Datasets:
ShutongFeng
commited on
Commit
•
9b592da
1
Parent(s):
0315ca1
Update emowoz.py
Browse filesImprove usage: load EmoWOZ/MultiWOZ/DialMage at one go
emowoz.py
CHANGED
@@ -91,15 +91,9 @@ class EmoWOZ(datasets.GeneratorBasedBuilder):
|
|
91 |
# data = datasets.load_dataset('my_dataset', 'first_domain')
|
92 |
# data = datasets.load_dataset('my_dataset', 'second_domain')
|
93 |
BUILDER_CONFIGS = [
|
94 |
-
datasets.BuilderConfig(name="
|
95 |
-
datasets.BuilderConfig(name="
|
96 |
-
datasets.BuilderConfig(name="
|
97 |
-
datasets.BuilderConfig(name="train-multiwoz", version=VERSION, description="This part contains the training set of user-emotion-annotated dialogues from MultiWOZ"),
|
98 |
-
datasets.BuilderConfig(name="dev-multiwoz", version=VERSION, description="This part contains the development set of user-emotion-annotated dialogues from MultiWOZ"),
|
99 |
-
datasets.BuilderConfig(name="test-multiwoz", version=VERSION, description="This part contains the test set of user-emotion-annotated dialogues from MultiWOZ"),
|
100 |
-
datasets.BuilderConfig(name="train-dialmage", version=VERSION, description="This part contains the training set of user-emotion-annotated dialogues from human-machine interaction (DialMAGE)"),
|
101 |
-
datasets.BuilderConfig(name="dev-dialmage", version=VERSION, description="This part contains the development set of user-emotion-annotated dialogues from human-machine interaction (DialMAGE)"),
|
102 |
-
datasets.BuilderConfig(name="test-dialmage", version=VERSION, description="This part contains the test set of user-emotion-annotated dialogues from human-machine interaction (DialMAGE)"),
|
103 |
]
|
104 |
|
105 |
# DEFAULT_CONFIG_NAME = "first_domain" # It's not mandatory to have a default configuration. Just use one if it make sense.
|
@@ -182,13 +176,12 @@ class EmoWOZ(datasets.GeneratorBasedBuilder):
|
|
182 |
with open(split_filepath, encoding="utf-8") as f:
|
183 |
data_split = json.load(f)
|
184 |
|
185 |
-
|
186 |
-
if subset == 'all':
|
187 |
dial_ids = data_split[split]['multiwoz'] + data_split[split]['dialmage']
|
188 |
else:
|
189 |
-
dial_ids = data_split[split][
|
190 |
|
191 |
-
# resolve the duplicate key in the training set of emowoz/data-split.json
|
192 |
dial_ids = list(set(dial_ids))
|
193 |
|
194 |
for key in dial_ids:
|
|
|
91 |
# data = datasets.load_dataset('my_dataset', 'first_domain')
|
92 |
# data = datasets.load_dataset('my_dataset', 'second_domain')
|
93 |
BUILDER_CONFIGS = [
|
94 |
+
datasets.BuilderConfig(name="all", version=VERSION, description="This part contains all user-emotion-annotated dialogues from EmoWOZ"),
|
95 |
+
datasets.BuilderConfig(name="multiwoz", version=VERSION, description="This part contains all user-emotion-annotated dialogues from MultiWOZ"),
|
96 |
+
datasets.BuilderConfig(name="dialmage", version=VERSION, description="This part contains all user-emotion-annotated dialogues from human-machine interactions (DialMAGE)"),
|
|
|
|
|
|
|
|
|
|
|
|
|
97 |
]
|
98 |
|
99 |
# DEFAULT_CONFIG_NAME = "first_domain" # It's not mandatory to have a default configuration. Just use one if it make sense.
|
|
|
176 |
with open(split_filepath, encoding="utf-8") as f:
|
177 |
data_split = json.load(f)
|
178 |
|
179 |
+
if self.config.name == 'all':
|
|
|
180 |
dial_ids = data_split[split]['multiwoz'] + data_split[split]['dialmage']
|
181 |
else:
|
182 |
+
dial_ids = data_split[split][self.config.name]
|
183 |
|
184 |
+
# resolve the one duplicate key in the training set of emowoz/data-split.json
|
185 |
dial_ids = list(set(dial_ids))
|
186 |
|
187 |
for key in dial_ids:
|