001/**
002 * Copyright (c) 2011, The University of Southampton and the individual contributors.
003 * All rights reserved.
004 *
005 * Redistribution and use in source and binary forms, with or without modification,
006 * are permitted provided that the following conditions are met:
007 *
008 *   *  Redistributions of source code must retain the above copyright notice,
009 *      this list of conditions and the following disclaimer.
010 *
011 *   *  Redistributions in binary form must reproduce the above copyright notice,
012 *      this list of conditions and the following disclaimer in the documentation
013 *      and/or other materials provided with the distribution.
014 *
015 *   *  Neither the name of the University of Southampton nor the names of its
016 *      contributors may be used to endorse or promote products derived from this
017 *      software without specific prior written permission.
018 *
019 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
020 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
021 * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
022 * DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR
023 * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
024 * (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
025 * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON
026 * ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
027 * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
028 * SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
029 */
030package org.openimaj.hadoop.tools.image.indexing;
031
032import java.io.ByteArrayInputStream;
033import java.io.ByteArrayOutputStream;
034import java.io.DataOutputStream;
035import java.io.File;
036import java.io.IOException;
037import java.net.URI;
038import java.util.List;
039
040import org.apache.hadoop.conf.Configured;
041import org.apache.hadoop.filecache.DistributedCache;
042import org.apache.hadoop.fs.Path;
043import org.apache.hadoop.io.BytesWritable;
044import org.apache.hadoop.io.Text;
045import org.apache.hadoop.mapreduce.Job;
046import org.apache.hadoop.mapreduce.Mapper;
047import org.apache.hadoop.mapreduce.lib.output.MultipleOutputs;
048import org.apache.hadoop.mapreduce.lib.output.SequenceFileOutputFormat;
049import org.apache.hadoop.util.Tool;
050import org.apache.hadoop.util.ToolRunner;
051import org.kohsuke.args4j.CmdLineException;
052import org.kohsuke.args4j.CmdLineParser;
053import org.kohsuke.args4j.Option;
054import org.openimaj.feature.local.list.MemoryLocalFeatureList;
055import org.openimaj.hadoop.mapreduce.TextBytesJobUtil;
056import org.openimaj.hadoop.sequencefile.SequenceFileUtility;
057import org.openimaj.image.feature.local.keypoints.Keypoint;
058import org.openimaj.image.indexing.vlad.VLADIndexerData;
059
060/**
061 * Indexer for Product-quantised VLAD-PCA features. Consumes existing
062 * local-features and requires a {@link VLADIndexerData} to provide the data.
063 * 
064 * @author Jonathon Hare (jsh2@ecs.soton.ac.uk)
065 */
066public class HadoopPqPcaVladIndexer extends Configured implements Tool {
067        /**
068         * {@link Mapper} for extracting PQ-PCA-VLAD features from sets of local
069         * features. Also outputs the raw PCA-VLAD features.
070         * 
071         * @author Jonathon Hare (jsh2@ecs.soton.ac.uk)
072         */
073        static class PqPcaVladMapper extends Mapper<Text, BytesWritable, Text, BytesWritable> {
074                static enum COUNTERS {
075                        EMIT, NULL;
076                }
077
078                private VLADIndexerData indexer;
079                private MultipleOutputs<Text, BytesWritable> mos;
080
081                @Override
082                protected void setup(Context context) throws IOException, InterruptedException
083                {
084                        indexer = VLADIndexerData.read(new File("vlad-data.bin"));
085                        mos = new MultipleOutputs<Text, BytesWritable>(context);
086                }
087
088                @Override
089                protected void map(Text key, BytesWritable value, Context context)
090                                throws IOException, InterruptedException
091                {
092                        final List<Keypoint> keys = MemoryLocalFeatureList.read(new ByteArrayInputStream(value.getBytes()),
093                                        Keypoint.class);
094
095                        final float[] vladData = indexer.extractPcaVlad(keys);
096
097                        if (vladData == null) {
098                                context.getCounter(COUNTERS.NULL).increment(1L);
099                                System.out.println("VLAD is null; keys has length " + keys.size());
100                                return;
101                        }
102
103                        final byte[] pqVladData = indexer.getProductQuantiser().quantise(vladData);
104
105                        mos.write("pcavlad", key, floatToBytes(vladData));
106
107                        context.write(key, new BytesWritable(pqVladData));
108                        context.getCounter(COUNTERS.EMIT).increment(1L);
109                }
110
111                BytesWritable floatToBytes(float[] arr) throws IOException {
112                        final ByteArrayOutputStream baos = new ByteArrayOutputStream();
113                        final DataOutputStream dos = new DataOutputStream(baos);
114                        for (final float f : arr)
115                                dos.writeFloat(f);
116
117                        return new BytesWritable(baos.toByteArray());
118                }
119
120                @Override
121                protected void cleanup(Context context) throws IOException,
122                                InterruptedException
123                {
124                        super.cleanup(context);
125                        mos.close();
126                }
127        }
128
129        @Option(
130                        name = "--dont-compress-output",
131                        required = false,
132                        usage = "Don't compress sequencefile records.",
133                        metaVar = "BOOLEAN")
134        private boolean dontcompress = false;
135
136        @Option(
137                        name = "--remove",
138                        aliases = "-rm",
139                        required = false,
140                        usage = "Remove the existing output location if it exists.",
141                        metaVar = "BOOLEAN")
142        private boolean replace = false;
143
144        @Option(name = "--input", aliases = "-i", required = true, usage = "Input local features file.", metaVar = "STRING")
145        private String input;
146
147        @Option(name = "--output", aliases = "-o", required = true, usage = "Output pca-vlad file.", metaVar = "STRING")
148        private String output;
149
150        @Option(name = "--indexer-data", aliases = "-id", required = true, usage = "Indexer data file.", metaVar = "STRING")
151        private String indexerData;
152
153        @Override
154        public int run(String[] args) throws Exception {
155                final CmdLineParser parser = new CmdLineParser(this);
156
157                try {
158                        parser.parseArgument(args);
159                } catch (final CmdLineException e) {
160                        System.err.println(e.getMessage());
161                        System.err.println("Usage: hadoop jar HadoopImageIndexer.jar [options]");
162                        parser.printUsage(System.err);
163                        return -1;
164                }
165
166                final Path[] paths = SequenceFileUtility.getFilePaths(input, "part");
167                final Path outputPath = new Path(output);
168
169                if (outputPath.getFileSystem(this.getConf()).exists(outputPath) && replace)
170                        outputPath.getFileSystem(this.getConf()).delete(outputPath, true);
171
172                final Job job = TextBytesJobUtil.createJob(paths, outputPath, null, this.getConf());
173                job.setJarByClass(this.getClass());
174                job.setMapperClass(PqPcaVladMapper.class);
175                job.setNumReduceTasks(0);
176
177                MultipleOutputs.addNamedOutput(job, "pcavlad", SequenceFileOutputFormat.class, Text.class, BytesWritable.class);
178
179                DistributedCache.createSymlink(job.getConfiguration());
180                DistributedCache.addCacheFile(new URI(indexerData + "#vlad-data.bin"), job.getConfiguration());
181
182                SequenceFileOutputFormat.setCompressOutput(job, !dontcompress);
183                job.waitForCompletion(true);
184
185                return 0;
186        }
187
188        /**
189         * Main method
190         * 
191         * @param args
192         * @throws Exception
193         */
194        public static void main(String[] args) throws Exception {
195                ToolRunner.run(new HadoopPqPcaVladIndexer(), args);
196        }
197}