schema.org markup for video objects

Gemfile

@@ -42,6 +42,7 @@ end
 group :development do
   # Access an IRB console on exception pages or by using <%= console %> anywhere in the code.
   gem 'web-console', '>= 3.3.0'
+  gem 'byebug'
 end
 
 group :development, :test do

Gemfile.lock

@@ -89,6 +89,7 @@ GEM
     bundler-audit (0.9.1)
       bundler (>= 1.2.0, < 3)
       thor (~> 1.0)
+    byebug (11.1.3)
     cancancan (3.5.0)
     capistrano (3.18.0)
       airbrussh (>= 1.0.0)
@@ -442,6 +443,7 @@ PLATFORMS
 DEPENDENCIES
   addressable
   bootsnap (>= 1.1.0)
+  byebug
   cancancan
   capistrano (~> 3.0)
   capistrano-bundler

lib/metadata/schema_dot_org.rb

@@ -1,4 +1,5 @@
 module Metadata
+  # rubocop:disable Metrics/ClassLength
   class SchemaDotOrg
     def self.call(cocina_json)
       new(cocina_json).call
@@ -17,23 +18,21 @@ def call
         "@context": 'http://schema.org',
         "@type": schema_type,
         "name": title_name,
-        "identifier": identifier,
-        "description": description,
-        "isAccessibleForFree": access,
-        "license": license,
-        "url": url,
-        "creator": creators
-      }.compact
+        "description": description
+      }.merge(format_specific_fields)
+        .compact
     end
 
     def schema_type?
-      dataset?
+      dataset? || video?
     end
 
     private
 
     def schema_type
-      'Dataset' if dataset?
+      return 'Dataset' if dataset?
+
+      'Video' if video?
     end
 
     def dataset?
@@ -44,15 +43,38 @@ def dataset?
       false
     end
 
+    def video?
+      # Only return video metadata if world-downloadable.
+      video = JsonPath.new("$.description.form[?(@['value'] == 'moving image' && @['type'] == 'resource type')]").on(@cocina_json)
+      return true if video.any? && access?
+
+      false
+    end
+
     def title_name
       # title.value or concatenated title.structuredValue 1) for title with status "primary" if present 2) for first title
-      # required for Datasets
+      # required for Datasets and Videos
       titles = JsonPath.new("$.description.title[?(@['status' == 'primary'])].structuredValue[*].value").on(@cocina_json)
       return titles.join('\n') unless titles.empty?
 
       JsonPath.new('$.description.title[0].value').first(@cocina_json)
     end
 
+    def format_specific_fields
+      if dataset?
+        return { "identifier": identifier,
+                 "isAccessibleForFree": access?,
+                 "license": license,
+                 "url": url,
+                 "creator": creators }
+      elsif video?
+        return { "thumbnailUrl": thumbnail,
+                 "uploadDate": upload_date,
+                 "embedUrl": embed_url }
+      end
+      {}
+    end
+
     def description
       # description.note where type=summary or type=abstract, concatenating with \n if multiple
       # required for Datasets
@@ -75,7 +97,7 @@ def identifier
       [URI.join('https://doi.org', identifier).to_s]
     end
 
-    def access
+    def access?
       # true if access.download = "world"
       return true if JsonPath.new("$.access[?(@['download'] == 'world')]").first(@cocina_json)
 
@@ -140,5 +162,90 @@ def orcid(contributor)
 
       URI.join('https://orcid.org/', orcid).to_s
     end
+
+    def embed_url
+      response = Faraday.get(iframe_url)
+      return unless response.success?
+
+      iframe_html = Nokogiri::HTML(get_html(response.body))
+      iframe_html.css('iframe')[0]['src']
+    end
+
+    def get_html(response_body)
+      JsonPath.new('$.html').first(response_body)
+    end
+
+    def iframe_url
+      oembed_url_template.expand(format: 'json', url: embeddable_url).to_s
+    end
+
+    def oembed_url_template
+      Addressable::Template.new(Settings.embed.url_template)
+    end
+
+    def embeddable_url
+      format(Settings.embed.url, druid: bare_druid)
+    end
+
+    def bare_druid
+      druid.delete_prefix('druid:')
+    end
+
+    def druid
+      JsonPath.new('$.externalIdentifier').first(@cocina_json)
+    end
+
+    def thumbnail
+      # required for Videos
+      # structural.contains.filename with hasMimeType = "image/jp2" where structural.contains has type https://cocina.sul.stanford.edu/models/resources/video",
+      video = JsonPath.new("$.structural.contains[*][?(@['type'] == 'https://cocina.sul.stanford.edu/models/resources/video')]").on(@cocina_json)
+      filename = JsonPath.new("$[*].structural.contains[*][?(@['hasMimeType'] == 'image/jp2')].filename").first(video)
+      return if filename.blank?
+
+      URI.join(Settings.stacks.url, "file/#{druid}/#{filename}").to_s
+    end
+
+    def upload_date
+      # required for Videos
+      # event.date.value or event.date.structuredValue.value with event.date.type "publication" and event.date.status "primary"
+      # first event.date.value or event.date.structuredValue.value with event.date.type "publication"
+      events = JsonPath.new('$.description.event[*]').on(@cocina_json)
+      JsonPath.new("$[*].date[*][?(@['type'] == 'publication' && @['status'] == 'primary')].value").first(events) ||
+        JsonPath.new("$[*].date[*][?(@['type'] == 'publication' && @['status'] == 'primary')].structuredValue[*].value").first(events) ||
+        JsonPath.new("$[*].date[*][?(@['type'] == 'publication')].value").first(events) ||
+        JsonPath.new("$[*].date[*][?(@['type'] == 'publication')].structuredValue[*].value").first(events) ||
+        no_date_type(events) || no_event_type(events)
+    end
+
+    def no_date_type(events)
+      # first event.date.value or event.date.structuredValue.value with event.type "publication" and event.date.type null
+      return unless events.any?
+
+      dates = JsonPath.new("$.[?(@['type']) == 'publication')].date[*].[?(@['value'])]").on(events)
+      structured_dates = JsonPath.new("$.[?(@['type']) == 'publication')].date[*].structuredValue[*]").on(events)
+      dates.concat(structured_dates)
+      return unless dates.any?
+
+      date_value(dates)
+    end
+
+    def no_event_type(events)
+      # first event.date.value or event.date.structuredValue with event.type null and event.date.type null
+      return unless events.any?
+
+      events.select! { |event| event.key?('type') == false }
+      dates = JsonPath.new('$[*].date[*]').on(events)
+      return unless dates.any?
+
+      date_value(dates)
+    end
+
+    def date_value(dates)
+      dates.select! { |date| date.key?('type') == false }
+      return unless dates.any?
+
+      dates.first.fetch('value')
+    end
   end
+  # rubocop:enable Metrics/ClassLength
 end