Datasets:
Commit
•
5ee2de3
1
Parent(s):
6958faa
Delete loading script
Browse files- aquamuse.py +0 -154
aquamuse.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 |
-
"""AQuaMuSe is a novel scalable approach to automatically mine dual query based multi-document summarization datasets for extractive and abstractive summaries using question answering dataset (Google Natural Questions) and large document corpora (Common Crawl)"""
|
16 |
-
|
17 |
-
|
18 |
-
import os
|
19 |
-
from os import listdir
|
20 |
-
from os.path import isfile, join
|
21 |
-
|
22 |
-
import tensorflow as tf
|
23 |
-
|
24 |
-
import datasets
|
25 |
-
|
26 |
-
|
27 |
-
_CITATION = """\
|
28 |
-
@misc{kulkarni2020aquamuse,
|
29 |
-
title={AQuaMuSe: Automatically Generating Datasets for Query-Based Multi-Document Summarization},
|
30 |
-
author={Sayali Kulkarni and Sheide Chammas and Wan Zhu and Fei Sha and Eugene Ie},
|
31 |
-
year={2020},
|
32 |
-
eprint={2010.12694},
|
33 |
-
archivePrefix={arXiv},
|
34 |
-
primaryClass={cs.CL}
|
35 |
-
}
|
36 |
-
"""
|
37 |
-
|
38 |
-
_DESCRIPTION = """AQuaMuSe is a novel scalable approach to automatically mine dual query based multi-document summarization datasets for extractive and abstractive summaries using question answering dataset (Google Natural Questions) and large document corpora (Common Crawl)"""
|
39 |
-
|
40 |
-
_HOMEPAGE = "https://github.com/google-research-datasets/aquamuse"
|
41 |
-
|
42 |
-
_LICENSE = ""
|
43 |
-
|
44 |
-
zipped_data_url = "https://github.com/google-research-datasets/aquamuse/raw/main/v2/aquamuse_v2.zip"
|
45 |
-
|
46 |
-
|
47 |
-
class Aquamuse(datasets.GeneratorBasedBuilder):
|
48 |
-
"""Dataset for Query-based Multi-Document Summarization"""
|
49 |
-
|
50 |
-
VERSION = datasets.Version("2.3.0")
|
51 |
-
|
52 |
-
BUILDER_CONFIGS = [
|
53 |
-
datasets.BuilderConfig(
|
54 |
-
name="abstractive", version=VERSION, description="Abstractive query-based multi-document summarization"
|
55 |
-
),
|
56 |
-
datasets.BuilderConfig(
|
57 |
-
name="extractive", version=VERSION, description="Extractive query-based multi-document summarization"
|
58 |
-
),
|
59 |
-
]
|
60 |
-
|
61 |
-
# DEFAULT_CONFIG_NAME = "abstractive" # It's not mandatory to have a default configuration. Just use one if it make sense.
|
62 |
-
|
63 |
-
def _info(self):
|
64 |
-
features = datasets.Features(
|
65 |
-
{
|
66 |
-
"query": datasets.Value("string"),
|
67 |
-
"input_urls": datasets.Sequence(datasets.Value("string")),
|
68 |
-
"target": datasets.Value("string"),
|
69 |
-
}
|
70 |
-
)
|
71 |
-
|
72 |
-
return datasets.DatasetInfo(
|
73 |
-
description=_DESCRIPTION,
|
74 |
-
features=features,
|
75 |
-
supervised_keys=None,
|
76 |
-
homepage=_HOMEPAGE,
|
77 |
-
license=_LICENSE,
|
78 |
-
citation=_CITATION,
|
79 |
-
)
|
80 |
-
|
81 |
-
def _split_generators(self, dl_manager):
|
82 |
-
"""Returns SplitGenerators."""
|
83 |
-
|
84 |
-
if self.config.name == "abstractive":
|
85 |
-
data_dir = dl_manager.download_and_extract(zipped_data_url)
|
86 |
-
return [
|
87 |
-
datasets.SplitGenerator(
|
88 |
-
name=datasets.Split.TRAIN,
|
89 |
-
# These kwargs will be passed to _generate_examples
|
90 |
-
gen_kwargs={
|
91 |
-
"filepath": os.path.join(data_dir, "v2.3/abstractive/train/"),
|
92 |
-
"split": "train",
|
93 |
-
},
|
94 |
-
),
|
95 |
-
datasets.SplitGenerator(
|
96 |
-
name=datasets.Split.TEST,
|
97 |
-
# These kwargs will be passed to _generate_examples
|
98 |
-
gen_kwargs={
|
99 |
-
"filepath": os.path.join(data_dir, "v2.3/abstractive/test/"),
|
100 |
-
"split": "test",
|
101 |
-
},
|
102 |
-
),
|
103 |
-
datasets.SplitGenerator(
|
104 |
-
name=datasets.Split.VALIDATION,
|
105 |
-
# These kwargs will be passed to _generate_examples
|
106 |
-
gen_kwargs={
|
107 |
-
"filepath": os.path.join(data_dir, "v2.3/abstractive/dev/"),
|
108 |
-
"split": "dev",
|
109 |
-
},
|
110 |
-
),
|
111 |
-
]
|
112 |
-
|
113 |
-
else:
|
114 |
-
data_dir = dl_manager.download_and_extract(zipped_data_url)
|
115 |
-
return [
|
116 |
-
datasets.SplitGenerator(
|
117 |
-
name=datasets.Split.TRAIN,
|
118 |
-
# These kwargs will be passed to _generate_examples
|
119 |
-
gen_kwargs={
|
120 |
-
"filepath": os.path.join(data_dir, "v2.3/extractive/train/"),
|
121 |
-
"split": "train",
|
122 |
-
},
|
123 |
-
),
|
124 |
-
datasets.SplitGenerator(
|
125 |
-
name=datasets.Split.TEST,
|
126 |
-
# These kwargs will be passed to _generate_examples
|
127 |
-
gen_kwargs={
|
128 |
-
"filepath": os.path.join(data_dir, "v2.3/extractive/test/"),
|
129 |
-
"split": "test",
|
130 |
-
},
|
131 |
-
),
|
132 |
-
datasets.SplitGenerator(
|
133 |
-
name=datasets.Split.VALIDATION,
|
134 |
-
# These kwargs will be passed to _generate_examples
|
135 |
-
gen_kwargs={
|
136 |
-
"filepath": os.path.join(data_dir, "v2.3/extractive/dev/"),
|
137 |
-
"split": "dev",
|
138 |
-
},
|
139 |
-
),
|
140 |
-
]
|
141 |
-
|
142 |
-
def _generate_examples(self, filepath, split):
|
143 |
-
"""Yields examples."""
|
144 |
-
filepath = [join(filepath, f) for f in listdir(filepath) if isfile(join(filepath, f))]
|
145 |
-
filepath = sorted(filepath)
|
146 |
-
raw_dataset = tf.data.TFRecordDataset(filepath)
|
147 |
-
for id_, raw_record in enumerate(raw_dataset):
|
148 |
-
example = tf.train.Example()
|
149 |
-
example.ParseFromString(raw_record.numpy())
|
150 |
-
yield id_, {
|
151 |
-
"query": example.features.feature["query"].bytes_list.value[0].decode(),
|
152 |
-
"input_urls": example.features.feature["input_urls"].bytes_list.value[0].decode().split("<EOD>"),
|
153 |
-
"target": example.features.feature["target"].bytes_list.value[0].decode(),
|
154 |
-
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|