Skip to content

Instantly share code, notes, and snippets.

@nicksellen
Last active October 5, 2018 16:36
Show Gist options
  • Star 0 You must be signed in to star a gist
  • Fork 0 You must be signed in to fork a gist
  • Save nicksellen/a70ff34960ab39388b59b8c41de4bfe6 to your computer and use it in GitHub Desktop.
Save nicksellen/a70ff34960ab39388b59b8c41de4bfe6 to your computer and use it in GitHub Desktop.
# Run this with:
#
# docker exec -i <idofcontainer> rails runner - < local.rb >/tmp/localfiles
#
# The first line will be a warning message, discard this, then you have a nice list of local files!
# It only takes the original attachments, not the small size.
MediaAttachment.local.find_in_batches(batch_size: 200) do |items|
items.each do |item|
puts item.file.url(:original).gsub(/.*\/social-coop-live\//, '')
end
end
Account.local.find_in_batches(batch_size: 200) do |accounts|
accounts.each do |account|
avatar = account.avatar_original_url.gsub(/.*\/social-coop-live\//, '')
puts avatar if avatar != '/avatars/original/missing.png'
header = account.header_original_url.gsub(/.*\/social-coop-live\//, '')
puts header if header != '/headers/original/missing.png'
end
end
from multiprocessing import Process, Queue
import boto3
opts = {
'service_name':'s3',
'aws_access_key_id':'FOO',
'aws_secret_access_key':'BAR',
'endpoint_url':'https://ams3.digitaloceanspaces.com',
}
bucket_name = 'social-coop-media'
process_count = 10
session = boto3.session.Session()
s3_client = session.client(**opts)
s3 = session.resource(**opts)
bucket = s3.Bucket(bucket_name)
DONE = '___DONE___'
processed = dict()
for item in open('aclprocessed').read().splitlines():
processed[item] = True
def process(q):
while True:
key = q.get()
if key == DONE:
break
s3_client.put_object_acl(Bucket=bucket_name, Key=key, ACL='public-read')
print('processed', key)
if __name__ == '__main__':
q = Queue(process_count)
processes = []
for n in range(process_count):
p = Process(target=process, args=(q,))
p.start()
processes.append(p)
for obj in bucket.objects.all():
if obj.key in processed:
print('skipped', obj.key)
else:
q.put(obj.key)
for n in range(process_count):
q.put(END_KEY)
for p in processes:
p.join()
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment