system HF staff commited on
Commit
58040f3
1 Parent(s): d1bc6f8

Update files from the datasets library (from 1.16.0)

Browse files

Release notes: https://github.com/huggingface/datasets/releases/tag/1.16.0

Files changed (2) hide show
  1. README.md +1 -0
  2. mlqa.py +19 -20
README.md CHANGED
@@ -1,4 +1,5 @@
1
  ---
 
2
  languages:
3
  - en
4
  paperswithcode_id: mlqa
 
1
  ---
2
+ pretty_name: MLQA (MultiLingual Question Answering)
3
  languages:
4
  - en
5
  paperswithcode_id: mlqa
mlqa.py CHANGED
@@ -117,27 +117,23 @@ class Mlqa(datasets.GeneratorBasedBuilder):
117
  # dl_manager is a datasets.download.DownloadManager that can be used to
118
  # download and extract URLs
119
  if self.config.name.startswith("mlqa-translate-train"):
120
- dl_file = dl_manager.download_and_extract(self.config.data_url)
121
  lang = self.config.name.split(".")[-1]
122
  return [
123
  datasets.SplitGenerator(
124
  name=datasets.Split.TRAIN,
125
  # These kwargs will be passed to _generate_examples
126
  gen_kwargs={
127
- "filepath": os.path.join(
128
- os.path.join(dl_file, "mlqa-translate-train"),
129
- "{}_squad-translate-train-train-v1.1.json".format(lang),
130
- )
131
  },
132
  ),
133
  datasets.SplitGenerator(
134
  name=datasets.Split.VALIDATION,
135
  # These kwargs will be passed to _generate_examples
136
  gen_kwargs={
137
- "filepath": os.path.join(
138
- os.path.join(dl_file, "mlqa-translate-train"),
139
- "{}_squad-translate-train-dev-v1.1.json".format(lang),
140
- )
141
  },
142
  ),
143
  ]
@@ -154,7 +150,7 @@ class Mlqa(datasets.GeneratorBasedBuilder):
154
  gen_kwargs={
155
  "filepath": os.path.join(
156
  os.path.join(dl_file, "MLQA_V1/test"),
157
- "test-context-{}-question-{}.json".format(l1, l2),
158
  )
159
  },
160
  ),
@@ -163,33 +159,36 @@ class Mlqa(datasets.GeneratorBasedBuilder):
163
  # These kwargs will be passed to _generate_examples
164
  gen_kwargs={
165
  "filepath": os.path.join(
166
- os.path.join(dl_file, "MLQA_V1/dev"), "dev-context-{}-question-{}.json".format(l1, l2)
167
  )
168
  },
169
  ),
170
  ]
171
  else:
172
  if self.config.name.startswith("mlqa-translate-test"):
173
- dl_file = dl_manager.download_and_extract(self.config.data_url)
174
  lang = self.config.name.split(".")[-1]
175
  return [
176
  datasets.SplitGenerator(
177
  name=datasets.Split.TEST,
178
  # These kwargs will be passed to _generate_examples
179
  gen_kwargs={
180
- "filepath": os.path.join(
181
- os.path.join(dl_file, "mlqa-translate-test"),
182
- "translate-test-context-{}-question-{}.json".format(lang, lang),
183
- )
184
  },
185
  ),
186
  ]
187
 
188
- def _generate_examples(self, filepath):
189
  """Yields examples."""
190
- # TODO(mlqa): Yields (key, example) tuples from the dataset
191
- with open(filepath, encoding="utf-8") as f:
192
- data = json.load(f)
 
 
 
 
 
193
  for examples in data["data"]:
194
  for example in examples["paragraphs"]:
195
  context = example["context"]
 
117
  # dl_manager is a datasets.download.DownloadManager that can be used to
118
  # download and extract URLs
119
  if self.config.name.startswith("mlqa-translate-train"):
120
+ archive = dl_manager.download(self.config.data_url)
121
  lang = self.config.name.split(".")[-1]
122
  return [
123
  datasets.SplitGenerator(
124
  name=datasets.Split.TRAIN,
125
  # These kwargs will be passed to _generate_examples
126
  gen_kwargs={
127
+ "filepath": f"mlqa-translate-train/{lang}_squad-translate-train-train-v1.1.json",
128
+ "files": dl_manager.iter_archive(archive),
 
 
129
  },
130
  ),
131
  datasets.SplitGenerator(
132
  name=datasets.Split.VALIDATION,
133
  # These kwargs will be passed to _generate_examples
134
  gen_kwargs={
135
+ "filepath": f"mlqa-translate-train/{lang}_squad-translate-train-dev-v1.1.json",
136
+ "files": dl_manager.iter_archive(archive),
 
 
137
  },
138
  ),
139
  ]
 
150
  gen_kwargs={
151
  "filepath": os.path.join(
152
  os.path.join(dl_file, "MLQA_V1/test"),
153
+ f"test-context-{l1}-question-{l2}.json",
154
  )
155
  },
156
  ),
 
159
  # These kwargs will be passed to _generate_examples
160
  gen_kwargs={
161
  "filepath": os.path.join(
162
+ os.path.join(dl_file, "MLQA_V1/dev"), f"dev-context-{l1}-question-{l2}.json"
163
  )
164
  },
165
  ),
166
  ]
167
  else:
168
  if self.config.name.startswith("mlqa-translate-test"):
169
+ archive = dl_manager.download(self.config.data_url)
170
  lang = self.config.name.split(".")[-1]
171
  return [
172
  datasets.SplitGenerator(
173
  name=datasets.Split.TEST,
174
  # These kwargs will be passed to _generate_examples
175
  gen_kwargs={
176
+ "filepath": f"mlqa-translate-test/translate-test-context-{lang}-question-{lang}.json",
177
+ "files": dl_manager.iter_archive(archive),
 
 
178
  },
179
  ),
180
  ]
181
 
182
+ def _generate_examples(self, filepath, files=None):
183
  """Yields examples."""
184
+ if self.config.name.startswith("mlqa-translate"):
185
+ for path, f in files:
186
+ if path == filepath:
187
+ data = json.loads(f.read().decode("utf-8"))
188
+ break
189
+ else:
190
+ with open(filepath, encoding="utf-8") as f:
191
+ data = json.load(f)
192
  for examples in data["data"]:
193
  for example in examples["paragraphs"]:
194
  context = example["context"]