Change tootctl media remove-orphans to work for all classes (#13316)

Change `tootctl media lookup` to not use an interactive prompt
This commit is contained in:
Eugen Rochko 2020-03-26 01:56:41 +01:00 committed by GitHub
parent e187537dfd
commit 0c8945e5ff
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
4 changed files with 98 additions and 41 deletions

View File

@ -15,6 +15,8 @@ class ActivityPub::TagManager
def url_for(target) def url_for(target)
return target.url if target.respond_to?(:local?) && !target.local? return target.url if target.respond_to?(:local?) && !target.local?
return unless target.respond_to?(:object_type)
case target.object_type case target.object_type
when :person when :person
target.instance_actor? ? about_more_url(instance_actor: true) : short_account_url(target) target.instance_actor? ? about_more_url(instance_actor: true) : short_account_url(target)

View File

@ -184,19 +184,6 @@ class MediaAttachment < ApplicationRecord
audio? || video? audio? || video?
end end
def variant?(other_file_name)
return true if file_file_name == other_file_name
return false if file_file_name.nil?
formats = file.styles.values.map(&:format).compact
return false if formats.empty?
extension = File.extname(other_file_name)
formats.include?(extension.delete('.')) && File.basename(other_file_name, extension) == File.basename(file_file_name, File.extname(file_file_name))
end
def to_param def to_param
shortcode shortcode
end end

View File

@ -45,6 +45,7 @@ module Mastodon
end end
option :start_after option :start_after
option :prefix
option :dry_run, type: :boolean, default: false option :dry_run, type: :boolean, default: false
desc 'remove-orphans', 'Scan storage and check for files that do not belong to existing media attachments' desc 'remove-orphans', 'Scan storage and check for files that do not belong to existing media attachments'
long_desc <<~LONG_DESC long_desc <<~LONG_DESC
@ -58,6 +59,7 @@ module Mastodon
reclaimed_bytes = 0 reclaimed_bytes = 0
removed = 0 removed = 0
dry_run = options[:dry_run] ? ' (DRY RUN)' : '' dry_run = options[:dry_run] ? ' (DRY RUN)' : ''
prefix = options[:prefix]
case Paperclip::Attachment.default_options[:storage] case Paperclip::Attachment.default_options[:storage]
when :s3 when :s3
@ -69,7 +71,7 @@ module Mastodon
loop do loop do
objects = begin objects = begin
begin begin
bucket.objects(start_after: last_key, prefix: 'media_attachments/files/').limit(1000).map { |x| x } bucket.objects(start_after: last_key, prefix: prefix).limit(1000).map { |x| x }
rescue => e rescue => e
progress.log(pastel.red("Error fetching list of files: #{e}")) progress.log(pastel.red("Error fetching list of files: #{e}"))
progress.log("If you want to continue from this point, add --start-after=#{last_key} to your command") if last_key progress.log("If you want to continue from this point, add --start-after=#{last_key} to your command") if last_key
@ -79,16 +81,21 @@ module Mastodon
break if objects.empty? break if objects.empty?
last_key = objects.last.key last_key = objects.last.key
attachments_map = MediaAttachment.where(id: objects.map { |object| object.key.split('/')[2..-2].join.to_i }).each_with_object({}) { |attachment, map| map[attachment.id] = attachment } record_map = preload_records_from_mixed_objects(objects)
objects.each do |object| objects.each do |object|
attachment_id = object.key.split('/')[2..-2].join.to_i path_segments = object.key.split('/')
filename = object.key.split('/').last model_name = path_segments.first.classify
attachment_name = path_segments[1].singularize
record_id = path_segments[2..-2].join.to_i
file_name = path_segments.last
record = record_map.dig(model_name, record_id)
attachment = record&.public_send(attachment_name)
progress.increment progress.increment
next unless attachments_map[attachment_id].nil? || !attachments_map[attachment_id].variant?(filename) next unless attachment.blank? || !attachment.variant?(file_name)
begin begin
object.delete unless options[:dry_run] object.delete unless options[:dry_run]
@ -110,17 +117,24 @@ module Mastodon
root_path = ENV.fetch('RAILS_ROOT_PATH', File.join(':rails_root', 'public', 'system')).gsub(':rails_root', Rails.root.to_s) root_path = ENV.fetch('RAILS_ROOT_PATH', File.join(':rails_root', 'public', 'system')).gsub(':rails_root', Rails.root.to_s)
Find.find(File.join(root_path, 'media_attachments', 'files')) do |path| Find.find(File.join(*[root_path, prefix].compact)) do |path|
next if File.directory?(path) next if File.directory?(path)
key = path.gsub("#{root_path}#{File::SEPARATOR}", '') key = path.gsub("#{root_path}#{File::SEPARATOR}", '')
attachment_id = key.split(File::SEPARATOR)[2..-2].join.to_i path_segments = key.split(File::SEPARATOR)
filename = key.split(File::SEPARATOR).last model_name = path_segments.first.classify
attachment = MediaAttachment.find_by(id: attachment_id) record_id = path_segments[2..-2].join.to_i
attachment_name = path_segments[1].singularize
file_name = path_segments.last
next unless PRELOAD_MODEL_WHITELIST.include?(model_name)
record = model_name.constantize.find_by(id: record_id)
attachment = record&.public_send(attachment_name)
progress.increment progress.increment
next unless attachment.nil? || !attachment.variant?(filename) next unless attachment.blank? || !attachment.variant?(file_name)
begin begin
size = File.size(path) size = File.size(path)
@ -213,25 +227,66 @@ module Mastodon
say("Settings:\t#{number_to_human_size(SiteUpload.sum(:file_file_size))}") say("Settings:\t#{number_to_human_size(SiteUpload.sum(:file_file_size))}")
end end
desc 'lookup', 'Lookup where media is displayed by passing a media URL' desc 'lookup URL', 'Lookup where media is displayed by passing a media URL'
def lookup def lookup(url)
prompt = TTY::Prompt.new path = Addressable::URI.parse(url).path
path_segments = path.split('/')[2..-1]
model_name = path_segments.first.classify
record_id = path_segments[2..-2].join.to_i
url = prompt.ask('Please enter a URL to the media to lookup:', required: true) unless PRELOAD_MODEL_WHITELIST.include?(model_name)
say("Cannot find corresponding model: #{model_name}", :red)
exit(1)
end
attachment_id = url record = model_name.constantize.find_by(id: record_id)
.split('/')[0..-2] record = record.status if record.respond_to?(:status)
.grep(/\A\d+\z/)
.join('')
if url.split('/')[0..-2].include? 'media_attachments' unless record
model = MediaAttachment.find(attachment_id).status say('Cannot find corresponding record', :red)
prompt.say(ActivityPub::TagManager.instance.url_for(model)) exit(1)
elsif url.split('/')[0..-2].include? 'accounts' end
model = Account.find(attachment_id)
prompt.say(ActivityPub::TagManager.instance.url_for(model)) display_url = ActivityPub::TagManager.instance.url_for(record)
else
prompt.say('Not found') if display_url.blank?
say('No public URL for this type of record', :red)
exit(1)
end
say(display_url, :blue)
rescue Addressable::URI::InvalidURIError
say('Invalid URL', :red)
exit(1)
end
private
PRELOAD_MODEL_WHITELIST = %w(
Account
Backup
CustomEmoji
Import
MediaAttachment
PreviewCard
SiteUpload
).freeze
def preload_records_from_mixed_objects(objects)
preload_map = Hash.new { |hash, key| hash[key] = [] }
objects.map do |object|
segments = object.key.split('/').first
model_name = segments.first.classify
record_id = segments[2..-2].join.to_i
next unless PRELOAD_MODEL_WHITELIST.include?(model_name)
preload_map[model_name] << record_id
end
preload_map.each_with_object({}) do |(model_name, record_ids), model_map|
model_map[model_name] = model_name.constantize.where(id: record_ids).each_with_object({}) { |record, record_map| record_map[record.id] = record }
end end
end end
end end

View File

@ -24,6 +24,19 @@ module Paperclip
flush_deletes flush_deletes
end end
end end
def variant?(other_filename)
return true if original_filename == other_filename
return false if original_filename.nil?
formats = styles.values.map(&:format).compact
return false if formats.empty?
other_extension = File.extname(other_filename)
formats.include?(other_extension.delete('.')) && File.basename(other_filename, other_extension) == File.basename(original_filename, File.extname(original_filename))
end
end end
end end