Merge branch 'develop' into model-selection-bayestree

gtsam/discrete/tests/testAlgebraicDecisionTree.cpp

@@ -593,6 +593,55 @@ TEST(ADT, zero) {
   EXPECT_DOUBLES_EQUAL(0, anotb(x11), 1e-9);
 }
 
+/// Example ADT from 0 to 11.
+ADT exampleADT() {
+  DiscreteKey A(0, 2), B(1, 3), C(2, 2);
+  ADT f(A & B & C, "0 6  2 8  4 10    1 7  3 9  5 11");
+  return f;
+}
+/* ************************************************************************** */
+// Test sum
+TEST(ADT, Sum) {
+  ADT a = exampleADT();
+  double expected_sum = 0;
+  for (double i = 0; i < 12; i++) {
+    expected_sum += i;
+  }
+  EXPECT_DOUBLES_EQUAL(expected_sum, a.sum(), 1e-9);
+}
+
+/* ************************************************************************** */
+// Test normalize
+TEST(ADT, Normalize) {
+  ADT a = exampleADT();
+  double sum = a.sum();
+  auto actual = a.normalize(sum);
+
+  DiscreteKey A(0, 2), B(1, 3), C(2, 2);
+  DiscreteKeys keys = DiscreteKeys{A, B, C};
+  std::vector<double> cpt{0 / sum, 6 / sum,  2 / sum, 8 / sum,
+                          4 / sum, 10 / sum, 1 / sum, 7 / sum,
+                          3 / sum, 9 / sum,  5 / sum, 11 / sum};
+  ADT expected(keys, cpt);
+  EXPECT(assert_equal(expected, actual));
+}
+
+/* ************************************************************************** */
+// Test min
+TEST(ADT, Min) {
+  ADT a = exampleADT();
+  double min = a.min();
+  EXPECT_DOUBLES_EQUAL(0.0, min, 1e-9);
+}
+
+/* ************************************************************************** */
+// Test max
+TEST(ADT, Max) {
+  ADT a = exampleADT();
+  double max = a.max();
+  EXPECT_DOUBLES_EQUAL(11.0, max, 1e-9);
+}
+
 /* ************************************************************************* */
 int main() {
   TestResult tr;

gtsam/geometry/geometry.i

@@ -856,7 +856,7 @@ class Cal3_S2Stereo {
   gtsam::Matrix K() const;
   gtsam::Point2 principalPoint() const;
   double baseline() const;
-  Vector6 vector() const;
+  gtsam::Vector6 vector() const;
   gtsam::Matrix inverse() const;
 };
 

gtsam/hybrid/GaussianMixture.cpp

@@ -24,7 +24,6 @@
 #include <gtsam/hybrid/GaussianMixtureFactor.h>
 #include <gtsam/hybrid/HybridValues.h>
 #include <gtsam/inference/Conditional-inst.h>
-#include <gtsam/linear/GaussianBayesNet.h>
 #include <gtsam/linear/GaussianFactorGraph.h>
 
 namespace gtsam {
@@ -70,43 +69,6 @@ GaussianMixture::GaussianMixture(
     : GaussianMixture(continuousFrontals, continuousParents, discreteParents,
                       Conditionals(discreteParents, conditionals)) {}
 
-/* *******************************************************************************/
-GaussianBayesNetTree GaussianMixture::asGaussianBayesNetTree() const {
-  auto wrap = [](const GaussianConditional::shared_ptr &gc) {
-    if (gc) {
-      return GaussianBayesNet{gc};
-    } else {
-      return GaussianBayesNet();
-    }
-  };
-  return {conditionals_, wrap};
-}
-
-/* *******************************************************************************/
-GaussianFactorGraphTree GaussianMixture::asGaussianFactorGraphTree() const {
-  auto wrap = [](const GaussianBayesNet &gbn) {
-    return GaussianFactorGraph(gbn);
-  };
-  return {this->asGaussianBayesNetTree(), wrap};
-}
-
-/*
-*******************************************************************************/
-GaussianBayesNetTree GaussianMixture::add(
-    const GaussianBayesNetTree &sum) const {
-  using Y = GaussianBayesNet;
-  auto add = [](const Y &graph1, const Y &graph2) {
-    auto result = graph1;
-    if (graph2.size() == 0) {
-      return GaussianBayesNet();
-    }
-    result.push_back(graph2);
-    return result;
-  };
-  const auto tree = asGaussianBayesNetTree();
-  return sum.empty() ? tree : sum.apply(tree, add);
-}
-
 /* *******************************************************************************/
 // TODO(dellaert): This is copy/paste: GaussianMixture should be derived from
 // GaussianMixtureFactor, no?
@@ -122,6 +84,14 @@ GaussianFactorGraphTree GaussianMixture::add(
   return sum.empty() ? tree : sum.apply(tree, add);
 }
 
+/* *******************************************************************************/
+GaussianFactorGraphTree GaussianMixture::asGaussianFactorGraphTree() const {
+  auto wrap = [](const GaussianConditional::shared_ptr &gc) {
+    return GaussianFactorGraph{gc};
+  };
+  return {conditionals_, wrap};
+}
+
 /* *******************************************************************************/
 size_t GaussianMixture::nrComponents() const {
   size_t total = 0;
@@ -347,15 +317,8 @@ AlgebraicDecisionTree<Key> GaussianMixture::logProbability(
 AlgebraicDecisionTree<Key> GaussianMixture::errorTree(
     const VectorValues &continuousValues) const {
   auto errorFunc = [&](const GaussianConditional::shared_ptr &conditional) {
-    // Check if valid pointer
+    return conditional->error(continuousValues) +  //
-    if (conditional) {
+           logConstant_ - conditional->logNormalizationConstant();
-      return conditional->error(continuousValues) +  //
-             logConstant_ - conditional->logNormalizationConstant();
-    } else {
-      // If not valid, pointer, it means this conditional was pruned,
-      // so we return maximum error.
-      return std::numeric_limits<double>::max();
-    }
   };
   DecisionTree<Key, double> error_tree(conditionals_, errorFunc);
   return error_tree;
@@ -363,32 +326,10 @@ AlgebraicDecisionTree<Key> GaussianMixture::errorTree(
 
 /* *******************************************************************************/
 double GaussianMixture::error(const HybridValues &values) const {
-  // Check if discrete keys in discrete assignment are
-  // present in the GaussianMixture
-  KeyVector dKeys = this->discreteKeys_.indices();
-  bool valid_assignment = false;
-  for (auto &&kv : values.discrete()) {
-    if (std::find(dKeys.begin(), dKeys.end(), kv.first) != dKeys.end()) {
-      valid_assignment = true;
-      break;
-    }
-  }
-
-  // The discrete assignment is not valid so we return 0.0 erorr.
-  if (!valid_assignment) {
-    return 0.0;
-  }
-
   // Directly index to get the conditional, no need to build the whole tree.
   auto conditional = conditionals_(values.discrete());
-  if (conditional) {
+  return conditional->error(values.continuous()) +  //
-    return conditional->error(values.continuous()) +  //
+         logConstant_ - conditional->logNormalizationConstant();
-           logConstant_ - conditional->logNormalizationConstant();
-  } else {
-    // If not valid, pointer, it means this conditional was pruned,
-    // so we return maximum error.
-    return std::numeric_limits<double>::max();
-  }
 }
 
 /* *******************************************************************************/

gtsam/hybrid/GaussianMixture.h

@@ -72,12 +72,6 @@ class GTSAM_EXPORT GaussianMixture
    */
   GaussianFactorGraphTree asGaussianFactorGraphTree() const;
 
-  /**
-   * @brief Convert a DecisionTree of conditionals into
-   * a DecisionTree of Gaussian Bayes nets.
-   */
-  GaussianBayesNetTree asGaussianBayesNetTree() const;
-
   /**
    * @brief Helper function to get the pruner functor.
    *
@@ -221,7 +215,8 @@ class GTSAM_EXPORT GaussianMixture
    * @return AlgebraicDecisionTree<Key> A decision tree on the discrete keys
    * only, with the leaf values as the error for each assignment.
    */
-  AlgebraicDecisionTree<Key> errorTree(const VectorValues &continuousValues) const;
+  AlgebraicDecisionTree<Key> errorTree(
+      const VectorValues &continuousValues) const;
 
   /**
    * @brief Compute the logProbability of this Gaussian Mixture.
@@ -255,15 +250,6 @@ class GTSAM_EXPORT GaussianMixture
    * @return GaussianFactorGraphTree
    */
   GaussianFactorGraphTree add(const GaussianFactorGraphTree &sum) const;
-
-  /**
-   * @brief Merge the Gaussian Bayes Nets in `this` and `sum` while
-   * maintaining the decision tree structure.
-   *
-   * @param sum Decision Tree of Gaussian Bayes Nets
-   * @return GaussianBayesNetTree
-   */
-  GaussianBayesNetTree add(const GaussianBayesNetTree &sum) const;
   /// @}
 
  private:

gtsam/hybrid/GaussianMixtureFactor.cpp

@@ -28,86 +28,11 @@
 
 namespace gtsam {
 
-/**
- * @brief Helper function to correct the [A|b] matrices in the factor components
- * with the normalizer values.
- * This is done by storing the normalizer value in
- * the `b` vector as an additional row.
- *
- * @param factors DecisionTree of GaussianFactor shared pointers.
- * @param varyingNormalizers Flag indicating the normalizers are different for
- * each component.
- * @return GaussianMixtureFactor::Factors
- */
-GaussianMixtureFactor::Factors correct(
-    const GaussianMixtureFactor::Factors &factors, bool varyingNormalizers) {
-  if (!varyingNormalizers) {
-    return factors;
-  }
-
-  // First compute all the sqrt(|2 pi Sigma|) terms
-  auto computeNormalizers = [](const GaussianMixtureFactor::sharedFactor &gf) {
-    auto jf = std::dynamic_pointer_cast<JacobianFactor>(gf);
-    // If we have, say, a Hessian factor, then no need to do anything
-    if (!jf) return 0.0;
-
-    auto model = jf->get_model();
-    // If there is no noise model, there is nothing to do.
-    if (!model) {
-      return 0.0;
-    }
-    // Since noise models are Gaussian, we can get the logDeterminant using the
-    // same trick as in GaussianConditional
-    double logDetR =
-        model->R().diagonal().unaryExpr([](double x) { return log(x); }).sum();
-    double logDeterminantSigma = -2.0 * logDetR;
-
-    size_t n = model->dim();
-    constexpr double log2pi = 1.8378770664093454835606594728112;
-    return n * log2pi + logDeterminantSigma;
-  };
-
-  AlgebraicDecisionTree<Key> log_normalizers =
-      DecisionTree<Key, double>(factors, computeNormalizers);
-
-  // Find the minimum value so we can "proselytize" to positive values.
-  // Done because we can't have sqrt of negative numbers.
-  double min_log_normalizer = log_normalizers.min();
-  log_normalizers = log_normalizers.apply(
-      [&min_log_normalizer](double n) { return n - min_log_normalizer; });
-
-  // Finally, update the [A|b] matrices.
-  auto update = [&log_normalizers](
-                    const Assignment<Key> &assignment,
-                    const GaussianMixtureFactor::sharedFactor &gf) {
-    auto jf = std::dynamic_pointer_cast<JacobianFactor>(gf);
-    if (!jf) return gf;
-    // If there is no noise model, there is nothing to do.
-    if (!jf->get_model()) return gf;
-    // If the log_normalizer is 0, do nothing
-    if (log_normalizers(assignment) == 0.0) return gf;
-
-    GaussianFactorGraph gfg;
-    gfg.push_back(jf);
-
-    Vector c(1);
-    c << std::sqrt(log_normalizers(assignment));
-    auto constantFactor = std::make_shared<JacobianFactor>(c);
-
-    gfg.push_back(constantFactor);
-    return std::dynamic_pointer_cast<GaussianFactor>(
-        std::make_shared<JacobianFactor>(gfg));
-  };
-  return factors.apply(update);
-}
-
 /* *******************************************************************************/
 GaussianMixtureFactor::GaussianMixtureFactor(const KeyVector &continuousKeys,
                                              const DiscreteKeys &discreteKeys,
-                                             const Factors &factors,
+                                             const Factors &factors)
-                                             bool varyingNormalizers)
+    : Base(continuousKeys, discreteKeys), factors_(factors) {}
-    : Base(continuousKeys, discreteKeys),
-      factors_(correct(factors, varyingNormalizers)) {}
 
 /* *******************************************************************************/
 bool GaussianMixtureFactor::equals(const HybridFactor &lf, double tol) const {
@@ -129,9 +54,7 @@ bool GaussianMixtureFactor::equals(const HybridFactor &lf, double tol) const {
 /* *******************************************************************************/
 void GaussianMixtureFactor::print(const std::string &s,
                                   const KeyFormatter &formatter) const {
-  std::cout << (s.empty() ? "" : s + "\n");
+  HybridFactor::print(s, formatter);
-  std::cout << "GaussianMixtureFactor" << std::endl;
-  HybridFactor::print("", formatter);
   std::cout << "{\n";
   if (factors_.empty()) {
     std::cout << "  empty" << std::endl;

gtsam/hybrid/GaussianMixtureFactor.h

@@ -82,13 +82,10 @@ class GTSAM_EXPORT GaussianMixtureFactor : public HybridFactor {
    * their cardinalities.
    * @param factors The decision tree of Gaussian factors stored as the mixture
    * density.
-   * @param varyingNormalizers Flag indicating factor components have varying
-   * normalizer values.
    */
   GaussianMixtureFactor(const KeyVector &continuousKeys,
                         const DiscreteKeys &discreteKeys,
-                        const Factors &factors,
+                        const Factors &factors);
-                        bool varyingNormalizers = false);
 
   /**
    * @brief Construct a new GaussianMixtureFactor object using a vector of
@@ -97,16 +94,12 @@ class GTSAM_EXPORT GaussianMixtureFactor : public HybridFactor {
    * @param continuousKeys Vector of keys for continuous factors.
    * @param discreteKeys Vector of discrete keys.
    * @param factors Vector of gaussian factor shared pointers.
-   * @param varyingNormalizers Flag indicating factor components have varying
-   * normalizer values.
    */
   GaussianMixtureFactor(const KeyVector &continuousKeys,
                         const DiscreteKeys &discreteKeys,
-                        const std::vector<sharedFactor> &factors,
+                        const std::vector<sharedFactor> &factors)
-                        bool varyingNormalizers = false)
       : GaussianMixtureFactor(continuousKeys, discreteKeys,
-                              Factors(discreteKeys, factors),
+                              Factors(discreteKeys, factors)) {}
-                              varyingNormalizers) {}
 
   /// @}
   /// @name Testable
@@ -114,8 +107,9 @@ class GTSAM_EXPORT GaussianMixtureFactor : public HybridFactor {
 
   bool equals(const HybridFactor &lf, double tol = 1e-9) const override;
 
-  void print(const std::string &s = "", const KeyFormatter &formatter =
+  void print(
-                                            DefaultKeyFormatter) const override;
+      const std::string &s = "GaussianMixtureFactor\n",
+      const KeyFormatter &formatter = DefaultKeyFormatter) const override;
 
   /// @}
   /// @name Standard API

gtsam/hybrid/HybridBayesNet.cpp

@@ -220,16 +220,15 @@ GaussianBayesNet HybridBayesNet::choose(
 /* ************************************************************************* */
 HybridValues HybridBayesNet::optimize() const {
   // Collect all the discrete factors to compute MPE
-  DiscreteFactorGraph discrete_fg;
+  DiscreteBayesNet discrete_bn;
-
   for (auto &&conditional : *this) {
     if (conditional->isDiscrete()) {
-      discrete_fg.push_back(conditional->asDiscrete());
+      discrete_bn.push_back(conditional->asDiscrete());
     }
   }
 
   // Solve for the MPE
-  DiscreteValues mpe = discrete_fg.optimize();
+  DiscreteValues mpe = DiscreteFactorGraph(discrete_bn).optimize();
 
   // Given the MPE, compute the optimal continuous values.
   return HybridValues(optimize(mpe), mpe);

gtsam/hybrid/HybridFactor.h

@@ -13,7 +13,6 @@
  *  @file HybridFactor.h
  *  @date Mar 11, 2022
  *  @author Fan Jiang
- *  @author Varun Agrawal
  */
 
 #pragma once
@@ -34,8 +33,6 @@ class HybridValues;
 
 /// Alias for DecisionTree of GaussianFactorGraphs
 using GaussianFactorGraphTree = DecisionTree<Key, GaussianFactorGraph>;
-/// Alias for DecisionTree of GaussianBayesNets
-using GaussianBayesNetTree = DecisionTree<Key, GaussianBayesNet>;
 
 KeyVector CollectKeys(const KeyVector &continuousKeys,
                       const DiscreteKeys &discreteKeys);

gtsam/hybrid/HybridGaussianFactorGraph.cpp

@@ -279,37 +279,21 @@ GaussianFactorGraphTree removeEmpty(const GaussianFactorGraphTree &sum) {
 using Result = std::pair<std::shared_ptr<GaussianConditional>,
                          GaussianMixtureFactor::sharedFactor>;
 
-/**
+// Integrate the probability mass in the last continuous conditional using
- * Compute the probability q(μ;m) = exp(-error(μ;m)) * sqrt(det(2π Σ_m)
+// the unnormalized probability q(μ;m) = exp(-error(μ;m)) at the mean.
- * from the residual error at the mean μ.
+//   discrete_probability = exp(-error(μ;m)) * sqrt(det(2π Σ_m))
- * The residual error contains no keys, and only
- * depends on the discrete separator if present.
- */
 static std::shared_ptr<Factor> createDiscreteFactor(
     const DecisionTree<Key, Result> &eliminationResults,
     const DiscreteKeys &discreteSeparator) {
-  auto logProbability = [&](const Result &pair) -> double {
+  auto probability = [&](const Result &pair) -> double {
     const auto &[conditional, factor] = pair;
     static const VectorValues kEmpty;
     // If the factor is not null, it has no keys, just contains the residual.
     if (!factor) return 1.0;  // TODO(dellaert): not loving this.
-
+    return exp(-factor->error(kEmpty)) / conditional->normalizationConstant();
-    // Logspace version of:
-    // exp(-factor->error(kEmpty)) / conditional->normalizationConstant();
-    // We take negative of the logNormalizationConstant `log(1/k)`
-    // to get `log(k)`.
-    return -factor->error(kEmpty) + (-conditional->logNormalizationConstant());
   };
 
-  AlgebraicDecisionTree<Key> logProbabilities(
+  DecisionTree<Key, double> probabilities(eliminationResults, probability);
-      DecisionTree<Key, double>(eliminationResults, logProbability));
-
-  // Perform normalization
-  double max_log = logProbabilities.max();
-  AlgebraicDecisionTree probabilities = DecisionTree<Key, double>(
-      logProbabilities,
-      [&max_log](const double x) { return exp(x - max_log); });
-  probabilities = probabilities.normalize(probabilities.sum());
 
   return std::make_shared<DecisionTreeFactor>(discreteSeparator, probabilities);
 }
@@ -372,12 +356,6 @@ hybridElimination(const HybridGaussianFactorGraph &factors,
   // Perform elimination!
   DecisionTree<Key, Result> eliminationResults(factorGraphTree, eliminate);
 
-  // Create the GaussianMixture from the conditionals
-  GaussianMixture::Conditionals conditionals(
-      eliminationResults, [](const Result &pair) { return pair.first; });
-  auto gaussianMixture = std::make_shared<GaussianMixture>(
-      frontalKeys, continuousSeparator, discreteSeparator, conditionals);
-
   // If there are no more continuous parents we create a DiscreteFactor with the
   // error for each discrete choice. Otherwise, create a GaussianMixtureFactor
   // on the separator, taking care to correct for conditional constants.
@@ -387,6 +365,12 @@ hybridElimination(const HybridGaussianFactorGraph &factors,
           : createGaussianMixtureFactor(eliminationResults, continuousSeparator,
                                         discreteSeparator);
 
+  // Create the GaussianMixture from the conditionals
+  GaussianMixture::Conditionals conditionals(
+      eliminationResults, [](const Result &pair) { return pair.first; });
+  auto gaussianMixture = std::make_shared<GaussianMixture>(
+      frontalKeys, continuousSeparator, discreteSeparator, conditionals);
+
   return {std::make_shared<HybridConditional>(gaussianMixture), newFactor};
 }
 

gtsam/hybrid/tests/testGaussianMixtureFactor.cpp

@@ -22,13 +22,9 @@
 #include <gtsam/discrete/DiscreteValues.h>
 #include <gtsam/hybrid/GaussianMixture.h>
 #include <gtsam/hybrid/GaussianMixtureFactor.h>
-#include <gtsam/hybrid/HybridBayesNet.h>
-#include <gtsam/hybrid/HybridGaussianFactorGraph.h>
 #include <gtsam/hybrid/HybridValues.h>
 #include <gtsam/inference/Symbol.h>
 #include <gtsam/linear/GaussianFactorGraph.h>
-#include <gtsam/nonlinear/PriorFactor.h>
-#include <gtsam/slam/BetweenFactor.h>
 
 // Include for test suite
 #include <CppUnitLite/TestHarness.h>
@@ -60,6 +56,7 @@ TEST(GaussianMixtureFactor, Sum) {
   auto b = Matrix::Zero(2, 1);
   Vector2 sigmas;
   sigmas << 1, 2;
+  auto model = noiseModel::Diagonal::Sigmas(sigmas, true);
 
   auto f10 = std::make_shared<JacobianFactor>(X(1), A1, X(2), A2, b);
   auto f11 = std::make_shared<JacobianFactor>(X(1), A1, X(2), A2, b);
@@ -109,8 +106,7 @@ TEST(GaussianMixtureFactor, Printing) {
   GaussianMixtureFactor mixtureFactor({X(1), X(2)}, {m1}, factors);
 
   std::string expected =
-      R"(GaussianMixtureFactor
+      R"(Hybrid [x1 x2; 1]{
-Hybrid [x1 x2; 1]{
  Choice(1) 
  0 Leaf :
   A[x1] = [
@@ -182,8 +178,7 @@ TEST(GaussianMixtureFactor, Error) {
   continuousValues.insert(X(2), Vector2(1, 1));
 
   // error should return a tree of errors, with nodes for each discrete value.
-  AlgebraicDecisionTree<Key> error_tree =
+  AlgebraicDecisionTree<Key> error_tree = mixtureFactor.errorTree(continuousValues);
-      mixtureFactor.errorTree(continuousValues);
 
   std::vector<DiscreteKey> discrete_keys = {m1};
   // Error values for regression test
@@ -196,240 +191,8 @@ TEST(GaussianMixtureFactor, Error) {
   DiscreteValues discreteValues;
   discreteValues[m1.first] = 1;
   EXPECT_DOUBLES_EQUAL(
-      4.0, mixtureFactor.error({continuousValues, discreteValues}), 1e-9);
+      4.0, mixtureFactor.error({continuousValues, discreteValues}),
-}
+      1e-9);
-
-/* ************************************************************************* */
-// Test components with differing means
-TEST(GaussianMixtureFactor, DifferentMeans) {
-  DiscreteKey m1(M(1), 2), m2(M(2), 2);
-
-  Values values;
-  double x1 = 0.0, x2 = 1.75, x3 = 2.60;
-  values.insert(X(1), x1);
-  values.insert(X(2), x2);
-  values.insert(X(3), x3);
-
-  auto model0 = noiseModel::Isotropic::Sigma(1, 1e-0);
-  auto model1 = noiseModel::Isotropic::Sigma(1, 1e-0);
-  auto prior_noise = noiseModel::Isotropic::Sigma(1, 1e-0);
-
-  auto f0 = std::make_shared<BetweenFactor<double>>(X(1), X(2), 0.0, model0)
-                ->linearize(values);
-  auto f1 = std::make_shared<BetweenFactor<double>>(X(1), X(2), 2.0, model1)
-                ->linearize(values);
-  std::vector<GaussianFactor::shared_ptr> factors{f0, f1};
-
-  GaussianMixtureFactor mixtureFactor({X(1), X(2)}, {m1}, factors, true);
-  HybridGaussianFactorGraph hfg;
-  hfg.push_back(mixtureFactor);
-
-  f0 = std::make_shared<BetweenFactor<double>>(X(2), X(3), 0.0, model0)
-           ->linearize(values);
-  f1 = std::make_shared<BetweenFactor<double>>(X(2), X(3), 2.0, model1)
-           ->linearize(values);
-  std::vector<GaussianFactor::shared_ptr> factors23{f0, f1};
-  hfg.push_back(GaussianMixtureFactor({X(2), X(3)}, {m2}, factors23, true));
-
-  auto prior = PriorFactor<double>(X(1), x1, prior_noise).linearize(values);
-  hfg.push_back(prior);
-
-  hfg.push_back(PriorFactor<double>(X(2), 2.0, prior_noise).linearize(values));
-
-  auto bn = hfg.eliminateSequential();
-  HybridValues actual = bn->optimize();
-
-  HybridValues expected(
-      VectorValues{
-          {X(1), Vector1(0.0)}, {X(2), Vector1(0.25)}, {X(3), Vector1(-0.6)}},
-      DiscreteValues{{M(1), 1}, {M(2), 0}});
-
-  EXPECT(assert_equal(expected, actual));
-
-  {
-    DiscreteValues dv{{M(1), 0}, {M(2), 0}};
-    VectorValues cont = bn->optimize(dv);
-    double error = bn->error(HybridValues(cont, dv));
-    // regression
-    EXPECT_DOUBLES_EQUAL(1.77418393408, error, 1e-9);
-  }
-  {
-    DiscreteValues dv{{M(1), 0}, {M(2), 1}};
-    VectorValues cont = bn->optimize(dv);
-    double error = bn->error(HybridValues(cont, dv));
-    // regression
-    EXPECT_DOUBLES_EQUAL(1.77418393408, error, 1e-9);
-  }
-  {
-    DiscreteValues dv{{M(1), 1}, {M(2), 0}};
-    VectorValues cont = bn->optimize(dv);
-    double error = bn->error(HybridValues(cont, dv));
-    // regression
-    EXPECT_DOUBLES_EQUAL(1.10751726741, error, 1e-9);
-  }
-  {
-    DiscreteValues dv{{M(1), 1}, {M(2), 1}};
-    VectorValues cont = bn->optimize(dv);
-    double error = bn->error(HybridValues(cont, dv));
-    // regression
-    EXPECT_DOUBLES_EQUAL(1.10751726741, error, 1e-9);
-  }
-}
-
-/* ************************************************************************* */
-/**
- * @brief Test components with differing covariances.
- * The factor graph is
- *     *-X1-*-X2
- *          |
- *          M1
- */
-TEST(GaussianMixtureFactor, DifferentCovariances) {
-  DiscreteKey m1(M(1), 2);
-
-  Values values;
-  double x1 = 1.0, x2 = 1.0;
-  values.insert(X(1), x1);
-  values.insert(X(2), x2);
-
-  double between = 0.0;
-
-  auto model0 = noiseModel::Isotropic::Sigma(1, 1e2);
-  auto model1 = noiseModel::Isotropic::Sigma(1, 1e-2);
-  auto prior_noise = noiseModel::Isotropic::Sigma(1, 1e-3);
-
-  auto f0 =
-      std::make_shared<BetweenFactor<double>>(X(1), X(2), between, model0);
-  auto f1 =
-      std::make_shared<BetweenFactor<double>>(X(1), X(2), between, model1);
-  std::vector<NonlinearFactor::shared_ptr> factors{f0, f1};
-
-  // Create via toFactorGraph
-  using symbol_shorthand::Z;
-  Matrix H0_1, H0_2, H1_1, H1_2;
-  Vector d0 = f0->evaluateError(x1, x2, &H0_1, &H0_2);
-  std::vector<std::pair<Key, Matrix>> terms0 = {{Z(1), gtsam::I_1x1 /*Rx*/},
-                                                //
-                                                {X(1), H0_1 /*Sp1*/},
-                                                {X(2), H0_2 /*Tp2*/}};
-
-  Vector d1 = f1->evaluateError(x1, x2, &H1_1, &H1_2);
-  std::vector<std::pair<Key, Matrix>> terms1 = {{Z(1), gtsam::I_1x1 /*Rx*/},
-                                                //
-                                                {X(1), H1_1 /*Sp1*/},
-                                                {X(2), H1_2 /*Tp2*/}};
-  gtsam::GaussianMixtureFactor gmf(
-      {X(1), X(2)}, {m1},
-      {std::make_shared<JacobianFactor>(X(1), H0_1, X(2), H0_2, -d0, model0),
-       std::make_shared<JacobianFactor>(X(1), H1_1, X(2), H1_2, -d1, model1)},
-      true);
-
-  // Create FG with single GaussianMixtureFactor
-  HybridGaussianFactorGraph mixture_fg;
-  mixture_fg.add(gmf);
-
-  // Linearized prior factor on X1
-  auto prior = PriorFactor<double>(X(1), x1, prior_noise).linearize(values);
-  mixture_fg.push_back(prior);
-
-  auto hbn = mixture_fg.eliminateSequential();
-  // hbn->print();
-
-  VectorValues cv;
-  cv.insert(X(1), Vector1(0.0));
-  cv.insert(X(2), Vector1(0.0));
-
-  // Check that the error values at the MLE point μ.
-  AlgebraicDecisionTree<Key> errorTree = hbn->errorTree(cv);
-
-  DiscreteValues dv0{{M(1), 0}};
-  DiscreteValues dv1{{M(1), 1}};
-
-  // regression
-  EXPECT_DOUBLES_EQUAL(9.90348755254, errorTree(dv0), 1e-9);
-  EXPECT_DOUBLES_EQUAL(0.69314718056, errorTree(dv1), 1e-9);
-
-  DiscreteConditional expected_m1(m1, "0.5/0.5");
-  DiscreteConditional actual_m1 = *(hbn->at(2)->asDiscrete());
-
-  EXPECT(assert_equal(expected_m1, actual_m1));
-}
-
-/* ************************************************************************* */
-/**
- * @brief Test components with differing covariances
- * but with a Bayes net P(Z|X, M) converted to a FG.
- */
-TEST(GaussianMixtureFactor, DifferentCovariances2) {
-  DiscreteKey m1(M(1), 2);
-
-  Values values;
-  double x1 = 1.0, x2 = 1.0;
-  values.insert(X(1), x1);
-  values.insert(X(2), x2);
-
-  double between = 0.0;
-
-  auto model0 = noiseModel::Isotropic::Sigma(1, 1e2);
-  auto model1 = noiseModel::Isotropic::Sigma(1, 1e-2);
-  auto prior_noise = noiseModel::Isotropic::Sigma(1, 1e-3);
-
-  auto f0 =
-      std::make_shared<BetweenFactor<double>>(X(1), X(2), between, model0);
-  auto f1 =
-      std::make_shared<BetweenFactor<double>>(X(1), X(2), between, model1);
-  std::vector<NonlinearFactor::shared_ptr> factors{f0, f1};
-
-  // Create via toFactorGraph
-  using symbol_shorthand::Z;
-  Matrix H0_1, H0_2, H1_1, H1_2;
-  Vector d0 = f0->evaluateError(x1, x2, &H0_1, &H0_2);
-  std::vector<std::pair<Key, Matrix>> terms0 = {{Z(1), gtsam::I_1x1 /*Rx*/},
-                                                //
-                                                {X(1), H0_1 /*Sp1*/},
-                                                {X(2), H0_2 /*Tp2*/}};
-
-  Vector d1 = f1->evaluateError(x1, x2, &H1_1, &H1_2);
-  std::vector<std::pair<Key, Matrix>> terms1 = {{Z(1), gtsam::I_1x1 /*Rx*/},
-                                                //
-                                                {X(1), H1_1 /*Sp1*/},
-                                                {X(2), H1_2 /*Tp2*/}};
-  auto gm = new gtsam::GaussianMixture(
-      {Z(1)}, {X(1), X(2)}, {m1},
-      {std::make_shared<GaussianConditional>(terms0, 1, -d0, model0),
-       std::make_shared<GaussianConditional>(terms1, 1, -d1, model1)});
-  gtsam::HybridBayesNet bn;
-  bn.emplace_back(gm);
-
-  gtsam::VectorValues measurements;
-  measurements.insert(Z(1), gtsam::Z_1x1);
-  // Create FG with single GaussianMixtureFactor
-  HybridGaussianFactorGraph mixture_fg = bn.toFactorGraph(measurements);
-
-  // Linearized prior factor on X1
-  auto prior = PriorFactor<double>(X(1), x1, prior_noise).linearize(values);
-  mixture_fg.push_back(prior);
-
-  auto hbn = mixture_fg.eliminateSequential();
-
-  VectorValues cv;
-  cv.insert(X(1), Vector1(0.0));
-  cv.insert(X(2), Vector1(0.0));
-
-  // Check that the error values at the MLE point μ.
-  AlgebraicDecisionTree<Key> errorTree = hbn->errorTree(cv);
-
-  DiscreteValues dv0{{M(1), 0}};
-  DiscreteValues dv1{{M(1), 1}};
-
-  // regression
-  EXPECT_DOUBLES_EQUAL(9.90348755254, errorTree(dv0), 1e-9);
-  EXPECT_DOUBLES_EQUAL(0.69314718056, errorTree(dv1), 1e-9);
-
-  DiscreteConditional expected_m1(m1, "0.5/0.5");
-  DiscreteConditional actual_m1 = *(hbn->at(2)->asDiscrete());
-
-  EXPECT(assert_equal(expected_m1, actual_m1));
 }
 
 /* ************************************************************************* */

gtsam/hybrid/tests/testHybridNonlinearFactorGraph.cpp

@@ -510,7 +510,6 @@ factor 0:
   b = [ -10 ]
   No noise model
 factor 1: 
-GaussianMixtureFactor
 Hybrid [x0 x1; m0]{
  Choice(m0) 
  0 Leaf :
@@ -535,7 +534,6 @@ Hybrid [x0 x1; m0]{
 
 }
 factor 2: 
-GaussianMixtureFactor
 Hybrid [x1 x2; m1]{
  Choice(m1) 
  0 Leaf :
@@ -677,8 +675,6 @@ factor 6:  P( m1 | m0 ):
 size: 3
 conditional 0: Hybrid  P( x0 | x1 m0)
  Discrete Keys = (m0, 2), 
- logNormalizationConstant: 1.38862
-
  Choice(m0) 
  0 Leaf p(x0 | x1)
   R = [ 10.0499 ]
@@ -696,8 +692,6 @@ conditional 0: Hybrid  P( x0 | x1 m0)
 
 conditional 1: Hybrid  P( x1 | x2 m0 m1)
  Discrete Keys = (m0, 2), (m1, 2), 
- logNormalizationConstant: 1.3935
-
  Choice(m1) 
  0 Choice(m0) 
  0 0 Leaf p(x1 | x2)
@@ -731,8 +725,6 @@ conditional 1: Hybrid  P( x1 | x2 m0 m1)
 
 conditional 2: Hybrid  P( x2 | m0 m1)
  Discrete Keys = (m0, 2), (m1, 2), 
- logNormalizationConstant: 1.38857
-
  Choice(m1) 
  0 Choice(m0) 
  0 0 Leaf p(x2)

gtsam/hybrid/tests/testMixtureFactor.cpp

@@ -18,9 +18,6 @@
 
 #include <gtsam/base/TestableAssertions.h>
 #include <gtsam/discrete/DiscreteValues.h>
-#include <gtsam/hybrid/HybridBayesNet.h>
-#include <gtsam/hybrid/HybridGaussianFactorGraph.h>
-#include <gtsam/hybrid/HybridNonlinearFactorGraph.h>
 #include <gtsam/hybrid/MixtureFactor.h>
 #include <gtsam/inference/Symbol.h>
 #include <gtsam/slam/BetweenFactor.h>

gtsam/linear/VectorValues.h

@@ -263,7 +263,7 @@ namespace gtsam {
     /** equals required by Testable for unit testing */
     bool equals(const VectorValues& x, double tol = 1e-9) const;
 
-    /// Check exact equality.
+    /// Check equality.
     friend bool operator==(const VectorValues& lhs, const VectorValues& rhs) {
       return lhs.equals(rhs);
     }

gtsam/linear/tests/testGaussianBayesNet.cpp

@@ -80,6 +80,8 @@ TEST(GaussianBayesNet, Evaluate1) {
                        smallBayesNet.at(0)->logNormalizationConstant() +
                            smallBayesNet.at(1)->logNormalizationConstant(),
                        1e-9);
+  EXPECT_DOUBLES_EQUAL(log(constant), smallBayesNet.logNormalizationConstant(),
+                       1e-9);
   const double actual = smallBayesNet.evaluate(mean);
   EXPECT_DOUBLES_EQUAL(constant, actual, 1e-9);
 }