edu.cmu.tetrad.bayes.BayesPm Maven / Gradle / Ivy
///////////////////////////////////////////////////////////////////////////////
// For information as to what this class does, see the Javadoc, below. //
// Copyright (C) 1998, 1999, 2000, 2001, 2002, 2003, 2004, 2005, 2006, //
// 2007, 2008, 2009, 2010, 2014, 2015, 2022 by Peter Spirtes, Richard //
// Scheines, Joseph Ramsey, and Clark Glymour. //
// //
// This program is free software; you can redistribute it and/or modify //
// it under the terms of the GNU General Public License as published by //
// the Free Software Foundation; either version 2 of the License, or //
// (at your option) any later version. //
// //
// This program is distributed in the hope that it will be useful, //
// but WITHOUT ANY WARRANTY; without even the implied warranty of //
// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the //
// GNU General Public License for more details. //
// //
// You should have received a copy of the GNU General Public License //
// along with this program; if not, write to the Free Software //
// Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA //
///////////////////////////////////////////////////////////////////////////////
package edu.cmu.tetrad.bayes;
import edu.cmu.tetrad.data.DataUtils;
import edu.cmu.tetrad.data.DiscreteVariable;
import edu.cmu.tetrad.data.VariableSource;
import edu.cmu.tetrad.graph.*;
import edu.cmu.tetrad.util.Pm;
import edu.cmu.tetrad.util.RandomUtil;
import edu.cmu.tetrad.util.TetradLogger;
import org.apache.commons.math3.util.FastMath;
import java.io.IOException;
import java.io.ObjectInputStream;
import java.io.ObjectOutputStream;
import java.io.Serial;
import java.util.*;
/**
* Implements a discrete Bayes parametric model--that is, a DAG together with a map from the nodes in the graph to a set
* of discrete variables, specifying the number of categories for each variable and the name of each category for each
* variable. This is all the information one needs to know in order to determine the parametric form of a Bayes net up
* to actual values of parameters. Specific values for the Bayes net are stored in a BayesIM object (see).
*
* @author josephramsey
* @version $Id: $Id
* @see edu.cmu.tetrad.graph.Dag
* @see BayesIm
*/
public final class BayesPm implements Pm, VariableSource {
@Serial
private static final long serialVersionUID = 23L;
/**
* The underlying graph that's being parameterized.
*
* @serial Cannot be null.
*/
private final Graph dag;
/**
* The map from nodes to variables.
*
* @serial Cannot be null.
*/
private final Map nodesToVariables;
//=========================CONSTRUCTORS=============================//
/**
* Construct a new BayesPm using the given DAG, assigning each variable two values named "value1" and "value2"
* unless nodes are discrete variables with categories already defined.
*
* @param graph Ibid.
*/
public BayesPm(Graph graph) {
if (graph == null) {
throw new NullPointerException("The graph must not be null.");
}
this.dag = new EdgeListGraph(graph);
this.nodesToVariables = new HashMap<>();
boolean allDiscreteVars = true;
for (Node node : graph.getNodes()) {
if (!(node instanceof DiscreteVariable)) {
allDiscreteVars = false;
break;
}
}
if (!allDiscreteVars) {
initializeValues(2, 2);
} else {
for (Node node : this.dag.getNodes()) {
this.nodesToVariables.put(node, (DiscreteVariable) node);
}
}
}
/**
* Constructs a new BayesPm using a given DAG, using as much information from the old BayesPm as possible.
*
* @param graph Ibid.
* @param oldBayesPm Ibid.
*/
public BayesPm(Graph graph, BayesPm oldBayesPm) {
this(graph, oldBayesPm, 2, 2);
}
/**
* Constructs a new BayesPm from the given DAG, assigning each variable a random number of values between
* lowerBound and
* upperBound. Uses a fixed number of values if lowerBound ==
* upperBound. The values are named "value1" ... "valuen".
*
* @param graph Ibid.
* @param lowerBound Ibid.
* @param upperBound Ibid.
*/
public BayesPm(Graph graph, int lowerBound, int upperBound) {
if (graph == null) {
throw new NullPointerException("The graph must not be null.");
}
this.dag = new EdgeListGraph(graph);
this.nodesToVariables = new HashMap<>();
initializeValues(lowerBound, upperBound);
}
/**
* Constructs a new BayesPm from the given DAG, using as much information from the old BayesPm as possible. For
* variables not in the old BayesPm, assigns each variable a random number of values between
* lowerBound and upperBound. Uses a fixed number
* of values if lowerBound == upperBound. The values are named "value1" ... "valuen".
*
* @param graph Ibid.
* @param oldBayesPm Ibid.
* @param lowerBound Ibid.
* @param upperBound Ibid.
*/
public BayesPm(Graph graph, BayesPm oldBayesPm, int lowerBound,
int upperBound) {
// Should be OK wrt variable mismatch problems. jdramsey 2004/1/21
if (graph == null) {
throw new NullPointerException("The graph must not be null.");
}
if (oldBayesPm == null) {
throw new NullPointerException("BayesPm must not be null.");
}
if (graph.getNumNodes() == 0) {
throw new IllegalArgumentException(
"The graph must have at least " + "one node in it.");
}
this.dag = new EdgeListGraph(graph);
this.nodesToVariables = new HashMap<>();
copyAvailableInformationFromOldBayesPm(oldBayesPm, lowerBound,
upperBound);
}
/**
* Copy constructor.
*
* @param bayesPm Ibid.
*/
public BayesPm(BayesPm bayesPm) {
this.dag = bayesPm.dag;
this.nodesToVariables = new HashMap<>();
for (Node node : bayesPm.nodesToVariables.keySet()) {
DiscreteVariable variable = bayesPm.nodesToVariables.get(node);
DiscreteVariable newVariable = new DiscreteVariable(variable);
newVariable.setNodeType(node.getNodeType());
this.nodesToVariables.put(node, newVariable);
}
}
/**
* Generates a simple exemplar of this class to test serialization.
*
* @return Ibid.
*/
public static BayesPm serializableInstance() {
return new BayesPm(Dag.serializableInstance());
}
//=========================PUBLIC METHODS=============================//
/**
* Returns the parameter names.
*
* @return Ibid.
*/
public static List getParameterNames() {
List parameters = new ArrayList<>();
parameters.add("minCategories");
parameters.add("maxCategories");
return parameters;
}
private static int pickNumVals(int lowerBound, int upperBound) {
if (lowerBound < 2) {
throw new IllegalArgumentException(
"Lower bound must be >= 2: " + lowerBound);
}
if (upperBound < lowerBound) {
throw new IllegalArgumentException(
"Upper bound for number of categories must be >= lower " + "bound.");
}
int difference = upperBound - lowerBound;
RandomUtil randomUtil = RandomUtil.getInstance();
return randomUtil.nextInt(difference + 1) + lowerBound;
}
/**
* Returns the DAG.
*
* @return the DAG as a Graph.
*/
public Graph getDag() {
return this.dag;
}
/**
* Returns the number of values for the given node.
*
* @param node Ibid.
* @return the number of values for the given node.
*/
public int getNumCategories(Node node) {
DiscreteVariable variable = this.nodesToVariables.get(node);
if (variable == null) {
return 0;
}
return variable.getNumCategories();
}
/**
* Returns the index'th value for the given node.
*
* @param node Ibid.
* @param index Ibid.
* @return the index'th value for the given node.
*/
public String getCategory(Node node, int index) {
DiscreteVariable variable = this.nodesToVariables.get(node);
if (variable != null) {
return variable.getCategory(index);
}
for (DiscreteVariable _node : this.nodesToVariables.values()) {
if (_node == null) {
continue;
}
if (_node.getName().equals(node.getName())) {
return _node.getCategory(index);
}
}
throw new IllegalStateException();
}
/**
* Returns the index of the given category for the given node.
*
* @param node Ibid.
* @param category Ibid.
* @return the index of the given category for the given node.
*/
public int getCategoryIndex(Node node, String category) {
DiscreteVariable variable = this.nodesToVariables.get(node);
return variable.getIndex(category);
}
/**
* Sets the number of values for the given node to the given number.
*
* @param node Ibid.
* @param numCategories Ibid.
*/
public void setNumCategories(Node node, int numCategories) {
if (!this.nodesToVariables.containsKey(node)) {
throw new IllegalArgumentException("Node not in BayesPm: " + node);
}
if (numCategories < 1) {
throw new IllegalArgumentException(
"Number of categories must be >= 1: " + numCategories);
}
DiscreteVariable variable = this.nodesToVariables.get(node);
List oldCategories = variable.getCategories();
List newCategories = new LinkedList<>();
int min = FastMath.min(numCategories, oldCategories.size());
for (int i = 0; i < min; i++) {
newCategories.add(oldCategories.get(i));
}
for (int i = min; i < numCategories; i++) {
String proposedName = DataUtils.defaultCategory(i);
if (newCategories.contains(proposedName)) {
throw new IllegalArgumentException("Default name already in " +
"list of categories: " + proposedName);
}
newCategories.add(proposedName);
}
mapNodeToVariable(node, newCategories);
}
/**
* {@inheritDoc}
*
* Will return true if the argument is a BayesPm with the same graph and variables.
*/
public boolean equals(Object o) {
if (o == null) {
return false;
}
if (!(o instanceof BayesPm bayesPm)) {
return false;
}
return bayesPm.dag.equals(this.dag) && bayesPm.nodesToVariables.equals(this.nodesToVariables);
}
/**
* Sets the number of values for the given node to the given number.
*
* @param node Ibid.
* @param categories Ibid.
*/
public void setCategories(Node node, List categories) {
mapNodeToVariable(node, categories);
}
/**
* getVariables.
*
* @return a {@link java.util.List} object
*/
public List getVariables() {
List variables = new LinkedList<>();
for (Node node : this.nodesToVariables.keySet()) {
variables.add(this.nodesToVariables.get(node));
}
return variables;
}
/**
* Returns the variable names.
*
* @return Ibid.
*/
public List getVariableNames() {
List variables = getVariables();
List names = new ArrayList<>();
for (Node variable : variables) {
DiscreteVariable discreteVariable = (DiscreteVariable) variable;
names.add(discreteVariable.getName());
}
return names;
}
/**
* Returns the variable for the given node.
*
* @param node Ibid.
* @return Ibid.
*/
public Node getVariable(Node node) {
return this.nodesToVariables.get(node);
}
/**
* Returns the measured nodes.
*
* @return the list of measured variableNodes.
*/
public List getMeasuredNodes() {
List measuredNodes = new ArrayList<>();
for (Node variable : getVariables()) {
if (variable.getNodeType() == NodeType.MEASURED) {
measuredNodes.add(variable);
}
}
return measuredNodes;
}
/**
* Prints out the list of values for each node.
*
* @return Ibid.
*/
public String toString() {
StringBuilder buf = new StringBuilder();
for (Node node1 : this.nodesToVariables.keySet()) {
buf.append("\n");
buf.append((node1));
buf.append(": ");
DiscreteVariable variable = this.nodesToVariables.get((node1));
for (int j = 0; j < variable.getNumCategories(); j++) {
buf.append(variable.getCategory(j));
if (j < variable.getNumCategories() - 1) {
buf.append(", ");
}
}
}
return buf.toString();
}
/**
* Returns the node by the given name.
*
* @param nodeName Ibid.
* @return Ibid.
*/
public Node getNode(String nodeName) {
return this.dag.getNode(nodeName);
}
/**
* Returns the node at the given index.
*
* @param index Ibid.
* @return Ibid.
*/
public Node getNode(int index) {
return getVariables().get(index);
}
/**
* Returns the node index.
*
* @return -1.
*/
public int getNodeIndex() {
return -1;
}
/**
* Returns the number of nodes.
*
* @return Ibid.
*/
public int getNumNodes() {
return this.dag.getNumNodes();
}
//=========================PRIVATE METHODS=============================//
private void copyAvailableInformationFromOldBayesPm(BayesPm oldbayesPm,
int lowerBound, int upperBound) {
Graph newGraph = getDag();
Graph oldGraph = oldbayesPm.getDag();
for (Node node1 : newGraph.getNodes()) {
if (oldGraph.containsNode(node1)) {
copyOldValues(oldbayesPm, node1, node1, lowerBound, upperBound);
} else {
setNewValues(node1, lowerBound, upperBound);
}
}
for (Node node2 : newGraph.getNodes()) {
if (oldGraph.containsNode(node2)) {
Node _node2 = this.dag.getNode(node2.getName());
DiscreteVariable oldNode2 = oldbayesPm.nodesToVariables.get(_node2);
oldNode2.setNodeType(node2.getNodeType());
this.nodesToVariables.put(_node2, oldNode2);
} else {
setNewValues(node2, lowerBound, upperBound);
}
}
}
private void copyOldValues(BayesPm oldBayesPm, Node oldNode, Node node,
int lowerBound, int upperBound) {
List values = new ArrayList<>();
List oldNames = new LinkedList<>();
List oldNodes = oldBayesPm.getDag().getNodes();
for (Node oldNode1 : oldNodes) {
oldNames.add(oldNode1.getName());
}
int numVals;
if (oldNames.contains(node.getName())) {
Node oldNode2 = oldBayesPm.getDag().getNode(node.getName());
numVals = oldBayesPm.getNumCategories(oldNode2);
} else {
numVals = BayesPm.pickNumVals(lowerBound, upperBound);
}
int min = FastMath.min(oldBayesPm.getNumCategories(oldNode), numVals);
for (int i = 0; i < min; i++) {
values.add(oldBayesPm.getCategory(oldNode, i));
}
for (int i = min; i < numVals; i++) {
String proposedName = DataUtils.defaultCategory(i);
if (values.contains(proposedName)) {
throw new IllegalArgumentException("Default name already in " +
"list of values: " + proposedName);
}
values.add(proposedName);
}
mapNodeToVariable(node, values);
}
private void setNewValues(Node node, int lowerBound, int upperBound) {
if (node == null) {
throw new NullPointerException("Node must not be null.");
}
List valueList = new ArrayList<>();
for (int i = 0; i < BayesPm.pickNumVals(lowerBound, upperBound); i++) {
valueList.add(DataUtils.defaultCategory(i));
}
mapNodeToVariable(node, valueList);
}
private void mapNodeToVariable(Node node, List categories) {
if (categories.size() != new HashSet<>(categories).size()) {
throw new IllegalArgumentException("Duplicate variable names.");
}
DiscreteVariable variable =
new DiscreteVariable(node.getName(), categories);
variable.setNodeType(node.getNodeType());
this.nodesToVariables.put(node, variable);
}
private void initializeValues(int lowerBound, int upperBound) {
for (Node node : this.dag.getNodes()) {
setNewValues(node, lowerBound, upperBound);
}
}
/**
* Writes the object to the specified ObjectOutputStream.
*
* @param out The ObjectOutputStream to write the object to.
* @throws IOException If an I/O error occurs.
*/
@Serial
private void writeObject(ObjectOutputStream out) throws IOException {
try {
out.defaultWriteObject();
} catch (IOException e) {
TetradLogger.getInstance().log("Failed to serialize object: " + getClass().getCanonicalName()
+ ", " + e.getMessage());
throw e;
}
}
/**
* Reads the object from the specified ObjectInputStream. This method is used during deserialization
* to restore the state of the object.
*
* @param in The ObjectInputStream to read the object from.
* @throws IOException If an I/O error occurs.
* @throws ClassNotFoundException If the class of the serialized object cannot be found.
*/
@Serial
private void readObject(ObjectInputStream in) throws IOException, ClassNotFoundException {
try {
in.defaultReadObject();
} catch (IOException e) {
TetradLogger.getInstance().log("Failed to deserialize object: " + getClass().getCanonicalName()
+ ", " + e.getMessage());
throw e;
}
}
}