@Test
public void testExtractorMapperCSV() throws Exception {
FeatureExtractorMapper mapper = new FeatureExtractorMapper();
Configuration conf = getConfiguration();
conf.set("vector.implementation.class.name", "org.apache.mahout.math.RandomAccessSparseVector");
conf.set(FeatureExtractorKeySet.FEATURE_NAMES, RAW_CSV[0]);
conf.set(FeatureExtractorKeySet.SELECTED_DEPENDENT, DEPENDENT);
conf.set(FeatureExtractorKeySet.SELECTED_INDEPENDENT, INDEPENDENT);
conf.set(FeatureExtractorKeySet.SELECTED_INTERACTION, INTERACTION);
conf.set(FeatureExtractorKeySet.SEPARATOR, SEP_CSV);
DummyRecordWriter<Text, VectorWritable> writer = new DummyRecordWriter<Text, VectorWritable>();
Mapper<LongWritable, Text, Text, VectorWritable>.Context context =
DummyRecordWriter.build(mapper, conf, writer);
mapper.setup(context);
for (int i = 0; i < RAW_CSV.length; ++i) {
mapper.map(new LongWritable(i), new Text(RAW_CSV[i]), context);
}
assertEquals("Number of map results", 1, writer.getData().size());
assertEquals("Number of map results", 1, writer.getData().size());
for (int i = 0; i < writer.getValue(new Text("5")).size(); ++i) {
assertEquals(
"Features: ",
getFormatedOutput(writer.getValue(new Text("5")).get(i)),
getFormatedOutput(RAW_DATA[i]));
}
}
@Test
public void testAffinityMatrixInputMapper() throws Exception {
AffinityMatrixInputMapper mapper = new AffinityMatrixInputMapper();
Configuration conf = getConfiguration();
conf.setInt(Keys.AFFINITY_DIMENSIONS, RAW_DIMENSIONS);
// set up the dummy writer and the M/R context
DummyRecordWriter<IntWritable, MatrixEntryWritable> writer = new DummyRecordWriter<>();
Mapper<LongWritable, Text, IntWritable, MatrixEntryWritable>.Context context =
DummyRecordWriter.build(mapper, conf, writer);
// loop through all the points and test each one is converted
// successfully to a DistributedRowMatrix.MatrixEntry
for (String s : RAW) {
mapper.map(new LongWritable(), new Text(s), context);
}
// test the data was successfully constructed
assertEquals("Number of map results", RAW_DIMENSIONS, writer.getData().size());
Set<IntWritable> keys = writer.getData().keySet();
for (IntWritable i : keys) {
List<MatrixEntryWritable> row = writer.getData().get(i);
assertEquals("Number of items in row", RAW_DIMENSIONS, row.size());
}
}
@Test
public void testAffinitymatrixInputReducer() throws Exception {
AffinityMatrixInputMapper mapper = new AffinityMatrixInputMapper();
Configuration conf = getConfiguration();
conf.setInt(Keys.AFFINITY_DIMENSIONS, RAW_DIMENSIONS);
// set up the dummy writer and the M/R context
DummyRecordWriter<IntWritable, MatrixEntryWritable> mapWriter = new DummyRecordWriter<>();
Mapper<LongWritable, Text, IntWritable, MatrixEntryWritable>.Context mapContext =
DummyRecordWriter.build(mapper, conf, mapWriter);
// loop through all the points and test each one is converted
// successfully to a DistributedRowMatrix.MatrixEntry
for (String s : RAW) {
mapper.map(new LongWritable(), new Text(s), mapContext);
}
// store the data for checking later
Map<IntWritable, List<MatrixEntryWritable>> map = mapWriter.getData();
// now reduce the data
AffinityMatrixInputReducer reducer = new AffinityMatrixInputReducer();
DummyRecordWriter<IntWritable, VectorWritable> redWriter = new DummyRecordWriter<>();
Reducer<IntWritable, MatrixEntryWritable, IntWritable, VectorWritable>.Context redContext =
DummyRecordWriter.build(
reducer, conf, redWriter, IntWritable.class, MatrixEntryWritable.class);
for (IntWritable key : mapWriter.getKeys()) {
reducer.reduce(key, mapWriter.getValue(key), redContext);
}
// check that all the elements are correctly ordered
assertEquals("Number of reduce results", RAW_DIMENSIONS, redWriter.getData().size());
for (IntWritable row : redWriter.getKeys()) {
List<VectorWritable> list = redWriter.getValue(row);
assertEquals("Should only be one vector", 1, list.size());
// check that the elements in the array are correctly ordered
Vector v = list.get(0).get();
for (Vector.Element e : v.all()) {
// find this value in the original map
MatrixEntryWritable toCompare = new MatrixEntryWritable();
toCompare.setRow(-1);
toCompare.setCol(e.index());
toCompare.setVal(e.get());
assertTrue("This entry was correctly placed in its row", map.get(row).contains(toCompare));
}
}
}
@Test
public void testMatrixDiagonalizeReducer() throws Exception {
MatrixDiagonalizeMapper mapper = new MatrixDiagonalizeMapper();
Configuration conf = getConfiguration();
conf.setInt(Keys.AFFINITY_DIMENSIONS, RAW_DIMENSIONS);
// set up the dummy writers
DummyRecordWriter<NullWritable, IntDoublePairWritable> mapWriter = new DummyRecordWriter<>();
Mapper<IntWritable, VectorWritable, NullWritable, IntDoublePairWritable>.Context mapContext =
DummyRecordWriter.build(mapper, conf, mapWriter);
// perform the mapping
for (int i = 0; i < RAW_DIMENSIONS; i++) {
RandomAccessSparseVector toAdd = new RandomAccessSparseVector(RAW_DIMENSIONS);
toAdd.assign(RAW[i]);
mapper.map(new IntWritable(i), new VectorWritable(toAdd), mapContext);
}
// now perform the reduction
MatrixDiagonalizeReducer reducer = new MatrixDiagonalizeReducer();
DummyRecordWriter<NullWritable, VectorWritable> redWriter = new DummyRecordWriter<>();
Reducer<NullWritable, IntDoublePairWritable, NullWritable, VectorWritable>.Context redContext =
DummyRecordWriter.build(
reducer, conf, redWriter, NullWritable.class, IntDoublePairWritable.class);
// only need one reduction
reducer.reduce(NullWritable.get(), mapWriter.getValue(NullWritable.get()), redContext);
// first, make sure there's only one result
List<VectorWritable> list = redWriter.getValue(NullWritable.get());
assertEquals("Only a single resulting vector", 1, list.size());
Vector v = list.get(0).get();
for (int i = 0; i < v.size(); i++) {
assertEquals("Element sum is correct", rowSum(RAW[i]), v.get(i), 0.01);
}
}
/**
* Testing the mapper is fairly straightforward: there are two matrices to be processed
* simultaneously (cut matrix of sensitivities, and the affinity matrix), and since both are
* symmetric, two entries from each will be grouped together with the same key (or, in the case of
* an entry along the diagonal, only two entries).
*
* <p>The correct grouping of these quad or pair vertices is the only output of the mapper.
*
* @throws Exception
*/
@Test
public void testEigencutsAffinityCutsMapper() throws Exception {
EigencutsAffinityCutsMapper mapper = new EigencutsAffinityCutsMapper();
Configuration conf = new Configuration();
conf.setInt(EigencutsKeys.AFFINITY_DIMENSIONS, this.affinity.length);
// set up the writer
DummyRecordWriter<Text, VertexWritable> writer = new DummyRecordWriter<Text, VertexWritable>();
Mapper<IntWritable, VectorWritable, Text, VertexWritable>.Context context =
DummyRecordWriter.build(mapper, conf, writer);
// perform the maps
for (int i = 0; i < this.affinity.length; i++) {
VectorWritable aff = new VectorWritable(new DenseVector(this.affinity[i]));
VectorWritable sens = new VectorWritable(new DenseVector(this.sensitivity[i]));
IntWritable key = new IntWritable(i);
mapper.map(key, aff, context);
mapper.map(key, sens, context);
}
// were the vertices constructed correctly? if so, then for two 4x4
// matrices, there should be 10 unique keys with 56 total entries
assertEquals("Number of keys", 10, writer.getKeys().size());
for (int i = 0; i < this.affinity.length; i++) {
for (int j = 0; j < this.affinity.length; j++) {
Text key = new Text(Math.max(i, j) + "_" + Math.min(i, j));
List<VertexWritable> values = writer.getValue(key);
// if we're on a diagonal, there should only be 2 entries
// otherwise, there should be 4
if (i == j) {
assertEquals("Diagonal entry", 2, values.size());
for (VertexWritable v : values) {
assertFalse("Diagonal values are zero", v.getValue() > 0);
}
} else {
assertEquals("Off-diagonal entry", 4, values.size());
if (i + j == 3) { // all have values greater than 0
for (VertexWritable v : values) {
assertTrue("Off-diagonal non-zero entries", v.getValue() > 0);
}
}
}
}
}
}
@Test
public void testMatrixDiagonalizeMapper() throws Exception {
MatrixDiagonalizeMapper mapper = new MatrixDiagonalizeMapper();
Configuration conf = getConfiguration();
conf.setInt(Keys.AFFINITY_DIMENSIONS, RAW_DIMENSIONS);
// set up the dummy writers
DummyRecordWriter<NullWritable, IntDoublePairWritable> writer = new DummyRecordWriter<>();
Mapper<IntWritable, VectorWritable, NullWritable, IntDoublePairWritable>.Context context =
DummyRecordWriter.build(mapper, conf, writer);
// perform the mapping
for (int i = 0; i < RAW_DIMENSIONS; i++) {
RandomAccessSparseVector toAdd = new RandomAccessSparseVector(RAW_DIMENSIONS);
toAdd.assign(RAW[i]);
mapper.map(new IntWritable(i), new VectorWritable(toAdd), context);
}
// check the number of the results
assertEquals(
"Number of map results", RAW_DIMENSIONS, writer.getValue(NullWritable.get()).size());
}
/**
* Fairly straightforward: the task here is to reassemble the rows of the affinity matrix. The
* tricky part is that any specific element in the list of elements which does NOT lay on the
* diagonal will be so because it did not drop below the sensitivity threshold, hence it was not
* "cut".
*
* <p>On the flip side, there will be many entries whose coordinate is now set to the diagonal,
* indicating they were previously affinity entries whose sensitivities were below the threshold,
* and hence were "cut" - set to 0 at their original coordinates, and had their values added to
* the diagonal entry (hence the numerous entries with the coordinate of the diagonal).
*
* @throws Exception
*/
@Test
public void testEigencutsAffinityCutsReducer() throws Exception {
Configuration conf = new Configuration();
Path affinity = new Path("affinity");
Path sensitivity = new Path("sensitivity");
conf.set(EigencutsKeys.AFFINITY_PATH, affinity.getName());
conf.setInt(EigencutsKeys.AFFINITY_DIMENSIONS, this.affinity.length);
// since we need the working paths to distinguish the vertex types,
// we can't use the mapper (since we have no way of manually setting
// the Context.workingPath() )
Map<Text, List<VertexWritable>> data = buildMapData(affinity, sensitivity, this.sensitivity);
// now, set up the combiner
EigencutsAffinityCutsCombiner combiner = new EigencutsAffinityCutsCombiner();
DummyRecordWriter<Text, VertexWritable> comWriter =
new DummyRecordWriter<Text, VertexWritable>();
Reducer<Text, VertexWritable, Text, VertexWritable>.Context comContext =
DummyRecordWriter.build(combiner, conf, comWriter, Text.class, VertexWritable.class);
// perform the combining
for (Map.Entry<Text, List<VertexWritable>> entry : data.entrySet()) {
combiner.reduce(entry.getKey(), entry.getValue(), comContext);
}
// finally, set up the reduction writers
EigencutsAffinityCutsReducer reducer = new EigencutsAffinityCutsReducer();
DummyRecordWriter<IntWritable, VectorWritable> redWriter =
new DummyRecordWriter<IntWritable, VectorWritable>();
Reducer<Text, VertexWritable, IntWritable, VectorWritable>.Context redContext =
DummyRecordWriter.build(reducer, conf, redWriter, Text.class, VertexWritable.class);
// perform the reduction
for (Text key : comWriter.getKeys()) {
reducer.reduce(key, comWriter.getValue(key), redContext);
}
// now, check that the affinity matrix is correctly formed
for (IntWritable row : redWriter.getKeys()) {
List<VectorWritable> results = redWriter.getValue(row);
// there should only be 1 vector
assertEquals("Only one vector with a given row number", 1, results.size());
Vector therow = results.get(0).get();
for (Vector.Element e : therow.all()) {
// check the diagonal
if (row.get() == e.index()) {
assertEquals(
"Correct diagonal sum of cuts",
sumOfRowCuts(row.get(), this.sensitivity),
e.get(),
EPSILON);
} else {
// not on the diagonal...if it was an element labeled to be cut,
// it should have a value of 0. Otherwise, it should have kept its
// previous value
if (this.sensitivity[row.get()][e.index()] == 0.0) {
// should be what it was originally
assertEquals(
"Preserved element", this.affinity[row.get()][e.index()], e.get(), EPSILON);
} else {
// should be 0
assertEquals("Cut element", 0.0, e.get(), EPSILON);
}
}
}
}
}
/**
* This is by far the trickiest step. However, an easy condition is if we have only two vertices -
* indicating vertices on the diagonal of the two matrices - then we simply exit (since the
* algorithm does not operate on the diagonal; it makes no sense to perform cuts by isolating data
* points from themselves).
*
* <p>If there are four points, then first we must separate the two which belong to the affinity
* matrix from the two that are sensitivities. In theory, each pair should have exactly the same
* value (symmetry). If the sensitivity is below a certain threshold, then we set the two values
* of the affinity matrix to 0 (but not before adding the affinity values to the diagonal, so as
* to maintain the overall sum of the row of the affinity matrix).
*
* @throws Exception
*/
@Test
public void testEigencutsAffinityCutsCombiner() throws Exception {
Configuration conf = new Configuration();
Path affinity = new Path("affinity");
Path sensitivity = new Path("sensitivity");
conf.set(EigencutsKeys.AFFINITY_PATH, affinity.getName());
conf.setInt(EigencutsKeys.AFFINITY_DIMENSIONS, this.affinity.length);
// since we need the working paths to distinguish the vertex types,
// we can't use the mapper (since we have no way of manually setting
// the Context.workingPath() )
Map<Text, List<VertexWritable>> data = buildMapData(affinity, sensitivity, this.sensitivity);
// now, set up the combiner
EigencutsAffinityCutsCombiner combiner = new EigencutsAffinityCutsCombiner();
DummyRecordWriter<Text, VertexWritable> redWriter =
new DummyRecordWriter<Text, VertexWritable>();
Reducer<Text, VertexWritable, Text, VertexWritable>.Context redContext =
DummyRecordWriter.build(combiner, conf, redWriter, Text.class, VertexWritable.class);
// perform the combining
for (Map.Entry<Text, List<VertexWritable>> entry : data.entrySet()) {
combiner.reduce(entry.getKey(), entry.getValue(), redContext);
}
// test the number of cuts, there should be 2
assertEquals(
"Number of cuts detected",
4,
redContext.getCounter(EigencutsAffinityCutsJob.CUTSCOUNTER.NUM_CUTS).getValue());
// loop through all the results; let's see if they match up to our
// affinity matrix (and all the cuts appear where they should
Map<Text, List<VertexWritable>> results = redWriter.getData();
for (Map.Entry<Text, List<VertexWritable>> entry : results.entrySet()) {
List<VertexWritable> row = entry.getValue();
IntWritable key = new IntWritable(Integer.parseInt(entry.getKey().toString()));
double calcDiag = 0.0;
double trueDiag = sumOfRowCuts(key.get(), this.sensitivity);
for (VertexWritable e : row) {
// should the value have been cut, e.g. set to 0?
if (key.get() == e.getCol()) {
// we have our diagonal
calcDiag += e.getValue();
} else if (this.sensitivity[key.get()][e.getCol()] == 0.0) {
// no, corresponding affinity should have same value as before
assertEquals(
"Preserved affinity value",
this.affinity[key.get()][e.getCol()],
e.getValue(),
EPSILON);
} else {
// yes, corresponding affinity value should be 0
assertEquals("Cut affinity value", 0.0, e.getValue(), EPSILON);
}
}
// check the diagonal has the correct sum
assertEquals("Diagonal sum from cuts", trueDiag, calcDiag, EPSILON);
}
}