubTree predictedAST : predictedASTs) {
        Set<Sequence> B;
        try {
            List<Sequence> _B = new ArrayList<>();
            _B.add(new Sequence());
            predictedAST.updateSequences(_B, 999, 999);
            B = new HashSet<>(_B);
        } catch (DASTNode.TooManySequencesException | DASTNode.TooLongSequenceException e) {
            jaccard.add((float) 1);
            continue;
        }
        // A union B
        Set<Sequence> AunionB = new HashSet<>();
        AunionB.addAll(A);
        AunionB.addAll(B);
        // A intersect B
        Set<Sequence> AinterB = new HashSet<>();
        AinterB.addAll(A);
        AinterB.retainAll(B);
        jaccard.add(1 - ((float) AinterB.size()) / AunionB.size());
    }
    return Metric.aggregate(jaccard, aggregate);
}

Example 20 with DSubTree

use of edu.rice.cs.caper.bayou.core.dsl.DSubTree in project bayou by capergroup.

the class NumStatementsMetric method compute.

/**
 * Computes the minimum ratio of the difference between the number of
 * statements in the original vs predicted ASTs.
 */
@Override
public float compute(DSubTree originalAST, List<DSubTree> predictedASTs, String aggregate) {
    int original = originalAST.numStatements();
    List<Integer> diffs = new ArrayList<>();
    diffs.add(original);
    for (DSubTree predictedAST : predictedASTs) {
        int predicted = predictedAST.numStatements();
        int diff_predicted = Math.abs(predicted - original);
        diffs.add(diff_predicted);
    }
    float aggr_diff = Metric.aggregate(diffs, aggregate);
    return aggr_diff / original;
}