SubcommandClassify.java

package de.uni_passau.fim.se2.sa.readability.subcommands;

import de.uni_passau.fim.se2.sa.readability.utils.Classify;
import picocli.CommandLine.Command;
import picocli.CommandLine.Model.CommandSpec;
import picocli.CommandLine.Option;
import picocli.CommandLine.ParameterException;
import picocli.CommandLine.Spec;
import weka.classifiers.Evaluation;
import weka.core.Instances;

import java.io.File;
import java.util.concurrent.Callable;

@Command(
        name = "classify",
        description = "Classify readability of .jsnp Java snippets"
)
public class SubcommandClassify implements Callable<Integer> {

    @Spec
    CommandSpec spec;

    private File data;

    @Option(
            names = {"-d", "--data"},
            description = "The data .csv file to train the model on.",
            required = true
    )
    public void setDataFile(File dataFile) {
        if (!dataFile.exists() || !dataFile.isFile()) {
            throw new ParameterException(spec.commandLine(), "The data file does not exist or is not a file.");
        }
        data = dataFile;
    }

    public Integer call() {
        try {
            Instances dataset = Classify.loadDataset(data);
            Evaluation eval = Classify.trainAndEvaluate(dataset);
            printResults(eval);
            return 0;
        } catch (Exception e) {
            System.out.println(e.getMessage());
            return 1;
        }
    }

    /**
     * Prints the results of the classification evaluation.
     *
     * @param eval hosts the classification results.
     */
    private static void printResults(final Evaluation eval) {
        System.out.println(eval.toSummaryString());
        System.out.printf("%-20s%.2f%n", "Accuracy", eval.pctCorrect());
        System.out.printf("%-20s%.2f%n", "Area Under ROC", eval.areaUnderROC(0));
        System.out.printf("%-20s%.2f%n", "F-Score", eval.fMeasure(0));
    }
}