Updated interface and function names

andrewherren · andrewherren · commit 36aa10664570 · 2025-01-28T22:55:11.000-06:00
diff --git a/NAMESPACE b/NAMESPACE
@@ -25,16 +25,16 @@ export(createBCFModelFromJsonString)
 export(createCppJson)
 export(createCppJsonFile)
 export(createCppJsonString)
+export(createCppRNG)
 export(createForest)
-export(createForestContainer)
 export(createForestCovariates)
 export(createForestCovariatesFromMetadata)
 export(createForestDataset)
 export(createForestModel)
+export(createForestSamples)
 export(createOutcome)
 export(createPreprocessorFromJson)
 export(createPreprocessorFromJsonString)
-export(createRNG)
 export(createRandomEffectSamples)
 export(createRandomEffectsDataset)
 export(createRandomEffectsModel)
diff --git a/R/bart.R b/R/bart.R
@@ -536,7 +536,7 @@ bart <- function(X_train, y_train, leaf_basis_train = NULL, rfx_group_ids_train
     
     # Random number generator (std::mt19937)
     if (is.null(random_seed)) random_seed = sample(1:10000,1,F)
-    rng <- createRNG(random_seed)
+    rng <- createCppRNG(random_seed)
     
     # Sampling data structures
     feature_types <- as.integer(feature_types)
@@ -549,11 +549,11 @@ bart <- function(X_train, y_train, leaf_basis_train = NULL, rfx_group_ids_train
     
     # Container of forest samples
     if (include_mean_forest) {
-        forest_samples_mean <- createForestContainer(num_trees_mean, output_dimension, is_leaf_constant, FALSE)
+        forest_samples_mean <- createForestSamples(num_trees_mean, output_dimension, is_leaf_constant, FALSE)
         active_forest_mean <- createForest(num_trees_mean, output_dimension, is_leaf_constant, FALSE)
     }
     if (include_variance_forest) {
-        forest_samples_variance <- createForestContainer(num_trees_variance, 1, TRUE, TRUE)
+        forest_samples_variance <- createForestSamples(num_trees_variance, 1, TRUE, TRUE)
         active_forest_variance <- createForest(num_trees_variance, 1, TRUE, TRUE)
     }
     
diff --git a/R/bcf.R b/R/bcf.R
@@ -760,7 +760,7 @@ bcf <- function(X_train, Z_train, y_train, propensity_train = NULL, rfx_group_id
     
     # Random number generator (std::mt19937)
     if (is.null(random_seed)) random_seed = sample(1:10000,1,F)
-    rng <- createRNG(random_seed)
+    rng <- createCppRNG(random_seed)
     
     # Sampling data structures
     forest_model_mu <- createForestModel(forest_dataset_train, feature_types, num_trees_mu, nrow(X_train), alpha_mu, beta_mu, min_samples_leaf_mu, max_depth_mu)
@@ -770,12 +770,12 @@ bcf <- function(X_train, Z_train, y_train, propensity_train = NULL, rfx_group_id
     }
     
     # Container of forest samples
-    forest_samples_mu <- createForestContainer(num_trees_mu, 1, T)
-    forest_samples_tau <- createForestContainer(num_trees_tau, 1, F)
+    forest_samples_mu <- createForestSamples(num_trees_mu, 1, T)
+    forest_samples_tau <- createForestSamples(num_trees_tau, 1, F)
     active_forest_mu <- createForest(num_trees_mu, 1, T)
     active_forest_tau <- createForest(num_trees_tau, 1, F)
     if (include_variance_forest) {
-        forest_samples_variance <- createForestContainer(num_trees_variance, 1, TRUE, TRUE)
+        forest_samples_variance <- createForestSamples(num_trees_variance, 1, TRUE, TRUE)
         active_forest_variance <- createForest(num_trees_variance, 1, TRUE, TRUE)
     }
     
diff --git a/R/calibration.R b/R/calibration.R
@@ -18,7 +18,7 @@
 #' X <- matrix(runif(n*p), ncol = p)
 #' y <- 10*X[,1] - 20*X[,2] + rnorm(n)
 #' nu <- 3
-#' lambda <- calibrate_inverse_gamma_error_variance(y, X, nu = nu)
+#' lambda <- calibrateInverseGammaErrorVariance(y, X, nu = nu)
 #' sigma2hat <- mean(resid(lm(y~X))^2)
 #' mean(var(y)/rgamma(100000, nu, rate = nu*lambda) < sigma2hat)
 calibrateInverseGammaErrorVariance <- function(y, X, W = NULL, nu = 3, quant = 0.9, standardize = TRUE) {
diff --git a/R/forest.R b/R/forest.R
@@ -758,7 +758,7 @@ Forest <- R6::R6Class(
 #'
 #' @return `ForestSamples` object
 #' @export
-createForestContainer <- function(num_trees, output_dimension=1, is_leaf_constant=F, is_exponentiated=F) {
+createForestSamples <- function(num_trees, output_dimension=1, is_leaf_constant=F, is_exponentiated=F) {
     return(invisible((
         ForestSamples$new(num_trees, output_dimension, is_leaf_constant, is_exponentiated)
     )))
diff --git a/R/model.R b/R/model.R
@@ -170,7 +170,7 @@ ForestModel <- R6::R6Class(
 #'
 #' @return `CppRng` object
 #' @export
-createRNG <- function(random_seed = -1){
+createCppRNG <- function(random_seed = -1){
     return(invisible((
         CppRNG$new(random_seed)
     )))
diff --git a/_pkgdown.yml b/_pkgdown.yml
@@ -88,9 +88,9 @@ reference:
   - ForestModel
   - createForestModel
   - ForestSamples
-  - createForestContainer
+  - createForestSamples
   - CppRNG
-  - createRNG
+  - createCppRNG
   - calibrateInverseGammaErrorVariance
   - preprocessParams
   - computeMaxLeafIndex
diff --git a/man/calibrateInverseGammaErrorVariance.Rd b/man/calibrateInverseGammaErrorVariance.Rd
diff --git a/man/createCppRNG.Rd b/man/createCppRNG.Rd
diff --git a/man/createForestSamples.Rd b/man/createForestSamples.Rd
diff --git a/test/R/testthat/test-predict.R b/test/R/testthat/test-predict.R
@@ -11,7 +11,7 @@ test_that("Prediction from trees with constant leaf", {
     n <- nrow(X)
     p <- ncol(X)
     forest_dataset = createForestDataset(X)
-    forest_samples <- createForestContainer(num_trees, 1, T)
+    forest_samples <- createForestSamples(num_trees, 1, T)
     
     # Initialize a forest with constant root predictions
     forest_samples$add_forest_with_constant_leaves(0.)
@@ -65,7 +65,7 @@ test_that("Prediction from trees with univariate leaf basis", {
     n <- nrow(X)
     p <- ncol(X)
     forest_dataset = createForestDataset(X,W)
-    forest_samples <- createForestContainer(num_trees, 1, F)
+    forest_samples <- createForestSamples(num_trees, 1, F)
     
     # Initialize a forest with constant root predictions
     forest_samples$add_forest_with_constant_leaves(0.)
@@ -123,7 +123,7 @@ test_that("Prediction from trees with multivariate leaf basis", {
     p <- ncol(X)
     W = matrix(c(1,1,1,1,1,1,-1,-1,-1,1,1,1), byrow=F, nrow=6)
     forest_dataset = createForestDataset(X,W)
-    forest_samples <- createForestContainer(num_trees, output_dim, F)
+    forest_samples <- createForestSamples(num_trees, output_dim, F)
     
     # Initialize a forest with constant root predictions
     forest_samples$add_forest_with_constant_leaves(c(1.,1.))
diff --git a/test/R/testthat/test-residual.R b/test/R/testthat/test-residual.R
@@ -33,11 +33,11 @@ test_that("Residual updates correctly propagated after forest sampling step", {
     b_forest = 0
     
     # RNG
-    cpp_rng = createRNG(-1)
+    cpp_rng = createCppRNG(-1)
     
     # Create forest sampler and forest container
     forest_model = createForestModel(forest_dataset, feature_types, num_trees, n, alpha, beta, min_samples_leaf, max_depth)
-    forest_samples = createForestContainer(num_trees, 1, F)
+    forest_samples = createForestSamples(num_trees, 1, F)
     active_forest = createForest(num_trees, 1, F)
     
     # Initialize the leaves of each tree in the prognostic forest
diff --git a/vignettes/CustomSamplingRoutine.Rmd b/vignettes/CustomSamplingRoutine.Rmd
@@ -128,7 +128,7 @@ if (leaf_regression) {
 outcome <- createOutcome(resid)
 
 # Random number generator (std::mt19937)
-rng <- createRNG()
+rng <- createCppRNG()
 
 # Sampling data structures
 forest_model <- createForestModel(forest_dataset, feature_types, 
@@ -139,10 +139,10 @@ forest_model <- createForestModel(forest_dataset, feature_types,
 # container of forest samples (which is written to when 
 # a sample is not discarded due to burn-in / thinning)
 if (leaf_regression) {
-    forest_samples <- createForestContainer(num_trees, 1, F)
+    forest_samples <- createForestSamples(num_trees, 1, F)
     active_forest <- createForest(num_trees, 1, F)
 } else {
-    forest_samples <- createForestContainer(num_trees, 1, T)
+    forest_samples <- createForestSamples(num_trees, 1, T)
     active_forest <- createForest(num_trees, 1, T)
 }
 ```
@@ -317,7 +317,7 @@ if (leaf_regression) {
 outcome <- createOutcome(resid)
 
 # Random number generator (std::mt19937)
-rng <- createRNG()
+rng <- createCppRNG()
 
 # Sampling data structures
 forest_model <- createForestModel(forest_dataset, feature_types, 
@@ -328,10 +328,10 @@ forest_model <- createForestModel(forest_dataset, feature_types,
 # container of forest samples (which is written to when 
 # a sample is not discarded due to burn-in / thinning)
 if (leaf_regression) {
-    forest_samples <- createForestContainer(num_trees, 1, F)
+    forest_samples <- createForestSamples(num_trees, 1, F)
     active_forest <- createForest(num_trees, 1, F)
 } else {
-    forest_samples <- createForestContainer(num_trees, 1, T)
+    forest_samples <- createForestSamples(num_trees, 1, T)
     active_forest <- createForest(num_trees, 1, T)
 }
 
@@ -552,7 +552,7 @@ if (leaf_regression) {
 outcome <- createOutcome(resid)
 
 # Random number generator (std::mt19937)
-rng <- createRNG()
+rng <- createCppRNG()
 
 # Sampling data structures
 forest_model <- createForestModel(forest_dataset, feature_types, 
@@ -563,10 +563,10 @@ forest_model <- createForestModel(forest_dataset, feature_types,
 # container of forest samples (which is written to when 
 # a sample is not discarded due to burn-in / thinning)
 if (leaf_regression) {
-    forest_samples <- createForestContainer(num_trees, 1, F)
+    forest_samples <- createForestSamples(num_trees, 1, F)
     active_forest <- createForest(num_trees, 1, F)
 } else {
-    forest_samples <- createForestContainer(num_trees, 1, T)
+    forest_samples <- createForestSamples(num_trees, 1, T)
     active_forest <- createForest(num_trees, 1, T)
 }
 
@@ -786,7 +786,7 @@ if (leaf_regression) {
 outcome <- createOutcome(resid)
 
 # Random number generator (std::mt19937)
-rng <- createRNG()
+rng <- createCppRNG()
 
 # Sampling data structures
 forest_model <- createForestModel(forest_dataset, feature_types, 
@@ -797,10 +797,10 @@ forest_model <- createForestModel(forest_dataset, feature_types,
 # container of forest samples (which is written to when 
 # a sample is not discarded due to burn-in / thinning)
 if (leaf_regression) {
-    forest_samples <- createForestContainer(num_trees, 1, F)
+    forest_samples <- createForestSamples(num_trees, 1, F)
     active_forest <- createForest(num_trees, 1, F)
 } else {
-    forest_samples <- createForestContainer(num_trees, 1, T)
+    forest_samples <- createForestSamples(num_trees, 1, T)
     active_forest <- createForest(num_trees, 1, T)
 }
 ```
@@ -1132,7 +1132,7 @@ forest_dataset_tau <- createForestDataset(X_tau, tau_basis)
 outcome <- createOutcome(resid)
 
 # Random number generator (std::mt19937)
-rng <- createRNG()
+rng <- createCppRNG()
 
 # Sampling data structures
 forest_model_mu <- createForestModel(
@@ -1145,9 +1145,9 @@ forest_model_tau <- createForestModel(
 )
 
 # Container of forest samples
-forest_samples_mu <- createForestContainer(num_trees_mu, 1, T)
+forest_samples_mu <- createForestSamples(num_trees_mu, 1, T)
 active_forest_mu <- createForest(num_trees_mu, 1, T)
-forest_samples_tau <- createForestContainer(num_trees_tau, 1, F)
+forest_samples_tau <- createForestSamples(num_trees_tau, 1, F)
 active_forest_tau <- createForest(num_trees_tau, 1, F)
 
 # Initialize the leaves of each tree in the prognostic forest
diff --git a/vignettes/PriorCalibration.Rmd b/vignettes/PriorCalibration.Rmd
@@ -49,7 +49,7 @@ In this case, $\nu$ is set by default to 3 and $\lambda$ is calibrated as follow
 1. An "overestimate," $\hat{\sigma}^2$, of $\sigma^2$ is obtained via simple linear regression of $y$ on $X$
 2. $\lambda$ is chosen to ensure that $p(\sigma^2 < \hat{\sigma}^2) = q$ for some value $q$, typically set to a default value of 0.9.
 
-This is done in `stochtree` via the `calibrate_inverse_gamma_error_variance` function.
+This is done in `stochtree` via the `calibrateInverseGammaErrorVariance` function.
 
 ```{r}
 # Load library
@@ -81,7 +81,7 @@ y_train <- y[train_inds]
 
 # Calibrate the scale parameter for the variance term as in Chipman et al (2010)
 nu <- 3
-lambda <- calibrate_inverse_gamma_error_variance(y_train, X_train, nu = nu)
+lambda <- calibrateInverseGammaErrorVariance(y_train, X_train, nu = nu)
 ```
 
 Now we run a BART model with this variance parameterization
