Merge pull request #118 from tidy-finance/117-ensure-that-download_da…

…ta-fails-gracefully-if-the-resource-is-not-available-or-has-changed Add handle_download_error() function
tidy-finance · Nov 29, 2024 · b5c0988 · b5c0988
2 parents 495d542 + bfdcb09
commit b5c0988
Show file tree

Hide file tree

Showing 7 changed files with 85 additions and 8 deletions.
diff --git a/DESCRIPTION b/DESCRIPTION
@@ -1,7 +1,7 @@
 Type: Package
 Package: tidyfinance
 Title: Tidy Finance Helper Functions
-Version: 0.4.1.9003
+Version: 0.4.1.9004
 Authors@R: c(
     person("Christoph", "Scheuch", , "christoph.scheuch@gmail.com", role = c("aut", "cre", "cph"),
            comment = c(ORCID = "0009-0004-0423-6819")),

diff --git a/NEWS.md b/NEWS.md
@@ -1,5 +1,9 @@
 # tidyfinance (development version)
 
+## Bug fixes
+
+* `download_macro_predictors()`, `download_factors()`, and `download_osap()` now fail gracefully with informative messages instead of errors or warnings.
+
 ## Improvements
 
 * Updated `ccmxpf_linktable` to the new WRDS default `ccmxpf_lnkhist`.

diff --git a/R/download_data_factors.R b/R/download_data_factors.R
@@ -91,7 +91,16 @@ download_data_factors_ff <- function(
   factors_ff_types <- list_supported_types_ff()
   dataset <- factors_ff_types$dataset_name[factors_ff_types$type == type]
 
-  raw_data <- suppressMessages(frenchdata::download_french_data(dataset))
+  raw_data <- handle_download_error(
+    function() suppressMessages(frenchdata::download_french_data(dataset)),
+    fallback = tibble()
+  )
+
+  if (class(raw_data) != "french_dataset") {
+    cli::cli_inform("Returning an empty data set due to download failure.")
+    return(raw_data)
+  }
+
   raw_data <- raw_data$subsets$data[[1]]
 
   if (grepl("monthly", type, fixed = TRUE)) {
@@ -173,7 +182,19 @@ download_data_factors_q <- function(
 
   factors_q_types <- list_supported_types_q()
   dataset <- factors_q_types$dataset_name[factors_q_types$type == type]
-  raw_data <- suppressMessages(utils::read.csv(paste0(url, dataset)) |> as_tibble())
+
+  raw_data <- handle_download_error(
+    function(url) suppressWarnings(
+      suppressMessages(utils::read.csv(url)) |> as_tibble()
+    ),
+    paste0(url, dataset),
+    fallback = tibble()
+  )
+
+  if (nrow(raw_data) == 0) {
+    cli::cli_inform("Returning an empty data set due to download failure.")
+    return(raw_data)
+  }
 
   if (grepl("monthly", type, fixed = TRUE)) {
     processed_data <- raw_data |>

diff --git a/R/download_data_macro_predictors.R b/R/download_data_macro_predictors.R
@@ -23,7 +23,7 @@
 #'
 #' @examples
 #' \donttest{
-#'   macro_predictors_monthly <- download_data_macro_predictors("macro_predictors_monthly")
+#'  download_data_macro_predictors("macro_predictors_monthly")
 #' }
 #'
 download_data_macro_predictors <- function(
@@ -49,12 +49,36 @@ download_data_macro_predictors <- function(
   }
 
   if (grepl("monthly", type, fixed = TRUE)) {
-    raw_data <- as_tibble(read.csv(build_macro_predictors_url("Monthly")))
+
+    raw_data <- handle_download_error(
+      function() suppressWarnings(
+        as_tibble(read.csv(build_macro_predictors_url("Monthly")))
+      ),
+      fallback = tibble()
+    )
+
+    if (nrow(raw_data) == 0) {
+      cli::cli_inform("Returning an empty data set due to download failure.")
+      return(raw_data)
+    }
+
     processed_data <- raw_data |>
       mutate(date = ym(yyyymm))
   }
   if (grepl("quarterly", type, fixed = TRUE)) {
-    raw_data <- as_tibble(read.csv(build_macro_predictors_url("Quarterly")))
+
+    raw_data <- handle_download_error(
+      function() suppressWarnings(
+        as_tibble(read.csv(build_macro_predictors_url("Quarterly")))
+      ),
+      fallback = tibble()
+    )
+
+    if (nrow(raw_data) == 0) {
+      cli::cli_inform("Returning an empty data set due to download failure.")
+      return(raw_data)
+    }
+
     processed_data <- raw_data |>
       mutate(
         year = substr(yyyyq, 1, 4),

diff --git a/R/download_data_osap.R b/R/download_data_osap.R
@@ -34,7 +34,18 @@ download_data_osap <- function(
   }
 
   url <- paste0("https://drive.google.com/uc?export=download&id=", sheet_id)
-  raw_data <- as_tibble(read.csv(url))
+
+  raw_data <- handle_download_error(
+    function() suppressWarnings(
+      as_tibble(read.csv(url))
+    ),
+    fallback = tibble()
+  )
+
+  if (nrow(raw_data) == 0) {
+    cli::cli_inform("Returning an empty data set due to download failure.")
+    return(raw_data)
+  }
 
   processed_data <- raw_data |>
     mutate(date = ymd(date))

diff --git a/R/handle_download_error.R b/R/handle_download_error.R
@@ -0,0 +1,17 @@
+#' @keywords internal
+#'
+handle_download_error <- function(download_function, ..., fallback = NULL) {
+  tryCatch(
+    {
+      download_function(...)
+    },
+    error = function(e) {
+      cli::cli_inform(paste0(
+        "Failed to download or process the dataset. ",
+        "The resource may not be available, or the URL may have changed. ",
+        "Error message: {.message {e$message}}"
+      ))
+      fallback
+    }
+  )
+}
diff --git a/man/download_data_macro_predictors.Rd b/man/download_data_macro_predictors.Rd