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
|
from typing import Any, Dict, List, Optional
import fatcat_openapi_client
from fatcat_openapi_client import ApiClient, ContainerEntity
from fatcat_tools.normal import clean_str
from .common import EntityImporter
class ChoculaImporter(EntityImporter):
"""
Creates or updates container entities based on output of "chocula" script,
which munges/processes journal metadata from several sources, including
fatcat itself.
See guide for details on the many 'extra' fields used here.
"""
def __init__(self, api: ApiClient, **kwargs) -> None:
eg_desc = kwargs.get(
"editgroup_description",
"Automated import of container-level metadata from Chocula tool.",
)
eg_extra = kwargs.get("editgroup_extra", dict())
eg_extra["agent"] = eg_extra.get("agent", "fatcat_tools.ChoculaImporter")
super().__init__(api, editgroup_description=eg_desc, editgroup_extra=eg_extra, **kwargs)
def want(self, raw_record: Any) -> bool:
if not raw_record.get("ident") and not raw_record.get("_known_issnl"):
self.counts["skip-unknown-new-issnl"] += 1
return False
if raw_record.get("issnl") and raw_record.get("name"):
return True
return False
def parse_record(self, row: Dict[str, Any]) -> Optional[ContainerEntity]:
"""
row is a python dict (parsed from JSON).
returns a ContainerEntity (or None if invalid or couldn't parse)
"""
name = clean_str(row.get("name"))
if not name:
# Name is required (by schema)
return None
name = name.strip()
if name.endswith(", Proceedings of the"):
name = "Proceedings of the " + name.split(",")[0]
if name.endswith("."):
name = name[:-1]
extra = dict()
for k in (
"urls",
"webarchive_urls",
"country",
"sherpa_romeo",
"ezb",
"szczepanski",
"doaj",
"languages",
"ia",
"scielo",
"kbart",
"publisher_type",
"platform",
):
if row["extra"].get(k):
extra[k] = row["extra"][k]
container_type = None
if "proceedings" in name.lower():
container_type = "proceedings"
elif "journal " in name.lower():
container_type = "journal"
ce = ContainerEntity(
issnl=row["issnl"],
issnp=row["extra"].get("issnp"),
issne=row["extra"].get("issne"),
ident=row["ident"],
name=name,
container_type=container_type,
publisher=clean_str(row.get("publisher")),
wikidata_qid=row.get("wikidata_qid"),
extra=extra,
)
return ce
def try_update(self, ce: ContainerEntity) -> bool:
existing = None
if ce.ident:
try:
existing = self.api.get_container(ce.ident)
except fatcat_openapi_client.rest.ApiException as err:
if err.status != 404:
raise err
self.counts["exists"] += 1
self.counts["exists-not-found"] += 1
return False
if existing.state != "active":
self.counts["exists"] += 1
self.counts["exists-inactive"] += 1
return False
if not existing:
# check if existing by ISSN-L
try:
existing = self.api.lookup_container(issnl=ce.issnl)
except fatcat_openapi_client.rest.ApiException as err:
if err.status != 404:
raise err
if existing:
self.counts["exists"] += 1
self.counts["exists-by-issnl"] += 1
return False
# doesn't exist, always create
return True
# decide whether to update
do_update = False
if not self.do_updates:
self.counts["exists"] += 1
return False
if not existing.extra:
existing.extra = dict()
if ce.extra.get("urls") and set(ce.extra.get("urls", [])) != set(
existing.extra.get("urls", [])
):
do_update = True
if ce.extra.get("webarchive_urls") and set(ce.extra.get("webarchive_urls", [])) != set(
existing.extra.get("webarchive_urls", [])
):
do_update = True
for k in ("ezb", "szczepanski", "publisher_type", "platform"):
if ce.extra.get(k) and not existing.extra.get(k):
do_update = True
for k in ("kbart", "ia", "doaj"):
# always update these fields if not equal (chocula override)
if ce.extra.get(k) and ce.extra[k] != existing.extra.get(k):
do_update = True
if ce.publisher and not existing.publisher:
do_update = True
if ce.wikidata_qid and not existing.wikidata_qid:
do_update = True
if do_update:
existing.wikidata_qid = existing.wikidata_qid or ce.wikidata_qid
existing.publisher = existing.publisher or ce.publisher
existing.container_type = existing.container_type or ce.container_type
existing.issne = existing.issne or ce.issne
existing.issnp = existing.issnp or ce.issnp
for k in ("urls", "webarchive_urls"):
# be conservative about URL updates; don't clobber existing URL lists
# may want to make this behavior more sophisticated in the
# future, or at least a config flag
if ce.extra.get(k) and not existing.extra.get(k):
existing.extra[k] = ce.extra.get(k, [])
for k in (
"sherpa_romeo",
"ezb",
"szczepanski",
"doaj",
"ia",
"scielo",
"kbart",
"publisher_type",
"platform",
):
# always update (chocula over-rides)
if ce.extra.get(k):
existing.extra[k] = ce.extra[k]
for k in ("country",):
# only include if not set (don't clobber human edits)
if ce.extra.get(k) and not existing.extra.get(k):
existing.extra[k] = ce.extra[k]
if ce.extra.get("languages"):
if not existing.extra.get("languages"):
existing.extra["languages"] = ce.extra["languages"]
elif not ce.extra["languages"][0] in existing.extra["languages"]:
existing.extra["languages"].append(ce.extra["languages"][0])
self.api.update_container(self.get_editgroup_id(), existing.ident, existing)
self.counts["update"] += 1
return False
else:
self.counts["exists"] += 1
self.counts["exists-skip-update"] += 1
return False
# if we got this far, it's a bug
raise NotImplementedError
def insert_batch(self, batch: List[ContainerEntity]) -> None:
self.api.create_container_auto_batch(
fatcat_openapi_client.ContainerAutoBatch(
editgroup=fatcat_openapi_client.Editgroup(
description=self.editgroup_description, extra=self.editgroup_extra
),
entity_list=batch,
)
)
|