check-govuk-asset-content-types.rb

assets_checked = 0
assets_without_content_type = 0

percentage_of_assets_to_test = 0.1

assets = Asset.all
assets.each.with_index do |asset, index|
  next unless (rand * 100) < percentage_of_assets_to_test

  id = asset.id
  filename = asset.read_attribute(:file)
  url = "https://assets.publishing.service.gov.uk/media/#{id}/#{filename}"

  response = `curl -Is "#{url}"`
  response_code = (response[/HTTP\/1.1 (.*)/, 1] || '').chomp
  content_type = (response[/Content-Type: (.*)/, 1] || '').chomp

  puts [url, response_code, content_type].join(',')

  assets_checked += 1
  if content_type.blank? && response_code =~ /200/
    assets_without_content_type += 1
  end
end

puts "Assets checked: #{assets_checked}"
puts "Assets without Content-Type: #{assets_without_content_type}"