Initial pass at implementing test fixtures

src/HealthGPS.Tests/CMakeLists.txt

@@ -39,6 +39,7 @@ target_sources(
  "RelativeRiskLookup.Test.cpp"
  "RuntimeMetric.Test.cpp"
  "Scenario.Test.cpp"
+ "StaticLinearModel.Test.cpp"
  "TestMain.cpp"
  "WeightModel.Test.cpp"
  "CountryModule.h"

src/HealthGPS.Tests/StaticLinearModel.Test.cpp

@@ -0,0 +1,162 @@
+#include "pch.h"
+
+#include "HealthGPS/static_linear_model.h"
+
+// Create test fixture for a StaticLinearModel class instance
+class StaticLinearModelTestFixture : public::testing::Test {
+ public:
+
+ hgps::RiskFactorSexAgeTable expected;
+ std::vector<hgps::core::Identifier> names;
+ std::vector<hgps::LinearModelParams> models;
+ std::vector<double> lambda;
+ std::vector<double> stddev;
+ Eigen::Matrix2d correlation;
+ Eigen::MatrixXd cholesky;
+ std::vector<hgps::LinearModelParams> policy_models;
+ std::vector<hgps::core::DoubleInterval> policy_ranges;
+ Eigen::MatrixXd policy_cholesky;
+
+ StaticLinearModel* testModel;
+
+ void SetUp() override {
+ auto expected = hgps::RiskFactorSexAgeTable{};
+ std::vector<double> factor_means_male = {200.,40.,30.,2000.,1.,1000.,3.,50.};
+ std::vector<double> factor_means_female = {200.,40.,30.,2000.,1.,1000.,3.,50.};
+
+ std::unordered_map<hgps::core::Identifier, std::vector<double>> factors_male = {
+ {hgps::core::Identifier{"FoodCarbohydrate"}, factor_means_male},
+ {hgps::core::Identifier{"FoodFat"}, factor_means_male}
+ };
+
+ std::unordered_map<hgps::core::Identifier, std::vector<double>> factors_female = {
+ {hgps::core::Identifier{"FoodCarbohydrate"}, factor_means_female},
+ {hgps::core::Identifier{"FoodFat"}, factor_means_female}
+ };
+
+ expected.emplace_row(hgps::core::Gender::male, factors_male);
+ expected.emplace_row(hgps::core::Gender::female, factors_female);
+
+ std::vector<hgps::core::Identifier> names = {"FoodCarbohydrate", "FoodFat"};
+
+ // Define models
+ // We will define 2 RiskFactorModels: carb_model and fat_model 
+ std::vector<hgps::LinearModelParams> models;
+ hgps::LinearModelParams carb_model;
+ carb_model.intercept = -0.241505734056409; 
+ carb_model.coefficients = {
+ {"Gender", 0.00191412437431455},
+ {"Age", -0.000545257990453302},
+ {"Age2", 3.06434596341924e-6},
+ {"Age3", 5.24468215546687e-9},
+ {"Sector", 0.0967308121487847},
+ {"Income", 0.0810023788842083}
+ };
+
+ hgps::LinearModelParams fat_model;
+ fat_model.intercept = -0.693521187677;
+ fat_model.coefficients = {
+ {"Gender", 0.00385126434589752},
+ {"Age", -0.00433711587715954},
+ {"Age2", 3.11799231385975e-5},
+ {"Age3", 3.4291809335544e-8},
+ {"Sector", -0.0629531974852436},
+ {"Income", 0.373256996730791}
+ };
+
+ models.emplace_back(std::move(carb_model));
+ models.emplace_back(std::move(fat_model));
+
+ std::vector<double> lambda = {0.262626262626263, 0.141414141414141};
+ std::vector<double> stddev = {0.337810256621877, 0.436763527499362};
+
+ Eigen::Matrix2d correlation;
+ correlation << 1.0, 0.322065425, 0.322065425,1.0;
+ auto cholesky = Eigen::MatrixXd{Eigen::LLT<Eigen::Matrix2d>{correlation}.matrixL()};
+
+ Eigen::MatrixXd policy_covariance;
+ policy_covariance << 0.1,0.1,0.1,0.41;
+ auto policy_cholesky = Eigen::MatrixXd{Eigen::LLT<Eigen::MatrixXd>{policy_covariance}.matrixL()};
+
+ // Define policy models
+ std::vector<hgps::LinearModelParams> policy_models;
+ std::vector<hgps::core::DoubleInterval> policy_ranges;
+
+ // We need 2 policy models for the 2 risk factor models
+ hgps::LinearModelParams carb_policy_model;
+ carb_policy_model.intercept = -1.29385215316281;
+ carb_policy_model.coefficients = {
+ {"Gender", 0.00845121355575458},
+ {"Age", 0.000386003599279214},
+ {"Sector", 0.0621763530502617},
+ {"Income", 0.0560558815729017}
+ };
+ carb_policy_model.log_coefficients = {
+ {"FoodCarbohydrate", 0.577638667323293},
+ {"FoodFat", -0.0801300827856402},
+ {"FoodProtein", -0.480946893299471},
+ {"FoodSodium", -0.0396319735508116}
+ };
+ policy_ranges.emplace_back(hgps::core::DoubleInterval{-2.31063, 0.37526});
+
+ hgps::LinearModelParams fat_policy_model;
+ fat_policy_model.intercept = -0.734121205356728;
+ fat_policy_model.coefficients = {
+ {"Gender", 0.023422206650121},
+ {"Age", 0.000557687923688474},
+ {"Sector", 0.00166651997223223},
+ {"Income", -0.498755023022772}
+ };
+ fat_policy_model.log_coefficients = {
+ {"FoodCarbohydrate", 0.652615267035516},
+ {"FoodFat", 0.743457905132894},
+ {"FoodProtein", -1.2648185364167},
+ {"FoodSodium", -0.17181077457271}
+ };
+ policy_ranges.emplace_back(hgps::core::DoubleInterval{-4.87205, -0.10346});
+
+ policy_models.emplace_back(std::move(carb_policy_model));
+ policy_models.emplace_back(std::move(fat_policy_model));
+
+ const double info_speed = 0.1;
+
+ std::unordered_map<hgps::core::Identifier, std::unordered_map<hgps::core::Gender, double>> rural_prevalence;
+ rural_prevalence["Under18"] = {{hgps::core::Gender::female, 0.65},{hgps::core::Gender::male, 0.65}};
+ rural_prevalence["Over18"] = {{hgps::core::Gender::female, 0.6},{hgps::core::Gender::male, 0.6}};
+
+ std::unordered_map<hgps::core::Income, hgps::LinearModelParams> income_models;
+
+ hgps::LinearModelParams income_model_low;
+ income_model_low.intercept = 0.0;
+ income_model_low.coefficients = {};
+
+ hgps::LinearModelParams income_model_middle;
+ income_model_middle.intercept = -0.0023671151435328;
+ income_model_middle.coefficients = {
+ {"Gender", 0.0244784583947992},
+ {"Over18", 0.276669591410528},
+ {"Sector", -0.491794449118805}
+ };
+
+ hgps::LinearModelParams income_model_high;
+ income_model_high.intercept = 0.0187952471153142;
+ income_model_high.coefficients = {
+ {"Gender", 0.02018185354194},
+ {"Over18", 0.607099252829797},
+ {"Sector", -1.5870837069653}
+ };
+
+ income_models.emplace(hgps::core::Income::low, std::move(income_model_low));
+ income_models.emplace(hgps::core::Income::middle, std::move(income_model_middle));
+ income_models.emplace(hgps::core::Income::high, std::move(income_model_high));
+
+ const double physical_activity_stddev = 0.06;
+
+ };
+
+};
+
+TEST_F(StaticLinearModelTestFixture, InitialiseFactors) {
+ // Test logic goes here
+
+}