briard 2.8.2 → 2.9.1

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: b82034f66939dab07cc20186304bcdbffdde06c43abc4c614ae2bae68e4d0f72
-  data.tar.gz: b8f5d74bff45e5f016e7ec094860c632b01b66df5f104501b3c4bae6fc4adc8c
+  metadata.gz: 18ccea8677e7dd2e7c7fc253a23e060c8778f81780f55904ce5572c3d6c9d685
+  data.tar.gz: c9c42bd2d6fe73495da63b8280a2d06b172ad8013af605103d17c2c32523cc05
 SHA512:
-  metadata.gz: 4d143feff904aef51b688359e9888d83ac627566febea50fcfd7d5dec001c0634f268a3918aa12c4813e4f2c2a8acbb190c75a9ff7d6bbb580e26925ca292c2f
-  data.tar.gz: 4185d017bb16a13ee4e10a5e0b9335178e8b8bf36b6d951cc82976814176d0a254818235570c6806812f5e7b01b3026492044ca7923964185221e860abfc3ed5
+  metadata.gz: fac8077f5d561ad88c74be7f14da043e14a6dfb0a2ba36874905d39ac2f9a994adf4d7e21efd8885d1e9e8d7db21463970a5faf02b26366f3cb13efede9122bb
+  data.tar.gz: 48e8f3b859cf3a753eda735b579f6d6c43197fdc635ea2a215e12d0ef3078990ccf36ec0f215f09b7e846acf17427547b8acca172551b1d13bf94ab668ab5939

data/CHANGELOG.md

@@ -1,5 +1,9 @@
 # Changelog
 
+## [v2.8.2](https://github.com/front-matter/briard/tree/v2.8.2) (2022-11-23)
+
+[Full Changelog](https://github.com/front-matter/briard/compare/v2.8.0...v2.8.2)
+
 ## [v2.8.0](https://github.com/front-matter/briard/tree/v2.8.0) (2022-11-22)
 
 [Full Changelog](https://github.com/front-matter/briard/compare/v2.6.5...v2.8.0)

data/Gemfile.lock

@@ -1,7 +1,7 @@
 PATH
   remote: .
   specs:
-    briard (2.8.2)
+    briard (2.9.1)
       activesupport (>= 4.2.5, < 8.0)
       base32-url (>= 0.5.0, < 1)
       benchmark_methods (~> 0.7)
@@ -18,6 +18,7 @@ PATH
       gender_detector (~> 0.1.2)
       iso8601 (~> 0.9.1)
       json-ld-preloaded (~> 3.1, >= 3.1.3)
+      json_schemer (~> 0.2.23)
       jsonlint (~> 0.3.0)
       loofah (~> 2.19)
       maremma (>= 4.9.7, < 5)
@@ -75,6 +76,8 @@ GEM
       scanf (~> 1.0)
       sxp (~> 1.2)
       unicode-types (~> 1.7)
+    ecma-re-validator (0.4.0)
+      regexp_parser (~> 2.2)
     edtf (3.1.0)
       activesupport (>= 3.0, < 8.0)
     excon (0.71.1)
@@ -99,13 +102,14 @@ GEM
     haml (5.2.2)
       temple (>= 0.8.0)
       tilt
+    hana (1.3.7)
     hashdiff (1.0.1)
     htmlentities (4.3.4)
     i18n (1.12.0)
       concurrent-ruby (~> 1.0)
     iso8601 (0.9.1)
-    json (2.6.2)
-    json-canonicalization (0.3.0)
+    json (2.6.3)
+    json-canonicalization (0.3.1)
     json-ld (3.2.3)
       htmlentities (~> 4.3)
       json-canonicalization (~> 0.3)
@@ -116,12 +120,17 @@ GEM
     json-ld-preloaded (3.2.2)
       json-ld (~> 3.2)
       rdf (~> 3.2)
+    json_schemer (0.2.24)
+      ecma-re-validator (~> 0.3)
+      hana (~> 1.3)
+      regexp_parser (~> 2.0)
+      uri_template (~> 0.7)
     jsonlint (0.3.0)
       oj (~> 3)
       optimist (~> 3)
     latex-decode (0.4.0)
     link_header (0.0.8)
-    loofah (2.19.0)
+    loofah (2.19.1)
       crass (~> 1.0.2)
       nokogiri (>= 1.5.9)
     maremma (4.9.9)
@@ -139,27 +148,27 @@ GEM
       oj (>= 2.8.3)
       oj_mimic_json (~> 1.0, >= 1.0.1)
     matrix (0.4.2)
-    mini_portile2 (2.8.0)
-    minitest (5.16.3)
+    mini_portile2 (2.8.1)
+    minitest (5.17.0)
     multi_json (1.15.0)
     multipart-post (2.2.3)
     namae (1.1.1)
-    nokogiri (1.13.9)
+    nokogiri (1.13.10)
       mini_portile2 (~> 2.8.0)
       racc (~> 1.4)
     oj (3.13.23)
     oj_mimic_json (1.0.1)
     optimist (3.0.1)
     parallel (1.22.1)
-    parser (3.1.2.1)
+    parser (3.2.0.0)
       ast (~> 2.4.1)
     postrank-uri (1.0.24)
       addressable (>= 2.4.0)
       nokogiri (>= 1.8.0)
       public_suffix (>= 2.0.0, < 2.1)
     public_suffix (2.0.5)
-    racc (1.6.0)
-    rack (2.2.4)
+    racc (1.6.2)
+    rack (2.2.5)
     rack-test (2.0.2)
       rack (>= 1.3)
     rainbow (3.1.1)
@@ -196,34 +205,34 @@ GEM
       rspec-mocks (~> 3.12.0)
     rspec-core (3.12.0)
       rspec-support (~> 3.12.0)
-    rspec-expectations (3.12.0)
+    rspec-expectations (3.12.2)
       diff-lcs (>= 1.2.0, < 2.0)
       rspec-support (~> 3.12.0)
-    rspec-mocks (3.12.0)
+    rspec-mocks (3.12.2)
       diff-lcs (>= 1.2.0, < 2.0)
       rspec-support (~> 3.12.0)
     rspec-support (3.12.0)
     rspec-xsd (0.1.0)
       nokogiri (~> 1.6)
       rspec (~> 3)
-    rubocop (1.39.0)
+    rubocop (1.43.0)
       json (~> 2.3)
       parallel (~> 1.10)
-      parser (>= 3.1.2.1)
+      parser (>= 3.2.0.0)
       rainbow (>= 2.2.2, < 4.0)
       regexp_parser (>= 1.8, < 3.0)
       rexml (>= 3.2.5, < 4.0)
-      rubocop-ast (>= 1.23.0, < 2.0)
+      rubocop-ast (>= 1.24.1, < 2.0)
       ruby-progressbar (~> 1.7)
-      unicode-display_width (>= 1.4.0, < 3.0)
-    rubocop-ast (1.23.0)
+      unicode-display_width (>= 2.4.0, < 3.0)
+    rubocop-ast (1.24.1)
       parser (>= 3.1.1.0)
-    rubocop-performance (1.15.1)
+    rubocop-performance (1.15.2)
       rubocop (>= 1.7.0, < 2.0)
       rubocop-ast (>= 0.4.0)
     rubocop-rake (0.6.0)
       rubocop (~> 1.0)
-    rubocop-rspec (2.15.0)
+    rubocop-rspec (2.16.0)
       rubocop (~> 1.33)
     ruby-progressbar (1.11.0)
     ruby2_keywords (0.0.5)
@@ -242,9 +251,10 @@ GEM
     tilt (2.0.11)
     tzinfo (2.0.5)
       concurrent-ruby (~> 1.0)
-    unicode-display_width (2.3.0)
+    unicode-display_width (2.4.2)
     unicode-types (1.8.0)
     unicode_utils (1.4.0)
+    uri_template (0.7.0)
     vcr (3.0.3)
     webmock (3.18.1)
       addressable (>= 2.8.0)

data/briard.gemspec

@@ -35,6 +35,7 @@ Gem::Specification.new do |s|
   s.add_dependency 'iso8601', '~> 0.9.1'
   s.add_dependency 'json-ld-preloaded', '~> 3.1', '>= 3.1.3'
   s.add_dependency 'jsonlint', '~> 0.3.0'
+  s.add_dependency 'json_schemer', '~> 0.2.23'
   s.add_dependency 'loofah', '~> 2.19'
   s.add_dependency 'maremma', '>= 4.9.7', '< 5'
   s.add_dependency 'namae', '~> 1.0'

data/lib/briard/metadata.rb

@@ -163,9 +163,11 @@ module Briard
       exists? && errors.nil?
     end
 
-    # validate against DataCite schema, unless already errors in the reader
+    # Catch errors in the reader
+    # Then validate against JSON schema for internal metadata format
+    # Then validate against DataCite schema, unless already errors in the reader
     def errors
-      meta.fetch('errors', nil) || datacite_errors(xml: datacite, schema_version: schema_version)
+      meta.fetch('errors', nil) || json_schema_errors || datacite_errors(xml: datacite, schema_version: schema_version)
     end
 
     def descriptions

data/lib/briard/metadata_utils.rb

@@ -4,6 +4,7 @@ require_relative 'doi_utils'
 require_relative 'author_utils'
 require_relative 'crossref_utils'
 require_relative 'datacite_utils'
+require_relative 'schema_utils'
 require_relative 'utils'
 
 require_relative 'readers/bibtex_reader'
@@ -42,6 +43,7 @@ module Briard
     include Briard::AuthorUtils
     include Briard::CrossrefUtils
     include Briard::DataciteUtils
+    include Briard::SchemaUtils
     include Briard::Utils
 
     include Briard::Readers::BibtexReader

data/lib/briard/readers/datacite_reader.rb

@@ -310,7 +310,7 @@ module Briard
           'container' => set_container(meta),
           'publisher' => parse_attributes(meta.fetch('publisher', nil),
                                           first: true).to_s.strip.presence,
-          'agency' => 'datacite',
+          'agency' => 'DataCite',
           'funding_references' => funding_references,
           'dates' => dates,
           'publication_year' => parse_attributes(meta.fetch('publicationYear', nil),

data/lib/briard/readers/schema_org_reader.rb

@@ -43,12 +43,14 @@ module Briard
           link = doc.css("link[rel='canonical']")
           hsh['@id'] = link[0]['href'] if link.present?
 
-          # workaround if license included but not with schema.org
-          license = doc.at("meta[name='DCTERMS.license']")
+          # workaround if license not included with schema.org
+          license = doc.at("meta[name='dc.rights']")
           hsh['license'] = license['content'] if license.present?
 
           # workaround for html language attribute if no language is set via schema.org
-          lang = doc.at('html')['lang']
+          lang = doc.at("meta[name='dc.language']") || doc.at("meta[name='citation_language']")
+          lang = lang['content'] if lang.present?
+          lang = doc.at('html')['lang'] if lang.blank?
           hsh['inLanguage'] = lang if hsh['inLanguage'].blank?
 
           # workaround if issn not included with schema.org
@@ -57,6 +59,20 @@ module Briard
           hsh['isPartOf'] = { 'name' => name ? name['content'] : nil,
                               'issn' => issn ? issn['content'] : nil }.compact
 
+          # workaround if not all authors are included with schema.org (e.g. in Ghost metadata)
+          authors = doc.css("meta[name='citation_author']").map do |author|
+            { 'name' => author['content'] }
+          end
+          hsh['author'] = hsh['creator'] if hsh['author'].blank? && hsh['creator'].present?
+          hsh['author'] = authors if authors.length > Array.wrap(hsh['author']).length
+
+          # workaround if publisher not included with schema.org (e.g. Zenodo)
+          if hsh['publisher'].blank?
+            publisher = doc.at("meta[property='og:site_name']")
+            publisher = publisher['content'] if publisher.present?
+            hsh['publisher'] = { 'name' => publisher } 
+          end
+
           string = hsh.to_json if hsh.present?
         end
 
@@ -106,7 +122,7 @@ module Briard
         contributors = get_authors(from_schema_org_contributors(Array.wrap(meta.fetch('editor',
                                                                                       nil))))
         publisher = parse_attributes(meta.fetch('publisher', nil), content: 'name', first: true)
-
+        
         ct = schema_org == 'Dataset' ? 'includedInDataCatalog' : 'Periodical'
         container = if meta.fetch(ct, nil).present?
                       url = parse_attributes(from_schema_org(meta.fetch(ct, nil)), content: 'url',
@@ -125,12 +141,13 @@ module Briard
                       }.compact
                     elsif %w[BlogPosting Article].include?(schema_org)
                       issn = meta.dig('isPartOf', 'issn')
+                      url = meta.dig('publisher', 'url')
 
                       {
                         'type' => 'Blog',
                         'title' => meta.dig('isPartOf', 'name'),
-                        'identifier' => issn,
-                        'identifierType' => issn.present? ? 'ISSN' : nil
+                        'identifier' => issn.presence || url.presence,
+                        'identifierType' => issn.present? ? 'ISSN' : 'URL'
                       }.compact
                     else
                       {}
@@ -249,7 +266,7 @@ module Briard
                               [{ 'description' => sanitize(meta.fetch('description')),
                                  'descriptionType' => 'Abstract' }]
                             end,
-          'rights_list' => rights_list,
+          'rights_list' => rights_list.presence,
           'version_info' => meta.fetch('version', nil).to_s.presence,
           'subjects' => subjects,
           'language' => language,

data/lib/briard/schema_utils.rb

@@ -0,0 +1,15 @@
+# frozen_string_literal: true
+require 'json_schemer'
+require 'pathname'
+
+module Briard
+  module SchemaUtils
+    JSON_SCHEMA = schema = File.read(File.expand_path('../../resources/json-schema/briard_schema.json', __dir__))
+    
+    def json_schema_errors
+      schemer = JSONSchemer.schema(JSON_SCHEMA)
+      errors = schemer.validate(self.meta).to_a
+      errors.map {|err| JSONSchemer::Errors.pretty err }.presence
+    end
+  end
+end

data/lib/briard/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module Briard
-  VERSION = '2.8.2'
+  VERSION = '2.9.1'
 end