Update renovation.py

renovation.py

@@ -6,6 +6,7 @@ from io import BytesIO
 from datasets.tasks import ImageClassification
 import pyarrow as pa
 
+
 _HOMEPAGE = "https://huggingface.co/datasets/rshrott/renovation"
 
 _CITATION = """\
@@ -36,6 +37,7 @@ class RenovationQualityDataset(datasets.GeneratorBasedBuilder):
  description=_DESCRIPTION,
  features=datasets.Features(
  {
+ "image_file_path": datasets.Value("string"),
  "image": datasets.Value(pa.binary()),
  "label": datasets.features.ClassLabel(names=_NAMES),
  }
@@ -77,7 +79,7 @@ class RenovationQualityDataset(datasets.GeneratorBasedBuilder):
  response = requests.get(url)
  img = Image.open(BytesIO(response.content))
  return img
- 
+
  with open(filepath, "r") as f:
  reader = csv.reader(f)
  next(reader) # skip header
@@ -86,9 +88,9 @@ class RenovationQualityDataset(datasets.GeneratorBasedBuilder):
  rows = rows[:int(0.8 * len(rows))]
  elif split == 'validation':
  rows = rows[int(0.8 * len(rows)):int(0.9 * len(rows))]
- else: # test
+ else: # test
  rows = rows[int(0.9 * len(rows)):]
- 
+
  for id_, row in enumerate(rows):
  if len(row) < 2:
  print(f"Row with id {id_} has less than 2 elements: {row}")
@@ -96,8 +98,7 @@ class RenovationQualityDataset(datasets.GeneratorBasedBuilder):
  image = url_to_image(row[0])
  image_bytes = image.tobytes()
  yield id_, {
- 'image_file_path': str(row[0]), # Convert to string
+ 'image_file_path': row[0],
  'image': image_bytes,
  'label': row[1],
  }
-