Skip to content

Instantly share code, notes, and snippets.

@orangewolf

orangewolf/scratch.rb

Last active February 14, 2024 00:28
Show Gist options
  • Save orangewolf/908d94b2a1b13690390ec4e1fe6b9a3c to your computer and use it in GitHub Desktop.
Save orangewolf/908d94b2a1b13690390ec4e1fe6b9a3c to your computer and use it in GitHub Desktop.
Download ZIP
Raw
scratch.rb
f = Rails.root.join("tmp/imports/AMS1Importer_170000-180000/15-10wq0f0r.xml")
instantiation_identifier = nil
results = {}
# read each xml file
# get the physical_instantiation_local_identifiers and their matching cpb-aacip id
# get the digital_instantiation_local_identifiers and their matching cpb-aacip id
line = File.read(f)
pbcore_id = line.scan(/(cpb-aacip\/.+?)<\//).flatten.first
instantiation_identifiers = line.scan(/<instantiationIdentifier .+?>(.+?)<\/instantiationIdentifier>/mi).flatten
instantiation_identifiers.each do |instantiation_identifier|
results[instantiation_identifier] ||= []
results[instantiation_identifier] << pbcore_id
end
end
# digital_instantiation = DigitalInstantiation.where(local_instantiation_identifier: xml_di[:local_instantiation_identifier]).first&.attributes&.symbolize_keys
di = digital_instantiations.select do |d| d.date_uploaded.to_datetime.strftime("%Y-%m-%d %H:%M") != d.date_modified.to_datetime.strftime("%Y-%m-%d %H:%M") end
results.each do |key, value|
digital_instantiations = DigitalInstantiation.where(local_instantiation_identifier: key)
# find any di that the created and updated times differ by more than a small margin
di = digital_instantiations.select do |d| d.date_uploaded.to_datetime.strftime("%Y-%m-%d %H:%M") != d.date_modified.to_datetime.strftime("%Y-%m-%d %H:%M") end
# add the asset id to the result set if so
results[key] << found['id']
end
end
end
File.open('tmp/imports/child_snatcher.json', "w") do |f|
f.puts results.to_json
end
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment