You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@mahout.apache.org by sr...@apache.org on 2009/10/17 11:42:32 UTC
svn commit: r826208 - in /lucene/mahout/trunk/core/src:
main/java/org/apache/mahout/cf/taste/hadoop/
main/java/org/apache/mahout/cf/taste/impl/common/
main/java/org/apache/mahout/cf/taste/impl/model/file/
main/java/org/apache/mahout/cf/taste/impl/model...
Author: srowen
Date: Sat Oct 17 09:42:31 2009
New Revision: 826208
URL: http://svn.apache.org/viewvc?rev=826208&view=rev
Log:
More small code changes from code analysis
Modified:
lucene/mahout/trunk/core/src/main/java/org/apache/mahout/cf/taste/hadoop/SlopeOneDiffsToAveragesJob.java
lucene/mahout/trunk/core/src/main/java/org/apache/mahout/cf/taste/hadoop/SlopeOnePrefsToDiffsJob.java
lucene/mahout/trunk/core/src/main/java/org/apache/mahout/cf/taste/impl/common/LongPrimitiveArrayIterator.java
lucene/mahout/trunk/core/src/main/java/org/apache/mahout/cf/taste/impl/model/file/FileDataModel.java
lucene/mahout/trunk/core/src/main/java/org/apache/mahout/cf/taste/impl/model/jdbc/AbstractJDBCDataModel.java
lucene/mahout/trunk/core/src/main/java/org/apache/mahout/classifier/Classify.java
lucene/mahout/trunk/core/src/main/java/org/apache/mahout/classifier/bayes/datastore/HBaseBayesDatastore.java
lucene/mahout/trunk/core/src/main/java/org/apache/mahout/classifier/bayes/datastore/InMemoryBayesDatastore.java
lucene/mahout/trunk/core/src/main/java/org/apache/mahout/classifier/bayes/mapreduce/bayes/BayesClassifierDriver.java
lucene/mahout/trunk/core/src/main/java/org/apache/mahout/classifier/bayes/mapreduce/bayes/BayesClassifierMapper.java
lucene/mahout/trunk/core/src/main/java/org/apache/mahout/classifier/bayes/mapreduce/bayes/BayesThetaNormalizerMapper.java
lucene/mahout/trunk/core/src/main/java/org/apache/mahout/classifier/bayes/mapreduce/bayes/BayesThetaNormalizerReducer.java
lucene/mahout/trunk/core/src/main/java/org/apache/mahout/classifier/bayes/mapreduce/common/BayesFeatureDriver.java
lucene/mahout/trunk/core/src/main/java/org/apache/mahout/classifier/bayes/mapreduce/common/BayesTfIdfDriver.java
lucene/mahout/trunk/core/src/main/java/org/apache/mahout/classifier/bayes/mapreduce/common/BayesTfIdfMapper.java
lucene/mahout/trunk/core/src/main/java/org/apache/mahout/classifier/bayes/mapreduce/common/BayesTfIdfReducer.java
lucene/mahout/trunk/core/src/main/java/org/apache/mahout/classifier/bayes/mapreduce/common/BayesWeightSummerReducer.java
lucene/mahout/trunk/core/src/main/java/org/apache/mahout/classifier/bayes/mapreduce/common/JobExecutor.java
lucene/mahout/trunk/core/src/main/java/org/apache/mahout/clustering/canopy/Canopy.java
lucene/mahout/trunk/core/src/main/java/org/apache/mahout/clustering/kmeans/Cluster.java
lucene/mahout/trunk/core/src/main/java/org/apache/mahout/common/StringTuple.java
lucene/mahout/trunk/core/src/main/java/org/apache/mahout/common/TimingStatistics.java
lucene/mahout/trunk/core/src/main/java/org/apache/mahout/common/commandline/DefaultOptionCreator.java
lucene/mahout/trunk/core/src/main/java/org/apache/mahout/common/distance/SquaredEuclideanDistanceMeasure.java
lucene/mahout/trunk/core/src/main/java/org/apache/mahout/common/distance/TanimotoDistanceMeasure.java
lucene/mahout/trunk/core/src/main/java/org/apache/mahout/df/mapred/partial/PartialBuilder.java
lucene/mahout/trunk/core/src/main/java/org/apache/mahout/df/node/Leaf.java
lucene/mahout/trunk/core/src/main/java/org/apache/mahout/df/tools/Describe.java
lucene/mahout/trunk/core/src/main/java/org/apache/mahout/matrix/SparseColumnMatrix.java
lucene/mahout/trunk/core/src/main/java/org/apache/mahout/matrix/SparseRowMatrix.java
lucene/mahout/trunk/core/src/test/java/org/apache/mahout/cf/taste/impl/similarity/SpearmanCorrelationSimilarityTest.java
lucene/mahout/trunk/core/src/test/java/org/apache/mahout/clustering/canopy/VisibleCanopy.java
lucene/mahout/trunk/core/src/test/java/org/apache/mahout/clustering/lda/TestLDAInference.java
lucene/mahout/trunk/core/src/test/java/org/apache/mahout/clustering/lda/TestMapReduce.java
lucene/mahout/trunk/core/src/test/java/org/apache/mahout/common/distance/DefaultDistanceMeasureTest.java
lucene/mahout/trunk/core/src/test/java/org/apache/mahout/df/data/DataTest.java
lucene/mahout/trunk/core/src/test/java/org/apache/mahout/df/mapred/partial/PartialBuilderTest.java
lucene/mahout/trunk/core/src/test/java/org/apache/mahout/df/mapred/partial/PartialSequentialBuilder.java
lucene/mahout/trunk/core/src/test/java/org/apache/mahout/df/mapreduce/partial/Step0JobTest.java
lucene/mahout/trunk/core/src/test/java/org/apache/mahout/matrix/VectorTest.java
Modified: lucene/mahout/trunk/core/src/main/java/org/apache/mahout/cf/taste/hadoop/SlopeOneDiffsToAveragesJob.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/core/src/main/java/org/apache/mahout/cf/taste/hadoop/SlopeOneDiffsToAveragesJob.java?rev=826208&r1=826207&r2=826208&view=diff
==============================================================================
--- lucene/mahout/trunk/core/src/main/java/org/apache/mahout/cf/taste/hadoop/SlopeOneDiffsToAveragesJob.java (original)
+++ lucene/mahout/trunk/core/src/main/java/org/apache/mahout/cf/taste/hadoop/SlopeOneDiffsToAveragesJob.java Sat Oct 17 09:42:31 2009
@@ -46,7 +46,7 @@
public final class SlopeOneDiffsToAveragesJob extends Job {
/** Logger for this class. */
- private static Logger log = Logger.getLogger(SlopeOneDiffsToAveragesJob.class);
+ private static final Logger log = Logger.getLogger(SlopeOneDiffsToAveragesJob.class);
private SlopeOneDiffsToAveragesJob(Configuration jobConf) throws IOException {
super(jobConf);
Modified: lucene/mahout/trunk/core/src/main/java/org/apache/mahout/cf/taste/hadoop/SlopeOnePrefsToDiffsJob.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/core/src/main/java/org/apache/mahout/cf/taste/hadoop/SlopeOnePrefsToDiffsJob.java?rev=826208&r1=826207&r2=826208&view=diff
==============================================================================
--- lucene/mahout/trunk/core/src/main/java/org/apache/mahout/cf/taste/hadoop/SlopeOnePrefsToDiffsJob.java (original)
+++ lucene/mahout/trunk/core/src/main/java/org/apache/mahout/cf/taste/hadoop/SlopeOnePrefsToDiffsJob.java Sat Oct 17 09:42:31 2009
@@ -47,7 +47,7 @@
public final class SlopeOnePrefsToDiffsJob extends Job {
/** Logger for this class.*/
- private static Logger log = Logger.getLogger(SlopeOnePrefsToDiffsJob.class);
+ private static final Logger log = Logger.getLogger(SlopeOnePrefsToDiffsJob.class);
private SlopeOnePrefsToDiffsJob(Configuration jobConf) throws IOException {
super(jobConf);
Modified: lucene/mahout/trunk/core/src/main/java/org/apache/mahout/cf/taste/impl/common/LongPrimitiveArrayIterator.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/core/src/main/java/org/apache/mahout/cf/taste/impl/common/LongPrimitiveArrayIterator.java?rev=826208&r1=826207&r2=826208&view=diff
==============================================================================
--- lucene/mahout/trunk/core/src/main/java/org/apache/mahout/cf/taste/impl/common/LongPrimitiveArrayIterator.java (original)
+++ lucene/mahout/trunk/core/src/main/java/org/apache/mahout/cf/taste/impl/common/LongPrimitiveArrayIterator.java Sat Oct 17 09:42:31 2009
@@ -23,7 +23,7 @@
* While long[] is an Iterable, it is not an Iterable<Long>. This adapter class
* addresses that.
*/
-public final class LongPrimitiveArrayIterator implements LongPrimitiveIterator, SkippingIterator<Long> {
+public final class LongPrimitiveArrayIterator implements LongPrimitiveIterator {
private final long[] array;
private int position;
Modified: lucene/mahout/trunk/core/src/main/java/org/apache/mahout/cf/taste/impl/model/file/FileDataModel.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/core/src/main/java/org/apache/mahout/cf/taste/impl/model/file/FileDataModel.java?rev=826208&r1=826207&r2=826208&view=diff
==============================================================================
--- lucene/mahout/trunk/core/src/main/java/org/apache/mahout/cf/taste/impl/model/file/FileDataModel.java (original)
+++ lucene/mahout/trunk/core/src/main/java/org/apache/mahout/cf/taste/impl/model/file/FileDataModel.java Sat Oct 17 09:42:31 2009
@@ -222,7 +222,7 @@
if (line.length() > 0) {
processLine(line, data, delimiter);
int currentCount = count.incrementAndGet();
- if (currentCount % 100000 == 0) {
+ if (currentCount % 1000000 == 0) {
log.info("Processed {} lines", currentCount);
}
}
Modified: lucene/mahout/trunk/core/src/main/java/org/apache/mahout/cf/taste/impl/model/jdbc/AbstractJDBCDataModel.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/core/src/main/java/org/apache/mahout/cf/taste/impl/model/jdbc/AbstractJDBCDataModel.java?rev=826208&r1=826207&r2=826208&view=diff
==============================================================================
--- lucene/mahout/trunk/core/src/main/java/org/apache/mahout/cf/taste/impl/model/jdbc/AbstractJDBCDataModel.java (original)
+++ lucene/mahout/trunk/core/src/main/java/org/apache/mahout/cf/taste/impl/model/jdbc/AbstractJDBCDataModel.java Sat Oct 17 09:42:31 2009
@@ -618,7 +618,7 @@
* only release database resources after {@link #hasNext()} has been called and has returned <code>false</code>;
* callers should make sure to "drain" the entire set of data to avoid tying up database resources.</p>
*/
- private final class ResultSetIDIterator implements LongPrimitiveIterator, SkippingIterator<Long> {
+ private final class ResultSetIDIterator implements LongPrimitiveIterator {
private final Connection connection;
private final Statement statement;
Modified: lucene/mahout/trunk/core/src/main/java/org/apache/mahout/classifier/Classify.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/core/src/main/java/org/apache/mahout/classifier/Classify.java?rev=826208&r1=826207&r2=826208&view=diff
==============================================================================
--- lucene/mahout/trunk/core/src/main/java/org/apache/mahout/classifier/Classify.java (original)
+++ lucene/mahout/trunk/core/src/main/java/org/apache/mahout/classifier/Classify.java Sat Oct 17 09:42:31 2009
@@ -141,8 +141,8 @@
log.info("Loading model from: {}", params.print());
- Algorithm algorithm = null;
- Datastore datastore = null;
+ Algorithm algorithm;
+ Datastore datastore;
String classifierType = (String) cmdLine.getValue(typeOpt);
Modified: lucene/mahout/trunk/core/src/main/java/org/apache/mahout/classifier/bayes/datastore/HBaseBayesDatastore.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/core/src/main/java/org/apache/mahout/classifier/bayes/datastore/HBaseBayesDatastore.java?rev=826208&r1=826207&r2=826208&view=diff
==============================================================================
--- lucene/mahout/trunk/core/src/main/java/org/apache/mahout/classifier/bayes/datastore/HBaseBayesDatastore.java (original)
+++ lucene/mahout/trunk/core/src/main/java/org/apache/mahout/classifier/bayes/datastore/HBaseBayesDatastore.java Sat Oct 17 09:42:31 2009
@@ -76,9 +76,9 @@
BayesConstants.LABEL_THETA_NORMALIZER, label)));
}
for (String label : labels) {
- System.out.println(label + " "
+ System.out.println(label + ' '
+ getWeightFromHbase(BayesConstants.LABEL_THETA_NORMALIZER, label)
- + " " + thetaNormalizer + " "
+ + ' ' + thetaNormalizer + ' '
+ getWeightFromHbase(BayesConstants.LABEL_THETA_NORMALIZER, label)
/ thetaNormalizer);
}
@@ -91,7 +91,7 @@
throws InvalidDatastoreException {
if (keys.containsKey(name))
return keys.get(name);
- Result r = null;
+ Result r;
if (name.equals("labelWeight")) {
r = getRowFromHbase(BayesConstants.LABEL_SUM);
} else if (name.equals("thetaNormalizer")) {
Modified: lucene/mahout/trunk/core/src/main/java/org/apache/mahout/classifier/bayes/datastore/InMemoryBayesDatastore.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/core/src/main/java/org/apache/mahout/classifier/bayes/datastore/InMemoryBayesDatastore.java?rev=826208&r1=826207&r2=826208&view=diff
==============================================================================
--- lucene/mahout/trunk/core/src/main/java/org/apache/mahout/classifier/bayes/datastore/InMemoryBayesDatastore.java (original)
+++ lucene/mahout/trunk/core/src/main/java/org/apache/mahout/classifier/bayes/datastore/InMemoryBayesDatastore.java Sat Oct 17 09:42:31 2009
@@ -72,8 +72,8 @@
"thetaNormalizer", label)));
}
for (String label : labels) {
- System.out.println( label + " " +vectorGetCell(
- "thetaNormalizer", label) +" " +thetaNormalizer + " " + vectorGetCell(
+ System.out.println( label + ' ' +vectorGetCell(
+ "thetaNormalizer", label) + ' ' +thetaNormalizer + ' ' + vectorGetCell(
"thetaNormalizer", label)/thetaNormalizer);
}
}
Modified: lucene/mahout/trunk/core/src/main/java/org/apache/mahout/classifier/bayes/mapreduce/bayes/BayesClassifierDriver.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/core/src/main/java/org/apache/mahout/classifier/bayes/mapreduce/bayes/BayesClassifierDriver.java?rev=826208&r1=826207&r2=826208&view=diff
==============================================================================
--- lucene/mahout/trunk/core/src/main/java/org/apache/mahout/classifier/bayes/mapreduce/bayes/BayesClassifierDriver.java (original)
+++ lucene/mahout/trunk/core/src/main/java/org/apache/mahout/classifier/bayes/mapreduce/bayes/BayesClassifierDriver.java Sat Oct 17 09:42:31 2009
@@ -85,7 +85,8 @@
log.info("{}",matrix.summarize());
}
- private static final ConfusionMatrix readResult(FileSystem fs, Path pathPattern, Configuration conf, BayesParameters params) throws IOException{
+ private static ConfusionMatrix readResult(FileSystem fs, Path pathPattern, Configuration conf, BayesParameters params)
+ throws IOException {
StringTuple key = new StringTuple();
DoubleWritable value = new DoubleWritable();
Modified: lucene/mahout/trunk/core/src/main/java/org/apache/mahout/classifier/bayes/mapreduce/bayes/BayesClassifierMapper.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/core/src/main/java/org/apache/mahout/classifier/bayes/mapreduce/bayes/BayesClassifierMapper.java?rev=826208&r1=826207&r2=826208&view=diff
==============================================================================
--- lucene/mahout/trunk/core/src/main/java/org/apache/mahout/classifier/bayes/mapreduce/bayes/BayesClassifierMapper.java (original)
+++ lucene/mahout/trunk/core/src/main/java/org/apache/mahout/classifier/bayes/mapreduce/bayes/BayesClassifierMapper.java Sat Oct 17 09:42:31 2009
@@ -28,7 +28,6 @@
import org.apache.mahout.classifier.ClassifierResult;
import org.apache.mahout.classifier.bayes.algorithm.BayesAlgorithm;
import org.apache.mahout.classifier.bayes.algorithm.CBayesAlgorithm;
-import org.apache.mahout.classifier.bayes.common.BayesParameters;
import org.apache.mahout.classifier.bayes.datastore.HBaseBayesDatastore;
import org.apache.mahout.classifier.bayes.datastore.InMemoryBayesDatastore;
import org.apache.mahout.classifier.bayes.exceptions.InvalidDatastoreException;
@@ -98,10 +97,10 @@
public void configure(JobConf job) {
try {
log.info("Bayes Parameter" + job.get("bayes.parameters"));
- Parameters params = BayesParameters.fromString(job.get("bayes.parameters",""));
+ Parameters params = Parameters.fromString(job.get("bayes.parameters", ""));
log.info("{}", params.print());
- Algorithm algorithm = null;
- Datastore datastore = null;
+ Algorithm algorithm;
+ Datastore datastore;
if (params.get("dataSource").equals("hdfs")) {
Modified: lucene/mahout/trunk/core/src/main/java/org/apache/mahout/classifier/bayes/mapreduce/bayes/BayesThetaNormalizerMapper.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/core/src/main/java/org/apache/mahout/classifier/bayes/mapreduce/bayes/BayesThetaNormalizerMapper.java?rev=826208&r1=826207&r2=826208&view=diff
==============================================================================
--- lucene/mahout/trunk/core/src/main/java/org/apache/mahout/classifier/bayes/mapreduce/bayes/BayesThetaNormalizerMapper.java (original)
+++ lucene/mahout/trunk/core/src/main/java/org/apache/mahout/classifier/bayes/mapreduce/bayes/BayesThetaNormalizerMapper.java Sat Oct 17 09:42:31 2009
@@ -42,7 +42,8 @@
private Map<String, Double> labelWeightSum = null;
private double sigma_jSigma_k = 0.0;
private double vocabCount = 0.0;
- private double alpha_i = 1.0;
+ //private final double alpha_i = 1.0;
+
/**
* We need to calculate the thetaNormalization factor of each label
*
Modified: lucene/mahout/trunk/core/src/main/java/org/apache/mahout/classifier/bayes/mapreduce/bayes/BayesThetaNormalizerReducer.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/core/src/main/java/org/apache/mahout/classifier/bayes/mapreduce/bayes/BayesThetaNormalizerReducer.java?rev=826208&r1=826207&r2=826208&view=diff
==============================================================================
--- lucene/mahout/trunk/core/src/main/java/org/apache/mahout/classifier/bayes/mapreduce/bayes/BayesThetaNormalizerReducer.java (original)
+++ lucene/mahout/trunk/core/src/main/java/org/apache/mahout/classifier/bayes/mapreduce/bayes/BayesThetaNormalizerReducer.java Sat Oct 17 09:42:31 2009
@@ -47,8 +47,6 @@
private HTable table;
- private HBaseConfiguration HBconf;
-
boolean useHbase = false;
@Override
@@ -91,7 +89,7 @@
else
return;
- HBconf = new HBaseConfiguration(job);
+ HBaseConfiguration HBconf = new HBaseConfiguration(job);
table = new HTable(HBconf, job.get("output.table"));
} catch (IOException e) {
log.error("Unexpected error during configuration", e);
Modified: lucene/mahout/trunk/core/src/main/java/org/apache/mahout/classifier/bayes/mapreduce/common/BayesFeatureDriver.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/core/src/main/java/org/apache/mahout/classifier/bayes/mapreduce/common/BayesFeatureDriver.java?rev=826208&r1=826207&r2=826208&view=diff
==============================================================================
--- lucene/mahout/trunk/core/src/main/java/org/apache/mahout/classifier/bayes/mapreduce/common/BayesFeatureDriver.java (original)
+++ lucene/mahout/trunk/core/src/main/java/org/apache/mahout/classifier/bayes/mapreduce/common/BayesFeatureDriver.java Sat Oct 17 09:42:31 2009
@@ -20,7 +20,6 @@
import org.apache.hadoop.fs.FileSystem;
import org.apache.hadoop.fs.Path;
import org.apache.hadoop.io.DoubleWritable;
-import org.apache.hadoop.io.Text;
import org.apache.hadoop.mapred.FileInputFormat;
import org.apache.hadoop.mapred.FileOutputFormat;
import org.apache.hadoop.mapred.JobClient;
Modified: lucene/mahout/trunk/core/src/main/java/org/apache/mahout/classifier/bayes/mapreduce/common/BayesTfIdfDriver.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/core/src/main/java/org/apache/mahout/classifier/bayes/mapreduce/common/BayesTfIdfDriver.java?rev=826208&r1=826207&r2=826208&view=diff
==============================================================================
--- lucene/mahout/trunk/core/src/main/java/org/apache/mahout/classifier/bayes/mapreduce/common/BayesTfIdfDriver.java (original)
+++ lucene/mahout/trunk/core/src/main/java/org/apache/mahout/classifier/bayes/mapreduce/common/BayesTfIdfDriver.java Sat Oct 17 09:42:31 2009
@@ -108,7 +108,7 @@
if (params.get("dataSource").equals("hbase")) {
HBaseConfiguration hc = new HBaseConfiguration(new Configuration());
HTableDescriptor ht = new HTableDescriptor(output);
- HColumnDescriptor hcd = new HColumnDescriptor(BayesConstants.HBASE_COLUMN_FAMILY+":");
+ HColumnDescriptor hcd = new HColumnDescriptor(BayesConstants.HBASE_COLUMN_FAMILY+ ':');
hcd.setBloomfilter(true);
hcd.setInMemory(true);
hcd.setMaxVersions(1);
Modified: lucene/mahout/trunk/core/src/main/java/org/apache/mahout/classifier/bayes/mapreduce/common/BayesTfIdfMapper.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/core/src/main/java/org/apache/mahout/classifier/bayes/mapreduce/common/BayesTfIdfMapper.java?rev=826208&r1=826207&r2=826208&view=diff
==============================================================================
--- lucene/mahout/trunk/core/src/main/java/org/apache/mahout/classifier/bayes/mapreduce/common/BayesTfIdfMapper.java (original)
+++ lucene/mahout/trunk/core/src/main/java/org/apache/mahout/classifier/bayes/mapreduce/common/BayesTfIdfMapper.java Sat Oct 17 09:42:31 2009
@@ -41,10 +41,10 @@
private Map<String, Double> labelDocumentCounts = null;
- private static StringTuple vocabCount = new StringTuple(
+ private static final StringTuple vocabCount = new StringTuple(
BayesConstants.FEATURE_SET_SIZE);
- private static DoubleWritable one = new DoubleWritable(1.0d);
+ private static final DoubleWritable one = new DoubleWritable(1.0);
/**
* We need to calculate the Tf-Idf of each feature in each label
Modified: lucene/mahout/trunk/core/src/main/java/org/apache/mahout/classifier/bayes/mapreduce/common/BayesTfIdfReducer.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/core/src/main/java/org/apache/mahout/classifier/bayes/mapreduce/common/BayesTfIdfReducer.java?rev=826208&r1=826207&r2=826208&view=diff
==============================================================================
--- lucene/mahout/trunk/core/src/main/java/org/apache/mahout/classifier/bayes/mapreduce/common/BayesTfIdfReducer.java (original)
+++ lucene/mahout/trunk/core/src/main/java/org/apache/mahout/classifier/bayes/mapreduce/common/BayesTfIdfReducer.java Sat Oct 17 09:42:31 2009
@@ -46,9 +46,6 @@
.getLogger(BayesTfIdfReducer.class);
private HTable table;
- private HBaseConfiguration HBconf; //reloading configuration causes
- //a new htable session to get
- //created(from HBASE IRC)
boolean useHbase = false;
@@ -112,7 +109,7 @@
else
return;
- HBconf = new HBaseConfiguration(job);
+ HBaseConfiguration HBconf = new HBaseConfiguration(job);
table = new HTable(HBconf, job.get("output.table"));
Modified: lucene/mahout/trunk/core/src/main/java/org/apache/mahout/classifier/bayes/mapreduce/common/BayesWeightSummerReducer.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/core/src/main/java/org/apache/mahout/classifier/bayes/mapreduce/common/BayesWeightSummerReducer.java?rev=826208&r1=826207&r2=826208&view=diff
==============================================================================
--- lucene/mahout/trunk/core/src/main/java/org/apache/mahout/classifier/bayes/mapreduce/common/BayesWeightSummerReducer.java (original)
+++ lucene/mahout/trunk/core/src/main/java/org/apache/mahout/classifier/bayes/mapreduce/common/BayesWeightSummerReducer.java Sat Oct 17 09:42:31 2009
@@ -59,7 +59,7 @@
sum += values.next().get();
}
reporter.setStatus("Bayes Weight Summer Reducer: " + key + " => " + sum);
- char firstChar = key.toString().charAt(0);
+ //char firstChar = key.toString().charAt(0);
if (useHbase) {
if (key.stringAt(0).equals(BayesConstants.FEATURE_SUM)) { // sum of weight
// for all
Modified: lucene/mahout/trunk/core/src/main/java/org/apache/mahout/classifier/bayes/mapreduce/common/JobExecutor.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/core/src/main/java/org/apache/mahout/classifier/bayes/mapreduce/common/JobExecutor.java?rev=826208&r1=826207&r2=826208&view=diff
==============================================================================
--- lucene/mahout/trunk/core/src/main/java/org/apache/mahout/classifier/bayes/mapreduce/common/JobExecutor.java (original)
+++ lucene/mahout/trunk/core/src/main/java/org/apache/mahout/classifier/bayes/mapreduce/common/JobExecutor.java Sat Oct 17 09:42:31 2009
@@ -35,6 +35,9 @@
/** Logger for this class. */
private static final Logger log = LoggerFactory.getLogger(BayesDriver.class);
+ private JobExecutor() {
+ }
+
/**
* Execute a bayes classification job. Input and output path are parsed from
* the input parameters.
Modified: lucene/mahout/trunk/core/src/main/java/org/apache/mahout/clustering/canopy/Canopy.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/core/src/main/java/org/apache/mahout/clustering/canopy/Canopy.java?rev=826208&r1=826207&r2=826208&view=diff
==============================================================================
--- lucene/mahout/trunk/core/src/main/java/org/apache/mahout/clustering/canopy/Canopy.java (original)
+++ lucene/mahout/trunk/core/src/main/java/org/apache/mahout/clustering/canopy/Canopy.java Sat Oct 17 09:42:31 2009
@@ -18,7 +18,6 @@
package org.apache.mahout.clustering.canopy;
import org.apache.hadoop.io.Text;
-import org.apache.hadoop.io.Writable;
import org.apache.hadoop.mapred.JobConf;
import org.apache.hadoop.mapred.OutputCollector;
import org.apache.mahout.clustering.ClusterBase;
@@ -36,7 +35,7 @@
* application of some distance metric, and a point total which is the sum of all the points and is used to compute the
* centroid when needed.
*/
-public class Canopy extends ClusterBase implements Writable {
+public class Canopy extends ClusterBase {
// keys used by Driver, Mapper, Combiner & Reducer
public static final String DISTANCE_MEASURE_KEY = "org.apache.mahout.clustering.canopy.measure";
Modified: lucene/mahout/trunk/core/src/main/java/org/apache/mahout/clustering/kmeans/Cluster.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/core/src/main/java/org/apache/mahout/clustering/kmeans/Cluster.java?rev=826208&r1=826207&r2=826208&view=diff
==============================================================================
--- lucene/mahout/trunk/core/src/main/java/org/apache/mahout/clustering/kmeans/Cluster.java (original)
+++ lucene/mahout/trunk/core/src/main/java/org/apache/mahout/clustering/kmeans/Cluster.java Sat Oct 17 09:42:31 2009
@@ -17,7 +17,6 @@
package org.apache.mahout.clustering.kmeans;
import org.apache.hadoop.io.Text;
-import org.apache.hadoop.io.Writable;
import org.apache.hadoop.mapred.JobConf;
import org.apache.hadoop.mapred.OutputCollector;
import org.apache.mahout.clustering.ClusterBase;
@@ -31,7 +30,7 @@
import java.io.IOException;
import java.util.List;
-public class Cluster extends ClusterBase implements Writable {
+public class Cluster extends ClusterBase {
private static final String ERROR_UNKNOWN_CLUSTER_FORMAT="Unknown cluster format:\n";
@@ -86,7 +85,6 @@
*/
public static Cluster decodeCluster(String formattedString) {
final int beginIndex = formattedString.indexOf('{');
- final Cluster cluster;
if (beginIndex <= 0) {
throw new IllegalArgumentException(ERROR_UNKNOWN_CLUSTER_FORMAT + formattedString);
}
@@ -94,6 +92,7 @@
final String center = formattedString.substring(beginIndex);
final char firstChar = id.charAt(0);
final boolean startsWithV = firstChar == 'V';
+ final Cluster cluster;
if (firstChar == 'C' || startsWithV) {
final int clusterId = Integer.parseInt(formattedString.substring(1,
beginIndex - 2));
Modified: lucene/mahout/trunk/core/src/main/java/org/apache/mahout/common/StringTuple.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/core/src/main/java/org/apache/mahout/common/StringTuple.java?rev=826208&r1=826207&r2=826208&view=diff
==============================================================================
--- lucene/mahout/trunk/core/src/main/java/org/apache/mahout/common/StringTuple.java (original)
+++ lucene/mahout/trunk/core/src/main/java/org/apache/mahout/common/StringTuple.java Sat Oct 17 09:42:31 2009
@@ -21,20 +21,16 @@
import java.io.DataOutput;
import java.io.IOException;
import java.util.ArrayList;
-import java.util.Collection;
import java.util.Collections;
import java.util.List;
import org.apache.hadoop.hbase.util.Bytes;
-import org.apache.hadoop.io.Writable;
import org.apache.hadoop.io.WritableComparable;
/**
* An Ordered List of Strings which can be used in a Hadoop Map/Reduce Job
- *
- *
*/
-public class StringTuple implements Writable, WritableComparable<StringTuple> {
+public final class StringTuple implements WritableComparable<StringTuple> {
private List<String> tuple = new ArrayList<String>();
@@ -45,14 +41,16 @@
add(firstEntry);
}
- public StringTuple(Collection<String> entries) {
- for(String entry: entries)
+ public StringTuple(Iterable<String> entries) {
+ for (String entry : entries) {
add(entry);
+ }
}
-
+
public StringTuple(String[] entries) {
- for(String entry: entries)
+ for (String entry : entries) {
add(entry);
+ }
}
/**
@@ -107,7 +105,7 @@
@Override
public String toString() {
return tuple.toString();
- };
+ }
@Override
public int hashCode() {
@@ -155,14 +153,22 @@
@Override
public int compareTo(StringTuple otherTuple) {
- int min = Math.min(this.length(), otherTuple.length());
+ int thisLength = length();
+ int otherLength = otherTuple.length();
+ int min = Math.min(thisLength, otherLength);
for (int i = 0; i < min; i++) {
int ret = this.tuple.get(i).compareTo(otherTuple.stringAt(i));
- if (ret == 0)
- continue;
- return ret;
+ if (ret != 0) {
+ return ret;
+ }
+ }
+ if (thisLength < otherLength) {
+ return -1;
+ } else if (thisLength > otherLength) {
+ return 1;
+ } else {
+ return 0;
}
- return this.length() - otherTuple.length();
}
}
Modified: lucene/mahout/trunk/core/src/main/java/org/apache/mahout/common/TimingStatistics.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/core/src/main/java/org/apache/mahout/common/TimingStatistics.java?rev=826208&r1=826207&r2=826208&view=diff
==============================================================================
--- lucene/mahout/trunk/core/src/main/java/org/apache/mahout/common/TimingStatistics.java (original)
+++ lucene/mahout/trunk/core/src/main/java/org/apache/mahout/common/TimingStatistics.java Sat Oct 17 09:42:31 2009
@@ -19,95 +19,95 @@
import java.io.Serializable;
-public class TimingStatistics implements Serializable {
- private int nCalls;
- private long minTime;
- private long maxTime;
- private long sumTime;
- private double sumSquaredTime;
-
- /** Creates a new instance of CallStats */
- public TimingStatistics() {
- }
+public final class TimingStatistics implements Serializable {
- public TimingStatistics(int nCalls, long minTime, long maxTime, long sumTime,
- double sumSquaredTime) {
- this.nCalls = nCalls;
- this.minTime = minTime;
- this.maxTime = maxTime;
- this.sumTime = sumTime;
- this.sumSquaredTime = sumSquaredTime;
- }
-
- public int getNCalls() {
- return nCalls;
- }
-
- public long getMinTime() {
- return Math.max(0, minTime);
- }
-
- public long getMaxTime() {
- return maxTime;
- }
-
- public long getSumTime() {
- return sumTime;
- }
-
- public double getSumSquaredTime() {
- return sumSquaredTime;
- }
-
- public long getMeanTime() {
- if (nCalls == 0)
- return 0;
- else
- return sumTime / nCalls;
- }
-
- public long getStdDevTime() {
- if (nCalls == 0)
- return 0;
- double mean = getMeanTime();
- double meanSquared = mean * mean;
- double meanOfSquares = sumSquaredTime / nCalls;
- double variance = meanOfSquares - meanSquared;
- if (variance < 0)
- return 0; // might happen due to rounding error
- return (long) Math.sqrt(variance);
- }
-
- public String toString() {
- return "\n" +
- "nCalls = " + nCalls + ";\n" +
- "sumTime = " + getSumTime()/1000000000.0f + "s;\n" +
- "minTime = " + minTime/1000000.0f + "ms;\n" +
- "maxTime = " + maxTime/1000000.0f + "ms;\n" +
- "meanTime = " + getMeanTime()/1000000.0f + "ms;\n" +
- "stdDevTime = " + getStdDevTime()/1000000.0f + "ms;";
- }
-
- public Call newCall() {
- return new Call();
+ private int nCalls;
+ private long minTime;
+ private long maxTime;
+ private long sumTime;
+ private double sumSquaredTime;
+
+ /** Creates a new instance of CallStats */
+ public TimingStatistics() {
+ }
+
+ public TimingStatistics(int nCalls, long minTime, long maxTime, long sumTime,
+ double sumSquaredTime) {
+ this.nCalls = nCalls;
+ this.minTime = minTime;
+ this.maxTime = maxTime;
+ this.sumTime = sumTime;
+ this.sumSquaredTime = sumSquaredTime;
+ }
+
+ public int getNCalls() {
+ return nCalls;
+ }
+
+ public long getMinTime() {
+ return Math.max(0, minTime);
+ }
+
+ public long getMaxTime() {
+ return maxTime;
+ }
+
+ public long getSumTime() {
+ return sumTime;
+ }
+
+ public double getSumSquaredTime() {
+ return sumSquaredTime;
+ }
+
+ public long getMeanTime() {
+ return nCalls == 0 ? 0 : sumTime / nCalls;
+ }
+
+ public long getStdDevTime() {
+ if (nCalls == 0)
+ return 0;
+ double mean = getMeanTime();
+ double meanSquared = mean * mean;
+ double meanOfSquares = sumSquaredTime / nCalls;
+ double variance = meanOfSquares - meanSquared;
+ if (variance < 0)
+ return 0; // might happen due to rounding error
+ return (long) Math.sqrt(variance);
+ }
+
+ public String toString() {
+ return '\n' +
+ "nCalls = " + nCalls + ";\n" +
+ "sumTime = " + sumTime / 1000000000.0f + "s;\n" +
+ "minTime = " + minTime / 1000000.0f + "ms;\n" +
+ "maxTime = " + maxTime / 1000000.0f + "ms;\n" +
+ "meanTime = " + getMeanTime() / 1000000.0f + "ms;\n" +
+ "stdDevTime = " + getStdDevTime() / 1000000.0f + "ms;";
+ }
+
+ public Call newCall() {
+ return new Call();
+ }
+
+ public class Call {
+ private final long startTime = System.nanoTime();
+
+ private Call() {
}
-
- public class Call {
- private final long startTime = System.nanoTime();
- private Call() {}
-
- public void end() {
- long elapsed = System.nanoTime() - startTime;
- synchronized (TimingStatistics.this) {
- nCalls++;
- if (elapsed < minTime || nCalls == 1)
- minTime = elapsed;
- if (elapsed > maxTime)
- maxTime = elapsed;
- sumTime += elapsed;
- double elapsedFP = elapsed;
- sumSquaredTime += elapsedFP * elapsedFP;
- }
- }
+
+ public void end() {
+ long elapsed = System.nanoTime() - startTime;
+ synchronized (TimingStatistics.this) {
+ nCalls++;
+ if (elapsed < minTime || nCalls == 1)
+ minTime = elapsed;
+ if (elapsed > maxTime)
+ maxTime = elapsed;
+ sumTime += elapsed;
+ double elapsedFP = elapsed;
+ sumSquaredTime += elapsedFP * elapsedFP;
+ }
}
+ }
}
Modified: lucene/mahout/trunk/core/src/main/java/org/apache/mahout/common/commandline/DefaultOptionCreator.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/core/src/main/java/org/apache/mahout/common/commandline/DefaultOptionCreator.java?rev=826208&r1=826207&r2=826208&view=diff
==============================================================================
--- lucene/mahout/trunk/core/src/main/java/org/apache/mahout/common/commandline/DefaultOptionCreator.java (original)
+++ lucene/mahout/trunk/core/src/main/java/org/apache/mahout/common/commandline/DefaultOptionCreator.java Sat Oct 17 09:42:31 2009
@@ -112,9 +112,7 @@
* Returns a default command line option for help.
* */
public static Option helpOption(DefaultOptionBuilder obuilder) {
- Option helpOpt = obuilder.withLongName("help").
- withDescription("Print out help").withShortName("h").create();
- return helpOpt;
+ return obuilder.withLongName("help").withDescription("Print out help").withShortName("h").create();
}
}
Modified: lucene/mahout/trunk/core/src/main/java/org/apache/mahout/common/distance/SquaredEuclideanDistanceMeasure.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/core/src/main/java/org/apache/mahout/common/distance/SquaredEuclideanDistanceMeasure.java?rev=826208&r1=826207&r2=826208&view=diff
==============================================================================
--- lucene/mahout/trunk/core/src/main/java/org/apache/mahout/common/distance/SquaredEuclideanDistanceMeasure.java (original)
+++ lucene/mahout/trunk/core/src/main/java/org/apache/mahout/common/distance/SquaredEuclideanDistanceMeasure.java Sat Oct 17 09:42:31 2009
@@ -24,7 +24,6 @@
import java.util.Collection;
import java.util.Collections;
-import java.util.Iterator;
/**
* Like {@link EuclideanDistanceMeasure} but it does not take the square root. <p/> Thus, it is
Modified: lucene/mahout/trunk/core/src/main/java/org/apache/mahout/common/distance/TanimotoDistanceMeasure.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/core/src/main/java/org/apache/mahout/common/distance/TanimotoDistanceMeasure.java?rev=826208&r1=826207&r2=826208&view=diff
==============================================================================
--- lucene/mahout/trunk/core/src/main/java/org/apache/mahout/common/distance/TanimotoDistanceMeasure.java (original)
+++ lucene/mahout/trunk/core/src/main/java/org/apache/mahout/common/distance/TanimotoDistanceMeasure.java Sat Oct 17 09:42:31 2009
@@ -19,10 +19,7 @@
import org.apache.mahout.matrix.Vector;
-import java.util.HashSet;
import java.util.Iterator;
-import java.util.Set;
-
/**
* Tanimoto coefficient implementation.
@@ -48,7 +45,7 @@
double denominator = dot(a, a) + dot(b, b) - ab;
if(denominator < ab) { // correct for fp round-off: distance >= 0
denominator = ab;
- };
+ }
if(denominator > 0) {
// denom == 0 only when dot(a,a) == dot(b,b) == dot(a,b) == 0
return 1 - ab / denominator;
@@ -59,16 +56,15 @@
public double dot(Vector a, Vector b) {
Iterator<Vector.Element> it = a.iterateNonZero();
- Vector.Element el = null;
+ Vector.Element el;
Vector weights = getWeights();
double dot = 0;
- while(it.hasNext() && (el = it.next()) != null) {
- try {
- dot += el.get() * (a == b ? el.get() : b.getQuick(el.index())) * (weights == null ? 1.0 : weights.getQuick(el.index()));
- } catch (NullPointerException npe) {
- System.out.println(a.asFormatString() + "\n" + b.asFormatString() + "\n" + weights.asFormatString());
- throw npe;
+ while (it.hasNext() && (el = it.next()) != null) {
+ double value = el.get() * (a == b ? el.get() : b.getQuick(el.index()));
+ if (weights != null) {
+ value *= weights.getQuick(el.index());
}
+ dot += value;
}
return dot;
}
Modified: lucene/mahout/trunk/core/src/main/java/org/apache/mahout/df/mapred/partial/PartialBuilder.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/core/src/main/java/org/apache/mahout/df/mapred/partial/PartialBuilder.java?rev=826208&r1=826207&r2=826208&view=diff
==============================================================================
--- lucene/mahout/trunk/core/src/main/java/org/apache/mahout/df/mapred/partial/PartialBuilder.java (original)
+++ lucene/mahout/trunk/core/src/main/java/org/apache/mahout/df/mapred/partial/PartialBuilder.java Sat Oct 17 09:42:31 2009
@@ -149,6 +149,7 @@
protected static void processOutput(JobConf job, Path outputPath,
int[] firstIds, TreeID[] keys, Node[] trees, PredictionCallback callback)
throws IOException {
+ // TODO doesn't this assert throw NPE? looks like args are nullable
assert keys.length == trees.length : "keys.length != trees.length";
FileSystem fs = outputPath.getFileSystem(job);
Modified: lucene/mahout/trunk/core/src/main/java/org/apache/mahout/df/node/Leaf.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/core/src/main/java/org/apache/mahout/df/node/Leaf.java?rev=826208&r1=826207&r2=826208&view=diff
==============================================================================
--- lucene/mahout/trunk/core/src/main/java/org/apache/mahout/df/node/Leaf.java (original)
+++ lucene/mahout/trunk/core/src/main/java/org/apache/mahout/df/node/Leaf.java Sat Oct 17 09:42:31 2009
@@ -22,13 +22,12 @@
import java.io.IOException;
import java.util.StringTokenizer;
-import org.apache.hadoop.io.Writable;
import org.apache.mahout.df.data.Instance;
/**
* Represents a Leaf node
*/
-public class Leaf extends Node implements Writable {
+public class Leaf extends Node {
private int label;
protected Leaf() {
Modified: lucene/mahout/trunk/core/src/main/java/org/apache/mahout/df/tools/Describe.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/core/src/main/java/org/apache/mahout/df/tools/Describe.java?rev=826208&r1=826207&r2=826208&view=diff
==============================================================================
--- lucene/mahout/trunk/core/src/main/java/org/apache/mahout/df/tools/Describe.java (original)
+++ lucene/mahout/trunk/core/src/main/java/org/apache/mahout/df/tools/Describe.java Sat Oct 17 09:42:31 2009
@@ -86,7 +86,7 @@
if (cmdLine.hasOption(helpOpt)) {
CommandLineUtil.printHelp(group);
- System.exit(-1);
+ return;
}
String dataPath = cmdLine.getValue(pathOpt).toString();
@@ -101,7 +101,6 @@
} catch (OptionException e) {
System.err.println("Exception : " + e);
CommandLineUtil.printHelp(group);
- System.exit(-1);
}
}
Modified: lucene/mahout/trunk/core/src/main/java/org/apache/mahout/matrix/SparseColumnMatrix.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/core/src/main/java/org/apache/mahout/matrix/SparseColumnMatrix.java?rev=826208&r1=826207&r2=826208&view=diff
==============================================================================
--- lucene/mahout/trunk/core/src/main/java/org/apache/mahout/matrix/SparseColumnMatrix.java (original)
+++ lucene/mahout/trunk/core/src/main/java/org/apache/mahout/matrix/SparseColumnMatrix.java Sat Oct 17 09:42:31 2009
@@ -182,8 +182,7 @@
@Override
public void readFields(DataInput in) throws IOException {
super.readFields(in);
- int[] card = {in.readInt(), in.readInt()};
- this.cardinality = card;
+ this.cardinality = new int[]{in.readInt(), in.readInt()};
int colSize = in.readInt();
this.columns = new Vector[colSize];
for (int col = 0; col < colSize; col++) {
Modified: lucene/mahout/trunk/core/src/main/java/org/apache/mahout/matrix/SparseRowMatrix.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/core/src/main/java/org/apache/mahout/matrix/SparseRowMatrix.java?rev=826208&r1=826207&r2=826208&view=diff
==============================================================================
--- lucene/mahout/trunk/core/src/main/java/org/apache/mahout/matrix/SparseRowMatrix.java (original)
+++ lucene/mahout/trunk/core/src/main/java/org/apache/mahout/matrix/SparseRowMatrix.java Sat Oct 17 09:42:31 2009
@@ -178,8 +178,7 @@
@Override
public void readFields(DataInput in) throws IOException {
super.readFields(in);
- int[] card = {in.readInt(), in.readInt()};
- this.cardinality = card;
+ this.cardinality = new int[]{in.readInt(), in.readInt()};
int rowsize = in.readInt();
this.rows = new Vector[rowsize];
for (int row = 0; row < rowsize; row++) {
Modified: lucene/mahout/trunk/core/src/test/java/org/apache/mahout/cf/taste/impl/similarity/SpearmanCorrelationSimilarityTest.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/core/src/test/java/org/apache/mahout/cf/taste/impl/similarity/SpearmanCorrelationSimilarityTest.java?rev=826208&r1=826207&r2=826208&view=diff
==============================================================================
--- lucene/mahout/trunk/core/src/test/java/org/apache/mahout/cf/taste/impl/similarity/SpearmanCorrelationSimilarityTest.java (original)
+++ lucene/mahout/trunk/core/src/test/java/org/apache/mahout/cf/taste/impl/similarity/SpearmanCorrelationSimilarityTest.java Sat Oct 17 09:42:31 2009
@@ -17,7 +17,6 @@
package org.apache.mahout.cf.taste.impl.similarity;
-import org.apache.mahout.cf.taste.common.TasteException;
import org.apache.mahout.cf.taste.model.DataModel;
/** <p>Tests {@link SpearmanCorrelationSimilarity}.</p> */
Modified: lucene/mahout/trunk/core/src/test/java/org/apache/mahout/clustering/canopy/VisibleCanopy.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/core/src/test/java/org/apache/mahout/clustering/canopy/VisibleCanopy.java?rev=826208&r1=826207&r2=826208&view=diff
==============================================================================
--- lucene/mahout/trunk/core/src/test/java/org/apache/mahout/clustering/canopy/VisibleCanopy.java (original)
+++ lucene/mahout/trunk/core/src/test/java/org/apache/mahout/clustering/canopy/VisibleCanopy.java Sat Oct 17 09:42:31 2009
@@ -51,9 +51,7 @@
public String toString() {
String out = super.toString() + ": ";
for (Vector pt : points) {
- {
- out = pt.asFormatString();
- }
+ out = pt.asFormatString();
}
return out;
}
Modified: lucene/mahout/trunk/core/src/test/java/org/apache/mahout/clustering/lda/TestLDAInference.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/core/src/test/java/org/apache/mahout/clustering/lda/TestLDAInference.java?rev=826208&r1=826207&r2=826208&view=diff
==============================================================================
--- lucene/mahout/trunk/core/src/test/java/org/apache/mahout/clustering/lda/TestLDAInference.java (original)
+++ lucene/mahout/trunk/core/src/test/java/org/apache/mahout/clustering/lda/TestLDAInference.java Sat Oct 17 09:42:31 2009
@@ -64,7 +64,6 @@
double topicSmoothing = 50.0 / numTopics; // whatever
Matrix m = new DenseMatrix(numTopics, numWords);
double[] logTotals = new double[numTopics];
- double ll = Double.NEGATIVE_INFINITY; // TODO this is not updated in loop?
for (int k = 0; k < numTopics; ++k) {
double total = 0.0; // total number of pseudo counts we made
@@ -78,6 +77,7 @@
logTotals[k] = Math.log(total);
}
+ double ll = Double.NEGATIVE_INFINITY;
return new LDAState(numTopics, numWords, topicSmoothing, m, logTotals, ll);
}
Modified: lucene/mahout/trunk/core/src/test/java/org/apache/mahout/clustering/lda/TestMapReduce.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/core/src/test/java/org/apache/mahout/clustering/lda/TestMapReduce.java?rev=826208&r1=826207&r2=826208&view=diff
==============================================================================
--- lucene/mahout/trunk/core/src/test/java/org/apache/mahout/clustering/lda/TestMapReduce.java (original)
+++ lucene/mahout/trunk/core/src/test/java/org/apache/mahout/clustering/lda/TestMapReduce.java Sat Oct 17 09:42:31 2009
@@ -62,7 +62,6 @@
double topicSmoothing = 50.0 / numTopics; // whatever
Matrix m = new DenseMatrix(numTopics,numWords);
double[] logTotals = new double[numTopics];
- double ll = Double.NEGATIVE_INFINITY; // TODO this is not updated in loop?
for(int k = 0; k < numTopics; ++k) {
double total = 0.0; // total number of pseudo counts we made
for(int w = 0; w < numWords; ++w) {
@@ -75,6 +74,7 @@
logTotals[k] = Math.log(total);
}
+ double ll = Double.NEGATIVE_INFINITY;
return new LDAState(numTopics,numWords,topicSmoothing,m,logTotals,ll);
}
Modified: lucene/mahout/trunk/core/src/test/java/org/apache/mahout/common/distance/DefaultDistanceMeasureTest.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/core/src/test/java/org/apache/mahout/common/distance/DefaultDistanceMeasureTest.java?rev=826208&r1=826207&r2=826208&view=diff
==============================================================================
--- lucene/mahout/trunk/core/src/test/java/org/apache/mahout/common/distance/DefaultDistanceMeasureTest.java (original)
+++ lucene/mahout/trunk/core/src/test/java/org/apache/mahout/common/distance/DefaultDistanceMeasureTest.java Sat Oct 17 09:42:31 2009
@@ -61,7 +61,7 @@
compare(distanceMeasure, vectors);
}
- private void compare(DistanceMeasure distanceMeasure, Vector[] vectors) {
+ private static void compare(DistanceMeasure distanceMeasure, Vector[] vectors) {
double[][] distanceMatrix = new double[4][4];
for (int a = 0; a < 4; a++) {
@@ -87,7 +87,7 @@
assertTrue("Distance between vectors less than zero: "
+ distanceMatrix[a][b] + " = " + distanceMeasure.toString() +
".distance("+ vectors[a].asFormatString() + ", "
- + vectors[b].asFormatString() + ")",
+ + vectors[b].asFormatString() + ')',
distanceMatrix[a][b] >= 0);
if(vectors[a].plus(vectors[b]).norm(2) == 0 && vectors[a].norm(2) > 0) {
assertTrue("Distance from v to -v is equal to zero"
Modified: lucene/mahout/trunk/core/src/test/java/org/apache/mahout/df/data/DataTest.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/core/src/test/java/org/apache/mahout/df/data/DataTest.java?rev=826208&r1=826207&r2=826208&view=diff
==============================================================================
--- lucene/mahout/trunk/core/src/test/java/org/apache/mahout/df/data/DataTest.java (original)
+++ lucene/mahout/trunk/core/src/test/java/org/apache/mahout/df/data/DataTest.java Sat Oct 17 09:42:31 2009
@@ -278,6 +278,7 @@
sData = Utils.double2String(source);
dataset = DataLoader.generateDataset(descriptor, sData);
data = DataLoader.loadData(dataset, sData);
+ // TODO code1 is not used?
code1 = dataset.labelCode(Double.toString(label1));
int code2 = dataset.labelCode(Double.toString(label2));
Modified: lucene/mahout/trunk/core/src/test/java/org/apache/mahout/df/mapred/partial/PartialBuilderTest.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/core/src/test/java/org/apache/mahout/df/mapred/partial/PartialBuilderTest.java?rev=826208&r1=826207&r2=826208&view=diff
==============================================================================
--- lucene/mahout/trunk/core/src/test/java/org/apache/mahout/df/mapred/partial/PartialBuilderTest.java (original)
+++ lucene/mahout/trunk/core/src/test/java/org/apache/mahout/df/mapred/partial/PartialBuilderTest.java Sat Oct 17 09:42:31 2009
@@ -71,11 +71,8 @@
Writer writer = SequenceFile.createWriter(fs, job, outputFile,
TreeID.class, MapredOutput.class);
- try {
- for (int index = 0; index < numTrees; index++) {
- writer.append(keys[index], values[index]);
- }
- } finally {
+ for (int index = 0; index < numTrees; index++) {
+ writer.append(keys[index], values[index]);
}
writer.close();
Modified: lucene/mahout/trunk/core/src/test/java/org/apache/mahout/df/mapred/partial/PartialSequentialBuilder.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/core/src/test/java/org/apache/mahout/df/mapred/partial/PartialSequentialBuilder.java?rev=826208&r1=826207&r2=826208&view=diff
==============================================================================
--- lucene/mahout/trunk/core/src/test/java/org/apache/mahout/df/mapred/partial/PartialSequentialBuilder.java (original)
+++ lucene/mahout/trunk/core/src/test/java/org/apache/mahout/df/mapred/partial/PartialSequentialBuilder.java Sat Oct 17 09:42:31 2009
@@ -147,7 +147,7 @@
@Override
protected DecisionForest parseOutput(JobConf job, PredictionCallback callback)
throws IOException {
- int numMaps = job.getNumMapTasks();
+ //int numMaps = job.getNumMapTasks();
DecisionForest forest = processOutput(firstOutput.keys, firstOutput.values, callback);
Modified: lucene/mahout/trunk/core/src/test/java/org/apache/mahout/df/mapreduce/partial/Step0JobTest.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/core/src/test/java/org/apache/mahout/df/mapreduce/partial/Step0JobTest.java?rev=826208&r1=826207&r2=826208&view=diff
==============================================================================
--- lucene/mahout/trunk/core/src/test/java/org/apache/mahout/df/mapreduce/partial/Step0JobTest.java (original)
+++ lucene/mahout/trunk/core/src/test/java/org/apache/mahout/df/mapreduce/partial/Step0JobTest.java Sat Oct 17 09:42:31 2009
@@ -152,7 +152,7 @@
Job job = new Job();
job.setInputFormatClass(TextInputFormat.class);
- TextInputFormat.setInputPaths(job, dataPath);
+ FileInputFormat.setInputPaths(job, dataPath);
setMaxSplitSize(job.getConfiguration(), dataPath, numMaps);
Modified: lucene/mahout/trunk/core/src/test/java/org/apache/mahout/matrix/VectorTest.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/core/src/test/java/org/apache/mahout/matrix/VectorTest.java?rev=826208&r1=826207&r2=826208&view=diff
==============================================================================
--- lucene/mahout/trunk/core/src/test/java/org/apache/mahout/matrix/VectorTest.java (original)
+++ lucene/mahout/trunk/core/src/test/java/org/apache/mahout/matrix/VectorTest.java Sat Oct 17 09:42:31 2009
@@ -313,9 +313,7 @@
private static Vector randomSparseVector(Random rnd) {
SparseVector v1 = new SparseVector(50000);
for (int i = 0; i < 1000; i++) {
- {
- v1.setQuick(rnd.nextInt(50000), rnd.nextDouble());
- }
+ v1.setQuick(rnd.nextInt(50000), rnd.nextDouble());
}
return v1;
}
@@ -368,9 +366,7 @@
double[] values = {1.1, 2.2, 3.3};
Vector test = new SparseVector(3);
for (int i = 0; i < values.length; i++) {
- {
- test.set(i, values[i]);
- }
+ test.set(i, values[i]);
}
Map<String, Integer> bindings = new HashMap<String, Integer>();
bindings.put("Fee", 0);