aboutsummaryrefslogtreecommitdiffstats
path: root/fatcat_scholar/work_pipeline.py
blob: 720c696d959b8f97765f6dc02703ef4c36fb2d09 (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
import os
import io
import sys
import argparse
from typing import List, Dict, Tuple, Optional, Any, Sequence

import minio
import requests
import internetarchive
from dynaconf import settings
from fatcat_openapi_client import ReleaseEntity, FileEntity

from fatcat_scholar.api_entities import *
from fatcat_scholar.djvu import djvu_extract_leaf_texts
from fatcat_scholar.sandcrawler import (
    SandcrawlerPostgrestClient,
    SandcrawlerMinioClient,
)
from fatcat_scholar.issue_db import IssueDB, SimIssueRow, SimPubRow
from fatcat_scholar.schema import (
    DocType,
    IntermediateBundle,
)
from fatcat_scholar.sim_pipeline import truncate_pub_meta, truncate_issue_meta


def parse_pages(raw: str) -> Tuple[Optional[int], Optional[int]]:
    first_raw = raw.split("-")[0]
    if not first_raw.isdigit():
        return (None, None)
    first = int(first_raw)
    if not "-" in raw:
        return (first, first)
    last_raw = raw.split("-")[-1]
    if not last_raw.isdigit():
        return (first, first)
    last = int(last_raw)
    if last < first:
        last_munge = first_raw[0 : (len(first_raw) - len(last_raw))] + last_raw
        last = int(last_munge)
    if last < first:
        return (first, first)
    return (first, last)


def test_parse_pages() -> None:
    assert parse_pages("479-89") == (479, 489)
    assert parse_pages("466-7") == (466, 467)
    assert parse_pages("466-501") == (466, 501)
    assert parse_pages("466-401") == (466, 466)
    # TODO: or should it be (1, 1)?
    assert parse_pages("1") == (1, 1)
    # TODO: should we be doing strings instead of ints?
    assert parse_pages("iiv") == (None, None)


def fulltext_pref_list(releases: List[ReleaseEntity]) -> List[str]:
    """
    Returns a list of release idents in preference order (best first) to
    try and find fulltext for.
    """
    releases_sorted = sorted(
        releases,
        reverse=True,
        key=lambda r: (
            r.release_stage == "updated",
            r.release_stage == "published",
            r.volume is not None,
            r.container_id is not None,
            r.ext_ids.pmid is not None,
            r.release_stage == "submitted",
            r.release_type is not None,
            r.release_year,
            r.release_date,
            r.version,
        ),
    )
    return [r.ident for r in releases_sorted]


class WorkPipeline:
    def __init__(
        self,
        issue_db: IssueDB,
        sandcrawler_db_client: SandcrawlerPostgrestClient,
        sandcrawler_s3_client: SandcrawlerMinioClient,
    ):
        self.issue_db: IssueDB = issue_db
        self.ia_client = internetarchive.get_session()
        self.sandcrawler_db_client = sandcrawler_db_client
        self.sandcrawler_s3_client = sandcrawler_s3_client

    def fetch_file_grobid(self, fe: FileEntity, release_ident: str) -> Optional[Any]:
        """
        tei_xml: str
        release_ident: Optional[str]
        file_ident: Optional[str]
        """
        if not fe.sha1:
            return None
        if not fe.urls:
            return None
        grobid_meta = self.sandcrawler_db_client.get_grobid(fe.sha1)
        if not grobid_meta or grobid_meta["status"] != "success":
            return None
        # print(grobid_meta)
        try:
            grobid_xml = self.sandcrawler_s3_client.get_blob(
                bucket="sandcrawler",
                prefix="",
                folder="grobid",
                sha1hex=fe.sha1,
                extension=".tei.xml",
            )
            # print(grobid_xml)
        except minio.error.NoSuchKey:
            return None
        return dict(
            tei_xml=grobid_xml, release_ident=release_ident, file_ident=fe.ident,
        )

    def fetch_pdf_meta(
        self, fe: FileEntity, release_ident: str
    ) -> Optional[Dict[str, Any]]:
        """
        Fetches pdftext metadata from sandcrawler-db via postgrest HTTP
        interface.

        Returns a JSON object on success, or None if not found.

        raw_text: str
        release_ident: Optional[str]
        file_ident: Optional[str]
        """
        if not fe.sha1:
            return None
        pdf_meta = self.sandcrawler_db_client.get_pdf_meta(fe.sha1)
        if not pdf_meta or pdf_meta["status"] != "success":
            return None
        return dict(
            pdf_meta=pdf_meta, release_ident=release_ident, file_ident=fe.ident,
        )

    def fetch_file_pdftotext(self, fe: FileEntity, release_ident: str) -> Optional[Any]:
        """
        raw_text: str
        release_ident: Optional[str]
        file_ident: Optional[str]
        """
        if not fe.sha1:
            return None
        if not fe.urls:
            return None
        try:
            raw_text = self.sandcrawler_s3_client.get_blob(
                bucket="sandcrawler",
                prefix="",
                folder="text",
                sha1hex=fe.sha1,
                extension=".txt",
            )
            # print(raw_text)
        except minio.error.NoSuchKey:
            return None
        return dict(
            raw_text=raw_text, release_ident=release_ident, file_ident=fe.ident,
        )

    def lookup_sim(self, release: ReleaseEntity) -> Optional[SimIssueRow]:
        """
        Checks in IssueDB to see if this release is likely to have a copy in a
        SIM issue item.

        volume
        issue
        """
        if not (release.container_id and release.volume and release.issue):
            return None
        sim_pubid = self.issue_db.container2pubid(release.container_id)
        if not sim_pubid:
            return None

        return self.issue_db.lookup_issue(
            sim_pubid=sim_pubid, volume=release.volume, issue=release.issue
        )

    def fetch_sim(
        self,
        issue_db_row: SimIssueRow,
        issue_db_pub_row: SimPubRow,
        pages: str,
        release_ident: str,
    ) -> Optional[Any]:
        """
        issue_item
        pages: str
        page_texts: list
            page_num
            leaf_num
            raw_text
        release_ident: Optional[str]
        pub_item_metadata
        issue_item_metadata
        """

        first_page, last_page = parse_pages(pages)
        if first_page is None:
            return None

        # fetch full metadata from API
        issue_meta = self.ia_client.get_metadata(issue_db_row.issue_item)
        pub_meta = self.ia_client.get_metadata(issue_db_pub_row.pub_collection)

        leaf_index = dict()
        leaf_list = []
        if not "page_numbers" in issue_meta:
            # TODO: warn
            return None
        for entry in issue_meta["page_numbers"].get("pages", []):
            page_num = entry["pageNumber"]
            leaf_index[entry["leafNum"]] = page_num
            if not (page_num and page_num.isdigit()):
                continue
            page_num = int(page_num)
            if page_num >= first_page and page_num <= last_page:
                leaf_list.append(entry["leafNum"])

        if not leaf_list:
            return None

        page_texts: List[Dict[str, Any]] = []
        issue_item = self.ia_client.get_item(issue_db_row.issue_item)
        issue_item_djvu = issue_item.get_file(issue_db_row.issue_item + "_djvu.xml")

        # override 'close()' method so we can still read out contents
        djvu_bytes = io.BytesIO()
        djvu_bytes.close = lambda: None  # type: ignore
        assert issue_item_djvu.download(fileobj=djvu_bytes)
        djvu_bytes.seek(0)
        djvu_xml = io.StringIO(djvu_bytes.read().decode("UTF-8"))
        del djvu_bytes

        leaf_dict = djvu_extract_leaf_texts(djvu_xml, only_leaves=leaf_list)

        for leaf_num, raw_text in leaf_dict.items():
            page_texts.append(
                dict(
                    page_num=leaf_index.get(leaf_num),
                    leaf_num=leaf_num,
                    raw_text=raw_text,
                )
            )

        return dict(
            issue_item=issue_db_row.issue_item,
            pages=pages,
            page_texts=page_texts,
            release_ident=release_ident,
            pub_item_metadata=truncate_pub_meta(pub_meta),
            issue_item_metadata=truncate_issue_meta(issue_meta),
        )

    def process_release_list(self, releases: List[ReleaseEntity]) -> IntermediateBundle:
        """
        1. find best accessible fatcat file
            => fetch GROBID XML if available, else pdftotext if available
            => link to thumbnail if available
        2. find best SIM microfilm copy available
        """
        pref_idents = fulltext_pref_list(releases)
        release_dict = dict([(r.ident, r) for r in releases])

        # print(f"pref_idents={pref_idents}", file=sys.stderr)

        # find best accessible fatcat file
        grobid_fulltext: Optional[Any] = None
        pdf_meta: Optional[Any] = None
        pdftotext_fulltext: Optional[Any] = None
        for ident in pref_idents:
            release = release_dict[ident]
            if not release.files:
                continue
            for fe in release.files:
                if not fe.sha1 or fe.mimetype not in (None, "application/pdf"):
                    continue
                if not fe.urls:
                    continue
                grobid_fulltext = self.fetch_file_grobid(fe, ident)
                pdf_meta = self.fetch_pdf_meta(fe, ident)
                pdftotext_fulltext = None
                if pdf_meta and not grobid_fulltext:
                    pdftotext_fulltext = self.fetch_file_pdftotext(fe, ident)
                if grobid_fulltext or pdftotext_fulltext:
                    break
            if grobid_fulltext or pdftotext_fulltext:
                break

        # find best accessible SIM metadata and fulltext
        sim_fulltext: Optional[Any] = None
        sim_issue: Optional[Any] = None
        for ident in pref_idents:
            release = release_dict[ident]
            # print(f"{release.extra}\n{release.pages}", file=sys.stderr)
            if not release.pages:
                continue
            # TODO: in the future, will use release.extra.ia.sim.sim_pubid for lookup
            sim_issue = self.lookup_sim(release)
            # print(f"release_{release.ident}: sim_issue={sim_issue}", file=sys.stderr)
            if not sim_issue:
                continue
            sim_pub = self.issue_db.lookup_pub(sim_issue.sim_pubid)
            if not sim_pub:
                continue
            # XXX: control flow tweak?
            try:
                sim_fulltext = self.fetch_sim(
                    sim_issue, sim_pub, release.pages, release.ident
                )
            except requests.exceptions.ConnectionError as e:
                print(str(e), file=sys.stderr)
                continue
            except requests.exceptions.ReadTimeout as e:
                print(str(e), file=sys.stderr)
                continue
            if sim_fulltext:
                break

        return IntermediateBundle(
            doc_type=DocType.work,
            releases=releases,
            biblio_release_ident=pref_idents[0],
            grobid_fulltext=grobid_fulltext,
            pdftotext_fulltext=pdftotext_fulltext,
            pdf_meta=pdf_meta,
            sim_fulltext=sim_fulltext,
        )

    def run_releases(self, release_stream: Sequence[str]) -> None:
        """
        Iterates over the stream of releases, which are expected to be grouped
        (sorted) by work_ident.

        Collects releases under same work_ident into a batch and processes a
        work from that.

        TODO: what is the right API here? stream iterator? how should
        parallelism work?
        """
        batch = []
        batch_work_id = None
        for line in release_stream:
            if not line:
                continue
            release = entity_from_json(line, ReleaseEntity)
            if release.work_id == batch_work_id:
                batch.append(release)
                continue
            if batch:
                ib = self.process_release_list(batch)
                print(ib.json(exclude_none=True))
                batch_work_id = None
            batch = [
                release,
            ]
            batch_work_id = release.work_id

        if batch:
            ib = self.process_release_list(batch)
            print(ib.json(exclude_none=True, sort_keys=True))


def main() -> None:
    """
    Run this command like:

        python -m fatcat_scholar.work_pipeline
    """

    parser = argparse.ArgumentParser(
        formatter_class=argparse.ArgumentDefaultsHelpFormatter
    )
    subparsers = parser.add_subparsers()

    parser.add_argument(
        "--issue-db-file",
        help="sqlite3 database file to open",
        default="data/issue_db.sqlite",
        type=str,
    )
    parser.add_argument(
        "--sandcrawler-db-api",
        help="Sandcrawler Postgrest API endpoint",
        default=settings.SANDCRAWLER_DB_API,
        type=str,
    )
    parser.add_argument(
        "--sandcrawler-s3-api",
        help="Sandcrawler S3 (minio/seaweedfs) API endpoint",
        default=settings.SANDCRAWLER_S3_API,
        type=str,
    )

    sub = subparsers.add_parser(
        "run_releases", help="takes expanded release entity JSON, sorted by work_ident"
    )
    sub.set_defaults(func="run_releases")
    sub.add_argument(
        "json_file",
        help="release entities, as JSON-lines",
        nargs="?",
        default=sys.stdin,
        type=argparse.FileType("r"),
    )

    args = parser.parse_args()
    if not args.__dict__.get("func"):
        print("tell me what to do! (try --help)")
        sys.exit(-1)

    wp = WorkPipeline(
        issue_db=IssueDB(args.issue_db_file),
        sandcrawler_db_client=SandcrawlerPostgrestClient(
            api_url=args.sandcrawler_db_api
        ),
        sandcrawler_s3_client=SandcrawlerMinioClient(
            host_url=args.sandcrawler_s3_api,
            access_key=os.environ.get("MINIO_ACCESS_KEY"),
            secret_key=os.environ.get("MINIO_SECRET_KEY"),
        ),
    )

    if args.func == "run_releases":
        wp.run_releases(args.json_file)
    else:
        func = getattr(wp, args.func)
        func()


if __name__ == "__main__":
    main()