summaryrefslogtreecommitdiffstats
path: root/python/fatcat_tools/transforms/elasticsearch.py
blob: 331ca4e76c46a9a1ffdbd8f5b03b0e400504ca1e (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
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618

import datetime
from typing import Optional

import tldextract

from fatcat_openapi_client import ReleaseEntity, ContainerEntity


def check_kbart(year: int, archive: dict) -> Optional[bool]:
    if not archive or not archive.get('year_spans'):
        return None
    for span in archive['year_spans']:
        if year >= span[0] and year <= span[1]:
            return True
    return False

def test_check_kbart() -> None:

    assert check_kbart(1990, dict()) is None
    assert check_kbart(1990, dict(year_spans=[[2000, 2000]])) is False
    assert check_kbart(2000, dict(year_spans=[[2000, 2000]])) is True
    assert check_kbart(1950, dict(year_spans=[[1900, 1920], [1990, 2000]])) is False
    assert check_kbart(1950, dict(year_spans=[[1900, 1920], [1930, 2000]])) is True


def release_to_elasticsearch(entity: ReleaseEntity, force_bool: bool = True) -> dict:
    """
    Converts from an entity model/schema to elasticsearch oriented schema.

    This is a large/complex transform, so subsets are split out into helper
    functions.

    Returns: dict
    Raises exception on error (never returns None)
    """

    if entity.state in ('redirect', 'deleted'):
        return dict(
            ident = entity.ident,
            state = entity.state,
        )
    elif entity.state != 'active':
        raise ValueError("Unhandled entity state: {}".format(entity.state))

    # First, the easy ones (direct copy)
    release = entity
    t = dict(
        doc_index_ts=datetime.datetime.utcnow().isoformat()+"Z",
        ident = release.ident,
        state = release.state,
        revision = release.revision,
        work_id = release.work_id,
        title = release.title,
        subtitle = release.subtitle,
        original_title = release.original_title,
        release_type = release.release_type,
        release_stage = release.release_stage,
        withdrawn_status = release.withdrawn_status,
        language = release.language,
        volume = release.volume,
        issue = release.issue,
        pages = release.pages,
        number = release.number,
        license = release.license_slug,
        doi = release.ext_ids.doi,
        pmid = release.ext_ids.pmid,
        pmcid = release.ext_ids.pmcid,
        isbn13 = release.ext_ids.isbn13,
        wikidata_qid = release.ext_ids.wikidata_qid,
        core_id = release.ext_ids.core,
        arxiv_id = release.ext_ids.arxiv,
        jstor_id = release.ext_ids.jstor,
        ark_id = release.ext_ids.ark,
        mag_id = release.ext_ids.mag,
        dblp_id = release.ext_ids.dblp,
        doaj_id = release.ext_ids.doaj,
    )

    t.update(dict(
        is_oa = None,
        is_longtail_oa = None,
        is_preserved = None,
        in_web = False,
        in_dweb = False,
        in_ia = False,
        in_ia_sim = False,
        in_kbart = None,
        in_jstor = False,
        in_doaj= bool(release.ext_ids.doaj),
        in_shadows = False,
    ))

    release_year = release.release_year
    if release.release_date:
        # .isoformat() results in, eg, '2010-10-22' (YYYY-MM-DD)
        t['release_date'] = release.release_date.isoformat()
        if not release_year:
            release_year = release.release_date.year
    if release_year:
        t['release_year'] = release_year

    t['any_abstract'] = len(release.abstracts or []) > 0
    t['ref_count'] = len(release.refs or [])
    ref_release_ids = []
    for r in (release.refs or []):
        if r.target_release_id:
            ref_release_ids.append(r.target_release_id)
    t['ref_release_ids'] = ref_release_ids
    t['ref_linked_count'] = len(ref_release_ids)
    t['contrib_count'] = len(release.contribs or [])
    contrib_names = []
    contrib_affiliations = []
    creator_ids = []
    for c in (release.contribs or []):
        if c.creator and c.creator.display_ame:
            contrib_names.append(c.creator.display_ame)
        elif c.raw_name:
            contrib_names.append(c.raw_name)
        elif c.surname:
            contrib_names.append(c.surname)

        if c.creator_id:
            creator_ids.append(c.creator_id)
        if c.raw_affiliation:
            contrib_affiliations.append(c.raw_affiliation)
    t['contrib_names'] = contrib_names
    t['creator_ids'] = creator_ids
    t['affiliations'] = contrib_affiliations

    # TODO: mapping... probably by lookup?
    t['affiliation_rors'] = None

    if release.container:
        t.update(_rte_container_helper(release.container, release_year))

    # fall back to release-level container metadata if container not linked or
    # missing context
    if not t.get('publisher'):
        t['publisher'] = release.publisher
    if not t.get('container_name') and release.extra:
        t['container_name'] = release.extra.get('container_name')

    if release.ext_ids.jstor or (release.ext_ids.doi and release.ext_ids.doi.startswith('10.2307/')):
        t['in_jstor'] = True

    # transform file/fileset/webcapture related fields
    t.update(_rte_content_helper(release))

    if release.ext_ids.doaj:
        t['is_oa'] = True

    if release.license_slug:
        # TODO: more/better checks here, particularly strict *not* OA licenses
        if release.license_slug.startswith("CC-"):
            t['is_oa'] = True
        if release.license_slug.startswith("ARXIV-"):
            t['is_oa'] = True

    extra = release.extra or dict()
    if extra:
        if extra.get('is_oa'):
            # NOTE: not actually setting this anywhere... but could
            t['is_oa'] = True
        if extra.get('longtail_oa'):
            # sometimes set by GROBID/matcher
            t['is_oa'] = True
            t['is_longtail_oa'] = True
        if not t.get('container_name'):
            t['container_name'] = extra.get('container_name')
        if extra.get('crossref'):
            if extra['crossref'].get('archive'):
                # all crossref archives are KBART, I believe
                t['in_kbart'] = True
        # backwards compatible subtitle fetching
        if not t['subtitle'] and extra.get('subtitle'):
            if type(extra['subtitle']) == list:
                t['subtitle'] = extra['subtitle'][0]
            else:
                t['subtitle'] = extra['subtitle']

    t['first_page'] = None
    if release.pages:
        first = release.pages.split('-')[0]
        first = first.replace('p', '')
        if first.isdigit():
            t['first_page'] = first
        # TODO: non-numerical first pages

    t['ia_microfilm_url'] = None
    if t['in_ia_sim']:
        # TODO: determine URL somehow? I think this is in flux. Will probably
        # need extra metadata in the container extra field.
        # special case as a demo for now.
        if release.container_id == "hl5g6d5msjcl7hlbyyvcsbhc2u" \
                and release.release_year in (2011, 2013) \
                and release.issue \
                and release.issue.isdigit() \
                and t['first_page']:
            t['ia_microfilm_url'] = "https://archive.org/details/sim_bjog_{}-{:02d}/page/n{}".format(
                release.release_year,
                int(release.issue) - 1,
                t['first_page'],
            )

    t['doi_registrar'] = None
    if extra and t['doi']:
        for k in ('crossref', 'datacite', 'jalc'):
            if k in extra:
                t['doi_registrar'] = k
        if not 'doi_registrar' in t:
            t['doi_registrar'] = 'crossref'

    if t['doi']:
        t['doi_prefix'] = t['doi'].split('/')[0]

    if t['is_longtail_oa']:
        t['is_oa'] = True

    # optionally coerce all flags from Optional[bool] to bool
    if force_bool:
        for k in ('is_oa', 'is_longtail_oa', 'in_kbart', 'in_ia_sim',
                  'in_jstor', 'in_web', 'in_dweb', 'in_shadows'):
            t[k] = bool(t[k])

    t['in_ia'] = bool(t['in_ia'])
    t['is_preserved'] = bool(
        t['is_preserved']
        or t['in_ia']
        or t['in_kbart']
        or t['in_jstor']
        or t.get('pmcid')
        or t.get('arxiv_id')
    )

    if t['in_ia']:
        t['preservation'] = 'bright'
    elif t['is_preserved']:
        t['preservation'] = 'dark'
    elif t['in_shadows']:
        t['preservation'] = 'shadows_only'
    else:
        t['preservation'] = 'none'

    return t

def _rte_container_helper(container: ContainerEntity, release_year: Optional[int]) -> dict:
    """
    Container metadata sub-section of release_to_elasticsearch()
    """
    this_year = datetime.date.today().year
    t = dict()
    t['publisher'] = container.publisher
    t['container_name'] = container.name
    # this is container.ident, not release.container_id, because there may
    # be a redirect involved
    t['container_id'] = container.ident
    t['container_issnl'] = container.issnl
    t['container_type'] = container.container_type
    if container.extra:
        c_extra = container.extra
        if c_extra.get('kbart') and release_year:
            if check_kbart(release_year, c_extra['kbart'].get('jstor')):
                t['in_jstor'] = True
            if t.get('in_kbart') or t.get('in_jstor'):
                t['in_kbart'] = True
            for archive in ('portico', 'lockss', 'clockss', 'pkp_pln',
                            'hathitrust', 'scholarsportal', 'cariniana'):
                t['in_kbart'] = t.get('in_kbart') or check_kbart(release_year, c_extra['kbart'].get(archive))
                # recent KBART coverage is often not updated for the
                # current year. So for current-year publications, consider
                # coverage from *last* year to also be included in the
                # Keeper
                if not t.get('in_kbart') and release_year == this_year:
                    t['in_kbart'] = check_kbart(this_year - 1, c_extra['kbart'].get(archive))

        if c_extra.get('ia'):
            if c_extra['ia'].get('sim') and release_year:
                t['in_ia_sim'] = check_kbart(release_year, c_extra['ia']['sim'])
            if c_extra['ia'].get('longtail_oa'):
                t['is_longtail_oa'] = True
        if c_extra.get('sherpa_romeo'):
            if c_extra['sherpa_romeo'].get('color') == 'white':
                t['is_oa'] = False
        if c_extra.get('default_license') and c_extra.get('default_license').startswith('CC-'):
            t['is_oa'] = True
        if c_extra.get('doaj'):
            if c_extra['doaj'].get('as_of'):
                t['is_oa'] = True
                t['in_doaj'] = True
        if c_extra.get('road'):
            if c_extra['road'].get('as_of'):
                t['is_oa'] = True
        if c_extra.get('szczepanski'):
            if c_extra['szczepanski'].get('as_of'):
                t['is_oa'] = True
        if c_extra.get('country'):
            t['country_code'] = c_extra['country']
            t['country_code_upper'] = c_extra['country'].upper()
        if c_extra.get('publisher_type'):
            t['publisher_type'] = c_extra['publisher_type']
        if c_extra.get('discipline'):
            t['discipline'] = c_extra['discipline']
    return t

def _rte_content_helper(release: ReleaseEntity) -> dict:
    """
    File/FileSet/WebCapture sub-section of release_to_elasticsearch()

    The current priority order for "best_pdf_url" is:
    - internet archive urls (archive.org or web.archive.org)
    - other webarchive or repository URLs
    - any other URL
    """
    t = dict(
        file_count = len(release.files or []),
        fileset_count = len(release.filesets or []),
        webcapture_count = len(release.webcaptures or []),
    )

    any_pdf_url = None
    good_pdf_url = None
    best_pdf_url = None
    ia_pdf_url = None

    for f in release.files or []:
        if f.extra and f.extra.get('shadows'):
            t['in_shadows'] = True
        is_pdf = 'pdf' in (f.mimetype or '')
        for release_url in (f.urls or []):
            # first generic flags
            t.update(_rte_url_helper(release_url))

            # then PDF specific stuff (for generating "best URL" fields)
            if not f.mimetype and 'pdf' in release_url.url.lower():
                is_pdf = True
            if is_pdf:
                any_pdf_url = release_url.url
                if release_url.rel in ('webarchive', 'repository', 'repo'):
                    good_pdf_url = release_url.url
                if '//web.archive.org/' in release_url.url or '//archive.org/' in release_url.url:
                    best_pdf_url = release_url.url
                    ia_pdf_url = release_url.url

    # here is where we bake-in PDF url priority; IA-specific
    t['best_pdf_url'] = best_pdf_url or good_pdf_url or any_pdf_url
    t['ia_pdf_url'] = ia_pdf_url

    for fs in release.filesets or []:
        for url_obj in (fs.urls or []):
            t.update(_rte_url_helper(url_obj))

    for wc in release.webcaptures or []:
        for url_obj in (wc.archive_urls or []):
            t.update(_rte_url_helper(url_obj))

    return t

def _rte_url_helper(url_obj) -> dict:
    """
    Takes a location URL ('url' and 'rel' keys) and returns generic preservation status.

    Designed to work with file, webcapture, or fileset URLs.

    Returns a dict; should *not* include non-True values for any keys because
    these will be iteratively update() into the overal object.
    """
    t = dict()
    if url_obj.rel in ('webarchive', 'repository', 'archive', 'repo'):
        t['is_preserved'] = True
    if '//web.archive.org/' in url_obj.url or '//archive.org/' in url_obj.url:
        t['in_ia'] = True
    if url_obj.url.lower().startswith('http') or url_obj.url.lower().startswith('ftp'):
        t['in_web'] = True
    if url_obj.rel in ('dweb', 'p2p', 'ipfs', 'dat', 'torrent'):
        # not sure what rel will be for this stuff
        t['in_dweb'] = True
    if '//www.jstor.org/' in url_obj.url:
        t['in_jstor'] = True
    return t


def container_to_elasticsearch(entity, force_bool=True, stats=None):
    """
    Converts from an entity model/schema to elasticsearch oriented schema.

    Returns: dict
    Raises exception on error (never returns None)
    """

    if entity.state in ('redirect', 'deleted'):
        return dict(
            ident = entity.ident,
            state = entity.state,
        )
    elif entity.state != 'active':
        raise ValueError("Unhandled entity state: {}".format(entity.state))

    # First, the easy ones (direct copy)
    t = dict(
        doc_index_ts=datetime.datetime.utcnow().isoformat()+"Z",
        ident = entity.ident,
        state = entity.state,
        revision = entity.revision,

        name = entity.name,
        publisher = entity.publisher,
        container_type = entity.container_type,
        issnl = entity.issnl,
        wikidata_qid = entity.wikidata_qid,
    )

    if not entity.extra:
        entity.extra = dict()
    for key in ('country', 'languages', 'mimetypes', 'original_name',
                'first_year', 'last_year', 'aliases', 'abbrev', 'region',
                'discipline', 'publisher_type'):
        if entity.extra.get(key):
            t[key] = entity.extra[key]

    if entity.extra.get('dblp') and entity.extra['dblp'].get('prefix'):
        t['dblp_prefix'] = entity.extra['dblp']['prefix']

    if 'country' in t:
        t['country_code'] = t.pop('country')

    t['issns'] = []
    if entity.issnl:
        t['issns'].append(entity.issnl)
    for key in ('issnp', 'issne'):
        if entity.extra.get(key):
            t['issns'].append(entity.extra[key])

    in_doaj = None
    in_road = None
    is_oa = None
    is_longtail_oa = None
    any_kbart = None
    any_jstor = None
    any_ia_sim = None
    keepers = []

    extra = entity.extra
    if extra.get('doaj'):
        if extra['doaj'].get('as_of'):
            in_doaj = True
    if extra.get('road'):
        if extra['road'].get('as_of'):
            in_road = True
    if extra.get('szczepanski'):
        if extra['szczepanski'].get('as_of'):
            is_oa = True
    if extra.get('default_license'):
        if extra['default_license'].startswith('CC-'):
            is_oa = True
    t['sherpa_romeo_color'] = None
    if extra.get('sherpa_romeo'):
        t['sherpa_romeo_color'] = extra['sherpa_romeo'].get('color')
        if extra['sherpa_romeo'].get('color') == 'white':
            is_oa = False
    if extra.get('kbart'):
        any_kbart = True
        if extra['kbart'].get('jstor'):
            any_jstor = True
        for k, v in extra['kbart'].items():
            if v and isinstance(v, dict):
                keepers.append(k)
    if extra.get('ia'):
        if extra['ia'].get('sim'):
            any_ia_sim = True
        if extra['ia'].get('longtail_oa'):
            is_longtail_oa = True
    t['is_superceded'] = bool(extra.get('superceded'))

    t['keepers'] = keepers
    t['in_doaj'] = bool(in_doaj)
    t['in_road'] = bool(in_road)
    t['any_kbart'] = bool(any_kbart)
    if force_bool:
        t['is_oa'] = bool(in_doaj or in_road or is_oa)
        t['is_longtail_oa'] = bool(is_longtail_oa)
        t['any_jstor'] = bool(any_jstor)
        t['any_ia_sim'] = bool(any_ia_sim)
    else:
        t['is_oa'] = in_doaj or in_road or is_oa
        t['is_longtail_oa'] = is_longtail_oa
        t['any_jstor'] = any_jstor
        t['any_ia_sim'] = any_ia_sim

    # mix in stats, if provided
    if stats:
        t['releases_total'] = stats['total']
        t['preservation_bright'] = stats['preservation']['bright']
        t['preservation_dark'] = stats['preservation']['dark']
        t['preservation_shadows_only'] = stats['preservation']['shadows_only']
        t['preservation_none'] = stats['preservation']['none']
    return t


def _type_of_edit(edit):
    if edit.revision == None and edit.redirect_ident == None:
        return 'delete'
    elif edit.redirect_ident:
        # redirect
        return 'update'
    elif edit.prev_revision == None and edit.redirect_ident == None and edit.revision:
        return 'create'
    else:
        return 'update'


def changelog_to_elasticsearch(entity):
    """
    Note that this importer requires expanded fill info to work. Calling code
    may need to re-fetch editgroup from API to get the 'editor' field. Some of
    the old kafka feed content doesn't includes editor in particular.
    """

    editgroup = entity.editgroup
    t = dict(
        doc_index_ts=datetime.datetime.utcnow().isoformat()+"Z",
        index=entity.index,
        editgroup_id=entity.editgroup_id,
        timestamp=entity.timestamp.isoformat(),
        editor_id=editgroup.editor_id,
        username=editgroup.editor.username,
        is_bot=editgroup.editor.is_bot,
        is_admin=editgroup.editor.is_admin,
    )

    extra = editgroup.extra or dict()
    if extra.get('agent'):
        t['agent'] = extra['agent']

    containers = [_type_of_edit(e) for e in editgroup.edits.containers]
    creators = [_type_of_edit(e) for e in editgroup.edits.creators]
    files = [_type_of_edit(e) for e in editgroup.edits.files]
    filesets = [_type_of_edit(e) for e in editgroup.edits.filesets]
    webcaptures = [_type_of_edit(e) for e in editgroup.edits.webcaptures]
    releases = [_type_of_edit(e) for e in editgroup.edits.releases]
    works = [_type_of_edit(e) for e in editgroup.edits.works]

    t['containers'] = len(containers)
    t['new_containers'] = len([e for e in containers if e == 'create'])
    t['creators'] = len(creators)
    t['new_creators'] = len([e for e in creators if e == 'create'])
    t['files'] = len(files)
    t['new_files'] = len([e for e in files if e == 'create'])
    t['filesets'] = len(filesets)
    t['new_filesets'] = len([e for e in filesets if e == 'create'])
    t['webcaptures'] = len(webcaptures)
    t['new_webcaptures'] = len([e for e in webcaptures if e == 'create'])
    t['releases'] = len(releases)
    t['new_releases'] = len([e for e in releases if e == 'create'])
    t['works'] = len(works)
    t['new_works'] = len([e for e in works if e == 'create'])

    all_edits = containers + creators + files + filesets + webcaptures + releases + works

    t['created'] = len([e for e in all_edits if e == 'create'])
    t['updated'] = len([e for e in all_edits if e == 'update'])
    t['deleted'] = len([e for e in all_edits if e == 'delete'])
    t['total'] = len(all_edits)
    return t


def file_to_elasticsearch(entity):
    """
    Converts from an entity model/schema to elasticsearch oriented schema.

    Returns: dict
    Raises exception on error (never returns None)
    """

    if entity.state in ('redirect', 'deleted'):
        return dict(
            ident = entity.ident,
            state = entity.state,
        )
    elif entity.state != 'active':
        raise ValueError("Unhandled entity state: {}".format(entity.state))

    # First, the easy ones (direct copy)
    t = dict(
        doc_index_ts=datetime.datetime.utcnow().isoformat()+"Z",
        ident = entity.ident,
        state = entity.state,
        revision = entity.revision,
        release_ids = entity.release_ids,
        release_count = len(entity.release_ids),
        mimetype = entity.mimetype,
        size_bytes = entity.size,
        sha1 = entity.sha1,
        sha256 = entity.sha256,
        md5 = entity.md5,
    )

    parsed_urls = [tldextract.extract(u.url) for u in entity.urls]
    t['hosts'] = list(set(['.'.join([seg for seg in pu if seg]) for pu in parsed_urls]))
    t['domains'] = list(set([pu.registered_domain for pu in parsed_urls]))
    t['rels'] = list(set([u.rel for u in entity.urls]))

    t['in_ia'] = bool('archive.org' in t['domains'])
    t['in_ia_petabox'] = bool('archive.org' in t['hosts'])

    any_url = None
    good_url = None
    best_url = None
    for release_url in (entity.urls or []):
        any_url = release_url.url
        if release_url.rel in ('webarchive', 'repository'):
            good_url = release_url.url
        if '//web.archive.org/' in release_url.url or '//archive.org/' in release_url.url:
            best_url = release_url.url
    # here is where we bake-in priority; IA-specific
    t['best_url'] = best_url or good_url or any_url

    return t