edu.ucla.sspace.util.HashIndexer 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 2011 David Jurgens
 *
 * 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.util;

import java.lang.reflect.Array;

import java.util.Collection;
import java.util.Collections;
import java.util.HashMap;
import java.util.Iterator;
import java.util.Map;
import java.util.Set;

import gnu.trove.TDecorators;
import gnu.trove.iterator.TObjectIntIterator;
import gnu.trove.map.TObjectIntMap;
import gnu.trove.map.hash.TObjectIntHashMap;

/** 
 * A utility class for mapping a set of objects to unique indices based on
 * object equality.  The indices returned by this class will always begin at
 * {@code 0}.
 *
 *  This implementation provides faster {@link #index(Object)} performance
 * than {@link ObjectIndexer} but at the expensive of having amortized constant
 * time {@link #lookup(int)} and {@link #find(int)} performance.  This class
 * will have superior performance in use cases where all objects are indexed and
 * then {@code lookup} or {@code find} is called repeatedly whithout having
 * indexed any new objects.
 *
 * @see ObjectIndexer
 */
public class HashIndexer implements Indexer, java.io.Serializable {

    private static final long serialVersionUID = 1L;

    /**
     * A mapping from each item to its index.
     */
    private final TObjectIntHashMap indices;

    /**
     * The inverse mapping from indices to the object to which they are mapped.
     * This lookup is computed as-needed and cached so long as the index mapping
     * remains valid.
     */
    private transient T[] indexLookup;

    /**
     * Creates an empty {@code HashIndexer} with no mappings.
     */
    public HashIndexer() {
        indices = new TObjectIntHashMap();
    }

    /**
     * Creates an {@code HashIndexer} with indices for all of the provided items.
     */
    public HashIndexer(Collection items) {
        this();
        for (T item : items)
            index(item);
    }

    /**
     * Creates an {@code HashIndexer} with indices for all of the provided items.
     */
    public HashIndexer(Indexer indexed) {
        this();
        for (Map.Entry e : indexed)
            indices.put(e.getKey(), e.getValue());
    }

    /**
     * {@inheritDoc}
     */
    public void clear() {
        indices.clear();
    }

    /**
     * {@inheritDoc}
     */
    public boolean contains(T item) {
        return indices.containsKey(item);
    }

    /**
     * {@inheritDoc}
     */
    public Set items() {
        return Collections.unmodifiableSet(indices.keySet());
    }

    /**
     * {@inheritDoc}
     */
    public int find(T item) {
        return (indices.containsKey(item)) 
            ? indices.get(item) : -indices.size();
    }

    /**
     * {@inheritDoc}
     */
    public int highestIndex() {
        return indices.size() - 1;
    }

    /**
     * {@inheritDoc}
     */
    public int index(T item) {
        if (indices.containsKey(item)) 
            return indices.get(item);
        synchronized(indices) {
            // Double check that the item we are currently trying to index
            // wasn't added while we were blocking
            if (indices.containsKey(item)) 
                return indices.get(item);
            int index = indices.size();
            indices.put(item, index);
            return index;
        }
    }


    /**
     * {@inheritDoc}
     */
    public boolean indexAll(Collection items) {
        boolean changed = false;
        for (T item : items) {
            if (!indices.containsKey(item)) {
                synchronized(indices) {
                    // Double check that the item we are currently trying to
                    // index wasn't added while we were blocking
                    if (indices.containsKey(item)) 
                        continue;
                    int index = indices.size();
                    indices.put(item, index);
                    changed = true;
                }
            }
        }
        return changed;
    }

    /**
     * {@inheritDoc} The returned iterator does not support {@code remove}.
     */
    public Iterator> iterator() {
        // TODO: optimize this
        return Collections.unmodifiableSet(
            TDecorators.wrap(indices).entrySet()).iterator();
    }

    /**
     * {@inheritDoc}
     */
    public T lookup(int index) {
        if (index < 0 || index >= indices.size())
            return null;
        assert indices.size() > 0;
        // If the current lookup table is invalid or has yet to be computed,
        // recalcuate it
        while (indexLookup == null || indexLookup.length != indices.size()) {
            T t = indices.keySet().iterator().next();
            @SuppressWarnings("unchecked")
            T[] tmp = (T[])(Array.newInstance(t.getClass(), indices.size()));
            TObjectIntIterator iter = indices.iterator();
            while (iter.hasNext()) {
                iter.advance();
                tmp[iter.value()] = iter.key();
            }
            // Make the assignment at the end so that the lookup table
            // construction is atomic in the sense that the table is either
            // null, or is complete for some length k.  Assigning indexLookup
            // earlier creates the race condition where the null and length
            // checks pass, but another thread is concurrently filling the
            // table.
            indexLookup = tmp;
        }
        return indexLookup[index];
    }

    /**
     * {@inheritDoc}
     */
    public Map mapping() {
        throw new Error();
    }
    
    /**
     * {@inheritDoc}
     */
    public int size() {
        return indices.size();
    }

    /**
     * {@inheritDoc}
     */
    public String toString() {
        return indices.toString();
    }
}