aboutsummaryrefslogtreecommitdiffstats
path: root/python/fatcat_tools/importers/common.py
blob: 282f775c5ff54ad06ab9b7019cfd8f2661ef27c2 (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

import re
import sys
import csv
import json
import ftfy
import sqlite3
import itertools
import subprocess
from collections import Counter
import pykafka

import fatcat_client
from fatcat_client.rest import ApiException


SANE_MAX_RELEASES = 200
SANE_MAX_URLS = 100

def clean(thing, force_xml=False):
    """
    This function is appropriate to be called on any random, non-markup string,
    such as author names, titles, etc.

    It will try to clean up commong unicode mangles, HTML characters, etc.

    This will detect XML/HTML and "do the right thing" (aka, not remove
    entities like '&amp' if there are tags in the string), unless you pass the
    'force_xml' parameter, which might be appropriate for, eg, names and
    titles, which generally should be projected down to plain text.

    Also strips extra whitespace.
    """
    if not thing:
        return None
    fix_entities = 'auto'
    if force_xml:
        fix_entities = True
    fixed = ftfy.fix_text(thing, fix_entities=fix_entities).strip()
    if not fixed or len(fixed) <= 1:
        # wasn't zero-length before, but is now; return None
        return None
    return fixed

def test_clean():

    assert clean(None) == None
    assert clean('') == None
    assert clean('1') == None
    assert clean('123') == '123'
    assert clean('a&amp;b') == 'a&b'
    assert clean('<b>a&amp;b</b>') == '<b>a&amp;b</b>'
    assert clean('<b>a&amp;b</b>', force_xml=True) == '<b>a&b</b>'

DOMAIN_REL_MAP = {
    "archive.org": "archive",
    # LOCKSS, Portico, DuraSpace, etc would also be "archive"

    "arxiv.org": "repository",
    "babel.hathitrust.org": "repository",
    "cds.cern.ch": "repository",
    "deepblue.lib.umich.edu": "repository",
    "europepmc.org": "repository",
    "hal.inria.fr": "repository",
    "scielo.isciii.es": "repository",
    "www.dtic.mil": "repository",
    "www.jstage.jst.go.jp": "repository",
    "www.jstor.org": "repository",
    "www.ncbi.nlm.nih.gov": "repository",
    "ftp.ncbi.nlm.nih.gov": "repository",
    "www.scielo.br": "repository",
    "www.scielo.cl": "repository",
    "www.scielo.org.mx": "repository",
    "zenodo.org": "repository",

    "citeseerx.ist.psu.edu": "aggregator",
    "publisher-connector.core.ac.uk": "aggregator",
    "core.ac.uk": "aggregator",
    "static.aminer.org": "aggregator",
    "aminer.org": "aggregator",
    "pdfs.semanticscholar.org": "aggregator",
    "semanticscholar.org": "aggregator",
    "www.semanticscholar.org": "aggregator",

    "academic.oup.com": "publisher",
    "cdn.elifesciences.org": "publisher",
    "cell.com": "publisher",
    "dl.acm.org": "publisher",
    "downloads.hindawi.com": "publisher",
    "elifesciences.org": "publisher",
    "iopscience.iop.org": "publisher",
    "journals.plos.org": "publisher",
    "link.springer.com": "publisher",
    "onlinelibrary.wiley.com": "publisher",
    "works.bepress.com": "publisher",
    "www.biomedcentral.com": "publisher",
    "www.cell.com": "publisher",
    "www.nature.com": "publisher",
    "www.pnas.org": "publisher",
    "www.tandfonline.com": "publisher",

    "www.researchgate.net": "social",
    "academia.edu": "social",

    "wayback.archive-it.org": "webarchive",
    "web.archive.org": "webarchive",
    "archive.is": "webarchive",
}

def make_rel_url(raw_url, default_link_rel="web"):
    # this is where we map specific domains to rel types, and also filter out
    # bad domains, invalid URLs, etc
    rel = default_link_rel
    for domain, domain_rel in DOMAIN_REL_MAP.items():
        if "//{}/".format(domain) in raw_url:
            rel = domain_rel
            break
    return (rel, raw_url)

def test_make_rel_url():
    assert make_rel_url("http://example.com/thing.pdf")[0] == "web"
    assert make_rel_url("http://example.com/thing.pdf", default_link_rel="jeans")[0] == "jeans"
    assert make_rel_url("https://web.archive.org/web/*/http://example.com/thing.pdf")[0] == "webarchive"
    assert make_rel_url("http://cell.com/thing.pdf")[0] == "publisher"

class EntityImporter:
    """
    Base class for fatcat entity importers.

    The API exposed to record iterator is:

        push_record(raw_record)
        finish()

    The API that implementations are expected to fill in are:

        want(raw_record) -> boolean
        parse(raw_record) -> entity
        try_update(entity) -> boolean
        insert_batch([entity]) -> None

    This class exposes helpers for implementations:

        self.api
        self.get_editgroup_id()
        self.create_<entity>(entity) -> EntityEdit
            for related entity types
        self.push_entity(entity)
        self.counts['exists'] += 1
            if didn't update or insert because of existing)
        self.counts['update'] += 1
            if updated an entity
    """

    def __init__(self, api, **kwargs):

        eg_extra = kwargs.get('editgroup_extra', dict())
        eg_extra['git_rev'] = eg_extra.get('git_rev',
            subprocess.check_output(["git", "describe", "--always"]).strip()).decode('utf-8')
        eg_extra['agent'] = eg_extra.get('agent', 'fatcat_tools.EntityImporter')

        self.api = api
        self.bezerk_mode = kwargs.get('bezerk_mode', False)
        self.edit_batch_size = kwargs.get('edit_batch_size', 100)
        self.editgroup_description = kwargs.get('editgroup_description')
        self.editgroup_extra = kwargs.get('editgroup_extra')
        self.reset()

        self._issnl_id_map = dict()
        self._orcid_id_map = dict()
        self._orcid_regex = re.compile("^\\d{4}-\\d{4}-\\d{4}-\\d{3}[\\dX]$")
        self._doi_id_map = dict()

    def reset(self):
        self.counts = Counter({'skip': 0, 'insert': 0, 'update': 0, 'exists': 0})
        self._edit_count = 0
        self._editgroup_id = None
        self._entity_queue = []
        self._edits_inflight = []

    def push_record(self, raw_record):
        """
        Returns nothing.
        """
        if (not raw_record) or (not self.want(raw_record)):
            self.counts['skip'] += 1
            return
        entity = self.parse_record(raw_record)
        if not entity:
            self.counts['skip'] += 1
            return
        if self.bezerk_mode:
            self.push_entity(entity)
            return
        if self.try_update(entity):
            self.push_entity(entity)
        return

    def parse_record(self, raw_record):
        # implementations should fill this in
        raise NotImplementedError

    def finish(self):
        if self._edit_count > 0:
            self.api.accept_editgroup(self._editgroup_id)
            self._editgroup_id = None
            self._edit_count = 0
            self._edits_inflight = []

        if self._entity_queue:
            self.insert_batch(self._entity_queue)
            self.counts['insert'] += len(self._entity_queue)
            self._entity_queue =  []

        self.counts['total'] = 0
        for key in ('skip', 'insert', 'update', 'exists'):
            self.counts['total'] += self.counts[key]
        return self.counts

    def get_editgroup_id(self, edits=1):
        if self._edit_count >= self.edit_batch_size:
            self.api.accept_editgroup(self._editgroup_id)
            self._editgroup_id = None
            self._edit_count = 0
            self._edits_inflight = []

        if not self._editgroup_id:
            eg = self.api.create_editgroup(
                fatcat_client.Editgroup(
                    description=self.editgroup_description,
                    extra=self.editgroup_extra))
            self._editgroup_id = eg.editgroup_id

        self._edit_count += edits
        return self._editgroup_id

    def create_container(self, entity):
        eg_id = self.get_editgroup_id()
        self.counts['inserted.container'] += 1
        return self.api.create_container(eg_id, entity)

    def create_release(self, entity):
        eg_id = self.get_editgroup_id()
        self.counts['inserted.release'] += 1
        return self.api.create_release(eg_id, entity)

    def create_file(self, entity):
        eg_id = self.get_editgroup_id()
        self.counts['inserted.file'] += 1
        return self.api.create_file(eg_id, entity)

    def updated(self):
        """
        Implementations should call this from try_update() if the update was successful
        """
        self.counts['update'] += 1

    def push_entity(self, entity):
        self._entity_queue.append(entity)
        if len(self._entity_queue) >= self.edit_batch_size:
            self.insert_batch(self._entity_queue)
            self.counts['insert'] += len(self._entity_queue)
            self._entity_queue = []

    def want(self, raw_record):
        """
        Implementations can override for optional fast-path to drop a record.
        Must have no side-effects; returns bool.
        """
        return True

    def parse(self, raw_record):
        """
        Returns an entity class type, or None if we should skip this one.

        May have side-effects (eg, create related entities), but shouldn't
        update/mutate the actual entity.
        """
        raise NotImplementedError

    def try_update(self, raw_record):
        """
        Passed the output of parse(). Should try to find an existing entity and
        update it (PUT), decide we should do nothing (based on the existing
        record), or create a new one.

        Implementations must update the exists/updated/skip counts
        appropriately in this method.

        Returns boolean: True if the entity should still be inserted, False otherwise
        """
        raise NotImplementedError

    def insert_batch(self, raw_record):
        raise NotImplementedError

    def is_orcid(self, orcid):
        return self._orcid_regex.match(orcid) is not None

    def lookup_orcid(self, orcid):
        """Caches calls to the Orcid lookup API endpoint in a local dict"""
        if not self.is_orcid(orcid):
            return None
        if orcid in self._orcid_id_map:
            return self._orcid_id_map[orcid]
        creator_id = None
        try:
            rv = self.api.lookup_creator(orcid=orcid)
            creator_id = rv.ident
        except ApiException as ae:
            # If anything other than a 404 (not found), something is wrong
            assert ae.status == 404
        self._orcid_id_map[orcid] = creator_id # might be None
        return creator_id

    def is_doi(self, doi):
        return doi.startswith("10.") and doi.count("/") >= 1

    def lookup_doi(self, doi):
        """Caches calls to the doi lookup API endpoint in a local dict"""
        assert self.is_doi(doi)
        doi = doi.lower()
        if doi in self._doi_id_map:
            return self._doi_id_map[doi]
        release_id = None
        try:
            rv = self.api.lookup_release(doi=doi)
            release_id = rv.ident
        except ApiException as ae:
            # If anything other than a 404 (not found), something is wrong
            assert ae.status == 404
        self._doi_id_map[doi] = release_id # might be None
        return release_id

    def is_issnl(self, issnl):
        return len(issnl) == 9 and issnl[4] == '-'

    def lookup_issnl(self, issnl):
        """Caches calls to the ISSN-L lookup API endpoint in a local dict"""
        if issnl in self._issnl_id_map:
            return self._issnl_id_map[issnl]
        container_id = None
        try:
            rv = self.api.lookup_container(issnl=issnl)
            container_id = rv.ident
        except ApiException as ae:
            # If anything other than a 404 (not found), something is wrong
            assert ae.status == 404
        self._issnl_id_map[issnl] = container_id # might be None
        return container_id

    def read_issn_map_file(self, issn_map_file):
        print("Loading ISSN map file...")
        self._issn_issnl_map = dict()
        for line in issn_map_file:
            if line.startswith("ISSN") or len(line) == 0:
                continue
            (issn, issnl) = line.split()[0:2]
            self._issn_issnl_map[issn] = issnl
            # double mapping makes lookups easy
            self._issn_issnl_map[issnl] = issnl
        print("Got {} ISSN-L mappings.".format(len(self._issn_issnl_map)))

    def issn2issnl(self, issn):
        if issn is None:
            return None
        return self._issn_issnl_map.get(issn)


class RecordPusher:
    """
    Base class for different importer sources. Pretty trivial interface, just
    wraps an importer and pushes records in to it.
    """

    def __init__(self, importer, **kwargs):
        self.importer = importer

    def run(self):
        """
        This will look something like:

            for line in sys.stdin:
                record = json.loads(line)
                self.importer.push_record(record)
            print(self.importer.finish())
        """
        raise NotImplementedError


class JsonLinePusher(RecordPusher):

    def __init__(self, importer, json_file, **kwargs):
        self.importer = importer
        self.json_file = json_file

    def run(self):
        for line in self.json_file:
            if not line:
                continue
            record = json.loads(line)
            self.importer.push_record(record)
        counts = self.importer.finish()
        print(counts)
        return counts


class CsvPusher(RecordPusher):

    def __init__(self, importer, csv_file, **kwargs):
        self.importer = importer
        self.reader = csv.DictReader(csv_file, delimiter=kwargs.get('delimiter', ','))

    def run(self):
        for line in self.reader:
            if not line:
                continue
            self.importer.push_record(line)
        counts = self.importer.finish()
        print(counts)
        return counts


class LinePusher(RecordPusher):

    def __init__(self, importer, text_file, **kwargs):
        self.importer = importer
        self.text_file = text_file

    def run(self):
        for line in self.text_file:
            if not line:
                continue
            self.importer.push_record(line)
        counts = self.importer.finish()
        print(counts)
        return counts


class SqlitePusher(RecordPusher):

    def __init__(self, importer, db_file, table_name, where_clause="", **kwargs):
        self.importer = importer
        self.db = sqlite3.connect(db_file, isolation_level='EXCLUSIVE')
        self.db.row_factory = sqlite3.Row
        self.table_name = table_name
        self.where_clause = where_clause

    def run(self):
        cur = self.db.execute("SELECT * FROM {} {};".format(
            self.table_name, self.where_clause))
        for row in cur:
            self.importer.push_record(row)
        counts = self.importer.finish()
        print(counts)
        return counts


class KafkaJsonPusher(RecordPusher):

    def __init__(self, importer, kafka_hosts, kafka_env, topic_suffix, group, **kwargs):
        self.importer = importer
        self.consumer = make_kafka_consumer(
            kafka_hosts,
            kafka_env,
            topic_suffix,
            group,
        )

    def run(self):
        count = 0
        for msg in self.consumer:
            if not msg:
                continue
            record = json.loads(msg.value.decode('utf-8'))
            self.importer.push_record(record)
            count += 1
            if count % 500 == 0:
                print("Import counts: {}".format(self.importer.counts))
        # TODO: should catch UNIX signals (HUP?) to shutdown cleanly, and/or
        # commit the current batch if it has been lingering
        counts = self.importer.finish()
        print(counts)
        return counts


def make_kafka_consumer(hosts, env, topic_suffix, group):
    topic_name = "fatcat-{}.{}".format(env, topic_suffix).encode('utf-8')
    client = pykafka.KafkaClient(hosts=hosts, broker_version="1.0.0")
    consume_topic = client.topics[topic_name]
    print("Consuming from kafka topic {}, group {}".format(topic_name, group))

    consumer = consume_topic.get_balanced_consumer(
        consumer_group=group.encode('utf-8'),
        managed=True,
        auto_commit_enable=True,
        auto_commit_interval_ms=30000, # 30 seconds
        compacted_topic=True,
    )
    return consumer