aboutsummaryrefslogtreecommitdiffstats
path: root/python/tests/import_datacite.py
blob: 15650375346d698b8d7dd16fb5f91d881c5a3077 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
"""
Test datacite importer.
"""

import collections
import datetime
import pytest
import gzip
from fatcat_tools.importers import DataciteImporter, JsonLinePusher
from fatcat_tools.importers.datacite import find_original_language_title, parse_datacite_titles, parse_datacite_dates, clean_doi, index_form_to_display_name, lookup_license_slug
from fatcat_tools.transforms import entity_to_dict
from fixtures import api
import json


@pytest.fixture(scope="function")
def datacite_importer(api):
    with open('tests/files/ISSN-to-ISSN-L.snip.txt', 'r') as issn_file:
        yield DataciteImporter(api, issn_file, extid_map_file='tests/files/example_map.sqlite3',
                               bezerk_mode=True)

@pytest.fixture(scope="function")
def datacite_importer_existing(api):
    with open('tests/files/ISSN-to-ISSN-L.snip.txt', 'r') as issn_file:
        yield DataciteImporter(api, issn_file, extid_map_file='tests/files/example_map.sqlite3',
                               bezerk_mode=False)

@pytest.mark.skip(reason="larger datacite import slows tests down")
def test_datacite_importer_huge(datacite_importer):
    last_index = datacite_importer.api.get_changelog(limit=1)[0].index
    with gzip.open('tests/files/datacite_1k_records.jsonl.gz', 'rt') as f:
        datacite_importer.bezerk_mode = True
        counts = JsonLinePusher(datacite_importer, f).run()
    assert counts['insert'] == 998
    change = datacite_importer.api.get_changelog_entry(index=last_index+1)
    release = datacite_importer.api.get_release(change.editgroup.edits.releases[0].ident)
    assert len(release.contribs) == 3


def test_find_original_language_title():
    """
    Original language might be included, in various ways.
    """
    Case = collections.namedtuple('Case', 'about input result')
    cases = [
        Case('defaults to None', {}, None),
        Case('ignore unknown keys', {'broken': 'kv'}, None),
        Case('just a title', {'title': 'Noise Reduction'}, None),
        Case('same title should be ignored', {
            'title': 'Noise Reduction',
            'original_language_title': 'Noise Reduction'
        }, None),
        Case('empty subdict is ignored', {
            'title': 'Noise Reduction',
            'original_language_title': {},
        }, None),
        Case('unknown subdict keys are ignored', {
            'title': 'Noise Reduction',
            'original_language_title': {'broken': 'kv'},
        }, None),
        Case('original string', {
            'title': 'Noise Reduction',
            'original_language_title': 'Подавление шума',
        }, 'Подавление шума'),
        Case('language tag is ignored, since its broken', {
            'title': 'Noise Reduction',
            'original_language_title': {
                'language': 'ja',
                '__content__': 'Noise Reduction'
            },
        }, None),
        Case('do not care about language', {
            'title': 'Noise Reduction',
            'original_language_title': {
                'language': 'ja',
                '__content__': 'Rauschunterdrückung',
            },
        }, 'Rauschunterdrückung'),
        Case('ignore excessive questionmarks', {
            'title': 'Noise Reduction',
            'original_language_title': {
                'language': 'ja',
                '__content__': '???? However',
            },
        }, None),
    ]

    for case in cases:
        result = find_original_language_title(case.input)
        assert result == case.result

def test_parse_datacite_titles():
    """
    Given a list of titles, find title, original_language_title and subtitle.
    Result is a 3-tuple of title, original_language_title, subtitle.
    """
    Case = collections.namedtuple('Case', 'about input result')
    cases = [
        Case('handle None', None, (None, None, None)),
        Case('empty list', [], (None, None, None)),
        Case('empty item', [{}], (None, None, None)),
        Case('broken keys', [{'broken': 'kv'}], (None, None, None)),
        Case('title only', [{'title': 'Total carbon dioxide'}],
             ('Total carbon dioxide', None, None),
        ),
        Case('title and subtitle', [
            {'title': 'Total carbon dioxide'},
            {'title': 'Station TT043_7-9', 'titleType': 'Subtitle'},
        ],
             ('Total carbon dioxide', None, 'Station TT043_7-9'),
        ),
        Case('title, subtitle order does not matter', [
            {'title': 'Station TT043_7-9', 'titleType': 'Subtitle'},
            {'title': 'Total carbon dioxide'},
        ],
             ('Total carbon dioxide', None, 'Station TT043_7-9'),
        ),
        Case('multiple titles, first wins', [
            {'title': 'Total carbon dioxide'},
            {'title': 'Meeting Heterogeneity'},
        ],
             ('Total carbon dioxide', None, None),
        ),
        Case('multiple titles, plus sub', [
            {'title': 'Total carbon dioxide'},
            {'title': 'Meeting Heterogeneity'},
            {'title': 'Station TT043_7-9', 'titleType': 'Subtitle'},
        ],
             ('Total carbon dioxide', None, 'Station TT043_7-9'),
        ),
        Case('multiple titles, multiple subs', [
            {'title': 'Total carbon dioxide'},
            {'title': 'Meeting Heterogeneity'},
            {'title': 'Station TT043_7-9', 'titleType': 'Subtitle'},
            {'title': 'Some other subtitle', 'titleType': 'Subtitle'},
        ],
             ('Total carbon dioxide', None, 'Station TT043_7-9'),
        ),
        Case('title, original, sub', [
            {'title': 'Total carbon dioxide', 'original_language_title': 'Всего углекислого газа'},
            {'title': 'Station TT043_7-9', 'titleType': 'Subtitle'},
        ],
             ('Total carbon dioxide', 'Всего углекислого газа', 'Station TT043_7-9'),
        ),
        Case('title, original same as title, sub', [
            {'title': 'Total carbon dioxide', 'original_language_title': {
                '__content__': 'Total carbon dioxide',
            }},
            {'title': 'Station TT043_7-9', 'titleType': 'Subtitle'},
        ],
             ('Total carbon dioxide', None, 'Station TT043_7-9'),
        ),
        Case('title, original dict, sub', [
            {'title': 'Total carbon dioxide', 'original_language_title': {
                '__content__': 'Всего углекислого газа',
            }},
            {'title': 'Station TT043_7-9', 'titleType': 'Subtitle'},
        ],
             ('Total carbon dioxide', 'Всего углекислого газа', 'Station TT043_7-9'),
        ),
    ]

    for case in cases:
        result = parse_datacite_titles(case.input)
        assert result == case.result, case.about

def test_parse_datacite_dates():
    """
    Test datacite date parsing.
    """
    Case = collections.namedtuple('Case', 'about input result')
    cases = [
        Case('None is None', None, (None, None, None)),
        Case('empty list is None', [], (None, None, None)),
        Case('empty item is None', [{}], (None, None, None)),
        Case('year only yields year only', [{'date': '2019'}], (None, None, 2019)),
        Case('first wins', [{'date': '2019'}, {'date': '2020'}], (None, None, 2019)),
        Case('skip bogus year', [{'date': 'abc'}, {'date': '2020'}], (None, None, 2020)),
        Case('first with type', [
            {'date': '2019', 'dateType': 'Accepted'}, {'date': '2020'}
        ], (None, None, 2019)),
        Case('full date', [
            {'date': '2019-12-01', 'dateType': 'Valid'},
        ], (datetime.date(2019, 12, 1), 12, 2019)),
        Case('date type prio', [
            {'date': '2000-12-01', 'dateType': 'Valid'},
            {'date': '2010-01-01', 'dateType': 'Updated'},
        ], (datetime.date(2000, 12, 1), 12, 2000)),
        Case('date type prio, Available > Updated', [
            {'date': '2010-01-01', 'dateType': 'Updated'},
            {'date': '2000-12-01', 'dateType': 'Available'},
        ], (datetime.date(2000, 12, 1), 12, 2000)),
        Case('allow different date formats, Available > Updated', [
            {'date': '2010-01-01T10:00:00', 'dateType': 'Updated'},
            {'date': '2000-12-01T10:00:00', 'dateType': 'Available'},
        ], (datetime.date(2000, 12, 1), 12, 2000)),
        Case('allow different date formats, Available > Updated', [
            {'date': '2010-01-01T10:00:00Z', 'dateType': 'Updated'},
            {'date': '2000-12-01T10:00:00Z', 'dateType': 'Available'},
        ], (datetime.date(2000, 12, 1), 12, 2000)),
        Case('allow fuzzy date formats, Available > Updated', [
            {'date': '2010', 'dateType': 'Updated'},
            {'date': '2000 Dec 01', 'dateType': 'Available'},
        ], (datetime.date(2000, 12, 1), 12, 2000)),
        Case('fuzzy year only', [
            {'date': 'Year 2010', 'dateType': 'Issued'},
        ], (None, None, 2010)),
        Case('fuzzy year and month', [
            {'date': 'Year 2010 Feb', 'dateType': 'Issued'},
        ], (None, 2, 2010)),
        Case('fuzzy year, month, day', [
            {'date': 'Year 2010 Feb 24', 'dateType': 'Issued'},
        ], (datetime.date(2010, 2, 24), 2, 2010)),
        Case('ignore broken date', [
            {'date': 'Febrrr 45', 'dateType': 'Updated'},
        ], (None, None, None)),
    ]
    for case in cases:
        result = parse_datacite_dates(case.input)
        assert result == case.result, case.about

def test_datacite_importer(datacite_importer):
    last_index = datacite_importer.api.get_changelog(limit=1)[0].index
    with open('tests/files/datacite_sample.jsonl', 'r') as f:
        datacite_importer.bezerk_mode = True
        counts = JsonLinePusher(datacite_importer, f).run()
    assert counts['insert'] == 1
    assert counts['exists'] == 0
    assert counts['skip'] == 0

    # fetch most recent editgroup
    change = datacite_importer.api.get_changelog_entry(index=last_index+1)
    eg = change.editgroup
    assert eg.description
    assert "datacite" in eg.description.lower()
    assert eg.extra['git_rev']
    assert "fatcat_tools.DataciteImporter" in eg.extra['agent']

    last_index = datacite_importer.api.get_changelog(limit=1)[0].index
    with open('tests/files/datacite_sample.jsonl', 'r') as f:
        datacite_importer.bezerk_mode = False
        datacite_importer.reset()
        counts = JsonLinePusher(datacite_importer, f).run()
    assert counts['insert'] == 0
    assert counts['exists'] == 1
    assert counts['skip'] == 0
    assert last_index == datacite_importer.api.get_changelog(limit=1)[0].index

def test_datacite_dict_parse(datacite_importer):
    with open('tests/files/datacite_sample.jsonl', 'r') as f:
        raw = json.load(f)
        r = datacite_importer.parse_record(raw)
        # ensure the API server is ok with format
        JsonLinePusher(datacite_importer, [json.dumps(raw)]).run()

        print(r.extra)
        assert r.title == "Triticum turgidum L. subsp. durum (Desf.) Husn. 97090"
        assert r.publisher == "International Centre for Agricultural Research in Dry Areas"
        assert r.release_type == "article"
        assert r.release_stage == "published"
        assert r.license_slug == None
        assert r.original_title == None
        assert r.ext_ids.doi == "10.18730/8dym9"
        assert r.ext_ids.isbn13 == None
        assert r.language == "en"
        assert r.subtitle == None
        assert r.release_date == None
        assert r.release_year == 1986
        assert 'subtitle' not in r.extra
        assert 'subtitle' not in r.extra['datacite']
        assert 'funder' not in r.extra
        assert 'funder' not in r.extra['datacite']
        # matched by ISSN, so shouldn't be in there
        #assert extra['container_name'] == "International Journal of Quantum Chemistry"
        assert r.extra['datacite']['subjects'] == [{'subject': 'Plant Genetic Resource for Food and Agriculture'}]
        assert len(r.abstracts) == 1
        assert len(r.abstracts[0].content) == 421
        assert len(r.contribs) == 2
        assert r.contribs[0].raw_name == "GLIS Of The ITPGRFA"
        assert r.contribs[0].given_name == None
        assert r.contribs[0].surname == None
        assert len(r.refs) == 0

def test_datacite_conversions(datacite_importer):
    """
    Datacite JSON to release entity JSON representation. The count is hardcoded
    for now.
    """
    datacite_importer.debug = True
    for i in range(31):
        src = 'tests/files/datacite/datacite_doc_{0:02d}.json'.format(i)
        dst = 'tests/files/datacite/datacite_result_{0:02d}.json'.format(i)
        with open(src, 'r') as f:
            re = datacite_importer.parse_record(json.load(f))
            result = entity_to_dict(re)
        with open(dst, 'r') as f:
           expected = json.loads(f.read())

        assert result == expected, 'output mismatch in {}'.format(dst)

def test_index_form_to_display_name():
    Case = collections.namedtuple('Case', 'input output')
    cases = [
        Case('', ''),
        Case('ABC', 'ABC'),
        Case('International Space Station', 'International Space Station'),
        Case('Jin, Shan', 'Shan Jin'),
        Case('Volkshochschule Der Bundesstadt Bonn', 'Volkshochschule Der Bundesstadt Bonn'),
        Case('Solomon, P. M.', 'P. M. Solomon'),
        Case('Sujeevan Ratnasingham', 'Sujeevan Ratnasingham'),
        Case('Paul Stöckli (1906-1991), Künstler', 'Paul Stöckli (1906-1991), Künstler'),
    ]

    for c in cases:
        assert c.output == index_form_to_display_name(c.input)


def test_lookup_license_slug():
    Case = collections.namedtuple('Case', 'input output')
    cases = [
        Case('https://opensource.org/licenses/MIT', 'MIT'),
        Case('creativecommons.org/licenses/by-nc-nd/3.0/', 'CC-BY-NC-ND'),
        Case('http://creativecommons.org/licences/by-nc-sa/4.0', 'CC-BY-NC-SA'),
        Case('http://creativecommons.org/licenses/by-nc-nd/2.5/co', 'CC-BY-NC-ND'),
        Case('http://creativecommons.org/licenses/by-nd/4.0/legalcode', 'CC-BY-ND'),
        Case('http://creativecommons.org/licenses/by/2.0/uk/legalcode', 'CC-BY'),
        Case('http://creativecommons.org/publicdomain/zero/1.0/legalcode', 'CC-0'),
        Case('http://doi.wiley.com/10.1002/tdm_license_1.1', 'WILEY-TDM-1.1'),
        Case('http://homepage.data-planet.com/terms-use', 'SAGE-DATA-PLANET'),
        Case('http://www.springer.com/tdm', 'SPRINGER-TDM'),
        Case('https://archaeologydataservice.ac.uk/advice/termsOfUseAndAccess.xhtml', 'ADS-UK'),
        Case('https://archaeologydataservice.ac.uk/advice/termsOfUseAndAccess', 'ADS-UK'),
        Case('https://creativecommons.org/public-domain/cc0', 'CC-0'),
        Case('https://creativecommons.org/publicdomain/zero/1.0', 'CC-0'),
        Case('https://creativecommons.org/share-your-work/public-domain/cc0', 'CC-0'),
        Case('https://www.elsevier.com/tdm/userlicense/1.0', 'ELSEVIER-USER-1.0'),
        Case('https://www.gnu.org/licenses/gpl-3.0.html', 'GPL-3.0'),
        Case('http://rightsstatements.org/page/InC/1.0?language=en', 'RS-INC'),
        Case('http://onlinelibrary.wiley.com/termsAndConditions', 'WILEY'),
        Case('https://publikationen.bibliothek.kit.edu/kitopen-lizenz', 'KIT-OPEN'),
        Case('http://journals.sagepub.com/page/policies/text-and-data-mining-license', 'SAGE-TDM'),
    ]

    for c in cases:
        got = lookup_license_slug(c.input)
        assert c.output == got, '{}: got {}, want {}'.format(c.input, got, c.output)