io.github.ericmedvet.jgea.problem.classification.BinaryTextFlaggingProblem Maven / Gradle / Ivy
Go to download
Show more of this group Show more artifacts with this name
Show all versions of jgea.problem Show documentation
Show all versions of jgea.problem Show documentation
Problem (benchmarks and templates) for jgea.
The newest version!
/*-
* ========================LICENSE_START=================================
* jgea-problem
* %%
* Copyright (C) 2018 - 2024 Eric Medvet
* %%
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
* =========================LICENSE_END==================================
*/
package io.github.ericmedvet.jgea.problem.classification;
import io.github.ericmedvet.jgea.problem.extraction.string.RegexGrammar;
import io.github.ericmedvet.jnb.datastructure.Pair;
import java.util.*;
import java.util.regex.Pattern;
import java.util.stream.Collectors;
import java.util.stream.Stream;
public class BinaryTextFlaggingProblem extends GrammarBasedTextFlaggingProblem {
private static final String[] REGEXES = new String[] {"101010...010101", "11111...11111", "(11110000)++"};
private static final String ALPHABET = "01";
public BinaryTextFlaggingProblem(
int size,
int length,
long seed,
int folds,
int i,
ClassificationFitness.Metric learningErrorMetric,
ClassificationFitness.Metric validationErrorMetric,
RegexGrammar.Option... options) {
super(
new TreeSet<>(ALPHABET.chars().mapToObj(c -> (char) c).collect(Collectors.toSet())),
new LinkedHashSet<>(Arrays.asList(options)),
buildData(REGEXES, ALPHABET, length, size, new Random(seed)),
folds,
i,
learningErrorMetric,
validationErrorMetric);
}
private static List> buildData(
String[] regexes, String alphabet, int length, int size, Random random) {
List positives = new ArrayList<>();
List negatives = new ArrayList<>();
List patterns = Stream.of(regexes).map(Pattern::compile).toList();
while ((positives.size() < size) || (negatives.size() < size)) {
StringBuilder sb = new StringBuilder();
while (sb.length() < length) {
sb.append(alphabet.charAt(random.nextInt(alphabet.length())));
}
if (patterns.stream().anyMatch((Pattern p) -> (p.matcher(sb).find()))) {
if (positives.size() < size) {
positives.add(sb.toString());
}
} else {
if (negatives.size() < size) {
negatives.add(sb.toString());
}
}
}
// return
List> data = new ArrayList<>();
data.addAll(positives.stream().map(s -> new Pair<>(s, Label.FOUND)).toList());
data.addAll(negatives.stream().map(s -> new Pair<>(s, Label.NOT_FOUND)).toList());
return data;
}
}