Move arrow to Suggests to appease CRAN

.gitignore

@@ -4,4 +4,6 @@
 *.png
 *.rec
 *.csv
+*.Rcheck/*
 inst/tools/starvz_fast_csv_split
+starvz_*.tar.gz

DESCRIPTION

@@ -44,14 +44,13 @@ Imports: methods,
     data.tree,
     RColorBrewer,
     zoo,
-    Rcpp,
-    arrow (>= 3.0.0)
+    Rcpp
 License: GPL-3
 Encoding: UTF-8
 SystemRequirements: C++, arrow package with gzip codec, StarPU
 LazyData: true
 LinkingTo: Rcpp (>= 1.0.6), BH
-RoxygenNote: 7.1.1
+RoxygenNote: 7.2.3
 Collate:
     'RcppExports.R'
     'starvz_data.R'
@@ -83,9 +82,11 @@ Collate:
     'read_functions.R'
     'write_functions.R'
 Suggests:
+    arrow (>= 3.0.0),
     testthat,
     flexmix,
     car,
     viridis,
     ggrepel,
     mclust
+Additional_repositories: https://p3m.dev/cran/2024-02-02

INSTALL.org

@@ -246,7 +246,7 @@ While the ~arrow~ installation is straightforward in many systems
 (pre-compiled binaries), it may be hard to obtain an installation with
 compression support in Debian systems.  See [[https://issues.apache.org/jira/browse/ARROW-8556][this link for more
 information]]. A possible fix is do the following (install everything by
-hand [[https://cran.r-project.org/web/packages/arrow/vignettes/install.html][following a combination of these instructions]]). See below for
+hand [[https://arrow.apache.org/docs/r/articles/install.html][following a combination of these instructions]]). See below for
 details.
 
 *** Install the Apache Arrow C++
@@ -291,16 +291,12 @@ your R session:
 
 #+begin_src R :results output :session :exports both
 library(arrow)
-arrow_available()
 arrow::CompressionType
 #+end_src
 
 Confirm the output is something like this:
 
 #+BEGIN_EXAMPLE
-> arrow_available()
-[1] TRUE
-
 > arrow::CompressionType
 $UNCOMPRESSED
 [1] 0

NAMESPACE

@@ -68,13 +68,6 @@ import(Rcpp)
 import(dplyr)
 import(ggplot2)
 importFrom(RColorBrewer,brewer.pal)
-importFrom(arrow,ParquetWriterProperties)
-importFrom(arrow,arrow_available)
-importFrom(arrow,codec_is_available)
-importFrom(arrow,read_feather)
-importFrom(arrow,read_parquet)
-importFrom(arrow,write_feather)
-importFrom(arrow,write_parquet)
 importFrom(data.tree,Prune)
 importFrom(data.tree,Set)
 importFrom(data.tree,as.Node)

R/phase1.R

@@ -18,7 +18,7 @@ NULL
 #' @return ggplot object with all starvz plots
 #' @family phase1 functions
 #'
-#' @examples
+#' @examplesIf requireNamespace("arrow", quietly = TRUE)
 #' \donttest{
 #' example_folder <- system.file("extdata", "lu_trace", package = "starvz")
 #' starvz_phase1(directory = example_folder)
@@ -262,8 +262,9 @@ hl_y_paje_tree <- function(where = ".") {
   entities.csv <- paste0(where, "/entities.csv")
 
   if (file.exists(entities.feather)) {
+    check_arrow()
     starvz_log(paste("Reading ", entities.feather))
-    dfe <- read_feather(entities.feather)
+    dfe <- arrow::read_feather(entities.feather)
   } else if (file.exists(entities.csv)) {
     starvz_log(paste("Reading ", entities.csv))
     dfe <- starvz_suppressWarnings(read_csv(entities.csv,

R/phase1_parse_csv.R

@@ -626,8 +626,9 @@ pmtool_bounds_csv_parser <- function(where = ".", ZERO = 0) {
   entities.csv <- paste0(where, "/pmtool.csv")
 
   if (file.exists(entities.feather)) {
+    check_arrow()
     starvz_log(paste("Reading ", entities.feather))
-    pm <- read_feather(entities.feather)
+    pm <- arrow::read_feather(entities.feather)
   } else if (file.exists(entities.csv)) {
     starvz_log(paste("Reading ", entities.csv))
     pm <- starvz_suppressWarnings(read_csv(entities.csv,
@@ -795,8 +796,9 @@ task_handles_parser <- function(where = ".") {
   entities.feather <- paste0(where, "/task_handles.feather")
 
   if (file.exists(entities.feather)) {
+    check_arrow()
     starvz_log(paste("Reading ", entities.feather))
-    ret <- read_feather(entities.feather)
+    ret <- arrow::read_feather(entities.feather)
     return(ret)
   }
 

R/phase2_util.R

@@ -1,8 +1,13 @@
 #' @include starvz_data.R
 
 check_arrow <- function() {
-  if (!arrow_available()) {
-    starvz_warn("R package arrow was not property installed, use: install_arrow()")
+  if (!requireNamespace("arrow", quietly = TRUE)) {
+    msg <- paste(
+      "The 'arrow' package is required but is not available. Install it with:",
+      'install.packages("arrow", repos = c("https://p3m.dev/cran/2024-02-02", getOption("repos")))',
+      sep = "\n"
+    )
+    stop(msg, call. = FALSE)
   }
 }
 

R/read_functions.R

@@ -5,7 +5,7 @@ starvz_read_some_feather <- function(directory = ".", tables = c("application"))
   l2 <- lapply(tables, function(table) {
     table_file <- file.path(directory, paste0(table, ".feather"))
     if (file.exists(table_file)) {
-      read_parquet(table_file)
+      arrow::read_parquet(table_file)
     } else {
       starvz_log(paste("The file:", table_file, " does not exist on that directory. Ignore."))
       NULL
@@ -16,6 +16,7 @@ starvz_read_some_feather <- function(directory = ".", tables = c("application"))
 }
 
 starvz_read_feather <- function(directory = ".") {
+  check_arrow()
   filenames <- list.files(
     path = directory,
     pattern = "*.feather",
@@ -31,15 +32,16 @@ starvz_read_feather <- function(directory = ".") {
 the_fast_reader_function <- starvz_read_feather
 
 starvz_read_some_parquet <- function(directory = ".", tables = c("application")) {
-  if (!codec_is_available("gzip")) {
+  check_arrow()
+  if (!arrow::codec_is_available("gzip")) {
     starvz_warn("R package arrow does not have 'gzip' codec, try using arrow::install_arrow()")
     return(list())
   } else {
     l1 <- list(Origin = directory)
     l2 <- lapply(tables, function(table) {
       table_file <- file.path(directory, paste0(table, ".parquet"))
       if (file.exists(table_file)) {
-        read_parquet(table_file)
+        arrow::read_parquet(table_file)
       } else {
         starvz_log(paste("The file:", table_file, " does not exist on that directory. Ignore."))
         NULL
@@ -51,6 +53,7 @@ starvz_read_some_parquet <- function(directory = ".", tables = c("application"))
 }
 
 starvz_read_parquet <- function(directory = ".") {
+  check_arrow()
   filenames <- list.files(
     path = directory,
     pattern = "*.parquet",
@@ -99,7 +102,7 @@ starvz_read_some <- function(directory = ".", tables = c("application"), config_
 #' @return The starvz_data with all tables
 #' @usage starvz_read(directory = ".",
 #'                config_file = NULL, selective = TRUE)
-#' @examples
+#' @examplesIf requireNamespace("arrow", quietly = TRUE)
 #' starvz_read("folder_to_parquet_files/")
 #' starvz_read(
 #'   directory = "folder_to_parquet_files/",

R/starvz_data.R

@@ -16,7 +16,6 @@
 #' @importFrom stringr str_replace str_replace_all str_to_title
 #' @importFrom purrr list_modify map map2
 #' @importFrom patchwork wrap_plots
-#' @importFrom arrow arrow_available read_feather read_parquet write_feather write_parquet codec_is_available ParquetWriterProperties
 #' @importFrom lpSolve lp
 #' @importFrom data.tree as.Node Prune Set
 #' @importFrom gtools mixedorder mixedsort

R/write_functions.R

@@ -8,7 +8,7 @@ starvz_write_feather <- function(data, directory = ".") {
       starvz_log(filename)
       if (!is.null(data[[x]])) {
         if (is.data.frame(data[[x]])) {
-          write_feather(data[[x]], filename)
+          arrow::write_feather(data[[x]], filename)
         } else {
           starvz_log(paste(filename, "must be a data frame."))
         }
@@ -23,7 +23,7 @@ starvz_write_parquet <- function(data, directory = ".") {
   if (is.null(data)) {
     return(NULL)
   }
-  if (!codec_is_available("gzip")) {
+  if (!arrow::codec_is_available("gzip")) {
     starvz_warn("R package arrow does not have 'gzip' codec, try using arrow::install_arrow()")
   } else {
     invisible(data %>% list_modify("Origin" = NULL) %>% names() %>%
@@ -32,7 +32,7 @@ starvz_write_parquet <- function(data, directory = ".") {
         starvz_log(filename)
         if (!is.null(data[[x]])) {
           if (is.data.frame(data[[x]])) {
-            write_parquet(data[[x]], filename, compression = "gzip")
+            arrow::write_parquet(data[[x]], filename, compression = "gzip")
           } else {
             starvz_log(paste(filename, "must be a data frame."))
           }

tests/testthat/test_lu.R

@@ -8,6 +8,7 @@ test_that("starvz_plot works", {
   expect_equal(class(pl), c("patchwork", "gg", "ggplot"))
 
   skip_on_cran()
+  skip_if_not_installed("arrow")
 
   other_data <- starvz_sample_lu
 
@@ -49,7 +50,7 @@ test_that("starvz_plot works", {
   other_data <- starvz_phase1(system.file("extdata", "lu_trace", package = "starvz"), lu_colors, state_filter = 2, whichApplication = "lu")
   other_data <- starvz_read(system.file("extdata", "lu_trace", package = "starvz"), system.file("extdata", "config.yaml", package = "starvz"))
   result <- all.equal(nrow(other_data$Application), nrow(starvz_sample_lu$Application))
-  if (codec_is_available("gzip")) {
+  if (arrow::codec_is_available("gzip")) {
     expect_equal(result, TRUE)
   }