edu.ucla.sspace.vector.ViewDoubleAsDoubleSparseVector Maven / Gradle / Ivy

Go to download

Show more of this group Show more artifacts with this name
Show all versions of sspace-wordsi Show documentation

The S-Space Package is a collection of algorithms for building Semantic Spaces as well as a highly-scalable library for designing new distributional semantics algorithms. Distributional algorithms process text corpora and represent the semantic for words as high dimensional feature vectors. This package also includes matrices, vectors, and numerous clustering algorithms. These approaches are known by many names, such as word spaces, semantic spaces, or distributed semantics and rest upon the Distributional Hypothesis: words that appear in similar contexts have similar meanings.

The newest version!

/*
 * Copyright 2009 Keith Stevens 
 *
 * This file is part of the S-Space package and is covered under the terms and
 * conditions therein.
 *
 * The S-Space package is free software: you can redistribute it and/or modify
 * it under the terms of the GNU General Public License version 2 as published
 * by the Free Software Foundation and distributed hereunder to you.
 *
 * THIS SOFTWARE IS PROVIDED "AS IS" AND NO REPRESENTATIONS OR WARRANTIES,
 * EXPRESS OR IMPLIED ARE MADE.  BY WAY OF EXAMPLE, BUT NOT LIMITATION, WE MAKE
 * NO REPRESENTATIONS OR WARRANTIES OF MERCHANT- ABILITY OR FITNESS FOR ANY
 * PARTICULAR PURPOSE OR THAT THE USE OF THE LICENSED SOFTWARE OR DOCUMENTATION
 * WILL NOT INFRINGE ANY THIRD PARTY PATENTS, COPYRIGHTS, TRADEMARKS OR OTHER
 * RIGHTS.
 *
 * You should have received a copy of the GNU General Public License
 * along with this program. If not, see .
 */

package edu.ucla.sspace.vector;

import edu.ucla.sspace.util.DoubleEntry;

import java.io.Serializable;

import java.util.Arrays;


/**
 * A view of a sparse {@link DoubleVector} that that allows the backing data to
 * be resized and also viewed from an offset.  Furthermore, this class allows
 * the viewed data to be immutable, where all mutating operations throw {@link
 * UnsupportedOperationException}.
 * 
 * @author Keith Stevens
 * @authod David Jurgens
 */
class ViewDoubleAsDoubleSparseVector extends DoubleVectorView
        implements SparseDoubleVector {

    private static final long serialVersionUID = 1L;

    /**
     * A {@link SparseVector} reference to the {@link DoubleVector} data backing
     * this view.
     */
    private final SparseVector sparseVector;

    /**
     * Creates a new {@link DoubleVector} view of the data in the provided
     * {@link DoubleVector}.
     *
     * @param v the {@code DoubleVector} to view as containing double data.
     */
    public > 
            ViewDoubleAsDoubleSparseVector(T v) {
        this(v, 0, v.length(), false);
    }

    /**
     * Creates a new, optionally immutable {@link DoubleVector} view of the data
     * in the provided {@link DoubleVector}.
     *
     * @param v the {@code DoubleVector} to view as containing double data.
     * @param isImmutable {@code true} if this view should not allow mutating
     *        operations to change the state of the backing vector
     */
    public > 
            ViewDoubleAsDoubleSparseVector(T v, boolean isImmutable) {
        this(v, 0, v.length(), isImmutable);
    }

    /**
     * Creates a new {@link DoubleVector} sub-view of the data in the provided
     * {@link DoubleVector} using the offset and length to specify a viewing
     * region.
     *
     * @param v the {@code Vector} whose data is reflected in this view.
     * @param offset the index of {@code v} at which the first index of this
     *               view starts
     * @param length the length of this view.
     */
    public > 
            ViewDoubleAsDoubleSparseVector(T v, int offset, int length) {
        this(v, offset, length, false);
    }

    /**
     * Creates a new, optionally immutable {@link DoubleVector} sub-view of the
     * data in the provided {@link DoubleVector} using the offset and length to
     * specify a viewing region.
     *
     * @param v the {@code Vector} whose data is reflected in this view.
     * @param offset the index of {@code v} at which the first index of this
     *               view starts
     * @param length the length of this view.
     * @param isImmutable {@code true} if this view should not allow mutating
     *        operations to change the state of the backing vector.
     */
    public >
           ViewDoubleAsDoubleSparseVector(T v, int offset, int length, 
                                          boolean isImmutable) {
        super(v, offset, length, isImmutable);
        sparseVector = v;
    }

    /**
     * {@inheritDoc}
     */
    public int[] getNonZeroIndices() {
        if (vectorOffset == 0 && vectorLength == sparseVector.length())
            return sparseVector.getNonZeroIndices();
        // If the sparse vector is a sub-view, calculate which indices are
        // reflected in this view
        else {
            int inRange = 0;
            int[] indices = sparseVector.getNonZeroIndices();
            for (int nz : indices) {
                if (nz >= vectorOffset && nz < vectorOffset + vectorLength)
                    inRange++;
            }
            int[] arr = new int[inRange];
            int idx = 0;
            for (int nz : indices) {
                if (nz >= vectorOffset && nz < vectorOffset + vectorLength)
                    arr[idx++] = nz;
            }
            return arr; 
        }
    }

    /**
     * {@inheritDoc}
     */
    @SuppressWarnings("unchecked") @Override 
    public double magnitude() {
        // Check whether the current magnitude is valid and if not, recompute it
        if (magnitude < 0) {
            double m = 0;
            // Special case if we can iterate in time linear to the number of
            // non-zero values
            if (sparseVector instanceof Iterable) {
                for (DoubleEntry e : (Iterable)sparseVector) {
                    int idx = e.index();
                    if (idx >= vectorOffset 
                            && idx < vectorOffset + vectorLength) {
                        double d = e.value();
                        m += d * d;
                    }
                }
            }
            else {
                for (int nz : sparseVector.getNonZeroIndices()) {
                    if (nz >= vectorOffset 
                            && nz < vectorOffset + vectorLength) {
                        double d = doubleVector.get(nz);
                        m += d * d;
                    }
                }
            }
            magnitude = Math.sqrt(m);
        }
        return magnitude;
    }
}