/**
* Portions Copyright 2006 DFKI GmbH.
* Portions Copyright 2001 Sun Microsystems, Inc.
* Portions Copyright 1999-2001 Language Technologies Institute,
* Carnegie Mellon University.
* All Rights Reserved. Use is subject to license terms.
*
* Permission is hereby granted, free of charge, to use and distribute
* this software and its documentation without restriction, including
* without limitation the rights to use, copy, modify, merge, publish,
* distribute, sublicense, and/or sell copies of this work, and to
* permit persons to whom this work is furnished to do so, subject to
* the following conditions:
*
* 1. The code must retain the above copyright notice, this list of
* conditions and the following disclaimer.
* 2. Any modifications must be clearly marked as such.
* 3. Original authors' names are not deleted.
* 4. The authors' names are not used to endorse or promote products
* derived from this software without specific prior written
* permission.
*
* DFKI GMBH AND THE CONTRIBUTORS TO THIS WORK DISCLAIM ALL WARRANTIES WITH
* REGARD TO THIS SOFTWARE, INCLUDING ALL IMPLIED WARRANTIES OF
* MERCHANTABILITY AND FITNESS, IN NO EVENT SHALL DFKI GMBH NOR THE
* CONTRIBUTORS BE LIABLE FOR ANY SPECIAL, INDIRECT OR CONSEQUENTIAL
* DAMAGES OR ANY DAMAGES WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR
* PROFITS, WHETHER IN AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS
* ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF
* THIS SOFTWARE.
*/
package marytts.tools.voiceimport;
import java.io.BufferedReader;
import java.io.File;
import java.io.FileInputStream;
import java.io.FileNotFoundException;
import java.io.FileReader;
import java.io.FileWriter;
import java.io.IOException;
import java.io.PrintWriter;
import java.io.StringReader;
import java.util.HashMap;
import java.util.HashSet;
import java.util.Map;
import java.util.Scanner;
import java.util.Set;
import java.util.SortedMap;
import java.util.TreeMap;
import java.util.Vector;
import marytts.cart.CART;
import marytts.cart.Node;
import marytts.cart.LeafNode.PdfLeafNode;
import marytts.cart.io.HTSCARTReader;
import marytts.features.FeatureDefinition;
import marytts.features.FeatureVector;
import marytts.htsengine.HMMData;
import marytts.htsengine.PhoneTranslator;
import marytts.htsengine.HMMData.PdfFileFormat;
import marytts.modules.phonemiser.AllophoneSet;
import marytts.unitselection.data.FeatureFileReader;
import marytts.unitselection.data.UnitFileReader;
import marytts.unitselection.select.JoinCostFeatures;
import marytts.util.io.General;
import marytts.util.math.MathUtils;
public class JoinModeller extends VoiceImportComponent {
private DatabaseLayout db = null;
private int percent = 0;
private HMMData htsData = null; /* for using the function readFeatureList() */
private Vector<String> featureList = null;
private Map<String, String> feat2shortFeat = new HashMap<String, String>();
private int numberOfFeatures = 0;
private float[] fw = null;
private String[] wfun = null;
FileWriter statsStream = null;
FileWriter mmfStream = null;
FileWriter fullStream = null;
public final String JOINCOSTFEATURESFILE = "JoinModeller.joinCostFeaturesFile";
public final String UNITFEATURESFILE = "JoinModeller.unitFeaturesFile";
public final String UNITFILE = "JoinModeller.unitFile";
public final String STATSFILE = "JoinModeller.statsFile";
public final String MMFFILE = "JoinModeller.mmfFile";
public final String FULLFILE = "JoinModeller.fullFile";
public final String CXCHEDFILE = "JoinModeller.cxcJoinFile";
public final String JOINTREEFILE = "JoinModeller.joinTreeFile";
public final String CNVHEDFILE = "JoinModeller.cnvJoinFile";
public final String TRNCONFFILE = "JoinModeller.trnFile";
public final String CNVCONFFILE = "JoinModeller.cnvFile";
public final String HHEDCOMMAND = "JoinModeller.hhedCommand";
public final String FEATURELISTFILE = "JoinModeller.featureListFile";
public final String TRICKYPHONESFILE = "JoinModeller.trickyPhonesFile";
public JoinModeller() {
}
public String getName() {
return "JoinModeller";
}
public SortedMap<String, String> getDefaultProps(DatabaseLayout db) {
this.db = db;
if (props == null) {
props = new TreeMap<String, String>();
String filedir = db.getProp(db.FILEDIR);
String maryExt = db.getProp(db.MARYEXT);
props.put(JOINCOSTFEATURESFILE, filedir + "joinCostFeatures" + maryExt);
props.put(UNITFEATURESFILE, filedir + "halfphoneFeatures" + maryExt);
props.put(UNITFILE, filedir + "halfphoneUnits" + maryExt);
props.put(STATSFILE, filedir + "stats" + maryExt);
props.put(MMFFILE, filedir + "join_mmf" + maryExt);
props.put(FULLFILE, filedir + "fullList" + maryExt);
props.put(FEATURELISTFILE, filedir + "/mary/featureListFile.txt");
props.put(TRICKYPHONESFILE, filedir + "/mary/trickyPhones.txt");
props.put(CXCHEDFILE, filedir + "cxc_join.hed");
props.put(JOINTREEFILE, filedir + "join_tree.inf");
props.put(CNVHEDFILE, filedir + "cnv_join.hed");
props.put(TRNCONFFILE, filedir + "trn.cnf");
props.put(CNVCONFFILE, filedir + "cnv.cnf");
props.put(HHEDCOMMAND, "/project/mary/marcela/sw/HTS_2.0.1/htk/bin/HHEd");
}
return props;
}
protected void setupHelp() {
props2Help = new TreeMap<String, String>();
props2Help.put(JOINCOSTFEATURESFILE, "file containing all halfphone units and their join cost features");
props2Help.put(UNITFEATURESFILE, "file containing all halfphone units and their target cost features");
props2Help.put(UNITFILE, "file containing all halfphone units");
props2Help.put(STATSFILE, "output file containing statistics of the models in HTK stats format");
props2Help
.put(MMFFILE,
"output file containing one state HMM models, HTK format, representing join models (mean and variances are calculated in this class)");
props2Help.put(FULLFILE, "output file containing the full list of HMM model names");
props2Help.put(FEATURELISTFILE, "feature list for making fullcontext names and questions");
props2Help.put(TRICKYPHONESFILE, "list of aliases for tricky phones, so HTK-HHEd command can handle them.");
props2Help
.put(CXCHEDFILE,
"HTK hed file used by HHEd, load stats file, contains questions for decision tree-based context clustering and outputs result in join-tree.inf");
props2Help.put(CNVHEDFILE, "HTK hed file used by HHEd to convert trees and mmf into hts_engine format");
props2Help.put(TRNCONFFILE, "HTK configuration file for context clustering");
props2Help.put(CNVCONFFILE, "HTK configuration file for converting to hts_engine format");
props2Help.put(HHEDCOMMAND, "HTS-HTK HHEd command, HTS version minimum HTS_2.0.1");
}
public boolean compute() throws IOException, Exception {
System.out.println("\n---- Training join models\n");
FeatureFileReader unitFeatures = FeatureFileReader.getFeatureFileReader(getProp(UNITFEATURESFILE));
JoinCostFeatures joinFeatures = new JoinCostFeatures(getProp(JOINCOSTFEATURESFILE));
UnitFileReader units = new UnitFileReader(getProp(UNITFILE));
FeatureDefinition def = unitFeatures.getFeatureDefinition();
// for creating a contexttranslator object we need to check if there is
PhoneTranslator contextTranslator = null;
// Check if there are tricky phones
if (HMMVoiceMakeData.checkTrickyPhones(db.getAllophoneSet(), getProp(TRICKYPHONESFILE)))
contextTranslator = new PhoneTranslator(new FileInputStream(getProp(TRICKYPHONESFILE)));
else
contextTranslator = new PhoneTranslator(null);
featureList = new Vector<String>();
readFeatureList(getProp(FEATURELISTFILE), def);
statsStream = new FileWriter(getProp(STATSFILE));
mmfStream = new FileWriter(getProp(MMFFILE));
fullStream = new FileWriter(getProp(FULLFILE));
// output HTK model definition and dummy state-transition matrix, macro ~t "trP_1"
int numFeatures = joinFeatures.getNumberOfFeatures();
mmfStream.write("~o\n" + "<VECSIZE> " + numFeatures + " <USER><DIAGC>\n" + "~t \"trP_1\"\n<TRANSP> 3\n"
+ "0 1 0\n0 0 1\n0 0 0\n");
if (unitFeatures.getNumberOfUnits() != joinFeatures.getNumberOfUnits())
throw new IllegalStateException("Number of units in unit and join feature files does not match!");
if (unitFeatures.getNumberOfUnits() != units.getNumberOfUnits())
throw new IllegalStateException("Number of units in unit file and unit feature file does not match!");
int numUnits = unitFeatures.getNumberOfUnits();
// ----FeatureDefinition def = unitFeatures.getFeatureDefinition();
int iPhoneme = def.getFeatureIndex("phone");
int nPhonemes = def.getNumberOfValues(iPhoneme);
int iLeftRight = def.getFeatureIndex("halfphone_lr");
byte vLeft = def.getFeatureValueAsByte(iLeftRight, "L");
byte vRight = def.getFeatureValueAsByte(iLeftRight, "R");
int iEdge = def.getFeatureIndex("edge");
byte vNoEdge = def.getFeatureValueAsByte(iEdge, "0");
byte vStartEdge = def.getFeatureValueAsByte(iEdge, "start");
byte vEndEdge = def.getFeatureValueAsByte(iEdge, "end");
Map<String, Set<double[]>> uniqueFeatureVectors = new HashMap<String, Set<double[]>>();
// Now look at all pairs of adjacent units
FeatureVector fvNext = unitFeatures.getFeatureVector(0);
for (int i = 0; i < numUnits - 1; i++) {
percent = 100 * (i + 1) / numUnits;
FeatureVector fv = fvNext;
byte edge = fv.getByteFeature(iEdge);
fvNext = unitFeatures.getFeatureVector(i + 1);
byte edgeNext = fvNext.getByteFeature(iEdge);
if (edge == vNoEdge && edgeNext == vNoEdge) {
// TODO: do we need this?
int phone = fv.getFeatureAsInt(iPhoneme);
assert 0 <= phone && phone < nPhonemes;
byte lr = fv.getByteFeature(iLeftRight);
int phoneNext = fvNext.getFeatureAsInt(iPhoneme);
byte lrNext = fvNext.getByteFeature(iLeftRight);
String pair = def.getFeatureValueAsString(iPhoneme, phone) + "_" + def.getFeatureValueAsString(iLeftRight, lr)
+ "-" + def.getFeatureValueAsString(iPhoneme, phoneNext) + "_"
+ def.getFeatureValueAsString(iLeftRight, lrNext);
// Compute the difference vector
float[] myRightFrame = joinFeatures.getRightJCF(i);
float[] nextLeftFrame = joinFeatures.getLeftJCF(i + 1);
double[] difference = new double[myRightFrame.length];
for (int k = 0, len = myRightFrame.length; k < len; k++) {
if (k == (len - 1)
&& (myRightFrame[k] == Float.POSITIVE_INFINITY || nextLeftFrame[k] == Float.POSITIVE_INFINITY)) {
// TODO: find out in JoinCostFeatureFileWriter why Infinity values occur, and fix it.
difference[k] = 0.0;
System.out.println("WARNING: numUnit=" + i + " myRightFrame[k]=" + myRightFrame[k] + " nextLeftFrame[k]="
+ nextLeftFrame[k]);
} else
difference[k] = ((double) myRightFrame[k]) - nextLeftFrame[k];
}
// Group the units with the same feature vectors
String contextName = contextTranslator.features2LongContext(def, fv, featureList);
Set<double[]> unitsWithFV = uniqueFeatureVectors.get(contextName);
if (unitsWithFV == null) {
unitsWithFV = new HashSet<double[]>();
uniqueFeatureVectors.put(contextName, unitsWithFV);
}
unitsWithFV.add(difference);
}
}
int numUniqueFea = 1;
int n, i, j, m;
for (String fvString : uniqueFeatureVectors.keySet()) {
double[][] diffVectors = uniqueFeatureVectors.get(fvString).toArray(new double[0][]);
n = diffVectors.length;
// Compute means and variances of the features across difference vectors
double[] means;
double[] variances;
if (n == 1) {
means = diffVectors[0];
variances = MathUtils.zeros(means.length);
} else {
means = MathUtils.mean(diffVectors, true);
variances = MathUtils.variance(diffVectors, means, true);
}
assert means.length == numFeatures : "expected to have " + numFeatures + " features, got " + means.length;
fullStream.write(fvString + "\n");
statsStream.write(numUniqueFea + " \"" + fvString + "\" " + n + " " + n + "\n");
mmfStream.write("~h " + "\"" + fvString + "\"\n");
/* Stream 1 */
mmfStream.write("<BEGINHMM>\n<NUMSTATES> 3\n<STATE> 2\n");
mmfStream.write("<MEAN> " + numFeatures + "\n");
for (i = 0; i < means.length; i++)
mmfStream.write(means[i] + " ");
mmfStream.write("\n<VARIANCE> " + numFeatures + "\n");
for (i = 0; i < variances.length; i++)
mmfStream.write(variances[i] + " ");
mmfStream.write("\n~t \"trP_1\"\n<ENDHMM>\n");
numUniqueFea++;
}
fullStream.close();
statsStream.close();
mmfStream.close();
Process proc = null;
String cmdLine = null;
BufferedReader procStdout = null;
String line = null;
String filedir = db.getProp(db.FILEDIR);
System.out.println(uniqueFeatureVectors.keySet().size() + " unique feature vectors, " + numUnits + " units");
System.out.println("Generated files: " + getProp(STATSFILE) + " " + getProp(MMFFILE) + " " + getProp(FULLFILE));
System.out.println("\n---- Creating tree clustering command file for HHEd\n");
PrintWriter pw = new PrintWriter(new File(getProp(CXCHEDFILE)));
pw.println("// load stats file");
pw.println("RO 000 \"" + getProp(STATSFILE) + "\"");
pw.println();
pw.println("TR 0");
pw.println();
pw.println("// questions for decision tree-based context clustering");
for (String f : featureList) {
int featureIndex = def.getFeatureIndex(f);
String[] values = def.getPossibleValues(featureIndex);
for (String v : values) {
if (f.endsWith("phone")) {
v = contextTranslator.replaceTrickyPhones(v);
} else if (f.endsWith("sentence_punc") || f.endsWith("punctuation")) {
v = contextTranslator.replacePunc(v);
}
pw.println("QS \"" + f + "=" + v + "\" {*|" + f + "=" + v + "|*}");
}
pw.println();
}
pw.println("TR 3");
pw.println();
pw.println("// construct decision trees");
pw.println("TB 000 join_ {*.state[2]}");
pw.println();
pw.println("TR 1");
pw.println();
pw.println("// output constructed tree");
pw.println("ST \"" + getProp(JOINTREEFILE) + "\"");
pw.close();
System.out.println("\n---- Tree-based context clustering for joinModeller\n");
// here the input file is join_mmf.mry and the output is join_mmf.mry.clustered
cmdLine = getProp(HHEDCOMMAND) + " -A -C " + getProp(TRNCONFFILE) + " -D -T 2 -p -i -H " + getProp(MMFFILE)
+ " -m -a 1.0 -w " + getProp(MMFFILE) + ".clustered" + " " + getProp(CXCHEDFILE) + " " + getProp(FULLFILE);
General.launchProc(cmdLine, "HHEd", filedir);
System.out.println("\n---- Creating conversion-to-hts command file for HHEd\n");
pw = new PrintWriter(new File(getProp(CNVHEDFILE)));
pw.println("TR 3");
pw.println();
pw.println("// load trees for joinModeller");
pw.println("LT \"" + getProp(JOINTREEFILE) + "\"");
pw.println();
pw.println("// convert loaded trees for hts_engine format");
pw.println("CT \"" + filedir + "\"");
pw.println();
pw.println("// convert mmf for hts_engine format");
pw.println("CM \"" + filedir + "\"");
pw.close();
System.out.println("\n---- Converting mmfs to the hts_engine file format\n");
// the input of this command are: join_mmf.mry and join_tree.inf and the output: trees.1 and pdf.1
cmdLine = getProp(HHEDCOMMAND) + " -A -C " + getProp(CNVCONFFILE) + " -D -T 1 -p -i -H " + getProp(MMFFILE)
+ ".clustered" + " " + getProp(CNVHEDFILE) + " " + getProp(FULLFILE);
General.launchProc(cmdLine, "HHEd", filedir);
// the files trees.1 and pdf.1 produced by the previous command are renamed as tree-joinModeller.inf and joinModeller.pdf
cmdLine = "mv " + filedir + "trees.1 " + filedir + "tree-joinModeller.inf";
General.launchProc(cmdLine, "mv", filedir);
cmdLine = "mv " + filedir + "pdf.1 " + filedir + "joinModeller.pdf";
General.launchProc(cmdLine, "mv", filedir);
System.out.println("\n---- Created files: tree-joinModeller.inf, joinModeller.pdf");
return true;
}
/**
* Provide the progress of computation, in percent, or -1 if that feature is not implemented.
*
* @return -1 if not implemented, or an integer between 0 and 100.
*/
public int getProgress() {
return percent;
}
/**
* This function reads a feature list file originally used to train the HMMs, in HMM voices. The file format is simple, one
* feature after another like in the ../mary/features.txt An example of feature list is in ../mary/featuresHmmVoice.txt Since
* the features are provided by the user, it should be checked that the features exist
*
* @param featureListFile
* featureListFile
* @param feaDef
* feaDef
* @throws Exception
* Exception
*/
private void readFeatureList(String featureListFile, FeatureDefinition feaDef) throws Exception {
String line;
int i;
// System.out.println("featureListFile: " + featureListFile);
Scanner s = null;
try {
s = new Scanner(new BufferedReader(new FileReader(featureListFile))).useDelimiter("\n");
String fea;
System.out.println("The following are other context features used for training Hmms: ");
while (s.hasNext()) {
fea = s.nextLine();
// Check if the feature exist
if (feaDef.hasFeature(fea)) {
featureList.add(fea);
// System.out.println(" " + fea);
} else
throw new Exception("Error: feature \"" + fea + "\" in feature list file: " + featureListFile
+ " does not exist in FeatureDefinition.");
}
if (s != null) {
s.close();
}
} catch (FileNotFoundException e) {
System.out.println("readFeatureList: " + e.getMessage());
}
System.out.println("readFeatureList: loaded " + featureList.size() + " context features from " + featureListFile);
} /* method ReadFeatureList */
public static void main(String[] args) throws IOException, InterruptedException {
try {
/* configure log info */
org.apache.log4j.BasicConfigurator.configure();
/* These trees are for halphone features */
String joinPdfFile = "/project/mary/marcela/HMM-voices/DFKI_German_Poker/mary_files_old/joinModeller.pdf";
String joinTreeFile = "/project/mary/marcela/HMM-voices/DFKI_German_Poker/mary_files_old/tree-joinModeller.inf";
// String contextFile = "/project/mary/marcela/HMM-voices/DFKI_German_Poker/phonefeatures/m0001.pfeats";
/* halphone example */
String contextFile = "/project/mary/marcela/unitselection-halfphone.pfeats";
Scanner context = new Scanner(new BufferedReader(new FileReader(contextFile)));
String strContext = "";
while (context.hasNext()) {
strContext += context.nextLine();
strContext += "\n";
}
context.close();
// System.out.println(strContext);
FeatureDefinition def = new FeatureDefinition(new BufferedReader(new StringReader(strContext)), false);
String trickyPhonesFile = "/project/mary/marcela/HMM-voices/DFKI_German_Poker/mary_files_old/trickyPhones.txt";
String allophonesFile = "/project/mary/marcela/openmary/lib/modules/en/us/lexicon/allophones.en_US.xml";
// Check if there are tricky phones
if (!(HMMVoiceMakeData.checkTrickyPhones(AllophoneSet.getAllophoneSet(allophonesFile), trickyPhonesFile)))
trickyPhonesFile = null;
// Check if there are tricky phones, and create a PhoneTranslator object
PhoneTranslator phTranslator = new PhoneTranslator(new FileInputStream(trickyPhonesFile));
CART[] joinTree = null;
HTSCARTReader htsReader = new HTSCARTReader();
int numStates = 1; // just one state in the joinModeller
int vectorSize = 0;
try {
joinTree = htsReader.load(numStates, new FileInputStream(joinTreeFile), new FileInputStream(joinPdfFile),
PdfFileFormat.join, def, phTranslator);
vectorSize = htsReader.getVectorSize();
} catch (Exception e) {
IOException ioe = new IOException("Cannot load join model trees from " + joinTreeFile);
ioe.initCause(e);
throw ioe;
}
/* Given a context feature model name, find its join PDF mean and variance */
/* first, find an index in the tree and then find the mean and variance that corresponds to that index in joinPdf */
double[] mean = new double[vectorSize];
double[] variance = new double[vectorSize];
// String modelName =
// "0^f-t+_=0||str=0|pos_syl=3|pos_type=final|pos=NN|snt_punc=.|snt_numwrds=7|wrds_snt_stt=2|wrds_snt_end=4|wrd_numsyls=3|syls_wrd_stt=2|syls_wrd_end=0|wrd_numsegs=9|segs_wrd_stt=8|segs_wrd_end=0|syl_numsegs=4|segs_syl_stt=3|segs_syl_end=0|syls_p_str=2|syls_n_str=1|p_punc=0|n_punc=.|wrds_p_punc=2|wrds_n_punc=4|wrd_freq=0|";
// String modelName =
// "0^i:-o:+_=0||str=0|pos_syl=0|pos_type=final|pos=FM|snt_punc=.|snt_numwrds=17|wrds_snt_stt=16|wrds_snt_end=0|wrd_numsyls=4|syls_wrd_stt=3|syls_wrd_end=0|wrd_numsegs=7|segs_wrd_stt=6|segs_wrd_end=0|syl_numsegs=1|segs_syl_stt=0|segs_syl_end=0|syls_p_str=2|syls_n_str=0|p_punc=:|n_punc=.|wrds_p_punc=4|wrds_n_punc=0|wrd_freq=0|";
// String modelName =
// "aI^R-a:+t=@||wrds_n_punc=0|syl_numsegs=2|segs_wrd_stt=5|syls_wrd_stt=2|str=0|wrds_snt_end=0|pos_syl=1|p_punc=0|wrd_freq=3|snt_punc=dt|segs_wrd_end=3|wrds_snt_stt=4|wrds_p_punc=4|n_punc=dt|segs_syl_end=0|syls_wrd_end=1|pos_type=mid|syls_n_str=0|POS=VVPP|snt_numwrds=5|wrd_numsyls=4|wrd_numsegs=9|syls_p_str=1|segs_syl_stt=1||";
// String feaLine =
// "24 0 1 0 0 1 2 3 1 0 30 2 2 1 1 0 0 0 0 43 2 2 4 1 0 0 0 0 1 1 0 1 3 0 0 2 2 0 0 0 0 0 1 1 1 0 0 5 1 3 5 1 3 10 1 9 1 3 1 1 3 0 2 6 1 4 2 1 0 1 3 1 7 9 0 1 6 0 1 1 3 0";
// String feaLine =
// "23 0 2 0 0 2 0 0 0 0 0 0 0 0 0 0 0 0 0 34 2 4 2 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0";
/* halfphone example */
String feaLine = "8 3 2 0 1 0 6 1 0 0 4 2 38 1 0 0 1 1 1 2 2 35 2 3 1 1 0 0 0 0 1 2 1 0 1 0 0 13 0 0 0 0 0 0 0 0 1 0 0 6 0 5 6 0 5 9 0 9 0 5 0 1 1 0 0 3 2 0 3 2 0 0 1 0 7 9 0 2 0 0 0 0 6 8";
FeatureVector fv;
context = new Scanner(new BufferedReader(new FileReader(contextFile)));
while (context.hasNext()) {
feaLine = context.nextLine();
if (feaLine.trim().equals(""))
break;
}
while (context.hasNext()) {
feaLine = context.nextLine();
if (feaLine.trim().equals(""))
break;
}
while (context.hasNext()) {
feaLine = context.nextLine();
// System.out.println(feaLine);
fv = def.toFeatureVector(0, feaLine);
Node node = joinTree[0].interpretToNode(fv, 1);
assert node instanceof PdfLeafNode : "The node must be a PdfLeafNode.";
mean = ((PdfLeafNode) node).getMean();
variance = ((PdfLeafNode) node).getVariance();
System.out.print("mean: ");
for (int i = 0; i < vectorSize; i++)
System.out.print(mean[i] + " ");
System.out.print("\nvariance: ");
for (int i = 0; i < vectorSize; i++)
System.out.print(variance[i] + " ");
System.out.println();
}
} catch (Exception e) {
System.err.println("Exception: " + e.getMessage());
}
}
}