uniol.apt.analysis.processmining.InvariantsMapper.java Source code

Introduction

Here is the source code for uniol.apt.analysis.processmining.InvariantsMapper.java

Source

/*-
 * APT - Analysis of Petri Nets and labeled Transition systems
 * Copyright (C) 2016  Uli Schlachter
 *
 * 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.,
 * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
 */

package uniol.apt.analysis.processmining;

import java.util.ArrayList;
import java.util.Collection;
import java.util.Collections;
import java.util.HashSet;
import java.util.List;
import java.util.Set;

import org.apache.commons.collections4.Transformer;

import uniol.apt.adt.ts.ParikhVector;
import uniol.apt.analysis.processmining.algebra.ArrayMatrix;
import uniol.apt.analysis.processmining.algebra.Matrix;
import uniol.apt.analysis.processmining.algebra.SmithNormalForm;

import static uniol.apt.util.DebugUtil.debug;
import static uniol.apt.util.DebugUtil.debugFormat;

/**
 * @author Uli Schlachter
 */
public class InvariantsMapper implements Transformer<ParikhVector, List<Integer>> {
    private final List<String> alphabetList;
    private final List<List<Integer>> newEventWeights = new ArrayList<>();

    /**
     * Define an invariant mapper over the given alphabet with the given words being invariant. All {@link
     * ParikhVector}s that should later be mapped must be over the given alphabet. For example, with the only
     * invariant being "a","b", the words "c" and "a","c","b" will be mapped to the equal results.
     * @param alphabet The alphabet over which Parikh vectors will be given.
     * @param invariantWords A collection of words which should be invariant.
     */
    public InvariantsMapper(Set<String> alphabet, Collection<List<String>> invariantWords) {
        alphabetList = new ArrayList<>(alphabet);

        if (alphabetList.isEmpty())
            return;

        Set<ParikhVector> invariants = new HashSet<>();

        for (List<String> word : invariantWords) {
            ParikhVector wordPV = new ParikhVector(word);
            invariants.add(wordPV);
            if (!alphabet.containsAll(wordPV.getLabels()))
                throw new IllegalArgumentException();
        }

        Matrix transformation;
        List<Integer> diagonals;

        if (invariants.isEmpty()) {
            // This would need a matrix with dimension zero. Instead, give the result explicitly.
            debug("No invariants giving, using identity mapping");
            transformation = ArrayMatrix.createIdentityMatrix(alphabetList.size(), alphabetList.size());
            diagonals = Collections.emptyList();
        } else {
            // Turn the invariants into a matrix where each column is the Parikh vector of an invariant
            Matrix matrix = ArrayMatrix.createIdentityMatrix(alphabetList.size(), invariants.size());
            int invariantNumber = 0;
            for (ParikhVector invariant : invariants) {
                int eventNumber = 0;
                for (String event : alphabetList) {
                    matrix.set(eventNumber, invariantNumber, invariant.get(event));
                    eventNumber++;
                }
                invariantNumber++;
            }

            // Calculate the Smith normal form of the above matrix
            SmithNormalForm nf = new SmithNormalForm(matrix);
            transformation = nf.getLeftHandMatrixInverse();
            diagonals = nf.getDiagonalEntries();
            debugFormat("Calculated smith normal form: %s*(%s)*%s", nf.getLeftHandMatrix(), diagonals,
                    nf.getRightHandMatrix());
            debugFormat("Inverse matrices are %s and %s", nf.getLeftHandMatrixInverse(),
                    nf.getRightHandMatrixInverse());
        }

        // Each non-zero diagonal entry corresponds to a "vanishing dimension", i.e. some new-event who will be
        // ignored in the following. Each zero diagonal entry produces a new event.

        for (int index = 0; index < alphabetList.size(); index++) {
            int diag = index < diagonals.size() ? diagonals.get(index) : 0;
            if (diag != 0) {
                // haven't reached the zero entries yet
                continue;
            }

            List<Integer> weights = new ArrayList<>();
            int eventIndex = 0;
            for (String event : alphabetList) {
                weights.add(transformation.get(index, eventIndex));
                eventIndex++;
            }
            newEventWeights.add(weights);
            debugFormat("New event e_%d has weights %s in %s", index, weights, alphabetList);
        }
    }

    @Override
    public List<Integer> transform(ParikhVector pv) {
        if (!alphabetList.containsAll(pv.getLabels()))
            throw new IllegalArgumentException();
        List<Integer> result = new ArrayList<>();
        for (List<Integer> eventWeights : newEventWeights) {
            int count = 0;
            for (int index = 0; index < alphabetList.size(); index++)
                count += eventWeights.get(index) * pv.get(alphabetList.get(index));
            result.add(count);
        }
        return result;
    }
}

// vim: ft=java:noet:sw=8:sts=8:ts=8:tw=120