skytnt commited on
Commit
7bd707d
1 Parent(s): 106d446

Update fbanimehq.py

Browse files
Files changed (1) hide show
  1. fbanimehq.py +21 -7
fbanimehq.py CHANGED
@@ -1,6 +1,5 @@
1
  import os
2
  import datasets
3
- import PIL
4
  from datasets import DownloadManager, DatasetInfo
5
 
6
  _DESCRIPTION = """\
@@ -8,16 +7,31 @@ FBAnimeHQ is a dataset with high-quality full-body anime girl images in a resolu
8
  """
9
  _HOMEPAGE = "https://huggingface.co/datasets/skytnt/fbanimehq"
10
  _URL_BASE = "https://huggingface.co/datasets/skytnt/fbanimehq/resolve/main/data/"
11
- _URLS = [
 
12
  _URL_BASE + "fbanimehq-00.zip",
13
  _URL_BASE + "fbanimehq-01.zip",
14
  _URL_BASE + "fbanimehq-02.zip",
15
  _URL_BASE + "fbanimehq-03.zip",
16
  _URL_BASE + "fbanimehq-04.zip",
17
  ]
 
 
 
 
 
 
 
 
 
 
18
 
19
 
20
  class FBAnimeHQ(datasets.GeneratorBasedBuilder):
 
 
 
 
21
 
22
  def _info(self) -> DatasetInfo:
23
  return datasets.DatasetInfo(
@@ -33,14 +47,14 @@ class FBAnimeHQ(datasets.GeneratorBasedBuilder):
33
  )
34
 
35
  def _split_generators(self, dl_manager: DownloadManager):
36
- downloaded_files = dl_manager.download_and_extract(_URLS)
37
- return [datasets.SplitGenerator(name=datasets.Split.TRAIN, gen_kwargs={"filepath": downloaded_files})]
38
 
39
- def _generate_examples(self, filepath):
40
- for path in filepath:
41
  all_fnames = {os.path.relpath(os.path.join(root, fname), start=path)
42
  for root, _dirs, files in os.walk(path) for fname in files}
43
  image_fnames = sorted(fname for fname in all_fnames
44
- if os.path.splitext(fname)[1].lower() in PIL.Image.EXTENSION)
45
  for image_fname in image_fnames:
46
  yield image_fname, {"image": os.path.join(path, image_fname)}
 
1
  import os
2
  import datasets
 
3
  from datasets import DownloadManager, DatasetInfo
4
 
5
  _DESCRIPTION = """\
 
7
  """
8
  _HOMEPAGE = "https://huggingface.co/datasets/skytnt/fbanimehq"
9
  _URL_BASE = "https://huggingface.co/datasets/skytnt/fbanimehq/resolve/main/data/"
10
+ _EXTENSION = [".png", ".jpg"]
11
+ _URLS_HQ = [
12
  _URL_BASE + "fbanimehq-00.zip",
13
  _URL_BASE + "fbanimehq-01.zip",
14
  _URL_BASE + "fbanimehq-02.zip",
15
  _URL_BASE + "fbanimehq-03.zip",
16
  _URL_BASE + "fbanimehq-04.zip",
17
  ]
18
+ _URLS_SMALL = [
19
+ _URL_BASE + "fbanimesmall.zip",
20
+ ]
21
+
22
+
23
+ class FBAnimeHQConfig(datasets.BuilderConfig):
24
+
25
+ def __init__(self, urls, **kwargs):
26
+ super(FBAnimeHQConfig, self).__init__(**kwargs)
27
+ self.urls = urls
28
 
29
 
30
  class FBAnimeHQ(datasets.GeneratorBasedBuilder):
31
+ BUILDER_CONFIGS = [
32
+ FBAnimeHQConfig(name="hq", urls=_URLS_HQ),
33
+ FBAnimeHQConfig(name="small", urls=_URLS_SMALL)
34
+ ]
35
 
36
  def _info(self) -> DatasetInfo:
37
  return datasets.DatasetInfo(
 
47
  )
48
 
49
  def _split_generators(self, dl_manager: DownloadManager):
50
+ dirs = dl_manager.download_and_extract(self.config.urls)
51
+ return [datasets.SplitGenerator(name=datasets.Split.TRAIN, gen_kwargs={"dirs": dirs})]
52
 
53
+ def _generate_examples(self, dirs):
54
+ for path in dirs:
55
  all_fnames = {os.path.relpath(os.path.join(root, fname), start=path)
56
  for root, _dirs, files in os.walk(path) for fname in files}
57
  image_fnames = sorted(fname for fname in all_fnames
58
+ if os.path.splitext(fname)[1].lower() in _EXTENSION)
59
  for image_fname in image_fnames:
60
  yield image_fname, {"image": os.path.join(path, image_fname)}