Create GenMix50k.py
Browse files- GenMix50k.py +88 -0
GenMix50k.py
ADDED
@@ -0,0 +1,88 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import os
|
2 |
+
import json
|
3 |
+
import datasets as ds
|
4 |
+
|
5 |
+
_DESCRIPTION = """
|
6 |
+
Description of the dataset.
|
7 |
+
"""
|
8 |
+
|
9 |
+
_CITATION = """
|
10 |
+
Citation for the dataset.
|
11 |
+
"""
|
12 |
+
|
13 |
+
_LICENCE = """
|
14 |
+
License information for the dataset.
|
15 |
+
"""
|
16 |
+
|
17 |
+
_FEATURES = ds.Features({
|
18 |
+
"x": ds.Value(dtype="string"),
|
19 |
+
"y": ds.Value(dtype="string")
|
20 |
+
})
|
21 |
+
|
22 |
+
class GenMix50kConfig(ds.BuilderConfig):
|
23 |
+
def __init__(self, **kwargs):
|
24 |
+
super(GenMix50kConfig, self).__init__(**kwargs)
|
25 |
+
self.version = ds.Version("1.0.3")
|
26 |
+
self.features = _FEATURES
|
27 |
+
self.citation = _CITATION
|
28 |
+
|
29 |
+
class GenMix50k(ds.GeneratorBasedBuilder):
|
30 |
+
BUILDER_CONFIGS = [
|
31 |
+
GenMix50kConfig(name="translation"),
|
32 |
+
GenMix50kConfig(name="dialogue"),
|
33 |
+
GenMix50kConfig(name="summarization")
|
34 |
+
]
|
35 |
+
|
36 |
+
def _info(self) -> ds.DatasetInfo:
|
37 |
+
"""Returns the dataset metadata."""
|
38 |
+
return ds.DatasetInfo(
|
39 |
+
description=_DESCRIPTION,
|
40 |
+
features=self.config.features,
|
41 |
+
citation=self.config.citation,
|
42 |
+
license=_LICENCE,
|
43 |
+
supervised_keys=None
|
44 |
+
)
|
45 |
+
|
46 |
+
def _split_generators(self, dl_manager: ds.DownloadManager):
|
47 |
+
"""Returns SplitGenerators"""
|
48 |
+
base_url = "https://huggingface.co/datasets/moon23k/GenMix50k/resolve/main/"
|
49 |
+
data_files = {
|
50 |
+
"translation": {
|
51 |
+
"train": base_url + "translation/train.json",
|
52 |
+
"validation": base_url + "translation/valid.json",
|
53 |
+
"test": base_url + "translation/test.json"
|
54 |
+
},
|
55 |
+
"dialogue": {
|
56 |
+
"train": base_url + "dialogue/train.json",
|
57 |
+
"validation": base_url + "dialogue/valid.json",
|
58 |
+
"test": base_url + "dialogue/test.json"
|
59 |
+
},
|
60 |
+
"summarization": {
|
61 |
+
"train": base_url + "summarization/train.json",
|
62 |
+
"validation": base_url + "summarization/valid.json",
|
63 |
+
"test": base_url + "summarization/test.json"
|
64 |
+
}
|
65 |
+
}
|
66 |
+
|
67 |
+
downloaded_files = dl_manager.download_and_extract(data_files[self.config.name])
|
68 |
+
|
69 |
+
return [
|
70 |
+
ds.SplitGenerator(
|
71 |
+
name=ds.Split.TRAIN,
|
72 |
+
gen_kwargs={"filepath": downloaded_files["train"]},
|
73 |
+
),
|
74 |
+
ds.SplitGenerator(
|
75 |
+
name=ds.Split.VALIDATION,
|
76 |
+
gen_kwargs={"filepath": downloaded_files["validation"]},
|
77 |
+
),
|
78 |
+
ds.SplitGenerator(
|
79 |
+
name=ds.Split.TEST,
|
80 |
+
gen_kwargs={"filepath": downloaded_files["test"]},
|
81 |
+
),
|
82 |
+
]
|
83 |
+
|
84 |
+
def _generate_examples(self, filepath):
|
85 |
+
with open(filepath, encoding="utf-8") as f:
|
86 |
+
for idx, line in enumerate(f):
|
87 |
+
example = json.loads(line)
|
88 |
+
yield idx, example
|