Add Dataset small_coralnet (#240)

* Improve messages and translations with 'split' infromation
Initial addition of whoi_small_coralnet

* fix md5 values and add tests

* update translations

Author: cregouby

NAMESPACE

@@ -189,6 +189,7 @@ export(transform_to_tensor)
 export(transform_vflip)
 export(vision_make_grid)
 export(whoi_plankton_dataset)
+export(whoi_small_coralnet_dataset)
 export(whoi_small_plankton_dataset)
 importFrom(grDevices,dev.off)
 importFrom(graphics,polygon)

NEWS.md

@@ -5,7 +5,7 @@
 * Added `lfw_people_dataset()` and `lfw_pairs_dataset()` for loading Labelled Faces in the Wild (LFW) datasets (@DerrickUnleashed, #203).
 * Added `places365_dataset()`for loading the Places365 dataset (@koshtiakanksha, #196).
 * Added `pascal_segmentation_dataset()`, and `pascal_detection_dataset()` for loading the Pascal Visual Object Classes datasets (@DerrickUnleashed, #209).
-* Added `whoi_plankton_dataset()`, and `whoi_small_plankton_dataset()` (@cregouby, #236).
+* Added `whoi_plankton_dataset()`, `whoi_small_plankton_dataset()`, and  `whoi_small_coral_dataset()` (@cregouby, #236).
 
 ## New models
 

R/dataset-flowers.R

@@ -99,7 +99,7 @@ flowers102_dataset <- dataset(
     meta <- readRDS(file.path(self$processed_folder, glue::glue("{self$split}.rds")))
     self$img_path <- meta$img_path
     self$labels <- meta$labels
-    cli_inform("Split {.val {self$split}} of dataset {.cls {class(self)[[1]]}} loaded with {length(self$img_path)} samples.")
+    cli_inform("Split {.val {self$split}} of dataset {.cls {class(self)[[1]]}} loaded with {self$.length()} samples.")
   },
 
   .getitem = function(index) {

R/dataset-plankton.R

@@ -2,7 +2,13 @@
 #' WHOI-Plankton Dataset
 #'
 #' The WHOI-Plankton and WHOI-Plankton small are **image classification** datasets
-#' of submarine plankton small grayscale images of varying size, classified into 100 classes.
+#' from the Woods Hole Oceanographic Institution (WHOI) of microscopic marine plankton.
+#' https://hdl.handle.net/10.1575/1912/7341
+#' Images were collected in situ by automated submersible imaging-in-flow cytometry
+#' with an instrument called Imaging FlowCytobot (IFCB). They are small grayscale images
+#' of varying size.
+#' Images are classified into 100 classes, with an overview  available in
+#' [project Wiki page](https://whoigit.github.io/whoi-plankton/)
 #' Dataset size is 957k and 58k respectively, and each provides a train / val / test split.
 #'
 #' @inheritParams eurosat_dataset
@@ -66,6 +72,7 @@ whoi_small_plankton_dataset <- torch::dataset(
       install.packages("prettyunits")
     }
 
+    self$split <- match.arg(split, c("train", "val", "test"))
     self$transform <- transform
     self$target_transform <- target_transform
     self$archive_url <- self$resources[self$resources$split == split,]$url
@@ -74,7 +81,7 @@ whoi_small_plankton_dataset <- torch::dataset(
     self$split_file <- sapply(self$archive_url, \(x) file.path(rappdirs::user_cache_dir("torch"), class(self)[1], sub("\\?download=.*", "", basename(x))))
 
     if (download) {
-      cli_inform("Dataset {.cls {class(self)[[1]]}} (~{.emph {self$archive_size}}) will be downloaded and processed if not already available.")
+      cli_inform("Split {.val {self$split}} of dataset {.cls {class(self)[[1]]}} (~{.emph {self$archive_size}}) will be downloaded and processed if not already available.")
       self$download()
     }
 
@@ -106,7 +113,12 @@ whoi_small_plankton_dataset <- torch::dataset(
 
   .getitem = function(index) {
     df <- self$.data[index,]$to_data_frame()
-    x <- df$image$bytes %>% unlist() %>% as.raw() %>% png::readPNG()
+    x_raw <- df$image$bytes %>% unlist() %>% as.raw()
+    if (tolower(tools::file_ext(df$image$path)) == "jpg") {
+      x <- jpeg::readJPEG(x_raw)
+    } else {
+      x <- png::readPNG(x_raw)
+    }
     y <- df$label + 1L
 
     if (!is.null(self$transform))
@@ -127,7 +139,7 @@ whoi_small_plankton_dataset <- torch::dataset(
 
 #' WHOI-Plankton Dataset
 #'
-#' @inheritParams whoi_plankton_dataset#'
+#' @inheritParams whoi_plankton_dataset
 #' @rdname whoi_plankton_dataset
 #' @export
 whoi_plankton_dataset <- torch::dataset(
@@ -162,3 +174,32 @@ whoi_plankton_dataset <- torch::dataset(
     size = c(rep(450e6, 4), rep(490e6, 13), rep(450e6, 2))
   )
 )
+
+
+#' Coralnet Dataset
+#'
+#' Small Coralnet dataset is an image **classification dataset**
+#' of very large submarine coral reef images annotated into 3 classes
+#' and produced by  [CoralNet](https://coralnet.ucsd.edu),
+#' a resource for benthic images classification.
+#'
+#' @inheritParams whoi_plankton_dataset
+#' @export
+whoi_small_coralnet_dataset <- torch::dataset(
+  name = "whoi_small_coralnet",
+  inherit = whoi_small_plankton_dataset,
+  archive_size = "2.1 GB",
+  resources = data.frame(
+    split = c("test", rep("train", 4), "val"),
+    url = c("https://huggingface.co/datasets/nf-whoi/coralnet-small/resolve/main/data/test-00000-of-00001.parquet?download=true",
+            paste0("https://huggingface.co/datasets/nf-whoi/coralnet-small/resolve/main/data/train-0000",0:3,"-of-00004.parquet?download=true"),
+            "https://huggingface.co/datasets/nf-whoi/coralnet-small/resolve/main/data/validation-00000-of-00001.parquet?download=true"),
+    md5 = c("f9a3ce864fdbeb5f1f3d243fe1285186",
+            "82269e2251db22ef213e438126198afd",
+            "82d2cafbad7740e476310565a2bcd44e",
+            "f4dd2d2effc1f9c02918e3ee614b85d3",
+            "d66ec691a4c5c63878a9cfff164a6aaf",
+            "7ea146b9b2f7b6cee99092bd44182d06"),
+    size = c(430e6, rep(380e6, 4), 192e6)
+  )
+)

_pkgdown.yml

@@ -53,6 +53,7 @@ reference:
    - ends_with("1_dataset")
    - ends_with("2_dataset")
    - ends_with("5_dataset")
+   - ends_with("kton_dataset")
 - subtitle: for Object Detection
   descr: >
     Dataset having items with "y" as a named list of bounding-box and

inst/po/fr/LC_MESSAGES/R-torchvision.mo

@@ -1,8 +1,8 @@
 msgid ""
 msgstr ""
 "Project-Id-Version: torchvision 0.7.0.9000\n"
-"POT-Creation-Date: 2025-08-10 20:26+0200\n"
-"PO-Revision-Date: 2025-08-10 20:35+0200\n"
+"POT-Creation-Date: 2025-08-15 10:15+0200\n"
+"PO-Revision-Date: 2025-08-15 10:18+0200\n"
 "Last-Translator: Christophe Regouby <[email protected]>\n"
 "Language-Team: \n"
 "Language: fr\n"
@@ -46,12 +46,12 @@ msgstr ""
 #: dataset-lfw.R:246 dataset-mnist.R:86 dataset-mnist.R:234 dataset-mnist.R:421
 #: dataset-oxfordiiitpet.R:71 dataset-oxfordiiitpet.R:283
 #: dataset-oxfordiiitpet.R:347 dataset-pascal.R:136 dataset-pascal.R:294
-#: dataset-places365.R:98 dataset-plankton.R:71
+#: dataset-places365.R:98
 msgid ""
 "Dataset {.cls {class(self)[[1]]}} (~{.emph {self$archive_size}}) will be "
 "downloaded and processed if not already available."
 msgstr ""
-"Le jeu de données {.cls {class(self)[[1]]}} de taille (~{.emph "
+"Le jeu de données {.cls {class(self)[[1]]}} (de taille ~{.emph "
 "{self$archive_size}}) sera téléchargé et traité s'il n'est pas déjà "
 "disponible."
 
@@ -60,7 +60,7 @@ msgstr ""
 #: dataset-flowers.R:97 dataset-lfw.R:113 dataset-lfw.R:251 dataset-mnist.R:91
 #: dataset-mnist.R:239 dataset-mnist.R:426 dataset-oxfordiiitpet.R:76
 #: dataset-oxfordiiitpet.R:288 dataset-oxfordiiitpet.R:352 dataset-pascal.R:141
-#: dataset-pascal.R:299 dataset-places365.R:103 dataset-plankton.R:76
+#: dataset-pascal.R:299 dataset-places365.R:103 dataset-plankton.R:89
 msgid "Dataset not found. You can use `download = TRUE` to download it."
 msgstr ""
 "Jeu de données introuvable. Veuillez ajouter `download = TRUE` pour le "
@@ -69,7 +69,7 @@ msgstr ""
 #: dataset-caltech.R:84 dataset-caltech.R:208 dataset-lfw.R:135
 #: dataset-lfw.R:286 dataset-oxfordiiitpet.R:95 dataset-oxfordiiitpet.R:302
 #: dataset-oxfordiiitpet.R:366 dataset-pascal.R:149 dataset-pascal.R:317
-#: dataset-plankton.R:83
+#: dataset-plankton.R:93
 msgid ""
 "{.cls {class(self)[[1]]}} dataset loaded with {self$.length()} images across "
 "{length(self$classes)} classes."
@@ -79,15 +79,15 @@ msgstr ""
 
 #: dataset-caltech.R:117 dataset-coco.R:187 dataset-eurosat.R:81
 #: dataset-flickr.R:136 dataset-lfw.R:146 dataset-oxfordiiitpet.R:107
-#: dataset-pascal.R:158 dataset-plankton.R:90
+#: dataset-pascal.R:158 dataset-plankton.R:100
 msgid "Downloading {.cls {class(self)[[1]]}}..."
 msgstr "Téléchargement de {.cls {class(self)[[1]]}}..."
 
 #: dataset-caltech.R:125 dataset-cifar.R:111 dataset-coco.R:193
 #: dataset-eurosat.R:88 dataset-fer.R:129 dataset-flowers.R:136
 #: dataset-lfw.R:157 dataset-lfw.R:172 dataset-mnist.R:121 dataset-mnist.R:264
 #: dataset-mnist.R:453 dataset-oxfordiiitpet.R:115 dataset-pascal.R:168
-#: dataset-places365.R:182 dataset-plankton.R:95 models-vit.R:49
+#: dataset-places365.R:182 dataset-plankton.R:105 models-vit.R:49
 msgid "Corrupt file! Delete the file in {archive} and try again."
 msgstr "Fichier corrompu. Supprimez le fichier {archive} et recommencez."
 
@@ -160,7 +160,7 @@ msgstr "Extraction de l'archive {.cls {class(self)[[1]]}} terminée..."
 #: dataset-eurosat.R:98
 msgid "Downloading {.cls {class(self)[[1]]}} split file: {self$split_url}"
 msgstr ""
-"Téléchargement du fichier de split de {.cls {class(self)[[1]]}} depuis "
+"Téléchargement du fichier de partition de {.cls {class(self)[[1]]}} depuis "
 "{self$split_url} ..."
 
 #: dataset-eurosat.R:102
@@ -214,18 +214,18 @@ msgstr ""
 #: dataset-flowers.R:102
 msgid ""
 "Split {.val {self$split}} of dataset {.cls {class(self)[[1]]}} loaded with "
-"{length(self$img_path)} samples."
+"{self$.length()} samples."
 msgstr ""
-"Le sous-ensemble {.val {self$split}} du jeu de données {.cls {class(self)"
-"[[1]]}} chargé avec {length(self$img_path)} images."
+"La partition {.val {self$split}} du jeu de données {.cls {class(self)[[1]]}} "
+"est chargée avec {length(self$img_path)} images."
 
 #: dataset-flowers.R:126
 msgid ""
 "Split {.val {self$split}} of dataset {.cls {class(self)[[1]]}} is already "
 "processed and cached."
 msgstr ""
-"Le sous-ensemble {.val {self$split}} du jeu de données {.cls {class(self)"
-"[[1]]}} a déjà été préparé et mis en cache."
+"La partition {.val {self$split}} du jeu de données {.cls {class(self)[[1]]}} "
+"a déjà été préparé et mis en cache."
 
 #: dataset-flowers.R:140
 msgid "{.cls {class(self)[[1]]}} Extracting images and processing dataset..."
@@ -262,9 +262,19 @@ msgstr "Partage non valide : {self$split}"
 #: dataset-places365.R:163
 msgid "Downloading {.cls {class(self)[[1]]}} split '{self$split}'..."
 msgstr ""
-"Téléchargement du fichier de partage '{self$split}' de {.cls {class(self)"
+"Téléchargement du fichier de partition '{self$split}' de {.cls {class(self)"
 "[[1]]} ..."
 
+#: dataset-plankton.R:84
+msgid ""
+"Split {.val {self$split}} of dataset {.cls {class(self)[[1]]}} (~{.emph "
+"{self$archive_size}}) will be downloaded and processed if not already "
+"available."
+msgstr ""
+"La partition {.val {self$split}} du jeu de données {.cls {class(self)[[1]]}} "
+"(de taille ~{.emph {self$archive_size}}) sera téléchargée et traitée si elle "
+"n'est pas déjà disponible."
+
 #: extension.R:2
 msgid ""
 "has_ops() Not implemented yet. https://github.com/pytorch/vision/blob/"

man/whoi_plankton_dataset.Rd

@@ -1,7 +1,7 @@
 msgid ""
 msgstr ""
 "Project-Id-Version: torchvision 0.7.0.9000\n"
-"POT-Creation-Date: 2025-08-10 20:26+0200\n"
+"POT-Creation-Date: 2025-08-15 10:15+0200\n"
 "PO-Revision-Date: YEAR-MO-DA HO:MI+ZONE\n"
 "Last-Translator: FULL NAME <EMAIL@ADDRESS>\n"
 "Language-Team: LANGUAGE <[email protected]>\n"
@@ -38,27 +38,27 @@ msgstr ""
 msgid "deprecated"
 msgstr ""
 
-#: dataset-caltech.R:61 dataset-cifar.R:52 dataset-coco.R:85 dataset-coco.R:305 dataset-eurosat.R:57 dataset-fer.R:63 dataset-fgvc.R:91 dataset-flickr.R:69 dataset-flickr.R:231 dataset-flowers.R:92 dataset-lfw.R:108 dataset-lfw.R:246 dataset-mnist.R:86 dataset-mnist.R:234 dataset-mnist.R:421 dataset-oxfordiiitpet.R:71 dataset-oxfordiiitpet.R:283 dataset-oxfordiiitpet.R:347 dataset-pascal.R:136 dataset-pascal.R:294 dataset-places365.R:98 dataset-plankton.R:71
+#: dataset-caltech.R:61 dataset-cifar.R:52 dataset-coco.R:85 dataset-coco.R:305 dataset-eurosat.R:57 dataset-fer.R:63 dataset-fgvc.R:91 dataset-flickr.R:69 dataset-flickr.R:231 dataset-flowers.R:92 dataset-lfw.R:108 dataset-lfw.R:246 dataset-mnist.R:86 dataset-mnist.R:234 dataset-mnist.R:421 dataset-oxfordiiitpet.R:71 dataset-oxfordiiitpet.R:283 dataset-oxfordiiitpet.R:347 dataset-pascal.R:136 dataset-pascal.R:294 dataset-places365.R:98
 msgid ""
 "Dataset {.cls {class(self)[[1]]}} (~{.emph {self$archive_size}}) will be "
 "downloaded and processed if not already available."
 msgstr ""
 
-#: dataset-caltech.R:66 dataset-caltech.R:187 dataset-coco.R:90 dataset-coco.R:310 dataset-eurosat.R:64 dataset-fer.R:70 dataset-flowers.R:97 dataset-lfw.R:113 dataset-lfw.R:251 dataset-mnist.R:91 dataset-mnist.R:239 dataset-mnist.R:426 dataset-oxfordiiitpet.R:76 dataset-oxfordiiitpet.R:288 dataset-oxfordiiitpet.R:352 dataset-pascal.R:141 dataset-pascal.R:299 dataset-places365.R:103 dataset-plankton.R:76
+#: dataset-caltech.R:66 dataset-caltech.R:187 dataset-coco.R:90 dataset-coco.R:310 dataset-eurosat.R:64 dataset-fer.R:70 dataset-flowers.R:97 dataset-lfw.R:113 dataset-lfw.R:251 dataset-mnist.R:91 dataset-mnist.R:239 dataset-mnist.R:426 dataset-oxfordiiitpet.R:76 dataset-oxfordiiitpet.R:288 dataset-oxfordiiitpet.R:352 dataset-pascal.R:141 dataset-pascal.R:299 dataset-places365.R:103 dataset-plankton.R:89
 msgid "Dataset not found. You can use `download = TRUE` to download it."
 msgstr ""
 
-#: dataset-caltech.R:84 dataset-caltech.R:208 dataset-lfw.R:135 dataset-lfw.R:286 dataset-oxfordiiitpet.R:95 dataset-oxfordiiitpet.R:302 dataset-oxfordiiitpet.R:366 dataset-pascal.R:149 dataset-pascal.R:317 dataset-plankton.R:83
+#: dataset-caltech.R:84 dataset-caltech.R:208 dataset-lfw.R:135 dataset-lfw.R:286 dataset-oxfordiiitpet.R:95 dataset-oxfordiiitpet.R:302 dataset-oxfordiiitpet.R:366 dataset-pascal.R:149 dataset-pascal.R:317 dataset-plankton.R:93
 msgid ""
 "{.cls {class(self)[[1]]}} dataset loaded with {self$.length()} images across "
 "{length(self$classes)} classes."
 msgstr ""
 
-#: dataset-caltech.R:117 dataset-coco.R:187 dataset-eurosat.R:81 dataset-flickr.R:136 dataset-lfw.R:146 dataset-oxfordiiitpet.R:107 dataset-pascal.R:158 dataset-plankton.R:90
+#: dataset-caltech.R:117 dataset-coco.R:187 dataset-eurosat.R:81 dataset-flickr.R:136 dataset-lfw.R:146 dataset-oxfordiiitpet.R:107 dataset-pascal.R:158 dataset-plankton.R:100
 msgid "Downloading {.cls {class(self)[[1]]}}..."
 msgstr ""
 
-#: dataset-caltech.R:125 dataset-cifar.R:111 dataset-coco.R:193 dataset-eurosat.R:88 dataset-fer.R:129 dataset-flowers.R:136 dataset-lfw.R:157 dataset-lfw.R:172 dataset-mnist.R:121 dataset-mnist.R:264 dataset-mnist.R:453 dataset-oxfordiiitpet.R:115 dataset-pascal.R:168 dataset-places365.R:182 dataset-plankton.R:95 models-vit.R:49
+#: dataset-caltech.R:125 dataset-cifar.R:111 dataset-coco.R:193 dataset-eurosat.R:88 dataset-fer.R:129 dataset-flowers.R:136 dataset-lfw.R:157 dataset-lfw.R:172 dataset-mnist.R:121 dataset-mnist.R:264 dataset-mnist.R:453 dataset-oxfordiiitpet.R:115 dataset-pascal.R:168 dataset-places365.R:182 dataset-plankton.R:105 models-vit.R:49
 msgid "Corrupt file! Delete the file in {archive} and try again."
 msgstr ""
 
@@ -158,7 +158,7 @@ msgstr ""
 #: dataset-flowers.R:102
 msgid ""
 "Split {.val {self$split}} of dataset {.cls {class(self)[[1]]}} loaded with "
-"{length(self$img_path)} samples."
+"{self$.length()} samples."
 msgstr ""
 
 #: dataset-flowers.R:126
@@ -198,6 +198,13 @@ msgstr ""
 msgid "Downloading {.cls {class(self)[[1]]}} split '{self$split}'..."
 msgstr ""
 
+#: dataset-plankton.R:84
+msgid ""
+"Split {.val {self$split}} of dataset {.cls {class(self)[[1]]}} (~{.emph "
+"{self$archive_size}}) will be downloaded and processed if not already "
+"available."
+msgstr ""
+
 #: extension.R:2
 msgid ""
 "has_ops() Not implemented yet. https://github.com/pytorch/vision/blob/"