jekyll-jsonld.rb

# frozen_string_literal: true

require 'json'
require './_plugins/gtn'
require './_plugins/gtn/git'
require './_plugins/util'

module Jekyll
  module Filters

    # Generate JSON-LD metadata for the GTN.
    module JsonldFilter
      GTN = {
        '@type': 'Organization',
        'http://purl.org/dc/terms/conformsTo': {
          # Bioschemas profile
          '@id': 'https://bioschemas.org/profiles/Organization/0.2-DRAFT-2019_07_19',
          '@type': 'Organization'
        },
        id: 'https://training.galaxyproject.org',
        email: '[email protected]',
        name: 'Galaxy Training Network',
        legalName: 'Galaxy Training Network',
        alternateName: 'GTN',
        url: 'https://training.galaxyproject.org',
        logo: 'https://training.galaxyproject.org/training-material/assets/images/GTNLogo1000.png',
        keywords: %w[galaxy bioinformatics training fair accessible],
        status: 'active',
        foundingDate: Gtn::Git.discover['founding_date'].to_s,
      }.freeze

      A11Y = {
        accessMode: %w[textual visual],
        accessModeSufficient: %w[textual visual],
        # "accessibilityAPI": ,
        accessibilityControl: %w[fullKeyboardControl fullMouseControl],
        accessibilityFeature: %w[alternativeText tableOfContents],
        # "accessibilityHazard": [],
        accessibilitySummary: 'The text aims to be as accessible as possible. Image descriptions will vary per ' \
                              'tutorial, from images being completely inaccessible, to images with good descriptions ' \
                              'for non-visual users.',
      }.freeze

      EDU_ROLES = {
        'use' => 'Students',
        'admin-dev' => 'Galaxy Administrators',
        'basics' => 'Students',
        'data-science' => 'Data-Science Students',
        'instructors' => 'Instructors',
      }

      ##
      # Generate the Dublin Core metadata for a material.
      # Parmaeters:
      # +material+:: The material to generate the metadata for.
      # +site+:: The site object.
      # Returns:
      # A string containing the metadata.
      #
      # Example:
      #  {{ material | generate_dublin_core: site }}
      #  => <meta name="DC.identifier" content="..." />
      def generate_dublin_core(material, site)
        return if material.key?('data') && material['data'].fetch('type', 'none') != 'tutorial_hands_on'

        attributes = [
          ['DC.identifier', site['github_repository']],
          ['DC.type', 'text'],
          ['DC.title', material['title']],
          ['DC.publisher', 'Galaxy Training Network'],
          ['DC.date', Gtn::ModificationTimes.obtain_time(material['path'])]
        ]

        attributes += Gtn::Contributors.get_authors(material).map do |user|
          ['DC.creator', Gtn::Contributors.fetch_name(site, user)]
        end

        attributes.map { |a, b| "<meta name=\"#{a}\" content=\"#{b}\">" }.join("\n")
      end

      ##
      # Generate the JSON-LD metadata for a person
      # Parameters:
      # +id+:: The id of the person.
      # +contributor+:: The contributor object from CONTRIBUTORS.yaml.
      # +site+:: The site object.
      # Returns:
      # +Hash+:: The JSON-LD metadata.
      #
      # Example:
      #  generate_person_jsonld("hexylena", site['data']['contributors']['hexylena'], site)
      #  => {
      #    "@context": "https://schema.org",
      #    "@type": "Person",
      #    "http://purl.org/dc/terms/conformsTo": {
      #      # Bioschemas profile
      #      "@id": "https://bioschemas.org/profiles/Person/0.2-DRAFT-2019_07_19",
      #      "@type": "Person"
      #    },
      #    "url": "https://training.galaxyproject.org/hall-of-fame/hexylena/",
      #    "mainEntityOfPage": "https://training.galaxyproject.org/hall-of-fame/hexylena/",
      #    "name": "hexylena",
      #    "image": "https://avatars.githubusercontent.com/hexylena",
      #    "description": "A contributor to the GTN project.",
      #    "memberOf": [...],
      #    "identifier": "https://orcid.org/0000-0002-6601-2165",
      #    "orcid": "https://orcid.org/0000-0002-6601-2165"
      #  }
      #
      def generate_person_jsonld(id, contributor, site)
        member_of = Gtn::Contributors.fetch_contributor(site, id)['affiliations'] || []
        member_of = member_of.map do |org_id|
          org = Gtn::Contributors.fetch_contributor(site, org_id)
          generate_org_jsonld(org_id, org, site)
        end

        person = {
          '@context': 'https://schema.org',
          '@type': 'Person',
          'http://purl.org/dc/terms/conformsTo': {
            '@id': 'https://bioschemas.org/profiles/Person/0.3-DRAFT',
            '@type': 'CreativeWork'
          },
          # I guess these are identical?
          url: "#{site['url']}#{site['baseurl']}/hall-of-fame/#{id}/",
          mainEntityOfPage: "#{site['url']}#{site['baseurl']}/hall-of-fame/#{id}/",
          name: Gtn::Contributors.fetch_name(site, id),
          image: "https://avatars.githubusercontent.com/#{id}",
          # No clue what to put here it's a person.
          description: if contributor.nil?
                         'A contributor to the GTN project.'
                       else
                         contributor.fetch('bio',
                                           'A contributor to the GTN project.')
                       end,
          memberOf: [GTN] + member_of,
        }
        if !contributor.nil? && contributor.key?('orcid') && contributor['orcid']
          person['identifier'] = "https://orcid.org/#{contributor['orcid']}"
        end

        person
      end

      ##
      # Generate the JSON-LD metadata for an organisation
      # Parameters:
      # +id+:: The id of the org.
      # +contributor+:: The contributor object from ORGANISATIONS.yaml.
      # +site+:: The site object.
      # Returns:
      # +Hash+:: The JSON-LD metadata.
      def generate_org_jsonld(id, contributor, site)
        organization = {
          '@context': 'https://schema.org',
          '@type': 'Organization',
          'http://purl.org/dc/terms/conformsTo': {
            '@id': 'https://bioschemas.org/profiles/Organization/0.3-DRAFT',
            '@type': 'CreativeWork'
          },
          id: "#{site['url']}#{site['baseurl']}/hall-of-fame/#{id}/",
          name: Gtn::Contributors.fetch_name(site, id),
          description: 'An organization supporting the Galaxy Training Network',
        }

        organization['url'] = contributor['url'] if contributor.key?('url') && contributor['url']

        organization
      end

      ##
      # Generate the JSON-LD metadata for a funding organisation
      # Parameters:
      # +id+:: The id of the person.
      # +contributor+:: The contributor object from ORGANISATIONS.yaml.
      # +site+:: The site object.
      # Returns:
      # +Hash+:: The JSON-LD metadata.
      def generate_funder_jsonld(id, contributor, site)
        {
          '@context': 'https://schema.org',
          '@type': 'Organization',
          'http://purl.org/dc/terms/conformsTo': {
            '@id': 'https://bioschemas.org/profiles/Organization/0.3-DRAFT',
            '@type': 'CreativeWork'
          },
          name: Gtn::Contributors.fetch_name(site, id),
          description: contributor.fetch('funding_statement', 'An organization supporting the Galaxy Training Network'),
          url: contributor.fetch('url', "https://training.galaxyproject.org/training-material/hall-of-fame/#{id}/"),
          logo: contributor.fetch('avatar', "https://github.com/#{id}.png"),
        }
      end

      ##
      # Generate the JSON-LD metadata for a grant
      # Parameters:
      # +id+:: The id of the grant.
      # +contributor+:: The contributor object from GRANTS.yaml.
      # +site+:: The site object.
      # Returns:
      # +Hash+:: The JSON-LD metadata.
      def generate_grant_jsonld(id, contributor, site)
        organization = {
          '@context': 'https://schema.org',
          '@type': 'Grant',
          identifier: contributor['funding_id'],
          url: Gtn::Contributors.fetch_funding_url(contributor) || contributor['url'],
          funder: generate_funder_jsonld(id, contributor, site)
        }

        organization['startDate'] = contributor['start_date'] if contributor.key?('start_date')
        organization['endDate'] = contributor['end_date'] if contributor.key?('end_date')

        organization
      end

      ##
      # Generate the JSON-LD metadata for a person, funder, or organisation as JSON.
      # Parameters:
      # +id+:: The id of the person.
      # +site+:: The site object.
      # +json+:: Should the output be rendered as JSON (only really used in contributor page.)
      # Returns:
      # +String+:: The JSON-LD metadata.
      def to_pfo_jsonld(id, site, json: true)
        contributor = Gtn::Contributors.fetch_contributor(site, id)
        d = if Gtn::Contributors.person?(site, id)
              generate_person_jsonld(id, contributor, site)
            elsif Gtn::Contributors.grant?(site, id)
              generate_grant_jsonld(id, contributor, site)
            else
              generate_org_jsonld(id, contributor, site)
            end

        if json
          JSON.pretty_generate(d)
        else
          d
        end
      end

      ##
      # Generate the JSON-LD metadata for a news article (blog)
      # Parameters:
      # +page+:: The page object.
      # +site+:: The +Jekyll::Site+ site object.
      # Returns:
      # +Hash+:: The JSON-LD metadata.
      def generate_news_jsonld(page, site)
        authors = Gtn::Contributors.get_authors(page.to_h).map do |x|
          to_pfo_jsonld(x, site, json: false)
        end

        data = {
          '@context': 'https://schema.org',
          '@type': 'BlogPosting',
          url: "#{site['url']}#{site['baseurl']}#{page['url']}",
          name: page['title'],
          headline: page.excerpt[0..100].gsub(/\n/, ' '), # TODO: remove html tags.
          keywords: page['tags'] || [],
          description: page.excerpt[0..100].gsub(/\n/, ' '), # TODO: remove html tags
          articleBody: page.content, # TODO: remove html tags
          datePublished: page.date,
          dateModified: Gtn::ModificationTimes.obtain_time(page.path),
          author: authors,
          publisher: GTN,
          mainEntityOfPage: {
            '@type': 'WebPage',
            '@id': "#{site['url']}#{page['url']}"
          },
          image: {
            '@type': 'ImageObject',
            width: 60,
            height: 60,
            url: "#{site['baseurl']}/assets/images/GTN-60px.png"
          }
        }
        data.update(A11Y)

        JSON.pretty_generate(data)
      end

      ##
      # Generate the JSON-LD metadata for an event
      # Parameters:
      # +page+:: The page object.
      # +site+:: The +Jekyll::Site+ site object.
      # Returns:
      # +Hash+:: The JSON-LD metadata.
      #
      # Examples:
      #   {{ page | generate_event_jsonld: site }}
      def generate_event_jsonld(page, site)
        organisers = Gtn::Contributors.get_organisers(page.to_h).map do |x|
          to_pfo_jsonld(x, site, json: false)
        end
        instructors = Gtn::Contributors.get_instructors(page.to_h).map do |x|
          to_pfo_jsonld(x, site, json: false)
        end
        funders = Gtn::Contributors.get_funders(site, page.to_h).map do |x|
          to_pfo_jsonld(x, site, json: false)
        end
        funding = Gtn::Contributors.get_grants(site, page.to_h).map do |x|
          to_pfo_jsonld(x, site, json: false)
        end

        materials = []
        if page['program']
          page['program'].each do |section|
            if !section.key? 'tutorials'
              next
            end

            section['tutorials'].each do |tutorial|
              if tutorial.key?('custom')
                next
              end

              material = Gtn::TopicFilter.fetch_tutorial_material(site, tutorial['topic'], tutorial['name'])
              materials.push(material)
            end
          end
        end
        materials.compact!

        # Extract EDAM terms from all materials
        edam_terms = materials.map do |material|
          material.fetch('edam_ontology', []).map do |term|
            {
              '@type': 'DefinedTerm',
              '@id': "http://edamontology.org/#{term}",
              inDefinedTermSet: 'http://edamontology.org',
              termCode: term,
            }
          end
        end.flatten.uniq

        learning_objectives = materials.map do |material|
          material.fetch('objectives', [])
        end.flatten.compact

        # TODO: add topic edam terms too? Not sure.
        parts = []
        materials.each do |material|
          mat = generate_material_jsonld(material, site['data'][material['topic_name']], site)
          if !mat.nil? && !mat.empty?
            parts.push(mat)
          end
        end

        if page['program']
          syllab = page['program'].reject { |s| s['section'].nil? }.map do |section|
            {
              '@type': 'Syllabus',
              name: section['section'],
              description: section.fetch('description', nil),
            }
          end
        end

        data = {
          '@context': 'https://schema.org',
          '@type': 'Course',
          url: "#{site['url']}#{site['baseurl']}#{page['url']}",
          name: page['title'],
          keywords: page['tags'] || [],
          description: page['description'],

          about: edam_terms, # TeSS, "scientific topics".
          audience: page['audience'], # TeSS: target audience
          # If 'online' is present in the mode, the course is online.
          # Will fail on "this is NOT an online course"
          # Acceptable.
          courseMode: page['mode'],
          startDate: page['date_start'],
          endDate: page['date_end'],
          organizer: organisers, # TeSS only, US spelling, non-standard

          location: page['location'], # TODO, TeSS location
          teaches: learning_objectives, # TeSS, "learning objectives"
          # timeRequired: 'P1D', # TeSS, "duration", TODO: calculate from start/end date, not implemented in scraper currently.

          availableLanguage: ['en'], # TODO: support other languages
          inLanguage: ['en'], # TODO: support other languages
          # courseCode
          # coursePrerequisites
          # educationalCredentialAwarded
          # financialAidEligible
          # hasCourseInstance
          # numberOfCredits
          # occupationalCredentialAwarded
          # syllabusSections
          # totalHistoricalEnrollment

          # assesses
          # competencyRequired
          # educationalAlignment
          # educationalLevel
          # educationalUse
          # learningResourceType
          # teaches

          funder: funders, # Org or person
          funding: funding, # Grant
          publisher: GTN,
          provider: GTN,
          syllabusSections: syllab,
          # Session materials
          # TODO: not currently parsed by TeSS, google just complains about it, so we're leaving it out.
          # hasPart: parts,
        }

        begin
          data['dateModified'] = Gtn::ModificationTimes.obtain_time(page.path)
          data['datePublished'] = Gtn::PublicationTimes.obtain_time(page.path)
        rescue StandardError
          data['dateModified'] = Gtn::ModificationTimes.obtain_time(page['path'])
          data['datePublished'] = Gtn::PublicationTimes.obtain_time(page['path'])
        end

        if page['cover']
          data['image'] = if page['cover'] =~ /^http/
                            [page['cover']]
                          else
                            ["#{site['url']}#{site['baseurl']}#{page['cover']}"]
                          end
        end

        # We CANNOT guarantee A11Y
        # data.update(A11Y)
        if page['cost'] and page['cost'].downcase == 'free'
          data['isAccessibleForFree'] = true
          offer = {
            '@type': 'Offer',
            price: 0,
            priceCurrency: 'EUR',
            category: 'Free',
            isAccessibleForFree: true,
          }
        elsif page['cost']
          data['isAccessibleForFree'] = false
          offer = {
            '@type': 'Offer',
            price: page['cost'].split[0],
            priceCurrency: page['cost'].split[1],
            isAccessibleForFree: false,
            category: 'Paid',
            # TODO: this can be more advanced but we need to collect start/end times, and timezone.
          }
        end

        # TODO: this is wrong in a whole host of scenarios like incl weekends.
        course_days = (page.fetch('date_end', page['date_start']) - page['date_start']).to_i
        if course_days < 1
          course_days = 1
        end
        data['hasCourseInstance'] = [
          {
            '@type': 'CourseInstance',
            courseMode: page['mode'],
            # courseWorkload: "A daily course running from #{page['date_start']} to #{page['date_end']}",
            offers: offer,
            instructor: instructors,
            isAccessibleForFree: data['isAccessibleForFree'],
            courseSchedule: {
              '@type': 'Schedule',
              startDate: page['date_start'],
              endDate: page.fetch('date_end', page['date_start']),
              repeatCount: course_days,
              repeatFrequency: 'daily', # Contrary to schema.org spec, this is what Google wants.
            },
            courseWorkload: "P#{course_days}D",
          }
        ]

        data['offers'] = [offer]

        if page.key?('location') && page['location'].keys.length > 1
          data['location'] = {
            '@type': 'Place',
            name: page['location']['name'],
            address: {
              '@type': 'PostalAddress',
              streetAddress: page['location'].fetch('address', nil),
              addressLocality: page['location'].fetch('city', nil),
              addressRegion: page['location'].fetch('region', nil),
              postalCode: page['location'].fetch('postcode', nil),
              addressCountry: page['location'].fetch('country', nil)
            }
          }
        end

        JSON.pretty_generate(data)
      end

      ##
      # Generate the JSON-LD metadata for a learning pathway
      # Parameters:
      # +page+:: The page object.
      # +site+:: The +Jekyll::Site+ site object.
      # Returns:
      # +Hash+:: The JSON-LD metadata.
      #
      # Examples:
      #   {{ page | generate_learning_pathway_jsonld: site }}
      def generate_learning_pathway_jsonld(page, site)
        materials = []
        page['pathway'].each do |section|
          if !section.key? 'tutorials'
            next
          end

          section['tutorials'].each do |tutorial|
            if tutorial.key?('custom')
              next
            end

            material = Gtn::TopicFilter.fetch_tutorial_material(site, tutorial['topic'], tutorial['name'])
            materials.push(material)
          end
        end
        materials.compact!

        # Extract EDAM terms from all materials
        edam_terms = materials.map do |material|
          material.fetch('edam_ontology', []).map do |term|
            {
              '@type': 'DefinedTerm',
              '@id': "http://edamontology.org/#{term}",
              inDefinedTermSet: 'http://edamontology.org',
              termCode: term,
            }
          end
        end.flatten.uniq

        learning_objectives = materials.map do |material|
          material.fetch('objectives', [])
        end.flatten.compact

        funders = materials.map do |material|
          Gtn::Contributors.get_funders(site, material).map do |x|
            to_pfo_jsonld(x, site, json: false)
          end
        end.flatten.uniq.compact

        funding = materials.map do |material|
          Gtn::Contributors.get_grants(site, material).map do |x|
            to_pfo_jsonld(x, site, json: false)
          end
        end.flatten.uniq.compact

        # TODO: add topic edam terms too? Not sure.
        parts = []
        materials.each do |material|
          mat = generate_material_jsonld(material, site['data'][material['topic_name']], site)
          if !mat.nil? && !mat.empty?
            parts.push(mat)
          end
        end

        syllab = page['pathway'].reject { |s| s['section'].nil? }.map do |section|
          {
            '@type': 'Syllabus',
            name: section['section'],
            description: section.fetch('description', nil),
          }
        end

        data = {
          '@context': 'https://schema.org',
          '@type': 'Course',
          url: "#{site['url']}#{site['baseurl']}#{page['url']}",
          name: "Learning Pathway #{page['title']}",
          keywords: page['tags'] || [],
          description: page['description'],
          about: edam_terms, # TeSS, "scientific topics".
          audience: page['audience'], # TeSS: target audience
          teaches: learning_objectives, # TeSS, "learning objectives"
          availableLanguage: ['en'], # TODO: support other languages
          inLanguage: ['en'], # TODO: support other languages
          # courseCode
          # coursePrerequisites
          # educationalCredentialAwarded
          # financialAidEligible
          # hasCourseInstance
          # numberOfCredits
          # occupationalCredentialAwarded
          # syllabusSections
          # totalHistoricalEnrollment

          # assesses
          # competencyRequired
          # educationalAlignment
          # educationalLevel
          # educationalUse
          # learningResourceType
          # teaches

          funder: funders, # Org or person
          funding: funding, # Grant
          publisher: GTN,
          provider: GTN,
          syllabusSections: syllab,
          # Session materials
          # TODO: not currently parsed by TeSS, google just complains about it, so we're leaving it out.
          # hasPart: parts,
        }

        begin
          data['dateModified'] = Gtn::ModificationTimes.obtain_time(page.path)
          data['datePublished'] = Gtn::PublicationTimes.obtain_time(page.path)
        rescue StandardError
          data['dateModified'] = Gtn::ModificationTimes.obtain_time(page['path'])
          data['datePublished'] = Gtn::PublicationTimes.obtain_time(page['path'])
        end

        if page['cover']
          data['image'] = if page['cover'] =~ /^http/
                            [page['cover']]
                          else
                            ["#{site['url']}#{site['baseurl']}#{page['cover']}"]
                          end
        end

        # We CANNOT guarantee A11Y
        # data.update(A11Y)
        data['isAccessibleForFree'] = true
        offer = {
          '@type': 'Offer',
          price: 0,
          priceCurrency: 'EUR',
          category: 'Free',
          isAccessibleForFree: true,
        }
        data['offers'] = [offer]

        # TODO: this is basically just wrong.
        data['hasCourseInstance'] = [
          {
            '@type': 'CourseInstance',
            courseMode: 'online',
            offers: offer,
            isAccessibleForFree: data['isAccessibleForFree'],
          }
        ]

        JSON.pretty_generate(data)
      end

      ##
      # Convert a material to JSON-LD, intended to be used in Jekyll Liquid templates.
      # Parameters:
      # +material+:: The material object.
      # +topic+:: The topic object.
      # +site+:: The +Jekyll::Site+ site object.
      #
      # Returns:
      # +String+:: The JSON-LD metadata.
      #
      # Examples:
      #   {{ material | to_jsonld: topic, site }}
      def to_jsonld(material, topic, site)
        JSON.pretty_generate(generate_material_jsonld(material, topic, site))
      end

      ##
      # Convert a material to JSON-LD. Intended to be used by the filters which you should call in templates.
      #
      # Parameters:
      # +material+:: The material object.
      # +topic+:: The topic object.
      # +site+:: The +Jekyll::Site+ site object.
      #
      # Returns:
      # +Hash+:: The JSON-LD metadata.
      def generate_material_jsonld(material, topic, site)
        langCodeMap = {
          "en" => 'English',
          "es" => 'Español',
          "fr" => 'Français',
        }

        eduLevel = {
          'Introductory' => 'Beginner',
          'Intermediate' => 'Intermediate',
          'Advanced' => 'Advanced'
        }
        return '{}' if !topic

        topic_desc = {
          '@type': 'CreativeWork',
          name: (topic['title']).to_s,
          description: (topic['summary']).to_s,
          url: "#{site['url']}#{site['baseurl']}/topics/#{topic['name']}/"
        }

        # aggregate everything
        data = {
          # Properties from Course
          '@context': 'http://schema.org',
          '@type': 'LearningResource',

          # Required for BioSchemas
          'http://purl.org/dc/terms/conformsTo': {
            '@id': 'https://bioschemas.org/profiles/TrainingMaterial/1.0-RELEASE',
            '@type': 'CreativeWork'
          },

          # Properties from CreativeWork
          # "about" described below
          #
          # "accountablePerson":,
          # "aggregateRating":,
          # "alternativeHeadline":,
          # "associatedMedia":,
          audience: {
            '@type': 'EducationalAudience',
            educationalRole: EDU_ROLES[topic['type']]
          },
          # "audio":,
          # "award":,
          # "author" described below
          # "character":,
          citation: [
            {
              '@type': 'CreativeWork',
              name: 'Galaxy Training: A Powerful Framework for Teaching!',
              url: 'https://doi.org/10.1371/journal.pcbi.1010752'
            },
            {
              '@type': 'CreativeWork',
              name: 'Community-Driven Data Analysis Training for Biology',
              url: 'https://doi.org/10.1016/j.cels.2018.05.012'
            }
          ],
          # "comment":,
          # "commentCount":,
          # "contentLocation":,
          # "contentRating":,
          # "contentReferenceTime":,
          # "contributor" described below
          # copyrightHolder: GTN,
          # copyrightNotice: m
          # "copyrightYear":,
          # "correction":,
          # "creator":,
          # "dateCreated":,
          # "datePublished":,
          discussionUrl: site['gitter_url'],
          # "editor":,
          # "educationalAlignment":,
          # "educationalUse":,
          # "encoding":,
          # "encodingFormat":,
          # "exampleOfWork":,
          # "expires":,
          # "funder": funding,
          # "genre":,
          # "hasPart" described below
          headline: (material['title']).to_s,
          # "interactionStatistic":,
          interactivityType: 'mixed',
          isAccessibleForFree: true,
          # "isBasedOn":,
          isFamilyFriendly: true,
          # "isPartOf" described below
          # "keywords": described below
          # "learningResourceType" described below
          license: 'https://spdx.org/licenses/CC-BY-4.0.html',
          # "locationCreated":,
          # "mainEntity":,
          # "material":,
          # "mentions" described below
          # "offers":,
          # "position":,
          producer: GTN,
          provider: GTN,
          # "publication":,
          # "publisher":,
          # "publisherImprint":,
          # "publishingPrinciples":,
          # "recordedAt":,
          # "releasedEvent":,
          # "review":,
          # "schemaVersion":,
          # "sdDatePublished":,
          # "sdLicense":,
          # "sdPublisher":,
          sourceOrganization: GTN,
          # "spatialCoverage":,
          # "sponsor":,
          # "temporalCoverage":,
          # "text":,
          # "thumbnailUrl":,
          # "timeRequired" described below
          # "translationOfWork":,
          # "translator": Google Translate???,
          # "typicalAgeRange":,
          # "version":,
          # "video":,
          # "workExample":,
          # "workTranslation":,

          # Properties from Thing
          # "additionalType":,
          # "alternateName":,
          # "description" described below
          # "disambiguatingDescription":,
          # "image":,
          # "mainEntityOfPage":,
          # "name" described below
          # "potentialAction":,
          # "sameAs":,
          # "subjectOf":,
          # "url" described below
          workTranslation: [],
          creativeWorkStatus: material['draft'] ? 'Draft' : 'Active',
        }

        if material.key?('pub_date')
          data['dateModified'] = material['mod_date']
          data['datePublished'] = material['pub_date']
        else
          begin
            data['dateModified'] = Gtn::ModificationTimes.obtain_time(material.path)
            data['datePublished'] = Gtn::PublicationTimes.obtain_time(material.path)
          rescue StandardError
            data['dateModified'] = Gtn::ModificationTimes.obtain_time(material['path'])
            data['datePublished'] = Gtn::PublicationTimes.obtain_time(material['path'])
          end
        end

        if material.key?('copyright')
          # copyrightHolder: GTN,
          data['copyrightNotice'] = material['copyright']
        else
          # I'm not sure this is accurate.
          data['copyrightHolder'] = GTN
        end

        funders = Gtn::Contributors.get_funders(site, material).map do |x|
          to_pfo_jsonld(x, site, json: false)
        end
        grants = Gtn::Contributors.get_grants(site, material).map do |x|
          to_pfo_jsonld(x, site, json: false)
        end

        data['funder'] = funders
        data['funding'] = grants

        data['identifier'] = "https://gxy.io/GTN:#{material['short_id']}" if material.key?('short_id')

        data.update(A11Y)

        actual_material = Gtn::TopicFilter.fetch_tutorial_material(site, material['topic_name'], material['tutorial_name'])

        # info depending if tutorial, hands-on or slide level
        # parts = []
        # data['hasPart'] = parts

        mentions = []
        description = []

        data['isPartOf'] = topic_desc

        data['abstract'] = material
          .fetch('content', '')
          .strip
          .split("\n")
          .first

        if ! data['abstract'].nil?
          data['abstract'] = data['abstract']
            .gsub(/\{\{\s*site.baseurl\s*\}\}/, url_prefix(site))
            .gsub(/\[{{\s*site.url\s*}}/, '[' + url_prefix(site))
            .gsub(/{% link (topics[^%]*).md %}/, url_prefix(site) + '\1.html')
            .gsub(/{% link (topics[^%]*).html %}/, url_prefix(site) + '\1.html')
            .gsub(/\s*\(?{%\s*cite [^}]+\s*%}\)?/, '')
            .gsub('{{ site.github_repository }}', safe_site_config(site, 'github_repository', 'https://example.com'))
            .gsub(/{% snippet ([^%]*) %}/, '')
            .gsub(/{% include ([^%]*) %}/, '')
        end

        description.push("## Abstract\n\n#{data['abstract']}\n\n")

        if (material['name'] == 'tutorial.md') || (material['name'] == 'slides.html')

          if material['name'] == 'tutorial.md'
            data['learningResourceType'] = 'e-learning'
            description.push("## About This Material\n\nThis is a Hands-on Tutorial from the GTN which is usable either for individual self-study, or as a teaching material in a classroom.\n\n")
          else
            data['learningResourceType'] = 'slides'
          end

          data['name'] = material['title']
          data['url'] = "#{site['url']}#{site['baseurl']}#{material['url']}"

          # Requires https://github.com/galaxyproject/training-material/pull/4271
          data['version'] = Gtn::ModificationTimes.obtain_modification_count(material['path'])

          # Time required
          if material.key?('time_estimation') && !material['time_estimation'].nil?
            data['timeRequired'] = "PT#{material['time_estimation'].upcase}"
          end

          # Description with questions, objectives and keypoints
          if material.key?('questions') && !material['questions'].nil? && material['questions'].length.positive?
            questions = material['questions'].join("\n - ")
            description.push("## Questions this #{material['type']} will address\n\n - #{questions}\n\n")
          end
          if material.key?('objectives') && !material['objectives'].nil? && material['objectives'].length.positive?
            objectives = material['objectives'].map{|x| "- #{x}"}.join("\n")
            description.push("## Learning Objectives\n\n#{objectives}\n\n")
            data['teaches'] = objectives
          end
          if material.key?('keypoints') && !material['keypoints'].nil? && material['keypoints'].length.positive?
            keypoints = material['keypoints'].join("\n - ")
            description.push("## Key Points\n\n - #{keypoints}\n\n")
          end

          # Keywords
          data['keywords'] = [topic['title']] + (material['tags'] || [])
          # Zenodo links
        end

        # Mentions are 'external resources' in TeSS.
        # This could be expanded with
        # - supported servers
        # - tools and resources used (e.g. Galaxy) or tools linked to the TS.
        # - slides (if tutorial) and tutorial (if slides)
        # - other materials in the same topic?
        if actual_material.key?('workflows')
          mentions.push({
                          '@type': 'Thing',
                          url: "#{site['url']}#{site['baseurl']}#{material['dir']}workflows/",
                          name: "Associated Workflows"
                        })
        end

        # Notebooks
        if actual_material.key?('notebook')
          if actual_material['notebook']['language'] != 'r'
            # Python, Bash, SQL (all via jupyter)
            url = "#{site['url']}#{site['baseurl']}#{material['dir']}#{material['topic_name']}-#{material['tutorial_name']}.ipynb"
            mentions.push({
                            '@type': 'Thing',
                            url: url,
                            name: "Jupyter Notebook (with Solutions)"
                          })
            mentions.push({
                            '@type': 'Thing',
                            url: url.gsub(/\.ipynb$/, '-course.ipynb'),
                            name: "Jupyter Notebook (without Solutions)"
                          })
          elsif actual_material['notebook']['language'] == 'r' # Actual R
            url = "#{site['url']}#{site['baseurl']}#{material['dir']}#{material['topic_name']}-#{material['tutorial_name']}.Rmd"
            mentions.push({
                            '@type': 'Thing',
                            url: url,
                            name: "Quarto/RMarkdown Notebook"
                          })
            mentions.push({
                            '@type': 'Thing',
                            url: "https://bio.tools/tool/rstudio",
                            name: "RStudio"
                          })
          end
        end

        # Tools
        uses_tools = false
        (actual_material['tools'] || []).each do |tool|
          if site.data['tool-meta'].nil?
            next
          end

          toolmeta = site.data['tool-meta'][tool]
          if toolmeta.nil?
            next
          end

          if toolmeta['bio.tools'].length.positive?
            mentions.push({
                            '@type': 'Thing',
                            url: "https://bio.tools/tool/#{toolmeta['bio.tools']}",
                            name: toolmeta.fetch('bio.tools_name', toolmeta['name'])
                          })
          end
          uses_tools = true
        end
        if uses_tools
          mentions.push({
                          '@type': 'Thing',
                          url: "https://bio.tools/tool/galaxy",
                          name: "Galaxy"
                        })
        end

        # Zenodo link out
        if actual_material.key?('zenodo_link') && ! actual_material['zenodo_link'].nil?
          if actual_material['zenodo_link'].length.positive?
            mentions.push({
                            '@type': 'Thing',
                            url: (actual_material['zenodo_link']).to_s,
                            name: "Associated Training Datasets"
                          })
          end
        end

        if description.empty?
          description.push(material.fetch('content', '').strip.split("\n").first)
        end
        data['description'] = description.join("\n")

        data['inLanguage'] = if material.key?('lang')
                               {
                                 '@type': 'Language',
                                 name: langCodeMap[material['lang']],
                                 alternateName: material['lang']
                               }
                             else
                               {
                                 '@type': 'Language',
                                 name: 'English',
                                 alternateName: 'en'
                               }
                             end

        # Course requirements (material + topic)
        reqs = []
        reqs.push(*topic['requirements']) if topic.key?('requirements')
        reqs.push(*material['requirements']) if material.key?('requirements')
        if !reqs.empty?
          coursePrerequisites = []
          reqs.each do |req|
            if req['type'] == 'internal'
              if req.key?('tutorials')
                (req['tutorials']).each do |tuto|
                  (site['pages']).each do |page|
                    if ((page['name'] == 'tutorial.md') || (page['name'] == 'slides.html')) &&
                       ((page['topic_name'] == req['topic_name']) && (page['tutorial_name'] == tuto))
                      # slides
                      if page['name'] == 'slides.html'
                        coursePrerequisites.push(
                          {
                            '@context': 'http://schema.org',
                            '@type': 'LearningResource',
                            url: "#{site['url']}#{site['baseurl']}/topics/#{req['topic_name']}/" \
                                 "tutorials/#{tuto}/slides.html",
                            name: (page['title']).to_s,
                            description: "Slides for '#{page['title']}' tutorial",
                            learningResourceType: 'slides',
                            interactivityType: 'expositive',
                            provider: GTN
                          }
                        )
                        if page['hands_on_url']
                          coursePrerequisites.push(
                            {
                              '@context': 'http://schema.org',
                              '@type': 'LearningResource',
                              url: (page['hands_on_url']).to_s,
                              learningResourceType: 'e-learning',
                              interactivityType: 'expositive',
                            }
                          )
                        end
                      end
                      # hands-on
                      if page['name'] == 'tutorial.md'
                        coursePrerequisites.push(
                          {
                            '@context': 'http://schema.org',
                            '@type': 'LearningResource',
                            url: "#{site['url']}#{site['baseurl']}/topics/#{req['topic_name']}/tutorials" \
                                 "/#{tuto}/tutorial.html",
                            name: (page['title']).to_s,
                            description: "Hands-on for '#{page['title']}' tutorial",
                            learningResourceType: 'e-learning',
                            interactivityType: 'expositive',
                            provider: GTN
                          }
                        )
                      end
                    end
                  end
                end
              else
                coursePrerequisites.push(
                  {
                    '@context': 'http://schema.org',
                    '@type': 'LearningResource',
                    url: "#{site['url']}#{site['baseurl']}/topics/#{req['topic_name']}/",
                    name: (site['data'][req['topic_name']]['title']).to_s,
                    description: (site['data'][req['topic_name']]['title']).to_s,
                    provider: GTN
                  }
                )
              end
            elsif req['type'] == 'external'
              coursePrerequisites.push({
                                         '@type': 'CreativeWork',
                                         url: (req['link']).to_s,
                                         name: (req['title']).to_s
                                       })
            else
              coursePrerequisites.push((req['title']).to_s)
            end
          end
          data['competencyRequired'] = coursePrerequisites.uniq
        end

        # Add contributors/authors
        if material.key?('contributors') || material.key?('contributions')
          authors = Gtn::Contributors.get_authors(material).map do |x|
            generate_person_jsonld(x, Gtn::Contributors.fetch_contributor(site, x), site)
          end

          data['author'] = authors
        end

        # Add non-author contributors
        if material.key?('contributions')
          data['contributor'] = Gtn::Contributors.get_non_authors(material).map do |x|
            generate_person_jsonld(x, site['data']['contributors'][x], site)
          end
        end

        about = []
        about.push(topic_desc)
        edam_terms = topic.fetch('edam_ontology', []) | material.fetch('edam_ontology', [])

        about += edam_terms.map do |term|
          {
            '@type': 'DefinedTerm',
            '@id': "http://edamontology.org/#{term}",
            inDefinedTermSet: 'http://edamontology.org',
            termCode: term,
            # "name": ,
            url: 'https://bioportal.bioontology.org/ontologies/EDAM/?p=classes&conceptid=' \
                 "http%3A%2F%2Fedamontology.org%2F#{term}"
          }
        end

        data['about'] = about

        data['educationalLevel'] = material.key?('level') ? eduLevel[material['level']] : 'Beginner'
        data['mentions'] = mentions

        data
      end
    end
  end
end

Liquid::Template.register_filter(Jekyll::Filters::JsonldFilter)