Cyrile commited on
Commit
5a48d5c
1 Parent(s): 3fb3a01

Upload aftdb.py

Browse files
Files changed (1) hide show
  1. aftdb.py +51 -27
aftdb.py CHANGED
@@ -43,14 +43,13 @@ class AFTConfig(datasets.BuilderConfig):
43
  """Builder Config for AFTdb"""
44
 
45
  def __init__(self, nb_files_figure, nb_files_table, **kwargs):
46
- """BuilderConfig for AFTdb."""
47
  super().__init__(version=datasets.__version__, **kwargs)
48
  self.nb_files_figure = nb_files_figure
49
  self.nb_files_table = nb_files_table
50
 
51
 
52
  class AFT_Dataset(datasets.GeneratorBasedBuilder):
53
- """Arxiv Figure Table (AFT) dataset"""
54
 
55
  BUILDER_CONFIGS = [
56
  AFTConfig(
@@ -59,7 +58,7 @@ class AFT_Dataset(datasets.GeneratorBasedBuilder):
59
  "Dataset containing scientific article figures associated "
60
  "with their caption, summary, and article title."
61
  ),
62
- data_dir="./{type}",
63
  nb_files_figure=_NB_TAR_FIGURE,
64
  nb_files_table=None
65
  ),
@@ -71,7 +70,7 @@ class AFT_Dataset(datasets.GeneratorBasedBuilder):
71
  "representation of the table, including its caption, summary, "
72
  "and article title."
73
  ),
74
- data_dir="./{type}",
75
  nb_files_figure=None,
76
  nb_files_table=_NB_TAR_TABLE
77
  ),
@@ -83,7 +82,7 @@ class AFT_Dataset(datasets.GeneratorBasedBuilder):
83
  "textual representation of the table, including its caption, "
84
  "summary, and article title."
85
  ),
86
- data_dir="./{type}",
87
  nb_files_figure=_NB_TAR_FIGURE,
88
  nb_files_table=_NB_TAR_TABLE
89
  )
@@ -136,32 +135,57 @@ class AFT_Dataset(datasets.GeneratorBasedBuilder):
136
  data_dir=self.config.data_dir.format(type='table'),
137
  nb_files=self.config.nb_files_table
138
  )
 
139
  if dl_manager.is_streaming:
140
  downloaded_files = dl_manager.download(all_path)
141
- downloaded_files['train'] = [
142
- dl_manager.iter_archive(ii) for ii in downloaded_files['train']
143
- ]
144
- downloaded_files['test'] = [
145
- dl_manager.iter_archive(ii) for ii in downloaded_files['test']
146
- ]
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
147
  else:
148
- downloaded_files = dl_manager.download_and_extract(all_path)
149
- return [
150
- datasets.SplitGenerator(
151
- name=datasets.Split.TRAIN,
152
- gen_kwargs={
153
- 'filepaths': downloaded_files['train'],
154
- 'is_streaming': dl_manager.is_streaming
155
- }
156
- ),
157
- datasets.SplitGenerator(
158
- name=datasets.Split.TEST,
159
- gen_kwargs={
160
- "filepaths": downloaded_files['test'],
161
- 'is_streaming': dl_manager.is_streaming
162
- }
 
 
 
 
 
 
163
  )
164
- ]
165
 
166
  def _generate_examples(self, filepaths, is_streaming):
167
  if is_streaming:
 
43
  """Builder Config for AFTdb"""
44
 
45
  def __init__(self, nb_files_figure, nb_files_table, **kwargs):
 
46
  super().__init__(version=datasets.__version__, **kwargs)
47
  self.nb_files_figure = nb_files_figure
48
  self.nb_files_table = nb_files_table
49
 
50
 
51
  class AFT_Dataset(datasets.GeneratorBasedBuilder):
52
+ """Arxiv Figure Table database (AFTdb)"""
53
 
54
  BUILDER_CONFIGS = [
55
  AFTConfig(
 
58
  "Dataset containing scientific article figures associated "
59
  "with their caption, summary, and article title."
60
  ),
61
+ data_dir="./data/arxiv_dataset/{type}", # A modiféer sur Huggingface Hub
62
  nb_files_figure=_NB_TAR_FIGURE,
63
  nb_files_table=None
64
  ),
 
70
  "representation of the table, including its caption, summary, "
71
  "and article title."
72
  ),
73
+ data_dir="./data/arxiv_dataset/{type}", # A modiféer sur Huggingface Hub
74
  nb_files_figure=None,
75
  nb_files_table=_NB_TAR_TABLE
76
  ),
 
82
  "textual representation of the table, including its caption, "
83
  "summary, and article title."
84
  ),
85
+ data_dir="./data/arxiv_dataset/{type}", # A modiféer sur Huggingface Hub
86
  nb_files_figure=_NB_TAR_FIGURE,
87
  nb_files_table=_NB_TAR_TABLE
88
  )
 
135
  data_dir=self.config.data_dir.format(type='table'),
136
  nb_files=self.config.nb_files_table
137
  )
138
+ dataset = []
139
  if dl_manager.is_streaming:
140
  downloaded_files = dl_manager.download(all_path)
141
+ dataset.append(
142
+ datasets.SplitGenerator(
143
+ name=datasets.Split.TRAIN,
144
+ gen_kwargs={
145
+ 'filepaths': [
146
+ dl_manager.iter_archive(ii)
147
+ for ii in downloaded_files['train']
148
+ ],
149
+ 'is_streaming': dl_manager.is_streaming
150
+ }
151
+ )
152
+ )
153
+ dataset.append(
154
+ datasets.SplitGenerator(
155
+ name=datasets.Split.TEST,
156
+ gen_kwargs={
157
+ "filepaths": [
158
+ dl_manager.iter_archive(ii)
159
+ for ii in downloaded_files['test']
160
+ ],
161
+ 'is_streaming': dl_manager.is_streaming
162
+ }
163
+ )
164
+ )
165
  else:
166
+ dataset.append(
167
+ datasets.SplitGenerator(
168
+ name=datasets.Split.TRAIN,
169
+ gen_kwargs={
170
+ 'filepaths': dl_manager.download_and_extract(
171
+ all_path['train']
172
+ ),
173
+ 'is_streaming': dl_manager.is_streaming
174
+ }
175
+ )
176
+ )
177
+ dataset.append(
178
+ datasets.SplitGenerator(
179
+ name=datasets.Split.TEST,
180
+ gen_kwargs={
181
+ "filepaths": dl_manager.download_and_extract(
182
+ all_path['test']
183
+ ),
184
+ 'is_streaming': dl_manager.is_streaming
185
+ }
186
+ )
187
  )
188
+ return dataset
189
 
190
  def _generate_examples(self, filepaths, is_streaming):
191
  if is_streaming: