Automatically fix hangul encoding

ValueExample.java

import java.io.UnsupportedEncodingException;
import java.util.AbstractMap;
import java.util.ArrayList;
import java.util.Arrays;
import java.util.HashMap;
import java.util.List;
import java.util.Map;

public class ValueExample {

  public static Map.Entry<String, String> guessAndCorrectEncoding(final String broken) {
    return Arrays.asList("utf-8", "euc-kr", "ksc5601", "x-windows-949", "iso-8859-1").stream()
        .map(charset -> {
          String encoded = null;
          try {
            encoded = new String(broken.getBytes(charset), "utf-8");
          } catch (UnsupportedEncodingException e) {
          }
          return new AbstractMap.SimpleEntry<String, String>(charset, encoded);
        }).filter(encoded -> !containsNonKeyboardCharacters(encoded.getValue()))
        .max((s1, s2) -> compareDouble(getEntropy(s2.getValue()), getEntropy(s1.getValue()))).get();
  }

  private static int compareDouble(double d1, double d2) {
    final double epsilon = 1E-5;
    double diff = d2 - d1;
    if (diff < epsilon) {
      return 0;
    }
    return diff > 0 ? 1 : -1;
  }

  private static boolean containsNonKeyboardCharacters(String s) {
    final int length = s.length();
    for (int codepoint, offset = 0; offset < length; offset += Character.charCount(codepoint)) {
      codepoint = s.codePointAt(offset);
      if (codepoint < 32)
        return true;
      if (codepoint > 126 && codepoint < 12593)
        return true;
      if (codepoint > 12687 && codepoint < 44032)
        return true;
      if (codepoint > 55204)
        return true;
    }
    return false;
  }

  private static double getEntropy(String s) {
    List<Integer> cps = getCodePoints(s);

    Map<Integer, Integer> freq = new HashMap<>();

    for (Integer cp : cps) {
      freq.compute(cp, (k, v) -> (v == null ? 0 : v) + 1);
    }

    double entropy = 0.0;

    for (Map.Entry<Integer, Integer> entry : freq.entrySet()) {
      double p = 1.0 * entry.getValue() / cps.size();
      entropy -= p * Math.log(p) / Math.log(2);
    }

    return entropy;
  }

  private static List<Integer> getCodePoints(String s) {
    List<Integer> cp = new ArrayList<>();

    final int length = s.length();
    for (int codepoint, offset = 0; offset < length; offset += Character.charCount(codepoint)) {
      codepoint = s.codePointAt(offset);

      cp.add(codepoint);
    }

    return cp;
  }

}