summaryrefslogtreecommitdiffstats
path: root/python/fatcat_tools/importers/matched.py
blob: 0b77bcf0de08ffa9f7fe5b4f7267e15b62ad8b58 (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

import sys
import json
import sqlite3
import itertools
import fatcat_client
from .common import FatcatImporter

#row = row.split('\t')
#assert len(row) == 2
#sha1 = row[0].replace('sha1:')
#sha1 = base64.b16encode(base64.b32decode(sha1)).lower()
#print(sha1)
#dois = [d.lower() for d in json.loads(row[1])]

class MatchedImporter(FatcatImporter):
    """
    Importer for "file to crossref DOI" matches.

    These matches are currently generated by Internet Archive hadoop jobs
    written in scala (part of the 'sandcrawler' repo/project), but could be
    generated by other parties as well.

    Input format is JSON with keys:
    - dois (list)
    - sha1 (hex)
    - md5 (hex)
    - sha256 (hex)
    - size (int)
    - cdx (list of objects)
        - dt
        - url
    - mimetype
    - urls (list of strings... or objects?)

    Future handlings/extensions:
    - core_id, wikidata_id, pmcid, pmid: not as lists
    """

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

        eg_desc = kwargs.get('editgroup_description',
            "Import of large-scale file-to-release match results. Source of metadata varies.")
        eg_extra = kwargs.get('editgroup_extra', dict())
        eg_extra['agent'] = eg_extra.get('agent', 'fatcat_tools.MatchedImporter')
        super().__init__(api,
            editgroup_description=eg_desc,
            editgroup_extra=eg_extra)
        self.default_link_rel = kwargs.get("default_link_rel", "web")
        self.default_mime = kwargs.get("default_mime", None)
        self.skip_file_updates = kwargs.get("skip_file_updates", False)

    def make_url(self, raw):
        rel = self.default_link_rel
        # TODO: this is where we could map specific domains to rel types,
        # and also filter out bad domains, invalid URLs, etc
        if "//archive.org/" in raw or "//arxiv.org/" in raw:
            # TODO: special-case the arxiv.org bulk mirror?
            rel = "repository"
        elif "//web.archive.org/" in raw or "//archive.is/" in raw:
            rel = "webarchive"
        return fatcat_client.FileEntityUrls(url=raw, rel=rel)

    def parse_matched_dict(self, obj):
        sha1 = obj['sha1']
        dois = [d.lower() for d in obj.get('dois', [])]

        # lookup sha1, or create new entity
        fe = None
        if not self.skip_file_updates:
            try:
                fe = self.api.lookup_file(sha1=sha1)
            except fatcat_client.rest.ApiException as err:
                if err.status != 404:
                    raise err
        if fe is None:
            fe = fatcat_client.FileEntity(
                sha1=sha1,
                release_ids=[],
                urls=[],
            )

        # lookup dois
        re_list = set()
        for doi in dois:
            try:
                re = self.api.lookup_release(doi=doi)
            except fatcat_client.rest.ApiException as err:
                if err.status != 404:
                    raise err
                re = None
            if re is None:
                print("DOI not found: {}".format(doi))
            else:
                re_list.add(re.ident)
        if len(re_list) == 0:
            return None
        if fe.release_ids == set(re_list):
            return None
        re_list.update(fe.release_ids)
        fe.release_ids = list(re_list)

        # parse URLs and CDX
        existing_urls = [feu.url for feu in fe.urls]
        for url in obj.get('url', []):
            if url not in existing_urls:
                url = self.make_url(url)
                if url != None:
                    fe.urls.append(url)
        for cdx in obj.get('cdx', []):
            original = cdx['url']
            wayback = "https://web.archive.org/web/{}/{}".format(
                cdx['dt'],
                original)
            if wayback not in existing_urls:
                fe.urls.append(
                    fatcat_client.FileEntityUrls(url=wayback, rel="webarchive"))
            if original not in existing_urls:
                url = self.make_url(original)
                if url != None:
                    fe.urls.append(url)

        if obj.get('size') != None:
            fe.size = int(obj['size'])
        fe.sha256 = obj.get('sha256', fe.sha256)
        fe.md5 = obj.get('md5', fe.sha256)
        if obj.get('mimetype') is None:
            if fe.mimetype is None:
                fe.mimetype = self.default_mime
        else:
            fe.mimetype = obj.get('mimetype')
        return fe

    def create_row(self, row, editgroup_id=None):
        obj = json.loads(row)
        fe = self.parse_matched_dict(obj)
        if fe is not None:
            if fe.ident is None:
                self.api.create_file(fe, editgroup_id=editgroup_id)
                self.counts['insert'] += 1
            else:
                self.api.update_file(fe.ident, fe, editgroup_id=editgroup_id)
                self.counts['update'] += 1

    def create_batch(self, batch, editgroup_id=None):
        """Reads and processes in batches (not API-call-per-line)"""
        objects = [self.parse_matched_dict(json.loads(l))
                   for l in batch if l != None]
        new_objects = [o for o in objects if o != None and o.ident == None]
        update_objects = [o for o in objects if o != None and o.ident != None]
        for obj in update_objects:
            self.api.update_file(obj.ident, obj, editgroup_id=editgroup_id)
        if len(new_objects) > 0:
            self.api.create_file_batch(new_objects, autoaccept="true", editgroup_id=editgroup_id)
        self.counts['update'] += len(update_objects)
        self.counts['insert'] += len(new_objects)