albertvillanova HF staff commited on
Commit
a6e91a9
1 Parent(s): 21f3804

Delete loading script

Browse files
Files changed (1) hide show
  1. blbooks.py +0 -234
blbooks.py DELETED
@@ -1,234 +0,0 @@
1
- # coding=utf-8
2
- # Copyright 2022 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
- import gzip
16
- import json
17
- from datetime import datetime
18
- from functools import lru_cache
19
- from typing import Dict, List
20
-
21
- import datasets
22
- from datasets.tasks import LanguageModeling
23
-
24
-
25
- _CITATION = """\
26
- @misc{BritishLibraryBooks2021,
27
- author = {British Library Labs},
28
- title = {Digitised Books. c. 1510 - c. 1900. JSONL (OCR derived text + metadata)},
29
- year = {2021},
30
- publisher = {British Library},
31
- howpublished={https://doi.org/10.23636/r7w6-zy15}
32
- """
33
-
34
- _DESCRIPTION = """\
35
- A dataset comprising of text created by OCR from the 49,455 digitised books, equating to 65,227 volumes (25+ million pages), published between c. 1510 - c. 1900.
36
- The books cover a wide range of subject areas including philosophy, history, poetry and literature.
37
- """
38
-
39
- _BASE_URL = "https://bl.iro.bl.uk/downloads/"
40
-
41
-
42
- _DATA_URLS = {
43
- "1510_1699": _BASE_URL + "61f58234-b370-422f-8591-8f98e46c2757?locale=en",
44
- "1700_1799": _BASE_URL + "78b4a8ec-395e-4383-831c-809faff85ad7?locale=en",
45
- "1800_1809": _BASE_URL + "91ae15cb-e08f-4abf-8396-e4742d9d4e37?locale=en",
46
- "1810_1819": _BASE_URL + "6d1a6e17-f28d-45b9-8f7a-a03cf3a96491?locale=en",
47
- "1820_1829": _BASE_URL + "ec764dbd-1ed4-4fc2-8668-b4df5c8ec451?locale=en",
48
- "1830_1839": _BASE_URL + "eab68022-0418-4df7-a401-78972514ed20?locale=en",
49
- "1840_1849": _BASE_URL + "d16d88b0-aa3f-4dfe-b728-c58d168d7b4d?locale=en",
50
- "1850_1859": _BASE_URL + "a6a44ea8-8d33-4880-8b17-f89c90e3d89a?locale=en",
51
- "1860_1869": _BASE_URL + "2e17f00f-52e6-4259-962c-b88ad60dec23?locale=en",
52
- "1870_1879": _BASE_URL + "899c3719-030c-4517-abd3-b28fdc85eed4?locale=en",
53
- "1880_1889": _BASE_URL + "ec3b8545-775b-47bd-885d-ce895263709e?locale=en",
54
- "1890_1899": _BASE_URL + "54ed2842-089a-439a-b751-2179b3ffba28?locale=en",
55
- }
56
-
57
- _ALL = list(_DATA_URLS.values())
58
- _1800_1899 = [
59
- _DATA_URLS.get(subset)
60
- for subset in [
61
- "1800_1809",
62
- "1810_1819",
63
- "1820_1829",
64
- "1830_1839",
65
- "1840_1849",
66
- "1850_1859",
67
- "1860_1869",
68
- "1870_1879",
69
- "1880_1889",
70
- "1890_1899",
71
- ]
72
- ]
73
- _1700_1799 = [_DATA_URLS.get(subset) for subset in ["1700_1799"]]
74
- _1510_1699 = [_DATA_URLS.get(subset) for subset in ["1510_1699"]]
75
-
76
- URL = "https://doi.org/10.23636/r7w6-zy15"
77
-
78
- features = datasets.Features(
79
- {
80
- "record_id": datasets.Value("string"),
81
- "date": datasets.Value("timestamp[s]"),
82
- "raw_date": datasets.Value("string"),
83
- "title": datasets.Value("string"),
84
- "place": datasets.Value("string"),
85
- "empty_pg": datasets.Value("bool"),
86
- "text": datasets.Value("string"),
87
- "pg": datasets.Value("int32"),
88
- "mean_wc_ocr": datasets.Value("float32"),
89
- "std_wc_ocr": datasets.Value("float64"),
90
- "name": datasets.Value("string"),
91
- "all_names": datasets.Value("string"),
92
- "Publisher": datasets.Value("string"),
93
- "Country of publication 1": datasets.Value("string"),
94
- "all Countries of publication": datasets.Value("string"),
95
- "Physical description": datasets.Value("string"),
96
- "Language_1": datasets.Value("string"),
97
- "Language_2": datasets.Value("string"),
98
- "Language_3": datasets.Value("string"),
99
- "Language_4": datasets.Value("string"),
100
- "multi_language": datasets.Value("bool"),
101
- }
102
- )
103
-
104
-
105
- class BritishLibraryBooksConfig(datasets.BuilderConfig):
106
- """BuilderConfig for BritishLibraryBooks."""
107
-
108
- def __init__(self, data_urls, citation, url, skip_empty=False, **kwargs):
109
- """BuilderConfig for BritishLibraryBooks.
110
-
111
- Args:
112
- data_url: `string`, url to download the zip file from.
113
- citation: `string`, citation for the data set.
114
- url: `string`, url for information about the data set.
115
- skip_empty: `bool`, whether to skip empty pages.
116
- **kwargs: keyword arguments forwarded to super.
117
- """
118
-
119
- super(BritishLibraryBooksConfig, self).__init__(version=datasets.Version("1.0.2"), **kwargs)
120
- self.url: str = url
121
- self.data_urls: List[str] = data_urls
122
- self.citation: str = citation
123
- self.skip_empty: bool = skip_empty
124
-
125
-
126
- class BritishLibraryBooks(datasets.GeneratorBasedBuilder):
127
- """The BritishLibraryBooks dataset."""
128
-
129
- BUILDER_CONFIGS = [
130
- BritishLibraryBooksConfig(
131
- name="1500_1899",
132
- description="All periods of" + _DESCRIPTION,
133
- data_urls=_ALL,
134
- citation=_CITATION,
135
- url=URL,
136
- skip_empty=True,
137
- ),
138
- BritishLibraryBooksConfig(
139
- name="1800_1899",
140
- description="A subset covering texts published during the 1800-1899 of" + _DESCRIPTION,
141
- data_urls=_1800_1899,
142
- citation=_CITATION,
143
- url=URL,
144
- skip_empty=True,
145
- ),
146
- BritishLibraryBooksConfig(
147
- name="1700_1799",
148
- description="Subset covering 1700-1799 of" + _DESCRIPTION,
149
- data_urls=_1700_1799,
150
- citation=_CITATION,
151
- url=URL,
152
- skip_empty=True,
153
- ),
154
- BritishLibraryBooksConfig(
155
- name="1510_1699",
156
- description="Subset covering 1510-1699 of " + _DESCRIPTION,
157
- data_urls=_1510_1699,
158
- citation=_CITATION,
159
- url=URL,
160
- skip_empty=True,
161
- ),
162
- ]
163
-
164
- DEFAULT_CONFIG_NAME = "1500_1899"
165
-
166
- def _info(self):
167
- return datasets.DatasetInfo(
168
- description=_DESCRIPTION,
169
- features=features,
170
- supervised_keys=None,
171
- homepage="https://www.bl.uk/collection-guides/digitised-printed-books",
172
- citation=_CITATION,
173
- task_templates=[LanguageModeling(text_column="text")],
174
- )
175
-
176
- def _split_generators(self, dl_manager: datasets.DownloadManager):
177
- urls_to_download = self.config.data_urls
178
- downloaded_archives = dl_manager.download(urls_to_download)
179
- downloaded_archives = [dl_manager.iter_archive(archive) for archive in downloaded_archives]
180
- return [datasets.SplitGenerator(name=datasets.Split.TRAIN, gen_kwargs={"data_dirs": downloaded_archives})]
181
-
182
- @lru_cache(maxsize=512)
183
- def _parse_date(self, date):
184
- if date is not None:
185
- date = datetime.strptime(str(date), "%Y")
186
- return date
187
-
188
- def _parse_data(self, data: Dict) -> Dict:
189
- mean_wc_ocr = data["mean_wc_ocr"]
190
- mean_wc_ocr = float(mean_wc_ocr) if mean_wc_ocr else None
191
- std_wc_ocr = data["std_wc_ocr"]
192
- std_wc_ocr = float(data["std_wc_ocr"]) if std_wc_ocr else None
193
- date = data["date"]
194
- if date is not None:
195
- date = datetime.strptime(str(date), "%Y")
196
- return {
197
- "record_id": data["record_id"],
198
- "date": date,
199
- "raw_date": data["raw_date"],
200
- "title": data["title"],
201
- "place": data["place"],
202
- "text": data["text"],
203
- "pg": int(data["pg"]),
204
- "mean_wc_ocr": data["mean_wc_ocr"],
205
- "std_wc_ocr": std_wc_ocr,
206
- "name": data["Name"],
207
- "all_names": data["All names"],
208
- "Publisher": data["Publisher"],
209
- "Country of publication 1": data["Country of publication 1"],
210
- "all Countries of publication": data["All Countries of publication"],
211
- "Physical description": data["Physical description"],
212
- "Language_1": data["Language_1"],
213
- "Language_2": data["Language_2"],
214
- "Language_3": data["Language_3"],
215
- "Language_4": data["Language_4"],
216
- "multi_language": data["multi_language"],
217
- }
218
-
219
- def _generate_examples(self, data_dirs):
220
- skip_empty = self.config.skip_empty
221
- id_ = 0
222
- for data_dir in data_dirs:
223
- for path, file in data_dir:
224
- if not path.endswith(".gz"):
225
- continue
226
- with gzip.open(file) as json_l:
227
- for row in json_l:
228
- data = json.loads(row)
229
- empty_pg = data["empty_pg"]
230
- if skip_empty and empty_pg:
231
- continue
232
- parsed_data = self._parse_data(data)
233
- yield id_, {**parsed_data, **{"empty_pg": empty_pg}}
234
- id_ += 1