Update script to hub

Boat_dataset.py

@@ -6,22 +6,19 @@ import os
 
 import datasets
 
-
 _CITATION = """\
 @InProceedings{huggingface:dataset,
 title = {Boat dataset},
-author={Tzu-Chi Chen, Inc.
-},
+author={Tzu-Chi Chen, Inc.},
 year={2024}
 }
 """
 
-
 _DESCRIPTION = """\
-This dataset is designed to solve object detection task.
+This dataset is designed to solve an object detection task with images of boats.
 """
 
-_HOMEPAGE = "https://huggingface.co/datasets/zhuchi76/Boat_dataset/resolve/main"
+_HOMEPAGE = "https://huggingface.co/datasets/zhuchi76/Boat_dataset"
 
 _LICENSE = ""
 
@@ -69,8 +66,8 @@ class BoatDataset(datasets.GeneratorBasedBuilder):
         # Download all files and extract them
         downloaded_files = dl_manager.download_and_extract(_URLS)
 
-        # Paths to files and archives
-        image_archive_path = downloaded_files["images"]
+        # Extract the image archive
+        image_dir = dl_manager.extract(downloaded_files["images"])
         classes_file_path = downloaded_files["classes"]
         annotations_paths = downloaded_files["anno"]
 
@@ -82,7 +79,7 @@ class BoatDataset(datasets.GeneratorBasedBuilder):
             datasets.SplitGenerator(
                 name=datasets.Split.TRAIN,
                 gen_kwargs={
-                    "image_archive_path": image_archive_path,
+                    "image_dir": image_dir,
                     "annotations_file": annotations_paths["train"],
                     "classes": classes,
                     "split": "train",
@@ -91,7 +88,7 @@ class BoatDataset(datasets.GeneratorBasedBuilder):
             datasets.SplitGenerator(
                 name=datasets.Split.VALIDATION,
                 gen_kwargs={
-                    "image_archive_path": image_archive_path,
+                    "image_dir": image_dir,
                     "annotations_file": annotations_paths["val"],
                     "classes": classes,
                     "split": "val",
@@ -100,7 +97,7 @@ class BoatDataset(datasets.GeneratorBasedBuilder):
             datasets.SplitGenerator(
                 name=datasets.Split.TEST,
                 gen_kwargs={
-                    "image_archive_path": image_archive_path,
+                    "image_dir": image_dir,
                     "annotations_file": annotations_paths["test"],
                     "classes": classes,
                     "split": "val_real",
@@ -108,23 +105,18 @@ class BoatDataset(datasets.GeneratorBasedBuilder):
             ),
         ]
 
-    def _generate_examples(self, dl_manager, image_archive_path, annotations_file, classes, split):
-        # Open the archive using iter_archive to access files without extracting them
-        with dl_manager.iter_archive(image_archive_path) as archive:
-            # Convert path: file-like object pairs to a dict for quick access
-            images_dict = {os.path.basename(path): file for path, file in archive}
-
+    def _generate_examples(self, image_dir, annotations_file, classes, split):
         # Process annotations
         with open(annotations_file, encoding="utf-8") as f:
             for key, row in enumerate(f):
                 try:
                     data = json.loads(row.strip())
-                    file = images_dict.get(data["file_name"])
-                    if not file:
-                        continue  # Skip if file is not found in the archive
+                    file_path = os.path.join(image_dir, data["file_name"])
+                    if not os.path.isfile(file_path):
+                        continue  # Skip if file is not found in the directory
                     yield key, {
                         "image_id": data["image_id"],
-                        "image_path": file,  # You can save the path or file object depending on your processing needs
+                        "image_path": file_path,  # Provide the full path to the image
                         "width": data["width"],
                         "height": data["height"],
                         "objects": {