org.openimaj.text.nlp.language.LanguageModel Maven / Gradle / Ivy

Go to download
Show more of this group Show more artifacts with this name
Show all versions of nlp Show documentation
The OpenIMAJ NLP Library contains a text pre-processing pipeline which goes from raw unstructured text to part of speech tagged stemmed text.
The newest version!
/**
 * Copyright (c) 2011, The University of Southampton and the individual contributors.
 * All rights reserved.
 *
 * Redistribution and use in source and binary forms, with or without modification,
 * are permitted provided that the following conditions are met:
 *
 *   * 	Redistributions of source code must retain the above copyright notice,
 * 	this list of conditions and the following disclaimer.
 *
 *   *	Redistributions in binary form must reproduce the above copyright notice,
 * 	this list of conditions and the following disclaimer in the documentation
 * 	and/or other materials provided with the distribution.
 *
 *   *	Neither the name of the University of Southampton nor the names of its
 * 	contributors may be used to endorse or promote products derived from this
 * 	software without specific prior written permission.
 *
 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
 * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
 * DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR
 * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
 * (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
 * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON
 * ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
 * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
 * SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 */
package org.openimaj.text.nlp.language;


import gnu.trove.map.hash.TIntObjectHashMap;
import gnu.trove.procedure.TIntObjectProcedure;

import java.io.DataInput;
import java.io.DataOutput;
import java.io.IOException;
import java.util.Arrays;
import java.util.List;
import java.util.Map;
import java.util.Map.Entry;

import no.uib.cipr.matrix.DenseMatrix;

import org.openimaj.io.ReadWriteableBinary;
import org.openimaj.io.wrappers.Readable2DArrayBinary;
import org.openimaj.io.wrappers.ReadableArrayBinary;
import org.openimaj.io.wrappers.Writeable2DArrayBinary;
import org.openimaj.io.wrappers.WriteableArrayBinary;
import org.openimaj.math.matrix.MatrixUtils;


/**
 * The data used by {@link LanguageDetector}
 *
 * @author Sina Samangooei ([email protected])
 *
 *
 */

public class LanguageModel implements ReadWriteableBinary{
	DenseMatrix naiveBayesPC; // N x 1
	DenseMatrix naiveBayesPTC; // N x M
	String[] naiveBayesClasses; // the language classes
	TIntObjectHashMap tk_output;
	int[] tk_nextmove;
	int naiveBayesNFeats;

	/**
	 * do nothing
	 */
	public LanguageModel(){}

	/**
	 * @param languageModel
	 */
	@SuppressWarnings("unchecked")
	public LanguageModel(Map languageModel){
		List nb_pc_list = (List) languageModel.get("nb_pc");
		double[][] nb_pc_darr = new double[1][nb_pc_list.size()];
		int i = 0;
		for (double value : nb_pc_list) {
			nb_pc_darr[0][i++] = value;
		}
		naiveBayesPC = new DenseMatrix(nb_pc_darr);

		List> nb_ptc_list = (List>) languageModel.get("nb_ptc");
		double[][] nb_ptc_darr = new double[nb_ptc_list.size()][nb_ptc_list.get(0).size()];
		i = 0;
		for (List row: nb_ptc_list) {
			int j = 0;
			for(double val : row){
				nb_ptc_darr[i][j++] = val;
			}
			i++;
		}
		naiveBayesPTC = new DenseMatrix(nb_ptc_darr);

		this.naiveBayesNFeats = (naiveBayesPTC.numColumns() * naiveBayesPTC.numRows()) / naiveBayesPC.numColumns();

		List nb_classes_list = (List)languageModel.get("nb_classes");
		naiveBayesClasses = nb_classes_list.toArray(new String[nb_classes_list.size()]);

		tk_output = new TIntObjectHashMap();
		Map> tk_output_map = (Map>) languageModel.get("tk_outp");
		for (Entry> entry : tk_output_map .entrySet()) {
			i = 0;
			int[] entryArr = new int[entry.getValue().size()];
			for (double entryVal : entry.getValue()) {
				entryArr[i++] = (int) entryVal;
			}
			tk_output.put(Integer.parseInt(entry.getKey()),entryArr );
		}
		List tk_nextmove_list = (List) languageModel.get("tk_nextmove");
		tk_nextmove = new int[tk_nextmove_list.size()];
		i = 0;
		for (double val : tk_nextmove_list) {
			tk_nextmove[i++] = (int)val;
		}
	}

	@Override
	public void writeBinary(final DataOutput out) throws IOException {
		new Writeable2DArrayBinary(MatrixUtils.mtjToDoubleArray(naiveBayesPC)).writeBinary(out);
		new Writeable2DArrayBinary(MatrixUtils.mtjToDoubleArray(naiveBayesPTC)).writeBinary(out);
		WriteableArrayBinary stringWriter = new WriteableArrayBinary(naiveBayesClasses) {
			@Override
			protected void writeValue(String v, DataOutput out) throws IOException {
				out.writeUTF(v);
			}
		};
		stringWriter.writeBinary(out);
		out.writeInt(tk_output.size());
		this.tk_output.forEachEntry(new TIntObjectProcedure() {

			@Override
			public boolean execute(int key, int[] value) {
				try {
					out.writeInt(key);
					out.writeInt(value.length);
					for (int i : value) {
						out.writeInt(i);
					}
				} catch (IOException e) {
					return false;
				}
				return true;
			}
		});
		out.writeInt(this.tk_nextmove.length);
		for (int nextmove : this.tk_nextmove) {
			out.writeInt(nextmove);
		}
	}

	@Override
	public byte[] binaryHeader() {
		return "LANGMODEL".getBytes();
	}

	@Override
	public void readBinary(DataInput in) throws IOException {
		Readable2DArrayBinary matrixReader = new Readable2DArrayBinary(null);
		matrixReader.readBinary(in);
		naiveBayesPC = new DenseMatrix(matrixReader.value);

		matrixReader.readBinary(in);
		naiveBayesPTC = new DenseMatrix(matrixReader.value);

		this.naiveBayesNFeats = (naiveBayesPTC.numColumns() * naiveBayesPTC.numRows()) / naiveBayesPC.numColumns();

		ReadableArrayBinary readableClasses = new ReadableArrayBinary(null){

			@Override
			protected String readValue(DataInput in) throws IOException {
				return in.readUTF();
			}

			@Override
			protected String[] createEmpty(int sz) throws IOException {
				return new String[sz];
			}
		};
		readableClasses.readBinary(in);
		this.naiveBayesClasses = readableClasses.value;

		int nTKOut = in.readInt();
		this.tk_output = new TIntObjectHashMap(nTKOut);
		for (int i = 0; i < nTKOut; i++) {
			int key = in.readInt();
			int length = in.readInt();
			int[] data = new int[length];
			for (int j = 0; j < length; j++) {
				data[j] = in.readInt();
			}
			this.tk_output.put(key, data);
		}
		int nextMoveLength = in.readInt();
		this.tk_nextmove = new int[nextMoveLength];
		for (int i = 0; i < nextMoveLength; i++) {
			this.tk_nextmove[i] = in.readInt();
		}
	}

	@Override
	public boolean equals(Object other){
		if(!(other instanceof LanguageModel)) return false;
		final LanguageModel that = (LanguageModel) other;

		boolean equal = true;
		equal = Arrays.deepEquals(this.naiveBayesClasses, that.naiveBayesClasses); if(!equal) return false;
		equal = this.naiveBayesNFeats == that.naiveBayesNFeats; if(!equal) return false;
		equal = Arrays.deepEquals(MatrixUtils.mtjToDoubleArray(this.naiveBayesPC),MatrixUtils.mtjToDoubleArray(that.naiveBayesPC)); if(!equal) return false;
		equal = Arrays.deepEquals(MatrixUtils.mtjToDoubleArray(this.naiveBayesPTC),MatrixUtils.mtjToDoubleArray(that.naiveBayesPTC)); if(!equal) return false;
		equal = Arrays.equals(this.tk_nextmove,that.tk_nextmove); if(!equal) return false;
		equal = this.tk_output.forEachEntry(new TIntObjectProcedure() {

			@Override
			public boolean execute(int key, int[] value) {
				return Arrays.equals(value, that.tk_output.get(key));
			}
		});if(!equal) return false;
		return equal;
	}
}