🪿 black
Browse files- many_emotions.py +46 -34
many_emotions.py
CHANGED
@@ -51,7 +51,6 @@ _CLASS_NAMES = [
|
|
51 |
|
52 |
|
53 |
class EmotionsDatasetConfig(datasets.BuilderConfig):
|
54 |
-
|
55 |
def __init__(self, features, label_classes, **kwargs):
|
56 |
super().__init__(**kwargs)
|
57 |
self.features = features
|
@@ -63,46 +62,59 @@ class EmotionsDataset(datasets.GeneratorBasedBuilder):
|
|
63 |
EmotionsDatasetConfig(
|
64 |
name="raw",
|
65 |
label_classes=_SUB_CLASSES,
|
66 |
-
features=["text", "label", "dataset", "license"]
|
67 |
),
|
68 |
EmotionsDatasetConfig(
|
69 |
name="split",
|
70 |
label_classes=_SUB_CLASSES,
|
71 |
-
features=["text", "label", "dataset", "license", "language"]
|
72 |
-
)
|
73 |
]
|
74 |
|
75 |
DEFAULT_CONFIG_NAME = "split"
|
76 |
|
77 |
def _info(self):
|
78 |
-
|
79 |
-
|
80 |
-
|
81 |
-
|
82 |
-
|
83 |
-
|
84 |
-
|
85 |
-
|
86 |
-
|
87 |
-
|
88 |
-
)
|
89 |
-
)
|
90 |
|
91 |
-
def _split_generators(
|
|
|
|
|
92 |
splits = []
|
93 |
if self.config.name == "raw":
|
94 |
-
downloaded_files = dl_manager.download_and_extract(
|
|
|
|
|
95 |
for lang in _LANGUAGES:
|
96 |
-
splits.append(
|
97 |
-
|
98 |
-
|
99 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
100 |
else:
|
101 |
for split in ["train", "validation", "test"]:
|
102 |
-
downloaded_files = dl_manager.download_and_extract(
|
103 |
-
|
104 |
-
|
105 |
-
|
|
|
|
|
|
|
|
|
|
|
106 |
return splits
|
107 |
|
108 |
def _generate_examples(self, filepaths, dataset, license=None, language=None):
|
@@ -114,19 +126,19 @@ class EmotionsDataset(datasets.GeneratorBasedBuilder):
|
|
114 |
if language != "all":
|
115 |
example = {
|
116 |
"id": example["id"],
|
117 |
-
|
118 |
-
|
119 |
-
|
120 |
-
|
|
|
|
|
121 |
}
|
122 |
label = _CLASS_NAMES[example["label"]]
|
123 |
if label == "no emotion":
|
124 |
label = "neutral"
|
125 |
elif label == "happiness":
|
126 |
label = "joy"
|
127 |
-
example.update({
|
128 |
-
"label": label
|
129 |
-
})
|
130 |
yield example["id"], example
|
131 |
else:
|
132 |
for i, filepath in enumerate(filepaths):
|
@@ -137,5 +149,5 @@ class EmotionsDataset(datasets.GeneratorBasedBuilder):
|
|
137 |
|
138 |
|
139 |
if __name__ == "__main__":
|
140 |
-
dataset = load_dataset("ma2za/many_emotions", name="
|
141 |
print()
|
|
|
51 |
|
52 |
|
53 |
class EmotionsDatasetConfig(datasets.BuilderConfig):
|
|
|
54 |
def __init__(self, features, label_classes, **kwargs):
|
55 |
super().__init__(**kwargs)
|
56 |
self.features = features
|
|
|
62 |
EmotionsDatasetConfig(
|
63 |
name="raw",
|
64 |
label_classes=_SUB_CLASSES,
|
65 |
+
features=["text", "label", "dataset", "license"],
|
66 |
),
|
67 |
EmotionsDatasetConfig(
|
68 |
name="split",
|
69 |
label_classes=_SUB_CLASSES,
|
70 |
+
features=["text", "label", "dataset", "license", "language"],
|
71 |
+
),
|
72 |
]
|
73 |
|
74 |
DEFAULT_CONFIG_NAME = "split"
|
75 |
|
76 |
def _info(self):
|
77 |
+
features = {
|
78 |
+
"id": datasets.Value("string"),
|
79 |
+
"text": Value(dtype="string", id=None),
|
80 |
+
"label": ClassLabel(names=_SUB_CLASSES, id=None),
|
81 |
+
"dataset": Value(dtype="string", id=None),
|
82 |
+
"license": Value(dtype="string", id=None),
|
83 |
+
}
|
84 |
+
if self.config.name == "split":
|
85 |
+
features.update({"language": ClassLabel(names=_LANGUAGES, id=None)})
|
86 |
+
return datasets.DatasetInfo(features=datasets.Features(features))
|
|
|
|
|
87 |
|
88 |
+
def _split_generators(
|
89 |
+
self, dl_manager: datasets.DownloadManager
|
90 |
+
) -> List[datasets.SplitGenerator]:
|
91 |
splits = []
|
92 |
if self.config.name == "raw":
|
93 |
+
downloaded_files = dl_manager.download_and_extract(
|
94 |
+
["data/many_emotions.json.gz"]
|
95 |
+
)
|
96 |
for lang in _LANGUAGES:
|
97 |
+
splits.append(
|
98 |
+
datasets.SplitGenerator(
|
99 |
+
name=lang,
|
100 |
+
gen_kwargs={
|
101 |
+
"filepaths": downloaded_files,
|
102 |
+
"language": lang,
|
103 |
+
"dataset": "raw",
|
104 |
+
},
|
105 |
+
)
|
106 |
+
)
|
107 |
else:
|
108 |
for split in ["train", "validation", "test"]:
|
109 |
+
downloaded_files = dl_manager.download_and_extract(
|
110 |
+
[f"data/split_dataset_{split}.jsonl.gz"]
|
111 |
+
)
|
112 |
+
splits.append(
|
113 |
+
datasets.SplitGenerator(
|
114 |
+
name=split,
|
115 |
+
gen_kwargs={"filepaths": downloaded_files, "dataset": "split"},
|
116 |
+
)
|
117 |
+
)
|
118 |
return splits
|
119 |
|
120 |
def _generate_examples(self, filepaths, dataset, license=None, language=None):
|
|
|
126 |
if language != "all":
|
127 |
example = {
|
128 |
"id": example["id"],
|
129 |
+
"text": example[
|
130 |
+
"text" if language == "en" else language
|
131 |
+
],
|
132 |
+
"label": example["label"],
|
133 |
+
"dataset": example["dataset"],
|
134 |
+
"license": example["license"],
|
135 |
}
|
136 |
label = _CLASS_NAMES[example["label"]]
|
137 |
if label == "no emotion":
|
138 |
label = "neutral"
|
139 |
elif label == "happiness":
|
140 |
label = "joy"
|
141 |
+
example.update({"label": label})
|
|
|
|
|
142 |
yield example["id"], example
|
143 |
else:
|
144 |
for i, filepath in enumerate(filepaths):
|
|
|
149 |
|
150 |
|
151 |
if __name__ == "__main__":
|
152 |
+
dataset = load_dataset("ma2za/many_emotions", name="raw")
|
153 |
print()
|