Create mhr_recognize_datesets.py

mhr_recognize_datesets.py

@@ -0,0 +1,66 @@
+import datasets
+import pandas as pd
+from pathlib import Path
+
+_DESCRIPTION = """\
+Monster Hunter Rise images and labels.
+"""
+
+
+_DATA_URL = {
+    "whole_image": "https://huggingface.co/datasets/miojizzy/mhr_recognize_datasets/data/whole/train.zip",
+    "whole_label": "https://huggingface.co/datasets/miojizzy/mhr_recognize_datasets/data/whole/label.zip",
+    "hole_image": "https://huggingface.co/datasets/miojizzy/mhr_recognize_datasets/data/hole/train.zip",
+    "hole_label": "https://huggingface.co/datasets/miojizzy/mhr_recognize_datasets/data/hole/label.zip",
+    "skill_image": "https://huggingface.co/datasets/miojizzy/mhr_recognize_datasets/data/skill/train.zip",
+    "skill_label": "https://huggingface.co/datasets/miojizzy/mhr_recognize_datasets/data/skill/label.zip",
+}
+
+class MHRRecognizeDatasetsConfig(datasets.BuilderConfig):
+    def __init__(self, name, **kwargs):
+        super(MHRRecognizeDatasetsConfig, self).__init__(
+		version=datasets.Version("1.0.0"),
+		name=name,
+		description=_DESCRIPTION,
+		**kwargs,
+	    )
+
+class MHRRecognizeDatasets(datasets.GeneratorBasedBuilder):
+    DEFAULT_WRITER_BATCH_SIZE = 8
+    
+    BUILDER_CONFIGS = [
+        MHRRecognizeDatasetsConfig("whole"),
+        MHRRecognizeDatasetsConfig("hole"),
+        MHRRecognizeDatasetsConfig("skill"),
+    ]
+
+    def _info(self):
+        return datasets.DatasetInfo(
+            description=_DESCRIPTION,
+            features=datasets.Features(
+                {
+                    "image": datasets.Image(),
+                    "label": -1,
+                }
+            )
+        )
+
+    def _split_generators(self, dl_manager):
+        """Returns SplitGenerators."""
+        download_files = dl_manager.download_and_extract(_DATA_URL)
+        print(download_files)
+        return [
+            datasets.SplitGenerator(
+                name=datasets.Split.TRAIN,
+                gen_kwargs=download_files,
+            )
+            ]
+
+    def _generate_examples(self, image_dir, metadata):
+        """Yields examples."""
+        metadata_df = pd.read_csv(metadata, index_col=0)
+        print(metadata_df)
+        for idx, row in metadata_df.iterrows():
+            image_path = Path(image_dir) / row["file_name"]
+            sample_dict = {'image': str(image_path), 'label': row}
+            yield idx, sample_dict