How can I programmatically give Cloud ML access to a bucket?

Question

The Cloud ML instructions show how to grant Cloud ML access to a bucket using gsutil. Is there a way to do this programmatically in python?

Answer 1

You can use the google-cloud python libraries to programmatically give access to a GCS bucket to Cloud ML.

from google.cloud import storage
b = client.get_bucket(BUCKET)

# Grant WRITE access to the service account.
b.acl.user(SERVICE_ACCOUNT).grant_write()
b.acl.save()

# Change the default object permissions to give read access to the service account
b.default_object_acl.user(SERVICE_ACCOUNT).grant_read()
b.default_object_acl.save()

# Grant read access to all existing objects
for o in b.list_blobs():
  o.acl.user(SERVICE_ACCOUNT).grant_read()
  o.acl.save()

For instructions on obtaining the service account programmatically see this question