Merge pull request #50 from probcomp/061324-emilyaf-bernoulli-bool

Change BetaBernoulli SampleType from double to bool.

cxx/distributions/beta_bernoulli.cc

@@ -7,21 +7,21 @@
 
 #include "util_math.hh"
 
-void BetaBernoulli::incorporate(const double& x) {
+void BetaBernoulli::incorporate(const bool& x) {
   assert(x == 0 || x == 1);
   ++N;
   s += x;
 }
 
-void BetaBernoulli::unincorporate(const double& x) {
+void BetaBernoulli::unincorporate(const bool& x) {
   assert(x == 0 || x == 1);
   --N;
   s -= x;
   assert(0 <= s);
   assert(0 <= N);
 }
 
-double BetaBernoulli::logp(const double& x) const {
+double BetaBernoulli::logp(const bool& x) const {
   assert(x == 0 || x == 1);
   double log_denom = log(N + alpha + beta);
   double log_numer = x ? log(s + alpha) : log(N - s + beta);
@@ -34,9 +34,9 @@ double BetaBernoulli::logp_score() const {
   return v1 - v2;
 }
 
-double BetaBernoulli::sample() {
+bool BetaBernoulli::sample() {
   double p = exp(logp(1));
-  std::vector<int> items{0, 1};
+  std::vector<bool> items{false, true};
   std::vector<double> weights{1 - p, p};
   int idx = choice(weights, prng);
   return items[idx];

cxx/distributions/beta_bernoulli.hh

@@ -9,7 +9,7 @@
 
 // TODO(thomaswc, emilyaf): Change BetaBernoulli to use bool instead of
 // double.
-class BetaBernoulli : public Distribution<double> {
+class BetaBernoulli : public Distribution<bool> {
  public:
   double alpha = 1;  // hyperparameter
   double beta = 1;   // hyperparameter
@@ -26,15 +26,15 @@ class BetaBernoulli : public Distribution<double> {
     beta_grid = log_linspace(1e-4, 1e4, 10, true);
   }
 
-  void incorporate(const double& x);
+  void incorporate(const bool& x);
 
-  void unincorporate(const double& x);
+  void unincorporate(const bool& x);
 
-  double logp(const double& x) const;
+  double logp(const bool& x) const;
 
   double logp_score() const;
 
-  double sample();
+  bool sample();
 
   void transition_hyperparameters();
 };

cxx/tests/test_hirm_animals.cc

@@ -75,25 +75,25 @@ int main(int argc, char** argv) {
 
   // Marginally normalized.
   int persiancat = enc["animal"]["persiancat"];
-  auto p0_black_persiancat = hirm.logp({{"black", {persiancat}, 0.}});
-  auto p1_black_persiancat = hirm.logp({{"black", {persiancat}, 1.}});
+  auto p0_black_persiancat = hirm.logp({{"black", {persiancat}, false}});
+  auto p1_black_persiancat = hirm.logp({{"black", {persiancat}, true}});
   assert(abs(logsumexp({p0_black_persiancat, p1_black_persiancat})) < 1e-10);
 
   // Marginally normalized.
   int sheep = enc["animal"]["sheep"];
-  auto p0_solitary_sheep = hirm.logp({{"solitary", {sheep}, 0.}});
-  auto p1_solitary_sheep = hirm.logp({{"solitary", {sheep}, 1.}});
+  auto p0_solitary_sheep = hirm.logp({{"solitary", {sheep}, false}});
+  auto p1_solitary_sheep = hirm.logp({{"solitary", {sheep}, true}});
   assert(abs(logsumexp({p0_solitary_sheep, p1_solitary_sheep})) < 1e-10);
 
   // Jointly normalized.
   auto p00_black_persiancat_solitary_sheep =
-      hirm.logp({{"black", {persiancat}, 0.}, {"solitary", {sheep}, 0.}});
+      hirm.logp({{"black", {persiancat}, false}, {"solitary", {sheep}, false}});
   auto p01_black_persiancat_solitary_sheep =
-      hirm.logp({{"black", {persiancat}, 0.}, {"solitary", {sheep}, 1.}});
+      hirm.logp({{"black", {persiancat}, false}, {"solitary", {sheep}, true}});
   auto p10_black_persiancat_solitary_sheep =
-      hirm.logp({{"black", {persiancat}, 1.}, {"solitary", {sheep}, 0.}});
+      hirm.logp({{"black", {persiancat}, true}, {"solitary", {sheep}, false}});
   auto p11_black_persiancat_solitary_sheep =
-      hirm.logp({{"black", {persiancat}, 1.}, {"solitary", {sheep}, 1.}});
+      hirm.logp({{"black", {persiancat}, true}, {"solitary", {sheep}, true}});
   auto Z = logsumexp({
       p00_black_persiancat_solitary_sheep,
       p01_black_persiancat_solitary_sheep,

cxx/tests/test_irm_two_relations.cc

@@ -80,10 +80,10 @@ int main(int argc, char** argv) {
     assert(l.size() == 2);
     auto x1 = l.at(0);
     auto x2 = l.at(1);
-    auto p0 = std::get<T_r>(irm.relations.at("R1"))->logp({x1, x2}, 0.);
-    auto p0_irm = irm.logp({{"R1", {x1, x2}, 0.}});
+    auto p0 = std::get<T_r>(irm.relations.at("R1"))->logp({x1, x2}, false);
+    auto p0_irm = irm.logp({{"R1", {x1, x2}, false}});
     assert(abs(p0 - p0_irm) < 1e-10);
-    auto p1 = std::get<T_r>(irm.relations.at("R1"))->logp({x1, x2}, 1.);
+    auto p1 = std::get<T_r>(irm.relations.at("R1"))->logp({x1, x2}, true);
     auto Z = logsumexp({p0, p1});
     assert(abs(Z) < 1e-10);
     assert(abs(exp(p0) - expected_p0[x1].at(x2)) < .1);
@@ -94,10 +94,10 @@ int main(int argc, char** argv) {
     auto x1 = l.at(0);
     auto x2 = l.at(1);
     auto x3 = l.at(2);
-    auto p00 = irm.logp({{"R1", {x1, x2}, 0.}, {"R1", {x1, x3}, 0.}});
-    auto p01 = irm.logp({{"R1", {x1, x2}, 0.}, {"R1", {x1, x3}, 1.}});
-    auto p10 = irm.logp({{"R1", {x1, x2}, 1.}, {"R1", {x1, x3}, 0.}});
-    auto p11 = irm.logp({{"R1", {x1, x2}, 1.}, {"R1", {x1, x3}, 1.}});
+    auto p00 = irm.logp({{"R1", {x1, x2}, false}, {"R1", {x1, x3}, false}});
+    auto p01 = irm.logp({{"R1", {x1, x2}, false}, {"R1", {x1, x3}, true}});
+    auto p10 = irm.logp({{"R1", {x1, x2}, true}, {"R1", {x1, x3}, false}});
+    auto p11 = irm.logp({{"R1", {x1, x2}, true}, {"R1", {x1, x3}, true}});
     auto Z = logsumexp({p00, p01, p10, p11});
     assert(abs(Z) < 1e-10);
   }

cxx/tests/test_misc.cc

@@ -75,7 +75,7 @@ int main(int argc, char** argv) {
     auto value = std::get<2>(i);
     auto item = std::get<1>(i);
     printf("incorporating %s ", relation.c_str());
-    printf("%1.f ", std::get<double>(value));
+    printf("%d ", std::get<bool>(value));
     int counter = 0;
     T_items items_code;
     for (auto const& item : std::get<1>(i)) {

cxx/util_distribution_variant.cc

@@ -20,7 +20,7 @@ ObservationVariant observation_string_to_value(
     case DistributionEnum::normal:
       return std::stod(value_str);
     case DistributionEnum::bernoulli:
-      return std::stod(value_str);
+      return static_cast<bool>(std::stoi(value_str));
     case DistributionEnum::categorical:
       return std::stoi(value_str);
     case DistributionEnum::bigram:

cxx/util_distribution_variant.hh

@@ -28,7 +28,7 @@ template <typename DistributionType>
 class Relation;
 
 // Set of all distribution sample types.
-using ObservationVariant = std::variant<double, int, std::string>;
+using ObservationVariant = std::variant<double, int, bool, std::string>;
 
 using DistributionVariant =
     std::variant<BetaBernoulli*, Bigram*, DirichletCategorical*, Normal*>;