davanstrien HF staff commited on
Commit
4f8d358
1 Parent(s): fec92d0

Upload ethos.py

Browse files
Files changed (1) hide show
  1. ethos.py +153 -0
ethos.py ADDED
@@ -0,0 +1,153 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # coding=utf-8
2
+ """British Library EThos dataset"""
3
+
4
+ import csv
5
+ from datetime import datetime
6
+ import datasets
7
+ from datasets.features import Features
8
+ from datasets.tasks import LanguageModeling
9
+
10
+ _CITATION = """\
11
+ @misc{british library_genre,
12
+ title={UK Doctoral Thesis Metadata from EThOS},
13
+ url={UK Doctoral Thesis Metadata from EThOS},
14
+ author={{British Library} and {Rosie, Heather}},
15
+ year={2021}}
16
+ """
17
+
18
+
19
+ _DESCRIPTION = """\
20
+ The data in this collection comprises the bibliographic metadata for all UK doctoral theses listed in EThOS, the UK's national thesis service.
21
+ We estimate the data covers around 98% of all PhDs ever awarded by UK Higher Education institutions, dating back to 1787.
22
+ Thesis metadata from every PhD-awarding university in the UK is included.
23
+ """
24
+
25
+
26
+ _HOMEPAGE = "https://doi.org/10.23636/ybpt-nh33"
27
+ _LICENSE = "CC BY 4.0 Attribution"
28
+
29
+
30
+ _URL = "https://bl.iro.bl.uk/downloads/3043a513-72af-4fad-864d-1da5ee1e8ed1?locale=en"
31
+
32
+
33
+ features = Features(
34
+ {
35
+ "Title": datasets.Value("string"),
36
+ "DOI": datasets.Value("string"),
37
+ "Author": datasets.Value("string"),
38
+ "Author ISNI": datasets.Value("string"),
39
+ "ORCID": datasets.Value("string"),
40
+ "Institution": datasets.Value("string"),
41
+ "Institution ISNI": datasets.Value("string"),
42
+ "Date": datasets.Value("timestamp[s]"),
43
+ "Qualification": datasets.Value("string"),
44
+ "Abstract": datasets.Value("string"),
45
+ "Subject Discipline": datasets.ClassLabel(
46
+ names=[
47
+ "Physical Sciences",
48
+ "Biological Sciences",
49
+ "Engineering & Technology",
50
+ "Mathematics & Statistics",
51
+ "Agriculture & Veterinary Sciences",
52
+ "Medicine & Health",
53
+ "Computer Science",
54
+ "Philosophy, Psychology & Religious Studies",
55
+ "Business & Administrative Studies",
56
+ "Education",
57
+ "Language & Literature",
58
+ "Social, Economic & Political Studies",
59
+ "Architecture, Building & Planning",
60
+ "History & Archaeology",
61
+ "Creative Arts & Design",
62
+ "Law",
63
+ "Sport & Recreation",
64
+ "Librarianship & Information Science",
65
+ "Music",
66
+ " ",
67
+ ]
68
+ ),
69
+ "Supervisor(s)": datasets.Value("string"),
70
+ "Funder(s)": datasets.Value("string"),
71
+ "EThOS URL": datasets.Value("string"),
72
+ "IR URL": datasets.Value("string"),
73
+ }
74
+ )
75
+
76
+
77
+ class Ethos(datasets.GeneratorBasedBuilder):
78
+
79
+ VERSION = datasets.Version("1.1.0")
80
+ BUILDER_CONFIGS = [
81
+ datasets.BuilderConfig(
82
+ name="all",
83
+ version=VERSION,
84
+ description="",
85
+ ),
86
+ datasets.BuilderConfig(
87
+ "skip_empty_abstracts",
88
+ version=VERSION,
89
+ description="EThOs skipping entires with no abstract",
90
+ ),
91
+ ]
92
+
93
+ DEFAULT_CONFIG_NAME = "all"
94
+
95
+ def _info(self):
96
+ return datasets.DatasetInfo(
97
+ description=_DESCRIPTION,
98
+ features=features,
99
+ supervised_keys=None,
100
+ homepage=_HOMEPAGE,
101
+ license=_LICENSE,
102
+ citation=_CITATION,
103
+ task_templates=[LanguageModeling(text_column="Abstract")],
104
+ )
105
+
106
+ def _split_generators(self, dl_manager):
107
+ """Returns SplitGenerators."""
108
+ data_file = dl_manager.download(_URL)
109
+ return [
110
+ datasets.SplitGenerator(
111
+ name=datasets.Split.TRAIN,
112
+ # These kwargs will be passed to _generate_examples
113
+ gen_kwargs={
114
+ "filepath": data_file,
115
+ },
116
+ ),
117
+ ]
118
+
119
+ def _generate_examples(self, filepath):
120
+ """Yields examples as (key, example) tuples."""
121
+ if self.config.name == "skip_empty_abstracts":
122
+ skip = True
123
+ else:
124
+ skip = False
125
+ with open(filepath, encoding="latin-1") as f:
126
+ reader = csv.DictReader(f)
127
+ id_ = 0
128
+ for row in reader:
129
+ abstract = row["Abstract"]
130
+ if skip and len(abstract) < 2:
131
+ continue
132
+ try:
133
+ date = datetime.strptime(row["Date"], "%Y")
134
+ except ValueError:
135
+ date = None
136
+ id_ += 1
137
+ yield id_, {
138
+ "Title": row["Title"],
139
+ "DOI": row["DOI"],
140
+ "Author": row["Author"],
141
+ "Author ISNI": row["Author ISNI"],
142
+ "ORCID": row["ORCID"],
143
+ "Institution": row["Institution"],
144
+ "Institution ISNI": row["Institution ISNI"],
145
+ "Date": date,
146
+ "Qualification": row["Qualification"],
147
+ "Abstract": abstract,
148
+ "Subject Discipline": row["Subject Discipline"],
149
+ "Supervisor(s)": row["Supervisor(s)"],
150
+ "Funder(s)": row["Funder(s)"],
151
+ "EThOS URL": row["EThOS URL"],
152
+ "IR URL": row["IR URL"],
153
+ }