Skip to content

Commit 2371ca9

Browse files
committed
oddpub and grobid/bibr
1 parent 8ba3f8e commit 2371ca9

3 files changed

Lines changed: 169 additions & 0 deletions

File tree

metacheck_convert.json

Lines changed: 27 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,27 @@
1+
[
2+
{
3+
"id": "Grobid FULL 0.9.0",
4+
"service": "grobid",
5+
"version": "0.9.0",
6+
"url": "https://grobidOrg-grobid.hf.space"
7+
},
8+
{
9+
"id": "Grobid CRF 0.9.0",
10+
"service": "grobid",
11+
"version": "0.9.0",
12+
"url": "https://grobidOrg-grobid-crf.hf.space"
13+
},
14+
{
15+
"id": "Metacheck grobid 0.9.0",
16+
"service": "grobid",
17+
"version": "0.9.0",
18+
"url": "https://grobid.metacheck.app"
19+
},
20+
{
21+
"id": "Platform Bibr 10.0",
22+
"service": "bibr",
23+
"version": "10.0",
24+
"url": "https://platform.metacheck.app"
25+
"api_key": "SCIVRS_API_KEY"
26+
}
27+
]

oddpub.json

Lines changed: 83 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,83 @@
1+
{
2+
"$schema": "https://json-schema.org/draft/2020-12/schema",
3+
"$id": "https://scienceverse.org/schema/oddpub.json",
4+
"title": "Open Practices via ODDPub",
5+
"description": "Open data and code detection via ODDPub (https://github.com/quest-bih/oddpub). This describes individual papers -- papers may have more than one open data or code source, which are described together here with arrays for open_data_category and by combining sentences for das/cas/open_data_statements/open_code_statements.",
6+
7+
"type": "object",
8+
9+
"properties": {
10+
"article": {
11+
"description": "A reference to the paper being processed",
12+
"type": ["string"]
13+
},
14+
"is_open_data": {
15+
"description": "Is open data detected",
16+
"type": ["boolean"]
17+
},
18+
"open_data_category": {
19+
"description": "What type of data are available",
20+
"type": "array",
21+
"items": {
22+
"type": "string",
23+
"enum": [
24+
"field-specific repository",
25+
"general-purpose repository",
26+
"dataset",
27+
"supplemental table or data",
28+
"supplementary raw/full data with specific file format",
29+
"dataset on Github",
30+
"upon request",
31+
"re-use",
32+
"unknown/misspecified source"
33+
]
34+
}
35+
},
36+
"is_reuse": {
37+
"description": "Is data reused from another source",
38+
"type": ["boolean"]
39+
},
40+
"is_open_code": {
41+
"description": "Is open code detected",
42+
"type": ["boolean"]
43+
},
44+
"is_code_supplement": {
45+
"description": "Is there code in supplemental materials",
46+
"type": ["boolean"]
47+
},
48+
"is_code_reuse": {
49+
"description": "Is code reused from another source",
50+
"type": ["boolean"]
51+
},
52+
"is_open_data_das": {
53+
"description": "Is there a data availability statement indicating open data",
54+
"type": ["boolean"]
55+
},
56+
"is_open_code_cas": {
57+
"description": "Is there a code availability statement indicating open code",
58+
"type": ["boolean"]
59+
},
60+
"das": {
61+
"description": "The data availability statement",
62+
"type": ["string"]
63+
},
64+
"open_data_statements": {
65+
"description": "Any statements about open data in the full tet",
66+
"type": ["string"]
67+
},
68+
"cas": {
69+
"description": "The code availability statement",
70+
"type": ["string"]
71+
},
72+
"open_code_statements": {
73+
"description": "Any statements about open code in the full text",
74+
"type": ["string"]
75+
}
76+
},
77+
78+
"required": [
79+
"article"
80+
],
81+
82+
"additionalProperties": false
83+
}

open_practices.json

Lines changed: 59 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,59 @@
1+
{
2+
"$schema": "https://json-schema.org/draft/2020-12/schema",
3+
"$id": "https://scienceverse.org/schema/open_practices.json",
4+
"title": "Open Practices via MetaCheck",
5+
"description": "Open practices detection via MetaCheck. This describes individual practices -- papers may have more than one practice, described by an array of open_practices objects.",
6+
"type": "object",
7+
"properties": {
8+
"paper_id": {
9+
"description": "A unique ID for each paper",
10+
"type": ["string"]
11+
},
12+
"practice": {
13+
"description": "The practice type",
14+
"type": ["string"],
15+
"enum": [
16+
"data",
17+
"code",
18+
"materials",
19+
"registration",
20+
"other"
21+
]
22+
}
23+
"is_open": {
24+
"description": "Is the practice open",
25+
"type": ["boolean"]
26+
},
27+
"category": {
28+
"description": "The category of supporting materials",
29+
"type": "array",
30+
"items": {
31+
"type": ["string", "null"],
32+
"enum": [
33+
"field-specific repository",
34+
"general-purpose repository",
35+
"supplementary",
36+
"re-use",
37+
"website",
38+
"upon request",
39+
"unknown"
40+
]
41+
}
42+
},
43+
"location": {
44+
"description": "The location of supporting materials (e.g., a URL)",
45+
"type": ["string", "null"]
46+
},
47+
"text_id": {
48+
"description": "The text IDs of any supporting statements in the full text",
49+
"type": "array",
50+
"items": { "type": "integer" }
51+
}
52+
},
53+
"required": [
54+
"paper_id",
55+
"practice",
56+
"is_open"
57+
],
58+
"additionalProperties": false
59+
}

0 commit comments

Comments
 (0)