diff --git a/.gitmodules b/.gitmodules index 1012f4edb..0725523e7 100644 --- a/.gitmodules +++ b/.gitmodules @@ -2,24 +2,24 @@ path = deps/fast_double_parser url = https://github.com/lemire/fast_double_parser.git branch = master - shallow = true + shallow = true [submodule "deps/fmt"] path = deps/fmt url = https://github.com/fmtlib/fmt.git branch = master - shallow = true + shallow = true [submodule "deps/boost_math"] path = deps/boost_math url = https://github.com/boostorg/math.git branch = master - shallow = true + shallow = true [submodule "deps/eigen"] path = deps/eigen url = https://gitlab.com/libeigen/eigen.git - branch = 3.4 - shallow = true + branch = 5.0 + shallow = true [submodule "deps/pybind11"] path = deps/pybind11 url = https://github.com/pybind/pybind11.git branch = v2.12 - shallow = true + shallow = true diff --git a/DESCRIPTION b/DESCRIPTION index 8e609128c..31198850e 100644 --- a/DESCRIPTION +++ b/DESCRIPTION @@ -11,6 +11,7 @@ Authors@R: person("Pedro", "Lima", role = "ctb"), person("Entejar", "Alam", role = "ctb"), person("David", "Wachsmuth", role = "ctb"), + person("Michael", "Chirico", role = "ctb"), person("stochtree", "contributors", role = c("cph")), person("Eigen", "contributors", role = c("cph"), comment = "C++ source uses the Eigen library for matrix operations, see inst/COPYRIGHTS"), person("xgboost", "contributors", role = c("cph"), comment = "C++ tree code and related operations include or are inspired by code from the xgboost library, see inst/COPYRIGHTS"), diff --git a/MANIFEST.in b/MANIFEST.in index 82d41ade6..56e63070c 100644 --- a/MANIFEST.in +++ b/MANIFEST.in @@ -53,21 +53,22 @@ include deps/eigen/Eigen/SparseCore include deps/eigen/Eigen/SparseQR include deps/eigen/Eigen/misc include deps/eigen/Eigen/plugins -recursive-include deps/eigen/Eigen/src/Cholesky *.h -recursive-include deps/eigen/Eigen/src/Core *.h -recursive-include deps/eigen/Eigen/src/Dense *.h -recursive-include deps/eigen/Eigen/src/Eigenvalues *.h -recursive-include deps/eigen/Eigen/src/Geometry *.h -recursive-include deps/eigen/Eigen/src/Householder *.h -recursive-include deps/eigen/Eigen/src/IterativeLinearSolvers *.h -recursive-include deps/eigen/Eigen/src/Jacobi *.h -recursive-include deps/eigen/Eigen/src/LU *.h -recursive-include deps/eigen/Eigen/src/OrderingMethods *.h -recursive-include deps/eigen/Eigen/src/QR *.h -recursive-include deps/eigen/Eigen/src/SVD *.h -recursive-include deps/eigen/Eigen/src/Sparse *.h -recursive-include deps/eigen/Eigen/src/SparseCholesky *.h -recursive-include deps/eigen/Eigen/src/SparseCore *.h -recursive-include deps/eigen/Eigen/src/SparseQR *.h -recursive-include deps/eigen/Eigen/src/misc *.h -recursive-include deps/eigen/Eigen/src/plugins *.h \ No newline at end of file +include deps/eigen/Eigen/Version +recursive-include deps/eigen/Eigen/src/Cholesky *.h *.inc +recursive-include deps/eigen/Eigen/src/Core *.h *.inc +recursive-include deps/eigen/Eigen/src/Dense *.h *.inc +recursive-include deps/eigen/Eigen/src/Eigenvalues *.h *.inc +recursive-include deps/eigen/Eigen/src/Geometry *.h *.inc +recursive-include deps/eigen/Eigen/src/Householder *.h *.inc +recursive-include deps/eigen/Eigen/src/IterativeLinearSolvers *.h *.inc +recursive-include deps/eigen/Eigen/src/Jacobi *.h *.inc +recursive-include deps/eigen/Eigen/src/LU *.h *.inc +recursive-include deps/eigen/Eigen/src/OrderingMethods *.h *.inc +recursive-include deps/eigen/Eigen/src/QR *.h *.inc +recursive-include deps/eigen/Eigen/src/SVD *.h *.inc +recursive-include deps/eigen/Eigen/src/Sparse *.h *.inc +recursive-include deps/eigen/Eigen/src/SparseCholesky *.h *.inc +recursive-include deps/eigen/Eigen/src/SparseCore *.h *.inc +recursive-include deps/eigen/Eigen/src/SparseQR *.h *.inc +recursive-include deps/eigen/Eigen/src/misc *.h *.inc +recursive-include deps/eigen/Eigen/src/plugins *.h *.inc \ No newline at end of file diff --git a/NEWS.md b/NEWS.md index 7c63cf1a6..aedb28a75 100644 --- a/NEWS.md +++ b/NEWS.md @@ -6,11 +6,17 @@ * Fix the move-type and node-selection probabilities in the MCMC grow / prune Metropolis-Hastings acceptance ratios, so that the tree sampler targets the correct posterior over trees [#426](https://github.com/StochasticTree/stochtree/pull/426) * Fix ordering of missing (`NaN`) covariate values when presorting features for the grow-from-root (GFR) sampler, which now sorts `NaN`s last, consistent with how numeric splits route them [#427](https://github.com/StochasticTree/stochtree/pull/427) * Fix the conditional posterior of the random effects variance components, which was missing factors of one half in its inverse gamma shape and scale [#428](https://github.com/StochasticTree/stochtree/pull/428) +* Fix an uninitialized starting value when initializing cloglog forests in R, which could crash or corrupt the outcome data for binary and ordinal models with a cloglog link [#420](https://github.com/StochasticTree/stochtree/pull/420) Fixes in [#425](https://github.com/StochasticTree/stochtree/pull/425), [#426](https://github.com/StochasticTree/stochtree/pull/426), and [#428](https://github.com/StochasticTree/stochtree/pull/428) change sampler output, as does [#427](https://github.com/StochasticTree/stochtree/pull/427) for covariates with missing values, so results for a given random seed will differ from previous versions. ## Documentation and Other Maintenance +* Remove unused variables flagged by compiler warnings [#417](https://github.com/StochasticTree/stochtree/pull/417) +* Mark `Log::Fatal` as `[[noreturn]]`, which silences "may be used uninitialized" compiler warnings [#418](https://github.com/StochasticTree/stochtree/pull/418) +* Explicitly initialize variables in the R forest initialization and sampling code [#421](https://github.com/StochasticTree/stochtree/pull/421) +* Use shallow clones and canonical `.git` URLs for git submodules [#422](https://github.com/StochasticTree/stochtree/pull/422) +* Updated Eigen version to 5.0.1, which also fixes builds on PowerPC without VSX support [#424](https://github.com/StochasticTree/stochtree/pull/424) * Add absolute tolerance to several python floating point comparison tests [#430](https://github.com/StochasticTree/stochtree/pull/430). * Add exact-posterior tests of the MCMC tree sampler in R and Python, which check that a long chain on a small problem reproduces the analytically enumerated posterior over trees [#431](https://github.com/StochasticTree/stochtree/pull/431) diff --git a/cran-bootstrap.R b/cran-bootstrap.R index d23067141..69ef56204 100644 --- a/cran-bootstrap.R +++ b/cran-bootstrap.R @@ -131,9 +131,13 @@ if (pkgdown_build) { if (include_tests) { test_files_src <- list.files("test/R", recursive = TRUE, full.names = TRUE) # Exclude fixture JSON files: large test-only snapshots, not needed on CRAN - test_files_src <- test_files_src[!grepl("/fixtures/.*\\.json$", test_files_src)] + test_files_src <- test_files_src[ + !grepl("/fixtures/.*\\.json$", test_files_src) + ] # Exclude backward-compat tests: they depend on jsonlite and fixture files, not suitable for CRAN - test_files_src <- test_files_src[!grepl("test-serialization-compat\\.R$", test_files_src)] + test_files_src <- test_files_src[ + !grepl("test-serialization-compat\\.R$", test_files_src) + ] test_files_dst <- file.path(cran_dir, gsub("test/R", "tests", test_files_src)) pkg_core_files <- c(pkg_core_files, test_files_src) pkg_core_files_dst <- c(pkg_core_files_dst, test_files_dst) @@ -431,7 +435,8 @@ eigen_modules <- c( "SparseLU", "SparseQR", "misc", - "plugins" + "plugins", + "Version" ) eigen_files_to_vendor_src <- c() eigen_files_to_vendor_dst <- c() @@ -515,11 +520,16 @@ if (all(file.exists(eigen_files_to_vendor_src))) { } # Clean up pragmas that suppress warnings in Eigen and JSON headers -# File 1: Eigen "DisableStupidWarnings" header -cran_eigen_suppress_warnings <- file.path( +# Eigen files: the "DisableStupidWarnings" header, plus the AltiVec (PowerPC) +# packet math header, which gained a GCC diagnostic block in Eigen 5 +cran_eigen_pragma_files <- file.path( cran_dir, - "src/include/Eigen/src/Core/util/DisableStupidWarnings.h" + c( + "src/include/Eigen/src/Core/util/DisableStupidWarnings.h", + "src/include/Eigen/src/Core/arch/AltiVec/PacketMath.h" + ) ) +for (cran_eigen_suppress_warnings in cran_eigen_pragma_files) { eigen_suppress_warnings_lines <- readLines(cran_eigen_suppress_warnings) for (i in 1:length(eigen_suppress_warnings_lines)) { line <- eigen_suppress_warnings_lines[i] @@ -555,3 +565,4 @@ for (i in 1:length(eigen_suppress_warnings_lines)) { ) } writeLines(eigen_suppress_warnings_lines, cran_eigen_suppress_warnings) +} diff --git a/deps/eigen b/deps/eigen index 9df21dc8b..bc3b39870 160000 --- a/deps/eigen +++ b/deps/eigen @@ -1 +1 @@ -Subproject commit 9df21dc8b4b576a7aa5c0094daa8d7e8b8be60f0 +Subproject commit bc3b39870ecb690a623a3f49149a358b95c5781d diff --git a/include/stochtree/random_effects.h b/include/stochtree/random_effects.h index b322a5604..f869f4487 100644 --- a/include/stochtree/random_effects.h +++ b/include/stochtree/random_effects.h @@ -155,7 +155,7 @@ class MultivariateRegressionRandomEffectsModel { group_parameters_ = group_parameters; } void SetGroupParameter(Eigen::VectorXd& group_parameter, int32_t group_id) { - group_parameters_(Eigen::all, group_id) = group_parameter; + group_parameters_(Eigen::placeholders::all, group_id) = group_parameter; } void SetWorkingParameterCovariance(Eigen::MatrixXd& working_parameter_covariance) { working_parameter_covariance_ = working_parameter_covariance; @@ -211,7 +211,7 @@ class MultivariateRegressionRandomEffectsModel { std::int32_t group_ind; for (int i = 0; i < n; i++) { group_ind = tracker.CategoryNumber(group_labels[i]); - output[i] = X(i, Eigen::all) * alpha_diag * group_parameters_(Eigen::all, group_ind); + output[i] = X(i, Eigen::placeholders::all) * alpha_diag * group_parameters_(Eigen::placeholders::all, group_ind); } } @@ -238,7 +238,7 @@ class MultivariateRegressionRandomEffectsModel { std::int32_t group_ind; for (int i = 0; i < n; i++) { group_ind = tracker.CategoryNumber(group_labels[i]); - new_pred = X(i, Eigen::all) * alpha_diag * group_parameters_(Eigen::all, group_ind); + new_pred = X(i, Eigen::placeholders::all) * alpha_diag * group_parameters_(Eigen::placeholders::all, group_ind); new_resid = residual.GetElement(i) - new_pred; residual.SetElement(i, new_resid); tracker.SetPrediction(i, new_pred); diff --git a/man/stochtree-package.Rd b/man/stochtree-package.Rd index 9c285fff0..78d39e5dd 100644 --- a/man/stochtree-package.Rd +++ b/man/stochtree-package.Rd @@ -33,6 +33,7 @@ Other contributors: \item Pedro Lima [contributor] \item Entejar Alam [contributor] \item David Wachsmuth [contributor] + \item Michael Chirico [contributor] \item stochtree contributors [copyright holder] \item Eigen contributors (C++ source uses the Eigen library for matrix operations, see inst/COPYRIGHTS) [copyright holder] \item xgboost contributors (C++ tree code and related operations include or are inspired by code from the xgboost library, see inst/COPYRIGHTS) [copyright holder] diff --git a/src/random_effects.cpp b/src/random_effects.cpp index 9af61a870..5d03eb13b 100644 --- a/src/random_effects.cpp +++ b/src/random_effects.cpp @@ -54,7 +54,7 @@ void RandomEffectsTracker::ResetFromSample(MultivariateRegressionRandomEffectsMo for (int i = 0; i < n; i++) { group_ind = CategoryNumber(group_labels[i]); prev_pred = GetPrediction(i); - new_pred = X(i, Eigen::all) * alpha_diag * xi(Eigen::all, group_ind); + new_pred = X(i, Eigen::placeholders::all) * alpha_diag * xi(Eigen::placeholders::all, group_ind); new_resid = residual.GetElement(i) - new_pred + prev_pred; residual.SetElement(i, new_resid); SetPrediction(i, new_pred); @@ -123,7 +123,7 @@ void MultivariateRegressionRandomEffectsModel::SampleGroupParameters(RandomEffec for (int i = 0; i < num_groups; i++) { posterior_mean = GroupParameterMean(dataset, residual, rfx_tracker, global_variance, i); posterior_covariance = GroupParameterVariance(dataset, residual, rfx_tracker, global_variance, i); - group_parameters_(Eigen::all, i) = normal_sampler_.SampleEigen(posterior_mean, posterior_covariance, gen); + group_parameters_(Eigen::placeholders::all, i) = normal_sampler_.SampleEigen(posterior_mean, posterior_covariance, gen); } } @@ -155,9 +155,9 @@ Eigen::VectorXd MultivariateRegressionRandomEffectsModel::WorkingParameterMean(R Eigen::MatrixXd xi = group_parameters_; for (int i = 0; i < num_groups; i++) { observation_indices = rfx_tracker.NodeIndicesInternalIndex(i); - X_group = X(observation_indices, Eigen::all); - y_group = y(observation_indices, Eigen::all); - xi_group = xi(Eigen::all, i); + X_group = X(observation_indices, Eigen::placeholders::all); + y_group = y(observation_indices, Eigen::placeholders::all); + xi_group = xi(Eigen::placeholders::all, i); posterior_denominator += ((xi_group).asDiagonal() * X_group.transpose() * X_group * (xi_group).asDiagonal()) / global_variance; posterior_numerator += (xi_group).asDiagonal() * X_group.transpose() * y_group / global_variance; } @@ -178,9 +178,9 @@ Eigen::MatrixXd MultivariateRegressionRandomEffectsModel::WorkingParameterVarian Eigen::MatrixXd xi = group_parameters_; for (int i = 0; i < num_groups; i++) { observation_indices = rfx_tracker.NodeIndicesInternalIndex(i); - X_group = X(observation_indices, Eigen::all); - y_group = y(observation_indices, Eigen::all); - xi_group = xi(Eigen::all, i); + X_group = X(observation_indices, Eigen::placeholders::all); + y_group = y(observation_indices, Eigen::placeholders::all); + xi_group = xi(Eigen::placeholders::all, i); posterior_denominator += ((xi_group).asDiagonal() * X_group.transpose() * X_group * (xi_group).asDiagonal()) / (global_variance); } return posterior_denominator.inverse(); @@ -195,8 +195,8 @@ Eigen::VectorXd MultivariateRegressionRandomEffectsModel::GroupParameterMean(Ran Eigen::MatrixXd posterior_denominator = group_parameter_covariance_.inverse(); Eigen::VectorXd posterior_numerator = Eigen::VectorXd::Zero(num_components); std::vector observation_indices = rfx_tracker.NodeIndicesInternalIndex(group_id); - Eigen::MatrixXd X_group = X(observation_indices, Eigen::all); - Eigen::VectorXd y_group = y(observation_indices, Eigen::all); + Eigen::MatrixXd X_group = X(observation_indices, Eigen::placeholders::all); + Eigen::VectorXd y_group = y(observation_indices, Eigen::placeholders::all); posterior_denominator += ((alpha).asDiagonal() * X_group.transpose() * X_group * (alpha).asDiagonal()) / (global_variance); posterior_numerator += (alpha).asDiagonal() * X_group.transpose() * y_group / global_variance; return posterior_denominator.inverse() * posterior_numerator; @@ -211,8 +211,8 @@ Eigen::MatrixXd MultivariateRegressionRandomEffectsModel::GroupParameterVariance Eigen::MatrixXd posterior_denominator = group_parameter_covariance_.inverse(); // Eigen::VectorXd posterior_numerator = Eigen::VectorXd::Zero(num_components); std::vector observation_indices = rfx_tracker.NodeIndicesInternalIndex(group_id); - Eigen::MatrixXd X_group = X(observation_indices, Eigen::all); -// Eigen::VectorXd y_group = y(observation_indices, Eigen::all); + Eigen::MatrixXd X_group = X(observation_indices, Eigen::placeholders::all); +// Eigen::VectorXd y_group = y(observation_indices, Eigen::placeholders::all); posterior_denominator += ((alpha).asDiagonal() * X_group.transpose() * X_group * (alpha).asDiagonal()) / (global_variance); // posterior_numerator += (alpha).asDiagonal() * X_group.transpose() * y_group; return posterior_denominator.inverse(); diff --git a/test/cpp/test_random_effects.cpp b/test/cpp/test_random_effects.cpp index 39b686c3a..4db4926e2 100644 --- a/test/cpp/test_random_effects.cpp +++ b/test/cpp/test_random_effects.cpp @@ -134,8 +134,8 @@ TEST(RandomEffects, Construction) { Eigen::MatrixXd sigma(test_dataset.rfx_basis_cols, test_dataset.rfx_basis_cols); alpha << 1.5; xi << 2, 4; - Eigen::VectorXd xi0 = xi(Eigen::all, 0); - Eigen::VectorXd xi1 = xi(Eigen::all, 1); + Eigen::VectorXd xi0 = xi(Eigen::placeholders::all, 0); + Eigen::VectorXd xi1 = xi(Eigen::placeholders::all, 1); sigma << 1; model.SetWorkingParameter(alpha); model.SetGroupParameter(xi0, 0); @@ -148,8 +148,8 @@ TEST(RandomEffects, Construction) { // Change values and push a second "sample" to the container alpha << 2.0; xi << 1, 3; - xi0 = xi(Eigen::all, 0); - xi1 = xi(Eigen::all, 1); + xi0 = xi(Eigen::placeholders::all, 0); + xi1 = xi(Eigen::placeholders::all, 1); sigma << 1; model.SetWorkingParameter(alpha); model.SetGroupParameter(xi0, 0); @@ -200,9 +200,9 @@ TEST(RandomEffects, Computation) { Eigen::MatrixXd sigma(test_dataset.rfx_basis_cols, test_dataset.rfx_basis_cols); alpha << 1., 1.; xi << 1., 1., 1., 1., 1., 1.; - Eigen::VectorXd xi0 = xi(Eigen::all, 0); - Eigen::VectorXd xi1 = xi(Eigen::all, 1); - Eigen::VectorXd xi2 = xi(Eigen::all, 2); + Eigen::VectorXd xi0 = xi(Eigen::placeholders::all, 0); + Eigen::VectorXd xi1 = xi(Eigen::placeholders::all, 1); + Eigen::VectorXd xi2 = xi(Eigen::placeholders::all, 2); sigma << 1, 0, 0, 1; model.SetWorkingParameter(alpha); model.SetGroupParameter(xi0, 0); @@ -256,8 +256,8 @@ TEST(RandomEffects, Predict) { Eigen::MatrixXd sigma(test_dataset.rfx_basis_cols, test_dataset.rfx_basis_cols); alpha << 1.5; xi << 2, 4; - Eigen::VectorXd xi0 = xi(Eigen::all, 0); - Eigen::VectorXd xi1 = xi(Eigen::all, 1); + Eigen::VectorXd xi0 = xi(Eigen::placeholders::all, 0); + Eigen::VectorXd xi1 = xi(Eigen::placeholders::all, 1); sigma << 1; model.SetWorkingParameter(alpha); model.SetGroupParameter(xi0, 0); @@ -270,8 +270,8 @@ TEST(RandomEffects, Predict) { // Change values and push a second "sample" to the container alpha << 2.0; xi << 1, 3; - xi0 = xi(Eigen::all, 0); - xi1 = xi(Eigen::all, 1); + xi0 = xi(Eigen::placeholders::all, 0); + xi1 = xi(Eigen::placeholders::all, 1); sigma << 1; model.SetWorkingParameter(alpha); model.SetGroupParameter(xi0, 0); @@ -318,8 +318,8 @@ TEST(RandomEffects, Serialization) { Eigen::MatrixXd sigma(test_dataset.rfx_basis_cols, test_dataset.rfx_basis_cols); alpha << 1.5; xi << 2, 4; - Eigen::VectorXd xi0 = xi(Eigen::all, 0); - Eigen::VectorXd xi1 = xi(Eigen::all, 1); + Eigen::VectorXd xi0 = xi(Eigen::placeholders::all, 0); + Eigen::VectorXd xi1 = xi(Eigen::placeholders::all, 1); sigma << 1; model.SetWorkingParameter(alpha); model.SetGroupParameter(xi0, 0); @@ -332,8 +332,8 @@ TEST(RandomEffects, Serialization) { // Change values and push a second "sample" to the container alpha << 2.0; xi << 1, 3; - xi0 = xi(Eigen::all, 0); - xi1 = xi(Eigen::all, 1); + xi0 = xi(Eigen::placeholders::all, 0); + xi1 = xi(Eigen::placeholders::all, 1); sigma << 1; model.SetWorkingParameter(alpha); model.SetGroupParameter(xi0, 0);