DicoTiar
commited on
Commit
•
9b5b745
1
Parent(s):
7581327
loading script default
Browse files
story.py
ADDED
@@ -0,0 +1,131 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import csv
|
2 |
+
import os
|
3 |
+
|
4 |
+
import datasets
|
5 |
+
|
6 |
+
_CITATION = """\
|
7 |
+
@Dataset{wisdomify:storyteller,
|
8 |
+
title = {Korean proverb definitions and examples},
|
9 |
+
author={Jongyoon Kim, Yubin Kim, Yongtaek Im
|
10 |
+
},
|
11 |
+
year={2021}
|
12 |
+
}
|
13 |
+
"""
|
14 |
+
|
15 |
+
_DESCRIPTION = """\
|
16 |
+
This dataset is designed to provide forward and reverse dictionary of Korean proverbs.
|
17 |
+
"""
|
18 |
+
|
19 |
+
# TODO: Add a link to an official homepage for the dataset here
|
20 |
+
_HOMEPAGE = ""
|
21 |
+
|
22 |
+
# TODO: Add the licence for the dataset here if you can find it
|
23 |
+
_LICENSE = ""
|
24 |
+
|
25 |
+
# TODO: Add link to the official dataset URLs here
|
26 |
+
# If it is dropbox link, you must set 1 for query parameter "dl".
|
27 |
+
_URLs = {
|
28 |
+
'definition': "https://www.dropbox.com/s/4uh564afaimtob3/definition.zip?dl=1",
|
29 |
+
'example': "https://www.dropbox.com/s/adlt9n6x5gjs0a6/example.zip?dl=1",
|
30 |
+
}
|
31 |
+
|
32 |
+
|
33 |
+
class Story(datasets.GeneratorBasedBuilder):
|
34 |
+
# version must be "x.y.z' form
|
35 |
+
VERSION = datasets.Version("0.0.0")
|
36 |
+
|
37 |
+
BUILDER_CONFIGS = [
|
38 |
+
datasets.BuilderConfig(name="definition", version=VERSION, description="definition"),
|
39 |
+
datasets.BuilderConfig(name="example", version=VERSION, description="example"),
|
40 |
+
]
|
41 |
+
|
42 |
+
# This config is applied when user load dataset without "name".
|
43 |
+
DEFAULT_CONFIG_NAME = "definition"
|
44 |
+
|
45 |
+
def _info(self):
|
46 |
+
# This method specifies the datasets.DatasetInfo object which contains information
|
47 |
+
# and typings for the dataset
|
48 |
+
|
49 |
+
if self.config.name == "definition":
|
50 |
+
# These are the features of your dataset like images, labels ...
|
51 |
+
features = datasets.Features(
|
52 |
+
{
|
53 |
+
"wisdom": datasets.Value("string"),
|
54 |
+
"def": datasets.Value("string"),
|
55 |
+
|
56 |
+
}
|
57 |
+
)
|
58 |
+
elif self.config.name == "example":
|
59 |
+
features = datasets.Features(
|
60 |
+
{
|
61 |
+
"wisdom": datasets.Value("string"),
|
62 |
+
"eg": datasets.Value("string"),
|
63 |
+
}
|
64 |
+
)
|
65 |
+
else:
|
66 |
+
raise NotImplementedError(f"Wrong name: {self.config.name}")
|
67 |
+
|
68 |
+
return datasets.DatasetInfo(
|
69 |
+
description=_DESCRIPTION,
|
70 |
+
features=features,
|
71 |
+
supervised_keys=None,
|
72 |
+
homepage=_HOMEPAGE,
|
73 |
+
license=_LICENSE,
|
74 |
+
citation=_CITATION,
|
75 |
+
)
|
76 |
+
|
77 |
+
def _split_generators(self, dl_manager):
|
78 |
+
"""Returns SplitGenerators."""
|
79 |
+
# This method is used when user loads dataset.
|
80 |
+
# dl_manager can be used to download and extract the dataset
|
81 |
+
# and also can set split depending onf the configuration
|
82 |
+
|
83 |
+
# Downloading data with _URLs
|
84 |
+
downloaded_files = dl_manager.download_and_extract(_URLs[self.config.name])
|
85 |
+
|
86 |
+
dtp = 'def' if self.config.name == "definition" else 'eg'
|
87 |
+
|
88 |
+
train_path = os.path.join(downloaded_files, f'train_wisdom2{dtp}.tsv')
|
89 |
+
val_path = os.path.join(downloaded_files, f'val_wisdom2{dtp}.tsv')
|
90 |
+
test_path = os.path.join(downloaded_files, f'test_wisdom2{dtp}.tsv')
|
91 |
+
|
92 |
+
return [
|
93 |
+
# These gen_kwargs will be passed to _generate_examples
|
94 |
+
datasets.SplitGenerator(
|
95 |
+
name=datasets.Split.TRAIN,
|
96 |
+
gen_kwargs={"filepath": train_path, "split": "train"},
|
97 |
+
),
|
98 |
+
datasets.SplitGenerator(
|
99 |
+
name=datasets.Split.VALIDATION,
|
100 |
+
gen_kwargs={"filepath": val_path, "split": "validation"},
|
101 |
+
),
|
102 |
+
datasets.SplitGenerator(
|
103 |
+
name=datasets.Split.TEST,
|
104 |
+
gen_kwargs={"filepath": test_path, "split": "test"},
|
105 |
+
),
|
106 |
+
]
|
107 |
+
|
108 |
+
def _generate_examples(self, filepath, split):
|
109 |
+
# method parameters are unpacked from `gen_kwargs` as given in `_split_generators`
|
110 |
+
""" Yields examples as (key, example) tuples. """
|
111 |
+
# This method handles input defined in _split_generators to yield (key, example) tuples from the dataset.
|
112 |
+
# The `key` is here for legacy reason (tfds) and is not important in itself.
|
113 |
+
|
114 |
+
with open(filepath, encoding="utf-8") as f:
|
115 |
+
tsv_reader = csv.reader(f, delimiter="\t")
|
116 |
+
for id_, row in enumerate(tsv_reader):
|
117 |
+
if id_ == 0:
|
118 |
+
continue # first row shows column info
|
119 |
+
|
120 |
+
if self.config.name == "definition":
|
121 |
+
yield id_, {
|
122 |
+
"wisdom": row[0],
|
123 |
+
"def": row[1],
|
124 |
+
}
|
125 |
+
elif self.config.name == "example":
|
126 |
+
yield id_, {
|
127 |
+
"wisdom": row[0],
|
128 |
+
"eg": row[1],
|
129 |
+
}
|
130 |
+
else:
|
131 |
+
raise NotImplementedError(f"Wrong name: {self.config.name}")
|