Add more checking functions:

* check_for_free_state_error()
* check_initial_values_correct_class()
* check_missing_infinite_values()
* check_timeout()
* check_sampling_implemented()
* check_truncation_implemented()
* add internal `are_intials()` function

Author: njtierney

R/checkers.R

@@ -1399,14 +1399,16 @@ check_initial_values_match_chains <- function(initial_values,
                                               n_chains,
                                               call = rlang::caller_env()){
 
-  if (!is.initials(initial_values) && is.list(initial_values)) {
+  initials <- initial_values
+  not_initials_but_list <- !is.initials(initials) && is.list(initials)
+  if (not_initials_but_list) {
     # if the user provided a list of initial values, check elements and length
-    are_initials <- vapply(initial_values, is.initials, FUN.VALUE = FALSE)
+    all_initials <- all(are_initials(initials))
 
-    n_sets <- length(initial_values)
+    n_sets <- length(initials)
 
     initial_values_do_not_match_chains <- n_sets != n_chains
-    if (initial_values_do_not_match_chains && all(are_initials)) {
+    if (initial_values_do_not_match_chains && all_initials) {
       cli::cli_abort(
         message = c(
           "The number of provided initial values does not match chains",
@@ -1437,6 +1439,29 @@ check_initial_values_correct_dim <- function(target_dims,
 
 }
 
+check_initial_values_correct_class <- function(initial_values,
+                                               call = rlang::caller_env()){
+
+  initials <- initial_values
+  not_initials_but_list <- !is.initials(initials) && is.list(initials)
+  not_initials_not_list <- !is.initials(initials) && !is.list(initials)
+  # if the user provided a list of initial values, check elements and the
+  # length
+  all_initials <- all(are_initials(initials))
+  not_all_initials <- !all_initials
+
+  if (not_initials_but_list && not_all_initials || not_initials_not_list) {
+    cli::cli_abort(
+      message = c(
+        "{.arg initial_values} must be an initials object created with \\
+        {.fun initials}, or a simple list of initials objects"
+      ),
+      call = call
+    )
+  }
+
+}
+
 check_nodes_all_variable <- function(nodes,
                                      call = rlang::caller_env()){
   types <- lapply(nodes, node_type)
@@ -1921,16 +1946,78 @@ check_has_representation <- function(repr,
 check_is_greta_array <- function(x,
                                  arg = rlang::caller_arg(x),
                                  call = rlang::caller_env()){
-  # only for greta arrays
   if (!is.greta_array(x)) {
     cli::cli_abort(
       message = c(
         "{.arg {arg}} must be {.cls greta_array}",
-        "Object was is {.cls {class(x)}}"
+        "{.arg {arg}} is: {.cls {class(x)}}"
+      ),
+      call = call
+    )
+  }
+}
+
+check_missing_infinite_values <- function(x,
+                                          optional,
+                                          call = rlang::caller_env()){
+  contains_missing_or_inf <- !optional & any(!is.finite(x))
+  if (contains_missing_or_inf) {
+    cli::cli_abort(
+      message = c(
+        "{.cls greta_array} must not contain missing or infinite values"
+      ),
+      call = call
+    )
+  }
+}
+
+check_truncation_implemented <- function(tfp_distribution,
+                                         distribution_node,
+                                         call = rlang::caller_env()){
+
+  cdf <- tfp_distribution$cdf
+  quantile <- tfp_distribution$quantile
+
+  is_truncated <- is.null(cdf) | is.null(quantile)
+  if (is_truncated) {
+    cli::cli_abort(
+      message = c(
+        "Sampling is not yet implemented for truncated \\
+            {.val {distribution_node$distribution_name}} distributions"
+      ),
+      call = call
+    )
+  }
+
+}
+
+check_sampling_implemented <- function(sample,
+                                       distribution_node,
+                                       call = rlang::caller_env()){
+  if (is.null(sample)) {
+    cli::cli_abort(
+      "Sampling is not yet implemented for \\
+            {.val {distribution_node$distribution_name}} distributions"
+    )
+  }
+}
+
+check_timeout <- function(it,
+                          maxit,
+                          call = rlang::caller_env()){
+  # check we didn't time out
+  if (it == maxit) {
+    cli::cli_abort(
+      message = c(
+      "Could not determine the number of independent models in a reasonable \\
+      amount of time",
+      "Iterations = {.val {it}}",
+      "Maximum iterations = {.cal {maxit}}"
       ),
       call = call
     )
   }
+
 }
 
 

R/dag_class.R

@@ -769,13 +769,7 @@ dag_class <- R6Class(
         }
       }
 
-      # check we didn't time out
-      if (it == maxit) {
-        cli::cli_abort(
-          "could not determine the number of independent models in a \\
-          reasonable amount of time"
-        )
-      }
+      check_timeout(it, maxit)
 
       # find the cluster IDs
       n <- nrow(r)
@@ -812,12 +806,7 @@ dag_class <- R6Class(
 
       sample <- tfp_distribution$sample
 
-      if (is.null(sample)) {
-        cli::cli_abort(
-          "sampling is not yet implemented for \\
-            {.val {distribution_node$distribution_name}} distributions"
-        )
-      }
+      check_sampling_implemented(sample, distribution_node)
 
       truncation <- distribution_node$truncation
 
@@ -833,14 +822,7 @@ dag_class <- R6Class(
 
         cdf <- tfp_distribution$cdf
         quantile <- tfp_distribution$quantile
-
-        is_truncated <- is.null(cdf) | is.null(quantile)
-        if (is_truncated) {
-          cli::cli_abort(
-            "sampling is not yet implemented for truncated \\
-            {.val {distribution_node$distribution_name}} distributions"
-          )
-        }
+        check_truncation_implemented(tfp_distribution, distribution_node)
 
         # generate a random uniform sample of the correct shape and transform
         # through truncated inverse CDF to get draws on truncated scale

R/distribution.R

@@ -43,6 +43,7 @@
 
   node <- get_node(greta_array)
 
+  # TODO revisit checking functions here
   # only for greta arrays without distributions
   ## TODO provide more detail on the distribution already assigned
   ## This might come up when the user accidentally runs assignment

R/extract_replace_combine.R

@@ -405,6 +405,7 @@ abind.greta_array <- function(...,
     along <- max(1, min(n + 1, ceiling(along)))
   }
 
+  # TODO revisit checking functions here
   along_outside_0_n <- !(along %in% 0:n)
   if (along_outside_0_n) {
     cli::cli_abort(
@@ -530,6 +531,7 @@ length.greta_array <- function(x) {
 
   dims <- dims %||% length(x)
 
+  # TODO revisit logic / checking functions here
   if (length(dims) == 0L) {
     cli::cli_abort(
       "length-0 dimension vector is invalid"

R/greta_array_class.R

@@ -78,12 +78,8 @@ as.greta_array.array <- function(x, optional = FALSE, original_x = x, ...) {
 # finally, reject if there are any missing values, or set up the greta_array
 #' @export
 as.greta_array.numeric <- function(x, optional = FALSE, original_x = x, ...) {
-  contains_missing_or_inf <- !optional & any(!is.finite(x))
-  if (contains_missing_or_inf) {
-    cli::cli_abort(
-        "{.cls greta_array} must not contain missing or infinite values"
-    )
-  }
+  check_missing_infinite_values(x, optional)
+
   as.greta_array.node(data_node$new(x),
     optional = optional,
     original_x = original_x,

R/inference.R

@@ -561,6 +561,8 @@ to_free <- function(node, data) {
   lower <- node$lower
   upper <- node$upper
 
+  # TODO
+  # replace these with more informative errors related to the range of values
   unsupported_error <- function() {
     cli::cli_abort(
       "Some provided initial values are outside the range of values their \\
@@ -661,7 +663,6 @@ parse_initial_values <- function(initials, dag) {
 # correct length, with nice error messages
 prep_initials <- function(initial_values, n_chains, dag) {
 
-  # TODO: Tidy up the logic here for errors and messages
   # if the user passed a single set of initial values, repeat them for all
   # chains
   if (is.initials(initial_values)) {
@@ -673,33 +674,9 @@ prep_initials <- function(initial_values, n_chains, dag) {
     )
   }
 
-  not_initials_but_list <- !is.initials(initial_values) && is.list(initial_values)
-  if (not_initials_but_list) {
-
-    # if the user provided a list of initial values, check elements and the
-    # length
-    are_initials <- vapply(initial_values, is.initials, FUN.VALUE = FALSE)
-
-    if (all(are_initials)) {
-      check_initial_values_match_chains(initial_values, n_chains)
-    }
-    if (!all(are_initials)) {
-      initial_values <- NULL
-    }
-  }
-  if (!not_initials_but_list) {
-    initial_values <- NULL
-  }
-
-  # error on a bad object
-  if (is.null(initial_values)) {
-    cli::cli_abort(
-      c(
-        "{.arg initial_values} must be an initials object created with \\
-        {.fun initials}, or a simple list of initials objects"
-      )
-    )
-  }
+  # TODO: revisit logic here for errors and messages
+  check_initial_values_match_chains(initial_values, n_chains)
+  check_initial_values_correct_class(initial_values)
 
   # convert them to free state vectors
   initial_values <- lapply(

R/inference_class.R

@@ -804,6 +804,15 @@ sampler <- R6Class(
         )
       ) # closing cleanly
 
+      # if it's fine, batch_results is the output
+      # if it's a non-numerical error, it will error
+      # if it's a numerical error, batch_results will be an error object
+      self$check_for_free_state_error(result, n_samples)
+
+      result
+    },
+
+    check_for_free_state_error = function(result, n_samples){
       # if it's fine, batch_results is the output
       # if it's a non-numerical error, it will error
       # if it's a numerical error, batch_results will be an error object
@@ -827,7 +836,7 @@ sampler <- R6Class(
           # won't be valid if we just restart, so we need to error here,
           # informing the user how to run one sample at a time
           cli::cli_abort(
-            c(
+            message = c(
               "TensorFlow hit a numerical problem that caused it to error",
               "{.pkg greta} can handle these as bad proposals if you rerun \\
               {.fun mcmc} with the argument {.code one_by_one = TRUE}.",
@@ -839,9 +848,8 @@ sampler <- R6Class(
 
         }
       }
-
-      result
     },
+
     sampler_parameter_values = function() {
 
       # random number of integration steps

R/utils.R

@@ -1166,3 +1166,11 @@ outside_version_range <- function(provided, range) {
 }
 
 pretty_dim <- function(x) paste0(dim(x), collapse = "x")
+
+are_initials <- function(x){
+  vapply(
+    X = x,
+    FUN = is.initials,
+    FUN.VALUE = logical(1)
+    )
+}

tests/testthat/_snaps/calculate.md

@@ -59,7 +59,7 @@
 
 # calculate errors if a distribution cannot be sampled from
 
-    sampling is not yet implemented for "hypergeometric" distributions
+    Sampling is not yet implemented for "hypergeometric" distributions
 
 # calculate errors nicely if nsim is invalid
 

tests/testthat/_snaps/distributions.md

@@ -81,13 +81,21 @@
 
 # wishart distribution errors informatively
 
-    `Sigma` must be a square 2D greta array
-    However, `Sigma` has dimensions "3x3x3"
+    Code
+      wishart(3, b)
+    Condition
+      Error in `initialize()`:
+      ! `Sigma` must be a square 2D greta array
+      However, `Sigma` has dimensions "3x3x3"
 
 ---
 
-    `Sigma` must be a square 2D greta array
-    However, `Sigma` has dimensions "3x2"
+    Code
+      wishart(3, c)
+    Condition
+      Error in `initialize()`:
+      ! `Sigma` must be a square 2D greta array
+      However, `Sigma` has dimensions "3x2"
 
 # lkj_correlation distribution errors informatively
 

tests/testthat/_snaps/iid_samples.md

@@ -1,8 +1,8 @@
 # distributions without RNG error nicely
 
-    sampling is not yet implemented for "hypergeometric" distributions
+    Sampling is not yet implemented for "hypergeometric" distributions
 
 ---
 
-    sampling is not yet implemented for truncated "f" distributions
+    Sampling is not yet implemented for truncated "f" distributions
 

tests/testthat/_snaps/simulate.md

@@ -4,7 +4,7 @@
 
 # simulate errors if a distribution cannot be sampled from
 
-    sampling is not yet implemented for "hypergeometric" distributions
+    Sampling is not yet implemented for "hypergeometric" distributions
 
 # simulate errors nicely if nsim is invalid
 

tests/testthat/_snaps/syntax.md

@@ -37,6 +37,10 @@
 
 # distribution() errors informatively
 
-    `distribution()` expects object of type <greta_array>
-    object was not a <greta_array>, but <array>
+    Code
+      distribution(y)
+    Condition
+      Error in `distribution()`:
+      ! `greta_array` must be <greta_array>
+      `greta_array` is: <array>
 

tests/testthat/test_syntax.R

@@ -95,7 +95,8 @@ test_that("distribution() errors informatively", {
 
   y <- randn(3)
 
-  expect_snapshot_error(
+  expect_snapshot(
+    error = TRUE,
     distribution(y)
   )
 })