Datasets:

ohashi56225 commited on
Commit
5aeb3eb
1 Parent(s): c0cc6a5

Update jmultiwoz.py

Browse files
Files changed (1) hide show
  1. jmultiwoz.py +40 -43
jmultiwoz.py CHANGED
@@ -38,7 +38,8 @@ _CITATION = """\
38
  _DESCRIPTION = """\
39
  JMultiWOZ is a large-scale Japanese multi-domain task-oriented dialogue dataset. The dataset is collected using
40
  the Wizard-of-Oz (WoZ) methodology, where two human annotators simulate the user and the system. The dataset contains
41
- 4,264 dialogues across 6 domains, including restaurant, hotel, attraction, shopping, taxi, and weather.
 
42
  """
43
 
44
  # TODO: Add a link to an official homepage for the dataset here
@@ -51,11 +52,12 @@ _LICENSE = "CC BY-ND 4.0"
51
  # The HuggingFace Datasets library doesn't host the datasets but only points to the original files.
52
  # This can be an arbitrary nested dict/list of URLs (see below in `_split_generators` method)
53
  _URLS = {
54
- "original_zip": "https://github.com/ohashi56225/jmultiwoz-evaluation/raw/master/dataset/JMultiWOZ_1.0.zip",
 
55
  }
56
 
57
 
58
- def _flatten_value(values):
59
  if not isinstance(values, list):
60
  return values
61
  flat_values = [
@@ -99,13 +101,13 @@ class JMultiWOZDataset(datasets.GeneratorBasedBuilder):
99
  "slot": datasets.Value("string"),
100
  "value": datasets.Value("string"),
101
  }),
102
- "db_result": datasets.Sequence({
103
  "candidate_entities": datasets.Sequence(datasets.Value("string")),
104
  "active_entity": datasets.Sequence({
105
  "slot": datasets.Value("string"),
106
  "value": datasets.Value("string"),
107
  })
108
- }),
109
  "book_result": datasets.Sequence({
110
  "domain": datasets.Value("string"),
111
  "success": datasets.Value("string"),
@@ -152,7 +154,6 @@ class JMultiWOZDataset(datasets.GeneratorBasedBuilder):
152
  # These kwargs will be passed to _generate_examples
153
  gen_kwargs={
154
  "dialogues": [dialogues[dialogue_name] for dialogue_name in split_list["train"]],
155
- "split": "train",
156
  },
157
  ),
158
  datasets.SplitGenerator(
@@ -160,7 +161,6 @@ class JMultiWOZDataset(datasets.GeneratorBasedBuilder):
160
  # These kwargs will be passed to _generate_examples
161
  gen_kwargs={
162
  "dialogues": [dialogues[dialogue_name] for dialogue_name in split_list["dev"]],
163
- "split": "dev",
164
  },
165
  ),
166
  datasets.SplitGenerator(
@@ -168,13 +168,12 @@ class JMultiWOZDataset(datasets.GeneratorBasedBuilder):
168
  # These kwargs will be passed to _generate_examples
169
  gen_kwargs={
170
  "dialogues": [dialogues[dialogue_name] for dialogue_name in split_list["test"]],
171
- "split": "test"
172
  },
173
  ),
174
  ]
175
 
176
  # method parameters are unpacked from `gen_kwargs` as given in `_split_generators`
177
- def _generate_examples(self, dialogues, split):
178
  # TODO: This method handles input defined in _split_generators to yield (key, example) tuples from the dataset.
179
  # The `key` is for legacy reasons (tfds) and is not important in itself, but must be unique for each example.
180
 
@@ -216,48 +215,46 @@ class JMultiWOZDataset(datasets.GeneratorBasedBuilder):
216
  "dialogue_state": {
217
  "belief_state": [],
218
  "book_state": [],
219
- "db_result": [],
220
  "book_result": [],
221
  },
222
  }
223
- if turn["speaker"] == "USER":
224
- continue
 
 
 
 
 
 
225
 
226
- for domain, slots in turn["dialogue_state"]["belief_state"].items():
227
- for slot, value in slots.items():
228
- example_turn["dialogue_state"]["belief_state"].append({
229
- "domain": domain,
 
 
 
 
 
 
 
 
 
 
 
230
  "slot": slot,
231
- "value": value,
232
- })
233
-
234
- for domain, slots in turn["dialogue_state"]["book_state"].items():
235
- for slot, value in slots.items():
236
- example_turn["dialogue_state"]["book_state"].append({
237
  "domain": domain,
238
- "slot": slot,
239
- "value": value,
240
  })
241
 
242
- candidate_entities = turn["dialogue_state"]["db_result"]["candidate_entities"]
243
- active_entity = turn["dialogue_state"]["db_result"]["active_entity"]
244
- if not active_entity:
245
- active_entity = {}
246
- example_turn["dialogue_state"]["db_result"].append({
247
- "candidate_entities":candidate_entities,
248
- "active_entity": [{
249
- "slot": slot,
250
- "value": _flatten_value(value),
251
- } for slot, value in active_entity.items()]
252
- })
253
-
254
- for domain, result in turn["dialogue_state"]["book_result"].items():
255
- example_turn["dialogue_state"]["book_result"].append({
256
- "domain": domain,
257
- "success": result["success"],
258
- "ref": result["ref"],
259
- })
260
-
261
  example["turns"].append(example_turn)
262
 
263
  yield id_, example
 
38
  _DESCRIPTION = """\
39
  JMultiWOZ is a large-scale Japanese multi-domain task-oriented dialogue dataset. The dataset is collected using
40
  the Wizard-of-Oz (WoZ) methodology, where two human annotators simulate the user and the system. The dataset contains
41
+ 4,246 dialogues across 6 domains, including restaurant, hotel, attraction, shopping, taxi, and weather. Available
42
+ annotations include user goal, dialogue state, and utterances.
43
  """
44
 
45
  # TODO: Add a link to an official homepage for the dataset here
 
52
  # The HuggingFace Datasets library doesn't host the datasets but only points to the original files.
53
  # This can be an arbitrary nested dict/list of URLs (see below in `_split_generators` method)
54
  _URLS = {
55
+ # "original_zip": "https://github.com/ohashi56225/jmultiwoz-evaluation/raw/master/dataset/JMultiWOZ_1.0.zip",
56
+ "original_zip": "JMultiWOZ_1.0.zip"
57
  }
58
 
59
 
60
+ def _flatten_value(values) -> str:
61
  if not isinstance(values, list):
62
  return values
63
  flat_values = [
 
101
  "slot": datasets.Value("string"),
102
  "value": datasets.Value("string"),
103
  }),
104
+ "db_result": {
105
  "candidate_entities": datasets.Sequence(datasets.Value("string")),
106
  "active_entity": datasets.Sequence({
107
  "slot": datasets.Value("string"),
108
  "value": datasets.Value("string"),
109
  })
110
+ },
111
  "book_result": datasets.Sequence({
112
  "domain": datasets.Value("string"),
113
  "success": datasets.Value("string"),
 
154
  # These kwargs will be passed to _generate_examples
155
  gen_kwargs={
156
  "dialogues": [dialogues[dialogue_name] for dialogue_name in split_list["train"]],
 
157
  },
158
  ),
159
  datasets.SplitGenerator(
 
161
  # These kwargs will be passed to _generate_examples
162
  gen_kwargs={
163
  "dialogues": [dialogues[dialogue_name] for dialogue_name in split_list["dev"]],
 
164
  },
165
  ),
166
  datasets.SplitGenerator(
 
168
  # These kwargs will be passed to _generate_examples
169
  gen_kwargs={
170
  "dialogues": [dialogues[dialogue_name] for dialogue_name in split_list["test"]],
 
171
  },
172
  ),
173
  ]
174
 
175
  # method parameters are unpacked from `gen_kwargs` as given in `_split_generators`
176
+ def _generate_examples(self, dialogues):
177
  # TODO: This method handles input defined in _split_generators to yield (key, example) tuples from the dataset.
178
  # The `key` is for legacy reasons (tfds) and is not important in itself, but must be unique for each example.
179
 
 
215
  "dialogue_state": {
216
  "belief_state": [],
217
  "book_state": [],
218
+ "db_result": {},
219
  "book_result": [],
220
  },
221
  }
222
+ if turn["speaker"] == "SYSTEM":
223
+ for domain, slots in turn["dialogue_state"]["belief_state"].items():
224
+ for slot, value in slots.items():
225
+ example_turn["dialogue_state"]["belief_state"].append({
226
+ "domain": domain,
227
+ "slot": slot,
228
+ "value": value,
229
+ })
230
 
231
+ for domain, slots in turn["dialogue_state"]["book_state"].items():
232
+ for slot, value in slots.items():
233
+ example_turn["dialogue_state"]["book_state"].append({
234
+ "domain": domain,
235
+ "slot": slot,
236
+ "value": value,
237
+ })
238
+
239
+ candidate_entities = turn["dialogue_state"]["db_result"]["candidate_entities"]
240
+ active_entity = turn["dialogue_state"]["db_result"]["active_entity"]
241
+ if not active_entity:
242
+ active_entity = {}
243
+ example_turn["dialogue_state"]["db_result"] = {
244
+ "candidate_entities":candidate_entities,
245
+ "active_entity": [{
246
  "slot": slot,
247
+ "value": _flatten_value(value),
248
+ } for slot, value in active_entity.items()]
249
+ }
250
+
251
+ for domain, result in turn["dialogue_state"]["book_result"].items():
252
+ example_turn["dialogue_state"]["book_result"].append({
253
  "domain": domain,
254
+ "success": result["success"],
255
+ "ref": result["ref"],
256
  })
257
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
258
  example["turns"].append(example_turn)
259
 
260
  yield id_, example