com.expleague.ml.io.HMMConversionPack Maven / Gradle / Ivy

Go to download

Show more of this group Show more artifacts with this name
Show all versions of jmll

Various ML methods implemented by myself and my students

There is a newer version: 1.4.9

package com.expleague.ml.io;

import com.expleague.commons.func.types.ConversionDependant;
import com.expleague.commons.func.types.ConversionPack;
import com.expleague.commons.func.types.ConversionRepository;
import com.expleague.commons.func.types.TypeConverter;
import com.expleague.commons.math.vectors.Vec;
import com.expleague.commons.seq.CharSeqReader;
import com.expleague.commons.seq.CharSeqTools;
import com.expleague.commons.seq.ReaderChopper;
import com.expleague.commons.seq.regexp.Alphabet;
import com.expleague.ml.models.hmm.HiddenMarkovModel;

/**
 * User: solar
 * Date: 25.09.17
 */
public class HMMConversionPack implements ConversionPack {
  public static class To implements TypeConverter , ConversionDependant {
    private ConversionRepository repository;

    @Override
    public void setConversionRepository(final ConversionRepository repository) {
      this.repository = repository;
    }

    @Override
    public CharSequence convert(final HiddenMarkovModel hmm) {
      StringBuilder result = new StringBuilder();
      result.append(hmm.states()).append("\n")
          .append(repository.convert(hmm.alpha(), CharSequence.class)).append("\n")
          .append(repository.convert(hmm.betta(), CharSequence.class));
      return result;
    }
  }

  public static class From implements TypeConverter , ConversionDependant {
    private ConversionRepository repository;

    @Override
    public void setConversionRepository(final ConversionRepository repository) {
      this.repository = repository;
    }

    @SuppressWarnings("unchecked")
    @Override
    public HiddenMarkovModel convert(final CharSequence source) {
      final ReaderChopper chopper = new ReaderChopper(new CharSeqReader(source));
      final int states = CharSeqTools.parseInt(chopper.chopQuiet('\n'));
      final Alphabet alpha = repository.convert(chopper.chopQuiet('\n'), Alphabet.class);
      final Vec betta = repository.convert(chopper.restQuiet(), Vec.class);
      return new HiddenMarkovModel(alpha, states, betta);
    }
  }

  @Override
  public Class> to() {
    return To.class;
  }

  @Override
  public Class> from() {
    return From.class;
  }
}