author-page.rb

# frozen_string_literal: true

require './_plugins/gtn/mod'
require './_plugins/gtn'

module Jekyll
  module Generators
    ##
    # This class generates the GTN's author pags
    class AuthorPageGenerator < Generator
      safe true

      ##
      # This extracts the contributions and pushes them on to an existing
      # datastructure, modifying it in the process. It's pretty gross.
      #
      # Params
      # +t+:: The tutorial, slide, or news item
      # +datastructure+:: The hash of contributors that the author information should be pushed onto
      # +flat+:: Whether the datastructure is a flat array or a nested array
      #
      # Returns
      # +datastructure+:: The modified datastructure
      def pusher(t, datastructure, flat)
        if t.data.key?('contributors')
          if flat
            t.data['contributors'].each { |c| datastructure[c].push(t) }
          else
            t.data['contributors'].each { |c| datastructure[c].push([t, nil]) }
          end
        elsif t.data.key?('contributions')
          t.data['contributions'].each do |contribution_type, contributor|
            contributor.each do |c|
              if flat
                datastructure[c].push(t)
              else
                datastructure[c].push([t, contribution_type])
              end
            end
          end
        end

        t.data['maintainers'].each { |c| datastructure[c].push([t, 'maintainer']) } if t.data.key?('maintainers')
        t.data['funding'].each { |c| datastructure[c].push([t, 'funding']) } if t.data.key?('funding')

        datastructure
      end

      ##
      # This generates the author pages
      # Params
      # +site+:: The site object
      def generate(site)
        return unless site.layouts.key? 'contributor_index'

        dir = 'hall-of-fame'

        # pre-calculating this hash saves about 4.9 seconds off the previous
        # build time of 5 seconds.
        tutorials_by_author = Hash.new { |hash, key| hash[key] = [] }
        learning_pathways_by_author = Hash.new { |hash, key| hash[key] = [] }
        slides_by_author = Hash.new { |hash, key| hash[key] = [] }
        news_by_author = Hash.new { |hash, key| hash[key] = [] }
        events_by_author = Hash.new { |hash, key| hash[key] = [] }
        videos_by_author = Hash.new { |hash, key| hash[key] = [] }
        faqs_by_author = Hash.new { |hash, key| hash[key] = [] }
        has_philosophy = Hash.new { false }

        prs_by_author = Hash.new { |hash, key| hash[key] = [] }
        reviews_by_author = Hash.new { |hash, key| hash[key] = [] }

        site.data['github'].each do |num, pr|
          prs_by_author[pr['author']['login']] << [num, pr['mergedAt']]

          pr['reviews'].each do |review|
            reviews_by_author[review['author']['login']] << [num, review['submittedAt'], review['state']]
          end
        end

        site.pages.each do |t|
          # Skip Symlinks
          if t.data['symlink']
            next
          end

          # Tutorials
          pusher(t, tutorials_by_author, false) if t['layout'] == 'tutorial_hands_on'

          # Slides
          if !%w[base_slides introduction_slides tutorial_slides].index(t['layout']).nil?
            pusher(t, slides_by_author, false)
          end

          pusher(t, events_by_author, false) if t['layout'] == 'event' or t['layout'] == "event-external"

          pusher(t, faqs_by_author, false) if t['layout'] == 'faq'

          t.data.fetch('recordings', []).each do |r|
            r.fetch('captioners', []).each { |ent| videos_by_author[ent].push([t, 'captioner', r]) }
            r.fetch('speakers', []).each { |ent| videos_by_author[ent].push([t, 'speaker', r]) }
          end

          pusher(t, learning_pathways_by_author, false) if t['layout'] == 'learning-pathway'

          # Philosophies
          has_philosophy[t.data['username']] = true if t['layout'] == 'training_philosophy' && !t.data['username'].nil?
        end

        site.posts.docs.each do |t|
          # News
          pusher(t, news_by_author, true) if t['layout'] == 'news'
        end

        Gtn::Contributors.list(site).each_key do |contributor|
          # Using PageWithoutAFile instead of a custom class which reads files
          # from disk each time, saves some time, but it is unclear how much
          # due to how the previous was accounted. But assuming 0.040s per page * 193 should be about 8 seconds.
          page2 = PageWithoutAFile.new(site, '', File.join(dir, contributor), 'index.html')
          page2.content = nil
          name = Gtn::Contributors.fetch_contributor(site, contributor).fetch('name', contributor)

          # Their tutorials
          page2.data['contributor'] = contributor
          page2.data['personname'] = name
          page2.data['title'] = "GTN Contributor: #{name}"
          page2.data['layout'] = 'contributor_index'

          page2.data['tutorials'] = tutorials_by_author[contributor].group_by{|x| x[0] }.map{|k, v| [k, v.map{|vv| vv[1]}.compact]}
          page2.data['slides'] = slides_by_author[contributor].group_by{|x| x[0] }.map{|k, v| [k, v.map{|vv| vv[1]}.compact]}
          page2.data['news'] = news_by_author[contributor]
          page2.data['learning_pathways'] = learning_pathways_by_author[contributor]
          page2.data['events'] = events_by_author[contributor].group_by{|x| x[0] }.map{|k, v| [k, v.map{|vv| vv[1]}.compact]}
          page2.data['videos'] = videos_by_author[contributor].group_by{|x| x[0] }.map{|k, v| [k, v.map{|vv| vv[1]}.uniq.compact]}
          page2.data['faqs'] = faqs_by_author[contributor].group_by{|x| x[0] }.map{|k, v| [k, v.map{|vv| vv[1]}.uniq.compact]}

          page2.data['tutorials_count'] = tutorials_by_author[contributor].length
          page2.data['slides_count'] = slides_by_author[contributor].length
          page2.data['news_count'] = news_by_author[contributor].length
          page2.data['learning_pathways_count'] = learning_pathways_by_author[contributor].length
          page2.data['events_count'] = events_by_author[contributor].length
          page2.data['videos_count'] = videos_by_author[contributor].length
          page2.data['faqs_count'] = faqs_by_author[contributor].length

          page2.data['editors'] = Gtn::TopicFilter.enumerate_topics(site).select do |t|
            t.fetch('editorial_board', []).include?(contributor)
          end
          # Also their learning pathways
          page2.data['editors'] += site.pages.select do |t|
            t['layout'] == 'learning-pathway' && t.data.fetch('editorial_board', []).include?(contributor)
          end
          page2.data['editor_count'] = page2.data['editors'].length

          page2.data['has_philosophy'] = has_philosophy[contributor]

          countable_reviews = reviews_by_author[contributor]
            .reject{|x| x[1].nil?} # Group by PRs.
            .group_by{|x| x[0]}.map{|x, r| r.sort_by{|r1| r1[1]}.max}.sort_by{|w| w[1]}.reverse

          page2.data['gh_prs_count'] = prs_by_author[contributor].count
          page2.data['gh_reviews_count'] = countable_reviews.count

          page2.data['gh_prs_recent'] = prs_by_author[contributor]
            .reject{|x| x[1].nil?}.sort_by { |x| x[1] }.reverse.take(5)
            .map{|x| x[0]}
          page2.data['gh_reviews_recent'] = countable_reviews.take(5)
            .map{|x| x[0]}

          site.pages << page2
        end
      end
    end
  end
end

Jekyll::Hooks.register :site, :post_read do |site|
  if Jekyll.env == 'production'
    Jekyll.logger.info "[GTN/Reviewers] Ingesting GitHub reviewer metadata"
    start_time = Time.now
    # Maps a lowercase version of their name to the potential mixed-case version
    contribs_lower = site.data['contributors'].map{|k, v| [k.downcase, k]}.to_h

    # Annotate the from github metadata
    gh_reviewers_by_path = Hash.new { |hash, key| hash[key] = [] }
    # Hash of PRs by path
    site.data['github'].each do |num, pr|
      # Within a PR we have some reviews, let's get that set organised:
      reviewers = pr['reviews'].map do |review|
        # Just "people"
        contribs_lower.fetch(review['author']['login'].downcase, review['author']['login'])
      end.uniq.reject{|x| x == 'github-actions'}

      pr['files'].select{|p| p['path'] =~ /.(md|html)$/}.each do |file|
        real_path = Gtn::PublicationTimes.chase_rename(file['path'])
        gh_reviewers_by_path[real_path] += reviewers
        gh_reviewers_by_path[real_path].uniq!
      end
    end

    # For all of our pages, if the path is mentioned above, then, tag it.
    site.pages.select{|t| gh_reviewers_by_path.key?(t.path)}.each do |t|
      if t['layout'] == 'tutorial_hands_on' or !%w[base_slides introduction_slides tutorial_slides].index(t['layout']).nil?
        if t.data.key?('contributors')
          # Automatically 'upgrade' to new structure
          t.data['contributions'] = {
            'authorship' => t.data['contributors'],
            'reviewing' => gh_reviewers_by_path[t.path]
          }
          t.data.delete('contributors')
        elsif t.data.key?('contributions')
          if t.data['contributions'].key?('reviewing')
            t.data['contributions']['reviewing'] += gh_reviewers_by_path[t.path]
          else
            t.data['contributions']['reviewing'] = gh_reviewers_by_path[t.path]
          end
          t.data['contributions']['reviewing'].uniq!
        end
      end
    end

    site.posts.docs.select{|t| gh_reviewers_by_path.key?(t.path)}.each do |t|
      if t['layout'] == 'news'
        if t.data.key?('contributors')
          t.data['contributions'] = {
            'authorship' => t.data['contributors'],
            'reviewing' => gh_reviewers_by_path[t.path]
          }
          t.data.delete('contributors')
        elsif t.data.key?('contributions')
          if t.data['contributions'].key?('reviewing')
            t.data['contributions']['reviewing'] += gh_reviewers_by_path[t.path]
          else
            t.data['contributions']['reviewing'] = gh_reviewers_by_path[t.path]
          end
          t.data['contributions']['reviewing'].uniq!
        end
      end
    end
    Jekyll.logger.info "[GTN/Reviewers] Complete in #{Time.now - start_time} seconds"
  end
end