bolognese 1.0.12 → 1.0.13
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +4 -4
- data/Gemfile.lock +1 -1
- data/lib/bolognese/readers/datacite_reader.rb +23 -9
- data/lib/bolognese/utils.rb +5 -1
- data/lib/bolognese/version.rb +1 -1
- data/spec/fixtures/datacite-multiple-rights.xml +59 -0
- data/spec/fixtures/datacite_schema_3.xml +58 -0
- data/spec/readers/datacite_reader_spec.rb +34 -2
- data/spec/utils_spec.rb +1 -1
- metadata +4 -2
checksums.yaml
CHANGED
|
@@ -1,7 +1,7 @@
|
|
|
1
1
|
---
|
|
2
2
|
SHA256:
|
|
3
|
-
metadata.gz:
|
|
4
|
-
data.tar.gz:
|
|
3
|
+
metadata.gz: 1bc5619b1c9dc7435a062e1e9e143c190594087d80e7cb7dac27df816804cd88
|
|
4
|
+
data.tar.gz: 0bc05be427ae2be8b3e9adcebd2eb25d2660409425da9d15c98c319f936fa3a0
|
|
5
5
|
SHA512:
|
|
6
|
-
metadata.gz:
|
|
7
|
-
data.tar.gz:
|
|
6
|
+
metadata.gz: 9115ce5569d2d3a3e2a9b189fa0e2901bb72837704ff8174b742ed05cd4cbb8ad84a7a79e8f0e7635469f0aec480e750d823a80f8f0fe57750dc69de27db54cd
|
|
7
|
+
data.tar.gz: 9238067f30a00dba50cd61a22b4c64b9d3e04f155a2531a3edb2b58cda1d906741bf664f019e4029dbe9e221f881c21e2d0a8e9d6dcc1ceb8a1f59fff6c4abbc
|
data/Gemfile.lock
CHANGED
|
@@ -85,24 +85,38 @@ module Bolognese
|
|
|
85
85
|
}.compact
|
|
86
86
|
|
|
87
87
|
titles = Array.wrap(meta.dig("titles", "title")).map do |r|
|
|
88
|
-
if r.
|
|
88
|
+
if r.blank?
|
|
89
|
+
nil
|
|
90
|
+
elsif r.is_a?(String)
|
|
89
91
|
{ "title" => sanitize(r) }
|
|
90
92
|
else
|
|
91
93
|
{ "title" => sanitize(r["__content__"]), "titleType" => r["titleType"], "lang" => r["lang"] }.compact
|
|
92
94
|
end
|
|
93
|
-
end
|
|
95
|
+
end.compact
|
|
94
96
|
|
|
95
97
|
alternate_identifiers = Array.wrap(meta.dig("alternateIdentifiers", "alternateIdentifier")).map do |r|
|
|
96
98
|
{ "alternateIdentifierType" => r["alternateIdentifierType"], "alternateIdentifier" => r["__content__"].presence }.compact
|
|
97
99
|
end.compact
|
|
98
|
-
descriptions = Array.wrap(meta.dig("descriptions", "description")).
|
|
99
|
-
|
|
100
|
-
|
|
100
|
+
descriptions = Array.wrap(meta.dig("descriptions", "description")).map do |r|
|
|
101
|
+
if r.blank?
|
|
102
|
+
nil
|
|
103
|
+
elsif r.is_a?(String)
|
|
104
|
+
{ "description" => sanitize(r), "descriptionType" => "Abstract" }
|
|
105
|
+
elsif r.is_a?(Hash)
|
|
106
|
+
{ "description" => sanitize(r["__content__"]), "descriptionType" => r["descriptionType"], "lang" => r["lang"] }.compact
|
|
107
|
+
end
|
|
108
|
+
end.compact
|
|
101
109
|
rights_list = Array.wrap(meta.dig("rightsList", "rights")).map do |r|
|
|
102
|
-
|
|
103
|
-
|
|
110
|
+
if r.blank?
|
|
111
|
+
nil
|
|
112
|
+
elsif r.is_a?(String)
|
|
113
|
+
{ "rights" => r }
|
|
114
|
+
elsif r.is_a?(Hash)
|
|
115
|
+
{ "rights" => r["__content__"], "rightsUri" => normalize_url(r["rightsURI"]), "lang" => r["lang"] }.compact
|
|
116
|
+
end
|
|
117
|
+
end.compact
|
|
104
118
|
subjects = Array.wrap(meta.dig("subjects", "subject")).map do |k|
|
|
105
|
-
if k.
|
|
119
|
+
if k.blank?
|
|
106
120
|
nil
|
|
107
121
|
elsif k.is_a?(String)
|
|
108
122
|
{ "subject" => sanitize(k) }
|
|
@@ -116,7 +130,7 @@ module Bolognese
|
|
|
116
130
|
"dateType" => parse_attributes(d, content: "dateType"),
|
|
117
131
|
"dateInformation" => parse_attributes(d, content: "dateInformation")
|
|
118
132
|
}.compact
|
|
119
|
-
end
|
|
133
|
+
end.compact
|
|
120
134
|
dates << { "date" => meta.fetch("publicationYear", nil), "dateType" => "Issued" } if meta.fetch("publicationYear", nil).present? && get_date(dates, "Issued").blank?
|
|
121
135
|
sizes = Array.wrap(meta.dig("sizes", "size"))
|
|
122
136
|
formats = Array.wrap(meta.dig("formats", "format"))
|
data/lib/bolognese/utils.rb
CHANGED
|
@@ -455,9 +455,13 @@ module Bolognese
|
|
|
455
455
|
def normalize_url(id)
|
|
456
456
|
return nil unless id.present?
|
|
457
457
|
|
|
458
|
+
# handle info URIs
|
|
459
|
+
return id if id.to_s.start_with?("info")
|
|
460
|
+
|
|
458
461
|
# check for valid HTTP uri
|
|
459
462
|
uri = Addressable::URI.parse(id)
|
|
460
|
-
|
|
463
|
+
|
|
464
|
+
return nil unless uri && uri.host && %w(http https ftp).include?(uri.scheme)
|
|
461
465
|
|
|
462
466
|
# clean up URL
|
|
463
467
|
PostRank::URI.clean(id)
|
data/lib/bolognese/version.rb
CHANGED
|
@@ -0,0 +1,59 @@
|
|
|
1
|
+
<?xml version="1.0" encoding="UTF-8"?>
|
|
2
|
+
<resource xmlns="http://datacite.org/schema/kernel-3" xmlns:dim="http://www.dspace.org/xmlns/dspace/dim" xmlns:dryad="http://purl.org/dryad/terms/" xmlns:dspace="http://www.dspace.org/xmlns/dspace/dim" xmlns:mets="http://www.loc.gov/METS/" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://datacite.org/schema/kernel-3 http://schema.datacite.org/meta/kernel-3/metadata.xsd">
|
|
3
|
+
<identifier identifierType="DOI">10.5061/DRYAD.8515</identifier>
|
|
4
|
+
<version>1</version>
|
|
5
|
+
<creators>
|
|
6
|
+
<creator>
|
|
7
|
+
<creatorName>Ollomo, Benjamin</creatorName>
|
|
8
|
+
</creator>
|
|
9
|
+
<creator>
|
|
10
|
+
<creatorName>Durand, Patrick</creatorName>
|
|
11
|
+
</creator>
|
|
12
|
+
<creator>
|
|
13
|
+
<creatorName>Prugnolle, Franck</creatorName>
|
|
14
|
+
</creator>
|
|
15
|
+
<creator>
|
|
16
|
+
<creatorName>Douzery, Emmanuel J. P.</creatorName>
|
|
17
|
+
</creator>
|
|
18
|
+
<creator>
|
|
19
|
+
<creatorName>Arnathau, Céline</creatorName>
|
|
20
|
+
</creator>
|
|
21
|
+
<creator>
|
|
22
|
+
<creatorName>Nkoghe, Dieudonné</creatorName>
|
|
23
|
+
</creator>
|
|
24
|
+
<creator>
|
|
25
|
+
<creatorName>Leroy, Eric</creatorName>
|
|
26
|
+
</creator>
|
|
27
|
+
<creator>
|
|
28
|
+
<creatorName>Renaud, François</creatorName>
|
|
29
|
+
</creator>
|
|
30
|
+
</creators>
|
|
31
|
+
<titles>
|
|
32
|
+
<title>Data from: A new malaria agent in African hominids.</title>
|
|
33
|
+
</titles>
|
|
34
|
+
<publisher>Dryad Digital Repository</publisher>
|
|
35
|
+
<publicationYear>2011</publicationYear>
|
|
36
|
+
<subjects>
|
|
37
|
+
<subject>Phylogeny</subject>
|
|
38
|
+
<subject>Malaria</subject>
|
|
39
|
+
<subject>Parasites</subject>
|
|
40
|
+
<subject>Taxonomy</subject>
|
|
41
|
+
<subject>Mitochondrial genome</subject>
|
|
42
|
+
<subject>Africa</subject>
|
|
43
|
+
<subject>Plasmodium</subject>
|
|
44
|
+
</subjects>
|
|
45
|
+
<resourceType resourceTypeGeneral="Dataset">DataPackage</resourceType>
|
|
46
|
+
<alternateIdentifiers>
|
|
47
|
+
<alternateIdentifier alternateIdentifierType="citation">Ollomo B, Durand P, Prugnolle F, Douzery EJP, Arnathau C, Nkoghe D, Leroy E, Renaud F (2009) A new malaria agent in African hominids. PLoS Pathogens 5(5): e1000446.</alternateIdentifier>
|
|
48
|
+
</alternateIdentifiers>
|
|
49
|
+
<relatedIdentifiers>
|
|
50
|
+
<relatedIdentifier relatedIdentifierType="DOI" relationType="HasPart">10.5061/DRYAD.8515/1</relatedIdentifier>
|
|
51
|
+
<relatedIdentifier relatedIdentifierType="DOI" relationType="HasPart">10.5061/DRYAD.8515/2</relatedIdentifier>
|
|
52
|
+
<relatedIdentifier relatedIdentifierType="DOI" relationType="IsReferencedBy">10.1371/JOURNAL.PPAT.1000446</relatedIdentifier>
|
|
53
|
+
<relatedIdentifier relatedIdentifierType="PMID" relationType="IsReferencedBy">19478877</relatedIdentifier>
|
|
54
|
+
</relatedIdentifiers>
|
|
55
|
+
<rightsList>
|
|
56
|
+
<rights>info:eu-repo/semantics/openAccess</rights>
|
|
57
|
+
<rights rightsURI="info:eu-repo/semantics/openAccess">Open Access</rights>
|
|
58
|
+
</rightsList>
|
|
59
|
+
</resource>
|
|
@@ -0,0 +1,58 @@
|
|
|
1
|
+
<?xml version="1.0" encoding="UTF-8"?>
|
|
2
|
+
<resource xmlns="http://datacite.org/schema/kernel-3" xmlns:dim="http://www.dspace.org/xmlns/dspace/dim" xmlns:dryad="http://purl.org/dryad/terms/" xmlns:dspace="http://www.dspace.org/xmlns/dspace/dim" xmlns:mets="http://www.loc.gov/METS/" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://datacite.org/schema/kernel-3 http://schema.datacite.org/meta/kernel-3/metadata.xsd">
|
|
3
|
+
<identifier identifierType="DOI">10.5061/DRYAD.8515</identifier>
|
|
4
|
+
<version>1</version>
|
|
5
|
+
<creators>
|
|
6
|
+
<creator>
|
|
7
|
+
<creatorName>Ollomo, Benjamin</creatorName>
|
|
8
|
+
</creator>
|
|
9
|
+
<creator>
|
|
10
|
+
<creatorName>Durand, Patrick</creatorName>
|
|
11
|
+
</creator>
|
|
12
|
+
<creator>
|
|
13
|
+
<creatorName>Prugnolle, Franck</creatorName>
|
|
14
|
+
</creator>
|
|
15
|
+
<creator>
|
|
16
|
+
<creatorName>Douzery, Emmanuel J. P.</creatorName>
|
|
17
|
+
</creator>
|
|
18
|
+
<creator>
|
|
19
|
+
<creatorName>Arnathau, Céline</creatorName>
|
|
20
|
+
</creator>
|
|
21
|
+
<creator>
|
|
22
|
+
<creatorName>Nkoghe, Dieudonné</creatorName>
|
|
23
|
+
</creator>
|
|
24
|
+
<creator>
|
|
25
|
+
<creatorName>Leroy, Eric</creatorName>
|
|
26
|
+
</creator>
|
|
27
|
+
<creator>
|
|
28
|
+
<creatorName>Renaud, François</creatorName>
|
|
29
|
+
</creator>
|
|
30
|
+
</creators>
|
|
31
|
+
<titles>
|
|
32
|
+
<title>Data from: A new malaria agent in African hominids.</title>
|
|
33
|
+
</titles>
|
|
34
|
+
<publisher>Dryad Digital Repository</publisher>
|
|
35
|
+
<publicationYear>2011</publicationYear>
|
|
36
|
+
<subjects>
|
|
37
|
+
<subject>Phylogeny</subject>
|
|
38
|
+
<subject>Malaria</subject>
|
|
39
|
+
<subject>Parasites</subject>
|
|
40
|
+
<subject>Taxonomy</subject>
|
|
41
|
+
<subject>Mitochondrial genome</subject>
|
|
42
|
+
<subject>Africa</subject>
|
|
43
|
+
<subject>Plasmodium</subject>
|
|
44
|
+
</subjects>
|
|
45
|
+
<resourceType resourceTypeGeneral="Dataset">DataPackage</resourceType>
|
|
46
|
+
<alternateIdentifiers>
|
|
47
|
+
<alternateIdentifier alternateIdentifierType="citation">Ollomo B, Durand P, Prugnolle F, Douzery EJP, Arnathau C, Nkoghe D, Leroy E, Renaud F (2009) A new malaria agent in African hominids. PLoS Pathogens 5(5): e1000446.</alternateIdentifier>
|
|
48
|
+
</alternateIdentifiers>
|
|
49
|
+
<relatedIdentifiers>
|
|
50
|
+
<relatedIdentifier relatedIdentifierType="DOI" relationType="HasPart">10.5061/DRYAD.8515/1</relatedIdentifier>
|
|
51
|
+
<relatedIdentifier relatedIdentifierType="DOI" relationType="HasPart">10.5061/DRYAD.8515/2</relatedIdentifier>
|
|
52
|
+
<relatedIdentifier relatedIdentifierType="DOI" relationType="IsReferencedBy">10.1371/JOURNAL.PPAT.1000446</relatedIdentifier>
|
|
53
|
+
<relatedIdentifier relatedIdentifierType="PMID" relationType="IsReferencedBy">19478877</relatedIdentifier>
|
|
54
|
+
</relatedIdentifiers>
|
|
55
|
+
<rightsList>
|
|
56
|
+
<rights rightsURI="http://creativecommons.org/publicdomain/zero/1.0/" />
|
|
57
|
+
</rightsList>
|
|
58
|
+
</resource>
|
|
@@ -111,7 +111,7 @@ describe Bolognese::Metadata, vcr: true do
|
|
|
111
111
|
expect(subject.creator).to eq([{"type"=>"Person", "name"=>"Kristian Garza", "givenName"=>"Kristian", "familyName"=>"Garza"}])
|
|
112
112
|
expect(subject.titles).to eq([{"title"=>"Analysis Tools For Crossover Experiment Of Ui Using Choice Architecture"}])
|
|
113
113
|
expect(subject.descriptions.first["description"]).to start_with("This tools are used to analyse the data produced by the Crosssover Experiment")
|
|
114
|
-
expect(subject.rights_list).to eq([{"rightsUri"=>"https://creativecommons.org/licenses/by-nc-sa/4.0",
|
|
114
|
+
expect(subject.rights_list).to eq([{"rights"=>"Creative Commons Attribution-NonCommercial-ShareAlike", "rightsUri"=>"https://creativecommons.org/licenses/by-nc-sa/4.0"},{"rights"=>"Open Access", "rightsUri"=>"info:eu-repo/semantics/openAccess"}])
|
|
115
115
|
expect(subject.dates).to eq([{"date"=>"2016-03-27", "dateType"=>"Issued"}])
|
|
116
116
|
expect(subject.publication_year).to eq("2016")
|
|
117
117
|
expect(subject.related_identifiers.length).to eq(1)
|
|
@@ -418,6 +418,31 @@ describe Bolognese::Metadata, vcr: true do
|
|
|
418
418
|
expect(subject.schema_version).to eq("http://datacite.org/schema/kernel-4.0")
|
|
419
419
|
end
|
|
420
420
|
|
|
421
|
+
it "Schema 3 from string" do
|
|
422
|
+
input = fixture_path + "datacite_schema_3.xml"
|
|
423
|
+
subject = Bolognese::Metadata.new(input: input)
|
|
424
|
+
expect(subject.valid?).to be true
|
|
425
|
+
expect(subject.identifier).to eq("https://doi.org/10.5061/dryad.8515")
|
|
426
|
+
expect(subject.types["schemaOrg"]).to eq("Dataset")
|
|
427
|
+
expect(subject.types["resourceType"]).to eq("DataPackage")
|
|
428
|
+
expect(subject.types["resourceTypeGeneral"]).to eq("Dataset")
|
|
429
|
+
expect(subject.types["ris"]).to eq("DATA")
|
|
430
|
+
expect(subject.types["citeproc"]).to eq("dataset")
|
|
431
|
+
expect(subject.creator.length).to eq(8)
|
|
432
|
+
expect(subject.creator.last).to eq("familyName"=>"Renaud", "givenName"=>"François", "name"=>"François Renaud", "type"=>"Person")
|
|
433
|
+
expect(subject.titles).to eq([{"title"=>"Data from: A new malaria agent in African hominids."}])
|
|
434
|
+
expect(subject.alternate_identifiers).to eq([{"alternateIdentifier"=>
|
|
435
|
+
"Ollomo B, Durand P, Prugnolle F, Douzery EJP, Arnathau C, Nkoghe D, Leroy E, Renaud F (2009) A new malaria agent in African hominids. PLoS Pathogens 5(5): e1000446.",
|
|
436
|
+
"alternateIdentifierType"=>"citation"}])
|
|
437
|
+
expect(subject.publication_year).to eq("2011")
|
|
438
|
+
expect(subject.related_identifiers.length).to eq(4)
|
|
439
|
+
expect(subject.related_identifiers.last).to eq("relatedIdentifier"=>"19478877", "relatedIdentifierType"=>"PMID", "relationType"=>"IsReferencedBy")
|
|
440
|
+
expect(subject.rights_list).to eq([{"rightsUri"=>"http://creativecommons.org/publicdomain/zero/1.0"}])
|
|
441
|
+
expect(subject.publisher).to eq("Dryad Digital Repository")
|
|
442
|
+
expect(subject.source).to eq("DataCite")
|
|
443
|
+
expect(subject.schema_version).to eq("http://datacite.org/schema/kernel-3")
|
|
444
|
+
end
|
|
445
|
+
|
|
421
446
|
it "Schema 3.0 from string" do
|
|
422
447
|
input = fixture_path + "datacite-example-complicated-v3.0.xml"
|
|
423
448
|
subject = Bolognese::Metadata.new(input: input)
|
|
@@ -550,7 +575,7 @@ describe Bolognese::Metadata, vcr: true do
|
|
|
550
575
|
expect(subject.dates).to eq([{"date"=>"2008-11-01", "dateType"=>"Valid"}, {"date"=>"2008", "dateType"=>"Issued"}])
|
|
551
576
|
expect(subject.publication_year).to eq("2008")
|
|
552
577
|
expect(subject.periodical).to eq("title"=>"23rd European Photovoltaic Solar Energy Conference and Exhibition, 1-5 September 2008, Valencia, Spain; 3353-3356", "type"=>"Periodical")
|
|
553
|
-
expect(subject.descriptions
|
|
578
|
+
expect(subject.descriptions[1]["description"]).to start_with("Aim of this paper is the presentation")
|
|
554
579
|
expect(subject.publisher).to eq("WIP-Munich")
|
|
555
580
|
expect(subject.source).to eq("DataCite")
|
|
556
581
|
expect(subject.schema_version).to eq("http://datacite.org/schema/kernel-2.2")
|
|
@@ -660,6 +685,13 @@ describe Bolognese::Metadata, vcr: true do
|
|
|
660
685
|
expect(subject.schema_version).to eq("http://datacite.org/schema/kernel-3")
|
|
661
686
|
end
|
|
662
687
|
|
|
688
|
+
it "multiple rights" do
|
|
689
|
+
input = fixture_path + "datacite-multiple-rights.xml"
|
|
690
|
+
subject = Bolognese::Metadata.new(input: input)
|
|
691
|
+
expect(subject.valid?).to be true
|
|
692
|
+
expect(subject.rights_list).to eq([{"rights"=>"info:eu-repo/semantics/openAccess"}, {"rights"=>"Open Access", "rightsUri"=>"info:eu-repo/semantics/openAccess"}])
|
|
693
|
+
end
|
|
694
|
+
|
|
663
695
|
it "missing creator" do
|
|
664
696
|
input = fixture_path + "datacite_missing_creator.xml"
|
|
665
697
|
subject = Bolognese::Metadata.new(input: input, regenerate: true)
|
data/spec/utils_spec.rb
CHANGED
|
@@ -215,7 +215,7 @@ describe Bolognese::Metadata, vcr: true do
|
|
|
215
215
|
it "uri" do
|
|
216
216
|
url = "info:eu-repo/semantics/openAccess"
|
|
217
217
|
response = subject.normalize_url(url)
|
|
218
|
-
expect(response).to
|
|
218
|
+
expect(response).to eq("info:eu-repo/semantics/openAccess")
|
|
219
219
|
end
|
|
220
220
|
end
|
|
221
221
|
|
metadata
CHANGED
|
@@ -1,14 +1,14 @@
|
|
|
1
1
|
--- !ruby/object:Gem::Specification
|
|
2
2
|
name: bolognese
|
|
3
3
|
version: !ruby/object:Gem::Version
|
|
4
|
-
version: 1.0.
|
|
4
|
+
version: 1.0.13
|
|
5
5
|
platform: ruby
|
|
6
6
|
authors:
|
|
7
7
|
- Martin Fenner
|
|
8
8
|
autorequire:
|
|
9
9
|
bindir: bin
|
|
10
10
|
cert_chain: []
|
|
11
|
-
date: 2018-11-
|
|
11
|
+
date: 2018-11-18 00:00:00.000000000 Z
|
|
12
12
|
dependencies:
|
|
13
13
|
- !ruby/object:Gem::Dependency
|
|
14
14
|
name: maremma
|
|
@@ -674,12 +674,14 @@ files:
|
|
|
674
674
|
- spec/fixtures/datacite-geolocation-empty.xml
|
|
675
675
|
- spec/fixtures/datacite-metadata-sample-complicated-v2.2.xml
|
|
676
676
|
- spec/fixtures/datacite-multiple-language.xml
|
|
677
|
+
- spec/fixtures/datacite-multiple-rights.xml
|
|
677
678
|
- spec/fixtures/datacite-xml-lang.xml
|
|
678
679
|
- spec/fixtures/datacite.json
|
|
679
680
|
- spec/fixtures/datacite.xml
|
|
680
681
|
- spec/fixtures/datacite_dataset.xml
|
|
681
682
|
- spec/fixtures/datacite_malformed_creator.xml
|
|
682
683
|
- spec/fixtures/datacite_missing_creator.xml
|
|
684
|
+
- spec/fixtures/datacite_schema_3.xml
|
|
683
685
|
- spec/fixtures/datacite_software.json
|
|
684
686
|
- spec/fixtures/datacite_software_missing_comma.json
|
|
685
687
|
- spec/fixtures/datacite_software_overlapping_keys.json
|