Update dataset.py

dataset.py

@@ -1,5 +1,5 @@
 import os
-from datasets import DatasetInfo, GeneratorBasedBuilder, Split, Features, Image, ClassLabel
+from datasets import DatasetInfo, GeneratorBasedBuilder, Features, Image, ClassLabel, Split
 from huggingface_hub import hf_hub_download
 
 class UIR25MReference(GeneratorBasedBuilder):
@@ -7,11 +7,11 @@ class UIR25MReference(GeneratorBasedBuilder):
 
     VERSION = "1.0.0"
 
-    # Directory to store pretrained models
+    # Directory where pretrained models will be downloaded
     PRETRAINED_DIR = os.path.expanduser("./pretrained_models")
     os.makedirs(PRETRAINED_DIR, exist_ok=True)
 
-    # List of pretrained model files
+    # List of pretrained model files with Hugging Face URLs
     PRETRAINED_MODELS = [
         "nafnet_maskdcpt_12d.pth",
         "nafnet_maskdcpt_5d.pth",
@@ -21,22 +21,22 @@ class UIR25MReference(GeneratorBasedBuilder):
         "swinir_maskdcpt_5d.pth"
     ]
 
-    HF_REPO = "Jiakui/MaskDCPT"
+    PRETRAINED_REPO = "Jiakui/MaskDCPT"
 
-    @staticmethod
-    def download_pretrained_models():
+    def download_pretrained_models(self):
+        """Download all pretrained models to local directory."""
         downloaded_files = []
-        for filename in UIR25MReference.PRETRAINED_MODELS:
-            target_path = os.path.join(UIR25MReference.PRETRAINED_DIR, filename)
+        for filename in self.PRETRAINED_MODELS:
+            target_path = os.path.join(self.PRETRAINED_DIR, filename)
             if not os.path.exists(target_path):
-                print(f"Downloading {filename}...")
+                print(f"Downloading pretrained model: {filename}")
                 hf_hub_download(
-                    repo_id=UIR25MReference.HF_REPO,
+                    repo_id=self.PRETRAINED_REPO,
                     filename=f"pretrained_models/{filename}",
-                    local_dir=UIR25MReference.PRETRAINED_DIR
+                    local_dir=self.PRETRAINED_DIR
                 )
             else:
-                print(f"{filename} already exists, skipping.")
+                print(f"{filename} already exists, skipping download.")
             downloaded_files.append(target_path)
         return downloaded_files
 
@@ -51,10 +51,10 @@ class UIR25MReference(GeneratorBasedBuilder):
                 "degradation_type": ClassLabel(names=[
                     "noise", "blur", "compression", "haze", "low_light",
                     "degradation_5", "degradation_6", "degradation_7",
-                    "degradation_8", "degradation_9", "degradation_10",
-                    "degradation_11", "degradation_12", "degradation_13",
-                    "degradation_14", "degradation_15", "degradation_16",
-                    "degradation_17", "degradation_18", "degradation_19"
+                    "degradation_8", "degradation_9",
+                    "degradation_10", "degradation_11", "degradation_12",
+                    "degradation_13", "degradation_14", "degradation_15",
+                    "degradation_16", "degradation_17", "degradation_18", "degradation_19"
                 ])
             }),
             task_categories=["image-to-image"],
@@ -62,30 +62,36 @@ class UIR25MReference(GeneratorBasedBuilder):
         )
 
     def _split_generators(self, dl_manager):
-        data_dir = os.path.expanduser(dl_manager.download_and_extract("https://github.com/MILab-PKU/MaskDCPT.git"))
-        return [
-            self.SplitGenerator(
-                name=Split.TRAIN,
-                gen_kwargs={"images_dir": os.path.join(data_dir, "train")}
-            ),
-            # Optionally add test split
-        ]
+        """
+        Leverage automatic split detection:
+        Expects directories named `train`, `test`, `validation` in external path
+        """
+        data_dir = dl_manager.download_and_extract("https://github.com/MILab-PKU/MaskDCPT.git")
+        splits = []
+        for split_name in ["train", "test", "validation"]:
+            split_path = os.path.join(data_dir, split_name)
+            if os.path.exists(split_path):
+                splits.append(
+                    self.SplitGenerator(
+                        name=getattr(Split, split_name.upper()),
+                        gen_kwargs={"images_dir": split_path}
+                    )
+                )
+        return splits
 
     def _generate_examples(self, images_dir):
-        for idx, fname in enumerate(os.listdir(images_dir)):
-            if fname.endswith(".png") or fname.endswith(".jpg"):
-                low_quality_path = os.path.join(images_dir, fname)
-                high_quality_path = os.path.join(images_dir, "high_quality", fname)
-                degradation_type = 0  # Replace with actual mapping from metadata
+        """Yield examples for Hugging Face dataset."""
+        low_dir = os.path.join(images_dir, "low_quality")
+        high_dir = os.path.join(images_dir, "high_quality")
+
+        # Expect a mapping from low to high quality images and degradation type metadata
+        for idx, fname in enumerate(os.listdir(low_dir)):
+            if fname.lower().endswith((".png", ".jpg", ".jpeg")):
+                low_path = os.path.join(low_dir, fname)
+                high_path = os.path.join(high_dir, fname)
+                degradation_type = 0  # Replace with actual metadata mapping if available
                 yield idx, {
-                    "low_quality": low_quality_path,
-                    "high_quality": high_quality_path,
+                    "low_quality": low_path,
+                    "high_quality": high_path,
                     "degradation_type": degradation_type
-                }
-
-# Example: download pretrained models if needed
-if __name__ == "__main__":
-    downloaded_models = UIR25MReference.download_pretrained_models()
-    print("Downloaded pretrained models:")
-    for path in downloaded_models:
-        print(path)
+                }