Datasets:
GEM
/

Tasks:
Other
Modalities:
Text
Languages:
English
ArXiv:
Tags:
question-generation
License:
Abinaya Mahendiran commited on
Commit
88991a8
1 Parent(s): c1bf92d

Updated data loader script

Browse files
Files changed (1) hide show
  1. squad_v2.py +15 -67
squad_v2.py CHANGED
@@ -28,24 +28,14 @@ _DESCRIPTION = """\
28
  """
29
 
30
  _URLS = {
31
- "challenge_sets": {
32
- "source_data_split": {
33
- "train": "./challenge_sets/source_data_split/train.json",
34
- "test": "./challenge_sets/source_data_split/test.json",
35
- "validation": "./challenge_sets/source_data_split/validation.json",
36
- },
37
- "question_type_split": {
38
- "train": "./challenge_sets/question_type_split/challenge_question_filter_train.json",
39
- "test": "./challenge_sets/question_type_split/challenge_question_filter_test.json",
40
- "validation": "./challenge_sets/question_type_split/challenge_question_filter_validation.json",
41
- },
42
- "answerable_question_split": {
43
- "train": "./challenge_sets/answerable_question_split/challenge_answer_filter_train.json",
44
- "test": "./challenge_sets/answerable_question_split/challenge_answer_filter_test.json",
45
- "validation": "./challenge_sets/answerable_question_split/challenge_answer_filter_validation.json",
46
- },
47
- }
48
- }
49
 
50
 
51
  class SquadV2Config(datasets.BuilderConfig):
@@ -67,10 +57,10 @@ class SquadV2(datasets.GeneratorBasedBuilder):
67
  VERSION_1 = datasets.Version("1.0.0")
68
 
69
  BUILDER_CONFIGS = [
70
- SquadV2Config(name="challenge_sets", version=VERSION_1, description="SQuAD2.0 challenge set version"),
71
  ]
72
 
73
- DEFAULT_CONFIG_NAME = "challenge_sets"
74
 
75
  def _info(self):
76
  # TODO(squad_v2): Specifies the datasets.DatasetInfo object
@@ -116,74 +106,32 @@ class SquadV2(datasets.GeneratorBasedBuilder):
116
  datasets.SplitGenerator(
117
  name=datasets.Split.TRAIN,
118
  gen_kwargs={
119
- "filepath": downloaded_files["source_data_split"]["train"],
120
  "split": "train",
121
  },
122
  ),
123
  datasets.SplitGenerator(
124
  name=datasets.Split.VALIDATION,
125
  gen_kwargs={
126
- "filepath": downloaded_files["source_data_split"]["validation"],
127
  "split": "validation",
128
  },
129
  ),
130
  datasets.SplitGenerator(
131
  name=datasets.Split.TEST,
132
  gen_kwargs={
133
- "filepath": downloaded_files["source_data_split"]["test"],
134
- "split": "test",
135
- },
136
- ),
137
- datasets.SplitGenerator(
138
- name="question_type_train",
139
- gen_kwargs={
140
- "filepath": downloaded_files["question_type_split"]["train"],
141
- "split": "train",
142
- },
143
- ),
144
- datasets.SplitGenerator(
145
- name="question_type_test",
146
- gen_kwargs={
147
- "filepath": downloaded_files["question_type_split"]["test"],
148
- "split": "test",
149
- },
150
- ),
151
- datasets.SplitGenerator(
152
- name="question_type_validation",
153
- gen_kwargs={
154
- "filepath": downloaded_files["question_type_split"]["validation"],
155
- "split": "validation",
156
- },
157
- ),
158
- datasets.SplitGenerator(
159
- name="answerable_question_train",
160
- gen_kwargs={
161
- "filepath": downloaded_files["answerable_question_split"]["train"],
162
- "split": "train",
163
- },
164
- ),
165
- datasets.SplitGenerator(
166
- name="answerable_question_test",
167
- gen_kwargs={
168
- "filepath": downloaded_files["answerable_question_split"]["test"],
169
  "split": "test",
170
  },
171
  ),
172
- datasets.SplitGenerator(
173
- name="answerable_question_validation",
174
- gen_kwargs={
175
- "filepath": downloaded_files["answerable_question_split"]["validation"],
176
- "split": "validation",
177
- },
178
- ),
179
  ]
180
 
181
  def _generate_examples(self, filepath, split):
182
  """Yields examples."""
183
  # TODO(squad_v2): Yields (key, example) tuples from the dataset
184
  with open(filepath, encoding="utf-8") as f:
185
- for id_, row in enumerate(f):
186
- data = json.loads(row)
187
  # Features currently used are "context", "question", and "answers".
188
  # Others are extracted here for the ease of future expansions.
189
  yield id_, {
 
28
  """
29
 
30
  _URLS = {
31
+ "gem_data_split":
32
+ {
33
+ "train": "./gem_data_split/train.json",
34
+ "test": "./gem_data_split/test.json",
35
+ "validation": "./gem_data_split/validation.json",
36
+ },
37
+ }
38
+
 
 
 
 
 
 
 
 
 
 
39
 
40
 
41
  class SquadV2Config(datasets.BuilderConfig):
 
57
  VERSION_1 = datasets.Version("1.0.0")
58
 
59
  BUILDER_CONFIGS = [
60
+ SquadV2Config(name="gem_data_split", version=VERSION_1, description="SQuAD2.0 - GEM version 1"),
61
  ]
62
 
63
+ DEFAULT_CONFIG_NAME = "gem_data_split"
64
 
65
  def _info(self):
66
  # TODO(squad_v2): Specifies the datasets.DatasetInfo object
 
106
  datasets.SplitGenerator(
107
  name=datasets.Split.TRAIN,
108
  gen_kwargs={
109
+ "filepath": downloaded_files["gem_data_split"]["train"],
110
  "split": "train",
111
  },
112
  ),
113
  datasets.SplitGenerator(
114
  name=datasets.Split.VALIDATION,
115
  gen_kwargs={
116
+ "filepath": downloaded_files["gem_data_split"]["validation"],
117
  "split": "validation",
118
  },
119
  ),
120
  datasets.SplitGenerator(
121
  name=datasets.Split.TEST,
122
  gen_kwargs={
123
+ "filepath": downloaded_files["gem_data_split"]["test"],
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
124
  "split": "test",
125
  },
126
  ),
 
 
 
 
 
 
 
127
  ]
128
 
129
  def _generate_examples(self, filepath, split):
130
  """Yields examples."""
131
  # TODO(squad_v2): Yields (key, example) tuples from the dataset
132
  with open(filepath, encoding="utf-8") as f:
133
+ data = json.load(f)
134
+ for id_, row in data:
135
  # Features currently used are "context", "question", and "answers".
136
  # Others are extracted here for the ease of future expansions.
137
  yield id_, {