WIP GobiertoData add dcat catalog endpoint

app/controllers/gobierto_data/api/v1/datasets_controller.rb

@@ -11,6 +11,14 @@ class DatasetsController < BaseController
         skip_before_action :authenticate_in_site, only: [:new, :create, :update, :destroy]
         skip_before_action :set_admin_with_token, except: [:new, :create, :update, :destroy]
 
+        # GET /api/v1/data/catalog.xml
+        def catalog
+          @catalog = DatasetPresenter.new(current_site).build_catalog
+          respond_to do |format|
+              format.xml
+            end
+        end
+
         # GET /api/v1/data/datasets
         # GET /api/v1/data/datasets.json
         # GET /api/v1/data/datasets.csv

app/models/gobierto_data/dataset.rb

@@ -10,12 +10,15 @@ class Dataset < ApplicationRecord
     include GobiertoData::Favoriteable
     include GobiertoAttachments::Attachable
     include GobiertoCommon::Collectionable
+    include GobiertoCommon::HasCustomFieldRecords
 
     belongs_to :site
     has_many :queries, dependent: :destroy, class_name: "GobiertoData::Query"
     has_many :visualizations, dependent: :destroy, class_name: "GobiertoData::Visualization"
 
     scope :sorted, -> { order(data_updated_at: :desc) }
+    scope :visibles, -> { where(visibility_level: "active") }
+    scope :by_site, ->(site_id) { where(site_id: site_id) }
 
     translates :name
 

app/presenters/gobierto_data/dataset_presenter.rb

@@ -0,0 +1,76 @@
+module GobiertoData
+  class DatasetPresenter
+    include ActionView::Helpers::UrlHelper
+
+    attr_reader :site
+
+    def initialize(site)
+      @site = site
+    end
+
+    def build_catalog
+      catalog = {
+        identifier_uri: url_helpers.gobierto_data_root_url(host: site.domain),
+        title:          "dcat catalog for #{site}",
+        description:    "this is catalog published by #{@catalog_identifier_uri} which contains datasets",
+        issued:         site.created_at,
+        modified:       GobiertoData::Dataset.where(site_id: site.id).maximum(:created_at) || site.created_at,
+        languages:      site["configuration_data"]["available_locales"],
+        homepage:       url_helpers.gobierto_data_root_url(host: site.domain),
+        license_url:    "https://opendatacommons.org/licenses/odbl/",
+        datasets:       build_datasets_for_catalog
+      }
+    end
+
+    private
+
+    def build_datasets_for_catalog
+      datasets = []
+      Dataset.where(site_id: site.id).visibles.each do |dataset|
+        datasets <<  build_dataset_for_catalog(dataset)
+      end
+      datasets
+    end
+
+    def build_dataset_for_catalog(dataset)
+      {
+        url: url_helpers.gobierto_data_datasets_url(host: site.domain, id: dataset.slug),
+        title: dataset.name,
+        description: description_custom_field_record(dataset),
+        keywords: [],
+        issued: dataset.created_at,
+        modified: dataset.updated_at,
+        languages: [site_locale],
+        license_url: "https://opendatacommons.org/licenses/odbl/",
+        publisher:  site.name,
+        publisher_mbox: site.reply_to_email,
+        distributions: build_distribution_for_catalog(dataset)
+      }
+    end
+
+    def build_distribution_for_catalog(dataset)
+      [
+        {
+          format: 'application/csv',
+          download_url: url_helpers.download_gobierto_data_api_v1_dataset_url(dataset, host: site.domain)
+        }
+      ]
+    end
+
+    def url_helpers
+      Rails.application.routes.url_helpers
+    end
+
+    def site_locale
+      site.configuration_data["default_locale"]
+    end
+
+    def description_custom_field_record(dataset)
+      if dataset.custom_field_record_with_uid("description")
+        dataset.custom_field_record_with_uid("description").payload["description"][site_locale]
+      else
+        ""
+      end
+    end
+  end
+end

app/views/gobierto_data/api/v1/datasets/catalog.xml.erb

@@ -0,0 +1,50 @@
+<?xml version="1.0" encoding="utf-8" ?>
+<rdf:RDF
+ xmlns:time="http://www.w3.org/2006/time#"
+ xmlns:dct="http://purl.org/dc/terms/"
+ xmlns:dcat="http://www.w3.org/ns/dcat#"
+ xmlns:foaf="http://xmlns.com/foaf/0.1/"
+ xmlns:xsd="http://www.w3.org/2001/XMLSchema#"
+ xmlns:rdfs="http://www.w3.org/2000/01/rdf-schema#"
+ xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#"><% cache @catalog do %>
+ <dcat:Catalog rdf:about="https://data.some.org/catalog">
+   <dct:identifier><%= @catalog[:identifier_uri] %></dct:identifier>
+   <dct:title><%= @catalog[:title] %></dct:title>
+   <dct:description><%= @catalog[:description] %></dct:description>
+   <dct:issued rdf:datatype="http://www.w3.org/2001/XMLSchema#dateTime"><%= @catalog[:issued] %></dct:issued>
+   <dct:modified rdf:datatype="http://www.w3.org/2001/XMLSchema#dateTime"><%= @catalog[:modified] %></dct:modified>
+   <%= @catalog[:languages].map { |locale| "<dct:language>#{locale}</dct:language>" }.join.html_safe %>
+   <foaf:homepage rdf:resource="<%= @catalog[:homepage] %>"/>
+   <dct:license rdf:resource="<%= @catalog[:license_url] %>"/>
+   <% @catalog[:datasets].each do |dataset| %><dcat:dataset>
+     <dcat:Dataset rdf:about="<%= dataset[:url]%>">
+       <dct:identifier><%= dataset[:url] %></dct:identifier>
+       <dct:title><%= dataset[:title] %></dct:title>
+       <dct:description><%= dataset[:description] %></dct:description>
+       <%= dataset[:keywords].map { |keyword| "<dct:keyword>#{locale}</dct:keyword>" }.join.html_safe %>
+       <dct:issued rdf:datatype="http://www.w3.org/2001/XMLSchema#dateTime"><%= dataset[:issued] %></dct:issued>
+       <dct:modified rdf:datatype="http://www.w3.org/2001/XMLSchema#dateTime"><%= dataset[:modified] %></dct:modified>
+       <%= dataset[:languages].map { |lang| "<dct:language>#{lang}</dct:language>" }.join.html_safe %>
+       <dct:license rdf:resource="<%= dataset[:title] %>"/>
+       <dct:publisher>
+         <foaf:Organization>
+             <foaf:name><%= dataset[:publisher] %></foaf:name>
+             <foaf:mbox><%= dataset[:publisher_mbox] %></foaf:mbox>
+         </foaf:Organization>
+       </dct:publisher>
+       <% dataset[:distributions].each do |distribution| %>
+       <dcat:distribution>
+         <dcat:Distribution>
+           <dct:identifier><%= @identifier_uri %></dct:identifier>
+           <dct:title><%= dataset[:title] %> in CSV format</dct:title>
+           <dct:description><%= dataset[:description] %></dct:description>
+           <dcat:downloadURL rdf:datatype="http://www.w3.org/2001/XMLSchema#anyURI"><%= distribution[:download_url] %></dcat:downloadURL>
+           <dcat:mediaType><%= distribution[:format] %></dcat:mediaType>
+           <dct:license rdf:resource="<%= dataset[:license_url] %>"/>
+         </dcat:Distribution>
+       </dcat:distribution>
+       <% end %>
+     </dcat:Dataset>
+   </dcat:dataset><% end %>
+ </dcat:Catalog><% end %>
+ </rdf:RDF>

config/routes.rb

@@ -664,6 +664,7 @@
               resource :favorite, only: [:create, :destroy]
               resources :favorites, only: [:index]
             end
+            get "catalog" => "datasets#catalog"
           end
         end
       end

test/fixtures/gobierto_common/custom_field_records.yml

@@ -391,6 +391,11 @@ users_dataset_md_with_translations_custom_field_record:
     "md-with-translations" => { en: "Paragraph with translations!", es: "¡Párrafo con traducciones!" }
    }.to_json %>
 
+users_dataset_md_with_translations_custom_field_record:
+  item: users_dataset (GobiertoData::Dataset)
+  custom_field: madrid_data_datasets_custom_field_description
+  payload: <%= {"description": {"en": "This dataset contains the council' persons related.", "es": "Este dataset contiene las personas relaccionadas con el ayuntamiento."}}.to_json %>
+
 ## Users
 
 peter_custom_field_issue:

test/fixtures/gobierto_common/custom_fields.yml

@@ -344,6 +344,14 @@ madrid_data_datasets_custom_field_md_with_translations:
   field_type: <%= GobiertoCommon::CustomField.field_types[:localized_paragraph] %>
   uid: md-with-translations
 
+madrid_data_datasets_custom_field_description:
+  site: madrid
+  class_name: GobiertoData::Dataset
+  position: 4
+  name_translations: <%= {"en": "Description", "es": "Descripción"}.to_json %>
+  field_type: <%= GobiertoCommon::CustomField.field_types[:localized_paragraph] %>
+  uid: description
+
 madrid_custom_field_human_resources_table_plugin:
   site: madrid
   class_name: GobiertoPlans::Node

test/fixtures/sites.yml

@@ -34,6 +34,8 @@ madrid:
   organization_address: Fake St., 123
   organization_document_number: 0123456789A
   visibility_level: <%= Site.visibility_levels["active"] %>
+  created_at: <%= 1.year.ago %>
+  updated_at: <%= 1.week.ago %>
 
 santander:
   title_translations: <%= { 'en' => 'Transparencia Ciudadana', 'es' => 'Transparencia Ciudadana' }.to_json %>

test/presenters/gobierto_data/dataset_presenter_test.rb

@@ -0,0 +1,57 @@
+# frozen_string_literal: true
+
+require "test_helper"
+
+module GobiertoData
+  class DatasetPresenterTest < ActiveSupport::TestCase
+
+    def setup
+      super
+      @site = sites(:madrid)
+      @subject = DatasetPresenter.new(@site)
+      @other_site = sites(:huesca)
+      @other_subject = DatasetPresenter.new(@other_site)
+    end
+
+    def test_structure_catalog_building_do_not_show_draft_datasets
+      catalog = @subject.build_catalog
+      datasets_published = GobiertoData::Dataset.by_site(@site.id).visibles.size
+      assert_equal datasets_published, catalog[:datasets].size
+    end
+
+    def test_structure_catalog_building_using_site_with_datasets
+      catalog = @subject.build_catalog
+      assert catalog.has_key? :identifier_uri
+      assert catalog.has_key? :title
+      assert catalog.has_key? :description
+      assert catalog.has_key? :issued
+      assert catalog.has_key? :modified
+      assert catalog.has_key? :languages
+      assert catalog.has_key? :homepage
+      assert catalog.has_key? :license_url
+      assert catalog.has_key? :datasets
+      catalog[:datasets].each do |dataset|
+        assert dataset.has_key? :url
+        assert dataset.has_key? :title
+        assert dataset.has_key? :description
+        assert dataset.has_key? :keywords
+        assert dataset.has_key? :issued
+        assert dataset.has_key? :modified
+        assert dataset.has_key? :languages
+        assert dataset.has_key? :license_url
+        assert dataset.has_key? :publisher
+        assert dataset.has_key? :publisher_mbox
+        assert dataset.has_key? :distributions
+        dataset[:distributions].each do |distribution|
+          assert distribution.has_key? :format
+          assert distribution.has_key? :download_url
+        end
+      end
+    end
+
+    def test_structure_catalog_building_using_site_without_datasets
+      catalog = @other_subject.build_catalog
+      assert_equal 0, catalog[:datasets].size
+    end
+  end
+end