Merge pull request #1146 from pan-training/oai_pmh

OAI-PMH 2.0 Implementation to export Materials

Author: fbacall

Gemfile

@@ -34,6 +34,7 @@ gem 'kt-paperclip'
 gem 'linkeddata'
 gem 'maxmind-db'
 gem 'money-rails'
+gem 'oai'
 gem 'omniauth_openid_connect'
 gem 'omniauth-rails_csrf_protection'
 gem 'pg'

Gemfile.lock

@@ -398,6 +398,11 @@ GEM
       mini_portile2 (~> 2.8.2)
       racc (~> 1.4)
     numerizer (0.1.1)
+    oai (1.3.0)
+      builder (>= 3.1.0)
+      faraday (< 3)
+      faraday-follow_redirects (>= 0.3.0, < 2)
+      rexml
     omniauth (2.1.1)
       hashie (>= 3.4.6)
       rack (>= 2.2.3)
@@ -853,6 +858,7 @@ DEPENDENCIES
   minitest
   minitest-reporters
   money-rails
+  oai
   omniauth-rails_csrf_protection
   omniauth_openid_connect
   pg

app/controllers/oai_controller.rb

@@ -0,0 +1,22 @@
+# The controller for actions related to OAI-PMH
+class OaiController < ApplicationController
+  # CSRF token authentication causes problems with OAI-PMH POST requests and OAI-PMH POST is safe because it returns static public content
+  skip_before_action :verify_authenticity_token, only: [:index]
+
+  # GET /oai-pmh
+  def index
+    provider = TrainingProvider.new
+    response = provider.process_request(oai_params.to_h)
+
+    # add XSLT prefix
+    response.sub!(/<\?xml[^>]+\?>/, "\\0\n<?xml-stylesheet type=\"text/xsl\" href=\"/oai2xhtml.xsl\"?>")
+
+    render body: response, content_type: 'text/xml'
+  end
+
+  private
+
+  def oai_params
+    params.permit(:verb, :identifier, :metadataPrefix, :set, :from, :until, :resumptionToken)
+  end
+end

app/helpers/oai_helper.rb

@@ -0,0 +1,2 @@
+module OaiHelper
+end

app/models/material.rb

@@ -1,4 +1,8 @@
 require 'rails/html/sanitizer'
+require 'json/ld'
+require 'rdf'
+require 'rdf/rdfxml'
+require 'builder'
 
 class Material < ApplicationRecord
   include PublicActivity::Common
@@ -186,4 +190,58 @@ def duplicate
   def archived?
     status == 'archived'
   end
+
+  def to_rdf
+    jsonld_str = to_bioschemas[0].to_json
+
+    graph = RDF::Graph.new
+    JSON::LD::Reader.new(jsonld_str) do |reader|
+      reader.each_statement { |stmt| graph << stmt }
+    end
+
+    rdfxml_str = graph.dump(:rdfxml, prefixes: { sdo: 'http://schema.org/', dc: 'http://purl.org/dc/terms/' })
+    rdfxml_str.sub(/\A<\?xml.*?\?>\s*/, '') # remove XML declaration because this is used inside OAI-PMH response
+  end
+
+  def to_oai_dc
+    xml = ::Builder::XmlMarkup.new
+    xml.tag!('oai_dc:dc',
+             'xmlns:oai_dc' => 'http://www.openarchives.org/OAI/2.0/oai_dc/',
+             'xmlns:dc' => 'http://purl.org/dc/elements/1.1/',
+             'xmlns:xsi' => 'http://www.w3.org/2001/XMLSchema-instance',
+             'xsi:schemaLocation' => 'http://www.openarchives.org/OAI/2.0/oai_dc/ http://www.openarchives.org/OAI/2.0/oai_dc.xsd') do
+      xml.tag!('dc:title', title)
+      xml.tag!('dc:description', description)
+      authors.each { |a| xml.tag!('dc:creator', a) }
+      contributors.each { |a| xml.tag!('dc:contributor', a) }
+      xml.tag!('dc:publisher', content_provider.title) if content_provider
+
+      xml.tag!('dc:format', 'text/html')
+      xml.tag!('dc:language', 'en')
+      xml.tag!('dc:rights', licence) if licence.present?
+
+      [date_published, date_created, date_modified].compact.each do |d|
+        xml.tag!('dc:date', d.iso8601)
+      end
+
+      if doi.present?
+        doi_iri = doi.start_with?('http://', 'https://') ? doi : "https://doi.org/#{doi}"
+        xml.tag!('dc:identifier', doi_iri)
+      else
+        xml.tag!('dc:identifier', url)
+      end
+
+      (keywords + scientific_topics.map(&:uri) + operations.map(&:uri)).each do |s|
+        xml.tag!('dc:subject', s)
+      end
+
+      xml.tag!('dc:type', 'http://purl.org/dc/dcmitype/Text')
+      xml.tag!('dc:type', 'https://schema.org/LearningResource')
+      resource_type.each { |t| xml.tag!('dc:type', t) }
+
+      xml.tag!('dc:relation', "#{TeSS::Config.base_url}#{Rails.application.routes.url_helpers.material_path(self)}")
+      xml.tag!('dc:relation', content_provider.url) if content_provider&.url
+    end
+    xml.target!
+  end
 end

app/views/static/home/_counters.html.erb

@@ -12,7 +12,6 @@
 
           <div id='resource_count' class='resource-counter-number'> <%= @count_strings[feature] %> </div>
           <div class='resource-counter-text'> <%= feature == 'events' ? 'Upcoming Events' : feature.titleize %> </div>
-
         <% end %>
       </li>
     <% end %>

config/environment.rb

@@ -1,5 +1,5 @@
 # Load the Rails application.
-require_relative "application"
+require_relative 'application'
 
 # Initialize the Rails application.
 Rails.application.initialize!

config/initializers/oai_provider.rb

@@ -0,0 +1,29 @@
+# Configure OAI-PMH library
+# see comments in: https://github.com/code4lib/ruby-oai/blob/54ea6f7f5b1e2c1be5d0a7cc61cb696b5e653d8a/lib/oai/provider.rb#L98
+require 'oai'
+require 'uri'
+
+class OAIRDF < OAI::Provider::Metadata::Format
+  def initialize
+    @prefix = 'rdf'
+    @schema = 'http://www.openarchives.org/OAI/2.0/rdf.xsd'
+    @namespace = 'http://www.w3.org/1999/02/22-rdf-syntax-ns#'
+    @element_namespace = 'rdf'
+  end
+end
+
+class TrainingProvider < OAI::Provider::Base
+  repository_name TeSS::Config.site['title']
+  repository_url "#{TeSS::Config.base_url}/oai-pmh"
+  record_prefix "oai:#{URI(TeSS::Config.base_url).host}"
+  admin_email TeSS::Config.contact_email
+  sample_id '142' # so that example id is oai:domain:142
+
+  register_format(OAIRDF.instance)
+end
+
+Rails.application.config.after_initialize do
+  TrainingProvider.source_model OAI::Provider::ActiveRecordWrapper.new(Material.where(visible: true))
+rescue ActiveRecord::ActiveRecordError
+  Rails.logger.debug 'There is no database yet or some other error, so the OAI-PMH endpoint is not configured.'
+end

config/routes.rb

@@ -177,6 +177,8 @@
 
   get 'up' => 'health_check#show'
 
+  match 'oai-pmh', to: "oai#index", via: [:get, :post]
+
   # The priority is based upon order of creation: first created -> highest priority.
   # See how all your routes lay out with "rake routes".