/**
* Licensed to the Apache Software Foundation (ASF) under one or more
* contributor license agreements. See the NOTICE file distributed with
* this work for additional information regarding copyright ownership.
* The ASF licenses this file to You under the Apache License, Version 2.0
* (the "License"); you may not use this file except in compliance with
* the License. You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
package org.apache.mahout.classifier.bayes.mapreduce.common;
import java.io.IOException;
import java.util.Map;
import com.google.common.collect.Maps;
import org.apache.hadoop.io.DefaultStringifier;
import org.apache.hadoop.io.DoubleWritable;
import org.apache.hadoop.mapred.JobConf;
import org.apache.hadoop.mapred.MapReduceBase;
import org.apache.hadoop.mapred.Mapper;
import org.apache.hadoop.mapred.OutputCollector;
import org.apache.hadoop.mapred.Reporter;
import org.apache.hadoop.util.GenericsUtil;
import org.apache.mahout.common.StringTuple;
import org.apache.mahout.math.map.OpenObjectDoubleHashMap;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;
/**
* Naive Bayes Tfidf Mapper. Calculates per document statistics
*
*/
public class BayesTfIdfMapper extends MapReduceBase implements
Mapper<StringTuple,DoubleWritable,StringTuple,DoubleWritable> {
private static final Logger log = LoggerFactory.getLogger(BayesTfIdfMapper.class);
private static final StringTuple VOCAB_COUNT = new StringTuple(BayesConstants.FEATURE_SET_SIZE);
private static final DoubleWritable ONE = new DoubleWritable(1.0);
private final OpenObjectDoubleHashMap<String> labelDocumentCounts = new OpenObjectDoubleHashMap<String>();
/**
* We need to calculate the Tf-Idf of each feature in each label
*
* @param key
* The label,feature pair (can either be the freq Count or the term Document count
*/
@Override
public void map(StringTuple key,
DoubleWritable value,
OutputCollector<StringTuple,DoubleWritable> output,
Reporter reporter) throws IOException {
if (key.length() == 3) {
if (key.stringAt(0).equals(BayesConstants.WEIGHT)) {
reporter.setStatus("Bayes TfIdf Mapper: Tf: " + key);
output.collect(key, value);
} else if (key.stringAt(0).equals(BayesConstants.DOCUMENT_FREQUENCY)) {
String label = key.stringAt(1);
Double labelDocumentCount = labelDocumentCounts.get(label);
double logIdf = Math.log(labelDocumentCount / value.get());
key.replaceAt(0, BayesConstants.WEIGHT);
output.collect(key, new DoubleWritable(logIdf));
reporter.setStatus("Bayes TfIdf Mapper: log(Idf): " + key);
} else {
throw new IllegalArgumentException("Unrecognized Tuple: " + key);
}
} else if (key.length() == 2) {
if (key.stringAt(0).equals(BayesConstants.FEATURE_COUNT)) {
output.collect(VOCAB_COUNT, ONE);
reporter.setStatus("Bayes TfIdf Mapper: vocabCount");
} else {
throw new IllegalArgumentException("Unexpected Tuple: " + key);
}
}
}
@Override
public void configure(JobConf job) {
try {
this.labelDocumentCounts.clear();
Map<String,Double> labelDocCountTemp = Maps.newHashMap();
DefaultStringifier<Map<String,Double>> mapStringifier = new DefaultStringifier<Map<String,Double>>(job,
GenericsUtil.getClass(labelDocCountTemp));
String labelDocumentCountString =
job.get("cnaivebayes.labelDocumentCounts", mapStringifier.toString(labelDocCountTemp));
labelDocCountTemp = mapStringifier.fromString(labelDocumentCountString);
for (Map.Entry<String, Double> stringDoubleEntry : labelDocCountTemp.entrySet()) {
this.labelDocumentCounts.put(stringDoubleEntry.getKey(), stringDoubleEntry.getValue());
}
} catch (IOException ex) {
log.warn(ex.toString(), ex);
}
}
}