aboutsummaryrefslogtreecommitdiffstats
path: root/python/sandcrawler/persist.py
blob: ee153ab21aa6857f48f5ec06e9e528ada6c10c6b (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

"""
cdx
- read raw CDX, filter
- push to SQL table

ingest-file-result
- read JSON format (batch)
- cdx SQL push batch (on conflict skip)
- file_meta SQL push batch (on conflict update)
- ingest request push batch (on conflict skip)
- ingest result push batch (on conflict update)

grobid
- reads JSON format (batch)
- grobid2json
- minio push (one-by-one)
- grobid SQL push batch (on conflict update)
- file_meta SQL push batch (on conflict update)
"""

import os
from typing import Optional, AnyStr
import xml.etree.ElementTree

from sandcrawler.workers import SandcrawlerWorker
from sandcrawler.db import SandcrawlerPostgresClient
from sandcrawler.minio import SandcrawlerMinioClient
from sandcrawler.grobid import GrobidClient
from sandcrawler.pdfextract import PdfExtractResult
from sandcrawler.ingest_html import HtmlMetaRow


class PersistCdxWorker(SandcrawlerWorker):

    def __init__(self, db_url, **kwargs):
        super().__init__()
        self.db = SandcrawlerPostgresClient(db_url)
        self.cur = self.db.conn.cursor()

    def process(self, record, key=None):
        """
        Only do batches (as transactions)
        """
        raise NotImplementedError

    def push_batch(self, batch):
        self.counts['total'] += len(batch)
        # filter to full CDX lines, no liveweb
        cdx_batch = [r for r in batch if r.get('warc_path') and ("/" in r['warc_path'])]
        resp = self.db.insert_cdx(self.cur, cdx_batch)
        if len(cdx_batch) < len(batch):
            self.counts['skip'] += len(batch) - len(cdx_batch)
        self.counts['insert-cdx'] += resp[0]
        self.counts['update-cdx'] += resp[1]
        self.db.commit()
        return []

class PersistIngestFileResultWorker(SandcrawlerWorker):

    def __init__(self, db_url, **kwargs):
        super().__init__()
        self.db = SandcrawlerPostgresClient(db_url)
        self.cur = self.db.conn.cursor()

    def process(self, record, key=None):
        """
        Only do batches (as transactions)
        """
        raise NotImplementedError

    def request_to_row(self, raw):
        """
        Converts ingest-request JSON schema (eg, from Kafka) to SQL ingest_request schema

        if there is a problem with conversion, return None
        """
        # backwards compat hacks; transform request to look like current schema
        if raw.get('ingest_type') == 'file':
            raw['ingest_type'] = 'pdf'
        if (not raw.get('link_source')
                and raw.get('base_url')
                and raw.get('ext_ids', {}).get('doi')
                and raw['base_url'] == "https://doi.org/{}".format(raw['ext_ids']['doi'])):
            # set link_source(_id) for old ingest requests
            raw['link_source'] = 'doi'
            raw['link_source_id'] = raw['ext_ids']['doi']
        if (not raw.get('link_source')
                and raw.get('ingest_request_source', '').startswith('savepapernow')
                and raw.get('fatcat', {}).get('release_ident')):
            # set link_source(_id) for old ingest requests
            raw['link_source'] = 'spn'
            raw['link_source_id'] = raw['fatcat']['release_ident']

        for k in ('ingest_type', 'base_url', 'link_source', 'link_source_id'):
            if not k in raw:
                self.counts['skip-request-fields'] += 1
                return None
        if raw['ingest_type'] not in ('pdf', 'xml', 'html'):
            self.counts['skip-ingest-type'] += 1
            return None
        request = {
            'ingest_type': raw['ingest_type'],
            'base_url': raw['base_url'],
            'link_source': raw['link_source'],
            'link_source_id': raw['link_source_id'],
            'ingest_request_source': raw.get('ingest_request_source'),
            'request': {},
        }
        # extra/optional fields
        if raw.get('release_stage'):
            request['release_stage'] = raw['release_stage']
        if raw.get('fatcat', {}).get('release_ident'):
            request['request']['release_ident'] = raw['fatcat']['release_ident']
        for k in ('ext_ids', 'edit_extra', 'rel'):
            if raw.get(k):
                request['request'][k] = raw[k]
        # if this dict is empty, trim it to save DB space
        if not request['request']:
            request['request'] = None
        return request
            

    def file_result_to_row(self, raw: dict) -> Optional[dict]:
        """
        Converts ingest-result JSON schema (eg, from Kafka) to SQL ingest_file_result schema

        if there is a problem with conversion, return None and set skip count
        """
        for k in ('request', 'hit', 'status'):
            if not k in raw:
                self.counts['skip-result-fields'] += 1
                return None
        if not 'base_url' in raw['request']:
            self.counts['skip-result-fields'] += 1
            return None
        ingest_type = raw['request'].get('ingest_type')
        if ingest_type == 'file':
            ingest_type = 'pdf'
        if ingest_type not in ('pdf', 'xml', 'html'):
            self.counts['skip-ingest-type'] += 1
            return None
        if raw['status'] in ("existing", ):
            self.counts['skip-existing'] += 1
            return None
        result = {
            'ingest_type': ingest_type,
            'base_url': raw['request']['base_url'],
            'hit': raw['hit'],
            'status': raw['status'],
        }
        terminal = raw.get('terminal')
        if terminal:
            result['terminal_url'] = terminal.get('terminal_url') or terminal.get('url')
            result['terminal_dt'] = terminal.get('terminal_dt')
            result['terminal_status_code'] = terminal.get('terminal_status_code') or terminal.get('status_code') or terminal.get('http_code')
            if result['terminal_status_code']:
                result['terminal_status_code'] = int(result['terminal_status_code'])
            result['terminal_sha1hex'] = terminal.get('terminal_sha1hex')
            if len(result['terminal_url']) > 2048:
                # postgresql13 doesn't like extremely large URLs in b-tree index
                self.counts['skip-huge-url'] += 1
                return None
        return result

    def result_to_html_meta(self, record: dict) -> Optional[HtmlMetaRow]:
        html_body = record.get('html_body')
        file_meta = record.get('file_meta')
        if not (file_meta and html_body):
            return None
        return HtmlMetaRow(
            sha1hex=file_meta["sha1hex"],
            status=record.get('status'),
            scope=record.get('scope'),
            has_teixml=bool(html_body and html_body['status'] == 'success'),
            has_thumbnail=False,  # TODO
            word_count=(html_body and html_body.get('word_count')) or None,
            biblio=record.get('html_biblio'),
            resources=record.get('html_resources'),
        )

    def push_batch(self, batch):
        self.counts['total'] += len(batch)

        if not batch:
            return []

        results = [self.file_result_to_row(raw) for raw in batch]
        results = [r for r in results if r]

        requests = [self.request_to_row(raw['request']) for raw in batch if raw.get('request')]
        requests = [r for r in requests if r]

        if requests:
            resp = self.db.insert_ingest_request(self.cur, requests)
            self.counts['insert-requests'] += resp[0]
            self.counts['update-requests'] += resp[1]
        if results:
            resp = self.db.insert_ingest_file_result(self.cur, results, on_conflict="update")
            self.counts['insert-results'] += resp[0]
            self.counts['update-results'] += resp[1]

        # these schemas match, so can just pass through
        cdx_batch = [r['cdx'] for r in batch if r.get('hit') and r.get('cdx')]
        revisit_cdx_batch = [r['revisit_cdx'] for r in batch if r.get('hit') and r.get('revisit_cdx')]
        cdx_batch.extend(revisit_cdx_batch)
        # filter to full CDX lines, with full warc_paths (not liveweb)
        cdx_batch = [r for r in cdx_batch if r.get('warc_path') and ("/" in r['warc_path'])]
        if cdx_batch:
            resp = self.db.insert_cdx(self.cur, cdx_batch)
            self.counts['insert-cdx'] += resp[0]
            self.counts['update-cdx'] += resp[1]

        file_meta_batch = [r['file_meta'] for r in batch if r.get('hit') and r.get('file_meta')]
        if file_meta_batch:
            resp = self.db.insert_file_meta(self.cur, file_meta_batch, on_conflict="nothing")
            self.counts['insert-file_meta'] += resp[0]
            self.counts['update-file_meta'] += resp[1]

        html_meta_batch = [self.result_to_html_meta(r) for r in batch if r.get('hit') and r.get('html_body')]
        if html_meta_batch:
            resp = self.db.insert_html_meta(self.cur, html_meta_batch, on_conflict="update")
            self.counts['insert-html_meta'] += resp[0]
            self.counts['update-html_meta'] += resp[1]

        self.db.commit()
        return []

class PersistIngestRequestWorker(PersistIngestFileResultWorker):

    def __init__(self, db_url, **kwargs):
        super().__init__(db_url=db_url)

    def process(self, record, key=None):
        """
        Only do batches (as transactions)
        """
        raise NotImplementedError

    def push_batch(self, batch):
        self.counts['total'] += len(batch)

        if not batch:
            return []

        requests = [self.request_to_row(raw) for raw in batch]
        requests = [r for r in requests if r]

        if requests:
            resp = self.db.insert_ingest_request(self.cur, requests)
            self.counts['insert-requests'] += resp[0]
            self.counts['update-requests'] += resp[1]

        self.db.commit()
        return []

class PersistGrobidWorker(SandcrawlerWorker):

    def __init__(self, db_url, **kwargs):
        super().__init__()
        self.grobid = GrobidClient()
        self.s3 = SandcrawlerMinioClient(
            host_url=kwargs.get('s3_url', 'localhost:9000'),
            access_key=kwargs['s3_access_key'],
            secret_key=kwargs['s3_secret_key'],
            default_bucket=kwargs['s3_bucket'],
        )
        self.s3_only = kwargs.get('s3_only', False)
        self.db_only = kwargs.get('db_only', False)
        assert not (self.s3_only and self.db_only), "Only one of s3_only and db_only allowed"
        if not self.s3_only:
            self.db = SandcrawlerPostgresClient(db_url)
            self.cur = self.db.conn.cursor()
        else:
            self.db = None
            self.cur = None

    def process(self, record, key=None):
        """
        Only do batches (as transactions)
        """
        raise NotImplementedError

    def push_batch(self, batch):
        self.counts['total'] += len(batch)

        # filter out bad "missing status_code" timeout rows
        missing = [r for r in batch if not r.get('status_code')]
        if missing:
            self.counts['skip-missing-status'] += len(missing)
            batch = [r for r in batch if r.get('status_code')]

        for r in batch:
            if r['status_code'] != 200 or not r.get('tei_xml'):
                self.counts['s3-skip-status'] += 1
                if r.get('error_msg'):
                    r['metadata'] = {'error_msg': r['error_msg'][:500]}
                continue

            assert len(r['key']) == 40
            if not self.db_only:
                resp = self.s3.put_blob(
                    folder="grobid",
                    blob=r['tei_xml'],
                    sha1hex=r['key'],
                    extension=".tei.xml",
                )
                self.counts['s3-put'] += 1

            # enhance with teixml2json metadata, if available
            try:
                metadata = self.grobid.metadata(r)
            except xml.etree.ElementTree.ParseError as xml_e:
                r['status'] = 'bad-grobid-xml'
                r['metadata'] = {'error_msg': str(xml_e)[:1024]}
                continue
            if not metadata:
                continue
            for k in ('fatcat_release', 'grobid_version'):
                r[k] = metadata.pop(k, None)
            if r.get('fatcat_release'):
                r['fatcat_release'] = r['fatcat_release'].replace('release_', '')
            if metadata.get('grobid_timestamp'):
                r['updated'] = metadata['grobid_timestamp']
            r['metadata'] = metadata

        if not self.s3_only:
            resp = self.db.insert_grobid(self.cur, batch, on_conflict="update")
            self.counts['insert-grobid'] += resp[0]
            self.counts['update-grobid'] += resp[1]

            file_meta_batch = [r['file_meta'] for r in batch if r.get('file_meta')]
            resp = self.db.insert_file_meta(self.cur, file_meta_batch, on_conflict="update")
            self.counts['insert-file-meta'] += resp[0]
            self.counts['update-file-meta'] += resp[1]

            self.db.commit()

        return []


class PersistGrobidDiskWorker(SandcrawlerWorker):
    """
    Writes blobs out to disk.

    This could be refactored into a "Sink" type with an even thinner wrapper.
    """

    def __init__(self, output_dir):
        super().__init__()
        self.output_dir = output_dir

    def _blob_path(self, sha1hex, extension=".tei.xml"):
        obj_path = "{}/{}/{}{}".format(
            sha1hex[0:2],
            sha1hex[2:4],
            sha1hex,
            extension,
        )
        return obj_path

    def process(self, record, key=None):

        if record.get('status_code') != 200 or not record.get('tei_xml'):
            return False
        assert(len(record['key'])) == 40
        p = "{}/{}".format(self.output_dir, self._blob_path(record['key']))
        os.makedirs(os.path.dirname(p), exist_ok=True)
        with open(p, 'w') as f:
            f.write(record.pop('tei_xml'))
        self.counts['written'] += 1
        return record


class PersistPdfTrioWorker(SandcrawlerWorker):

    def __init__(self, db_url, **kwargs):
        super().__init__()
        self.db = SandcrawlerPostgresClient(db_url)
        self.cur = self.db.conn.cursor()

    def process(self, record, key=None):
        """
        Only do batches (as transactions)
        """
        raise NotImplementedError

    def push_batch(self, batch):
        self.counts['total'] += len(batch)

        batch = [r for r in batch if 'pdf_trio' in r and r['pdf_trio'].get('status_code')]
        for r in batch:
            # copy key (sha1hex) into sub-object
            r['pdf_trio']['key'] = r['key']
        pdftrio_batch = [r['pdf_trio'] for r in batch]
        resp = self.db.insert_pdftrio(self.cur, pdftrio_batch, on_conflict="update")
        self.counts['insert-pdftrio'] += resp[0]
        self.counts['update-pdftrio'] += resp[1]

        file_meta_batch = [r['file_meta'] for r in batch if r['pdf_trio']['status'] == "success" and r.get('file_meta')]
        resp = self.db.insert_file_meta(self.cur, file_meta_batch)
        self.counts['insert-file-meta'] += resp[0]
        self.counts['update-file-meta'] += resp[1]

        self.db.commit()
        return []


class PersistPdfTextWorker(SandcrawlerWorker):
    """
    Pushes text file to blob store (S3/seaweed/minio) and PDF metadata to SQL table.

    Should keep batch sizes small.
    """

    def __init__(self, db_url, **kwargs):
        super().__init__()
        self.s3 = SandcrawlerMinioClient(
            host_url=kwargs.get('s3_url', 'localhost:9000'),
            access_key=kwargs['s3_access_key'],
            secret_key=kwargs['s3_secret_key'],
            default_bucket=kwargs['s3_bucket'],
        )
        self.s3_only = kwargs.get('s3_only', False)
        self.db_only = kwargs.get('db_only', False)
        assert not (self.s3_only and self.db_only), "Only one of s3_only and db_only allowed"
        if not self.s3_only:
            self.db = SandcrawlerPostgresClient(db_url)
            self.cur = self.db.conn.cursor()
        else:
            self.db = None
            self.cur = None

    def process(self, record, key=None):
        """
        Only do batches (as transactions)
        """
        raise NotImplementedError

    def push_batch(self, batch):
        self.counts['total'] += len(batch)

        parsed_batch = []
        for r in batch:
            parsed_batch.append(PdfExtractResult.from_pdftext_dict(r))

        for r in parsed_batch:
            if r.status != 'success' or not r.text:
                self.counts['s3-skip-status'] += 1
                if r.error_msg:
                    r.metadata = {'error_msg': r.error_msg[:500]}
                continue

            assert len(r.sha1hex) == 40
            if not self.db_only:
                resp = self.s3.put_blob(
                    folder="text",
                    blob=r.text,
                    sha1hex=r.sha1hex,
                    extension=".txt",
                )
                self.counts['s3-put'] += 1

        if not self.s3_only:
            resp = self.db.insert_pdf_meta(self.cur, parsed_batch, on_conflict="update")
            self.counts['insert-pdf-meta'] += resp[0]
            self.counts['update-pdf-meta'] += resp[1]

            file_meta_batch = [r.file_meta for r in parsed_batch if r.file_meta]
            resp = self.db.insert_file_meta(self.cur, file_meta_batch, on_conflict="update")
            self.counts['insert-file-meta'] += resp[0]
            self.counts['update-file-meta'] += resp[1]

            self.db.commit()

        return []


class PersistThumbnailWorker(SandcrawlerWorker):
    """
    Pushes text file to blob store (S3/seaweed/minio) and PDF metadata to SQL
    table.

    This worker *must* be used with raw kakfa mode; thumbnails are *not*
    wrapped in JSON like most sandcrawler kafka messages.
    """

    def __init__(self, **kwargs):
        super().__init__()
        self.s3 = SandcrawlerMinioClient(
            host_url=kwargs.get('s3_url', 'localhost:9000'),
            access_key=kwargs['s3_access_key'],
            secret_key=kwargs['s3_secret_key'],
            default_bucket=kwargs['s3_bucket'],
        )
        self.s3_extension = kwargs.get('s3_extension', ".jpg")
        self.s3_folder = kwargs.get('s3_folder', "pdf")

    def process(self, blob: bytes, key: Optional[str] = None):
        """
        Processing raw messages, not decoded JSON objects
        """

        if isinstance(key, bytes):
            key = key.decode('utf-8')
        assert key is not None and len(key) == 40 and isinstance(key, str)
        assert isinstance(blob, bytes)
        assert len(blob) >= 50

        resp = self.s3.put_blob(
            folder=self.s3_folder,
            blob=blob,
            sha1hex=key,
            extension=self.s3_extension,
        )
        self.counts['s3-put'] += 1


class GenericPersistDocWorker(SandcrawlerWorker):
    """
    Pushes blobs from Kafka to S3.

    Objects are assumed to be JSON-wrapped strings.
    """

    def __init__(self, **kwargs):
        super().__init__()
        self.s3 = SandcrawlerMinioClient(
            host_url=kwargs.get('s3_url', 'localhost:9000'),
            access_key=kwargs['s3_access_key'],
            secret_key=kwargs['s3_secret_key'],
            default_bucket=kwargs['s3_bucket'],
        )
        self.s3_extension = kwargs.get('s3_extension', ".unknown")
        self.s3_folder = kwargs.get('s3_folder', "unknown")
        self.doc_key = "unknown"

    def process(self, record: dict, key: Optional[AnyStr] = None) -> None:

        if record.get('status') != 'success' or not record.get(self.doc_key):
            return

        assert key is not None
        if isinstance(key, bytes):
            key_str = key.decode('utf-8')
        elif isinstance(key, str):
            key_str = key
        assert len(key_str) == 40
        if 'sha1hex' in record:
            assert key_str == record['sha1hex']

        resp = self.s3.put_blob(
            folder=self.s3_folder,
            blob=record[self.doc_key].encode('utf-8'),
            sha1hex=key_str,
            extension=self.s3_extension,
        )
        self.counts['s3-put'] += 1


class PersistXmlDocWorker(GenericPersistDocWorker):
    """
    Pushes TEI-XML file to blob store (S3/seaweed/minio). Does not talk to
    sandcrawler database (SQL).
    """

    def __init__(self, **kwargs):
        super().__init__(**kwargs)
        self.s3_extension = kwargs.get('s3_extension', ".jats.xml")
        self.s3_folder = kwargs.get('s3_folder', "xml_doc")
        self.doc_key = "jats_xml"


class PersistHtmlTeiXmlWorker(GenericPersistDocWorker):
    """
    Pushes TEI-XML file to blob store (S3/seaweed/minio). Does not talk to
    sandcrawler database (SQL).
    """

    def __init__(self, **kwargs):
        super().__init__(**kwargs)
        self.s3_extension = kwargs.get('s3_extension', ".tei.xml")
        self.s3_folder = kwargs.get('s3_folder', "html_body")
        self.doc_key = "tei_xml"