Merge pull request #75 from datacite/create-crossref-doi-import-task

Create crossref doi import task

Author: wendelfabianchinsamy

Gemfile

@@ -43,6 +43,7 @@ group :development, :test do
   gem "bundler-audit", require: false
   gem "brakeman", require: false
   gem "rspec-rails", "~> 7.0.0"
+  gem "parallel", "~> 1.27"
 end
 
 group :test do

Gemfile.lock

@@ -222,7 +222,7 @@ GEM
       bigdecimal (>= 3.0)
       ostruct (>= 0.2)
     ostruct (0.6.1)
-    parallel (1.26.3)
+    parallel (1.27.0)
     parser (3.3.7.1)
       ast (~> 2.4.1)
       racc
@@ -406,6 +406,7 @@ DEPENDENCIES
   logstash-logger (~> 0.26.1)
   mysql2 (~> 0.5.3)
   oj (~> 3.16, >= 3.16.11)
+  parallel (~> 1.27)
   rack-cors
   rails (~> 7.1.3, >= 7.1.3.2)
   rspec-rails (~> 7.0.0)

lib/tasks/event.rake

@@ -0,0 +1,28 @@
+# frozen_string_literal: true
+
+namespace :event do
+  desc "Import Crossref DOIs for events created within a date range"
+  # example command: IMPORT_DATE=2025-06-11 bundle exec rake event:import_crossref_event_dois
+  task import_crossref_event_dois: :environment do
+    date = Date.parse(ENV["IMPORT_DATE"])
+    start_date = date.beginning_of_day
+    end_date = (date + 1).beginning_of_day
+
+    puts("Import date: #{date}")
+    puts("Start date: #{start_date}")
+    puts("End date: #{end_date}")
+
+    events = Event
+      .where(source_id: ["crossref", "datacite-crossref"])
+      .where(created_at: start_date...end_date)
+
+    puts("Number of events: #{events.count}")
+
+    Parallel.each(events, in_threads: 20) do |event|
+      SqsUtilities.send_events_other_doi_job_message({
+        subj_id: event.subj_id,
+        obj_id: event.obj_id,
+      })
+    end
+  end
+end