Datasets:
Commit
·
9adcec4
1
Parent(s):
622f3dd
rename script to match dataset name
Browse files
TyDiP.py
ADDED
@@ -0,0 +1,118 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
# coding=utf-8
|
2 |
+
# Copyright 2020 The TensorFlow Datasets Authors and the HuggingFace Datasets Authors.
|
3 |
+
#
|
4 |
+
# Licensed under the Apache License, Version 2.0 (the "License");
|
5 |
+
# you may not use this file except in compliance with the License.
|
6 |
+
# You may obtain a copy of the License at
|
7 |
+
#
|
8 |
+
# http://www.apache.org/licenses/LICENSE-2.0
|
9 |
+
#
|
10 |
+
# Unless required by applicable law or agreed to in writing, software
|
11 |
+
# distributed under the License is distributed on an "AS IS" BASIS,
|
12 |
+
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
13 |
+
# See the License for the specific language governing permissions and
|
14 |
+
# limitations under the License.
|
15 |
+
|
16 |
+
# Lint as: python3
|
17 |
+
"""TyDiP: A Multilingual Politeness Dataset"""
|
18 |
+
|
19 |
+
|
20 |
+
import csv
|
21 |
+
from dataclasses import dataclass
|
22 |
+
import datasets
|
23 |
+
from datasets.tasks import TextClassification
|
24 |
+
|
25 |
+
|
26 |
+
_CITATION = """\
|
27 |
+
@inproceedings{srinivasan-choi-2022-tydip,
|
28 |
+
title = "{T}y{D}i{P}: A Dataset for Politeness Classification in Nine Typologically Diverse Languages",
|
29 |
+
author = "Srinivasan, Anirudh and
|
30 |
+
Choi, Eunsol",
|
31 |
+
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2022",
|
32 |
+
month = dec,
|
33 |
+
year = "2022",
|
34 |
+
address = "Abu Dhabi, United Arab Emirates",
|
35 |
+
publisher = "Association for Computational Linguistics",
|
36 |
+
url = "https://aclanthology.org/2022.findings-emnlp.420",
|
37 |
+
pages = "5723--5738",
|
38 |
+
}"""
|
39 |
+
|
40 |
+
_DESCRIPTION = """\
|
41 |
+
The TyDiP dataset is a dataset of requests in conversations between wikipedia editors
|
42 |
+
that have been annotated for politeness. The splits available below consists of only
|
43 |
+
requests from the top 25 percentile (polite) and bottom 25 percentile (impolite) of
|
44 |
+
politeness scores. The English train set and English test set that are
|
45 |
+
adapted from the Stanford Politeness Corpus, and test data in 9 more languages
|
46 |
+
(Hindi, Korean, Spanish, Tamil, French, Vietnamese, Russian, Afrikaans, Hungarian)
|
47 |
+
was annotated by us.
|
48 |
+
"""
|
49 |
+
|
50 |
+
_LANGUAGES = ("en", "hi", "ko", "es", "ta", "fr", "vi", "ru", "af", "hu")
|
51 |
+
|
52 |
+
|
53 |
+
# The HuggingFace Datasets library doesn't host the datasets but only points to the original files.
|
54 |
+
# This can be an arbitrary nested dict/list of URLs (see below in `_split_generators` method)
|
55 |
+
# _URL = "https://huggingface.co/datasets/Genius1237/TyDiP/resolve/main/data/binary/"
|
56 |
+
_URL = "https://huggingface.co/datasets/Genius1237/TyDiP/raw/main/data/binary/"
|
57 |
+
_URLS = {
|
58 |
+
'en': {
|
59 |
+
'train': _URL + 'en_train_binary.csv',
|
60 |
+
'test': _URL + 'en_test_binary.csv'
|
61 |
+
},
|
62 |
+
} | {lang: {'test': _URL + '{}_test_binary.csv'.format(lang)} for lang in _LANGUAGES[1:]}
|
63 |
+
|
64 |
+
|
65 |
+
@dataclass
|
66 |
+
class TyDiPConfig(datasets.BuilderConfig):
|
67 |
+
"""BuilderConfig for TyDiP."""
|
68 |
+
lang: str = None
|
69 |
+
|
70 |
+
|
71 |
+
class MultilingualLibrispeech(datasets.GeneratorBasedBuilder):
|
72 |
+
"""TyDiP dataset."""
|
73 |
+
|
74 |
+
BUILDER_CONFIGS = [
|
75 |
+
TyDiPConfig(name=lang, lang=lang) for lang in _LANGUAGES
|
76 |
+
]
|
77 |
+
|
78 |
+
def _info(self):
|
79 |
+
return datasets.DatasetInfo(
|
80 |
+
description=_DESCRIPTION,
|
81 |
+
features=datasets.Features(
|
82 |
+
{
|
83 |
+
"text": datasets.Value("string"),
|
84 |
+
"labels": datasets.ClassLabel(num_classes=2, names=[0, 1]),
|
85 |
+
}
|
86 |
+
),
|
87 |
+
supervised_keys=("text", "labels"),
|
88 |
+
homepage=_URL,
|
89 |
+
citation=_CITATION,
|
90 |
+
task_templates=[TextClassification(text_column="text", label_column="labels")],
|
91 |
+
)
|
92 |
+
|
93 |
+
def _split_generators(self, dl_manager):
|
94 |
+
splits = []
|
95 |
+
if self.config.lang == 'en':
|
96 |
+
file_path = dl_manager.download_and_extract(_URLS['en']['train'])
|
97 |
+
splits.append(
|
98 |
+
datasets.SplitGenerator(
|
99 |
+
name=datasets.Split.TRAIN, gen_kwargs={"data_file": file_path}
|
100 |
+
))
|
101 |
+
file_path = dl_manager.download_and_extract(_URLS[self.config.lang]['test'])
|
102 |
+
splits.append(
|
103 |
+
datasets.SplitGenerator(
|
104 |
+
name=datasets.Split.TEST, gen_kwargs={"data_file": file_path}
|
105 |
+
)
|
106 |
+
)
|
107 |
+
return splits
|
108 |
+
|
109 |
+
def _generate_examples(self, data_file):
|
110 |
+
"""Generate examples from a TyDiP data file"""
|
111 |
+
with open(data_file) as f:
|
112 |
+
csv_reader = csv.reader(f)
|
113 |
+
for i, row in enumerate(csv_reader):
|
114 |
+
if i != 0:
|
115 |
+
yield i - 1, {
|
116 |
+
"text": row[0],
|
117 |
+
"labels": int(float(row[1]) > 0),
|
118 |
+
}
|