HELIO-4704 Attempt to speed up sitemap generation

config/sitemap.rb

@@ -36,28 +36,44 @@
   # add '/static/about_daily', changefreq: 'monthly'
   # add '/static/about_project', changefreq: 'monthly'
   # add '/static/rights', changefreq: 'monthly'
+  ActiveFedora::SolrService.query("+has_model_ssim:Monograph AND -(press_sim:demo OR press_sim:heliotrope OR press_sim:monitoringservicetarget)",
+                                  fl: ['id',
+                                       'date_modified_dtsi',
+                                       'visibility_ssi',
+                                       'tombstone_ssim',
+                                       'representative_id_ssim',
+                                       'file_set_ids_ssim'],
+                                  rows: 100_000).each do |doc|
+    next unless doc['visibility_ssi'] == 'open'
+    next if doc['tombstone_ssim']&.first == 'yes'
 
-  docs = ActiveFedora::SolrService.query("+has_model_ssim:Monograph AND -(press_sim:demo OR press_sim:heliotrope OR press_sim:monitoringservicetarget)", rows: 100_000)
-  docs.each do |d|
-    next unless d['visibility_ssi'] == 'open'
-    next if d['tombstone_ssim']&.first == 'yes'
+    file_set_ids = doc['file_set_ids_ssim']
+    cover_id = doc['representative_id_ssim']&.first
+    rep_file_set_ids = FeaturedRepresentative.where(work_id: doc['id'], file_set_id: file_set_ids).pluck(:file_set_id)
 
-    cover_id = d['representative_id_ssim']&.first
+    file_set_docs = []
+    # We have books with so many file_sets that the query is too big for solr, so we break them up to manage query size if needed
+    if file_set_ids.present?
+      until file_set_ids.empty?
+        file_set_docs << ActiveFedora::SolrService.query("{!terms f=id}#{file_set_ids.shift(999).join(",")}", fl: ['id', 'visibility_ssi'], rows: 1000)
+      end
+      file_set_docs = file_set_docs.flatten
+    end
 
-    d['file_set_ids_ssim']&.each do |fsid|
-      fs = ActiveFedora::SolrService.query("{!terms f=id}#{fsid}", rows: 1).first
+    file_set_docs.each do |fs|
       next unless fs.present? && fs['visibility_ssi'] == 'open'
       # Monograph cover and "featured representative" file_set URLs don't need to be in sitemaps.
       # Crawlers cannot parse content from CSB, anyway. They can only read the page title.
-      rep = FeaturedRepresentative.where(work_id: d['id'], file_set_id: fsid).first
-      next if rep&.kind.present? || fsid == cover_id
+      next if fs['id'] == cover_id
+      next if rep_file_set_ids.any?(fs['id'])
 
       # the majority of FileSets won't be featured reps at all, so get a 'normal' url
-      url = Rails.application.routes.url_helpers.hyrax_file_set_path(fsid)
-      add url, lastmod: d['date_modified_dtsi'], priority: 0.5, changefreq: 'monthly'
+      url = Rails.application.routes.url_helpers.hyrax_file_set_path(fs['id'])
+      add url, lastmod: doc['date_modified_dtsi'], priority: 0.5, changefreq: 'monthly'
     end
+
     # monographs are always in sitemaps (unless they're in Draft)
-    url = Rails.application.routes.url_helpers.hyrax_monograph_path(d['id'])
-    add url, lastmod: d['date_modified_dtsi'], priority: 1, changefreq: 'monthly'
+    url = Rails.application.routes.url_helpers.hyrax_monograph_path(doc['id'])
+    add url, lastmod: doc['date_modified_dtsi'], priority: 1, changefreq: 'monthly'
   end
 end