Fixing mask archive path

Signed-off-by: Jiri Podivin <[email protected]>

metadata_semantic_test.csv

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:44dc86e46441a8ead49263facb19293912482a992d02d3a07754a8855994779e
-size 148053
+oid sha256:e142e80f70774bb055767f4484704f89dd7edf2dfc0db456ef4625cfe1949cf5
+size 148063

metadata_semantic_train.csv

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:4931badcc1fd59cc6d64bbecaba36b100205b7d25edbe456a8fd7395371d10eb
-size 591777
+oid sha256:cb2a941de28584d065e441b56bbc25aaa2870e66cf3d2c506c7d2c2b4a1a3250
+size 591787

plantorgans.py

@@ -16,7 +16,7 @@ _BASE_URL = "https://huggingface.co/datasets/jpodivin/plantorgans/resolve/main/"
 _TRAIN_URLS = [_BASE_URL + f"sourcedata_labeled.tar.{i:02}" for i in range(0, 8)]
 _TEST_URLS = [_BASE_URL + f"sourcedata_labeled.tar.{i:02}" for i in range(8, 12)]
 _MASKS_URLS = [_BASE_URL + f"masks.tar.0{i}" for i in range(0, 2)]
-_SEMANTIC_MASKS_URLS = [_BASE_URL + f"semantic_masks.tar.0{i}" for i in range(0, 2)]
+_SEMANTIC_MASKS_URLS = "semantic_masks.tar.gz"
 
 _SEMANTIC_METADATA_URLS = {
     'train': 'https://huggingface.co/datasets/jpodivin/plantorgans/resolve/main/metadata_semantic_train.csv',
@@ -28,6 +28,7 @@ _PANOPTIC_METADATA_URLS = {
     'test': 'https://huggingface.co/datasets/jpodivin/plantorgans/resolve/main/metadata_test.csv'
 }
 
+
 class PlantOrgansConfig(datasets.BuilderConfig):
     """Builder Config for PlantOrgans"""
  
@@ -70,14 +71,7 @@ class PlantOrgans(datasets.GeneratorBasedBuilder):
                     "image": datasets.Image(),
                     "mask": datasets.Image(),
                     "image_name": datasets.Value(dtype="string"),
-                    "class": datasets.ClassLabel(
-                        names=['Fruit', 'Leaf', 'Flower', 'Stem']),
                 })
-        if self.config.name == 'instance_segmentation_full':
-            features['score'] = datasets.Value(dtype="double")
-        else:
-            features['class'] = datasets.ClassLabel(
-                    names=['Fruit', 'Leaf', 'Flower', 'Stem'])
         return datasets.DatasetInfo(
             description=_DESCRIPTION,
             features=features,
@@ -104,7 +98,7 @@ class PlantOrgans(datasets.GeneratorBasedBuilder):
         if self.config.name == 'instance_segmentation_full':
             metadata_urls = _PANOPTIC_METADATA_URLS
             mask_urls = _MASKS_URLS
-            mask_glob = '/_masks/**.png'
+            mask_glob = '/masks/**.png'
         else:
             metadata_urls = _SEMANTIC_METADATA_URLS
             mask_urls = _SEMANTIC_MASKS_URLS
@@ -155,6 +149,8 @@ class PlantOrgans(datasets.GeneratorBasedBuilder):
         
         # Get all common about images and masks from csv
         metadata = pd.read_csv(metadata_path)
+        metadata['image'] = metadata['image_path'].apply(lambda x: str(Path(x).parts[-1]))
+        metadata['mask'] = metadata['mask_path'].apply(lambda x: str(Path(x).parts[-1]))
 
         # Merge dataframes
         metadata = metadata.merge(masks_paths, on='mask', how='inner')
@@ -168,10 +164,5 @@ class PlantOrgans(datasets.GeneratorBasedBuilder):
                 'mask': r['mask_path'],
                 'image': r['image_path'],
                 'image_name': Path(r['image_path']).parts[-1],
-                'class': r['class']
             }
-            if self.config.name == 'instance_segmentation_full':
-                example['score'] = r['score']
-            else:
-                example['class'] = r['class']
             yield i, example