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 2010/01/21 20:51:51 UTC
svn commit: r901839 [3/3] - in /lucene/mahout/trunk:
collections-codegen-plugin/src/main/java/org/apache/mahout/collection_codegen/
core/src/main/java/org/apache/mahout/cf/taste/hadoop/item/
core/src/main/java/org/apache/mahout/cf/taste/impl/eval/ core...
Modified: lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/jet/stat/Descriptive.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/jet/stat/Descriptive.java?rev=901839&r1=901838&r2=901839&view=diff
==============================================================================
--- lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/jet/stat/Descriptive.java (original)
+++ lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/jet/stat/Descriptive.java Thu Jan 21 19:51:44 2010
@@ -470,8 +470,7 @@
v += (delta1 * delta1 - v) / (i + 1);
}
- double r1 = q / v;
- return r1;
+ return q / v;
}
/** Returns the largest member of a data sequence. */
Modified: lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/jet/stat/quantile/DoubleBufferSet.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/jet/stat/quantile/DoubleBufferSet.java?rev=901839&r1=901838&r2=901839&view=diff
==============================================================================
--- lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/jet/stat/quantile/DoubleBufferSet.java (original)
+++ lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/jet/stat/quantile/DoubleBufferSet.java Thu Jan 21 19:51:44 2010
@@ -241,18 +241,6 @@
* @return outputValues a list filled with the values at triggerPositions
*/
protected double[] getValuesAtPositions(DoubleBuffer[] buffers, long[] triggerPositions) {
- //if (buffers.length==0)
- //{
- // throw new IllegalArgumentException("Oops! buffer.length==0.");
- //}
-
- //log.info("triggers="+cern.it.util.Arrays.toString(positions));
-
- //new DoubleArrayList(outputValues).fillFromToWith(0, outputValues.length-1, 0.0f);
- //delte the above line, it is only for testing
-
- //cern.it.util.Log.println("\nEntering getValuesAtPositions...");
- //cern.it.util.Log.println("hitPositions="+cern.it.util.Arrays.toString(positions));
// sort buffers.
for (int i = buffers.length; --i >= 0;) {
@@ -294,9 +282,6 @@
// fill all output values with equi-distant elements.
long counter = 0; //current position in sorted sequence
while (j < triggerPositionsLength) {
- //log.info("\nj="+j);
- //log.info("counter="+counter);
- //log.info("nextHit="+nextHit);
// determine buffer with smallest value at cursor position.
double minValue = Double.POSITIVE_INFINITY;
@@ -321,7 +306,6 @@
counter += minBuffer.weight();
while (counter > nextHit && j < triggerPositionsLength) {
outputValues[j++] = minValue;
- //log.info("adding to output="+minValue);
if (j < triggerPositionsLength) {
nextHit = triggerPositions[j];
}
@@ -330,7 +314,6 @@
// that element has now been treated, move further.
cursors[minBufferIndex]++;
- //log.info("cursors="+cern.it.util.Arrays.toString(cursors));
} //end while (j<k)
Modified: lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/jet/stat/quantile/DoubleQuantileEstimator.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/jet/stat/quantile/DoubleQuantileEstimator.java?rev=901839&r1=901838&r2=901839&view=diff
==============================================================================
--- lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/jet/stat/quantile/DoubleQuantileEstimator.java (original)
+++ lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/jet/stat/quantile/DoubleQuantileEstimator.java Thu Jan 21 19:51:44 2010
@@ -140,8 +140,7 @@
if (this.bufferSet != null) {
copy.bufferSet = (DoubleBufferSet) copy.bufferSet.clone();
if (this.currentBufferToFill != null) {
- int index;
- for (index = 0; index < bufferSet.buffers.length; index ++) {
+ for (int index = 0; index < bufferSet.buffers.length; index ++) {
if (bufferSet.buffers[index].equals(currentBufferToFill)) {
copy.currentBufferToFill = copy.bufferSet.buffers[index];
return copy;
Modified: lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/jet/stat/quantile/KnownDoubleQuantileEstimator.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/jet/stat/quantile/KnownDoubleQuantileEstimator.java?rev=901839&r1=901838&r2=901839&view=diff
==============================================================================
--- lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/jet/stat/quantile/KnownDoubleQuantileEstimator.java (original)
+++ lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/jet/stat/quantile/KnownDoubleQuantileEstimator.java Thu Jan 21 19:51:44 2010
@@ -215,7 +215,6 @@
throw new IllegalStateException("Oops! illegal missing values.");
}
- //log.info("adding "+missingValues+" infinity elements...");
this.addInfinities(missingValues, partial);
//determine beta (N + Infinity values = beta * N)
Modified: lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/jet/stat/quantile/QuantileCalc.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/jet/stat/quantile/QuantileCalc.java?rev=901839&r1=901838&r2=901839&view=diff
==============================================================================
--- lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/jet/stat/quantile/QuantileCalc.java (original)
+++ lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/jet/stat/quantile/QuantileCalc.java Thu Jan 21 19:51:44 2010
@@ -325,13 +325,13 @@
for (int p : quantiles) {
log.info("------------------------------");
- log.info("computing for p = " + p);
+ log.info("computing for p = {}", p);
for (long N : sizes) {
log.info(" ------------------------------");
- log.info(" computing for N = " + N);
+ log.info(" computing for N = {}", N);
for (double delta : deltas) {
log.info(" ------------------------------");
- log.info(" computing for delta = " + delta);
+ log.info(" computing for delta = {}", delta);
for (double epsilon : epsilons) {
double[] returnSamplingRate = new double[1];
long[] result;
@@ -343,14 +343,14 @@
long b = result[0];
long k = result[1];
- log.info(" (e,d,N,p)=(" + epsilon + ',' + delta + ',' + N + ',' + p + ") --> ");
+ log.info(" (e,d,N,p)=({},{},{},{}) --> ", new Object[] {epsilon, delta, N, p});
log.info("(b,k,mem");
if (known_N) {
log.info(",sampling");
}
- log.info(")=(" + b + ',' + k + ',' + (b * k / 1024));
+ log.info(")=({},{},{}", new Object[] {b, k, (b * k / 1024)});
if (known_N) {
- log.info("," + returnSamplingRate[0]);
+ log.info(",{}", returnSamplingRate[0]);
}
log.info(")");
}
Modified: lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/jet/stat/quantile/UnknownDoubleQuantileEstimator.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/jet/stat/quantile/UnknownDoubleQuantileEstimator.java?rev=901839&r1=901838&r2=901839&view=diff
==============================================================================
--- lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/jet/stat/quantile/UnknownDoubleQuantileEstimator.java (original)
+++ lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/jet/stat/quantile/UnknownDoubleQuantileEstimator.java Thu Jan 21 19:51:44 2010
@@ -188,8 +188,8 @@
new Comparator<DoubleBuffer>() {
@Override
public int compare(DoubleBuffer o1, DoubleBuffer o2) {
- int l1 = ((Buffer) o1).level();
- int l2 = ((Buffer) o2).level();
+ int l1 = o1.level();
+ int l2 = o2.level();
return l1 < l2 ? -1 : l1 == l2 ? 0 : 1;
}
}
Modified: lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/list/AbstractList.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/list/AbstractList.java?rev=901839&r1=901838&r2=901839&view=diff
==============================================================================
--- lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/list/AbstractList.java (original)
+++ lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/list/AbstractList.java Thu Jan 21 19:51:44 2010
@@ -44,7 +44,7 @@
@author wolfgang.hoschek@cern.ch
@version 1.0, 09/24/99
- @see ArrayList
+ @see java.util.ArrayList
@see java.util.Vector
@see java.util.Arrays
*/
@@ -53,7 +53,7 @@
public abstract int size();
public boolean isEmpty() {
- return 0 == size();
+ return size() == 0;
}
/**
Modified: lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/list/AbstractObjectList.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/list/AbstractObjectList.java?rev=901839&r1=901838&r2=901839&view=diff
==============================================================================
--- lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/list/AbstractObjectList.java (original)
+++ lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/list/AbstractObjectList.java Thu Jan 21 19:51:44 2010
@@ -66,18 +66,6 @@
this.replaceFromWith(index, collection);
}
-
- /**
- * Removes from the receiver all elements whose index is between <code>from</code>, inclusive and <code>to</code>,
- * inclusive. Shifts any succeeding elements to the left (reduces their index). This call shortens the list by
- * <tt>(to - from + 1)</tt> elements.
- *
- * @param fromIndex index of first element to be removed.
- * @param toIndex index of last element to be removed.
- * @throws IndexOutOfBoundsException if <tt>(from<0 || from>to || to>=size()) && to!=from-1</tt>.
- */
- public abstract void removeFromTo(int fromIndex, int toIndex);
-
/**
* Replaces the part of the receiver starting at <code>from</code> (inclusive) with all the elements of the specified
* collection. Does not alter the size of the receiver. Replaces exactly <tt>Math.max(0,Math.min(size()-from,
Modified: lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/list/ObjectArrayList.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/list/ObjectArrayList.java?rev=901839&r1=901838&r2=901839&view=diff
==============================================================================
--- lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/list/ObjectArrayList.java (original)
+++ lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/list/ObjectArrayList.java Thu Jan 21 19:51:44 2010
@@ -120,8 +120,7 @@
@Override
public Object clone() {
// overridden for performance only.
- ObjectArrayList<T> clone = new ObjectArrayList<T>((T[]) elements.clone());
- return clone;
+ return new ObjectArrayList<T>((T[]) elements.clone());
}
/**
@@ -193,7 +192,7 @@
if (otherObj == null) {
return false;
}
- ObjectArrayList other = (ObjectArrayList) otherObj;
+ ObjectArrayList<?> other = (ObjectArrayList<?>) otherObj;
if (size() != other.size()) {
return false;
}
Modified: lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/map/OpenHashMap.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/map/OpenHashMap.java?rev=901839&r1=901838&r2=901839&view=diff
==============================================================================
--- lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/map/OpenHashMap.java (original)
+++ lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/map/OpenHashMap.java Thu Jan 21 19:51:44 2010
@@ -111,7 +111,7 @@
@Override
@SuppressWarnings("unchecked")
public Object clone() {
- OpenHashMap copy = (OpenHashMap) super.clone();
+ OpenHashMap<K,V> copy = (OpenHashMap<K,V>) super.clone();
copy.table = copy.table.clone();
copy.values = copy.values.clone();
copy.state = copy.state.clone();
@@ -362,20 +362,15 @@
@Override
public V put(K key, V value) {
int i = indexOfInsertion(key);
- V previous = null;
if (i < 0) { //already contained
i = -i - 1;
- previous = (V) this.values[i];
+ V previous = (V) this.values[i];
this.values[i] = value;
return previous;
}
if (this.distinct > this.highWaterMark) {
int newCapacity = chooseGrowCapacity(this.distinct + 1, this.minLoadFactor, this.maxLoadFactor);
- /*
- log.info("grow rehashing ");
- log.info("at distinct="+distinct+", capacity="+table.length+" to newCapacity="+newCapacity+" ...");
- */
rehash(newCapacity);
return put(key, value);
}
@@ -442,13 +437,12 @@
@SuppressWarnings("unchecked")
@Override
public V remove(Object key) {
- V removed = null;
int i = indexOfKey((K)key);
if (i < 0) {
return null;
- } else {// key not contained
- removed = (V)values[i];
}
+ // key not contained
+ V removed = (V) values[i];
this.state[i] = REMOVED;
//this.values[i]=0; // delta
@@ -456,12 +450,6 @@
if (this.distinct < this.lowWaterMark) {
int newCapacity = chooseShrinkCapacity(this.distinct, this.minLoadFactor, this.maxLoadFactor);
- /*
- if (table.length != newCapacity) {
- log.info("shrink rehashing ");
- log.info("at distinct="+distinct+", capacity="+table.length+" to newCapacity="+newCapacity+" ...");
- }
- */
rehash(newCapacity);
}
@@ -539,8 +527,8 @@
}
private class MapEntry implements Map.Entry<K,V> {
- private K key;
- private V value;
+ private final K key;
+ private final V value;
MapEntry(K key, V value) {
this.key = key;
@@ -569,7 +557,7 @@
*/
@Override
public Set<java.util.Map.Entry<K,V>> entrySet() {
- final OpenHashSet<Map.Entry<K,V>> entries = new OpenHashSet<Map.Entry<K,V>>();
+ final Set<Entry<K, V>> entries = new OpenHashSet<Map.Entry<K,V>>();
forEachPair(new ObjectObjectProcedure<K,V>() {
@Override
@@ -586,7 +574,7 @@
*/
@Override
public Set<K> keySet() {
- final OpenHashSet<K> keys = new OpenHashSet<K>();
+ final Set<K> keys = new OpenHashSet<K>();
forEachKey(new ObjectProcedure<K>() {
@Override
@@ -627,7 +615,7 @@
if (! (obj instanceof OpenHashMap)) {
return false;
}
- final OpenHashMap o = (OpenHashMap) obj;
+ final OpenHashMap<K,V> o = (OpenHashMap<K,V>) obj;
if (o.size() != size()) {
return false;
}
@@ -650,19 +638,19 @@
@Override
public String toString() {
final StringBuilder sb = new StringBuilder();
- sb.append("{");
+ sb.append('{');
forEachPair(new ObjectObjectProcedure<K,V>() {
@Override
public boolean apply(K key, V value) {
- sb.append("[");
+ sb.append('[');
sb.append(key);
sb.append(" -> ");
sb.append(value);
sb.append("] ");
return true;
}});
- sb.append("}");
+ sb.append('}');
return sb.toString();
}
}
Modified: lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/map/PrimeFinder.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/map/PrimeFinder.java?rev=901839&r1=901838&r2=901839&view=diff
==============================================================================
--- lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/map/PrimeFinder.java (original)
+++ lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/map/PrimeFinder.java Thu Jan 21 19:51:44 2010
@@ -112,14 +112,6 @@
private PrimeFinder() {
}
- /** Tests correctness. Try from=1000, to=10000 from=200, to=1000 from=16, to=1000 from=1000, to=Integer.MAX_VALUE */
- protected static void main(String[] args) {
- int from = Integer.parseInt(args[0]);
- int to = Integer.parseInt(args[1]);
-
- statistics(from, to);
- }
-
/**
* Returns a prime number which is <code>>= desiredCapacity</code> and very close to <code>desiredCapacity</code>
* (within 11% if <code>desiredCapacity >= 1000</code>).
@@ -156,7 +148,7 @@
if (deviation > maxDeviation) {
maxDeviation = deviation;
- log.info("new maxdev @" + i + "@dev=" + maxDeviation);
+ log.info("new maxdev @{}@dev={}", i, maxDeviation);
}
accDeviation += deviation;
@@ -164,8 +156,8 @@
long width = 1 + (long) to - (long) from;
double meanDeviation = accDeviation / width;
- log.info("Statistics for [" + from + ',' + to + "] are as follows");
- log.info("meanDeviation = " + (float) meanDeviation * 100 + " %");
- log.info("maxDeviation = " + (float) maxDeviation * 100 + " %");
+ log.info("Statistics for [{},{}] are as follows", from, to);
+ log.info("meanDeviation = {} %", (float) meanDeviation * 100);
+ log.info("maxDeviation = {} %", (float) maxDeviation * 100);
}
}
Modified: lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/map/QuickOpenIntIntHashMap.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/map/QuickOpenIntIntHashMap.java?rev=901839&r1=901838&r2=901839&view=diff
==============================================================================
--- lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/map/QuickOpenIntIntHashMap.java (original)
+++ lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/map/QuickOpenIntIntHashMap.java Thu Jan 21 19:51:44 2010
@@ -88,7 +88,6 @@
if (decrement == 0) {
decrement = 1;
}
- //log.info("insert search for (key,value)=("+key+","+value+") at i="+i+", dec="+decrement);
// stop if we find a removed or free slot, or if we find the key itself
// do NOT skip over removed slots (yes, open addressing is like that...)
@@ -103,7 +102,6 @@
i += length;
}
}
- //if (comparisons-comp>0) log.info("probed "+(comparisons-comp)+" slots.");
if (stat[i] == FULL) {
// key already contained at slot i.
this.values[i] = value;
@@ -113,10 +111,6 @@
if (this.distinct > this.highWaterMark) {
int newCapacity = chooseGrowCapacity(this.distinct + 1, this.minLoadFactor, this.maxLoadFactor);
-
- //log.info("grow rehashing ");
- //log.info("at distinct="+distinct+", capacity="+table.length+" to newCapacity="+newCapacity+" ...");
-
rehash(newCapacity);
return put(key, value);
}
@@ -135,7 +129,6 @@
Note: Under the great majority of insertions t<=1, so the loop is entered very infrequently.
*/
while (t > 1) {
- //log.info("t="+t);
int key0 = tab[p0];
hash = HashFunctions.hash(key0) & 0x7FFFFFFF;
decrement = (hash / length) % length;
@@ -151,8 +144,6 @@
p0 = pc;
t--;
} else { // free or removed slot found, now move...
- //log.info("copying p0="+p0+" to pc="+pc+", (key,val)=("+tab[p0]+","+values[p0]+"), saving "+(t-1)+" probes.");
- //this.totalProbesSaved += (t - 1);
tab[pc] = key0;
stat[pc] = FULL;
values[pc] = values[p0];
@@ -161,7 +152,6 @@
}
}
- //log.info("inserting at i="+i);
this.table[i] = key;
this.values[i] = value;
if (this.state[i] == FREE) {
Modified: lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/matrix/DoubleFactory2D.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/matrix/DoubleFactory2D.java?rev=901839&r1=901838&r2=901839&view=diff
==============================================================================
--- lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/matrix/DoubleFactory2D.java (original)
+++ lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/matrix/DoubleFactory2D.java Thu Jan 21 19:51:44 2010
@@ -249,7 +249,6 @@
* { null, identity(3).viewColumnFlip() },
* { identity(3).viewRowFlip(), null }
* };
- * log.info("\n"+make(parts3));
* </pre>
* </td> <td><tt>9 x 6 matrix<br> 1 0 0 0 0 0<br>
* 0 1 0 0 0 0<br> 0 0 1 0 0 0<br>
@@ -267,7 +266,6 @@
* { A, _, A, _ },
* { _, A, _, B }
* };
- * log.info("\n"+make(parts4));
* </pre>
* </td> <td><tt>4 x 8 matrix<br> 1 2 0 0 1 2 0 0<br>
* 3 4 0 0 3 4 0 0<br> 0 0 1 2 0 0 3 2<br>
@@ -279,7 +277,6 @@
* { make(4,4,2), null, make(4,3,3) },
* { null, make(2,3,4), null }
* };
- * log.info("\n"+Factory2D.make(parts2));
* </pre>
* </td> <td><tt>IllegalArgumentException<br> A[0,1].cols != A[2,1].cols<br> (2 != 3)</tt></td> </tr> </table>
*
@@ -423,10 +420,6 @@
* { _, D, _ }
* };
* decompose(parts,matrix);
- * log.info("\nA = "+A);
- * log.info("\nB = "+B);
- * log.info("\nC = "+C);
- * log.info("\nD = "+D);
* </pre>
* </td> <td><tt>8 x 9 matrix<br> 9 9 9 9 1 1 9 9 9<br>
* 9 9 9 9 1 1 9 9 9<br> 2 2 2 2 9 9 3 3 3<br>
Modified: lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/matrix/doublealgo/Formatter.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/matrix/doublealgo/Formatter.java?rev=901839&r1=901838&r2=901839&view=diff
==============================================================================
--- lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/matrix/doublealgo/Formatter.java (original)
+++ lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/matrix/doublealgo/Formatter.java Thu Jan 21 19:51:44 2010
@@ -290,7 +290,7 @@
}
/** Converts a given cell to a String; no alignment considered. */
- protected String form(DoubleMatrix1D matrix, int index, Former formatter) {
+ protected static String form(DoubleMatrix1D matrix, int index, Former formatter) {
return formatter.form(matrix.get(index));
}
@@ -316,7 +316,7 @@
}
/** Returns the index of the decimal point. */
- protected int indexOfDecimalPoint(String s) {
+ protected static int indexOfDecimalPoint(String s) {
int i = s.lastIndexOf('.');
if (i < 0) {
i = s.lastIndexOf('e');
Modified: lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/matrix/impl/AbstractFormatter.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/matrix/impl/AbstractFormatter.java?rev=901839&r1=901838&r2=901839&view=diff
==============================================================================
--- lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/matrix/impl/AbstractFormatter.java (original)
+++ lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/matrix/impl/AbstractFormatter.java Thu Jan 21 19:51:44 2010
@@ -320,12 +320,11 @@
String str = buf.toString();
for (int i = size; --i >= 0;) {
blanksCache[i] = str.substring(0, i);
- //log.info(i+"-"+blanksCache[i]+"-");
}
}
/** Returns a short string representation describing the shape of the matrix. */
- public static String shape(AbstractMatrix1D matrix) {
+ public static String shape(AbstractMatrix matrix) {
//return "Matrix1D of size="+matrix.size();
//return matrix.size()+" element matrix";
//return "matrix("+matrix.size()+")";
Modified: lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/matrix/impl/DelegateDoubleMatrix1D.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/matrix/impl/DelegateDoubleMatrix1D.java?rev=901839&r1=901838&r2=901839&view=diff
==============================================================================
--- lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/matrix/impl/DelegateDoubleMatrix1D.java (original)
+++ lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/matrix/impl/DelegateDoubleMatrix1D.java Thu Jan 21 19:51:44 2010
@@ -22,11 +22,11 @@
/*
* The elements of the matrix.
*/
- private DoubleMatrix2D content;
+ private final DoubleMatrix2D content;
/*
* The row this view is bound to.
*/
- private int row;
+ private final int row;
DelegateDoubleMatrix1D(DoubleMatrix2D newContent, int row) {
super(null);
Modified: lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/matrix/impl/DenseDoubleMatrix2D.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/matrix/impl/DenseDoubleMatrix2D.java?rev=901839&r1=901838&r2=901839&view=diff
==============================================================================
--- lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/matrix/impl/DenseDoubleMatrix2D.java (original)
+++ lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/matrix/impl/DenseDoubleMatrix2D.java Thu Jan 21 19:51:44 2010
@@ -23,7 +23,7 @@
* columnOf(index)==index%columns rowOf(index)==index/columns i.e. {row0 column0..m}, {row1 column0..m}, ..., {rown
* column0..m}
*/
- protected double[] elements;
+ protected final double[] elements;
/**
* Constructs a matrix with a copy of the given values. <tt>values</tt> is required to have the form
Modified: lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/matrix/impl/DenseDoubleMatrix3D.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/matrix/impl/DenseDoubleMatrix3D.java?rev=901839&r1=901838&r2=901839&view=diff
==============================================================================
--- lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/matrix/impl/DenseDoubleMatrix3D.java (original)
+++ lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/matrix/impl/DenseDoubleMatrix3D.java Thu Jan 21 19:51:44 2010
@@ -21,7 +21,7 @@
* row0..m}, ..., {sliceN row0..m} with each row storead as {row0 column0..m}, {row1 column0..m}, ..., {rown
* column0..m}
*/
- protected double[] elements;
+ protected final double[] elements;
/**
* Constructs a matrix with a copy of the given values. <tt>values</tt> is required to have the form
Modified: lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/matrix/impl/RCDoubleMatrix2D.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/matrix/impl/RCDoubleMatrix2D.java?rev=901839&r1=901838&r2=901839&view=diff
==============================================================================
--- lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/matrix/impl/RCDoubleMatrix2D.java (original)
+++ lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/matrix/impl/RCDoubleMatrix2D.java Thu Jan 21 19:51:44 2010
@@ -410,19 +410,6 @@
throw new InternalError();
}
- /*
- forEachNonZero(
- new IntIntDoubleFunction() {
- public double apply(int i, int j, double value) {
- zElements[zi + zStride*i] += value * yElements[yi + yStride*j];
- //z.setQuick(row,z.getQuick(row) + value * y.getQuick(column));
- //log.info("["+i+","+j+"]-->"+value);
- return value;
- }
- }
- );
- */
-
int[] idx = indexes.elements();
double[] vals = values.elements();
Modified: lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/matrix/impl/SparseDoubleMatrix2D.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/matrix/impl/SparseDoubleMatrix2D.java?rev=901839&r1=901838&r2=901839&view=diff
==============================================================================
--- lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/matrix/impl/SparseDoubleMatrix2D.java (original)
+++ lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/matrix/impl/SparseDoubleMatrix2D.java Thu Jan 21 19:51:44 2010
@@ -484,27 +484,12 @@
j = tmp;
}
zElements[zi + zStride * i] += value * yElements[yi + yStride * j];
- //log.info("["+i+","+j+"]-->"+value);
return true;
}
}
);
- /*
- forEachNonZero(
- new IntIntDoubleFunction() {
- public double apply(int i, int j, double value) {
- if (transposeA) { int tmp=i; i=j; j=tmp; }
- zElements[zi + zStride*i] += value * yElements[yi + yStride*j];
- //z.setQuick(row,z.getQuick(row) + value * y.getQuick(column));
- //log.info("["+i+","+j+"]-->"+value);
- return value;
- }
- }
- );
- */
-
- if (alpha != 1) {
+ if (alpha != 1.0) {
z.assign(Functions.mult(alpha));
}
return z;
Modified: lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/matrix/impl/SparseDoubleMatrix3D.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/matrix/impl/SparseDoubleMatrix3D.java?rev=901839&r1=901838&r2=901839&view=diff
==============================================================================
--- lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/matrix/impl/SparseDoubleMatrix3D.java (original)
+++ lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/matrix/impl/SparseDoubleMatrix3D.java Thu Jan 21 19:51:44 2010
@@ -19,7 +19,7 @@
/*
* The elements of the matrix.
*/
- protected AbstractIntDoubleMap elements;
+ protected final AbstractIntDoubleMap elements;
/**
* Constructs a matrix with a copy of the given values. <tt>values</tt> is required to have the form
Modified: lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/matrix/impl/TridiagonalDoubleMatrix2D.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/matrix/impl/TridiagonalDoubleMatrix2D.java?rev=901839&r1=901838&r2=901839&view=diff
==============================================================================
--- lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/matrix/impl/TridiagonalDoubleMatrix2D.java (original)
+++ lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/matrix/impl/TridiagonalDoubleMatrix2D.java Thu Jan 21 19:51:44 2010
@@ -85,9 +85,7 @@
}
values = new double[l + d + u]; // {lower, diagonal, upper}
- int[] dimensions = {0, l, l + d, l + d + u, 0, 0,
- 0}; // {lowerStart, diagonalStart, upperStart, values.length, lowerNonZeros, diagonalNonZeros, upperNonZeros}
- dims = dimensions;
+ dims = new int[]{0, l, l + d, l + d + u, 0, 0, 0};
}
/**
@@ -452,14 +450,12 @@
j = tmp;
}
zElements[zi + zStride * i] += value * yElements[yi + yStride * j];
- //z.setQuick(row,z.getQuick(row) + value * y.getQuick(column));
- //log.info("["+i+","+j+"]-->"+value);
return value;
}
}
);
- if (alpha != 1) {
+ if (alpha != 1.0) {
z.assign(Functions.mult(alpha));
}
return z;
Modified: lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/matrix/impl/WrapperDoubleMatrix1D.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/matrix/impl/WrapperDoubleMatrix1D.java?rev=901839&r1=901838&r2=901839&view=diff
==============================================================================
--- lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/matrix/impl/WrapperDoubleMatrix1D.java (original)
+++ lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/matrix/impl/WrapperDoubleMatrix1D.java Thu Jan 21 19:51:44 2010
@@ -109,7 +109,7 @@
*/
@Override
public DoubleMatrix1D viewFlip() {
- DoubleMatrix1D view = new WrapperDoubleMatrix1D(this) {
+ return new WrapperDoubleMatrix1D(WrapperDoubleMatrix1D.this) {
@Override
public double getQuick(int index) {
return content.get(size - 1 - index);
@@ -120,7 +120,6 @@
content.set(size - 1 - index, value);
}
};
- return view;
}
/**
Modified: lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/matrix/impl/WrapperDoubleMatrix2D.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/matrix/impl/WrapperDoubleMatrix2D.java?rev=901839&r1=901838&r2=901839&view=diff
==============================================================================
--- lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/matrix/impl/WrapperDoubleMatrix2D.java (original)
+++ lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/matrix/impl/WrapperDoubleMatrix2D.java Thu Jan 21 19:51:44 2010
@@ -160,7 +160,7 @@
if (columns == 0) {
return this;
}
- DoubleMatrix2D view = new WrapperDoubleMatrix2D(this) {
+ return new WrapperDoubleMatrix2D(WrapperDoubleMatrix2D.this) {
@Override
public double getQuick(int row, int column) {
return content.get(row, columns - 1 - column);
@@ -171,7 +171,6 @@
content.set(row, columns - 1 - column, value);
}
};
- return view;
}
/**
Modified: lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/matrix/linalg/Algebra.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/matrix/linalg/Algebra.java?rev=901839&r1=901838&r2=901839&view=diff
==============================================================================
--- lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/matrix/linalg/Algebra.java (original)
+++ lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/matrix/linalg/Algebra.java Thu Jan 21 19:51:44 2010
@@ -88,7 +88,7 @@
}
/** Returns the condition of matrix <tt>A</tt>, which is the ratio of largest to smallest singular value. */
- public double cond(DoubleMatrix2D A) {
+ public static double cond(DoubleMatrix2D A) {
return svd(A).cond();
}
@@ -97,7 +97,7 @@
*
* @return the determinant.
*/
- public double det(DoubleMatrix2D A) {
+ public static double det(DoubleMatrix2D A) {
return lu(A).det();
}
@@ -162,7 +162,7 @@
* @return the inner product.
* @throws IllegalArgumentException if <tt>x.size() != y.size()</tt>.
*/
- public double mult(DoubleMatrix1D x, DoubleMatrix1D y) {
+ public static double mult(DoubleMatrix1D x, DoubleMatrix1D y) {
return x.zDotProduct(y);
}
@@ -175,7 +175,7 @@
* @return <tt>z</tt>; a new vector with <tt>z.size()==A.rows()</tt>.
* @throws IllegalArgumentException if <tt>A.columns() != y.size()</tt>.
*/
- public DoubleMatrix1D mult(DoubleMatrix2D A, DoubleMatrix1D y) {
+ public static DoubleMatrix1D mult(DoubleMatrix2D A, DoubleMatrix1D y) {
return A.zMult(y, null);
}
@@ -188,7 +188,7 @@
* @return <tt>C</tt>; a new matrix holding the results, with <tt>C.rows()=A.rows(), C.columns()==B.columns()</tt>.
* @throws IllegalArgumentException if <tt>B.rows() != A.columns()</tt>.
*/
- public DoubleMatrix2D mult(DoubleMatrix2D A, DoubleMatrix2D B) {
+ public static DoubleMatrix2D mult(DoubleMatrix2D A, DoubleMatrix2D B) {
return A.zMult(B, null);
}
@@ -202,7 +202,7 @@
* @return A (for convenience only).
* @throws IllegalArgumentException if <tt>A.rows() != x.size() || A.columns() != y.size()</tt>.
*/
- public DoubleMatrix2D multOuter(DoubleMatrix1D x, DoubleMatrix1D y, DoubleMatrix2D A) {
+ public static DoubleMatrix2D multOuter(DoubleMatrix1D x, DoubleMatrix1D y, DoubleMatrix2D A) {
int rows = x.size();
int columns = y.size();
if (A == null) {
@@ -223,7 +223,7 @@
}
/** Returns the one-norm of vector <tt>x</tt>, which is <tt>Sum(abs(x[i]))</tt>. */
- public double norm1(DoubleMatrix1D x) {
+ public static double norm1(DoubleMatrix1D x) {
if (x.size() == 0) {
return 0;
}
@@ -245,12 +245,12 @@
}
/** Returns the two-norm of matrix <tt>A</tt>, which is the maximum singular value; obtained from SVD. */
- public double norm2(DoubleMatrix2D A) {
+ public static double norm2(DoubleMatrix2D A) {
return svd(A).norm2();
}
/** Returns the Frobenius norm of matrix <tt>A</tt>, which is <tt>Sqrt(Sum(A[i,j]<sup>2</sup>))</tt>. */
- public double normF(DoubleMatrix2D A) {
+ public static double normF(DoubleMatrix2D A) {
if (A.size() == 0) {
return 0;
}
@@ -258,7 +258,7 @@
}
/** Returns the infinity norm of vector <tt>x</tt>, which is <tt>Max(abs(x[i]))</tt>. */
- public double normInfinity(DoubleMatrix1D x) {
+ public static double normInfinity(DoubleMatrix1D x) {
// fix for bug reported by T.J.Hunt@open.ac.uk
if (x.size() == 0) {
return 0;
@@ -306,7 +306,7 @@
* @return the modified <tt>A</tt> (for convenience only).
* @throws IndexOutOfBoundsException if <tt>indexes.length != A.size()</tt>.
*/
- public DoubleMatrix1D permute(DoubleMatrix1D A, int[] indexes, double[] work) {
+ public static DoubleMatrix1D permute(DoubleMatrix1D A, int[] indexes, double[] work) {
// check validity
int size = A.size();
if (indexes.length != size) {
@@ -339,7 +339,7 @@
*
* @return the new permuted selection view.
*/
- public DoubleMatrix2D permute(DoubleMatrix2D A, int[] rowIndexes, int[] columnIndexes) {
+ public static DoubleMatrix2D permute(DoubleMatrix2D A, int[] rowIndexes, int[] columnIndexes) {
return A.viewSelection(rowIndexes, columnIndexes);
}
@@ -510,7 +510,7 @@
}
/** Returns the effective numerical rank of matrix <tt>A</tt>, obtained from Singular Value Decomposition. */
- public int rank(DoubleMatrix2D A) {
+ public static int rank(DoubleMatrix2D A) {
return svd(A).rank();
}
@@ -539,7 +539,7 @@
*
* @return X; a new independent matrix; solution if A is square, least squares solution otherwise.
*/
- public DoubleMatrix2D solve(DoubleMatrix2D A, DoubleMatrix2D B) {
+ public static DoubleMatrix2D solve(DoubleMatrix2D A, DoubleMatrix2D B) {
return (A.rows() == A.columns() ? (lu(A).solve(B)) : (qr(A).solve(B)));
}
@@ -627,7 +627,7 @@
* @throws IndexOutOfBoundsException if <tt>fromColumn<0 || toColumn-fromColumn+1<0 || toColumn>=A.columns() ||
* fromRow<0 || toRow-fromRow+1<0 || toRow>=A.rows()</tt>
*/
- public DoubleMatrix2D subMatrix(DoubleMatrix2D A, int fromRow, int toRow, int fromColumn, int toColumn) {
+ public static DoubleMatrix2D subMatrix(DoubleMatrix2D A, int fromRow, int toRow, int fromColumn, int toColumn) {
return A.viewPart(fromRow, fromColumn, toRow - fromRow + 1, toColumn - fromColumn + 1);
}
@@ -651,8 +651,8 @@
* </pre>
*/
public String toString(DoubleMatrix2D matrix) {
- final org.apache.mahout.math.list.ObjectArrayList names = new ObjectArrayList();
- final org.apache.mahout.math.list.ObjectArrayList values = new ObjectArrayList();
+ final ObjectArrayList<String> names = new ObjectArrayList<String>();
+ final ObjectArrayList<String> values = new ObjectArrayList<String>();
// determine properties
names.add("cond");
@@ -731,7 +731,7 @@
Swapper swapper = new Swapper() {
@Override
public void swap(int a, int b) {
- Object tmp = names.get(a);
+ String tmp = names.get(a);
names.set(a, names.get(b));
names.set(b, tmp);
tmp = values.get(a);
@@ -744,7 +744,7 @@
// determine padding for nice formatting
int maxLength = 0;
for (int i = 0; i < names.size(); i++) {
- int length = ((String) names.get(i)).length();
+ int length = (names.get(i)).length();
maxLength = Math.max(length, maxLength);
}
@@ -988,7 +988,7 @@
}
/** Returns the sum of the diagonal elements of matrix <tt>A</tt>; <tt>Sum(A[i,i])</tt>. */
- public double trace(DoubleMatrix2D A) {
+ public static double trace(DoubleMatrix2D A) {
double sum = 0;
for (int i = Math.min(A.rows(), A.columns()); --i >= 0;) {
sum += A.getQuick(i, i);
@@ -1007,7 +1007,7 @@
*
* @return a new transposed view.
*/
- public DoubleMatrix2D transpose(DoubleMatrix2D A) {
+ public static DoubleMatrix2D transpose(DoubleMatrix2D A) {
return A.viewDice();
}
@@ -1016,7 +1016,7 @@
*
* @return <tt>A</tt> (for convenience only).
*/
- protected DoubleMatrix2D trapezoidalLower(DoubleMatrix2D A) {
+ protected static DoubleMatrix2D trapezoidalLower(DoubleMatrix2D A) {
int rows = A.rows();
int columns = A.columns();
for (int r = rows; --r >= 0;) {
Modified: lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/matrix/linalg/Property.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/matrix/linalg/Property.java?rev=901839&r1=901838&r2=901839&view=diff
==============================================================================
--- lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/matrix/linalg/Property.java (original)
+++ lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/matrix/linalg/Property.java Thu Jan 21 19:51:44 2010
@@ -19,6 +19,7 @@
import org.apache.mahout.math.matrix.DoubleMatrix2D;
import org.apache.mahout.math.matrix.DoubleMatrix3D;
import org.apache.mahout.math.matrix.impl.AbstractFormatter;
+import org.apache.mahout.math.matrix.impl.AbstractMatrix2D;
/** @deprecated until unit tests are in place. Until this time, this class/interface is unsupported. */
@Deprecated
@@ -62,7 +63,7 @@
*
* @throws IllegalArgumentException if <tt>A.rows() < A.columns()</tt>.
*/
- public void checkRectangular(DoubleMatrix2D A) {
+ public void checkRectangular(AbstractMatrix2D A) {
if (A.rows() < A.columns()) {
throw new IllegalArgumentException("Matrix must be rectangular: " + AbstractFormatter.shape(A));
}
@@ -73,7 +74,7 @@
*
* @throws IllegalArgumentException if <tt>A.rows() != A.columns()</tt>.
*/
- public void checkSquare(DoubleMatrix2D A) {
+ public void checkSquare(AbstractMatrix2D A) {
if (A.rows() != A.columns()) {
throw new IllegalArgumentException("Matrix must be square: " + AbstractFormatter.shape(A));
}
@@ -327,10 +328,8 @@
}
}
- /**
- */
- protected static String get(ObjectArrayList list, int index) {
- return ((String) list.get(index));
+ protected static String get(ObjectArrayList<String> list, int index) {
+ return (list.get(index));
}
/**
@@ -523,7 +522,7 @@
}
/** A matrix <tt>A</tt> is <i>square</i> if it has the same number of rows and columns. */
- public boolean isSquare(DoubleMatrix2D A) {
+ public boolean isSquare(AbstractMatrix2D A) {
return A.rows() == A.columns();
}
@@ -832,8 +831,8 @@
* </pre>
*/
public String toString(DoubleMatrix2D A) {
- final org.apache.mahout.math.list.ObjectArrayList names = new ObjectArrayList();
- final org.apache.mahout.math.list.ObjectArrayList values = new ObjectArrayList();
+ final ObjectArrayList<String> names = new ObjectArrayList<String>();
+ final ObjectArrayList<String> values = new ObjectArrayList<String>();
// determine properties
names.add("density");
@@ -1059,7 +1058,7 @@
Swapper swapper = new Swapper() {
@Override
public void swap(int a, int b) {
- Object tmp = names.get(a);
+ String tmp = names.get(a);
names.set(a, names.get(b));
names.set(b, tmp);
tmp = values.get(a);
Modified: lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/matrix/linalg/Smp.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/matrix/linalg/Smp.java?rev=901839&r1=901838&r2=901839&view=diff
==============================================================================
--- lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/matrix/linalg/Smp.java (original)
+++ lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/matrix/linalg/Smp.java Thu Jan 21 19:51:44 2010
@@ -140,8 +140,7 @@
if (blocksB == null) {
return null;
}
- DoubleMatrix2D[][] blocks = {blocksA, blocksB};
- return blocks;
+ return new DoubleMatrix2D[][]{blocksA, blocksB};
}
protected DoubleMatrix2D[] splitStridedNN(DoubleMatrix2D A, int threshold, long flops) {
Modified: lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/matrix/linalg/SmpBlas.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/matrix/linalg/SmpBlas.java?rev=901839&r1=901838&r2=901839&view=diff
==============================================================================
--- lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/matrix/linalg/SmpBlas.java (original)
+++ lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/matrix/linalg/SmpBlas.java Thu Jan 21 19:51:44 2010
@@ -216,7 +216,6 @@
@Override
public void run() {
seqBlas.dgemm(transposeA, transposeB, alpha, AA, BB, beta, CC);
- //log.info("Hello "+offset);
}
};
}
@@ -287,7 +286,6 @@
@Override
public void run() {
seqBlas.dgemv(transposeA, alpha, AA, x, beta, yy);
- //log.info("Hello "+offset);
}
};
}
Modified: lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/set/AbstractSet.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/set/AbstractSet.java?rev=901839&r1=901838&r2=901839&view=diff
==============================================================================
--- lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/set/AbstractSet.java (original)
+++ lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/set/AbstractSet.java Thu Jan 21 19:51:44 2010
@@ -50,9 +50,9 @@
protected double maxLoadFactor;
// these are public access for unit tests.
- static public final int defaultCapacity = 277;
- static public final double defaultMinLoadFactor = 0.2;
- static public final double defaultMaxLoadFactor = 0.5;
+ public static final int defaultCapacity = 277;
+ public static final double defaultMinLoadFactor = 0.2;
+ public static final double defaultMaxLoadFactor = 0.5;
/**
* Chooses a new prime table capacity optimized for growing that (approximately) satisfies the invariant <tt>c *
@@ -128,7 +128,7 @@
* @param desiredCapacity the capacity desired by the user.
* @return the capacity which should be used for a hashtable.
*/
-protected int nextPrime(int desiredCapacity) {
+ protected int nextPrime(int desiredCapacity) {
return PrimeFinder.nextPrime(desiredCapacity);
}
Modified: lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/set/OpenHashSet.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/set/OpenHashSet.java?rev=901839&r1=901838&r2=901839&view=diff
==============================================================================
--- lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/set/OpenHashSet.java (original)
+++ lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/set/OpenHashSet.java Thu Jan 21 19:51:44 2010
@@ -94,7 +94,7 @@
@SuppressWarnings("unchecked")
@Override
public Object clone() {
- OpenHashSet copy = (OpenHashSet) super.clone();
+ OpenHashSet<T> copy = (OpenHashSet<T>) super.clone();
copy.table = copy.table.clone();
copy.state = copy.state.clone();
return copy;
@@ -105,6 +105,7 @@
*
* @return <tt>true</tt> if the receiver contains the specified key.
*/
+ @Override
@SuppressWarnings("unchecked")
public boolean contains(Object key) {
return indexOfKey((T)key) >= 0;
@@ -264,16 +265,6 @@
}
}
-
- /**
- * Associates the given key with the given value. Replaces any old <tt>(key,someOtherValue)</tt> association, if
- * existing.
- *
- * @param key the key the value shall be associated with.
- * @param value the value to be associated.
- * @return <tt>true</tt> if the receiver did not already contain such a key; <tt>false</tt> if the receiver did
- * already contain such a key - the new value has now replaced the formerly associated value.
- */
@SuppressWarnings("unchecked")
@Override
public boolean add(Object key) {
@@ -284,10 +275,6 @@
if (this.distinct > this.highWaterMark) {
int newCapacity = chooseGrowCapacity(this.distinct + 1, this.minLoadFactor, this.maxLoadFactor);
- /*
- log.info("grow rehashing ");
- log.info("at distinct="+distinct+", capacity="+table.length+" to newCapacity="+newCapacity+" ...");
- */
rehash(newCapacity);
return add(key);
}
@@ -360,12 +347,6 @@
if (this.distinct < this.lowWaterMark) {
int newCapacity = chooseShrinkCapacity(this.distinct, this.minLoadFactor, this.maxLoadFactor);
- /*
- if (table.length != newCapacity) {
- log.info("shrink rehashing ");
- log.info("at distinct="+distinct+", capacity="+table.length+" to newCapacity="+newCapacity+" ...");
- }
- */
rehash(newCapacity);
}
@@ -460,7 +441,7 @@
if (!(obj instanceof OpenHashSet)) {
return false;
}
- final OpenHashSet other = (OpenHashSet) obj;
+ final OpenHashSet<T> other = (OpenHashSet<T>) obj;
if (other.size() != size()) {
return false;
}
Modified: lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/stats/LogLikelihood.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/stats/LogLikelihood.java?rev=901839&r1=901838&r2=901839&view=diff
==============================================================================
--- lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/stats/LogLikelihood.java (original)
+++ lucene/mahout/trunk/math/src/main/java/org/apache/mahout/math/stats/LogLikelihood.java Thu Jan 21 19:51:44 2010
@@ -21,6 +21,10 @@
* Utility methods for working with log-likelihood
*/
public class LogLikelihood {
+
+ private LogLikelihood() {
+ }
+
/**
* Calculate the Shannon entropy.
* @param elements TODO FILL IN HERE
@@ -34,7 +38,7 @@
double result = 0.0;
for (int x : elements) {
if (x < 0) {
- throw new IllegalArgumentException("Should not have negative count for entropy computation: (" + x + ")");
+ throw new IllegalArgumentException("Should not have negative count for entropy computation: (" + x + ')');
}
int zeroFlag = (x == 0 ? 1 : 0);
result += x * Math.log((x + zeroFlag) / sum);
Modified: lucene/mahout/trunk/math/src/test/java-templates/org/apache/mahout/math/map/OpenKeyTypeObjectHashMapTest.java.t
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/math/src/test/java-templates/org/apache/mahout/math/map/OpenKeyTypeObjectHashMapTest.java.t?rev=901839&r1=901838&r2=901839&view=diff
==============================================================================
--- lucene/mahout/trunk/math/src/test/java-templates/org/apache/mahout/math/map/OpenKeyTypeObjectHashMapTest.java.t (original)
+++ lucene/mahout/trunk/math/src/test/java-templates/org/apache/mahout/math/map/OpenKeyTypeObjectHashMapTest.java.t Thu Jan 21 19:51:44 2010
@@ -66,8 +66,7 @@
if (obj == null) return false;
if (getClass() != obj.getClass()) return false;
TestClass other = (TestClass) obj;
- if (x != other.x) return false;
- return true;
+ return x == other.x;
}
${keyType} x;
Modified: lucene/mahout/trunk/math/src/test/java-templates/org/apache/mahout/math/map/OpenObjectValueTypeHashMapTest.java.t
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/math/src/test/java-templates/org/apache/mahout/math/map/OpenObjectValueTypeHashMapTest.java.t?rev=901839&r1=901838&r2=901839&view=diff
==============================================================================
--- lucene/mahout/trunk/math/src/test/java-templates/org/apache/mahout/math/map/OpenObjectValueTypeHashMapTest.java.t (original)
+++ lucene/mahout/trunk/math/src/test/java-templates/org/apache/mahout/math/map/OpenObjectValueTypeHashMapTest.java.t Thu Jan 21 19:51:44 2010
@@ -63,12 +63,11 @@
if (obj == null) return false;
if (getClass() != obj.getClass()) return false;
NotComparableKey other = (NotComparableKey) obj;
- if (x != other.x) return false;
- return true;
+ return x == other.x;
}
}
- private NotComparableKey[] ncKeys = {
+ private final NotComparableKey[] ncKeys = {
new NotComparableKey(101),
new NotComparableKey(99),
new NotComparableKey(2),
Modified: lucene/mahout/trunk/math/src/test/java/org/apache/mahout/math/SortingTest.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/math/src/test/java/org/apache/mahout/math/SortingTest.java?rev=901839&r1=901838&r2=901839&view=diff
==============================================================================
--- lucene/mahout/trunk/math/src/test/java/org/apache/mahout/math/SortingTest.java (original)
+++ lucene/mahout/trunk/math/src/test/java/org/apache/mahout/math/SortingTest.java Thu Jan 21 19:51:44 2010
@@ -100,7 +100,7 @@
}
static class ForSorting implements Comparable<ForSorting> {
- private Integer i;
+ private final Integer i;
ForSorting(int i) {
this.i = i;
@@ -127,7 +127,7 @@
@Test
public void testBinarySearch() {
- byte[] bytes = new byte[] {-5, -2, 0, 100, 103};
+ byte[] bytes = {-5, -2, 0, 100, 103};
int x = Sorting.binarySearchFromTo(bytes, (byte) -6, 0, 4);
assertEquals(-1, x);
x = Sorting.binarySearchFromTo(bytes, (byte) 0, 0, 4);
@@ -137,7 +137,7 @@
x = Sorting.binarySearchFromTo(bytes, (byte) 0, 3, 4);
assertEquals(-4, x);
- char[] chars = new char[] {1, 2, 5, 100, 103};
+ char[] chars = {1, 2, 5, 100, 103};
x = Sorting.binarySearchFromTo(chars, (char) 0, 0, 4);
assertEquals(-1, x);
x = Sorting.binarySearchFromTo(chars, (char) 1, 0, 4);
@@ -147,7 +147,7 @@
x = Sorting.binarySearchFromTo(chars, (char) 0, 3, 4);
assertEquals(-4, x);
- short[] shorts = new short[] {-5, -2, 0, 100, 103};
+ short[] shorts = {-5, -2, 0, 100, 103};
x = Sorting.binarySearchFromTo(shorts, (short) -6, 0, 4);
assertEquals(-1, x);
x = Sorting.binarySearchFromTo(shorts, (short) 0, 0, 4);
@@ -157,7 +157,7 @@
x = Sorting.binarySearchFromTo(shorts, (short) 0, 3, 4);
assertEquals(-4, x);
- int[] ints = new int[] {-5, -2, 0, 100, 103};
+ int[] ints = {-5, -2, 0, 100, 103};
x = Sorting.binarySearchFromTo(ints, (int) -6, 0, 4);
assertEquals(-1, x);
x = Sorting.binarySearchFromTo(ints, (int) 0, 0, 4);
@@ -167,7 +167,7 @@
x = Sorting.binarySearchFromTo(ints, (int) 0, 3, 4);
assertEquals(-4, x);
- long[] longs = new long[] {-5, -2, 0, 100, 103};
+ long[] longs = {-5, -2, 0, 100, 103};
x = Sorting.binarySearchFromTo(longs, (long) -6, 0, 4);
assertEquals(-1, x);
x = Sorting.binarySearchFromTo(longs, (long) 0, 0, 4);
@@ -177,7 +177,7 @@
x = Sorting.binarySearchFromTo(longs, (long) 0, 3, 4);
assertEquals(-4, x);
- float[] floats = new float[] {-5, -2, 0, 100, 103};
+ float[] floats = {-5, -2, 0, 100, 103};
x = Sorting.binarySearchFromTo(floats, (float) -6, 0, 4);
assertEquals(-1, x);
x = Sorting.binarySearchFromTo(floats, (float) 0, 0, 4);
@@ -187,7 +187,7 @@
x = Sorting.binarySearchFromTo(floats, (float) 0, 3, 4);
assertEquals(-4, x);
- double[] doubles = new double[] {-5, -2, 0, 100, 103};
+ double[] doubles = {-5, -2, 0, 100, 103};
x = Sorting.binarySearchFromTo(doubles, (double) -6, 0, 4);
assertEquals(-1, x);
x = Sorting.binarySearchFromTo(doubles, (double) 0, 0, 4);
@@ -311,10 +311,10 @@
@Test
public void testQuickSortExternals() {
- int stuff[] = randomInts();
+ int[] stuff = randomInts();
final Integer[] bigInts = new Integer[stuff.length];
for (int x = 0; x < stuff.length; x ++) {
- bigInts[x] = Integer.valueOf(stuff[x]);
+ bigInts[x] = stuff[x];
}
Sorting.quickSort(0, stuff.length, new IntComparator() {
Modified: lucene/mahout/trunk/math/src/test/java/org/apache/mahout/math/map/OpenHashMapTest.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/math/src/test/java/org/apache/mahout/math/map/OpenHashMapTest.java?rev=901839&r1=901838&r2=901839&view=diff
==============================================================================
--- lucene/mahout/trunk/math/src/test/java/org/apache/mahout/math/map/OpenHashMapTest.java (original)
+++ lucene/mahout/trunk/math/src/test/java/org/apache/mahout/math/map/OpenHashMapTest.java Thu Jan 21 19:51:44 2010
@@ -135,8 +135,8 @@
}
private static class Pair implements Comparable<Pair> {
- String k;
- String v;
+ final String k;
+ final String v;
Pair(String k, String v) {
this.k = k;
Modified: lucene/mahout/trunk/utils/src/main/java/org/apache/mahout/utils/SequenceFileDumper.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/utils/src/main/java/org/apache/mahout/utils/SequenceFileDumper.java?rev=901839&r1=901838&r2=901839&view=diff
==============================================================================
--- lucene/mahout/trunk/utils/src/main/java/org/apache/mahout/utils/SequenceFileDumper.java (original)
+++ lucene/mahout/trunk/utils/src/main/java/org/apache/mahout/utils/SequenceFileDumper.java Thu Jan 21 19:51:44 2010
@@ -33,7 +33,6 @@
import org.apache.hadoop.mapred.JobClient;
import org.apache.hadoop.mapred.JobConf;
import org.apache.hadoop.mapred.jobcontrol.Job;
-import org.apache.mahout.utils.strings.StringUtil;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;
@@ -95,7 +94,7 @@
} else {
writer = new OutputStreamWriter(System.out);
}
- writer.append("Input Path: ").append(String.valueOf(path)).append(StringUtil.LINE_SEP);
+ writer.append("Input Path: ").append(String.valueOf(path)).append('\n');
int sub = Integer.MAX_VALUE;
if (cmdLine.hasOption(substringOpt)) {
@@ -104,7 +103,8 @@
boolean countOnly = cmdLine.hasOption(countOpt);
Writable key = (Writable) reader.getKeyClass().newInstance();
Writable value = (Writable) reader.getValueClass().newInstance();
- writer.append("Key class: ").append(String.valueOf(reader.getKeyClass())).append(" Value Class: ").append(String.valueOf(value.getClass())).append(StringUtil.LINE_SEP);
+ writer.append("Key class: ").append(String.valueOf(reader.getKeyClass())).append(" Value Class: ")
+ .append(String.valueOf(value.getClass())).append('\n');
writer.flush();
long count = 0;
if (countOnly == false) {
@@ -112,16 +112,16 @@
writer.append("Key: ").append(String.valueOf(key));
String str = value.toString();
writer.append(": Value: ").append(str.length() > sub ? str.substring(0, sub) : str);
- writer.write(StringUtil.LINE_SEP);
+ writer.write('\n');
writer.flush();
count++;
}
- writer.append("Count: ").append(String.valueOf(count)).append(StringUtil.LINE_SEP);
+ writer.append("Count: ").append(String.valueOf(count)).append('\n');
} else {
while (reader.next(key, value)) {
count++;
}
- writer.append("Count: ").append(String.valueOf(count)).append(StringUtil.LINE_SEP);
+ writer.append("Count: ").append(String.valueOf(count)).append('\n');
}
writer.flush();
if (cmdLine.hasOption(outputOpt)) {
Modified: lucene/mahout/trunk/utils/src/main/java/org/apache/mahout/utils/clustering/ClusterDumper.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/utils/src/main/java/org/apache/mahout/utils/clustering/ClusterDumper.java?rev=901839&r1=901838&r2=901839&view=diff
==============================================================================
--- lucene/mahout/trunk/utils/src/main/java/org/apache/mahout/utils/clustering/ClusterDumper.java (original)
+++ lucene/mahout/trunk/utils/src/main/java/org/apache/mahout/utils/clustering/ClusterDumper.java Thu Jan 21 19:51:44 2010
@@ -59,15 +59,13 @@
public final class ClusterDumper {
private static final Logger log = LoggerFactory.getLogger(ClusterDumper.class);
- private static final String LINE_SEP = System.getProperty("line.separator");
-
- String seqFileDir;
- String pointsDir;
- String termDictionary;
- String outputFile;
- int subString = Integer.MAX_VALUE;
- Map<String, List<String>> clusterIdToPoints = null;
+ private final String seqFileDir;
+ private final String pointsDir;
+ private String termDictionary;
+ private String outputFile;
+ private int subString = Integer.MAX_VALUE;
+ private Map<String, List<String>> clusterIdToPoints = null;
private boolean useJSON = false;
public ClusterDumper(String seqFileDir, String pointsDir) throws IOException {
@@ -121,15 +119,15 @@
ClusterBase value = (ClusterBase) reader.getValueClass().newInstance();
while (reader.next(key, value)){
Vector center = value.getCenter();
- String fmtStr = useJSON == false ? VectorHelper.vectorToString(center, dictionary) : center.asFormatString();
+ String fmtStr = useJSON ? center.asFormatString() : VectorHelper.vectorToString(center, dictionary);
writer.append("Id: ").append(String.valueOf(value.getId())).append(":").append("name:")
- .append(center.getName()).append(":").append(fmtStr.substring(0, Math.min(subString, fmtStr.length()))).append(LINE_SEP);
+ .append(center.getName()).append(":").append(fmtStr.substring(0, Math.min(subString, fmtStr.length()))).append('\n');
if (dictionary != null) {
String topTerms = getTopFeatures(center, dictionary, 10);
writer.write("\tTop Terms: ");
writer.write(topTerms);
- writer.write(LINE_SEP);
+ writer.write('\n');
}
List<String> points = clusterIdToPoints.get(String.valueOf(value.getId()));
@@ -142,7 +140,7 @@
writer.append(", ");
}
}
- writer.write(LINE_SEP);
+ writer.write('\n');
}
writer.flush();
}
@@ -338,7 +336,7 @@
int index = vectorTerms.get(i).index;
String dictTerm = dictionary[index];
if (dictTerm == null) {
- log.error("Dictionary entry missing for "+ index);
+ log.error("Dictionary entry missing for {}", index);
continue;
}
topTerms.add(dictTerm);
Modified: lucene/mahout/trunk/utils/src/main/java/org/apache/mahout/utils/vectors/VectorDumper.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/utils/src/main/java/org/apache/mahout/utils/vectors/VectorDumper.java?rev=901839&r1=901838&r2=901839&view=diff
==============================================================================
--- lucene/mahout/trunk/utils/src/main/java/org/apache/mahout/utils/vectors/VectorDumper.java (original)
+++ lucene/mahout/trunk/utils/src/main/java/org/apache/mahout/utils/vectors/VectorDumper.java Thu Jan 21 19:51:44 2010
@@ -51,7 +51,6 @@
public final class VectorDumper {
private static final Logger log = LoggerFactory.getLogger(VectorDumper.class);
- private static final String LINE_SEP = System.getProperty("line.separator");
private VectorDumper() {
}
@@ -124,9 +123,9 @@
writer.write(iterator.key().toString());
writer.write("\t");
}
- String fmtStr = useJSON == true ? vector.asFormatString() : (dictionary != null ? VectorHelper.vectorToString(vector, dictionary) : vector.asFormatString());
+ String fmtStr = useJSON ? vector.asFormatString() : (dictionary != null ? VectorHelper.vectorToString(vector, dictionary) : vector.asFormatString());
writer.write(fmtStr);
- writer.write(LINE_SEP);
+ writer.write('\n');
i++;
}
writer.flush();
Modified: lucene/mahout/trunk/utils/src/main/java/org/apache/mahout/utils/vectors/arff/Driver.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/utils/src/main/java/org/apache/mahout/utils/vectors/arff/Driver.java?rev=901839&r1=901838&r2=901839&view=diff
==============================================================================
--- lucene/mahout/trunk/utils/src/main/java/org/apache/mahout/utils/vectors/arff/Driver.java (original)
+++ lucene/mahout/trunk/utils/src/main/java/org/apache/mahout/utils/vectors/arff/Driver.java Thu Jan 21 19:51:44 2010
@@ -32,7 +32,6 @@
import org.apache.hadoop.io.SequenceFile;
import org.apache.mahout.math.RandomAccessSparseVector;
import org.apache.mahout.common.CommandLineUtil;
-import org.apache.mahout.utils.strings.StringUtil;
import org.apache.mahout.utils.vectors.io.JWriterVectorWriter;
import org.apache.mahout.utils.vectors.io.SequenceFileVectorWriter;
import org.apache.mahout.utils.vectors.io.VectorWriter;
@@ -109,7 +108,7 @@
throw new IllegalArgumentException("maxDocs must be >= 0");
}
String outDir = cmdLine.getValue(outputOpt).toString();
- log.info("Output Dir: " + outDir);
+ log.info("Output Dir: {}", outDir);
String outWriter = null;
if (cmdLine.hasOption(outWriterOpt)) {
outWriter = cmdLine.getValue(outWriterOpt).toString();
@@ -131,11 +130,11 @@
} else {
writeFile(outWriter, outDir, input, maxDocs, model);
}
- log.info("Dictionary Output file: " + dictOut);
+ log.info("Dictionary Output file: {}", dictOut);
BufferedWriter writer = new BufferedWriter(new OutputStreamWriter(new FileOutputStream(dictOut), Charset.forName("UTF8")));
Map<String, Integer> labels = model.getLabelBindings();
for (Map.Entry<String, Integer> entry : labels.entrySet()) {
- writer.append(entry.getKey()).append(delimiter).append(String.valueOf(entry.getValue())).append(StringUtil.LINE_SEP);
+ writer.append(entry.getKey()).append(delimiter).append(String.valueOf(entry.getValue())).append('\n');
}
writer.close();
}
@@ -148,7 +147,7 @@
private static void writeFile(String outWriter, String outDir, File file,
long maxDocs, ARFFModel arffModel) throws IOException {
- log.info("Converting File: " + file);
+ log.info("Converting File: {}", file);
ARFFModel model = new MapBackedARFFModel(arffModel.getWords(), arffModel.getWordCount() + 1,
arffModel.getNominalMap());
ARFFVectorIterable iteratable = new ARFFVectorIterable(file, model);
@@ -168,7 +167,7 @@
long numDocs = vectorWriter.write(iteratable, maxDocs);
vectorWriter.close();
- log.info("Wrote: " + numDocs + " vectors");
+ log.info("Wrote: {} vectors", numDocs);
}
private static VectorWriter getSeqFileWriter(String outFile) throws IOException {
Modified: lucene/mahout/trunk/utils/src/main/java/org/apache/mahout/utils/vectors/lucene/ClusterLabels.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/utils/src/main/java/org/apache/mahout/utils/vectors/lucene/ClusterLabels.java?rev=901839&r1=901838&r2=901839&view=diff
==============================================================================
--- lucene/mahout/trunk/utils/src/main/java/org/apache/mahout/utils/vectors/lucene/ClusterLabels.java (original)
+++ lucene/mahout/trunk/utils/src/main/java/org/apache/mahout/utils/vectors/lucene/ClusterLabels.java Thu Jan 21 19:51:44 2010
@@ -69,9 +69,9 @@
public class ClusterLabels {
class TermInfoClusterInOut implements Comparable<TermInfoClusterInOut> {
- public String term;
- public int inClusterDF;
- public int outClusterDF;
+ public final String term;
+ public final int inClusterDF;
+ public final int outClusterDF;
public double logLikelihoodRatio;
public TermInfoClusterInOut(String term, int inClusterDF, int outClusterDF) {
@@ -95,14 +95,13 @@
}
private static final Logger log = LoggerFactory.getLogger(ClusterLabels.class);
- private static final String LINE_SEP = System.getProperty("line.separator");
public static final int DEFAULT_MIN_IDS = 50;
public static final int DEFAULT_MAX_LABELS = 25;
- private String seqFileDir;
- private String pointsDir;
- private String indexDir;
- private String contentField;
+ private final String seqFileDir;
+ private final String pointsDir;
+ private final String indexDir;
+ private final String contentField;
private String idField;
private Map<String, List<String>> clusterIdToPoints = null;
private String output;
@@ -137,14 +136,14 @@
List<String> ids = clusterIdToPoints.get(clusterID);
List<TermInfoClusterInOut> termInfos = getClusterLabels(clusterID, ids);
if (termInfos != null) {
- writer.write(LINE_SEP);
+ writer.write('\n');
writer.write("Top labels for Cluster " + clusterID + " containing " + ids.size() + " vectors");
- writer.write(LINE_SEP);
+ writer.write('\n');
writer.write("Term \t\t LLR \t\t In-ClusterDF \t\t Out-ClusterDF ");
- writer.write(LINE_SEP);
+ writer.write('\n');
for (TermInfoClusterInOut termInfo : termInfos) {
writer.write(termInfo.term + "\t\t" + termInfo.logLikelihoodRatio + "\t\t" + termInfo.inClusterDF + "\t\t" + termInfo.outClusterDF);
- writer.write(LINE_SEP);
+ writer.write('\n');
}
}
}
@@ -163,18 +162,18 @@
* @throws CorruptIndexException
* @throws IOException
*/
- protected List<TermInfoClusterInOut> getClusterLabels(String clusterID, List<String> ids) throws CorruptIndexException, IOException {
+ protected List<TermInfoClusterInOut> getClusterLabels(String clusterID, List<String> ids) throws IOException {
if (ids.size() < minNumIds) {
- log.info("Skipping small cluster " + clusterID + " with size: " + ids.size());
+ log.info("Skipping small cluster {} with size: {}", clusterID, ids.size());
return null;
}
- log.info("Processing Cluster " + clusterID + " with " + ids.size() + " documents");
+ log.info("Processing Cluster {} with {} documents", clusterID, ids.size());
Directory dir = FSDirectory.open(new File(this.indexDir));
IndexReader reader = IndexReader.open(dir, false);
- log.info("# of documents in the index " + reader.numDocs());
+ log.info("# of documents in the index {}", reader.numDocs());
Set<String> idSet = new HashSet<String>();
idSet.addAll(ids);
@@ -266,7 +265,7 @@
bitset.set(i);
}
}
- log.info("Created bitset for in-cluster documents : " + bitset.cardinality());
+ log.info("Created bitset for in-cluster documents : {}", bitset.cardinality());
return bitset;
}
Modified: lucene/mahout/trunk/utils/src/main/java/org/apache/mahout/utils/vectors/lucene/Driver.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/utils/src/main/java/org/apache/mahout/utils/vectors/lucene/Driver.java?rev=901839&r1=901838&r2=901839&view=diff
==============================================================================
--- lucene/mahout/trunk/utils/src/main/java/org/apache/mahout/utils/vectors/lucene/Driver.java (original)
+++ lucene/mahout/trunk/utils/src/main/java/org/apache/mahout/utils/vectors/lucene/Driver.java Thu Jan 21 19:51:44 2010
@@ -182,7 +182,7 @@
iterable = new LuceneIterable(reader, idField, field, mapper, norm);
}
String outFile = cmdLine.getValue(outputOpt).toString();
- log.info("Output File: " + outFile);
+ log.info("Output File: {}", outFile);
VectorWriter vectorWriter;
if (cmdLine.hasOption(outWriterOpt)) {
@@ -199,11 +199,11 @@
long numDocs = vectorWriter.write(iterable, maxDocs);
vectorWriter.close();
- log.info("Wrote: " + numDocs + " vectors");
+ log.info("Wrote: {} vectors", numDocs);
String delimiter = cmdLine.hasOption(delimiterOpt) ? cmdLine.getValue(delimiterOpt).toString() : "\t";
File dictOutFile = new File(cmdLine.getValue(dictOutOpt).toString());
- log.info("Dictionary Output file: " + dictOutFile);
+ log.info("Dictionary Output file: {}", dictOutFile);
BufferedWriter writer = new BufferedWriter(new OutputStreamWriter(new FileOutputStream(dictOutFile), Charset.forName("UTF8")));
JWriterTermInfoWriter tiWriter = new JWriterTermInfoWriter(writer, delimiter, field);
tiWriter.write(termInfo);
Modified: lucene/mahout/trunk/utils/src/main/java/org/apache/mahout/utils/vectors/text/DictionaryVectorizer.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/utils/src/main/java/org/apache/mahout/utils/vectors/text/DictionaryVectorizer.java?rev=901839&r1=901838&r2=901839&view=diff
==============================================================================
--- lucene/mahout/trunk/utils/src/main/java/org/apache/mahout/utils/vectors/text/DictionaryVectorizer.java (original)
+++ lucene/mahout/trunk/utils/src/main/java/org/apache/mahout/utils/vectors/text/DictionaryVectorizer.java Thu Jan 21 19:51:44 2010
@@ -96,7 +96,7 @@
* @param output
* output directory where {@link org.apache.mahout.math.RandomAccessSparseVector}'s of the document are
* generated
- * @param analyzer
+ * @param analyzerClass
* the Lucene {@link Analyzer} used to tokenize the UTF-8
* @param minSupport
* the minimum frequency of the feature in the entire corpus to be
@@ -248,10 +248,7 @@
* @throws URISyntaxException
*/
private static void createVectorFromPartialVectors(List<Path> partialVectorPaths,
- String output) throws IOException,
- InterruptedException,
- ClassNotFoundException,
- URISyntaxException {
+ String output) throws IOException {
Configurable client = new JobClient();
JobConf conf = new JobConf(DictionaryVectorizer.class);
@@ -311,7 +308,7 @@
* location of the chunk of features and the id's
* @param output
* output directory were the partial vectors have to be created
- * @param analyzer
+ * @param analyzerClass
* The Lucene {@link Analyzer} for tokenizing the text
* @throws IOException
* @throws InterruptedException
@@ -321,10 +318,7 @@
private static void makePartialVectors(String input,
Path dictionaryFilePath,
Class<? extends Analyzer> analyzerClass,
- Path output) throws IOException,
- InterruptedException,
- ClassNotFoundException,
- URISyntaxException {
+ Path output) throws IOException {
Configurable client = new JobClient();
JobConf conf = new JobConf(DictionaryVectorizer.class);
@@ -365,19 +359,10 @@
/**
* Count the frequencies of words in parallel using Map/Reduce. The input
* documents have to be in {@link SequenceFile} format
- *
- * @param input
- * @param output
- * @param analyzer
- * @throws IOException
- * @throws InterruptedException
- * @throws ClassNotFoundException
*/
private static void startWordCounting(Path input,
Class<? extends Analyzer> analyzerClass,
- Path output) throws IOException,
- InterruptedException,
- ClassNotFoundException {
+ Path output) throws IOException {
Configurable client = new JobClient();
JobConf conf = new JobConf(DictionaryVectorizer.class);
Modified: lucene/mahout/trunk/utils/src/main/java/org/apache/mahout/utils/vectors/text/PartialVectorGenerator.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/utils/src/main/java/org/apache/mahout/utils/vectors/text/PartialVectorGenerator.java?rev=901839&r1=901838&r2=901839&view=diff
==============================================================================
--- lucene/mahout/trunk/utils/src/main/java/org/apache/mahout/utils/vectors/text/PartialVectorGenerator.java (original)
+++ lucene/mahout/trunk/utils/src/main/java/org/apache/mahout/utils/vectors/text/PartialVectorGenerator.java Thu Jan 21 19:51:44 2010
@@ -45,11 +45,11 @@
public class PartialVectorGenerator extends MapReduceBase implements
Reducer<Text,Text,Text,VectorWritable> {
private Analyzer analyzer;
- private Map<String,int[]> dictionary = new HashMap<String,int[]>();
+ private final Map<String,int[]> dictionary = new HashMap<String,int[]>();
private FileSystem fs; // local filesystem
private URI[] localFiles; // local filenames from the distributed cache
- private VectorWritable vectorWritable = new VectorWritable();
+ private final VectorWritable vectorWritable = new VectorWritable();
public void reduce(Text key,
Iterator<Text> values,
Modified: lucene/mahout/trunk/utils/src/main/java/org/apache/mahout/utils/vectors/text/PartialVectorMerger.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/utils/src/main/java/org/apache/mahout/utils/vectors/text/PartialVectorMerger.java?rev=901839&r1=901838&r2=901839&view=diff
==============================================================================
--- lucene/mahout/trunk/utils/src/main/java/org/apache/mahout/utils/vectors/text/PartialVectorMerger.java (original)
+++ lucene/mahout/trunk/utils/src/main/java/org/apache/mahout/utils/vectors/text/PartialVectorMerger.java Thu Jan 21 19:51:44 2010
@@ -34,7 +34,7 @@
public class PartialVectorMerger extends MapReduceBase implements
Reducer<Text,VectorWritable,Text, VectorWritable> {
- private VectorWritable vectorWritable = new VectorWritable();
+ private final VectorWritable vectorWritable = new VectorWritable();
@Override
public void reduce(Text key,
Modified: lucene/mahout/trunk/utils/src/main/java/org/apache/mahout/utils/vectors/text/TermCountMapper.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/utils/src/main/java/org/apache/mahout/utils/vectors/text/TermCountMapper.java?rev=901839&r1=901838&r2=901839&view=diff
==============================================================================
--- lucene/mahout/trunk/utils/src/main/java/org/apache/mahout/utils/vectors/text/TermCountMapper.java (original)
+++ lucene/mahout/trunk/utils/src/main/java/org/apache/mahout/utils/vectors/text/TermCountMapper.java Thu Jan 21 19:51:44 2010
@@ -32,11 +32,9 @@
import org.apache.hadoop.mapred.OutputCollector;
import org.apache.hadoop.mapred.Reporter;
import org.apache.lucene.analysis.Analyzer;
-import org.apache.lucene.analysis.Token;
import org.apache.lucene.analysis.TokenStream;
import org.apache.lucene.analysis.standard.StandardAnalyzer;
import org.apache.lucene.analysis.tokenattributes.TermAttribute;
-import org.apache.mahout.common.parameters.ClassParameter;
/**
* TextVectorizer Term Count Mapper. Tokenizes a text document and outputs the
Modified: lucene/mahout/trunk/utils/src/test/java/org/apache/mahout/utils/vectors/SequenceFileVectorIterableTest.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/utils/src/test/java/org/apache/mahout/utils/vectors/SequenceFileVectorIterableTest.java?rev=901839&r1=901838&r2=901839&view=diff
==============================================================================
--- lucene/mahout/trunk/utils/src/test/java/org/apache/mahout/utils/vectors/SequenceFileVectorIterableTest.java (original)
+++ lucene/mahout/trunk/utils/src/test/java/org/apache/mahout/utils/vectors/SequenceFileVectorIterableTest.java Thu Jan 21 19:51:44 2010
@@ -30,12 +30,29 @@
import java.io.File;
public class SequenceFileVectorIterableTest extends MahoutTestCase {
- public void testIterable() throws Exception {
+
+ private File tmpLoc;
+ private File tmpFile;
+
+ @Override
+ public void setUp() throws Exception {
+ super.setUp();
File tmpDir = new File(System.getProperty("java.io.tmpdir"));
- File tmpLoc = new File(tmpDir, "sfvit");
+ tmpLoc = new File(tmpDir, "sfvit");
+ tmpLoc.deleteOnExit();
tmpLoc.mkdirs();
- File tmpFile = File.createTempFile("sfvit", ".dat", tmpLoc);
+ tmpFile = File.createTempFile("sfvit", ".dat", tmpLoc);
+ tmpFile.deleteOnExit();
+ }
+ @Override
+ public void tearDown() throws Exception {
+ tmpFile.delete();
+ tmpLoc.delete();
+ super.tearDown();
+ }
+
+ public void testIterable() throws Exception {
Path path = new Path(tmpFile.getAbsolutePath());
Configuration conf = new Configuration();
FileSystem fs = FileSystem.get(conf);
@@ -49,10 +66,10 @@
SequenceFileVectorIterable sfvi = new SequenceFileVectorIterable(seqReader);
int count = 0;
for (Vector vector : sfvi) {
- System.out.println("Vec: " + vector.asFormatString());
+ //System.out.println("Vec: " + vector.asFormatString());
count++;
}
seqReader.close();
- assertEquals(count + " does not equal: " + 50, 50, count);
+ assertEquals(50, count);
}
}
Modified: lucene/mahout/trunk/utils/src/test/java/org/apache/mahout/utils/vectors/arff/ARFFVectorIterableTest.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/utils/src/test/java/org/apache/mahout/utils/vectors/arff/ARFFVectorIterableTest.java?rev=901839&r1=901838&r2=901839&view=diff
==============================================================================
--- lucene/mahout/trunk/utils/src/test/java/org/apache/mahout/utils/vectors/arff/ARFFVectorIterableTest.java (original)
+++ lucene/mahout/trunk/utils/src/test/java/org/apache/mahout/utils/vectors/arff/ARFFVectorIterableTest.java Thu Jan 21 19:51:44 2010
@@ -21,7 +21,6 @@
import org.apache.mahout.math.DenseVector;
import org.apache.mahout.math.RandomAccessSparseVector;
import org.apache.mahout.math.Vector;
-import org.apache.mahout.utils.strings.StringUtil;
import java.text.DateFormat;
import java.util.Iterator;
@@ -31,16 +30,16 @@
public void testValues() throws Exception {
StringBuilder builder = new StringBuilder();
- builder.append("%comments").append(StringUtil.LINE_SEP).append("@RELATION Mahout").append(StringUtil.LINE_SEP)
- .append("@ATTRIBUTE foo numeric").append(StringUtil.LINE_SEP)
- .append("@ATTRIBUTE bar numeric").append(StringUtil.LINE_SEP)
- .append("@ATTRIBUTE timestamp DATE \"yyyy-MM-dd HH:mm:ss\"").append(StringUtil.LINE_SEP)
- .append("@ATTRIBUTE junk string").append(StringUtil.LINE_SEP)
- .append("@ATTRIBUTE theNominal {c,b,a}").append(StringUtil.LINE_SEP)
- .append("@DATA").append(StringUtil.LINE_SEP)
- .append("1,2, \"2009-01-01 5:55:55\", foo, c").append(StringUtil.LINE_SEP)
- .append("2,3").append(StringUtil.LINE_SEP)
- .append("{0 5,1 23}").append(StringUtil.LINE_SEP);
+ builder.append("%comments").append('\n').append("@RELATION Mahout").append('\n')
+ .append("@ATTRIBUTE foo numeric").append('\n')
+ .append("@ATTRIBUTE bar numeric").append('\n')
+ .append("@ATTRIBUTE timestamp DATE \"yyyy-MM-dd HH:mm:ss\"").append('\n')
+ .append("@ATTRIBUTE junk string").append('\n')
+ .append("@ATTRIBUTE theNominal {c,b,a}").append('\n')
+ .append("@DATA").append('\n')
+ .append("1,2, \"2009-01-01 5:55:55\", foo, c").append('\n')
+ .append("2,3").append('\n')
+ .append("{0 5,1 23}").append('\n');
ARFFModel model = new MapBackedARFFModel();
ARFFVectorIterable iterable = new ARFFVectorIterable(builder.toString(), model);
assertEquals("Mahout", iterable.getModel().getRelation());
Modified: lucene/mahout/trunk/utils/src/test/java/org/apache/mahout/utils/vectors/io/VectorWriterTest.java
URL: http://svn.apache.org/viewvc/lucene/mahout/trunk/utils/src/test/java/org/apache/mahout/utils/vectors/io/VectorWriterTest.java?rev=901839&r1=901838&r2=901839&view=diff
==============================================================================
--- lucene/mahout/trunk/utils/src/test/java/org/apache/mahout/utils/vectors/io/VectorWriterTest.java (original)
+++ lucene/mahout/trunk/utils/src/test/java/org/apache/mahout/utils/vectors/io/VectorWriterTest.java Thu Jan 21 19:51:44 2010
@@ -17,7 +17,6 @@
package org.apache.mahout.utils.vectors.io;
-import junit.framework.TestCase;
import org.apache.hadoop.fs.Path;
import org.apache.hadoop.fs.FileSystem;
import org.apache.hadoop.conf.Configuration;
@@ -36,12 +35,28 @@
public class VectorWriterTest extends MahoutTestCase {
- public void testSFVW() throws Exception {
+ private File tmpLoc;
+ private File tmpFile;
+
+ @Override
+ public void setUp() throws Exception {
+ super.setUp();
File tmpDir = new File(System.getProperty("java.io.tmpdir"));
- File tmpLoc = new File(tmpDir, "sfvwt");
+ tmpLoc = new File(tmpDir, "sfvwt");
+ tmpLoc.deleteOnExit();
tmpLoc.mkdirs();
- File tmpFile = File.createTempFile("sfvwt", ".dat", tmpLoc);
+ tmpFile = File.createTempFile("sfvwt", ".dat", tmpLoc);
+ tmpFile.deleteOnExit();
+ }
+ @Override
+ public void tearDown() throws Exception {
+ tmpFile.delete();
+ tmpLoc.delete();
+ super.tearDown();
+ }
+
+ public void testSFVW() throws Exception {
Path path = new Path(tmpFile.getAbsolutePath());
Configuration conf = new Configuration();
FileSystem fs = FileSystem.get(conf);
@@ -72,7 +87,6 @@
StringBuffer buffer = strWriter.getBuffer();
assertNotNull(buffer);
assertTrue(buffer.length() > 0);
- System.out.println("Buffer: " + buffer);
}
}