commonmeta-ruby 3.0.9 → 3.2.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 20990b6d7886c09275212fc7d92b3a0450a38e9f6f903f011ad560c58dce13b5
-  data.tar.gz: 768648c8a29a23744b6f5f371e5950af7e7a038864d7c4b9686ae533a2441e81
+  metadata.gz: 1ff3685620f0d6399524eac22d162f7b32391fa1424783a30e6c2f914d30deea
+  data.tar.gz: bf986a52084af3c28fb163a59022b1ede266b37b77bfd0bf8365c35fd98893b3
 SHA512:
-  metadata.gz: c7d38f03cc79506b8dc32c39fcbe0d353c0da9fde8da49c7e73c4c5bcf513dbb80129a39533a5b69a11bd88b12a257e8901050e5956be6b52f9c0e774f04a8f8
-  data.tar.gz: 12a2a786ae08af6d8d796886eb183b415e37fefa43cd98c7407acf14ca7afe56f7b448aaa2c463e1706e47e9b1bff9fcfae89736e1e3376908af8ef93dcfe5b8
+  metadata.gz: 447cc3b60622046224caf62ee2afa70f35f1b47c229c7407f126a3efcc6815cf129386590c6f1114c894111b657326fc35b442427c84907058bd809236a5afff
+  data.tar.gz: 69715e0926e03fc2441b8236a6e6beb0440b69dad8bde707f8d097a3d1922b71b3c80ac4c6286d8fc1fa5d489cc29be02250be6c4a6aadd0747d91bef1251b78

data/.gitignore

@@ -10,9 +10,6 @@
 /test/version_tmp/
 /tmp/
 
-# Used by dotenv library to load environment variables.
-.env
-
 ## Specific to RubyMotion:
 .dat*
 .repl_history

data/Gemfile.lock

@@ -1,7 +1,7 @@
 PATH
   remote: .
   specs:
-    commonmeta-ruby (3.0.9)
+    commonmeta-ruby (3.2.0)
       activesupport (>= 4.2.5, < 8.0)
       addressable (~> 2.8.1, < 2.8.2)
       base32-url (>= 0.5.0, < 1)
@@ -9,12 +9,11 @@ PATH
       builder (~> 3.2, >= 3.2.4)
       citeproc-ruby (~> 2.0)
       csl-styles (~> 2.0)
-      dotenv (~> 2.1, >= 2.1.1)
       edtf (~> 3.0, >= 3.0.4)
       gender_detector (~> 2.0)
       http (~> 5.1, >= 5.1.1)
       json-ld-preloaded (~> 3.2, >= 3.2.2)
-      json_schemer (~> 0.2.23)
+      json_schemer (~> 1.0.1)
       jsonlint (~> 0.3.0)
       loofah (~> 2.19)
       namae (~> 1.0)
@@ -26,7 +25,7 @@ PATH
 GEM
   remote: https://rubygems.org/
   specs:
-    activesupport (7.0.4.3)
+    activesupport (7.0.5)
       concurrent-ruby (~> 1.0, >= 1.0.2)
       i18n (>= 1.6, < 2)
       minitest (>= 5.1)
@@ -58,15 +57,12 @@ GEM
     docile (1.4.0)
     domain_name (0.5.20190701)
       unf (>= 0.0.5, < 1.0.0)
-    dotenv (2.8.1)
-    ebnf (2.3.2)
+    ebnf (2.3.3)
       htmlentities (~> 4.3)
       rdf (~> 3.2)
       scanf (~> 1.0)
       sxp (~> 1.2)
       unicode-types (~> 1.8)
-    ecma-re-validator (0.4.0)
-      regexp_parser (~> 2.2)
     edtf (3.1.1)
       activesupport (>= 3.0, < 8.0)
     ffi (1.15.5)
@@ -74,9 +70,8 @@ GEM
       ffi (>= 1.0.0)
       rake
     gender_detector (2.0.0)
-    haml (6.1.1)
-      temple (>= 0.8.2)
-      thor
+    haml (5.2.2)
+      temple (>= 0.8.0)
       tilt
     hana (1.3.7)
     hashdiff (1.0.1)
@@ -89,25 +84,24 @@ GEM
     http-cookie (1.0.5)
       domain_name (~> 0.5)
     http-form_data (2.3.0)
-    i18n (1.12.0)
+    i18n (1.14.0)
       concurrent-ruby (~> 1.0)
     json (2.6.3)
-    json-canonicalization (0.3.1)
-    json-ld (3.2.3)
+    json-canonicalization (0.3.2)
+    json-ld (3.2.5)
       htmlentities (~> 4.3)
-      json-canonicalization (~> 0.3)
+      json-canonicalization (~> 0.3, >= 0.3.2)
       link_header (~> 0.0, >= 0.0.8)
       multi_json (~> 1.15)
-      rack (~> 2.2)
-      rdf (~> 3.2, >= 3.2.9)
+      rack (>= 2.2, < 4)
+      rdf (~> 3.2, >= 3.2.10)
     json-ld-preloaded (3.2.2)
       json-ld (~> 3.2)
       rdf (~> 3.2)
-    json_schemer (0.2.24)
-      ecma-re-validator (~> 0.3)
+    json_schemer (1.0.1)
       hana (~> 1.3)
       regexp_parser (~> 2.0)
-      uri_template (~> 0.7)
+      simpleidn (~> 0.2)
     jsonlint (0.3.0)
       oj (~> 3)
       optimist (~> 3)
@@ -116,19 +110,19 @@ GEM
     llhttp-ffi (0.4.0)
       ffi-compiler (~> 1.0)
       rake (~> 13.0)
-    loofah (2.20.0)
+    loofah (2.21.3)
       crass (~> 1.0.2)
-      nokogiri (>= 1.5.9)
+      nokogiri (>= 1.12.0)
     matrix (0.4.2)
     minitest (5.18.0)
     multi_json (1.15.0)
     namae (1.1.1)
-    nokogiri (1.14.2-arm64-darwin)
+    nokogiri (1.15.2-arm64-darwin)
       racc (~> 1.4)
-    oj (3.14.2)
+    oj (3.15.0)
     optimist (3.0.1)
-    parallel (1.22.1)
-    parser (3.2.2.0)
+    parallel (1.23.0)
+    parser (3.2.2.1)
       ast (~> 2.4.1)
     postrank-uri (1.1)
       addressable (>= 2.4.0)
@@ -136,23 +130,24 @@ GEM
       public_suffix (>= 4.0.0, < 5)
     public_suffix (4.0.7)
     racc (1.6.2)
-    rack (2.2.6.4)
+    rack (3.0.7)
     rack-test (2.1.0)
       rack (>= 1.3)
     rainbow (3.1.1)
     rake (13.0.6)
-    rdf (3.2.9)
+    rdf (3.2.10)
       link_header (~> 0.0, >= 0.0.8)
     rdf-aggregate-repo (3.2.1)
       rdf (~> 3.2)
-    rdf-rdfa (3.2.1)
+    rdf-rdfa (3.2.2)
       haml (>= 5.2, < 7)
       htmlentities (~> 4.3)
       rdf (~> 3.2)
       rdf-aggregate-repo (~> 3.2)
       rdf-vocab (~> 3.2)
       rdf-xsd (~> 3.2)
-    rdf-rdfxml (3.2.0)
+    rdf-rdfxml (3.2.1)
+      haml (~> 5.2)
       htmlentities (~> 4.3)
       rdf (~> 3.2)
       rdf-rdfa (~> 3.2)
@@ -160,20 +155,20 @@ GEM
     rdf-turtle (3.2.1)
       ebnf (~> 2.3)
       rdf (~> 3.2)
-    rdf-vocab (3.2.4)
+    rdf-vocab (3.2.6)
       rdf (~> 3.2, >= 3.2.4)
     rdf-xsd (3.2.1)
       rdf (~> 3.2)
       rexml (~> 3.2)
-    regexp_parser (2.7.0)
+    regexp_parser (2.8.0)
     rexml (3.2.5)
     rspec (3.12.0)
       rspec-core (~> 3.12.0)
       rspec-expectations (~> 3.12.0)
       rspec-mocks (~> 3.12.0)
-    rspec-core (3.12.1)
+    rspec-core (3.12.2)
       rspec-support (~> 3.12.0)
-    rspec-expectations (3.12.2)
+    rspec-expectations (3.12.3)
       diff-lcs (>= 1.2.0, < 2.0)
       rspec-support (~> 3.12.0)
     rspec-mocks (3.12.5)
@@ -183,7 +178,7 @@ GEM
     rspec-xsd (0.1.0)
       nokogiri (~> 1.6)
       rspec (~> 3)
-    rubocop (1.49.0)
+    rubocop (1.52.0)
       json (~> 2.3)
       parallel (~> 1.10)
       parser (>= 3.2.0.0)
@@ -193,18 +188,21 @@ GEM
       rubocop-ast (>= 1.28.0, < 2.0)
       ruby-progressbar (~> 1.7)
       unicode-display_width (>= 2.4.0, < 3.0)
-    rubocop-ast (1.28.0)
+    rubocop-ast (1.29.0)
       parser (>= 3.2.1.0)
-    rubocop-capybara (2.17.1)
+    rubocop-capybara (2.18.0)
       rubocop (~> 1.41)
-    rubocop-performance (1.16.0)
+    rubocop-factory_bot (2.23.1)
+      rubocop (~> 1.33)
+    rubocop-performance (1.18.0)
       rubocop (>= 1.7.0, < 2.0)
       rubocop-ast (>= 0.4.0)
     rubocop-rake (0.6.0)
       rubocop (~> 1.0)
-    rubocop-rspec (2.19.0)
+    rubocop-rspec (2.22.0)
       rubocop (~> 1.33)
       rubocop-capybara (~> 2.17)
+      rubocop-factory_bot (~> 2.22)
     ruby-progressbar (1.13.0)
     scanf (1.0.0)
     simplecov (0.22.0)
@@ -213,11 +211,13 @@ GEM
       simplecov_json_formatter (~> 0.1)
     simplecov-html (0.12.3)
     simplecov_json_formatter (0.1.4)
-    sxp (1.2.3)
+    simpleidn (0.2.1)
+      unf (~> 0.1.4)
+    sxp (1.2.4)
       matrix (~> 0.4)
       rdf (~> 3.2)
-    temple (0.10.0)
-    thor (1.2.1)
+    temple (0.10.2)
+    thor (1.2.2)
     tilt (2.1.0)
     tzinfo (2.0.6)
       concurrent-ruby (~> 1.0)
@@ -226,7 +226,6 @@ GEM
     unf_ext (0.0.8.2)
     unicode-display_width (2.4.2)
     unicode-types (1.8.0)
-    uri_template (0.7.0)
     vcr (6.1.0)
     webmock (3.18.1)
       addressable (>= 2.8.0)

data/commonmeta.gemspec

@@ -25,13 +25,12 @@ Gem::Specification.new do |s|
   s.add_dependency 'builder', '~> 3.2', '>= 3.2.4'
   s.add_dependency 'citeproc-ruby', '~> 2.0'
   s.add_dependency 'csl-styles', '~> 2.0'
-  s.add_dependency 'dotenv', '~> 2.1', '>= 2.1.1'
   s.add_dependency 'edtf', '~> 3.0', '>= 3.0.4'
   s.add_dependency 'gender_detector', '~> 2.0'
   s.add_dependency 'http', '~> 5.1', '>= 5.1.1'
   s.add_dependency 'json-ld-preloaded', '~> 3.2', '>= 3.2.2'
   s.add_dependency 'jsonlint', '~> 0.3.0'
-  s.add_dependency 'json_schemer', '~> 0.2.23'
+  s.add_dependency 'json_schemer', '~> 1.0.1'
   s.add_dependency 'loofah', '~> 2.19'
   s.add_dependency 'namae', '~> 1.0'
   s.add_dependency 'postrank-uri', '~> 1.1'

data/lib/commonmeta/crossref_utils.rb

@@ -2,9 +2,6 @@
 
 module Commonmeta
   module CrossrefUtils
-    # To configure the writing of Crossref metadata, use environmental
-    # variables CROSSREF_DEPOSITOR_NAME, CROSSREF_DEPOSITOR_EMAIL and CROSSREF_REGISTRANT,
-    # e.g. in a .env file
     def write_crossref_xml
       @crossref_xml ||= Nokogiri::XML::Builder.new(encoding: 'UTF-8') do |xml|
         xml.doi_batch(crossref_root_attributes) do
@@ -13,10 +10,10 @@ module Commonmeta
             xml.doi_batch_id(SecureRandom.uuid)
             xml.timestamp(Time.now.utc.strftime('%Y%m%d%H%M%S'))
             xml.depositor do
-              xml.depositor_name(ENV.fetch('CROSSREF_DEPOSITOR_NAME', nil))
-              xml.email_address(ENV.fetch('CROSSREF_DEPOSITOR_EMAIL', nil))
+              xml.depositor_name(depositor)
+              xml.email_address(email)
             end
-            xml.registrant(ENV.fetch('CROSSREF_REGISTRANT', nil))
+            xml.registrant(registrant)
           end
           xml.body do
             insert_crossref_work(xml)
@@ -133,7 +130,7 @@ module Commonmeta
 
     def insert_citation_list(xml)
       xml.citation_list do
-        references.each do |ref|
+        Array.wrap(references).each do |ref|
           xml.citation('key' => ref['key']) do
             xml.journal_article(ref['journal_title'])
             xml.author(ref['author'])

data/lib/commonmeta/metadata.rb

@@ -45,9 +45,6 @@ module Commonmeta
             'state' => options[:state],
             'provider_id' => options[:provider_id],
             'client_id' => options[:client_id],
-            'depositor' => options[:depositor],
-            'email' => options[:email],
-            'registrant' => options[:registrant],
             'content_url' => options[:content_url]
           }
           string = File.read(options[:input])
@@ -62,9 +59,6 @@ module Commonmeta
           'state' => options[:state],
           'provider_id' => options[:provider_id],
           'client_id' => options[:client_id],
-          'depositor' => options[:depositor],
-          'email' => options[:email],
-          'registrant' => options[:registrant],
           'content_url' => options[:content_url],
           'creators' => options[:creators],
           'contributors' => options[:contributors],
@@ -94,11 +88,11 @@ module Commonmeta
       @client_id = hsh.to_h['client_id'].presence
       @content_url = hsh.to_h['content_url'].presence
 
-      # options that come from the submission, needed
+      # options that come from the cli, needed
       # for crossref doi registration
-      @depositor = hsh.to_h['depositor'].presence
-      @email = hsh.to_h['email'].presence
-      @registrant = hsh.to_h['registrant'].presence
+      @depositor = options[:depositor]
+      @email = options[:email]
+      @registrant = options[:registrant]
 
       # set attributes directly
       read_options = options.slice(

data/lib/commonmeta/metadata_utils.rb

@@ -13,6 +13,7 @@ require_relative 'readers/codemeta_reader'
 require_relative 'readers/crossref_reader'
 require_relative 'readers/crossref_xml_reader'
 require_relative 'readers/datacite_reader'
+require_relative 'readers/json_post_reader'
 require_relative 'readers/npm_reader'
 require_relative 'readers/ris_reader'
 require_relative 'readers/schema_org_reader'
@@ -46,6 +47,7 @@ module Commonmeta
     include Commonmeta::Readers::CrossrefXmlReader
     include Commonmeta::Readers::CslReader
     include Commonmeta::Readers::DataciteReader
+    include Commonmeta::Readers::JsonPostReader
     include Commonmeta::Readers::NpmReader
     include Commonmeta::Readers::RisReader
     include Commonmeta::Readers::SchemaOrgReader

data/lib/commonmeta/readers/json_post_reader.rb

@@ -0,0 +1,78 @@
+# frozen_string_literal: true
+
+module Commonmeta
+  module Readers
+    module JsonPostReader
+      def get_json_post(id: nil, **_options)
+        return { "string" => nil, "state" => "not_found" } unless id.present?
+
+        url = normalize_id(id)
+        response = HTTP.get(url)
+        return { "string" => nil, "state" => "not_found" } unless response.status.success?
+  
+        { "string" => response.body.to_s }
+      end
+
+      def read_json_post(string: nil, **options)
+        read_options = ActiveSupport::HashWithIndifferentAccess.new(options.except(:doi, :id, :url,
+                                                                                   :sandbox, :validate, :ra))
+
+        meta = string.present? ? JSON.parse(string) : {}
+
+        id = options[:doi] ? normalize_doi(options[:doi]) : normalize_id(meta.fetch("id", nil))
+        url = normalize_url(meta.fetch("url", nil))
+        type = "Article"
+        creators = if meta.fetch("authors", nil).present?
+            get_authors(from_json_post(Array.wrap(meta.fetch("authors"))))
+          else
+            [{ "type" => "Organization", "name" => ":(unav)" }]
+          end
+        titles = [{ "title" => meta.fetch("title", nil) }]
+        publisher = { "name" => meta.dig("blog", "title") }
+
+        date = {}
+        date["published"] = get_iso8601_date(meta.dig("date_published")) if meta.dig("date_published").present?
+        date["updated"] = get_iso8601_date(meta.dig("date_modified")) if meta.dig("date_modified").present?
+
+        license = if meta.dig("blog", "license").present?
+            hsh_to_spdx("rightsURI" => meta.dig("blog", "license"))
+          end
+        home_page_url = normalize_url(meta.dig("blog", "home_page_url"))
+        container = if meta.dig("blog", "title").present?
+            { "type" => "Periodical",
+              "title" => meta.dig("blog", "title"),
+              "identifier" => home_page_url,
+              "identifierType" => "URL" }
+          end
+
+        descriptions = if meta.fetch("summary", nil).present?
+            [{ "description" => sanitize(meta.fetch("summary", nil)),
+               "descriptionType" => "Abstract" }]
+          else
+            []
+          end
+        language = meta.fetch("language", nil) || meta.fetch("blog", "language", nil)
+        state = id.present? || read_options.present? ? "findable" : "not_found"
+        subjects = Array.wrap(meta.fetch("tags", nil)).reduce([]) do |sum, subject|
+          sum += name_to_fos(subject)
+
+          sum
+        end
+
+        { "id" => id,
+          "type" => type,
+          "url" => url,
+          "titles" => titles,
+          "creators" => creators,
+          "publisher" => publisher,
+          "container" => container,
+          "date" => date,
+          "language" => language,
+          "descriptions" => descriptions,
+          "license" => license,
+          "subjects" => subjects.presence,
+          "state" => state }.compact.merge(read_options)
+      end
+    end
+  end
+end

data/lib/commonmeta/utils.rb

@@ -468,6 +468,8 @@ module Commonmeta
         "cff"
       elsif %r{\A(http|https):/(/)?github\.com/(.+)\z}.match?(id)
         "cff"
+      elsif %r{\A(http|https):/(/)?rogue-scholar\.org/api/posts/(.+)\z}.match?(id)
+        "json_post"
       else
         "schema_org"
       end
@@ -875,6 +877,12 @@ module Commonmeta
       }
     end
 
+    def from_json_post(element)
+      mapping = { "url" => "id" }
+
+      map_hash_keys(element: element, mapping: mapping)
+    end
+
     def from_csl(element)
       Array.wrap(element).map do |a|
         if a["literal"].present?
@@ -1068,6 +1076,8 @@ module Commonmeta
 
       return iso8601_time.split(".").first + "Z" if iso8601_time.to_s.include? "."
 
+      return iso8601_time.split("+").first + "Z" if iso8601_time.to_s.include? "+"
+
       iso8601_time
     end
 
@@ -1084,6 +1094,15 @@ module Commonmeta
       dd.fetch("date", nil)
     end
 
+    def get_link(links, link_type)
+      ll = Array.wrap(links).find { |d| d["rel"] == link_type } || {}
+      ll.fetch("href", nil)
+    end
+
+    def rogue_scholar_api_url(id, _options = {})
+      "https://rogue-scholar.org/api/posts/#{id}"
+    end
+
     # convert commonmeta dates to DataCite format
     def get_dates_from_date(date)
       return nil if date.nil?

data/lib/commonmeta/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module Commonmeta
-  VERSION = '3.0.9'
+  VERSION = '3.2.0'
 end

data/lib/commonmeta.rb

@@ -1,6 +1,5 @@
 # frozen_string_literal: true
 
-require 'dotenv/load'
 require 'active_support/all'
 require 'nokogiri'
 ActiveSupport::XmlMini.backend = 'Nokogiri'

data/spec/author_utils_spec.rb

@@ -40,6 +40,16 @@ describe Commonmeta::Metadata, vcr: true do
       expect(subject.is_personal_name?(name: author['name'])).to be true
     end
 
+    it 'has unknown given name' do
+      author = { 'name' => 'Rintze Zelle' }
+      expect(subject.is_personal_name?(name: author['name'])).to be false
+    end
+
+    it 'has middle initial' do
+      author = { 'name' => 'Martin H. Fenner' }
+      expect(subject.is_personal_name?(name: author['name'])).to be true
+    end
+
     it 'has no info' do
       author = { 'name' => 'M Fenner' }
       expect(subject.is_personal_name?(name: author['name'])).to be false

data/spec/cli_spec.rb

@@ -11,6 +11,7 @@ describe Commonmeta::CLI do
   describe "convert from id", vcr: true do
     SCHEMA_ORG_REGEX = /"@context": "http:\/\/schema.org"/
     CROSSREF_XML_REGEX = /<doi_batch/
+    DEPOSITOR_REGEX = /<depositor_name>test<\/depositor_name>/
     DATACITE_REGEX = /"types": {/
     BIBTEX_REGEX = /@article{https:\/\/doi.org\/10.7554\/elife.01567/
     BIBTEX_DATA_REGEX = /@misc{https:\/\/doi.org\/10.5061\/dryad.8515/
@@ -141,6 +142,12 @@ describe Commonmeta::CLI do
         expect { subject.convert id }.to output(DATACITE_REGEX).to_stdout
       end
 
+      it 'to crossref_xml' do
+        subject.options = { to: "crossref_xml", depositor: "test", email: "info@example.org", registrant: "test" }
+        expect { subject.convert id }.to output(CROSSREF_XML_REGEX).to_stdout
+        expect { subject.convert id }.to output(DEPOSITOR_REGEX).to_stdout
+      end
+
       # TODO
       # it 'to bibtex' do
       #   subject.options = { to: "bibtex" }
@@ -166,6 +173,12 @@ describe Commonmeta::CLI do
         subject.options = { to: "datacite" }
         expect { subject.convert file }.to output(DATACITE_REGEX).to_stdout
       end
+
+      it 'to crossref_xml' do
+        subject.options = { to: "crossref_xml", depositor: "test", email: "info@example.org", registrant: "test" }
+        expect { subject.convert file }.to output(CROSSREF_XML_REGEX).to_stdout
+        expect { subject.convert file }.to output(DEPOSITOR_REGEX).to_stdout
+      end
     end
 
     context "crossref_xml", vcr: true do
@@ -209,8 +222,9 @@ describe Commonmeta::CLI do
       end
 
       it 'to crossref_xml' do
-        subject.options = { to: "crossref_xml" }
+        subject.options = { to: "crossref_xml", depositor: "test", email: "info@example.org", registrant: "test" }
         expect { subject.convert file }.to output(CROSSREF_XML_REGEX).to_stdout
+        expect { subject.convert file }.to output(DEPOSITOR_REGEX).to_stdout
       end
 
       it 'to datacite' do
@@ -246,6 +260,12 @@ describe Commonmeta::CLI do
         expect { subject.convert file }.to output(/@article{https:\/\/doi.org\/10.5438\/4k3m-nyvg/).to_stdout
       end
 
+      it 'to crossref_xml' do
+        subject.options = { to: "crossref_xml", depositor: "test", email: "info@example.org", registrant: "test" }
+        expect { subject.convert file }.to output(CROSSREF_XML_REGEX).to_stdout
+        expect { subject.convert file }.to output(DEPOSITOR_REGEX).to_stdout
+      end
+
       it 'to datacite invalid' do
         file = fixture_path + "datacite_missing_creator.xml"
         subject.options = { to: "datacite", show_errors: "true" }