Add tar_format_nanoparquet()

DESCRIPTION

@@ -54,6 +54,7 @@ Imports:
 Suggests:
   curl (>= 4.3),
   knitr (>= 1.28),
+  nanoparquet,
   quarto (>= 1.4),
   rmarkdown (>= 2.1),
   testthat (>= 3.0.0),

NAMESPACE

@@ -54,6 +54,7 @@ export(tar_force)
 export(tar_force_change)
 export(tar_format_aws_feather)
 export(tar_format_feather)
+export(tar_format_nanoparquet)
 export(tar_fst)
 export(tar_fst_dt)
 export(tar_fst_tbl)
@@ -91,6 +92,9 @@ export(tar_map2_size)
 export(tar_map2_size_raw)
 export(tar_map_rep)
 export(tar_map_rep_raw)
+export(tar_nanoparquet_convert)
+export(tar_nanoparquet_read)
+export(tar_nanoparquet_write)
 export(tar_parquet)
 export(tar_plan)
 export(tar_qs)
@@ -201,6 +205,7 @@ importFrom(targets,tar_deparse_safe)
 importFrom(targets,tar_dir)
 importFrom(targets,tar_envir)
 importFrom(targets,tar_exist_meta)
+importFrom(targets,tar_format)
 importFrom(targets,tar_group)
 importFrom(targets,tar_load)
 importFrom(targets,tar_meta)

R/tar_format_nanoparquet.R

@@ -0,0 +1,96 @@
+#' @title Nanoparquet format
+#' @export
+#' @keywords storage formats
+#' @description Nanoparquet storage format for data frames.
+#'   Uses [nanoparquet::read_parquet()] and [nanoparquet::write_parquet()]
+#'   to read and write data frames returned by targets in a pipeline.
+#'   Note: attributes such as `dplyr` row groupings and `posterior`
+#'   draws info are dropped during the writing process.
+#' @return A [targets::tar_format()] storage format specification string
+#'   that can be directly supplied to the `format` argument of
+#'   [targets::tar_target()] or [targets::tar_option_set()].
+#' @param compression Character string, compression type for saving the
+#'   data. See the `compression` argument of [nanoparquet::write_parquet()]
+#'   for details.
+#' @param class Character vector with the data frame subclasses to assign.
+#'   See the `class` argument of [nanoparquet::parquet_options()] for details.
+#' @examples
+#' if (identical(Sys.getenv("TAR_LONG_EXAMPLES"), "true")) {
+#' targets::tar_dir({ # tar_dir() runs code from a temporary directory.
+#' targets::tar_script({
+#'   library(targets)
+#'   libary(tarchetypes)
+#'   list(
+#'     tar_target(
+#'       name = data,
+#'       command = airquality,
+#'       format = tar_format_nanoparquet()
+#'     )
+#'   )
+#' })
+#' tar_make()
+#' tar_read(data)
+#' })
+#' }
+tar_format_nanoparquet <- function(compression = "snappy", class = "tbl") {
+  rlang::check_installed("nanoparquet")
+  read <- function(path) {}
+  body(read) <- substitute(
+    tarchetypes::tar_nanoparquet_read(path, class),
+    env = list(class = class)
+  )
+  write <- function(object, path) {}
+  body(write) <- substitute(
+    tarchetypes::tar_nanoparquet_write(object, path, compression),
+    env = list(compression = compression)
+  )
+  convert <- function(object) {}
+  body(convert) <- substitute(
+    tarchetypes::tar_nanoparquet_convert(object, class),
+    env = list(class = class)
+  )
+  targets::tar_format(read = read, write = write, convert = convert)
+}
+
+#' @title Nanoparquet read method
+#' @export
+#' @keywords internal
+#' @description Internal function.
+#' @param path Path to the data.
+#' @param class S3 classes to assign to the returned object.
+tar_nanoparquet_read <- function(path, class) {
+  nanoparquet::read_parquet(
+    file = path,
+    options = nanoparquet::parquet_options(
+      class = class,
+      use_arrow_metadata = TRUE
+    )
+  )
+}
+
+#' @title Nanoparquet write method
+#' @export
+#' @keywords internal
+#' @description Internal function.
+#' @param object R object to save.
+#' @param path Path to the data.
+#' @param compression Compression type.
+tar_nanoparquet_write <- function(object, path, compression) {
+  nanoparquet::write_parquet(
+    x = object,
+    file = path,
+    compression = compression,
+    options = nanoparquet::parquet_options(write_arrow_metadata = TRUE)
+  )
+}
+
+#' @title Nanoparquet convert method
+#' @export
+#' @keywords internal
+#' @description Internal function.
+#' @param object R object to convert.
+#' @param class S3 classes to assign to the returned object.
+tar_nanoparquet_convert <- function(object, class) {
+  class(object) <- c(class, "data.frame")
+  object
+}

R/tar_package.R

@@ -27,7 +27,7 @@
 #'   tar_assert_scalar tar_assert_target tar_assert_target_list
 #'   tar_assert_true tar_assert_unique tar_assert_unique_targets
 #'   tar_cue tar_definition tar_deparse_language tar_deparse_safe
-#'   tar_dir tar_envir tar_exist_meta tar_group tar_load tar_meta
+#'   tar_dir tar_envir tar_exist_meta tar_format tar_group tar_load tar_meta
 #'   tar_option_get tar_option_set tar_path_target tar_read
 #'   tar_runtime_object tar_script tar_seed_create tar_seed_set
 #'   tar_target tar_target_raw tar_test tar_tidy_eval

_pkgdown.yml

@@ -28,6 +28,9 @@ reference:
 - title: Target factories for storage formats
   contents:
   - 'tar_formats'
+- title: Storage formats
+  contents:
+  - 'tar_format_nanoparquet'
 - title: Simple files
   contents:
   - 'tar_file_read'