Datasets:
Commit
•
26d179e
1
Parent(s):
d8ceda6
Delete loading script
Browse files
un_ga.py
DELETED
@@ -1,154 +0,0 @@
|
|
1 |
-
# coding=utf-8
|
2 |
-
# Copyright 2020 The HuggingFace Datasets Authors and the current dataset script contributor.
|
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 |
-
"""United nations general assembly resolutions: A six-language parallel corpus"""
|
16 |
-
|
17 |
-
|
18 |
-
import os
|
19 |
-
|
20 |
-
import datasets
|
21 |
-
|
22 |
-
|
23 |
-
_CITATION = """\
|
24 |
-
@inproceedings{title = "United Nations General Assembly Resolutions: a six-language parallel corpus",
|
25 |
-
abstract = "In this paper we describe a six-ways parallel public-domain corpus consisting of 2100 United Nations General Assembly Resolutions with translations in the six official languages of the United Nations, with an average of around 3 million tokens per language. The corpus is available in a preprocessed, formatting-normalized TMX format with paragraphs aligned across multiple languages. We describe the background to the corpus and its content, the process of its construction, and some of its interesting properties.",
|
26 |
-
author = "Alexandre Rafalovitch and Robert Dale",
|
27 |
-
year = "2009",
|
28 |
-
language = "English",
|
29 |
-
booktitle = "MT Summit XII proceedings",
|
30 |
-
publisher = "International Association of Machine Translation",
|
31 |
-
}"""
|
32 |
-
|
33 |
-
_HOMEPAGE = "http://opus.nlpl.eu/UN.php"
|
34 |
-
|
35 |
-
|
36 |
-
_LICENSE = ""
|
37 |
-
|
38 |
-
_VALID_LANGUAGE_PAIRS = {
|
39 |
-
("ar", "en"): "http://opus.nlpl.eu/download.php?f=UN/v20090831/moses/ar-en.txt.zip",
|
40 |
-
("ar", "es"): "http://opus.nlpl.eu/download.php?f=UN/v20090831/moses/ar-es.txt.zip",
|
41 |
-
("ar", "fr"): "http://opus.nlpl.eu/download.php?f=UN/v20090831/moses/ar-fr.txt.zip",
|
42 |
-
("ar", "ru"): "http://opus.nlpl.eu/download.php?f=UN/v20090831/moses/ar-ru.txt.zip",
|
43 |
-
("ar", "zh"): "http://opus.nlpl.eu/download.php?f=UN/v20090831/moses/ar-zh.txt.zip",
|
44 |
-
("en", "es"): "http://opus.nlpl.eu/download.php?f=UN/v20090831/moses/en-es.txt.zip",
|
45 |
-
("en", "fr"): "http://opus.nlpl.eu/download.php?f=UN/v20090831/moses/en-fr.txt.zip",
|
46 |
-
("en", "ru"): "http://opus.nlpl.eu/download.php?f=UN/v20090831/moses/en-ru.txt.zip",
|
47 |
-
("en", "zh"): "http://opus.nlpl.eu/download.php?f=UN/v20090831/moses/en-zh.txt.zip",
|
48 |
-
("es", "fr"): "http://opus.nlpl.eu/download.php?f=UN/v20090831/moses/es-fr.txt.zip",
|
49 |
-
("es", "ru"): "http://opus.nlpl.eu/download.php?f=UN/v20090831/moses/es-ru.txt.zip",
|
50 |
-
("es", "zh"): "http://opus.nlpl.eu/download.php?f=UN/v20090831/moses/es-zh.txt.zip",
|
51 |
-
("fr", "ru"): "http://opus.nlpl.eu/download.php?f=UN/v20090831/moses/fr-ru.txt.zip",
|
52 |
-
("fr", "zh"): "http://opus.nlpl.eu/download.php?f=UN/v20090831/moses/fr-zh.txt.zip",
|
53 |
-
("ru", "zh"): "http://opus.nlpl.eu/download.php?f=UN/v20090831/moses/ru-zh.txt.zip",
|
54 |
-
}
|
55 |
-
|
56 |
-
_VERSION = "2.0.0"
|
57 |
-
|
58 |
-
_DESCRIPTION = """\
|
59 |
-
United nations general assembly resolutions: A six-language parallel corpus.
|
60 |
-
This is a collection of translated documents from the United Nations originally compiled into a translation memory by Alexandre Rafalovitch, Robert Dale (see http://uncorpora.org).
|
61 |
-
6 languages, 15 bitexts
|
62 |
-
total number of files: 6
|
63 |
-
total number of tokens: 18.87M
|
64 |
-
total number of sentence fragments: 0.44M
|
65 |
-
"""
|
66 |
-
|
67 |
-
_BASE_NAME = "UN.{}-{}.{}"
|
68 |
-
|
69 |
-
|
70 |
-
class UnGaConfig(datasets.BuilderConfig):
|
71 |
-
"""BuilderConfig for United nations general assembly resolutions: A six-language parallel corpus"""
|
72 |
-
|
73 |
-
def __init__(self, language_pair=(None, None), **kwargs):
|
74 |
-
"""BuilderConfig for United nations general assembly resolutions: A six-language parallel corpus.
|
75 |
-
The first language in `language_pair` should consist of two strings joined by
|
76 |
-
an underscore (e.g. "en-tr").
|
77 |
-
Args:
|
78 |
-
language_pair: pair of languages that will be used for translation.
|
79 |
-
**kwargs: keyword arguments forwarded to super.
|
80 |
-
"""
|
81 |
-
name = "%s_to_%s" % (language_pair[0], language_pair[1])
|
82 |
-
|
83 |
-
description = ("Translation dataset from %s to %s or %s to %s.") % (
|
84 |
-
language_pair[0],
|
85 |
-
language_pair[1],
|
86 |
-
language_pair[1],
|
87 |
-
language_pair[0],
|
88 |
-
)
|
89 |
-
super(UnGaConfig, self).__init__(
|
90 |
-
name=name, description=description, version=datasets.Version(_VERSION, ""), **kwargs
|
91 |
-
)
|
92 |
-
|
93 |
-
# Validate language pair.
|
94 |
-
assert language_pair in _VALID_LANGUAGE_PAIRS, (
|
95 |
-
"Config language pair (%s, " "%s) not supported"
|
96 |
-
) % language_pair
|
97 |
-
|
98 |
-
self.language_pair = language_pair
|
99 |
-
|
100 |
-
|
101 |
-
class UnGa(datasets.GeneratorBasedBuilder):
|
102 |
-
|
103 |
-
BUILDER_CONFIGS = [
|
104 |
-
UnGaConfig(
|
105 |
-
language_pair=pair,
|
106 |
-
)
|
107 |
-
for pair in _VALID_LANGUAGE_PAIRS.keys()
|
108 |
-
]
|
109 |
-
|
110 |
-
BUILDER_CONFIG_CLASS = UnGaConfig
|
111 |
-
|
112 |
-
def _info(self):
|
113 |
-
return datasets.DatasetInfo(
|
114 |
-
description=_DESCRIPTION,
|
115 |
-
features=datasets.Features(
|
116 |
-
{
|
117 |
-
"id": datasets.Value("string"),
|
118 |
-
"translation": datasets.Translation(languages=tuple(self.config.language_pair)),
|
119 |
-
},
|
120 |
-
),
|
121 |
-
supervised_keys=None,
|
122 |
-
homepage=_HOMEPAGE,
|
123 |
-
citation=_CITATION,
|
124 |
-
)
|
125 |
-
|
126 |
-
def _split_generators(self, dl_manager):
|
127 |
-
download_url = _VALID_LANGUAGE_PAIRS.get(tuple(self.config.language_pair))
|
128 |
-
path = dl_manager.download_and_extract(download_url)
|
129 |
-
return [
|
130 |
-
datasets.SplitGenerator(
|
131 |
-
name=datasets.Split.TRAIN,
|
132 |
-
gen_kwargs={"datapath": path},
|
133 |
-
)
|
134 |
-
]
|
135 |
-
|
136 |
-
def _generate_examples(self, datapath):
|
137 |
-
lang1, lang2 = self.config.language_pair
|
138 |
-
lang1_file = _BASE_NAME.format(lang1, lang2, lang1)
|
139 |
-
lang2_file = _BASE_NAME.format(lang1, lang2, lang2)
|
140 |
-
lang1_path = os.path.join(datapath, lang1_file)
|
141 |
-
lang2_path = os.path.join(datapath, lang2_file)
|
142 |
-
|
143 |
-
with open(lang1_path, encoding="utf-8") as f1, open(lang2_path, encoding="utf-8") as f2:
|
144 |
-
for sentence_counter, (x, y) in enumerate(zip(f1, f2)):
|
145 |
-
x = x.strip()
|
146 |
-
y = y.strip()
|
147 |
-
result = (
|
148 |
-
sentence_counter,
|
149 |
-
{
|
150 |
-
"id": str(sentence_counter),
|
151 |
-
"translation": {lang1: x, lang2: y},
|
152 |
-
},
|
153 |
-
)
|
154 |
-
yield result
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|