/*
* LensKit, an open source recommender systems toolkit.
* Copyright 2010-2014 LensKit Contributors. See CONTRIBUTORS.md.
* Work on LensKit has been funded by the National Science Foundation under
* grants IIS 05-34939, 08-08692, 08-12148, and 10-17697.
*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU Lesser General Public License as
* published by the Free Software Foundation; either version 2.1 of the
* License, or (at your option) any later version.
*
* This program is distributed in the hope that it will be useful, but WITHOUT
* ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
* FOR A PARTICULAR PURPOSE. See the GNU General Public License for more
* details.
*
* You should have received a copy of the GNU General Public License along with
* this program; if not, write to the Free Software Foundation, Inc., 51
* Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
*/
package org.grouplens.lenskit.knn.item.model;
import com.google.common.collect.FluentIterable;
import it.unimi.dsi.fastutil.longs.*;
import org.grouplens.lenskit.collections.LongKeyDomain;
import org.grouplens.lenskit.collections.LongUtils;
import org.grouplens.lenskit.core.Transient;
import org.grouplens.lenskit.cursors.Cursor;
import org.grouplens.lenskit.data.dao.ItemDAO;
import org.grouplens.lenskit.data.dao.ItemEventDAO;
import org.grouplens.lenskit.data.event.Event;
import org.grouplens.lenskit.data.event.Rating;
import org.grouplens.lenskit.data.event.Ratings;
import org.grouplens.lenskit.data.history.ItemEventCollection;
import org.grouplens.lenskit.transform.normalize.ItemVectorNormalizer;
import org.grouplens.lenskit.vectors.MutableSparseVector;
import org.grouplens.lenskit.vectors.SparseVector;
import org.grouplens.lenskit.vectors.VectorEntry;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;
import javax.inject.Inject;
import javax.inject.Provider;
import java.util.List;
/**
* Builder for {@link ItemItemBuildContext} that normalizes per-item, not per-user. More efficient
* when using e.g. item-based normalization. Right now it only works for rating data.
*
* @author <a href="http://www.grouplens.org">GroupLens Research</a>
*/
public class ItemwiseBuildContextProvider implements Provider<ItemItemBuildContext> {
private static final Logger logger = LoggerFactory.getLogger(ItemwiseBuildContextProvider.class);
private final ItemEventDAO itemEventDAO;
private final ItemDAO itemDAO;
private final ItemVectorNormalizer normalizer;
/**
* Construct a new build context provider.
* @param edao The item-event DAO.
* @param idao The item DAO.
* @param norm The item vector normalizer. This is applied to item rating vectors. You should
* take care to use a compatible normalizer for the item scorer (e.g. if this uses
* a {@link org.grouplens.lenskit.transform.normalize.MeanCenteringVectorNormalizer},
* then you should use {@link org.grouplens.lenskit.baseline.ItemMeanRatingItemScorer}
* for the user vector normalization in the scorer).
*/
@Inject
public ItemwiseBuildContextProvider(@Transient ItemEventDAO edao, @Transient ItemDAO idao,
@Transient ItemVectorNormalizer norm) {
itemEventDAO = edao;
itemDAO = idao;
normalizer = norm;
}
/**
* Constructs and returns a new ItemItemBuildContext.
*
* @return a new ItemItemBuildContext.
*/
@Override
public ItemItemBuildContext get() {
logger.info("constructing build context");
logger.debug("using normalizer {}", normalizer);
logger.debug("Building item data");
Long2ObjectMap<LongList> userItems = new Long2ObjectOpenHashMap<LongList>(1000);
Long2ObjectMap<SparseVector> itemVectors = new Long2ObjectOpenHashMap<SparseVector>(1000);
Cursor<ItemEventCollection<Event>> itemCursor = itemEventDAO.streamEventsByItem();
try {
for (ItemEventCollection<Event> item: itemCursor) {
if (logger.isTraceEnabled()) {
logger.trace("processing {} ratings for item {}", item.size(), item);
}
List<Rating> ratings = FluentIterable.from(item)
.filter(Rating.class)
.toList();
MutableSparseVector vector = Ratings.itemRatingVector(ratings);
normalizer.normalize(item.getItemId(), vector, vector);
for (VectorEntry e: vector) {
long user = e.getKey();
LongList uis = userItems.get(user);
if (uis == null) {
// lists are nice and fast, we only see each item once
uis = new LongArrayList();
userItems.put(user, uis);
}
uis.add(item.getItemId());
}
itemVectors.put(item.getItemId(), vector.freeze());
}
} finally {
itemCursor.close();
}
Long2ObjectMap<LongSortedSet> userItemSets = new Long2ObjectOpenHashMap<LongSortedSet>();
for (Long2ObjectMap.Entry<LongList> entry: userItems.long2ObjectEntrySet()) {
userItemSets.put(entry.getLongKey(), LongUtils.packedSet(entry.getValue()));
}
LongKeyDomain items = LongKeyDomain.fromCollection(itemVectors.keySet(), true);
SparseVector[] itemData = new SparseVector[items.domainSize()];
for (int i = 0; i < itemData.length; i++) {
long itemId = items.getKey(i);
itemData[i] = itemVectors.get(itemId);
}
logger.debug("item data completed");
return new ItemItemBuildContext(items, itemData, userItemSets);
}
}