davanstrien HF staff commited on
Commit
a7fa5c2
1 Parent(s): 1c5de89
Files changed (2) hide show
  1. README.md +1 -1
  2. amazonian_fish_classifier_data.py +14 -9
README.md CHANGED
@@ -44,6 +44,6 @@ dataset_info:
44
  - name: train
45
  num_bytes: 578234
46
  num_examples: 3068
47
- download_size: 330399200
48
  dataset_size: 578234
49
  ---
 
44
  - name: train
45
  num_bytes: 578234
46
  num_examples: 3068
47
+ download_size: 330476983
48
  dataset_size: 578234
49
  ---
amazonian_fish_classifier_data.py CHANGED
@@ -13,8 +13,9 @@
13
  # limitations under the License.
14
  """TODO."""
15
 
 
16
  import os
17
- from pathlib import Path
18
  import datasets
19
 
20
  _CITATION = """TODO"""
@@ -24,13 +25,14 @@ _DESCRIPTION = """\
24
  TODO
25
  """
26
 
27
- _HOMEPAGE = "https://doi.org/10.25573/data.17314730.v1"
28
 
29
  _LICENSE = "CC BY 4.0"
30
 
31
 
32
  _URLS = {
33
  "images": "https://smithsonian.figshare.com/ndownloader/files/31975544",
 
34
  }
35
 
36
 
@@ -93,19 +95,22 @@ class AmazonianFish(datasets.GeneratorBasedBuilder):
93
 
94
  def _split_generators(self, dl_manager):
95
  images = dl_manager.download_and_extract(_URLS["images"])
 
 
 
96
  return [
97
  datasets.SplitGenerator(
98
  name=datasets.Split.TRAIN,
99
  gen_kwargs={
100
  "images": os.path.join(images, "training_images"),
 
101
  },
102
  ),
103
  ]
104
 
105
- def _generate_examples(self, images):
106
- id_ = 0
107
- for example in Path(images).rglob("*.jpg"):
108
- if example.name.startswith("._"):
109
- continue
110
- id_ += 1
111
- yield id_, {"image": str(example), "label": example.parent.name}
 
13
  # limitations under the License.
14
  """TODO."""
15
 
16
+
17
  import os
18
+ import pandas as pd
19
  import datasets
20
 
21
  _CITATION = """TODO"""
 
25
  TODO
26
  """
27
 
28
+ _HOMEPAGE = "https://doi.org/10.25573/data.17314730.v1"
29
 
30
  _LICENSE = "CC BY 4.0"
31
 
32
 
33
  _URLS = {
34
  "images": "https://smithsonian.figshare.com/ndownloader/files/31975544",
35
+ "labels": "https://smithsonian.figshare.com/ndownloader/files/31975646",
36
  }
37
 
38
 
 
95
 
96
  def _split_generators(self, dl_manager):
97
  images = dl_manager.download_and_extract(_URLS["images"])
98
+ labels = dl_manager.download(_URLS["labels"])
99
+ df = pd.read_csv(labels)
100
+ labels = df.to_dict(orient="records")
101
  return [
102
  datasets.SplitGenerator(
103
  name=datasets.Split.TRAIN,
104
  gen_kwargs={
105
  "images": os.path.join(images, "training_images"),
106
+ "labels": labels,
107
  },
108
  ),
109
  ]
110
 
111
+ def _generate_examples(self, images, labels):
112
+ for id_, example in enumerate(labels):
113
+ yield id_, {
114
+ "image": os.path.join(images, example["Genus"], example["Image_name"]),
115
+ "label": example["Genus"],
116
+ }