Datasets:

Modalities:
Text
Languages:
Hebrew
ArXiv:
Libraries:
Datasets
imvladikon commited on
Commit
d4b59cd
1 Parent(s): 9f0b37f

Create parashoot.py

Browse files
Files changed (1) hide show
  1. parashoot.py +134 -0
parashoot.py ADDED
@@ -0,0 +1,134 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #!/usr/bin/env python3
2
+ # -*- coding: utf-8 -*-
3
+ import json
4
+ import os
5
+
6
+ import datasets
7
+ from datasets.tasks import QuestionAnsweringExtractive
8
+
9
+
10
+ logger = datasets.logging.get_logger(__name__)
11
+
12
+
13
+ _CITATION = """\
14
+ @inproceedings{keren2021parashoot,
15
+ title={ParaShoot: A Hebrew Question Answering Dataset},
16
+ author={Keren, Omri and Levy, Omer},
17
+ booktitle={Proceedings of the 3rd Workshop on Machine Reading for Question Answering},
18
+ pages={106--112},
19
+ year={2021}
20
+ }
21
+ """
22
+
23
+ _DESCRIPTION = """
24
+ A Hebrew question and answering dataset in the style of SQuAD, based on articles scraped from Wikipedia. The dataset contains a few thousand crowdsource-annotated pairs of questions and answers, in a setting suitable for few-shot learning.
25
+ """
26
+
27
+ _URLS = {
28
+ "train": "data/train.tar.gz",
29
+ "validation": "data/dev.tar.gz",
30
+ "test": "data/test.tar.gz",
31
+ }
32
+
33
+
34
+ class ParashootConfig(datasets.BuilderConfig):
35
+ """BuilderConfig for Parashoot."""
36
+
37
+ def __init__(self, **kwargs):
38
+ """BuilderConfig for Parashoot.
39
+ Args:
40
+ **kwargs: keyword arguments forwarded to super.
41
+ """
42
+ super(ParashootConfig, self).__init__(**kwargs)
43
+
44
+
45
+ class Parashoot(datasets.GeneratorBasedBuilder):
46
+ """Parashoot: The Hebrew Question Answering Dataset. Version 1.1."""
47
+
48
+ BUILDER_CONFIGS = [
49
+ ParashootConfig(
50
+ version=datasets.Version("1.1.0", ""),
51
+ description=_DESCRIPTION,
52
+ ),
53
+ ]
54
+
55
+ def _info(self):
56
+ return datasets.DatasetInfo(
57
+ description=_DESCRIPTION,
58
+ features=datasets.Features(
59
+ {
60
+ "id": datasets.Value("string"),
61
+ "title": datasets.Value("string"),
62
+ "context": datasets.Value("string"),
63
+ "question": datasets.Value("string"),
64
+ "answers": datasets.features.Sequence(
65
+ {
66
+ "text": datasets.Value("string"),
67
+ "answer_start": datasets.Value("int32"),
68
+ }
69
+ ),
70
+ }
71
+ ),
72
+ # No default supervised_keys (as we have to pass both question
73
+ # and context as input).
74
+ supervised_keys=None,
75
+ homepage="https://github.com/omrikeren/ParaShoot",
76
+ citation=_CITATION,
77
+ task_templates=[
78
+ QuestionAnsweringExtractive(
79
+ question_column="question",
80
+ context_column="context",
81
+ answers_column="answers",
82
+ )
83
+ ],
84
+ )
85
+
86
+ def _split_generators(self, dl_manager):
87
+ downloaded_files = dl_manager.download_and_extract(_URLS)
88
+
89
+ return [
90
+ datasets.SplitGenerator(
91
+ name=datasets.Split.TRAIN,
92
+ gen_kwargs={
93
+ "filepath": downloaded_files["train"],
94
+ "basename": "train.jsonl",
95
+ },
96
+ ),
97
+ datasets.SplitGenerator(
98
+ name=datasets.Split.VALIDATION,
99
+ gen_kwargs={
100
+ "filepath": downloaded_files["validation"],
101
+ "basename": "dev.jsonl",
102
+ },
103
+ ),
104
+ datasets.SplitGenerator(
105
+ name=datasets.Split.TEST,
106
+ gen_kwargs={
107
+ "filepath": downloaded_files["test"],
108
+ "basename": "test.jsonl",
109
+ },
110
+ ),
111
+ ]
112
+
113
+ def _generate_examples(self, filepath, basename):
114
+ """This function returns the examples in the raw (text) form."""
115
+ logger.info("generating examples from = %s", filepath)
116
+ key = 0
117
+ with open(os.path.join(filepath, basename), encoding="utf-8") as f:
118
+ for line in f:
119
+ article = json.loads(line)
120
+ title = article.get("title", "")
121
+ context = article["context"]
122
+ answer_starts = article["answers"]["answer_start"]
123
+ answers = article["answers"]["text"]
124
+ yield key, {
125
+ "title": title,
126
+ "context": context,
127
+ "question": article["question"],
128
+ "id": article["id"],
129
+ "answers": {
130
+ "answer_start": answer_starts,
131
+ "text": answers,
132
+ },
133
+ }
134
+ key += 1