cleaning AWS S3 with batch

aws_clean.py

#!/usr/bin/python

import sys, getopt
import subprocess
import string
import pdb

# install AWS cli https://aws.amazon.com/cli/
# usage example :
# python aws_clean.py -p "jongla/broadcast" -b "coverimagedevsg"
def main(argv):
   helpMsg = 'aws_clean.py -b <bucket> -p <prefix>'
   AWS_LIST = 'aws s3api list-objects --bucket %s --prefix %s --query Contents[*].Key' 
   AWS_DELETE = 'aws s3 rm s3://%s/%s'
   
   prefix = ''
   bucket = ''
   try:
      opts, args = getopt.getopt(argv,"hp:b:",["prefix="])
   except getopt.GetoptError:
      print helpMsg
      sys.exit(2)
   for opt, arg in opts:
      if opt == '-h':
         print helpMsg
         sys.exit()
      elif opt in ("-p", "--prefix"):
         prefix = arg
      elif opt in ("-b", "--bucket"):
         bucket = arg
   
   if bucket == '' or prefix == '':
      print helpMsg
      sys.exit()
   
   outlist = subprocess.check_output(string.split(AWS_LIST %  (bucket, prefix), ' '))
   print outlist
   cont = raw_input('\twant to continue ? --> ') 
   if cont.lower()[0] != 'y':
      print '...aborted'
      sys.exit()
    
   preplist = map(lambda x: x.strip(), string.split(outlist, '\n'))
   flist = map(lambda x: x.strip('",'), filter(lambda x: len(x) > 1 and x[0] == '"', preplist))
   for name in flist:
      #pdb.set_trace()
      print subprocess.check_output(string.split(AWS_DELETE % (bucket, name), ' '), stderr=subprocess.STDOUT)

if __name__ == "__main__":
   main(sys.argv[1:])