ERA Open Access report

open_access.csv

open_access.rb

CSV.open('open_access.csv', 'wb') do |csv|
  csv << ['year', 'count', 'community', 'collection', 'collection url']
  open_access_items = Item.select(:member_of_paths, :record_created_at)
                          .where(visibility: JupiterCore::VISIBILITY_PUBLIC)
                          .group_by {|item| item.record_created_at.year }
  open_access_items.each do |year, items|
    items.map(&:member_of_paths).flatten.tally.each do |member_of_path, count|
      community_id, collection_id = member_of_path.split('/')
      csv << [year, count, Community.find(community_id).title, Collection.find(collection_id).title,
              Rails.application.routes.url_helpers.community_collection_url(community_id, collection_id)]
    end
  end
end

record_created_at is the date created in ERA - see https://docs.google.com/document/d/1-d0eM0kn79h1DC6X4aO_0IRzq7qKoDcymqdcW7RASV8/edit?tab=t.0#heading=h.q9yzq2h9giy5

Looks good! Had a quick look at this as my day is ending, a deeper dive could be done.

I think a final tally of the total item count by year (throughout all collections) may be beneficial although that could easily be done after the fact in sheets/excel if needed.

Some optimisation could be done, however it may not be worthwhile for a one-time script. Something that comes to mind would be utilising ActiveRecord's in_batches/find_in_batches for line 3/6 however that would require testing (and setting the batch size as the default is 1000, which wouldn't test well in the test environment.)

Given that this is a one-time script and the time constraint I don't see any issues with this as is.