BNLearner.h

Go to the documentation of this file.

/****************************************************************************
 *   This file is part of the aGrUM/pyAgrum library.                        *
 *                                                                          *
 *   Copyright (c) 2005-2025 by                                             *
 *       - Pierre-Henri WUILLEMIN(_at_LIP6)                                 *
 *       - Christophe GONZALES(_at_AMU)                                     *
 *                                                                          *
 *   The aGrUM/pyAgrum library is free software; you can redistribute it    *
 *   and/or modify it under the terms of either :                           *
 *                                                                          *
 *    - the GNU Lesser General Public License as published by               *
 *      the Free Software Foundation, either version 3 of the License,      *
 *      or (at your option) any later version,                              *
 *    - the MIT license (MIT),                                              *
 *    - or both in dual license, as here.                                   *
 *                                                                          *
 *   (see https://agrum.gitlab.io/articles/dual-licenses-lgplv3mit.html)    *
 *                                                                          *
 *   This aGrUM/pyAgrum library is distributed in the hope that it will be  *
 *   useful, but WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED,          *
 *   INCLUDING BUT NOT LIMITED TO THE WARRANTIES MERCHANTABILITY or FITNESS *
 *   FOR A PARTICULAR PURPOSE  AND NONINFRINGEMENT. IN NO EVENT SHALL THE   *
 *   AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER *
 *   LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,        *
 *   ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR  *
 *   OTHER DEALINGS IN THE SOFTWARE.                                        *
 *                                                                          *
 *   See LICENCES for more details.                                         *
 *                                                                          *
 *   SPDX-FileCopyrightText: Copyright 2005-2025                            *
 *       - Pierre-Henri WUILLEMIN(_at_LIP6)                                 *
 *       - Christophe GONZALES(_at_AMU)                                     *
 *   SPDX-License-Identifier: LGPL-3.0-or-later OR MIT                      *
 *                                                                          *
 *   Contact  : info_at_agrum_dot_org                                       *
 *   homepage : http://agrum.gitlab.io                                      *
 *   gitlab   : https://gitlab.com/agrumery/agrum                           *
 *                                                                          *
 ****************************************************************************/
 

#ifndef GUM_LEARNING_BN_LEARNER_H
#define GUM_LEARNING_BN_LEARNER_H
 
#include <algorithm>
#include <sstream>
#include <vector>
 
#include <agrum/agrum.h>
 
#include <agrum/BN/learning/BNLearnUtils/IBNLearner.h>
#include <agrum/BN/learning/priors/DirichletPriorFromBN.h>
 
namespace gum {
  namespace learning {
    class BNLearnerListener;

    template < typename GUM_SCALAR >
    class BNLearner final: public IBNLearner {
      public:
      // ##########################################################################
      // ##########################################################################


      BNLearner(const std::string&                filename,
                const std::vector< std::string >& missingSymbols = {"?"},
                const bool                        induceTypes    = true);


      BNLearner(const DatabaseTable& db);

      BNLearner(const std::string&                 filename,
                const gum::BayesNet< GUM_SCALAR >& src,
                const std::vector< std::string >&  missing_symbols = {"?"});

      BNLearner(const BNLearner&);

      BNLearner(BNLearner&&);

      virtual ~BNLearner();

 
      // ##########################################################################
      // ##########################################################################

      BNLearner& operator=(const BNLearner&);

      BNLearner& operator=(BNLearner&&) noexcept;


      BayesNet< GUM_SCALAR > learnBN();

      BayesNet< GUM_SCALAR > learnParameters(const DAG& dag, bool takeIntoAccountScore = true);

      BayesNet< GUM_SCALAR > learnParameters(const BayesNet< GUM_SCALAR >& bn,
                                             bool takeIntoAccountScore = true);

      BayesNet< GUM_SCALAR > learnParameters(bool take_into_account_score = true);

      std::string toString() const;

      std::vector< std::tuple< std::string, std::string, std::string > > state() const;

      void copyState(const BNLearner< GUM_SCALAR >& learner);
 
      //=== === add return to certain methods in order to chain command
      BNLearner< GUM_SCALAR >& setInitialDAG(const DAG& dag) {
        IBNLearner::setInitialDAG(dag);
        return *this;
      }

      BNLearner< GUM_SCALAR >& useEM(const double epsilon, const double noise = default_EM_noise) {
        IBNLearner::useEM(epsilon, noise);
        return *this;
      }

      BNLearner< GUM_SCALAR >& useEMWithRateCriterion(const double epsilon,
                                                      const double noise = default_EM_noise) {
        IBNLearner::useEMWithRateCriterion(epsilon, noise);
        return *this;
      }

      BNLearner< GUM_SCALAR >& useEMWithDiffCriterion(const double epsilon,
                                                      const double noise = default_EM_noise) {
        IBNLearner::useEMWithDiffCriterion(epsilon, noise);
        return *this;
      }

      BNLearner< GUM_SCALAR >& forbidEM() {
        IBNLearner::forbidEM();
        return *this;
      }

      BNLearner< GUM_SCALAR >& EMsetEpsilon(const double eps) {
        IBNLearner::EMsetEpsilon(eps);
        return *this;
      }

      BNLearner< GUM_SCALAR >& EMdisableEpsilon() {
        IBNLearner::EMdisableEpsilon();
        return *this;
      }

      BNLearner< GUM_SCALAR >& EMenableEpsilon() {
        IBNLearner::EMenableEpsilon();
        return *this;
      }

      BNLearner< GUM_SCALAR >& EMsetMinEpsilonRate(const double rate) {
        IBNLearner::EMsetMinEpsilonRate(rate);
        return *this;
      }

      BNLearner< GUM_SCALAR >& EMdisableMinEpsilonRate() {
        IBNLearner::EMdisableMinEpsilonRate();
        return *this;
      }

      BNLearner< GUM_SCALAR >& EMenableMinEpsilonRate() {
        IBNLearner::EMenableMinEpsilonRate();
        return *this;
      }

      BNLearner< GUM_SCALAR >& EMsetMaxIter(const Size max) {
        IBNLearner::EMsetMaxIter(max);
        return *this;
      }

      BNLearner< GUM_SCALAR >& EMdisableMaxIter() {
        IBNLearner::EMdisableMaxIter();
        return *this;
      }

      BNLearner< GUM_SCALAR >& EMenableMaxIter() {
        IBNLearner::EMenableMaxIter();
        return *this;
      }

      BNLearner< GUM_SCALAR >& EMsetMaxTime(const double timeout) {
        IBNLearner::EMsetMaxTime(timeout);
        return *this;
      }

      BNLearner< GUM_SCALAR >& EMdisableMaxTime() {
        IBNLearner::EMdisableMaxTime();
        return *this;
      }

      BNLearner< GUM_SCALAR >& EMenableMaxTime() {
        IBNLearner::EMenableMaxTime();
        return *this;
      };

      BNLearner< GUM_SCALAR >& EMsetPeriodSize(const Size p) {
        IBNLearner::EMsetPeriodSize(p);
        return *this;
      }
 
      // for pyagrum support
      using IBNLearner::EMPeriodSize;

      BNLearner< GUM_SCALAR >& EMsetVerbosity(const bool v) {
        IBNLearner::EMsetVerbosity(v);
        return *this;
      }
 
      BNLearner< GUM_SCALAR >& useScoreAIC() {
        IBNLearner::useScoreAIC();
        return *this;
      }
 
      BNLearner< GUM_SCALAR >& useScoreBD() {
        IBNLearner::useScoreBD();
        return *this;
      }
 
      BNLearner< GUM_SCALAR >& useScoreBDeu() {
        IBNLearner::useScoreBDeu();
        return *this;
      }
 
      BNLearner< GUM_SCALAR >& useScoreBIC() {
        IBNLearner::useScoreBIC();
        return *this;
      }
 
      BNLearner< GUM_SCALAR >& useScoreK2() {
        IBNLearner::useScoreK2();
        return *this;
      }
 
      BNLearner< GUM_SCALAR >& useScoreLog2Likelihood() {
        IBNLearner::useScoreLog2Likelihood();
        return *this;
      }
 
      BNLearner< GUM_SCALAR >& useNoPrior() {
        IBNLearner::useNoPrior();
        return *this;
      }
 
      BNLearner< GUM_SCALAR >& useBDeuPrior(double weight = 1.0) {
        IBNLearner::useBDeuPrior(weight);
        return *this;
      }
 
      BNLearner< GUM_SCALAR >& useSmoothingPrior(double weight = 1) {
        IBNLearner::useSmoothingPrior(weight);
        return *this;
      }
 
      BNLearner< GUM_SCALAR >& useDirichletPrior(const std::string& filename, double weight = 1) {
        IBNLearner::useDirichletPrior(filename, weight);
        return *this;
      }
 
      BNLearner< GUM_SCALAR >& useDirichletPrior(const gum::BayesNet< GUM_SCALAR >& bn,
                                                 double                             weight = 1) {
        if (weight < 0) { GUM_ERROR(OutOfBounds, "the weight of the prior must be positive") }
 
        _prior_bn_ = gum::BayesNet< GUM_SCALAR >(bn);
        priorType_ = BNLearnerPriorType::DIRICHLET_FROM_BAYESNET;
        _setPriorWeight_(weight);
 
        checkScorePriorCompatibility();
        return *this;
      }
 
      BNLearner< GUM_SCALAR >& useGreedyHillClimbing() {
        IBNLearner::useGreedyHillClimbing();
        return *this;
      }
 
      BNLearner< GUM_SCALAR >& useLocalSearchWithTabuList(Size tabu_size   = 100,
                                                          Size nb_decrease = 2) {
        IBNLearner::useLocalSearchWithTabuList(tabu_size, nb_decrease);
        return *this;
      }
 
      BNLearner< GUM_SCALAR >& useK2(const Sequence< NodeId >& order) {
        IBNLearner::useK2(order);
        return *this;
      }
 
      BNLearner< GUM_SCALAR >& useK2(const std::vector< NodeId >& order) {
        IBNLearner::useK2(order);
        return *this;
      }
 
      BNLearner< GUM_SCALAR >& useMIIC() {
        IBNLearner::useMIIC();
        return *this;
      }
 
      BNLearner< GUM_SCALAR >& useNMLCorrection() {
        IBNLearner::useNMLCorrection();
        return *this;
      }
 
      BNLearner< GUM_SCALAR >& useMDLCorrection() {
        IBNLearner::useMDLCorrection();
        return *this;
      }
 
      BNLearner< GUM_SCALAR >& useNoCorrection() {
        IBNLearner::useNoCorrection();
        return *this;
      }
 
      BNLearner< GUM_SCALAR >& setMaxIndegree(Size max_indegree) {
        IBNLearner::setMaxIndegree(max_indegree);
        return *this;
      }
 
      BNLearner< GUM_SCALAR >& setSliceOrder(const NodeProperty< NodeId >& slice_order) {
        IBNLearner::setSliceOrder(slice_order);
        return *this;
      }
 
      BNLearner< GUM_SCALAR >&
          setSliceOrder(const std::vector< std::vector< std::string > >& slices) {
        IBNLearner::setSliceOrder(slices);
        return *this;
      }
 
      BNLearner< GUM_SCALAR >& setForbiddenArcs(const ArcSet& set) {
        IBNLearner::setForbiddenArcs(set);
        return *this;
      }
 
      BNLearner< GUM_SCALAR >& addForbiddenArc(const Arc& arc) {
        IBNLearner::addForbiddenArc(arc);
        return *this;
      }
 
      BNLearner< GUM_SCALAR >& addForbiddenArc(NodeId tail, NodeId head) {
        IBNLearner::addForbiddenArc(tail, head);
        return *this;
      }
 
      BNLearner< GUM_SCALAR >& addForbiddenArc(const std::string& tail, const std::string& head) {
        IBNLearner::addForbiddenArc(tail, head);
        return *this;
      }
 
      BNLearner< GUM_SCALAR >& eraseForbiddenArc(const Arc& arc) {
        IBNLearner::eraseForbiddenArc(arc);
        return *this;
      }
 
      BNLearner< GUM_SCALAR >& eraseForbiddenArc(NodeId tail, NodeId head) {
        IBNLearner::eraseForbiddenArc(tail, head);
        return *this;
      }
 
      BNLearner< GUM_SCALAR >& eraseForbiddenArc(const std::string& tail, const std::string& head) {
        IBNLearner::eraseForbiddenArc(tail, head);
        return *this;
      }
 
      BNLearner< GUM_SCALAR >& addMandatoryArc(const Arc& arc) {
        IBNLearner::addMandatoryArc(arc);
        return *this;
      }
 
      BNLearner< GUM_SCALAR >& addMandatoryArc(NodeId tail, NodeId head) {
        IBNLearner::addMandatoryArc(tail, head);
        return *this;
      }
 
      BNLearner< GUM_SCALAR >& addMandatoryArc(const std::string& tail, const std::string& head) {
        IBNLearner::addMandatoryArc(tail, head);
        return *this;
      }
 
      BNLearner< GUM_SCALAR >& eraseMandatoryArc(const Arc& arc) {
        IBNLearner::eraseMandatoryArc(arc);
        return *this;
      }
 
      BNLearner< GUM_SCALAR >& eraseMandatoryArc(NodeId tail, NodeId head) {
        IBNLearner::eraseMandatoryArc(tail, head);
        return *this;
      }
 
      BNLearner< GUM_SCALAR >& eraseMandatoryArc(const std::string& tail, const std::string& head) {
        IBNLearner::eraseMandatoryArc(tail, head);
        return *this;
      }
 
      BNLearner< GUM_SCALAR >& addPossibleEdge(const Edge& edge) {
        IBNLearner::addPossibleEdge(edge);
        return *this;
      }
 
      BNLearner< GUM_SCALAR >& addPossibleEdge(NodeId tail, NodeId head) {
        IBNLearner::addPossibleEdge(tail, head);
        return *this;
      }
 
      BNLearner< GUM_SCALAR >& addPossibleEdge(const std::string& tail, const std::string& head) {
        IBNLearner::addPossibleEdge(tail, head);
        return *this;
      }
 
      BNLearner< GUM_SCALAR >& erasePossibleEdge(const Edge& edge) {
        IBNLearner::erasePossibleEdge(edge);
        return *this;
      }
 
      BNLearner< GUM_SCALAR >& erasePossibleEdge(NodeId tail, NodeId head) {
        IBNLearner::erasePossibleEdge(tail, head);
        return *this;
      }
 
      BNLearner< GUM_SCALAR >& erasePossibleEdge(const std::string& tail, const std::string& head) {
        IBNLearner::erasePossibleEdge(tail, head);
        return *this;
      }
 
      BNLearner< GUM_SCALAR >& setMandatoryArcs(const ArcSet& set) {
        IBNLearner::setMandatoryArcs(set);
        return *this;
      }
 
      BNLearner< GUM_SCALAR >& setPossibleEdges(const EdgeSet& set) {
        IBNLearner::setPossibleEdges(set);
        return *this;
      }
 
      BNLearner< GUM_SCALAR >& setPossibleSkeleton(const UndiGraph& skeleton) {
        IBNLearner::setPossibleSkeleton(skeleton);
        return *this;
      }
 
      BNLearner< GUM_SCALAR >& addNoParentNode(NodeId node) {
        IBNLearner::addNoParentNode(node);
        return *this;
      }
 
      BNLearner< GUM_SCALAR >& addNoParentNode(const std::string& name) {
        IBNLearner::addNoParentNode(name);
        return *this;
      }
 
      BNLearner< GUM_SCALAR >& eraseNoParentNode(NodeId node) {
        IBNLearner::eraseNoParentNode(node);
        return *this;
      }
 
      BNLearner< GUM_SCALAR >& eraseNoParentNode(const std::string& name) {
        IBNLearner::eraseNoParentNode(name);
        return *this;
      }
 
      BNLearner< GUM_SCALAR >& addNoChildrenNode(NodeId node) {
        IBNLearner::addNoChildrenNode(node);
        return *this;
      }
 
      BNLearner< GUM_SCALAR >& addNoChildrenNode(const std::string& name) {
        IBNLearner::addNoChildrenNode(name);
        return *this;
      }
 
      BNLearner< GUM_SCALAR >& eraseNoChildrenNode(NodeId node) {
        IBNLearner::eraseNoChildrenNode(node);
        return *this;
      }
 
      BNLearner< GUM_SCALAR >& eraseNoChildrenNode(const std::string& name) {
        IBNLearner::eraseNoChildrenNode(name);
        return *this;
      }
 
      bool isConstraintBased() const { return IBNLearner::isConstraintBased(); }
 
      bool isScoreBased() const { return IBNLearner::isScoreBased(); }
 
      protected:
      void createPrior_();
 
      private:
      BayesNet< GUM_SCALAR > _prior_bn_;

      NodeProperty< Sequence< std::string > > _labelsFromBN_(const std::string&            filename,
                                                             const BayesNet< GUM_SCALAR >& src);

      void _checkDAGCompatibility_(const DAG& dag);

      BayesNet< GUM_SCALAR > _learnParameters_(const DAG& dag, bool takeIntoAccountScore);

      std::pair< std::shared_ptr< ParamEstimator >, std::shared_ptr< ParamEstimator > >
          _initializeEMParameterLearning_(const DAG& dag, bool takeIntoAccountScore);

      BayesNet< GUM_SCALAR > _learnParametersWithEM_(const DAG& dag, bool takeIntoAccountScore);

      BayesNet< GUM_SCALAR > _learnParametersWithEM_(const BayesNet< GUM_SCALAR >& bn,
                                                     bool takeIntoAccountScore);
    };

    template < typename GUM_SCALAR >
    std::ostream& operator<<(std::ostream& output, const BNLearner< GUM_SCALAR >& learner);
  } /* namespace learning */
} /* namespace gum */

#include <agrum/BN/learning/BNLearner_tpl.h>
 
#endif /* GUM_LEARNING_BN_LEARNER_H */